The Everglades Experiments: Lessons for Ecosystem Restoration (Ecological Studies, 201)

Ecological Studies, Vol. 201 The Everglades Experiments

Edited by

M.M. Caldwell, Washington, USA G. Heldmaier, Marburg, Germany R.B. Jackson, Durham, USA O.L. Lange, Würzburg, Germany H.A. Mooney, Stanford, USA E.-D. Schulze, Jena, Germany U. Sommer, Kiel, Germany

Ecological Studies Volumes published since 2006 are listed at the end of this book.

Curtis J. Richardson

The Everglades Experiments Lessons for Ecosystem Restoration

Curtis J. Richardson Nicholas School of the Environment and Earth Sciences Duke University Wetland Center Levine Science Research Center Duke University Durham, NC 27708-0333 USA [email protected]

ISBN 978-0-387-98796-5

e-ISBN 978-0-387-68923-4

Library of Congress Control Number: 2007938761 © 2008 Springer Science+Business Media, LLC All rights reserved. This work may not be translated or copied in whole or in part without the written permission of the publisher (Springer Science+Business Media, LLC, 233 Spring Street, New York, NY-10013, USA), except for brief excerpts in connection with reviews or scholarly analysis. Use in connection with any form of information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed is forbidden. The use in this publication of trade names, trademarks, service marks, and similar terms, even if they are not identified as such, is not to be taken as an expression of opinion as to whether or not they are subject to proprietary rights. Printed on acid-free paper 9 8 7 6 5 4 3 2 1 springer.com

This book is dedicated to my family – my wife Carol, son John, and daughter Suzanne – who for more than a decade provided me with the love and support I needed to traverse the scientific and political mires of Florida’s Everglades.

Preface

In the late 1960s, I worked as a graduate teaching assistant in plant ecology for the late Dr. John Henry Davis at the University of Florida. On one of our visits to the Everglades, he mentioned to me that he had been studying problems of the Everglades since the early 1930s, and that rapid growth in Florida, unless checked, was about to doom the Everglades. He hoped his vegetation survey of the Everglades and his vegetation map could someday be used to help restore the Everglades to some semblance of what it had been prior to the turn of the century. These long-forgotten discussions with Dr. Davis were rekindled when, during a wetland conference in Orlando, Florida in the late 1980s, I was asked what might be responsible for the reported massive invasion of cattails that had been noted during the past decade in the Everglades. Several hypotheses were presented at the meeting, including some preliminary data on the significant inputs of nutrients from agricultural lands and Lake Okeechobee to the north. The shifts in the hydrologic conditions and flow patterns of the existing Everglades were also mentioned. Because of the extensive work on phosphorus and nutrient retention then being done at the Duke University Wetland Center, I was asked in early 1989 to do a preliminary survey and analysis of the ecological status of the Everglades. From this early work, carried out by Dr. Chris Craft and myself, it was apparent that the Everglades had undergone radical changes in both water flow and water quality since my early visits to the Everglades in the late 1960s. This led us to develop and focus our research on three key questions. (1) What are the effects of increased nutrient and water inputs on the native plant and animal communities? (2) What is the long-term nutrient storage capacity of the Everglades? (3) How can water management in the Everglades be improved to maintain the natural communities? Our early studies showed that the multipurpose management objectives that had been maintained by the US Army Corps of Engineers, the South Florida Water Management District, and the State of Florida since the 1950s had resulted in major alterations in hydrologic and nutrient regimes throughout the Everglades. Moreover, the long-term ecological effects of the changes in hydroperiod and increased nutrient loadings during the past three decades have not been quantified under experimental conditions. While numerous reports on water monitoring and several excellent volumes on the Everglades regarding the ecological effects have been published vii

viii

Preface

(Davis and Ogden 1994; Porter and Porter 2002), no detailed synthesis of long-term experimental work in the Everglades has been available. Clearly, the restoration of the Everglades first requires a rigorous analysis of ecosystem structure and function to determine which factors and interaction of factors are responsible for the plant and animal community alterations reported for the Everglades. During a 14-year period (1989–2003), the faculty and students at the Duke University Wetland Center and its partner institutions conducted extensive experimental research on the effects of water, nutrients, and fire on the Everglades communities. This volume is a synthesis of the key findings and summary of the experiments conducted during this period by faculty and students at the Duke University Wetland Center and scientists from the University of Louisville, Michigan State University, Indiana University, University of Nevada Reno, SUNY Plattsburg of New York, Brigham Young University, Portland State University, the U.S. Forest Service, and the Hydrobotanical Institute, Czech Republic. Many thanks go to Lisa Blumenthal Rattray, my administrative assistant for nearly 10 years, for her tireless editing of our annual Everglades reports, the starting point for a lot of this book. The data presented herein would not be possible without the thousands of analyses of soil, water, and plant samples carried out by the analytical technicians at the Duke University Wetland Center in Durham, NC. Key personnel include Paul Heine, lab manager, analytical chemist, and QA/QC specialist; Wes Willis, analytical chemist and QA/QC officer; and Julie Rice, analytical technician. At the Duke Florida lab, I would like to thank Dr. Panchabi Vaithiyanathan whose scientific insights into Everglades biogeochemistry along with his outstanding experimental designs for our field research were indispensable. Bob Johnson and Jeff Johnson are to be commended for their skilled lab management over the 14-year period, and key field technicians John Zahina, Kevin Nicholas, and Lea Karppi collected massive amounts of data under often very difficult field conditions. Dr. Mengchi Ho provided invaluable assistance with statistical analysis, database management, and many of the graphics found in this volume. Dr. Jan Stevenson provided invaluable early work on periphyton analysis from the dosing and gradient studies, and provided some unprocessed samples for our macroinvertebrate chapter. Species identifications for macroinvertebrates were verified by J. Epler (Coleoptera, Diptera, Ephemeroptera, Hemiptera, Trichoptera, Hydracarina), R. Maddocks (Ostracoda), M. Milligan (Decapoda, Hirudinea, Oligochaeta, other miscellaneous taxa), J. Daigle (Odonata), F. Thompson (Gastropoda), and M. Larsen (Cladocera, Copepoda). W. Loftus verified fish identifications of lake chubsucker, seminole killifish, and redfin pickerel. Soil core work was aided by Molie Polk and Chad Gorham, and analyses for paleoecological research and microfossil extractions were performed by Daniel Jones with help from Rachel McCaskill, Jennifer Jensen, Laura Pyle, and Elizabeth Sklad. We are indebted to Claire Schelske, Jaye Cable, Bill Burnett, and Jason Lynch for help with radioisotope dating procedures and analyses. Thanks to David Higdon and Jackie Huvane for help with multivariate analyses, Eric Edlund for help and advice with Calpalyn, and Tom Pummer and the Radiology Department at Palms West Hospital in Loxahatchee, FL for X-rays of the soil cores.

Preface

ix

We would like to thank Scott Struck (Indiana University) for the original design of the figures in Chap. 3. The study reported in Chapter 18 was supported by grant MSM 600 766 5801 from the Czech Republic Ministry of Education. Finally, to the dozens of graduate and undergraduate students who worked tirelessly on mountains of samples in the lab and entered data into spreadsheets for countless hours, I can say that your contributions to saving the Everglades were priceless. Importantly, the work summarized in this volume covers both the structural responses and the functional responses of the Everglades ecosystem via experimental and gradient studies on microbial activity, algal responses, macroinvertebrate populations, macrophyte populations, and productivity in response to alterations to nutrients in soil and water, hydrologic changes, and fire. No studies were conducted to any degree on bird or fish populations. I am grateful to Assistant Editor Janet Slobodien and Senior Production Editor Anne Meagher of Springer-Verlag, Sumathi Srinivasan of SPi Technologies, and the numerous outside reviewers of the manuscript chapters, who aided greatly with improving the quality of the book. The quality of this book was greatly enhanced by the extensive scientific technical editing done by Dr. Jan Vymazal and the detailed and professional technical editing done by Dr. Randy Neighbarger. However, any errors or omissions are my responsibility. I also wish to thank the Everglades Area Agricultural Environmental Protection District for financial support to collect the data that made this volume possible. Many thanks go to the South Florida Water Management District for providing access to our Everglades research sites. Finally, it is hoped that this body of work will provide an ecological basis to aid future scientists, managers, and decision makers in restoring our nation’s only major subtropical wetland. Durham, NC

Curtis J. Richardson

Contents

Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

vii

Contributors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

xv

1

Overview of Gradient Studies and Experiments . . . . . . . . . . . . . . . . .

Part I 2

1

Everglades Ecosystem

Ecological Status of the Everglades: Environmental and Human Factors that Control the Peatland Complex on the Landscape . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

13

3

Soil Characteristics of the Everglades Peatland . . . . . . . . . . . . . . . . . .

59

4

Vegetation and Algae of the Everglades Fen . . . . . . . . . . . . . . . . . . . . .

73

Part II

Nutrient and Hydrologic Gradient Studies

5

Introduction to the Gradient Studies . . . . . . . . . . . . . . . . . . . . . . . . . .

97

6

Enrichment Gradients in WCA-2A and Northern WCA-3A: Water, Soil, Plant Biomass, and Nutrient Storage Responses . . . . . . .

103

7

Geologic Settings and Hydrology Gradients in the Everglades . . . . .

167

8

Effects of Hydrologic Management Decisions on Everglades Tree Islands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

191

Macrophyte Community Responses in the Everglades with an Emphasis on Cattail (Typha domingensis) and Sawgrass (Cladium jamaicense) Interactions along a Gradient of Long-Term Nutrient Additions, Altered Hydroperiod and Fire . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

215

9

xi

xii

Contents

10

Algal Responses to Long-Term Nutrient Additions . . . . . . . . . . . . . .

261

11

Macroinvertebrate Responses to a Gradient of Long-Term Nutrient Additions, Altered Hydroperiod, and Fire. . . . . . . . . . . . . .

277

Historical Changes in Water Quality and Vegetation in WCA-2A Determined by Paleoecological Analyses . . . . . . . . . . . . . .

321

Carbon Cycling and Dissolved Organic Matter Export in the Northern Everglades . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

351

12

13

Part III Everglades Experiments A. Phosphorus Dosing 14

15

16

17

18

19

Introduction to a Mesocosm Approach for Establishment of Phosphorus Gradient Experiments . . . . . . . . . . . . . . . . . . . . . . . . .

375

Water Quality, Soil Chemistry, and Ecosystem Responses to P Dosing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

385

Macrophyte Slough Community Response to Experimental Phosphorus Enrichment and Periphyton Removal . .

417

Decomposition of Litter and Peat in the Everglades: The Influence of P Concentrations . . . . . . . . . . . . . . . . . . . . . . . . . . . .

441

Experimental Assessment of Phosphorus Effects on Algal Assemblages in Dosing Mesocosms . . . . . . . . . . . . . . . . . . . . . . . . . . .

461

Macroinvertebrate and Fish Responses to Experimental P Additions in Everglades Sloughs . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

477

B. Fertilizer Experiments 20

Plant Community Response to Long-Term N and P Fertilization . .

505

C. Disturbance Experiments 21

The Effects of Disturbance, Phosphorus, and Water Level on Plant Succession in the Everglades . . . . . . . . . . . . . . . . . . . . . . . . .

531

D. Germination Experiments 22

Establishment and Seedling Growth of Sawgrass and Cattail from the Everglades . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

547

Contents

Part IV 23

24

25

Modeling Ecosystem Responses to Phosphorus Additions

Long-Term Phosphorus Assimilative Capacity (PAC) in the Everglades . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

567

Spatial Distributions of Total Phosphorus and Phosphorus Accretion Rates in Everglades Soils . . . . . . . . . . . . . . . . . . . . . . . . . .

579

An Ecological Basis for Establishment of a Phosphorus Threshold for the Everglades Ecosystem . . . . . . . . . . . . . . . . . . . . . .

595

Part V 26

xiii

Lessons for Restoration of the Everglades

An Ecological Approach for Restoration of the Everglades Fen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

621

Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

643

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

679

Contributors

M. Lee Barber National Guard Bureau, Arlington, VA, USA, [email protected] Marek Bastl Department of Botany, Faculty of Biological Sciences, University of South Bohemia, Branišovská 331, 370 05 České Budějovice, Czech Republic, [email protected] Mark B. Bush Florida Institute of Technology, College of Science, 150 West University Boulevard, Melbourne, FL 32901, USA, [email protected] Sherri R. Cooper Biology Department, Bryn Athyn College, Box 707, Benaade Hall, Bryn Athyn, PA 19009, USA, [email protected] Christopher B. Craft School of Public and Environmental Affairs, Indiana University, 1315 E. 10th St., Bloomington, IN 47405-2100, USA, [email protected] Alisa Dickson Army Compatible Use Buffers and National Environmental Policy Act, National Guard Bureau Environmental Programs Division, 111 South George Mason Dr, Arlington, VA 22204, USA, [email protected] Michelle Goman Department of Earth and Atmospheric Sciences, Snee Hall, Cornell University, Ithaca, NY 14853-1504, USA, [email protected] Sarah C. Goslee USDA-ARS, Building 3702 Curtin Road, University Park, PA 16802, USA, [email protected] Patrick N. Halpin Nicholas School of the Environment and Earth Sciences, Duke University, Box 90328, Durham, NC 27708, USA, [email protected]

xv

xvi

Contributors

Mengchi Ho Nicholas School of the Environment and Earth Sciences, Duke University Wetland Center, Box 90333, Durham, NC 27708, USA, [email protected] Kirsten Hofmockel University of Michigan School of Natural Resources and Environment, Dana Building, 440 Church Street, Ann Arbor, MI 48109, USA, [email protected] Jacqueline K. Huvane Quintiles Transnational Corp., Research Triangle Park, NC, USA, [email protected] Jeffrey Johnson Duke University Wetland Center, Haverhill, FL 33425, USA, [email protected] Robert R. Johnson South Florida Water Management District, Everglades Systems Research, 2850 NW 6th Ave., Boca Raton, FL 33431, USA, [email protected] Jan Kaštovský Department of Botany, Faculty of Biological Sciences, University of South Bohemia Branišovská 331, 370 05 České Budějovice, Czech Republic [email protected] Ryan S. King Center for Reservoir and Aquatic Systems Research, Department of Biology, Baylor University, P.O. Box 97388, Waco, TX 76798-7388, USA, [email protected] Jaroslava Komárková Institute of Hydrobiology, Czech Academy of Sciences, Na Sádkách 7, 370 05 České Budějovice, Czech Republic [email protected] James W. Pahl State of Louisiana Department of Natural Resources, P.O. Box 44027, Baton Rouge, LA 70804, USA, [email protected] Song S. Qian Nicholas School of the Environment and Earth Sciences, Duke University, Box 90328, Durham, NC 27708, USA, [email protected] Robert G. Qualls Natural Resources and Environmental Science, University of Nevada, Reno, Reno, NV 89557, USA, [email protected] Klára Rˇeháková Biology Centre of the Czech Academy of Sciences, Na Sádkách 7, 370 05 České Budějovice, Czech Republic, [email protected]

Contributors

xvii

Curtis J. Richardson Nicholas School of the Environment and Earth Sciences, Duke University Wetland Center, Box 90333, Durham, NC 27708, USA, [email protected] Edwin A. Romanowicz Center for Earth and Environmental Science, State University of New York, Plattsburgh, NY 12901-2681, USA, [email protected] Craig A. Stow NOAA/Great Lakes Environmental Research Laboratory, 2205 Commonwealth Blvd., Ann Arbor, MI 48105 USA, [email protected] P.V. Sundareshwar Institute of Atmospheric Sciences, South Dakota School of Mines and Technology, 501 East St. Joseph Street, Rapid City, SD 57701, USA, [email protected] Panchabi Vaithiyanathan Divers Alert Network, The Peter B. Bennett Center, 6 West Colony Place, Durham, NC 27705, USA, [email protected] Jan Vymazal ENKI, o.p.s., Dukelská 145, 379 01 Trˇebonˇ, Czech Republic, [email protected] John G. Zahina South Florida Water Management District, P.O. Box 24680, West Palm Beach, FL 33416-4680, USA, [email protected]

Color Plates

Plate 1. Aerial view of dense housing development built adjacent to the Everglades in southern Florida. Note the water drainage canal moving water into the Atlantic Ocean and the Sawgrass Expressway that separate the massive development from the natural Everglades. (Photo by Curtis J. Richardson) (See Page 1, 40)

Plate 2. A dense community of sawgrass (Cladium jamaicense) typical of the peatland fen found throughout undisturbed portions of the Everglades. (Photo by Curtis J. Richardson) (See Page 74)

Plate 3. Wet prairie (often called “flats”) is a common vegetation type in the northern Everglades. Note the tree islands in the background. (Photo by Curtis J. Richardson) (See Pages 74, 78)

Plate 4. Sloughs or open water areas found primarily in the northeast and south-central Everglades. Note the water lily plants (Nymphaea odorata) in the slough surrounded by a monoculture of sawgrass. (Photo by Curtis J. Richardson) (See Pages 74, 78)

Plate 5. Aerial view of tree islands in the southern Everglades. Note the surrounding sloughs and sawgrass stands with ponds (small circular deep water areas) scattered throughout and often created by fire and maintained by alligator activity. (Photo by Curtis J. Richardson) (See Pages 74, 79, 80)

Plate 6. A dense stand of cattails (Typha domingensis) in a heavily P-enriched area in northern WCA-2A along our C-gradient study transect. Note: Two other invasive species—water lettuce (Pistia stratiotes) and water hyacinth (Eichhornia crassipes)—are shown in the foreground. (Photo by Curtis J. Richardson) (See Page 154)

Plate 7. An aerial view of a mesocosm flume experiment with 4 dosing channels, a walled control, and an unwalled control in an unenriched slough in southern WCA-2A. (Photo by Curtis J. Richardson) (See Page 376, 385, 387 and 390)

Plate 8. Fertilizer experiment plots in a slough site in WCA-2B were dosed with nitrogen and phosphorus. (Photo by Curtis J. Richardson) (See Page 507)

1

Overview of Gradient Studies and Experiments Curtis J. Richardson

1.1

Introduction

Marjory Stoneman Douglas captured the public imagination with the simple but illustrative statement, “It is a river of grass.” Her seminal 1947 book The Everglades: River of Grass focused the world’s attention on the ecosystem and the dangers it faced from human encroachment with that image. Those words saved the Everglades from total destruction. But scientifically the Everglades is not a river of grass; rather, the Everglades is a fen – an alkaline peatland that is one of the great wonders of the world. The distance between these two characterizations did not seem large in the middle of the twentieth century, but today it is a chasm with potentially disastrous consequences. The Army Corps of Engineers Comprehensive Everglades Restoration Plan (CERP 1998, 2005) is a proposal to pump or flow massive amounts of water into the Everglades while creating 87,668 ha of storage reservoirs. One could argue that the Corps’ literal acceptance of the scientifically incorrect “river of grass” metaphor has set us off on a path that will not restore the historic Everglades but replace it with a totally managed marsh ecosystem that no longer hydrologically or ecologically functions as a rainfall-driven peatland ecosystem (see Chap. 2 for a full discussion of peatland hydrodynamics and a scientifically based classification of the Everglades). That is not to say that the hundreds of dedicated scientists, engineers, and water managers who are working tirelessly on Everglades restoration under CERP have not made great progress; they have. However, their work has become even more difficult as the rapidly expanding human population and agricultural activity make ever-increasing water demands on the presently diked and channelized Everglades landscape (see Plate 1). Nevertheless, the question that scientists and policy makers working toward a true restoration of the Everglades must ask is whether CERP is based on sound ecological principles and an understanding of the range of peatland ecosystem processes that need to be restored. To that end, I and my fellow contributors have not written a book about rivers or grass. Instead, this volume’s purpose is to contribute to the restoration of the Everglades fen ecosystem. It is our hope that our research will provide insights to aid in this restoration process as well as raise questions about current approaches 1

2

C.J. Richardson

or suggest new adaptive management scenarios that might be attempted to complete this difficult restoration. As the title suggests, this book is focused on experimental research findings and long-term gradient studies in the Everglades that can provide a scientific basis to better understand this complex wetland ecosystem’s structure and functioning. It presents nearly 16 years of experimental research and data synthesis done in the Everglades by researchers, graduate students, and faculty at the Duke University Wetland Center as well as scientists from 11 other institutions. While the project began in 1989 and continued into 2005, the core data analyzed in this volume comes from work done primarily from 1989 until 2003. Our research looked at biogeochemical cycling, hydrologic processes, soil chemistry, and peat accretion, factors which control or can be related to plant community responses, algal populations and diversity dynamics, macroinvertebrate populations, and ecosystem-level processes such as productivity and decomposition. Additional key work was done on the paleoecological analyses of diatoms and pollen of the Everglades to determine historic conditions and communities, which can be used to help set restoration goals. During those years, the 18 Ph.D. research faculties, 16 graduate students, and 23 research staff who worked on the project have written over 150 refereed articles, book chapters, and reports about our research. Now, this book is the synthesis of our key findings. This volume presents new and revised information on topics such as (1) the role of nutrients vs. hydroperiod on plant communities in the Everglades peat-based ecosystem, (2) sawgrass germination and plant biomass (above- and belowground) nutrient and litter accumulations along a nutrient gradient, (3) a model of the spatial extent of elevated P in the Everglades soils, (4) precipitation chemistry, (5) longterm analysis of tree island loss, (6) paleoecological assessments of historic pH shifts and algal community changes, (7) predictive models of P assimilative capacity, (8) P thresholds for varying trophic levels, and (9) P storage rates as it relates to Stormwater Treatment Areas (STAs). Each chapter also discusses how these scientific findings might provide “lessons for restoration.” In fact, a number of the CERP goal indicators (SFWMD 2005) are directly addressed by our studies. For example, we provide scientific data on sheetflow and hydropattern (Chaps. 7 and 8) that may aid in refinement of hydrologic indicators and flow goals. Water quality indicators that cover TP/SRP relationships as well as periphyton mat cover and structure are found in Chaps. 6, 9, 12, 15, and – especially – 25. The CERP’s biological or landscape-level indicators include tree islands and the extent of cattail habitat changes over time, and these are covered in detail in Chaps. 8 and 9, respectively. The book is organized in five sections. Part I (Chaps. 2–4) provides a brief overview of the historical and current ecological status of the Everglades, soils characteristics, and vegetation and algae of the Everglades. These chapters are not intended to be a comprehensive review, but they do stress our need for understanding changes in the hydrologic, biogeochemical, and soil processes as well as the natural succession that occurred in the Everglades peatlands. For more detailed summaries, readers can consult earlier excellent background synthesis volumes on the Everglades by

1 Overview of Gradient Studies and Experiments

3

Gleason (1974a, 1984) and Davis and Ogden (1994a), and a more recent volume on the Everglades and Florida Bay by Porter and Porter (2002). Readers would also be remiss if they did not take the time to read J.H. Davis’s classic 1943 volume The Natural Features of Southern Florida to get an assessment of the system prior to massive agriculture and urban development effects. Part II presents our gradient studies to assess long-term effects of nutrient additions on key ecosystem components. Here long-term (6–15 year) studies are presented on water and soil chemistry (Chap. 6), geologic and hydrologic settings (Chap. 7), tree islands losses (Chap. 8), macrophyte responses to nutrient additions, altered hydroperiod and fire (Chap. 9), algal responses (Chap. 10), invertebrate responses (Chap. 11), paleoecological analyses of vegetation and water quality changes (Chap. 12), and carbon cycling and dissolved organic matter export (Chap. 13). Part III focuses on our 6-year P dosing experiment. Topics include experimental design and operation (Chap. 14), water and soil chemistry responses to P dosing (Chap. 15), plant community response (Chap. 16), decomposition rate changes (Chap. 17), algal responses (Chap. 18), and macroinvertebrate and fish responses (Chap. 19). Fertilizer and disturbance experiments were done to assess concurrent N and P addition additions as well as vegetation removal, and results are given in Chaps. 20 and 21, respectively. Germination experiments (Chap. 22) on sawgrass and cattail were done to determine factors controlling fecundity, seed germination, and in turn give some insights on the reestablishment of these species. Part IV presents models of biotic and ecosystem-level responses to P additions. Long-term P assimilative capacity is covered in Chap. 23 and spatial distributions of P in the soils of the Everglades in Chap. 24. The ecological basis for a P threshold is detailed in Chap. 25 using a Bayesian predictive model, which includes an assessment of risk. In Part V, Chap. 26 provides an ecological primer for the restoration of the Everglades, presenting a further analysis of the “lessons for restoration” that evolved from our findings. Our research focused on biogeochemical cycling, hydrologic processes, soil chemistry, and peat accretion, factors which control or can be related to plant community responses, algal populations and diversity dynamics, macroinvertebrate populations, and ecosystem-level processes such as productivity and decomposition. Additional key work was done on the paleoecological analyses of diatoms and pollen of the Everglades to determine historic conditions and communities, which can be used to help set restoration goals. This book is, however, limited in scope and does not provide data on bird population ecology nor does it deal with mammals or reptiles. Fish data are confined to one specific experiment. Our data from the Loxahatchee Wildlife Refuge (WCA-1) and Everglades National Park (ENP) are also limited due to the very restricted access provided to our researchers by the Federal Government. While some monitoring work is presented, it was not a central focus of our efforts, and the reader is referred to the annual reports of the work of the South Florida Water Management District (SFWMD 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006), and the CERP website (http://www.evergladesplan.org/) for comprehensive monitoring data and reports.

4

1.2

C.J. Richardson

Project Overview

The Everglades is a large, heterogeneous ecosystem in which vegetation–animal– environment linkages have received much attention. However, little is known of the environmental thresholds, which cause a shift in whole ecosystem stable states (Gunderson et al. 2002). We know that communities in this wetland ecosystem have been affected by a variety of anthropogenic influences in the past several decades, which has led to a surge of recent studies designed to infer the causes of observed changes (Davis 1994; Davis and Ogden 1994a; Richardson et al. 1999; Porter and Porter 2002; King et al. 2004). Landscape fragmentation, cultural eutrophication, and introduction of toxic substances have also occurred in the Everglades, and they have been shown to cause changes in both structure and function of aquatic ecosystems resulting in alternative stable states (Holling 1973; Scheffer et al. 1993; Gunderson et al. 2002). These disturbances have not often been studied along welldefined nutrient and hydrologic gradients. Multipurpose management objectives have caused these altered hydrologic and nutrient regimes throughout the Water Conservation Areas (WCAs) and portions of the ENP (Toth 1988; Walters et al. 1992; Reddy et al. 1993; Craft and Richardson 1993a; Qualls and Richardson 1995; Richardson and Qian 1999; Davis and Ogden 1994a; McCally 1999; Porter and Porter 2002; Lodge 2005). The long-term ecological effects of changes in hydroperiod and increased nutrient loadings during the past three decades in the WCAs have not been studied under experimental conditions until recently. Our research emphasized the experimental approach (with controls) to support our long-term observational studies that were established in 1990. In addition, we utilized different scales of study (laboratory, microcosm, mesocosms, field-scale gradient studies and experiments, and watershed-level studies) to address the following four central questions: 1. What are the effects of increased nutrients (especially P), fire, and water inputs on the native plant and animal communities in the WCAs? 2. What is the long-term nutrient storage capacity of the WCAs? 3. What are the effects of changes in water level and hydroperiod on Everglades structure and function? 4. How can water and nutrients be managed to maintain and restore the natural communities of the Everglades? To address these questions, Duke University Wetland Center researchers developed an integrated ecosystem research program of detailed multiple-scale experiments and long-term observational studies that would provide information useful to restoration efforts in the Everglades ecosystem (Fig. 1.1). The concept was that microcosm and mesocosm experiments would help explain the mechanisms controlling the alternate stable states and community shifts we were finding along the hydrologic and nutrient gradients. In addition, the landscape-level work provided key information on factors controlling nutrient biogeochemistry, the importance of rainfall inputs to overall nutrient budgets, the relationship between hydrology and

1 Overview of Gradient Studies and Experiments

5

Fig. 1.1 An organizational chart of the studies and experiments that comprised the Duke University Wetland Center Integrated Everglades Research Program. Studies ranged from microcosm experiments to large-scale ecosystem gradient studies. Research contributing directly to Everglades restoration include studies of organism responses to P additions, hydrologic shifts, disturbance studies as well as numerous process-level studies focused on understanding biogeochemical fluxes

nutrients and plant community, and macroinvertebrate responses as well as natural and management factors controlling hydrologic, carbon, and nutrient budgets. A key focus of our effort was the role of P additions on trophic-level responses for bacteria, algae, macrophyte, macroinvertebrate, and community and ecosystem-level responses as well as determining P storage, P assimilative capacity, and P thresholds (Fig. 1.1). Concurrently, we measured the changes in ion chemistry for other key nutrients like nitrogen as well as key cations (Ca, K, Mg, Na, etc.) and anions (SO4, NO3, Cl, etc.) plus trace metals and mercury.

1.3

Database Information Supporting the Research

Data collected and analyzed by the Duke University Wetland Center in the past 14 years from a series of long-term studies conducted in the Everglades have been compiled in an extensive database stored in ACCESS at the Duke Wetland Center. The database includes water quality measurements, soils chemistry, peat accretion data, long-term frequency count and cover data on plants, algae counts, and biovolume as well as macroinvertebrate data collected along human-made nutrient and hydrologic gradients. Similar data are available for a replicated mesocosm

6

C.J. Richardson

Table 1.1 Overview of Duke Everglades data records for experimental and observational sites in the Water Conservation Areas Area study Period Records Sites WCA-2A Dosing experiment Eastern gradients Spatial analysis Western gradient WCA-2B

1990–2000 1989–2003 1997–2001 1997–2003

119,102 19,909 296 21,07

2 sites, 12 channels 3 transects, 24 locations 74 locations 1 transect, 7 locations

Fertilizer experiment Disturbance experiment WCA-3A/B

1990–1994 1992–1999

16,279 7,389

3 sites, 27 plots 1 site, 15 plots

Northeastern gradients 1998–2003 5,275 2 transects, 20 locations Spatial analysis 2001–2002 3,136 51 locations Fire study 2002–2003 14,834 10 locations Physical impact study 2003 264 8 locations Data inventories show the study areas, period of study, records, and number of site locations

P threshold experiment conducted in the undisturbed area of the Everglades. A disturbance study where all vegetation was removed provides data on plant succession water quality and hydrology. Data on the effects of both N and P fertilization on plant growth and nutrient status are also available. This archived database has been collected following EPA QA/QC guidelines. Importantly, the database includes ecological metric data from the fastest (periphyton), intermediate (macroinvertebrates, sawgrass, cattail, etc.) and slowest processes (peat accretion, nutrient accretion in the peat) from both the natural state and the alternative states. In addition, data are available from a series of explicit spatial analysis sampling studies. The database used in this volume is available upon request provided that requesters give the appropriate credit to the investigators. An earlier version of the database was filed with the Florida Department of Environmental Protection. Table 1.1 provides an overview of the study site locations, type of study, number of study sites, dates of sampling, and number of records.

1.4

Location of Gradient and Experimental Study Sites

A map of south Florida depicting the WCAs of the Everglades, the Everglades National Park (ENP), Lake Okeechobee (LO), Everglades Agricultural Area (EAA), and surrounding wetlands like the Big Cypress Swamp (BCS) are presented in Fig. 1.2. The research locations for the gradient studies in both WCA-2A and WCA-3A are shown along with the sites for the dosing, fertilizer, and disturbance experiments. Site location maps for additional gradient studies of changes in water quality, soils, and vegetation in WCA-2A and WCA-3A appear later in this volume in the appropriate chapters. All sampling site GPS coordinates are provided in the database.

1 Overview of Gradient Studies and Experiments

7

Fig. 1.2 A general map showing the location of the major wetland, agricultural, and lake areas in south Florida. Specific areas include the Everglades National Park, Water Conservation Areas, Big Cypress Swamp, Loxahatchee Wildlife Refuge, Everglades Agricultural Area, and Lake Okeechobee. The relative study locations of our research sites are as follows (1) nutrient gradients are depicted by open circles, in WCA-2A and WCA-3A, (2) square represents the P dosing experiment location in lower WCA-2A, and (3) the disturbance and fertilizer experimental areas are shown in WCA-2B by a star and rectangles, respectively, on the exploded portion of the diagram

1.5

The Past, the Present, and Lessons for Restoration

The predisturbance community structure of the northern Everglades ecosystem was historically sustained by fire and hydropattern (Loveless 1959; Craighead 1971) while limited by phosphorus (P) (Steward and Ornes 1975a, b; Davis 1991; Noe et al. 2001) and nutrient release related to fire events (Duever et al. 1976; Richardson and Huvane 2001). In the lowest elevations with the longest hydroperiods, deepwater sloughs containing Nuphar lutea (L.) (yellow pond lily; all taxonomy according to the Integrated Taxonomic Information System) and Nymphaea odorata Ait. (American white water lily) predominate. With increasing elevation and decreasing hydroperiod, a diverse wet prairie mix of flood-tolerant grasses and forbs transforms to a fen dominated by Cladium mariscus (L.) Pohl ssp. jamaicense (Crantz) Kükenth (sawgrass). At the highest elevations, lowest hydroperiods, and under low fire frequency, structurally and taxonomically diverse hardwood and shrub tree islands occur. While still evident in the central and southern Everglades, the heterogeneous mosaic of vegetation communities

8

C.J. Richardson

described by Davis (1943) has been altered by the disruption of natural environmental variation (Porter and Porter 2002) and degradation of water quality (SFWMD 1992; Davis and Ogden 1994a; Richardson et al. 1999). While modifications to hydrology, fire frequency and intensity, and other environmental factors are suggested to play a role in the alteration of biotic distributions in the Everglades, P-enriched runoff from the EAA has been identified as the primary stressor (SFWMD 1992; Koch and Reddy 1992; Debusk et al. 1994; Richardson et al. 1999). The extensive canal-and-levee system that compartmentalizes the remnant Everglades also serves as a conduit for P, Na, and SO4 from the EAA, and water-control structures along the canals function as point sources of these and other ions to downstream portions of the wetland ecosystem. In areas near watercontrol structures, P has been found to be at least partially responsible for transforming Cladium stands and primarily responsible for changing open-water sloughs to dense stands of invasive Typha domingensis Pers. (southern cattail, Typha) (Davis 1991; Urban et al. 1993; Newman et al. 1998; King et al. 2004). Typha distribution and growth are positively correlated with both soil and water total P and limited in areas with low P (Craft and Richardson 1997; Doren et al. 1997; Miao and Sklar 1998; Miao et al. 2000). Mesocosm studies also have demonstrated that Typha is more competitive than Cladium under high-P conditions (e.g., Newman et al. 1996). However, fertilizer experiments have been unable to show that adding P alone necessarily results in competitive exclusion of Cladium (Craft et al. 1995; Chiang et al. 2000). These experimental findings suggest that other factors, such as hydropattern (e.g., Toth 1987; Urban et al. 1993; Newman et al. 1996), fire (e.g., Gunderson and Snyder 1994; Urban et al. 1993; Newman et al. 1998), or cations such as sodium (Craft and Richardson 1997) may be important synergists in Typha expansion. This volume presents detailed information to further elucidate the mechanisms controlling shifts in plant and animal communities. Because the autecology of both Typha and Cladium has received most of the attention from researchers, few studies have examined patterns of whole ecosystems in response to these human influences. The general observation has been that high-P areas are dominated by monotypic stands of dense Typha (e.g., Jensen 1995; Rutchey and Vilchek 1999), resulting in a relatively homogeneous landscape pattern when compared to the predisturbance ridge-and-slough mosaic (Obeysekera and Rutchey 1997; Wu et al. 1997). However, this observation has been largely based on photointerpretation of satellite imagery, which is limited in spatial resolution and is not appropriate for assessing fine-scale pattern in species composition (Obeysekera and Rutchey 1997; Richardson et al. 1997a). Recent field studies (Chap. 9) have indicated that many other macrophyte species coexist with Typha in high-P areas, and that biotic diversity due to exotic invasions is actually greater near canals than in interior-wetland locations (Doren et al. 1997; Vaithiyanathan and Richardson 1999). However, little is known about the shifts in patterns of these communities, the environmental factors that are responsible for generating these patterns, or the influence of reductions in nutrient inputs and shifts in hydrologic conditions on changes in alternate ecosystem stable states (Carpenter et al. 1999; Gunderson et al. 2002).

1 Overview of Gradient Studies and Experiments

9

Additionally, intercorrelations and spatial autocorrelation of multiple factors have made it difficult to isolate the linkages between specific environmental variables and observed changes in vegetation patterns (King et al. 2004). Fortunately, the researchers at the Duke University Wetland Center have maintained a long-term series of ecological studies at key locations along the nutrient and hydrologic gradients in both the natural and altered states. This provides the basis for our ability to present in the following chapters the interactions among fast and slow responding species and determine ecological metrics indicating alternative stable states and species and community thresholds. Importantly, lessons that may aid in the restoration of the Everglades based on our scientific findings are briefly presented at the end of appropriate chapters and are fully discussed in the final restoration chapter. It is hoped that this information will be utilized to help restore the remaining Everglades to the functioning fen system that it once was as well as provide a baseline for measuring future changes that will occur under current and future water and nutrient management plans. Many questions remain on how best to accomplish the extremely difficult task of sustaining a functioning Everglades. For example, one central question is how to create hydrologic and nutrient regimes that could lead to the reduction of cattail monocultures in localized areas and facilitate the replacement of disturbed communities with a diversity of natural communities. We have started on a path to significantly reduce P loadings from agriculture via the construction >16,000 ha of the STAs in former agricultural areas. Will the proposed area be enough to reach the criterion of 10 µg l−1 of TP? Recent findings from the SFWMD (2004, 2005, 2006) suggest that more STAs are needed to reach this goal (Chap. 2). We also provide additional data in this volume as to the processes and rates controlling the storage of P. Interestingly, early work by our research group suggested that at least 24,000 ha would be needed to reduce P concentrations to appropriate levels (Richardson and Craft 1993), but the political will and economic ability to acquire this much land initially was lacking. Another important issue to be considered is whether the reservoir of stored P in the already contaminated soils will prevent or delay the return of native species. Our sorption/desorption studies provide some clues that may help answer this question (Chap. 6). The CERP plan proposes an extensive series of alterations and modifications of the current hydrologic flow and storage system (Fig. 8.1). Does the newly developed flow scheme have any chance of restoring the Everglades’s historic condition as a functioning peatland? Our analysis would suggest that it might, but only in selected areas. A case in point is Water Conservation Area 1. Over the last 1,000 years, WCA-1 had evolved into a rainfall-driven nutrient poor (ombrotrophic) bog system (acid pH water dominates the interior). Now calcium- and nutrient-laden water being routed into this system under current restoration plans is greatly altering the chemistries and, in turn, the communities in this ecosystem. The reasons behind routing the water into WCA-1 are many, but the scientific data do not support the addition of any surface waters to any ombrotrophic bog if true peatland development and restoration are the goal (see Chap. 2 for a further discussion of peatland hydrodynamics and development in the Everglades). Such problems arise because

10

C.J. Richardson

the specific ecosystem processes needed to restore each community type in the Everglades were not considered and also because there is no plan based on a full understanding of peatland development, hydrodynamics, and succession (Moore and Bellamy 1974; Gleason 1984; Rydin and Jeglum 2006; see Chap. 2). Information from our research on the mechanisms controlling animal and plant community stability and succession will hopefully provide some additional insights to aid in the proper design and development of the adaptive management plans for the Everglades. Additionally, our information will provide some new understanding concerning the long-term response of the Everglades to reduced nutrients and increased water inputs providing the hydrologic regimes are not radically altered in the future. Finally, there is no question that our primary goal should be to maintain biotic diversity and peatland hydrodynamics as well as ecosystem function within the wetland communities of the Everglades. This can only be accomplished if we understand the effects of water, nutrients, and fire on the Everglades ecosystem and use scientific findings to better our restoration efforts. We hope this volume contributes to that goal in some small way.

Part I

Everglades Ecosystem

2

Ecological Status of the Everglades: Environmental and Human Factors that Control the Peatland Complex on the Landscape Curtis J. Richardson and Jacqueline K. Huvane

2.1

Introduction

The Everglades was an almost impenetrable wall of sawgrass “plains” and reptile-infested waters according to the early Spanish and American explorers (Ives 1856; Lodge 1994). Its name may have come from the term “Never Glades” as first used by Vignoles (1823). Originally called Pa-hay-okee (“grassy lake”) by the resident Native Americans, the Everglades was later popularized and put forward as a threatened environment that needed federal protection by Marjory Stoneman Douglas’s seminal 1947 book The Everglades: River of Grass. Her wonderful “river of grass” metaphor has unfortunately led to a simplistic view of the complexities of the Everglades ecosystem, how it functions on the landscape, and how its diversity of communities should be managed to sustain this subtropical wetland (McCally 1999). It is often referred to as the “Everglades marsh or swamp” by local residents, biologists, and engineers; however, it is correctly identified as a fen (Richardson 2000; Keddy 2000; Rydin and Jeglum 2006; Grunwald 2006). In more generic terms, the entire wetland would be referred to as a peatland by wetland ecologists in North America or as a mire by those in Europe. A mire is a wet terrain dominated by living peat-forming plants and is often used in botanical and ecological investigations of vegetation types. “Peatland” is a more universal term used to define a terrain covered by peat, usually to a minimum depth of 30–40 cm. Even if the site is drained, it is still a peatland, but if it loses its original peat-forming plants it is no longer considered a mire (Sjors 1948; Rydin and Jeglum 2006). The terms peatland and mire are therefore not used interchangeably by peatland ecologists in Europe and North America. Here we use “peatland” to generally represent the complex diversity of community types found on peat soils in the Everglades and “fen” in a more strict sense to represent alkaline peat or calcium mineral-based ecosystems found over vast portions of the Everglades landscape. Future detailed research on groundwater flows and geochemistry will be needed to distinguish which specific locations within the Everglades function as true groundwater-influenced fens vs. peatland types with surface- or rainfall-dominated inputs (discussed later in Sect. 2.5). Thus, the Everglades should not be classified as a swamp because it is not a forest-dominated wetland, and it is not technically a marsh because marshes 13

14

C.J. Richardson and J.K. Huvane

are characterized by standing or slow-moving water with submerged, floatingleaved, or emergent plant cover rooted primarily in mineral soil with nutrient-rich overlying waters (Rydin and Jeglum 2006). These are important distinctions when one considers how different marshes and swamps are from peatlands in terms of their hydrologic controls, biogeochemistry, rates of peat accretion, plant and animal communities, and successional development. To maintain some continuity of terms in this volume with historic usage we do use the word marsh to refer to specific community types like cattail marsh. Nevertheless, the terms “Everglades peatland” or “fen” by themselves do not reveal the vital and multifaceted hydrologic connections and nutrient sources that historically existed between the Everglades and surface water runoff coming from Lake Okeechobee via the Kissimmee River, the close connections of groundwater and surface waters in the region due to the karst limestone underlying the wetlands, and most importantly the seasonal influence of the key water source – rainfall (Parker et al. 1955; see Chap. 7). The Everglades peatland complex was created by blocked drainage due to development of limestone substrata of various porosities overlain on a flat basement rock and confined by sandy ridges that developed from sea level rise and fall about 5,000 years before present (YBP) (Gleason and Stone 1994). The thin layer layers of porous rock that formed during earlier glacial periods absorbed, stored, and transmitted water at different rates, a characteristic crucial to the formation of a myriad of different plant communities found in the Everglades mire even today (McCally 1999). For example, the landscape, while dominated by sawgrass, is interlaced with periphyton-rush sloughs, wet prairies, and ponds, and it is dotted with tree islands and willow heads. The proportion of each community type varies greatly along a north-to-south hydrologic gradient (see Chaps. 4, 7–9, and 12). Most of these plant community associations evolved under low phosphorus (P) concentrations because the main source of water was rainfall with extremely low P concentrations (Redfield and Urban 1997). The exception to communities evolving under low P concentrations are tree islands and the vegetation around alligator holes (Davis 1943; Loveless 1959; Steward and Ornes 1975b; Craft and Richardson 1993a; Sklar and van der Valk 2002) as well as plant communities adjacent to Lake Okeechobee with its high historical TP concentrations >30 µg l−1 (Walker 2000). Another factor maintaining P limitations in the Everglades, unlike northern peatlands or fens, is the nitrogen-fixing blue-green algae community, or periphyton, found in open-water sloughs. Because of the periphyton community’s high rates of nitrogen fixation, Everglades soils are exceptionally high in nitrogen (2–4% by weight); thus, very high N:P ratios (>100) exist, further driving the system to severe P limitations (Richardson et al. 1999). To fully understand the Everglades ecosystem, it is necessary to understand how human interventions over the last one hundred years have dramatically altered the natural Everglades development processes that started more than 5,000 years ago. Thus, the objectives of this chapter are to provide the reader with a basic foundation for understanding how the Everglades ecosystem complex has developed and to provide an analysis of factors controlling ecosystem functions in the Everglades today. It is not our intent to review in detail the geological formations and processes

2 Ecological Status of the Everglades: Environmental and Human Factors

15

that have led to the development of the Everglades, as so many great articles and volumes have been written on this topic (e.g., Brooks 1968; Gleason 1974a; Perkins 1977; Gleason and Stone 1994). To accomplish our goal and help interpret the specific research and restoration lessons presented in the chapters that follow we (1) present a brief review of Everglades peatland formation and characterize the wetland processes that led to development of this vast peatland complex, (2) provide a proper classification of the Everglades system that might help in the development of a more appropriate restoration management framework, (3) review the factors controlling ecosystem structure and succession of communities found within the peatland complex today as compared to historical conditions, and (4) compile and synthesize historical and current data on some key elements of precipitation trends, hydrologic shifts, and nutrient inputs on a landscape scale.

2.2

Formation of the Everglades: The Historical Everglades Prior to Major Anthropogenic Impacts

One of the key benefits of examining the long-term history of the Everglades is that it is possible to learn about natural variations in the system prior to the industrial era as well as determine what environmental factors controlled the formation and development of the Everglades. Knowledge of the rates and magnitudes of change, as well as of recovery rates from disturbances, is critical to future restoration plans. Restoration plans that incorporate natural variation and known responses to disturbance are also likely to be more ecologically and economically feasible. The Everglades mineral substrate formed a large basin or trough during the Pleistocene, and shallow marine sediments were deposited, primarily during the Sangamon interglacial stage 125,000 YBP (Davis 1943; Parker and Cooke 1944; Gleason 1984). The retreat of the northern U.S. glaciers 18,000–16,000 YBP, blockage of drainage from the Everglades due to rising sea level, a change to a subtropical climate, and the concurrent increase in rainfall allowed for the development of the Everglades as we know it. Three limestone formations underlie the Everglades. The Miami Formation is found in the southern Everglades National Park (ENP) region; the Anastasia Formation, comprised of sandy calcareous sandstone, is found in the northeast area; and the Fort Thompson Formation, which underlies the northern half to a depth of 50 m, is mostly marine and freshwater marls, limestone, and sandstone (Enos and Perkins 1977). A geological study of the bedrock that underlies the Everglades shows a differentiation in permeability from north to south. Low-permeability limestone underlies the northern portion of the Everglades basin around Lake Okeechobee and extends into the northern half of WCA-3 and into the western portions of WCA-2. In the southern section of WCA-3 and the southeastern section of WCA-2B, there is an abrupt shift to highly permeable limestone (Gleason et al. 1974; Perkins 1977). This has important ramifications for the movement and storage of water, peat development, and the establishment of plant communities. Moreover, construction

16

C.J. Richardson and J.K. Huvane

of any water storage areas in the lower eastern part of the Everglades would be subject to severe water loss unless extensive and expensive efforts were made to line the reservoirs due to the high permeability of the underlying bedrock. According to Gleason et al. (1974), bedrock configuration established the drainage directions prior to peat deposition in the Everglades. For example, Lake Okeechobee flowed through a channel eastward to the area now known as WCA-1, and a deep depression bisected the lower Everglades and created a southwest flow gradient toward Florida Bay (Fig. 2.1). These patterns appear to have changed little over the course of time. For example, Gleason et al. (1974) note that tree island orientation is correlated with drainage directions expected from bedrock topography. The only detailed vegetation map of the Everglades came from early survey work of Davis (1943) and was based on his extensive field observations in the late 1930s (Fig. 2.2). Although, the mapping was done prior to any massive increase in farming in the Everglades Agriculture Area (EAA), many of the large canals had been dug and peat subsidence had started according to his field notes. His map provides distributions

Fig. 2.1 Bedrock map of the Everglades prior to peat development based on kriging of USGS depth measurements and isopleths maps (Parker et al. 1955; Parker and Cooke 1944). Darker shades represent higher regions (bedrock plateau south of Lake Okeechobee, etc.) and lighter shades represent depressions or troughs in the bedrock (e.g., in WCA-1 and in lower WCA-3A, WCA-3B and in the northern portion of the ENP where Taylor slough is now found). Also shown are basal dates of peat from 14C measurements (McDowell et al. 1969; Gleason et al. 1974; Craft and Richardson 1998)

2 Ecological Status of the Everglades: Environmental and Human Factors

17

Fig. 2.2 Historic map of the vegetation communities in the Everglades based on the map of Davis (1943). The map has been redrawn and simplified from the original map, and the boundaries of the current Water Conservation Areas (WCA-1, WCA-2A, and WCA-3A), the Everglades National Park (ENP), and the Everglades Agricultural Areas have been added

18

C.J. Richardson and J.K. Huvane

of all the major plant communities and indicates that higher densities of sawgrass stands existed in the northern Everglades. He clearly mapped wet prairies and ombrotrophic (rainfall-driven) areas in WCA-1. The map also demarks slough and ponds areas, mangrove swamps, and vast areas of tree islands as well as pointing out a large stand of ferns and cattails (> 8,000 ha or 20,000 acres) in what is now the EAA. The reason for the prevalence of cattails, which are commonly found in P-enriched areas, is unknown because farming had not started to any extent; however, Davis (1943) suggested that it may have been related to fires occurring shortly before the mapping, which would have released P from the burned peat.

2.2.1

Peat Formation

Paleoecological studies based on the examination of peat or soil cores have been used to provide information about the Everglades system since its formation (Gleason et al. 1974; Willard et al. 2001; see Chap. 12). Radiometric dating analyses such as 14C and 210Pb allow researchers to date the soil strata in the cores. Peat type and pollen remains provide a picture of what the vegetation of the area was like at each time period. Diatom remains provide information concerning water quality in the past, particularly past pH and nutrient levels. A view of the general location of the numerous cores that have been analyzed in the Everglades by various researchers and summarized in this section is given in Fig. 2.3. Peat formation in the Everglades began around 5,000 YBP in the northern Everglades and around 2,000–3,000 YBP further south (Gleason and Stone 1994). These dates are the result of 14C dating of basal peats in numerous peat cores. Thus, the Everglades is geologically a relatively young ecosystem. The rate of peat accumulation from north to south has always been of interest, and peat cores collected from four sites in the Everglades – one each from Loxahatchee, WCA-2, WCA-3A, and the ENP – indicate an interesting trend. Accelerated mass spectrometry 14C dates for samples from depths between about 40 and 50 cm indicate that there has been more peat accumulation in the north. For instance, the date estimated for the 36–38 cm depth from a core in ENP was about 2,400 YBP. However, for the core collected from WCA-2A the date estimated for the interval 33–48 cm was only about 800 YBP (C.J. Richardson, unpublished data). This suggests different patterns of peat initiation and much higher rates of accretion in the northern Everglades compared to the southern. Another indication of this trend was an analysis of peat depths in WCA-2A vs. WCA-3A. Ninety depth probes to bedrock taken during our grid soil surveys in 2001 (Chap. 6) throughout WCA-2A averaged 144.5 ± 37.3 cm in depth, with a median value of 144 cm and a range from 62 to 252 cm (Fig. 2.4). Several deep peat areas were found in the northwest and southeast portions of WCA-2A. Peat was generally deeper in the northern than the southern part of WCA-2A. Fifty-nine depth probes taken in 2000–2003 throughout WCA-3A averaged 80.1 ± 36.3 cm in depth, with a median value of 76.5 cm and a range from 16 to 180 cm. Peat depths were shallowest in the northern part of WCA-3A

2 Ecological Status of the Everglades: Environmental and Human Factors

19

Fig. 2.3 Location and sources of peat core sites in Water Conservation Areas and the Everglades National Park used to assess historic communities

compared to the southeast corner of WCA-3A (Fig. 2.5). Severe fires in the north along with dry, peat-oxidizing conditions may have contributed to the shallower peat depths in the north. Thus, peat depths in the more northern WCA-2A were 1.8 times those found in the southern Everglades in WCA-3A and clearly demonstrated a north-to-south trend in peat accumulation over the past 5,000 years (Chap. 3), corroborating the evidence from the radiometric dating of peat cores. The soils of the Everglades are recent Holocene Histosols and Inceptisols (Gunderson and Loftus 1993). The soils are primarily peats and mucks that had accumulated to a depth of nearly 4 m in the north but are less than 20-cm deep in portions of the ENP (Stephens and Johnson 1951; see Chap. 3). Thus, historic northern peat depths are nearly twice what is currently found. The deepest peats in the southern Everglades are found in depressions and major water flows, such as the Shark River slough. Gleason (1984) dated the basal peats and found that peat deposition began as early as 5,490 ± 90 YBP, but most peats date from 2,000 to

20

C.J. Richardson and J.K. Huvane

Fig. 2.4 Isopleths of peat depth to bedrock in WCA-2A taken during 2001. The map is based on 90 sample depth measurements

4,500 YBP. However, the tree islands are more recent formations and date only from 1,300 YBP (Gleason and Stone 1994; Craft and Richardson 1998). The other dominant and oldest soil type is a calcitic mud, an Inceptisol, formed by cyanobacteria (blue-green algae) that reprecipitate calcium carbonate or marl (CaCO3) originally derived from the limestone substrate (Browder et al. 1994). It is found underlying most of the peatlands and has been dated at 6,470 YBP (Gleason and Stone 1994). It is also sometimes found in layers within the peat, indicating periods of short seasonal hydroperiod as compared to the longer period of flooding required for peat formation by macrophytes.

2 Ecological Status of the Everglades: Environmental and Human Factors

21

Fig. 2.5 Isopleths of peat depth to bedrock in WCA-3A taken during 2001. The map is based on 59 sample depth measurements

2.2.2

Peat Analysis as a Tool to Determine Environmental Changes

Data obtained from the examination of a number of peat core stratigraphies indicate that between 2,000 and 3,000 YBP there was probably a period of reduced seasonal flooding in the central to southern Everglades (Fig. 2.6; Gleason et al. 1974). Data from sites in the north were less conclusive, indicating only that an environmental change had occurred. Several peat cores (Gleason et al. 1974; Gleason and Stone 1994; Stone 2000) indicate that vegetation shifts at a given coring site were common over the last 4,000–5,000 years in the Everglades (Fig. 2.6). Unfortunately, many of these cores have only a basal date, making it difficult to determine exactly when

22

C.J. Richardson and J.K. Huvane

Fig. 2.6 Peat stratigraphy of tree islands from WCA-1 and WCA-2, modified from Gleason et al. (1974)

these vegetation shifts occurred. In a peat core collected from a tree island in the northern region of WCA-2A (core 7), there was a bottom 14C date of 4,235 YBP in a sapropel layer. Below this dated horizon was a calcite mud layer. Calcite mud layers in peat cores are thought to represent drier periods that favored the precipitation of calcium carbonate. Above the sapropel layer is a thick layer of sawgrass–water lily peat that probably spans a period of about 2,000 years (between about 3,500 and 1,500 YBP). Above this layer is a thinner sawgrass layer, probably spanning the period between about 500 and 1,500 YBP, followed by a proto-hammock layer

2 Ecological Status of the Everglades: Environmental and Human Factors

23

(Fig. 2.6). Although there was no cattail peat found in this core, cattail pollen has been found in the fen (sawgrass/water lily) peat toward the bottom of the core, suggesting that cattail was abundant during the initial stages of peat formation (Stone 2000). Peat cores typically have alternating layers of sawgrass and water lily community indicators over much of their lengths, as shown in core 9 from WCA-1 (Gleason et al. 1974; Gleason and Stone 1994; Stone 2000). It is not known if these vegetation shifts are the result of climatic or hydrologic changes or whether they simply reflect lateral shifts in the vegetation mosaic (Stone 2000). However, of critical importance for restoration and management plans, the data suggest that these Everglades community types were capable of rapid reestablishment when conditions were favorable.

2.3

Indicators of Changed Hydroperiods and Drainage Effects

A study of pollen stratigraphies by Willard et al. (2001) found that between 1,200 and 2,000 YBP, there was an interval of longer hydroperiods and relatively deep water in the northern Everglades as evidenced by the pollen profiles from western WCA-1 and northern WCA-2A. There were shallow water conditions and probably droughts between about 1,200 and 800 YBP. After 800 YBP, there was a return to wetter conditions in the northern Everglades. Pollen indicative of slough vegetation was common until about ad 1930, when there was a shift to taxa indicative of shallow water (Willard et al. 2001). During the last century, there have been rapid changes in the plant communities in the Everglades. Both pollen and peat type show that vegetation patterns over the last 500 years were relatively stable in comparison to the major changes that took place in the twentieth century (Gleason and Stone 1994; Willard et al. 2001). In general, the landscape was less fragmented and less compartmentalized prior to the twentieth century. More recently, there has been a shift in vegetation and loss of sloughs and tree islands (Chaps. 8 and 9). The pollen data suggest that anthropogenic changes to the hydrology of the Everglades had an impact on the vegetation as early as ad 1930, as evidenced by increases in weedy annuals (Willard et al. 2001). Although anthropogenic impacts are often thought to have occurred mainly in the northern sections of the Water Conservation Areas (WCAs), pollen profiles from the southern areas also reveal the effects of drainage (Bartow et al. 1996; see Chap. 12). These changes in the pollen stratigraphy likely reflect changes in the drainage patterns. For instance, by ad 1917, four major canals had been dug, and by ad 1928, the Tamiami Trail construction was complete (Light and Dineen 1994). The next period of development did not occur until the early 1950s, when the eastern perimeter levee was constructed, as well as water control structures that allowed discharge from what was to become WCA-2 into WCA-3. In the 1960s, the WCAs were established, and water control structures that regulated flow from WCA-1 to WCA-2 were constructed (Light and Dineen 1994). Concomitant changes in vegetation occurred between ad 1950 and ad 1960 when there were increases in sawgrass, indicating a drier hydrologic regime. Willard et al.

24

C.J. Richardson and J.K. Huvane

(2001) assert that the drying was induced by anthropogenic changes in the area (e.g., building of canals and levees and the establishment of the WCAs) because rainfall patterns during this period show a higher than average precipitation. Therefore, the drier conditions were not simply the result of climatic changes in precipitation. Bartow et al. (1996) also found that there was a shift to pollen indicative of drier communities after about ad 1950 in cores from the northern and southern portions of WCA-2A. Cooper and Goman (see Chap. 12) saw a similar shift toward drier conditions after about ad 1960 based on pollen from cores in WCA-2A. Another important impact of anthropogenic changes to the natural hydrology of the system is the loss of tree islands. Prior to ca. ad 1950, tree islands were more common in WCA-2A (Chap. 8). The decline has been attributed to water impoundment during the 1960s that resulted in the drowning of tree islands. A study of soil cores collected from existing tree islands indicates that tree islands had formed by about ad 1200. Tree islands were also found to have substantially higher soil TP (concentrations often exceeding 2,000 µg g−1 of TP in the soil profile) and higher elevation than the surrounding fen areas (Willard et al. 2002). This increased P in the tree island soils is due in large part to the roosting of birds on the islands. Thus, tree islands are P hotspots where the birds move P from low P habitats such as sloughs via fish and insect harvesting and deposit it on tree island soils as concentrated bird droppings or guano. These transfers have a significant effect on the increased growth patterns in vegetation surrounding tree islands, especially downstream. Moreover, when these islands burn, P is released downstream, further increasing vegetation growth and diversity (C.J. Richardson, unpublished data). Thus, the decline in tree islands has important implications for the Everglades ecosystem.

2.4 Evidence of the Recent Effects of Anthropogenic Disturbance 2.4.1

Paleoecological Evidence

Although many of the major impacts in the Everglades occurred after ad 1950, anthropogenic effects are evident earlier in the century as demonstrated by both pollen and diatoms in soil cores profiles. Paleoecological studies of the Everglades that span millennia as opposed to recent decades show that vegetative communities have shifted in the past prior to anthropogenic impact; however, these communities typically persisted over long time periods (centuries). Willard et al. (2001) also found that weedy annuals increased around ad 1930, probably due to drainage and even nutrient inputs. Public outcries also suggest that the landscape experienced substantial anthropogenic perturbations prior to ad 1950. Small (1929) published a book entitled From Eden to Sahara: Florida’s Tragedy in which he decried the destruction of natural communities in south Florida. Evidence of nutrient-rich agricultural runoff and runoff from Lake Okeechobee is documented in soil cores from WCA-2A. Cores collected from areas near the outflow of canals from the EAA show increases in Typha pollen after ad 1960 (Fig. 2.7;

2 Ecological Status of the Everglades: Environmental and Human Factors Fig. 2.7 Pollen profile showing changes in vegetation from early 1800s until 1995 in a peat core from an enriched part of the C transect (10C-1) in northern WCA-2A (from Bartow et al. 1996) 25

26

C.J. Richardson and J.K. Huvane

Chap. 12; Willard et al. 2001). Furthermore, analyses of diatoms from soil cores from WCA-2A show that diatom species indicative of eutrophic conditions start to increase after ad 1960 (Slate and Stevenson 2000; see Chap. 12). These data suggest that diatom communities may serve as early warning indicators of eutrophication. Soil cores collected in various areas of the Everglades have provided details concerning how vegetation has changed over time. Cores collected from the nutrientenriched sites (10-C1) contained two distinct zones, the recent zone 1 (0–17 cm) and the older zone 2 (17–45 cm) (Fig. 2.7). The 0–17 cm depth was predominantly comprised of Chenopodia–Amaranthaceae (Cheno-ams) and cattail (Typha spp.). Cheno-ams, such as pigweed, are indicators of terrestrial environments and frequently colonize disturbed areas such as fallow agricultural lands. Cattail often is found in nutrient-enriched areas or areas characterized by extended hydroperiod and/ or higher water levels in the Everglades where it has displaced slough communities or weakened sawgrass areas. The great increase in cattails in the 1960s and 1970s closely follows the artificially higher water levels maintained by the south Florida Water Management District during this period (SFWMD 1992, Fig. 8.4). The increase of other terrestrial species, such as ragweed (Ambrosia), dog-fennel (Eupatorium type spp.), beach-elder (Iva imbricata Walt.), and an unclassified member of the Asteraceae family, suggest the area was drier during the 1980s than in the past. Combined, these trends indicate major shifts in hydroperiod patterns for the site. Below 17 cm, the pollen is predominantly Alismataceae and pine (Pinus spp.). The Alismataceae includes emergent aquatic vegetation, such as duck-potato and arrowheads, which are common in slough communities where surface water is generally present most of the year (Fig. 2.7). Water lily pollen, another component of the slough, also is greater below 17 cm. The large amount of pine pollen below 17 cm is attributable to original pine forests that existed on the Atlantic Coastal Ridge, north of Big Cypress National Preserve, and on the sandy flatlands northeast of the WCAs (McPherson et al. 1976). Pine is a prolific pollen producer, and the pollen is easily dispersed by the wind; therefore the pine pollen in the soil cores may not be indicative of local vegetation (Brown and Cohen 1985). The pollen of several species such as buttonbush (Cephalanthus occidentalis), grasses, and marsh ferns (Thelypteris-type ferns) are present in relatively constant amounts throughout the core. The presence of these species throughout the profile suggests that the large proportion of Cheno-ams and cattail are not diluting the other pollen types in the core. Sawgrass (Cyperaceae) is also present throughout the core, but decreases sharply around 13 cm, suggesting its displacement by other species, especially cattail and Cheno-ams. The distribution of pollen in a core from the unenriched site (Fig. 2.8) lacks some trends present in the core from the nutrient-enriched site. Care must be taken in comparing these two sites because the peat accretion rate at the unenriched site is much lower. Changes in this core correspond to a period of intensive efforts to regulate the hydrology of the remaining Everglades, including construction of canals and levees along the eastern perimeter of the Everglades and construction of the WCAs (Light and Dineen 1994). However, there also appears to be a shift in the distribution of pollen types 8 cm below the surface corresponding to the late 1940s

2 Ecological Status of the Everglades: Environmental and Human Factors Fig. 2.8 Pollen profile showing changes in vegetation from early 1800s until 1995 in a peat core from an unenriched part of the C transect (10C-6) in southern WCA-2A (from Bartow et al. 1996)

27

28

C.J. Richardson and J.K. Huvane

and early 1950s. This 0–8 cm zone has a large amount of pollen from Cheno-ams, ragweed, and buttonbush, as well as a large but decreasing amount of wax myrtle (Myrica cerifera). Of note is the fact that these species really increased in the pollen count in the early 1950s. As mentioned previously, Cheno-ams and ragweed are indicators of terrestrial environments. Buttonbush is a plant species that exists under stable hydrologic conditions (BAPM 1988). The large proportion of buttonbush pollen in this zone in the past few decades may reflect a different hydroperiod caused by impoundment of WCA-2A in 1961. Jones (1948) observed that wax myrtle has been increasing in areas dominated by sawgrass during and after the construction of the canal system. It has been hypothesized that this increase in wax myrtle is a result of a lower water table and reduced fire frequency (Loveless 1959). In the recent years, this species appears to have decreased in importance, possibly due to shifting hydroperiod pattern, i.e., alternating periods of extremely high and low water levels controlled by the District (Fig. 8.4). Below 8 cm (prior to about 1950), there is an increase in pollen representative of aquatic vegetation such as arrowheads and water lily. The increase in aquatic vegetation at depth in this core and in the core from the enriched site suggests that anthropogenic hydroperiod alterations have resulted in a reduction in the extent of aquatic (slough) communities during the past 50 years. This decrease in slough habitat suggests that the present-day Everglades are drier than in the past, presumably as a result of anthropogenic drainage activities. Of interest is the periodic but regular increase and decrease of sawgrass and Typha spp. during the past 200 years. This may reflect long-term weather patterns and resulting changes in hydrologic conditions. As in the enriched core, pine pollen is greatest near the bottom of the core and decreases toward the top. This decrease is most likely explained by logging operations that occurred on the eastern seaboard as the urban communities began to rapidly expand as well as in mid-Florida as the land began to be developed for agriculture (Watts 1975). Several additional arboreal species, Australian pine (Casuarina spp.), and Brazilian pepper (Schinus terebinthifolius) as well as the wetland species melaleuca (Melaleuca quinquenervia) are known as exotic weed species throughout southern Florida, but they are virtually absent in either core except at the upper 5 cm for Melaleuca. These species thrive on dry upland areas or in association with tree islands and are relatively sparse in WCA-2A, comprising less than 10% of the fen area (SFWMD 1992). Melaleuca, however, has become a major weedy pest species throughout the Everglades and now is found on thousands of hectares as a pure monoculture (Bodle et al. 1994; SFWMD 2006).

2.4.2

Changes in Water Quality as a Result of Canal Digging

The examination of diatom remains revealed another effect that the construction of canals had on water quality. Two studies found a higher incidence of the acid-loving diatom genus Eunotia prior to ad 1960 (Slate and Stevenson 2000; see Chap. 12). This indicates that conditions were more acidic prior to about ad 1960. The change

2 Ecological Status of the Everglades: Environmental and Human Factors

29

is probably related to the deepening of drainage canals that took place in the late 1950s, which exposed the underlying limestone bedrock, thus reducing acidity and likely increasing calcareous periphyton mat formation. This suggests that the extensive periphyton mats that we see today in the northern Everglades may not have been present prior to the digging of the large canals.

2.4.3

Summary of the Environmental History of the Everglades

A summary timeline of general environmental changes in the northern, central, and southern Everglades based on the information presented earlier is presented in Fig. 2.9. The paleoecological analysis clearly shows that the Everglades is a dynamic ecosystem that has shown major shifts in plant communities over the last 5,000 years. Peat initiation started earliest in the northern Everglades and last in the south according to basal peat dating. There are also inherent differences between the northern and the southern Everglades as evidenced by bedrock geology, peat type, hydrology, vegetation, and diatoms. The northern part of the Everglades has historically shown higher peat accretion rates compared to the southern areas. The most dramatic changes have taken place during the twentieth century and are associated with development, particularly agricultural practices and the construction of drainage canals. These changes have included a loss of slough areas and tree islands, and an increase in pH. Changes in the northern Everglades affected by agricultural runoff resulted in higher incidences of eutrophic diatoms. Later changes included the replacement of sawgrass and slough communities by Typha (Fig. 2.7). A loss of acidophilic diatoms occurred after about ad 1960, probably related to canal digging that exposed limestone bedrock (Chap. 12). These shifts in flora and an understanding of their causes have important implications for restoration. For example, the plans that are focused on maintaining a calcareous periphyton mat community in areas that were formerly acidic, and therefore probably had few periphyton mats, may not aim for the appropriate community target if historic community restoration is the goal. Furthermore, as the bedrock becomes less exposed over time by soil/ peat accumulation, there may be less calcium and the pH of the water may decline, resulting in a loss of calcareous periphyton cover.

2.5

Classification

As mentioned earlier, the Everglades would be classified as a fen or peatland by most of the world’s peatland ecologists. It is classified as a Palustrine SYSTEM, CLASS Emergent Wetland, SUBCLASS Persistent, WATER REGIME, Semipermanently Flooded, WATER CHEMISTRY Fresh-Circumneutral, SOIL, Organic according to the USFWS wetland classification (Cowardin et al. 1979). However, a more comprehensive functional peatland classification is needed to

30

C.J. Richardson and J.K. Huvane

Fig. 2.9 A summary timeline diagram depicting the overall trend in the vegetation history of the northern, central, and southern Everglades from peat initiation around 5,000 years ago up to the present. The diagram is based on composite information from cores taken throughout the Everglades (see Fig. 2.3)

incorporate the sources of water and provide some indication of potential differences in its nutrient status across the landscape. Several complex environmental gradients caused by the geological substrate, hydrologic flows, and nutrient inputs were responsible for the distinct formation of

2 Ecological Status of the Everglades: Environmental and Human Factors

31

the Everglades peat types and the north-to-south decrease in peat depths. The degree of wetness and aeration, along with gradients of calcium, salinity, pH, and nutrients, controlled the development of the plant communities (Chap. 12). In addition it has long been recognized that the origin of mire water is a major factor controlling peatland development (Du Rietz 1954). These formation features help define its classification. For example, the Everglades is often described as a rainfall-driven ecosystem, which would normally classify it as an ombrogenous peatland (nourished only by precipitation). However, due to the size and complexity of the Everglades it cannot be easily placed into a single classification. To emphasize the chemical source on productivity, wetland ecologists today refer to ombrogenous sites as ombrotrophic. In the past thousand years before drainage ditches and peat oxidation took place, ombrotrophy may have been true for many portions of the Everglades, but is relevant today only for the raised center portion of the Loxahatchee (WCA1A). Today, massive subsidence of peat has taken place due to drainage from deeply cut canals and ditches. Large portions of the Everglades are now nourished by waters that have passed over or through calcareous mineral parent soils and are then released through canal gate structures (Stephens and Johnson 1951; see Chap. 7). Chemistry profiles and gradients found within the Everglades (Chap. 6) make it clear that portions of the Everglades are currently nourished by mineral groundwater and would be referred to as minerogenous or in modern terms are minerotrophic systems (Richardson 1995; Vaithiyanathan and Richardson 1997a). Thus, the current Everglades ecosystem is not simply rainfall driven. The Everglades peatland complex was historically a mixture of defined hydrologic system types. For example, to further define the peatland water flow regime, the term minerogenous is further divided into major hydrologic systems know as topogenous, soligenous, or limnogenous peatlands (von Post and Granlund 1926; Sjors 1948). Topogenous peatlands have flat water tables located in basins with no outlet or a single outlet and inlet. Soligenous peatlands have a slope with directional water flow through the peat or surface. Limnogenous peatlands are located along lakes and streams and are flooded by these waters (Rydin and Jeglum 2006). To simplify terminology, many peatland ecologists would follow the convention of using the term fen for minerogenous and bog for the ombrogenous types. Thus, historically the Everglades peatland complex would have had several types of dominant hydrologic systems. For example, historically a limnogenous peatland that was historically located along the southern edge of Lake Okeechobee no longer exists due to the building of the Hoover dike in the 1930s (Fig. 2.10a). The main hydrologic system for the Everglades would have been classified as soligenous with a minor slope and water flowing generally south (Parker et al. 1955). The center portions of WCA-1 would have originally been topogenous (Fig. 2.1), eventually evolving into the ombrotrophic system it is today (Fig. 2.10b). It is important to realize that the dominant water sources for various parts of the Everglades naturally evolved but have also been greatly altered by the vast system of canals and dikes that have been installed since the early 1900s. Today most of the Everglades would be loosely classified as soligenous, but in reality it is totally a managed system and should be reclassified as managenous (managed water flow) (Fig. 2.10b).

32

C.J. Richardson and J.K. Huvane

Fig. 2.10 A comparison of hydrologic peatland classifications in the historic Everglades with the present-day Everglades. Shown are the general categories of ombrotrophy and hydrologic flow, systems that control types of peatland formation

The Everglades may also be classified based on nutrient gradients. The terms oligotrophic, mesotrophic, and eutrophic have been adapted from limnology and used to further explain gradients of increasing productivity due to increasing nutrient availability, especially N and P. The oligotrophic class is somewhat broader than ombrotrophic and includes minerotrophic sites with low pH and underlain by nutrient-poor sandy soils as found in pocosins in North Carolina (Bridgham and Richardson 1993). However, there are sites like the Everglades that would be classified as oligotrophic (low productivity) in minerotrophic conditions with very high pH and Ca content, because P has become limiting due to adsorption to Ca and high N-to-P ratios (Richardson and Vaithiyanathan 1995; Richardson et al. 1999; see Chap. 6). These differences in water source and nutrient conditions for the Everglades peatland (Fig. 2.10a,b) show that an understanding of the hydrologic equivalence concept proposed by Bedford (1996) will be essential to any restoration effort. The hydrologic equivalence concept proposes that hydrological conditions similar to those of the original ecosystem type must be restored on the landscape to restore equivalent ecosystem functions. Landscape hydrologic equivalence will have to be considered in the restoration of the modern Everglades if we ever hope to maintain the diversity of Everglades habitats and communities. For example, care must be taken to maintain the ombrogenous portions of the Everglades like WCA-1A, reestablish

2 Ecological Status of the Everglades: Environmental and Human Factors

33

limnogenous peatlands south of Lake Okeechobee, recreate conditions for soligenous peatlands where topographically possible, and reduce managenous water flow conditions (water pumping and release across narrow outlets) within portion of the Glades. Unfortunately, hydrologic equivalence concepts for peatlands have not been considered in the current restoration plans (CERP 1999).

2.6 2.6.1

Factors Controlling Succession in Everglades Communities Ranking of Factors

Historically, the primary factor controlling the long-term development of Everglades plant communities is climate. The amount and seasonal distribution of water from year to year controlled the hydrologic dynamics of the fen system. The hydrologic conditions in turn controlled the fire patterns. The native seed bank was responsible for the regeneration of endemic plant communities once they were disturbed or altered (van der Valk and Rosburg 1997). Massive landscape development in the past 100 years has resulted in regulated hydroperiods (i.e., the number of days that the Everglades ecosystem has standing water at or near the surface) and altered hydropatterns (the distribution of water within the wetland as noted earlier), which in turn have changed fire frequency patterns and fire intensity. Increased P and N loadings from agriculture and urban runoff and introduced exotic species in the early 1900s have all significantly affected plant and animal communities of the Everglades and the WCAs of today (Craft and Richardson 1993b; Davis and Ogden 1994b; DeBusk et al. 1994; Qualls and Richardson 1995; Vaithiyanathan and Richardson 1997a). Importantly, it has been demonstrated by numerous studies that P is the limiting plant nutrient in the Everglades (Steward and Ornes 1983; Koch and Reddy 1992; Richardson and Vaithiyanathan 1995; Craft and Richardson 1997; Richardson et al. 1999; Noe et al. 2001). Thus, the increase in P concentrations as a result of agricultural runoff may have a dramatic effect on local plant communities in addition to the hydrologic changes associated with water management efforts. The main difficulty for ecologists is in separating the influence of primary climatedriven factors like rainfall, hydroperiod, and fire from the secondary human factors of drainage and flooding, nutrient additions, site disturbance, and exotic species invasions. Moreover, the influence of anthropogenic inputs of nutrients and water varies greatly in each portion of the Everglades, depending on proximity to canal input structures, mode of delivery (i.e., point or nonpoint source) and whether water delivery is seasonally pulsed or continuously released. In other regions like WCA-3A, vast stands of exotic species, such as M. quinquenervia and S. terebinthifolius (Brazilian pepper) provide a seed source for the ever-increasing spread of these species, although intensive and expensive control measures are underway by state and federal agencies. To better understand the current status of the main factors controlling the Everglades plant communities, we present a scaled model of impacts (Fig. 2.11). The general scaling is only given as a basis to understand what is now affecting the

34

C.J. Richardson and J.K. Huvane

Fig. 2.11 Primary factors controlling Everglades wetland plant community succession. The size of the arrows depicts the magnitude of impact. The percentage of the Everglades area that is impacted by each factor is based on area surveys in 2000 and was used to scale the factors

overall Everglades since specific areas of the Glades may be influenced to a larger or smaller degree by any particular factor like exotic species or nutrients. The scaling for each factor is based on the total area that is currently affected by hydrologic management, exotic species, nutrient inputs, etc. (Chaps. 6, 9, and 24). The current status or impact of each factor is scaled from 0 to 10 and indicates that the most important factor controlling the vegetation communities at the present time is the altered hydrologic regimes. As Bedford (1996) has noted, hydrologic shifts are the dominant factor responsible for wetland changes. As a result of the altered drainage patterns that resulted from water management in the Everglades, climate is no longer the dominant factor controlling plant community succession since altered drainage patterns are so dominant throughout the Everglades. Other factors in order of importance are the invasion of exotic species, phosphorus additions, disturbance, and fire. Here disturbance refers to the planned new water structure changes, new water inputs, and Stormwater Treatment Area (STA) releases. One could argue that the scaling of some of these factors should be different, as in the case of fire influences in some years or at some locations. However, our scaling system weights longer overall relative impact factors and shows that current hydrologic management controls everything, including the amount and intensity of fires much more than in the historic past. In addition, fire’s effect is now reduced, especially since fire is an event managed in many areas of the Everglades by the State of Florida Fish and Game Commission (Chap. 9). It is true that massive fires still occur in the Everglades, but they are now often caused by the excessive drying out of areas due to the managed shifting of water from one area of the Everglades to another. The constant manipulation and general lowering of Lake Okeechobee water levels and flows as compared to historic conditions also keep parts of the Everglades far drier than in the past while other areas are maintained with waters deeper than historical levels. This indicates that location and the timing of fires in the Everglades have been affected by water management practices.

2 Ecological Status of the Everglades: Environmental and Human Factors

35

Some may argue that P is the main factor controlling most areas of the Everglades, but when one examines the amount of area impacted by P enrichment, it is not the case. While P enrichment can have a great impact on plant community structure, the actual area of the Everglades that is affected by P enrichment is not extensive. In fact, less area is impacted by P than that by invading exotic species. Qian and Richardson (Chap. 24) found, for example, that WCA-1, WCA-3, and the ENP have 81, 91, and 94% of their area with soil P concentrations less than 500 mg kg−1, a value above this indicating enrichment beyond historic levels (Bruland et al. 2006). Further supporting this view of the limited role of P was the recent finding of Bruland et al. (2006) who reported only 263 ha (0.11%) of WCA-3A displayed soils above 500 mg kg−1. Thus, our working hypothesis regarding controlling factors is that the restoration of the Everglades community complex is dependent primarily on the creation of natural hydroperiods and hydropatterns that must include periods of drought and fire. These conditions must be based on the ecological requirements of the dominant species of each community on the landscape. Of course, the reduction of P input to historic levels is critical to the restoration effort, as is the removal of the exotic species.

2.6.2

Succession

Succession in the Everglades has been summarized by Gunderson and Loftus (1993) (Fig. 2.12). They demonstrate that succession of Everglades communities is influenced mostly by disturbance to the hydrology and, in turn, fire frequency and

Fig. 2.12 Succession patterns in the Everglades (modified from Richardson 2000)

36

C.J. Richardson and J.K. Huvane

intensity. Plant communities (Chap. 4) are found along an elevational gradient that translates directly into a hydrologic gradient, which controls fire intensity and frequency. The gradual build-up of marl soil or peat via accretion (1–2 mm year−1; Craft and Richardson 1993a) results in the gradual increase in elevation, which changes the hydroperiod for the species. Ponds are the wettest sites, and soil accretion eventually allows them to develop into wet prairie communities, then willow heads and even sawgrass if not severely burned. Frequent light fires have little effect on this successional sequence. Severe fires burn the peat soil and lower the sites, which results in a reversal of this sequence and moves the communities back to wetter habitats. The lack of fire during drought or drainage allows for the invasion of upland macrophytes, scrub, and hardwood species. Alligator activity also acts to change the hydrology and nutrient status of areas and can result in pond development and maintenance (Kushlan 1974, 1987). More recent studies have demonstrated the importance of tree islands in the Everglades and revealed that they are phosphorus “hot spots” on the landscape, i.e., they act as a reservoir of P on the landscape due to the transfer of P from low-concentration surrounding areas by roosting birds and predators (Sklar and van der Valk 2002). The storage and release of high P concentrations from the tree islands have important implications for the ecological successional patterns of the Everglades that are not well understood. We do know that the southern tail ends of tree islands are often areas of higher productivity due to the release of P and that burning of tree islands also releases large amounts of P to downstream areas (C.J. Richardson, unpublished data). The successional dynamics of the Everglades is thus mainly controlled by the interaction of climatic patterns (droughts and rainfall) and human alterations on hydroperiod, which in turn influences fire frequency and the degree of fire intensity as well as the transfer and release of P on the landscape. An understanding of the long-term climatic patterns is thus important to understanding changes within the Everglades.

2.7 2.7.1

Climate and Hydrology Climate

The subtropical climate of south Florida has hot humid summers, mild winters, and a distinct wet season with 80% of the rainfall falling from mid-May to October (MacVicar and Lin 1984). The Everglades has more in common with tropical climates in that a wet/dry season is probably more important to vegetation composition than winter/summer differences in temperature. Daily temperatures average above 27°C from April to October in the northern part of the Everglades and from March to November in the south. Average daily temperatures are above 10°C even in winter, but freezing temperatures do occasionally occur. The key component of climate controlling vegetation patterns and succession is the amount of precipitation. A 110-year weighted average analysis of annual rainfall over south Florida (1895– 2005) shows distinct drought and heavy rainfall periods when compared to the

2 Ecological Status of the Everglades: Environmental and Human Factors

37

Fig. 2.13 Annual precipitation (nine-point smoothing) in the Florida Everglades based on data from 1895 to 2005. The average rainfall from 1895 to 2005 is shown with a horizontal line (data source from NOAA at http://www.ncdc.noaa.gov/onlineprod/drought/xmgr.html)

long-term average annual rainfall of 1,320 mm year−1 (Fig. 2.13). The Everglades underwent distinct periods of drought beginning in the early 1900s lasting through the mid-1920s. There was a similar drought period during the 1970s and 1980s up until about 1990. A long-term wet period began in the 1940s and lasted through the 1960s, broken only briefly in the 1950s. The highest period of rainfall, totaling 1,450 mm, was seen in 1995. Importantly, the EAA, which drains partially into WCA-2A, and the ENP, the southern most remnant of the original Glades, have received annual rainfall consistently below the historical rainfall for southern Florida. During the period 1970–1985, the EAA and ENP received less rainfall 80 and 67% of the time, respectively (SFWMD 1992). From the period 1970 to 1989, the EAA received less rainfall than average 14 out of 19 years. Extensive droughts with rainfall 250 mm below normal per year existed for six of those years, and in

38

C.J. Richardson and J.K. Huvane

1988–1989 they reached 510 mm below normal (SFWMD 1992). However, since the 1990s the Everglades have experienced dramatic increases in rainfall, with highest levels occurring in the mid-1990s (Fig. 2.13). These data indicate that the Everglades has experienced reduced rainfall for extended periods followed by significant rainfall periods that significantly altered the plant-growing environment. These rainfall patterns, when combined with effects of dikes and canal drainage, have resulted in severe drying and flooding of portions of the Everglades with a resultant shift in plant communities as noted in Sect. 2.3. Annual rainfall is the main driver of hydrology, but hurricanes (sustained winds of 120 km h−1) are an important reoccurring event (every 3 years) in south Florida. Hurricanes can produce great wind damage and significant increases in annual rainfall and storm surges (Gunderson and Loftus 1993). Thus, extreme hydrologic events like hurricanes and droughts have had significant effects on the water budgets for south Florida and the Everglades. A severe drought occurs on average every 10 years. El Niño weather patterns result in greater than average rainfall in central and south Florida, while La Niña patterns have the opposite effect (Abtew et al. 2006). These shifts in rainfall patterns have also influenced the yearly nutrient loadings and concentrations (especially P) entering the WCAs (SFWMD 2004, 2005, 2006). Evapotranspiration (ET) is also an extremely important component of the Everglades. It has been estimated that 70–100% of rainfall exits the Everglades this way (Dohrenwend 1977; Fennema et al. 1994). Evapotranspiration varies across south Florida where lakes, impoundments, and flooded wetlands evaporation losses equal potential ET (Abtew et al. 2006). Higher ET occurs in the southern part of the Everglades compared to areas north of Lake Okeechobee (Fig. 2.14a). However,

Fig. 2.14 (a) Spatial pattern of potential evapotranspiration (PET) isohyetal lines for south Floridabased work of Abtew et al. (2003). (b) Monthly average seasonal PET for 2001 in the SFWMD (data from Abtew et al. 2003)

2 Ecological Status of the Everglades: Environmental and Human Factors

39

the temporal variation in ET varies little in south Florida compared to annual rainfall variations (Figs. 2.13 and 2.14b) (Abtew et al. 2003). The combination of severe variation in rainfall along with water conservation measures has had a great impact on the Everglades landscape.

2.7.2

Hydrologic Shifts on the Landscape

In 1898, Willoughby reported in his book Across the Everglades: A Canoe Journey of Exploration that there were vast amounts of standing water, high dense sawgrass, and numerous upwelling of water from shallow pools in the bedrock. He stated, “All this moving water cannot be accounted for by the rain alone, and the water is too hard for rainwater, so that in all probability more comes from below than above” (Willoughby 1898). This, coupled with the documentation of waterfalls pouring out of the Everglades, upwellings from numerous springs at the edge of the Everglades, and freshwater bubbling up in Biscayne Bay in the early 1900s, clearly indicates an Everglades that maintained a large hydrologic freshwater head on the landscape and originally relied heavily on base flow, a much different hydrology than the one we see today. The role Lake Okeechobee played in supplying water to the Everglades was also not well understood. Historically, lake levels in excess of 20 ft. (6 m) were measured in the lake in the 1850s and as late as the early 1900s, and it was reported that when lake levels exceed 22 ft. (approximately 20.6 ft. NGVD) water would spill over the soil bank on the southern part of the lake into the Everglades (Steinman et al. 2002). Before major alterations and the building of Hoover Dike around the southern part of the lake, Ives (1856) reported in a military survey that at least eight rivers ran directly into the Everglades for 2 or 3 miles (3.2–4.8 km) and disappeared (McCally 1999). Thus, some believed that the periodic overflow of Lake Okeechobee was not the source of water that maintained sheet flow, but rather the rivers that according to early surveys continually supplied the northern Everglades. These discrepancies clearly indicate how poorly we understood the hydrologic relationship of the Kissimmee–Lake Okeechobee–Everglades complex. Moreover, the importance of surface and groundwater interactions in the Everglades was not really appreciated until the USGS report by Parker et al. (1955), who detailed studies on surface and groundwater flows and storage. Parker clearly showed for the first time the complexities of the hydrologic system that controlled the Everglades and that the extensive canal and dike system installed since the early 1900s (Chap. 7) had significantly altered water storage, surface and groundwater interactions, flow of water, and water depths throughout the Everglades. It is not possible in this chapter to cover in detail the water plans and schedules that have been implemented or proposed over the past 50 years or give a full account of the planned system changes. However, it is possible to present some concept of how highly managed the Everglades hydrology has become, how limited our control over hydrologic flow and storage becomes under extremes of low and high rainfall events (e.g., hurricanes), and how restricted our movement (i.e., newly approved P release criteria) and storage of nutrient-enriched waters have grown to be.

40

C.J. Richardson and J.K. Huvane

For those wishing more extensive reviews of hydrology and water delivery plans and schedules they are available in Davis and Ogden (1994a), Porter and Porter (2002), the CERP (1999), and updates on the modified delivery schedules at the USACE Website (http://www.saj.usace.army.mil/dp/mwdenp-c111/index.htm). Water flow in the Everglades was historically extremely slow due to the fact that, with virtually no elevation gradient (30 cm in 8 km) and high vegetational resistances to water flow, water moved sluggishly as surface water flow (0–1 cm s−1) and even slower as groundwater (Rosendahl and Rose 1981; see Chap. 7). Of ecological significance is the impact of major human-made controls over water flow, direction of flows, and hydroperiods, which shifted the system from a soligenous peatland hydrology to a highly managenous (managed) peatland hydrology as noted earlier. For example, during the past 50 years Lake Okeechobee, an integral part of the northern historic Everglades water flow system released nearly 50% of its water into the Atlantic Ocean and Gulf of Mexico due to construction of the St. Lucie and Caloosahatchee canals, with only a fraction of historic flow now going to the ENP and Florida Bay (Fennema et al. 1994). Thus, the entire annual water budget for the Everglades has been greatly altered and the management tightly controlled by the Corps of Engineers mainly for flood control, water supply, freshwater conservation, and wildlife starting with the Corps’ comprehensive plan in 1948. This plan was enacted when Congress passed a law (PL 80-858) for a massive central and southern Florida Project for flood control and other purposes (Secretary of the Army 1949). The implementation of the Corps’ 1948 plan and more recent SFWMD plans (1990, 1992, 2006) and Federal plans (CERP 1999) has resulted in a complex and often conflicting array of seasonally and annually revised (often revised during the year due to hurricane or drought conditions, etc.) water schedules for Lake Okeechobee, the WCAs and the ENP. When viewed from solely an engineering perspective, the plans are often considered to be one of the most successful landscape water management schemes in the world because it has kept millions of people dry during the wet season, supplies drinking and agriculture water even during droughts, reduces flood and hurricane impacts, and provides wetlands with water. By 2000, the central and southern Florida (C&SF) Project had over 1,000 miles (1,609 km) of canals, 720 miles (1,158 km) of levees, and approximately 200 water control structure that cover 16 counties and an area of over 18,000 square miles (6,948 km2) from Orlando to the Florida Reefs (Fig. 8.3). But this maize of structures has had severe negative effects on Everglades ecology (see Plate 1). In simple terms, the water structures and delivery plans did not protect the structural complexity or functions needed to sustain the minerogenous Everglades peatland complex. This is no surprise since they were not designed with this knowledge or purpose. However, CERP restudy plans were designed on the principle of restoring the natural Everglades ecosystem while fulfilling the human water needs (e.g., flood control and a steady supply of water) on the landscape. These multiple goals are not easy to balance considering the year-to-year weather variations and ever increasing human water demands and water quality regulations. The difficulty of managing this wetland/lake complex starts with Lake Okeechobee, originally a primary source of Everglades water. Prior to 1930, Lake

2 Ecological Status of the Everglades: Environmental and Human Factors

41

Okeechobee expanded and contracted depending on rainfall and inflows. After construction on the Hoover Dike was begun on the lake’s southern borders in 1938 and completed all around the lake by 1960, water was confined and lake levels and outflows were totally regulated under a series of guidelines and a prescribed regulation schedule. The large littoral and marsh areas that extended north, south, and west of the lake were cut off from water by the dike, thus removing a large nutrient sink for the lake’s excess nutrients. The main human-induced threats to the ecological health of Lake Okeechobee are now deemed to be excess nutrient loadings, especially P, altered hydroperiod, and invasion of exotic species (Steinman et al. 2002). Today outflow is highly regulated, as is the lake level under the current WS/ E (water supply/environmental) schedule where, for example, on May 31 (start of the wet season) lake water level above 13.5 ft. (4.05 m) would require water release until levels drop below this trigger level. The trigger for releases increases slowly from 13.5 ft. (4.05 m) on May 31 to 15.5 ft. (4.65 m) on September 30 (Steinman et al. 2002). However, recent hurricane events and droughts have greatly altered these schedules, and as recently as 2005 and 2006 there was great controversy over the increased pulsed releases of high P-laden water into the estuaries. Thus, maintaining water levels and standard release schedules are very difficult for Lake Okeechobee water managers, further complicating downstream Everglades water regimes. Further information on the WS/E schedule and regulations, the multitude of SFWMD temporary deviation release schedules, and guidelines for pulsed releases for each zone (ABCD) in the lake can be found at the SFWMD Website (http://www.sfwmd.gov/org/pld/hsm/reg_app/lok_reg/index.html). The effects of dramatic shifts in the water flow at the landscape scale can be easily appreciated by comparing flows under natural conditions and current water management plans using a Minard-type diagram of the historic surface and groundwater flows based on predictions from the Natural System Model (NSM; SFWMD 1998). In the diagram, the width of the lines shows the amount of water flowing along key points; the direction of flow is shown as well (Tufte 1983). The historical NSM model was run with no canals and dikes in place and then compared with flows measured in the mid-1990s with dikes and flow pumps and gates in operation (Fig. 2.15; Larson 1994). Under historic conditions, a balanced and similar annual volume (~1,481 × 106 m3 or 1,200 k acre-ft. year−1) of water was found leaving the Kissimmee Basin flowing into the ENP via the Tamiami Trail. Historically on average only 503 × 106 m3 (408 k acre-ft. 503 × 106 m3) of water left Lake Okeechobee annually because of high ET rates in the lake coupled with restricted flow south due to a natural soil berm, dense sawgrass, and no direct outlets to the Caloosahatchee or St. Lucie Rivers. The central glades had approximately 1.49 × 106 m3 (1,200 k acre-ft.) of water, of which approximately half or 814 × 106 m3 (660 k acre-ft.) exited the Everglades to the Lower East Coast (LEC) yearly (Fig. 2.15). The historical total discharge for the LEC to the Atlantic was estimated by the NSM to be 1,987 × 106 m3 year−1 (1,600 k acre-ft.). By 1994, the annual Everglades water budget was highly regulated, and LEC flows dramatically increased from 1,987 × 106 m3 to 4,579 × 106 m3 (1,611–3,712 k acre-ft.) as freshwater water was being transported to the Atlantic Ocean via a

42

C.J. Richardson and J.K. Huvane

Fig. 2.15 Minard-type graphic of the historic (before 1880) and modern average annual water flows (based on 1993 SFWMD LEC report data and Larson 1994). The line widths are proportional to the volume of the water flows. Values are given as 106 m3 year−1

complex series of canals and pumping stations at the expense of flows into the ENP (Fig. 2.15). Importantly, water inputs into the ENP were less than half of historic inputs, and flows shifted east to the LEC had more than doubled. The transportation of freshwater to the Atlantic Ocean was orchestrated through directives from the Corps and SFWMD to keep the urban LEC from flooding. The Kissimmee basin due to dredging, canals, and extensive agricultural and cattle land development also released 52% more water (1,233 × 106 m3 or 1,765 acre-ft.) into Lake Okeechobee in 1994 than in past centuries; in addition, this greater volume of water was at times extremely high in P concentrations (>500 µg l−1) (Walker 2000). The high concentrations of P in lake water have been a problem during the past 35 years and currently prevent the release of excess water directly to the Everglades due to the current 10 µg l−1 TP water criterion approved by the US EPA in 2005 (SFWMD 2006). Lake releases to both the St. Lucie and Caloosahatchee Rivers under the modern flow regime are ~40% of the lake inflow and, when combined with flows to the EAA, reach 56% of the lake of inputs (Fig. 2.15). Inflows in 1994 into WCA-1A, WCA-2A, and WCA-3A were 439 × 106 m3 (356 k acre-ft.), 687 × 106 m3 (557 k acre-ft.), and 1,432 × 106 m3 (1,161 k acre-ft.), respectively. Similar shifts in water budgets in and out of ENP and the WCAs were estimated from 1965 to 1995 under

2 Ecological Status of the Everglades: Environmental and Human Factors

43

both natural conditions (NSM version 4.5) and under the 1995 base managed conditions (SFWMD version 3.5) by Sklar et al. (2002). They also noted striking differences between natural and managed surface and groundwater flows. Ten years later in 2004, flows into the WCAs were surprisingly comparable to those reported a decade earlier since outflows from Lake Okeechobee were 3,229 × 106 m3 (2,618 k acre-ft.) or nearly three times higher than those reported in 1994, in part due to the effects of numerous hurricanes in late 2004 (SFWMD 2005). Importantly, outflows into the St. Lucie and Caloosahatchee Rivers increased 2.2 times and 3.2 times compared to 1994 while flows to the ENP were similar. The influence of dramatic increases of water from three major hurricanes (Charley, Frances, and Jeanne) and a remnant of hurricane Ivan that all hit south Florida in the fall of 2004 was still seen in massive alterations to 2005 water management allocations. Input and outputs to Lake Okeechobee were increased 1.5 times above recent values (4,319 × 106 m3 or 3,502 k acre-ft.), and releases were two times (3,556 × 106 m3 or 2,883 k acre-ft.) modern records (SFWMD 2006). As a result, the SFWMD had to release exceptionally high volumes of water, 2,469 × 106 m3 (2,002 k acre-ft.), to the St. Caloosahatchee estuary and higher than normal volumes, 872 × 106 m3 (707 k acre-ft.), to the Lucie estuary. As mentioned earlier, these nutrient-laden freshwater waters continue to plague the estuaries on both coasts, and recent water volume increases have exacerbated the problem. These increased inputs to the Caloosahatchee River have changed estuarine salinity, flows, and nutrient inputs, all of which can affect estuarine fishes, manatees, benthic communities, oysters, and clams as well as submerged aquatic vegetation and the remaining mangrove forests (Barnes 2005). Water inflows into WCA-1A in 2005 were 588 × 106 m3 (477 k acre-ft.), values below recent average levels. However, inflows into WCA-2A were 1,209 × 106 m3 (980 k acre-ft.), nearly double 1994 values and much higher than recent records, while flows into WCA-3A of 1,686 × 106 m3 (1,367 k acre-ft.) were similar to recent values (SFWMD 2006). These reallocations resulted in variations in water conditions in the northern vs. the southern Everglades, but surprisingly the ENP received 65% lower inputs even during this wet period (SFWMD 2006). These differences in water allocations over space and time emphasize the difficulty and challenges in maintaining a balanced flow pattern that will sustain and restore the Everglades and meet human needs.

2.7.3

Future Hydrologic Plans

To help overcome some of the major water management problems of past plans, the central and southern Florida Project Comprehensive Review Study (The Restudy) was undertaken by the Corps of Engineers under the Water Resource Development Acts of 1992 and 1996. The Corps was tasked with developing a comprehensive plan to restore and preserve the south Florida natural ecosystem while enhancing water supplies and maintaining flood protection. Restoration of the Everglades ecosystem was the key purpose of the plan, but as required by law the plan also

44

C.J. Richardson and J.K. Huvane

provided for necessary water-related needs of the region, including urban and agricultural water supply and flood protection. The Restudy Plan emphasized four key problems (1) the reduction in total area of the Everglades ecosystem by 50%; (2) the reduction of water flows to the Everglades by 70%; (3) the deterioration of water quality; and (4) the reduction and damage of natural habitats such that 68 Everglades species were endangered. The plan was to be based on scientific research to correctly time the return of the right quantity and quality of water to each habitat and develop a flexible and adaptive approach that was multiagency/ multidisciplinary in nature. Key to the plan was the removal of 400 km of dikes and levees, the construction of new filter wetlands, and the use of hundreds of underground aquifer storage and recovery (ASR) wells over a 20-year period. The benefits were that 80% of the “new or retrieved” water was to be sent to the ecosystem and 94% of predrainage flows returned to the ENP while maintaining flood control and water supply for a sustainable south Florida. The implementation plan was to achieve ecosystem restoration as soon as possible and by 2010 have more than 50% of the hydrologic restoration completed. The overall plan was to take over 30 years at an estimated cost of $7.8 billion, with costs to be shared by the State of Florida and the Federal Government. According to news releases from the Corps in 1999 and the CERP Website (http://www.evergladesplan.org/indes.aspx), sound science and peer review were integral parts of the plan’s adaptive management approach of continuously monitoring and making changes when necessary to achieve maximum benefits. The CERP plan was designed to restore more natural flow to the Everglades complex, and increase water volume to the ENP without drowning tree islands in the northern and central WCAs (Kloor 2000). Highlights of the plan, when implemented, proposed flows and allocations that would result in a 20% reduction per year of LEC losses to the Atlantic Ocean, from 3,641 × 106 m3 to 4,578 × 106 m3 (3,172–2,953 k acre-ft.) and 442 × 106 m3 (358 k acre-ft.) of new environmental water allocated to the ENP. Flows of 2,025 × 106 m3 year−1 (1,642 k acre-ft.) into Lake Okeechobee were projected to be near 1994 levels, but outflows to the Caloosahatchee were doubled from 519 × 106 m3 to 1,029 × 106 m3 year−1 (421– 834 k acre-ft.). EAA makeup water from the lake in the amount of 203 × 106 m3 year−1 (165 k acre-ft.) was also planned for additions to the WCAs. Almost immediately, the plan was under attack from environmentalists and scientists who were concerned that too little water was being allocated to the ENP, although under the plan more water is allocated than in the past (Kloor 2000). Another key concern was that moving extra water to the park would come at the expense of the central Everglades ecology. These areas would have to bear the increased flow, which in all likelihood would damage the tree island habitats (Chap. 8) and lead to a loss of key species. Leading the objections were the Miccosukee Tribe, who have over 100,000 ha of holdings in the central Everglades and view the tree islands as key to their hunting and ceremonies. The Miccosukee also worried that the extra water would be laden with excess nutrients (Kloor 2000). A case in point is the EAA makeup water, which is currently too high in nutrients to meet the current standards. Finally, the concept of utilizing several hundred ASR wells

2 Ecological Status of the Everglades: Environmental and Human Factors

45

storage was highly criticized as an untested method of water storage and retrieval under the geologic and hydrologic conditions of south Florida. The first step in the CERP process was the Water Resources Development Act of 2000, which was signed into law by President Clinton on 11 December 2000. Ten identified restoration projects with a combined budget of $1,400,000,000 were approved. They ranged from construction of 20,200 ha (50,000 acres) of agricultural area water storage reservoirs to levee seepage management projects in WCA3A and WCA-3B. However, of key importance was the focus on STA projects around C-9, C-11 impoundment, and Taylor Creek/Nubbin Slough; the raising of the bridge on the eastern portion of the Tamiami Trail; and the filling of the Miami Canal. The C-111 N spreader canal project was to be implemented to improve water distribution and connectivity and sheet flow in the southern Everglades. One hundred million dollars was approved for adaptive assessment and monitoring. Most of these projects were slated to start construction between 2004 and 2005. With an emphasis on delivering more water to the ENP to more closely mimic historic conditions, the USACE devised a modified water delivery (MWD) plan for the Everglades National Park and south Dade Canals (C-111). The plans are outlined in detail on the USACE Website (http://www.saj.usace.army.mil/dp/mwdenp-c111/ index.htm) along with development schedules for new projects to the existing central and southern Florida (C&SF) Project. These projects were required to enable water deliveries for the restoration of more natural hydrologic conditions in ENP. These improvements are to enable the reestablishment of the historic Shark River Slough flow-way from WCA-3A through WCA-3B to ENP. However, of major concern in the delivery of water to the ENP is the loss of endangered species habitat for the Cape Sable seaside sparrow (Ammodramus maritimus mirabilis), Everglades snail kite (Rostrhamus sociabilis plumbeus) and wood stork (Mycteria americana). In December 2006, the USACE released its Final Supplemental Environmental Impact Statement (FSEIS) for the Interim Operation Plan (IOP) for the protection of the Cape Sable seaside sparrow (CSSS) and other species, with a recommendation that alternative plan 7R was the best water release schedule according to model predictions of lower impacts for the CSSS, snail kite, wood stork, and their critical habitats (USACE 2006). This alternative also allows for emergency operations under high rainfall conditions. Despite all the management guidelines and the recent alterations to the system, questions still remain on whether water delivery schedules will be adequate to maintain ecosystem integrity of the ENP and maintain populations of the endangered species. However, an assessment of the water allocated through the 12 structures and S333 to ENP provides us with insight into how water deliveries to the ENP have changed over time, how closely mandated regulation schedules have been followed, the effects of extreme climatic events on delivery schedules, and whether water deliveries can meet restoration needs on an annual basis (Fig. 2.16). Clearly, water management deliveries to northern parts of the ENP have changed dramatically over the past 30 years. In 1970, the U.S. Congress passed a law (PL 91–282) that mandated flow to the park was to be 388 × 106 m3 year−1 (315,000 acre-ft.), with 320 × 106 m3 (260,000 acre-ft.) allocated for Shark River Slough (Light and Dineen

46

C.J. Richardson and J.K. Huvane

Fig. 2.16 Annual flows of water into the ENP into the northern ENP south of Tamiami Trail and east and west of the L-67 extension via S333 (net to ENP), and the sum of the S12ABCD structures (data from the SFWMD 2006; T.K. MacVicar, personal communication)

1994). As noted earlier, the ENP originally received as much water as the LEC, more than 1,400 × 106 m3 year−1 (1.1 million acre-ft.) according to the NSM (Fig. 2.15). These volumes are supported by recent simulation runs (NSM46) by the USACE indicating that 1,693 × 106 m3 (1,373 k acre-ft.) of water flowed into the northern ENP south of Tamiami Trail and east and west of the L-67 extension (USACE, unpublished January 2006 model run using model SFWMM version 5.6). The first striking feature of the 27-year record is that flows varied greatly from year to year; moreover, flows only exceeded 1 million acre-ft. year−1 (1,233 × 106 m3) once from 1978 until 1992 (Fig. 2.16). Flows to the ENP were the lowest in 1989 (0.8 × 106 m3 or 69 k acre-ft.) due to the extensive drought that year. Ironically, this was my first year (senior author) of sampling in the Everglades, and virtually no water passed through the 12 structures. Surface water sampling was impossible, and extensive fires burned throughout the ENP. The highest flows were in 1995 and reached 2,837 × 106 m3 (2.3 million acre-ft.) during a year of exceptionally high rainfall (Fig. 2.13). Thus, the ENP has been kept exceptionally dry during some periods due to the lack of water and then drowned in wet years even though a mandated water delivery schedule was in place. Since 2002, the ENP has been receiving water under the alternative 7R schedule of the MWD plan, under which deliveries should average closer to 979 × 106 m3 year−1 (794 k acre-ft.) according to the SFWMM model (USACE, unpublished January 2006 modeling). Flows in 2003 were near 1 million acre-ft. (810 × 106 m3) and in 2005 reached 1,566 × 106 m3 (1.27 million acre-ft.). In 2006, an exceedingly dry rainfall year, the park received only 573 × 106 m3 (465 k acre-ft.). Thus, while the shifts in water delivery have been less dramatic under the MWD schedule now in place than earlier delivery schedules, year-to-year variations in rainfall still highly influence release volumes due to a

2 Ecological Status of the Everglades: Environmental and Human Factors

47

lack of upstream water storage reservoirs. Currently, water continues to be pumped to the ocean and estuaries, and this pumping will continue until water reservoirs are constructed, a plan now under way with the State of Florida’s Accelerator-8 funding. Unfortunately, it still appears that lower amounts of water will be delivered to the ENP in drought years due to human and agricultural allocations. These shifts in allocations are of major concern and more recently the USACE has developed under its adaptive management plan a new series of alternative model runs to more closely mimic historic flows. It has for example put forward alternative plan 5 with annual flows in the northern ENP of 1,202 × 106 m3 (975 k acre-ft.) (http://www.saj. usace.army.mil/dp/mwdenp-c111/index.htm). Thus, the plans are in place to modify and update the MWDs to sustain the ENP, but whether the water is available each year to meet these guidelines is still controlled to a large degree by climatic conditions and upstream human demands on water allocations. The correct timing and volumes of future water delivery schedules are not the only aspects of water delivery that need to be restored to maintain the original minerogenous portions of the Everglades peatland complex (limnogenous/soligenous/topogenous zones). Unfortunately, peatland hydrodynamics were not taken into account in the management plans; thus, the normal successional patterns and development of the Everglades fen will forever be altered. In the future, the Everglades will be maintained mostly as a managed or managenous peatland system. With only 50% of the original Everglades remaining and hundreds of control structures in place some say this is the only choice available. Peatland ecologists would argue that we have the opportunity with adaptive management to test alternative peatland restoration techniques and restore key components of the former Everglades. For example, restoration experiments on tree island habitats (SFWMD 2006) are showing some success, but larger scale work is needed on alternate flow regimes and delivery system effects on the peatlands themselves. Hopefully, the ombrogenous interior portion of WCA1A will be maintained by not allowing surface water flows into the region so that the normal succession stage of marsh–fen–bog development can continue. Finally, a major continuing problem for water managers in the future will be trying to balance conditions to maintain soligenous peatlands conditions for the central Everglades habitats while being pressured continuously to alter hydrologic levels and flows for survival of endangered species at specific locations or for human water needs. By not maintaining the variety of specific hydrologic, nutrient, and fire conditions that shaped the diversity of Everglades habitats, endangered species arguments will continue to mount and ecosystem management will become more and more manipulated to the detriment of the natural community structure and diversity.

2.8

Nutrients: Rainfall and Anthropogenic Inputs

The effects of nutrients additions on plant and animal communities, the development of nutrient gradients, and nutrient storage in plants and soils are covered in detail in the following chapters so no attempt is made to cover these topics in this chapter. Here we provide more of a landscape picture of nutrient additions, primarily P because

48

C.J. Richardson and J.K. Huvane

it is the limiting nutrient. We focus on the two main sources of nutrients, rainfall contributions, and runoff from the surrounding land uses. Unfortunately, little information had been collected on subsurface additions or upflux of nutrients, although we provide nutrient profiles and information on soil nutrient storage and releases in Chap. 6, and water column profiles of nutrients in the Everglades for some ions have been analyzed in the interior of WCA-2A (Vaithiyanathan and Richardson 1997a). Historically, the Everglades received nutrients primarily from rainfall, surface flow, and recycling within the system, especially after fire (Davis 1943; Swift and Nicholas 1987; see Chaps. 6 and 9). It is a P-limited system that has evolved plant species that can survive under TP water concentrations as low as 5 µg l−1 (Koch and Reddy 1992; Richardson et al. 1995; Richardson and Vaithiyanathan 1995; Richardson et al. 1999).

2.8.1

Nutrients in Rainfall and Runoff

There are no reliable long-term historical records of the amount of nutrients in rainfall in south Florida. An analysis of rainfall nutrient loadings over the past few decades suggests that dry fall represents a major source of N and P into the Everglades, but numbers vary greatly in time and space (SFWMD 1992; Peters and Reese 1995). The flow-weighted mean for various stations in south Florida in the 1980s ranged from 20 to 220 µg l−1 for total P with loads of 31–66 mg m2 year−1 (Hendry et al. 1981), but many consider these numbers too high due to collector contamination. There is concern when comparing data from different sources due to differences in sampling. For instance, you may be comparing data where a different type of sampler was used, data where either wet deposition (rainfall events) or dry fall (settling of particles) only was collected, or data where wet and dry fall were combined or collected in bulk samplers. These sampling variations can affect results and conclusions and make it difficult to pool data from different studies. Cloud deposition, although considered important, is difficult to measure and is often not included. Another factor to recognize is the source of nutrients and whether the measuring systems are close to the ground or near objects that result in recycling of local materials. Because of these difficulties it has been suggested that a P deposition and nutrient monitoring program should focus only on wet deposition values since the chemistry is more certain than dry deposition or bulk samples (Redfield and Urban 1997). Given all these problems, it is not surprising that there is a lack of good data. However, virtually everyone agrees that a better estimate of rainfall P and nutrient input is needed to accurately estimate the nutrient budget for the Everglade. This is especially important for determining nutrient loading to undisturbed or ombrotrophic areas of the Everglades. In 1997, the Director of the SFWMD, in a conference entitled “Atmospheric Deposition into South Florida,” declared that the atmospheric deposition of nutrients must be considered in all future actions to restore the Everglades (Redfield and Urban 1997). Concerns raised at the conference included problems with methods

2 Ecological Status of the Everglades: Environmental and Human Factors

49

of collection, sample contamination, best estimates of mass loads of P to the Everglades, the need for establishment of monitoring programs, and how to separate local sources from long distance sources. Few of these problems have been addressed a decade later but some findings help place the importance of rainfall nutrients into perspective with respect to restoration. For example, simulations found significant effects of varying concentrations of rainfall P on water column P concentrations, especially in remote areas (Chen and Fontaine 1997). They reported that when 10 µg l−1 of P was in the rainfall the average water column P concentrations ranged from 2 µg l−1 in WCA-1A to 57 µg l−1 in WCA-2B. By comparison, when rainfall was estimated to be 100 µg l−1 WCA-1A and WCA-2B had 18 and 64 µg l−1 of P in the water column, respectively. Of concern was the ENP, where water column P increased from 6 to 26 µg l−1 under the two simulation scenarios. Raising further alarm was that fact that these concentrations, when used to calculate the loading rates, indicated that a rainfall concentration of 10 µg l−1 of P resulted in 883 metric tons of P in the EPA (the Everglades Protection Area includes all WCAs and the ENP and covers of 9,151 km2), around half of the 1,682 metric tons of P from surface loading via structures. When rainfall P concentrations were assumed to be 30 µg l−1 the ratio of P load from rainfall to P load from structures increased by 1.6, meaning that rainfall contributed 60% more P into the EPA than structural inputs and runoff into the Everglades (Chen and Fontaine 1997). Although the concentrations used in the simulations cover too wide a range according to more recent data, they clearly show that rainfall contributes a significant part of the P budget of the Everglades (Redfield and Urban 1997). To provide a better estimate of P rainfall contributions to the Everglades and overcome the concerns regarding elevated P concentrations in samplers due to contamination from bird droppings, insects, animal parts, and dust, a statistical screening technique was devised by the SFWMD whereby outliers were removed to obtain a better estimate of wet deposition from their samplers by using a Kalman filtering algorithm to provide a minimum error of variance estimation (Ahn 1997). Results from Aerochem wet/dry rainfall collectors at 15 sites run by the SFWMD since 1992 provided a mean P concentration of 10.9 µg l−1 with a standard deviation of 13.4 µg l−1. Highest values ranged from 21.7 µg l−1 at S-127 to 5.6 µg l−1 at L-67A stations (Ahn 1997). When combined with the SFWMD District’s average rainfall of 1,346 mm of rainfall the estimated annual P load was 14.8 mg m2 year−1. These values are close to the concentration and load values (15.4 µg l−1 and 14.8 mg m2 year−1) used by Walker (2000) in his modeling of rainfall P inputs into Lake Okeechobee. More recent models for determining P dynamics in the STAs have used P concentration and load values of 10 µg l−1 and 20 mg m2 year−1, respectively (Walker and Kadlec 2006). These P loading rates are close to average reported bulk sampler values of 23.8 mg m2 year−1 for 17 forested ecosystems in the U.S. and Europe (Likens and Bormann 1985) and 15 lake site values of 27.8 mg m2 year−1 in Michigan (Eisenreich et al. 1977) but much lower than wet/dry bucket values reported earlier for Florida by Hendry et al. (1981). A more recent detailed study in Japan, however, suggests that P inputs (wet and dry) dropped by nearly half from 25.3 to 13.1 mg m2 year−1 when local contributions were excluded (Tsukuda et al. 2006). In calculating

50

C.J. Richardson and J.K. Huvane

the total P loadings for the Everglades EPA in 2004 the SFWMD used P loading values that ranged from 20 mg m2 year−1 in the ENP to 35 mg m2 year−1 values in WCA-1A and WCA-2A (SFWMD 2005). These range of values resulted in an annual atmospheric P loading of 193 MT, or 63.5% of the total loading to the EPA. To obtain a more accurate picture of the major nutrients and ion rainfall inputs into the Everglades at our P dosing sites (Fig. 1.2 and Chaps. 14 and 15), we monitored nutrient input in Aerochem wet/dry rainfall collectors from September 1993 until September 1997 in two oligotrophic slough areas in WCA-2A (Richardson and Vaithiyanathan 1997a). Sample buckets were sampled and acid washed weekly, and wire bird guards and gauze were used to reduce contamination from sources besides atmospheric deposition (e.g., bird droppings, insects). Samples were carefully screened in the field for any signs of contamination, and any contaminated samples were discarded. In addition, a statistical screening for outliers (>2 standard deviations) was used to remove additional samples before analysis. The geometric mean total P concentration in rainfall from 1993 to 1997 (n = 149 sampling dates) was 10.5 µg l−1, almost identical values to those reported by Ahn (1997) and used in recent modeling efforts. Orthophosphate values averaged 4.6 µg l−1 during 1993–1995 and 1997. Our results for P deposition showed strong seasonal and interannual variations, which reflect inputs from local fires, large-scale burning of sugarcane fields to the north, as well as seasonal and annual differences in rainfall patterns. For example, for some unknown reason (i.e., lab and field blanks were all within lab standards) P in rainfall was significantly higher in 1996 than all the other years we measured. Our annual wet rainfall geometric mean P deposition from 1993 until 1997 was 23.6 mg m2 year−1, which was similar to values measured at various locations in south Florida and around the globe, which suggests that our value is a realistic estimate of Everglades P loadings (Table 2.1). Another way to determine if these P loading values are realistic was to compare measured atmospheric P deposition flux with rates of P accumulation in peat estimated from 137Cs (~ last 30-year period of accumulation) and 210Pb (~ last 125-year period of accumulation) dating of soil cores (Craft and Richardson 1993a; see Chap. 3) taken in ombrotrophic areas of the Everglades (Table 2.1). The center of the Loxahatchee refuge is a topographically high area, and it only receives nutrients from rainfall (ombrotrophic); thus, P accumulation rates should closely follow precipitation inputs. Our rainfall P loadings of 23.6 mg m2 year−1 are close to the long-term P storage measured in the center of the ombrotrophic Loxahatchee Wildlife Refuge (WCA-1A) during the past 100 years (30 mg m2 year−1) but are double the 10 mg m2 year−1 accumulation rates found in the past 30 years. However, the more recent estimate of P accretion rates with the 137 Cs technique is more susceptible to error due to the difficulty in measuring such low peat accretion rates accurately. In general, both estimates clearly show a low atmospheric loading rate for P, although dry deposition was not included. The ENP, an area receiving the lowest levels of P runoff input due to P filtration by wetlands to the north and much reduced flow of water into the Everglades during the past few decades, displays three times the accumulation rate (30 mg m2 year−1) of P in the past 30 years than the Loxahatchee as a result of both input sources (runoff and rainfall). Interestingly, accumulation rates over the past 100-plus years are three times higher in the ENP than more recent estimates of inputs. This probably

2 Ecological Status of the Everglades: Environmental and Human Factors

51

Table 2.1 A comparison of annual atmospheric P flux density deposition rates for short-term and long-term P storage in Everglades soils based on 137Cs and 210Pb dating techniques, respectively (see Chap. 3) Location

Reference

WCA-2A Entire Everglades

U.S. forests Czech Republic Location

This study (1993–1997) Ranges used in models for SFWMD (see citations in text) Likens and Borman (1985) Kopacek et al. (1997) Reference

WCA-2A (unenriched)b,c WCA-3A (unenriched)b,c WCA-1 (ombrotrophic)b,c ENPb,c WCA-2A (enriched)

This study This study This study This study This study

P deposition rates (mg m2 year−1) 23.6 14.8–35

24 15–24a P accumulation rates (mg m2 year−1) b 60 –80c 60b–80c 10b–30c 30b–90c 460b

a

Two locations for 16 years Cs-137 (Craft and Richardson 1998) c Pb-210 (see Chap. 3) b

reflects the much larger volume of water with low nutrient concentrations historically entering the ENP in the last century (Fig. 2.15) compared to current conditions; even given the low P concentrations, higher water volume would result in higher mass P loadings. The so-called unenriched areas of WCA-2A and WCA-3A show a doubling of P inputs compared to the ENP during the past 30 years due to increased P in runoff from agriculture (Table 2.1). Values for these areas 100 years ago are closer to the ENP accumulation rates, which suggest a more uniform rate of P input in the past across the Everglades and a lack of agriculture input to the north. By contrast the eutrophic areas in the WCAs are clearly dominated by runoff inflow where areas just south of input structures often have over 4,000 mg m2 year−1 of P coming through the gates as in WCA-2A in the north (Richardson and Qian 1999) and accumulate on average 460 mg m2 year−1 of P, thus potentially releasing over 3,500 mg m2 year−1 of P downstream (Table 2.1). Here rainfall contributes less than 1% of the P load to eutrophic areas. Thus, a massive reduction of nutrient mass loadings in runoff into the Everglades is needed to recreate historic conditions and reduce impacts to plant and animal communities.

2.8.2

Nutrients in Runoff

Agricultural runoff from the EAA and Lake Okeechobee both contribute water with significantly higher concentrations of N and P than is typically found in rainfall and in the Everglades (Craft and Richardson 1993b; Davis and Ogden 1994a; Walker

52

C.J. Richardson and J.K. Huvane

2000). A landscape analysis of the P gradient for south Florida showed that the dairy and cattle regions northeast of Lake Okeechobee had by far the highest total P input concentrations between 1973 and 1999, and the P load averaged 498 MT per year. The lake P concentration increased from ~40 µg l−1 P in 1973 to ~100 µg l−1 P by 1999 (Walker 2000). The average P concentration in water leaving the EAA farmland in the early 1990s was 150 µg l−1 P and was reduced to 115 µg l−1 P in the canals and edges of WCA-1 (SFWMD 1992). Water flowing out of WCA-2A into WCA-3A often contained 40 µg l−1 P. By the time surface waters reached the structures above the ENP, concentrations were 10 µg l−1 P. A trend analysis of inflow, interior, and outflow P concentrations over the past 27 years reveals interesting patterns of higher P inputs into the northern WCAs and much lower P inputs into the ENP (Fig. 2.17). Over 70 µg l−1 P flowed into WCA-1

Fig. 2.17 A comparison of phosphorus input, interior, and outflow concentrations from 1978 to 2005 for WCA-1, WCA-2A, WCA-3A, and the ENP. All values are the annual geometric mean of total phosphorus (data from the SFWMD 2003, 2004, 2005, 2006)

2 Ecological Status of the Everglades: Environmental and Human Factors

53

on average from 1978 until 2000, with some inputs reaching maximum concentrations over 1,400 µg l−1 P (SFWMD 2006). From 2001 until 2004, inflow P concentrations decreased significantly in both WCA-1 and WCA-2A. In 2005, input concentrations increased in WCA-1 nearly back to values prior to any best management practice (BMP) implementations. The SFWMD attributes this rise to excess rainfall and runoff due to high hurricane activity in late 2004 and 2005 (SFWMD 2006). Inputs of P into WCA-3A from 1978 until 2000 averaged 36 µg l−1 P and had dropped to 24 µg l−1 P by 2005. Low-level mean P inputs (8.9 µg l−1 P) into the ENP were recorded over the period 1978–2000, and values rose very slightly in 2005 to 9.1 µg l−1 P. Interior P values were lowest in WCA-3A and ENP throughout the study period, averaging around 10 and 5 µg l−1 P, respectively. The SFWMD (2006) reported that interior P concentrations rose during low rainfall periods and droughts and were diluted during the wet season in all areas. WCA-1 had interior values that averaged 10 µg l−1 P from 1978 until 2000 and then rose in 2005 to 12.1 µg l−1 P. However, interior WCA-2A values were always the highest, averaging 17.1 µg l−1 P from 1978 to 2000 then rising slightly to 17.9 µg l−1 P in 2005. The rises in 2005 P interior values was attributed to periods of excessive rainfall and periods of soil P release during drought conditions (SFWMD 2006). However, the fact that WCA-2A interior site P concentrations have virtually remained the same even though inputs have been reduced by >50% may be due to the reflux of the large residual amount of resident P in soil at this site (Reddy and Rao 1983; see Chaps. 3 and 6). According to the SFWMD (2006) report, approximately 85% of the P samples collected in 2005 in the entire EPA (all sites and areas) had values below 50 µg l−1 P, 51% were below 15 µg l−1 P, and 32% were at or below 10 µg l−1 P. These data indicate that the interior sites in the ENP and WCA-3A meet the 5-year geometric mean criterion of less than or equal to 10 µg l−1 P across all sites in three of five years, are annually less than or equal to 11 µg l−1 P across all stations, and are less than or equal to 15 µg l−1 P annually at all individual stations (SFWMD 2006). However, current inflow P concentrations into WCA-1, WCA-2A, and WCA-3A are still far in excess of the approved P criterion, and interior values of WCA-2A have not changed (Fig. 2.17). More troubling are the high concentrations of P that are still flowing out of the WCAs and toward the ENP. The northern WCAs release far higher P concentrations than WCA-3A, but all values are well above the USEPAapproved P criterion even though farm BMPs have been in place for over a decade and STAs are now in operation. Of importance but often not addressed is the amount of SRP or orthophosphate (i.e., considered more readily available for uptake by organism) vs. TP in input, interior, and outflow areas. From 1978 to 2003, SRP comprised 34, 31, 26, and 29% of TP input concentrations into WCA-1, WCA-2A, WCA-3A, and ENP, respectively (SFWMD 2006). Inflow SRP concentrations compared to outflow concentrations showed considerable reductions in values for some of the WCAs (23–17 µg l−1 SRP in WCA-1, 18–5.2 µg l−1 SRP in WCA-2A, and 9–2.8 µg l−1 SRP in WCA-3A) but showed little change in the SRP to TP outflow ratios which remained at 31, 25, and 25%, for WCA-1, WCA-2A, and WCA-3A, respectively. Interior concentrations of SRP were quite low and averaged 1.6 µg l−1 SRP in WCA-1, 3.7 µg l−1 SRP in

54

C.J. Richardson and J.K. Huvane

WCA-2A, 1.7 µg l−1 SRP in WCA-3A, and 2.7 µg l−1 SRP in ENP over the 1978– 2003 period. Nitrogen, another important nutrient in the Everglades, often receives less interest since P is the limiting nutrient. Everglades soils often have more than 3% N by weight, and water values of NH4–N are very high (Chaps. 3 and 6). However, a brief assessment of the total nitrogen (assessed as Kjeldahl N) shows a similar north-tosouth gradient as P with higher input concentrations in the north and values decreasing in the south. Average annual long-term input concentrations (1978–2003) ranged from 3.4 mg l−1 going into WCA-1 to 1.2 mg l−1 flowing into the ENP (SFWMD 2006). Interior concentrations also followed a north-to-south gradient with longterm values averaging 1.6, 2.4, 1.5, and 1.3 mg l−1 in WCA-1, WCA-2A, WCA-3A, and ENP, respectively. Of more concern than the N concentrations alone are the shifts in N:P ratios in the various components of the Everglades because high ratios of this key index indicate more severe P limitations in the Everglades (Richardson et al. 1999). For example, inflow N:P ratios to WCA-1, WCA-2A, WCA-3A, and ENP averaged over the 1978–2003 period were 51, 52, 59, and 141, respectively, and were quite different from 2005 inflow ratios values of 32, 92, 70, and 108 for each respective area. These data suggest that shifts in the N:P ratio are taking place due to the upstream BMPs on farmland as well as the use of the STAs to remove P. Long-term, these shifts should improve and maintain P-limiting conditions in the interior of the fens. Interior N:P ratios for WCA-1, WCA-2A, WCA-3A, and ENP averaged over 1978–2003 were 119, 130, 160, and 257, respectively (SFWMD 2006). These general trends remained the same in 2005, indicating that P was extremely limiting in all interior locations, but especially in ENP. Phosphorus loadings from the EAA have been implicated in the replacement of sawgrass by cattail in WCA-1 (Loxahatchee National Wildlife Refuge) and WCA2A in the early 1980s (Toth 1987, 1988; Belanger et al. 1989; Urban et al. 1993). Belanger et al. (1989) asserted that additions of nutrient-enriched water to WCA2A have contributed to the invasion of a monotypic cattail community. The high P levels in vegetation, soils, and surface waters of the cattail-dominated areas of WCA-2A suggest that P may be primarily responsible for the invasion of cattails in WCA-2A (Belanger et al. 1989; Richardson and Craft 1993; Vaithiyanathan et al. 1995, 1997; Craft and Richardson 1997; see Chap. 9). Thus, the control of cattail expansion and community shifts will require an understanding of the effectiveness of the Everglades BMP regulatory P reduction program since its implementation in 1996 (Fig. 2.18). Over 1,600 MT of P have been prevented from entering the EPA since 1996 and the predicted P load has been reduced from the baseline period (1978–1988, prior to any treatments or BMPs) average annual loading of 444 MT and P concentrations 173 µg l−1 P to 182 MT and 124 µg l−1 P by 2005 (SFWMD 2006). Of note is the sharp reduction in mass P loadings since full implementation of the BMPs in 1996, except for 2000 and 2005. These increases have been attributed to rainfall and weather events by the SFWMD but they were far below the predicted loads of over 400 MT in those years. Thus the BMP program has had a dramatic effect on reductions of P going into the EPA, but loads are still variable from year to year and will be greatly influenced by the P removal effectiveness of the STAs.

2 Ecological Status of the Everglades: Environmental and Human Factors

55

Fig. 2.18 Total phosphorus loads from the Everglades Agricultural Area (EAA) to the Everglades Protection Area (EPA) from 1980 until 2005. Best management practices (BMPs) are denoted as well as the first compliance year (data from SFWMD 2006)

The major hope for reducing P loads into the EPA in the future was the use of STAs to treat EAA, upstream and Lake Okeechobee waters prior to their release. The design criteria for these STAs were carefully planned and modeled to determine the correct sizing of the STAs by scientists at the SFWMD with assistance of engineering experts Robert Kadlec and William Walker (Walker 1995; Walker and Kadlec 2006). To date six STAs covering over 41,000 acres (16,564 ha) have been built, the earliest in operation since 1994–1995. They were designed initially around a 1.3–1.5 g m2 year−1 loading rate (Table 2.2). Since their inception they have been estimated to reduced P loadings by 617 MT into the EPA, and in 2005 all the STAs combined removed 189 MT of P, or 71% of the 268 MT of loadings (SFWMD 2006). The flow-weighted mean inflow ranged from 247 µg l−1 P in STA-1W to 78 µg l−1 P in STA-6 in 2005. Overall mean inflow P averaged 147 µg l−1 P and outflow concentration 41 µg l−1 P in 2005. Unfortunately, the actual loading rates to some of the STAs during operation have greatly exceeded their design loading criteria by more than a factor of 2, and outflow P concentrations have also greatly exceeded desired levels (Table 2.2). In fact, in 2005 and 2006 four of the STAs exceeded 50 µg l−1 P, and several had output values over 100 µg l−1 P. Importantly, only three of the STAs (STA-2, STA-34, and STA-6) were loaded with P near their design criteria, and only these sites came close to outflow concentration of 20 µg l−1 P over the entire period. Moreover, two STAs (STA-2 and STA-6) doubled their concentration output in 2005, which was again attributed to high rainfall and runoff inputs (Table 2.2). In terms of P reductions, both the BMPs and the STAs have resulted in a significant decrease of P to the EPA. However, EAA outflow P concentrations continue to

56

C.J. Richardson and J.K. Huvane

Table 2.2 Summary of Stormwater Treatment Areas (STAs), inflow loads, and outflow P concentrations in 2006 (Walker and Kadlec 2006) Inflow P load (g m2 year−1)

Outflow P (µg l−1)

Operating period

STA

Design ECP

Design LTP

Operation period

Last 12 months

Operation period

Last 12 months

First

Last

STA-1W STA-1E STA-2 STA-34 STA-5 STA-6 All

1.4 1.4 1.3 1.3 1.5 1.3 1.3

1.0 1.1 1.1 1.1 1.1 1.1 1.1

2.7 0.6 1.3 1.2 2.7 1.4 1.5

1.7 0.8 1.8 1.5 2.5 1.6 1.6

66 174 19 19 99 21 43

125 122 28 24 100 26 48

July-01 May-04 July-01 Dec-03 May-01 Aug-00 Aug-00

July-06 July-06 July-06 May-06 June-06 June-06 June-06

Design loading criteria are shown for each the STAs between 1.0 and 1.5 g m2 year−1, but they have been exceeded in five of the six treatment areas in the last 12 months and on average for all sites

remain high, and STA reductions have not consistently reached the low concentrations 10–15 µg l−1 P that had been hoped for by many scientists. While P mass loadings are significantly reduced by more than 50% for the EAA, even in wet years like 2005, P concentrations remain too high for major improvements in the receiving waters as noted in Fig. 2.17. If the present trend continues and no additional STAs are built, the Everglades will continue to receive unacceptable concentrations and loads of P for the foreseeable future. This will have significant consequences for the native biota and ecosystem structure and function as explored in following chapters.

2.9

Conclusions and Lessons for Restoration

One of the gaps in our knowledge concerns how the diversity of communities was formed in the Everglades landscape complex. Paleoecological studies indicate a northto-south trend in peat accumulation over the past 5,000 years, with considerable periods of longer and shorter hydroperiods long before drainage canals altered hydrologic flow and water levels. Numerous changes in vegetation have occurred over the past few thousand years, but recent alterations have resulted in an invasion of upland and exotic species as well as massive increase in cattail due to increased P mass loadings. The most dramatic changes have taken place in the twentieth century, which has resulted in loss of slough areas and tree islands, as well as extensive cattail invasions in the northern Everglades. A loss of acidophilic diatoms in recent times in WCA-2A indicate that this area was probably more acidic and had fewer calcareous periphyton mats than are present today. The Everglades should be classified as a fen or, in more generic terms, a peatland – not a marsh or swamp. The hydrogeologic factors controlling the formation of fens are totally different from that of marshes and swamps, and this has important hydrologic consequences for restoration of the Everglades. Restoring landscape

2 Ecological Status of the Everglades: Environmental and Human Factors

57

hydrologic equivalence will be essential to successful restoration of the modern Everglades if we ever hope to maintain the diversity of Everglades habitats and communities. For example, care must be taken to maintain the ombrogenous portions of the Everglades like WCA-1, reestablish limnogenous peatlands south of Lake Okeechobee, recreate conditions for soligenous peatlands where topographically possible, and reduce managenous water flow conditions (water pumping and release across narrow outlets) within portions of the Everglades. Succession in the Everglades is mostly controlled by hydrologic conditions and, in turn, fire frequency and intensity. The impact of excessive nutrients, especially P, is critical in some areas of the Everglades as is the invasion of exotic species. Historic hydrologic flow patterns and volumes of water have been greatly altered as evidenced by the reductions in flow to the ENP and increased flow of freshwater to the oceans and Gulf of Mexico. We have only recently begun to understand the hydrologic relationship of the Kissimmee–Lake Okeechobee–Everglades complex. Moreover, the importance of surface and groundwater interactions in the Everglades is still not fully understood, and the influence of canals and pumping stations on community responses are for the most part unknown. The difficulty of managing the wetland/lake complex starts with the myriad of Lake Okeechobee regulation schedules and farm-use plans for water, which has a ripple effect on all downstream water conditions in the Everglades. The CERP plan was designed to restore more natural flow to the Everglades complex and increase water volume to the ENP without drowning tree islands in the northern and central WCAs, but difficulties abound in meeting regulated flow conditions when droughts or extreme wet seasons occur. The shifts in water delivery have been less dramatic under the modified water delivery schedule now in place than with earlier delivery schedules but year-to-year variations in rainfall still highly influence release volumes due to a lack of upstream water storage reservoirs. Currently, water continues to be pumped to the ocean and estuaries and will continue until water reservoirs are constructed. The implementation of the CERP called for 50% of the hydrologic restorations to be completed by 2010; however, that deadline will not be met. Moreover, it still appears that lower amounts of water will be delivered to ENP in drought years due to human and agricultural allocations. Unfortunately, the correct timing and volumes of future water delivery schedules as well as the mode of delivery needed to restore the original minerogenous Everglades fen and peatland complex (limnogenous/soligenous/topogenous zones) have been ignored, and thus the normal successional patterns and development of the Everglades peatland will forever be altered. In the future the Everglades will be maintained mostly as a managed or managenous peatland system. Phosphorus in rainfall is a major contributor to the overall P budget for the Everglades but plays a lesser role in areas where there is runoff from agriculture. However, the accuracy of P measurements in rainfall is not adequate at the present time, thus estimates of total P loadings to the Everglades are in question. Research to improve estimates is badly needed. Data on nutrient concentrations and mass loadings for N and P are readily available and show good long-term trends of decreased inputs from agriculture and adjacent

58

C.J. Richardson and J.K. Huvane

areas due to farmland BMP reductions and STA uptake and storage. Of importance but often not addressed is the amount of SRP or orthophosphate vs. total P in input, interior, and outflow areas. Another area that needs to be monitored carefully is the shifts in N:P ratios in the various components of the Everglades since high ratios of this key index indicate P limitation, a condition representative of the historic Everglades. Recent data suggest that shifts in the N:P ratio are taking place due to the upstream BMPs on farmland as well as the use of the STAs to remove P. Longterm, these shifts should improve and maintain P-limiting conditions in the interior of the fens. According to the SFWMD (2006) report, the entire EPA in 2005 had approximately 85% of the P samples collected with values below 50 µg l−1 P, 51% below 15 µg l−1 P, and 32% at or below 10 µg l−1 P. These data indicate that the interior sites in the ENP and WCA-3A met the USEPA P criterion; however, current inflow P concentrations into WCA-1A, WCA-2A, and WCA-3A are still far in excess of the approved P criterion, and interior values of WCA-2A have not changed over the past 27 years. In terms of P reductions both the BMPs and the STAs have resulted in a significant reduction of P to the EPA. However, EAA outflow P concentrations continue to remain high, and STA reductions have not consistently reached the low concentrations 10–15 µg l−1 P that are needed to sustain the Everglades. While P mass loadings are currently significantly reduced by more than 50% compared to the 1970s and 1980s for the EAA, even in wet years P concentrations remain too high for major improvements in the receiving waters. If the present trend continues and no additional STAs are built to reduce current P overloading, then the Everglades will continue to receive unacceptable concentrations and loads of P for the foreseeable future. This will have significant consequences for the native biota and ecosystem structure and function as noted in a number of the following chapters.

3

Soil Characteristics of the Everglades Peatland Christopher B. Craft and Curtis J. Richardson

3.1

Introduction

The Everglades is a 700,000 ha subtropical wetland whose origin dates to approximately 5,000 years BP when the rate of sea level rise slowed and peat began to accumulate in the shallow embayment of the south Florida peninsula (Gleason and Stone 1994). For the next 4,900 years, the Everglades was a net sink for organic carbon as peat accreted to depths of 1–3 m over much of the embayment and up to 4 m in areas south of Lake Okeechobee. During the past 100 years, however, as the population of south Florida swelled from approximately 20,000 to more than 4,000,000 people, efforts to drain the Everglades led to loss of nearly 65% of the original acreage (Kushlan 1989). Furthermore, the historical hydrology and low nutrient regimes characteristic of the extant Everglades have been modified by the vast network of canals, levees, and pumping stations (Walters et al. 1992) and, in northern areas, by nutrient-enriched drainage from the Everglades Agricultural Area (EAA; Davis 1994). This chapter describes some of the chemical and physical characteristics of the Everglades peatlands including the distribution of organic C and nutrients (N, P), historical and recent rates of peat accretion, nutrient accumulation, and the effects of anthropogenic hydrologic alteration and nutrient enrichment on these processes. Excellent reviews of the origin and development of the Everglades peatland have been written by Gleason et al. (1984) and Gleason and Stone (1994). The effects of eutrophication, especially P, on Everglades plant community structure and ecosystem processes have been described extensively (Belanger et al. 1989; Davis 1989, 1991; Koch and Reddy 1992; Craft and Richardson 1993a, 1995; Reddy et al. 1993; Qualls and Richardson 1995; Miao and Sklar 1998; Vaithiyanathan and Richardson 1999; Chiang et al. 2000; Qualls and Richardson 2000). Likewise, review articles describing efforts to drain the Everglades and the ecological consequences of these efforts are described by Light and Dineen (1994) and in this book’s Chaps. 2 and 8. The reader is referred to these papers to gain an in-depth understanding of the geological development and the effects of human activities on the Everglades peatlands.

59

60

3.2

C.B. Craft and C.J. Richardson

Methods

We focused on chemical and physical characteristics of the peat, including organic C, N, and P in the rooting zone; peat accretion; and organic C, N, and P accumulation. Twenty peat cores (30–50 cm deep) were collected along a north-to-south gradient from the Loxahatchee National Wildlife Refuge (Loxahatchee NWR), through Water Conservation Areas (WCAs) 2A, 2B, and 3A, to Everglades National Park (ENP; Fig. 3.1). The north-to-south gradient roughly parallels the historical pattern of sheetflow of water that overflowed from Lake Okeechobee and moved southeast and then southwest into Florida Bay (Fennema et al. 1994). With the exception of northern WCA-2A, which is impacted by nutrient-enriched drainage from the EAA, cores were collected from areas currently unaffected by agricultural nutrient loadings. At all locations, soils were collected from sawgrass (Cladium jamaicense), the dominant plant community of the Everglades (Gunderson 1994). Ten cores were collected from unenriched interior areas far removed from levees, canals, and water control structures: Loxahatchee NWR (n = 1), unenriched WCA-2A (n = 3), WCA-2B (n = 2), WCA-3A (3A) (n = 3), and ENP (n = 1) (Fig. 3.1). Ten cores were collected in areas proximal to canals and levees in enriched WCA-2A (n = 4) and in WCA-3A (AA, 12A, 12C, n = 2 for each location) (Fig. 3.1). Cores were sectioned in 2 cm depth increments and analyzed for bulk density, organic C, N, and P. Depth increments from all 20 cores were analyzed for 137Cs to calculate recent (30 years) rates of peat accretion and nutrient accumulation. Lead210 was measured in depth increments of eight cores (Loxahatchee NWR, WCA2A (enriched, n = 2), WCA-2A (unenriched, n = 3), WCA-3A, ENP) to quantify peat accretion and nutrient accumulation during the past 100 years. In several cores, subsurface peat was analyzed for 14C to quantify historical (> 500 years) rates of peat accretion. Detailed descriptions of the analytical methods and results are published in Craft and Richardson (1993a, b, 1998).

3.3 3.3.1

Results and Discussion Everglades Peat Types

Much of the Everglades is underlain by peat. Two types, Everglades peat and Loxahatchee peat, encompass over 7,000 km2 (Table 3.1). Other peat types include Okeechobee muck (130 km2) and Okeelanta peaty muck (105 km2), which are found along the southern and eastern shore of Lake Okeechobee in the EAA (Gleason and Stone 1994). Gandy peat (77 km2) is found associated with tree islands, usually atop Everglades or Loxahatchee peat.

3 Soil Characteristics of the Everglades Peatland

61

Fig. 3.1 Sampling locations where peat cores were collected along the north-to-south (e.g., downstream) gradient through the Everglades

Of the two most abundant peat types, Everglades peat is thinner, higher in mineral content, and less decomposed than Loxahatchee peat. It forms mostly from

62

C.B. Craft and C.J. Richardson

Table 3.1 Selected characteristics of the dominant types of peat in the Everglades USDA classificationa

Areab (km2)

Thicknessb (m) Originb

pHc

4,420 0.5–2.0 Sawgrass Euic hyperther5.5–6.5 mic typic (Cladium medihemist jamaicense) (Pahokee series)d Loxahatchee Euic hyperthermic 2,950 2.0–3.0 Water lily 5.0–6.5 peat typic medis(Nymphaea aprist (Terra odorata) Ceia series) a SCS (1978) b Gleason and Stone (1994) c Jones (1948) d Also includes Terra Ceia, Lauderhill, and Dania series Everglades peat

Organic contentc (%) 85–92

92

partial decomposition of sawgrass, the dominant wetland plant species. Everglades peat underlies much of the central and southern Everglades (Gleason and Stone 1994). Loxahatchee peat is thicker, higher in organic content, N, and acidity and more decomposed than Everglades peat (Table 3.1; Gleason and Stone 1994). Loxahatchee peat forms from vegetation of sloughs, especially water lily. Loxahatchee peat underlies WCA-1 (Loxahatchee NWR) and northeastern areas of WCA-2A (Gleason and Stone 1994).

3.3.2

Surface Soil Bulk Density, Organic Carbon, Nitrogen, and Phosphorus

Bulk density (0–24 cm depth) exhibited a general increase from north to south along the gradient of historical sheetflow (Fig. 3.2a). Bulk density was 0.06 ± 0.01 g cm−3 in Loxahatchee NWR, 0.10 ± 0.01 g cm−3 in the WCAs, and 0.17 ± 0.06 g cm−3 in ENP. Increased bulk density from north to south corresponds to increased mineral content that is the result of decreasing thickness of the peat from north to south (see Table 3.1). As the peat thins from north to south, mineral material from the underlying bedrock mixes with the peat, resulting in an increase in bulk density and inorganic constituents (Al, Fe) along the north-to-south gradient (Craft and Richardson 1997). Soil organic C, which varied inversely with bulk density, generally decreased along the north-to-south gradient (Fig. 3.2b). Organic C was 51 ± 4% in Loxahatchee NWR, 44 ± 2% in the WCAs, and 43 ± 5% in ENP.

3 Soil Characteristics of the Everglades Peatland

63

Fig. 3.2 Surface soil (a) bulk density, (b) organic carbon, (c) nitrogen, and (d) phosphorus along the north-to-south gradient through the Everglades. Bars represent mean values (±1 standard error) of 2 cm depth increments from 0 to 24 cm. Values at locations AA (0–20 cm) and 12A (0–12 cm) represent the entire peat layer down to bedrock. En is the nutrient-enriched area of WCA-2A, Un is the unenriched area

In contrast to bulk density and organic C, there was no apparent trend in N or P along the north-to-south gradient (Fig. 3.2c,d). Nitrogen concentrations were highest in Loxahatchee NWR (4.1 ± 0.5%) and areas of WCA-3A (4.2 ± 0.2% at 3A and 12A) and lowest in WCA-2B (3 ± 0.1%) and ENP (2.9 ± 0.9%). Soil P was highly variable from north to south through the Everglades (Fig. 3.2d). Average phosphorus concentration was much higher in northern WCA-2A (1,250 µg g−1) where nutrient-enriched agricultural drainage enters from pumping stations along the Hillsboro canal. The extreme northern and southern areas of the Everglades, Loxahatchee NWR, and ENP had the lowest P concentrations (Fig. 3.2d). Soil P was only 100 ± 50 µg g−1 in interior areas of the Loxahatchee NWR and 200 ± 50 µg g−1 in ENP. Loxahatchee NWR and ENP receive much of their water and nutrient inputs from rainfall (Light and Dineen 1994; Fennema et al. 1994). In contrast, depending on their proximity to canals, the WCAs receive at least some water and nutrients and, in some cases, substantial quantities of surface runoff from the network of canals that dissect the WCAs. Other locations in the WCAs had soil P concentrations that were intermediate (620 ± 60 µg g−1) between the nutrient-enriched area of WCA-2A and Loxahatchee NWR/ENP.

64

3.3.3

C.B. Craft and C.J. Richardson

Bulk Density, Organic Carbon, Nitrogen, and Phosphorus as a Function of Depth

Bulk density, organic C, and nitrogen changed little with depth in the upper 30 cm of peat (Fig. 3.3a–c). Published studies from the Everglades report similar trends in bulk density, C, and N with depth (Craft and Richardson 1993a, b, 1998; Reddy et al. 1993). In the Everglades, peat is derived mostly from deposition of plant biomass (sawgrass, water lily) that contains a fixed proportion of C to N (Craft et al. 1995; Schlesinger 1997). As a result, accumulating peat typically

Fig. 3.3 Distribution of soil (a) bulk density, (b) organic carbon, (c) nitrogen, and (d) phosphorus with depth in the upper 30 cm of peat collected from Everglades National Park. Note that bulk density, organic C, and N are relatively constant with depth but most phosphorus is concentrated near the surface, in the zone of greatest biological activity

3 Soil Characteristics of the Everglades Peatland

65

contains relatively constant concentrations of C and N with depth. In contrast to C and N, phosphorus exhibited a marked decrease with depth throughout the upper 30 cm of peat (Fig. 3.3d). Published studies throughout the Everglades also reported decreasing P with depth (Craft and Richardson 1993a, b, 1998; Reddy et al. 1993). High levels of P in surface peat often are evidence of wetland P limitation because the limiting nutrient, in this case P, is concentrated and maintained in the zone of greatest biological activity (Richardson and Marshall 1986; Walbridge 1991).

3.3.4

Thickness and Accretion of Peat

As discussed previously, peat thickness decreased from north to south through the Everglades (Table 3.2). Peat thickness ranged from 3 m in Loxahatchee NWR to less than 0.5 m in WCA-2B, WCA-3A, and ENP. In spite of the differences in the thickness of the peat, recent (137Cs) rates of peat accretion were relatively uniform (Table 3.2). With the exception of the enriched area of WCA-2A, recent rates of accretion ranged from 0.8 to 3.2 mm year−1. Recent peat accretion was lowest in the Loxahatchee NWR and ENP, 0.8 mm year−1. Loxahatchee NWR and ENP receive most water and nutrient inputs from rainfall so that hydroperiod is shorter and nutrient inputs are lower at these locations as compared to the WCAs. Low nutrient inputs, especially P, lead to reduced NPP, organic C inputs to the soil and, hence, peat accretion at these locations.

Table 3.2 Peat thickness and peat accretion along a north-to-south gradient in the Everglades Location Loxahatchee NWR WCA-2A (nutrient enriched) WCA-2A (unenriched)

Sample number

n = 1 core n = 4 for 137Cs, n = 2 for 210Pba n = 3 for 137Cs, n = 3 for 210Pb WCA-2B n = 2 for 137Cs WCA-3A (AA) n = 2 for 137Cs WCA-3A (3A) n = 3 for 137Cs, n = 1 for 210Pb WCA-3A (12A) n = 2 for 137Cs WCA-3A (12C) n = 2 for 137Cs Everglades National Park n = 1 core Mean (excluding enriched) n = 20 for 137Cs, n = 8 for 210Pb

a

Peat thickness (m)

Accretion rate (mm year−1) 137

210

3.0 2.4

0.8 5.3 ± 0.9

1.1 0.9 5.8 ± 1.4 –

2.5

2.0 ± 0.6

2.0 ± 0.1 0.6

0.5 0.3 0.6

2.4 ± 0.4 2.0 ± 0 1.7 ± 0.3

– – 1.4

– – 0.2

0.1 0.4 0.4 –

2.8 ± 0 3.2 ± 0.4 0.8 2.0 ± 0.3

– – 1.9 1.6 ± 0.2

– – 0.2 0.5 ± 0.2

Cs

Accretion rates based on 210Pb encompass the period 1962–1994

Pb

14

C

66

C.B. Craft and C.J. Richardson

Recent peat accretion was highest in the enriched area of WCA-2A, 5.3 ± 0.9 mm year−1. Previously published studies also reported higher rates of peat accretion in northern WCA-2A as compared to other locations in response to P loadings from agricultural drainage (Craft and Richardson 1993a, b, 1998; Reddy et al. 1993). At the enriched location, recent (30 years) rates of peat accretion based on 210Pb (5.8 ± 1.4 mm year−1) were similar to rates determined using 137Cs (5.3 ± 0.9 mm year−1, Table 3.2). Unenriched areas of WCA-2A and WCA-3A also possessed similar rates of peat accretion based on 137Cs (2.0 ± 0.9 mm year−1 for WCA-2A, 2.4 ± 0.4 mm year−1 for WCA-3A) and 210Pb (2.0 ± 0.1 mm year−1 for WCA-2A, 1.4 mm year−1 for WCA-3A). Carbon-14 dating of peat collected from similar depths (37–46 cm) of four cores revealed a gradient of decreasing peat accretion and increasing age of the peat along the north-to-south gradient. Subsurface peat (46 cm) in the Loxahatchee NWR and WCA-2A was deposited only 500 and 800 years BP, respectively (Table 3.3). In WCA-3A, peat collected from the same depth was deposited more than 2,000 years BP. Carbon-14 dating of basal peat (37 cm) collected from ENP indicated that, at this location, peat began to accumulate more than 2,500 years ago. Carbon-14 dating of basal peat from the EAA, Loxahatchee NWR, and northern WCA-2A indicates that peat accumulation began earlier in those areas, 4,500–4,800 years BP (McDowell et al. 1969; Gleason and Stone 1994). In southern WCA-2A, the onset of peat accumulation began 2,000–3,000 years BP. The large difference in 14C age of peat collected from similar depths from the Loxahatchee NWR, WCA-3A, and ENP suggests that historical environmental factors differed greatly between the northern and southern Everglades. The underlying bedrock of southern Everglades consists of Miami limestone that is more permeable than the Fort Thompson formation that underlies areas of the northern Everglades (Gleason et al. 1984). The porous nature of the Miami limestone perhaps contributes to reduced hydroperiod and, consequently, increased fire frequency in the southern Everglades, leading to reduced peat accretion and thinner peat at these locations. Carbon-14-based peat accretion ranged from 0.9 mm year−1 in the Loxahatchee NWR to 0.2 mm year−1 in interior areas of WCA-3A and ENP. The average rate of accretion based on 14C was 0.5 ± 0.2 mm year−1. In the far northern Everglades, in what is now the Everglades Agricultural Area, 14C dating of basal (3.5 m depth) and subsurface peat (1.3 m depth) yielded accretion rates of 0.8 and Table 3.3 14C age of subsurface peat along a north-to-south gradient in the Everglades (n = 1 per location) 14 C age (years BP) Location Loxahatchee NWRa WCA-2A (unenriched)a WCA-3A (3A)a Everglades National Parkb a Subsurface peat (46 cm depth) b Basal peat (37 cm depth)

530 ± 50 830 ± 60 2,060 ± 60 2,550 ± 60

3 Soil Characteristics of the Everglades Peatland

67

1.5 mm year−1, respectively (McDowell et al. 1969). With the exception of Loxahatchee NWR, 14C peat accretion was much lower than long-term rates in the EAA area or our recent accretion measurements. In contrast to 137Cs and 210 Pb that measure accretion during past 30 and 100 years, respectively, the 14C method measures peat accretion over a period of centuries to millennia (Table 3.3). Thus, the 14C method accounts for long-term decomposition, subsidence, and fire, all of which contribute to reduce vertical accretion over time. The similarity in 137Cs, 210Pb, and 14C peat accretion (0.9 ± 0.1 mm year−1) in Loxahatchee NWR suggests that the rate of peat accretion in the least disturbed portion of northern Everglades has been relatively stable over time, perhaps due to fact that this is a rainfall-driven area (in the interior), and hydroperiod and fire frequency are more uniform compared to other hydrologically disturbed areas in the northern Everglades.

3.3.5

Organic Carbon, Nitrogen, and Phosphorus Accumulation

Like peat accretion, there was no clear gradient in organic C accumulation from north to south. Organic C accumulation was lowest in Loxahatchee NWR and ENP (28 ± 9 g m−2 year−1) and highest in the enriched area of northern WCA-2A (192 ± 30 g m−2 year−1) (Fig. 3.4a). Low recent organic C accumulation in Loxahatchee NWR and ENP probably reflects the isolated nature of these rainfall-driven locations as compared to the WCAs, where hydrology has been altered by canals and water control structures. Recent (30 years) organic C accumulation in enriched WCA-2A (192 g m−2 year−1) based on 137Cs was more than double that of unenriched areas (86 ± 35 g m−2 year−1). Previous studies of nutrient-enriched areas of northern WCA-2A also reported much higher organic C accumulation as compared to unenriched areas of WCA-2A (Craft and Richardson 1993b; Reddy et al. 1993) as well as other areas in the WCAs (Craft and Richardson 1998). Long-term (100 years) organic C accumulation based on 210Pb generally was comparable to recent (137Cs) organic C accumulation (Fig. 3.4a). Long-term organic C accumulation was lowest in Loxahatchee NWR (36 g m−2 year−1) and highest in the enriched area of WCA-2A (189 g m−2 year−1). In Loxahatchee NWR and ENP, long-term organic C accumulation was two to three times higher than recent accumulation (Fig. 3.4a), perhaps reflecting extended hydroperiod at these interior locations prior to anthropogenic drainage efforts during the twentieth century. With the exception of enriched WCA-2A, recent organic C accumulation was higher in areas of southern WCA-3A (12A, 12C) as compared to other locations (Fig. 3.4a). At these locations, hydroperiod was extended by construction of an east–west highway and levee across the Everglades during the twentieth century. Consequently, water pools along the north side of the levee in WCA-3A, resulting in extended hydroperiod (SFWMD 1992) and enhanced peat accretion and organic C accumulation at these locations (Craft and Richardson 1993a). Similarly, the combination of extended hydroperiod and increased P loadings accelerated organic

68

C.B. Craft and C.J. Richardson

Fig. 3.4 Recent (30 years, 137Cs) and long-term (100 years, 210Pb) (a) organic carbon, (b) nitrogen, and (c) phosphorus accumulation along the north-to-south gradient through the Everglades. Bars represent mean values (±1 standard error) of two to four peat cores collected from the same location (see Table 3.1 for details). En is the nutrient-enriched area of WCA-2A, Un is the unenriched area. At the enriched location, accumulation rates based on 210Pb encompass the period 1962–1994

3 Soil Characteristics of the Everglades Peatland

69

C accumulation in northern WCA-2A as compared to other areas of the Everglades (Craft and Richardson 1993a). Nitrogen accumulation exhibited the same pattern as peat accretion and organic C accumulation. Recent (137Cs) N accumulation was lowest in Loxahatchee NWR and ENP (2.3 ± 0.8 g m−2 year−1) and highest in the enriched area of WCA-2A (14.1 ± 2.4 g m−2 year−1) (Fig. 3.4b). Extended hydroperiod in southern WCA-3A (12A, 12C) also resulted in higher N accumulation as compared to other locations. Long-term (210Pb) N accumulation was similar to recent accumulation, ranging from 2.9 g m−2 year−1 in Loxahatchee NWR to 13.5 ± 2.5 g m−2 year−1 in the enriched area of WCA2A. Like organic C, recent N accumulation in ENP (3.1 g m−2 year−1) was much lower than long-term accumulation (10.2 g m−2 year−1). In ENP, low N (and C) accumulation during the past 35 years may reflect reduced hydroperiod caused by drainage activities in the WCAs that reduced water flows to ENP (Light and Dineen 1994). Phosphorus accumulation along the north-to-south gradient varied more than organic C and N accumulation (Fig. 3.4c). In unenriched areas, recent (137Cs) P accumulation ranged from 0.01 g m−2 year−1 in Loxahatchee NWR to 0.20 ± 0.03 g m−2 year−1 at locations 12A and 12C in southern WCA-3A. In unenriched areas, P accumulation was lowest in areas that primarily are rainfall-fed (e.g., Loxahatchee NWR and ENP) and highest in areas of extended hydroperiod and enhanced peat accretion, 12A and 12C in southern WCA-3A. Long-term (210Pb) accumulation of P in unenriched areas was low, 0.06 ± 0.01 g m−2 year−1, as compared to recent accumulation (Fig. 3.4c). Phosphorus accumulation in northern WCA-2A (enriched) was much higher than in unenriched areas of the Everglades (Fig. 3.4c). Since 1963, 137Cs-based P accumulation in the enriched area (0.46 ± 0.05 g m−2 year−1) was three times higher as compared to unenriched areas of the WCAs (0.15 ± 0.02 g m−2 year−1) and 20 times higher than in Loxahatchee NWR and ENP (0.02 ± 0.01 g m−2 year−1). Likewise, 210 Pb-based P accumulation in enriched WCA-2A since the early 1960s (0.40 ± 0.0 g m−2 year−1) was seven times higher as compared to unenriched areas of the Everglades (0.06 ± 0.01 g m−2 year−1). Increased P accumulation in northern WCA-2A reflects increased loading of P and water to WCA-2A from the EAA during the past 50 years (Craft and Richardson 1993). Construction activities in the early 1960s resulted in complete impoundment by levees and by installation of four water control structures that replaced sheetflow across the fen (Light and Dineen 1994). Collectively, these hydrologic alterations accelerated eutrophication of WCA-2A while sparing areas downstream of WCA-2A from excessive phosphorus.

3.3.6

Effects of Increased Water and Nutrient Loading on Peat Accretion and Nutrient Accumulation

The effect of anthropogenic water and nutrient loadings on Everglades plant community structure and ecosystem processes has received much attention. Much of this research has focused on WCA-2A, particularly northern areas that receive vast quantities of nutrient-laden drainage water from the EAA. Between 1978 and 1987, an average

70

C.B. Craft and C.J. Richardson

of 459,000,000 m3 of water, 1,814 metric tons of N, and 60 metric tons of P was released annually into northern WCA-2A through four water control structures along the Hillsboro canal (SFWMD 1992). The combination of extended hydroperiod and increased P, the primary limiting nutrient in the Everglades (Steward and Ornes 1975b; Belanger et al. 1989; Craft et al. 1995), has led to a shift in plant community composition as cattail encroached into sawgrass and slough communities (Belanger et al. 1989; Urban et al. 1993; Davis 1994; Jensen et al. 1995; Bartow et al. 1996; Vaithiyanathan and Richardson 1999; Richardson et al. 1999). Furthermore, P enrichment has contributed to increased net primary production (Davis 1989, 1991; Miao and Sklar 1998), decomposition (Davis 1991; Qualls and Richardson 2000) and enrichment of soil P pools (Koch and Reddy 1992; DeBusk et al. 1994; Qualls and Richardson 1995; Richardson et al. 1999) in northern WCA-2A. The effects of increased water and P loading also enhanced peat and nutrient accumulation in northern WCA-2A. Since the early 1960s, peat accretion in the enriched area of WCA-2A more than doubled (5.3 ± 0.9 mm year−1) as compared to unenriched areas in southern WCA-2A (Table 3.4). Prior to 1962, peat accretion in the enriched area (1.9 ± 0.2 mm year−1) was similar to unenriched areas in southern WCA-2A (2.0 mm year−1). Like P, organic C, and N accumulation in the enriched

Table 3.4 Effects of increased water and phosphorus loadings on peat accretion and nutrient accumulation in nutrient-enriched and -unenriched areas of WCA-2A Accretion rate (mm year−1)

Organic C (g m−2 N (g m−2 year−1) year−1)

P (g m−2 year−1)

1964–1990 (137Cs)a 1964–1990 (137Cs)b 1962–1990 (210Pb)a 1900–1962 (210Pb)a Unenriched

5.3 ± 0.9

192 ± 30

14.1 ± 2.4

0.46 ± 0.05

6.0 ± 1.1

215 ± 38

14.1 ± 2.3

0.53 ± 0.14

5.8 ± 1.4

189 ± 23

13.5 ± 2.5

0.40 ± 0.0

1.9 ± 0.2

111 ± 36

6.7 ± 2.2

0.06 ± 0.0

1964–1990 (137Cs)c 1964–1990 (137Cs)d 1900–1990 (210Pb)c Mean enriched (this study) Mean unenriched (this study)

2.0 ± 0.6

86 ± 35

5.6 ± 3.0

0.09 ± 0.03

3.6 ± 0.5

124 ± 21

7.7 ± 1.4

0.15 ± 0.02

2.0 ± 0.1

112 ± 6

7.0 ± 0.3

0.07 ± 0.01

5.6 ± 0.2

190 ± 2

13.8 ± 0.3

0.43 ± 0.03

6.3 ± 0.7

0.08 ± 0.01

Time interval (years) Enriched

a

2.0 ± 0.0

99 ± 13

n = 4 for 137Cs, n = 2 for 210Pb Reddy et al. (1993); (n = 6, one core each was collected from stations 10, 11A, 11B, 12, 13, and 14) c n = 3 for 137Cs, n = 3 for 210Pb d Reddy et al. (1993); (n = 3, one core each was collected from stations 15–17) b

3 Soil Characteristics of the Everglades Peatland

71

area since 1962 was twice that of unenriched areas of WCA-2A or the enriched area prior to 1962 (Table 3.4). Reddy et al. (1993) reported similar increases in peat accretion (6.0 ± 1.1 mm year−1) and organic C (215 ± 38 g m−2 year−1) and N accumulation (14.1 ± 2.3 g m−2 year−1) in the enriched area using the 137Cs method (Table 3.4). Phosphorus accumulation in the enriched area since the early 1960s was five to seven times higher (0.43 ± 0.03 g m−2 year−1) as compared to the period 1900–1960 (0.06 ± 0.0 g m−2 year−1) or to unenriched areas of southern WCA-2A (0.09 ± 0.03 g m−2 year−1). Reddy et al. (1993) also reported much higher P accumulation in the enriched northern area (0.53 ± 0.14 g m−2 year−1) as compared to southern WCA-2A (0.15 ± 0.02 g m−2 year−1). Thus, in addition to obvious shifts in plant species composition, increased water and P loading to northern WCA-2A have led to less obvious changes in soil processes such as peat accretion and nutrient accumulation.

3.3.7

Comparison with Other North American Peatlands

Organic C and nutrient accumulation in unenriched areas of the Everglades was compared to other North American peatlands (Table 3.5). Organic C accumulation in the Everglades (98 ± 13 g m−2 year−1), a southern fen (e.g., groundwater-fed, Table 3.5 Comparison of organic C, N, and P accumulation in unenriched areas of the Everglades with other North American peatlands Organic C (g m−2 Nitrogen (g m−2 year−1) year−1)

Phosphorus (g m−2 year−1)

81 ± 14 124 ± 21 90 ± 20 98 ± 13 38 ± 10 49 ± 15 43 ± 6

6.5 ± 1.2 7.7 ± 1.4 6.4 ± 1.6 6.7 ± 0.4 2.6 ± 0.7 3.5 ± 1.1 3.1 ± 0.5

0.12 ± 0.03 0.15 ± 0.02 0.06 ± 0.01 0.11 ± 0.03 0.07 ± 0.02 0.09 ± 0.04 0.08 ± 0.01

82 Okefenokee (210Pb)d (GA) 127 Pocosin (210Pb)e (NC) Northern bogs (210Pb)f (MA, MD, 79 ± 4 PA, WV, MN) Mean (fens) 71 ± 28 Mean (bogs) 96 ± 16

3.8 3.0 2.1 ± 0.3

0.15 0.06 0.11 ± 0.01

4.9 ± 1.2 3.0 ± 0.5

0.10 ± 0.02 0.11 ± 0.03

Peatland Fens Everglades (137Cs)a (FL) Everglades (137Cs)b (FL) Everglades (210Pb)a (FL) Mean Northern fens (137Cs)c (MI) Northern fens (210Pb)c (MI) Mean Bogs

a

This study Reddy et al. (1993); (n = 3, one core each was collected from stations 15–17 of WCA-2A) c C.B. Craft, unpublished data (n = 2 fens) d Schlesinger (1978) (n = 1 core) e C.B. Craft, unpublished data (n = 1 core) f Hemond (1980, 1983) and Wieder et al. (1994) cited in Craft and Richardson (1998) b

72

C.B. Craft and C.J. Richardson

circumneutral pH), was greater than northern fens (43 ± 6 g m−2 year−1) but similar to precipitation-fed, acidic bogs (96 ± 16 g m−2 year−1). Nitrogen accumulation was much higher in the Everglades (6.7 ± 0.4 g m−2 year−1) as compared to bogs (3.0 ± 0.5 g m−2 year−1) and northern fens (3.1 ± 0.5 g m−2 year−1). Higher rates of N accumulation in the Everglades reflect increased N fixation by cyanobacteria that are abundant in the open water sloughs (Swift and Nicholas 1987; Craft and Richardson 1993a). Overall, fens and bogs had comparable P accumulation, 0.10 ± 0.02 g m−2 year−1 for fens and 0.11 ± 0.03 g m−2 year−1 for bogs. However, long-term (210Pb measured) accumulation of P in the Everglades (0.06 ± 0.01 g m−2 year−1) is among the lowest reported North American peatlands (Table 3.5). Low P accumulation in the Everglades is a consequence of the historical low P inputs and reflects the current role of P as the primary limiting nutrient in this oligotrophic peatland.

3.4

Conclusions and Lessons

The Everglades is a subtropical fen wetland underlain by deep (2–3 m) organic soils in the north that decrease in thickness and increase in mineral content to the south, along the direction of historical water flow. Recent (137Cs) peat accretion (0.8 ± 0 mm year−1) and organic C (28 ± 9 g m−2 year−1), N (2.3 ± 0.8 g m−2 year−1), and P (0.02 ± 0.01 g m−2 year−1) accumulation are lowest in interior areas such as the Loxahatchee NWR and ENP that are primarily rainfall-fed. Anthropogenic alteration of the hydrologic cycle has reduced organic C, N, and P accumulation in overdrained interior areas and increased accumulation in areas near canals where inundation is extended. Anthropogenic P loading from agricultural drainage has led to dramatically higher peat accretion (5.6 ± 0.2 mm year−1) and organic C (190 ± 2 g m−2 year−1), N (13.8 ± 0.3 g m−2 year−1), and P (0.43 ± 0.03 g m−2 year−1) accumulation in eutrophic areas of the Everglades, especially northern WCA-2A. Unenriched areas of the Everglades possess some of the lowest rates of P accumulation (≤ 0.06 g m−2 year−1) of North American peatlands. The dramatic response of peat accretion and nutrient accumulation to P loadings reflects the importance of P in regulating ecosystem-level processes in this P-limited oligotrophic peatland. Our initial long-term P accretion rates in the enriched areas provided a basis for estimating the number of storm water treatment areas (STAs) that would be needed to effectively remove P from the water column and store P permanently in the soil, mainly as organic P. Richardson and Craft (1993) calculated that more than 30,000 ha would be required to reduce P to levels close to background concentrations. An increased amount of STA wetland area is now being considered for construction by the SFWMD to meet the required 10 µg l−1 criterion (SFWMD 2006).

4

Vegetation and Algae of the Everglades Fen Curtis J. Richardson, Jan Vymazal, and John G. Zahina

4.1

Introduction

A basic understanding of the factors controlling plant community structure and succession in the Everglades is essential to the development of any water or nutrient management plans for maintaining Everglade plant communities. Six factors contributing to the development of the plant communities that exist within the Everglades are hydrology, climate, fire, phosphorus, disturbance, and exotics (Richardson et al. 2000; see Chaps. 2 and 26). Historically, climate was the primary factor controlling succession in Everglades communities, with fire and hydrology next in importance. However, human activities in the Everglades have caused a shift of influence among the various factors. More recently, nutrient additions, exotic species invasions, and disturbance (e.g., airboat trails, swamp buggy tracks) have become increasingly important. The region’s hydroperiod (i.e., the number of days that the wetland ecosystem has standing water at or near the surface) has been dramatically altered – decreased in some areas and increased in others – over the past 60 years (Davis and Ogden 1994a). Flooding in many areas, caused by extensive diking in the Everglades, has decreased fire’s importance in determining succession patterns. Thus, if our understanding of Everglades plant community structure is to be useful for the development of effective management plans, it must adequately document anthropogenic changes that have taken place in those communities. As noted in Chap. 2, we classify the Everglades as a fen or alkaline peatland, but the dominant plant communities (cattail marsh, wet prairie, etc.) are identified by their historic names in this volume. The original major macrophyte vegetation communities of the Everglades (Table 4.1) consisted of large areas of sawgrass, slough, wet prairie, and smaller areas of tree islands (Davis 1943; Loveless 1959). The current communities show a considerable loss of wetland types as well as a 49% loss of total area. Tragically, there has been a 100% loss of the swamp forest, peripheral wet prairies, and cypress stands. Even the largest communities of pure sawgrass plains have been reduced by 74% (Table 4.1). The objectives of this chapter are to briefly review the species composition of plant communities in the Florida Everglades as well as present an assessment of aboveand belowground biomass for each plant community. A more complete survey was 73

74

C.J. Richardson et al.

Table 4.1 The aerial coverage of wetland communities in the Everglades of the south Florida Landscape type Predrainage area (ha) Current area (ha) Percent loss Swamp forest 60,000 0 100 Sawgrass plains 238,000 63,000 74 Slough/tree islands/sawgrass 311,000 271,000 13 mosaic Sawgrass-dominated 179,000 94,000 47 Peripheral wet prairies 117,000 0 100 Cypress stands 12,000 0 100 Southern marl-forming marshes 249,000 190,000 24 Total coverage 1,666,000 618,000 49 The percentage loss is based upon predrainage estimates and current estimates for each particular plant community type. Several of the community types have been combined from the original data, since it was impossible to separate which components were utilized in the aerial surveys (from Davis et al. 1994)

also done for several conservation areas, and a detailed list of vegetation in Water Conservation Areas 2 and 3 is presented in Appendix 4.1.

4.2

Plant Communities of the Everglades

A complete listing of the plant species characteristic of each vegetation community in the Everglades is shown on Table 4.2. The dominant species for each community are noted, as well as their growth form. A typical cross-sectional profile of these plant communities with annual water depths is shown in Fig. 4.1 and provides a basis for comparing the importance of water levels on plant community species composition. Photographic Plates 2–5 provide the reader with a visual basis for comparing plant community structure.

4.2.1

Sawgrass

Sawgrass (Cladium jamaicense) is the dominant vegetation community found throughout the freshwater Everglades fen (Fig. 4.1a,b, Plate 2). Sawgrass grows to 2–3 m in height on deep peat but only 0.5 m on shallow peat. It prefers sites with a fairly constant water depth of 10–20 cm (Toth 1987; Gunderson 1990). Its presence in the Everglades is due to its ability to survive fire, low soil nutrient content, and occasional freezing (Stewart and Ornes 1975b). Sawgrass does not survive well in highly variable (>30 cm) water regimes (Toth 1987). The current diking and flooding in portions of WCA-2 as well as in other parts of the Everglades have resulted in the loss of this community due to deep and fluctuating water levels. Sawgrass

4 Vegetation and Algae of the Everglades Fen

75

Table 4.2 Characteristic plant taxa in the vegetation communities of the Everglades system Community

Species

Common name

Growth form

Ponds, slough

(D) Nymphaea odorata Ait. (D) Nymphoides aquatica (S.G. Gmel.) Kuntze Nuphar lutea (L.) Sibth. & J.E. Smith Sagittaria lancifolia L. Pontederia cordata L. Bacopa caroliniana (Walt.) Robins Utricularia foliosa L. Utricularia purpurea Walt.

White water lily Floating hearts

FL FL

Spatterdock

FL

Leafy bladderwort Purple bladderwort

S S

(D) Cladium jamaicense Crantz. Justicia angusta (Chapm.) Small Eleocharis cellulosa Torr. Typha domingensis Pers. (D) Cladium jamaicense Crantz. Crinum americanum L. Peltandra virginica (L.) Schott & Endl Hymenocallis latifolia (Herb.) M. Roem. Aeschynomene pratensis Small Ipomoea sagittata Poir.

Sawgrass Pineland water-willow Gulfcoast spikerush Southern cattail Sawgrass Southern swamp-lily Green arum

E E E E E E E

Mangrove spider-lily

E

Netted shy-leaf Arrow-leaf morningglory

E V

(D) Eleocharis cellulosa Torr. (D) Eleocharis elongata Champ. (D) Rhynchospora tracyi Britt. Rhynchospora inundata (Oakes) Fern. (D) Panicum hemitomon Schult. Paspalidium geminatum (Forssk.) Stapf Crinum americanum L. Bacopa caroliniana (Walt.) Robins Sagittaria lancifolia L. Oxypolis filiformis (Walt.) Britt. (D) Cladium jamaicense Crantz. (D) Muhlenbergia filipes Schizachyrium rhizomatum (Swallen) Gould Dichromena colorata (L.) Hitchc. Schoenus nigricans L. Aristida purpurascens Poir. in Lam.

Coastal spikerush Water spikerush Tracy’s beakrush Horned beakrush

E E E E

Maidencane Water panicum

E E

Southern swamp-lily Water hyssop

E S

Lance-leaf arrowhead Water dropwort Sawgrass Muhly grass South Florida bluestem White-top sedge

E E E E E

Black sedge Arrowfeather grass

E E

Sawgrass “prairie” Tall stature

Intermediate stature

Wet prairies (peat) Eleocharis marshes Rhynchospora flats

Panicum flats

Wet prairies (marl)

Lance-leaf arrowhead E Pickerelweed E Water hyssop S

E

(continued)

76

C.J. Richardson et al.

Table 4.2 (continued) Community

Species

Panicum tenerum Beyr. Rhynchospora divergens Champ. ex M.A. Curtis Rhynchospora microcarpa Balw. ex Gray Bayhead/swamp forest (D) Persea borbonia (L.) Spreng. (D) Magnolia virginiana L. (D) Ilex cassine L. (D) Salix caroliniana Michx. (D) Myrica cerifera L. Chrysobalanus icaco L. Blechnum serrulatum L.C. Richard Acrostichum danaeifolium Langsd. & Fisch. Acer rubrum L. Rhizophora mangle L. Pond apple forest (D) Annona glabra L. Fraxinus caroliniana Mill. Commelina gigas Small Cucurbita okeechobeensis (Small) Bailey Sambucus simpsonii Rehd. ex Sarg. Ficus aurea Nutt. Baccharis halimifolia L. Tillandsia spp. Willow heads

Cypress forests

Hardwood hammock

(D) Salix caroliniana Michx. Myrica cerifera L. Cephalanthus occidentalis L. Cladium jamaicense Crantz. Taxodium ascendens Brongn. Cladium jamaicense Crantz. Schizachyrium rhizomatum (Swallen) Dichromena colorata (L.) Hitchc. (D) Ficus aurea Nutt. (D) Bursera simaruba (L.) Sarg. (D) Quercus virginiana P. Mill. (D) Lysiloma latisiliquum Sabal palmetto (Walt.) Lodd. ex J.S. Shult. & Shult Celtis laevigata Willd. Morus rubra L. Citrus spp. Diospyros virginiana L.

Common name

Growth form

Bluejoint panicum Spreading beakrush

E E

Southern beakrush

E

Red bay

T

Sweet bay Dahoon holly Carolina willow Wax myrtle Coco plum Swamp fern

T T T T SH F

Leather fern

F

Red maple (north) Red mangrove (south) Pond apple Pop ash Climbing dayflower Okeechobee gourd

T T T T V V

Common elderberry

T

Florida strangler fig Saltbush Air plants Orchids Carolina willow Wax myrtle Buttonbush Sawgrass Pond cypress Sawgrass Gould South Florida bluestem White-top sedge

T SH EP EP T T SH E T E E

Florida strangler fig Gumbo limbo Live oak Wild tamarind Cabbage palm

T T T T T

Hackberry, Sugarberry Red mulberry Citrus Common persimmon

T T T T

G

(continued)

Table 4.2 (continued) Community

Successional shrub

Species

Common name

Growth form

Swietenia mahogani (L.) Jacq.

West Indian mahogany Paurotis palm

T

Florida royal palm

T

Jamaican nettletree Saltbush Wax myrtle Bracken fern

T SH T F

Acoelorraphe wrightii (Griseb. & Wendl.) Wendl ex Becc. Roystonea regia (Kunth) O.F. Cook Trema micranthum (L.) Blume Baccharis halimifolia L. Myrica cerifera L. Pteridium aquilinum

T

E emergent, EP epiphyte, F fern, FL floating-leaved, G grass, S submergent, SH shrub, T tree, V vine (modified from Gunderson and Loftus 1993) Dominant species are indicated by D, nomenclature follows Long and Lakela (1971), Avery and Loope (1980), and Tobe et al. (1998)

Fig. 4.1 Idealized profile of Everglades communities from hummock (a) to sloughs (b) (from Gundersen and Loftus 1993)

78

C.J. Richardson et al.

occurs either in almost pure stands or mixed with a wide variety of other plants, e.g., bulltongue (Sagittaria lancifolia), maidencane (Panicum hemitomon), pickerelweed (Pontederia cordata (Muhl.) Torr.), or cattail (Typha spp.) (Loveless 1959). Estimates of the extent of mixed sawgrass areas range from 65 to 70% of the remaining Everglades fen (Kushlan 1987; Loveless 1959; Schomer and Drew 1982; Steward and Ornes 1975b; Davis 1994). Davis et al. (1994) estimated that pure sawgrass-dominated areas currently make up only 38% of 417,000 ha of historic sawgrass plains and sawgrass-dominated areas (Table 4.1). The sawgrass/tree island/slough mosaic has not changed much in aerial extent.

4.2.2

Wet Prairies

Wet prairies are among the common vegetation types in the northern Everglades (Fig. 4.1b, Plate 3). Often referred to as “flats,” these freshwater communities are characterized by low stature and emergent plant species, and they are found in the northern and central Everglades in conjunction with tree islands (Goodrick 1984; Gunderson and Loftus 1993). Wet prairies exist on both peat and marl soil. Variation in species composition is shown on Table 4.2. The wet prairies in the south found on calcitic mud or marl occur on higher and drier sites but are wet 3–7 months of the year (Davis 1943; Gunderson and Loftus 1993). The water depth of these areas is generally less than sloughs but deeper than sawgrass; thus, the vegetation seldom burns (Fig. 4.1b). Loveless (1959) described three well-defined wet prairie associations in the northern Everglades (1) Rhynchospora flats, (2) Panicum flats, and (3) Eleocharis flats. These plant associations are composed primarily of Tracey’s horned rush (Rhynchospora tracyi), gulfcoast spikerush (Eleocharis cellulosa) – both sedges – and the wetland panic grass, maidencane (Panicum hemitomon). However, many other plant species may also be present on these flats, depending upon hydrological conditions, the season of the year, and soil type. Wet prairies usually dry out on an annual basis and are transition zone between sawgrass areas and sloughs (Goodrick 1984). Wet prairies require seasonal inundation with standing water present for 6–10 months of the year (Schomer and Drew 1982). Seasonal drying of the moist soil conditions in these communities is required for seed germination and establishment of new seedlings (Dineen 1972).

4.2.3

Sloughs

Sloughs are open water marsh areas, found primarily in the northeast and south-central Everglades, that are dominated by floating-leaved aquatic plants with some emergent plants of low stature (Davis 1943; Loveless 1959) (Fig. 4.1b, Table 4.2, Plate 4). Sloughs are among the most widespread community types in the Everglades.

4 Vegetation and Algae of the Everglades Fen

79

Aquatic sloughs represent the lowest elevation of the Everglades ecosystem, except for ponds. They have deep water levels averaging 30 cm annually and longer inundation periods than other Everglades wetland communities (Gunderson and Loftus 1993). Sloughs occur throughout the Everglades, with the largest pond– slough systems occurring in the Everglades National Park (ENP; Shark River and Taylor Sloughs) and portions of the northern Everglades (McPherson et al. 1976). Sloughs are narrow drainage channels that are water-filled, or at least wet, most of the year. The “valleys” of these channels average only a few cm to 60 cm below the elevation of adjacent fen areas. Not as extensive as they once were, some sloughs apparently have been replaced by either sawgrass or wax myrtle and willow stands. Cattail has also filled many of the sloughs in the natural-enriched areas of the northern Everglades (Rader and Richardson 1992; Urban et al. 1993; Craft and Richardson 1997). This reduction in slough areas has also been due to artificial drainage and the increase in sawgrass in the southern Everglades (Loveless 1959; Davis et al. 1994). Sloughs are easily recognized by their water drainage patterns and by characteristic plant species, such as white water lily (Nymphaea odorata), floating hearts (Nymphoides peltata), bladderworts (Utricularia spp.), spikerushes (Eleocharis spp.), spatterdock (Nuphar lutea), or water hyssop (Bacopa caroliniana) (Davis 1943; Loveless 1959; Van Meter-Kasanof 1973; Gunderson and Loftus 1993) (Table 4.2). Sloughs and wet prairies are ecologically important in the Everglades landscape. During the dry season, sloughs serve as important feeding areas and habitats for Everglades wildlife. As the higher elevation wet prairies dry out, sloughs provide refuge for aquatic invertebrates and fish. This high concentration of aquatic life, in turn, makes sloughs important feeding areas for Everglades wading bird populations. When the fen is reflooded, the animals that have survived in the sloughs repopulate the fen as water level rises (Loveless 1959). The slough/wet prairie sawgrass mosaic covers 271,000 ha (44%) of the remaining 618,000 ha area of the Everglades (Davis et al. 1994) (Table 4.1). The plant species diversity tends to be higher in sloughs and wet prairie communities than in pure sawgrass and cattail marsh communities (SFWMD 1992; Craft et al. 1995). The abundance of macroinvertebrates, fish, and wading birds is also higher in sloughs than in sawgrass and cattail marshes (SFWMD 1992; Rader and Richardson 1992, 1994; Davis and Ogden 1994a).

4.2.4

Ponds

Ponds are small open water areas that are scattered throughout most of the Everglades and represent the deepest water regime (Fig. 4.1b, Table 4.2, Plate 5). They occur in bedrock depressions where fire has burned away the peat (Loveless 1959). Alligator activity often maintains open water in the ponds, which the locals call “alligator holes” for this reason. Ponds are wet except in the driest years; thus they

80

C.J. Richardson et al.

are important habitats for animals, especially birds. These holes have borders of water lilies (Nymphaea spp.), spatterdock (Nuphar spp.), pickerelweed (Pontederia cordata), and woody species such as Carolina willow (Salix caroliniana) or water primrose (Ludwigia peruviana (L.) Hara) (Gunderson and Loftus 1993).

4.2.5

Tree Islands (Bayhead/Swamp Forests)

Everglades broadleaf, hardwood forests are locally called tree islands. The term refers to a variety of tree clusters that stand above a matrix of shorter vegetation (Craighead 1984). Tree islands occur throughout the entire region but are most abundant in the central part of WCA-1 (Loveless 1959). Tree islands may be either bayhead (swamp forests) or hammocks (upland forest), or a combination of the two (Davis 1943; Gunderson and Loftus 1993) (Fig. 4.1a, Plate 5). The dominant species in each type of forest is shown on Table 4.2. Red bay (Persea borbonia), swamp bay (Magnolia virginiana), dahoon holly (Ilex cassine), Carolina willow (Salix caroliniana), and wax myrtle (Myrica cerifera) dominate the swamp forests. The large tree islands have a teardrop shape with the main axis paralleling the flow of water. The small islands (≅100 m2) are usually round. The forests are found on the highest sites in the Everglades (Fig. 4.1a) on a peat classified as Gandy peat (Davis 1943; Loveless 1959). The sites are wet 2–6 months year−1, but in drought conditions these systems are very susceptible to burning (Gunderson and Loftus 1993). The soil P nutrient content of tree islands is usually much higher (>1,000 vs. 500 mg kg−1 of P) than the surrounding landscape (C.J. Richardson, unpublished data).

4.2.6

Willow Heads, Cypress Forests, Pond Apple Forests, and Hardwood Upland Hammocks

These forest types comprise only a small area of the Everglades. They include interesting communities with distinct species (Table 4.2). The pond apple forest (Annona glabra) existed primarily south of Lake Okeechobee in a band 5 km wide (Davis 1943). The land has been totally developed for agriculture and now the species only exists in small, scattered stands. Willow heads exist throughout the Everglades in monotypic stands (Loveless 1959). They exist in fire-disturbed areas, as well as around alligator holes. The upland hardwood hammocks are dominated by broadleaf hardwood trees of both temperate and tropical origin (Table 4.2, Fig. 4.1a). Dominant trees include live oak (Quercus virginiana), gumbo limbo (Bursera simaruba), sabal palm (Sabal palmetto), and strangler fig (Ficus aurea). The cypress forests are found only in the southwestern Everglades and are dominant in the adjacent Big Cypress National Preserve. Pond cypress (Taxodium ascendens) are very

4 Vegetation and Algae of the Everglades Fen

81

short and occur as widely scattered individuals displaying very stunted growth. They are often called dwarf or hat rack cypress and seldom reach heights over 3–5 m.

4.2.7

Periphyton

Periphyton and algal mats are seldom thought of as “valuable” ecological resources or even listed among vegetation community species. Several authors, however, have pointed out an important role of the calcareous periphyton and algal mats characteristic of sloughs and wet prairies (Pan et al. 2000). Components of the periphyton/algal mat (especially diatoms) are high-quality food for some animals (Browder et al. 1981; Browder et al. 1994). Photosynthesis by the algae in sloughs can raise daytime dissolved oxygen concentrations and pH much higher (7.5 to >10) than in adjacent sawgrass marshes (Belanger et al. 1989; Rader and Richardson 1992). Also, the calcareous periphyton deposits marl (calcitic mud), the second most common soil sediment type (190,000 ha) in the Everglades (Gleason and Spackman 1974; Davis et al. 1994). Three types of calcareous periphyton are known in the southern Everglades: calcareous blue-green, calcareous diatom-rich, and calcareous green (Browder et al. 1994). A defining feature of all calcareous periphyton is their high inorganic component, no less than 49% by mass (Browder et al. 1981, 1982). Based on statistical comparison by Browder et al. (1981), blue-green periphyton is considered to be any calcareous periphyton comprised of greater than 80% blue-green algae, by cell volume, during the latter part of the wet season. The term diatom-rich periphyton is used to label calcareous periphyton communities with a substantial diatom component. Periphyton communities are considered to be diatom-rich if their cell volume consists of less than 80% blue-green algae and diatoms make up a greater proportion of cell volume than greens (Browder et al. 1984). Green periphyton contains less than 80% cell volume as blue-green algae and a larger volume of green algae and desmids than diatoms. To conform to the definition given by Van Meter-Kasanof (1973), the green algal component includes desmids, particularly Pleurotaenium. The desmids in calcareous green periphyton are not the same acid water species as those described by Gleason and Spackman (1974) and Swift and Nicholas (1987). Rather, they are a less specialized group that tolerates hard water (Browder et al. 1994). Water quality and hydroperiod are perhaps the major factors influencing species composition and growth rates of Everglades periphyton communities (Van Meter 1965; Gleason 1972; Gleason and Spackman 1974; Wood and Maynard 1974; Swift 1981, 1984; Swift and Nicholas 1987; Browder et al. 1981; Flora et al. 1988; Rader and Richardson 1992; McCormick et al. 1998; Pan et al. 2000; Vymazal et al. 2000, 2001a). Periphyton is abundant in areas of the Everglades that retain the historic oligotrophic conditions of the fen. In these areas periphyton biomass on an area basis can reach values that are comparable to or higher than that of macrophytes.

82

C.J. Richardson et al.

Gleason and Spackman (1974) reported total periphyton biomass (excluding the epipelon) in WCA-1 in the range of 40–225 g AFDM m−2 and up to 447 g DM m−2. Van Meter (1965) reported DM up to 351 g m−2 from Taylor and Shark River sloughs in ENP. Also Browder et al. (1982) reported up to 2,682 g DM m−2 and 526 g AFDM m−2 from open water areas of southern Everglades ENP. The highest rates of periphyton productivity in the Everglades usually occur in sloughs and Eleocharis stands. McCormick et al. (1998) found periphyton biomass in the range of 100–1,600 g AFDM m−2 in oligotrophic sloughs and in stands of Eleocharis, but periphyton biomass was low in adjacent sawgrass (Cladium jamaicense) stands (7–52 g AFDM m−2). Also J. Vymazal and C.J. Richardson (unpublished results) found periphyton biomass in southern WCA-2A sloughs up to 3,300 g DM m−2 and 1,340 g AFDM m−2. Numerous results from the Everglades indicate that AFDM values comprise usually <50% of dry matter and very often <20% of dry matter values (e.g., Wood and Maynard 1974; Browder et al. 1994; Vymazal and Richardson 1995). The CaCO3 content of the calcareous periphyton is high and may be as high as 80% of the biomass (Van Meter-Kasanof 1973; Wood and Maynard 1974; Vymazal and Richardson 1995). Calcareous periphyton shows a preference for some plant species over others (Van Meter 1965; Gleason and Spackman 1974; Browder et al. 1994; Vymazal and Richardson 1995). Plants that are readily coated by periphyton include Eleocharis cellulosa, E. elongata (spikerush), Utricularia purpurea (purple-flowered bladderwort), Bacopa caroliniana (lemon bacopa), and Rhynchospora spp. (beakrush). Plants that have very little periphyton encrusting include Nymphaea odorata (white water lily) and Eriocaulon compressum (pipewort). Vymazal and Richardson (1995) reported that periphyton biomass growing on Eleocharis elongata was about 20 times higher on stem surface basis as compared to periphyton biomass growing on peduncles of Nymphaea odorata in southern WCA-2A. Van MeterKasanof (1973) reported that periphyton cylinder on stems in the southern Everglades ranged from 0 to 50 mm in diameter. Gleason and Spackman (1974) reported periphyton cylinder diameter values up to 63 mm, while Wood and Maynard (1974) reported algal mats as thick as 100 mm in parts of the southern Everglades. Periphyton communities are a major component of the detrital-based Everglades food web providing organic food matter and habitat for a wide variety of grazing invertebrates and foraging fish (Craighead 1971; Carter et al. 1973; Wood and Maynard 1974; Browder et al. 1981). Periphyton photosynthesis and respiration play an important role in controlling diurnal pH, dissolved oxygen, carbon dioxide, and calcium concentration within fen surface waters (Gleason 1972; Gleason and Spackman 1974; Wilson 1974). Algal photosynthesis accounts for a large portion of calcium precipitation within the fen and is responsible for the formation of marl soils within the southern Everglades (Gleason 1972; Gleason and Spackman 1974; Vymazal and Richardson 1995). Gleason and Spackman (1974) concluded that the aerial extent of calcareous periphyton is impressive, being found abundantly in all of the WCAs, the ENP, the Big Cypress National Preserve, and the inland prairies of the southern coast, the largest expanses being found in WCA-3 and the ENP.

4 Vegetation and Algae of the Everglades Fen

4.3

83

Vegetation and Plant Communities of Water Conservation Areas 2 and 3

Water Conservation Areas (WCAs) 2A and 2B of the northern Everglades are remnants of the vast sawgrass fen that once stretched from Lake Okeechobee to the southern tip of the Florida peninsula. WCA-2A covers an area of 450 km2 and WCA-2B covers 96 km2. They are located adjacent to the Everglades Agricultural Area (EAA), the Loxahatchee National Wildlife Refuge (also referred to as WCA-1), and WCA-3 (Zahina and Richardson 1997) (Fig. 1.2). The plant communities found in WCA-2A, WCA-2B, and WCA-3A can be broadly defined by the following habitat types, ranked from highest to lowest mean water depth: deep water (canals), slough, flat (wet prairies), marsh, peat battery, tree island, transitional levee, dry levee, and epiphytic (Fig. 4.2, Zahina and Richardson 1997). Again we employ these traditionally used terms to describe the plant communities and habitats but, as noted earlier, classify the Everglades as a fen (Chap. 2). Distinct quantitative differences in soil surface elevations exist among plant communities and result from temporal, generally cyclic variations of the hydrologic regime. Overlap often occurs between these communities, and in some situations it is difficult to categorize a location as one specific type; however, biomass can differ greatly (Zahina and Richardson 1997). An analysis of above- and belowground plant biomass in various habitats in WCA-2A and WCA-3A shows that a large range in the amount of biomass exists in various community types (Table 4.3). The deep water and enriched marsh habitats have nearly double the amount of total standing biomass (above- and belowground) as that of the unenriched marsh. Enrichment of the marsh with phosphorus increases the plant marsh biomass to levels found in the most productive deep water habitats. Importantly, the enriched community also has nearly a 30% increase in belowground biomass as compared to the unenriched marsh. The sloughs have the lowest amount of biomass per unit area of all the habitats. This may be in part because they are dominated by algal productivity and that while annual production is high, standing crop at any one time is low.

Fig. 4.2 Major plant communities of WCA-2A, WCA-2B, and WCA-3A

84

C.J. Richardson et al.

Table 4.3 Examples of plant biomass in various habitats in WCA-2A and WCA-3A (J. Vymazal and C.J. Richardson, unpublished data) Biomass (g DM m−2) Plant community

Aboveground

Belowground

Total

N

Marsh (unenricheda) 1,851 2,011 3,862 14 2,171 2,881 5,053 12 Marsh (enricheda) Slough 357 613 970 6 Deep water 3,069 3,253 6,322 5 Flats 1,294 1,737 3,031 7 Major plant components for each plant community are unenriched marsh: Cladium jamaicense; enriched marsh: Cladium jamaicense, Typha domingensis, Mikania scandens; slough: Eleocharis cellulosa, E. elongata, Nymphaea odorata, Chara sp., Utricularia spp.; deep water: Phragmites australis, Typha domingensis, Polygonum densiflorum, Eichhornia crassipes; and flats: Panicum hemitomon, P. repens, Pontederia cordata, Sagittaria lancifolia, Rhynchospora tracyi a Unenriched sites <600 mg P kg−1, enriched sites >600 mg P kg−1 in the top 10 cm of soil

4.3.1

Deep Water

Deep water is found in and adjacent to canals (Fig. 4.2). This human-made habitat is usually parallel to the levee (except for the stub canal in western WCA-2A). Most are peripheral to the WCAs. This habitat never dries out and usually maintains an average water depth greater than 0.75 m. Vegetation is most abundant along the canal slopes, rather than along the bottom of the water body (Zahina and Richardson 1997). Vegetation found in this habitat reflects the trophic status of water (in most cases high nutrient concentrations) as well as hydrologic conditions (faster water movement, especially adjacent to water gates) and includes many species of emergent, submerged, and free-floating macrophytes. A complete listing of the species we found in WCA-2A and WCA-3A during our 12 years of surveys is given in Appendix 4.1. Commonly occurring species include: Emergent species. Phragmites australis (Cav.) Trin. ex Steud. – Common reed, Colocasia esculenta (L.) Schott – Wild taro, Zizaniopsis miliacea (Michx.) Doell & Asch. – Southern wild rice, Typha domingensis, Polygonum densiflorum Meisn. – Smartweed, Polygonum hydropiperoides Michx. – Wild water pepper Submerged species. Hydrilla verticillata (L.) – Hydrilla, Vallisneria americana Michx. – Tape grass Free-floating species. Eichhornia crassipes (Mart.) Solms. – Water hyacinth, Pistia stratiotes L. – Water lettuce, Lemna valdiviana Phil. – Duckweed

4.3.2

Sloughs

Sloughs are the deepest natural habitats found in the WCAs. Sloughs dry out only during the most severe drought events. The dominant species in unenriched areas are

4 Vegetation and Algae of the Everglades Fen

85

Nymphaea odorata and Utricularia spp. Often, Eleocharis cellulosa, E. elongata, macroalga Chara spp., and seasonal calcareous algal mats are present. In areas that have been impacted by nutrients and altered hydroperiod conditions, other species usually occur (e.g., Typha domingensis). Sloughs are uncommon in the northern sections of WCA-2A and WCA-2B and common in the southern sections, mostly due to pooling of water in the southern ends of the WCAs (Zahina and Richardson 1997).

4.3.3

Flats

Flats are intermediate habitats between sloughs and marsh. The dominant macrophyte species are usually Eleocharis cellulosa and E. elongata, with abundant occurrence of species such as Panicum hemitomon, P. repens, Pontederia cordata, Peltandra virginica, Sagittaria lancifolia, Rhynchospora tracyi, or Hymenocallis latifolia. Eleocharis flats are prevalent in the south and central sections of the WCAs. However, E. elongata appears to be sensitive to nutrient enrichment and thus may be absent in nutrient-enriched areas (Vaithiyanathan and Richardson 1999). Flat communities become less common closer to the S-10 gate structure. The extensive Rhynchospora flats noted in earlier surveys (Davis 1943; Loveless 1959) are gone.

4.3.4

Marshes

Marsh habitat in unenriched areas is usually dominated by sawgrass (note: sawgrass is misnamed and is a true sedge), with other species present throughout the community matrix. These include, among others, Eleocharis cellulosa, Panicum hemitomon, and Sagittaria lancifolia. Salix caroliniana and Myrica cerifera become established less frequently. In the nutrient-enriched northeastern section of WCA-2A, a marsh dominated by Typha spp. has become established more recently and is interspersed with such plants as Acrostichum danaeifolium, Mikania scandens (L.) Wild. – Hempvine, Polygonum densiflorum, Rumex verticillatus L. – Swamp dock, and Sarcostemma clausum (Jacq.) Schult. – White vine. Several studies indicate that the invasion of Typha marsh in areas that were previously slough or a Cladium fen is a result of P-loading and hydrologic alterations (Urban et al. 1993; Craft et al. 1995; Richardson et al. 1999). In general, sawgrass is the dominant habitat in the unenriched northern parts of the WCAs.

4.3.5

Peat Batteries

Peat batteries are special habitats that are created when a portion of the peat soil becomes dislodged from a slough bottom and forms a floating mass. This natural phenomenon occurs in open water-unenriched sloughs. Recolonization is accomplished

86

C.J. Richardson et al.

by succession through a set of intermediate species (e.g., Amaranthus australis (A. Gray) Sauer. – Southern water hemp, Cyperus spp. – Flat sedges, Eleocharis flavescens (Poir.) Urban – Yellow spikerush, Polygonum spp. – Smartweeds, and Typha spp. – Cattails), usually resulting in a climax Cladium community or in the establishment of a tree island (Gunderson 1994; Zahina and Richardson 1997). Most of the intermediate species are found only in disturbed sites.

4.3.6

Tree Islands

Tree islands are found throughout both WCAs and are of three basic types depending on the dominant tree species: Melaleuca, willow, and bayheads. Melaleuca (Melaleuca quinquenervia (Cav.) Blake – Punk tree, Bottlebrush tree, Paperbark tree, or Cajeput) heads are most abundant throughout WCA-2B. These heads are generally monotypic communities, sometimes with a fern understory. Willow heads often contain varying mixture of Salix caroliniana and Myrica cerifera. Usually, these tree islands are small remnants of bayheads, which were common throughout the WCAs earlier in the century. Other species common in willow heads are Acrostichum danaeifolium (fern), Cephalanthus occidentalis (shrub), Ipomoea sagittata (vine), and Sarcostemma clausum (vine). Only a few bayheads persist within the extreme northwest section of WCA-2A today. Bayheads once were common throughout WCA-2 and were still visible on photographs from early 1950s. Persea borbonia, Ilex cassine, and Myrica cerifera dominate these communities. Other species common within the bayheads include numerous ferns such as Acrostichum danaeifolium, Blechnum serrulatum L.C. Rich – Swamp fern, Thelypteris spp. – Marsh ferns as well as Hibiscus grandiflorus Michx. – Swamp hibiscus, Big rose mallow, Ipomoea sagittata, Ludwigia spp. – Water primroses, and Sarcostemma clausum. Tree islands are still found in large numbers in WCA-3A and the ENP.

4.3.7

Transitional Levee

This habitat is a steep-sloped, elongated transitional zone between the deep water and dry levee communities. The dynamic habitat was created by the construction of the levee–canal system that surrounds most of the WCAs. It is characterized by widely fluctuating water levels and a high degree of disturbance. At times a portion of this zone may be completely inundated for up to several months and then it may change to dry ground in a matter of days. These fluctuations are a result of water management practices and storm events. Typical species include Annona glabra, Bacopa monnieri (L.) Pennell. – Smooth water hyssop, Cladium jamaicense, Eleocharis spp., Ludwigia repens Forst. – Red ludwigia, Nymphaea odorata, Phragmites australis, and Typha spp. The species present are often discretely stratified from drier to deeper zones. Thin peat soil may persist in the lower strata, but this suddenly gives way to a loose, sandy shell-rock substrate (Zahina and Richardson 1997).

4 Vegetation and Algae of the Everglades Fen

4.3.8

87

Dry Levee

Dry levee habitat is found on the dry slope embankments and atop the periphery levees of the WCAs. Unlike other habitats, there are no organic soils present. The substrate consists entirely of a well-drained, sandy shell-rock. The characteristic flora includes a high number of grass species (e.g., Andropogon glomeratus (Walter) BSP – Bushy broom grass, Andropogon virginicus L. – Broom grass, Aristida spiciformis Ell. – Bottlebush three-awngrass, Setaria geniculata (Lamm.) Beauv. – Knotroot bristlegrass) and Bidens alba (L.) DC. – White beggar-ticks, Sida spp., Passiflora suberosa, and Sarcostemma clausum. This community is also the most diverse in terms of species richness.

4.3.9

Epiphytic

Epiphytic habitats are restricted to tree islands or transitional levee areas where arborescent species have an opportunity to attain enough maturity to support the few species capable of a nonterrestrial lifestyle. Very few vascular epiphytes are found today in the WCAs, those being primarily of the family Bromeliaceae.

4.3.10

Summary of Taxonomic Groups

Because of its connection to the continental United States and close proximity to the Caribbean basin, the flora of the Everglades contains a mixture of both tropical and temperate species, as well as a number of endemic taxa. More than 330 species from 84 families have been collected by the Wetland Center Staff. Collection vouchers for most species are kept at the Duke University Wetland Center. Of the identified taxa, 11% are nonnative. The dominant families are the Poaceae (15%), while the Asteraceae comprised (11%), the Cyperaceae (10%), and the Fabaceae (4%). Surprisingly, the family distribution in WCA-2A is somewhat comparable with the ENP, which covers an area over ten times that of WCA-2A. The total area of the ENP is 20% larger than all the WCAs combined (4,363 km2 vs 3554 km2). Of the some 830 species enumerated within the park (Avery and Loope 1983), 17% of the total taxa are exotic species compared to our 11%. The dominant families were the Poaceae (12%), Asteraceae (9%), Fabaceae (7%), and Cyperaceae (5%) with the top 3 families being the same dominant families found in the WCAs.

4.4

Conclusions and Lessons for Restoration

The loss of several of the historical plant communities or landscape types in the drained Everglades presents a difficult challenge for future restoration efforts. For example, these losses make it very difficult under the current proposed hydrologic regimes to restore the swamp forest or maintain large areas of peripheral wet prairies

88

C.J. Richardson et al.

due to inadequate hydroperiod conditions for the dominant species in these community types. The pond apple community is also another area of concern for restoration and efforts are underway to determine if this can be restored, although the proper soil conditions probably no longer exist along Lake Okeechobee for its reestablishment in the original community location (Chap. 26). Also of great concern is the large increase of nonnative species found in the Everglades (see Appendix 4.1). To have more than 10% of the identified taxa as nonnative presents another challenge for restoring the native plant communities. Of specific concern is Hempvine (Mikania scandens), an invasive that is virtually blanketing plant communities of the northern areas of WCA-2A, including cattaildominated areas (Chap. 9). The Australian Melaleuca quinquenervia is also a major foreign invasive and in some areas covers nearly 100% of landscape. Brazilian Pepper (Schinus terebinthifolius) also has dominated the landscape and outcompeted native species in the more southern areas of the Everglades (the hole in the donut being one example of a massive invasion). The key to restoring the Everglades landscape and the diversity of plant communities will be in determining the hydrologic regimes necessary to maintain the dominant plant species in each community type. Unfortunately, little experimental information exists upon which to base the hydrologic conditions necessary to maintain most of the species, especially for those communities of higher plant diversity. It is clear that increased phosphorus additions will result in increased biomass, but mostly of invasive species (Chap. 9). Finally, the restoration of the plant communities will depend on a more complete ecological understanding of the successional dynamics of plant communities, peat formation, development of peat batteries and tree islands, effects of fire on formation of sloughs and deeper water habitats, and – most importantly – reestablishing the hydrodynamics of the Everglades peatlands (Chaps. 2 and 26).

Appendix 4.1 A listing of the plant species identified during the Duke University Wetland Center’s 12 years of community surveys primarily in Water Conservation Areas 2A, 2B, and 3A. The following plant list cannot be considered authoritatively comprehensive because an intensive survey was not possible for the entire area. The following list contains a representation of the major taxa that are found in the WCAs. Nonnative taxa are marked with an asterisk (*). Nomenclature and vernacular names follow USDA, NRCS (2007), The PLANTS Database, National Plant Data Center, Baton Rouge, LA, USA. Ferns Azollaceae Azolla caroliniana Willd. – Carolina mosquito fern Blechnaceae Blechnum serrulatum L.C. Rich. – toothed midsorus fern

4 Vegetation and Algae of the Everglades Fen

89

Osmundaceae Osmunda regalis L. var. spectabilis (Willd.) Gray – royal fern Polypodiaceae Phlebodium aureum (L.) J. Sm. – golden polypody Psilotaceae Psilotum nudum (L.) Beauv. – whisk fern Pteridaceae Acrostichum danaeifolium Langsd. & Fisch. – inland leather fern * Pteris vittata L. – ladder brake Salviniaceae * Salvinia minima Baker – water spangles Thelypteridaceae Thelypteris interrupta (Willd.) K. Iwats. – Willdenow’s maiden fern Thelypteris kunthii (Desv.) Morton – Kunth’s maiden fern Conifers Cupressaceae Taxodium ascendens Brongn. – pond cypress Taxodium distichum (L.) L.C. Rich. – bald cypress Monocots Alismataceae Sagittaria lancifolia L. – bluetongue arrowhead Sagittaria latifolia Willd. – broadleaf arrowhead Araceae * Colocasia esculenta (L.) Schott – coco yam Peltandra virginica (L.) Schott – green arrow arum Pistia stratiotes L. – water lettuce Arecaceae Sabal palmetto (Walt.) Lodd. ex J.A. & J.H. Schultes – cabbage palmetto Bromeliaceae Tillandsia balbisiana J.A. & J.H. Schultes – northern needleleaf Commelinaceae Commelina diffusa Burm. f. – climbing dayflower Cyperaceae Cladium mariscus (L.) Pohl ssp. jamaicense (Crantz) Kükenth – Jamaican swamp sawgrass Cyperus haspan L. – haspan flatsedge Cyperus odoratus L. – fragrant flatsedge Eleocharis cellulosa Torr. – Gulfcoast spikerush Eleocharis elongata Chapman – slim spikerush Eleocharis interstincta (Vahl) Roemer & J.A. Schultes – knotted spikerush Rhynchospora colorata (L.) H. Pfeiffer – starrush white-top

90

C.J. Richardson et al.

Schoenoplectus tabernaemontani (K.C. Gmel.) Palla – softstem bulrush Hydrocharitaceae Limnobium spongia (Bosc.) L.C. Rich. ex Steud. – American spongeplant * Hydrilla verticillata (L. f.) Royle – waterthyme Vallisneria americana Michx. – American eelgrass Lemnaceae Lemna minuta Kunth – least duckweed Spirodela polyrrhiza (L.) Schleid. – common duckmeat Wolffiella gladiata (Hegelm.) Hegelm. – Florida mudmidget Liliaceae Crinum americanum L. – seven sisters Poaceae Andropogon glomeratus (Walt.) B.S.P. – bushy bluestem Andropogon virginicus L. – broomsedge bluestem Cenchrus spinifex Cav. – coast sandspur Eragrostis elliottii S. Wats. – field lovegrass Heteropogon contortus (L.) Beauv. ex Roemer & J.A. Schultes – tanglehead * Melinis repens (Willd.) Zizka – rose Natal grass Panicum hemitomon J.A. Schultes – maidencane * Panicum repens L. – torpedo grass Phragmites australis (Cav.) Trin. ex Steud. – common reed Saccharum giganteum (Walt.) Pers. – sugarcane plumegrass Setaria parviflora (Poir.) Kerguélen – marsh bristlegrass Spartina alterniflora Loisel. – smooth cordgrass Zizaniopsis miliacea (Michx.) Doell & Aschers. – giant cutgrass Pontederiaceae * Eichhornia crassipes (Mart.) Solms – common water hyacinth Pontederia cordata L. – pickerelweed Typhaceae Typha domingensis Pers. – southern cattail Typha latifolia L. – broadleaf cattail Dicots Aceraceae Acer rubrum L. – red maple Amaranthaceae * Alternanthera philoxeroides (Mart.) Griseb. – alligatorweed Amaranthus australis (Gray) Sauer. – southern amaranth Anacardiaceae * Schinus terebinthifolius Raddi – Brazilian pepper Annonaceae Annona glabra L. – pond apple Apiaceae Cicuta maculata L. – spotted water hemlock

4 Vegetation and Algae of the Everglades Fen

91

Hydrocotyle umbellata L. – manyflower marshpennywort Aquifoliaceae Ilex cassine L. – dahoon Asclepiadaceae Funastrum clausum (Jacq.) Schlechter – white twinevine Asteraceae Ageratum conyzoides L. – tropical whiteweed Ambrosia artemisiifolia L. – annual ragweed Ampelaster carolinianus (Walt.) Nesom – climbing aster Baccharis glomeruliflora Pers. – silverling Bidens alba (L.) DC. var. radiata (Schultz-Bip.) Ballard ex T.E. Melchert. – romerillo Bidens laevis (L.) B.S.P. – smooth beggartick Cirsium horridulum Michx. – yellow thistle Conoclinium coelestinum (L.) DC. – blue mistflower Erechtites hieraciifolia (L.) Raf. ex DC. – American burnweed Eupatorium capillifolium (Lam.) Small – dogfennel Eupatorium serotinum Michx. – lateflowering thoroughwort Mikania scandens (L.) Willd. – climbing hempvine Pluchea odorata (L.) Cass. – sweetscent Solidago sempervirens L. – seaside goldenrod * Sonchus oleraceus L. – common sowthistle * Sphagneticola trilobata (L.) Pruski – Bay Biscayne creeping-oxeye Symphyotrichum elliottii (Torr. & Gray) Nesom – Elliott’s aster * Tridax procumbens L. – coatbuttons * Youngia japonica (L.) DC. – Oriental false hawksbeard Boraginaceae Heliotropium polyphyllum Lehm. – pineland heliotrope Brassicaceae Descurainia pinnata (Walt.) Britt. – western tansymustard Burseraceae Bursera simaruba (L.) Sarg. – gumbo limbo Caprifoliaceae Sambucus nigra L. canadensis (L.) R. Bolli – common elderberry Casuarinaceae * Casuarina equisetifolia L. – beach sheoak Ceratophyllaceae Ceratophyllum demersum L. – coon’s tail Chrysobalanaceae Chrysobalanus icaco L. – icaco coco plum Clusiaceae Hypericum hypericoides (L.) Crantz – St. Andrew’s cross Convolvulaceae Ipomoea alba L. – tropical white morning-glory Ipomoea sagittata Poir. – saltmarsh morning-glory

92

C.J. Richardson et al.

Curcubitaceae Melothria pendula L. – Guadeloupe cucumber Euphorbiaceae * Bischofia javanica Blume – Javanese bishopwood Chamaesyce hypericifolia (L.) Millsp. – graceful sandmat Chamaesyce maculata (L.) Small – spotted sandmat Euphorbia heterophylla L. – Mexican fireplant * Phyllanthus tenellus Roxb. – Mascarene Island leaf-flower * Ricinus communis L. – castorbean Fabaceae Acacia macracantha Humb. & Bonpl. ex Willd. – porknut Apios americana Medik. – groundnut Desmodium paniculatum (L.) DC. – panicleleaf ticktrefoil Senna obtusifolia (L.) Irwin & Barneby – Java-bean * Senna occidentalis (L.) Link – septicweed Vigna luteola (Jacq.) Benth. – hairypod cowpea Haloragaceae Proserpinaca palustris L. – marsh mermaidweed Lamiaceae Hyptis alata (Raf.) Shinners – clustered bushmint Trichostema dichotomum L. – forked bluecurls Lauraceae Persea borbonia (L.) Spreng. – red bay Lentibulariaceae Utricularia gibba L. – humped bladderwort Utricularia foliosa L. – leafy bladderwort Utricularia purpurea Walt. – eastern purple bladderwort Loganiaceae Mitreola petiolata (J.F. Gmel.) Torr. & Gray – lax hornpod Lythraceae Lythrum alatum Pursh – winged lythrum Malvaceae Hibiscus grandiflorus Michx. – swamp rose mallow Kosteletzkya virginica (L.) K. Presl ex Gray – Virginia saltmarsh mallow Sida cordifolia L. – llima Moraceae Ficus aurea Nutt. – Florida strangler fig Myricaceae Morella cerifera (L.) Small – wax myrtle Myrsinaceae Myrsine cubana A. DC. – Guianese colicwood Myrtaceae * Melaleuca quinquenervia (Cav.) Blake – punk tree * Psidium guajava L. – guava

4 Vegetation and Algae of the Everglades Fen

93

Nymphaeaceae Nuphar lutea (L.) Sm. ssp. advena (Ait.) Kartesz & Gandhi – yellow pond-lily Nymphaea odorata Ait. – American white water lily Onagraceae Ludwigia leptocarpa (Nutt.) Hara – anglestem primrose-willow * Ludwigia peruviana (L.) Hara – Peruvian primrose-willow Ludwigia repens J.R. Forst. – creeping primrose-willow Polygonaceae Polygonum glabrum Willd. – denseflower knotweed Polygonum hydropiperoides Michx. – swamp smartweed Polygonum punctatum Ell. – dotted smartweed Rumex verticillatus L. – swamp dock Rubiaceae Cephalanthus occidentalis L. – common buttonbush Salicaceae Salix caroliniana Michx. – coastal plain willow Saururaceae Saururus cernuus L. – lizard’s tail Scrophulariaceae Bacopa caroliniana (Walt.) B.L. Robins. – blue water hyssop Bacopa monnieri (L.) Pennell – herb of grace Solanaceae * Capsicum annuum L. – cayenne pepper Physalis pubescens L. – husk tomato Sterculiaceae Waltheria indica L. – uhaloa Ulmaceae Trema micranthum (L.) Blume – Jamaican nettletree Urticaceae Boehmeria cylindrica (L.) Sw. – smallspike false nettle Verbenaceae Callicarpa americana L. – American beautyberry Lantana camara L. – lantana Phyla nodiflora (L.) Greene – Turkey tangle fogfruit Vitaceae Parthenocissus quinquefolia (L.) Planch. – Virginia creeper Vitis rotundifolia Michx. – muscadine

Part II

Nutrient and Hydrologic Gradient Studies

5

Introduction to the Gradient Studies Curtis J. Richardson

5.1

Introduction

By 1989, the astounding expansion rate of cattail (Typha domingensis) due to phosphorus pollution was proclaimed to be the major environmental concern in the Everglades. The local media reported that 4 acres (1.6 ha) per day were being lost to cattail invasions and that native sawgrass (Cladium jamaicense) and open water sloughs would soon be gone due to eutrophication problems from upstream agriculture. With the loss of slough habitats would come a further decrease in habitat diversity and biodiversity. However, earlier ecological studies along phosphorus gradients in areas of WCA-2A (Gleason 1974b; Dineen 1974) had shown that ecological responses were also influenced by other ion additions such as Na, Ca, N, and S; thus, the shifts in species and changes in water and soil chemistry might not be totally attributed to P additions. Moreover, these additional ion inputs could confound any interpretation of biotic responses along the gradient. For this reason, Duke University researchers took special care to assess both nutrient enrichment gradients and unenriched areas when establishing baseline or reference conditions. To determine if the rates of cattail invasions were accurate and to assess the causes of this invasion, we established a series of permanent research sites to assess vegetation, soils, and water chemistry changes along known nutrient and hydrologic gradients in WCA-2A and WCA-3A. The long history of nutrient enrichment in these Water Conservation Areas provided an opportunity to examine the longterm effects of nutrient inputs on soil and water chemistry (Chap. 6) and the resulting responses of macrophytes (Chap. 9), algae (Chap. 10), and invertebrates (Chap. 11). In addition, hydrologic conditions along the gradients had been greatly altered over the past 30 years (Chap. 8) and Everglades research had not adequately taken into account the effects of flow regime and water level changes on biota. To address this gap, we analyzed both the geologic settings of WCA-2A and the influence of its hydrologic control structure on water movement to establish historical flow patterns along the main nutrient gradients as well as determine an overall water budget (Chap. 7). Our historical analysis also showed that hydrology played an important role in the loss of tree islands in WCA-2A over the past 50 years (Chap. 8). We undertook historical paleoecological analysis of vegetation patterns along these 97

98

C.J. Richardson

transects to learn how vegetation patterns have changed over the past 1,300 years as well as to determine if vegetation patterns changed in conjunction with alterations along the gradient (Chap. 12). The annual rate of carbon cycling and export from the northern EAA to the southern ENP was quantified to determine the annual rate of transfers of C and nutrients as dissolved organic matter (Chap. 13). These studies were focused on addressing the effects of increased nutrients and altered hydrology on the native plants and animals (Chap. 11 macroinvertebrates) of the Everglades. Importantly, our studies analyzed all the major nutrients and ions along the gradient (Chap. 6) in an attempt to determine how these inputs influenced community changes in conjunction with hydrologic shifts.

5.2

General Objectives

Our objectives with the gradient research were to (1) assess long-term soil and water gradients in both WCA-2A and WCA-3A, (2) quantify vegetation compositional changes in WCA-2A over a 12-year period along hydrologic and nutrient gradients, (3) determine how species changes were related to soil and water nutrients, hydrologic conditions, and fire, (4) evaluate plant community species composition, biomass, and nutrient storage throughout WCA-2A and WCA-3A from a series of field vegetation surveys, (5) determine algal and macroinvertebrate responses along a nutrient gradient, (6) measure changes in tree island patterns over the past 50 years in WCA-2A in relation to hydrologic alterations, and (7) assess and relate historical changes in nutrient status with changes in vegetation in WCA-2A.

5.3 5.3.1

Methods Location and Physical Layout of Gradient Sites

Water Conservation Area 2 is bounded on the northeast by both a dike and the Hillsboro Canal (Fig. 5.1). Our first permanent plots were established in WCA-2A in September 1989 along three transects ≈10 km in length south of the water control structures 10A, 10C, and 10D, which are periodically opened to release nutrientladen water in the fen of WCA-2A (Fig. 5.1). The location of the transects was first determined after a general helicopter survey of northern and central WCA-2A was done to establish the specific areas where cattail was dominant, transition zones, and areas of open water sloughs and sawgrass. A north-to-south vegetation gradient was found, and the transects were extended from pure cattail areas closest to the WCA-2A gates into areas 7–10 km south that were native sawgrass and sloughs. Six plots were established along each transect starting ≈ 0.75 km south of the control structures. The transects were roughly parallel to the dominant water flow from the drainage canal gates on the southern side of the Hillsboro Canal. Plots were

5 Introduction to the Gradient Studies

99

Fig. 5.1 Map showing the long-term nutrient gradient sites along 10A, 10C, 10D, and western (W) transects in WCA-2A adjacent to STA2. Also shown are central and eastern transects in WCA-3A, located south of STA3–4, as well as two southern locations, 3AS 1 and 3AS 2

located about 1.6 km apart along each transect (Table 5.1). The sites were permanently marked and the GPS locations recorded so that sampling could be conducted at the same location over the 12-year survey period. An additional gradient was established in 1998 in the unenriched part of western WCA-2A to establish baseline conditions for vegetation, biomass, nutrient stocks, soils, and water quality prior to the construction of Stormwater Treatment Areas STA2 and STA3–4 (Fig. 5.1 and Table 5.1). The permanent plots were established in a similar gradient pattern and with the same methods used in the 1989 plots. To assess changes in WCA-3A, we established two permanent transects in the eastern and central part of northern WCA-3A (Fig. 5.1, Table 5.1) in 1998 (prior to the extensive fire of 1999) Here ten plots extended >10 km into the interior of the fen south of Alligator Alley. Plots were analyzed for soil nutrients, water chemistry, and vegetation cover as well as biomass and nutrient stock at selected locations. A few additional plots were established in southern WCA-3A for biomass and nutrient stock estimates. Soils were also sampled in both WCA-2A and WCA-3A on a grid basis to determine total P concentrations for the conservation areas. Sampling was conducted in the early 1990s and in 2000. The dates and methods for these surveys are covered in detail in Chap. 6.

100

C.J. Richardson Table 5.1 Duke GPS research site locations for long-term monitoring in Water Conservation Areas 2A and 3A Sites Latitude (N) Longitude (W) Gradient 10A Gate location 10A1 10A2 10A3 10A4 10A5 10A6

26:21.56 26:20.62 26:19.68 26:18.70 26:17.60 26:16.70 26:15.69

80:18.78 80:18.49 80:18.46 80:18.53 80:18.40 80:18.53 80:18.45

26:22.27 26:21.51 26:20.40 26:19.53 26:18.55 26:17.50 26:16.60

80:21.13 80:21.25 80:21.33 80:21.41 80:21.44 80:21.41 80:21.41

26:23.32 26:22.50 26:21.60 26:20.58 26:19.55 26:18.60 26:17.52

80:22.91 80:22.98 80:22.93 80:22.82 80:22.85 80:22.92 80:22.87

Gradient 10C Gate location 10C1 10C2 10C3 10C4 10C5 10C6 Gradient 10D Gate location 10D1 10D2 10D3 10D4 10D5 10D6

Location western WCA-2A W2A 1 W2A 2 W2A 3 W2A 4 W2A 5 W2A 6

26:23.30 26:22.40 26:21.50 26:21.00 26:20.15 26:19.30

80:29.45 80:29.10 80:28.30 80:28.00 80:27.15 80:26.45

26:20.00 26:19.00 26:18.00 26:17.00 26:16.00 26:15.00 26:14.00 26:13.00 26:12.00 26:11.00

80:32.00 80:32.00 80:32.00 80:32.00 80:32.00 80:32.00 80:32.00 80:32.00 80:32.00 80:32.00

26:20.00 26:19.00 26:18.00

80:37.30 80:37.30 80:37.30

Location WCA-3A East transect 3E 1 3E 2 3E 3 3E 4 3E 5 3E 6 3E 7 3E 8 3E 9 3E 10 Central transect 3C 1 3C 2 3C 3

(continued)

5 Introduction to the Gradient Studies

101

Table 5.1 (continued) Sites Latitude (N)

Longitude (W)

3C 4 3C 5 3C 6 3C 7 3C 8 3C 9 3C 10

26:17.00 26:16.00 26:15.00 26:14.00 26:13.00 26:12.00 26:11.00

80:37.30 80:37.30 80:37.30 80:37.30 80:37.30 80:37.30 80:37.30

26:07.33 25:46.51

80:38.00 80:44.07

South sites 3AS 1 3AS 2

5.4

Conclusions

The research conducted at the permanent stations along the gradients in both WCA2A and WCA-3A provided us with over a decade’s worth of response data that made it possible to assess rates of biotic change and determine the factors controlling that change. In addition, our research provided the magnitude of data needed to develop biotic thresholds. Importantly, years of studying changing nutrient and water conditions have given us insights into changes in biotic responses and biogeochemical rates, insights we hope will help in the restoration of the Everglades’s most disturbed areas. The following chapters in this section provide detailed information on studies along nutrient, hydrologic, and fire gradients in the Everglades.

6

Enrichment Gradients in WCA-2A and Northern WCA-3A: Water, Soil, Plant Biomass, and Nutrient Storage Responses Curtis J. Richardson, James W. Pahl, Jan Vymazal, Panchabi Vaithiyanathan, Robert G. Qualls, P.V. Sundareshwar, M. Lee Barber, and Jeffrey Johnson

6.1

Introduction

Nutrient gradients exist in most wetlands with concentrations being higher along the edge or near input structures (Keddy 2000; Richardson and Qian 1999). Water and elemental gradients have existed in the Everglades since its inception nearly 5,000 years ago according to paleoecological records (Chap. 12) and early records of flow and water chemistry from Lake Okeechobee (Davis 1943; Gleason and Stone 1994; Walker 1995; see Chap. 2). However, massive alterations in the Everglades drainage, starting in the late 1890s and culminating during the 1970s, resulted in nutrient, ion, and water flow gradients that have significantly altered plant and animal communities (Craft and Richardson 1993b; Davis 1994; Richardson and Vaithiyanathan 1995; van der Valk and Rosburg 1997; Miao and Sklar 1998; Richardson et al. 1999; King et al. 2004; see Chaps. 8–11). Plant production and storage of nutrients have also been greatly changed as a result of these anthropogenic inputs but few studies have quantified plant production and above- and belowground storage of nutrients along these enrichment gradients (Craft and Richardson 1993a,b; Miao and Sklar 1998). Moreover, the legacy of historical accumulation of phosphorus (P) in the northern Everglades may continue to impact ecosystem functioning (e.g., water column P levels), even after the elimination of the anthropogenic impact (i.e., external phosphorus loads). This is because P availability in wetland soils is controlled by complex in situ biotic and abiotic exchange processes and is greatly influenced by residual stocks of P in the soil (Richardson and Craft 1993). Owing to its solubility, phosphorus can accumulate in wetland soils and cycles readily between soils, water column, and plant components of the ecosystem. These processes dictate if a wetland acts as a source or sink for a nutrient. Thus, it becomes important to understand the magnitude and the rate of flux of P from impacted wetland soils to predict whether a desired management plan will result in a reduction or increase in surface water P concentrations. Surface waters flowing southward from Lake Okeechobee are likewise contaminated with additional agricultural chemicals such as excess sulfur (S) as they pass through the Everglades Agricultural Area (EAA) and into the northern portions of Water 103

104

C.J. Richardson et al.

Conservation Areas (WCAs) 2A and 3A (Barber 2003; SFWMD 2006). Elemental sulfur (S0) is widely applied within the EAA to acidify the circumneutral soils (acidity is produced as the sulfur is oxidized to sulfate), contributing to elevated levels of sulfate and sulfide in the surface and pore waters of the portions of the Everglades that receive agricultural runoff (Gilmour 2003). Because of the relationship between available sulfate and mercury (Hg), the conditions created by eutrophic runoff have direct bearing on Hg biogeochemistry in the Everglades system (Vaithiyanathan et al. 1996). Increases in cation (e.g., Ca2+, Na+, and Mg2+) and anion (e.g., Cl− and NO−3) concentrations have been documented throughout the greater Everglades ecosystem (SFWMD 2004, 2005, 2006). In particular, concerns have been raised that elevated calcium (Ca) and sodium (Na) have altered the historic distribution of algal populations (e.g., the relationship between diatom-dominated assemblages and the calcareous periphyton mats; see Chap. 12). Additionally, increased levels of heavy metals and pesticides have been found near input structures, and gradients of these chemicals have developed in portions of the Everglades (Barber 2003; SFWMD 2004, 2005, 2006). In this chapter we will: 1. Detail the methods used for water, soil, and plant elemental analyses reported in this volume as well as provide data quality objectives and method detection limits for parameters measured in the laboratory by the Duke University Wetland Center (DUWC). 2. Provide a summary of nutrients and ion gradients in water that were measured by DUWC personnel from 1990 to 2003 in WCA-2A and 1998–2002 in WCA-3A. 3. Assess P concentrations in soils along nutrient gradients and present a historical trend analysis. 4. Address the problem of efflux of P from enriched soils along the WCA-2A nutrient gradient. 5. Present gradient trends for Hg and some trace metals in WCA-2A. 6. Quantify the changes in plant biomass and nutrient storage along enrichment gradients.

6.2

Sampling Locations and Methods

6.2.1

Surface Water Quality Along the WCA-2A and WCA-3A Nutrient Enrichment Gradients

6.2.1.1

Site Establishment

Transects in WCA-2A and WCA-3A were established as described in Sect. 5.3 of Chap. 5. The three transects in eastern WCA-2A were established in late 1989 to allow us to quantify changes in the ecological condition of the WCA-2A wetlands

6 Enrichment Gradients in WCA-2A and Northern WCA-3A

105

resulting from inputs of eutrophic surface water from the Hillsboro Canal. The total sampling effort along gradients is shown in Table 1.1. The 1990–2003 sampling effort was concentrated on the 10C transect, which was sampled 50 times during the 13-year study. In comparison, the 10A transect was sampled 24 times from 1990 to 2003, and the 10D transect was sampled 20 times. In 1997, we established a fourth transect in western WCA-2A to help define temporal responses within what was presumed to be an area reflective of the “background condition” (Fig. 5.1) and to provide a reference point for comparison to the results from the 10A/C/D nutrient-enriched transects in eastern WCA-2A. This also provided a baseline prior to water discharges from the Stormwater Treatment Areas (STAs). Although the transect originated at the L-6 canal, which forms the northeastern border of WCA-2A, there was no gate structure to force L-6 water into WCA-2A, as with the 10A–D transects. This transect was sampled 13 times between 1997 and 2003, but the canal itself was not sampled until 2002. The two transects established in northern WCA-3A in 1998 (Fig. 5.1) were intended to determine if a significant gradient in surface water and soil enrichment existed from either the L-5 or L-35 canals or through the various water control structures that hydrologically link WCA-2A and WCA-2B to WCA-3A (Fig. 8.3). In contrast to the WCA-2A transects, the central (C) and eastern (E) transects in WCA-3A were set at a length of 16 km (approx. 10 miles). Although transects were oriented south from the northern border of WCA-3A (the L-5 levee), the L-5 canal itself was never sampled since no input structures flowed directly into the northern part of WCA-3A. The 3A-C transect was sampled 12 times between 1997 and 2003, while the E-transect was sampled nine times between 1998 and 2003.

6.2.1.2

Surface Water Sampling and Analysis Protocols

Surface water was sampled from an airboat using either a long-handled polyethylene dip bucket or a centrifugal pump. The precision, accuracy, and calibration range for each element is presented in Table 6.1 along with the MDL, PQL, and unit for each analysis. The EPA-approved methods for cations and anions along with the instruments used to complete the analyses are given in Table 6.2. The standard methods or EPA protocol is given for each element or assay. It should be noted that all phosphorus values were measured and reported as PO4-P. The terms orthophosphate (PO4) and soluble reactive phosphorus (SRP) are used the same throughout this volume to indicate where phosphate was added or measured. A detailed presentation of methods as well as procedures for all lab and field protocols using blanks, spikes, and standards are given in detail in our Quality Assurance Plan certified by the Florida Department of Environmental Protection (FDEP; License No. 950381). Grab samples for total nitrogen (N) and total P were placed into 50-ml tubes acidified with 9N H2SO4, samples for SRP were filtered through a 0.45-µm nylon filter (Nylon Acrodisc GF Syringe Filter, Pall Life Sciences, Ann Arbor, MI) into 50-ml tubes, samples for NH+4-N and (NO−3+ NO−2)-N were filtered through a 0.45-µm nylon filter into 50-ml tubes acidified with 9N H2SO4, and samples for Ca, Mg, Na,

106

C.J. Richardson et al.

Table 6.1 Data quality objectives for parameters measured in the laboratory by DUWC Method protocol

Technique

Parameter

EPA 310.1 Titrimetric Alkalinity EPA 350.1 Colorimetricc Ammonia EPA 353.2 Colorimetricd Nitrogen, total (as NO3 + NO2)e EPA 353.2 Colorimetricd Nitrate–nitrite SM 3500- SM 3111 Bf Calcium Ca B SM 3500- SM 3111 B Potassium KB SM 3500- SM 3111 B Magnesium Mg B SM 3500- SM 3111 B Sodium Na D SM 4500- Colorimetricg Phosphorus, PE dissolved reactive SM 4500- Colorimetricg Phosphorus, PE total

Precision Accuracy Calibration (%RSD)a (%R)b range MDL PQL Unit 10 10 10

90–110 90–110 90–110

pH 4–7 0–1 0–2

0.1 0.4 mg l−1 0.009 0.036 mg l−1 0.015 0.060 mg l−1

10 10

90–110 80–120

0–2 0–6

0.008 0.032 mg l−1 0.02 0.08 mg l−1

10

90–110

0–6

0.01 0.04

10

90–110

0–1.5

0.002 0.008 mg l−1

10

90–110

0–3

0.04 0.16

10

90–110

0–0.2

0.002 0.008 mg l−1

10

90–110g

0–0.2

0.002 0.008 mg l−1

mg l−1

mg l−1

Method detection limits are calculated using the US EPA method. Historically, the PQL is defined by DUWC as four times the MDL a Calculation of %RSD is based on duplicate sample readings b Calculation of %R is based on recoveries from certified reference materials c Automated phenate d Automated cadmium reduction e Determination of total nitrogen is surface and wastewater by alkaline digestion and flow injection analysis. For a detailed description, see Quality Assurance Plan (FDEP No. 950381) f SM 3111 B = direct air–acetylene atomic absorption spectrometry g Manual ascorbic acid reduction

and K were filtered through a 0.45-µm nylon filter into 50-ml tubes acidified with 8N HNO3. Relevant blanks were processed to control for potential sources of contamination. In Sect. 6.2.1.3, we discuss how comparison of the sample data with results from the blanks was done in conjunction with our QA/QC analysis for all data in this volume. To more accurately measure dissolved organic P (DOP) and particulate P (PP), we modified the standard approach of Wetzel and Likens (1990), which calculates the following DOP = TDP − SRP,

(6.1)

where TDP is the total dissolved P, and PP = total P − TDP.

(6.2)

6 Enrichment Gradients in WCA-2A and Northern WCA-3A

107

Table 6.2 Methods of analysis of water samples and soil extracts Parameter

Methodology

Orthophosphate (PO4-P, or soluble reactive phosphate (SRP) )

Ascorbic acid-molybdate blue method (Wetzel and Likens 1990) manual method, 5-cm cell on a Beckman spectrophotometer or Technicon TRAACS method Digestion of persulfate reagent alone followed by addition of sample and the ascorbic-molybdate reagent as above Persulfate digestion (Koroleff 1983) of filtered samples followed by SRP colormetric method Persulfate digestion of shaken unfiltered samples (Wetzel and Likens 1990) followed by SRP colormetric method TDP–(PO4-P) TP–TDP Autoanalyzer phenate method (Technicon TRAACS method) Autoanalyzer diazotization method (Technicon TRAACS method) Autoanalyzer cadmium reduction followed by diazotization of NO−2 for water samples (Technicon 1988). For high level extracts and digests: autoanalyzer hydrazine reduction followed by diazotization of NO−2 (Technicon TRAACS method 1988) (NOx-N)–(NO2-N) Persulfate digestion of filtered water samples (Koroleff 1983) followed by (NOx-N) method Same procedure as above using shaken unfiltered samples TDN–(NOx-N)–(NH4-N) TN–TDN Automated combustion on a Pt catalyst followed by infrared spectroscopy of the liberated CO2. Total inorganic carbon is subtracted to obtain DOC. Shimadzu TOC-700 analyzer Shimadzu TOC-700 analyzer, heating at over 100°C on acid medium and infrared spectroscopy of the liberated CO2 (Note: we assume that over 90% of inorganic C is in the HCO−3 form since the pH of most samples was generally between 7.4 and 8.6, but small percentages were H2CO3 or CO−32 depending on pH) Atomic absorption spectroscopy using a PerkinElmer Model 5100 PC A.A. spectrophotometer Ion chromatography using a Dionex Ion model 300 Chromatograph Ion-selective electrode (Orion Research) or ion chromatography

Orthophosphate in samples to be digested for TDP Total dissolved P (TDP) Total P in water (TP)

Dissolved organic phosphorus (DOP) Particulate phosphorus in water (PP) Ammonia (NH4-N) Nitrite (NO2-N) Nitrate + nitrite (NOx-N)

Nitrate (NO3-N) Total dissolved N (TDN) Total N in water samples (TN) Dissolved organic N (DON) Suspended particulate N (PN) Dissolved organic carbon (DOC)

Dissolved inorganic carbon (bicarbonate C)

Base cations (Ca, Na, Mg, K) Anions (Cl−, SO−42) Bromide (Br−)

108

C.J. Richardson et al.

However, there can be considerable error in the calculated difference when PO4 concentrations are high relative to either TDP or total P. We found increased sensitivity and reproducibility of DOP and particulate P measurements by adding the sample to a predigested persulfate reagent and then adding the color reagent, thus measuring PO4 while controlling for the matrix of the original sample. Total P was then measured as above.

6.2.1.3

Handling and Statistical Analysis of Water Quality Data

Data handling details can be a significant issue with environmental studies. For example, the analysis of samples from the “bottom” or tail end of gradients is typically complicated by concentrations at the low end of standard analytical equipment, and thus by high signal-to-noise ratios. It is therefore critical that such data are handled in an objective and responsible, but also importantly a rational, manner. Data from all individual surface water-sampling events (gradient and dosing studies) were first evaluated via the EPA’s “5× rule,” which states to “…consider site sample results as positive only if the concentration of the chemical in the site sample exceeds five times the maximum amount detected in any blank. Treat samples containing less than five times the amount in any blank as nondetects…” (US EPA 1989). Accordingly, data for all analytes and from each sampling event were compared to the relevant blanks from that event, and tallied as either the numeric value for samples exceeding five times the lowest blank, or assigned the status of “below detection limit” (BDL) for data below the five times threshold. Note that this differs from the criteria for database inclusion established by FDEP, but an assessment of differences revealed no significance in values under either method during comparison tests (Richardson et al. 2003). Our minimum detection limit (MDL) removal criteria followed the International Union of Pure and Applied Chemistry (IUPAC 1978), and we used Draper and Smith’s (1981) approach for calculation of fiducial limits as detailed in Richardson et al. (2003). Data analysis followed the guidelines established in US EPA (2000), which directs the handling of BDL data within a dataset depending upon the percentage of the entire dataset classified as censored data. Specifically, US EPA (2000) suggests that if a dataset has <15% BDL data, the BDLs can be substituted with a minimum value such as the MDL or (1/2) × MDL, while higher percentages of censored data require maximum likelihood estimations (MLE) or percentile expression. Considerations of the proper handling of BDL data, or censored data in general, have led to a large number of peer-reviewed papers in the recent past such as Sharma and Argawal (2003), which advocates the use of MLE. For purposes of this volume, we replaced all BDL identifiers in the gradient surface water dataset covering the period 1990–2003 with a value equal to an annualized MDL determined from that year’s QA/QC samples (blanks). From the resulting datasets, we calculated parameter-specific geometric means expressed as a function of the transect-specific station or dosing locations. Surface water data in this chapter were aggregated into three subsets, reflecting the overall management/restoration

6 Enrichment Gradients in WCA-2A and Northern WCA-3A

109

efforts applied to the overall northern Everglades system: samples collected Prior to Restoration (1990–1995); samples collected following the initial adoption of best management practices (BMPs) in the EAA (1996–2001); and samples collected after the construction and initial operation of the eastern group of STA wetlands on the perimeter of the WCA-2A and WCA-3A (2002–2003).

6.2.1.4

Examination of Landscape Effects on Surface Water Dissolved Oxygen

We also conducted a survey of surface water dissolved oxygen (DO) in the sloughs and sawgrass marshes in the oligotrophic areas of WCA-2A to characterize and evaluate differences in the DO concentration in the sloughs and sawgrass marshes. The information generated in this study will serve as a reference to understand the differences in the DO profiles between the open water sloughs and the emergent marsh environment. This baseline information would allow us to accurately evaluate the influence of P enrichment on the water column DO concentration in the impacted areas of Everglades. Hydrolab recorders were deployed in the field for 2 weeks and DO measurements were taken on an hourly basis. In this study we examined the DO profile in the sloughs and sawgrass marshes of oligotrophic areas of WCA-2A. Three sloughs and three sawgrass marshes were examined in this study. Hydrolabs were installed in the center of the slough and in the marshes approximately 3–6 m (10–20 ft.) from the slough margins at marsh 1 and marsh 2 locations and within 1.7 m (5 ft.) at the marsh 3 location. The macrophyte population at each study site was surveyed and the surface water samples were collected on the day of installation and on the day of retrieval and analyzed for nutrients.

6.3 6.3.1

Soil Chemistry Influence of Nutrient Enrichment on Phosphorus Fractionation of Soils from WCA-2A

Soil cores were collected from WCA-2A in June 1989 and 1990 to assess nutrient loading effects on P fractions from areas receiving agricultural drainage. Cores were taken along a gradient from the nutrient-impacted area in WCA-2A south through unimpacted areas. Twenty cores were collected from WCA-2A following the methods of Craft and Richardson (1993b) and Richardson and Craft (1993). The enriched area was dominated by cattail (Typha domingensis). Cores were collected from the oligotrophic sawgrass (Cladium jamaicense) plains and sloughs 10 km south of the nutrient-impacted area. Soil cores were sectioned into 1.5-cm depth increments and air-dried. Sections were weighed, ground with a

110

C.J. Richardson et al.

mortar and pestle, and sieved through a 2-mm mesh diameter screen. Soils were analyzed at each depth interval for 137Cs activity to estimate accretion rates. Cs137 was measured for 8–24 h using a high purity germanium gamma ray counter (EG&G Ortec, Oak Ridge, TN) coupled with a 2048 multichannel analyzer (The Nucleus, Inc., Oak Ridge, TN). Depth increments also were analyzed for bulk density, total N and P according to the methods described in Craft and Richardson (1993a,b). Where appropriate, analyses were expressed on an ovendried (75°C) weight basis. We used the Hedley et al. (1982) fractionation method omitting the initial resin extraction step (Qualls and Richardson 1995). Field moist samples were sequentially extracted for 16 h with 30 ml of (a) 0.5 M bicarbonate after chloroform treatment of the soil to lyse microbial cells, (b) 0.1 M NaOH, (c) 0.1 M NaOH after sonication, and (d) 1 M HCl, and finally the insoluble residual was digested in nitric–perchloric acid. A second parallel subsample extracted by bicarbonate after chloroform treatment was multiplied by a conversion factor of 2.38 after Walbridge (1991) to estimate microbial biomass P. Designation of each fraction as labile or refractory was based on tests by Hedley et al. (1982) and Chang and Stewart (1957).

6.3.2

Spatial Extent of Soil Physical Properties and Phosphorus Content in WCA-2A

This long-term study investigated the extent to which soil P concentrations along the DUWC gradients and in WCA-2A as a whole, measured in 1997 and 2001, changed from values from Reddy et al. (1991) investigations of the spatial patterns of P enrichment. To maximize between-study data comparability, this study sampled the same spatial grid of 74 sites on 7 different north–south transects spaced at approximately 3.2-km (2-mile) intervals (Fig. 6.1) as was used by Reddy et al. (1991), and all sampling similarly took place in a fen environment, avoiding tree islands and sloughs. In 1991, soil cores were taken to 40 cm by driving a 7.5-cm (inside diameter) PVC pipe into the peat and dividing the resulting core into 10-cm increments for composite analysis (Reddy et al. 1991). To determine bulk density, a subsample was removed from each composite section and weighed, then dried at 70°C for 48 h. Percent dry weight and water content were calculated from soil wet and dry weights. Dry bulk density of the soil was computed from percent dry weight of the soil subsample and wet weight and volume of the original soil core section. Percent dry weight was extrapolated from the subsamples to determine dry weight of the entire 10-cm depth increment of known volume. Bulk density was expressed as g dry weight per cm3 of soil. Our 1997 and 2001 studies used the same sampling technique, but cores were taken to a depth of 20 cm only. We analyzed bulk density by drying the entire upper 10 cm of the core at 70°C to a steady weight, and bulk density was computed based on the whole sample weight.

6 Enrichment Gradients in WCA-2A and Northern WCA-3A

111

Fig. 6.1 Locations from the Reddy et al. (1991) grid of 74 stations sampled by the DUWC in 1997 and 2001 for determining the spatial distribution of WCA-2A ecosystem properties

Our digestion methodology for determining total soil P in 1997 and 2001 differed from Reddy et al. (1991). Each entire oven-dried core was passed through a 0.5-mm sieve, ground, and mixed thoroughly. Approximately 130 mg of the ground, mixed, and sieved core was wet-ashed using HNO3 and HClO4 on a digestion block following the Canadian Society of Soil Science Method 23.2.3.3 for total P (CSSS 1993). U.S. National Bureau of Standards (USNBS) soil standards were used to determine digestion efficiency and accuracy. The concentration of P in the resulting solution

112

C.J. Richardson et al.

was measured spectrophotometrically by the Murphy–Riley phosphomolybdate blue complex reaction on a TRAACS 800 autoanalyzer. Isopleths of soil P concentrations were developed by using a kriging program (SURFER) following Vaithiyanathan and Richardson (1999).

6.3.3

Analysis of Trace Metals and Mercury in WCA-2A Soils Along a Phosphorus Enrichment Gradient

Surface soil samples were collected from the six fen stations along the 10C transect in WCA-2A in 1993 (Fig. 1.2) to measure accumulation rates of nutrients and heavy metals. Soil cores were obtained using a 56.25-cm2 inside area (7.5 cm per side square) by 60-cm deep stainless steel peat box corer. For trace metal and total P analysis, approximately 100 mg of the ground soil was digested in HNO3/HClO4 as described earlier. Concentrations of Cu and Zn were determined by flame atomic absorption spectrophotometry with a Perkin-Elmer Model 5100 PC atomic absorption spectrophotometer (Perkin-Elmer, Norwalk, CT), while Pb and Cd concentrations were determined using flameless atomic absorption using a Perkin-Elmer Model 5100 ZL spectrophotometer. Total C and N were measured in 10–15 mg samples using a Perkin-Elmer 2400 CHNS analyzer. Buffalo river sediment (NBS #2704) and peach leaves (NBS #1547) were employed as standards for C and N, respectively. The standard material pine needles (NBS #1575) and Buffalo river sediment were digested and analyzed for trace metals and P using the same method used for soil samples. The measured values were within 4% of the actual value. Additional soil cores were collected in 1996 and 1997 for Pb and Hg analysis (Barber 2003). In May 1996, single cores were collected from the WCA-2A 10C transect 1.4, 5.1, and 8.8 km south of the water control structure. In June 1997, nine cores were collected along the same transect from stations 3.5, 5.1, and 8.8 km south of 10C, and four cores were collected from a single site in WCA2B (Fig. 1.2). Vegetative cover at the sampling sites ranged from cattail (T. domingensis) monoculture at the northernmost site in WCA-2A, through mixed cattail–sawgrass (C. jamaicense L.) in the intermediately enriched areas, to sawgrass–slough communities in the south (Vaithiyanathan and Richardson 1999). The vegetation at the WCA-2B site was similar to the WCA-2A site located 8.8 km south of 10C. To minimize collection problems caused by macrophyte roots and rhizomes, cores were collected from areas only partially covered with emergent macrophytes. Soil cores were taken to 50 cm for Hg analysis using an aluminum cylinder 10.2 cm in diameter. Cores were capped at both ends with rubber stoppers and transported to the laboratory. Accumulations of carbonate floc material were separated from the underlying peat and frozen separately. The peat cores were sectioned into 15-cm increments, double bagged in polyethylene zip-closure bags, and frozen. For Hg analysis, each core segment was thawed, sectioned into 1-cm increments, homogenized, and frozen in acid-washed glass jars with Teflon-lined lids until

6 Enrichment Gradients in WCA-2A and Northern WCA-3A

113

analysis. A full description of Hg analytical methods and quantification standards are available in Barber (2003).

6.3.4

Investigation of Soil Efflux Along the WCA-2A C Gradient

This study was initiated to characterize the potential for soil-to-surface water efflux of P in an environment of improved water quality of surface water inputs into WCA-2A. We specifically incubated intact soil cores collected along the 10C transect gradient to quantify the magnitude of internal P load from WCA-2A wetland soils to the water column. Cores were collected at DUWC stations C1, C3, and C6 (see Fig. 5.1) using a 14-cm diameter × 80-cm tall polycarbonate coring tube, driven into the soil to a depth of approximately 40 cm and taking appropriate precautions to minimize compaction. An inflatable bulb was used to seal the upper end of the tube and the core tubes were carefully pulled up. The bottom of the core tubes were sealed using rubber plumbing end caps (Fernco, Inc., Davison, MI). We also collected a triplicate set of smaller diameter cores from each location to estimate the total P in the soils prior to the commencement of our experiment. Cores were transported back to the DUWC field station in Loxahatchee, Florida and maintained at field conditions until the commencement of the experiment. The experiment was a fully randomized design with a 3 × 4 factorial treatment arrangement (station within the soil P gradient × treatment) with three replicates of each treatment combination. To control for the response of the water column itself to P dosing, a reference set of nine core tubes with no soil (three from each site) received only the three treatments of spiked surface water. Care was taken to isolate the sides of the soil cores from light while permitting the surface of the soil cores to be exposed to light by covering the soil core stands with a black plastic sheet. The cores were incubated at room temperature and the top ends of the core tubes were covered using plastic wrap to minimize evaporation. The overlying water column was maintained at a height in the core tubes of 32 cm, the approximate water depth in the field at the time of collection, equaling a volume of 5.35 l. Consistent with the recognized gradient in soil P enrichment in WCA-2A, our database on soil and surface water from these sites indicated that the total soil P was highest at the C1 location and lowest at the C6 location (Richardson et al. 2000), with the C3 location showing intermediate levels of soil P. For example, mean total soil P in the upper 10 cm of soil for samples collected between 1990 and 1998 ranged from 1,840 mg P kg−1 at C1 to 928 mg P kg−1 at C3 to 443 mg P kg−1 at C6 (Richardson et al. 2000b). Surface water total P concentrations also showed a similar trend (Vaithiyanathan and Richardson 1997a). The sediment and surface water profiles of nutrient parameters at C6 were representative of “background” or undisturbed regions of the northern Everglades. Hence, for the purpose of this experiment, surface water from C6 was used as background water for our treatments. Surface water from C6 equaled a “0” treatment P load scenario, although the actual concentration was 10 µg l−1. With regards to the other two treatments, the surface water

114

C.J. Richardson et al.

from C6 was spiked with a P concentration of 10 and 35 µg l−1 (total P concentration thus equaled 20 and 45 µg l−1) simulating a medium and high P load scenario, respectively. Phosphorus spikes were as Na2HPO4 and values reflect the mean filtered total P concentrations in the surface water at each location. The overlying water column in each core tube was sampled with replacement by withdrawing 160 ml of water using a syringe from the overlying water column. The sampling was set up such that a permanent “L”-shaped tube was placed to facilitate sampling of surface water at a constant distance from the sediment–water interface. This volume was replaced with 160 ml of corresponding background water depending on the treatment (background water used initially plus appropriate P additions). The samples were analyzed for total P and all applicable QA/QC protocols were followed as described earlier. To estimate the short-term efflux rates, we sampled the overlying water columns on day 0, 1, 2, 5, 12, and 15. Subsequently, this sampling was continued weekly for a period of 124 days. This provided an estimate of long-term P efflux rates from these soils and the duration needed for attaining equilibrium with the overlying water column. Phosphorus efflux rates were calculated after correcting for changes in total P concentrations in the water column in control (no soil added) core tubes. The linear portion of the dataset was regressed against the surface water TP concentration over the appropriate time period and multiplied by the ratio of surface water volume in the core tubes to the soil surface area of the soil core, to calculate the P efflux as follows J=

dC V , dt A

(6.3)

where J is the flux of total P (TP) (mg m−2 day−1), C the concentration of TP in the floodwater (mg l−1) corrected for changes in TP concentrations in surface water in control incubations, t the time (day), V the floodwater (surface water) volume (l), and A the soil surface area of soil cores (m2). Upon the completion of the water-sampling regime, the soil cores were extruded from the polycarbonate tubes and total P in soil samples measured in first 10 cm at 1-cm increments and next 40 cm at 5-cm increments.

6.4

6.4.1

Plant Biomass, Nutrient Storage, and Metal Storage Along Gradients Nutrient and Heavy Metal Content in WCA-2A and WCA-2B Plant Tissue

Concurrent with the soil chemistry sampling described in Sect. 6.3, six individual plants of C. jamaicense were randomly selected from each of the six fen stations along the 10-C transect. Above- and belowground plant materials were collected, transported back to the laboratory, and separated into live and dead leaf and root

6 Enrichment Gradients in WCA-2A and Northern WCA-3A

115

tissue. Plant tissue was air-dried and ground through a 2-mm mesh sieve. Heavy metal and nutrient content of plant tissues were determined as described earlier for soil samples.

6.4.2

Nutrient Content of Plant Tissues from Throughout the Northern Everglades

A total of 24 sampling sites in WCA-2A were selected in November 1998 (Fig. 5.1) for evaluation of sawgrass (C. jamaicense) and cattail (T. domingensis) above- and belowground biomass with respect to soil nutrient status. Similar samples were taken in October 2000 at selected locations in WCA-3A. The sampling sites were selected to cover the wide range of soil P concentrations that partially control the occurrence, biomass, and nutrient storage of macrophytes in WCA-2A (Craft and Richardson 1997; Vaithiyanathan and Richardson 1999; Doren et al. 1997). The sampling sites were located along the same four WCA-2A transects as described earlier. In addition, five sites were sampled for plant biomass (Fig. 5.1), three in northern WCA-3A along the C-transect and two located (S1 and S2) in the south. Aboveground biomass was harvested from the sampling sites by clipping four randomly selected 0.25-m2 quadrants. The aboveground vegetation was separated by species and by live and dead (standing litter) material. Belowground biomass was also sampled along the 10C and western transects in WCA-2A and from sites in WCA-3A. All plant material to a depth of 20 cm was recovered and carefully washed with water in the field and then with water in the laboratory. Plant material was dried at 75°C to a constant weight to estimate dry mass. Concentrations of N and P in plant biomass were determined using the same digestion and analysis methods as for soils noted earlier.

6.5

Results and Discussion

6.5.1

Surface Water Quality Along the WCA-2A and WCA-3A Nutrient Enrichment Gradients

6.5.1.1

Phosphorus

Excess phosphorus (P) is the primary pollutant of concern within the northern Everglades, due to the evidence that it fosters loss of ecosystem diversity at multiple trophic levels and in particular that it leads to the loss of the native ridge-and-slough ecosystem structure (Richardson et al. 1999; Childers et al. 2003) by providing a physiological advantage to T. domingensis and other species (Vaithiyanthan and Richardson 1999; Lorenzen et al. 2001; Kuhn et al. 2002; Lissner et al. 2003).

116

C.J. Richardson et al.

Thus, our initial focus on both surface water quality and soil nutrient status will be on the discussion of P patterns in those two compartments and P storage in plants. We will also present temporal and spatial patterns of other important nutrients and trace elements later in this chapter. Within WCA-2A, there were very strong gradients in both soluble reactive P (SRP) and total P in the 10C and 10D transects from 1990 to 1995, prior to the onset of any large-scale nutrient management efforts (Fig. 6.2). The differential loading of total P to the two transects, with typically higher P loadings through the 10D gate (Fig. 6.3), may explain the higher concentrations of total P and SRP along the D-transect. The 10D gate is “upstream” of the 10C gate with regards to surface water flows from the EAA through the Hillsboro Canal, and thus first receives water discharges from the canal (Fig. 8.3), and P concentrations along the Dtransect were accordingly very high. Waters in excess of 50 µg l−1 total P flowed from the 10D gate as far as 3.2 km downstream.

Fig. 6.2 Geometric means for surface water phosphate (as soluble reactive P; left) and total P (right) concentrations on the DUWC D- and C-transects in WCA-2A, segregated into three restoration management periods during the time period 1990–2003. The asymmetric confidence intervals shown were calculated from the log10-based SE

6 Enrichment Gradients in WCA-2A and Northern WCA-3A

117

Fig. 6.3 Estimated mean annual total P loads to WCA-2A through the 10C and 10D gates for the time period studied. Note that the water year (WY) load estimates do not correspond exactly to the geometric mean concentration data shown in other figures, which are based on calendar year groupings. Estimates of total P loads were made by applying geometric means for annual total P at gate inflow monitoring stations (FDEP 2003) to WY annualized flow estimates based on SFWMD DBHYDRO data accessed on 27 March 2007 (http://my.sfwmd.gov/portal/page?_ pageid=2235,4688582&_dad=portal&_schema=PORTAL)

Although efforts to develop BMPs in the EAA have been more successful than originally anticipated (Adorisio et al. 2006), the strength of the gradients following initial implementation of BMPs in the EAA unexpectedly increased along both transects (Fig. 6.2), most dramatically along the D-transect where values for both SRP and total P were in excess of 200 µg l−1 up to 1.5 km south of the gate structure. The more severe P gradient along both the C- and D-transects that developed during the 1996–2001 period of study appears to be partially independent of the gross loading of total P through the Hillsboro Canal (Fig. 6.3). Although P concentrations were likewise high downstream of the 10C gate structure, there was some suggestion within the C-transect that the fen itself may have been a source of P in addition to canal input, consistent with the theory of soil efflux discussed later in this chapter. However, we cannot also discount evaporative concentration effects on P concentrations following surface water movement from the canals to the fen, especially in the dry season. Values were significantly lower for both SRP and total P in both the D- and C-transects in the last time period examined (2002–2003), and there was no evident gradient in SRP in either the D- or C-transect, or in the D-transect for total P. There was a dampened total P gradient in the C-transect, where values dissipated gradually into the interior of the fen (Fig. 6.2, see also Fig. 2.17). Our measurements generally follow input trends reported by the SFWMD (2006) with lower 2002–2003 concentrations closely following the TP and SRP input values (Chap. 2).

118

C.J. Richardson et al.

Our sampling of the western transect in WCA-2A (Fig. 5.1) was more restrictive, with data available only from 1997 to 2003 (Fig. 6.4). Prior to 2002, concentrations of total P along the western transect were typically at or near the calculated MDL. In contrast to the lower P concentrations in the surface water of the D- and C-transects, values in surface waters along the western transect increased significantly between 2000 and 2002 nearer the L-6 canal, with mean total P remaining elevated in 2003. Interestingly, this increase was contemporaneous with the development of STA-2, which discharges to the L-6 canal at the G335 water control structure (Fig. 8.3). Flow from L-6 can enter western WCA-2A north of G335 through the G336 A–F structure cluster or the gapped perimeter levee south of the STA. The possibility thus arises that the operation of STA-2 may actually have resulted in an enrichment of the surface waters of western and central WCA-2A. Since our sampling ceased in spring 2003, we cannot comment on the persistence of this trend; however, P output concentrations increased in 2006 from this STA (see Table 2.2). In WCA-3A, there were no clear trends in SRP for either the C- or E-transects for either time period examined (Fig. 6.5). For the 1996–2001 time period, there was also no clear interior pattern in total P for either transect; however, total P was elevated in both transects at stations close to the L-5 canal when measured for the 2002–2003 time period. STA-3/4 was not online at this time, and there are no other overt factors which would explain this pattern, with the possible exception of higher total P surface water coming from the either the L-6 or L-18 canals through the S7 control structure. As with the WCA-2A western transect, though, it is uncertain if these elevated total P concentrations represent a temporal anomaly or a trend. It is clear, however, that with regards to WCA-2A and northern WCA-3A, high P inputs at control structures did result in elevated P gradients well into the interior of the WCAs that persisted long after inputs were lowered (see also Chap. 2; Fig. 2.17).

6.5.1.2

Nitrogen

The Everglades is typically not considered an N-limited system (Craft et al. 1995), although that may be an overly broad statement that may not hold for all ecosystem components (Amador and Jones 1993; Scott et al. 2005). Prior to the onset of nutrient management activities, there was a definable gradient in mean total N on the WCA-2A D-transect (Fig. 6.6), with a flatter but still obvious trend in the Ctransect. Those trends were magnified in the 1996–2001 sampling period, where

Fig. 6.4 Geometric means for total P in the DUWC western transect in WCA-2A, organized on an annual basis from 1997 to 2003 (excluding 2001) by distance downstream (km) from the L-6 canal. The black line in each graph indicates the year-specific DUWC Analytical Lab minimum detection limit (MDL) for Everglades total P and “MDL” indicates that all replicates were below detection limit and assigned a value equal to the MDL following QA/QC of the data (see text). “ND” indicates no samples were taken from a specific station in the year indicated. The asymmetric confidence intervals shown were calculated from the log10-based SE

6 Enrichment Gradients in WCA-2A and Northern WCA-3A

119

120

C.J. Richardson et al.

Fig. 6.5 Geometric means for surface water phosphate (as soluble reactive P; left) and total P (right) concentrations on the DUWC C- and E-transects in WCA-3A, segregated into two restoration management periods during the time period 1996–2003. The asymmetric confidence intervals shown were calculated from the log10-based SE

obvious gradients in total N were evident in both transects due to increases in total N close to the Hillsboro Canal. For the 2002–2003 post-STA sampling period, however, while there was still an evident total N gradient in the D-transect, the C-transect again showed a flatter total N distribution in surface waters. For the WCA-2A western transect, the temporal pattern of mean total N (Fig. 6.7) was opposite that seen for total P. Early in our sampling (1998–1999), there was a consistent pattern of increasing total N moving into the interior of the WCA. However, that gradient was not seen in our sampling from 2000 to 2003; primarily due to increases in total N at transect stations closer to and in the L-6 canal. South Florida Water Management District records for gross compartmentalization of total N in the northern Everglades generally show higher total N in inflows than in the interior of the WCAs (Payne et al. 2006). Without samples from the L-6 (0-km station) prior to 2002, we cannot say if the increases in total N close to the canal were related to the canal itself, or whether they may have been the result of surface water N mineralized from organic matter by fires. Likewise, the pattern of mean total N in the WCA-3A C-transect for the time period 1996–2001 suggested increasing total N in surface water moving toward the interior of northeastern WCA-3A (Fig. 6.8). In contrast, the pattern for 2002–2003 was an opposite gradient of decreasing total N moving south from the L-5 canal following large increases in total N concentrations in the two northernmost stations. Total N concentration values were nearly three times higher when measured in 2002 and 2003. This is mainly due to higher N input concentrations. However, it may be that the unusually high precipitation in south Florida in the latter half of the decade (Fig. 2.13) kept the upper reaches of the transects wet and prevented the release of

6 Enrichment Gradients in WCA-2A and Northern WCA-3A

121

Fig. 6.6 Geometric means for surface water total N concentrations on the DUWC D- and Ctransects in WCA-2A, segregated into three restoration management periods during the time period 1990–2003. The asymmetric confidence intervals shown were calculated from the log10based SE

N from the often burned soil. The north-central portion of WCA-3A is typically more mesic than the remainder of the WCA, and fen fires are common in that area. High total N concentrations in surface water can be expected under such circumstances (SFWMD 2005, 2006), as N mineralization has been shown to be higher under a more xeric environment associated with extended drawdowns and low water periods (Wilson et al. 1999). These patterns were not mirrored in the E-transect, where there were no trends for either sampling period.

122

C.J. Richardson et al.

Fig. 6.7 Geometric means for total N in the DUWC western transect in WCA-2A, organized on an annual basis from 1998 to 2003 (excluding 2001) by distance downstream (km) from the L-6 canal. “ND” indicates no samples were taken from a specific station in the year indicated

6.5.1.3

Ionic Chemical Species

Our sampling suggested that very little in the way of either a Na or Cl gradient existed along either the D- or C-transects in WCA-2A during the 1990–1995 sampling period (Fig. 6.9). However, significant gradients developed for both Na and

6 Enrichment Gradients in WCA-2A and Northern WCA-3A

123

Fig. 6.8 Geometric means for surface water total N concentrations on the DUWC C- and E-transects in WCA-3A, segregated into two restoration management periods during the time period 1996–2003. The asymmetric confidence intervals shown were calculated from the log10based SE

Cl as the study proceeded into the latter management periods, first showing in the C-transect during the 1996–2001 interval and in the D-transect during the last 2 years of our study, when the Na and Cl gradients on the C-transect had also become more well defined. Neither Na nor Cl showed any definable patterns in distribution along the WCA-2A western transect (data not shown). Values for all stations were between 110 and 160 mg l−1 for Na and between 135 and 250 mg l−1 for Cl. The patterns of Na and Cl in eastern WCA-2A surface water seen during the study may have been due to high Na and Cl from the canals being concentrated in the surface water of the fen due to evaporation. The source of at least some of the Na and Cl in WCA-2A is likely the release of connate seawater trapped in aquifers below WCA-2A following canal maintenance activities. The digging of wells and canals has led to the release of that connate seawater into surface waters (Parker et al. 1955; Chen et al. 2006), and in fact there is a significant connate aquifer below most of the EAA and western WCA-2A, with concentrations of Cl in the groundwater >100 mg l−1 (Parker et al. 1955).

124

C.J. Richardson et al.

Fig. 6.9 Geometric means for surface water sodium (left) and chloride (right) concentrations on the DUWC D- and C-transects in WCA-2A, segregated into three restoration management periods during the time period 1990–2003. The asymmetric confidence intervals shown were calculated from the log10-based SE

Interestingly, neither the L-5 canal nor the S7 gates appear to have been as significant a source of Na and Cl to northern WCA-3A as the EAA was to WCA-2A (Fig. 6.10), where there was the suggestion of an increasing gradient toward the interior in surface water Na in the C-transect in WCA-3A during the 2002–2003 sampling period. This pattern is typical of a fen, where groundwater fluxes influence surface water chemistry (Keddy 2000; Mitsch and Gosselink 2000). Patterns of Na and Cl in the E-transect were more complex. There, both Na and Cl increased to 5.7 km downstream, then subsequently decreased and remained relatively steady through 15 km downstream. It is possible that this pattern results from dilution from the 11A–C gate complex inflowing with the L-38 canal and WCA-2A surface water into northeastern WCA-3A. The exchange between Everglades groundwater and surface water has also been recently demonstrated (Wetzel et al. 2005; Harvey et al. 2006; Ritter and Muñoz-Carpena 2006), and perhaps we also observed a dilution of surface water with low-Na/Cl groundwater (Chap. 7). Calcium is considered an important cationic species to the Everglades ecosystem due to its ability to bond with phosphate in the surface water and subsequently precipitate onto the soil surface (Chap. 15). Its abundance in the system is due to the CaCO3 bedrock of the system and high concentrations in the connate groundwater,

6 Enrichment Gradients in WCA-2A and Northern WCA-3A

125

Fig. 6.10 Geometric means for surface water sodium (top) and chloride (bottom) concentrations on the DUWC C- and E-transects in WCA-3A, during the 2002–2003 sampling period. The asymmetric confidence intervals shown were calculated from the log10-based SE

and concentrations often exceed saturation values and have led to the development of calcareous periphyton mats (Chaps. 12 and 15). As with Na and Cl, mean Ca in surface waters demonstrated a similar development of definable gradients along both the D- and C-transects in WCA-2A over time in both transects (Fig. 6.11). In comparison, there were no significant trends in mean surface water Ca in the WCA-2A western transect (data not shown) or along either of the WCA-3A transects (Fig. 6.12). Overall surface water Ca concentrations in WCA-3A were consistent with those along the C- and D-transects in WCA-2A, but concentrations along the WCA-2A western transect were slightly lower, with station means between 55 and 85 mg l−1. Sulfate values were highly variable in both WCA-2A (Fig. 6.11) and WCA-3A (Fig. 6.12). In WCA-2A, sulfate in the D-transect surface water decreased toward the interior during the 1990–1995 sampling period, but there was an opposite gradient from 1996 to 2001, and significantly lower values and no apparent gradient during 2002–2003, when surface water concentrations decreased significantly to near MDL levels at most stations. In comparison, the C-transect showed no strong trends

126

C.J. Richardson et al.

Fig. 6.11 Geometric means for surface water calcium (left) and sulfate (right) concentrations on the DUWC D- and C-transects in WCA-2A, segregated into three restoration management periods during the time period 1990–2003. The asymmetric confidence intervals shown were calculated from the log10-based SE

for either the 1990–1995 or 2002–2003 sampling periods, but a significant gradient during the intervening 1996–2001 period was largely driven by high concentrations close to the 10C gate. Weaver et al. (2007) report that the arithmetic mean of sulfate concentrations in surface waters entering WCA-2A was 52 ± 1.4 mg l−1 for the 1978–2004 time period, illustrating that very high sulfate inputs did occur during our period of study. Of interest is the great decrease in sulfate in our C- and Dtransects during the 2002–2003 period as compared to earlier periods both at the input gate and in interior stations. It will be important to determine if this trend continues and, if so, whether it is due to agricultural BMPs, STA removal, or a short-term temporal response. Along the WCA-2A western transect, there was a degree of variability in surface water sulfate concentrations between 2002 and 2003, with significant increases in concentrations along all fen stations in 2003 as compared to 2002 levels, which were all at or near MDL status that year (Fig. 6.13). This was combined with a noticeable drop in concentration within the L-6 canal itself. Also, note that the concentrations of sulfate in the interior of western WCA-2A transect were consistent with, if not slightly higher than concentrations along the C- and D-transects to the east during the same 2002–2003 period.

6 Enrichment Gradients in WCA-2A and Northern WCA-3A

127

Fig. 6.12 Geometric means for surface water calcium (top) and sulfate (bottom) concentrations on the DUWC C- and E-transects in WCA-3A, during the 2002–2003 sampling period. The asymmetric confidence intervals shown were calculated from the log10-based SE

Fig. 6.13 Geometric means for surface water sulfate in the DUWC western transect in WCA-2A, organized on an annual basis from 2002 and 2003 by distance downstream (km) from the L-6 canal

128

C.J. Richardson et al.

In WCA-3A, there were opposite trends in surface water sulfate concentrations between the C- and E-transects (Fig. 6.12), largely mirroring the Na and Cl data described earlier. In general, surface water concentrations of sulfate in WCA-3A were much lower than those along the WCA-2A transects, although they were consistent with WCA-2A C- and D-transect concentrations for the 2002–2003 period. As with Na and Cl, Bates et al. (2002) implicated the canal system of the EAA as a major contributor of sulfate to the Everglades fens. Although they identified agricultural fertilizer as the major source, they could not rule out significant groundwater release, peat oxidation, and extra EAA sources. We measured silicon (Si) concentrations in the surface water of the WCA-2A C-transect from 1995 to 2000, and along the D-transect in 1997 (Fig. 6.14). In 1995, there was no consistent pattern in surface water Si concentrations in the C-transect. For the 1996–2000 period, however, a significant gradient developed along the C-transect, primarily due to a decrease in surface water concentrations toward the interior of WCA-2A. Silicon is an important element for benthic and pelagic diatom growth, and is also important for grasses and other macrophytes.

Fig. 6.14 Geometric means for surface water silicon concentrations on the DUWC D- and Ctransects in WCA-2A, segregated into two restoration management periods during the time period 1990–2001. The asymmetric confidence intervals shown were calculated from the log10-based SE

6 Enrichment Gradients in WCA-2A and Northern WCA-3A

6.5.1.4

129

Examination of Community Effects on Surface Water Dissolved Oxygen

The SFWMD and FDEP have utilized DO as a measure of ecosystem health (McCormick et al. 1997, 1999, 2000). In the Everglades a number of studies have examined the changes in the DO profiles along the P enrichment gradient in WCA2A (Belanger et al. 1989; McCormick et al. 1997). These studies have not focused on the influence of the plant communities (e.g., open water sloughs vs. sawgrass marsh, note here we use the terms slough and marsh to depict community types in the Everglades fen) on the water column DO concentration and have attributed the changes in the DO concentration almost exclusively to P enrichment. To examine the role of vegetation, we studied daily trends in DO oligotrophic areas of WCA-2A. The nutrient chemistry of the surface water is summarized in Table 6.3, while data from concurrently conducted macrophyte surveys are presented in Table 6.4. The DO concentration exhibited a strong diurnal variation in the sloughs and sawgrass marshes in the oligotrophic areas of WCA-2A (Figs. 6.15 and 6.16). The diurnal variation in the DO concentration in the sloughs varied from 2 mg l−1 in the evening to 10 mg l−1 during midday (Fig. 6.15). In the sawgrass marsh, the DO concentration ranged only from 1 to 6 mg l−1 (Fig. 6.16). The relatively high DO concentration in the sloughs reflects the intense photosynthetic activity of periphyton in these open water systems. The DO concentration appeared considerably diminished in the emergent sawgrass marshes reflecting the decrease in algal photosynthesis from shading effects. Values in the sawgrass decreased during the growing season except at marsh 3. The relatively high values of DO at the marsh 3 location were most likely related to the close proximity of the site to the slough margin (<1.5 m), but no reason is known for the late-season increase. The median values of DO in the sloughs (5.5–6.0 mg l−1) were nearly two to three times higher compared to the sawgrass marshes (2–3 mg l−1) (Fig. 6.17). The State of Florida’s criteria for oxygen are 5 mg l−1, a level that the sawgrass community would not often meet (SFWMD 2004, 2005, 2006). This finding is not atypical, in that wetland oxygen concentrations often do not meet state water standards that are usually developed for rivers and lakes. Table 6.3 Water column SRP and TP concentration in the sloughs and sawgrass marshes in the oligotrophic areas of WCA-2A (April–May 2000) TP (µg l−1) Site SRP (µg l−1) S1 1.8 ± 1.1 S2 1.8 ± 0.5 S3 1.8 ± 0.8 M1 1.6 ± 0.6 M2 2.2 ± 0.7 M3 2.2 ± 0.5 Slough 1.8 ± 0.8 Marsh 2.0 ± 0.6 Values shown are means ±1 SD

7.6 ± 2.1 6.2 ± 0.5 6.1 ± 1.0 7.3 ± 2.1 7.4 ± 1.3 8.3 ± 1.7 6.6 ± 1.2 7.7 ± 1.7

130

Table 6.4 Macrophyte species distribution in the sloughs and marshes of the oligotrophic areas in WCA-2A (April 2000) Periphyton Eleocharis cellu- Eleocharis elon- Utricularia Utricularia Utricularia Nymphaea Chara zey- Cladium jamaicense Location mat (% cover) losa (stems m−2) gate (stems m−2) purpurea (n) fibrosa (n) foliosa (n) odorata (n) lanica (n) (stems m−2) S1 S2 S3 M1 M2 M3

75 ± 26 80 ± 13 74 ± 22 0 0 0

8.7 ± 4.7 2.3 ± 2.5 14.0 ± 16.0 0 0 0

24.0 ± 15.7 12.7 ± 5.9 3.0 ± 5.2 0 0 0

1.3 ± 1.5 5.3 ± 4.0 2.0 ± 2.0 0 0 0

0 0 1.7 ± 2.9 0 0 0

0 0 0.3 ± 0.6 0 0 0

0 1.7 ± 1.5 0 0 0 0

0 0.7 ± 1.2 0.7 ± 1.2 0 0 0

0 0 0 196 ± 28 132 ± 14 164 ± 24

C.J. Richardson et al.

6 Enrichment Gradients in WCA-2A and Northern WCA-3A

131

Fig. 6.15 Diurnal variation in the dissolved oxygen (DO) concentration in the oligotrophic Everglades sloughs in WCA-2A

132

C.J. Richardson et al.

Fig. 6.16 Diurnal variation in the dissolved oxygen (DO) concentration in the oligotrophic Everglades sawgrass community

6 Enrichment Gradients in WCA-2A and Northern WCA-3A

133

Fig. 6.17 Box plots of dissolved oxygen (DO) concentration in the sloughs and the adjacent sawgrass community in the oligotrophic Everglades

Our results clearly show that major differences in the DO concentration between the sloughs and the emergent sawgrass must be taken into account to evaluate the natural differences between DO concentrations in wetland communities. However, it is also clear from our P dosing studies (Chap. 15) that increased P concentrations diminish water column oxygen concentrations and effect overall ecosystem net productivity even in open water systems. This trend was also found earlier along the WCA-2A P gradient by Rader and Richardson (1992).

6.6 6.6.1

Influence of Enrichment on Soils Fractionation of Soil Phosphorus in WCA-2A

To quantify and separate long-term soil P storage from short-term recycling processes, we measured analytically the various labile (bioavailable) and refractory (unavailable) soil P fractions along our study gradient to determine if enrichment altered the stored forms of P and their bioavailability (Fig. 6.18). In the enriched zone the P soil concentration in the top 5 cm was @1,500 mg kg−1, nearly three times that found in the oligotrophic zone. The refractory fractions made up 80% of the P stored in the enriched zone but only 58% in the oligotrophic area (Table 6.5). Long-term storage was mainly as humic organic P and insoluble P. Iron and Al controlled very little of the P as compared to >10% controlled by Ca precipitation in both the enriched and oligotrophic sites. In the oligotrophic portion of the Everglades gradient (7–11 km) labile fractions comprise @42% of the stored P but only

134

C.J. Richardson et al.

Fig. 6.18 Labile and refractory fractions of soil (0–5 cm depth) phosphorus along a eutrophication gradient in WCA-2A in the Everglades of Florida (redrawn from Richardson et al. 1997)

20% in the enriched zone. Thus, labile P fractions are decreased in the enriched P zone. Maintaining P availability is important to organisms in the oligotrophic zone since this area of the gradient has been shown to be P limited (Craft and Richardson 1993b). Using 137Cs-dating techniques, we estimated the average annual storage rate of each form of P in the enriched and oligotrophic areas over the past 26 years (Table 6.5). This revealed two important findings. First, microbial absorption accounted for 35% of the P sequestered in the oligotrophic zone but was reduced to 13% in the enriched cattail zone. Second, the average long-term total refractory P sequestered in the cattail zone (0.43 g m−2 year−1) was nearly eight times that stored in the oligotrophic area, but well below 1 g m−2 year−1. We conclude from the fractionation analyses that P enrichment greatly increases long-term total soil P storage up to approximately 0.5 g m−2 year−1, but not all of it is refractory. Also, P additions can significantly reduce the importance of short-term microbial P cycling in wetlands.

6.6.2

Historical Trends in Soil P Concentrations in WCA-2A and WCA-3A

The results of the 2001 study vs. the 1991 and 1997 studies are summarized in Table 6.6. Although there was a consistent increase in mean soil total P concentration of more than 22% during the course of the study (1991–2001), no significant

6 Enrichment Gradients in WCA-2A and Northern WCA-3A

135

Table 6.5 Average annual accretion rate of P fractions since 1964 (137Cs peak) in enriched and oligotrophic areas along a gradient in the Everglades (from Richardson et al. 1997a)

P fraction Labile (short-term storage) Bicarbonate extractable (organic P) Microbial biomass Exchangeable inorganic P Labile subtotal Refractory (long-term storage) Insoluble organic P Humic organic P Ca-bound P Occluded Fe/Albound inorganic P Surface Fe/Al-bound inorganic P Refractory subtotal Total

Dominant plant community Background (sawgrass/slough) Enricheda (cattail) % Above Cs Accretion rate Accretion rate (g m−2 year−1) peak (g m−2 year−1) % Above Cs peak

0.021 ± 0.006

3.8

0.003 ± 0.001

2.8

0.069 ± 0.014 0.017 ± 0.004

13.0 3.1

0.038 ± 0.010 0.005 ± 0.001

34.8 4.6

0.107 ± 0.020

19.9

0.046 ± 0.010

42.2

0.099 ± 0.029 0.221 ± 0.065 0.074 ± 0.022 0.020 ± 0.005

18.6 41.6 13.9 3.6

0.013 ± 0.005 0.027 ± 0.011 0.011 ± 0.004 0.006 ± 0.001

11.9 24.8 10.1 5.5

0.013 ± 0.003

2.4

0.006 ± 0.001

5.5

0.427 ± 0.120 0.530 ± 0.140

80.8 100.0

0.063 ± 0.020 0.110 ± 0.030

57.8 100.0

Accretion rate is calculated by summing the P content of each form down to the depth corresponding to the 137Cs peak and dividing by 26 years. The enriched areas and the oligotrophic areas are represented by the average rates at four stations a All accretion rates were significantly greater (p < 0.05 using t-tests) in the enriched area compared with the oligotrophic area. Standard deviations are in parentheses. The percentage of each P fraction is also shown

Table 6.6 Comparison of soil phosphorus and bulk densities observed in 1991, 1997, and 2001 in WCA-2A Bulk density (g cm−3) Soil total P content (µg cm−3) Year Soil total P (mg kg−1) 1991 671 ± 48 A 0.068 ± 0.003 B 45.4 ± 3.9 B 1997 716 ± 51 A 0.097 ± 0.003 A 65.9 ± 4.9 A 2001 823 ± 61 A 0.070 ± 0.003 B 52.2 ± 3.2 B p > F 0.122 All samples were collected from top 10-cm mineral soil. Entries are mean ±1 SE. Means with the same letter are not discernible at α = 0.05 based on Tukey criterion

differences were found among the three dates when all 74 sites were compared. In contrast, bulk density did show a difference with mean values in 1997 being much higher than in either 1991 or 2001. The grouped mean bulk density for all plots was 43% higher in 1997 than in 1991, which was statistically significant (a = 0.05),

136

Table 6.7 Comparison of bulk density and soil total phosphorus results from Reddy et al. (1991) and DUWC (2001) studies Bulk density (g cm−3) Soil total phosphorus (µg cm−3) Soil total phosphorus (mg kg−1) Transect

1991

2001

p>t

1991

2001

p>t

1991

2001

p>t

637 ± 127 833 ± 134 0.115 0.060 ± 0.005 0.051 ± 0.009 0.437 36.5 ± 5.3 42.1 ± 9.7 0.603 I (10A) 777 ± 144 879 ± 163 0.335 0.052 ± 0.003 0.054 ± 0.006 0.717 41.4 ± 8.9 49.7 ± 11.0 0.365 II (10C) III (10D) 1,021 ± 197 1,043 ± 231 0.652 0.058 ± 0.003 0.058 ± 0.006 0.859 55.5 ± 9.8 52.1 ± 10.8 0.748 482 ± 56 637 ± 55 0.063 0.059 ± 0.005 0.079 ± 0.008 0.022 26.1 ± 2.0 46.8 ± 4.8 <0.001 IV 499 ± 37 615 ± 73 0.127 0.075 ± 0.003 0.084 ± 0.006 0.132 37.5 ± 3.5 49.6 ± 5.4 0.006 V 571 ± 98 886 ± 257 0.093 0.082 ± 0.005 0.078 ± 0.009 0.727 49.2 ± 12.5 57.7 ± 6.8 0.395 VI 985 ± 156 1,262 ± 255 0.078 0.107 ± 0.014 0.075 ± 0.010 0.160 108.4 ± 24.2 82.8 ± 10.9 0.193 VII 671 ± 48 823 ± 61 <0.001 0.068 ± 0.003 0.070 ± 0.003 0.642 45.4 ± 3.9 52.2 ± 3.2 0.027 Overall A t-test was performed for paired samples for each transect and for WCA-2A overall sampling sites with significant differences shown in bold

C.J. Richardson et al.

6 Enrichment Gradients in WCA-2A and Northern WCA-3A

137

but 1991 and 2001 soil bulk densities were similar. The difference in bulk density in 1997 compared to the other two time periods resulted in 45% higher soil total P content found per unit volume of soil compared to that found in 1991. The reason for the higher bulk density in 1997 is unknown but may be related to greater soil compaction due to drier conditions, since sampling procedures were consistent during the study. Although soil total P per volume was 15% higher in 2001 than in 1991, statistically similar mean bulk densities for those years suggest that a direct comparison between 1991 and 2001 soil total P is reasonable. No significant differences were found along the individual system-wide WCA-wide transects (Fig. 6.1) in terms of total P concentrations and bulk densities except for bulk density along transect IV, which ran north–south through the center of WCA-2A (Table 6.7). A review of the individual transects indicates that on a weight basis 55 of the 74 sites showed a higher P concentration in the soil in 2001 than in 1991. On a volume basis only 11 sites were higher in 2001 and surprisingly most of the sites were not adjacent to the main inflow 10A/C/D structures where most of the P has historically entered. To assess where changes in P concentrations occurred, the mean soil TP on both a weight and volume basis was analyzed within four zones: highly enriched, intermediate enriched, moderately enriched, and oligotrophic zones. Figure 6.19 displays kriged mapping results from 1991, 1997, and 2001 in WCA-2A for TP soil concentrations on a volume basis, respectively. Only the oligotrophic zone showed a significant increase in TP per unit volume, while the moderate and oligotrophic zones demonstrated a significant increase in TP in terms of weight (data not shown). The TP per unit volume shows an overall increase over the 10-year period (1991–2001) (Fig. 6.19). The highly enriched areas seemed to have decreased slightly in some areas of WCA-2A, especially in the north and southwest. However, these data are highly influenced by the changes in the bulk density of the soil, and thus care must be taken when suggesting an increase or decrease in overall areas as modeled by kriging. In summary these data suggest that the areas that have received increased loadings in the past are not increasing

Fig. 6.19 Maps of soil total phosphorus concentration on volume basis in three annual surveys. Isopleths are derived from kriging

138

C.J. Richardson et al.

significantly in P but the oligotrophic areas are experiencing a residual movement of P into these areas from the past loadings, an important point since the current loadings have been significantly reduced to WCA-2A (Fig. 6.3). We found a pattern of declining soil total P with distance from the Hillsboro Canal (Fig. 6.20), where mean values 8.8 km down-gradient were below the 500 mg P kg−1 threshold (Chaps. 9 and 24). There were no corresponding gradients in either soil total N or total C along the same transect (Fig. 6.20). There were no significant N, P, or C gradients found along the western WCA-2A transect (data not shown). There was also a definable gradient of decreasing total P in soils of the WCA-3A transect when sampled in 1998–1999 (Fig. 6.21). The gradient was much steeper

Fig. 6.20 Total phosphorus (top), nitrogen (middle), and carbon (bottom) in soils sampled from DUWC stations along the WCA-2A 10C-transect. Values are means ±1 SE (sampled from 1990 to 2001)

6 Enrichment Gradients in WCA-2A and Northern WCA-3A

139

Fig. 6.21 Total P (by concentration, top), bulk density (middle), and volumetric total P content (bottom) in soils from the WCA-3A C- and E-transects sampled in 1998/1999. Shown are means ±1 SE

140

C.J. Richardson et al.

and shorter than in WCA-2A, reaching more consistent “background” levels 3.7 km inland from the northern border of the WCA-3A. There was a noticeable but transient increase in soil total P 7.3 km into the interior of WCA-3A on transect C, likely due to lateral P loadings from Miami Canal. The levee for Miami Canal in northern WCA-3A has numerous gaps, and as shown in Fig. 8.3, the S339 water control structure is intended to push water flows into the surrounding fens when closed (Lietz 2000). There were no clear trends in soil P along the WCA-3A E-transect (Fig. 6.21), consistent with our findings from water quality surveys described earlier. There were also no clear trends in either total N or total C for the WCA-3A transects (data not shown), similar to our findings for the WCA-2A 10C transect. Mean soil C along the E-transect was 465 mg kg−1, with a range from 443 to 486 mg kg−1, and mean soil N was 30 mg kg−1, with a range from 28 to 35 mg kg−1. These values are consistent with those determined for the 10C transect in WCA-2A (Fig. 6.20), suggesting that soil N and C content were largely unaffected by the documented surface water enrichment gradients.

6.6.3

Trace and Heavy Metal Responses

6.6.3.1

Copper, Zinc, Cadmium, and Lead

We first examined the copper (Cu) and zinc (Zn) distribution and accumulation rates in the alkaline peat soils of WCA-2A since it receives agricultural runoff from the EAA. Concentrations of Cu and Zn varied in the surface soils (0–5 cm) and were strongly correlated with soil P (r2 = 0.96 and 0.97 for Cu and Zn, respectively). Copper and Zn enrichment also varied with depth in the soil profile (Fig. 6.22a, b). Concentrations of 5–20 µg Cu g−1 and 10–40 µg Zn g−1 were relatively high in the upper layers of the soil profile and decreased with depth at all locations to values < 5 µg g−1. Copper concentrations were much higher at the enriched site (C1) than at the oligotrophic C6 site. Copper and Zn levels in the soils reported in this study are comparable to the results reported by Delfino et al. (1993) for the Everglades soils. The Cu enrichment extended to depths of 18 cm at the location nearest to the Hillsboro Canal (C1) but decreased at depths of < 5 cm at a distance of 8 km (C6). Pore water in the eutrophic soils ranged from 0 to 10 µg Cu l−1 and 5 to 25 µg Zn l−1. These metals probably originate in the EAA where both metals are applied as micronutrients (Sanchez 1990; SFWMD 1992; Coale 1994). Zinc may also enter the Everglades by atmospheric deposition from incineration facility emissions (Delfino et al. 1993). In contrast to Cu and Zn, lead (Pb) and cadmium (Cd) did not exhibit major increases from the enriched to the oligotrophic locations (Fig. 6.22c, d). In fact, lead shows a decrease in the upper profile at the enriched site. This may be due to leaching or more likely is a dilution factor due to increased peat accretion at this site. The presence of lead is likely linked to the combustion of lead-containing

6 Enrichment Gradients in WCA-2A and Northern WCA-3A

141

Fig. 6.22 Depth profiles for (a) copper, (b) zinc, (c) cadmium, and (d) lead in WCA-2A at the P-enriched 10C1 site and oligotrophic 10C6 site

fuels; however, confounding that assumption are observations that concentrations of 51 mg Pb g−1 and 33 mg Zn g−1 in peats measured after 1985 exceeded those measured pre-1985 (39 mg Pb g−1 and 15 mg Zn g−1; Delfino et al. 1993). Cadmium may also provide an indication of emissions from incineration facilities, but we found low-Cd concentrations in the soil samples with no distinct pattern and these concentrations were near our analytical detection limits. Recent metal accumulation rates were determined by comparing data with peat accumulation 137Cs dating (Table 6.8). These rates indicate increased accumulation rates for Cu, Zn, and – to some degree – Cd at the enriched site. These increased rates of metal accumulation at the enriched location may reflect increased anthropogenic loadings to the Everglades during the past 40 years. Lead showed the reverse trend with the oligotrophic site having nearly twice the accumulation rate. Although metal concentrations in the enriched area are not high enough to be toxic to plant species, they may have a significant impact on the food chain through bioaccumulation (Delfino et al. 1993). Phosphorus content in live leaf and root tissue of C. jamaicense in WCA-2A was highest at our C1 station (1,034 and 605 µg total P g−1, respectively) and decreased linearly with distance away from the inflow structures (Table 6.9). However, sawgrass plants growing in eutrophic soils did not accumulate more Cu and Zn than plants growing in oligotrophic soils (Table 6.9), and concentrations of Cu and Zn in sawgrass tissue were in the lower range of values reported for wetland plants from other ecosystems (Vymazal 1995). The fact that the native sawgrass does not take up more Cu and Zn at the P-enriched site when compared to that at the oligotrophic

142

C.J. Richardson et al.

Table 6.8 Metal accumulation rates in enriched and oligotrophic areas of WCA-2A Cadmium Copper (mg m−2 Lead (mg m−2 −2 −1 year−1) Zinc (mg m−2 year−1) Location (mg m year ) year−1) 10-C1, enriched 0.09 10-C6, oligotrophic 0.07

9.39 2.64

8.93 15.01

9.39 5.57

Table 6.9 Tissue P, Cu, and Zn concentrations in live leaf and root tissue of Cladium jamaicense along the N–S gradient in WCA-2A Root (µg g−1) Leaf (µg g−1) Site

D (km)

P

Cu

Zn

P

Cu

Zn

C1 C2 C4 C5 C6

1.4 3.5 6.9 8.8 10.5

1,034 928 475 325 272

0.83 0.73 0.07 1.03 0.54

6.2 7.9 3.9 8.2 9.3

605 389 433 112 113

0.54 1.18 0.58 1.60 2.68

10.0 7.6 9.1 12.7 20.6

sites suggests that trace metal accumulation in macrophytes is insignificant in the alkaline peat soils of the Everglades. Circumneutral pH of the soils (7.0–7.1) and the tendency of Cu and Zn to strongly bind with soil organic matter (Gambrell 1994) may have favored immobilization of these metals in the Everglades soils. Low availability of Cu and Zn in submerged organic soils has also been reported from plant nutrition work in croplands (Mikkelsen and Brandon 1975; Schueneman and Sanchez 1994). Our results do not support earlier studies (Delfino et al. 1993) that the emissions from incineration facilities are the primary source of Cu and Zn in the Everglades soils. Accumulation of Cu, Zn, and other trace metals from agricultural practices have also been reported for the agricultural soils from other parts of North America (Holmgren et al. 1993; Mermutt et al. 1996) and Europe (Angelone and Bini 1992). The WCA-2A in the northern Everglades may therefore not represent a unique situation since other wetlands receiving agricultural runoff show similar contamination with trace metals. Current trace metal contributions from agricultural sources to the Everglades need to be compared to urban and industrial sources and their significance and long-term impact needs to be verified.

6.6.3.2

Mercury

A mercury (Hg) contamination problem in south Florida was first recognized in the late 1980s with reports of elevated Hg concentrations in fish (> 0.5 µg g−1) and other wildlife (Ware et al. 1990). Although the highest levels of mercury were measured in fish from the Everglades ecosystem, this problem is widespread throughout the State of Florida. For example, concentrations in largemouth bass tissue were found

6 Enrichment Gradients in WCA-2A and Northern WCA-3A

143

to range from 0.04 to 1.33 µg Hg g−1 (Lange et al. 1993). This contamination prompted the state public health authorities to post-advisories limiting or banning consumption of edible freshwater fish in Everglades and Big Cypress National Preserve. It has also possibly adversely affected this region’s wildlife. Alterations in reproductive patterns and poisoning in wading birds have been attributed to Hg contamination (Frederick and Spalding 1994). The sources and processes responsible for Hg contamination have been recently attributed mostly to atmospheric deposition of global and local origins and the release of Hg stored in rapidly oxidized peat deposits in the EAA (Rood et al. 1995). Another plausible cause is the alteration of processes that control the amount of Hg that is available for accumulation by aquatic biota to favor formation of methyl mercury (MeHg). The in situ synthesis and degradation of MeHg are a key issue in any Hg contamination problem because Hg in freshwater fish occurs almost exclusively in the methylated form (Grieb et al. 1990; Bloom 1992). Microbial-mediated transformations are largely responsible for the methylation of Hg [II] (Gilmour et al. 1992) and degradation of MeHg (Oremland et al. 1991), but chemical methylation reactions are also known (Weber 1993). Changes in the ecosystem that directly or indirectly increase methylation or decrease MeHg degradation may result in an elevated concentration of MeHg, its biomagnification in the aquatic food web, and a public health problem. Lake acidification due to acid deposition (Winfrey and Rudd 1990; Gilmour and Henry 1991) and increased oxidation of organic matter in newly flooded hydroelectric reservoirs (Jackson 1988) are examples where ecosystem alterations may result in increased MeHg production. More recently Gilmour et al. (1998) and Cleckner et al. (1999) have shown that sulfate-reducing bacteria may be responsible for the increased methylation of Hg. As shown earlier in this chapter, sulfate enrichment also occurs in the northern portions of the Everglades, but input concentrations have decreased after STAs have come online (Figs. 6.11 and 6.12). To assess trends in mercury, we collected soil cores from three sites from the 10C transect in WCA-2A (C1, C3, and C6) in 1996 and found that Hg concentrations were most similar in the deepest portions of the cores (between 125 and 150 ng Hg g−1 at 15-cm depth, Fig. 6.23). Mercury concentrations measured at site C1 were relatively uniform between the surface and 15-cm depth, with the exception of a small spike in mercury concentration to 228 ng Hg g−1 between 5- and 7-cm depth. The soil core collected at C3 showed a trend of increasing Hg concentration with decreasing soil depth, but the highest concentration (305 ng Hg g−1) occurred between 4- and 6-cm depth. Similarly, mercury concentration increased with decreasing depth in the C6 core, reaching a maximum concentration of 380 ng Hg g−1 at the peat surface, but once again there was a spike in concentration within the core to 250 ng Hg g−1 between 4 and 5 cm. Cores collected from the single sampling site in WCA-2B showed clear differences from the cores collected along the 10C transect in WCA-2A (Fig. 6.24). The mean mercury concentration was still lowest in the deepest segments of the core, but unlike the cores from WCA-2A, this minimum was approximately 100 ng g−1, while the maximum total mercury concentration for WCA-2B cores was only

144

C.J. Richardson et al.

Fig. 6.23 Total Hg concentration (ng Hg g−1 dry soil) in cores (depth in cm) collected in 1996 from sites C1, C3, and C6 in WCA-2A (from Barber 2003)

200 ng Hg g−1 compared to the maximum of 300 ng Hg g−1 at C5. Further, Hg concentrations remained below 150 ng Hg g−1 to within 8 cm of the peat surface. In addition, the maximum Hg concentrations occurred just below the peat surface rather than deeper within the cores. Mercury, lead, and phosphorus accumulation rates were calculated using element concentrations in combination with soil bulk density measurements determined by companion studies (Craft and Richardson 1993a,b). Peat accretion rates used in the calculations were chosen to be representative of previously published values for soils collected from the same area as cores collected for this study (Table 6.10). Because the peat accretion rates used in these calculations (and those from the

6 Enrichment Gradients in WCA-2A and Northern WCA-3A

145

Fig. 6.24 Total Hg concentrations (ng Hg g−1 dry wt) in peat cores (depth in cm) collected in 1997 along the nutrient enrichment gradient in WCA-2A and in oligotrophic WCA-2B (from Barber 2003)

Table 6.10 Hg, Pb, and P accumulation rates in WCA-2A from 1964 to 1997 (±1 SE) WCA-2A sampling station Chemical species

C2

C3

C5

75 ± 1.0a 84 ± 3.9a 58 ± 6.4b Hg (µg m−2 year−1) Pb (mg m−2 year−1) 6.2 ± 0.8a 9.9 ± 0.9b 7.1 ± 0.2a,b P (mg m−2 year−1) 438 ± 99a 280 ± 49a,b 75 ± 17c Standard error estimates are based only upon variability in measured element concentrations. Values sharing superscripts are not significantly different from one another (ANOVA with Student–Newman–Keuls pairwise multiple comparisons analysis, p < 0.05, from Barber 2003)

Everglades literature in general) are determined by 137Cs dating, they represent an average yearly peat accretion rate for the period since 1964. For this reason, our calculated element accumulation rates have validity only when integrated over the same period of 33 years. Calculated elemental accumulation rates suggest that there are indeed spatial differences in the accumulation of Hg, Pb, and P in WCA-2A soils. Mercury accumulation rates at sites C2 and C3, respectively, were significantly higher in the northern portion of WCA-2A than those measured at site C5 further south (ANOVA, p ≤ 0.05; Table 6.10). The location of these sites also roughly corresponds to degree of P enrichment (Figs. 6.18 and 6.19). Phosphorus concentrations collected at the enriched sites (C2 and C3) were similar, but accumulation rates were significantly

146

C.J. Richardson et al.

greater among the C2 and C3 sites and the oligotrophic C5 site. The spatial differences in Pb accumulation along the nutrient enrichment gradient are less clear, with site C3 having the highest concentrations between the three sampling locations. A number of other studies have examined mercury inputs into WCA-2A. Stober et al. (1995) found that, while total mercury in canal waters decreased from north to south (mean = 1.74 ng Hg l−1, range 0.27–15.5 ng Hg l−1), total Hg in canal sediments were higher in the south than in the north (mean = 67 µg Hg kg−1, range 6.4–460 µg Hg kg−1). Rood et al. (1995) reported a mean accumulation rate of 59 µg Hg m−2 year−1 (range 35–95 µg Hg m−2 year−1) for the post-1985 period for WCA-2A. Further, inspection of the data from Delfino et al. (1993) on which the Rood et al. (1995) publication is based reveals that mercury accumulation rates were higher in the north of WCA-2A than in the south, at least to the extent a trend can be inferred from two cores collected at opposite ends of the nutrient enrichment gradient. Vaithiyanathan et al. (1996) reported a mean accumulation rate of 30 µg Hg m−2 year−1 in a single core collected 1.4 km south of the 10C water control structure, and 37 µg Hg m−2 year−1 for a second core collected 10.5 km south of 10C. The absolute Hg concentrations measured by Vaithiyanathan et al. (1996) are considerably less than those found by this study, but review of the soil digestion method used in that study (digest in 3:1 mixture of HCl and HNO3 at 95°C for 5 min followed by oxidation in 1.4% KMnO4 at 95°C for 45 min) suggests that their digestion of the highly organic Everglades peat may have been less than complete. The mercury accumulation rates calculated in this study are within the range presented by Rood et al. (1995). In summary our mercury accumulation data are derived from a sampling protocol that suggests that total mercury accumulation rates from sites highly impacted by agricultural runoff are higher than that for the less-impacted site further south. It is important to note that the Hg accumulation rates reported here are markedly higher than the atmospheric deposition of Hg reported for the Everglades region. A 4-year study of mercury deposition in South Florida reports bulk fluxes of 14–22 µg Hg m−2 year−1 for a series of eight sampling locations in and around the Everglades (Guentzel et al. 2001). The authors of this study further conclude that Hg deposition is relatively uniform across the South Florida landscape, and that the contributions of urban Hg sources are minor. The spatial trend in Hg accumulation reported here, along with the twofold to threefold difference between measured atmospheric deposition of Hg and accumulation rates in peat as determined by this study, and Rood et al. (1995) reveal a clear gap in our understanding of the processes governing Hg biogeochemistry and accumulation in the Everglades. If atmospheric Hg inputs are indeed uniform within the Everglades watershed, any spatial differences in Hg accumulation rates are due to factors other than atmospheric deposition. If atmospheric deposition of Hg is indeed occurring at the levels reported by Guentzel et al. (2001), there must be some other source entering the system. Kang et al. (2000) documented not only the accumulation of Hg in Florida Bay sediments to a level greater than what would be expected from atmospheric deposition alone, but also decreasing Hg accumulation with increasing distance from the freshwater inputs of Taylor and Shark Slough. Their conclusion is that freshwater runoff from the lower Everglades is a source of Hg to Florida Bay sediments.

6 Enrichment Gradients in WCA-2A and Northern WCA-3A

147

Downstream transport of Hg has also been documented for other systems (Balogh et al. 1997; Balogh et al. 1999; Sanudo-Wilhelmy and Gill 1999). The results of our study clearly show that Hg accumulation is greater in northern WCA-2A and that the areas where accumulation rates are highest are those most impacted by the agricultural runoff. In the absence of more data on the effects of nutrient enrichment and plant community structure on biogeochemical processes such as the binding/ scavenging of atmospheric Hg by plant leaves and the volatilization of gaseous Hg from soil and water, the conclusion that agricultural runoff is or has been a source of Hg to WCA-2A is a reasonable one. These findings suggest the need for further evaluation of drainage from the EAA as a historic or possibly current source of Hg to the Everglades ecosystem. The impact of treated waters released from STAs also needs to be factored into future assessments of Hg release. Other disturbed wetlands – particularly drained or partially drained peatlands – also need to be studied as potential sources of Hg to downstream systems. However, it should be noted that increased Hg concentrations in the soil have not been found to be strongly correlated with MeHg concentrations, a toxic form of Hg (Krabbenhoft et al. 2006a,b). However, in another study, Vaithiyanathan et al. (1996) tested the hypothesis that P eutrophication increased Hg methylation by obtaining total Hg concentrations and accumulation rates and estimating the potential for microbial methylation and methyl mercury degradation in peat soils collected along the P gradient in WCA-2. A negative correlation was observed between total Hg and P concentrations in soils (r2 = 0.64) and was explained by increased peat accretion rates in a nutrientenriched area (7.1–7.5 mm year−1) as compared to an oligotrophic area (1.92– 2.50 mm year−1), estimated using 210Pb and 137Cs dating. Total Hg accumulation rates (post-1964) were comparable for the enriched and oligotrophic sites (29–30 and 29–37 µg Hg m−2 year−1, respectively). Thus, calculations of total Hg accumulation rates are confounded by differences in peat accretion rates in the Everglades. Potential rates for both methylation (2.3–48.6 ng Hg g−1 day−1) and demethylation (6.5–113.2 ng Hg g−1 day−1) were higher in samples from WCA-2A than in samples collected in an area of the Everglades that had never been exposed to nutrients (Vaithiyanathan et al. 1996). However, trends suggesting the relationships of these activities to the P gradient in WCA-2A were not detected, and the ratio of methylation to demethylation (M/D) did not correlate with soil P concentrations. These results suggest that P eutrophication did not affect the potential for net methyl mercury formation in peat soils. Relating M/D to P concentrations further suggested that eutrophication had no effect on net MeHg formation (Fig. 6.25). However, eutrophication causes a variety of alterations that might have opposing effects on MeHg formation. For example, labile organic matter stimulates microbial activities but decreases the potential for methylation (Miskimmin et al. 1992) possibly by reducing Hg availability for methylation. Likewise, anaerobic conditions stimulate methylation when methylating bacterial populations are established, yet the increased production of S2− would inhibit methylation (Winfrey and Rudd 1990). Our results suggest that the effects of eutrophication on the variable factors and numerous interactions that influence the potential for MeHg production did not produce a clear trend to suggest a cause–effect

148

C.J. Richardson et al.

Fig. 6.25 Relationships between M/D and total P concentration in surface soils collected along a eutrophication gradient in WCA-2A. Ratios between the means of triplicate methylation (M) and demethylation (D) are reported for August 1993 samples (maximum variability was 41%) (redrawn from Vaithiyanthan et al. 1996)

relationship. However, Hg methylation and MeHg degradation might have been altered by exposure to other contaminants in agricultural drainage water and/or eutrophication as suggested by higher rates in WCA-2A as compared to other areas of the Everglades (Vaithiyanathan et al. 1996). A more detailed examination of the isolated processes that directly or indirectly influence methylation and demethylation is needed to delineate how eutrophication affects MeHg accumulation in the Everglades. This is especially true since sulfate enrichment has been linked to increased activity of sulfate-reducing bacteria, which may be responsible for the increased methylation of Hg (Gilmour et al. 1998; Cleckner et al. 1999). Interestingly, results of a pilot study in the canals of the Everglades ecosystem (Stober et al. 1995), as mentioned earlier, have revealed clear north–south gradients (high to low) in total P, total Hg, and MeHg in water and a gradient reversal from south to north for total Hg in sediments. It needs to be verified (from sedimentation rates in the canals) whether in situ Hg dilution due to increased accretion rates may account for the reported south–north gradient of decreasing total Hg concentrations in canal sediments. Recent results from the ACME study (Krabbenhoft et al. 2006a,b) have clearly shown links between MeHg abundance and several key ecosystem factors, including atmospheric Hg loading, hydroperiod maintenance, sulfate loading from EAA runoff, and dissolved organic carbon (DOC) levels in surface water. Of these factors, they suggest that only atmospheric Hg loading will not be affected by restoration efforts, and their results clearly show that decisions regarding possible water delivery and land use changes in the Everglades are equally, if not more, important in controlling MeHg levels now and in the future. In addition, Krabbenhoft et al. (2006a,b) point out that long-term records of MeHg

6 Enrichment Gradients in WCA-2A and Northern WCA-3A

149

from a site in central WCA-3A show a sharp decline in MeHg levels (> 90%) since about 2000, which was concurrent with a commensurate decline in sulfate inputs. These results challenge other recent studies concluding that declines in fish Hg levels are the result of reductions in Hg deposition. Controlled field dosing experiments have confirmed this observation, and question the assumption that this ecosystem is naturally high in MeHg, and suggest that changes to water quality and water flows from the restoration will have a great influence on MeHg exposure levels to indigenous wildlife and humans in south Florida. Finally, in a new summary of recent studies by the USGS, Krabbenhoft et al. (2006a,b) summarize key findings to date on Hg accumulation, causes for methylation and demethylation as well as the role of P and S in the production of MeHg. Although more research is ongoing and many questions remain, they report the following findings: (1) MeHg bioaccumulation is driven by internal MeHg production, mainly in surface sediments. (2) MeHg concentrations in all matrices (sediment, surface water, pore water, and biota) are maximal in the central Everglades (southern WCA-2A, WCA-2B, and north-central WCA-3). (3) MeHg is somewhat lower in more pristine areas like Everglades National Park and WCA-1, and it is much lower in the most eutrophic areas of WCA-2A and the constructed nutrient-retention wetlands. (4) The spatial MeHg pattern is not driven primarily by inorganic Hg concentration, although there is weak but significant relationship between Hg and MeHg concentrations in surface sediments. (5) Photochemical reduction and photodemethylation are important mechanisms for removal of mercury and destruction of MeHg, respectively, over much of the Everglades. (6) Sulfur inputs from areas north of the Everglades have a large impact on MeHg production, but the magnitude and even direction of the impact vary with the sulfate and sulfide concentration. (7) Phosphate and nitrate generally have no direct effect on MeHg production rates in sediment cores. (8) Anaerobic microbial processes, including sulfate reduction, are key components of microbial organic carbon decomposition in Everglades sediment. (9) Microbial dissimilatory sulfate reduction (rather than assimilation by plants) appears to be the most important mechanism for reduced sulfur storage in Everglades peat. (10) Natural fires and extended periods of peat exposure can greatly exacerbate MeHg production (for example 10× increases in sediment MeHg levels), and this phenomenon appears to be driven by sediment oxidation and release of sulfate after reinundation. (11) Bioaccumulation of MeHg in Gambusia appears to be facilitated by the movement of benthic invertebrates (insects and zooplankton) into the water column, and less importantly by direct grazing on surface sediments and periphyton.

150

C.J. Richardson et al.

(12) Methylation occurs only in periphyton “mats” where microbial sulfur cycling occurs, and is most common in the less-calcareous periphyton found in eutrophic areas. Clearly a number of major questions remain on factors controlling MeHg production and uptake and Hg inputs into the Everglades, but evidence of the important role of sulfate-reducing bacteria and the importance of sulfate additions to MeHg production in the system is unmistakably discernible from the USGS studies and our gradient research (Figs. 6.11 and 6.12).

6.7

Investigation of Soil Efflux Along the WCA-2A C Gradient

Under historically high total P loadings, the soils from Everglades WCA-2A sequestered P from the overlying water column (Richardson and Vaithiyanathan 1995), which resulted in a dramatic increase in P accumulation rates in soils downstream of the ten gate structures (Fig. 6.18). The rates of P accumulation in those soils increased from 0.06 g m−2 year−1 during the period 1900–1960 to 0.46 g m−2 year−1 during 1966–1990 (Craft and Richardson 1993a; Richardson et al. 1995; see Chap. 3). As shown in Sect. 6.6, the sites closest to the inflow structures sequestered the greatest amount of P from the surface water column, and a southward-oriented gradient of decreasing soil total P was established. Beginning in the 1990s, intensive efforts were made to reduce the concentrations of total P in the surface water inputs to WCA-2A from Hillsboro Canal (Fig. 6.3). Since sorption–desorption processes of soil P are a function of the concentration gradient between surface waters and the underlying soil, a decline in surface water P concentration at the most-enriched locations can potentially lead to efflux of P from the soil (Reddy and Rao 1983). To test if the enriched soils along the 10C gradient could transit from a net sink to an internal source of total P load to the surface water column upon reduction of external P loading, we developed an experiment with varying levels of surface water additions to gradient soils with varying soil P concentrations. There were marked differences in the rate or extent of total P efflux from the three locations along the P enrichment gradient (C1, C3, and C6), though patterns were generally consistent across the three treatments regardless of the P concentration of the overlying water column (Fig. 6.26). The highest rates of total P efflux to the water column occurred from the wetland soils that were collected from the most heavily impacted (C1) location, while the soils from the “background” location (C6) did not yield any measurable TP efflux under any of the surface water treatments. The soils from C3 generally supported total P efflux rates intermediate between the C1 and C6 soils. When incubated under a “background” surface water column, the short-term (15 day) mean TP efflux rate was calculated to be 6.96, 0.46, and 0.067 mg total P m−2 day−1 for cores from sites C1, C3, and C6, respectively. These rates agree well with published rates of TP efflux from similar locations along the WCA-2A transect

6 Enrichment Gradients in WCA-2A and Northern WCA-3A

151

Fig. 6.26 Changes in total P concentrations of surface water overlying soil cores taken from within the DUWC C-transect of the WCA-2A P enrichment gradient. Cores were incubated for 124 days with surface water from DUWC site C6 (top), C6 water augmented with 10 µg l−1 total P (middle), or C6 water augmented with 35 µg l−1 total P (bottom). Values shown are means (n = 3) ±1 SE

(Fisher and Reddy 2001). Interestingly, the efflux rates integrated over a longer period of time yielded similar estimates for the C1 location but doubled the efflux rates for the C3 location. Incubation of these wetland soil cores in background water yielded linear TP efflux rates of 5.98 mg total P m−2 day−1 for a period of 47 days for C1 cores and 1.26 mg total P m−2 day−1 for a period of 110 days for C3 cores. The C6 cores, as expected, did not show any discernable trend in TP efflux even after 124 days of incubation. This clearly demonstrates that the soils from the heavily impacted (C1) location tend to release a higher amount of TP and achieve a steady state more rapidly than moderately enriched (C3) soils, and that the C3 soils will require a longer duration to reach a steady-state condition, albeit yield lower amounts of TP than the C1 soils. A total P concentration of background + 10 µg l−1 reflects the ambient surface water concentration at the C3 location. Incubating soil cores from all three locations

152

C.J. Richardson et al.

with background + 10 µg total P l−1 resulted in efflux rates of 5.02 mg total P m−2 day−1 for C1 soils. In contrast, soils from C3 did not demonstrate a significant efflux of total P (0.35 mg m−2 day−1) even after 124 days of incubation. Interestingly, when incubated with background + 10 µg l−1 water, the oligotrophic C6 soils removed total P from the water column as indicated by a negative slope of the regression total P vs. time, consistent with the historic process of soil P accumulation that occurred under conditions of high surface water P concentrations prior to active water quality management. It should be noted that under these conditions the surface water overlying the highly enriched (C1) soils achieved a steady state of approximately 900 µg l−1 after 45 days, nearly identical in time and magnitude to the steady state that resulted under incubation with background water only. In contrast, under a surface water column of “background” water + 35 µg l−1, C1 cores exhibited an unusual biphasic pattern where the consistent 45-day plateau occurred at a much lower concentration than under the other two surface water treatments, but then increased from day 81 to 108 (Fig. 6.26). However, a linear regression fitted to the C1 flux curve including all the time points yields a good fit (r2 = 0.92, p < 0.0001, n = 18) and a total P flux rate of 2.98 mg m−2 day−1, a 51.5% decline in the C1 total P efflux rate from that under the “background” surface water. The total P efflux rate from the C3 cores was 1.24 mg m−2 day−1. Our results from the intact core incubation study clearly show that the rates of total P efflux differ among soils along the P enrichment gradient (Fig. 6.27). The highest rate of total P efflux was from the most P-enriched soils (C1 soils), followed by C3 soils, while the oligotrophic site did not show any discernible trends in total

Fig. 6.27 Changes in estimated P efflux rates from locations along a P gradient in WCA-2A. Estimates are based on intact soil core incubations as a function of initial surface water P concentrations

6 Enrichment Gradients in WCA-2A and Northern WCA-3A

153

P efflux. These rates were also sensitive to the initial surface water total P concentration in the overlying water column. For instance, raising the surface water total P concentration from background levels (“0” total P load) to background + 35 µg l−1 total P load decreased the total P efflux rate from the most-impacted soils from 5.98 to 2.9 mg m−2 day−1, a 51.5% drop. These rates can also provide an estimate of the duration of sustained total P efflux from the most heavily and medium-impacted soil cores. Since most of the externally loaded total P is sequestered in the top 0–20 cm depth fraction of the fen (Richardson and Vaithiyanathan 1995), for C1 cores we can use a conservative depth integrated (0–30 cm) soil total P concentration of 1,552 mg kg−1 and a bulk density of 0.049 g cm−3 to calculate the duration for which the total P efflux to the water column would be sustained. Assuming that approximately 25% of soil total P is mobile (Fisher and Reddy 2001), the three different total P efflux rate estimates from our experiments yield estimates of sustained total P efflux for a period of approximately 2.6, 3.1, and 5.2 years for background surface water total P concentration, background + 10 µg l−1, and background + 35 µg l−1, respectively. These rates are comparable to others published in the literature (e.g., Fisher and Reddy 2001). Similarly, for soils from the C3 site (medium impact), with an approximate soil total P concentration of 966 mg kg−1 (for top 30 cm) and a bulk density of 0.06 g cm−3, we estimate a duration of sustained total P efflux that ranges from 9.5 years (background water total P concentration) to 34 years with an initial total P concentration in the water column of background water + 10 µg l−1. It is interesting that the total P efflux rates from C3 soils were identical for incubations with background water and background water + 35 µg l−1 (Fig. 6.27). At an initial surface water total P concentration of background + 10 µg l−1, these soils exhibit a dramatic reduction in total P efflux rate which drops to 0.35 mg m−2 day−1 thus resulting in a long period of sustained TP efflux, but at a lower rate. The effects of this long-term release on the communities at this location are unknown. Importantly, these experiments were performed as intact core incubations. These studies constituted a “closed system” where the diffusive flux of total P was dictated by the concentration in the overlying water column at any given time. This, however, does not reflect the “free-flowing” surface water as is the case in the field, and there is a need to design an experimental approach to approximate natural conditions. Outputs from a preliminary simulation model (not shown) suggest that the duration of sustained total P efflux from these soils is extremely sensitive to surface water flow rates. Our preliminary model also suggests that the C1 soils will act as an internal P source for WCA-2A area and will lead to an increase in surface water and soil TP concentration at C3. Outputs also suggest that, depending on the flow rate of surface water, the soil and surface water TP concentration at C6 will also increase (owing to upstream inputs). Information on TP efflux under varying flow rates is essential to refine these scenarios and provides more precise information on the trend and the magnitude of total P efflux from impacted soils along the WCA-2A locations as influenced by water flow rates. Such an experiment would also unequivocally demonstrate if the soil total P gradient will migrate further downstream upon reduction in external

154

C.J. Richardson et al.

total P load to the WCA-2A in the Everglades. We also suggest that this be done in an area with a periphyton mat in place, since the effects of soil total P efflux on native communities under flowing water conditions are unknown. It is hypothesized that the presence of a periphyton mat will reduce water column total P and prevent downstream movement of P, and may also result in much lower total P values in the water column.

6.8

6.8.1

Plant Biomass and Nutrient Storage Along the C Nutrient Gradient and Control Sites in WCA-2A and WCA-3A WCA-2A Plant Biomass and Nutrient Storage

A biomass and nutrient study was completed along the WCA-2A 10C transect in 1998 to assess plant growth and plant nutrient storage capacity in relation to the nutrient gradients. The DUWC western transect in WCA-2A (Fig. 5.1) was also harvested to provide a baseline for nutrient and biomass conditions in areas with no direct canal hydrologic or nutrients inputs. Cattail (T. domingensis) biomass was dominant in the more eutrophic sites while sawgrass (C. jamaicense) was dominant in less-impacted or oligotrophic sites along the C-transect (Fig. 6.28, see Plate 6). Total aboveground biomass of two dominant species was inversely related along the C-transect. No sawgrass was found at C1 and cattail biomass declined along the transect and was not present at sites C5 and C6. Aboveground biomass was close

Fig. 6.28 Changes in aboveground biomass of Cladium jamaicense and Typha domingensis along the WCA-2A 10C-transect, measured in 1998

6 Enrichment Gradients in WCA-2A and Northern WCA-3A

155

to 2,000 g m−2 at all sites along the gradient, but at vegetation transition sites C2 and C3 biomass was split evenly between the two species. The relationship between soil (0–10 cm) total P and aboveground biomass of sawgrass and cattail displays a significant (p < 0.05) but different relationship between soil total P and total and live biomass (Fig. 6.29a, b). Cattail was positively correlated to soil total P while sawgrass was negatively correlated to soil TP. A stronger relationship existed for total aboveground biomass (i.e., live biomass + standing litter) than live biomass only. This finding was in agreement with frequency

Fig. 6.29 Relationship between soil total phosphorus (TP) and (a) total (live + standing litter) and (b) live aboveground biomass of sawgrass (Cladium jamaicense) and cattail (Typha domingensis) in WCA-2A

156

C.J. Richardson et al.

of occurrence of both species in WCA-2A (Craft and Richardson 1997; see Chap. 9). Craft and Richardson (1997) also reported that cattail was highly correlated with soil total P (R = 0.74) while percent frequency of sawgrass was highly but negatively correlated (R = −0.83). These findings also support previous studies, which indicate that cattail encroachment into sawgrass communities in northern WCA-2A is caused, to a large extent, by P enrichment of the peat (Urban et al. 1993; DeBusk et al. 1994). The fact that cattail is favored over sawgrass by elevated nutrients was also reported by Newman et al. (1996), DeBusk et al. (1994), Miao and Sklar (1998), and Doren et al. (1997). In contrast, several authors reported an increase of sawgrass biomass in relation to P additions or higher soil P concentrations. Chiang et al. (2000) reported up to three times higher aboveground sawgrass biomass after 2 years of P additions as compared to control stands. In years 3 and 4 of the experiment, aboveground biomass was three to seven times higher than controls. Moreover, the authors did not observe any encroachment of cattail into sawgrass stands that had been fertilized for 4 years and had no hydrologic alterations. Davis (1989) reported that sawgrass and cattail in enriched areas of WCA-2A had higher aboveground biomass than plants in oligotrophic areas. However, earlier greenhouse studies of less than 20 months did not show any effects of N and P additions on sawgrass aboveground biomass (Steward and Ornes 1975b; Sutter 1992). Taking into account long-term results from the Everglades, WCA-2A gradient studies and greenhouse research suggest that full cattail encroachment into healthy sawgrass stands with significantly increased biomass will require a matter of decade or more rather than several years, unless hydrologic shifts are sufficient enough to drown out sawgrass or some other disturbance such as fire decreases sawgrass densities to allow for cattail invasions (Toth 1987, 1988; see Chaps. 20 and 21). At all sites along the gradient, the live and standing dead aboveground biomass was quite similar and amounted to about 2,000 g m−2 (Fig. 6.30a). However, the amount of “old litter,” i.e., detached litter on the soil surface, differs along the gradients and was more than 2,000 g m−2 at the most oligotrophic sites C6. The biomass of old litter actually exceeded that of standing biomass at site C6, most probably due to slow decomposition process under oligotrophic water conditions. On the other hand, “old litter” biomass formed only about 30% of the total biomass at site C1 due to eutrophic conditions and faster decomposition (Qualls and Richardson 2000). Live and standing dead above- and belowground dry mass ground was similar all along the undisturbed western gradient (Figs. 5.1 and 6.31a) and surprisingly aboveground values were close to dry mass found along the nutrient-enriched C gradient (Fig. 6.30a). The major difference noted was in the amount of old litter that accumulated along most of the western gradient, often greater than 2,000 g m−2, the same as found at the oligotrophic plots at C6. This suggests that nutrient enrichment increases decomposition and reduces the amount of litter biomass that is available to transfer to the soil compartment. The biomass data also supported the macrophyte frequency study results (Chap. 9) that exotic invasive vines thrive in eutrophic sites. For example, Mikania scandens biomass comprised more than 15% of the total biomass at C1. Surprisingly,

6 Enrichment Gradients in WCA-2A and Northern WCA-3A

157

Fig. 6.30 Total above- and belowground biomass (g m−2, top), nitrogen (middle), and phosphorus (bottom) standing stock (both in g m−2) of dominant vegetation along the C-transect in WCA-2A. For plant composition, see Figs. 9.5 and 9.6

Mikania belowground biomass was nearly three to six times higher than aboveground biomass (e.g., 307 vs. 886 g m−2 at site C1 and 67 vs. 386 g m−2 at site C2). The total amount of belowground biomass found at the enriched sites C1, C2, and C3 is nearly double that found in the oligotrophic sites (C4–C6, Fig. 6.30a). Belowground biomass was nearly double live aboveground biomass at all sites. This increased biomass may be due to the increased amount of cattail root and rhizome biomass production found at C1–C3. Total above- and belowground cattail biomass including standing dead reached 4,258 g m−2 at C2.

158

C.J. Richardson et al.

Cattail rhizomes are more robust compared to sawgrass (Toth 1987, 1988; Miao and Sklar 1998). Nitrogen and phosphorus standing stocks (amounts of nutrients stored in plant compartments) in plants along the 10C transect reached their highest levels at site C2 with standing stocks at other sites being comparable (Fig. 6.30b) for aboveground N storage. Belowground standing nitrogen stock was higher at sites C1, C2, and C3 as compared to other three sites in less-enriched areas. Site C2 stored nearly 90 g m−2 of N (total of above- and belowground), nearly two to three times the amount of N stored at the C4–C6 sites. Phosphorus standing stock (Fig. 6.30c) also showed the highest value at site C2, with site C1 having somewhat comparable values. The P standing stock exhibited a decreasing storage trend toward the “background” sites at C5 and C6. Nearly two to four times as much P was stored belowground in the enriched sites as compared to the oligotrophic sites with P storage reaching 4 g m−2 at C2 (Fig. 6.30c). Live aboveground plants stored around 1 g m−2 at C1 and C2 but only a third of that by C6. Total cattail aboveground P storage (live 0.64 g m−2 + standing litter 0.15 g m−2) accounted for 0.79 g m−2 of P (»80% of aboveground live storage) at C2 while belowground storage was three times greater at 2.56 g m−2, resulting in 3.35 g m−2 of total P storage for cattail alone at this location. The largest standing stock of N along the oligotrophic western transect in WCA-2A was found in the old litter, which peaked at 17.9 g m−2 at site W1 (Fig. 6.31b). Belowground N storage averaged nearly 13 g m−2 with aboveground live averaging only 10 g m−2 of N. Nitrogen storage values on the western transect were similar to those found at the oligotrophic C4–C6 locations where total N stocks above- and belowground averaged around 35 g m−2. In the oligotrophic western transect, both live aboveground P and belowground plant storage were approximately 0.2 g m−2 each and the old litter storing an additional 0.1 g m−2, except at W1 (Fig. 6.31c). The average above- and belowground plant storage of 0.5 g m−2 P in undisturbed areas is nearly five times higher than the long-term soil annual P accretion rates for these oligotrophic areas (Chap. 3). Both nitrogen and phosphorus standing stocks exhibited the same trends with highest stocks at site W1, which receives slightly higher concentrations of nutrients since it is closest to the dike.

6.8.2

Plant Biomass and Nutrient Storage in WCA-3A

The biomass distribution for five locations (Fig. 5.1) in WCA-3A sites is shown in Fig. 6.32a. The total biomass was quite low at sites C1, C5, and C10 on our central transect. At sites S1 and S2, enormous amounts of old litter were present on the sediment floor. At site S2 more than 7,500 g dry mass m−2 was measured, while over 2,000 was present at S1 in the southern part of WCA-3A. These extremely high amounts of old litter in the southern two locations compared to the C-transect reflect the fire history of the region. The sites at the C-transect experienced a series

6 Enrichment Gradients in WCA-2A and Northern WCA-3A

159

Fig. 6.31 Total above- and belowground biomass (g m−2, top), nitrogen (middle), and phosphorus (bottom) standing stock (both in g m−2) of dominant vegetation along the western transect in WCA-2A. For plant composition, see Figs. 9.5 and 9.6

of major fires in the late 1990s due to the dryness of the northern part of WCA-3A compared to the southern region. Belowground biomass was near 2,000 g m−2 at all locations tested in WCA-3A, except at C1 where values were below 1,000 g m−2. These mass values were similar to those found at C4–C6 in WCA-2A but well below those found in enriched areas. The high amounts of old litter present at S1 and S2 increased N standing stock to nearly 60 g m−2 aboveground (Fig. 6.32b). It had less of an effect on P standing stock (Fig. 6.32c). Phosphorus standing stock at site C5 was comparable to S1 and was due to the presence of plants with high tissue

160

C.J. Richardson et al.

Fig. 6.32 Total above- and belowground biomass (g m−2) of dominant vegetation, standing dead, and litter (top), nitrogen (middle), and phosphorus (bottom) standing stock at select sites in WCA3A sampled in 1999

P concentrations (Sagittaria lancifolia, Polygonum sp., and Panicum hemitomon). Surprisingly, the WCA-3A sites had higher (sometimes more than double as at site C5 and S1) aboveground standing stocks of P than the oligotrophic areas of WCA2A. This may reflect the higher diversity of species with higher P uptake as noted in Chap. 9, as well as the amount of plant available P recycled in these areas of shallower peat soil (i.e., plant roots are closer to the limestone bedrock and marl), especially during drying and rewetting of soil and after fire.

6 Enrichment Gradients in WCA-2A and Northern WCA-3A

6.9 6.9.1

161

Conclusions and Lessons for Restoration Water Chemistry

The northern Everglades has been witness to significant changes in community structure during the past several decades, primarily defined as a loss of the native ridgeand-slough architecture and the attendant rapid expansion of cattail-dominated marsh over large areas of the northern landscape. General consensus among the scientific community is that these changes were due to a combination of altered hydrologic patterns resulting from flood control efforts implemented throughout the twentieth century and significant long-term nutrient and chemical inputs sourced in surface water runoff from the Everglades Agricultural Area, Lake Okeechobee bypass waters, and municipal inputs surrounding the greater Everglades ecosystem. This chapter described the collective efforts of a number of DUWC researchers over a 13-year period (1990–2003) to quantify the water and soil nutrient enrichment and chemical gradients that resulted from those inputs, and changes in gradients that developed in response to the adoption of corrective landscape BMP management practices and the operation of STAs that started in the mid-1990s and early in 2000s, respectively. This chapter also details our US EPA quantitative methods for general water and soil analyses and method detection limits used in all our studies in this volume (Tables 6.1 and 6.2). Our data screening analysis utilized the US EPA’s “5 × rule” (US EPA 1989) for initial identification and removal of potential outliers. The most significant nutrient of management and scientific interest in the northern Everglades has been phosphorus, owing to the oligotrophic nature of the native Everglades peatland ecosystem and the biogeochemistry of both peat-based wetland soils and an underlying bedrock of calcium carbonate that provides a chemical/biological mechanism for rapidly immobilizing available phosphate. Our efforts began in 1990 when the P enrichment gradient in eastern WCA-2A was already well developed from years of eutrophic surface water inputs from the Hillsboro Canal through the 10A/C/D water control structures (Fig. 6.2). We observed a significant P gradient along transects oriented southward from the 10C and 10D gates throughout the 1990s, even though significant reductions in surface water P outputs from the EAA occurred due to the very successful development of BMPs by agriculture in conjunction with researchers from the University of Florida. By the last 2 years of our sampling efforts, however, measured concentrations of P in the surface waters of those same transects showed dramatic decreases from those of the previous decade. This may have been due to the continued efforts to reduce the P loads associated with EAA surface water outputs, and was also likely a result of the sequential development and implementation of the STAs, agricultural lands converted to surface water treatment wetlands that promise to sequester significant amounts of P (and other nutrients and chemicals) before they can impact the ecology of the remnant Everglades ecosystem in the WCAs. Interestingly, although the combination of BMP and STA development appears to have reduced surface water P concentrations in WCA-2A inflows via the Hillsboro Canal, an

162

C.J. Richardson et al.

enrichment of western WCA-2A surface waters occurred at the end of our study (Fig. 6.4). Given the high concentrations of total P that we measured in the L-6 canal in 2002, it appears that this new enrichment trend may have been related to the onset of discharges from STA-2 into that canal, which forms the western boundary of WCA-2A. A phosphorus gradient also existed from the edge toward the interior for P in WCA-3A, although inputs were much lower. Nitrogen patterns mirrored those of P, both spatially and temporally, in the surface waters of WCA-2A, but not so clearly in WCA-3A. The clear gradient in WCA-2A might not be expected, as the Everglades is typically not considered an N-limited system, and N-fertilization in the EAA is likewise limited. However, the amount of DON added to the system from peat oxidation and loss from the EAA is extremely large (Qualls and Richardson 2003). Significant gradients developed for both Na and Cl as the study proceeded into the latter management periods. The gradient patterns of Na and Cl in eastern WCA-2A surface water may have been due to the release of high Na and Cl from connate seawater trapped in aquifers below WCA-2A following canal digging and maintenance activities. Calcium in surface waters demonstrated a similar development of definable gradients from the canal edge along both the D- and C-transects in WCA-2A over time in both transects, while no clear trends occurred in the other transects for these ions. Sulfate content in the northern Everglades surface waters was high and variable during our studies (Figs. 6.11–6.13), with very low concentrations in the 10C Dtransects in WCA-2A after STAs were operational as compared to earlier input concentrations. Lowering concentrations of sulfate inputs are important because sulfur enrichment in the WCAs resulting from EAA runoff has recently become a matter of concern for managers for two primary reasons. First, elevated sulfate in the surface waters overlying Everglades peat threatens to increase organic matter oxidation due to the highly energetic biogeochemical cycling between sulfate and sulfides, and could lead to elevated peat loss, as has been suggested for coastal systems (Avery et al. 2001). Second, and more important from a public health standpoint, elevated sulfate concentration in surface water, and corresponding elevated activities of sulfatereducing bacteria, have been related to increased occurrence of methyl mercury in the landscape (Gilmour et al. 1998; Krabbenhoft et al. 2006a,b). Importantly, major differences were found in the DO concentration between unenriched sloughs and the emergent sawgrass marshes, showing the importance of understanding that undisturbed communities like sawgrass can have DO values well below the state standard of 5 mg l−1, i.e., the median values of DO in the sloughs (5.5–6.0 mg l−1) were nearly two to three times higher compared to the sawgrass marshes (2–3 mg l−1) (Fig. 6.17).

6.9.2

Soil Gradients

Soil P concentration closely followed surface water patterns in WCA-2A, but N and C showed no gradient trend. Soil N patterns were more variable in WCA-3A

6 Enrichment Gradients in WCA-2A and Northern WCA-3A

163

and the western transect of WCA-2A, but this is not surprising. These regions, particularly the western edge of the WCA-2A western transect and the northern reaches of the WCA-3A transects, are more mesic than the fens in eastern WCA-2A, and soil N is probably more dependent on recent fire history. A soil P fractionation procedure along the 10C transect in WCA-2A showed that refractory fractions made up 80% of the P stored in the enriched zone but only 58% in the oligotrophic area. Long-term storage was mainly as humic organic P and insoluble P. Elevated Cu and Zn concentrations along the soil gradient south of the ten structures probably originate in the EAA where both metals are applied as micronutrients. In contrast to Cu and Zn, lead (Pb) and cadmium (Cd) did not exhibit major increases from the enriched to the oligotrophic locations. Sawgrass showed no increase in plant tissue uptake of Cu and Zn along the P gradient. Mercury accumulation data suggest that total mercury accumulation rates from sites highly impacted by agricultural runoff are higher than that for the less-impacted sites further south, although the reasons are not fully elucidated. The spatial trend in Hg accumulation reported here along with the twofold to threefold difference between measured atmospheric deposition of Hg and accumulation rates in peat suggests that any spatial differences in Hg accumulation rates are due to runoff and accelerated peat accretion rates rather than atmospheric deposition.

6.9.3

Vegetation Gradients

Cattail biomass was positively related to soil TP along the 10 C gradients in WCA-2A while sawgrass was negatively correlated. Belowground biomass was more than doubled in the enriched areas of the gradient as compared to unenriched areas. Of key importance to the phosphorus and N cycles is storage of more than four times the P (4 g m−2) and N (55 g m−2) in belowground standing stock in the enriched areas compared to unenriched areas. These trends were not found in WCA-3A gradients; however, extremely high amounts of old litter were present on the sediment floor in lower WCA-3A, and more than 7,500 g dry mass m−2 was measured compared to other sites to the north. These extremely high amounts of old litter in the southern two locations as compared to the C-transect reflect the fire history of the region. Surprisingly, the WCA-3A sites had higher aboveground standing stocks of P than the oligotrophic areas of WCA-2A, which may be due to higher plant diversity in those plots. Patterns in plant N mirrored those of P in WCA-2A (Figs. 6.30 and 6.31), but not in WCA-3A, where tissue N and P were largely independent. Observations of elevated total N content in the L-6 canal surface water during 2002–2003 are consistent with P patterns and suggest that surface water N inflows into western WCA-2A were contributory to vegetation nutrient uptake. Further studies need to be done in this region of the northern Everglades, to determine if our data points represent a trend due to STA operations.

164

6.9.4

C.J. Richardson et al.

Soil P Efflux and Gradients

We need to return to the elevated surface water P concentrations that were observed along the WCA-2A C- and D-transects, because that data reinforce one of the most significant conclusions that we can make from the various studies described in this gradient chapter. Although improving surface water quality entering the northern Everglades is a necessary first step in reversing the ecological changes that have been attributed to P enrichment, our soil and vegetation studies indicate that the accumulated stock of P in the soil and vegetation represents a large reservoir that will continue to influence overall ecosystem functioning for some time. This P legacy will require a great deal of time to self-correct, even under improved surface water quality inputs, and the system that will emerge will differ from that which existed prior to the twentieth century. The historic development of a highly enriched plume of P-enriched soils in the northern Everglades (Figs. 6.19 and 6.20) resulted from the long-term input of severely eutrophic surface waters into an oligotrophic ecosystem that accumulated P mainly as organic P in peat and litter. This created the first real gradient of consequence following the introduction of EAA runoff into the northern Everglades fen, a diffusion gradient between the surface water and the soils, which favored the downward movement of P into the peats. The expansion of highly eutrophic soils southward and the establishment of the geographical gradient of enriched surface soils were then a result of those highly enriched soils slowly pushing the zone of effective peat sorption slowly southward as soil enrichment reduced the strength of the diffusion gradient. Following the same basic tenets of diffusion dynamics which led to the problem of soil enrichment in the first place, the current improvements in water quality have probably already begun to reverse the source–sink relationship that led to the historic loading of P into the peats and will result in the large-scale movement of P out of the soil and into the floodwaters (as we demonstrated in this chapter, see Fig. 6.26). This phenomenon of P efflux will likely accelerate as water quality increases (in terms of lower P concentrations) with future refinements and expansion of STA operations and further adoptions of additional BMPs within the EAA. This efflux will result in downstream transport of P into the mesotrophic soils of the existing transition zone, where at present only minor eutrophication still allows for native vegetation, in the absence of a disturbance event. Eventually, we expect that a new state of equilibrium will be established after the ecosystem has had a chance to equilibrate with the new average P load into the northern Everglades fens. We fully expect that new stable state to be defined by an overall soil P content that will be slightly higher than the 400 mg P kg−1 dry soil that has historically defined the background central WCA-2A, since we expect that, under the present management scenario, surface water inputs will stabilize near 20–30 µg P l−1 at best in the currently enriched areas (Chap. 2). Additional research on the soils of WCA-2A is needed to more precisely define both the temporal and spatial dynamics of the reconfiguration of soil P patterns within the northern Everglades. This is required, because it is likely that significant soil P redistribution will influence vegetation patterns within the unit for some time,

6 Enrichment Gradients in WCA-2A and Northern WCA-3A

165

and restoration of predisturbance vegetation patterns (i.e., removal of large-scale monoculture cattail stands and the return of the ridge-and-slough community structure) is a distinct goal of the overall Everglades restoration effort. We predict a transient increase in cattail-dominated communities contemporary with the initial movements of significant P concentrations downstream into the center of WCA-2A. However, as efflux-driven P redistribution progresses, soil P concentrations will continue to decline throughout the whole of WCA-2A, and eventually soil P content will drop below where cattail is physiologically favored over ridge-and-slough species such as sawgrass. Every effort needs to be made to recognize that this hypothesized expansion of cattail may occur as a natural result of the northern Everglades ecosystem establishing a P new equilibrium, and to incorporate that recognition into future management goals for the northern Everglades as a whole (Chap. 26). Although the expansion of P-enriched soils downstream into the background ridge-and-slough communities is likely to be inevitable, it is possible that proper management of the northern Everglades fens may preclude a significant, if transient, expansion of cattail. Specifically, the reintroduction of a significant fire management regime to the WCA-2A fens may aid in dampening the rate of cattail expansion. We therefore feel that present water quality improvements and the necessary future adoption of a regular fire management program in cattail-dominated areas in the Everglades hold the distinct promise of accelerated restoration of the natural ridge-and-slough community structure, and the removal of large monoculture cattail stands, especially in the northern Everglades (Chap. 26). More research on the spatial and temporal dynamics associated with the soil efflux phenomenon is also needed in order that the long-term but natural recovery of the ecosystem is well understood by all stakeholders.

7

Geologic Settings and Hydrology Gradients in the Everglades Edwin A. Romanowicz and Curtis J. Richardson

7.1

Introduction

The ecology of the Everglades is coupled closely to its hydrologic setting. The present hydrologic setting of the Everglades is very different from what it would have been without the human incursions of the last 100 years. Virtually all surfacewater flow in southern Florida is now managed. Many changes in the Everglades ecosystem have been documented, but the degree to which ecosystem changes can be ascribed to changes in hydrology is uncertain (Craft and Richardson 1993b; Davis and Ogden 1994b). One obstacle is the difficulty of determining south Florida’s hydrology prior to water management projects in the region. Without a detailed understanding of historical conditions it is difficult to quantify how the hydrology has changed in the past several decades. A second obstacle is the lack of data about which hydrologic characteristics (e.g., mean water depth, maximum or minimum water depth) elicit given responses from vegetation. The goal of this chapter is to review the geological strata that influence hydrologic flow and present detailed hydrologic investigations in Water Conservation Area 2A (WCA-2A) to characterize how man-made alterations (dikes, flood gates, etc.) affect the modern flow regime. We also report the results of several different studies at different temporal and spatial scales designed to study the hydrology in WCA-2A. From these results, we have current estimates of how water flows through heavily vegetated environments as compared to open water. We also demonstrate how different patterns of inundation and flooding within WCA-2A may account for some of the variations in vegetation communities. Importantly, results from our studies provide a basis for understanding how hydrological alterations effect water flow and hydrologic budgets. The results of this study hopefully will be used as a template to study hydrologic conditions in other areas of the Everglades as well as aid in restoration efforts by reducing hydrologic impacts to vegetation communities.

167

168

7.2 7.2.1

E.A. Romanowicz and C.J. Richardson

Geologic Settings of the Everglades Introduction

The geologic setting, climate, and the elevation of sea level are important elements that lead to the hydrologic conditions necessary for the formation of the Everglades ecosystem. The formation of the Everglades was initiated by poor drainage and precipitation in excess of evapotranspiration during a period of rising sea levels during the past 5,000 years. Spatial distribution of different ecosystems within the Everglades and Big Cypress coincide with bedrock and topography. South Florida ecosystems are dependent upon hydropattern and depth of inundation (as measured relative to the top of peat). These hydrologic properties are a function of the elevation of the top of bedrock relative to the equipotential surface of the surrounding water table. Geology is also an important factor in regional drainage of surface water and groundwater discharge and recharge.

7.2.2

Physiography

South Florida has very little topographic expression. The most prominent feature of topographic relief is the Atlantic Coastal Ridge (Fig. 7.1). Most of south Florida’s population is distributed along this ridge. The highest elevation along the ridge is only about 6–7 m (20 ft.) above mean sea level (MSL); however, this is high enough that the ridge is protected from coastal storms and can act as a physiographic boundary for the Everglades, confining surface water flow in the Everglades to west of the ridge. Much of the Everglades overlay three elongated bedrock depressions extending from Lake Okeechobee to Florida Bay: Loxahatchee Channel, Tamiami Basin, and Shark River Bedrock Slough (Jones 1948; Parker et al. 1955; Petuch 1986, Gleason and Stone 1994) (Fig. 7.1). There is little relief between the center of these depressions and the higher elevations to the sides (approximately 1.5 m), but the relief is sufficient to form channels through which much of the surface water in the Everglades flows. Within these channels there is increased flood duration and depth that promotes wetland vegetation patterns.

7.2.3

Geologic History

The geologic events that were most significant in the formation of the Everglades had their beginning about five million years ago in the Pliocene (Gleason and Stone 1994). Much of the geology of southern Florida is a result of marine transgressions and regressions (Davis 1943; Schmidt 1997). Five million years ago the coast of Florida was far inland from its present location (Williams et al. 1990). Within these

7 Geologic Settings and Hydrology Gradients in the Everglades

169

Fig. 7.1 Physiographic map of south Florida (Cooke 1939; Jones 1948; Parker et al. 1955; Petuch 1986; Gleason and Stone 1994; Schmidt 1997)

warm waters that covered the present day Florida peninsula there was an abundance of life from which precipitated calcium carbonate (CaCO3), the principal mineral making up limestone (Scott 1997). This carbonate deposition formed much of the bedrock on which the Everglades would later form. Several marine terraces are evidence of the ocean’s retreat across Florida to the present coastline. These terraces are topographic features that resulted from the erosion and sediment redeposition in

170

E.A. Romanowicz and C.J. Richardson

near-shore marine environments. The Everglades are located on the lowest of these marine terraces, the Silver Bluff Terrace (less than 1–3 m above MSL) and the Pamlico Terrace (2.5–7.6 m above MSL). The last marine transgression and regression in the region occurred during the Sangamon (Peorian) interglacial stage and the following Wisconsin glaciation (Davis 1943; Flint 1971). During the Sangamon interglacial stage, sea levels in Florida rose 7.6 m (25 ft.) above present-day sea level (Davis 1943), inundating any land below that elevation. Then during the last ice age (Wisconsin), sea levels dropped worldwide to as much as 7.6 m (25 ft.) below current levels, expanding the Florida Peninsula (Davis 1943; Schmidt 1997). In the last 25,000–35,000 years, sea levels have again risen in Florida, resulting in the present coastline. The sea level was 4–5 m below modern sea level when the basal peat in the Everglades began to form about 5,000 years ago (Gleason and Stone 1994). The rising sea level since the formation of the Everglades enhances poor drainage by raising the water table and decreasing overland drainage by reducing the hydraulic gradient. The Everglades overlay three Pleistocene lithologic units of the Biscayne aquifer system: Anastasia Formation, Miami Limestone, and Fort Thompson Formation. The Anastasia formation is composed of coquina, sand, sandy limestone, and marl. Water yields range from moderate to high (Klein and Causaras 1982). The Miami Formation is a sandy, oolitic limestone. Changes in the limestone facies from west to east show that the depositional environment changed. The western Miami Formation is bryozoan facies with bioclastic sands. To the east the Miami formation grades to oolitic calcarenite limestone that make up the rocks of the Atlantic Coastal Ridge (Hoffmeister and Multer 1968; Frazier and Schwimmer 1987; Miller 1997). Water yields are high in the Miami Formation (Klein and Causaras 1982; Miller 1997). The Fort Thompson Formation has alternating layers of marine and fresh-water limestone with high yields of water (Klein and Causaras 1982; Miller 1997). The limestone bedrock of south Florida is an important factor in the formation of the Everglades. The limestone is susceptible to chemical weathering. In the Everglades, the limestone has weathered to form basins in which the wet prairie, sloughs, saw grass, and tree islands are located. In the southern Everglades, irregularities in the bedrock topography form the core of the tree islands (Robertson 1953). To the west in Big Cypress, chemical weathering created depressions in which the cypress domes form (Duever et al. 1986).

7.2.4

Peat Depth and Water Flow

The Pleistocene Fort Thompson formation is overlain by the northern Everglades, and the Miami Formation is overlain by the southern Everglades (Davis 1943; Parker et al. 1955; Puri and Vernon 1964; Gleason and Stone 1994). A thin layer of peat, with thicknesses ranging between 1.5 m in the north and a few centimeters in the

7 Geologic Settings and Hydrology Gradients in the Everglades

171

south, covers the bedrock. The Fort Thompson Formation impedes the downward flow of water more than the Miami Formation; thus, water tended to pond in the northern Everglades (Davis 1943). The Everglades are underlain by three elongated bedrock depressions extending from Lake Okeechobee to Florida Bay (Loxahatchee Channel, Tamiami Basin, and Shark River Bedrock Slough) (Davis 1943; Jones 1948; Parker et al. 1955; Petuch 1986; Gleason and Stone 1994) (Fig. 7.1). Historically, surface water flowed south from Lake Okeechobee toward the Atlantic Coast Ridge. The Atlantic Coast Ridge deflected flow to the southwest, and much of the surface water flowed through Taylor Slough into Florida Bay. Lake Okeechobee is the source of the hydraulic head that “pushes” the water south through the Everglades. Historically, the hydraulic gradient ranged between 2.3 × 10−5 and 3.8 × 10−5 depending on the elevation of Lake Okeechobee (Parker et al. 1955). Because Lake Okeechobee has been artificially lowered, the hydraulic gradient has changed. The flow of water through the Everglades can be described best as “sheet flow.” Water moved very slowly through the Everglades because of the low hydraulic gradients and the resistance to flow offered by the thickly distributed vegetation. Rates of evapotranspiration and precipitation are nearly equal. Thus, water flowing south through the Everglades is rapidly evaporated and replaced by precipitation. As the sheet of surface water approached the southern end of the Florida peninsula, the water passed through the narrower channels of the Shark Valley Slough and the Taylor Slough which discharge into Florida Bay. Results from the Natural System Model (NSM, USACE 1999) show that the rate of surface water flow through the Shark River Slough was much greater than in other areas of the Everglades (Fennema et al. 1994). Flow in the fast-forming areas of the Everglades was nonexistent or very slow.

7.2.5

Climate

Climate is a critical factor in assessing the hydrogeologic setting of the Everglades. Florida’s climate is humid-subtropical (Jordan 1984). A humid-subtropical climate has both high annual precipitation and temperature. The amount of precipitation relative to evapotranspiration is an important part of the hydrologic budget. Between 1951 and 1980, the average annual rainfall for much of the area covered by the Everglades was about 124 cm (60 in.) (Jordan 1984). Over southeastern Florida, annual precipitation surpasses open pan evaporation by 10–20 cm year−1 (Winter and Woo 1990). In southwestern Florida, precipitation and open-pan evaporation are about equal (Winter and Woo 1990). The transition from excess average precipitation to comparable annual evaporation and precipitation rates occurs nearly coincident with the boundaries between the Everglades and Big Cypress. Precipitation and evaporation have seasonal cycles. Precipitation and evaporation rates trends for WCA-2A were studied using data from the meteorology station at pumping station S-7 (SFWMD 1995). WCA-2A receives about 60% of its total annual precipitation between April and September. Median monthly precipitation

172

E.A. Romanowicz and C.J. Richardson

rates (1973–2000) range between 3 cm month−1 in December and about 20 cm month−1 in June. Evaporation rates follow a similar trend. The greatest evaporation rates occur during May and June, with the greatest median evaporation rate in May. Evaporation rates range from a monthly median of 0.4 cm d−1 (January) to 0.8 cm d−1 (May). During the study period (1989–2000), WCA-2A had greater average precipitation and average evaporation than during the period 1973–2000. An average precipitation rate for 1973–2000 was 10 cm month−1. During 1989–2000, the average precipitation rate was 12 cm month−1. Average evaporation rates increased from 0.45 cm d−1 to 0.57 cm d−1 (1989–2000). These figures show a trend of increased annual precipitation since 1973.

7.3

Hydrology

Wetlands are typically highly stressed environments. In most wetlands, the hydrologic setting of the wetland contributes an important stress. Periodic flooding, changes in water depth, and nutrient transport are factors related to the hydrologic setting. Indirectly, the hydrology is linked to natural fire cycles that are important in many wetlands (Chap. 9). These stresses lead to the unique vegetation communities found in wetlands. Wetland vegetation is well adapted to survive in these environments. Terrestrial vegetation would fail to flourish with these stresses. Slight changes to the hydrologic setting of a wetland may impact the wetland significantly overstressing wetland vegetation or lessening stresses on nonwetland vegetation, which then becomes competitive. To assess the hydrologic patterns and flows in the northern Everglades, a detailed study was undertaken in WCA-2A.

7.3.1

Hydrologic Budget

The hydrologic budget quantifies hydrologic inputs and outputs to a wetland. Typically, hydrologic inputs include surface runoff (Ri), surface inflow (Si), groundwater (Gi), direct precipitation (P), and anthropogenic sources (Ai). Hydrologic outputs include evapotranspiration (ET), surface outflow (So), groundwater (Go), and anthropogenic outflows (Ao). The equation for the hydrologic budget is expressed as difference between inputs and outputs (7.1). Dstorage = Ri + Si + Gi + P + Ai – (ET + S˚ + G˚ + A˚ )

(7.1)

The change in storage (∆storage) is the change in the volume of water stored in the wetlands. If the hydrologic input exceeds output, the storage increases (∆storage > 0). If the output exceeds input, the storage will decrease (∆storage < 0). A wetland with a balanced hydrologic budget (input = output) will have no change in storage.

7 Geologic Settings and Hydrology Gradients in the Everglades

173

Storage is often measured as stage (water elevation). In this case, it is necessary to develop rating curves to show the relationship between stage and storage for specific wetlands.

7.3.2

Water Conservation Area 2A Hydrologic Budget

WCA-2A is essentially a wetland reservoir. The perimeter is lined entirely by a dike. All surface water flow is through regulated control structures. Surface water flows into WCA-2A from the north and generally flows out from the south (Fig. 7.2).

Fig. 7.2 Map of Water Conservation Area 2A, showing water control structures. Arrows show the direction of water flow through the structures

174

E.A. Romanowicz and C.J. Richardson

We identified and quantified hydrologic input and outputs to calculate a hydrologic budget for WCA-2A. This task was made simpler since all surface water inflows and outflows are regulated and monitored. Using data from the South Florida Water Management District (SFWMD 1995), annual hydrologic budgets were calculated for WCA-2A between 1973 and 1994 (Table 7.1). Surface water inflows to WCA-2A were used to calculate the hydrologic budget through gates S-10A, S-10C, S-10D, S-10E, and pumping station S-7 (Cooper and Roy 1991) (Fig. 7.2). Outflows used in the hydrologic budget are through gates S-143, S-144, S-145, S-146, S-11A, S-11B, and S-11C (Cooper and Roy 1991). Flow through these structures is regulated carefully so that flow occurs under very specific conditions. Evaporation rates were estimated using surface-water depth and land-use/vegetation coefficients from the Natural System Model (Fennema et al. 1994). Land-use/ vegetation coefficients were averaged for the year. Water depth was calculated using stage data from the 2A-17 guage (SFWMD 1995) (Fig. 7.3). Storage in WCA-2A was calculated from the mean stage at 2A-17 (SFWMD 1995). Direct precipitation was determined from daily accumulation rates from a precipitation guage at pumping station S-7 (Fig. 7.2) (SFWMD 1995). From the hydrologic budget (Table 7.1, Fig. 7.4), one can see that precipitation and evapotranspiration account for nearly equal percentages of the budget (22%

Table 7.1 Hydrologic budget for Water Conservation Area 2A (1973–1994) Input (i) Output (o) ∆Storage (i−o) %Difference (×108 m3) (×108 m3) (i−o)/(i+o) Year (×108 m3) 1973 6.06 7.42 1974 8.10 9.46 1975 6.46 5.89 1976 7.28 7.67 1977 7.19 7.82 1978 13.85 10.39 1979 10.11 11.92 1980 10.25 12.80 1981 7.11 7.11 1982 11.78 16.28 1983 12.60 17.44 1984 7.53 13.82 1985 8.98 15.60 1986 10.73 15.86 1987 7.90 12.57 1988 10.73 12.15 1989 5.32 6.09 1990 5.63 5.86 1991 9.66 14.90 1992 12.40 15.89 1993 13.22 14.92 1994 19.48 26.84 Mean 9.65 12.21 Median 9.32 12.36 i all measured inputs, o all measured outputs

−1.36 −1.36 0.57 −0.39 −0.63 3.46 −1.81 −2.55 0 −4.50 −4.84 −6.29 −6.62 −5.13 −4.67 −1.42 −0.077 −0.23 −5.24 −3.49 −1.70 −7.36 −2.52 −1.76

−10.08 −7.74 4.59 −2.65 −4.15 14.25 −8.23 −11.06 0 −16.04 −16.12 −29.44 −26.96 −19.28 −22.82 −6.19 −6.70 −1.97 −21.34 −12.31 −6.03 −15.89 −10.28 −9.16

Residence time (years) 1.6 1.3 2.2 1.6 1.7 0.9 1.0 0.7 1.2 0.5 0.5 0.6 0.5 0.5 0.8 0.7 1.4 1.4 0.6 0.6 0.6 0.6 0.98 0.75

7 Geologic Settings and Hydrology Gradients in the Everglades

175

Fig. 7.3 Stage at station 2A-17 (SFWMD 1998)

and 20%, respectively). Typically, surface-water outflows surpass inflows. This is expected since precipitation adds water to WCA-2A. Of interest is the fact that residence times of 1–2 years were found in the early 1970s, but by 1980 they had mostly decreased to 0.6 years. This corresponds to increased surface flow shown during the same period. Between 1973 and 1995 there was an 18% deficit in inflow versus outflow. There is some evidence that this deficit can be accounted for by a net inflow of groundwater. Profiles of dissolved chloride (Cl−) in the peat pore water from two sites in WCA-2A suggest that there is a groundwater flow in the northeast part of WCA-2A (Fig. 7.5). The profile in the northeast is consistent with diffusive and advective transport. Harvey (1996) reports from seepage studies in WCA2A that there appears to be a positive flux of groundwater to WCA-2A in the northern parts. Water chemistry of runoff from the Everglades Agriculture Area (Lutz 1977), surface water in WCA-2A, and groundwater in the conservation areas (Nealon 1984) has very similar percent milliequivalents of major cations (Ca, Na, K, Mg) and anions (Cl, HCO3, SO4). Although not conclusive, the similarity in the water chemistries support the idea that groundwater contributes to the surface water in WCA-2A. Moreover, there has not been sufficient change in stage to account for the budget deficit. The high yield of the bedrock on which the Everglades are located and the high potentiometric head in the aquifer (Parker et al. 1955) further support the conclusion that there is a net flux of groundwater to WCA-2A.

176

E.A. Romanowicz and C.J. Richardson

Fig. 7.4 Hydrologic budget of Water Conservation Area 2A

7.3.3

Hydraulic Conductivity

Hydraulic conductivity is a measure of the ease in which water flows through a porous media. Hydraulic conductivity differs from permeability because hydraulic conductivity is a function of the viscosity and density of the fluid flow through the media. Since the Everglades overlie the Biscayne Aquifer, the bedrock has a high hydraulic conductivity. We used in situ single-well slug tests to measure the hydraulic conductivity of the peat (Hvorslev 1951). Conductivity experiments were performed using piezometers constructed from 2-in. I.D. (5.0 cm) Schedule 80 PVC pipes with machined threaded couples to extend the pipes and add a drive tip. The drive tip had an 11-cm length screened interval. The piezometers were hand-driven into the peat to the desired depth relative

7 Geologic Settings and Hydrology Gradients in the Everglades

177

Fig. 7.5 Observed profiles of dissolved chloride in peat pore-water from two sites in WCA-2A. The differences in the profiles suggest that there may be groundwater flow into WCA-2A in the northeast

to the top of the surface water. Typically the piezometer was driven into the peat to the top of bedrock. After the piezometer was installed a pressure transducer attached to a continuous-recording digital data logger was lowered to the bottom of the piezometer. The piezometer was then filled with water to the top. The data logger recorded the hydraulic head recovery in the piezometer as the water in piezometer lowered to the ambient hydraulic head. Hydraulic head measurements were recorded every 2 s by the data logger. Hydraulic conductivities range between 1.6 × 10−4 and 6.7 × 10−2 cm s. This range of hydraulic conductivities is similar to the hydraulic conductivity of sand. From the hydraulic conductivities and difference between the hydraulic head in the piezometer and the surface of the water we calculated the Darcy vertical specific discharge (not accounting for porosity). The vertical specific discharges ranged from 7 × 10−4 to −1 × 10−3 ml s−1 cm−2 (positive indicates upward vertical flow). Again, these data provide further evidence that there exists a vertical upward flow of water.

7.3.4

Hydropatterns

Hydropattern describes the frequency of surface-water inundation, the depth, the duration, and timing of inundation. All of these are important factors in determining

178

E.A. Romanowicz and C.J. Richardson

the health and stability of the Everglades. The hydropattern in WCA-2A is regulated and no longer occurs as a natural response to precipitation and surface-water flow as it did before water control structures were built. Within WCA-2A, the hydropattern varies considerably. Variations in the hydropattern result from the relative rate of water flow through the water control structures, the topography of the peat surface, and equipotential surface of the water. Surface-water stage was monitored historically at several locations within WCA-2A (SFWMD 1992). However, the density of the monitoring network was not sufficient to monitor changes in the hydropatterns at a scale able to discern variations in the vegetation communities caused by those changes. Our studies do provide valuable data on flow rates and directions within Everglade communities. We studied the hydropattern using a network of 11 continuous-recording stage recorders equipped with 10-psi pressure transducers (error ± 0.1% full scale reading) (Fig. 7.6). The elevation of each site relative to the Natural Geodetic Vertical Datum 1988 (NVD) was determined with differential GPS (Keith and Schnars 1993).

Fig. 7.6 Site location map of surface-water stage monitoring stations in WCA-2A

7 Geologic Settings and Hydrology Gradients in the Everglades

179

These stations were monitored for a period of 2–3 years. The results of the monitoring were used to determine the equipotential surface of the surface water and to study how flow through the water control structures affected the equipotential surface. From this study, we completed a statistical model that could predict water elevation at our study sites using stage data from station 2A-17 and discharge through the water control structures (SFWMD 1995). The model was constructed to determine historical water depths throughout the study area. The South Florida Water Management District and the USGS maintain two stage recorders in WCA-2A. For one stage recorder (2A-17) there are data going back to the early 1950s. However, the distribution of stage recorders is insufficient to accurately monitor fluctuations in the equipotential surface. We made two assumptions in our model. First, we assumed that the equipotential surface in WCA-2A is affected by surface-water flows to and from WCA-2A through the water control structures. Second, we assumed that stage throughout WCA-2A is related to stage at station 2A-17. The use of 2A-17 is particularly important because it accounts for fluctuations in the equipotential surface due to precipitation and evapotranspiration. Importantly, we have a nearly continuous record of stage from 2A-17 back to 1951 (van Horn 1996). For the water control structures there is a nearly continuous record back to 1981 (van Horn 1996). Our model is limited by the flow data from the water control structures, so it can model hydrology no further back than 1981. We used a multiregression model using water flow data through water control structures S10-A, S10-C, S10-D (Fs10); S11-A, S11-B, S11-C (Fs11); S-143, S-144, S-145, S-146 (Fs140); S39 (Fs39); S7 spillway and pump (Fs7); and stage at 2A-17 (S17) (Fig. 7.2). Water flow data are mean daily flow rate in cubic feet per second (SFWMD 1995). The stage at 2A-17 is given in feet relative to sea level (Natural Geodetic Vertical Datum 1929). The general form of the modeled equation is Elevation = I + b1FS10 + b2FS11 + b3FS140 + b3FS39 + b4FS7 + b5S17.

(7.2)

In the equation, the modeled coefficients are b1–b5 and the intercept is I. For each of the 11 sites an equation with separate coefficients and intercept was modeled. We calculated coefficients, I and b1–b5 and standardized coefficients, R2 and adjusted R2 (adjusted for the number of independent variables) (Davis 1986; Schroeder et al. 1986). We checked model calibration by comparing predicted values of elevation at each site with observed elevations using mean error, mean absolute error, and root mean square (Anderson and Woessner 1992) (Table 7.2). From the mean error we can check the average residual between measured elevation and modeled elevation. Positive and negative residuals will tend to cancel each other. A small mean error would suggest that over extended periods of time the average performance of the model is accurate. The mean absolute error gives us the mean magnitude of the residual. The root mean square gives us the variance of the data. The water depths were calculated by subtracting the top of the peat elevation (Keith and Schnars 1993) from modeled water elevation. Peat accumulation (Craft and Richardson 1993b) rates were used to estimate changes in the elevation of the peat surface.

180

E.A. Romanowicz and C.J. Richardson

Table 7.2 Statistics of comparisons of modeled water elevations with observed water elevations in study area Site identifi- Number of cation data points

Mean error (×10−8 m)

Mean absoRoot mean lute error (m) square

R2

Adjusted R2

S-11 S-12 S-13 S-14 S-15 S-16 S-19 S-20 S-21 S-24 S-25

2.96 5.79 −2.29 1.52 −0.37 3.35 −3.02 −0.55 −0.67 −1.04 4.27

0.04 0.03 0.02 0.01 0.03 0.02 0.04 0.04 0.05 0.02 0.06

0.95 0.98 0.96 0.99 0.97 0.98 0.91 0.94 0.89 0.79 0.90

0.95 0.98 0.96 0.99 0.96 0.98 0.90 0.94 0.89 0.79 0.90

407 355 207 309 400 275 407 406 343 111 365

0.06 0.04 0.02 0.02 0.04 0.03 0.05 0.05 0.07 0.03 0.08

Seasonal variations in water depths during the year lagged behind precipitation patterns. Water depths at each of the study sites increased during late summer (July) and remain elevated through October. Shallowest water depths occur late spring and early summer (May to June), which were coincident with the greatest precipitation accumulation. The modeled water depth at each of the sites ranges over more than 1 m. Site-to-site comparisons of water depths show that WCA-2A does not behave as a simple reservoir. One would expect that water depths would be largely affected by the elevation of the top of the peat and a slight hydraulic gradient southward reflecting the assumed general southward flow of surface water. The elevation of the top of peat decreases from 3.4 m in the north to about 2.7 m (above MSL, Natural Geodetic Vertical Datum 1988) in the south. However, a comparison of the water depths and the site locations show that water depths are typically greater along the northern and southern boundaries and lower in the center of WCA-2A. Water flowing into WCA-2A from the north probably causes water depths at WCA-2A’s northern boundary to increase rapidly. This water gradually moves through WCA-2A and dissipates. As the water flows to the southern boundary, it ponds along the southern dike until the water can flow out of WCA-2A. From the model, we calculated recurrence intervals of water depths for WCA-2A. The recurrence interval describes the probability that water depth will exceed a specific depth in given number of days. For example, if the recurrence interval for a 30-cm water depth is 200 days, then one would expect that the probability on a given day the water depth is 30 cm or greater is 1/200. So on average, once every 200 days the water depth would be 30 cm or greater. The recurrence intervals show even more clearly the trends observed with the site-to-site water depth comparisons (Fig. 7.7). These maps (Fig. 7.7) show contours of predicted water depths with recurrence intervals of 10,100 and 1,000 days. In general, throughout the study area about 10% of the time (10-day recurrence interval), minimum water depths ranged between 50 and 75 cm. The minimum water depths 1% of the time (100-day recurrence interval) range between 90 and 140 cm. Finally, 0.1% of the time (1,000-day

7 Geologic Settings and Hydrology Gradients in the Everglades

181

Fig. 7.7 Depths (cm) with recurrence interval of (a) 10, (b) 100, and (c) 1,000 days calculated from statistical model

recurrence interval) minimum water depths range between 100 and 160 cm. Although the northwest part of WCA-2A was not instrumented for this study, the recurrence intervals suggest the region is drier than the eastern part. The 10-day and 100-day recurrence interval maps (Fig. 7.7) show that the center of the study area tends to have shallower water elevations. The northern and southern parts tend to have deeper water. Extreme events in the northeastern part of the study area with recurrence intervals of 1,000 days are deeper. The northwest and south parts of the study area do not have as deep water during these extreme events. Analysis of the frequency of water depths occurring in the study area shows that they result directly from inflow structure releases and the ponding to the south as water backs up before flowing out of WCA-2A. The northern and southern areas have greater variations in water depth than the central areas (near sites 2A-14 and 2A-16 (Fig. 7.6). The duration of flooding is an important attribute in a wetland’s hydrology. From the model, the mean duration of flood events was calculated. This is the mean number of consecutive days that the water depth was within a specific range of depth during the modeling timeframe (1981–1998). We looked at depths ≤5 cm and depths ≥30, 60, and 90 cm (Fig. 7.8). There is some redundancy since depths that are greater than 60 or 90 cm are also greater than 30 cm. These trends show again that the central part of the study area tends to be drier. The center of the study area had the greatest number of consecutive days with water depths less than 5 cm. The northern and southern areas had a few consecutive days, the north having the fewest. Looking at trends with increasing water depth, the center of the study area similarly had fewer consecutive days when water depths exceeded 30 cm (Fig. 7.8b). The north-central and south-central areas had the greatest number of days when water depths exceeded over 30 cm (60–65 days). The similar trend between the northern and southern parts of the study area vanishes at greater water depths. Looking at consecutive days with water depths greater than 60 and 90 cm, the northern parts have fewer consecutive days, with the greatest number of consecutive days occurring in the south (Fig. 7.8c, d). Although this may seem inconsistent with

182

E.A. Romanowicz and C.J. Richardson

Fig. 7.8 Maps of mean duration (days) of water depths less than or equal to (a) 5 cm, and greater than or equal to (b) 30, (c) 60, and (d) 90 cm

the trends observed from maps of the recurrence intervals, what we see is that the northern parts have greater and more rapid fluctuations in water depths than compared with the south. Water depths in the south were neither as deep as in the north nor did they fluctuate as much as the north. The center of the study had more shallow depths with greater frequency.

7 Geologic Settings and Hydrology Gradients in the Everglades

7.3.5

183

Equipotential Surface

The equipotential surface in WCA-2A is the elevation of the top of the water. It is expressed as an elevation relative to MSL (NGVD 1988). The elevation of the top of the water does not mean the same thing as water depth. Water depth is an important factor in many ecological applications; however, the elevation of the top of water expressed as an equipotential surface yields very important information about the direction of water flow. Water always flows from a high equipotential elevation to a lower equipotential elevation. The hydraulic gradient (change in water elevation over distance) yields very important information about relative flow rates (discharge) and the resistance to surface-water flow from the peat surface and vegetation. If the resistance to flow from vegetation is isotropic, then one would expect the direction of surface-water flow is parallel to the direction of maximum hydraulic gradient. We monitored the equipotential surface at the previously mentioned 11 monitoring stations from January 1997 through July 1999. Some stations experienced down time throughout the study period, while others were inaccessible for extended times because of low water levels. However, despite some data gaps, we compiled a detailed equipotential map using mean daily stage data calculated from data collected every 15 min. The equipotential surface in WCA-2A is dynamic. It fluctuates rapidly with changes in the relative flow rates to and from WCA-2A through the water control structures. We expected that flow in WCA-2A would be generally southward (Fig. 7.9a–d). Most of the time this was true, although in some cases the water flowed northeast to southwest (Fig. 7.9c). Other times water flowed from the northwest to the southeast (Fig. 7.9a). On some occasions, a ridge of high water formed from the northwest to the southeast through the center of WCA-2A. This ridge of high water bisected the surface-water flow into two regions. Water flowed east from the northeastern edge of the ridge and southwest from the southwestern edge of the ridge. The effect of the water control structures is evident from the equipotential maps (Fig. 7.9a–d). For instance, on January 2, 1997 flow was from WCA-2A through the S-39 gate, driving the flow to the southeast (Fig. 7.9a). On June 3, 1997, S-7 and S10-D (Fig. 8.3) were releasing water into WCA-2A. As a result, ponding occurred to the south as water elevation increased, decreasing the hydraulic gradient. Nearly 12 months later, on July 21, 1997, the flow of water moved from the northeast to the southwest as S-10 structures released water into WCA-2A and the S-11 structures removed water. Changes in the equipotential surface can occur rapidly with changes in flow rates through the water control structures and in the distribution of active and inactive water control structures. Typically, the elevation of the top of the water decreased 0.6 m over the north-tosouth length of the study area. At times, the hydraulic gradient was very small (January 3, 1998, Fig. 7.9d). The most significant feature of the equipotential map is the high ridge of water bisecting WCA-2A. Given a small hydraulic gradient, one would expect very little surface water flow. As the hydraulic gradient increases between the south and north, one would expect the flow to increase as well. It is not possible to always attribute the changes to hydraulic gradient to changes in flow rates.

184

E.A. Romanowicz and C.J. Richardson

Fig. 7.9 Maps of different configurations of the equipotential surface (m) in WCA-2A in the Everglades

It is possible that the hydraulic gradient changes not only as a function of flow but also as a function of depth, as resistance to flow changes for different water depth. Changes in flow directions have an impact on the transport of nutrients from the Everglades Agriculture Area through WCA-2A. When the ridge of high water is present, the movement of water east and possibly northeast will retard the southward transport of nutrients flowing into WCA-2A from the S-7 and S-10 water control structures. When the flow is southward, the movement of water will increase the advective transport of nutrients. However, since surface flows are so slow, it is likely that transport of nutrients is largely through diffusion.

7 Geologic Settings and Hydrology Gradients in the Everglades

7.3.6

185

Surface Water Flow

The flow of water in the Everglades is significantly reduced by vegetation, and the low hydraulic gradient that exists within the vegetation is due to increased flow resistance caused by the stems and debris. Often flow in wetlands is approximated assuming Manning’s flow through an open channel (Feng and Molz 1997). In a similar fashion, the flow in the Everglades has been approximated assuming a Manning’s flow (Fennema et al. 1994). It is questionable if an empirical relationship derived for open channel flow in a channel with a finite width is a reasonable approximation for sheet-flow in the Everglades. In Manning’s Equation, resistance to flow is a function of the channel cross-sectional geometry, roughness of the channel bottom and sides, and obstructions in the channel. Thus, to obtain a more reasonable estimate of flow throughout the Everglades we need to determine flow in both vegetated and open water areas. As water flows through vegetation, the vegetation resists the flow. This is evident by the decrease in water velocities around vegetation. The lowest velocities will be near the source of drag. Resistance may still affect water further from the source of drag because of the viscosity of water. Resistance to flow from vegetation depends on the type of vegetation. The shape and texture of the vegetation can affect resistance. Because the geometry of plants is not uniform over the entire height of the plant, resistance to flow can change as a function of water depth. Also, resistance can change over an area as vegetation undergoes developmental changes or changes the density. Resistance to flow can change in a given area over time. This resistance will affect the water velocity and discharge rate. Even under very low vegetation densities, the effect of vegetation is more important than bed-drag (Nepf 1999). It is likely that much of the water flows through connected open slough and airboat trails where there is less vegetation and less resistance to flow. The bottom of the water column in the Everglades is much less distinct than the bottom of the water column in most sediment or bedrock channels. The interface between the top of peat and the surface water is an indistinct transitional zone from open water to floating material as much as 10–20 cm thick. This zone is thick with suspended peat and other organic debris. We measured profiles of water velocity using a SonTek® ADV Acoustical Doppler velocity flow meter. This meter can measure velocities less than 0.1 mm s−1 using the Doppler effect on RADAR transmissions as the RADAR is reflected from particles suspended in the water column. The velocity meter was programmed to make 25 measurements a second. Typically, we measured water velocity along each point of the profile for about 90–120 s. To calculate the direction of flow we used directional statistics to calculate the mean flow direction and the arc describing the 95% confidence interval for the mean flow direction (Fisher 1953; McElhinny 1973, Fisher 1993; Swan and Sandilands 1995). We also calculated the mean resultant magnitude of flow. The velocity profiles revealed a complex interaction between hydraulic gradient driven flow and the transfer of momentum through the water column from the shear stress of wind moving along the surface of the water (Fig. 7.10). These interactions

Fig. 7.10 (a) Water velocity profiles for Site 12. (b) Water velocity profiles for Site 16

7 Geologic Settings and Hydrology Gradients in the Everglades

187

resulted in a flow system that was often bidirectional, with water along the surface flowing in directions different than deeper water. The average water velocities were approximately 0.25 cm s−1. Typically, the greatest velocities were near the surface (0.4 cm s−1). The lowest velocities were near the peat-water interface (0.2 cm s−1). Throughout the water column the velocities varied considerably. These velocity profiles were not a simple consistent profile and displayed a drag originating from the bottom of the water column (Fig. 7.10a–b). Most striking was the change in water flow direction. At Site 12, (Fig. 7.10a) the direction of water flow fluctuates rapidly with depth. Data from Site 16 (Fig. 7.10b) show a gradual change in flow direction from a northeasterly flow along the surface (presumably wind-driven) and a southerly flow deeper in the water (parallel to the hydraulic gradient). From the velocity profiles it appears that the movement of water is controlled by a combination of wind driven water and hydraulic gradient driven water (i.e., gates opening and closing). The transfer of wind momentum through the water column is probably responsible for the complex velocity profiles that we observed. From observed water velocity measurements, it would take approximately 80 days for water to flow through WCA-2A. Yet, we see more rapid responses throughout the conservation area due to discharges through water control structures. The only resolution to this apparent contradiction is that much of the water flows through open sloughs, airboat trails, and canals along the perimeter. Water movement through the open areas responds quickly to the changes in discharge into WCA-2A. As water levels in these open areas respond, water elevations change in the nearby vegetated areas. In WCA-2A we have a multiflow system. There is a rapid flow system through the open areas and a very sluggish flow system in the vegetated areas. It is unclear how different this is from the original Everglades. Presumably, the original Everglades had unconnected open sloughs through which water could flow easily; however, water still flowed through the Everglades very slowly because the open sloughs were unconnected (lacking, for example, airboat trails).

7.4

Conclusions and Lessons for Restoration

The goal of this hydrologic investigation was to characterize the modern flow regime in WCA-2A. The results of this study could then be used to determine the extent to which disturbance to the hydrology in the northern Everglades impacts vegetation communities. The most significant conclusion is that the water control structures around WCA-2A have significantly affected the hydrology. The current hydrologic setting does not mimic the historical natural hydrologic setting at all. Particularly, the seasonal variations in water depths are no longer coincident with precipitation patterns. Spatially, the hydroperiods show changes that probably would not have occurred under natural circumstances. We see clearly the effects of ponding behind southern dikes and the rapid rise of water depths along the water control structures at the northern boundaries. The deepest water inundation occurs along the north and

188

E.A. Romanowicz and C.J. Richardson

south perimeter of WCA-2A. The interior of WCA-2A was much drier with small fluctuations in water depths. Reasonably, one would probably have anticipated disturbances to the hydroperiod by the water control structures. Less obvious is the association between the equipotential surface and the magnitude and distribution of discharge through the water control structures. The surface elevation of water is controlled by the occurrence of flow through the water control structures. The hydrology in WCA-2A is impacted not only by the magnitude of flow through the water control structures but also by the distribution of open and closed water control structures. In extensive wetland systems such as the Everglades, the magnitude of the hydraulic gradient is very small; thus, even minute disturbances that affect water elevation can affect the direction and velocity of water flow. Another interesting finding from this study is the significance of wind in the movement of water. Velocity profiles showed that as much as the upper 20 cm of water in the water column may be moved due to the transfer of momentum from the wind through the water column. This might have a significant impact on the movement of nutrients and dissolved solids in the water column. The average flow of water was @ 200 m d−1 in the open water areas, which suggests that water entering the northern part of WCA-2A may take between 80 and 100 days to reach the south in open water channels. Residence times of 0.6 years for all of WCA-2A support a much slower flow rate for the entire area. Not surprisingly, the water flow in dense vegetation was nearly zero, with water leaving densely vegetated areas mostly by evapotranspiration. This indicates that the Everglades did not originally function as a “river of grass” as suggested in the phrase coined by Douglas (1947). This finding has major consequences for the proposed Corps of Engineers restoration plan (USACE 1999), which would have water flowing south in the Water Conservation Areas and into the Everglades National Park. Indeed, some of the findings from this research project suggest that the planned remediation effort proposed for the Everglades restoration may have adverse effects. Presently, the southward movement of water from the north is impeded by ponding and the changes in water flow directions due to discharge through the water control structures. If planned remediation efforts increase the flow of water to the south above current levels, it is possible that the movement of nutrients from the Everglades Agriculture Area may be enhanced, a matter of great concern. Moreover, the forced flow of water would have severe impacts on native communities like sawgrass (Chap. 9). This study has shown that the hydrodynamics in an extensive wetland system such as the Everglades is very complicated. The low hydraulic gradients are subject to disturbances introduced by slight ponding and wind flow, and we have learned from this study that the flow of surface water is not topographically driven. This hydrologic field investigation required two different types of data: hydrologic data for investigation of the hydrologic setting of the Everglades and hydrologic parameters to which wetland vegetation respond. Importantly, this required detailed information on elevation and water depth. Most of the information that has been collected by the State of Florida, the USGS, and the USACE is water depth data, which is

7 Geologic Settings and Hydrology Gradients in the Everglades

189

often easier to collect than information on elevation. However, elevation data are critical to determine water flow directions. It is very important in wetland systems such as the Everglades to collect accurate elevation data for equipotential maps, and this should be done for all areas prior to any attempt to restore hydrologic conditions in the Everglades. In this study, we were fortunate to have detailed data on water elevation and top-of-peat elevation, without which many of the more interesting dynamics of the flow in the Everglades would not have been observed.

8

Effects of Hydrologic Management Decisions on Everglades Tree Islands Kirsten Hofmockel, Curtis J. Richardson, and Patrick N. Halpin

8.1

Introduction

Pressures from increasing agriculture and growing human populations have resulted in alteration of many aspects of the Everglades wetland mosaic in south Florida. Water management has focused on extracting services from the wetland system to support the influx of human inhabitants and growing agricultural production. Management practices to ensure water supplies for human use, control floods, and minimize hurricane effects have caused ecosystem fragmentation and substantial reduction of the spatial extent of the Everglades. Nearly half of the ecosystem’s original 404,686 ha have been transferred to agricultural use and urban development. To satisfy the needs of local inhabitants, a system of levees and canals has been installed over the past 50 years (Chap. 2). In addition, creation of the Water Conservation Areas (WCAs) impounded large sections of the wetland between Lake Okeechobee and the Everglades National Park. The implications and impacts of such management actions to landscape level processes are relevant to current ecological management decisions and the future health of the Everglades. Few research studies have investigated the effects of altered ecosystem structure on tree island persistence. Worth (1983, 1988) performed a series of experiments addressing the effects of increased ponding depth and hydroperiod on tree island survival. These experiments evaluated vegetation patterns on tree islands before and after prolonged dry periods. The primary objective of their study was to restore native plant communities, consolidate sediments, and enhance bird use. The results indicated that drydown caused a decrease in aquatic vegetation and an increase in grasses and forbes on tree islands. Successful recruitment of woody seedlings occurred on islands higher than 3.65 m NGVD (National Geodetic Vertical Datum) (12 ft.) and in close proximity of mature seed-producing individuals (Worth, 1988). Soils were consolidated, and the populations of Everglades Kites and wading birds increased with decreasing water levels (Worth 191

192

K. Hofmockel et al.

1988). Worth’s studies successfully describe the effects of altered hydrology on tree island species and the ecological importance of tree islands in the Everglades landscape. More recent studies of the origin, classification, and ecology of tree islands were presented in an excellent volume by Sklar and van der Valk (2002). In that edited volume, the authors compare the elliptical fixed tree islands usually found on topographic high limestone ridges in WCA-2A, WCA-3A, and ENP with the smaller, round, pop-up, or battery tree islands found primarily in WCA-1 (Loxahatchee). Interestingly, both types have a center area of semitropical hardwoods that are intolerant of prolonged flooding, but the fixed tree islands often have a woody head followed by a southern extended tail of more water-tolerant aquatic plants (shrubs, ferns, and cattails). Tree islands are the only source of topographic variety in the Northern Everglades (Loveless 1959). They are clumps of shrubs and hardwood trees elevated 0.5–1.2 m above the surrounding peat (Gleason and Stone 1994; Sklar and van der Valk 2002). The shape and orientation of islands reflects predrainage flow direction and velocities. Islands are associated with the Loxahatchee peat region, occurring on bedrock troughs (Gleason and Stone 1994). Tree islands lead to the formation of Gandy peat, which is formed in the aerobic, exposed zone and is composed of root and leaf material common to forest vegetation of tree islands (Davis 1943; Gleason and Stone 1994). Formation of tree islands may be associated with multiple biogeomorphic processes. Small islands are formed by floating peat mats, often including chunks of peat released during the creation of alligator holes. The live oxygen-filled rhizomes, accompanied by marsh gas, make the peat buoyant. As vegetation grows on the mat, it may become stuck in the surrounding dense marsh and roots of the live vegetation, eventually reattaching the mat to the marsh floor (Gleason and Stone 1994). This formation process is generally limited to smaller tree islands. Such islands tend to be circular in shape and are 20–30 m in diameter. Larger tree islands are commonly associated with bedrock topographic highs that allow trees to become established and spread. Larger tree islands are elliptical in shape reaching 1,600 m in length (Gleason and Stone 1994; Sklar and van der Valk 2002). Tree islands are typically dominated by dahoon holly (Ilex cassine), wax myrtle (Myrica cerifera), and red bay (Persea borbonia) (Worth 1988). Other canopy species may include swamp bay (Magnolia virginiana), and pond apple (Annona glabra) (Gunderson 1994). In general, small islands are predominately red bay while elliptical islands support mainly dahoon holly. Islands typically have a sawgrass fringe. Perimeter species vary and include species such as cocoplum (Chrysobalanus icaco) and wax myrtle (M. cerifera) while the elevated centers of islands tend to be red bay and dahoon holly (Gleason and Stone 1994). Multiple researchers have evaluated structural changes in the Arthur R. Marshall Loxahatchee National Wildlife Refuge, also known as Water Conservation Area 1 (WCA-1) (Silveira 1996; Jordan et al. 1997; Brandt 1997). Research indicates that changes in tree island patterns in WCA-1 are more influenced by duration of

8 Effects of Hydrologic Management Decisions on Everglades Tree Islands

193

hydroperiod and depth than by amount of variation in hydroperiod and flow magnitudes (Brandt 1997). Regions closer to canals decreased in size and percent cover of tree islands while the interior area increased in size and percent cover. Sites with decreases in percent cover had the greatest mean ponding depth (0.32 m) and the lowest elevation (4.5 m) (Brandt 1997). Similar studies have not been performed in WCA-2A. Structural changes in WCA-2A have focused more on emergent vegetation. Short-term herbaceous changes in WCA-2A have been extracted from Landsat multispectral scanner data (1973, 1976, and 1982) and SPOT high-resolution visible multispectral data (1987 and 1991) (Jensen et al. 1995). The Institute for Ecological Economics created the Everglades Landscape Model, a spatially explicit model that tracks ecosystem change in recent years (Voinov et al. 1997). Until the current study no research has fully quantified long-term changes in tree islands in WCA-2A. Successfully managing natural resources requires monitoring the results of past activities. Due to the extent of the Everglades, large-scale indicators of ecological change are logistically favorable. Tree islands can be used as indicators of ecosystem change because they represent one of the major community types (slough, wet prairie, tree islands, sawgrass marsh) present in the northern Everglades. Tree islands are appropriate as a basic monitoring unit because they persist over time. Tree islands are a result of both microscale processes (peat accretion/deccretion) and macroscale processes (ponding depth, flow velocities), making them appropriate indicators of changes in wetland function over time.

8.2

Objectives

The objectives of this study were to quantify the effects of hydrologic management decisions on tree island landscape patterns in the northern Everglades and to relate these changes to ecological features important to the integrity of Everglades biota. Tree islands were used as the basic structural unit because they persist in the landscape through time and represent a major natural community of the northern Everglades ecosystem. In this study we compared tree island abundance, area, shape, pattern, and location between 1953, 1980, and 1995, and related these structural changes to altered functional attributes documented in the literature. The time series compares the landscape pattern from 1953, when a flow-through system was still intact in the northern Everglades, to the mosaic after complete impoundment of WCA-2A (1980 and 1995). In the initial time step, hydropattern reflected climatic conditions, while in the later time steps hydrologic flux and storage were subject to management schedules set by the South Florida Water Management District. Understanding the results of historic management decisions is critical to restoring the health and functions of this freshwater wetland system.

194

8.3 8.3.1

K. Hofmockel et al.

Background Natural Everglades Ecosystem

The historic Everglades was a hydrologically interrelated system of sawgrass sloughs, wet prairies, cypress swamps, mangrove swamps, coastal lagoons, and the Florida Bay (Davis 1943). The Kissimmee River meandered into Lake Okeechobee, which released its water into receiving Everglades marshes. Surface sheet flow from the fresh water marshes then ran to the south and west, feeding mangroves, salt marsh estuaries, Big Cypress Swamp, and ultimately Florida Bay (Light and Dineen 1994, Fig. 8.1). A deep layer of organic matter covers the flat, permeable limestone bed of this region. Predominate peat include Everglades, Loxahatchee, and Gandy peats (Davis 1943). The shallow elevation gradient of 1.57–3.16 cm km−1 (USACE 1994), coupled with deep overlying peat, allows for dynamic storage of water during wet periods and a slow release of excess water during dry periods. The South Florida peninsula regularly experiences severe conditions. Extreme events are an intrinsic and important component of the Everglades ecology. Natural disturbances include hurricanes, storms, drought, flood, fire, and freezes. Individual disturbances mutually influence each other (Gunderson 1994). For example, hurricanes and storms have a strong influence on the annual water input and may prohibit the other forms of disturbance such as fire. In the absence of hurricanes or major storms, other disturbances (e.g., drought or fire) may thrive. Fire may be promoted by freezing conditions that inhibit tropical invasions and augment fuel loading. Each disturbance represents an important component of the natural system. A disturbance regime is an unpredictable yet critical ingredient to the persistence and integrity of the Everglades.

Fig. 8.1 Historic, current, and planned water flow conditions under CERP in the Everglades Watershed (Source: www.Evergladesplan.org)

8 Effects of Hydrologic Management Decisions on Everglades Tree Islands

8.3.2

195

Managed Everglades System

Florida entered the union in 1845. Passage of the Swamplands Act of 1850 commenced alteration of the Everglades ecosystem by encouraging drainage and development of wetlands. Twenty million acres (8 million ha) were transferred to Florida for drainage and development (Blake 1980). In 1907, legislation created the 18,518 km2 Everglades Drainage District. By 1917 West Palm Beach Canal, Hillsboro Canal, North New River Canal and Miami Canal were created (totaling 378.9 km of canal) (Fig. 8.2). These four canals fragmented the Everglades from Lake Okeechobee to the Atlantic Ocean. Importantly, the creation of canals and dikes accelerated and redirected flow in the northern reaches of the Everglades towards the Atlantic Ocean and away from the southwestern portion of the Everglades. Severe weather patterns and major floods in south Florida resulted in further control of the hydrologic flow of the region. In 1930, the Army Corps of Engineers and the Okeechobee Flood Control District joined state and federal forces to increase lake discharge, improve canals and enlarge levees in an attempt to better regulate flow and reduce hurricane flood conditions. Additional canals and new levees inhibited overflow from Lake Okeechobee into the dense sawgrass peat beds to the south. The natural pulse of this freshwater system was subdued in an effort to further accommodate farming and inhabitance of the coastal ridge. During this period of development and water control, humans were rapidly migrating to Florida. Populations increased an order of magnitude from 22,961 in 1900 to 228,454 by 1930 (Light and Dineen 1994). Sugarcane production doubled from 410,000 tons to 873,000 tons between 1931 and 1941 (Light and Dineen 1994). It was not until 1947 that conservation of the Everglades became an important part of the picture with the opening of the Everglades National Park (ENP). During this same year Florida experienced extreme floods due to torrential rains and high seasonal water levels. Consequently in 1948 the Central and South Florida Project for Flood Control and Other Purposes was formed to provide flood control and water supply, prevent saltwater intrusion and protect fish and wildlife (USCOE 1994). The Army Corps of Engineers flood control design used levees, water storage areas, channel improvements and pumps to control water flow and storage. The four initial canals were deepened within the Everglades Agricultural Area (EAA) to increase flow of floodwaters entering the Water conservation Areas (WCAs). Additionally a series of levees were installed on the eastern perimeter of the northern Everglades, parallel to the Atlantic coastal ridge (Fig. 8.2). This became the eastern boundary of the WCAs. The levees were designed to prevent flow into urban areas. Simultaneously it altered flow into the receiving aquatic systems including the ENP. Fortunately the series of levees created an eastern boundary that has prevented further encroachment of agriculture and urban sprawl. Nonetheless agriculture and urban development continued to expand south of Lake Okeechobee. An additional 700,000 (282,000 ha) acres of rich wetland soils were transformed into the EAA and by 1972 over 200,999 acres (81,000 ha) were dedicated to sugarcane, producing 800,000 tons of sugar annually (Light and Dineen 1994). In response to flood control and advances in

196

K. Hofmockel et al.

Fig. 8.2 Map showing the major canal systems dug originally in 1917 along with the newly constructed eastern perimeter levee system constructed in the early 1950s (data from Light and Dineen 1994)

8 Effects of Hydrologic Management Decisions on Everglades Tree Islands

197

refrigeration and mosquito limitation, human populations grew from 750,000 to 4,000,000 between 1950 and 1990 in the region of Dade, Broward, Palm Beach and Monroe counties (Light and Dineen 1994). Between 1965 and 1973 large portions of the Central and Southern Florida project were completed to improve water supply, especially for the ENP (Fig. 8.3). With these structures all in place and with the water management practices of the US Army Corps of Engineers (USCOE) and the South Florida Management District annual wet-dry cycles inherent to the south Florida peninsula were seriously disrupted by the 1970’s (see Chap. 2). Extended inundation between 1960 and 1980 substantially impacted plant communities of WCA 2A (Fig. 8.4; Worth 1983, 1988; see Chap. 9). Long-term landscape scale implications were not the focus of water management decisions during the 1970 –1990 time period. Instead land and water use decisions focused on immediate, local issues, particularly episodic weather events and agricultural and urban development needs. Current management of the Everglades is now focusing on restoring ecosystem functions lost on account of fragmentation and impoundment of the natural system due to the extensive dike and canal systems now in place (Figs. 8.2 and 8.3, see Chap. 26). Proposed solutions are focused at a large scale, considering the effects of the EAA on the WCAs, the nationally protected ENP, and ultimately Florida Bay (Porter and Porter 2003, CERP 2005, Fig. 8.1, Chap. 2). Understanding the consequences of historic changes in landscape patterns and hydrologic conditions can help managers as they develop plans both to preserve the integrity of what remains of the historic Everglades wetland ecosystem and to restore the altered areas of the Glades. Understanding the consequences of historic changes in landscape patterns can also aid managers in developing plans to preserve the integrity of what remains of the historic Everglades wetland ecosystem.

8.3.3 Pattern and Process “Landscape ecology deals with the effects of the spatial configuration of mosaics on a wide variety of ecological phenomena” (Weins et al. 1993). Pattern is created in a landscape by the physical template, biological processes, and disturbance regimes. Understanding the spatial results of system dynamics allows evaluation of processes affecting large areas of land. Sampling large areas requires extensive field work. Linking spatial arrangement to ecological processes provides an alternative, efficient means of monitoring ecosystem response in large-scale systems. To conserve ecological processes, an ecosystem approach is essential (Franklin 1993). WCA-2A is only a small piece of the greater Everglades puzzle, yet the ecological processes of this region reach far beyond the boundaries of WCA-2A, influencing the larger, interconnected system. In this study, the spatial mosaic of tree islands was examined in WCA-2A in 1953, 1980, and 1995 to assess the effects of altered hydrologic patterns (Fig. 8.4). A priori there are multiple possible outcomes for landscape pattern change in WCA-2A over a time series (Fig. 8.5a–j).

198

K. Hofmockel et al.

Fig. 8.3 Major canal and levee systems used to compartmentalize the Everglades by 2000. Also shown are the main pumping stations, spillways, weirs, hurricane gates and gated culverts (source: SFWMD maps)

Fig. 8.4 The mean daily water elevations in WCA-2A 1952–1996 due to SFWMD water management schedules, which resulted in decades of extremely high and low water levels (Richardson et al. 1997b)

Fig. 8.5 Potential spatial relationships for changes in the Everglades due to changes in water flow patterns and diking in WCA-2A

200

K. Hofmockel et al.

The total number of tree islands may increase if new tree islands form faster than older tree islands disappear (Fig. 8.5c). Tree island loss will result if either the rate of disappearance is greater than the rate of formation, or if management decisions result in the end of new tree island formation (Fig. 8.5b). If both loss and gain of tree islands occur at equal rates, the total number of tree islands will remain constant (Fig. 8.5d), but the location and size of tree islands will vary with time as islands are lost and new islands are formed. Channelization of WCA-2A may have caused a change in the size distribution of tree islands (Fig. 8.5e–g). This would affect the total area of tree island habitat available to northern Everglades biota. If increased water depth affects small and large tree islands differently, the landscape may be dominated by a specific tree island size (Fig. 8.5f, g). If the impacts of channelization on tree islands are independent of island size, the landscape should affect both size classes equally (Fig. 8.5e). Whether a single large tree island or several small islands is most beneficial is speciesdependent. Given equal number of islands, the larger the habitat the larger the population the region can support (Weins 1993; Sklar and van der Valk 2002). Within WCA-2A, water impoundment may cause different tree island spatial distributions than open water flow would. Location of islands determines the connectivity of the landscape. Small clumps of islands may be functionally similar to single tree islands of equal area. Small islands alone are isolated from exchange with other islands due to large interisland distances (Fig. 8.5i, j). If channelization affects tree island location, specific clumps of islands may persist over time in isolation from the remaining affected landscape (Fig. 8.5h, i). If instead, the effects of impoundment are equal throughout the landscape, islands may be left in isolation with little regional clumping (Fig. 8.5j). Tree island location and proximity to each other will determine the connectivity of the landscape. On the basis of the outcome, tree island pattern can be related to ecosystem functional changes. In the case of WCA-2A, many functions have not been directly measured over time. For example, the role of tree islands in biogeochemical cycling and decomposition is absent in the literature. However, changes in tree island number, size, and shape can be related to community diversity and habitat. Spatiotemporal landscape pattern has been demonstrated to strongly affect wildlife habitat function (Harrison and Fahrig 1995; Sklar and van der Valk 2002). The Duke study was aimed to quantify the spatial characteristics of tree islands and to investigate the relationship between tree island patterns and documented functional changes in WCA-2A, particularly wildlife habitat.

8.4 8.4.1

Methods Site Description

This research focuses on Water Conservation Area 2A (WCA-2A), a 42,707-ha impoundment created between 1948 and 1961 and managed by the South Florida Water Management District. The predominant community types of WCA-2A

8 Effects of Hydrologic Management Decisions on Everglades Tree Islands

201

include open water sloughs, tree islands, sawgrass marshes, and wet prairies (Loveless 1959). WCA-2A contained both small and large tree islands prior to impoundment (Loveless 1959). They are a fundamental structure in the Everglades, particularly in the remnant northern portions now impounded in the Loxahatchee Wildlife Refuge (WCA-1) and in WCA-2A and WCA-3A. In geologic terms, tree islands are a recent landscape structure. Soil cores from WCA-1 date woody plant material characteristic of tree islands back to 1,300 years before present (Gleason and Stone 1994; Chap. 2).

8.4.2

Imagery Analysis of Tree Islands

Duke researchers took the following steps to identify tree island number, size, shape, and location over time: defined characteristics unique to tree island structures in black and white imagery; located tree islands in winter 1950, 1980, and 1995 island coverages using USGS hydrography and road maps; and determined summary statistics including total tree number of tree islands, island area, size distribution of islands, nearest neighbor distance, core to edge ratio, and orientation. Tree islands were identified in the U.S. Department of Agriculture Soil Conservation Service aerial photos flown by Park Aerial Surveys, Inc. of Louisville, KY. County index sheets and the original 60.96 × 81.28 cm (24 × 32 in.) photographs of Broward County, Florida and Palm Beach County Florida were used for tree island identification in 1953 (1:20,000) and 1980 (1:40,000). A tree island was defined as the entire structure including the sawgrass perimeters and tails. This analysis included bayheads and willowheads as tree islands. Initially, tree islands were identified in the winter 1995 TM imagery using bands 3, 4, and 5. Results were compared to published vegetation classification of WCA-2A (Rutchey and Vilcheck 1994). Comparison of the classified TM image to a subsection of the 1995 digital panchromatic aerial photo revealed the shape of tree islands was adequately distinct to identify historic tree islands in black and white imagery. It was not possible, using the black and white aerial photographs, to distinguish between islands supporting live vegetation and those no longer living but still elevated above the surrounding emergent vegetation. Consequently, identification of historic tree islands was limited to the structural feature, not the presence of live vegetation. However, after loss of living vegetation, tree islands would lose their distinct signature. Those structures clearly identified probably supported live vegetation at the time of photography or within the recent past. The spectral signatures of the tree islands varied among photos, depending on the angle of light and the tonal range of the image. The heads of the tree islands had a relatively uniform texture, although the absolute color varied from image to image. Tree island heads tend to have a salt-and-pepper appearance. The tree island tail is a feature limited to larger tree islands. The small tree islands could be identified by the texture and consolidated rounded shape, while medium to large tree islands were identified by the texture of the head and the tear drop shape of the elongated tail (Fig. 8.6).

202

K. Hofmockel et al.

Fig. 8.6 Subsection of 1953 aerial photo index sheet for Broward County

Table 8.1 Tree island size classes based on natural breaks in 1953 area distribution Size class Size range (m2) 1 2 3 4 5

Islands less than 31,683 Islands greater than 31,683 and less than 89,232 Islands greater than 89,232 and less than 203,341 Islands greater than 203,341 and less than 432,664 Islands greater than 432,664

Observed islands were traced from index sheets onto Mylar using the original photos to enhance island identification. WCA-2A was split between the Broward County and Palm Beach County in the 1953 photo series. Islands were located on each half, then digitized individually on the digitizing tablet. Tree island classes (1-5) were created to classify islands according to size ranging from islands less than 31,683 square meters (class 1) to islands greater than 432,664 square meters (class 5) (Table 8.1).

8 Effects of Hydrologic Management Decisions on Everglades Tree Islands

203

The two 1953 and single 1980 tree island data layers were projected into the Universal Transverse Mercator (UTM) coordinate system. Transformation was executed using 15 points with known UTM coordinates for the northern 1953 tree island data layer, 14 points for the south 1953 tree island map, and 15 points for the 1980 tree island map. The northern and southern 1953 tree island maps were combined in a geographical information system (GIS; Environmental Systems Research Institute, Inc. 1998). The northern section was adjusted to geographically match the more accurate southern section. Duplicate tree islands caused by overlap of the northern and southern maps were removed. The 1995 tree island map was created by merging together 20 compressed photos downloaded from the TerraServer Web site (http://terraserver.microsoft.com/). The most northern image was rectified to the 1995 TM image. Each consecutive image was stitched to the preceding image using features within the overlap space for georectification points. The final digital map was comprised of 19 individual images stitched together in ERDAS Imagine. The composite image was georectified to the 1995 satellite image using the georectification tool in ERDAS Imagine with 126 georectification points. This rectified composite photo was used for tree island identification in the 1995 time step. Islands were recognized using the same identification rules. Unlike the 1953 and 1980 scenes, the 1995 black and white digital composite did not have accompanying small-scale photos. The winter 1995 TM image was used instead to enhance tree island identification. Tree island number, size, and edge characteristics were determined by downloading the tree island characteristics from attribute tables associated with the tree island maps created in ARC (Environmental Systems Research Institute, Inc. 1998). Nearest neighbor distance and patch density were calculated using Fragstats, a landscape analysis program that outputs a variety of spatial statistics (McGarigal and Marks 1995). Patch density counts the number of islands within a 100-ha area. It quantifies the “clumpiness” or connectivity of the landscape (Fig. 8.5). Nearest neighbor distance measures the minimum distance between an island and any other individual island. This interisland distance measures dispersion of tree islands.

8.5

Results

In 1953, tree islands covered much of the landscape in WCA-2A. There were many different-sized tree islands with dense clusters in the northern region (Fig. 8.7). By 1980, there were nearly as many large islands in WCA-2A as in 1953 but virtually no small islands (Fig. 8.7). Most of the islands in the central region had disappeared by 1980. Only one clump of small islands in the south east corner and a few clumps in the north remained by 1980. Mostly large islands persisted through 1995 and the remaining small islands disappeared by the mid 1990s (Fig. 8.7). The landscape pattern of tree islands was significantly affected by water impoundment in the northern Everglades, with the number of islands clearly diminishing over

204

K. Hofmockel et al.

Fig. 8.7 Time series analysis of tree island spatial patterns in WCA-2A

time (Figs. 8.7 and 8.8a). All islands appear more isolated after complete impoundment. Fragmentation of a landscape often results in habitat loss and isolation of remaining habitat (Meffe and Carroll 1994). Habitat loss and isolation due to fragmentation is illustrated in WCA-2A over the period evaluated. The total number of tree islands (341) diminished by 77% between 1953 (when the flow-through system was still intact) and 1980 (after complete impoundment) (Fig. 8.8a). Between 1980 and 1995 the number of tree islands diminished an additional 10%, resulting in a total tree island loss of 87%. This caused an almost complete loss of upland patches in the northern Everglades over the time period examined. More than 30% of the total tree island area was lost between the flow-through system (1953) and after 15 years of impoundment (1995) (Fig. 8.8b). This translates into a loss of over 6,200,000 m2 of upland habitat due to impoundment (Table 8.2). Almost 70% of the total remaining area is represented in the surviving 13% of the tree islands. This is because primarily small islands disappeared from the system. As the total number of islands decreased over time, the size class distribution shifted (Fig. 8.8c). To better quantify the effects of water impoundment on tree island size, islands were grouped by size (Table 8.1). The resulting size classes were applied to each time step map (Table 8.3). In 1953, most of the islands were less than 20,000 m2 in size (Fig. 8.7 and Table 8.3). Largest loss of tree islands was in the small size class (Fig. 8.8c), with only 4 of the initial 227 small islands remaining in 1995 (Table 8.3). During the 1953 period, 66% of tree islands were in the smallest size class (less than 31,683 m2) (Fig. 8.8c, Table 8.3). Fewer than 9% of the remaining tree islands were of the smallest size class (less than 31,683 m2) by 1995 (Fig. 8.8c, Table 8.3). Ninety-eight percent of small (less than 31,683 m2) tree islands disappeared between 1953 and 1995 (Fig. 8.8c). As island size increased within a time period, the impact of water impoundment on tree island size decreased. Only 3% percent of all tree islands in 1953 were in the

Fig. 8.8 (a) Total number of tree islands in time series analysis of WCA-2A; (b) total area of tree islands in WCA-2A; (c) size class histograms for WCA-2A tree islands in 1953, 1980, and 1995. Size classes are in square meters with class breaks at (1) 31,683, (2) 89,232, (3) 203,341, (4) 432,664, (5) greater than 432,664 square meters

206

K. Hofmockel et al.

Table 8.2 Change in tree island shape in time series analysis of WCA-2A Total Total Average Average perimeter (m) area (m2) perimeter (m) Year area (m2)

Edge-toarea ratio

1953 1980 1995

0.017 0.010 0.008

20,944,231 14,298,552 14,726,878

356,168 149,657 118,667

61,240 178,732 320,150

1,041 1,871 2,580

Table 8.3 Tree island characteristics according to size class Number Total Year Class of islands class area (m2)

Mean island size (m2)

Total perimeter (m)

1953 1980 1995 1953 1980 1995 1953 1980 1995 1953 1980 1995 1953 1980 1995 1953 1980 1995

12,193 14,763 11,842 50,767 56,982 74,632 130,175 125,270 154,997 281,865 351,830 315,391 590,337 632,698 789,144 61,419 178,732 320,150

120,538 12,486 2081 75,158 23,806 12,792 54,644 30,450 25,951 53,701 35,724 23,258 52,001 47,191 54,585 356,043 149,657 118,667

1 1 1 2 2 2 3 3 3 4 4 4 5 5 5 Total Total Total

227 22 4 60 20 9 25 16 13 17 11 8 12 11 12 341 80 46

2,767,780 324,780 47,368 3,046,026 1,139,644 671,686 3,254,387 2,004,325 2,014,964 4,791,711 3,870,129 2,523,132 7,084,045 6,959,674 9,469,728 20,943,949 14,298,552 14,726,878

largest size class (greater than 423,664 m2). By 1995, the large tree islands comprised 26% of the remaining tree islands. There was no loss in the large tree island sized class over the time series evaluated (class 5, Table 8.3). The second largest class (class 4) lost more than 50% of its total number of islands. Patch density and nearest neighbor distance were calculated to determine connectivity of the landscape. Density of islands and distance between islands are inversely related (Fig. 8.9a, b). Over the time series the patch density decreased by 81% from an initial patch density of 16 islands 100 ha−1 to a final density of three islands 100 ha−1. The greatest change in patch density occurred between 1953 and 1980 with a 66% decrease. This reflects the predominant loss of clusters of small islands with a mean size of 1.21 ha. As a result, each remaining island became more isolated, with nearest neighbor distance increasing from 253 m in 1953 to 874 m in 1995 (Fig. 8.9b). This translates into a threefold increase in interpatch distances. The greatest change in nearest neighbor distance took place during the 1953–1980 period with a change from 253 m to 628 m. Total perimeter, or habitat edge, decreased by 66% over the time series evaluated (Fig. 8.9c). Only one-third of the initial transition zones remained in WCA-2A by the final time step in 1995. The perimeter to area ratio examines the loss of core habitat

8 Effects of Hydrologic Management Decisions on Everglades Tree Islands

207

Fig. 8.9 (a) Density of tree islands for a time series in WCA-2A; (b) interpatch distances for tree island time series in WCA-2A; (c) total length of tree island perimeter in time series of WCA-2A

208

K. Hofmockel et al.

versus edge habitat. The ratio was calculated for each individual tree island and averaged over each year. The edge-to-area ratio decreased by 38% between 1953 and 1980 and another 14% between 1980 and 1995, resulting in a 52% loss in total edge-to-area ratio (Table 8.2) and a predominance of large islands possessing big core areas relative to the island perimeter. Due to the nature of aerial photo interpretation, it is difficult to quantify all error in the time series spatial analysis. In qualitative terms, it is important to note that the raw data used for the 1953 and 1980 analyses were of much higher quality than that used in the 1995 evaluation. Original 1:20,000 and 1:40,000 photographs were available for analysis for the first two time steps. The 1995 analysis was performed on digital imagery. The digital imagery was less detailed than the USDA photos. Although the majority of the 20 images were from February 1995, some of the available digital images were taken in January 1996. All images are winter scenes; however, the ground conditions vary between years, making spectral interpretation more challenging, particularly for islands located at the intersection of two or more images. As a result, photointerpretation of the 1995 image is less accurate than for the 1953 and 1980 time steps. Differences in imagery quality contributed to inaccuracy in the number and area of islands in 1995 as well as the edge-to-area values. Additionally, the digital images are in raster format (grid cells rather than vectors), making round edges of tree islands appear angular and further inhibiting the accurate recording of island shape and area. Classified TM imagery reveals that the 1995 photointerpretation includes many islands no longer supporting live upland vegetation. According to classified TM imagery, most of the islands in south WCA-2A are currently sawgrass strands with only dead trees remaining. Live tree islands do still exist in northern WCA-2A. In all time steps there was error in the nearest neighbor distances and mean patch densities due to georectification error. This error is greatest in the northern portion of the 1953 image and smallest in the 1995 scene. During the earliest time step, the Everglades were still relatively undeveloped and lacked structures with known geocoordinates. This made finding an even distribution of known points for georectification quite difficult, and consequently the error is higher. However, a field check on several islands indicated that overall errors were minor when compared to the tree island losses found and the trends reported.

8.6

Discussion

The Everglades ecosystem was historically an immense freshwater wetland extending from the overflow of Lake Okeechobee to the tip of the Florida peninsula. WCA-2A consisted of four principal vegetation types: sawgrass marsh, wet prairies, slough aquatic communities, and tree island communities (Loveless 1959). Aside from tree islands, the physiognomy was homogeneous. Draining large sections of the natural system and installing flood control structures encouraged population growth in this region and altered the Everglades landscape

8 Effects of Hydrologic Management Decisions on Everglades Tree Islands

209

mosaic. Land-use changes (Fig. 8.3) reduced the spatial extent of the ecosystem (Fig. 8.1), compartmentalized the landscape, and disrupted the natural hydrology of the freshwater wetland system (see Chaps. 2 and 7). Previous analyses of this area indicate that implementation of water management control structures resulted in a more homogeneous landscape with only tree islands and sawgrass persisting (Worth 1988). Homogenizing landscape patterns generally results in reduced species diversity and reduced habitat diversity (Urban et al. 1987). More heterogeneous landscapes are better able to support disturbance by providing species with a diversity of habitat types at all times (Meffe and Carroll 1994). If a patch of a particular habitat is destroyed or altered by disturbance, species dependent on that habitat type may find refuge in repeated undisturbed patches of this habitat in a heterogeneous landscape. This study indicates that the area of total tree island habitat in WCA-2A was reduced by 30% after complete enclosure of WCA-2A (Fig. 8.8b). Tree islands support more species of vegetation and wildlife than all other habitats in the central Everglades (SFWMD 1999). Loss in the total number of tree islands (Fig. 8.8a) reduced the area supporting upland tree species that are otherwise not present in WCA-2A. This in turn affected wildlife that use the trees regularly, including the great blue heron, white ibis, great egret, and the Federally endangered wood stork (Hoffman et al. 1994). Studies in WCA-3 found more bird species in bayheads and willow heads than any other marsh habitat (Gawlik and Rocque 1998). Total bird abundance and species richness was largest on tree islands compared with marsh sites (SFWMD 1999; Sklar and van der Valk 2002). Habitat loss limits species dependent on multiple upland patches. Shallow receding water is associated with high reproductive and foraging success (SFWMD 1999). Loss of nesting sites or breeding habitat diminishes the population the region will support. The effects of lost habitat have been documented in snail kite (Rostrhamus sociabilis) populations in the WCAs. Snail kites require woody vegetation for nesting (Bennetts et al. 1994). In regions lacking suitable habitat, nesting failure has been greater than 95% (Bennetts et al. 1994). Dramatic decreases in snail kites documented in WCA-2A during the late 1980s and 1990s have been related to changes in hydrology (Fig. 8.10) but may in fact be more directly related to loss of tree island habitat (Figs. 8.7 and 8.8a–c). Terrestrial species are also affected, including white-tailed deer, muskrats, and rice rats. Lost habitat has especially affected species that use tree islands exclusively, such as tree snails (genus Liguus) (Robertson and Frederick 1994). Fortytwo species of reptiles and amphibians, 143 birds, 20 mammals, 29 butterflies, and 34 ant species have been associated with tree islands (Bass 1998). This study indicates that increased water depth and duration (Fig. 8.4) had the greatest affect on small tree islands (Fig. 8.8c). The WCA-2A landscape has become a vast region of cattail and saw/grass, some sloughs, and a few large tree islands (Fig. 8.7). This reduction in small islands and loss of connectivity in the landscape is demonstrated by the 83% decrease in tree island density (Fig. 8.9a) and the threefold increase in interpatch distance (Fig. 8.9b). Loss of many small islands results in a greater reduction in habitat connectivity than loss of a few isolated large islands (Fig. 8.5h–j).

210

K. Hofmockel et al.

Fig. 8.10 Minimum monthly water stages (top) and number of snail kites reported during annual surveys (bottom) for WCA-2A (Bennetts et al. 1994)

Loss of connectivity in the landscape increases dispersal distance and habitat isolation. Decreasing the ability of upland species to reproduce and spread negatively affects biological productivity and may prevent the development of new islands. Distribution of mobile organisms depends on ease of movement and habitat choice (Weins et al. 1993). Decreased connectivity makes travel between islands more

8 Effects of Hydrologic Management Decisions on Everglades Tree Islands

211

cumbersome for mobile species. This is particularly important during high water periods, when terrestrial species are in search of immediate refuge. Loss of primarily small tree islands with large edge-to-area ratios is reflected in 89% perimeter loss during the time series evaluated (Fig. 8.9c). Survival of only large tree islands is seen in a 52% reduction in the edge-to-area ratio. Loss of edge minimizes transition zones in the ecosystem. Tree islands are the sole source of topographic variability in WCA-2A (Loveless 1959). Topography creates a unique region of fluctuating hydrologic conditions. This transition zone is particularly important in providing wading bird foraging habitat during early season drydown (DeAngelis 1994). Shallow water edges are biologically productive regions important to the Everglades food web (DeAngelis 1994). During drought, central sloughs, ponds, and alligator holes become isolated regions of inundation and greatly increase the water edge area in the region. These pools concentrate the food supply and provide important feeding grounds during the dry season. During the wet season, tree islands augment the water edge area in the WCA-2A region, providing foraging and nesting sites for migratory birds (Holling et al. 1994). Many wading bird species (especially the endangered wood stork) are dependent on such areas for concentrating food supplies. Wading bird populations in the 1930s and 1940s were ten times current populations (Holling et al. 1994). Although no species have been lost (USACE 1994), the function of wildlife habitat has been affected by loss of tree islands. Perimeter reduction is severe (66% loss) in WCA-2A over the time series analyzed because impoundment disproportionately affected the smaller islands that have the highest perimeter to area ratio. In a similar study interpretation of twelve 1:60,000 aerial photos from November or December 1950 and twenty 1:40,000 color-infrared diapositives from December 1990 and January 1991 in WCA-1 revealed 1,144 tree islands in the 57,234-ha WCA-1 refuge area (Brandt 1997). The highest tree island density occurred in the north and east. In WCA-2A, greatest tree island density was in the north and southeast. To evaluate spatial and temporal patterns of the Loxahatchee Wildlife refuge, tree islands were clumped into size groups based on the distribution in WCA-1 (Brandt 1997). These size groups were compared to the size distribution in WCA-2A over the time series investigated (Table 8.4).

Table 8.4 Comparison of tree islands in WCA-1 and WCA-2A using size classes from Brandt (1997) WCA-1 1990–1991 WCA-2A 1953 WCA-2A 1980 WCA-2A 1995 Size class (ha) < 1.73 > 1.73 and < 3.82 > 3.82 and < 5.10 > 5.10 Total

Number

Percent

Number Percent

Number Percent Number Percent

983 34

93.18 3.22

170 72

49.56 20.99

12 13

15.19 16.46

0 2

0.00 4.55

6

0.57

19

5.54

3

3.80

1

2.27

32 1,055

3.03 100.00

82 343

23.91 100.00

51 79

64.56 100.00

41 44

93.18 100.00

212

K. Hofmockel et al.

Unlike WCA-1, the tree island size distribution at the most recent time step in WCA-2A has virtually no islands in the smaller size class and over 90% of the tree islands in the largest size class. It is evident that the size class distribution in WCA-2A was most similar to the 1991 size class distribution of WCA-1 during the 1953 time step, before complete impoundment and altered hydroperiods and flow regimes. Only in the 1953 time step do the majority of islands fall into the small island size class and medium values for the medium and extra large categories exist. This indicates that WCA-1 and WCA-2A are currently neither spatially nor hydrologically similar. WCA-1 is a national wildlife refuge located on a topographic high for the region and is primarily rainfall driven. Rainfall contributes 54% of the water budget in this wildlife refuge (SFWMD 1992). Moreover, WCA-2A is a water supply and flood control area with currently a very different hydrologic pattern than that of WCA-1 (Chap. 7) The massive loss of tree islands in WCA-2A reflects this altered hydrologic regime.

8.7

Future Management Plans

The most recent Comprehensive Plan for the Everglades addresses many of the current environmental and social issues of the south Florida region. The Comprehensive Plan intends to replumb the Everglades to mimic natural conditions. It is a $7.8 billion, 20year plan with annual operation, maintenance, and monitoring costs of $182 million (USACE 1999). Flow-through conditions similar to the 1953 time step of this study will not be restored in WCA-2A as the plan is currently written. However, there are ongoing modifications in design and construction of the Modified Water Deliveries to the Everglades National Park (US Army Corps of Engineers, 1999). WCA-2A will have better timing of water deliveries due to EAA storage reservoirs that will capture and manage flood releases from the EAA to the WCAs, but current plans will not allow for the restoration of the earlier tree island populations in WCA-2A. Flow-through conditions will be returned to Water Conservation Area 3 (WCA-3) through decompartmentalization. The Miami Canal in WCA-3A will be backfilled and sections of the Tamiami Trail will be replaced with bridges and culverts to recover sheet flow. Water storage will be augmented to support increasing human populations. Approximately 87,800 ha of new reservoirs and wetland water treatment areas will be created (USACE 1999). What effects these reservoirs will have on water flow into the WCAs is not fully understood at this stage.

8.8

Conclusions and Lessons for Restoration

Past management actions have severely impacted the Everglades landscape. Construction of canals and levees has changed the structure of the greater Everglades, fragmenting the northern freshwater marshes. Water impoundment has drastically reduced the quantity and quality of upland habitat in the northern Everglades,

8 Effects of Hydrologic Management Decisions on Everglades Tree Islands

213

especially in WCA-2A. Fragmentation of the landscape has primarily affected small islands, leaving the impacted landscape with only large isolated tree islands. The destruction of small islands with large edge-to-area ratios has negatively impacted wildlife, specifically wading birds that use hydrologic transition zones as foraging habitat. Additionally loss of small islands has decreased heterogeneity and reduced connectivity of the landscape. Although management activities have successfully stored and supplied water for agricultural, urban, and industrial use, they have not succeeded in meeting management objectives of enhancing fisheries and wildlife management. Improving WCA-2A fisheries and wildlife requires restoration of ecological processes, primarily hydrologic flow. Restoration requires hydrologic flow and storage reflecting climatic conditions. Disconnecting the vegetation from ambient climatic conditions and, in turn, the natural disturbance regime has severely disrupted the vegetation structure of WCA-2A. The link between structural changes and ecological processes is particularly evident in diminished avian populations in WCA-2A. The need to link the ecosystem with climatic conditions is supported by tree island analysis from WCA-1 (Brandt 1997), where landscape heterogeneity persists in the face of fragmentation. Unlike WCA-1, the hydrologic flux and storage of WCA-2A is determined not primarily by climate, but by water management schedules. The time series spatial analysis demonstrated flowthrough conditions of 1953 supported a heterogeneous landscape with tree islands throughout WCA-2A. Although the current Comprehensive Plan will reconnect flow in the western portions of the Everglades, the plan does not focus on reestablishing a flow-through system in WCA-2A. The only improvements to WCA-2A currently in the Comprehensive Plan are changes in the timing of water deliveries. Reconnecting the compartments of this watershed and returning to more natural conditions are necessary to improve vegetation structure and to restore conditions conducive to tree island formation. The current Comprehensive Plan for south Florida calls for the adequate monitoring of “all trophic levels, spatial and temporal scales among hydrological, ecological, water quality and physical components” (USACE 1999). Landscape interpretation is a useful tool for achieving this objective especially in south Florida, where local monitoring efforts are logistically complicated and expensive. This study demonstrates that changes in large-scale ecological processes due to water management decisions were reflected in reorganization and loss of the spatial pattern of tree islands in the WCA-2A landscape over the period evaluated, rendering spatial analysis of tree islands a useful tool for future monitoring efforts in the restoration of the Everglades. It is important to recognize that human population growth was the driving force of development in south Florida and fragmentation in the Everglades ecosystem. To successfully achieve management goals, future management plans must consider the carrying capacity of the Everglades ecosystem. South Florida’s current population of 6 million is projected to double over the next 50 years (Stevens 1999). Future demand for water resources is expected to exceed the available regional freshwater resources (USACE 1999). With current population growth the demand for fresh

214

K. Hofmockel et al.

water could increase from the current one billion gallons of water per day to two billion gallons per day by 2050 (USACE 1999). Successful restoration of this ecosystem is contingent upon satisfying many, often conflicting demands. The outcome of the Comprehensive Plan is of global importance. The Everglades is internationally recognized as an important and unique ecosystem. Restoration and preservation of its ecological integrity will set the precedence for future efforts to preserve large natural regions in the face of exponentially increasing human populations. Tree island density and patterns can be sentinels to the effectiveness of our restoration efforts.

9

Macrophyte Community Responses in the Everglades with an Emphasis on Cattail (Typha domingensis) and Sawgrass (Cladium jamaicense) Interactions along a Gradient of Long-Term Nutrient Additions, Altered Hydroperiod, and Fire Curtis J. Richardson, Ryan S. King, Jan Vymazal, Edwin A. Romanowicz, and James W. Pahl

9.1

Introduction

One of the major problems in the Everglades of south Florida during the past 100 years is the severe reduction or loss in area of communities like open-water sloughs, wet prairies, and sawgrass monocultures, known as sawgrass plains (Davis 1994; Richardson et al. 1999). These communities historically dominated large portions of the Everglades (Davis 1943; see Chap. 12). The majority of this earlier habitat loss, especially the sawgrass plains, was due to agricultural development south of Lake Okeechobee (Davis 1994). Approximately 50% of the original 900,000 ha Everglades has been converted to agricultural land and urban development. The remaining native plant communities are currently being further reduced in area, especially in the northern portions of the Everglades and along canal input structures in the southern Glades. This is caused by a number of interacting factors including massive influx of water and nutrients (especially P), disturbance of habitat, invasion of exotic species, altered hydroperiod due to pumping and diking and changed fire regimes (Craft and Richardson 1993a; Urban et al. 1993; Davis 1994; Richardson and Zahina 1995; Newman et al. 1996; Vaithiyanathan and Richardson 1997; King et al. 2004). Collectively, these studies suggest that multiple factors are responsible for the alteration and maintenance of plant community structure in the Everglades (Fig. 9.1). Historically, the primary factor controlling the long-term development of all Everglades communities is climate. The native seed bank was responsible for the regeneration of endemic plant communities once they were disturbed or altered (van der Valk and Rosburg 1997). Massive landscape development has resulted in regulated hydroperiods (i.e., the number of days that the Everglades ecosystem has standing water at or near the surface) and altered hydropatterns (the distribution of water within the wetland), which in turn have changed fire frequency patterns and fire intensity. Increased P and N loadings from agriculture and urban runoff and introduced 215

216

C.J. Richardson et al.

Fig. 9.1 Primary factors controlling Everglades wetland plant community succession. Natural controlling factors are shown in bold type and human altered factors are shown in outline type. The size of the arrows depicts the magnitude of impact. The percentage of the Everglades area that is impacted by each factor is shown and is based on area surveys in 2000

exotic species in the early 1900s have all significantly affected plant and animal communities of the Everglades and the Water Conservation Areas (WCAs) of today (Craft and Richardson 1993a; Davis and Ogden 1994a; DeBusk et al. 1994; Qualls and Richardson 1995; Vaithiyanathan and Richardson 1997). Importantly, it has been demonstrated by numerous studies that P is the limiting plant nutrient in the Everglades (Steward and Ornes 1983; Koch and Reddy 1992; Richardson and Vaithiyanathan 1995; Craft and Richardson 1997; Richardson et al. 1999; Noe et al. 2001). The main difficulty for ecologists is in separating the influence of primary climate-driven factors like rainfall, hydroperiod, and fire from the secondary human factors of drainage and flooding, nutrient additions, site disturbance, and exotic species invasions. Moreover, the influence of anthropogenic inputs of nutrients and water varies greatly in each portion of the Everglades depending on proximity to canal input structures, mode of delivery (i.e., point or nonpoint source), and whether water delivery is seasonally pulsed or continuously released. In other regions like WCA-3A vast stands of exotic species, such as Melaleuca quinquenervia and Schinus terebinthifolius (Brazilian pepper) provide a seed source for the everincreasing invasions by these species, although intensive and expensive control measures are underway by state and federal agencies.

9 Macrophyte Community Responses in the Everglades

217

The extensive canal-and-levee system that compartmentalizes the remnant Everglades serves as a conduit for P from the Everglades Agricultural Area (EAA), and water-control structures along the canals function as point sources of water and nutrients to downstream portions of the wetland ecosystem (Fig. 8.3). In areas near water-control structures, P has found to be mainly responsible for the transformation of Cladium jamaicense Crantz (sawgrass) stands and open-water sloughs to dense stands of invasive Typha domingensis Pers. (cattail) (Davis 1991; Urban et al. 1993; Newman et al. 1998). Typha distribution and growth is positively correlated with both soil and water total P and is limited in areas with low P (Craft and Richardson 1997; Doren et al. 1997; Miao and Sklar 1998; Miao et al. 2000). Mesocosm studies also have demonstrated that Typha is more competitive than Cladium under high-P conditions (e.g., Newman et al. 1996). However, field fertilizer experiments without hydrologic changes have been unable to show that adding P alone results in the invasion of Typha in healthy stands of Cladium (Craft et al. 1995; Chiang et al. 2000). These experimental findings suggest that other factors, such as hydropattern (e.g., Toth 1987; Urban et al. 1993; Newman et al. 1996), fire (e.g., Gunderson and Snyder 1994; Urban et al. 1993; Newman et al. 1998), or cations such as sodium (Craft and Richardson 1997) or sulfate (Chap. 6), may be important synergists in aiding Typha expansion. However, intercorrelations and spatial autocorrelation of multiple factors have made it difficult to isolate the linkages between specific environmental variables and observed changes in vegetation patterns. Water Conservation Areas 2A and 3A (WCA-2A and WCA-3A) of the northern and central Everglades are the focus of our vegetation studies. WCA-2A, the main focus of our nutrient gradient research efforts, is a remnant (43,280 ha) of the vast sawgrass marsh and slough community that once existed south of Lake Okeechobee (Fig. 5.1). The plant communities of this region and most portions of the Everglades were surveyed during the first half of the twentieth century (Harshburger 1914; Davis 1943; Loveless 1959). USGS aerial photographs of the area in 1953 show the northeastern section of WCA-2A to be a mosaic of sloughs and marsh, undifferentiated from other parts of the eastern portion of WCA-2A or the Loxahatchee National Wildlife Refuge to the north. A long-term analysis (15–20 years) of permanent vegetation plots that were established throughout the Everglades during the period 1965–1971 and resurveyed in 1984–1987 supports the invasion of cattail in the enriched areas of northern portions of WCA-2A (Davis 1994). However, the analysis documents a significant expansion, not reduction, of sawgrass communities into wet prairie/sloughs in other areas due to reduced hydroperiod during the measurement interval. Since the 1970s, the plant species composition in the northern-enriched areas of WCA-2A has shifted from a sawgrass and open slough-dominated fen wetland to a marsh with nearly 5,000 ha in pure cattail (T. domingensis) (Jensen et al. 1995). Davis (1989) suggested that the expansion of cattail in WCA-2A was accelerated by nutrient-enriched runoff that has been pumped into WCA-2A from the Hillsboro Canal. According to Davis (1991), cattail is able to assimilate more N and P than sawgrass. As a result, cattail is able to grow rapidly, further increasing its advantage of large reserves of energy, carbon, and nutrients. Thus, the ability of cattail to respond to yearly pulses of nutrients favors the

218

C.J. Richardson et al.

proliferation of cattail over sawgrass and other species in nutrient-enriched areas of the Everglades. An alternative hypothesis is that cattail invasions are the result of two interacting factors: altered hydrology and increased nutrients. This hypothesis was tested in pure sawgrass and slough stands that have been fertilized for six consecutive years (Chap. 20). Results indicate that N and P fertilization alone or in combination do not result in cattail invasions in pure sawgrass or slough stands. A third hypothesis is that cattail invasions first require a habitat disturbance, which removes or weakens the existing vegetation. A disturbance study established in 1991 supports this hypothesis (Chap. 21) since cattail only established in disturbed sites where native vegetation had been removed. However, cattail only grew and expanded in P-enriched plots. A study by Craft and Richardson (1997) also suggested that nutrient enrichment accelerates cattail encroachment, but increased soil concentrations of both P and Na were significantly related to increased cattail frequency. Thus, it is clear that the presence of P is required for a shift from the native vegetation to cattail but other controlling factors (i.e., hydrology or disturbance) are necessary to help initiate the invasions and expansion of cattail. However, once established in areas with high soil or water P concentrations cattail is the aggressor and will become the dominant species, provided ample water is supplied. The rate of cattail invasions and the timeframe for conversions are unclear. More information is needed on the dynamics of cattail expansions and contractions under reduced nutrient loadings and hydrologic shifts. WCA-3A, the largest of the conservation areas (199,608 ha), is the most pristine area of the Everglades north of the Everglades National Park (ENP). It has a large diversity of plant communities that range from mixed marl prairie and sawgrass to some of the best remaining slough–sawgrass–tree island mosaics (Davis et al. 1994). Tree islands, tropical hammocks, and willow heads are found throughout expansive marl prairies, especially in the south. However, the percentage of area covered by these islands is less than 5%. The northern portions of WCA-3A differ considerably from the southern areas in terms of drier hydrologic conditions and less marl soil. Also WCA-3A has experienced considerably less nutrient pollution from runoff, and the soils are considerably reduced in total phosphorus (TP) concentrations (Chap. 24). Davis et al. (1994) have argued that WCA-3A has not shown the increased encroachment of invasive species seen in WCA-2A because of less disturbance. To further assess changes in these areas and determine if nutrient gradients existed in WCA-3A, we also conducted a series of gradient studies for plant composition changes. Thus, long-term vegetation plots along nutrient and hydrologic gradients were used to assess changes in vegetation. These permanent plots are important since few studies have closely examined patterns of entire macrophyte communities in response to these human influences. The general observation has been that high-P areas are dominated by monotypic stands of dense Typha (e.g., Jensen et al. 1995, Rutchey and Vilchek 1999), resulting in a relatively homogeneous landscape pattern when compared to the reference Cladiumslough mosaic (Obeysekera and Rutchey 1997; Wu et al. 1997). However, this observation has been largely based on photointerpretation of satellite imagery, which is limited in spatial resolution and is not appropriate for assessing fine-scale patterns

9 Macrophyte Community Responses in the Everglades

219

in species composition (Obeysekera and Rutchey 1997; Richardson et al. 1997b). Recent field studies have indicated that many other macrophyte species coexist with Typha in high-P areas, and that diversity is actually greater near canals, due to the increase in invasive species, than in interior-wetland locations (Doren et al. 1997, Vaithiyanathan and Richardson 1999). However, little is known about the spatial patterns of these communities, the environmental factors that are responsible for generating these patterns, or the influence of such spatial patterns on autogenic environmental variation (e.g., biogeochemical properties of peat soils) (Noe et al. 2001). Thus, despite the large body of research that continues to expand our understanding of the Everglades ecosystem, no long-term field study has attempted to untangle the independent role of multiple abiotic factors in community-level changes in vegetation composition. Thus, we used vegetation data from long-term plots to evaluate linkages among spatial factors, environmental conditions, and vegetation composition along a gradient of anthropogenic influence in the northern Everglades (both WCA-2A and WCA-3A). To quantify the effects of both nutrient and hydroperiod alterations on plant community structure and to determine the rate of cattail and sawgrass invasions, we established a series of permanent long-term vegetation grids along a 10 km by 10-km nutrient and hydrologic gradient in WCA2A in 1990 and also in WCA-3A in 1998 (Fig. 5.1). Our objectives were to (1) present an assessment of plant community structure and species composition changes over a decade using long-term fixed plots along nutrient, hydrologic, and fire gradients (temporal study); (2) use a spatially intensive, multiscale sampling approach and multivariate analysis to examine vegetation–environment linkages along nutrient, hydrological, and fire gradients (spatial study); and (3) integrate results of these studies to determine to what degree species change can be related to these three anthropogenic factors. We contend that an understanding of the characteristic scales and patterns of vegetation–environment linkages is critical for the Everglades and other ecosystem restorations if these efforts are to effectively facilitate natural processes at multiple scales (Holling et al. 1994; Redfield 2000).

9.2 9.2.1

Methods Plants and Soil

In June–October from 1990 to 2002, macrophytes were surveyed every 2 years along three permanent 10-km transects (A1–6, C1–6, and D1–6) in WCA-2A to quantify plant species composition as well as soil and water nutrient enrichment south of the Hillsboro Canal (Fig. 5.1). (For details on soil and water sampling at these sites see Chap. 6.) The distribution of cattail, sawgrass, other species, and the amount of open bare ground was determined to estimate presence and absence of species, percent frequency of species, and cattail and sawgrass distributions over time as a result of nutrient enrichment and hydrologic conditions. At each site, four 10-m long transects were established, extending north, south, east, and west from a

220

C.J. Richardson et al.

central point. This method of surveying in each direction allowed us to subsample an area of 400 m2. On each transect, ten points (10 cm in length) were laid out in 1-m intervals. The total number of individual points sampled was 4,320. Species composition was determined using the point intercept method (Mueller-Dombois and Ellenberg 1974) and expressed as percent frequency of occurrence where Frequency = total number of points in which a species occurs × 100 Total number of transect intervals In addition, we added a western transect in WCA-2A in 1998 to assess changes over time in an area not subjected to nutrients. We also added two transects in northern WCA-3A in 1998 to assess changes in plant composition and nutrient storage in areas with reduced hydrologic flow.

9.2.2

Soil Analysis

Soils (0–10 cm) were collected during 1991–1998 from all sampling sites using a rectangular box corer (7.5 × 7.5 cm) with a removable side (C.J. Richardson, unpublished design), which minimizes the compression of organic soils. Samples were placed in polyethylene bags and stored on ice in a cooler during field collection, then refrigerated (4°C) upon return to the laboratory. All soil analyses were performed within 2 months after collection. The samples of soil were dried, ground, sieved through a 2 mm diameter mesh screen, and analyzed for total N and total P. Total P was determined by digesting 100 mg of soil in nitric/perchloric acid (Sommers and Nelson 1972) and measuring phosphate in the digest using a TRAACS 800 autoanalyzer. Total N and C were measured in 10–15 mg of dry samples using a Perkin-Elmer 2400 CHNS analyzer.

9.2.3

Hydrology

To assess how hydrology has affected the vegetation patterns in WCA-2A, it is important to characterize the hydropattern including water depth, flood duration, and periods of dryness as well as the frequency of occurrence of water depths throughout parts of WCA-2A. Unfortunately, these determinations require longterm information on water depth from numerous sites. The SFWMD’s station 2A-17 is the longest continuous (> 20 years) water elevation monitoring station in WCA2A, since it is the only long-term site available in this area. In order to study the hydropattern throughout WCA-2A, we utilized information from the 2A-17 station and interpolated (Chap. 7) among 11 water-level stations we installed in 1996 (Fig. 9.2). Using the temporal water-depth estimates, we considered metrics of (a) mean water depth, (b) frequency of exceeding or falling below certain depths (e.g., percentage

9 Macrophyte Community Responses in the Everglades

221

Fig. 9.2 Water-level stations installed in 1996 at permanent hydrologic sites. Sites (small filled circles) have been surveyed for elevation and established benchmarks placed within the wetland. The long-term plant survey transects (A, C, and D) established in 1990 to determine responses to nutrients and hydroperiod are denoted by large filled circles, squares, and triangles

of days > 150 cm), and (c) stability of water depth (e.g., metrics of variance or range in water depth). We evaluated all metrics using both short-term (1 year) and longterm (1981–1998) data. Short-term hydrology was expected to have a greater influence on subtle compositional patterns (Gunderson 1989), while long-term hydrology was expected to have a greater influence on coarse-scale species distributions (Urban et al. 1993; Busch et al. 1998; Shay et al. 1999). Preliminary results indicated that short-term mean water depth (hereafter, depth), long-term frequency of depth < −10 cm (index of severe dryness calculated as the percentage of days

222

C.J. Richardson et al.

during 1981–1998 in which water depth was < −10 cm; hereafter freq. < −10 cm), and long-term interquartile range of depth [a robust index of stability of water depth calculated as the range between the 25th and 75th percentiles of the distribution of daily water depths during 1981–1998; hereafter IQR (depth)] were most strongly related to vegetation. These three variables were not collinear, and each accounted for unique variation in vegetation composition along the anthropogenic influence gradient (p < 0.05, partial Mantel test – see Data Analysis). Thus, all three were retained for further analysis.

9.2.4

Depth Calculations

Soft-bottom elevations were estimated using inverse distance-weighted averaging from the five closest survey sites (Table 9.1). Since we are interpolating depth data back about 40 years, it was necessary to correct soft-bottom elevation for accumulation rates. Accumulation rates at the benchmark sites were estimated from Craft and Richardson 1993b (Table 9.1). In order to estimate the water elevation at each of the sites, the linear relationship between 2A-17 and each site was calculated. For example, 2A-14 is the closest survey site to the SFWMD’s site 2A-17. Water elevation at 2A-14 can be approximated using data from 2A-17, as follows: Stage2A-14 = 0.582 + 0.827 × Stage 2A-17. Stage at site 2A-17 was calculated for each day of the period of record from the stage data at 2A-17. Stage at each of the benchmarks was interpolated from each station by assuming the hydraulic gradient (change in elevation of top of surface-water over distance) in WCA-2A remains relatively constant over time and the equipotential lines trend east–west (Table 9.1). Subsequent work suggests that there are temporal variations to the hydraulic gradient and orientation of

Table 9.1 Soft-bottom elevation and ∆-stage in WCA-2A Station

Peat accumulation rate (cm year−1)

Soft-bottom elevation (m relative to MSL)

∆-Stage (cm relative to 2A-17)

2A-11 2A-12 2A-13 2A-14 2A-15 2A-16 2A-19 2A-20 2A-21 2A-24 2A-25

0.2 0.26 0.2 0.2 0.03 0.32 0.23 0.23 0.38 0.66 0.60

2.71 2.81 2.78 2.99 2.90 3.08 3.20 3.02 3.05 3.08 3.26

0 0 0 0 0 +4.27 +11.0 +10.4 +7.63 +15.56 +14.34

9 Macrophyte Community Responses in the Everglades

223

equipotential lines; however, these are not significant changes and probably do not significantly affect these results. Once the stage at each benchmark was calculated, the water depth was calculated by subtracting the soft-bottom elevation (corrected for peat accumulation rates) from the stage.

9.2.5

Recurrence Interval Calculations

The recurrence interval is an estimate of the probability that a given water depth will be exceeded within a certain interval of time. The recurrence interval is calculated by ranking the occurrences of water depth for each day for the period of record. The greatest water depth is ranked as 1 (M = 1). The most shallow water depth is ranked (N), where N is the total number of days of data. The recurrence interval for each water depth is calculated by R = (N + 1) / M. For each site (listed in Table 9.1) the water depth was calculated for each day for which data exists for 2A-17. The recurrence interval for each site was calculated for 5-, 10-, 30-, 60-, and 90-cm water depth. Contours were calculated for each depth using the spatial statistics module for S plus. Hydroperiod was also calculated for each site (listed in Table 9.1) for each day for which data exists for 2A-17. In addition, the average duration in days (both total and consecutive) of water depth < 5 and 10 cm, > 30, > 60, and > 90 cm was calculated for each site for the periods 1980– 1990, 1985–1990, 1990–1996, and individual years 1990–1996 and 1998. These data were then correlated to changes in vegetation frequency for sampled periods. These periods were selected since they represented the greatest changes in water levels and nutrient loadings.

9.2.6

Vegetation–Environment Linkages

We selected 14 stations along the C, D, and A transects as centroids for our sampling, all six stations from the central (C) transect and a random draw of four out of the six in each of the A and D transects. (King et al. 2004). In aggregate, five centroids were considered “impacted” (0–3 km), five “transition” (3–7 km), and four “reference” (> 7 km) based on prior classifications (Richardson et al. 1999). We used a stratified-cluster sampling design as described by Urban (2000). In this analysis, we hypothesized that variation in vegetation pattern was likely to occur on fine, local scales (e.g., tens of meters) as well as coarse or landscape (thousands of meters) scales. Conventional sampling designs such as random or stratified random would not have captured local patterns in variation if conducted with coarse- or landscape-scale separation distances among plots; similarly, at fine scales these

224

C.J. Richardson et al.

designs would require thousands of plots and thus would not be practical at a large spatial extent. Thus, this cluster approach provided a highly efficient method for considering multiple scales in spatial, environmental, and compositional variables (Fortin et al. 1989; Urban 2000). A single plot at each of the 14 stations served as a plot-cluster centroid. Eight additional plots were marked in a constellation, with four plots placed at 50-m distances from centroids and four others at 200-m distances, each in the four cardinal directions. Separation distances among plots within clusters ranged from 50 to 400 m, with a total of nine plots per cluster and 126 plots across the landscape. Distances among plot-clusters ranged from approximately 1,000–10,000 m. We chose a plot size of 10 m2, large enough to integrate across microhabitats and thus reduce noise, but not so large that they averaged across distinct patches of vegetation (Fortin et al. 1989). Plots were semicircular to facilitate sampling from the perimeter and to minimize disturbance. We used two complementary distance-based procedures to estimate relationships between vegetation composition and environmental variables. A complete and detailed description of our environmental and vegetation sampling design, site selections, and statistical methods are given by King et al. (2004). To complete the analysis we ordinated plots and species based on species composition using nonmetric multidimensional scaling (nMDS; Minchin 1987). Ordination provided a visual assessment of gradients in species composition among and within impact zones and was conducted to aid in the interpretation of partial Mantel tests, our second, more tactical approach to estimating partial spatial and environmental linkages to composition. We used Bray–Curtis dissimilarity (BCD) as the distance metric, a coefficient shown to be one of the most robust and ecologically interpretable (Faith et al. 1987). Once plots were ordinated, species sampling centroids were mapped into ordination space using weighted-averaging (Legendre and Legendre 1998). We performed ordinations for all plots (n = 126) as well as three subsets of plots corresponding to three zones (n = 45 for impacted and transition; n = 36 for reference). Partial Mantel tests were used to measure the partial correlation (Mantel r) between spatial, environmental, and vegetation distance matrices (Mantel 1967; Smouse et al. 1986). Fundamentally, the analysis examines whether plots that are environmentally similar are also compositionally similar (Urban et al. 2002). Mantel r coefficients are typically relatively small in magnitude (usually less than 0.5 for all but the strongest relationships) because the analysis considers the full rather than reduced dimensionality in multivariate data (e.g., Legendre and Fortin 1989; Leduc et al. 1992; Foster et al. 1999). Because it uses distance matrices, this approach allows the user to extract variation caused by spatial autocorrelation (Legendre 1993) as well as other environmental variables to yield pure-partial correlations – relationships that represent variation that cannot be explained by all other variables included in the analysis. We used spatial (space, canal) and environmental [C, N, P, Ca, K, Mg, Na, depth, IQR (water depth), freq. < −10 cm, fire] variables as individual predictors in the Mantel analysis. Individual variables were converted to distance matrices using

9 Macrophyte Community Responses in the Everglades

225

Euclidean distance (Legendre and Legendre 1998). An additional spatial matrix of Euclidean distances (m) was calculated using spatial coordinates (UTM) coordinates of every plot (space). Vegetation species composition was expressed using BCD, as in the nMDS analysis. We first examined simple relationships between space and individual environmental variables, canal and environmental variables, and each spatial or environmental variable and vegetation. We then examined the effect of spatial autocorrelation on vegetation–environment relationships by factoring out the effect of space. Finally, we estimated pure-partial vegetation–environment relationships. Here, the strength of a relationship between a predictor variable and vegetation was assessed after variation explained by all other variables had been removed (except for canal, because it was assumed to be causing variation in several environmental variables). We also examined the pure-partial space–vegetation relationship, which would indicate residual spatial pattern in vegetation that can be explained only by spatial processes such as dispersal or other unmeasured factors with spatial structure. Partial Mantel tests were conducted using data from all plots, three subsets of data corresponding to the impact zones, and coarse-scale averages of composition and the environment within the 14 plot-clusters. Significance (Bonferroni-corrected p ≤ 0.05) of Mantel r coefficients was assessed using 10,000 permutations. As a visual framework for these results, relationships among space, canal, environmental variables, and vegetation were synthesized using path diagrams, a schematic that depicts significant paths of relationships among variables (Leduc et al. 1992, Zmyslony and Gagnon 2000). Bootstrapped confidence limits (95% CLs) were estimated for partial Mantel r coefficients to allow for comparison of the strength of linkages among zones. Bootstrapping was conducted by resampling distance matrices at a level of 90%, with 1,000 resamples (Manly 1997, King and Richardson 2002). We contrasted pattern in vegetation among the three zones at both fine and coarse scales. Fine-scale pattern was examined within clusters using separation distances of 50–400 m, with 50-m intervals (e.g., 251–300 m separation distances = 300-m distance class). Coarse-scale pattern was assessed by comparing autocorrelation among clusters (within- vs. among-cluster variation). Distance classes were considered significantly different if CLs did not overlap (Manly 1997). Mantel tests and bootstrapping were performed using S-Plus 5.0 for Unix.

9.3

9.3.1

Environmental Factors and Anthropogenic Effects on Everglades Communities Climate

The key climatic factor (Fig. 9.1) controlling vegetation patterns and succession in the Everglades is the amount of precipitation. Precipitation directly influences hydroperiod and fire frequency in ombrotrophic (rainfall fed) wetlands like the

226

C.J. Richardson et al.

Fig. 9.3 The mean annual precipitation is shown for the northern Everglades from 1950 to 2005. Individual years above and below the long-term mean are depicted by bars above or below the mean annual average (1895–2005). Data source from NOAA at (http://www.ncdc.noaa.gov/ onlineprod/drought/xmgr.html)

Everglades, and hydroperiod is the most important determinant of wetland processes (Mitsch and Gosselink 2000). An analysis of annual rainfall over south Florida from 1915 to 1985 shows that for 80% of this time period annual rainfall was between 1,016 and 1,524 mm (SFWMD 1992). Approximately 66% of annual rainfall was recorded during the June to October wet seasons (SFWMD 1992). Average annual weighted precipitation for the Everglades from 1895 to 2005 is 1,320 mm (Fig. 9.3). A time series analysis of rainfall patterns over the past 55 years reveals that south Florida has experienced considerable periods of drought, especially during the 30-year period 1960–1990 (Fig. 9.3). Moreover, both the EAA, which drains partially into WCA-2A, and the ENP, the southernmost remnant of the original Everglades, have received annual rainfall consistently below the historical record. During the period 1970–1985, the EAA and ENP received less rainfall 80 and 67% of the time, respectively. From the period 1970–1989, the EAA received less rainfall than average 14 out of 19 years. Extensive droughts with rainfall 250 mm below normal per year existed for six of those years and in 1988–1989, they reached 510 mm below normal (SFWMD 1992). However, since 1990s the Glades have experienced dramatic increases in rainfall with highest levels falling in 1995 (Fig. 9.3). These data indicate that the Everglades has experienced reduced rainfall for extended periods followed by significant rainfall years, climatic conditions that will significantly alter the plant-growing environment. Additionally, the changes in water regulation schedules and operation of water-control structures have also significantly impacted the hydroperiod in the Everglades, particularly in portions of WCA-2A (Fennema et al. 1993; Davis and Ogden 1994a,b; SFWMD 2006). These shifts in rainfall patterns and water regulation schedules have also greatly influenced the yearly nutrient loadings and concentrations (especially P) entering the WCAs (SFWMD 2004, 2005, 2006).

9 Macrophyte Community Responses in the Everglades

9.3.2

227

Hydroperiod

Hydroperiod is directly tied to the amount of rainfall as well as the manmade pumping and drainage schedules that occur in the Everglades (Fig. 9.1). The SFWMD, in conjunction with the US Army Corps of Engineer’s water management practices, have dramatically altered the annual wet–dry cycle characteristic of the historic Everglades. For example, inflow structures along the northern levee in WCA-2A were completed in the early 1960s (Light and Dineen 1994; see Chap. 8). During the period 1952–1960, levees were being constructed around WCA-2A, but water levels were not well regulated or controlled. The central fen area in WCA-2A dried briefly in 1962 and in 1971, but remained continuously flooded between these years (Fig. 9.4). This extended period of inundation had significant impacts on the plant communities. Many of the original tree islands were drowned, wet prairie vegetation was eliminated, and a loss of sawgrass along slough edges was noted (Alexander and Crook 1973; Dineen 1972, 1974; SFWMD 1992; see Chap. 8). The SFWMD, which had maintained artificially high water levels in WCA-2A during portions of the 1960s and 1970s, then reduced water levels in the early 1980s to reduce the tree island death and changes in the periphyton community caused by

Fig. 9.4 The mean daily water elevations in WCA-2A at SFWMD’s Station 2A-17 are shown from 1952 to 2005. The dotted upper horizontal line depicts the mean water-level value during the 57 years of measurements. The lower dashed horizontal line depicts the average soil surface elevation of 3.35 m. The monthly average and moving average water levels are show by light and dark shaded graph lines

228

C.J. Richardson et al.

flooding (Dineen 1989; SFWMD 1992). Unfortunately, extensive droughts occurred during 1980–1981 (Fig. 9.3), resulting in dramatic reductions in water table levels in the early 1980s (Fig. 9.4) and some expansion of wet prairie and sawgrass communities occurred (Worth 1988). Droughts during the later portion of the 1980s also resulted in water elevations at or well below the surface of WCA-2A during portions of the year (Fig. 9.4). The proliferation of cattail during wet years and dieback during droughts suggest that pooling of nutrient-laden water in WCA2A prior to the lowering of the water levels in 1980 promoted cattail expansion into slough and sawgrass communities (Urban et al. 1993). Furthermore, prolonged high water periods have reportedly stimulated the spread of cattail in nutrientenrichment areas (Newman et al. 1996). However, the rate of plant frequency change has not been accurately quantified nor the controlling environmental variables assessed in the field.

9.3.3

Nutrients

A series of outflow structures have created a P-enrichment gradient that often extends 5–8 km into the interior of the northern WCAs as a result of total phosphorus (TP) loadings averaging 229 metric tons from 1978 to 1988 and 147 metric tons on average from 1995 to 2004 from agricultural runoff and Lake Okeechobee outflow despite the use of agricultural best management practices (BMPs) since 1996 and the construction of 16,000 ha of storm water treatment areas (SFWMD 2000, 2005, 2006). Although 2004 TP loadings were reduced to 82 metric tons, loading rates per unit area have often exceeded 4 g m2 year−1 at the edge of the WCA-2A fen (SFWMD 2005, Richardson and Qian 1999). In 2005, TP loadings from the EAA were more than double 2004 loadings, and inputs reached 182 metric tons of TP, levels not seen since 2000 (SFWMD 2006). Runoff enriched with 60 metric tons of P and 1,814 metric tons of TN per year was pumped into the northern part of WCA-2A in 459,000,000 m3 of water during the 1970s and 1980s (SFWMD 1992). This resulted in increased TP and SRP concentrations in surface water, soil pore water, and elevated soil P content along a 10-km enrichment gradient (Richardson et al. 1991, 1997b; Koch and Reddy 1992, Qualls and Richardson 1995, Richardson and Vaithiyanathan 1995). Total soil N, Mg, K, Al, Fe, and Mn, however, did not vary significantly downstream from inputs (Qualls and Richardson 1995; Craft and Richardson 1997). Concentrations of Na and Ca were much higher within 3 km of the input structures but decreased significantly with distance from the input structures (Craft and Richardson 1997). A computer-generated kreiged soil P contour map of WCA-2A reveals an increasing soil P gradient from 1991 to 2001 with concentrations that range from 1,500 mg kg−1 near the input structures to areas of 400 mg kg−1 in unenriched locations 10 km south of the input gates (Vaithiyanathan and Richardson 1999). However, a significant (p < 0.0001) pattern of decreasing water and soil TP concentrations is displayed over a 10-km downstream nutrient gradient (Chap. 6). The same pattern

9 Macrophyte Community Responses in the Everglades

229

exists for soil TP pore water with values ranging above 1,000 µg l−1 (0–15 cm) in highly enriched sites near the source to background concentrations averaging between 5 and 15 µg l−1 (0–15 cm) in unenriched areas (Vaithiyanathan and Richardson 1995, 1997). The distribution of plant nutrients and plant species along gradients have revealed a general pattern of increased cattail cover and higher TP tissue concentrations near the northern release structures than in areas more downstream of the N and P inputs (Koch and Reddy 1992; Craft and Richardson 1997; Richardson et al. 1999). Cattail has higher TP values in both roots and shoots than sawgrass, but both species respond positively to nutrient additions. The N/P ratios of sawgrass have been shown to vary from 9:1 at enriched sites to 63:1 at unenriched sites, but cattail N/P ratios varied only from 6:1 in the enriched areas to 10:1 in nutrient poor sites (Koch and Reddy 1992). These data suggest that P is limiting in unenriched areas of the Everglades, but these data do not give insights into all the factors causing changes in plant populations along the nutrient/hydrology/fire gradient. Soil and water quality gradients in WCA-3A are not as distinct as found in WCA-2A since loadings and concentration of nutrients into WCA-3A are historically lower than reported for WCA-2A (SFWMD 2005, 2006; see Chap. 6). For example, the geometric mean TP input concentration from 1978 to 2003 into WCA-2A averaged 55.9 µg l−1 as compared with 34.3 µg l−1 flowing into WCA-3A (SFWMD 2006). Interior TP water concentrations during that period were nearly half in WCA-3A as compared with WCA-2A and averaged 8.8 µg l−1 as compared with 16.8 µg l−1, respectively (see Fig. 2.16; Chap. 2).

9.3.4

Fire

Reductions in fire periodicity and intensity of fire regimes caused by water management practices have also contributed to the replacement of sawgrass by other plants (Loveless 1959, Hofstetter 1974, Fig. 9.1). In an excellent analysis of fire frequency and patterns, Gunderson and Loftus (1993) noted that the WCAs have had over 127 fires from a period of 1980 to 1990. The largest was approximately 34,000 ha, with many of the fires ranging from 8,000 to 15,000 ha. The frequency of small fires was 7 months to 1 year, with severe fires occurring every 10–14 years. Severe fires are often the result of lightning strikes in the Everglades during extended droughts. Factors influencing the timing and severity of a fire include the amount and type of fuel present, physical factors such as wind and moisture regime, and whether the fire is human-caused or lightning induced (Gunderson and Loftus 1993). The spatial and temporal aspects of fires in the Everglades are related to the vegetation characteristics. A composite map of the large wildfires caused by lightning and controlled burns by the state indicate that a large portion (> 50%) of WCA-2A has burned during the past 30 years (Richardson et al. 1997b, King et al. 2004). These fire events will have had major impacts on the plant populations, especially for sawgrass, a fire tolerant species, and cattail, a fire intolerant species (Hofstetter 1974).

230

C.J. Richardson et al.

Understanding the interaction of hydroperiod, hydropattern, fire, and nutrients is critical to the future management of the plant communities of the Everglades. Intense fires have also occurred in desiccated northern regions of WCA-3A since the late 1990s, and these have had dramatic effects on vegetation patterns as well.

9.4 9.4.1

Results and Discussion Macrophyte Species Distribution Patterns in WCA-2A

An analysis of 55 random survey vegetation plots completed in 1996 allowed us to assess the distributions of all species in WCA-2A. This survey covered all major regions of WCA-2A, including the drier western side as well as areas near the S-11 outlet structures (Richardson and Huvane 2001). A listing of all the major species found in WCA-2A is given in Chap. 4. Two species, sawgrass and cattail, dominated the survey. Sawgrass was found in 62% of all the plots and cattail in 47% of the plots. Cattail dominated the 20 plots in the northeastern corner of WCA-2A just south of the S-10 input structures. In contrast, sawgrass dominated the western side of WCA-2A as well as a number of interior sites. The overall average frequencies for sawgrass and cattail were 35.4 and 16.8%, respectively, for the entire survey. In all the plots where sawgrass was found, it averaged 57.3% of the plant frequencies. In some plots in the interior and in the western regions of WCA-2A, sawgrass comprised 100% of the plots. In the plots where cattail was found, it averaged 35% frequency but never reached densities above 73%. The enriched plots with cattail often had 10–15 other species associated with it. Nymphaea was found in 35% of the plots but only comprised 6.4% of the plant population for the region. However, in some sloughs Nymphaea comprised 72% of the macrophyte population. Other species that were found in more than 10% of the plots were Chara (11%), Sagittaria (25%), Mikania (24%), Utricularia (14.5%), and Eleocharis (18%). The highest density of cattail as noted earlier was found south of the S-10 input structures and along the southwestern outflow areas along the canals.

9.4.2

Plant Species Composition Along Nutrient Gradients in WCA-2A

The oldest detailed vegetation survey for this region (Davis 1943) suggests that a large portion of the area covered by our 10 km by 10-km sampling grid (Fig. 5.1) was dominated historically by a wet prairie/slough, tree island, sawgrass mosaic. We found sawgrass (C. jamaicense) to be the dominant emergent macrophyte along most of the sampling stations on all three 10-km transects during the 12-year sampling period (1990–2002) of our survey, especially on the undisturbed western

9 Macrophyte Community Responses in the Everglades

231

transect (Fig. 9.5). Relative sawgrass frequency was 61% for all sites combined, with cattail making up 17% and invading vines, Mikania or Sarcostemma, found at > 3% of all sample points. The largest number of species was found on the A transect. Of major importance is an assessment of the advance of cattail down the nutrient gradient formed by the release of nutrient-laden water from the A, C, and D control structures. In 1990, all three transects indicated a large cattail population at the 1.4- to 1.8-km sampling locations, with transect C being dominated by this species up to 3.5 km into the fen. Transects A and D had much lower populations of cattail and sawgrass or other species dominated the community. Transects A and D also had much more open bare area when compared with transect C. By 1992, transects A and D both showed an expansion of cattail frequency closer to the canals inputs, and cattail extended > 7 km along the D transect. Cattail showed a continued expansion into the fen during the 1990s as sawgrass populations decreased and open areas were reduced. By 2000 cattail had reached > 8 km into the fen along all three transects. However, in 2002 cattail frequency had decreased along transect A and was similar to the 2000 analysis on the C and D transects, thus suggesting that the P reductions into WCA-2A (Chap. 2) may be starting to reduce or stabilize cattail expansions. A careful analysis of the trends between the sampling intervals over the 12-year period also indicate that cattail and sawgrass populations expand and contract from sampling period to sampling period due to alterations in hydrologic and fire effects (Richardson et al. 1997b). Sawgrass formed dense stands at the three most distant (7–10 km) locations on all nutrient transects (A, C, and D) but was absent or present only in lower frequencies at the sampling sites closest (1.5 km) to the enriched areas near the Hillsboro Canal (Fig. 9.5). A Pearson’s correlation analysis indicates that sawgrass was positively correlated with distance from the canal and inversely correlated with soil P (p < 0.001). At P-enriched locations, cattail and other emergent were dominant (Fig. 9.5). For example, cattail in 1990 accounted for 35% of plant frequency at sites 10A-1, 60% at 10C-1, and < 5% at 10D-2 of the emergent vegetation. Twelve years later cattail plant frequency was greater than 60% at 10A-1 and 10C-1 and 40% at 10D-2. Thus, cattail had nearly doubled in amount in 12 years at site 10A-1, stayed the same at 10C-1 and increased many fold at 10D-2. While cattail was very abundant at the sampling sites closest to the nutrient input sites at 10A-1, 10C-1, and 10C-2 over the 12-year sampling period sawgrass was dominant at 10D-2 in 1990–1994, with cattail increasing in frequency from 1994 to 2002 (Fig. 9.5). The reason for the sawgrass dominance was due to a severe fire in 1989 at site 10D-2, which resulted in the complete destruction of the cattail population and a massive germination of sawgrass in 1990 (Richardson et al. 1997b). However, by 1996 cattail has reinvaded the area and the amount of sawgrass had been reduced. Our findings closely follow the 63% sawgrass cover reported in a 1986 survey for a square-mile plot located midway in WCA-2A (Davis 1994). Alexander and Crook (1973) originally established this permanent plot in 1971, when they found only 46% sawgrass and 18% cattail cover (as reported in Davis 1994). Davis found a reduction of cattail to only 5% in this plot by 1986. A time series analysis of remote sensing data for WCA-2A from 1973, 1976, 1982, and 1987–1991 revealed an

232

C.J. Richardson et al.

Fig. 9.5 Percent frequency of the dominant macrophytes along the A, C, and D nutrient gradient transects in WCA-2A from 1990 to 2002 and along the western transect from 1998 to 2002 (see Fig. 5.1). Note: all species comprising less than 10% of the total frequency are grouped into the other category and open refers to areas with no vegetation. The sampling locations A1 and D6, etc. correspond directly to the distances noted on the figures (A1 = 1.8 km, D6 = 10.7 km). Data from site D1 are missing from site after 1996 due to the extremely thick density of willow and cattail at this site, which prevented airboat accesses

increase of pure cattail from 920 to 4,900 ha over this time period (Jensen et al. 1995). Collectively, these data suggest that species composition and plant communities are dominated by a few species in the undisturbed areas and that certain portions of WCA-2A have experienced shifts in plant populations due to alterations in nutrient or hydrologic conditions. For example, we found cattail interlaced with vines, Scirpus, Salix, and exotic invasive species like Pistia primarily in the P-enriched areas along the first 3.5 km of the gradient. The overall cumulative frequency of Sagittaria spp. (arrowhead) (2.4%), Pontederia cordata (pickerel-weeds) (1.1%), and Nymphaea odorata (water lily) (1.9%) across all sites was low, and water lilies were generally more prevalent in the open-water slough areas of the unenriched southern portions of the transects (> 5 km from the P input sources). Sagittaria, Pontederia, Salvinia minima, and Scirpus validus (bulrush) were found primarily in the northern enriched slough areas. For example in 1990, bulrush was abundant at 10D-1 and 10C-2 (Fig. 9.6) where its frequency was 25 and 20%, respectively. Sarcostemma, a vine, was abundant at 10A-1 (25%) and 10C-1 (17.5%) while

9 Macrophyte Community Responses in the Everglades

233

Fig. 9.6 Percent frequency of major macrophytes along the C gradient transect in WCA-2A in 1990 vs. 2002. Vines are comprised primarily of Mikania or Sarcostemma

234

C.J. Richardson et al.

Sagittaria was a dominant species (50%) at 10A-2. Pigweed (Chenopodium sp.), an invading upland species, was found only along the D transect during the drought conditions experienced in 1989/1990. Cattail comprised the highest percentage of invading macrophytes in the northern slough areas during most years at distances < 3.5 km. In summary, an analysis of plot data of biannual changes in cattail frequency during the 12-year analysis suggests a consistent pattern exists across all three transects in terms of cattail increases or sawgrass decreases. In the enriched areas it is evident that cattail populations are generally lowest in 1990 and highest in the 2000–2002 period along the nutrient transects. Collectively, these data suggest that the cattail populations have increased in some locations as much as 75% in a 6-year period (e.g., site A2 from 1990 to 1996, Fig. 9.5) and decreased at other locations as much as 30% (e.g., site C2, Fig. 9.5). The increase in cattail at A2 was not due to a decrease in sawgrass but was represented by the loss of Sagittaria, Pontederia, and open/bare ground areas. Surprisingly, at C2 cattail decreased concurrently with an increase in sawgrass and bare/open cover. The amount of bare/open area varied by site and over time, with values reaching nearly 40% along A2 in 1990, and at D2 and D3 in 1996. The amount of open/bare areas found during the surveys reflects long-term open slough areas with limited macrophytes or areas that had recently been disturbed by fire. By contrast our 5-year analysis of a western transect established in 1998 to assess how vegetation changes in an undisturbed area with low nutrient and water inputs in WCA-2A indicated a sawgrass-dominated area with no cattails along the transect (Fig. 9.5). The amount of open-water bare area did show an expansion in 2002 due to increased rainfall during the early 2000–2002 period (Fig. 9.3). To understand the factors controlling these shifts in plant populations, we further analyzed plant frequencies in conjunction with nutrient concentrations in water and soil, water-depth changes, and fire frequency effects (see Sect. 9.4.5).

9.4.3

Plant Species Composition Along Nutrient/Hydrologic Gradients in WCA-3A

The central (C) and eastern (E) transects in WCA-3A were established in the summer of 1998 prior to the fire that burned the northern portions of WCA-3A in 1999 (Fig. 5.1). These plots were located to provide a monitoring record of plant community structure prior to the release of treated agricultural storm water from storm water areas (STAs) 3 and 4 into WCA-3A via the S-150 control structure and from WCA-2A through the S-11A, B, and C structures, as well as to provide a baseline to compare with future changes in the region. Our 1998 survey (Fig. 9.7) showed that sawgrass was the dominant species along both transects, especially along the Eastern Transect throughout the period of study. In contrast, cattail was nearly absent from the Eastern Transect in both 1998 and 1999, never exceeding 10% at any single station. Although we observed significant inclusions of Typha in the lower half of the Central Transect in 1998,

9 Macrophyte Community Responses in the Everglades

235

Fig. 9.7 Percent frequency of macrophytes along the central (C) and eastern (E) gradient transect in WCA-3A. Note: all species comprising less than 10% of the total frequency are grouped into the other category and open refers to areas with no vegetation

particularly at stations 3C-5 (7.3 km) and 3C-7 (11.1 km), those populations never exceeded a frequency of 30% and did not persist through 2002. A more stable population of Typha established at station 3C-8 (12.9 km) along the Central Transect in 1999, after the extensive El Niño-driven fires in the spring of 1999. The lack of significant and (more importantly) persistent cattail populations along the WCA-3A transects is noteworthy. We observed soil P concentrations in the northernmost stations of the 3C transect in excess of the 650 mg P kg−1 soil dry mass content (Chap. 24). This concentration was suggested by Wu et al. (1997) as the critical level above which Typha becomes dominant. However, in WCA-3A, we did not observe cattail populations associated with those soil chemistry observations, and we suggest that the increased fire frequency in northern WCA-3A precludes a significant cattail expansion regardless of soil nutrient loading. The cattail populations observed in the 3C transect (below 3C-4, see Fig. 9.7) were likely due to higher mean water levels resulting from an earlier peat fire in the center of northern WCA-3A that dropped relative soil elevations (T. Towles, FL FWCC, personal communication). Cattail is more tolerant of flooded soils than sawgrass (Toth 1988). During the study there was a significant presence of “other” species in both the central and eastern transects (Fig. 9.7), and these species actually dominated community structure at the northernmost stations (e.g., stations 3C-1 and 3E-1 in 2002). Unlike in WCA-2A, where the “other” plants were frequently vines such as Mikania scandens and Sarcostemma clausum (especially when cattail was dominant), in WCA-3A the “other” category was comprised exclusively of both woody

236

Fig. 9.8

C.J. Richardson et al.

Water depths along the central and eastern transects in WCA-3A

and herbaceous rooted emergent vegetation. For example, the 40% “other” species at 3E-1 in 2002 reflected the persistent Salix caroliniana Michx. (coastal plain willow) thickets that characterized the northeastern corner of WCA-3. Other species included the fleshy forbs Peltandra virginica (L.) Schott, Pontederia cordata L., and Sagittaria spp., as well as Thelypteris spp. (fern, either T. ovata R.P. St. John or T. palustris Schott var. pubescens (G. Lawson) Fernald) and Chara spp. (muskgrass or stonewort, an alga). Along the northern central transect, vegetation was typically much smaller in stature and more diverse than elsewhere within WCA-3A (J.W. Pahl, personal observation), reflective of the more mesic conditions (Fig. 9.8) and more frequent fire occurrence there than lower in the transect. Specifically, our hydrologic analysis for the period 1998–2003 along the WCA-3A transects indicated that the plots closest (0–4 km) to the dike were the driest sites and maintained the lowest water levels (10–20 cm), while water levels rose to range from 30 to 60 cm in depth > 6 km from the dike (Fig. 9.8). Interestingly, the frequency of “other” species, collectively averaged across both the 3C and 3E transects, was not correlated to any of the environmental variables measured contemporaneously with the vegetation surveys (e.g., water depth, surface water P or N) along the transect. The exception to this statement was a significant and negative correlation between “other” species and the concentration of Cl− measured in the surface water during the 2002 survey (Pearson Correlation Coefficient -0.47931, p = 0.0378, n = 19). Admittedly, correlations do not prove cause-and-effect, and since Cl− was not measured in the surface water in either 1998 or 1999 we cannot confirm if this relationship was consistent over the 5-year period of study. However, the potential for

9 Macrophyte Community Responses in the Everglades

237

Cl− to affect plant distribution is notable, given our observations of significant surface water Cl− gradients associated with our transects throughout the northern Everglades, especially the definitive Cl− gradient along the 3E transect in 2002 (Chap. 6). Reviewing our earlier discussion, these Cl− gradients may have resulted from the liberation of shallow connate saltwater aquifers by the construction of the EAA and northern Everglades drainage canal network (Slate 1998). Together with Cooper and Goman’s hypothesis (Chap. 12) that the digging of the drainage canals facilitated a change in conditions that favored the expansion of the periphyton, these results reinforce the proposal that the digging of the northern Everglades canal network had a significant effect on the community ecology of the region possibly exceeding the obvious disruptions in hydrologic flows.

9.4.4

Factors Controlling Plant Communities

9.4.4.1

Nutrients

To assess the possible relationship of nutrients and plant communities, we first completed a correlation analysis, which revealed a significant relationship between plant species and various soil and water nutrient concentrations (Table 9.2). Cattail was positively correlated (p < 0.001) to several indicators of increased P fertilization, including soil P at both 10- and 20-cm depth, pore water total dissolved nitrogen, and silica in the surface water. Surprisingly, cattail frequency was not related to surface water P concentrations or any other major cation. This may be due to the fact that the cattail populations in the northern parts of the nutrient gradient show considerable increases and in other cases decreases in frequency over the survey period due to fire effects (see Sect. 9.4.4.3). On the other hand, Vymazal and Richardson (2003) have reported that cattail aboveground biomass was strongly correlated with both soil and water P concentrations (Chap. 6). Conversely, sawgrass was negatively correlated to soil TP at both depths and surface water TP as well as soil pore water PO4 at 25- and 60-cm depths (Table 9.2). The same dependence was reported by Vymazal and Richardson (2003) for sawgrass aboveground biomass. Sawgrass was also negatively correlated to pore water Ca, Mg, Na, Cl, total dissolved N, total dissolved P, DOC, and HCO3 at varying depths. The latter negative correlations relate closely with the increased ions that are found in runoff from the agricultural lands to the north. The negative relationship of sawgrass to ions like P, Na, and Cl suggests this species is responding by decreasing population density with increased ion inputs, but no causal relationship can be established from correlation analysis alone. A number of other species comprising less than 10% of the population were highly correlated to increases in TP in the soil as well as an increase in ions in pore water (Table 9.2). Specifically, species like Polygonum were negatively correlated (p < 0.05) with soil P content, while species like Sagittaria were positively correlated (p < 0.001, data not shown). Salix (willow) was highly correlated to Ca, Mg, Na, K,

238

C.J. Richardson et al.

Table 9.2 Pearson correlation coefficients for plant species composition (percent frequency) of Typha, Cladium, and other species from 1990 to 1996 Variable Typha Cladium Other species Nutrients −0.60 (***) 0.67 (**) Soil P (0–10 cm) 0.72 (**) −0.51 (*) 0.61 (*) Soil P (10–20 cm) 0.74 (**) 0.33 NS −0.55 (**) 0.25 NS Surface H2O TP 0.03 NS −0.21 NS 0.03 zNS Pore water PO4 (12 cm) 0.31 NS −0.58 (**) 0.45 NS Pore water PO4 (25 cm) 0.49 NS −0.84 (***) 0.67 (*) Pore water PO4 (60 cm) 0.70 (*) Pore water Na (12 cm) −0.14 NS −0.65 (**) 0.44 NS Pore water Na (25 cm) −0.17 NS −0.65 (*) 0.71 (**) Pore water Na (60 cm) 0.04 NS −0.64 (*) 0.74 (**) Pore water Cl (12 cm) −0.26 NS −0.85 (**) 0.73 (**) Pore water Cl (25 cm) −0.20 NS −0.65 (*) 0.77 (**) Pore water Ca (12 cm) −0.08 NS −0.74 (**) 0.49 NS Pore water Ca (25 cm) −0.26 NS −0.67 (**) 0.75 (**) Pore water Mg (12 cm) 0.04 NS −0.71 (**) 0.44 NS Pore water Mg (25 cm) 0.16 NS −0.67 (*) −0.69 (**) 0.18 NS Pore water (TDN, 12 cm) 0.88 (*) 0.86 (*) 0.56 NS Pore water (DOC, 12 cm) 0.96 NS 0.96 (**) −0.34 NS 0.74 NS Silica (surface H2O) Hydrologic variables −0.81 (***) ND Sum of days ≥90 cm (10 years) 0.66 (*) 0.80 (**) ND Mean of days <5 cm −0.83 (**) (previous 10 years) Plant composition was related to water and soil nutrients and hydrologic conditions along a 10-km gradient. Only the main variables displaying a significant relationship with cattail or sawgrass are shown (n = 72). Numbers in bold indicate a significant relationship. NS nonsignificant; ND no data * p < 0.05, **p < 0.01, ***p < 0.001

Cl, NH4-N, and SO4-S. This is not surprising since woody species often require more Ca and other ions (Kramer and Kozlowski 1968). Willow is often found on highly disturbed sites and these sites exist just south of the inflow gates where additional inputs of these ions have been reported in the runoff and found stored in the soil (SFWMD 1992; Craft and Richardson 1997). Collectively, these analyses demonstrate that increased nutrients and ions have increased along with macrophyte species diversity south of the input structures, especially in the first 5 km. Concurrently, sawgrass and a number of slough species like Nymphaea have decreased in frequency. The relationship of soil TP (0–10 cm) to cattail and sawgrass frequency taken over a 6-year period of high nutrient input (1990–1996) suggests that a P threshold may exist for these two dominant macrophytes (Fig. 9.9a, b). Cattail expansion increases exponentially from an average of 3% in frequency below a soil concentration of 500 mg kg−1 TP to near 40% of the population above 1,000 mg kg−1 TP (Fig. 9.9a). Above 1,500 mg kg−1, cattails comprise approximately 80% of the species composition. Conversely, sawgrass populations decrease to near 5% in frequency above 1,250 mg kg−1 of soil TP, and they reach pure monocultures (> 90%) at locations with

9 Macrophyte Community Responses in the Everglades

239

Fig. 9.9 (a) Percent cattail and (b) sawgrass frequency compared with total soil phosphorus concentrations (mg kg−1) collected along three nutrient gradients in the upper 10 cm of soil in WCA-2A

TP soil concentrations between 400 and 600 mg kg−1 TP (Fig. 9.9b). These values closely follow the average soil TP concentrations reported by Debusk et al. (1994) for WCA-2A stands of sawgrass (473 ± 134 mg kg−1), mixed sawgrass and cattail (802 ± 444 mg kg−1), and pure cattail (1,338 ± 381 mg kg−1). A best-fit curve analysis suggests that cattail increases are positively related to soil TP (Fig. 9.9a) and sawgrass decreases are negatively related to increases in soil TP (Fig. 9.9b). These relationships suggest that one can simply predict cattail or sawgrass populations based solely on soil TP concentrations, with values much above 500–600 mg kg−1 TP resulting in increased cattail and decreased sawgrass populations (Chap. 24). However, a replot of the frequency changes in cattail populations related to soil TP concentrations over

240

C.J. Richardson et al.

Fig. 9.10 The change in frequency for cattail as related to soil phosphorus concentrations along three nutrient gradients sampled during each period covering 1990, 1992, 1994, and 1996 in WCA-2A

time indicate that this pattern is not solely controlled by P concentrations or that other conditions may mask plant population responses (Fig. 9.10). For example, cattail populations at 1,100–1,300 mg kg−1 varied from 0% in 1990 or 10% frequency in 1994 to over 60% in 1996. This change in cattail population may be related to other factors like fire or to changing water levels, and a series of fertilizer experiments alone have not shown the species shifts found along the nutrient gradient in WCA-2A (Craft et al. 1995; see Chap. 20). Importantly, no significant increase in cattail was found at 500 mg kg−1 of P or below (Fig. 9.10). The distribution patterns and relationships for the entire macrophyte communities of WCA-2A were also examined in WCA-2A, and results suggested that the loss and/or establishment of macrophyte species in the Everglades are strongly related to soil P status and changes in the hydrologic regimes (Vaithiyanathan and Richardson 1999). They compared macrophyte species distribution with the soil P concentrations and developed six major categories of P enrichment (delineated by soil P contours) that adequately described the macrophyte species distribution in the WCA-2A. The category 0 represented the unenriched area (< 500 mg kg−1 TP) whereas the P enrichment category of 5 denotes the most enriched area (> 1,300 mg kg−1 TP). Table 9.3 summarizes the soil TP levels in each category and our results on the macrophyte species changes along the gradient. Average surface water increases in TP and PO4-P levels (1986–1995) generally corresponded to the six soil P enrichment categories and suggest that soil P concentrations directly reflect water P loadings to each location along the gradient (Craft and Richardson 1993b, Richardson et al. 1997b). The total number of macrophyte species present within a category increased with P enrichment (Table 9.3) from 19 in the unenriched areas (category 0) to 42 in the most enriched areas (category 5). Although the total number of species increased, those species typical of the oligotrophic area showed a progressive

P-enrichment category

0

1

2

3

4

5

Soil TP range (mg kg−1) Surface water (mean ± SE) Grand mean (1986–1995)* TP (µg l−1) PO4-P (µg l−1) Number of study sites Number of macrophyte species Species common with 0 Species common with 1 Species common with 2 Species common with 3 Species common with 4 Species common with 5

< 500

500–700

700–900

900–1,100

1,100–1,300

> 1,300

15.0 ± 2.2 6.4 ± 0.4 12 19 19 – – – – –

28.0 ± 6.3 7.9 ± 0.6 9 28 18 28 – – – –

52.3 ± 10.0 22.4 ± 5.5 9 35 16 26 35 – – –

51.2 ± 5.9 17.1 ± 2.0 8 37 14 24 33 37 – –

101.3 ± 17.7 48.5 ± 7.1 7 31 11 21 27 29 31 –

130.4 ± 9.9 68.1 ± 6.8 7 42 11 19 26 30 28 42

9 Macrophyte Community Responses in the Everglades

Table 9.3 Phosphorus concentrations in soil and water, and macrophyte species distribution along a P-enrichment gradient in WCA-2A (from Vaithiyanathan and Richardson 1999)

*Data from Qualls and Richardson (1993), SFWMD, and Duke University Wetland Center represent the arithmetic mean of TP and PO4-P concentrations

241

242

C.J. Richardson et al.

decline with P enrichment. Only 11 of the 42 species (26%) observed in the most impacted area were found in the least impacted areas of the Everglades. Ninety-five percent of the species found in the unenriched areas were found in soils of category 1 (500–700 mg kg−1) and 84% were found in category 2. Nearly half of the macrophyte species widespread throughout the unenriched Everglades were absent from the two most enriched areas. These findings suggest that a number of macrophyte species have a lower soil TP threshold than the threshold found for sawgrass and cattail during our 1990–1996 studies (Fig. 9.9a, b). This indicates that neither cattail nor sawgrass are as sensitive to P loadings as some other native species in the Everglades. Moreover, a major change in macrophyte species composition apparently occurs once the soil P concentrations exceed the 700–900 mg kg−1 of TP. We conclude from these data that category 1 (500– 700 mg kg−1 of TP) may represent the soil P threshold zone for many species in the Everglades. A complete analysis of the species distribution and responses at each of the six categories was given by Vaithiyanathan and Richardson (1999) and supports these conclusions. An index of best indicator species is also presented, and Utricularia purpurea is among the plant species most sensitive to increased P concentrations. This finding is also supported by our P dosing study research (Chap. 16). However, the gradient time-series analysis of sawgrass and cattail populations vs. soil TP concentrations, together with the disturbance and P fertilizer experiments, collectively suggest that plant population patterns for these dominant species are also controlled by the effects of other factors like fire and hydrologic shifts rather than just soil TP. Factors like rapid water-level fluctuations and extended hydroperiod > 30 cm in depth have been shown to have a negative effect on sawgrass growth (Toth 1987; Raikes 1992; Sutter 1992; Newman et al. 1996), which may enhance the ability of cattail to invade areas with weakened sawgrass populations. Craft et al. (1995) also reported that mixed populations of sawgrass and cattail under both control and fertilized conditions existed primarily at sites with alternating deep and shallow hydrologic regimes. Our experimental studies (Chap. 20) suggest that cattail invasions first require disturbance and weakened sawgrass to establish stands. After the initial invasion, phosphorus then acts as a catalyst, causing cattail to increase in the open niches left by the reduced sawgrass populations.

9.4.4.2

Hydrology

We reconstructed a number of hydrologic indicators to assess if these hydrologic variables might be related to the vegetation responses found over a 6-year period (1990– 1996) of community analysis along the nutrient gradient. The hydrologic patterns developed included estimates at each sampling location (Fig. 9.2) for mean and median water levels, recurrence intervals, and average duration in consecutive days for water depths at 5, 30, 60, or 90 cm. We also calculated these depth averages for the periods 1980–1985, 1980–1990, 1990–1996, and for each year during the vegetation surveys. These data allowed us to characterize the hydropattern including water depth, flood duration, and periods of dryness as well as the frequency of occurrence

9 Macrophyte Community Responses in the Everglades

243

Fig. 9.11 Average duration of water depths for the periods of 1980–1990 and 1990–1996. These durations represent the average number of consecutive days with water at < 5, > 30, > 60, or > 90 cm at each location. Lines showing areas of similar depth are shown for the central areas of WCA-2A

of water depths throughout parts of WCA-2A. Among the most surprising but consistent patterns we discovered by these analyses were the general increases in water depth from north to south in WCA-2A (Fig. 9.11). This hydroperiod pattern was consistently found in both the 1980–1990 and 1990–1996 time periods. During both periods the northern sections of WCA-2A experienced periods of over 40 consecutive days of less than 5 cm of water depth whereas less than 10 days were found in the south. At deeper water depths the pattern became more pronounced, with areas in the south experiencing nearly 200 consecutive days of water depth greater than 30 cm and areas in the north generally less than 50 days (Fig. 9.11). The north-to-south time trends and patterns also existed at both the 60 and 90 cm depth (Fig. 9.11). An analysis of the current distribution patterns for cattail and sawgrass (Fig. 9.5) along this north-to-south gradient suggests that the highest density of cattail is found in the areas with the longest dry periods and sawgrass in areas with the longest consecutive days of deeper water (Fig. 9.11). On face value this finding seems to contradict our understanding of the requirements for each species, that cattail prefers deeper water and sawgrass shallower water depths. However, the drier periods when soils are slightly moist are often critical in the life cycle of some species, enhancing germination and establishment as well as growth and expansion (Chap. 22). In contrast, long-term deep water (e.g., depths > 60 cm) restricts the establishment of many species. The drier habitat found in the north may also help explain the increased diversity of species, including upland plants, found in this area (Fig. 9.6) as compared with the southern areas (Vaithiyanathan and Richardson 1999). Consistent with this

244

C.J. Richardson et al.

water-depth pattern is the increase in the amount of bare open-water area (slough), with increased water depths in the south (Figs. 9.6 and 9.11). These open areas are significantly related (p < 0.05) to the average duration of water depth greater than 30 cm during 1980–1990 period. The most consistent vegetational relationship found by the Pearson’s Correlation analysis over all the four sampling dates (1990, 1992, 1994, and 1996) was for the average number of days of duration of water depth less than 5 cm for the 1980–1990 time period. This suggests that the 10 years prior to 1990 were important in determining plant species composition from 1990 to 1996. Cattail density was negatively (p < 0.01) related (r = −0.83) and sawgrass positively (p < 0.005) related (r = 0.80) to the number of consecutive days at 5-cm depth during each of the 6 years of analyses. A further analysis of the relationship of cattail frequency to increased water depth was also found to be positive (r = 0.66) for average water duration at > 90-cm depth. Sawgrass by contrast showed a negative relationship (r = −0.81) to increased water depth. These findings, however, do not take into account the full life cycle of these species (germination, establishment, and growth requirements as mentioned earlier) nor do our studies assess the full health of the stands. The importance of moisture and anaerobic conditions for the germination and initial growth of both sawgrass and cattail is presented in detail in Chap. 22. Our findings suggest that the hydrologic conditions in WCA-2A have been greatly altered by the pumping and water-level management regimes maintained by the SFWMD (Figs. 7.3 and 9.11) and that these changed hydrologic conditions are conducive to maintaining cattail populations in the north and sawgrass to the south. They also indicate that maintaining low water levels in highly enriched P soils may result in the invasion of extensive cattail populations. Further evidence to support this contention was found during our hydrology/disturbance experimental research in WCA2B (Richardson and Zahina 1995; see Chap. 21). Here we found that cattail invaded all sites subjected to vegetation removal, soil disturbance, and P fertilization as long as water depths were maintained below 30 cm. While the dynamic interaction of hydrologic alterations with increased nutrient loadings is not well understood in terms of its effects on vegetation life cycles, it is clear that the past SFWMD water management plan for WCA-2A favored the encroachment and maintenance of cattail in the northern part of WCA-2A. The effects of reduced nutrient input of N and P and altered water flows into WCA-2A due to improved BMPs (SFWMD 2006) will now again alter the response of the plant and animal communities in the WCAs but the outcome is unknown. Importantly, our long-term vegetation plots may provide a baseline for assessing future changes in this area. The effect fire has on vegetation in conjunction with these factors is explored in Sect. 9.4.4.3.

9.4.4.3

Fire

It has been suggested that fire is important in maintaining the Everglades sawgrass communities (Alexander 1971, Hofstetter 1983). In July 1989, the area south of the S10-D gate was characterized by Larson and Associates (unpublished data) as containing 30 to > 50% cattail. Extrapolating Larson’s data to our locations, the 10D-2 site

9 Macrophyte Community Responses in the Everglades

245

Fig. 9.12 Changes in plant species composition at 10D-2 and 10D-3 in response to a winter freeze followed by a summer fire in WCA-2A in the northern Everglades. The relative density of plant species is shown for cattail, sawgrass, and other species

(3.2 km south of the gate) consisted of approximately 50% cattail while the 10D-3 site (5.1-km south) contained 30% cattail (Fig. 9.12a, b). In January 1990, we established permanent plots at these (and other) locations to investigate changes in soil and water chemistry as a result of nutrient and water loading to WCA-2A. At this time, the 10D-2 site was dominated by standing dead cattail while the 10D-3 site was dominated by sawgrass (Fig. 9.12a, b). A hard freeze in December 1989 following an extensive drought earlier that year (Fig. 9.3) resulted in dieback of the cattail, but sawgrass was unaffected. In August 1990, we established permanent vegetation transects at these (and other) locations to monitor changes in vegetation over time. Both 10D-2 and 10D-3 were burned in June 1990. Shortly after the burn, sawgrass was the dominant plant species at both 10D-2 and 10D-3 (Fig. 9.12a, b). The combination of a severe freeze (during the preceding winter) and the summer fire succeeded in killing much of the cattail at these locations in August 1990. A survey of 10D-2 and 10D-3 in February 1991, indicated that sawgrass was the dominant species (> 80%) although cattail had recolonized nearly 10% of both sites. A time-series comparison of the C3 (unburned) and D3 (burned) sites with sites at C2 (unburned) and D2 (burned), at locations 3 and 5 km from input gates, indicates

246

C.J. Richardson et al.

that soil TP is highly enriched in all locations. It ranged from 900 to 1,300 mg kg−1 of TP at 0–10 cm at the C3/D3 and C2/D2 sites, respectively, but with dissimilar amounts of sawgrass vs. cattail from 1990 to 1996 (Fig. 9.6). Cattail frequency ranged from 60% in 1990 at the C2 site (3.5 km) to a high of 80% in 1992 and then declined over the 1994–1996 period to only 35% (Figs. 9.5 and 9.6). However, by 2002 cattail again comprised 60% of the vegetation. Cattail populations were nonexistent at D2 (3.5 km) in 1990 due to the previously mentioned fire. However, the cattail frequency increased to 10% in 1992, then 25% in 1994, and reached 60% by 1996. From 1996 to 2002 cattail frequency stayed around 40%. Cattail populations at C3 remained near 10% during the 6-year period but the burned D3 site showed an increase of cattail from 0 to 20% by 1992, a value that did not change over the next 4 years (Fig. 9.5). These findings suggest that a hard freeze and/or fire may reverse the encroachment of cattail into P-enriched areas of the Everglades and allow reestablishment of sawgrass, as found along the D transect. Moreover, the reversals in cattail populations are greatly influenced by fire, drought, and freezing effects. This suggests that highly enriched cattail areas can be managed to revert to almost pure sawgrass after fire if growth conditions are suitable, but that cattail will reinvade in 5–6 years to previous levels unless fire management regimes are continued. However, the long-term consequences of the interactions of fire on plant community changes are not fully understood and further research is badly needed to understand these interactions. What is clear is that cattail and sawgrass populations can be managed by fire (Chap. 26), although the burning of cattail stands is difficult unless drought or reductions in water release are present prior to burning.

9.4.5

A Multivariate Assessment of Linkages Among Environmental Factors and Plant Communities

9.4.5.1

Environmental Variation Among Impact Zones

In 1998, we divided the C transect gradient into three zones of impact based on soil P concentrations (impacted by nutrients 0–3 km, a transition zone 3–7 km, and a reference zone > 7 km) to provide three distinct regions from which to assess key environmental impacts (King et al. 2004). Three of the 11 environmental variables analyzed using ANOVA differed among the landscape impact zones (Table 9.4). All three zones differed for P and IQR (water depth), while Na differed between impacted and reference zones. Values of Ca, water depth, and water freq. > −10 cm (below the surface depth) tended to be greatest in the reference zone but were not statistically different than other zones because of high heterogeneity among plots within clusters in that zone (reflecting the heterogeneity of the Cladium-slough mosaic). All 11 variables yielded a significant cluster effect, indicating coarse-scale spatial differences in the environment among clusters within one, two, or all of the three zones. A detailed analysis of the coarse and fine spatial differences among these sites is given in King et al. (2004). Sect. 9.4.5.2 represents a synthesis of the key findings from this analysis.

Landscape zone Variable

Code

Units

F(2,11)a c

p

Impacted (n = 45)b

Transition (n = 45)

Reference (n = 36)

Distance from canal Canal m – – 2,495 (869) 5,541 (914) 9,050 (924) 0.15 NS 435.0 (20.3) 435.0 (27.1) 428.0 (47.7) Total carbon (soil) C g kg−1 Total calcium (soil) Ca g kg−1 0.46 NS 37.1 (16.5) 42.8 (20.8) 47.0 (34.5) Total potassium (soil) K mg kg−1 0.06 NS 581.7 (29.5) 557.4 (58.4) 527.0 (28.3) Total magnesium (soil) Mg mg kg−1 0.11 NS 3,707 (111) 3,859 (140) 3,625 (146) Total sodium (soil) Na mg kg−1 4.31 0.032 2,900 (173) ab 2,162 (187) b 3,058 (160) a Total nitrogen (soil) N g kg−1 0.02 NS 29.3 (2.2) 29.0 (3.7) 29.2 (4.4) Total phosphorus (soil) P mg kg−1 102.30 ≤ 0.001 1,434 (174) a 1,198 (184) b 578 (152) c Water depth (1 year) Depth cm 3.74 NS 35.7 (8.3) 41.8 (9.6) 46.4 (10.4) Interquartile range, water depthd IQR(depth) cm 20.90 ≤ 0.001 28.2 (0.1) a 29.7 (0.2) b 33.6 (0.1) c Frequency, water depth < −10 cmd Freq. < −10 cm % 2.79 NS 3.1 (0.4) 3.1 (0.3) 6.0 (0.8) Fire indexd Fire Sume 0.46 NS 0.2 (0.4) 0.4 (0.5) 0.3 (0.5) a F ratios and associated p values correspond to the zone main effect (fixed effect). Cluster was a random effect nested within zone and was used as the error term. Cluster (random effect, F11,112 was significant ( p ≤ 0.05) for all variables b Mean (±1 SD) values are based on measurements collected at individual plots within landscape zones. LSD tests were used to compare means among levels of the zone effect when deemed significant from ANOVAs. Means with the same superscript letters do not differ ( p > 0.05) c ANOVA not conducted on Canal since it was not independent of the landscape zones d 1981–1998 e Sum of total number of fires/plot during 1981–1998, weighted as 1/log(t + 1) for each fire, where t = time (years) since fire

9 Macrophyte Community Responses in the Everglades

Table 9.4 Analysis of significant and nonsignificant variables based on a mixed-model nested ANOVA on selected environmental characteristics from impacted, transition, and reference zones (from King et al. 2004)

247

248

9.4.5.2

C.J. Richardson et al.

Macrophyte Species Distributions

Forty macrophyte species and three other cover types (open water, noncalcareous periphyton, and calcareous periphyton) were identified (Table 9.5). Eight species were significant indicators of the impacted zone, while six were indictors of the reference zone. Despite having the most species, no species or cover types were indicators of the transition zone, reaffirming it to be an ecological area of transition between the impacted and reference zones because it hosted patchy distributions of species common to both other zones (Table 9.5). T. domingensis and invasive vines M. scandens and S. clausum were the best indicators of the impacted zone; however, small floating species Lemna spp. and S. minima, as well as the understory herb Rumex verticillatus, also had strong affinities for sparse-canopied areas there. The invasive willow S. caroliniana was the only woody shrub with a significant indicator value; it was common in the impacted zone. C. jamaicense was a significant indicator of the reference zone despite being fairly common in the transition

Table 9.5 List of macrophyte species and cover types, including corresponding codes (see Figs. 9.2 and 9.3) and indicator values (IVs) for each landscape zone (from King et al. 2004) Species/cover type Acrostichum danaeifolium Langsd. & Fitch. Alternanthera philoxeroides (Mart.) Griseb. Amaranthus australis (Gray) Sauer. Aster sp. Bacopa sp. Calcareous periphyton mat Cephalanthus occidentalis L. Ceratophyllum demersum L. Chara sp. Cladium jamaicense Crantz. Crinum americanum L. Cyperus odoratus L. Eichornia crassipes (Mart.) Solms. Eleocharis cellulosa Torr. Eleocharis elongata Chapm. Hydrocotyle umbellata Lamark. Ipomoea sagittata Poir. Lemna sp. Limnobium spongia (Bosc.) Steud. Ludwigia leptocarpa (Nutt.) Ludwigia repens Forst.

Indicator value (IV) Impacted Transition Reference ACRODANA 0 6 0

p

ALTEPHIL

Code

NSa

7

0

0

NS

AMARAUST 1

1

0

NS

ASTER BACOPA CALMAT CEPHOCCI CERADEME CHARA CLADJAMA CRINAMER CYPEODOR EICHCRAS

0 0 2 3 4 5 28 1 0 0

0 0 37 1 0 19 46 8 0 0

NS NS £ 0.0001 NS NS 0.0055b £ 0.0001 0.0317b NS 0.0353b

ELEOCELL 0 ELEOELON 0 HYDRUMBE 11

1 0 1

13 29 0

0.0107b £ 0.0001 0.0190b

IPOMSAGI LEMNA LIMNSPON

1 30 0

1 4 4

2 0 0

NS £ 0.0001 NS

LUDWLEPT LUDWREPE

4 3

0 5

0 0

NS NS

2 2 0 5 0 0 6 0 2 9

(continued)

9 Macrophyte Community Responses in the Everglades

249

Table 9.5. (continued) Indicator value (IV) Species/cover type

Code

Impacted

Transition

Reference

p

Mikania scandens (L.) Willd. Nymphaea odorata Aiton. Open water – no cover Panicum repens L. Peltandra virginica (L.) Schott & Endl. Pistia stratiotes L. Poaceae sp. Polygonum densiflorum Meisn. Polygonum punctatum Ell. Pontederia cordata L. Rumex cf. verticillatus L. Sagittaria lancifolia L. Salix caroliniana Michx. Salvinia minima Baker. Sarcostemma clausum (Jacq.) Schult. Scirpus validus Vahl. Typha domingensis Pers. Utricularia fibrosa Walt. Utricularia foliosa L. Utricularia purpurea Walt. Wolfiella sp.

MIKASCAN NYMPODOR OPEN PANIREPE PELTVIRG

61 0 19 0 0

8 9 34 0 0

0 39 34 3 3

£ 0.0001 £ 0.0001 NS NS NS

PISTSTRA POACEAE POLYDENS POLYPUNC PONTCORD RUMEVERT SAGILANC SALICARO SALVMINI SARCCLAU

7 13 6 25 8 36 31 20 25 39

2 0 1 16 6 0 14 1 0 1

0 0 0 0 0 0 1 1 0 0

NS 0.0047b NS 0.0116b NS £ 0.0001 0.0011 £ 0.0001 £ 0.0001 ≤0.0001

SCIRVALI TYPHDOMI UTRIFIBR UTRIFOLI UTRIPURP WOLFIELL

3 58 0 0 0 4

1 31 10 6 0 0

0 2 37 14 48 0

NS £ 0.0001 £ 0.0001 0.0312b £ 0.0001 NS

a Indicates IVs that are not significant. IVs are percentage of perfect indication with significant (Bonferroni-corrected p ≤ 0.05) scores shown in bold b Not significant after Bonferroni correction

zone as well. Four slough community species also showed high fidelity to the reference zone: N. odorata (water lily), Eleocharis elongata (spikerush), and U. purpurea and Utricularia fibrosa (bladderworts). These same species were also key indicators of the unenriched areas in an earlier study in WCA-2A by Vaithiyanathan and Richardson (1999). Calcareous periphyton mat also was a significant indicator and was almost exclusively found in the reference zone.

9.4.5.3

Vegetation–Environment Linkages

Ordination of fine-scale composition resulted in two gradients: (1) a coarse-scale gradient significantly associated with Canal, P, IQR (depth), and depth, and (2) a fine-scale gradient related to freq. < −10 cm, fire, K, N, and Na (Fig. 9.13a–c). Canal was most strongly linked to variation in composition but was closely followed by P and IQR (depth) (Fig. 9.13a). Composition differed markedly among impact zones, as plots were sorted accordingly along nMDS Axis 1. Reference-zone vegetation, although tightly banded at one end of nMDS Axis 1, showed much fine-scale

250

C.J. Richardson et al.

Fig. 9.13 Nonmetric multidimensional scaling (nMDS) ordination of (a) individual plots, (b) species/cover-type centroids using fine-scale vegetation species composition, and (c) clusters using coarse-scale species composition (calculated as averages of the 14 plot-clusters). Symbols indicate membership among the three impact zones. Environmental vectors show the direction and magnitude of significant correlations (r values shown adjacent to vectors) within the ordination space. See Tables 9.4 and 9.5 for codes for environmental and species variables, respectively. * p ≤ 0.05; **p ≤ 0.001; ***p ≤ 0.0001 (from King et al. 2004). Reproduced from King et al. (2004), with kind permission from Springer Science and Business Media

9 Macrophyte Community Responses in the Everglades

251

Fig. 9.13 (continued)

variation as evidenced by great dispersion (β-diversity) along nMDS Axis 2. However, the transition zone showed the greatest overall variation in composition, with much spread along nMDS Axes 1 and 2, and several species sampling centroids were unevenly dispersed in ordination space. Remarkably few species were projected in the region corresponding to the transition zone; most centroids were clearly associated with the impacted or reference zones (Fig. 9.13a). In the reference zone, slough-species centroids closely corresponded to vectors of water depth, N, and K, while Cladium (CLADJAMA) was at the opposite end of this within-zone gradient and corresponded to freq. <−10 cm and Fire (Fig. 9.13b). Centroids were most tightly aggregated in the impacted zone, suggesting less distinction of discrete communities. Coarse-scale ordination of vegetation composition revealed that canal, P, IQR (depth), and depth were the primary correlates of coarse-scale pattern (Fig. 9.13c). Virtually all variation in composition occurred along nMDS Axis 1, and this gradient mirrored that of Axis 1 in the fine-scale ordination (Fig. 9.13a, b). Clusters were completely separated into distinct strata corresponding to the three zones of impact, which provided further support for the gradient/zone concept in the study area (Fig. 9.13c).

9.4.5.4

Partial Mantel Tests

Partial Mantel tests on fine-scale data from the full vegetation gradient indicated that most environmental variables were spatially autocorrelated (Fig. 9.14). However, only four variables, Na, P, IQR (depth), and Depth, were directly related to canal, corroborating results from ANOVA among the landscape zones (Table 9.4). Considering simple relationships, vegetation composition was significantly related to all but two variables (Na and fire). However, after extracting spatial autocorrelation (space), Ca and Mg were no longer significant (King et al. 2004). The variables

252

C.J. Richardson et al.

Fig. 9.14 Path diagram depicting the relationships among space, canal, individual environmental variables, and vegetation composition at (a) fine and (b) coarse scales along the vegetation gradient, as estimated using Mantel tests. Significant pure-partial relationships between variables and vegetation are shown by solid arrows, while relationships that were significant after accounting for spatial autocorrelation, but not as pure-partials, are shown as dotted arrows. Thickness of arrows is proportional to magnitude of relationship (from King et al. 2004). Reproduced from King et al. (2004), with kind permission from Springer Science and Business Media

that corresponded to nMDS Axis 1 (Fig. 9.13), Canal, P, and IQR (depth), were most closely tied to composition regardless of spatial dependencies. N, K, and C exhibited the strongest link to composition of the remaining fine-scale variables, although depth and freq. < −10 cm remained significantly correlated to vegetation as well. Pure-partial tests revealed that C, depth, and IQR (depth) could not account for unique variation in the vegetation composition (Fig. 9.14a). However, P remained highly significant. The relationship between freq. <−10 cm and vegetation actually improved slightly after variation from all other variables had been removed. N and K also remained highly significant as pure-partials. Confidence limits (95% CLs) generated by bootstrapping indicated that canal distance was the strongest factor (partial Mantel r = 0.26, lower 2.5% = 0.24, upper 2.5% = 0.28) – a

9 Macrophyte Community Responses in the Everglades

253

relationship that was indicative of variation in composition caused by the canaland-levee system but unexplained by any of the other measured spatial or environmental variables. Finally, a highly significant spatial residual suggested that spatial factors contributed to variation in vegetation pattern that could not be explained by the environment or canal. Mantel results from the coarse-scale analysis generally supported those of the coarse-scale nMDS ordination, as canal, P, and IQR (depth) were highly associated with vegetation (Fig. 9.14b). However, IQR (depth) was not significant as a purepartial. Canal and P were the only variables that significantly accounted for coarsescale variation that could not be accounted for by other variables. Although spatially autocorrelated, coarse-scale composition did not exhibit a significant spatial residual, indicating that coarse spatial patterns were mostly attributed to canal and the environment. Mantel analysis of data from just the impacted zone revealed that fewer environmental variables were dependent upon space or canal than in the full data set, reflecting greater homogeneity in the environment (Fig. 9.15a). Vegetation also had weaker linkages to the environment in this zone. N was the only environmental variable that was significant as a pure-partial. Space explained the most variation in composition in the impacted zone. Environmental variables measured in transition plots showed greater spatial dependency than in the impacted zone (Fig. 9.15b). Nevertheless, numerous environmental variables were linked to vegetation even after correcting for spatial autocorrelation and mutual correlations among variables. N and K had the strongest relationships to vegetation, although depth, freq. < −10 cm, and C also were significant. Space was the most influential factor, with lower confidence limits of its pure-partial Mantel coefficient exceeding the upper limits of any other variable (partial Mantel r = 0.48, lower 2.5% = 0.45, upper 2.5% = 0.50). In the reference zone, most environmental variables were spatially autocorrelated (Fig. 9.15c). However, depth, freq. < −10 cm, C, Ca, N, and P were significantly linked to vegetation regardless of spatial dependencies. Freq. < −10 cm, C, K, N, and P all remained linked to vegetation after variance explained by other variables had been removed. Vegetation was not spatially autocorrelated, nor was there a significant spatial residual, which indicated that the spatial distribution of macrophyte species was similar across the reference landscape. Based on 95% CLs of partial Mantel coefficients, space was significantly more important to vegetation patterns in transition and impacted zones than in the reference zone.

9.4.5.5

Vegetation–Environment Linkages: Allogenic or Autogenic?

Results from partial Mantel tests suggested that several of the spatial and environmental variables were independently related to vegetation composition. While our approach effectively removed variation accounted for by space or other variables and thus provided strong evidence for variables directly linked to vegetation patterns, it did not conclusively establish the nature of these relationships. That is, were environmental factors causing variation in vegetation composition (allogenic) or were existing patterns in composition causing observed environmental variation

254

C.J. Richardson et al.

Fine scale (a) Impacted

Vegetation

Depth C

N

P

Ca

Na

Mg

IQR(Depth)

K

Fire Freq< -10cm

Canal

Space

(b) Transition

Vegetation

Depth C

N

P

Ca

Na

Mg

IQR(Depth)

K

Fire Freq< -10cm

Canal

Space

(c) Reference

Vegetation

Depth C

N

P

Ca

Na

Mg

IQR(Depth)

K

Fire Freq< -10cm

Canal

Space

Fig. 9.15 Path diagrams depicting the relationships among space, canal, individual environmental variables, and fine-scale vegetation composition within (a) impacted (0–3 km), (b) transition (3– 7 km), and (c) reference zones (> 7 km), as estimated using Mantel tests. Significant pure-partial relationships between variables and vegetation are shown by solid arrows, while relationships that were significant after accounting for spatial autocorrelation, but not as pure-partials, are shown as dotted arrows. Thickness of arrows is proportional to magnitude of relationship (from King et al. 2004). Reproduced from King et al. (2004), with kind permission from Springer Science and Business Media

(autogenic)? The variables most closely related to the primary vegetation gradient were P, canal distance, and water depth, and they were allogenic factors. In the case of P, it is well established that Everglades vegetation is P-limited, with most species adapted to highly oligotrophic conditions (Richardson et al. 1999; Noe et al. 2001). Several species, particularly those comprising the slough community, have been

9 Macrophyte Community Responses in the Everglades

255

shown to be sensitive to elevated P in both observational (Vaithiyanathan and Richardson 1999) and experimental (Walker et al. 1989; Craft et al. 1995; Chiang et al. 2000; see Chap. 16) settings. Conversely, many species found to be most abundant in the impacted and transition zones are opportunistic, “weedy” species that are highly competitive in high-P environments (Davis 1994). Typha in particular is clearly most competitive under elevated nutrient conditions (Urban et al. 1993; Newman et al. 1996; Miao and Sklar 1998; Miao et al. 2000) and has been positively associated with P gradients in many locations in the Everglades (e.g., Doren et al. 1997). Thus, that P was significantly and independently linked to variation in vegetation pattern is not surprising and corroborates results from other studies that suggest P enrichment is a major source of perturbation to Everglades plant communities. Stability in water depth may have acted in concert with P to promote the establishment of invasive species such as Typha, although cattail does not seem to become established easily in water > 30 cm. Frequency of severe dryness and variability of water depth were significantly related to vegetation composition as noted earlier for cattail along this gradient, although water depth was not significant as a pure-partial coefficient. Several authors have suggested that hydropattern plays a role in the observed expansion of Typha in the Everglades, as Typha is highly competitive in deeper, more stable water conditions but intolerant of drought (Toth 1988; Urban et al. 1993; Newman et al. 1996). Cladium, on the other hand, is well adapted for dynamic hydrologic conditions but exhibits a diminished capacity to resist invasion by Typha under deep, stable water conditions (Toth 1987; Davis 1994; Newman et al. 1996). Our data showed that mean water depths from the previous year actually were greater with increasing distance from the canal, while variation in water depth was most stable near the canal inflow structures where Typha was most prolific. This does not contradict our earlier correlation analysis from the 1990 to 1996 period where we found that cattail becomes established during periods of low water in enriched areas but survives and flourishes in deeper water. Severe dryness tended to be experienced more frequently in the reference zone, and plots in this area also exhibited the greatest fluctuation in water levels. Concomitantly, Cladium dominated these drought-susceptible and hydrologically dynamic plots, as evidenced in the nMDS ordinations. The significant relationship between vegetation and hydrologic variability supports the hypothesis that modifications to hydropattern related to the canal-and-levee system have at least partially contributed to observed vegetation patterns. It further supports the hypothesis that hydrology may act synergistically with P to promote Typha and other invaders since experimental fertilizer studies as noted earlier have been unable to demonstrate that P enrichment alone results in competitive exclusion of Cladium (Craft et al. 1995; Chiang et al. 2000). Possibly the most interesting linkage revealed along the full vegetation gradient was the residual variation in composition that could only be explained by distance from the canal inflow structures. Canal had a direct effect on P and water depth, a strong indication that variation in these factors was due to water entering the wetland through canal inflow structures (SFWMD 1992; see Chap. 7). However, even after

256

C.J. Richardson et al.

variation from these and all other remaining variables was removed, Canal distance remained a significant correlate of composition; its partial correlation was greater than any other variable. Possibly, canal distance captured the synergistic effect of P and hydropattern that was not accounted for when those respective factors were considered individually. This may be partially a flow effect, i.e., water historically was not pumped and forced to flow in confined areas. In addition, this observation may have resulted from one or a combination of several other factors: (a) environmental variables not measured but directly influenced by the canal were acting to structure vegetation (e.g., micronutrients, sulfate; see Chap. 6); (b) proximity to the canal affected seed and/or floating plant dispersal (Vaithiyanathan and Richardson 1999); (c) a “front” of succeeding vegetation assemblages that began near the canal has progressed into the interior of the study area but asynchronously with environmental changes (i.e., level of P enrichment) (Wu et al. 1997; Richardson et al. 1999); and (d) other contagious spatial effects directly related to proximity to canal (e.g., herbivory or disease) (Richardson et al. 1999). Indeed, other potential factors beyond these surely exist and warrant examination in future research. At a minimum, it seems reasonable to conclude that influences directly attributable to canal-and-levee systems represent a serious threat to the long-term integrity of the Everglades and other large, wetland ecosystems (e.g., Delta Marsh) (Shay et al. 1999). While proximity to the canal inflow structures, elevated P, and modified hydropatterns were allogenic determinants of vegetation patterns along the anthropogenic influence gradient, the nature of the linkages between significant fine-scale environmental variables and vegetation was much less clear. Of particular interest was the significant linkage between N and vegetation composition. Several studies have shown that N additions do not stimulate growth in Everglades plant communities (Steward and Ornes 1975b; Walker et al. 1989; Craft et al. 1995; Chiang et al. 2000) and that all but the most heavily P-enriched locations are P-limited (Richardson et al. 1999). Nevertheless, N accounted for variation in vegetation along the full vegetation gradient and within all three impact zones as no other variable did. Within the transition and reference zones, highest N (and K) concentrations were associated with periphyton mats and slough community species, particularly the water lily N. odorata. Decomposing Nymphaea tissue has been shown to have greater N concentrations than more recalcitrant species like Cladium (Steward and Ornes 1975a; J. Vymazal, unpublished data). Periphyton mats are largely composed of blue-green algae, many of which are heterocystic N-fixers (Swift and Nicholas 1987; Browder et al. 1994), and also tend to have relatively high tissue N concentrations (Vymazal and Richardson 1995). Indeed, Gleason and Stone (1994) describe two principal sediment types in the Everglades that closely correspond to our observations: Everglades peat, generated by Cladium, and Loxahatchee peat, a product of Nymphaea slough communities. Thus, N may have been an excellent indicator of fine-scale composition because the vegetation itself may have been causing variation in N concentrations – an excellent illustration of the effect of pattern on process (Watt 1947). However, nitrogen may have played a more direct role in generating fine-scale compositional patterns in the impacted zone. It was the only environmental variable that was independently linked to vegetation composition in this zone. High soil P

9 Macrophyte Community Responses in the Everglades

257

concentrations adjacent to canal inflows have been suggested to result in some instances of N limitation (e.g., Richardson et al. 1999). This fact, coupled with the autogenic soil legacies described by Gleason and Stone (1994), suggests that remnant patches of soil containing elevated N may contribute to N-dependence in locations where P is no longer limiting. S. caroliniana (willow), an indicator of the impacted zone, was particularly abundant in areas of elevated N. Understory species (R. verticillatus, Lemna spp., S. minima) were closely associated with Salix stands, an indication that these species may have benefited from the reduction in canopy cover imposed by Typha (Grimshaw et al. 1997). The pre-impact physical template may have influenced observed fine-scale pattern in the impacted zone, and subsequently it may have contributed to the significant relationship between N and vegetation. A closer examination of spatial patterns of N and vegetation is needed to better evaluate its role in structuring fine-scale community composition in P-enriched areas of the Everglades. Considering ion–vegetation relationships, our results support the conclusions of Craft and Richardson (1997), who examined the potential relationship between Typha expansion and elevated Na, Ca, and Mg in soils near inflow structures. Using simple correlations, they showed that Typha was most strongly related to soil P and concluded that cations were not important to its distribution, even though Typha also was significantly correlated to Na. In our study, none of these cations was able to account for unexplained variation in vegetation. While Na and other cations cannot be completely ruled out as partial determinants, our results suggest that they may not be significant agents of pattern formation. However, the role of sulfate deserves further consideration since elevated levels of this anion, as well as toxic levels of hydrogen sulfide, have been measured along this gradient, greatly influencing redox conditions (C.J. Richardson, unpublished data; Qualls et al. 2001). Fire has also been shown to be an important disturbance in the maintaining of the mosaic of Everglades plant communities (Craighead 1971, Gunderson and Snyder 1994). A wide variety of vegetation responses to fire have been reported, primarily focusing on Typha and Cladium. These responses have depended largely on fire intensity and the presence of nutrient enrichment (e.g., Urban et al. 1993; Richardson et al. 1997a; Newman et al. 1998). In our study, fire was weakly but significantly correlated to composition in the ordination of the full vegetation gradient. Fire also covaried with freq. < −10 cm, as plots that were dry most frequently tended to burn more often. Additionally, Cladium was the dominant species in plots that were most susceptible to both drying and burning. However, fire was not linked to vegetation composition in the partial Mantel analysis. Thus, our results are equivocal. In general, these findings imply that fire covaries with other stronger patterns of environmental variation and thus is not able to explain a unique component of variation in the vegetation. It is important to note that our fire index only considered large fires, and only those that had occurred since 1981. Fires also were spatially contagious; thus, autocorrelation greatly inhibited our ability to detect fire effects. Due to the tremendous number of potential interactions with other variables and the high variability of fire frequency, intensity, and spatial extent, it is highly unlikely that simple patterns will emerge (e.g., Newman et al. 1998). For long-term

258

C.J. Richardson et al.

management and restoration of Everglades vegetation to succeed, a better understanding of the role of fire in community dynamics is clearly needed.

9.5

Conclusions and Lessons for Restoration

These studies have been able to link the importance of rainfall patterns, nutrient inputs, hydrologic regimes, and fire to vegetation distributions and patterns in the Everglades. Our surveys of vegetation patterns (1990–2002) and statistical analyses of hydrologic conditions, nutrient levels, fire, and disturbance show that plant populations in the Everglades oscillate greatly over short time periods due to impacts of all these variables but to varying degrees depending on scale and location along gradients. For example, the frequency of cattail was found to vary greatly (0–80% of the plant population) at P-enriched sites depending on the time since the last major fire. In this case dominant cattail populations vacillated between total dominance in the late 1980s in highly P-enriched areas near the input canals to near extirpation in 1990 following a fire that was preceded by a drought and a fatal winter freeze in the 1989. The current expansion of cattail and overall vegetation distribution patterns in WCA-2A and WCA-3A are the direct result of both increased phosphorus inputs, altered hydrologic regimes, and fire, especially during the past three decades. The most important nutrient controlling cattail density and distribution is phosphorus, but there are indications that other ions such as sodium may play some role for certain species. Moreover, it is evident that fire plays an important role in the revegetation of sawgrass, regardless of the elevated P soil nutrient regime. This is especially true for WCA-3A, where shallow water conditions in the north have exacerbated fires in the northern regions while elevated waters have dominated the vegetation patterns and litter buildup in the southern areas (Chap. 6). It is also apparent that P can accelerate the growth and greatly increase the density of cattail once it has become established at soil TP concentrations above 500–600 mg kg−1 (0–10 cm). Conversely, sawgrass populations decrease at higher soil P concentrations, but only above 1,000 mg kg−1 of TP (0–10 cm). However, the influence of altered hydrology apparently plays an important role in cattail establishment following the loss of sawgrass. Along the 10-km gradient, in WCA-2A, partial Mantel tests showed that nutrients (P, N, and K) and hydropattern (frequency of dryness) were independently linked to patterns in fine-scale vegetation composition, but P was the only environmental variable linked to patterns of coarse-scale composition. Regardless of scale, the effect of distance from canal inflows accounted for variation in vegetation that could not be explained by other variables. A significant residual effect of spatial proximity among sampling locations also was detected and was highly suggestive of dispersal or other spatial determinants of vegetation pattern. However, this pure spatial effect was significantly stronger in the transition and impacted zones than in the reference zone. Fine-scale environmental variables explained all of the spatial structure in vegetation in the reference zone. Collectively, these results indicate that allogenic spatial and

9 Macrophyte Community Responses in the Everglades

259

environmental factors related to the canal system have disrupted the coupling between pattern and process by altering fine-scale vegetation–environment linkages and spatial patterns characteristic of the natural Everglades ecosystem. For this reason, we need to elucidate the mechanisms for maintaining Everglades communities of sawgrass marsh and slough species in the face of encroachment by cattail and other undesirable plant species. The key to vegetation management for the Everglades will be to reduce P loadings to levels that do not exceed its P assimilative capacity (Richardson and Qian 1999; see Chap. 23). It has been suggested that this P threshold may be below 0.5 g m−2 year−1 for TP in the Everglades. Once P inputs are reduced below this level we have not seen an expansion of cattail or loss of native macrophyte species (Richardson et al. 1997a, Richardson and Qian 1999). Fire management is another mechanism that may be useful in preventing the spread of cattail into sawgrass communities. Prescribed burning of sawgrass once every 3–5 years during the wet season may be effective in maintaining sawgrass communities against invasion by cattail, even in highly P-enriched soils. However, a great deal of research is needed to effectively understand the role of fire in maintaining Everglades plant communities. Analysis of hydrologic patterns from 1985 to 1996 suggested that the most important hydrologic pattern influencing cattail frequency was the number of consecutive days that a site had water levels at or below 5 cm of water depth. This finding, when combined with our experimental work, suggests that low water is first required for cattail invasions and that elevated P in the soil then acts as a catalyst to increase cattail dominance and expansion in these areas. By contrast, sawgrass populations can be maintained in areas where water depths are above 30 cm as long as sufficient drying periods also exist. Our studies clearly indicate that populations of sawgrass can be maintained by careful management of the hydrologic and fire regimes, at all of the current soil nutrient regimes found in WCA-2A. Increased water depth > 60 cm for extended periods in low P soil areas will result in decreased sawgrass populations and an increase in open-water slough areas. Unfortunately, the eclectic water management and nutrient release regimes used by the SFWMD, in conjunction with Corps of Engineers approval, for WCA-2A during the past 30 years has fostered the invasion and spread of cattail. Current water management practices will reduce P inputs levels but will also continue to favor cattail over sawgrass and slough species since normal drying and wetting cycles are being greatly altered along with fire patterns. Furthermore, it has been suggested that cattail increases can simply be controlled by maintaining P inputs at background concentrations close to undisturbed levels. This approach, while appealing, simply does not take into account the other factors that control cattail populations and the other species found in the Everglades. The amount of residual P currently stored in the soil near input structures and the continual diffusion of a portion of this P to the pore water and the water column suggest that cattail growth will be a problem for decades to come, even with no further increase in P inputs (Fisher and Reddy 2001; see Chap. 6). Our results indicate that human impacts to the Everglades have resulted in more than just a replacement of Cladium with Typha. Collectively, these results indicate

260

C.J. Richardson et al.

that allogenic spatial and environmental factors related to the canal system have disrupted the coupling between pattern and process by altering fine-scale vegetation– environment linkages and spatial patterns characteristic of the natural Everglades ecosystem. These alterations certainly affect the spatial ecology of higher organisms, such as invertebrates, fish, birds, reptiles, and other wildlife in a variety of ways that may depend largely upon landscape connectivity and critical scales in their individual life histories (e.g., MacArthur and Wilson 1967, Levin 1976). The implications of this are great for restoration and management of vegetation, suggesting that serious attention needs to be given toward mimicking the characteristic spatial and temporal scales of pattern in the environment (DeAngelis 1994, Holling et al. 1994), as these patterns ultimately drive self-organization in landscapes (Phillips 1999). The field of landscape ecology has already begun to address many of these scaling issues for terrestrial wildlife (e.g., spotted owl). Such management approaches should be extended to aquatic systems as well. One simple management application could be the use of correlograms of vegetation and environmental spatial patterns from reference areas as models to guide restoration efforts in disturbed areas of the Everglades ecosystem (King et al. 2004). Principles of hierarchy theory (e.g., Allen and Star 1982, O’Neill et al. 1986) may provide a framework for such efforts. In conclusion, our studies indicate a need for more research on the spatial and temporal scales that are responsible for vegetation patterns and on the role these patterns play in the demographic processes of the flora and fauna across the Everglades.

10

Algal Responses to Long-Term Nutrient Additions ˇ eháková, Jan Vymazal, Jaroslava Komárková, Klára R Jan Kaštovský, and Marek Bastl

10.1

Introduction

Periphyton is an important part of the food web and is an integral part of the Everglades ecosystem (Wood and Maynard 1974; Browder et al. 1982). During the late summer and fall, thick mats floating at the surface, growing attached either to macrophytes or at the bottom, are among the most conspicuous components of unenriched open water habitats in the Everglades (Van Meter-Kasanof 1973; Gleason and Spackman 1974; Swift 1984). Changes in periphyton standing crop, species composition, and productivity along an eutrophication gradient in the Everglades indicate that phosphorus affects the composition and growth of the periphyton community (Swift and Nicholas 1987; Vymazal et al. 1994; McCormick and O’Dell 1996; McCormick et al. 1996, 1998; Pan et al. 2000). Experimental manipulations of nutrients demonstrate that phosphorus also has a negative effect on the development of the calcareous periphyton assemblage (Ornes and Steward 1973; Hall and Rice 1990; Stevenson and Richardson 1994, 1995). Enriched Everglades habitats develop periphyton that is dominated by green algae and eutrophication-tolerant diatoms and cyanobacteria (blue-green algae, Cyanoprokaryota) (Swift and Nicholas 1987; Raschke 1993). However, no studies have ever determined the ecological importance of each algal group in the Everglades by assessing changes in their biovolume along a nutrient gradient.

10.2

Objectives of the Study

The major objectives of this study were (1) to identify the algal taxa found along a nutrient gradient in the northern Everglades, (2) to determine if algal taxa change in response to nutrient concentrations, (3) to evaluate major algal components and taxonomic groups of the periphyton community along the nutrient gradient, and (4) to evaluate periphyton biovolume along the C-transect to determine the ecological importance of each algal component. 261

262

J. Vymazal et al.

10.3

Study Site

In 1990, Duke researchers established six research plots in WCA-2A along each of three transects running south from gates 10-A, 10-C, and 10-D (Chap. 5, Fig. 5.1) The transects, running north to south, were roughly parallel to the dominant water flow from the drainage canal gates located on the southern side of the Hillsboro canal. The algal response study presented here was conducted along the C-transect. The distances of sampling points C1 to C6 from Hillsboro Canal were 0.68, 2.40, 4.09, 6.14, 7.99, and 9.84 km, respectively. In Table 10.1, surface water chemistry from the sampling period of October to December 1999 is presented. The long-term water chemistry along the C-transect is presented in Chap. 6 of this book. Site C1 is dominated by cattail (Typha domingensis), while saw grass (Cladium jamaicense) is not present. Saw grass first occurs at C2, and its abundance increases until at C6 it forms a monotypic stand. Cattail abundance decreases with distance from the Hillsboro canal, and cattail was not recorded at all at sites C5 and C6 (Richardson et al. 1999). More detailed description of the macrophyte species growing at sampling sites, together with long-term surface water chemistry and soil chemistry, can be found in Chap. 9.

10.4

Materials and Methods

10.4.1

Sampling

Plexiglas slides (7.5×2.5 cm) were deployed at the sampling sites on October 20, 1999. They were suspended 10 cm below the water surface, connected to Styrofoam floaters on the surface and weighted beneath with lead sinks to maintain a vertical Table 10.1 (A) Concentrations of total phosphorus (TP), soluble orthophosphate (PO4–P), and nitrogen (µg L−1, SD in parentheses), and (B) calcium, magnesium, potassium, and sodium (mg L−1, SD in parentheses) at sites C1–C6 during the period of October–December 1999 TN A TP PO4–P C1 C2 C3 C4 C5 C6

51.8 (12.6) n = 19 33.3 (8.1) n = 11 13.6 (5.9) n = 11 6.2 (2.7) n = 19 5.8 (2.3) n = 11 4.4 (3.4) n = 19

26.2 (14.7) n = 5 11.6 (4.2) n = 5 4.7 (1.5) n = 5 3.7 (1.1) n = 5 3.0 (0.74) n = 5 3.3 (1.8) n = 5

1,578 (252) n = 13 1,839 (323) n = 5 1,699 (202) n = 5 1,182 (235) n = 13 1,456 (178) n = 5 1,244 (108) n = 13

B

Ca

Mg

K

Na

C1 C2 C3 C4 C5 C6

68.2 (2.6) n = 10 68.8 (3.9) n = 2 67.5 (2.2) n = 2 48.5 (7.6) n = 10 61.4 (8.5) n = 2 51.6 (2.0) n = 10

22.5 (1.8) n = 10 24.6 (0.75) n = 2 22.2 (0.65) n = 2 15.6 (2.6) n = 10 20.6 (1.4) n = 2 17.1 (1.7) n = 10

6.75 (0.64) n = 10 7.00 (0.06) n = 2 7.20 (0.35) n = 2 4.98 (0.7) n = 10 6.71 (0.43) n = 2 5.52 (0.32) n = 10

70.9 (7.7) n = 10 77.8 (5.8) n = 2 69.0 (0.2) n = 2 50.7 (8.1) n = 10 66.2 (4.6) n = 2 59.4 (4.1) n = 10

10 Algal Responses to Long-Term Nutrient Additions

263

position at that depth. Six slides were used at each site, positioned within a 30-m circle around the center pole of each site (C1–C6). At site C6, four additional slides (A, B, C, and D) were suspended in the water next to (1–6 m) each other to test variability in periphyton species composition at a particular sampling point. Slides were sampled on December 6 after a 47-day exposure, placed into 50-ml centrifuge vials with 1.5% solution of formaldehyde, and transported in ice coolers to the laboratory. In the laboratory, the slides were taken out of the vials with forceps; then, the periphyton was carefully removed from the slides with a razor blade and rubber squeegee and placed back into the vial. If any algae stuck to the razor blade and squeegee when the slides were scraped down, it was washed back into the vial with the rest of the sample. The content of the vial was then quantitatively transferred to a homogenizer and homogenized until large structures were broken down. Then, distilled water was added to make up the sample’s final volume, in most cases, 50 mL. Samples of periphyton growing on Plexiglas were evaluated differently. A suitable volume of the sample (ca. 2–4 mL) was transferred by a pipette with a larger inlet hole into a small vial. One drop of 10% citric acid was added to the sample and left overnight. If the carbonate was still present in the sample the next day, it was necessary to add one more drop of citric acid. The sample was then rinsed with water, acid water removed, and the sample with smaller volume compared to an initial volume and again transferred to a smaller homogenizer. The homogenized sample was then brought to an initial volume with water. To have suitable density of the sample it was necessary to either dilute or thicken the sample. The sample was then extensively shaken, and from the homogenous sample 2×20 µl was subsampled by micropipette. Both samples were spread with a cover slide on a microscope slide. Water was partially evaporated and then the cover glass was sealed with a fingernail polish. The counting of organisms was finished before the sample dried out. If the material was heterogeneous, it was necessary to take more subsamples, and then the average numbers were used. For calculation of biovolume, the following information was determined: initial sample volume, further dilution or condensation of the sample, volume of sample used for calculation in the microscope, real dimensions of the measuring abscissa, dimensions of the cover-slide, and real diameter of the view-area. The algal mat is an extremely diversified and complicated matrix for biomass estimation, especially estimating individual species or size forms. The mat consists of minute, bacterial-size cells living either solitarily or in colonies; large cells of diatoms or desmids; fine filaments of Leptolyngbya or Pseudanabaena; thick, very long filaments of Lyngbya and Oscillatoria; branched, relatively large species of Scytonema and Tolypothrix; and tissue-like species of Stigonema. As the commonly used methods for biomass estimation were inappropriate for our purpose, we developed a new method that combined the traditional way of counting individual cells with the line intercept method adopted for quantification of filamentous microorganisms (Komárková and Nedoma in preparation). Cumulative lengths of filaments in a sample were calculated from the number of intercepts between filaments and a test bar of known length. The line intercept method had originally been used for measuring root length (Newman 1966) and was adapted for measuring filamentous bacteria isolated on polycarbonate filters in epifluorescence (Nedoma et al. 2001).

264

J. Vymazal et al.

The product of the cumulative length and filament diameter reveals total biovolume of filaments. The method is reliable and statistically correct for quantification of long or dense filaments that cannot be measured individually. The cumulative length of filaments (CLF) in a sample can be calculated as follows CLF =

(π / 2) × ( N / T ) × ( AF ) × 10 −6 , VF

where CLF [µm mL−1] is the cumulative length of filaments, N/T (µm−1) is the ratio of the number of intercepts and the cumulative length of test bars, AF (µm2) is the effective filtration area of the filter examined, and VF (ml) is the volume of the water sample filtered (Nedoma et al. 2001). Random orientation and distribution of the measured filaments over the filter area is considered. Nedoma et al. (2001) found statistically highly significant agreement between the total lengths of filaments estimated by summing the lengths of individual filaments and by line intercept method using a model experiment with strings. The procedure used for counting was as follows: 1. Before counting, the sample was checked for large unicellular algal species, as those species must be counted separately using smaller magnification to avoid erratic results. 2. Filaments – at least 300 intersections of the dominant filament were measured with the grid. If the dominance was not clear, at least 500 intersections of all filaments were counted. 3. Particles – at least 400 (360–440) counts (cells) of a dominant species or at least 500 counts of all species were counted if dominance was not clear. Confidence limit p = 0.95, accuracy ∼=10%, (Wetzel and Likens 1990). The program’s output data are given in mg cm−2. Data on the biomass of algae and Cyanobacteria are expressed as fresh mass (FM), even if the original measurements were obtained as a biovolume (µm3 cm−2). Assuming that specific mass value of algae is 1, then 109 µm3 = 1 mg. To compare the data from plexislides with natural substrata, the counts from each sample were converted into an abundance scale used for natural periphyton.

10.4.2

Identification of Diatoms and Cyanoprokaryota (Blue-Green Algae)

Identification of Cyanoprokaryota (blue-greens) was made using the following literature: (Gardner 1927; Anagnostidis and Komárek 1988; Komárek 1989; Komárek and Anagnostidis 1998). For identification of Bacillariophyta (diatoms), monographs of Patrick and Reimer (1966) and Krammer and Lange-Bertalot (1988, 1991a,b) were used.

10 Algal Responses to Long-Term Nutrient Additions

10.4.3

265

Calculation Program for Biovolume Evaluation

The calculation program for biovolume evaluation distinguishes the following life forms: filaments with diameter < 3 µm, filaments with diameter > 3 µm, cells < 5 µm, cells > 5 µm, loose colonies, and dense colonies. Equations used to calculate biovolume for various cell shapes are listed in Table 10.2. For the purpose of the program, 11 taxonomic groups were selected (Table 10.3) with more detailed separation within Cyanoprokaryota.

Table 10.2 Equation used for biovolume estimation No. Shape description Equation

Example

1 Sphere 0.5236d3 Merismopedia glauca Gomphosphaeria semen vitis 2 Ellipsoid (rugby) 0.5236a2b Aphanothece comasii 3 Ellipsoid (UFO) 0.5236ab2 Gyrosigma sp. 4 Right prism (tall) a 2b Achnanthes linearis 5 Right prism (low) ab2 Cyclotella sp. 6 Cylinder (coin) 0.7854d2a Stichococcus sp. 7 Cylinder (pencil) 0.7854d2b Staurastrum , (branches) 8 Cone 0.2618a2b Closterium limneticum 9 Double cone 0.2618a22b Fragilaria tenera 10 Cone and prism projection 0.6309a2b Cocconeis placentula 11 Ellipsoid and prism projection 0.7618a2b (not found in this study) 12 Ellipsoid and cone projection 0.3972a2b Geitlerinema unigranulatum 13 Filament (volume per 1 µm) 0.7854d2 a = shorter dimension (width), b = longer dimension (length), d = diameter

Table 10.3 Taxonomic groups found in samples; number of species found at sampling points along the phosphorus gradient Group Common name No. of species Bacillariophyceae Diatoms Cyanoprokaryota – Oscillatoriales Filamentous blue-green algae Cyanoprokaryota – Chroococcales Coccal blue-green algae Conjugatophyceae Desmidsa Chlorophyceae Green algae (“true”) Cyanoprokaryota – Nostocales Filamentous N-fixing blue-green algae Cyanoprokaryota – Scytonematales Branched N-fixing blue-green algae Total Cyanoprokaryota Dinophyceae Dinoflagellates Xanthophyceae Yellow-green algae Chrysophyceae Golden algae Euglenophyceae Euglenoids Total a Includes species of the families Zygnemataceae and Desmidiaceae

59 51 36 25 22 6 3 96 2 1 1 1 207

266

10.5 10.5.1

J. Vymazal et al.

Results Taxonomic Groups

During the study, a total of 207 species were found on plexislides along the phosphorus gradient (Table 10.3). Among this total, there were 96 species of Cyanoprokaryota (blue-green algae, cyanobacteria), the most represented group. Within the Cyanoprokaryota, the order Oscillatoriales had the highest number of species (51). Species belonging to the order of Chroococcales were also abundant (36). Other orders of Cyanoprokaryota – Nostocales and Scytonematales – were less abundant, with six and three species, respectively. Other abundant groups were Bacillariophyceae (diatoms), with 59 species recorded; Chlorophyceae (22 species); and Conjugatophyceae (25 species). Species of other groups (Dinophyceae, Xanthophyceae, Euglenophyceae, and Chrysophyceae) were present only rarely (one or two species).

10.5.2

Total Fresh Biomass

The highest average fresh biomass was recorded at site C2 (32.57 mg cm−2 = 32.57×109 µm3 cm−2). This value was an order of magnitude higher than those recorded at other sampling sites, where fresh biomass values varied between 2.26 mg cm−2 at site C4 and 7.08 mg cm−2 at site C5 (Fig. 10.1a). The fresh biomass value at site C2 was significantly higher (p < 0.05) than values at all other sites. The high values were mostly caused by the abundance of the filamentous species Lyngbya (Cyanoprokaryota), Spirogyra, Mougeotia (Conjugatophyceae), and Oedogonium (Chlorophyceae). The high periphyton biomass at site C2 is in agreement with the high macrophyte biomass recorded at this site by Richardson and Vymazal, who found that macrophyte biomass at site C2 was the highest among all six sampling sites along the C-transect (Chap. 6, Fig. 6.30).

10.5.3

Fresh Biomass by Life Forms

The major part of the fresh biomass is formed by filaments with a diameter > 3 µm and cells > 5 µm (Fig. 10.1b). The percentage of total biomass made up of species of these types varied between 90.7% at site C4 and 96.5% at site C2, with an average value of 94.3%. However, data in Fig. 10.1b also indicate a clear shift between filaments and cells along the C-transect. At sites C1 and C2, the major part of the total biomass is formed by filaments with diameter > 3 µm; at sites C3, C4, and C5, the portion of filaments with diameter > 3 µm is approximately the same as that

10 Algal Responses to Long-Term Nutrient Additions

267

Fig. 10.1 Fresh biomass of periphyton growing on plexislides at sites C1 to C6. (a) Total biomass, (b) individual life forms expressed as percentage of total fresh mass, and (c) individual taxonomic groups expressed as percentage of total fresh biomass

268

J. Vymazal et al.

Fig. 10.1 (continued)

Table 10.4 Major representatives of selected life forms Life form Filaments with diameter < 3 µm Leptolyngbya spp.(CY-O), Phormidium spp.(CY-O), Pseudanabaena spp. (CY-O), Spirulina sp. (CY-O) Filaments with diameter > 3 µm Lyngbya spp. (CY-O), Phormidium spp. (CY-O), Spirogyra spp.(CON), Mougeotia spp.(CON), Oedogonium spp.(CHL) Cells < 5 µm Aphanothece spp. (from broken colonies) (CY-C) Cells > 5 µm Mastogloia smithii (BA), Nitzschia linearis (BA), Rhopalodia gibba (BA), Synedra ulna (BA), Navicula rhynchocephala (BA), Cymbella spp. (BA), Cosmarium spp. (CON), Staurastrum spp. (CON) Dense colonies Pediastrum duplex (CHL), Aphanothece comasii (CY-C), A. variabilis (CY-C) BA Bacillariophyceae; CHL Chlorophyceae; CON Conjugatophyceae; CY-C CyanoprokaryotaChroococcales; CY-O Cyanoprokaryota – Oscillatoriales

formed by cells > 5 µm; and at site C6, the major part of biomass is formed by cells > 5 µm. Other life forms (dense colonies, cells < 5 µm and filaments with diameter < 3 µm) were present only in small quantities. Major representatives of monitored life forms (?) are presented in Table 10.4.

10 Algal Responses to Long-Term Nutrient Additions

10.5.4

269

Fresh Biomass by Taxonomic Group

The fresh biomass of individual taxonomic groups expressed as percentage of total fresh biomass is given in Fig. 10.1c. The data indicate that species of the order Oscillatoriales made up the major portion of the total biomass at sites C1 (61.6%) and C2 (70.0%), while at less eutrophic sites C5 and C6, diatoms (Bacillariophyceae) formed the major portion of the biomass (43.1% and 71%, respectively). The portion of the biomass formed by species belonging to the class Chlorophyceae varied considerably less along the transect (from 5% at site C2 to 24.1% at site C3). Species belonging to the class Conjugatophyceae varied in biomass along the transect from 1.3% at site C6 to 36.5% at site C5, but with no clear pattern. Species of Cyanoprokaryotic orders, Nostocales and Chroococcales, formed only a small portion of the total biomass. Each group’s individual response is presented in the following sections.

10.5.5

Chlorophyceae

Biomass varied greatly from 0.11 mg cm−2 at site C4 to 1.49 mg cm−2 at site C3. However, the differences were not statistically significant. The percentage of total biomass that was formed by Chlorophyceae did not fluctuate as much as compared to fresh biomass values, varying between 5% at site C2 and 24.1% at site C3 with no statistical difference among sites (Fig. 10.1c). The major representatives of Chlorophyceae were filamentous species of the genus Oedogonium.

10.5.6

Conjugatophyceae

The highest fresh biomass of Conjugatophyceae was recorded at site C2 (11.9 mg cm−2), while the lowest biomass was recorded at site C6 (0.08 mg cm−2). Extremely high fluctuation of biomass at this site resulted in a finding of no significant difference in biomass among sites for this group. However, the percentage of total biomass formed by species belonging to the class Conjugatophyceae showed a completely different pattern as compared to biomass values (Fig. 10.1c). The highest percentage of Conjugatophyceae biomass was recorded at site C4 (36.5%), which was significantly different from values recorded at sites C3 (9.4%, p < 0.05) and C6 (1.3%, p < 0.01). The major representatives of the class Conjugatophyceae were filamentous species of the family Zygnemataceae (Mougeotia and Spirogyra) and species of the family Desmidiaceae (e.g., Cosmarium, Staurastrum, Pleurotaenium).

270

10.5.7

J. Vymazal et al.

Bacillariophyceae

The biomass of Bacillariophyceae showed a clear pattern and increased from 0.15 mg cm−2 at site C1 to 5.0 mg cm−2 at site C6. Biomass at site C6 was significantly higher than those at sites C1, C2, and C4 (p < 0.005) and C3 (p < 0.01). Percent of total biomass formed by diatom species followed the same pattern, with the lowest values at site C2 (2.6%) and the highest value at site C6 (71%) (Fig. 10.1c). The percentage of the total biomass at site C6 was significantly higher than those at sites C1, C2 (both p < 0.0001), C3, C4 (both p < 0.001), and C5 (p < 0.05). Site C2 was also significantly different from sites C3 (p < 0.05) and C5 (p < 0.005). In addition, site C1 was different from site C5 (p < 0.05). The major representatives of the class Bacillariophyceae were species Achnanthes sp., Cymbella spp., Fragilaria spp., Mastogloia smithii, Navicula lanceolata, Navicula rhynchocephala, Nitzschia linearis, Pinnularia sp. (spp.?), Rhopalodia gibba, and Synedra ulna.

10.5.8

Cyanoprokaryota

This division was divided into four orders: Oscillatoriales, Chroococcales, Nostocales, and Scytonematales. The species of the order Scytonematales occurred only on four slides at sites C4 and C5 and, therefore, this order will not be discussed in detail here. 10.5.8.1

Oscillatoriales

There is a clear decline in Oscillatoriales biomass along the transect. The highest biomass was recorded at site C2 (18.48 mg cm−2), and this value was significantly higher than those recorded at sites C1 (p < 0.01) and C3, C4, C5, and C6 (p < 0.005). Also, the percentage of the biomass formed by species of the order Oscillatoriales declined along the transect, with values at site C1 (61.6%) and C2 (70%) being significantly higher than those recorded at C3 (p < 0.05 and p < 0.01, respectively), C4, C5 (both p < 0.005 and p < 0.0005, respectively), and C6 (p < 0.0005). The lowest value recorded at site C6 (12%) was also significantly lower than value recorded at site C3 (37.6%). The major representatives of the order Oscillatoriales were species of the genera Lyngbya, Leptolyngbya, Phormidium, Pseudanabaena, and Spirulina. 10.5.8.2

Chroococcales

Despite the fact that there were 37 species of Chroococcales recorded on slides, the fresh biomass formed by this group was very small and varied only between 0.087 mg cm−2 at site C4 and 0.0018 mg cm−2 at site C3 (Fig. 10.1c). The highest proportion of the total biomass formed by species of the order Chroococcales was

10 Algal Responses to Long-Term Nutrient Additions

271

recorded at site C4 and amounted to only 4.9%. Values at other sites ranged between 0.03% at site C3 and 1.2% at site C1. There was no statistical difference in either biomass or percentage of the total biomass. The major representatives of the order Chroococcales were species of the genera Aphanocapsa, Aphanothece, Gloeothece, Chroococcus, and Merismopedia.

10.5.8.3

Nostocales

Species of the order Nostocales accounted for a small portion of the fresh biomass, varying between 0.099 mg cm−2 at site C2 and 0.006 mg cm−2 at site C1. The biomass at site C2 was significantly higher than those at sites C1 and C6 (both p < 0.01) and C3 and C4 (both p < 0.05) (Fig. 10.1c). The proportion of the total biomass formed by species of the order Nostocales increased from site C1 to C5 and dropped at site C6. However, the contribution of this group to the total biomass was very low, and it did not exceed 1.5% of the total biomass. The major representatives of the order Nostocales were species of the genera Anabaena, Cylindrospermum, Nodularia, Nostoc, and Trichormus.

10.5.9

Cyanoprokaryota – Total

The total biomass formed by species of Cyanoprokaryota (blue-green algae, cyanobacteria) ranged between 18.61 mg cm−2 at site C2 and 0.49 mg cm−2 at site C6. The proportion of the total biomass formed by species of Cyanoprokaryota was greatly influenced by the biomass of the order Oscillatoriales, which formed the major portion of the blue-green algal biomass (Fig. 10.1c). The highest proportion of the biomass was recorded at sites C1 and C2 (63.3% and 70.8%, respectively), while the lowest value was recorded at site C6 (12.7%). The significant differences also followed those found for the order Oscillatoriales. Cyanoprokaryota, together with diatoms (Bacillariophyceae), accounted for the major portion of the biomass at all sampling sites along the C-transect: C1 – 74.8%, C2 – 73.4%, C3 – 66.4%, C4 – 53.4%, C5 – 71.1%, and C6 – 83.7%.

10.6

Discussion

The results of this study show a clear shift in species composition along the nutrient gradient found along the C-transect. At the most eutrophic sites C1 and C2, biomass is dominated mostly by blue-green algae (most important genus Lyngbya), while diatoms dominate the biomass at the least eutrophic sites, C5 and C6. (The most important species are M. smithii (Fig. 10.2) and N. linearis.) Chlorophyceae and Conjugatophyceae, which also represent a significant portion of the biomass,

272

J. Vymazal et al.

Fig. 10.2 Percentage of total fresh mass (biovolume) of four species of Lyngbya and Mastogloia smithii along the C-transect in WCA-2A of the northern Everglades

do not exhibit any change pattern along the transect. Other groups were present in small percentages and did not contribute substantially to the periphyton biomass. The shift in species composition also correspondingly shifted the dominant life forms along the transect. At eutrophic sites C1 and C2, filamentous organisms prevailed (blue-green algae, green algae), while single-celled organisms (diatoms) dominated the noneutrophic sites C5 and C6 (Fig. 10.1b). Sampling sites C3 and C4 form a transition between eutrophic sites C1 and C2 and oligotrophic sites C5 and C6, but their character bears a greater similarity to unimpacted sites C5 and C6 (Vymazal et al. 2001a). The present study indicates that oligotrophic, low-nutrient sites are dominated by diatoms. This is in agreement with previous studies from the Everglades (e.g., Swift and Nicholas 1987; McCormick and O’Dell 1996). McCormick and O’Dell (1996) reported that in interior WCA-2A sites with water column TP concentrations between 5 and 7 µg L−1, diatoms formed 11–49% of the total biovolume in natural periphyton. In our study, diatoms formed 25.4–71% of the total biovolume (fresh biomass) at sites C4, C5, and C6 (Fig. 10.1c), with water column TP concentrations between 4.4 and 6.2 µg L−1. The higher percentage of diatoms found in our study may be influenced by the fact that biovolume was measured in periphyton growing on plexislides where considerably more diatom species were found as compared to periphyton growing on natural substrata (Vymazal et al. 2001b). In general, the plexislide favors diatom growth compared to natural substrata.

10 Algal Responses to Long-Term Nutrient Additions

273

In contrast to a study by McCormick and O’Dell (1996), who found 27–70% of the biovolume at oligotrophic sites formed by blue-green algae Scytonema hofmanii and Schizothrix calcicola, our study found only 12–21.9% of the total biovolume (fresh biomass) at oligotrophic sites C4, C5, and C6 to be blue-green algae. However, Scytonema spp. was abundant only in periphyton growing on natural substrata and was not found on Plexiglas (Vymazal et al. 2001b). In our study, S. calcicola was not found despite the fact that this species is considered one of the most frequent blue-green algae in the oligotrophic parts of the Everglades (Gleason and Spackman 1974; Swift and Nicholas 1987; McCormick and O’Dell 1996). However, it is important to realize that these authors consider S. calcicola sensu Drouet (1968, 1981). In Drouet’s classification scheme, several hundred species of several tens of genera were put under S. calcicola, giving this taxonomical approach little or no ecological value. In our study, the blue-green algae at sites C4, C5, and C6 were mostly represented by various species of Leptolyngbya and Phormidium, which would be classified as S. calcicola in Drouet’s system. The eutrophic sites C1 and C2 were dominated by wide-filamentous Cyanoprokaryota, namely four species of Lyngbya. These species were also abundant at site C3, but then the relative proportion of the total periphyton biomass represented by these species sharply decreased (Fig. 10.2). McCormick et al. (1998) reported that eutrophic sites at WCA-2A (TP concentrations of 44–103 µg L−1) had a higher proportion of green algae (e.g., Spirogyra sp. (spp?) ), and McCormick and O’Dell (1996) reported that green algae assemblages dominated by Spirogyra sp. formed up to 100% of the biovolume at sites with total P concentrations between about 10 and 30 µg L−1. In our study, we did not observe the dominance of green algae at eutrophic sites. Green algae (dominated by Oedogonium spp.) and Conjugatophyceae (dominated by Zygnematalean taxa Spirogyra spp., Mougeotia spp., and various desmids) were present along the whole transect with their maximum of 47% of the total biomass (biovolume) being found at site C4. However, the periphyton growing on natural substrata provided a very high abundance of Oedogonium, Spirogyra, and Mougeotia but never reached 100% biovolume or biomass (Vymazal et al. 2000, 2001b). Plexiglas (like plastic materials in general) is a less-selective material than glass and it is colonized earlier by higher succession communities. Although Plexiglas is more suitable as far as selectivity is concerned, it does not match the natural substrata and some extent of selectivity is always present. The comparison with periphyton species composition growing on natural substrate (Vymazal et al. 2000, 2001b) indicates clearly that Plexiglas overestimates the role of diatoms and also underestimates the presence of the genus Scytonema especially in monitoring (e.g., 1 or 2 months). A Plexiglas substratum also results in many green filamentous algal species being found along the whole transect, while in samples of natural periphyton a similar amount of green filamentous algae was present only at sites C1, C2, and C3, but at sites C4, C5, and C6 the number of green algae was much lower.

274

10.7

J. Vymazal et al.

Conclusions and Lessons for Restoration

A total of 207 algal species were found along the C-transect nutrient gradient. Species distributions were 96 Cyanoprokaryota (cyanobacteria, blue-green algae), 59 Bacillariophyceae (diatoms), 25 Conjugatophyceae, and 22 Chlorophyceae. The highest value of fresh biomass was recorded at site C2. This is in accordance with studies on macrophytes, which found the highest macrophyte productivity at this site (Chap. 6). The analysis of algal life forms revealed a clear shift in dominance from filamentous forms with filament diameter > 3 µm at sites C1 and C2 to single cells > 5 µm at site C6. The fresh mass of filaments with diameter > 3 µm formed 69.9% and 90.2% of the total fresh mass at sites C1 and C2, respectively, whereas at site C6 it formed only 20.4%. In contrast, cells > 5 µm formed 23.9%, 6.3%, and 75.4% at sites C1, C2, and C6, respectively. The major representatives of filaments > 3 µm were species of the Cyanoprokaryotic genera Lyngbya and Phormidium, genera belonging to the class Zygnematales (Conjugatophyceae) – Spirogyra and Mougeotia and genus Oedogonium (Chlorophyceae). The major representatives of cells > 5 µm were diatoms (e.g., M. smithii, N. linearis, R. gibba, S. ulna, Cymbella spp., N. rhynchocephala, Pinnularia sp., Fragilaria ulna) and desmids (e.g., Cosmarium spp. or Staurastrum spp.). Other life forms (filaments with diameter < 3 µm, cells < 5 µm, dense colonies) formed only 3.5–9.3% of the total fresh mass. Species belonging to 11 taxonomic groups were identified. There was a clear shift in periphyton species composition along the C-transect. At sites C1 and C2, the major portion of the total biomass was formed by species of the order Oscillatoriales (61.6% and 70.0%, respectively), while at the less eutrophic sites C5 and C6, diatoms (Bacillariophyceae) formed the major portion of the biomass (43.1% and 71.0%, respectively). The portion of the biomass formed by species belonging to the class Chlorophyceae varied considerably less along the transect (from 5% at site C2 to 24.1% at site C3), while species belonging to the class Conjugatophyceae varied in biomass along the transect (from 1.3% at site C6 to 36.5% at site C5) with no clear pattern. Species of Cyanoprokaryotic orders, Nostocales and Chroococcales, formed only a small portion of the total biomass. The results showed that the more eutrophic sites C1 and C2 are very similar. Typical species found at those sites were filamentous Cyanoprokaryota (cyanobacteria, blue-green algae) Lyngbya spp., Conjugatophyceae Spirogyra spp., and Mougeotia spp. Unimpacted sites C5 and C6 and most sites at C3 and C4 form a group substantially different from eutrophic sites C1 and C2. At sites C5 and C6, a large portion of the periphyton biomass is typically formed by diatoms (dominated by M. smithii), narrow filamentous blue-green algae (Leptolyngbya spp.), and green algae Oedogonium spp. Sites C3 and C4 are very heterogeneous and form a transition between the two distinct groups, but their character is more similar to sites C5 and C6. The use of Plexiglas slides to characterize algal communities in the Everglades presents an overestimate of the role of diatoms and also underestimates the presence of the genus Scytonema especially in short-term monitoring (e.g., 1 or 2 months).

10 Algal Responses to Long-Term Nutrient Additions

275

A Plexiglas substratum also results in many green filamentous algal species being found along a nutrient gradient in the Everglades as compared to analysis of natural substrates. Thus, the type of substrate must be taken into account when comparing algal communities and the use of biovolume is needed to assess the role of algal groups in community structure.

11

Macroinvertebrate Responses to a Gradient of Long-Term Nutrient Additions, Altered Hydroperiod, and Fire Ryan S. King and Curtis J. Richardson

11.1

Introduction

The Everglades has been the sentinel ecosystem for illustrating the deleterious effects of agricultural land practices, fire suppression, and hydrological alterations on freshwater wetlands. Numerous studies have illustrated that Everglades biota are adapted for survival under highly oligotrophic conditions (e.g., Browder 1982; Steward and Ornes 1975a,b; Swift and Nicholas 1987; Richardson et al. 1999) and are strongly P limited (reviewed by Noe et al. 2002). Moreover, the natural Everglades ecosystem has evolved under dynamic hydrological conditions, with strong annual wet–dry cycles that are critically coupled with large, periodic fires (e.g., Davis 1994). Thus, it is not surprising that anthropogenic modifications to the natural nutrient, hydrological, and fire regimes of the Everglades during the past few decades have had remarkable effects on biota across all levels of ecological organization (Davis and Ogden 1994). Although a variety of other human influences have been indicated as stressors to the Everglades, P-enriched runoff from the Everglades Agricultural Area (EAA) has been targeted as the chief offender (SFWMD 1992; Davis and Ogden 1994). The extensive canal and levee system that has compartmentalized the remnant Everglades has served as a conduit for P from the EAA and Lake Okeechobee, and water-control structures have been point sources of P to diked portions of the fen (SFWMD 1992, 2003, 2004, 2005, 2006). In areas near water-control structures, P is primarily responsible for the transformation of the natural pattern of Cladium jamaicense Crantz (sawgrass) stands and open-water sloughs to dense stands of invasive Typha domingensis Pers. (cattail) and other invasive vegetation (Davis 1991; Urban et al. 1993; Newman et al. 1998; Richardson et al. 1999). Phosphorus inputs have also had profound effects on other Everglades biota, including microbes (Grimshaw et al. 1997; Qualls and Richardson 2000), periphyton (Vymazal et al. 1994; McCormick and O’Dell 1996; McCormick et al. 1996; Pan et al. 2000), invertebrates (Rader and Richardson 1994; King and Richardson 2002, 2003), vegetation (Richardson et al. 1999; King et al. 2004), and fish (Jordan 1996; Turner et al. 1999). The question is not whether the Everglades is changed by nutrient additions and hydrologic shifts but rather if current indices 277

278

R.S. King and C.J. Richardson

or metrics of biotic response can provide an early detection system, especially for macroinvertebrates, which have been used as a fundamental index of ecosystem degradation and habitat loss in aquatic ecosystems (e.g., Rosenberg and Resh 1993; Karr and Chu 1997). Bioassessment using macroinvertebrates as ecosystem indicators has become a widely accepted technique for monitoring water quality and ecological health of aquatic systems (Rosenberg and Resh 1993). Attributes of macroinvertebrate assemblages provide considerable information regarding levels and sources of impairment imposed by human influence (e.g., Karr and Chu 1997). Bioassessment is especially effective in lotic systems and is used to monitor environmental quality in streams throughout the world (e.g., Reynoldson et al. 1995; Zamora-Muñoz and Alba-Tercedor 1996; Bailey et al. 1998; Barbour et al. 1999; Moss et al. 1999; Smith et al. 1999). Until recently, however, the use of biota to assess ecological condition of lentic habitats like wetlands had not received much attention (US EPA 1997a). In the USA, several states (e.g., Apfelbeck 1999; Gernes and Helgen 1999) along with the US Environmental Protection Agency (Danielson 1998) have recognized the need for biologically grounded wetland assessment methods. Most wetland assessment techniques in use today are based on functional indicators that do not explicitly measure biological condition (e.g., Brinson and Rheinhardt 1996) despite the mandate of Section 101(a) of the Clean Water Act to restore and maintain the chemical, physical, and biological integrity of the USA’s waters, which include wetlands. Such inconsistency with federal legislation has led to dissatisfaction with current wetland assessment methods (Kusler and Niering 1998) and a call for the development of methods that incorporate biological components, like macroinvertebrate assemblages, into assessment protocols (US EPA 1997a; King et al. 2000). Although interest in wetland bioassessment is currently high, no accepted assessment protocols for wetlands have been developed and published like those that exist for streams (e.g., Barbour et al. 1999). Wetlands are, however, structurally and functionally very different from streams (e.g., Richardson 1999). Even the definition of what constitutes a wetland is a source of confusion and contention (e.g., Cowardin et al. 1979; USACE 1987). While there are exceptions, typical stream and wetland habitats differ markedly in permanence of surface water (predominantly permanent in streams vs. seasonal/semipermanent in wetlands), hydrologic gradient (high in streams vs. low-to-none in wetlands), sources of energy (mostly allochthonous in streams vs. autochthonous in wetlands), habitat structure (riffle/ pool segments in streams vs. vegetated/unvegetated patches in wetlands), and water chemistry dynamics (comparatively stable water temperatures and dissolved oxygen (DO) in streams, dynamic fluxes in temperatures, and DO in wetlands) – many other differences can be added to this list. Thus, it is intuitive that structure and function of wetland invertebrate communities would differ from streams accordingly (Sharitz and Batzer 1999). Thus many of the models used to describe invertebrate community dynamics and bioassessments in streams have limited applicability to wetlands or, at least, require significant reevaluation before they can be used in the Everglades.

11 Macroinvertebrate Responses to a Gradient of Long-Term Nutrient Additions

279

In addition to poorly known sensitivities to anthropogenic stressors (Batzer and Wissinger 1996) and relatively few established metrics of human influence (Lemly and King 2000), wetland macroinvertebrate assemblages present difficulties in sampling and sample processing that are less prevalent in lotic bioassessments. However, detailed studies of the methods used to evaluate and assess macroinvertebrate communities and develop bioassessment sampling criteria in the Everglades are presented in King and Richardson (2002, 2003). In this chapter, we highlight some of these key findings and synthesize macroinvertebrate responses to nutrient additions and other environmental variables such as hydrology and fire. The three primary objectives of this phase of the research were to (1) evaluate the utility of wetland macroinvertebrate assemblages as an indicator group for bioassessment in the Everglades, with emphasis on the implications of differing laboratory methods of sample processing and levels taxonomic identification; (2) quantify the response of macroinvertebrate community biomass and species richness to P enrichment; and (3) identify major dimensions of community structure and the primary environmental factors controlling community organization. To address these objectives, two conceptual frameworks were used to guide the design of experiments and testing of hypotheses.

11.1.1

Subsidy–Stress Model

Odum et al. (1979) developed a conceptual model to describe ecological responses to system inputs (Fig. 11.1). These inputs may be usable (e.g., nutrients or energy) or acutely toxic (e.g., herbicides). Those inputs that are usable are initially hypothesized to result in a subsidy effect – a deviation above the system’s normal operating range – while those that are acutely toxic, a stress effect. However, increasing concentrations or levels of usable inputs may eventually result in a decrease in system performance. Termed the “subsidy–stress gradient,” this conceptual model may be usefully applied to predict community or ecosystem-level responses to P inputs in the Everglades since P is limiting and, therefore, represents a usable system input. How might inputs of P affect wetland invertebrate assemblages in an unproductive, naturally dynamic environment? Most taxa documented in the Everglades (Rader and Richardson 1992, 1994) and most wetland ecosystems are adapted for harsh, often temporary conditions, hence have limits of tolerance (sensu Shelford 1913) that may far exceed stress presented by changes such as depressed dissolved oxygen due to eutrophication. Intuitively, stimulation of primary producers (i.e., food resources) by P inputs would initially result in a subsidy effect for invertebrate assemblages on a community level (e.g., biomass, species richness). However, this response may not hold true at high concentrations of P if such inputs result in the expansion of dense, invasive vegetation and a reduction in quantity of high-energy food resources such as periphyton. Here, a stress response may be expected. At population (species) levels, however, P could result in a subsidy, stress, or subsidy– stress effect depending on a variety of factors such as niche breadth and opportunistic

280

R.S. King and C.J. Richardson

Fig. 11.1 Conceptual diagram of the subsidy–stress model illustrating hypothetical P–response relationship for invertebrate assemblages in the Everglades (modified from Odum et al. 1979)

nature of each taxon (e.g., specialist vs. generalist; Mihuc 1997). Thus, the subsidy–stress model is specific enough to be useful, but general enough to be inclusive of different types of perturbations that depend upon level of organization (Odum et al. 1979).

11.1.2

Hierarchy Theory

Hierarchy theory is a broad theory about the relationships between ecological processes and spatial and temporal scales and patterns observed across landscapes (Allen and Star 1982; O’Neill et al. 1986). More specifically, it is a conceptual framework that describes the ecological coupling of pattern at multiple scales – how pattern at lower-level scales can interact to give rise to pattern at higher levels. In a hierarchical system, lower-level units (e.g., patches of vegetation) can be thought of as small and relatively fast moving entities through time and space, while higher-level patterns are larger and slower (Urban et al. 1987). Lower-level units integrate to generate higher-level pattern, but higher-level pattern controls those at lower levels.

11 Macroinvertebrate Responses to a Gradient of Long-Term Nutrient Additions

281

In the Everglades, vegetation pattern in low-P, unimpacted areas has been described as a mosaic that varies markedly at fine scales but relatively little at coarse scales across the landscape (Jordan et al. 1997; King et al. 2004). This mosaic pattern undoubtedly plays a significant role in ecological processes and vice versa (Watt 1947). For example, fine-scale heterogeneity may be important for seasonal (Jordan 1996) or even diel (King and Wrubleski 1998) movements of invertebrates, while coarse patterns may influence dispersal across the landscape (e.g., Gilpin and Hanski 1991; Delettre and Morvan 2000; Palmer et al. 2000). Moreover, fine-scale heterogeneity may sustain unique, local assemblages and, consequently, increased species diversity (e.g., MacArthur and Wilson 1967). Therefore, alteration to this characteristic mosaic due to P inputs could be a substantial perturbation to invertebrate assemblages across all levels of the spatial hierarchy. Only a hierarchical perspective (sensu Urban et al. 1987) could reveal all of the possible implications of P enrichment to invertebrate assemblages across this large wetland landscape.

11.2 11.2.1

Methods Study Area and Sampling Design

Sampling was conducted in Water Conservation Area 2A (WCA-2A) in the northern Everglades (Fig. 11.2). A detailed description of the study area is presented in Chaps. 5 and 9, and King et al. (2004). The spatial component of this study used the sampling design described by King et al. (2004). All vegetation plots (n = 126 plots, n = 14 plot-clusters) were included in this component of the study. Data collection for the spatial study was conducted on 20–29 October 1998. In addition to the October 1998 collection, a temporal study was conducted as well, but space does not allow presentation of those results in this chapter (but see King 2001). In the temporal study, plots within 3 of the 14 clusters (one cluster in each of the Pimpacted, transition, and reference zones, respectively) were sampled during February 1999 (low water, dry season), July 1999 (immediately after reflooding following an extensive period of no surface water), and October 1999 (deep water, wet season, 1 year after first collection).

11.2.2

Abiotic Variables

Spatial, soil/sediment chemical, hydrological, and fire frequency variables were considered to be potentially important dimensions of vegetation and invertebrate assemblage organization along the P gradient (King et al. 2004). In October 1998, values of 14 spatio-environmental variables were estimated from each of 126 plots

282

R.S. King and C.J. Richardson

Fig. 11.2 Map of south Florida showing the location of Water Conservation Area 2A (WCA-2A); impacted, transition, and reference landscape zones; locations of S-10 water-control structures; centroids of sampling clusters; and plot-cluster sampling design. EAA Everglades Agricultural Area

in the spatial study (see Table 11.1). Greater details on the rationale and methods of measurement for all of these variables are provided in King et al. (2004). Of the abiotic variables, hydrology was the only one expected to change markedly over time (soil chemistry along the P gradient has remained similar over the past decade) (see Chap. 6; no fires occurred at the three temporal clusters during the study). Water depth at each plot within each cluster was estimated using a hydrological model developed by Romanowicz and Richardson (1997) (see Chap. 7). Depths (cm) were estimated daily through the end of the study period in October 1999 (King 2001).

11.2.3

Biotic Sampling

Vegetation species composition and cover was estimated at each plot in the spatial and temporal studies. Cover for each species was recorded using Braun-Blanquet

11 Macroinvertebrate Responses to a Gradient of Long-Term Nutrient Additions

283

Table 11.1 Mean (±1 SD) environmental characteristics of impacted, transition, and unimpacted landscape zones in WCA-2A of the Everglades (from King et al. 2004) Variable

ID

Units

Impacted (n = Transition (n 45) = 45)

Unimpacted (n = 36)

Distance from canal inflow structures Total carbon (soil)

Canal

m

2,495 ± 869

5,541 ± 914

C

g kg−1

435.8 ± 20.3

435.9 ± 27.2

Total calcium (soil) Total potassium (soil) Total magnesium (soil) Total sodium (soil)

Ca K Mg Na

g kg−1 g kg−1 g kg−1 mg kg−1

37.1 ± 1.7 0.6 ± 0.2 3.7 ± 0.8 3,058 ± 160

42.8 ± 2.1 0.6 ± 0.4 3.9 ± 0.9 2,900 ± 173

Total nitrogen (soil) N Total phosphorus (soil) P Water depth (1981–1998) xDepth

29.2 ± 2.2 g kg−1 mg kg−1 1,430 ± 172 cm 29.0 ± 8.7

29.0 ± 3.7 1,203 ± 181 32.3 ± 9.6

Water depth (1 year)

cm

35.7 ± 8.3

41.8 ± 9.6

28.2 ± 0.9

29.7 ± 1.5

9,050 ± 924 428.2 ± 47.7 47.0 ± 3.5 0.5 ± 0.2 3.6 ± 0.9 2,165 ± 113 29.2 ± 4.4 578 ± 151 31.2 ± 11.4 46.4 ± 10.4 33.6 ± 0.8

3.1 ± 3.0

3.1 ± 2.4

6.0 ± 4.5

0.2 ± 0.4

0.4 ± 0.5

0.3 ± 0.5

xDepth1y

Interquartile range, water IQR(Depth) cm depth (1981–1998) Frequency, water depth Freq. <−10 cm % <−10 cm (1981–1998) Fire index (frequency Fire Suma 1981–1998) a

Sum of total number of fires during 1981–1998, weighted as 1/log10(t + 1), where t is the time (years) since fire, for each fire

cover classes (Phillips 1959). Additional details on vegetation sampling are provided in King et al. (2004). Presence and abundance of periphyton was hypothesized to be an important determinant of invertebrate assemblage biomass and composition. Two periphyton abundance metrics were estimated: metaphyton (floating periphyton mats) cover and epiphyton (vegetatively attached periphyton) biomass accumulation. Metaphyton cover (hereafter, metaphyton) was estimated using Braun-Blanquet cover classes. Metaphyton samples were used to estimate molar C:N and C:P ratios of periphyton, important measures of food quality and potential elemental imbalance between consumers and their food (Sterner and Elser 2002). Macroinvertebrate sampling was based on a slight modification of protocols used by the state of Florida (FDEP 1996; FDEP SOP #BA-7) and the US EPA (US EPA 1997b; Barbour et al. 1999) for bioassessment. A D-framed dip net (0.3-m wide, 500-µm mesh) was used to collect ten sweeps of 0.5-m length within each plot (total area 1.5 m2). Because the initial sweep may have dislodged but missed organisms, the sweeping process was repeated rapidly two additional times over the same area (US EPA 1997b; Maxted et al. 2000). Contents of all ten sweeps were composited into a 500-µm mesh sieve bucket, rinsed to remove fine particulates, placed in 4-l heavy-duty storage bags, and put on ice for return to the laboratory. In the laboratory, samples were weighed for wet mass and preserved in 5% (v/v)

284

R.S. King and C.J. Richardson

buffered formalin stained with rose bengal. The method was quite repeatable and employed consistently throughout the study.

11.2.4

Macroinvertebrate Sample Processing: Subsampling and Taxonomic Resolution

Macroinvertebrate sample processing followed techniques recommended by FDEP (1996; FDEP SOP #BA-8) and Barbour et al. (1999). Samples were rinsed and homogenized in a 500-µm mesh sieve and large pieces of coarse particulate organic matter (CPOM) were discarded. Sieve contents were placed in a 20-cm wide × 45cm long subsampling pan, and gently spread evenly throughout. The subsampling pan was gridded with numbered 5 × 5 cm2 cells (36 cells total). Cells were selected for subsampling using a random number table. Individual cell contents were transferred to a petri dish marked with grooves into 1/8 sections. One 1/8-cell subsample of material was removed at a time, placed into a second petri dish, and a small amount of water was added to suspend all contents. Subsequently, invertebrates were picked from the subsample using a stereomicroscope at 10× magnification. The process was repeated until a target area or number of individuals was obtained. We selected three fixed-count (100, 200, and 300 individuals) and two fixed-area (10 and 25%) levels of subsampling for comparison. We chose fixed counts and fixed areas most commonly used in other bioassessment studies. We recognized that evaluations of fixed areas, by themselves, might be of limited utility to biologists because few have agreed on a standard sample size to be used (e.g., Courtemanch 1996; Larsen and Herlihy 1998). However, evaluated in the context of average numbers of individuals per subsample and average proportions of the total sample sorted, these fixed-area subsamples were similar to the fixed-count subsamples and allowed for valid comparisons among approaches. Upon reaching a specified number of individuals or area for a respective subsample level, specimens were placed in a vial containing 70% ethanol. Total area, time required to complete, and number of individuals were noted. Larger subsamples (e.g., 300 individuals) were actually an accumulation of specimens stored in several vials, each representing a previous stopping point for other subsamples. We implemented a supplementary large rare (LR) search as defined by Courtemanch (1996) once 300 individuals and 25% of the total sample were subsampled. However, rather than pick all LR taxa from a sample before subsampling, as recommended by Courtemanch (1996), we picked remaining LR taxa after all subsampling was completed because to remove them prior to subsampling would have altered the composition of subsamples and prevented a valid assessment of the use of the LR search as a supplementary procedure. We defined a priori all LR taxa so that individuals included as part of a larger subsample (e.g., 25%) could be added into the pool of LR individuals for smaller subsamples that included the LR search (e.g., 100 + LR). For example, a 100 + LR subsample might only represent

11 Macroinvertebrate Responses to a Gradient of Long-Term Nutrient Additions

285

5% of the total sample area for the fixed-count component. Subsequently, some LR taxa could be contained in the following subsamples of 200, 10%, 300, and 25% and would need to be counted in the final tally of additional LR organisms to be added to the 100 + LR subsample. Thus, any LR taxa in the 200, 10%, 300, and 25% subsamples would have to be added to the remaining LR search for the 100 + LR subsample to be accurate and valid. We classified large mollusks, hemipterans, hirudineans, coleopterans, decapods, all anisopteran odonates, and a few miscellaneous large taxa as LR taxa. We calculated densities for LR taxa on the basis of the total number of individuals per sample, not the fractional area of individual subsamples in which LR taxa were supplemented. We assembled macroinvertebrate data sets using the five basic levels of subsampling (100, 200, 300, 10%, and 25%), an integrated subsample requiring a minimum fixed count and fixed area in the same subsample (100 and 10%), and a fixed-count (100) and fixed-area (10%) subsample supplemented with the LR search. Data sets also were assembled using three levels of taxonomic resolution (family, genus, and species) for each subsampling level, thus totaling 24 sets. Each level of taxonomy connoted the lowest level achieved for most identifications. Data were densities (no. m−2) of each taxon for each of the 126 plots sampled. We evaluated the importance of identifying Chironomidae beyond family level by constructing three tiered data sets (1) non-Chironomidae family-level data tiered with species-level Chironomidae data, (2) non-Chironomidae genus-level data tiered with family-level Chironomidae data, and (3) non-Chironomidae specieslevel data tiered with family-level Chironomidae data. These tiered sets were compared with family-, genus-, and species-level data sets. A representative midsized subsample (200 count + LR) was used.

11.2.5

Macroinvertebrate Sample Processing: Biomass Estimation

Macroinvertebrate biomass was estimated for every plot in the spatial and temporal studies. Every individual of most taxa was measured to the nearest 0.5 mm. These measurements were used in taxon-specific length–mass regression equations to estimate individual dry mass (Kushlan et al. 1986; Meyer 1989; Sample et al. 1993; Benke et al. 1999). Biomass of taxa that either did not have published length–mass equations or were very small was estimated using a biovolume technique (Smit et al. 1993). For small taxa, particularly some Chironomidae and Oligochaeta, individuals were enumerated into taxon-specific size classes, based on length and width – these size classes were used to estimate dry mass using biovolume. Biomass of Gastropoda (other than Pomacea paludosa; Kushlan et al. 1986) was also estimated using biovolume since few length–mass equations were published to estimate flesh mass (excluding shell mass). Proximate geometric shapes and measured dimensions of tissue of individual gastropods were used to estimate biovolume and dry mass. Densities (no. m−2) of each taxon were used to calculate biomass (mg dry wt m−2) for each plot.

286

R.S. King and C.J. Richardson

Invertivorous fish were expected to be a possible determinant of invertebrate assemblage biomass and composition. Fish abundance and biomass were estimated using the same dip-net samples used to collect invertebrates. Because of dynamic hydrology, small, surface-oriented taxa dominate the fish assemblage in the Everglades (Jordan 1996; Turner et al. 1999), and the dip-net approach has been shown to be an effective technique for estimating abundance of these fishes (Rader and Richardson 1994). Fish were exhaustively picked from each sample, enumerated, measured for total length, and identified to species. Total length was used to estimate biomass (mg dry wt) (Kushlan et al. 1986). Taxa that were predaceous or omnivorous were classified as invertivorous. Densities and biomass (no. m−2 and mg dry wt m−2, respectively) of insectivorous fish were subsequently estimated for each plot.

11.2.6

Data Analysis: Effects of Subsampling and Taxonomic Resolution on Bioassessment

We compared the magnitude of assemblage–environment relationships among subsampling approaches and levels of taxonomic resolution using the multivariate Mantel test (Mantel 1967), which measures the correlation between distance matrices. Increasing magnitude in Mantel r, the test statistic, reflects a stronger correlation. Mantel r typically ranges from 0.1 to 0.3 for assemblage–environment relationships that are ecologically significant and infrequently exceeds 0.5 because the analysis is based on the full rather than reduced dimensionality (e.g., ordinationaxis scores) in the assemblage data (e.g., Leduc et al. 1992; Sanderson et al. 1995; Foster et al. 1999). We selected distance from canal inflow structures (hereafter, Canal) as predictor of macroinvertebrate assemblage composition because (1) it was a surrogate for a wide range of biogeochemical, hydrological, and habitat– structural variables that substantially change along this eutrophication gradient (Table 11.1) and (2) it was the best predictor of biological changes in this study area (King 2001). Canal (m) was converted to a distance matrix using Euclidean distance, whereas assemblage matrices used Bray–Curtis dissimilarity as the distance metric (Legendre and Legendre 1998). Bray–Curtis dissimilarity was selected because it is one of the most robust and ecologically interpretable distance metrics available (e.g., Faith et al. 1987; Legendre and Anderson 1999; Hawkins and Norris 2000). All macroinvertebrate density data were log10(x + 1) transformed prior to conversion to distance matrices to give greater weight to less-abundant taxa (Legendre and Legendre 1998). We estimated 95% confidence intervals (CIs) for each test statistic using bootstrapping, a resampling method (Manly 1997), rather than qualitatively comparing the magnitude of Mantel r statistics among subsamples and taxonomic levels. We resampled (with replacement) distance matrices at a level of 90%, with 1,000 resamples (Manly 1997). Mantel r statistics were considered significantly different if 95% CIs did not overlap (Manly 1997; Johnson 1999). We also evaluated whether Mantel r statistics were significantly different from 0 (p < 0.05) using 10,000 random

11 Macroinvertebrate Responses to a Gradient of Long-Term Nutrient Additions

287

permutations (Manly 1997); however, this test was merely an antecedent to the more relevant comparison of uncertainty (95% CI) among assemblage–environment correlations (Suter 1996; Germano 1999; Johnson 1999). Mantel tests and bootstrapping were done using S-Plus 5.0 for Unix (Mathsoft, Inc., Seattle, WA, USA).

11.2.7

Data Analysis: Diversity and Biomass in Relation to P

Biomass of the complete invertebrate assemblage, as well as of the common coarsetaxonomic groups (classes or orders), was plotted and regressed against distance from canal inflow structures and sediment total phosphorus (TP) to evaluate assemblage response relationships to the P gradient. Averages of biomass among all plots (n = 9) within each plot-cluster (n = 14) were used as replicates since these represented an estimate of the biomass across a large spatial area (weighted by vegetation pattern) rather than at individual plots (Allen and Wyleto 1983; Turner et al. 1999). Distance from canal (m) was based on the centroid of each cluster, while sediment TP was an average value from all plots within each plot-cluster. Since distance from canal produced results very similar to that of sediment TP for total biomass, all regressions were subsequently based only on sediment TP. All biomass data were log10(x) or log10(x + 1) transformed prior to averaging and analysis. Assemblage diversity was evaluated using species density (number of taxa/ fixed-area subsample) and species richness (total number of taxa/cluster of plots) (Hurlbert 1971; Larsen and Herlihy 1998). Species density was averaged among the nine plots per cluster, while species richness was the total accumulation of unique taxa among the nine plots per cluster. Species richness was estimated using data produced from the tiered 25% fixed-area/300 fixed-count + LR search subsampling approach (Vinson and Hawkins 1996). Species density and richness were regressed against sediment TP. To characterize diversity and distribution of invertebrate species at a landscape scale, species accumulation curves were generated using species-richness data from each plot. Accumulation curves were stratified by impact zone to examine differences in diversity among landscape regions of differing vegetation and nutrient status. Curves provided a visual assessment of average fine-scale (plot) as well as broad-scale (cluster and impact zone) species richness. Asymptotic species richness (Sjack) was also estimated for each curve using a firstorder jackknife procedure (Palmer 1990) to provide a better evaluation of the total expected number of species per zone. Accumulation curves and jackknife estimates were performed using PC-ORD 4.09.

11.2.8

Data Analysis: Abiotic and Biotic Drivers of Macroinvertebrate Community Structure

Two complementary procedures were used to determine the primary dimensions of invertebrate assemblage organization across the landscape. First, plots and species

288

R.S. King and C.J. Richardson

were ordinated based on species composition using nonmetric multidimensional scaling (nMDS; Minchin 1987). Ordination provided a visual assessment of gradients in species composition among and within impact zones. Bray–Curtis dissimilarity was used as the distance metric, a coefficient shown to be one of the most robust and ecologically meaningful (Faith et al. 1987). Log-transformed density data (log10[no. m−2]) for each taxon was used in calculation of dissimilarities rather than biomass data because preliminary results using densities indicated that it was slightly more robust (lower stress – an indicator of goodness-of-fit). Once plots were ordinated, species centroids were mapped into ordination space using weighted-averaging (Legendre and Legendre 1998). Ordinations were limited to two or three dimensions, as stress values were relatively low and exhibited small decreases at higher dimensionality. To relate abiotic and biotic variables to gradients in composition in nMDS ordinations, rotational vector fitting was used (Faith and Norris 1989). Vector fitting was performed on all ordinations. Abiotic and biotic values from each plot were used in fine-scale vector fitting, while average values from within each plot-cluster were used in the coarse-scale analysis. For vegetation data, dominant species (e.g., cover of Typha, Cladium) and structural groups (e.g., cover of unrooted floating species) were used as predictors. Density of invertivorous fish was used instead of biomass because it showed a stronger relationship to composition. Significance (Bonferroni-corrected p ≤ 0.05) of vectors was estimated using 10,000 random permutations. Ordination and vector fitting were performed using DECODA 2.05 (University of Melbourne, Parkville, Victoria, Australia). To assess sensitivities or affinities of invertebrate species among impact zones and help explain patterns of diversity, Indicator Species Analysis was used (INSPAN; Dufrêne and Legendre 1997). INSPAN is a nonparametric technique used to identify species with a high fidelity for a particular group or class, as defined by the user. The three impact zones were used as classes for the analysis. Significance (Bonferroni-corrected p ≤ 0.05) of indicator values was estimated using 10,000 random permutations (Manly 1997). INSPAN was performed using PC-ORD 4.09 (MjM Software, Gleneden Beach, OR, USA).

11.3 11.3.1

Results Subsample Characteristics and Taxonomic Structure

Over 78,000 individuals were identified across all 126 plots in the spatial study during October 1998, and additional 66,000 individuals were identified during the temporal study. A total of 93 families, 181 genera, and 272 unique taxa (species or morphospecies; lowest level of taxonomy achievable) were identified (see Table 11.2 for a complete list of macroinvertebrates identified in the Everglades gradient and dosing studies). Coleopterans, dipterans, gastropods, odonates, and oligochaetes were the

Gradient Group Amphipoda Amphipoda Anomopoda Anomopoda Anomopoda Anomopoda Anomopoda Anomopoda Anomopoda Anomopoda Anomopoda Anomopoda Anomopoda Arhynchobdellida Arhynchobdellida Arhynchobdellida Arhynchobdellida Arhynchobdellida Arynchobdellida Bivalvia Coleoptera Coleoptera Coleoptera Coleoptera Coleoptera Coleoptera Coleoptera

Family Crangonycitidae Hyalellidae Chydoridae Chydoridae Chydoridae Chydoridae Chydoridae Daphniidae Daphniidae Daphniidae Macrothricidae Macrothricidae Macrothricidae Erpobdellidae Erpobdellidae Erpobdellidae Hirudinidae Hirudinidae Haemopidae Sphaeriidae Curculionidae Dytiscidae Dytiscidae Dytiscidae Dytiscidae Dytiscidae Dytiscidae

Code CRANGONY HYALAZTE CAMPTOCE CHYDOSP1 CHYDOSP2 CHYDOSP3 CHYDORID CERIODAP DAPHNSP1 SIMOCEPH ILYOSPIN MACROTHR OPHRYOXU MOORMELA MOORMICR MOORTETR MACRDITE PHILOBDE HAEMSEPT SPHAERIU CURCULIO CELIANGU CELIIMIT CELISLOS CELINA.L CYBIFIMB DESMOPAC

Taxon Crangonyx nr. richmondensis Ellis Hyalella azteca (Saussure) Camptocercus sp. Chydoridae sp. 1 Chydoridae sp. 2 Chydoridae sp. 3 Chydoridae spp. Ceriodaphnia sp. Daphnidae sp. 1 Simocephalus sp. Ilyocryptus spinifer Herrick Macrothricidae sp. 1 Ophryoxus sp. Mooreobdella melanastoma Sawyer and Shelley Mooreobdella microstoma (Moore) Mooreobdella tetragon Sawyer and Shelley Macrobdella ditetra Moore Philobdella sp. Haemopsis septagon (Sawyer and Shelley) Sphaerium sp. Curculionidae sp. Celina angustata Aube Celina imitatrix Young Celina slossoni Mutchler Celina sp. larva Cybister fimbriolatus Wilke Desmopachria sp.

Spatial x x

x x x x x x x x x x x x x x x x x x x

Temporal Dosing x x x x x x x x x x x x x x x x x x x x

x

x x x x x x x

11 Macroinvertebrate Responses to a Gradient of Long-Term Nutrient Additions

Table 11.2 List of invertebrate taxa collected from the P gradient (spatial and temporal) and P-dosing studies in the Everglades

x

289

(continued)

290

Table 11.2 (continued) Gradient Group

Family

Code

Taxon

Spatial

Coleoptera Coleoptera Coleoptera Coleoptera Coleoptera Coleoptera Coleoptera Coleoptera Coleoptera Coleoptera Coleoptera Coleoptera Coleoptera Coleoptera Coleoptera Coleoptera Coleoptera Coleoptera Coleoptera Coleoptera Coleoptera Coleoptera Coleoptera Coleoptera Coleoptera Coleoptera Coleoptera Coleoptera

Dytiscidae Dytiscidae Dytiscidae Dytiscidae Dytiscidae Dytiscidae Dytiscidae Dytiscidae Dytiscidae Gyrinidae Gyrinidae Haliplidae Hydraenidae Hydrophilidae Hydrophilidae Hydrophilidae Hydrophilidae Hydrophilidae Hydrophilidae Hydrophilidae Hydrophilidae Hydrophilidae Hydrophilidae Hydrophilidae Hydrophilidae Hydrophilidae Hydrophilidae Hydrophilidae

HYDROPOR HYDROPOR HYDRPUST HYDROV.L LACCGENT LACCPROX LACCOP.L MATUOVAT NEOPORUS GYRIELEV GYRINU.L HALIPL.L HYDRAE.L BEROINFU BEROSU.L DERAALTU ENOCBLAT ENOCHAMI ENOCCONS ENOCOCHR ENOCPYGM ENOCSAYI ENOCHR.L HELOLARV HYDRCAST PHAEEXST PHAEMINO PHAENO.L

Hydroporus sp. Hydroporus sp. larva Hydrovatus pustulatus compressus Sharp Hydrovatus sp. larva Laccophilus gentilis gentilis LeConte Laccophilus proximus Say Laccophilus spp. larva Matus ovatus blatchleyi Leech Neoporus sp. Gyrinus elevatus LeConte Gyrinus sp. larva Haliplus sp. larva Hydraena sp. larva Berosus infuscatus LeConte Berosus sp. larva Derallus altus (LeConte) Enochrus blatchleyi (Fall) Enochrus consortus Green Enochrus hamiltoni (Horn) Enochrus ochraceus (Melsheimer) Enochrus pygmaeus pygmaeus (Fabricius) Enochrus sayi Gunderson Enochrus spp. larva Helobata larvalis (Horn) Hydrobiomorpha casta (Say) Phaenonotum exstriatum (Say) Phaenonotum minor Smetana Phaenonotum spp. larva

x x x x x x x x

x x

Temporal Dosing

x x x x x x x x

x x x x x x x x x

x x

x x

x x

x x x x x x

x x x

R.S. King and C.J. Richardson

x x x x x

TROPBLAT TROPLATE TROPIS.L LAMPYRID HYDROBLO HYDRREGI HYDCAN.L SUPHINFL SUPHGIBB SUPHPUNC SUPHISEL SCIRTES STAPHYLI ENTOMOBR ISOTOMA SMINTHRI SMINTHRU CALANOID CYCLPOID HARPACTI PSEUDOSI PROCALLE PROCFALL PALAPALU ATRICHOP BEZZSP1 BEZZSP2 BEZZSP3 CERATOPO CULICOID DASYNELE

Tropisternus blatchleyi blatchleyi d’Orchymont Tropisternus lateralis nimbatus (Say) Tropisternus spp. larva Lampyridae sp. larva Hydrocanthus oblongus Sharp Hydrocanthus regius Young Hydrocanthus spp. larva Suphis inflatus LeConte Suphisellus gibbulus (Aube) Suphisellus puncticollis Crotch Suphisellus spp. larva Scirtes sp. Staphylinidae sp. Entomobrya sp. Isotoma sp. Sminthurides sp. Sminthurus sp. Calanoida Cyclopoida Harpacticoida Pseudosida bidentata Herrick Procambarus cf. alleni (Faxon) Procambarus fallax (Hagen) Palaemonetes paludosus (Gibbes) Atrichopogon sp. Bezzia/Palpomyia gr. sp. 1 Bezzia/Palpomyia gr. sp. 2 Bezzia/Palpomyia gr. sp. 3 Ceratopogon sp. Culicoides sp. Dasyhelea sp.

x x x x x x x x x x x x x x x x x x x x x

x x x x x

x x x x x x x

x x x x x x x x x

x

x

x

291

Hydrophilidae Hydrophilidae Hydrophilidae Lampyridae Noteridae Noteridae Noteridae Noteridae Noteridae Noteridae Noteridae Scirtidae Staphylinidae Entomobryidae Isotomuridae Sminthuridae Sminthuridae Calanoida Cyclopoida Harpacticoida Sididae Cambaridae Cambaridae Palaemonidae Ceratopogonidae Ceratopogonidae Ceratopogonidae Ceratopogonidae Ceratopogonidae Ceratopogonidae Ceratopogoridae

11 Macroinvertebrate Responses to a Gradient of Long-Term Nutrient Additions

Coleoptera Coleoptera Coleoptera Coleoptera Coleoptera Coleoptera Coleoptera Coleoptera Coleoptera Coleoptera Coleoptera Coleoptera Coleoptera Collembola Collembola Collembola Collembola Copepoda Copepoda Copepopda Ctenopoda Decapoda Decapoda Decapoda Diptera Diptera Diptera Diptera Diptera Diptera Diptera

x x x x x x x x x x x x x

x x x x x x

(continued)

292

Table 11.2 (continued) Gradient Group

Family

Code

Taxon

Spatial

Temporal Dosing

Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera

Ceratopogonidae Ceratopogonidae Chironomidae Chironomidae Chironomidae Chironomidae Chironomidae Chironomidae Chironomidae Chironomidae Chironomidae Chironomidae Chironomidae Chironomidae Chironomidae Chironomidae Chironomidae Chironomidae Chironomidae Chironomidae Chironomidae Chironomidae Chironomidae Chironomidae Chironomidae Chironomidae Chironomidae Chironomidae

FORCMYIA PROBEZZI ABLAPELE ABLARHAM APEDELAS BEARTRUN CHMUSSP CHIRSTIG CLADOPEL CLADOTAN CORYSPD CRICSYLV DICROMOD DICROSIM DICROSP DICROSPA DJALPULC ENDONIGR ENDOHESP FITTSERT GLYPTOSP GOELDCAR GOELDNAT GOELDHOL KIEFFDUX KIEFFSPA LABRBECK LABRNEOP

Forcipomyia sp. Probezzia sp. Ablabesmyia peleensis (Walley) Ablabesmyia rhampe Sublette gr. Apedilum elaschitus Townes Beardius truncatus gr. sp. Reiss & Sublette Chironomus sp. Chironomus stigmaterus Say Cladopelma sp. Cladotanytarsus sp. Corynoneura sp. D Epler Cricotopus sylvestris Fabricius gr. Dicrotendipes modestus (Say) Dicrotendipes simpsoni Epler Dicrotendipes sp. Dicrotendipes sp. A Epler Djalmabatista pulchra (Johannsen) Endochironomus nigricans (Johannsen) Endotribelos hesperium (Sublette) Fittkauimyia serta (Roback) Glyptotendipes sp. Goeldichironomus carus (Townes) Goeldichironomus cf. natans Reiss Goeldichironomus holoprasinus (Goeldi) Kiefferulus dux/pungens gr. sp. Kiefferulus sp. A Epler Labrundinia becki Roback Labrundinia neopilosella Beck and Beck

x x x x x x x x x x x x x x x x

x x

x x x x x x x

x x x x x x x x x x x x

x x x x x x x x x

x x

x x x

x

R.S. King and C.J. Richardson

x x x

x x x x x x x x x x x

Chironomidae Chironomidae Chironomidae Chironomidae Chironomidae Chironomidae Chironomidae Chironomidae Chironomidae Chironomidae Chironomidae Chironomidae Chironomidae Chironomidae Chironomidae Chironomidae Chironomidae Chironomidae Chironomidae Chironomidae Chironomidae Chironomidae Chironomidae Chironomidae Chironomidae Chironomidae Chironomidae Chironomidae Chironomidae Chironomidae

LARSBERN LARSDECO LIMNOPHY MONOBOLI NANOALTE NILOTHAU PARACALA PARACDIR PARACMON PARAKSPC PARAMERI PARASPB PARASPC POLYHALT POLYILLI POLYPSPA POLYTRIG POLYTRIT PROCLAD PSEUDOCH TANYPCAR TANYLIMN TANYSP TANYSPE TANYSPF TANYSPG TANYSPJ TANYSPR TANYSPT XENOXENO

Larsia berneri Beck & Beck Larsia decolorata (Malloch) Limnophyes sp. Monopelopia boliekae Beck and Beck Nanocladius alternantherae Dendy & Sublette Nilothauma sp. Parachironomus alatus (Beck) Parachironomus directus (Dendy and Sublette) Parachironomus monochromus/tenuicaudatus gr. Parakiefferiella sp. C Epler Paramerina sp. Paratanytarsus sp. B Epler Paratanytarsus sp. C Epler Polypedilum halterale (Coquillett) gr. Polypedilum illinoense (Malloch) gr. Polypedilum sp. A Epler Polypedilum trigonus Townes Polypedilum tritum (Walker) Procladius (Holotanypus) sp. Pseudochironomus sp. Tanypus carinatus Sublette Tanytarsus limneticus Sublette Tanytarsus sp. Tanytarsus sp. E Epler Tanytarsus sp. F Epler Tanytarsus sp. G Epler Tanytarsus sp. J Epler Tanytarsus sp. R Epler Tanytarsus sp. T Epler Xenochironomus xenolabis (Kieffer)

x x x x x x x x x x x x x x x x x x x x x x x x x x x

x x x x x x x x x x x x x x x x x x x x x x x x x x x

x

x x x x x x x x x x x x x x x

x x

11 Macroinvertebrate Responses to a Gradient of Long-Term Nutrient Additions

Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera

(continued) 293

294

Table 11.2 (continued) Gradient Group

Family

Code

Taxon

Spatial

Temporal Dosing

Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Diptera Ectoprocta Ephemeroptera Ephemeroptera Gastropoda Gastropoda Gastropoda Gastropoda Gastropoda Gastropoda

Culicidae Culicidae Culicidae Culicidae Culicidae Culicidae Culicidae Dolichopodidae Ephydridae Muscidae Psychodidae Sciomyzidae Stratiomyiidae Tabanidae Tipulidae Tipulidae Tipulidae Tipulidae Tipulidae Plumatellidae Baetidae Caenidae Ancylidae Ancylidae Ancylidae Hydrobiidae Hydrobiidae Lymnaeidae

AEDES ANOPHSP1 ANOPHSP2 COQUPERT CULEX MANSOSP2 URANOTAE DOLICHOP EPHYDRA MUSCIDAE PSYCHODA SCIOMYZI ODONTOMY CHRYSOPS HELIUS LIMONIA ORMOSIA TIPULSP1 TIPULSP2 PLUMATEL CALLFLOR CAENDIMI FERRISSI HEBEEXCE LAEVPENI APHAOPAC LITTMONR FOSSMODI

Aedes sp. Anopheles sp. 1 Anopheles sp. 2 Coquillettidia perturbans (Walker) Culex sp. Mansonia titillans (Walker) Uranotaenia sapphirina (Osten Sacken) Dolichopodidae sp. cf. Ephydra sp. Muscidae sp. Psychoda/Threticus gr. sp. Sciomyzidae sp. Odontomyia sp. Chrysops sp. Helius sp. Limonia sp. Ormosia sp. Tipulidae sp. 1 Tipulidae sp. 2 Plumatella cf. repens (L.) Callibaetis floridanus Banks Caenis diminuta Walker Ferrissia sp. Hebetancylus excentricus (Morelit) Laevapex peninsulae (Pilsbry) Aphaostracon pachynotus Thompson Littoridinops monroensis (Frauenfeld) Fossaria modicella (Say)

x x x x x x x x

x x x x x x x x

x x x x x x x x

x x x x x x x x x

x x

x

x x x x

x x

R.S. King and C.J. Richardson

x x x x x x x x x

x x x x x x x

Lymnaeidae Physidae Physidae Pilidae Planorbidae Planorbidae Planorbidae Planorbidae Planorbidae Planorbidae Polygyridae Pupillidae Thiaridae Zonitidae Belostomatidae Belostomatidae Belostomatidae Corixidae Gerridae Hydrometridae Mesoveliidae Naucoridae Nepidae Notonectidae Pleidae Saldidae Veliidae Arrenuridae Arrenuridae Arrenuridae Arrenuridae

PSEUDCOL PHYSCUBE PHYSELLA POMAPALU GYRAPARV MICRDILA PLANDURY PLANSP1 PLANSCAL PLANTRIV POLYCERE VERTOVAT MELATUBE ZONIARBO BELOLUTA BELO.IMM BELOTEST TRICHORX RHEUVEGA HYDROMET MESOVELI PELOFEMO RANAAUST BUENOA PARAPLEA SALDIDAE MICRVELI ARREAPOP ARREZAPU ARRENSP1 ARRENSP2

Pseudosuccinea columella (Say) Physella cubensis (Pfieffer) Physella sp. Pomacea paludosa (Say) Gyraulus parvus (Say) Micromenetus dilatatus avus (Pilsbry) Planorbella duryi (Weatherby) Planorbella duryi/scalaris complex Planorbella scalaris (Jay) Planorbella trivolvis intertexta (Jeffreys) Polygyra cereolus (von Muhlfeld) Vertigo ovatus (Say) Melanoides tuberculata (Muller) Zonitoides arboreau (Say) Belostoma lutarium (Stal) Belostoma spp. immature Belostoma testaceum (Leidy) Trichocorixa sp. Rheumatobates cf. vegatus Drake and Harris Hydrometra sp. Mesovelia mulsaulti White Pelocoris femoratus (Palisot-Beauvois) Ranatra australis Hungerford Buenoa sp. Paraplea sp. Saldidae sp. Microvelia sp. Arrenurus nr. apopkensis Cook Arrenurus nr. zapus Cook Arrenurus sp. 1 Arrenurus sp. 2

x x x x x x x x x x x x x x x x x x x x x x x x x x x x

x x x x

x x x x x x x x x

x x

x x

x x x x x x x x x x x x x

x x x x x x x

x

x x

x

11 Macroinvertebrate Responses to a Gradient of Long-Term Nutrient Additions

Gastropoda Gastropoda Gastropoda Gastropoda Gastropoda Gastropoda Gastropoda Gastropoda Gastropoda Gastropoda Gastropoda Gastropoda Gastropoda Gastropoda Hemiptera Hemiptera Hemiptera Hemiptera Hemiptera Hemiptera Hemiptera Hemiptera Hemiptera Hemiptera Hemiptera Hemiptera Hemiptera Hydracarina Hydracarina Hydracarina Hydracarina

x x x 295

(continued)

296

Table 11.2 (continued) Gradient Group

Family

Code

Taxon

Spatial

Temporal Dosing

Hydracarina Hydracarina Hydracarina Hydracarina Hydracarina Hydracarina Hydracarina Hydracarina Hydracarina Hydracarina Hydroida Isopoda Lepidoptera Lepidoptera Lepidoptera Lepidoptera Lepidoptera Lepidoptera Lepidoptera Lepidoptera Lepidoptera Nemata Odonata Odonata Odonata Odonata Odonata Odonata

Arrenuridae Eylaidae Hydrodromidae Limnesiidae Limnocharidae Oxidae Sperchontidae Unionicolidae Unionicolidae Unionicolidae Hydridae Asellidae Nepticulidae Noctuidae Noctuidae Noctuidae Pyralidae Pyralidae Pyralidae Pyralidae Pyralidae Nemata Aeschnidae Aeschnidae Coenagrionidae Coenagrionidae Coenagrionidae Coenagrionidae

ARREZORU EYLAIS HYDRODRO LIMNESIA LIMNOCHA OXUS SPERCHON KOENIKEA NEUMANIA UNIONICO HYDRA CAECIDOT NEPTICUL NOCTUSP1 NOCTUSP2 SIMYRA ACENTRIA PARAPONY PYRALSP1 PYRALSP2 PYRALSP3 NEMATA ANAX CORYINGE ENALCIVI ENALPOLL ISCHHAST ISCHPOSI

Arrenurus zorus Cook Eylais sp. Hydrodroma sp. Limnesia sp. Limnochares sp. Oxus sp. Sperchon sp. Koenikea sp. Neumania sp. Unionicola sp. Hydra sp. Caecidotea sp. Nepticulidae sp. Noctuidae sp. 1 Noctuidae sp. 2 Simyra henrici (Grt.) Acentria sp. Paraponyx sp. Pyralidae sp. 1 Pyralidae sp. 2 Pyralidae sp. 3 Nemata Anax sp. Coryphaeschna ingens (Rambur) Enallagma civile Hagen Enallagma pollutum Hagen Ischnura hastata Say Ischnura posita

x x x x x x

x x

x x x x

x x x x x x x

x x x x x x x

x x x

x

x

x

R.S. King and C.J. Richardson

x x x x x x x x x x x x x

x x x x x x x x x

x x x x x

Coenagrionidae Coenagrionidae Libellulidae Libellulidae Libellulidae Libellulidae Libellulidae Enchytraeidae Lumbriculidae Naididae Naididae Naididae Naididae Naididae Naididae Naididae Naididae Naididae Naididae Naididae Naididae Naididae Opistocystidae Tubificidae Tubificidae Tubificidae Tubificidae Candonidae Cyprididae Cyprididae

ISCHNURA TELEBYER BRACGRAV CELIEPON ERYTHSIM LIBENEED PACHLONG ENCHYTRA ECLIPALU ALLOPECT BRATUNID DERODIGX DEROFURC DEROLODE DEROPECT DEROVAGA HAEMWALD PRISAEQU PRISLEID PRISLONG PRISTINE STYLLACU CRUSTRIB LIMNHOFF TUBIFSP1 TUBIFSP2 TUBIFSP3 CANDANNA CHLATEXA CHLAUNIS

Ischnura sp. Telebasis byersi Westfall Brachymesia gravida (Calvert) Celithemis eponina (Drury) Erythemis simplicicollis (Say) Libellula needhami Westfall Pachydiplax longipennis (Burmeister) Enchytraeidae Eclipidrilus palustris (Smith) Allonais pectinata (Stephenson) Bratislavia unidentata (Harman) Dero digitata (Muller) complex Dero furcata (Muller) Dero lodeni (Stephenson) Dero pectinata Aiyer Dero vaga (Leidy) Haemonais waldvogeli Bretscher Pristina aequiseta Bourne Pristina leidyi Smith Pristinella longisoma (Harman) Pristinella sp. Stylaria lacustris (L.) Crustipellis tribranchiata (Harman) Limnodrilus hoffmeisteri Clarapede Tubificidae imm. gr. 1 Tubificidae imm. gr. 2 Tubificidae imm. gr. 3 Candona annae Mehes Chlamydotheca texasiensis (Baird) Chlamydotheca unispinosa (Baird)

x x x x x x x x x x x x x x x x x x x x x x x x x x x x

x x x x x x x x x x x x x x x x x x x

x x x x x

x x x x x x x x

x x x x x x x x x

x x x x x

11 Macroinvertebrate Responses to a Gradient of Long-Term Nutrient Additions

Odonata Odonata Odonata Odonata Odonata Odonata Odonata Oligochaeta Oligochaeta Oligochaeta Oligochaeta Oligochaeta Oligochaeta Oligochaeta Oligochaeta Oligochaeta Oligochaeta Oligochaeta Oligochaeta Oligochaeta Oligochaeta Oligochaeta Oligochaeta Oligochaeta Oligochaeta Oligochaeta Oligochaeta Ostracoda Ostracoda Ostracoda

(continued) 297

298

Table 11.2 (continued) Gradient Group

Family

Code

Taxon

Spatial

Temporal Dosing

Ostracoda Ostracoda Ostracoda Polychaeta Porifera Porifera Rhynchobdellida Rhynchobdellida Rhynchobdellida Rhynchobdellida Trichoptera Trichoptera Trichoptera Trichoptera Trichoptera Tricladida

Cypridopsidae Cytheridae Cytheridae Neridae Spongillidae Spongillidae Glossiphoniidae Glossiphoniidae Glossiphoniidae Glossiphoniidae Hydroptilidae Leptoceridae Leptoceridae Leptoceridae Polycentropodidae Planariidae

CYPROKEE CYTHALOS HETEPUNC NAMALABI SPONCENO SPONGILL HELOFUSC HELOSTAG HELOTRIS PLACPAPI OXYETHIR OECECINE OECETSPE OECETSP CERNOTIN PLANARII

Cypridopsis okeechobei Furtos Cytheridella alosa (Tressler) Heterocypris punctata (Baird) Namalycastis abiuma (Muller) Spongilla cf cenota Penney and Racek Spongilla sp. Helobdella fusca (Castle) Helobdella stagnalis (Linnaeus) Helobdella triserialis (Blanchard) Placobdella papillifera (Verill) Oxyethira sp. Oecetis cinerascens (Hagen) Oecetis inconspicua complex sp. E Floyd Oecetis sp. Cernotina sp. Planariidae sp.

x x

x x x x

x

x x

x x x x x x x x x x x x

x x x x x x x x x

x x

x

R.S. King and C.J. Richardson

11 Macroinvertebrate Responses to a Gradient of Long-Term Nutrient Additions

299

Table 11.3 Comparison of selected properties of the eight subsampling approaches evaluated for wetland bioassessment (from King and Richardson 2002) Percentage of total sample

Number of individuals

Sorting time (min)

Subsample

Mean ± 1 SD

Mean ± 1 SD

Range

Mean ± 1 SD

100 count 200 count 300 count 10% area 25% area 100 count + LRa 10% area + LR 100 count and 10% area

7.7 ± 6.4 15.2 ± 12.7 22.6 ± 18.3 10.0 ± − 25.0 ± – NA NA 11.6 ± 4.3

102.7 ± 4.7 203.7 ± 7.3 304.6 ± 10.9 230.0 ± 178.3 573.1 ± 440.3 121.3 ± 14.3 247.7 ± 177.7 238.2 ± 170.9

92–118 191–224 283–326 23–1,036 62–2,558 100–214 36–1,044 97–1,036

94.2 ± 55.3 156.1 ± 100.0 206.7 ± 117.8 94.7 ± 51.5 250.7 ± 136.2 117.0 ± 54.0 117.2 ± 51.7 118.0 ± 64.5

NA not applicable Subsamples containing the large rare (LR) component were picked completely for all LR taxa (100% of sample area) in addition to the fixed-count or fixed-area component a

most diverse of the major taxonomic groups, and contributed most to the differences among the number of families, genera, and species identified. Chironomidae was the most diverse family in the spatial study, represented by 30 and 51 genera and species, respectively. Numbers of individuals showed tremendous variation among subsamples using the fixed-area approach (Table 11.3). Although 10% area averaged over twice the number of individuals as the 100 count, it produced as few as 23 individuals in one subsample, and had <100 individuals 27% of the time. Similarly, the 25% area averaged nearly twice the number of individuals as the 300 count despite averaging a similar percentage of the total sample subsampled. The LR search added an average of as many as 19 individuals to subsamples. Sorting times mirrored the percentage of total sample subsampled rather than number of individuals picked (Table 11.3). LR searches added an average of 23 min (100 + LR) to sorting time. The LR search added as many as 4 families, 9 genera, and 16 species, cumulatively, to any one level of subsampling (100 vs. 100 + LR). Frequencies of occurrence of many LR taxa increased as much as a factor of 10 by implementing the LR search, with the 100-count subsample performing the poorest of all in capturing LR taxa (King and Richardson 2002).

11.3.1.1

Effects of Subsampling and Taxonomic Resolution on Assemblage–Environment Relationships

Mantel r statistics were significantly different from 0 (p ≤ 0.0001), regardless of subsample or taxonomic level. However, the magnitude of these assemblage– environment correlations varied significantly (95% CI) among subsamples and taxonomic levels (Fig. 11.3). In particular, the greatest increase in assemblage– environment relationships with increasing subsample size was observed between

300

R.S. King and C.J. Richardson

100 and 200 counts – 100 performed significantly worse than 200, whereas 200 was not different from 300, regardless of taxonomic level. Differences in the magnitude of Mantel r values suggested that fixed-area subsamples generally were less efficient than fixed counts. Fixed counts of 200 and 300 individuals produced significantly greater Mantel r values than 10% area, despite averaging similar numbers of individuals (Fig. 11.3). Similarly, 25% area assemblage–environment relationships were not significantly greater than the less labor-intensive 300 count at the genus and species levels. Adding the LR search to 100-count and 10% area subsamples resulted in very slight increases in the strength of assemblage–environment relationships for all three levels of taxonomy (Fig. 11.3). LR taxa significantly increased the Mantel r value for 100-count data at the family level.

Fig. 11.3 Assemblage–environment correlations for each subsampling approach and level of taxonomic resolution, as estimated using Mantel tests. Significant differences in the magnitude of Mantel r values (bootstrapped 95% CI, error bars) among subsamples within taxonomic levels are indicated by the lower-case letters; Mantel r values with the same letters were not different. Among taxonomic levels, Mantel r values with overlapping 95% CI were not significantly different (all Mantel r values differed among the three levels of taxonomy within each level of subsampling)

11 Macroinvertebrate Responses to a Gradient of Long-Term Nutrient Additions

301

Differences in the magnitudes of assemblage–environment correlations were more apparent among levels of taxonomic resolution than among subsampling approaches. The family level was significantly inferior to both genus- and species-level data, regardless of subsampling approach (Fig. 11.3). Species-level data showed significantly stronger relationships to the environment than the genus level, although 95% CIs were only marginally separated within each level of subsampling. Chironomidae may have been largely responsible for the observed disparity in correlations to the environment among taxonomic levels (Fig. 11.4). Results from the tiered-taxonomic analysis revealed that tiering family-level data with specieslevel Chironomidae data yielded assemblage–environment correlations that were not different from those obtained by identifying all taxa to genus or species. Conversely, leaving Chironomidae identifications at just the family level but identifying other taxa to genus or species produced significantly worse assemblage–environment correlations than that of genus, species, and tiered family/Chironomidae-to-species data (King and Richardson 2002). Species accumulation curves indicated that species richness patterns among impact zones were scale dependent (Fig. 11.5). On a single-plot scale, the transition zone averaged more species than the impacted, while both of these zones averaged more than the reference zone. However, steepness of the accumulation curve was initially greater in the impacted zone than the transition, resulting in higher richness in this eutrophic region. The impacted-zone curve sharply flattened above 20 plots, while transition-zone plots continued to accumulate new species. Jackknife estimates of asymptotic richness indicated the intermediate-P, transition zone had the most species at a landscape scale, while impacted and reference zones were similar in total richness.

11.3.2

Biomass Responses to P

Macroinvertebrate assemblage biomass exhibited a significant unimodal response to both distance from canal (Fig. 11.6a) and sediment TP (Fig. 11.6b). These subsidy–stress relationships were nearly identical for both predictor variables; therefore, subsequent responses were evaluated only with sediment TP as a predictor. Twelve major taxonomic groups were evaluated for their specific responses to P. Eight groups showed subsidy–stress responses (Fig. 11.7a, b, e–g, i–k), three demonstrated significant subsidy responses (Fig. 11.7c, d, h), while one showed a significant stress response (Fig. 11.7l). Of the eight taxa responding with a subsidy– stress relationship, five were statistically significant. Decapoda made the greatest contribution to assemblage biomass, and revealed the most obvious subsidy–stress response to P enrichment (Fig. 11.7g). Represented by only two species, Palaemonetes paludosus and Procambarus fallax, their cumulative standing stocks increased markedly with intermediate (transition zone) P enrichment, but plummeted in localities within the eutrophic, impacted zone. P. paludosus, in particular, was rarely collected in high-P areas.

302

R.S. King and C.J. Richardson

Fig. 11.4 The effect of differing levels of taxonomic resolution on the environmental signal provided by the abundance of the family Chironomidae. Scatterplots of density (no. m−2) for all taxa in (a) Chironomidae, (b) two representative genera (Dicrotendipes and Tanytarsus), and (c) all six species within the genus Tanytarsus are shown as a function of distance from canal inflow structures in Water Conservation Area (WCA) 2A. Symbols indicate impacted, transition, and reference landscape zones

11 Macroinvertebrate Responses to a Gradient of Long-Term Nutrient Additions

303

Fig. 11.5 Invertebrate species accumulation curves (mean ± 1 SD) for impacted, transition, and unimpacted zones. Jackknife estimates of asymptotic richness (Sjack) are shown next to each curve

Taxa that were either primarily algivores or suggested to at least be partially dependent upon periphyton all showed subsidy–stress patterns along the P gradient. Additionally, two predaceous groups, Odonata and Hirudinea, exhibited this same response (Fig. 11.7f, k). However, two other predominantly predaceous groups, Hemiptera and Coleoptera, responded favorably to high levels of P (Fig. 11.7d, h). The only other major taxon to respond positively to high P was Isopoda, represented exclusively by Caecidotea sp. (Fig. 11.7c). This detritivorous taxon became most abundant in dense stands of Typha with large quantities of decaying CPOM. Finally, Trichoptera – represented by three families and at least five different species (Table 11.2) – was the only coarse taxon to show a stress response to P enrichment (Fig. 11.7l). Diversity responses mimicked the general pattern of biomass. Species density and richness showed subsidy–stress responses to P; however, these relationships were not statistically significant (Fig. 11.8a, b). However, species density clearly increased at intermediate levels of P when compared with low-P clusters (Fig. 11.8a). Species density and richness were variable in the high-P zone but tended to show a stress response above intermediate-P levels. Aside from vegetation, several biotic variables hypothesized to be determinants of invertebrate biomass also were significantly related to P. Food quality of periphyton, expressed as C:N ratio, decreased linearly with increasing sediment TP

304

R.S. King and C.J. Richardson

Fig. 11.6 Response of invertebrate assemblage biomass to (a) distance from canal inflow structures and (b) sediment total phosphorus (TP). Error bars indicate ±1 SE. Locations of plot-clusters used in the temporal study (C1, C4, C6) are indicated in (a)

11 Macroinvertebrate Responses to a Gradient of Long-Term Nutrient Additions

305

Fig. 11.7 (a–l) Response of biomass of 12 coarse-taxonomic invertebrate groups to sediment TP. Error bars indicate ±1 SE

(Fig. 11.9a). Despite the apparent increase in potential food quality with P, metaphyton cover decreased markedly above 1,200 mg kg−1 sediment TP (Fig. 11.9b). Invertivorous fish density and biomass were not significantly related to P, but both exhibited subsidy–stress relationships that were nearly significant (second-order polynomial, r2 = 0.38, p = 0.055 and r2 = 0.31, p = 0.120, respectively). Although it appeared that invertivorous fish density and biomass were greatest at intermediate levels of P, subsequent decreases at high levels of P were much less apparent than those exhibited by invertebrate biomass. Since small fish density showed the best relationship, it was retained as a potential predictor of invertebrate composition in the multivariate analysis.

306

R.S. King and C.J. Richardson

Fig. 11.8 Response of invertebrate (a) species density and (b) species richness to sediment TP. Error bars indicate ±1 SE

11 Macroinvertebrate Responses to a Gradient of Long-Term Nutrient Additions

307

Fig. 11.9 Response of (a) periphyton C:N ratio and (b) metaphyton cover to sediment TP. Error bars indicate ± 1 SE

11.3.3

Determinants of Assemblage Composition

Ordination of invertebrate species composition revealed distinct separation of plots among impact zones (Fig. 11.10a). The primary axis was a landscape-scale gradient significantly associated with spatial/abiotic variables such as Canal, P, and interquartile

308

R.S. King and C.J. Richardson

Fig. 11.10 Nonmetric multidimensional scaling ordination of fine-scale invertebrate assemblage composition (three-dimensional configuration; only two dimensions shown – third dimension added to reduce stress but it was not significantly related to abiotic or biotic variables). Significant (a, c) biotic and (b, d) abiotic (including canal) vectors are shown in relation to (a, b) plots and (c, d) indicator/important taxa in ordination space (see Table 11.2 for species codes; Tables 11.1

11 Macroinvertebrate Responses to a Gradient of Long-Term Nutrient Additions

309

Fig. 11.10 (continued) and 11.3 for abiotic and biotic variable codes). Magnitude of vector correlations (r) is shown in parentheses next to variable codes. Symbols indicate membership of plots to landscape zones. Codes for variables not defined previously: Veg(float) cover of unrooted floating vegetation; CNperi C:N ratio of periphyton

310

R.S. King and C.J. Richardson

range of water depth (IQR[Depth]) (Fig. 11.10b), and, in response, biotic variables such as cover of invasive vegetation (Typha, Mikania, Sarcostemma), cover of Cladium, C:N ratio of periphyton(CNperi) and abundance of invertivorous fish (Fish) (Fig. 11.10a). The second axis was a local, or fine-scale gradient driven primarily by water depth (freq. <−10 cm, xDepth1y, xDepth), but also related to soil chemistry (N, Na). Metaphyton, cover of floating vegetation, and cover of slough macrophyte species (Eleocharis, Utricularia, Nymphaea) all were related to this secondary axis.

Table 11.4 List of invertebrate taxa significantly associated with impacted, transition, or unimpacted landscape zones, as estimated using Indicator Species Analysis (INSPAN) Impacted

Transition

Anopheles sp. 1a

29.2

Caecidotea sp.d

52.9

Chironomus stigmaterusc

Unimpacted 49.3

33.8

Aphaostracon pachynotusb Bratislavia unidentatae Caenis diminutaf

Chlamydotheca unispinosah Entomobrya sp.k Cyclopoidai

36.8

Calanoidai

48.8 47.0

Desmopachria sp.l

30.3

Enochrus spp. (larvae)l

54.6

Zonitoides arboretumb

25.2

Goeldichironomus holoprasinusc Physella sp.b

66.8

Planorbella duryib

26.0

Polypedilum trigonusc

43.2

Scirtes sp.l Tanytarsus sp. Fc Uranotaenia sapphirinaa

35.2 27.4 41.6

37.8

Ablabesmyia rhampe gr. sp.c Beardius truncatusc

23.1

42.9

32.2

Bezzia/Palpomyia gr. sp. 2g Cernotina sp.j

Dero furcatae Dicrotendipes modestusc Dicrotendipes simpsonic Haemonais waldvogelie Laccophilus spp. (larvae)l Kiefferulus duxc

40.5 27.9

Cladotanytarsus sp.c Cypretta brevisaeptah

34.8 36.1

51.6

Enallagma civilem

60.7

54.4

16.7

38.7

Nanocladius alternantheraec Nilothauma sp.c

33.9

Oecetis sp. Ej

30.8

Laevapex peninsulaeb Micromenetus dilatatusb Pseudochironomus sp.c

45.8

Oxyethira sp.j

52.2

53.4

Parachironomus alatusc Parakiefferiella sp. Cc

24.6

Paraponyx sp.n Paratanytarsus sp. Bc Physella cubensisb Polypedilum halteralec Polypedilum sp. Ac Procladius sp.c Spongilla sp.o Tanytarsus sp. Rc

30.6 54.4 44.0 25.3 28.0 22.2 22.7 84.5

44.5 43.6

28.9

46.3

41.7

27.8

77.5

Indicator values (IVs; % of perfect indication) are shown next to each taxon. All taxa shown had IVs with p ≤ 0.0002 (Bonferroni-corrected p ≤ 0.05) Class/order – aDiptera:Culicidae; bGastropoda; cDiptera:Chironomidae; dIsopoda; eOligochaeta; f Ephemeroptera; gDiptera:Ceratopogonidae; hOstracoda; iCopepoda; jTrichoptera; kCollembola; l Coleoptera; mOdonata; nLepidoptera; oPorifera

11 Macroinvertebrate Responses to a Gradient of Long-Term Nutrient Additions

311

Centroids of species identified as indicators of the impacted zone were, accordingly, ordinated on the eutrophic end of nMDS Axis 1, near the ends of vectors for P and Typha (Fig. 11.10c, d; see codes in Table 11.2). At the oligotrophic end of the gradient, taxa were separated along Axis 2 according to vegetation and hydrological affinities – a continuum of species restricted primarily to deep-water slough habitats (e.g., Parakiefferiella sp. C, Oxyethira sp., Paratanytarsus sp. B, Cladotanytarsus sp.) to those almost exclusively found in shallower, dense stands of Cladium (e.g., Enallagma civile, Beardius truncatus, Oecetis sp. E). The two large-bodied decapods, P. paludosus and P. fallax, occupied slightly different locations in species space, with P. paludosus bordering the transition and unimpacted zones, while P. fallax was proximate to the center of the ordination. INSPAN analysis indicated that a number of these species had significant associations with specific impact zones (Table 11.4). The reference zone had the greatest number of indicator species (21) – over half of these were members of the family Chironomidae (Diptera). Three of the five trichopteran taxa collected also were indicators of this zone. Several species belonging to Gastropoda, Ostracoda, Lepidoptera, and Porifera were also sensitive and primarily found here. Several of the best indicators of the transition zone were either gastropods or naidid oligochaetes (Table 11.4), primarily grazers or collectors of periphyton. Impacted-zone indicators included many detritivorous taxa, such as Caecidotea sp., Scirtes sp., and filter-feeding Culicidae (Diptera), Anopheles sp. 1, and Uranotaenia sapphirina. Two chironomids often associated with organic pollution, Chironomus stigmaterus and Goeldichironomus holoprasinus, also were found mostly in this eutrophic region of the landscape.

11.4 11.4.1

Discussion Biomass Response to P

Results from the spatial study support the hypothesis that invertebrate assemblage biomass is resource limited, and this limitation is relaxed with P enrichment. The subsidy–stress model (sensu Odum et al. 1979; Fig. 11.1) served admirably as a theoretical framework for predicting assemblage-level responses along the P gradient (Fig. 11.6). We anticipated that a myriad of factors, mostly linked to changes in landscape pattern (i.e., vegetation), would cumulatively act as a stressor to standing stocks in high-P, eutrophic areas relative to areas of intermediate-P enrichment. Indeed, biomass showed a significant subsidy–stress relationship with P, and was lower in a high-P region of the wetland than an intermediate-P area on three of four collection dates. Moreover, this subsidy–stress pattern was evident for most of the major taxa collected. However, what factors were directly contributing to observed patterns along the P gradient?

312

R.S. King and C.J. Richardson

King (2001) hypothesized that changes in periphyton abundance and nutrient content would be an important direct determinant of invertebrate biomass because it has been suggested to be a significant pathway in food webs in wetlands (e.g., Murkin 1989; Keough et al. 1996; D.A. Wrubleski and N.E. Detenbeck, unpublished data; Wissinger 1999). Patterns of metaphyton cover and C:N ratios, across the landscape and through time, lend credence to this hypothesis. Metaphyton was reduced to very low cover in high-P areas during the spatial study, yet remained relatively high at intermediate levels of enrichment (Fig. 11.9b). Meanwhile, C:N ratios steadily declined with increased P, implying greater nutritional value of periphyton and its detritus (e.g., Sterner and Elser 2002). Consequently, localities of intermediate enrichment had relatively high quantities of periphyton but also higher protein (inferred) content than low-P areas, a fact that may have contributed to a subsidy effect for invertebrate standing stocks. Top-down regulation of invertebrates by invertivorous fish is a mechanism that can limit invertebrate biomass accumulation (e.g., Hairston et al. 1960; Oksanen et al. 1981). Turner et al. (1999) suggested that greater biomass and densities of small fish in eutrophic than oligotrophic areas of the Everglades may explain their finding of no increase in invertebrate biomass between these same two nutrient regimes. Although invertivorous fish abundance clearly is an important consideration, findings from this study do not provide sufficient evidence to imply that predation was the primary factor limiting invertebrate biomass in high-P areas (although it may have played a role in structuring composition). In the spatial study, biomass of invertivorous fish followed a similar subsidy–stress P–response curve to that of invertebrate biomass, while densities of small fish were generally similar or slightly lower at high-P locations than in the intermediate-P region of the gradient. In a temporal study (not reported here), fish abundance followed this similar pattern, with C1 and C4 exhibiting similar densities and biomass of fish, but greater than C6 (King 2001). Thus, it did not appear that invertebrate production was accumulating as invertivorous fish biomass to a degree that would explain patterns in invertebrate assemblage biomass. The fact that Turner et al. (1999) only sampled low- and high-P habitats, while not sampling the intermediate-P zone, may have contributed to their conclusion.

11.4.2

Diversity Relationships with P Enrichment

Estimation of species diversity (number of species) is dependent upon two important factors: sampling area (Arrhenius 1921) and a “sampling effect” related to the number of individuals sampled (Preston 1948; May 1975) – these factors are particularly influential at small spatial scales (Larsen and Herlihy 1998). Increasingly, scale of measurement has become recognized as one of the most important factors in relating diversity to nutrient or productivity gradients because of the confounding dependency of community density upon productivity (e.g., Oksanen 1996; Waide et al. 1999; Weiher 1999). In this study, the similarity between estimates of species

11 Macroinvertebrate Responses to a Gradient of Long-Term Nutrient Additions

313

density (mean plot-level α-diversity within clusters) and species richness (clusterlevel α-diversity) along the P gradient suggested that coarse-scale richness was at least partially due to fine-scale species density – clusters with higher mean densities of species tended to have higher total numbers of species (Fig. 11.8). Both of these diversity measures followed a subsidy–stress relationship with P enrichment in the spatial study, and predominantly subsidy–stress patterns among enrichment categories in the temporal study. Although not statistically significant in the spatial study, these patterns imply that the unimodal productivity–diversity relationship, often used as a model for plant communities, may also be a useful model for wetland invertebrate assemblages (Rosenzweig and Abramsky 1993). At larger spatial scales, however, species accumulation curves and jackknife estimates of richness within impact zones (surrogates for nutrient categories) indicated that the cluster-scale relationship between richness and P was partially confounded by community density. This was further illustrated through the examination of species accumulation within low-, intermediate-, and high-P clusters through time. Here, differences in richness between low- and intermediate-P areas appeared to be exaggerated at smaller scales (plots and clusters). Importantly, these curves illustrated that lower invertebrate biomass in the oligotrophic, unimpacted landscape resulted in a greater incidence of missing “rare” species at smaller scales relative to enriched locations (Fig. 11.5). A conservative interpretation of these results may be that a weak subsidy–stress relationship exists between diversity and nutrients in the Everglades, but the magnitude of this relationship is scale dependent. Results of this study conflict with those of Rader and Richardson (1994) regarding the response of species diversity to P enrichment. Sampling at many of the same locations used in this study, they concluded that P enrichment resulted in a dramatic subsidy effect for invertebrate species richness. A number of factors may have contributed to the discrepancy between studies. First, their sampling was limited to open-water (short-emergent, floating, and submergent vegetation only) patches along the P gradient – a habitat that was and is rare in high-P areas of the gradient. These patches, while likely harboring much periphyton, also may have represented an uncharacteristic refugium among the dense stands of invasive vegetation in such locations. Moreover, this stratification removed potentially relevant heterogeneity in the landscape and subsequently missed taxa that were found associated with other types of vegetation in this study. Second, they identified 137 invertebrate taxa, in aggregate, which was approximately 1/2 of the number identified in this study (272 taxa identified in spatial and temporal studies combined; King 2001). This may have been partly due to a sampling effect, as they identified approximately 11,000 individuals compared with over 144,000 in King (2001). Third, and directly related to the disparity of total taxa between studies, they used a dip net with 2-mm mesh compared with the standard 0.5-mm mesh used in this study and typically used for macroinvertebrate studies (e.g., Barbour et al. 1999). Numerous taxa, particularly many of the sensitive, indicator taxa from the unimpacted landscape (e.g., Chironomidae) were too small to be reliably

314

R.S. King and C.J. Richardson

collected with such a coarse device. Collectively, these three factors probably contributed to the differences between the Rader and Richardson (1994) findings and those of this study. Nevertheless, if limited to local-scale, open-water habitats, their conclusion that P additions increase species richness (i.e., species density) is probably robust. However, our results imply that this pattern should not be expected when considering the broad vegetation pattern of wetland landscapes impacted by nutrients.

11.4.3

Determinants of Assemblage Composition

Results from ordination implied that invertebrate assemblages along the P gradient were organized by two spatio-environmental dimensions (1) a coarse/landscapescale dimension best explained by distance from canal, sediment TP, variability of water depth, periphyton C:N ratio, and broad-scale vegetation pattern and (2) localscale dimension related to mean water depth, frequency of severe dry down, density of small invertivorous fish, cover of metaphyton, sediment Na, and fine-scale vegetation pattern. Since effectively summarized as two dimensions, this suggests that wetland invertebrate communities may be assembled in a predictable way in response to nutrient enrichment. However, the great diversity of significant pure– partial relationships between candidate spatial, abiotic, and biotic predictor variables and invertebrate composition suggests that this assembly is dependent upon numerous factors that may act independently or synergistically, vary among levels of nutrient enrichment, and vary across the spatial hierarchy (King 2001). Vegetation (expressed as cover-weighted species composition) was consistently the most important determinant of invertebrate species composition, regardless of scale or nutrient status. Reasons for this may be numerous, but may be best summarized simply: vegetation forms the physical template for other biota in wetlands. In this study, vegetation integrated numerous spatial and abiotic sources of variation directly attributable to the coarse/landscape-scale P gradient, yet manifested these sources of variability as both local- and coarse/landscape-scale variation in its compositional pattern (King et al. 2004; see Chap. 9). Consequently, associations between invertebrate species and specific plant assemblages were reflected on these same scales across the landscape. For example, the chironomid Beardius cf. truncatus was almost always collected in plots hosted by the macrophyte Cladium – distribution of Cladium varied both at local, fine scales (topographical variation) and at a coarse/landscape scale (associated with P enrichment). Vegetation also played a direct role in controlling cover of metaphyton and other periphyton that may have been important to many invertebrates as food (e.g., McCormick et al. 1998). Similarly, small fish densities were highly correlated to vegetation pattern, a trend that may have had predator–prey implications for invertebrates (Batzer and Resh 1991; Jordan 1996). Subsequently, vegetation explained variation directly related to its function as habitat to certain invertebrate taxa, and indirectly related to other biotic determinants such as food resources and predation.

11 Macroinvertebrate Responses to a Gradient of Long-Term Nutrient Additions

315

Therefore, vegetation may be expected to be the primary, direct factor driving invertebrate assemblage responses to nutrient enrichment in wetlands. Despite the strong influence of vegetation, several other variables explained residual variation in composition that could not be explained by vegetation alone (see King 2001). Considering sediment chemistry, both Na and P were significantly related to invertebrates on a fine scale (King 2001). The influence of Na was most apparent within the impacted landscape zone. Although predominantly a P gradient, cations such as Na have been shown to be elevated in canal water that enters the wetland (Craft and Richardson 1997). Insects, in particular, are sensitive to salinity and have an extremely limited distribution in estuarine and marine habitats (Williams and Feltmate 1992). Although no taxon typically associated with estuarine environments appeared related to Na, its small but significant correlation to the invertebrate assemblage suggests that it may have played a minor role in localized differences in composition. Similar to P, distance from canal was also an important correlate of composition. Because it was the source of the P gradient, canal had a direct effect on sediment TP, and thus an indirect effect on most biotic variables (King 2001; King et al. 2004). However, even after variation from these and all other variables was removed, canal remained a significant correlate of composition; in fact, the magnitude of its partial correlation was second only to vegetation. This residual dependency of invertebrate composition on canal mimics the same mysterious dependency exhibited by vegetation. In King et al. (2004), several plausible explanations are provided for this phenomenon; these may also apply to invertebrates here. Regardless of the mechanism, a safe conclusion is that the canal–levee system of south Florida plays a significant role in changes observed in structure of the Everglades ecosystem, and its influence is reflected in both vegetation and macroinvertebrate levels of organization. Spatial differences in hydrology were implied to be important to invertebrates. The most obvious pattern was related to local differences in water depth, particularly in the reference and transition zones. Here, open-water slough habitats were typically situated lower on the landscape than adjacent stands of Cladium. However, local pattern in vegetation was not perfectly related to water depth since both were significant pure–partial correlates of composition (King et al. 2004). Thus, hydrology probably also played a direct role in organizing species. For example, many taxa were consistently associated with plots with a high frequency of severe dry down (depth <−10 cm). This may have indicated a greater tolerance to drought resistance or that these taxa were more effective at recolonizing hydrologically unstable environments (Wiggins et al. 1980). Consistent with patterns observed in assemblage biomass, periphyton variables (metaphyton cover and C:N ratio of periphyton) were significantly related to composition. Metaphyton, particularly the calcareous form characteristic of low-P areas of the Everglades, has been indicated to be important not only as a food resource but also as a unique habitat to invertebrates (e.g., Browder 1982). Indeed, several indicator taxa such as Paraponyx sp., Tanytarsus sp. R, Cypretta brevisaepta, Parakiefferiella sp. C, and Cladotanytarsus sp. were almost always found associated

316

R.S. King and C.J. Richardson

with mats of calcareous periphyton in the unimpacted zone. Loss of these calcareous mats has been experimentally documented in response to P additions (e.g., Walker et al. 1989; Richardson et al. 2000; see Chap. 25), thus it can be inferred that loss of these mats may have been responsible for the reduction or elimination of some taxa deemed indicators of the low-P, unimpacted zone. The importance of calcareous metaphyton in structuring invertebrate assemblages is an area of research in need of direct investigation in the Everglades. Density of invertivorous fish was indicated to be a determinant of invertebrate assemblage composition. Although fish density or biomass did not appear to explain the subsidy–stress pattern of biomass along the P gradient, fish may have been structuring composition through selective predation of particular invertebrate taxa (top-down control). Alternatively, fish density may have been associated with greater numbers of particular taxa that were more readily available as food (bottomup control). For example, high densities of surface-feeding Gambusia holbrooki (mosquitofish) and Heterandria formosa (least killifish) were found among floating macrophytes in high-P areas; here, mosquito larvae such as Uranotaenia sapphirina, Mansonia titillans, and Coquillettidia perturbans were often in great abundance. Thus, it is difficult to know whether fish were influencing assemblage composition through predation, or responding to specific invertebrate assemblages. Both mechanisms are likely to be tightly coupled, and the extent to which one is predominant is quite likely to be dependent on an interaction among multiple factors through time and space (Batzer and Resh 1991; Jordan 1996).

11.4.4

Implications for Bioassessment

Two schools of thought persist in the scientific community regarding the best approach for bioassessment in streams (Reynoldson et al. 1997). The first school is the multimetric approach, an assessment framework that relies on an aggregated index of biological integrity (IBI; sensu Karr 1981) composed of multiple “metrics” to score sites. By definition, metrics are attributes that represent key elements of structure or function of biotic assemblages, and show a monotonic response to increasing levels of human influence or specific environmental stressors (Barbour et al. 1995). Typically, metrics are developed from one of four categories (1) taxonomic richness, (2) taxonomic structure, (3) feeding ecology, and (4) tolerance/ intolerance. Results from this study suggest that three of four categories (richness, structure, and feeding) may not be consistently effective for assessing detrimental nutrient enrichment in wetlands. The primary reason is that the vast majority of invertebrate assemblage attributes from these categories exhibited unimodal responses to P enrichment. Such responses are problematic for multimetric indexes, as sites at opposite ends of an environmental continuum would be considered equivalent for a metric responding in a unimodal fashion. The second school of bioassessment, the multivariate approach, may circumvent this unimodal-response problem. This approach relies on the identification of reference

11 Macroinvertebrate Responses to a Gradient of Long-Term Nutrient Additions

317

conditions (just as in the multimetric approach) to characterize the natural range of variability expected in minimally impacted aquatic habitats. However, rather than extracting coarse attributes from the assemblage to use as metrics, the multivariate approach uses taxonomic compositional data and multivariate data analysis techniques to assess a test site relative to a collection of reference sites (e.g., Reynoldson et al. 1997; Hawkins et al. 2000). On the basis of the findings in this study, this approach may be a more effective than multimetrics for assessing nutrient impacts. Changes in species-level composition in response to P were evident, as plots were clearly sorted according to impact zones in nMDS ordination space. Distance-based multivariate approaches like nMDS and Mantel tests are insensitive to the shape of species responses to environmental gradients, capturing a multitude of monotonic or unimodal responses along such gradients and expressing them as increased dissimilarity (β-diversity) relative to the reference condition. The central theme of these findings is that simple monotonic patterns in relation to nutrients are not likely to emerge for many of the invertebrate assemblage attributes commonly used in stream bioassessment today (e.g., species richness). This is not to say that the multimetric approach will not work for wetland systems. However, it implies that compositional metrics based on family-, genus-, or specieslevel taxonomy may be more effective than those based on coarse taxonomy since few wetland invertebrate groups are inherently sensitive to nutrient-related stressors (e.g., dissolved oxygen).

11.5

Conclusions and Lessons for Restoration

Our results suggest that effectiveness of subsampling depended more upon the minimum number of individuals retained than minimum area or proportion of the sample picked. Fixed-area subsamples were generally less efficient than fixed counts, with 200- and 300-individual fixed counts resulting in significantly greater assemblage–environment relationships and much higher accuracy in detecting impairment than 10% fixed area, despite averaging similar numbers of individuals. The greatest improvement with increasing subsample size was observed between fixed counts of 100 and 200 individuals; detecting impairment, in particular, was not markedly improved with subsample sizes >200 individuals. Supplementing subsamples with a LR search resulted in only very slight improvements in assemblage–environment relationships, but was effective in improving prediction accuracy, particularly for family-level data. However, family-level assemblage– environment relationships and abilities to detect impairment were inferior to genusand species-level data, regardless of subsample size. Species-level data performed best, primarily because of the large proportion (>20%) of total species belonging to Chironomidae. The potential importance of Chironomidae to wetland bioassessment was further revealed through an evaluation of a tiered-taxonomic approach, which showed that non-Chironomidae family-level data tiered with species-level Chironomidae data produced results very similar to those obtained using genus- or

318

R.S. King and C.J. Richardson

species-level data exclusively. Our results suggest that fixed counts ≥200 or integrated fixed-area/fixed-count approaches that consistently obtain a minimum of 200 individuals should be considered as minimum subsample sizes for wetlands. We additionally advocate LR searches and recommend genus- or species-level taxonomy, particularly for the Chironomidae. From a bioassessment perspective, the most significant conclusion may be that wetland invertebrate assemblages are sensitive to nutrient enrichment, and that they respond in predictable ways. However, many of the usual approaches and assemblage attributes used as metrics are not conducive to developing indices of nutrient impairment – subsidy–stress relationships using coarse levels of taxonomy, feeding ecology, or diversity were not satisfactory for this purpose. Rather, our results suggest that fine levels of taxonomic resolution (i.e., genus- or species-level data) may be necessary for bioassessment to be accurate. Compositional metrics that used species-level data were the most sensitive to P enrichment, and their responses were mediated largely through vegetation and periphyton. This begs the question: if invertebrate assemblage organization is tightly coupled to primary-producer response to nutrients, why not assess indicator groups such as periphyton or macrophytes instead? Clearly, periphyton and macrophytes are excellent indicators of nutrient status, and should be considered for wetland bioassessment. However, the appropriateness of an indicator group may be most dependent upon temporal scales of interest. In wetlands, microbes are the first to respond to enrichment, followed by periphyton, invertebrates, and finally macrophytes. Invertebrates have an advantage in bioassessment because of their dependence both on levels that respond faster (microbial, periphyton) and those that respond more slowly (vegetation) to pollution. Their intermediate position along this continuum integrates the effects of both episodic and cumulative stressors in aquatic systems. Indicators that respond quickly may also recover too quickly for detection if pollution is episodic, while slower indicators may not respond or recover quickly enough if water quality is cumulatively degraded or subsequently restored. Ultimately, the decision on the most appropriate indicator group or groups to use will depend on both the spatial and temporal scales of interest; our data suggest that invertebrate assemblages may be robust indicators in many situations. Anthropogenic inputs of nutrients into the Everglades are significant and ongoing, and our results suggest that the implications of such inputs go beyond just changes in primary productivity or proliferation of weedy species typically associated with enrichment. Rather, changes in pattern (e.g., arrangement of patches of vegetation) have profound effects on process (e.g., population dynamics through time and space), which subsequently affect pattern. It can be inferred that the spatial ecology of higher organisms, such as invertebrates, fish, birds, etc., can be significantly affected by these alterations in a variety of ways that may depend largely upon landscape connectivity and critical scales in their individual life histories. The field of landscape ecology has already begun to address many of these scaling issues for terrestrial wildlife (e.g., spotted owl); such management approaches could be extended to aquatic systems as well. Importantly, this research illustrates

11 Macroinvertebrate Responses to a Gradient of Long-Term Nutrient Additions

319

an important deficiency in aquatic research specifically that greater attention needs to be given toward evaluating the implications of spatial pattern and scale. Such research will be critical for the successful management and restoration of wetland and other aquatic ecosystems. Finally, one of the central conclusions of this research is that invertebrate assemblages are organized through a myriad of spatial, temporal, abiotic, and biotic pathways that are a product of differing nutrient levels and processes that operate at a variety of scales across the spatial hierarchy. From an ecological perspective, perhaps the most significant finding was the hierarchical spatial relationship between landscape pattern and invertebrate assemblage composition, and how this relationship varied among differing nutrient regimes. The finding that canals were a major influence on the macroinvertebrate community at the larger scale suggests that natural populations are highly influenced by these man made structures.

12

Historical Changes in Water Quality and Vegetation in WCA-2A Determined by Paleoecological Analyses Sherri R. Cooper, Michelle Goman, and Curtis J. Richardson

12.1

Introduction

The Florida Everglades is the largest subtropical peatland in North America. This extensive wetland supports a unique pattern of vegetation and wildlife. Before extensive anthropogenic influence of this nearly one million hectare freshwater wetland, dominant vegetation types included Cladium jamaicense Crantz (sawgrass) marshes, wet prairies, slough aquatic communities, and tree island communities (Davis 1943; Loveless 1959). After almost 5,000 years of relative environmental stability, the Florida Everglades has undergone major changes because of human intervention since the mid-nineteenth century (Gleason and Stone 1994). Changes in hydroperiod and water levels (since 1900) and recent nutrientenriched agricultural drainage (last several decades) have been implicated in changes in plant species composition and water quality through time, especially in Water Conservation Areas (WCAs) of the northern Everglades that receive runoff from the Everglades Agricultural Area (EAA). These changes include greater abundance of cattail, fewer tree islands (Chap. 8), and shorter hydroperiods. However, little is known about how early canal dredging and levee building (i.e., 1900–1950) affected the vegetation and water quality in comparison to more recent agricultural and other anthropogenic nutrient inputs (post-1950). To manage or “restore” the Everglades, it is important to know predisturbance vegetation composition and water quality, as well as the relative impacts of different factors affecting changes in the Everglades through time. One source for historical information on these changes lies within the stratigraphic record of peat and sediment accumulation within the Everglades. These data are accessible through the use of paleoecological methods. Paleoecology examines relationships between past organisms and the environment in which they lived. A clearer picture of the history of an ecosystem can often be obtained by combining various types of sedimentary evidence, including biological, chemical, and geological contributions (Birks and Birks 1980). This historical evidence can be compared with modern analogues to “reconstruct” past ecosystems. Diatom analysis is a useful paleoecological tool for studying water quality changes in aquatic environments, including wetlands. Diatoms make up one of the

321

322

S.R. Cooper et al.

three dominant algal groups in the periphyton of the Everglades, the other two being green algae and cyanobacteria. Diatoms have rapid immigration and replication rates, which allow them to respond quickly to changes in the aquatic environment. Diatoms are particularly useful for paleoecological research because their silica shell (frustule), which is often preserved in the stratigraphic record of sediments and soils, can be used for species-level identification. Changes in diatom composition over time have been interpreted as evidence of eutrophication (increased nutrient concentrations) in many paleoecological studies (e.g., Stoermer et al. 1992; Anderson et al. 1993; Cooper 1993, 1995; Dixit et al. 1993; Bennion et al. 1996). The use of diatoms in paleoecology has become an attractive tool to ecosystem managers for obtaining natural background or reference conditions for lakes (Smol 1992; Dixit and Smol 1994). A few paleoecological studies of diatoms have been completed in Everglades Water Conservation Area 2A (WCA-2A) (Slate 1998; Cooper et al. 1999; Slate and Stevenson 2000). Pollen analysis is another paleoecological tool frequently used to reconstruct changes in vegetation and community composition over time. Several palynological studies have been completed in the Everglades, including WCA-2A (Bartow et al. 1996; Willard 1997; Willard et al. 1997; Willard and Weimer 1997; Willard et al. 2004). The pollen and diatom results of this current study are compared with completed calibrations of pollen and diatoms to nutrient gradients in WCA-2A (Jensen et al. 1999; Cooper et al. 1999), and to other previous and ongoing paleoecological research in the Everglades. The objective of the paleoecological study discussed in this chapter is to determine historical changes in water quality (and quantity) and vegetation types over the past 100–500 years in relation to human-induced changes to the hydrology and nutrient inputs of the Everglades using paleoecological analyses of soil cores from WCA-2A. Samples have been analyzed for diatoms, pollen, total phosphorus (P), percent nitrogen (% N), percent organic carbon (% C), calcium and sodium ions (Ca and Na, respectively), and biogenic silica (BSi). These geochemical parameters provide an indication of loading of these elements to the system at different sites, as well as preservation of these elements within the soils. Concentrations of P, % N, and % C reflect both nutrient inputs and productivity within the system. BSi is primarily a measure of diatom production, but also includes chrysophyte cysts and sponge spicules, all of which are composed of amorphous silica. Biogenic silica can be measured separately from mineral silica (Conley 1988).

12.2

Historical Background

Europeans first landed in Florida in the sixteenth century, but Florida remained relatively unchanged for hundreds of years after. In 1821, the Adams-Onis treaty was ratified, by which Spain yielded Florida to the United States (Blake 1980). Florida became a state in May of 1845 with a population of approximately 50,000, about

12 Historical Changes in Water Quality and Vegetation in WCA-2A

323

90% of which still lived north of Gainesville (Snyder and Davidson 1994). It was not until 1905, when a “Board of Drainage Commissioners” was created by Governor Broward, that reclamation and drainage of the Everglades began in earnest (Blake 1980). By 1917, four channels were dug from Lake Okeechobee to the Atlantic Ocean. Construction was limited by funding and hampered by hurricanes in the 1920s, but by 1931 more improvements were made, for a total of 440 miles of canals excavated and 47 miles of levees built around southern Lake Okeechobee (Light and Dineen 1994). Between 1900 and 1930, the population along the lower East Coast of Florida increased almost tenfold. Land development was essentially abandoned in the 1930s and 1940s because of funding problems (Blake 1980; Light and Dineen 1994). During the 1930s and 1940s, conservationists began lobbying for protection of the southern Everglades. Consequently, the Florida legislature purchased land to create the Everglades National Park (ENP) in 1947, which included approximately 21% of the historic freshwater ecosystem (Blake 1980). Just after formation of the ENP, several major hurricanes struck south Florida between 1947 and 1948, flooding millions of acres of “reclaimed” land. The U.S. Army Corps of Engineers published the “Central and Southern Florida Project for Flood Control and Other Purposes,” a comprehensive plan to prevent future flooding in south Florida, in 1948. The Central and Southern Florida Flood Control District created under this plan soon began a series of changes to the Everglades, including the formation of the WCAs and the EAA. Construction and improvements continued from 1952 through 1979, and to the present. Between 1950 and 1990, human population increased in Dade, Broward, Palm Beach, and Monroe counties (southern Florida) from 750,000 to close to 4 million. In 1983, Florida Governor Bob Graham initiated the “Save our Everglades” program to protect and restore the Everglades (Light and Dineen 1994).

12.3 12.3.1

Methods Study Site

Soils of the Everglades are mainly organic peat with depths of 0.3–4 m (Gleason and Stone 1994). Surface water generally flows from north to south over the low relief landscape. WCA-2A is bordered by levees on all sides, and includes a small area of the original Everglades (Fig. 12.1). WCA-2A covers approximately 574 km2 of peat wetland, which receives nutrient-enriched water through the S-10 gates of the Hillsboro Canal along its northern border. A P gradient exists in WCA-2A that generally follows water flow (Reddy et al. 1991; Craft and Richardson 1993a,b; Qualls and Richardson 1995; Vaithiyanathan and Richardson 1997a). Soil cores were collected at one site in the P-enriched northern area of WCA-2A near the

324

S.R. Cooper et al.

Fig. 12.1 Map of Water Conservation Area 2A(WCA-2A) in the Florida Everglades (for location of WCA-2A, see maps in other chapters). The sites where soil cores were collected are labeled according to the core name within WCA-2A

Hillsboro Canal and at two more southerly locations in WCA-2A, in areas that are currently not as greatly affected by nutrient enrichment (Fig. 12.1).

12.3.2

Calibration Sets

Among the first steps in a paleoecological study are the development of reference collections of indicators such as pollen and diatoms and the determination of how these indicators are related to geological, chemical, and physical parameters. Surface sediments and soils collected at 31 sites throughout WCA-2A for a P-gradient seed bank germination study (Vaithiyanathan et al. 1997) were analyzed for fossil indicators including diatoms and pollen, and geochemical parameters (P, % C, % N, BSi, and Ca). Calibration models were developed that relate organisms to environmental parameters. These models were developed using weighted-averaging (WA) regression and calibration (with classical deshrinking) (Line et al. 1994). Optima and tolerances for each taxon were calculated along P, %N, and BSi gradients (for diatoms) and P gradients (for pollen). These relationships can then be used to reconstruct historical environmental parameters through time from fossil species identified in dated sediments or soils assuming that the same taxa occur in the past.

12 Historical Changes in Water Quality and Vegetation in WCA-2A

325

The calibration models developed for diatoms and pollen in WCA-2A are described in Cooper et al. (1999) and Jensen et al. (1999).

12.3.3

Core Collection

The soil cores were collected with a modified Livingstone piston coring device (Wright et al. 1965, 1984). The coring device was fitted with a custom-built serrated “cutting shoe” to aid in peat soil collection. The soil cores were 4 in. in diameter. Before being subsampled, the cores were X-rayed at Palms West Hospital in Loxahatchee, Florida, with a phototiming Philips Medio 50CP at 70 kV and 9–18 mA s (at the chest board). Coring tubes were cut open, and the peat soil was subsampled at 2 cm intervals. Overlying flocculant sediment composed primarily of periphyton detritus was separated and stored in a separate bag for analyses. This sediment varied in thickness at the different sites (Fig. 12.2). For the remainder of the data presented, depths for the peat soils were started at “0” after subtracting off the bulk of the recent periphyton detritus from the surface of the peat soil cores. Core EN1 peat soil depths remained unchanged because no significant amount of sediment was present. Sediment and soils were stored in a dark 4°C cold room in airtight plastic bags.

Fig. 12.2 Characterization of the five soil cores collected in WCA-2A of the Florida Everglades in May 1996. Latitude and longitude of each site are listed below the cartoon of each core. Depths in centimeters are shown along the left side of each cartoon

326

12.3.4

S.R. Cooper et al.

Dating of Soils

Soils were dated using a combination of methods, including 210Pb activity (within the range of 100–150 years) and 137Cs activity for recent sediment (35 years), using Ortec EG&G alpha and gamma spectrometers (Eakins and Morrison 1978; Schelske et al. 1994). 210Pb was determined using alpha spectroscopy of 210Po, the short-lived daughter of 210Pb. The extraction method used was a modification of Eakins and Morrison (1978) with a known spike of 209Po. 210Po was measured for 24 h by comparing alpha emissions between 210Po and the added 209Po. The data were analyzed using a “Constant Rate of Supply” (CRS) model (Binford 1990). 137Cs was determined using gamma spectroscopy, following the methods of Schelske et al. (1994). 137 Cs was measured for 24 h by counting gamma emissions at 661.62 kV. 137Cs fallout from the atmosphere reached a peak in 1963 ad as a result of atmospheric nuclear weapons testing (Olsson 1986). Several standards were used for both methods, including International Atomic Energy Agency reference material IAEA300 (radionuclides in Baltic Sea sediment) and Lake Rowell, Florida, sediments provided by Dr. Claire Schelske from the University of Florida in Gainesville.

12.3.5

Geochemistry

For this study, % C and % N of soils were measured on a Perkin-Elmer Series II CHNS/O analyzer 2400 (CSSS 1993). BSi was extracted following the methods described in Conley (1988) utilizing a weak Na2CO3 base to dissolve BSi in a sediment matrix, and measured on a Technicon Autoanalyzer II using a molybdate colorimetric method (Technicon Industrial Systems 1988). Total P was determined by pretreatment of soils using a nitric–perchloric acid digestion (CSSS 1993), and measured colorimetrically using a Bran & Luebbe TRAACS 800 Autoanalyzer by ascorbic acid reduction (US EPA 1983). Ca was measured from the same sample digest as P by atomic absorption spectrophotometry.

12.3.6

Diatoms

Diatoms were extracted at every 2 cm interval of the cores from 1 cm3 of wet soil using 25% H2O2, 25% HCl, and concentrated HNO3 with K2Cr2O7 (Funkhauser and Evitt 1959). Samples were then washed with distilled water, settled, and decanted until neutral pH was achieved. A measured volume of the diatom residue was permanently mounted on glass slides with Naphrax®. Diatoms were identified to species using light microscopy (Leica DM RB with Nomarski optics at 1000×) according to available taxonomic references (Krammer and Lange-Bertalot 1986–1991; Hustedt 1927–1930; Patrick and Reimer 1966, 1975). At least 400 diatom valves or one

12 Historical Changes in Water Quality and Vegetation in WCA-2A

327

complete slide of diatom valves (one sample for which 177 valves were present on one slide) were identified for each sample counted. Once counts were completed, percent abundance of each species for each sample was determined.

12.3.7

Pollen

Pollen extraction and acetolysis of 0.5 cm3 of soil followed the methods of Faegri and Iversen (1989) and Brush (1989). One tablet of Lycopodium spike was added to each sample prior to extraction and acetolysis for calculation of absolute pollen concentration (Stockmarr 1971). The extract was suspended in silicone oil and stored. Pollen grains were identified using light microscopy (Leitz Wetzlar at 400×). A variety of references were used including McAndrews et al. (1973) and Kapp (1969) as well as the Duke University Wetland Center’s Everglades pollen reference collection (Jensen et al. 1999). Horizontal traverses were made across each slide, and pollen and spores were counted and identified to family, genus, or species level when possible. At least 300 pollen grains were counted per slide; in some instances, despite traversing the whole slide this was not possible because of low pollen concentrations and obscuring plant debris on the slide (the minimum number of grains counted was 124). Analysis of longer counts suggests that the relationship between the major taxa do not differ. However, the occurrences of rarer taxa are increased in the longer counts (Moore et al. 1991). If the pollen was crumpled or obscured, it was noted as an indeterminate. Grains that were clearly visible but could not be identified were described, recorded as unknowns, and their slide coordinates noted. Pollen of Cladium was identified according to its overall size and the presence of a “beak” (Faegri and Iverson 1975). It is possible that some Cladium was incorporated into the Cyperaceae taxon grouping. The total pollen count is plotted in the form of both a percentage and an absolute pollen diagram using the U.C. Berkeley CalPalyn Program (Bauer et al. 1991).

12.3.8

Analysis of Soil Core Data

All geochemical data were graphed by depth and corrected for soil bulk density and calculated accretion rates to determine the accumulation rate for each parameter. Down core changes in diatom and pollen assemblages were initially analyzed by examining changes in the relative abundance of proposed indicator taxa (Cooper et al. 1999; Jensen et al. 1999). Shannon–Weaver diversity (H¢) was calculated for each diatom and pollen sample (Shannon and Weaver 1949). Shannon’s H¢ has been shown to be an effective indicator of diatom assemblage variation and is the most widely used diversity index in ecology (van Dam 1982; Washington 1984). Multivariate analyses were performed using cluster analysis. Cluster analysis allows the diatom and pollen data to be viewed in terms of the distance or dissimilarity

328

S.R. Cooper et al.

of assemblages between samples (depths in each core). An average-linkage cluster analysis method was used for this study, using all diatom taxa (196) and pollen and spore taxa (61) identified. The average-linkage method views the dissimilarity between clusters as the average of the dissimilarities between members of those clusters (Venables and Ripley 1994). This method produced a dendrogram with greater distances between samples when compared with the single-linkage or connected clustering method (Clarke 1993). Another multivariate technique, Detrended Correspondence Analyses (DCA), was also used to analyze the diatom data. The ordination was performed using the computer program CANOCO, version 3.1 (ter Braak 1990). Input for the CANOCO program included the percent abundance of diatom taxa that were present in two or more calibration samples with at least 1% abundance, or one sample with at least 5% abundance (Cooper et al. 1999). This reduced the diatom list from 196 to 71 taxa. DCA is an indirect ordination analysis that arranges samples along axes, based on species composition (ter Braak 1995). Ordination analyses indicate which samples are similar in species composition and the direction of change between samples.

12.4 12.4.1

Results Calibration Sets

Phosphorus (P) was found to be the most significant geochemical parameter in relation to the fossil indicators, and explained the most variation in the fossil assemblages (Cooper et al. 1999; Jensen et al. 1999). P was not found to be correlated with the other variables measured. WA regression and calibration models were developed for both pollen and diatoms in relation to soil P (Cooper et al. 1999; Jensen et al. 1999). The inference model developed from the diatom data for soil [P] in WCA-2A showed a very strong r 2apparent = 0.86 and RMSEapparent = 109 µg g−1. Bootstrap estimates of these values showed that the relationship is indeed strong, only lowering the r 2boot to 0.79, with an RMSEboot = 218 µg g−1 (Cooper et al. 1999). The P inference model based on pollen assemblages also had a very high correlation coefficient between measured and estimated soil [P], with r 2apparent = 0.77 and RMSEapparent = 336 µg g−1, and r 2boot = 0.72, with an RMSEboot = 255 µg g−1 (Jensen et al. 1999). % N, % C, Ca, and BSi were also found to be significant in explaining variation of surface soil diatom assemblages. From these analyses, we also determined diatom indicator species for low and high soil P levels (Cooper et al. 1999). Indicator species were identified based on three criteria (1) the species displayed a strong relationship to P, (2) the species displayed a narrow tolerance or deviation around the optimum (less than the average for all species), and (3) occurred commonly in the area (greater than 10% abundance at more than one site and present in at least 33% of sites) (Stevenson et al.

12 Historical Changes in Water Quality and Vegetation in WCA-2A

329

1991; Cooper et al. 1999). Five species were determined to be indicators of high soil [P] (~1,200 –1,400 µg g−1) in the calibration study: Nitzschia amphibia Grunow, N. palea (Kütz.) W. Smith, Gomphonema parvulum (Kütz.) Kützing, Amphora veneta (Grun.) Ross, and Navicula confervacea Kützing. Six species qualified as indicators of low soil [P] (~400–700 µg g−1): Amphora lineolata Ehrenberg, Mastogloia smithii Thwaites, Brachysira vitrea (Grun.) R. Ross in Hartley 1986, Cymbella sp. A, Fragilaria sp. A, and Achnanthes minutissima var. scotica (Carter) Lange-Bertalot (Cooper et al. 1999). The pollen calibration models and WA analysis of surface pollen assemblages in WCA-2A identified several potential P indicator taxa. Increasing soil [P] was associated with decreasing percentage abundances of Cyperaceae (dominated by Cladium jamaicense), Casuarina litorea L. (an introduced exotic), Myrica cerifera L., Nymphaea odorata Aiton, Sagittaria spp., and Asteraceae. In contrast, Typha spp., Polygonum spp., and Chenopodiaceae/Amaranthaceae (cheno/ams) pollen percentages increased in abundance with an increase in soil [P] (Jensen et al. 1999).

12.4.2

Soil Cores, Dating, and Soil Accumulation Rates

A total of five soil cores were collected at the three sites (Figs. 12.1 and 12.2). One core from each site within WCA-2A was analyzed in more detail, based on depth of core and apparent stratigraphy: these included EN1, U1, and U4 (Fig. 12.2). Cores analyzed were between 60 and 113 cm in length. Characterization of the cores collected includes notations on sand layers or obvious inconsistencies in the peat that may indicate strong storm activity or dry periods (Fig. 12.2). Only core U4 showed any nonuniformity in the X-ray images. Below the sand layer present (at about 73-cm depth on Fig. 12.2), the X-ray becomes less consistent, with a diagonal dark layer to the bottom of the core, apparently because of mixing or disturbance. Although there have been reports of Cesium mobility in the peat soils of the Everglades, the 210Pb and 137Cs dates and accumulation rates for each core in this study agreed fairly well. The graphs of activity of these radioisotopes for the three cores are shown in Figs. 12.3 and 12.4. On the basis of 137Cs alone, the accretion rates for the cores since 1963 ad averaged 7.0 mm year−1 for core EN1, 3.3 mm year−1 for core U1, and 1.2 mm year−1 for core U4. On the basis of 210Pb results, the average soil accumulation rates from approximately 1940 to the present were 7.2 mm year−1 for core EN1, 3.5 mm year−1 for U1, and 1.4 mm year−1 for U4. For the samples corresponding to the dates of approximately 1840–1940 ad, the rates were 1.4 mm year−1 for EN1, 1.4 mm year−1 for U1, and an unexpected 3.6 mm year−1 for U4. 210Pb analyses for core U4 were not completed in as much detail as for the other two cores, and the results may be less accurate. The CRS model predicts variable accumulation rates, based on 210Pb, so rates vary by depth or assigned date (Fig. 12.5). Before approximately 1850 ad, the soil accretion rates are assumed to

330

S.R. Cooper et al.

Fig. 12.3 Total 210Po and 137Cs radioisotope activity by depth for core EN1 collected in northern WCA-2A

Fig. 12.4 Total 210Po and 137Cs radioisotope activity by depth for cores U1 (left) and U4 (right) collected in southern WCA-2A

12 Historical Changes in Water Quality and Vegetation in WCA-2A

331

be constant and equal to the average accumulation rate from about 1850 to 1940 ad. These rates are extrapolated back in time, and agree well with other data, based on radiocarbon dates (Craft and Richardson 1993a,b, 1998).

12.4.3

Geochemistry

The soil accumulation rates and concentrations of P, N, and C are shown in Figs. 12.5 and 12.6. Accumulation of BSi, Na, and Ca are shown in Fig. 12.7. Accumulation rates were calculated from the concentration of each geochemical parameter measured for each sample multiplied by the bulk density and accretion rate determined for each sample interval. Input of these indicators to the soils is a function of both the concentration of the parameter and soil accretion rates. Preservation of these parameters in the soils reflects a balance between initial input with deposition and loss due to bacterially mediated decomposition (in the case of organic components) and diagenesis or movement within the soils (in the case of other indicators). The rates of decomposition (Chap. 17) or diagenesis are a function of the nature of the organic matter (i.e., its reactivity), the redox potential of the soils (including what oxidants are available), and the rates of transport of oxidants and ionic components. The confidence level in the results from core U4 are not as high as for the other two cores because of more uncertainty with the calculated accretion rates. As with soil accretion, concentration and accumulation of C, N, P, and Na show large increases in core

Fig. 12.5 Graphs of accretion rates (top left), concentration by depth (top right), and accumulation and concentration of P (along the y-axis) in relation to dated soils for cores EN1, U1, and U4 collected at three sites in WCA-2A. Dates assigned to soil samples are shown on the x-axis, from the most recent on the left back in time to the right

332

S.R. Cooper et al.

Fig. 12.6 Graphs of accumulation and concentration of C and N (along the y-axis) in relation to dated soils for cores EN1, U1, and U4. Dates assigned to soil samples are shown on the x-axis, from the most recent on the left back in time to the right

EN1 in the P-enriched area of WCA-2A in recent years (post-1950 ad). Smaller increases in the accumulation of these elements are found in cores U1 and U4. The exception to these trends is with Ca and BSi, which both show larger increases in accumulation in recent times in core U1 (discussed later). The results for core U4 are different in that accretion rates and C and N accumulations seem to have been higher around the turn of the century than in the past 50 years. However, all cores show an increase in P and Ca accumulations in the past 50 years compared with earlier times. The fact that they are much higher in the recent samples from core EN1 indicates that these increases, at least for this site, are not solely related to natural decomposition or diagenetic processes. The two cores collected in the unenriched areas of WCA-2A (U1 and U4) show greater increases in Ca and BSi accumulations in the past 50 years compared with previous (from 6 to 40 times for Ca, and six times the accumulation of BSi). The measurements for the EN1 core showed a sixfold increase in Ca and 0.7 times for BSi in the past 50 years. Core EN1 results show the greatest increases in accretion rate (five times) and P accumulation (five times [P] and 23 times P accumulation) in the past 50 years. In contrast, cores U1 and U4 show accretion rate changes of 2.5 times for U1 vs. 0.3 times for U4. These two cores collected in the less enriched areas of WCA-2A show twofold [P] and twofold to sixfold P accumulation over the past 50 years compared with earlier times. Recent soil accumulations of P, C, and N compared between the enriched core (EN1) and the unenriched sites (U1 and U4) show that average accumulations (and concentrations) are higher at the enriched site. Recent P accumulation at the enriched site is seven to ten times higher than at the unenriched sites, while C and

12 Historical Changes in Water Quality and Vegetation in WCA-2A

333

Fig. 12.7 Accumulation rates of calcium, sodium, and biogenic silica (along the y-axis) in relation to dated soils for cores EN1, U1, and U4 collected at three sites in WCA-2A. Dates assigned to soil samples are shown on the x-axis, from the most recent on the left back in time to the right

334

S.R. Cooper et al.

N accumulations are about two to four times higher for C and 1.5–4 times higher for N when compared with the unenriched sites.

12.4.4

Diatoms

A total of 8,325 diatom valves were identified from 20 core samples in soil cores EN1, U1 and U4, representing 196 taxa in 41 genera. Ninety-three percent of the species used in the calibration set were also found in relatively high abundance in these samples (especially recent samples). For all species identified, the core samples and calibration samples shared about 75% of species with each other, but differences became greater with depth in the cores. Eleven samples were analyzed for core EN1, seven for core U1, and two for core U4. Diatoms were not well preserved in some samples (particularly in core U4), so that counts were not possible for all depths. The genera with the most species represented included Navicula (34 species), Nitzschia (24 species), Eunotia (14 species), and Cymbella (12 species). Ten species showed greater than 5% abundance in more than one sample. These included Amphora lineolata, Brachysira vitrea, Cyclotella meneghiniana Kützing, Cymbella sp. A, Eunotia spp., Fragilaria sp. A, Frustulia rhomboides (Ehr.) De Toni, Gomphonema parvulum, Mastogloia smithii, and Nitzschia amphibia. Cluster analysis of diatom assemblages is shown in dendrogram form (Fig. 12.8). DCA of the diatom assemblages from the three cores are shown in Fig. 12.9. Clusters designated A through C1 are based on dissimilarities between assemblages (Euclidean distance between samples), and the abundance of dominant taxa. The clusters and the DCA both show that although recent diatom assemblages are very different between the more northern EN1 core collected in the P-enriched area and the U1 and U4 cores collected in less impacted areas of WCA-2A, the diatom assemblages are more similar down core at all sites. The cluster analysis shows that recent samples from EN1 are clustered together in Group A, which are most distant from all other sample groups (Fig. 12.8). Cluster group B contains the surface sample diatom assemblages from cores U1 and U4, whereas the remaining diatom assemblages from older samples at all sites are clustered together in Group C (with one outlier designated as group “C1”). The DCA shows the direction of change in the diatom assemblages along the two main ordination axes for the three cores from the most recent samples back in time. It is evident that the diatom assemblages were more similar before major human impacts to WCA-2A. The cluster groups are also overlain on this graph (Fig. 12.9). Diversity (Shannon’s H′) has generally declined over time in all three cores, indicating a drop in species richness and evenness in recent years (Fig. 12.10). This is especially pronounced for the EN1 core from the P-enriched area of WCA-2A where there has been a threefold drop in diversity of the diatom assemblages from before the turn of the century to the present. At the same time, there has been a huge decline in the abundance of Eunotia species to the present time (Fig. 12.10). It is possible that there has also been some differential preservation over time due to changes in pH, although the diatoms appear to be better preserved in the more

12 Historical Changes in Water Quality and Vegetation in WCA-2A

335

Fig. 12.8 Cluster analysis using Euclidean distance between diatom assemblages for core samples analyzed for diatoms from cores EN1, U1, and U4. Clusters A through C1 were determined based on Euclidean distance and abundance of dominant taxa. The labels on the figure indicate core and sample depth in centimeters (E core EN1, U core U1, UIV core U4). Dates are provided for each sample as year ad next to the cluster designations

Fig. 12.9 Detrended correspondence analysis ordination plot based on the diatom assemblages analyzed for cores EN1, U1, and U4 collected in WCA-2A. The labels on the figure indicate core and sample depth in centimeters (E core EN1, U core U1, UIV core U4). Arrows are drawn sequentially for each core from the top sample of the core to the deepest sample analyzed for each core, to indicate the direction of diatom assemblage change during the last 150 years (core EN1 = solid line arrows, core U1 = dotted line arrows, core U4 = dashed line arrow)

336

S.R. Cooper et al.

Fig. 12.10 Diatom diversity (Shannon’s H′), percent abundance of Eunotia species, and pollen diversity (Shannon’s H′), for samples analyzed from cores EN1 (filled circles), U1 (open squares), and U4 (shaded triangles) are shown graphed by time along the x-axis. Dates assigned to soil samples are shown from the most recent on the left

recent soils. In samples dated before 1960 ad, Eunotia species comprise at least 10% (average 27%) of the diatom assemblages from both cores EN1 and U1 (the dominant genus in most of these samples). In the recent samples, Eunotia species comprise less than 5%, with no valves seen at all in any of the samples since 1980 and 1990 for U1 and EN1, respectively. Because of the different diatom assemblages found down core at all sites compared with the surface samples, the weighted average regression and calibration models developed from diatom assemblages in the surface soils were not very useful in hindcasting geochemical parameters back in time. However, the diatom indicator species determined using these methods have been useful in assessing the changes seen at the different sites through time. Graphs of diatom species abundances in core samples that were determined to be indicators of high or low soil P in the calibration study (Cooper et al. 1999) are shown in Figs. 12.11 and 12.12. There has been a significant increase in abundance of the high soil P indicator species in recent times in the EN1 core. Only two of these

12 Historical Changes in Water Quality and Vegetation in WCA-2A

337

Fig. 12.11 Percent abundance of diatom taxa that were determined to be good indicators of high [P] (Cooper et al. 1999) for cores EN1 (filled circles), U1 (open squares), and U4 (shaded triangles) are shown graphed by time along the x-axis. Dates assigned to soil samples are shown from the most recent on the left. The diatom taxa shown include clockwise from top left: Nitzschia amphibia, Amphora veneta, Gomphonema parvulum, and N. palea. Please note that the y-axis scale is different for each taxon represented

species (Nitzschia palea and Gomphonema parvulum) are present in the other cores (only in samples analyzed from core U1), and then only in very low abundances. Of the five diatom species qualified as indicators of high soil P in the calibration study, four are shown. Navicula confervacea was present in only three out of the 20 core samples analyzed for diatoms, and is therefore not shown. Of the six diatom species that qualified as indicators of low soil P, only three were present in older samples of core EN1 (Mastogloia smithii, Brachysira vitrea, and Fragilaria sp. A). M. smithii and Amphora lineolata show significant increases in abundance in core U1 and U4 from the P-unenriched sites in WCA-2A in recent years, especially since 1960 ad. A. lineolata and M. smithii are listed as two of the most abundant diatom species associated with calcareous periphyton in WCA-2A (McCormick and O’Dell 1996; McCormick et al. 1996, 1998). Achnanthes minutissima var. scotica, which qualified as a low soil P indicator in the calibration study, showed higher abundances in surface samples in more eastern sites (Cooper et al. 1999) and is seen in samples from core U4, but not in samples from core U1. This species has been associated with small streams and flowing water conditions (Ludlam et al. 1996).

12.4.5

Pollen

A total of 61 pollen and spore taxa were identified from samples in cores EN1 and U1. Ten levels were analyzed from core U1 and 13 from core EN1. Pollen was generally well preserved. Data are plotted in the form of percentage and absolute

338

S.R. Cooper et al.

Fig. 12.12 Percent abundance of diatom taxa that were determined to be good indicators of low [P] (Cooper et al. 1999) for cores EN1 (filled circles), U1 (open squares), and U4 (shaded triangles) are shown graphed by time along the x-axis. Dates assigned to soil samples are shown from the most recent on the left. The diatom taxa shown include clockwise from top left: Amphora lineolata, Fragilaria sp. A, Cymbella sp. A, Achnanthes minutissima var. scotica, Mastogloia smithii, and Brachysira vitrea. Please note that the y-axis scale is different for each taxon represented

pollen concentrations (grains year−1 cm−2). Each core can be divided into two zones, based upon changing pollen associations. Cluster analysis of pollen assemblages from both cores U1 and EN1 is shown in a dendrogram (Fig. 12.13). Three clusters are apparent. The clusters show that recent pollen assemblages differ between the nutrient-enriched site and -unenriched site. Recent samples from EN1 are clustered in Group A, which like the diatom results, are more distant from all other sample groups. Cluster Groups B and C contain a mix of samples from both cores and from various depths within those cores. At core site EN1 pollen assemblages in Zone 2 (111–22 cm) are dominated by a combination of pollen of trees and shrubs (range 15–43%), especially Pinus spp., and pollen from wetland/aquatic taxa (13–53%) and taxa indicative of disturbance (11–40%), especially cheno/ams pollen (Figs. 12.14 and 12.15). In Zone 1 (22–0 cm),

12 Historical Changes in Water Quality and Vegetation in WCA-2A

339

Fig. 12.13 Cluster analysis using Euclidean distance between pollen assemblages for core samples analyzed for pollen from cores EN1 and U1. (Euclidean distance has been rescaled to 0–25 by the SPSS statistics package.) The labels on the figure indicate core and sample depth in centimeter (EN core EN1, U core U1). Dates are provided for each sample as year ad next to the cluster designations

the percentage importance of tree and shrub pollen declines markedly (Fig. 12.14), and pollen of wetland/aquatic taxa and taxa indicative of disturbance (Asteraceae and cheno/ams) are more important. Examination of the absolute diagram (Fig. 12.16) indicates that tree and shrub pollen remains more or less constant throughout the core. However, concentrations of wetland/aquatic pollen (specifically cheno/ams) increase dramatically. This is reflected in total pollen influx, which significantly increases in Zone 1, ranging from 86,000 to 34,000 grains year−1 cm−2, in comparison to Zone 2 influx, which was much lower (1,900–5,400 grains year−1 cm−2). Significantly, Typha spp. pollen increases in importance from 22 cm to the surface, while pollen of Nymphaeaceae declines from Zone 2. Cladium pollen also becomes important within Zone 1, but this may be a function of difficulties with identification. The pattern of changes in pollen input to EN1 is apparent in both the percentage and absolute pollen diagrams (Figs. 12.15 and 12.16).

340

S.R. Cooper et al.

Fig. 12.14 Summary percentage pollen diagrams for sites U1 and EN1. Note that the vertical axis is not to scale

At core site U1, the pollen diagram can also be divided into two zones (Figs. 12.17 and 12.18), with changes in pollen assemblages more marked in the absolute abundance diagram (Fig. 12.18). Zone 2 (41–14 cm) is dominated by disturbance taxa (range 27–54%), specifically by cheno/ams. Zone 1 (14–0 cm) is characterized by an overall decline in the importance of wetland/aquatic taxa (range 8–23%), but an increase in taxa of disturbance species (range 30–48%). There appears to be a decline in wetland/aquatic taxa from Zone 2 to Zone 1 as Nymphaeaceae, Ponteridaceae, and Sagittaria spp. become less important. Typha pollen is present at negligible concentrations (0.8%) at 3 cm. Tree and shrub pollen remains more or

12 Historical Changes in Water Quality and Vegetation in WCA-2A

341

Fig. 12.15 Percentage pollen diagram for EN1 showing significant taxa. Taxa are graphed as a percentage of total pollen and spores

less constant between the two zones (range 11–26%), except for one sample with 6% tree and shrub pollen (Fig. 12.14). There is a slight increase in total pollen influx from Zone 2 mean levels of ∼3,600 grains year−1 cm−2 to mean levels of ~5,300 grains year−1 cm−2 in Zone 1. Graphs of pollen diversity changes (Shannon’s H′) are presented in Fig. 12.10. Jensen (1998) found that diversity of surface pollen assemblages roughly reflected changes in soil [P], with the pollen spectrum from higher soil [P] sites ranging from H′ = 0.61 to 2.08, and from lower soil [P] sites ranging from H′ = 2.44 to 3.17. Average values for H′ in this study within the cores varied little between sites, U1: H′ = 2.18 ± 0.21 and EN1: H′ = 2.16 ± 0.39. However, the most recent samples from the enriched site indicate a slight declining trend in diversity, while the most recent samples from the unenriched site indicate a very slight increase in diversity. Several plant species (Vaithiyanathan and Richardson 1999) and pollen taxa (Jensen 1998; Jensen et al. 1999) are thought to be indicators of high or low soil [P]. Graphs of the abundances of these taxa in the core samples are shown in Fig. 12.19, except three taxa identified by Jensen (1998) and Jensen et al. (1999) as concentrations were too low (Polygonum spp. and Myrica spp.) or the taxon was not identified (Casuarina litorea). In particular, Typha spp. and Cladium

342

S.R. Cooper et al.

Fig. 12.16 Absolute pollen diagram for EN1 showing significant taxa. Taxa are graphed in relation to total pollen and spore influx (grains year−1 cm−2)

(Cyperaceae) are thought to reflect P enrichment and more pristine conditions, respectively. The trend for Typha clearly shows an increase in the percentage importance of the plant for the recent historic period at the enriched site. However, Cyperaceae (including Cladium) does not decline as might be expected in the P-enriched area, instead it shows an increase for the recent historic period. Relationships between other taxa identified as indicators of low or high [P] are also inconclusive. Cheno/am pollen, an indicator of high soil [P] (Jensen et al. 1999), does show an increase at the nutrient-enriched site, and a comparable decline at the unenriched site. Low soil [P] indicators Asteraceae and Nymphaea odorata show conflicting patterns. Asteraceae abundance increases to the present at both sites, while Nymphaea shows a decline at both sites to the present. However, the results for Asteraceae and cheno/ams cannot solely be interpreted in terms of changes in soil [P] as increases in abundance of both taxa have been linked to disturbance and changes in hydroperiod (Willard et al. 2001). Decreasing Sagittaria pollen was correlated with increasing soil P in surface calibration work (Jensen et al. 1999), while surface vegetation studies found it to have a widespread distribution throughout all P levels within WCA-2A today (Vaithiyanathan and Richardson 1999). Sagittaria pollen shows a marked decline in pollen abundance at both sites up to the present, except for a recent increase at the enriched site.

12 Historical Changes in Water Quality and Vegetation in WCA-2A

343

Fig. 12.17 Percentage pollen diagram for U1 showing significant taxa. Taxa are graphed as a percentage of total pollen and spores

12.5 12.5.1

Discussion Soil Cores, Dating, and Soil Accumulation Rates

Average soil accretion within WCA-2A cores is very similar to those reported by Craft and Richardson (1998) and Bartow et al. (1996). Craft and Richardson (1998) reported accretion rates of 4.4 and 7.2 mm year−1 in the recent soils (post1960) of cores collected in P-enriched sites of WCA-2A and an average 2.1 mm year−1 in one core collected at an unenriched site in WCA-2A. Bartow et al. (1996) measured peat accretion at an unenriched site averaging 2.05 mm year−1, while at an enriched site two distinct periods of peat accumulation were recognized, prior to 1953 accretion was low (1.5 mm year−1) followed by higher rates of accretion (7.2 mm year−1). It is apparent that soil accretion rates have increased in the northern part of WCA-2A in the past 50 years compared with unenriched areas. These changes are attributed to anthropogenic influences, primarily increased nutrient inputs from agricultural practices (EAA) via the Hillsboro Canal along the northern perimeter of WCA-2A. These changes may also be related to the changes in

Fig. 12.18 Absolute pollen diagram for U1 showing significant taxa. Taxa are graphed in relation to total pollen and spore influx (grains year−1 cm−2)

344 S.R. Cooper et al.

12 Historical Changes in Water Quality and Vegetation in WCA-2A

345

Fig. 12.19 Percent abundance of pollen taxa that were determined to be good indicators of high and low [P] (Jensen et al. 1999; Vaithiyanathan and Richardson 1999) for cores EN1 (filled circles) and U1 (open squares) are shown graphed by depth along the x-axis. The pollen taxa shown include clockwise from top left: Typha spp., Sagittaria spp., Cheno/ams, Cyperaceae family, Nymphaea spp., and Asteraceae family. Please note that the y-axis scale is different for each taxon represented

type of peat accumulating as cattails become more prevalent with higher nutrient inputs. Historical soil accumulation rates in northern WCA-2A (pre-1940 ad) appear to compare well with those in the unenriched sections of WCA-2A before major human impacts to the area. Although dating of core U4 was not accomplished with the same detail as core EN1 and U1, and calculated accretion rates are less accurate, it is possible that the observed increase in accretion rates and higher accumulations of C and N around the turn of the century are real. If so, what may have caused this change is not known. Diatoms were not well preserved in this core, and no counts are available for the time period in question.

346

12.5.2

S.R. Cooper et al.

Geochemistry

As with accretion rates, recent rates of C, N, and P accumulation were highest in core EN1 representing the nutrient-enriched area of WCA-2A. P accumulation in core EN1 averaged seven times higher than the rates calculated for cores U1 or U4 in samples dated since 1950 ad. This comparison agrees with the differences found by Craft and Richardson (1998) in a study of accumulation rates of P for six sites throughout the Everglades (including the ENP). They found accumulation rates of P at sites in northern P-enriched WCA-2A that were from three to 46 times higher than all other sites. The P gradient within WCA-2A has been well documented (Reddy et al. 1991; Craft and Richardson 1993a,b; Qualls and Richardson 1995; Vaithiyanathan and Richardson 1997a). Some of the other results of our study involve issues that are not thoroughly documented by earlier studies. The changes in accumulation rates of BSi and Ca are shown to be higher in recent years in core U1 when compared with EN1. BSi is a measure of primarily diatom remains found in the soil. Diatom productivity may be higher at the unenriched sites because of the higher area of slough habitats, fewer periods of dry downs, higher water levels since impoundment, and more calcareous periphyton (Richardson et al. 1997a). The presence of calcareous periphyton in the unenriched area sloughs of WCA-2A and the influence of P concentration, pH, and calcium carbonate (CaCO3) precipitation (Swift 1984; Vaithiyanathan et al. 1997) are most likely related to the recent increase in Ca accumulation in these areas when compared with the EN1 site. Slate (1998) also discovered similar trends and discussed possible sources of Ca to this area.

12.5.3

Diatom Assemblages

It is apparent that diatom assemblages found in WCA-2A were more uniform throughout the area before impoundment and the addition of nutrients via the Hillsboro Canal. This implies that water quality and hydrology was more uniform throughout the area now designated as WCA-2A before the turn of the century. As human impacts have increased and the P gradient has developed within WCA-2A, the diatom assemblages have diverged along this P gradient. These divergences also appear to be related to other changes in water quality and hydrology. Slate (1998) reports similar diatom assemblage change results in a related paleoecological study of soil cores from WCA-2A in which she quantified diatoms, chrysophyte cysts, and sponge gemmoscleres. One overall change in water quality in WCA-2A appears to be an increase in pH of the water column. This study, as well as Slate’s (1998), found that Eunotia species were much more prevalent (in fact, a dominant genus) throughout WCA-2A before 1960 ad. Almost all species of Eunotia are acidophilous (acid-loving)

12 Historical Changes in Water Quality and Vegetation in WCA-2A

347

species, having a pH optima for growth either at or below pH 6.0 (Cholnoky 1968). This includes the dominant species found in WCA-2A such as E. camelus Ehrenberg and E. arcus Ehrenberg. Slate (1998) postulates that CaCO3 input has increased with drainage and impoundment projects in WCA-2A because of erosion of limestone bedrock through which canals were deepened, raising the pH of the water column and buffering the system. This also implies that calcareous periphyton may be relatively new in WCA-2A, as was also alluded to in an earlier study by Gleason and Spackman (1974), who claimed that calcareous periphyton was consistently more prevalent south of Tamiami Trail in the Everglades. Another overall change is the decline in diversity of diatom assemblages at all sites over the past 40–50 years. This decline is more pronounced at core site EN1 than at the less P-enriched sites. In contrast, diversity of surface soil diatom assemblages found in the calibration study (Cooper et al. 1999) showed an increase with moderate P enrichment at different sites, with a decline at the highest soil P sites. The overall decline in diversity within the core samples over the past several decades may, therefore, be related to a synergistic effect of several factors; including nutrient enrichment, ion concentration changes, and hydrology (habitat) changes. Overall, it appears that there has been a significant shift in species dominance that diverges according to P enrichment. It appears that most of the diatom species identified in the calibration study (Cooper et al. 1999) as high soil [P] indicators have become more prevalent in the northern P-enriched areas while low [P] indicators have become more prevalent in the more southern-unenriched areas of WCA-2A. This is not what we expected to find in the course of this paleoecological study. It was hypothesized that diatom indicators of low soil [P] would be the dominant species at all sites in WCA-2A before significant anthropogenic influence on water quality and hydrology had occurred. Instead, it appears that water quality (e.g., pH, ion concentrations) and hydrology patterns have changed throughout WCA-2A, causing a shift in diatom assemblages that now reflect the larger differences in water quality within WCA-2A. The changes that appear to have the greatest impact are differences in [P] and its effect on Ca precipitation by periphyton species, and hydrology patterns (Romanowicz et al. 1996; Vaithiyanathan et al. 1997). The diatom species Achnanthes minutissima var. scotica, present at site U4 but not at U1, indicates that flow patterns may be different at these two sites (Ludlam et al. 1996). This is corroborated by findings that show slower flow patterns in the interior of WCA-2A than closer to the perimeter (Romanowicz et al. 1996). Three diatom species indicators of low soil [P] (Cooper et al. 1999) do appear to have been present in generally less than 10% abundance at all sites before significant anthropogenic influences to WCA-2A and could, therefore, be designated as “native” taxa: Mastogloia smithii, Fragilaria sp. A, and Brachysira vitrea. M. smithii have since significantly increased in abundance in the P-unenriched sites while declining in abundance at the P-enriched site over time. Fragilaria sp. A and B. vitrea appear to have declined in abundance through time at all sites where cores were collected. The decline in B. vitrea is probably not simply related to changes

348

S.R. Cooper et al.

in [P] at these sites, since the [P] optimum determined for this species was higher than for M. smithii (Cooper et al. 1999). The highest percent abundance of diatom indicators of high [P] have occurred in the past two decades. Slate (1998) also reports highest abundance of eutrophic diatoms in soils dated to the last two decades in two cores that she collected in the northern P-enriched area of WCA-2A. This result is most likely due to agricultural production of sugarcane, vegetables, and rice, and nutrient runoff to WCA-2A. Agricultural production in the area has steadily increased since the early 1900s into the 1980s (Snyder and Davidson 1994). These findings suggest that managers will not only have to choose which of a variety of issues they must address but they will also have to choose from among different scenarios of how and to what condition they want to restore the Everglades. The water quality and hydrology within WCA-2A was more uniform before impoundment than it is now. It was also different in terms of pH, Ca, and diatom assemblages in both the northern-enriched site and the southernunenriched sites of WCA-2A than present. The factors involved in changing water quality and diatom assemblages appear to be related to agriculture, canal building and dredging, and the alteration of hydrology patterns. Patterns of calcareous periphyton growth in the P-unenriched slough areas of WCA-2A may be a relatively new phenomenon, but worthy of maintaining. McCormick et al. (1998) discuss the benefits of periphyton as it is now found in areas of the WCAs. These include its usefulness as a food source for aquatic herbivores, habitat suitability for native Everglades species, and increased assimilative capacity of the wetland.

12.5.4

Pollen Assemblages

The results from both EN1 and U1 cores indicate significant changes in pollen assemblages occurring some time in the early 1960s, which compares well with the results from the diatom data. At the nutrient-enriched site EN1, commencing at approximately 1965 ad and continuing to the present, Typha spp. become prevalent in the record possibly reflecting [P] enrichment. The timing and increase in Typha pollen at EN1 is consistent with other pollen studies in the region, as is the decline in Nymphaeaceae pollen (Nymphaea and Nuphar luteum (Small) Beal.) (Bartow et al. 1996; Willard 1997). The increase in cheno/ams, Poaceae, and Asteraceae pollen abundance during this time period contributes support to the effects of anthropogenic disturbance in the region (Bartow et al. 1996; Vaithiyanathan and Richardson 1999). Total pollen influx dramatically increased during this time frame possibly reflecting increased productivity, this pattern is also recorded at other sites within the enriched region of WCA-2A (Bartow et al. 1996; Willard 1997). At site U1 the geochronological analysis and pollen data also indicate a change in pollen assemblages since approximately 1963 ad. Indeed, patterns in

12 Historical Changes in Water Quality and Vegetation in WCA-2A

349

pollen assemblages are very similar to that of EN1 except for the absence of Typha pollen. U1 also shows increases in the influx of disturbance indicators, especially Asteraceae, and a concomitant decrease in Nymphaeaceae pollen. Bartow et al. (1996) and Willard (1997) also note similar patterns. It is clear that while nutrient enrichment, specifically [P], is not affecting this portion of WCA-2A to the degree it is at site EN1, some form of disturbance (presumably anthropogenic) has occurred throughout the last 40 years. Bartow et al. (1996) suggest that changes in drainage characteristics of the region may be responsible for these trends. The pollen profiles of the lower sections of both cores (Zone 2) are similar. Notably, pollen from taxa found in sloughs, Nymphaeaceae and Sagittaria are present at low concentrations. Concentrations of Nymphaeaceae decline in both cores within Zone 1, while Sagittaria declines at U1 and its pollen influx increases at EN1. It is surprising that declines in these aquatic plants occur in recent deposits, particularly in the unenriched south, since water levels are generally higher in the south of WCA-2A today than prior to impoundment (Chap. 7). Tree and shrub pollen distribution throughout both cores does not change significantly from one zone to the next and probably reflects long distance transport into the WCA-2A region. Bartow et al. (1996) found that Pinus pollen declined toward the present because of logging activities beginning in the middle of the last century, a slight decline in Pinus occurs at this time in core U1, although it does not continue to decline into the twentieth century. No noticeable change in Pinus occurs at core site EN1. Analysis of the pollen diversity from both sites confirms the findings of the surface calibration survey (Jensen 1998), which found that pollen diversity was highest in regions of lower soil [P]. In the current study, the diversity at the enriched site has declined since the 1960s, while at the unenriched site diversity has increased since the turn of the century. This result might be expected as experimental work has determined that species diversity declines with fertilizer additions as more productive species competitively exclude less productive species (Rosenzweig 1971; Tilman 1993). In contrast, Vaithiyanathan and Richardson’s (1999) analysis of vegetation within WCA-2A determined an increase in macrophyte species within the nutrient-enriched area, although there was a loss of characteristic slough species such as Eleocharis spp. and Utricularia purpurea Walt. The palynological analysis presented here and that of other researchers (Bartow et al. 1996; Willard 1997) provides an important historical perspective on the changes that have occurred within Everglades plant communities. The pollen profiles complement the diatom research and indicate that changes in the nutrient composition and hydrology of the region have significantly impacted the vegetation. Not only has a major change in dominant plant communities occurred as a result of nutrient enrichment, Cladium to a Typha dominated marsh, but also changes in aquatic slough communities (decline in Nymphaeaceae and Sagittaria) has occurred throughout the region possibly as a result of changes in hydrology.

350

12.6

S.R. Cooper et al.

Conclusions and Lessons for Restoration

Paleoecological studies of the Everglades can provide a wealth of information about the current status of areas within the WCAs (e.g., calibration studies) as well as the history of vegetation cover, water quality, and hydrology patterns. This study of sites within WCA-2A shows that not only have water quality, vegetation, and diatom assemblages changed spatially due to changing nutrient inputs, they have also changed through time even at what many consider to be more “pristine” areas of the Everglades such as P-unenriched sections of WCA-2A. These changes may be due to overall changes in hydrology patterns and ion concentrations that have affected the pH of the overlying water and Ca availability. The most recognizable changes in vegetation, water quality, and sediment geochemistry have occurred in WCA-2A in the past 40–50 years. The magnitude (e.g., soil [P]) and type of changes (species shifts, diatom production, and Ca deposition) are different depending on P enrichment of the overlying waters. P has increased more in the north of WCA-2A over this time frame due to nutrient inputs from the EAA via the Hillsboro Canal gates. Overall there is an increase in disturbance species present throughout the area, and Typha in particular has taken advantage of human impacts to proliferate in the north. Although overall anthropogenic effects appear to be detrimental to the historic ecosystem of the northern Everglades, some effects of human disturbance, such as an increase in pH and Ca concentration – as long as P remains low – now allow periphyton growth in the slough areas of southern WCA-2A. Future research is needed to investigate further the effects of P and Ca changes on the Florida Everglades. Altered hydroperiod within the Everglades may also be responsible either directly or indirectly for some of the changes seen. The importance of hydroperiod changes in relation to P enrichment problems has not been adequately studied. More soil core analyses and broader calibration sets would facilitate understanding of the extent of human induced changes to the ecosystem. In particular, the pollen calibration set (Jensen et al. 1999) shows promise, but needs to be expanded and compared with additional core samples. Laboratory experiments to determine pH and [P] optima of diatom species under controlled situations would contribute to the diatom calibration studies undertaken in the field. With broader coverage and more detailed studies, it may be possible to develop models to predict post-restoration plant communities, based on planned restoration activities. Our findings suggest that managers will have to choose what condition they want to restore in the Everglades. The water quality and hydrology within WCA-2A was more uniform before impoundment than it is now. It was also different in terms of pH, Ca, and diatom assemblages in both the northern-enriched site and the southern-unenriched sites of WCA-2A than present. The dense patterns of calcareous periphyton growth in the P-unenriched slough areas of WCA-2A may be a relatively new phenomenon to the region because of the release of additional calcium from canal cuts in the bedrock. The factors involved in changing water quality and diatom assemblages appear to be related to agriculture, canal building and dredging, and the alteration of hydrology patterns. The future of algal communities in the managed Everglades depends on a better understanding of the relationship of proposed water management releases and water quality.

13

Carbon Cycling and Dissolved Organic Matter Export in the Northern Everglades Robert G. Qualls and Curtis J. Richardson

13.1

Introduction

Because much of the Everglades is a vast subtropical freshwater fen, its carbon cycle can be expected to differ from that of most ecosystems in several important ways. The accumulation of peat is evidence that net primary productivity (NPP) exceeds heterotrophic respiration over the long term. The Everglades are subtropical, and thus NPP proceeds year-round, although with a peak in the summer (Davis 1989). Consequently, with a release from nutrient limitations, the potential rates of NPP may be expected to be high compared with many ecosystems. However, nutrient limitation by P is severe, and – like in many peat soils – the long-term storage of P has depleted the sources of recyclable P. In addition, the vast extent of contiguous peatland probably serves to minimize the impact of runoff from surrounding natural upland geological sources of mineral P. Acidity is not a factor in slowing the carbon cycle as is the case in acidic peatlands like the Okefenokee. The occurrence of an annual dry season and periodic droughts creates great episodic bursts of aerobic mineralization and conditions conducive to fires. The sapric nature of Everglades peat is likely a result of the subtropical warmth and periodic aerobic conditions. The low clay content and lack of an underlying clay layer also preclude the protection by organomineral complexes of the humified peat from decomposition when it is drained. Consequently, subsidence and the acceleration of respiration in the warm humid climate are exceptionally severe in the drained peats of the Everglades Agricultural Area (EAA). Finally, the relatively moderate concentrations of sulfate and presence of emergent species of macrophytes that vent methane to the atmosphere allow methane emission to be a prominent feature of the carbon cycle (Chanton et al. 1993). Our discussion of the carbon cycle is restricted to the northern Everglades, which lack the extensive marl soils and mangroves found in the southern Everglades. The northern Everglades has been subject to two severe alterations that dramatically affect the carbon cycle: drainage in the EAA and phosphorus enrichment in areas bordering the EAA. Consequently, we will contrast the carbon cycle in the drained EAA, the P-enriched area of Water Conservation Area 2A (WCA-2A), and the unenriched area of WCA-2A. We have excluded the slough communities because of lack of data on NPP.

351

352

R.G. Qualls and C.J. Richardson

As with most efforts to compile a carbon budget for an ecosystem, data were gathered at different times, in different plots, and by different investigators. Despite this fact, such a synthesis reveals many insights. One objective of this chapter is to describe the most important aspects of the carbon cycle and how eutrophication and drainage affect the major fluxes. Another objective in presenting the fluxes carbon is to place the fluxes of dissolved organic carbon (DOC) in perspective. We will then focus in detail on experiments that show how the concentrations and export of DOC is controlled. In this discussion of DOC dynamics our specific objectives were: (1) To determine the influence of P enrichment on concentrations of DOC (2) To determine whether there is a net production and export of dissolved organic nutrients in the marshes of the Water Conservation Areas (WCAs) and the EAA (3) To determine the importance of biodegradation and solar radiation in the mineralization of C an N from dissolved organic matter (DOM)

13.2 13.2.1

Methods Carbon Budgets

Data on NPP, soil respiration, decomposition, peat accretion, and methane production were gathered from several studies listed in the footnotes of Figs. 13.1–13.3. Fluxes believed to be relatively small have, in some cases, been omitted from the cycle diagrams for simplicity. Some fluxes, such as CO2 respiration from peat, were calculated by difference.

13.2.2

Concentrations of DOC in WCA-2A

In WCA-2A, samples were taken along a nutrient-enrichment gradient approximately 10-km long along three transects in 18 plots (see Fig. 5.1). Samples of surface water and soil pore water at 12.5, 25, and 60 cm below the soil surface were gathered at 2-month intervals between January 1990 and April 1991. Soil pore water was collected in PVC pipe wells with 20-µm hole size nylon screen covering the bottom. Older pore water was pumped from the wells, and new pore water was allowed to infiltrate the well before sampling. Samples were immediately stored on ice and filtered through combusted glass fiber filters within about 12 h. Samples were analyzed for DOC, inorganic carbon, and phosphate within 5 days; NO–3 , NO–2 PO4–3 within 7 days; and total N and total P within 14 days. The DOC was measured by combustion at 650°C on a platinum catalyst with infrared detection of the CO2 evolved using a Shimadzu 500 Total Organic Carbon Analyzer (Shimadzu Corp., Columbia, MD). Other analytical methods are summarized in Qualls and Richardson (1995).

13 Carbon Cycling and Dissolved Organic Matter Export in the Northern Everglades

353

Fig. 13.1 Sources and calculation of fluxes and pools for the unenriched area of WCA-2A. Superscript letters in the diagram correspond to the footnotes: aNPP = net primary productivity (aboveground = 524 g C m−2 year−1 plus belowground = 62.9 g C m−2 year−1). Aboveground NPP from Davis (1989), cattail site A. Belowground NPP from Toth (cited in Davis 1991). Assumes dry mass = 45% C. Root litter input assumed to equal belowground NPP since the living biomass is not accreting. bLitterfall is assumed to equal belowground NPP since the living biomass is not accreting and in turns over several times per year. cDecomposition and CO2 efflux from litter during the first year of decomposition, data from Qualls and Richardson (Chap. 17), using C loss at 1 year in sawgrass litter from the control channel minus the initial content of soluble C. dDecomposition of peat older than 1 year = NPP − first-year litter decomposition − peat accretion rate (modified from Craft and Richardson 1993b to include belowground NPP and to subtract the firstyear decomposition). Because fragmentation was minimal all C lost in addition to the soluble C was assumed to lost as CO2. eCO2 efflux = decomposition rate of peat − methane flux. fMethane flux measured by Chanton et al. (1993) in sawgrass stands (including flux from vegetation) in WCA-3A using the average of sawgrass stands converted from mg CH4 day−1 to g C m−2 year−1. g Leaching of DOC from freshly senesced litter (from this chapter). hDOC imported and exported in canal water flowing into and out of WCA-2A (from this chapter). Note that since the import and export included all of WCA-2A, the export from the enriched area is not distinguished from the unenriched area although the export from the enriched portion is probably greater. iMicrobial decomposition of DOC (from this chapter). jAbiotic mineralization (from this chapter). kPlant biomass, from Davis (1989), sawgrass sites. lPool not measured but is variable over the year. mAssuming average 1-m depth of peat in unenriched area and a bulk density of 0.086 g cm−1 (Craft and Richardson 1993b) and %C = 47.3. o,pMicrobial biomass (DeBusk and Reddy 1998). qPeat accretion from Craft and Richardson (1993b)

354

R.G. Qualls and C.J. Richardson

Fig. 13.2 Sources and calculation of fluxes and pools for the P-enriched area of WCA-2A. Superscript letters in the diagram correspond to the footnotes: aNPP = net primary productivity (aboveground = 913 g C m−2 year−1 plus belowground = 146 g C m−2 year−1). Aboveground NPP from Davis (1989), sawgrass site. Belowground NPP from Toth (cited in Davis 1991). Assumes dry mass = 45% C. bRoot litter input assumed to equal belowground NPP since the living biomass is not accreting. cLitterfall is assumed to equal belowground NPP since the living biomass is not accreting and turns over several times per year. dDecomposition and CO2 efflux from litter during the first year of decomposition. Data from Qualls and Richardson (Chap. 17), using C loss at 1 year in cattail litter from the channel receiving the highest PO4 dose minus the initial content of soluble carbon. eDecomposition of peat older than 1 year = NPP − first-year litter decomposition − peat accretion rate (modified from Craft and Richardson 1993b to include belowground NPP and to subtract the first-year decomposition). Because fragmentation was minimal all C lost in addition to the soluble C was assumed to lost as CO2. fCO2 efflux = decomposition rate of peat − methane flux. gMethane flux measured by Chanton et al. (1993) in a cattail stand (including through vegetation) in WCA-3A using the cattail site with the highest biomass converted from mg CH4 day−1 to g C m−2 year−1. hLeaching of DOC from freshly senesced litter (from this chapter). i DOC imported and exported in canal water inflowing into WCA-2A (from this chapter). Note that since the import and export included all of WCA-2A, the export from the enriched area is not distinguished from the unenriched area although the export from the enriched portion is probably greater. jMicrobial decomposition of DOC (from this chapter). kAbiotic mineralization (from this chapter). lPlant biomass, from Davis (1989), average of cattail sites A and B. mPool not measured but is variable over the year. nAssuming average 2-m depth of peat in unenriched area and a bulk densities reported in Craft and Richardson (1993b). o,pMicrobial biomass in units of concentration (DeBusk and Reddy 1998). qPeat accretion from Craft and Richardson (1993b)

13 Carbon Cycling and Dissolved Organic Matter Export in the Northern Everglades

355

Fig. 13.3 Sources and calculation of fluxes and pools in the drained EAA. Superscript letters in the diagram correspond to the footnotes: many pools and fluxes are not estimated. e,f,qFrom Tate (1980) based on average rates of subsidence corrected for changes in bulk density. All loss assumed (as in Tate) to be lost as CO2 although some loss in soils with a high water table might produce CH4. iDOC import from Lake Okeechobee canals. DOC export in canals to Water Conservation Area, see Sect. 13.2. pFrom Tate (1980). qAverage concentration of DOC in farm canals (from this chapter)

13.2.3

Production of Soluble Organic Nutrients in Plant Litter

One source of DOM is the leaching of water-soluble organic matter from litter produced directly by plants. To measure this flux, recently senesced leaves were gathered from cattail (Typha domingensis), sawgrass (Cladium jamaicense), spikerush (Eleocharis cellulosa), and sugar cane (Saccharum saccharinum) during autumn. In order to harvest only material which had not been leached, only plant material above the water surface was gathered in a period when the leaves had not been leached by rain for at least 14 days. The samples of cattail and sawgrass leaves were gathered in both P-enriched and nonenriched areas of WCA-2A. Field moist samples were ground and shaken for 6 days in three successive aliquots of water adjusted to pH 7.8 and preserved with 40 mg l−1 HgCl2 to prevent microbial growth. Our objective was not to simulate natural leaching rates, but to estimate the total content of water extractable substances. The total DOC, dissolved organic N (DON), and dissolved organic P (DOP) extracted was measured and expressed

356

Table 13.1 Soluble C, N, and P extracted with water from freshly senesced plant litter. Reproduced from Table 2, Qualls and Richardson 2003, with kind permission of Springer Science and Business Media

Plant

Annual production Annual production Annual production % initial C % initial N % initial P Primary productivitya of soluble organic of soluble organic of soluble organic Pb extracted extracted extracted (g m−2 year−1) (g m−2 year−1) Cb (g m−2 year−1) Nb (g m−2 year−1)

Cattail, nonenriched area 10.8 Cattail, enriched area 9.6 Sawgrass, nonenriched area 5.6 Sawgrass, enriched area 6.0 Spikerush 8.4 Sugar cane 11.3

12.4 7.1 10.0 7.4 20.6 36.3

17 10 20 24 26 60

1,077 3,035 986 1,943 NDc ND

54 139 25.1 56.2 ND ND

4.3 13.4 4.5 13.9 ND ND

0.019 0.083 0.016 0.23 ND ND R.G. Qualls and C.J. Richardson

For C and N, only the dissolved organic form was included but for P, both the organic and inorganic forms leached from litter were included. Primary productivity data are taken from Davis (1989) a Expressed on a dry mass basis b Calculated as primary productivity × (g element extracted per g dry mass) c ND not determined

13 Carbon Cycling and Dissolved Organic Matter Export in the Northern Everglades

357

as a percentage of the initial total C, N, and P in the senesced tissue (Table 13.1). We used measurements of the NPP in WCA-2A by Davis (1989, 1991) to estimate production rates of soluble organic matter (C, N, or P) in litterfall: (net primary productivity) × (g of soluble organic element per g dry mass of senesced tissue) = annual production of soluble organic nutrients in plant litter.

13.2.4

Budget of DON for Water Conservation Areas

The South Florida Water Management District (SFWMD) maintained a budget of water, total N and total P for water entering and leaving the WCAs for the period 1978–1989 (SFWMD 1992). The water budget for the EAA was reported by Abtew and Khanal (1994). It was possible to calculate total organic N concentrations (including suspended particulate N) from these data. For each input or output point, during periods of water flow, we calculated the DON concentration as: (total Kjeldahl N–NH 3–estimated particulate N). DON and particulate N were measured separately only during the period 1978–1979, so for this period we calculated a regression between particulate N and turbidity. We then used turbidity as a predictor to estimate particulate N during the 11-year period (r = 0.69). Generally, less than 10% of the total N was particulate N, so the error in using a surrogate estimator was small. Flow weighted average DON concentrations were calculated for each input and output point, and then these were multiplied by the water flows reported by SFWMD (1992). Interpolation of concentrations to daily water flow measurements followed the methods outlined in SFWMD (1992). We also used the estimates of DON to indirectly estimate the import and export of DOC. In our samples analyzed for both DOC and DON, the correlation was very strong (R = 0.95) and was predicted as DOC = 23.8 x DON + 7.14.

13.3 13.3.1

Results and Discussion Overall Carbon Budgets

Aboveground NPP in the unenriched area of WCA-2A (587 − 62.9 = 524 g C m−2 year−1) (Fig. 13.1) was close to the mean for all nonforested wetlands of about 1,000 g dry mass m−2 year−1 (or about 450 g C m−2 year−1) summarized in a literature review by Brinson et al. (1981), although most of the values comprising this average were from marshes in colder climates. In contrast, the aboveground NPP for cattail in the enriched zone (Fig. 13.2) was about twice as high as this average but lay within the range of two values reported for Typha marshes (Brinson et al. 1981).

358

R.G. Qualls and C.J. Richardson

The values reported by Toth (cited in Davis 1989) for belowground NPP of 63 and 146 g C m−2 year−1 are at the low end of the range for belowground NPP in wetlands of 134–600 g dry mass m−2 year−1 (about 60–270 g C m−2 year−1) reviewed in Brinson et al. (1981). Given the difficulty of measuring fine root turnover, it is possible that these values for belowground NPP in WCA-2A are underestimates. The reported belowground NPP was 10.7% of total NPP in the unenriched area and 13.7% in the enriched area. Because over the long term we assume standing stock of living biomass is approximately constant and the cattail and sawgrass biomass turns over several times a year (Davis 1989), we assume that litterfall is equal to aboveground NPP and that the input of root litter is equal to belowground NPP. While we lack estimates of NPP for the EAA, it is likely to be very high in sugar cane agriculture. The calculation of the input of litter to the fields of the EAA is complicated by the burning of a large portion of the leaves and removal of the cane, but belowground NPP and root litter input may be substantial. Decomposition of litter in the first year after senescence is much higher in the enriched area (Figs. 13.1 and 13.2), and is caused by the influence of PO4 as a limiting factor to the decomposer microflora. In addition, cattail leaves decay faster than sawgrass leaves, even under similar conditions (Chap. 17). In the unenriched channel (used to represent the unenriched area), 27.5% of the C of sawgrass leaves was lost during the first year, but 5.6% of the litter was lost as soluble C, leaving only 21.9% as our estimate of the C respired. We assumed that losses due to fragmentation were negligible because the leaves appeared not to have been fragmented and the litter bags tended to retain fragments. We also assume that no C from the litter is lost to methane during the first year because of aerobic conditions in the first-year litter. The remainder was represented as input to the peat, although the age at which litter becomes peat is arbitrary. The length of time that litter remains standing above the water also influences the accuracy of this estimate. The nutrient-enriched Typha litter lost 61% of the C in litterfall, but 9.6% was water soluble and leached as DOC, leaving our estimate of 51.4% of the C being respired during the first year. Davis (1991) obtained generally similar decomposition rates during the first year of about 34% mass loss of sawgrass leaves and about 56% loss of cattail leaves in WCA-2A. Based on the carbon budgets (Figs. 13.1 and 13.2), we view the decomposition of litter in the first year or two as a critical phase in the accumulation of peat. Before it is buried, it is exposed to aerobic conditions as indicated by dissolved oxygen measurements and redox profiles (Qualls and Richardson 1993; Qualls et al. 2001; see Chap. 17). Consequently, the acceleration of decomposition while the litter is still in the aerobic zone results in a dramatically reduced proportion of the litterfall making the transition to peat, as depicted in Fig. 13.2. Root litter is shed directly in the anaerobic zone (as indicated by redox potential profiles; Qualls and Richardson 1993) so we hypothesize that it makes a much greater contribution to peat than is indicated by its relatively low contribution to NPP. The estimates of peat accretion for WCA-2A have been detailed elsewhere (Craft and Richardson 1993b). The estimates of gaseous CO2 plus methane loss from peat (as distinguished from first-year litter decomposition) are somewhat less

13 Carbon Cycling and Dissolved Organic Matter Export in the Northern Everglades

359

in the enriched area despite nearly twice as much NPP input (Figs. 13.1 and 13.2). This is because over half of the loss via decomposition occurs during the first year. Thus, most of the difference in decomposition and respiration of CO2 between the unenriched and enriched areas occurs during the early stages of decomposition in the aerobic zone. Because these estimates are calculated as the difference between other measurements, the error may be considerable. The effects of drainage, however, are dramatic, with losses of 3,300 g C m−2 year−1 via soil respiration in the EAA (Tate 1980), some seven times the efflux of CO2 plus CH4 from the unenriched area (Fig. 13.3). Tate made an interesting comparison in stating that in the EAA the peat lost each year took 50 years to accumulate. This analogy was based on peat accretion rates from other regions, so now we can refine this statement. If peat in the EAA prior to drainage accumulated at the same rate as in the recent period in the unenriched area of WCA-2A, then the amount lost each year took 38 years to accumulate. Tate (1980) also pointed out that the soil respiration rate from the EAA was greater than from any ecosystem reviewed by Schlesinger (1977). The mass balance approach used in the carbon budget appears to suggest that the decomposition of subsurface peat is not greater in the enriched zone of WCA-2A compared with the unenriched zone (Fig. 13.1 vs. 13.2). Aerobic incubations for 6 months of this subsurface peat, however, clearly show that P amendments increase the potential decomposition rates (Qualls and Richardson 1993; see Chap. 17). Amador and Jones (1993) incubated peat from the southern Everglades under aerobic slurry conditions and found that PO4 additions stimulated respiration of samples initially low in P concentration but did not stimulate respiration in soils with initially high P content. DeBusk and Reddy (1998) also monitored CO2 and CH4 evolution in lab incubations of samples taken from the enriched and unenriched areas. Under aerobic incubation conditions, peat from the enriched area in WCA-2A respired faster than peat from the unenriched area, and this was attributed both to the difference between the lower lignin concentration of cattail vs. sawgrass and to the higher initial P concentration of the peat. Under anaerobic conditions, litter and peat samples from the enriched area respired more than samples from the unenriched area, but again, the correlation of lignin content (cattail vs. sawgrass) and P content made it difficult to isolate the influence of initial P concentration, except in litter. Thus, it seems clear that P enrichment increases decomposition of surface and subsurface peat under aerobic conditions whether to peat is formed from cattail or sawgrass. Thus, during droughts when peat in the enriched zones is aerobic, the exogenous P in the pore water and the better substrate quality of the cattail peat may combine to dramatically accelerate decomposition of subsurface peat, just as is the case with the surface litter. The effect of P enrichment on anaerobic respiration of subsurface peat is not as clear, however. Microbial biomass C concentration, measured by DeBusk and Reddy (1998), was higher in the enriched area in litter and surface peat than is the unenriched area, but in peat from the 10- to 30-cm depth there was little difference (Figs. 13.1 and 13.2). They noted that the microbial biomass comprised 0.5–4% of the soil C in the unenriched area, a value typical of range of other soil types. Microbial biomass in Penriched litter, however, was higher than found in most soils. In contrast, Qualls and

360

R.G. Qualls and C.J. Richardson

Richardson (1995) found little or no correlation between microbial biomass P concentration in peat and distance along the P-enrichment gradient. However, in litter, Qualls and Richardson (2000) found large increases in microbial biomass P in response to experimental P additions (see also Chap. 17). The estimate of microbial biomass C in peat of the EAA (Tate 1980) is relatively low and was based on an estimate of biomass necessary to yield the observed respiration and may only reflect “active” biomass.

13.3.2

Methane Emission

The interest in estimating global budgets for methane emissions has stimulated several studies of methane in the Everglades as a representative of subtropical wetlands. The measurements we have used in our C budget come from the work of Chanton et al. (1993) in WCA-3, but the live plant biomass data in their plots appeared to be representative of the sawgrass community in the unenriched area of WCA-2A (Figs. 13.1 and 13.2). We also chose the cattail plot used by Chanton et al. (1993) with the closest biomass to that of Davis’s (1989) site in the enriched area of WCA-2A. Chanton et al. used a chamber that included both the water and emergent plants, and we thus regarded this as the most reliable measurement made in the area. The measure of emission rate from sawgrass (Fig. 13.1) by Chanton et al. (1993) was quite consistent with other studies in the southern Everglades, with means ranging from 26 to 45 mg CH4 m−2 day−1 (Chanton et al. 1993; Whiting et al. 1991; Bartlett et al. 1989). Exceptionally, tall sawgrass greater than 1 m in height had an emission rate of 71.9 mg CH4 m−2 day−1 in the study of Bartlett et al. (1989). One factor which would likely contribute to higher emission rates in the cattail community of the enriched area is that cattail exhibits pressurized ventilation of roots through aerenchyma and has been shown to emit relatively high amounts of methane while sawgrass relies on diffusion to ventilate roots (Chanton et al. 1993). In addition, Chanton et al. (1993) noted a positive relationship between methane emission and net ecosystem production. Further evidence of higher methane production in the enriched area was provided by Koch-Rose et al. (1994), who found higher methane concentrations in soil pore water and sediment bubbles in the enriched area. This higher methane concentrations might be caused by higher substrate (e.g., acetate) production in the enriched area associated with higher decomposition rates. Qualls and Richardson (1993) and Qualls et al. (2001) found only small differences in soil redox potential along the nutrient-enrichment gradient so these differences are unlikely to be caused by differences in redox potential. A large proportion of the methane produce in the Everglades peat is oxidized in the surface soil or in the overlying water. Happel et al. (1993) found that 91% of the methane produced in anaerobic soil was consumed by methane-oxidizing bacteria near the soil water interface. Along the nutrient-enrichment gradient, sharply decreased concentrations CH4 near the surface of the peat also suggested methane oxidation or rapid diffusion into plant roots. Consequently, the major route

13 Carbon Cycling and Dissolved Organic Matter Export in the Northern Everglades

361

for methane emission is through plants. Chanton et al. (1993) found that only 6% of the methane emission from the cattail plot was from the water surface, with the rest emitted through plants.

13.3.3

Dissolved Organic Carbon

Water draining from wetlands often contains high concentrations of darkly colored DOM. DOM carries not only carbon, but also nitrogen, phosphorus, and metals complexed by the carboxylic acid functional groups. DOM functions in several important ways in wetland ecosystems: 1. It is a major mode of export for N and P in many ecosystems. 2. Because it represents an important form of export, it plays a major role in determining the balance and accumulation of N, P, and perhaps even C over the long period of soil development. 3. Because natural DOM is a powerful agent for the complexation of metals, it plays an important role in metal toxicity and metal export, e.g., mercury (Mierle and Ingram 1991). 4. It provides a potential source of carbon for microbial growth (Wetzel et al. 1995). 5. It controls the pH of many wetland waters (Hemond 1980; McKnight et al. 1985). We believed that it was important to understand the geochemistry of DOM in the Everglades because there is a growing concern about the transport of nutrients and mercury (Strober et al. 1995) into the southern Everglades and Florida Bay as well as many other wetlands, and also because our understanding of the processes which control the concentration and export of DOM in wetlands is limited. Anthropogenic changes in the Everglades environment may affect the fluxes of DOM. The P enrichment of certain areas may have an influence on production of DOM. Over 90% of the nitrogen and a large fraction of the P in water are in the dissolved organic form (Qualls and Richardson 1993; Qualls and Richardson 2003). The Everglades, at a subtropical latitude, also experiences relatively high levels of solar UV radiation, which could increase in the future if stratospheric ozone continues to decline. Solar UV radiation has been shown to alter aquatic DOM (Keiber et al. 1990; Bushaw et al. 1996). The processes regulating the concentration and export of DOM are depicted in Fig. 13.5 and are discussed in detail in the following sections.

13.3.4

Distribution of Dissolved Organic Carbon in WCA-2A

Dissolved organic C decreased with distance south along the nutrient-enrichment gradient in WCA-2A, both in surface and soil pore water (Fig. 13.4). The highest concentrations of DOC in surface water occurred in the marsh near the Hillsboro

362

R.G. Qualls and C.J. Richardson

Fig. 13.4 DOC in surface water and soil pore water (12.5-cm depth below the soil surface) along a P and plant productivity gradient in WCA-2A. The “weighted distance” follows Qualls and Richardson (1995) and is not exactly equal to geographic distance from the source of input, the Hillsboro canal. The diamond data symbol represents inflowing water from the Hillsboro canal. Error bars indicate ±SE for 7–8 samplings over time at each plot. Regressions were significant (p < 0.05). DOC concentrations at the 25- and 60-cm depth also declined significantly with distance along the gradient (not shown). Reproduced from Fig. 7, Qualls and Richardson 2003, with kind permission of Springer Science and Business Media

Canal, not in the water entering from the Canal (represented by distance of 0), so the pattern of concentration did not simply reflect dilution of runoff entering from the canal. The pattern of DOC concentration follows a general gradient of plant production (Davis 1991) and soil P concentrations (Qualls and Richardson 1995). The concentrations of DOC in soil pore water tend to be 1.5–2 times those in the surface water. These concentrations of DOC are high compared with most natural waters (Thurman 1985).

13 Carbon Cycling and Dissolved Organic Matter Export in the Northern Everglades

363

The decrease in DOC concentrations along the north-to-south gradient in WCA-2A might be explained by several hypotheses that center on three aspects: production, regulation of concentration, and degradation. Higher concentrations of DOM may be caused by (1) greater plant production of soluble organic matter, (2) greater production of soluble organic byproducts of decomposition in areas where peat is being decomposed faster, (3) regulation of concentration by solubility or adsorption equilibrium with soil of varying composition, (4) less mineralization by solar radiation in areas where the water surface is more shaded, or (5) faster rates of biodegradation in some environments. Using the data for WCA-2A supplemented with concentration data from other areas of the Everglades (Qualls and Richardson 2003) we found that the DOC concentration in marsh surface water was significantly correlated to the log of our index of approximate plant biomass (r = 0.72). There was also a similar relationship for DOC concentration in shallow pore water (r = 0.80). There was also a significant, but poorer, relationship between the log of DOC concentration and the depth of peat (r = 0.55). Although this index of plant biomass was intended only as a crude comparison of plant biomass among a large number of sites, it suggests that higher plant productivity contributes to higher DOC concentrations. Davis (1989, 1991) found that within either sawgrass or cattail communities, NPP was proportional to aboveground biomass because leaf turnover was consistent, supporting our inferred association between biomass and NPP. We also have used our data on soluble organics and along with Davis’s (1989, 1991) aboveground primary production data, to estimate the average annual production of soluble organic nutrients in WCA-2A in plant leaf litter (Table 13.1). These fluxes are represented in Figs. 13.1 and 13.2 as coming from the litter pool. As such, our estimate is an underestimate of all soluble organic matter produced by plants because we did not consider soluble organic matter exuded from living leaves, algae, roots, and peat decomposition. Nevertheless, the production of soluble organic C alone (up to 139 g m−2 year−1) rivals NPP in many lakes (Schlesinger 1997). This greater plant production and species change in the enriched area result in a higher production of soluble organic C, N, and P in litterfall (Table 13.1; Figs. 13.1 and 13.2). Senescent cattail leaves contain more soluble organic C than sawgrass leaves, so the changes in species composition have also affected the production of soluble organic matter. There was only a 7% greater soluble C content in sawgrass leaves in the enriched area compared with those from the unenriched area. Thus, the most important factors influencing production of soluble organic C, N, and P originating from litter was the difference between cattail vs. sawgrass and the difference in NPP along the gradient. These patterns suggest an interesting interaction of the P, C, and N cycles in which agricultural runoff of phosphate causes increased primary production and a change in species composition, both of which act to increase the production of DOC, DON, and DOP. Over 90% of the P leached from the senescent leaves was PO4. The soluble P remaining in the leaves after resorption may have been largely hydrolyzed by phosphatase enzymes. This effect might help explain why the organic P content of the DOM is so much lower than the content in live leaf tissue. Similar results have been reported for autumn deciduous tree litter (Qualls and Haines 1991).

364

R.G. Qualls and C.J. Richardson

The extremely high agricultural plant production in the EAA might also contribute to the high concentrations of DOC in the EAA. Senesced leaves of sugar cane contained as much soluble organic matter as sawgrass leaves, but it is difficult to estimate production since part of the plant is burned during harvest and, consequently, this flux is not represented in Fig. 13.3. Higher plants certainly produce soluble organic matter directly, but another possible source of soluble organic nutrients is by the action of microbes decomposing solid peat. However, when peat was incubated for 21 days in HgCl2 to suppress microbial activity or non-poisoned solution, there was no difference in the DOC produced, suggesting microbial dissolution was a minor source (Qualls and Richardson 2003).

13.3.5

DON Budgets for the Marsh Areas of the Everglades

The SFWMD has not maintained routine measurements of DOC that directly allow calculation of import and export from the WCAs, but their measurements do allow calculation of budgets for DON. Because DON is highly correlated with DOC, we will use it as a surrogate. Despite the high concentrations of dissolved organic nutrients in the marshes and high production of soluble organic matter by plants and decomposers, there is no net export of dissolved organic nutrients from most of the marsh areas (Table 13.2). There is a large export of DON in runoff from the WCAs, but large amounts also flow into the areas. In a sense, the DON flowing in and out of WCA-2A might be construed as simply flowing through, but we will show that turnover is likely to be considerable. The WCA-1 and WCA-3A marshes are a sink for a Table 13.2 Budgets of DON, color, and water for the Everglades Agricultural Area (EAA) and Water Conservation Areas (WCA) 1, 2A, and 3A. Reproduced from Table 3, Qualls and Richardson 2003, with kind permission of Springer Science and Business Media Material EAA WCA-1 WCA-2A WCA-3A DON (g m−2 year−1)

Water (cm)

Import Export Net export Runoff in Runoff out Rain Evapotranspirationa

0.3 1.4 1.2 24.7 59.6 117.5 82.6 NA

3.0 2.5 −0.5 99.0 94.2 120.7 129.9 0.36

3.9 3.9 0 162.5 187.5 113.4 133.5 0.2

1.4 0.9 −0.8 77.0 48.1 110.7 133.2 0.73

Water residence time (year) 2,401 588 448 2,063 Area (km2) The budget covers the period 1 October 1978–30 September 1988 and is derived from data furnished by the South Florida Water Management District (SFWMD). The water budget and total N budget were reported directly by the SFWMD (1991). The water budget for the EAA was taken from Abtew and Khanal (1994). The DON budget was estimated from the total N budget as summarized in the text. Net export is export minus import a For EAA, calculated as rain + runoff in − runoff out. For WCA areas, calculated from pan evaporation

13 Carbon Cycling and Dissolved Organic Matter Export in the Northern Everglades

365

portion of the DON entering the areas. The drained soils of the EAA export less DON on an areal basis than the marshes, but they show a high net export because the source of water for the EAA is rainfall and water from Lake Okeechobee that is relatively low in DON concentration (Table 13.2). The relative amounts of runoff in and out of the EAA and WCAs also influence the import and export of DON. Rainfall in excess of evapotranspiration accounted for a greater net runoff out of water from the EAA, while WCA-2A was dominated by runoff moving through the area. The peat soil in the EAA has subsided by as much as 2 m since it was drained (Tate 1980). Most of this subsidence can be attributed to aerobic oxidation (Tate 1980), but Tate speculated that a substantial portion of the loss of peat could have been via leaching of DOC. Using our typical DOC/DON ratio of about 20 (Qualls and Richardson 1993), we estimated a net export of DOC of about 24 g m−2 (Fig. 13.3). This is only a small portion (0.7%) of the estimates of annual carbon loss of 3,300 g m−2 in the soil of the EAA (Tate 1980). It may seem paradoxical that there was no net export of DON from the marshes of WCA-2A, because we have the potential for high production of soluble organic matter in litterfall (Table 13.1). One explanation is that decomposition of the DOM balances its production. Another explanation is that high concentrations of DOM can suppress dissolution or desorption of additional DOM. Previous experiments showed that peat can desorb or release potentially soluble organic matter into solution but high initial DOC concentrations can suppress this desorption or even cause DOC to sorb to the solid peat (Qualls and Richardson 1993). Peat soil from a sugar cane field in the EAA exhibited the same tendency to desorb and sorb DOC as the marsh peat except it initially contained much more potentially soluble organic matter, releasing 45.5 mg l−1 DOC into the water with no added DOC. The equilibrium DOC concentration was much higher (about 97 mg l−1) however. This high equilibrium concentration corresponds to the high DOC concentrations found in pore water in other surface soil in the drained agricultural area. Sequential extraction of the peat from the 0- to 5-cm depth in WCA-2A indicated that approximately 2.7 mg C g−1 dry mass of peat, or 0.58% of the C was potentially soluble in water (Qualls and Richardson 1993). The pool of potentially soluble organic matter was less than 1% of the carbon in the peat sample but the very large amount of this organic soil underlying the water column corresponds to large pools of potentially soluble organic matter. The capacity of this pool to buffer DOC concentrations over long periods of time was indicated by persistently high concentrations leached even after a large number of extractions (Qualls and Richardson 1993). This sorption–desorption phenomenon in peat has a large capacity to buffer DOC concentrations and could explain the patterns of export of DOM from the ecosystems of the Everglades (Table 13.2), assuming that DON follows the same patterns as DOC (as indicated by a correlation between DON and DOC of R2 = 0.91; Qualls and Richardson 1993). Inputs of rain and irrigation water into the agricultural area from Lake Okeechobee with low DOC concentrations allow the net desorption of DOC and a high net export (Fig. 13.3). Surface runoff entering the marshes from the EAA already contains high concentrations of DOC and DON, and this suppresses the further net desorption of potentially soluble organic matter (perhaps

366

R.G. Qualls and C.J. Richardson

even resulting in a net sorption of DOC). There was no net production of DOC. Thus, there is no net export from the marsh areas (Figs. 13.1 and 13.2). The high production rates of soluble organic matter in the enriched area of WCA-2A probably add to the reservoir of potentially soluble organic matter and shift the equilibrium concentration higher compared with the less productive areas.

13.3.6

Abiotic Mineralization of DOC by Solar UV Radiation

We exposed filter sterilized DOC in water from the surface pore water to sunlight in autoclaved quartz tubes and found that 20.5% of the DOC was mineralized by sunlight in 21 days of exposure (Qualls and Richardson 1993, 2003). Controls covered with aluminum foil did not decline in DOC. At the same time, 7% of the DON was mineralized to ammonium. Subsequently, Bushaw et al. (1996) have also shown that ammonium can be released photochemically from aquatic DOM. They reported that between 7 and 16% of the DON was converted to ammonium over a 3-day incubation, results that are similar to ours given that our kinetic experiment showed that most of the photochemical release occurred within the first few days. Thus, this newly discovered pathway in the nitrogen cycle has been demonstrated in aquatic samples from different locations. Keiber et al. (1990) found that irradiation of natural DOM with sunlight for periods up to 50 h resulted in formation of low molecular weight aldehydes cleaved from humic matter. Most production was due to wavelengths between 290 and 315 nm. They found that the humic fraction was the most photochemically labile. The photochemical mineralization experiment was designed only to show the maximum potential effects on DOM at the water surface during summer sunlight. We consequently represent this abiotic DOC mineralization as a maximum flux in Figs. 13.1 and 13.2. These maximum effects in the water column will be reduced by (1) shading by plants and periphyton and (2) the limited transmission of light around 320 nm (generally 33–75% transmittance cm−1; Qualls and Richardson 1993) caused by the humic substances themselves. Increased water depth may not necessarily decrease the maximal reaction rates on an areal basis but will tend to dilute the effects as the water column mixes. Thus, from our experiment we derived a first order decay constant of 0.0109 day−1. Then, using an average concentration of 70 mg l−1 of DOC in the enriched area and an average water depth of 0.36 m we calculate a maximum rate as (13.1) for the enriched area (Fig. 13.2) 0.0109 day-1 x 365 day x 70 mgl-1 x 103 lm-1 depth x 0.36m depth = 90 g m-2 year -1

(13.1)

In our sunlight-exposed samples, there was a rapid bleaching, particularly in the 300–360 nm range (Qualls and Richardson 1993) and a lowering of the DOC/DON ratio. These effects suggested that an unusually low absorbance per unit DOC might serve as a “signature” of the influence of solar radiation in the field. Indeed,

13 Carbon Cycling and Dissolved Organic Matter Export in the Northern Everglades

367

surface and shallow pore water from the southern Everglades, areas with less shading by macrophytes, were more bleached at 320 nm and had a lower DOC/DON ratio than did water from areas with more shading (Qualls and Richardson 2003). Other hypotheses, however, might explain these patterns such as a greater percentage of algal derived nonhumic DOC.

13.3.7

Biodegradability of DOM

Experiments on water from WCA-2A showed the potential rate of biodegradation of DOC in water both in darkness and previously exposed to sunlight (Qualls and Richardson 1993, 2003). The organic matter that remained after exposure to sunlight underwent important alterations that enabled further mineralization by microorganisms. The unexposed DOM in shallow pore water was not very biodegradable. The addition of inorganic N and P did not stimulate the biodegradation of the DOM, but the DOC remaining after exposure to solar UV light was rendered somewhat more easily degraded by microorganisms. The natural water lost only a few percent of the initial DOC even after 184 days, even with additions of inorganic nutrients. Thus, while the inorganic nutrients present in agricultural runoff appeared to stimulate production of DOM, they did not stimulate decomposition of the DOC. The humic fractions isolated from dark controls and light-exposed samples lost less than 5% of their DOC to decomposition over 184 days. The hydrophilic acid fraction, however, was rendered far more susceptible to decomposition by prior exposure to sunlight. The UV light selectively destroyed the humic substances, so it may also increase the biodegradability of the remaining DOC by removing the inhibitory effects. The UV light also generated some smaller molecules from the hydrophilic acid fraction that are more accessible to bacterial enzymes. Overall, maximum potential solar UV radiation alone in the surface layer of Everglades water can initially degrade DOC much faster than microbial decomposition, but the two processes appear to work in tandem. Other studies have noted photolysis of aquatic DOC (Geller 1986; Keiber et al. 1990; Mopper et al. 1991; Bushaw et al. 1996) and a subsequent increase in biodegradability (Geller 1986; Keiber et al. 1990; Mopper et al. 1991; Wetzel et al. 1995). Wetzel et al. (1995) found that exposure of natural aquatic organic matter to a laboratory source of UV-B radiation resulted in generation of significant quantities of low molecular weight fatty acids and some degradation of lignin derived phenolic components of the natural macromolecules. Wetzel et al. (1995) have also found that solar UV can release alkaline phosphatase complexed to humic acid. From laboratory incubations at 24°C, a temperature close to the average water temperature in WCA-2A of 24.6°C, we obtained a first order decay constant of 0.0033 day−1 for water incubated in the dark but previously exposed to sunlight (in order to separate the processes of solar abiotic and biotic mineralization). Substituting this decay constant into (equation 13.1) we obtain 27.6 g m−2 year−1 as an estimated rate of biodegradation of DOC (Figs. 13.1 and 13.2). This can only be

368

R.G. Qualls and C.J. Richardson

considered as an approximation because several factors may vary in the field such as the surface area of microbial biofilms on plant surfaces. In addition, this underestimates the biodegradation of the labile sugar fraction in newly senesced litter because it is probably degraded locally near the fallen litter. Because the sum of biodegradation and abiotic mineralization exceeds the input of soluble organic matter in litter, while import and export balance, there must be additional sources of input of soluble organic matter such as direct leaching from live plants, root exudation, and desorption from solid peat. Because there is production and decomposition of DOC that is on the same order as the import, the balance of imports and exports suggest the role of equilibrium sorption/desorption in regulating concentration and export in the water.

13.3.8

Contribution of Organic Anions to the Ionic Balance

The carboxylic acids (R-COOH), which are part of the organic acids in DOM, were a significant percentage (averaging 4.5%) of the total anions in the water of WCA-2A (Qualls and Richardson 1993). While the concentration of organic anions was high (0.34 meq l−1), it was dwarfed by the relatively high ionic content of the calcareous water. The charge contribution of carboxylic acids was estimated by assuming humic substances contain 10.1 meq g−1 C and hydrophilic acids contain 13.2 meq g−1 C (for a peat bog water; McKnight et al. 1985). Shuman (1990) reviewed eight studies and found a remarkable consistency in the value for aquatic humic substances of 10.5 (±1.2 SD) meq g−1 C. This calculation is greatly simplified at the pH range of 7.1–8.1 because the carboxylic acid functional groups are completely ionized. We used our data on the percentage of the DOC in the humic and hydrophilic acids fractions to estimate 9.15 meq g−1 total DOC in water of WCA-2A. The accuracy of the estimate of the charge contribution of carboxylic acids was also supported by close charge balance of cation vs. anions in samples (sum of cation charge = 0.98 × sum of anion charge + 0.050, in meq l−1).

13.4

Conclusions and Lessons for Restoration

Carbon cycling in the unenriched area of the Everglades is characterized by a NPP which is relatively average for wetlands, slow decomposition of litter due to high lignin content of sawgrass (Chap. 17) and P deficiency, low peat accretion rates, the emission of mostly CO2 from peat because of this slow initial decomposition, and relatively low methane emission. Phosphorus enrichment results in a general acceleration of carbon cycling. Compared with the unenriched area, in the P-enriched area NPP and litterfall is almost doubled, production of soluble organic C in senescent litter is nearly tripled, decomposition rate of litter is nearly doubled due to

13 Carbon Cycling and Dissolved Organic Matter Export in the Northern Everglades

369

Fig. 13.5 Processes controlling the concentration and export of dissolved organic matter (and DOC). Reproduced from Figure 1, Qualls and Richardson 2003, with kind permission of Springer Science and Business Media

P enrichment and the low lignin content of cattail, total respiration from first-year litter is four times higher, peat accretion is higher, and methane emission is likely to be over four times as high. Furthermore, a much greater proportion of the total respiration of litter plus peat is channeled into respiration during the first year when litter is exposed to aerobic conditions. The effects of P enrichment on the carbon cycle are largely due to direct effects of P on plants and microbes as a limiting nutrient. Furthermore, the species change to cattail results in higher quality detritus, which affects decomposers, probably produces more fermentation products for methanogens, and allows efficient methane venting due to the characteristics of cattail. The effect of drainage in the EAA results in the rapid respiration of C accumulated over hundreds of years of NPP. Thus, emissions of CO2 are about seven times those of the unenriched area. The concentrations of dissolved organic nutrients in the Everglades appear to be controlled by local plant productivity and sorption equilibrium on the solid peat. Solar radiation has the potential to play a role in regulating concentration and export because it can mineralize significant amounts of the DOM. Although the Everglades water contains extraordinarily high concentrations of DOM, much of the DOM seems to be relatively resistant to mineralization by microbes. Natural

370

R.G. Qualls and C.J. Richardson

levels of solar radiation in summer sunlight near the water surface can potentially mineralize the dissolved organic nutrients much faster than microbes can. However, the solar radiation and microbial degradation work in tandem and complement one another to some extent. Despite high concentrations of dissolved organic nutrients in the marshes, high production of soluble organic matter by plants and decomposers, and considerable export, there is no net export of DON from most of the marsh areas. In fact, the EAA, with the lowest concentrations of DON in the water entering the area, is the only one to have a net export. High concentrations of DOM in runoff from “upstream areas” enter the marshes and suppress the net dissolution of more DOM. This phenomenon of sorption equilibrium can control dissolved organic nutrients over large areas. We hypothesize that these processes interact in the following way (Fig. 13.5). Potentially soluble organic matter is produced by plants, and possibly as a byproduct of microbial decomposition. Sorption and desorption from the peat serves as reservoir and buffer, strongly regulating concentration. High plant production of potentially soluble organic matter increases the concentration of sorbate and can shift the equilibrium to higher concentrations in water. Phosphorus enrichment may act indirectly to increase DOM concentrations by increasing plant productivity and encouraging the growth of species that produce more soluble organic matter. Microbial decomposition and solar radiation can mineralize some of the DOM but desorption and diffusion from the peat act to renew the concentrations in the surface water. Excessive drainage and phosphorus additions will result in significant shifts in the carbon cycle within the Everglades, primarily as accelerated losses of carbon as CO2 and increases in plant productivity and decomposition rates, respectively. These shifts will greatly alter ecosystem stability, peat accumulation rates, and export of DOC from the ecosystem.

Part III

Everglades Experiments

A. Phosphorus Dosing

14

Introduction to a Mesocosm Approach for Establishment of Phosphorus Gradient Experiments Curtis J. Richardson and Robert R. Johnson

14.1

Introduction

In the late 1980s and early 1990s, the major question being asked about the Everglades was: What concentration of phosphorus could be added to the ecosystem without resulting in an imbalance of flora and fauna? While ecological studies had been undertaken along known P gradients in areas like Water Conservation 2A (Koch and Reddy 1992; Craft and Richardson 1993b) it was also clear that the ecological responses along this gradient were also influenced by other nutrient additions, including Na, Ca, N, and S; thus, the shifts in species and changes in water and soil chemistry could not be totally attributed to P additions. Moreover, it became essential to determine if a P threshold existed in the Everglades, i.e., to determine at what total P (TP) concentration significant shifts in ecosystem structure and function occurred. This chapter outlines the design, construction, and operational features of a replicated dosing system created to establish a gradient of known water TP concentrations. Importantly, the dosing facility site was selected from among many sites tested for uniform low background concentrations of TP. The final selection of the two replicate undisturbed site locations in the lower part of WCA-2A was based on sites with uniform slough species distributions, low P concentrations (< 400 mg kg−1 TP) in the soil, and TP background water concentrations of ≈10 µg L−1 TP. Importantly, the dosing site was allowed to equilibrate from construction effects for 9 months prior to dosing trials. Test trials on the system were run to provide information on improving system water pumping efficiency and consistency as well as uniform phosphorus distributions and concentrations in the water column prior to the commencement of the 6-year P dosing experiment (1993–1998).

14.1.1

General Objectives

The general design objectives were to: (1) design an in situ experiment to determine the threshold level of phosphate additions necessary for changes to occur in water chemistry and soils and, in turn, algal, macrophyte, and invertebrate species as well as Everglades slough ecosystem structural and functional changes; (2) provide a 375

376

C.J. Richardson and R.R. Johnson

uniform gradient of TP concentrations ranging from background concentrations of 10 µg L−1 P to over 100 µg L−1 P by using both treated and untreated channels using four different levels (30, 50, 75, 150 µg L−1 SRP) of phosphorus input concentrations in the treated channels along with (untreated) background walled and unwalled controls, which received only background water; (3) replicate the experimental site in a similar habitat; and (4) record physical and chemical soil and water parameters: temperature, pH, oxygen, water depth, redox potential, rainfall, etc., at each site on a regular schedule (hours, days, weeks, months, or quarters) depending on the variable and relate these potential controlling variables to ecological responses measured at 1-m locations along the P gradient established in each channel.

14.2

Experimental Setup

We used a mesocosm-scale experiment at two sites to assess the biotic responses of the algal–plant–macroinvertebrate complex in a 6-year P dosing in a slough community located in an undisturbed area of the northern Everglades (see Plate 7). Two P dosing facilities (hereafter called “flumes”) were constructed in the unimpacted interior of WCA-2A in Autumn 1990 and then calibrated for over 1 year. The two sites were almost identical in macrophyte species composition and structure during the calibration year prior to treatment (Richardson et al. 2000). Flumes were constructed in open-water sloughs, a habitat that has been shown to be sensitive to P enrichment (McCormick and O’Dell 1996; Vaithiyanathan and Richardson 1998) and ecologically important (Turner et al. 1999). Each site had five walled flumes, 2-m wide × 10-m long, with walls approximately 90 cm in height above the slough substrate. An additional unwalled control area of identical size to that of the flumes was established on the west side of each dosing site to concurrently monitor the potential effects of placing walls around slough habitat. Flumes were oriented N-S and separated by 1 m, where permanent boardwalks were built to allow investigators access for sampling. Each flume channel was randomly assigned one of five soluble reactive phosphate (SRP as Na2HPO4) treatments and given an assigned channel designation: ∼5 µg L−1 (mean background concentration [walled and unwalled controls]; 0.25 g m−2 y−1), 30 channels, ∼22 µg L−1 (1.5 g m−2 y−1), 50 channels, ∼39 µg L−1 (2.75 g m−2 y−1), 75 channels, ∼57 µg L−1 (3.5 g m−2 y−1), and 150 channels, ∼126 µg L−1 (8.2 g m−2 y−1). SRP was dosed from the northern end of flumes via large mixing tanks. Flumes were dosed on a continuous schedule except during low- or high-water shutdowns or periodic maintenance to specific flumes. Dosing was applied from 30 November 1992 to 21 September 1998. Greater detail on the design and operation of the P dosing study is described by Richardson et al. (2000). The dosing system created a total phosphorus (TP) and SRP gradient down each flume. Water was sampled every 2 weeks for SRP, and TP analyses were done at each meter (1–8) down the length of each channel. At each distance, samples were

14 Introduction to a Mesocosm Approach for Establishment of Phosphorus

377

taken at mid-depth from two points 25 cm on each side of the central axis. When water depths were greater than 25 cm, samples were taken at 1/3 and 2/3 the depth in the water column to adequately represent the average concentration in the plane perpendicular to the channel axis. Thus, when water depth was greater than 25 cm, a total of four samples were taken at each meter’s distance down the channel. Between March 1993 and September 1998, once each month, water was sampled for a suite of chemical analyses. This included not only PO4 but also TP, dissolved organic P, particulate P, ammonia, nitrate + nitrite, dissolved organic N, particulate N, calcium, and potassium. The samples were taken at 2-, 4-, and 6-m distances down the channel length. For example, TP gradients ranged from a mean high of 75 µg L−1 at 1 m in the highest treatment flume to 15 µg L−1 at 8 m. Control channels averaged 10 µg L−1 TP throughout the flume. Sample collection methods, storage, analysis, and QA/QC procedures were in accordance with standard methods (APHA 1992). Biotic responses and water chemistry were measured at each meter location, and then selected metrics were compared to the geometric mean water column TP concentrations for either the previous 3-month (algal responses) or 6-month (macrophytes, macroinvertebrates, and community metrics) period. Other time periods (1-, 2-, 8-, and 12-month means) were also tested, but the 3-month and 6-month periods were chosen since they most closely represent the biological life of each trophic level, and values among other tested time periods proved to be more variable. Daily variations in water depth and temperature were measured at both sites using an Omni Data automated data logger water level system coupled with a Metri-Tape unit and a thermocouple probe placed 10 cm below the water surface. Standard methods and EPA approved methods were used for all analyses.

14.3 14.3.1

Methods Location and Physical Layout of Design

In 1991, two experimental sites were established on the southwest side of WCA2A, 1.5 miles (2 km) northeast of the fish camp known as Old Glory, at 80°22′30² latitude by 26°15′00² longitude. One site was approximately 200 m east of the other. Each site contained a dosing flume system constructed in the open-water slough community. The system consisted of one pump platform, one battery platform, one solar array tower, five walled channels, one nonwalled channel, and access walkways surrounding the experiment (Fig. 14.1). The channels were approximately 2-m wide by 10-m long. The depth of water in the channels during the year ranged from a few centimeters (during dry season) to over 1 m during the rainy season. The channels were oriented north–south, the same direction as natural water flow. There was a 1-m separation between channels with a suspended walkway to further insure isolation of the channel waters.

378

C.J. Richardson and R.R. Johnson

Fig. 14.1 The flume site layout for the P dosing system in the Everglades indicating treatment channels, controls, equipment locations, and walkways used in the experiments

14.3.2

Construction

The construction of the two sites was divided into six sections consisting of: (1) the channels, (2) the pump systems, (3) the access structures, (4) the power arrays, (5) wiring and plumbing between sections, and (6) signs and site marker systems (Fig. 14.1). The pump systems and pump/control platforms were constructed first, along with the location and assemblage of all materials for the construction of the channels. Upon completion of the pump/control platform boxes and final tests of the pumping systems and controls, the platforms and materials for the construction of the channels were flown by helicopter and/or transported by airboat to the sites. When the installation of the pump platforms was completed, the next step was to install the channels. To insure that the soils and plants in each channel would not be disturbed, specific pathways were marked out and personnel used only these pathways for access to the site. Steel pipe, 3.1 cm in diameter, was driven into the bedrock and the PVC sheets were inserted to a depth of 30 cm. They were then attached to the support pipes with plastic cable ties and secured on the ends (15-cm overlap) with glue to insure water-tight integrity. The installation of the mixing system consisted of installing a 1.2 m × 1.8 m (4′ × 6′) pipe frame platform north of the head of each of the walled channels and at the height of the walkways to insure a downhill flow of the water from the mixing tank to the head box. The pipe frame was covered with a sheet of plywood upon which the 378.5 L (100-gallon) mix tank was positioned. Water reached the mixing tank from the pump platform via a 2.5-cm hard plumbed PVC line. A head box was installed at the head of each channel to insure that water pumped into the channel was evenly distributed both horizontally and vertically

14 Introduction to a Mesocosm Approach for Establishment of Phosphorus

379

(Fig. 14.1). It was constructed of molded fiberglass with a perforated PVC sheet on the channel side. The head box was lowered into the channel and attached to the end pipes at the head of the channel. The head box was fitted with PVC sheets on the sides, firmly attached with self-drilling/tapping screws and foam pads to seal it to the channel. Fiberglass lids covered the tops of the end boxes and the feed pipe entered the box through a notch cut in one of them. The internal distribution riser provided a smooth flow of water to the full horizontal width of the box, thereby allowing even flow of water from the head box into the channel through a perforated face plate sheet comprised of > 2,000 evenly distributed holes. Flow distributions were checked in each channel using colored dye markers to assure a proper vertical and horizontal flow at the various depths that are found at the site throughout the year.

14.3.3

Design and Materials

14.3.3.1

Mixing Tanks

At the head of each channel, a tank was used to mix the concentrated phosphorus with slough water to bring it to the proper experimental concentration. Inside each mixing tank were magnetic reed switches that signaled the control system to turn the pumps on and off. The water exited the tank through a 1.2-cm PVC metering valve that controlled the flow rate.

14.3.3.2

Pump Platform

The pump/control platform, constructed of steel and plywood, was located east of each experimental site (Fig. 14.1). It housed: (1) five 12-Vdc Jabsco water pumps capable of delivery of 42.4 L min−1 (11.2 gpm), (2) four 12-Vdc peristaltic pump assemblies, (3) four 100-L phosphorus concentrate supply tanks, (4) two electrical control panels, (5) one 12-Vdc battery box (control voltage), (6) one Omnidata 900 series Easy Logger water temperature/depth computer unit, (7) one two-panel solar array on the lid (south side), and (9) a Metri-tape water level unit just outside the platform on the south side.

14.3.3.3

Battery Platform

The pump batteries were on a steel pipe frame platform located to the south of the pump platform (Fig. 14.1). Seven fiberglass battery boxes housing 20 deep-cycle 12-Vdc marine batteries were attached to the plywood base. These five separate battery systems (5 × 4) powered the pumps and were recharged through 12-Vdc ASC regulators by solar panels located south of the battery platform.

380

14.3.3.4

C.J. Richardson and R.R. Johnson

Solar Array

The solar array consisted of: (1) a pipe frame assembly, anchored into the bedrock, which supported a 5-in. (12.7 cm) steel stand pipe support frame; (2) the auto tracker frame, which mounted onto a steel support pipe; (3) ten solar panels wired in a 5 × 2 array; and (4) the wiring harness to bring power to the batteries.

14.3.3.5

Channels

Polyvinyl chloride (PVC) sheet stock 25 cm × 10 cm × 0.6 cm (10² × 4′ × 1/4²) was used to create the walls and 3.1-cm (1¼²) galvanized pipe fastened together with Kee Klamps were used to construct the frame. The pipe was driven through the peat to a depth of approximately 1 m (3–4 ft.) and into the limestone bedrock between 0.5 and 1.0 m (2–3 ft.). The pipes also provided the framework for the walkway supports. The PVC walls were sunk in the peat approximately 30 cm (1 ft.) to prevent leakage and were attached to the pipe with cable ties.

14.3.4

Systems Operation

14.3.4.1

Plumbing

The concept for the operation of the dosing system is shown in its plumbing design (Fig. 14.2). It used a self-priming, battery operated water pump to lift slough water through a screened well head and then transfer this water to a mixing tank at the head of each channel. Approximately 25 cm (10² ) below the intake side of the water pump was a fitting connecting the output of a peristaltic pump to the input of the water pump. The peristaltic pump removed a specified volume of concentrated phosphorus solution from a stock drum and pumped this into the intake side of the water pump. This action took place approximately 60 s after the water pump turned on to insure that water was flowing to the mixing tank prior to the introduction of the P concentrate. The water continued to flow after the peristaltic pump shut off to insure that all of the concentrate entered into the water line went to the mixing tank and had time to fully mix with the slough water. The mixing tanks held 378.5 L (100 gallons) of water. The flow from the mixing tank was continuous; therefore, only 246 L (65 gallons) of new water reached the tank at a time. The rest was separated by a baffle board, allowing the new water to mix with the concentrate prior to introduction to the slough. The concentrate consisted of approximately 250 mL of concentrated phosphorus solution requiring 7–10 s of peristaltic pumping out of a total water pumping time of 6 min. The water reached the mix tank at a rate of 42 L min−1 (11 gpm) and left at 1.89 L min−1 (1/2 gpm), allowing ample time for complete mixing to take place.

14 Introduction to a Mesocosm Approach for Establishment of Phosphorus

381

Fig. 14.2 The path of water flow into the flume system. Water exits the system into the headbox (not shown) and then downstream into the channels via the diffusion system

The water exited the mix tank via a metering ball valve. This valve allowed for very fine adjustments to be made to insure a steady rate of flow leaving the tank. The flow rates were checked and calibrated weekly. Upon leaving the tank and dropping approximately 60–120 cm (2–4 ft.) in height, the water was diverted laterally and released through a series of holes in a diffusion pipe into the head box. The south side of the head box was covered by a panel of perforated PVC that allowed the water to flow evenly from the box into the channel. The depth of the water in the channel was used to determine the total turnover rate when the flow to the head box was fixed at a specific rate. The maximum flow rate was determined by the power available to operate the pumps and recharge the system. A rate of 1.89 L min−1 (1/2 gpm) provided 2,725 L (720 gallons) of water per day. This allowed for a channel turnover of approximately once per day during the dry season and once every 7 days at the maximum depth that the walls permitted during the wet season. When the water rose over the tops of the walls during any portion of the year (tropical/hurricane disturbance) or when the water fell below 10 cm (drought), the system was shut down for that period of time. The shut-down insured that the waters were not over dosed, did not mix between channels, or cause erosion in the soil within a channel. To the south of the sites, plastic construction fencing was set in two rows 2–3 m (6–10 ft.) apart to keep wind and wave action on the open end of the channels to a minimum. The fencing helped protect the channel from south winds that could blow water and debris into the channel as well as from disturbance by airboat

382

C.J. Richardson and R.R. Johnson

wakes. It also assisted in keeping any large animals (e.g., alligators) out of the channels where they could damage experimental equipment and plants. The ability of the system to produce the desired uniform concentration of phosphorus throughout the channel was tested using a low concentration of sodium bromide as an inert tracer to test for distribution patterns. Test trials and head adjustments, along with pumping rate changes, allowed for a uniform mixing pattern to be determined. The factors affecting the test were the volume of water in the channel and the strength of the bromide solution. This test did not, however, take into account the effect of adsorption of phosphorus to the system itself; thus, a 9-month period of time to allow the system to equilibrate was used prior treatment.

14.3.4.2

Electrical

The location of the experiment in the Everglades made the design of a power system capable of operating the experiment a unique challenge. The movement of water is a costly proposition and is paid for in units of electrical power. The experiment location, 30 min away by airboat from the nearest power source, required a low-maintenance design rugged enough to take the environmental conditions of the south Florida Everglades. (Note: Hurricane Andrew destroyed our solar array tower). Each site had one control system, complete with its own batteries and solar chargers, and five separate pump systems. Each pump had its own power and regeneration system, but they all shared a common control system. The pump system for the walled control at both sites did not have a peristaltic pump as part of its system since no P was added. As water drained from the mixing tank it eventually closed the “on” magnetic reed switch in the tank, turning on the magnetic locking main control relay via coil contacts A and B, thereby activating power to two circuits. The first provided 12-Vdc power across the coil of the motor power relay, turning on the water pump. This, in turn, provided power across the peristaltic safety relay coil (insuring that concentrate could not be pumped unless the water pump is running). The second circuit that was powered by the main control relay was the peristaltic time delay relay coil. Upon power being applied to this relay, a delay circuit counter was energized and a period of 60 s was counted off before the second part of this relay was activated. Once the 60 s was reached, the peristaltic pump energized for a period of approximately 7–10 s, which allowed for the pumping of approximately 250 mL of concentrated phosphorus into the water line. The time delay relay then opened and waited for the power to be disconnected and then reapplied before operated again. The water pump continued to run until the mix tank reached a level where the water rise closed the “off” magnetic reed switch and turned off the main control relay by closing coil contacts and reversing the magnetic lock on the relay. This, in turn, de-energized the system until contacts were once again closed by the “on” magnetic reed switch. The 12-Vdc motors were powered by four 12-Vdc deep cycle, 102-A h marine batteries wired in parallel. The batteries were recharged by two Siemens M-55 solar modules wired in parallel and producing 106 W at 17 Vdc and 6.05 A. The power

14 Introduction to a Mesocosm Approach for Establishment of Phosphorus

383

from the solar array was passed through an ASC photovoltaic battery charge regulator and then on to the batteries. The solar panels were mounted on Zomeworks Passive Trackers to enable the panels to receive the greatest amount of sunlight available by tracking the sun while not using any of the system’s power to do so. The control circuits (coil and switch voltages except for safety relay) were all powered by a separate control power system. This system used two 12-Vdc batteries similar to the pump batteries, which are charged by two Hoxan H-4810 solar modules in parallel producing 17 V at 96 W and 5.6 A. These were fixed mounted on the top of the south side of the pump/control platform, and also fed through an ASC photovoltaic battery charge voltage regulator to insure proper charging.

14.4

Summary

The flumes as designed provided a replicated mesocosm-scale phosphorus dosing system that almost continuously delivered over a 6-year period a low-to-high range of phosphorus input concentrations in 12 channels in an undisturbed slough in the northern Everglades. Low levels of soil water column SRP and TP as well as total soil and vegetation TP were initially measured at the site. A vegetation survey prior to treatment found that both sites were similar and representative of an undisturbed Everglades slough in the Northern Everglades.

15

Water Quality, Soil Chemistry, and Ecosystem Responses to P Dosing Curtis J. Richardson, Panchabi Vaithiyanathan, Robert G. Qualls, Mark B. Bush, Craig A. Stow, and Mengchi Ho

15.1

Introduction

Our mesocosm study consisted of flumes or channels established in 1991 in two unimpacted slough communities in unenriched areas of WCA-2A (see Figs. 14.1 and 14.2, Chap. 14, Plate 7). The word “mesocosm” here refers to an experimental enclosure where only phosphorus concentrations were chemically manipulated. The enclosures were large enough to contain essentially all important components of the biological community. The advantages of using such an approach – as opposed to using “microcosms” in small containers – is that one can utilize an existing natural community with minimal disturbance and subject it to the natural environmental conditions except for the experimental treatment (Odum 1984). The phosphorus dosing research was designed to determine the P concentration required to significantly alter periphyton, macrophyte, and macroinvertebrate communities as well as ecosystem structure and function. The two sloughs were initially similar in macrophyte species composition – both being dominated by Gulf Coast and viviparous spikerushes (Eleocharis cellulosa and E. elongata, respectively), white water lily (Nymphaea odorata), purple bladderwort (Utricularia purpurea), and macroalga Chara sp. (stonewort). Twelve flumes (four treatment channels and an unwalled and walled control in each slough) were used to manipulate surface water orthophosphorus (PO4), total phosphorus (TP) concentrations, and P loadings to establish gradients in the channels from 1993 to 1998. Importantly, it must be recognized that PO4 is not biologically equivalent to TP. Total P may include a large proportion of particulate or organic P fractions which are not as readily available to organisms but its concentration in the water column is related to the amount of PO4 that the organisms have been exposed to in each channel. Orthophosphate (referred to as PO4 or soluble reactive phosphorus (SRP) in this volume, as noted in Chap. 6) is the most available form of P that organisms can readily absorb (Wetzel 2001). Phosphate is always measured and reported as the ion PO4-P in this volume. Our approach is similar to that of Tilman (1990) in establishing nutrient enrichment gradients. Each of the different average input concentrations over the 6 years established a different and decreasing PO4 and TP concentration gradient down the length of each channel, creating uniform P zones that could be sampled at discrete

385

386

C.J. Richardson et al.

intervals (1 m) for water column PO4 and TP concentrations as well as other ions. The PO4 concentrations represent the residual PO4 in the water column after uptake, sedimentation, and precipitation (Richardson and Craft 1993). It is important to note that the concentrations at various distances in the channel proportionally follow the upstream dosing in the channel mixing tank (especially in the first 2 m) but do not represent the initial mixing tank dosing concentrations. The distinction here is that the amount of PO4 remaining in the water column above the control channels’ PO4 concentration represents the amount of PO4 in excess of short-term uptake and storage processes (see P model of Richardson and Craft 1993). This means that a distinction must be made between the PO4 doses that was applied in each channel and the residual PO4 found in the water at any location. However, both of these variables relate to “P threshold” values. The P loadings to each channel are directly reflected in the residual P in both organism and sediments, while elevated PO4 in the water column above background only exists at those locations where uptake, sorption, and precipitation mechanisms are overloaded, which usually occurred in the upper portions of the channels (Richardson et al. 2003). We postulated that the highest doses of PO4 would result in the highest amount of TP, due to cumulative luxury uptake of PO4 and storage as organic P by biota. Extremely high flow rates might overcome the dominance of uptake to create a longer TP and PO4 gradient, but high flow could also create unnaturally high rates of diffusion downstream. Our mesocosm approach of establishing trends along an experimental P enrichment gradient is compatible with the large-scale release of PO4 and TP found in south of the S-10 structures on the Hillsboro Canal or near other inflow structures flowing into the Everglades. However, phosphorus speciation is different in the dosing channels vs. the WCA’s canal releases, since the canals discharge nearly a 50:50 ratio of PO4 to TP and the dosing facility discharges mostly PO4 (Qualls and Richardson 1995; Richardson et al. 1999; see Chap. 6). Thus, the ratio of available P to total P (i.e., PO4 to TP) is much higher in the channels (~80:20). That means that the amount of PO4 available to organisms is higher in the flume study and P thresholds based on this work would be a conservative estimate of system responses to total P dose. In addition, the nutrient gradients in WCA-2A and WCA-3A contain high concentrations of nitrogen species as well as major cations, chloride, and sulfate which also influence species distributions (Vaithiyanathan and Richardson 1997a); thus, the mesocosm experiment gives a more realistic organism response to P concentrations. The purpose of this chapter is to characterize the water and soil chemistry with a strong emphasis on phosphorus dynamics and trends. Information on water depth variations, diel oxygen, and pH provides supporting information for determining both P responses and effects. Ecosystem responses to P additions are presented via the P effects on diel oxygen production, respiration, gross primary productivity (GPP), and community respiration. In addition, measurements of alkaline phosphatase were done to assess P limitations in the channels and to assure that P additions affected the entire channel. Individual plant, algal, macroinvertebrate, and fish responses are presented in later chapters in this volume (Chaps. 16, 18, and 19). The influence of P on decomposition is given in Chap. 17. The assessment of a P threshold is covered in detail in the modeling section of Chap. 25.

15 Water Quality, Soil Chemistry, and Ecosystem Responses to P Dosing

15.2 15.2.1

387

Methods Dosing Operations

Water from the slough was periodically pumped into a mixing tank located at the head of each channel (Chap. 14). Float switches in the tank activated the pumps when the water level fell to a predetermined minimum level and turned them off when the tank was full. A concentrated solution of Na2HPO4 and NaBr (about 250 ml) was pumped into a T-junction on the influent side of each water pump by a peristaltic pump for a 7–10 s period while the water was being pumped into the mixing tank. Consequently, a PO4 concentrate was mixed with natural slough water usually containing < 5 µg l−1 PO4 in a regulated ratio to produce a given concentration in the mixing tank. Each different PO4 input required a different concentrate (Table 15.1). Concentrates were made up in deionized water in the laboratory and were delivered to the holding tanks as needed. The same concentration of NaBr was added to all concentrates to act as a conservative tracer to assess for uptake vs. dilution or channel leakage. A nearly constant flow of PO4-enriched or background slough water into the head of each channel produced a continuous drain rate of about 1.89 min−1 from the mixing tanks into a distribution pipe that drained into a perforated head box. As water drained down to a level corresponding to 45% of the 379 l capacity, the switches activated the pumps and refilled the mixing tank. Since the tank drained continuously and the hydraulic head fluctuation was small, there was a nearly constant flow of water into the head of the channel. The flow gauges for each tank were calibrated weekly or after shutdown for any period. It was important to distribute the flow into the channel evenly in both the vertical and horizontal directions. A T-shaped distribution tube leading into a box with a perforated PVC plate was designed for this purpose. The perforated plate covered the entire cross section of the channel, and water trickled into the channel through the perforations. The downstream ends of the channels were open to water flow, but the openings were restricted by PVC baffle panels to avoid excessive mixing of outside water and to keep floating plants and alligators from entering (see Plate 7). The fixed flow of water into the dosing channels resulted in hydraulic residence times that varied with water level (Table 15.2). The dosing period for both dosing sites was similar in that pumping was scheduled to run continuously once the experiment started. Occasionally, a channel pump Table 15.1 PO4-P and Br concentrations in concentrated and diluted channel input water in the Duke dosing system Channel name

Diluted channel input Br (µM)

Concentrated PO4-P (mg l−1)

Concentrated Br (mM)

30 50 75 150

88.8 88.8 88.8 88.8

29.8 46.7 65.0 125.0

77 77 77 77

388

C.J. Richardson et al.

Table 15.2 Channel volume and average hydraulic residence time at various water depths assuming 1.89 l min−1 influent flow rate Length (cm) Width (cm) Depth (cm) Volume (l) Residence time (day) 800 800 800 800 800 800 800

200 200 200 200 200 200 200

10 20 30 40 50 60 70

1,600 3,200 4,800 6,400 8,000 9,600 11,200

0.59 1.18 1.76 2.35 2.94 3.53 4.12

or flow system would malfunction for several days and would require replacement, and monthly maintenance and calibrations required the system to be off for a few hours. In addition, there were several periods when no pumping occurred at the site for several weeks due to hurricanes or severe electrical storms. We also did not pump when water levels exceeded 120 cm (excessive water depths) or when drought or summer conditions reduced water depths to <10 cm. These periods of shutdown were recorded, and water chemistry measurements were always recorded as occurring during pumping or nonpumping periods. All dosing loads were calculated for the pumping periods. The period of pumping was from 23 November 1992 to 21 September 1998, which totaled 2,129 days. Site 1 pumped on average 62% of the time over the 6 years and Site 2 for 57% during the entire study period. However, there were some years when dosing days were higher or lower than the average. Importantly, our pumping schedule exceeded the number of days most of the SFWMD flow gates are open in any 1 year, and our shutdown times were also far below the number of consecutive days stations like the S-10 structures are closed in WCA-2A (SFWMD 1999). Thus, our system provided a realistic dosing schedule and P loadings to the slough communities. Finally, all major studies and analyses were completed only during periods of extensive continuous dosing.

15.2.2

Water Sampling

Water was sampled every 2 weeks (November 1992 and October 1998) for PO4, TP, and Br analyses at 0–7 m down the length of each channel, except during droughts or hurricanes. At each distance, samples were taken at mid-depth from two points 25 cm on each side of the central axis. When water depths were greater than 25 cm, samples were taken at 1/3 and 2/3 the depth in the water column to adequately represent the average concentration in the plane perpendicular to the channel axis. Thus, when water depth was greater than 25 cm, a total of four samples were taken at each meter’s distance down the channel. Between March 1993 and December 1998, water was also sampled once monthly for a suite of additional chemical analyses. This includes not only PO4 (also reported as SRP) but also TP, dissolved organic P (DOP), particulate P (PP), ammonia

15 Water Quality, Soil Chemistry, and Ecosystem Responses to P Dosing

389

(NH4-N), nitrate + nitrite (reported here as NO3-N), dissolved organic N (DON), particulate N (PN), calcium (Ca), potassium (K), chloride (Cl), and sulfate (SO4-S) at 1, 2, 4, and 6 m distances down the channel length. Samples collected on the two sides and two depths at a particular distance were composites. Simultaneously, samples collected from 1, 3, 5, and 7 m down the channel were analyzed only for PO4-P and Bromide (Br) in accordance with the regular biweekly sampling. In the spring of 1995, we began supplemental total P analyses at 0.5 m and each meter downstream biweekly, in addition to our previous scheme of measuring TP at 2, 4, and 6 m distance down the channel to assess P speciation (TP, PP, and DOP as well as PO4). Daily variations in water depth and temperature were measured at both sites using an Omni Data automated data logger water level system coupled with a Metri-Tape unit and a thermocouple probe placed 10 cm below the water surface. Oxygen, temperature, conductivity, and pH measurements were made with YSI units, which were calibrated weekly and placed at meter 2 in each channel. Thus, meter 2 PO4 and TP concentrations were used to compare with O2, respiration, and productivity estimates. GPP and community respiration were estimated from the diel dissolved oxygen (DO) curve method (Greeson 1977; Wetzel 2001). Gross productivity and respiration values reported in this study are uncorrected for diffusion. A weather station was used to continuously record means at 60-min intervals for water temperature, soil temperature, and solar irradiance (measured with a LiCor LI190SB quantum sensor). In situ periphyton mat pH and temperature measurements were made with an Orion model 250A pH meter. DO concentration was measured using a microelectrode OM-4 oxygen meter and dissolved calcium (Ca2+) concentration was measured using an Orion model 93-20 calcium electrode. The technique of R.G. Wetzel (personal communication), a modification of Pettersson and Jansson (1978), was used to determine alkaline phosphatase activity in all samples. Detailed methods and specific sampling protocols are given in Bush and Richardson (1995) and in Chap. 6. The specific methods used in the water chemistry analysis for the various constituents are given in Table 6.2. A detailed presentation of methods as well as procedures for lab and field protocols using blanks, spikes, and standards are given in detail in our FDEP QA/QC approved plan (DUWC 1995) and are outlined in Chap. 6.

15.2.3

Phosphorus Uptake, Dilution, and Channel Flow

The Br ion is a commonly used conservative tracer because it is very soluble, not taken up by organisms in significant amounts, and relatively unreactive with soil sorption sites. In this study, we added NaBr along with Na2HPO4 to measure the dilution of added solution. By measuring the difference between dilution of Br and the reduction in PO4 concentration, we could distinguish dilution from uptake of PO4. We tested two models of flow down the channels and derived an equation for measuring PO4 uptake. Definitions for some of the terminology can be found in “Chemical reaction engineering” (Levenspeil 1972).

390

C.J. Richardson et al.

Model 1. Model 1 is a channel closed at one end and open at the other. Ao represents the flow rate into the channel. The channel is divided into i segments (1 m segments for our channels). Ai represents the downstream (axial) flow rate from segment i to segment i + 1. An represents the downstream flow rate from the end of the channel. B represents upstream inputs of water from the end of the channel, caused, for example, by eddy dispersion. Bi represents upstream exchanges between segments caused by axial dispersion. Di and Ei represent exchanges of water with the environment through the walls or from the sediments. For model 1, we assumed simple open channel flow with no axial dispersion (plug flow). Thus B = 0 and Bi = 0. However, we assumed exchange of water with the environment by Di and Ei vectors and that Di = Ei. In other words, there is no net influx or efflux of water into the channels except the input Ao. Our equation for distinguishing uptake from dilution using a conservative tracer is (derived from Eq. (2) in Qualls et al. 1989) PO4 removed by uptake = ([BriPO4o/Bro]−PO4i)Ao / area of segment, where Bro or PO4o is the concentration in mixing tank minus background, Bri or PO4i is the concentration at distance i down the channel minus background concentration, Ao is the input flow rate, and area of segment = 2 m2 if i = 1 m lengths. Expressed more formally: PO4 uptake rate at any point in the channel = ([BriPO4i-1 /Bri-1]-PO4i)/Ai / area, n

PO4 uptake rate from point of input to i = ∫0 [((BriPO4i-1/Bri-1)-PO4i)A0 /widh]dl, where l equals distance down the channel. Model 2. In model 2, we considered advection down the channel with dispersion. At the end of the channel eddy dispersion resulted in entrainment of exogenous water. Again we assumed that exchange Di = Ei. Thus An = B + Ao. In this case the above equations also apply for PO4 uptake except the Ao should be substituted for Ai. Since PO4 uptake proved to be first order with respect to concentration, we considered each 1 m segment as a “continuously stirred tank reactor” and utilized the well-known equations for a first order chemical reaction in a series of continuously stirred tank reactors (Levenspeil 1972).

15.3 15.3.1

Results and Discussion Background, Biogeochemical Interactions, and Biotic Responses to PO4 Dosing

The sloughs sites (Figs. 1.2 and 14.1, Plate 7) selected for the dosing research had low P concentrations in the soil ≈250 mg kg−1 (0–15 cm depth) and background water conditions considered typical for unenriched areas in the northern Everglades.

15 Water Quality, Soil Chemistry, and Ecosystem Responses to P Dosing

391

Table 15.3 Mean monthly water chemistry in two sloughs selected in WCA-2A in 1991–1992 prior to initiation of the phosphorus dosing experiment Site 1 Site 2 Mean 6.75 Dissolved O2 (mg l−1) 875 Conductivity (µmho cm−1) Ph 8.45 T (°C) 23.8 5.8 PO4-P (µg l−1) 186 NH4-N (µg l−1) 23.8 NO3-N (µg l−1) 123 Na (mg l−1) 9.7 K (mg l−1) 48 Ca (mg l−1) 26 Mg (mg l−1) Samples were collected between 11 a.m. and 12 p.m.

SD

Mean

SD

0.67 144 0.27 6.8 3.9 63 11.4 10 0.4 1.4 1.3

7.11 952 8.27 23.8 6.5 204 17.9 142 10.3 49 27

0.58 170 0.28 7.3 3.9 8.5 1.8 15.7 1.6 5.5 4.2

Table 15.3 presents water quality within the slough from July 1991 to August 1992 prior to dosing trials in November/December of 1992. The values are, in general, similar to those reported by Swift (1981) and Swift and Nicholas (1987) for interior WCA-2A locations with the exceptions of pH and dissolved oxygen. Higher values for both these parameters measured in our study may be influenced by the midday sampling time we used. Seasonal water temperatures ranged from a low of 13°C in December 1992 to 33°C in September 1993. The two sites were almost identical in terms of their water chemistry. Slightly higher concentrations of phosphorus and nitrogen were found in our study site than the concentrations reported by Swift (1981) and Swift and Nicholas (1987), but the values were low and represented background concentrations for this section of the northern Everglades during the drought and higher nutrient loading conditions found in the early 1990s. The average PO4 concentrations applied to the dosing channels and the mean P loadings are given in Fig. 15.1. The concentration values ranged from a low of 5.4 µg l−1 of PO4-P in the walled controls to 127 µg l−1 PO4-P in the 150 channels. Both sites were comparable in terms of their long-term phosphate doses. The TP concentrations were nearly double the phosphate additions at the lower doses but only 10–20% higher in the higher dosed channels (Figs. 15.1 and 15.2). This was due to bacterial and some algal uptake in the mixing tank, which created higher TP values. (Note: Tanks were black inside and sealed from sunlight to prevent organism uptake as much as possible.) The 6-year average TP loadings were comprised of ≈ 80% PO4 and were very similar for each treatment at Sites 1 and 2. Concentrations and loadings of P in the dosing channels covered the range of values reported for the P gradient south of the Hillsboro Canal (Koch and Reddy 1992; Qualls and Richardson 1995; Richardson and Qian 1999). A comparison of the control channels’ TP values from 1993 to 1998 indicates that geometric mean TP values were close to 10 µg l−1 TP for most years but some years were below and some slightly above (Fig. 15.3). Of interest is the range of

392

C.J. Richardson et al.

Fig. 15.1 The 6-year mean phosphate (PO4-P) concentration and PO4-P mass loadings, which were added to each channel in the dosing experiment during the 1993–1998 period. Also shown is the mass loading of phosphate (in g m−2 year−1) for the 6-year study period (WC walled control, and 30, 50, 75, and 150 represent the channel names and not the dose). Note: channel order shown here does not represent the random channel treatment order assigned in the field

Fig. 15.2 The 6-year mean total phosphate (TP) concentration and TP mass loadings, which were added to each channel in the dosing experiment during the 1993–1998 period. Also shown is the mass loading of TP (in g m−2 year−1) for the 6-year study period (WC walled control, and 30, 50, 75, and 150 represent the channel names and not the dose)

values from near 27 to as low as 3 µg l−1 of TP. This reflects the changes in water depth that are found both seasonally and from year to year in the Everglades (Fig. 15.4). These water level changes are identical in nature to the yearly and seasonal trends reported by the SFWMD for their central station in WCA-2A

15 Water Quality, Soil Chemistry, and Ecosystem Responses to P Dosing

393

Fig. 15.3 The mean annual total phosphorus concentrations from 1993 to 1998 in the control channels from the Duke dosing experiment. Note that the ERC- and FDEP-approved total phosphorus criterion of 10 P µg l−1 is shown as a reference line, as well as the box plot of the individual values for each year

Fig. 15.4 The mean average water depth (cm) at both Site 1 and Site 2 in the Duke dosing study from September 1992 to August 1998. Mean daily average values were calculated and utilized in the computation of the average water depth

(SFWMD 1999). The highest TP values were found in 1993, 1994, and 1997, years with extended periods of low water levels. In contrast, the lowest TP values (mean below or at 10 µg l−1 TP) were found in the wettest years; in 1995 and 1998. The median TP values were mostly below 10 µg l−1 TP but values from the driest growing season (March–September) in 1994 were slightly above 10 µg l−1. In addition, concentrations of TP, total dissolved phosphorus (TDP), TN, TDN, NH4-N, and NO3-N in the surface water of the sloughs were significantly related to the water

394

C.J. Richardson et al.

depth fluctuations (p < 0.0001) (Vaithiyanathan and Richardson 1998). The concentration levels were relatively low at high water depths and vice versa. The water column PO4-P concentration showed no systematic seasonal trend and was not strongly related (p = 0.15) to the water level (Richardson et al. 1997c). The geometric mean SRP and TP concentration averaged over site and for each dosing channel treatment at each location over the 6-year period indicates that the treatments successfully created a range of phosphorus concentrations that decreased down each channel except for the control channels (Fig. 15.5). Soluble reactive P increased dramatically in the 75 and 150 channels over the amount found in the 30

Fig. 15.5 The surface water SRP and total phosphorus concentration in the dosing channels at various distances from the influent head boxes in the Duke dosing experiment in WCA-2A. The averages represent the geometric mean total phosphorus found at Site 1 and Site 2 during the dosing period (1993–1998). Note: 30, 50, 75, and 150 are channel names

15 Water Quality, Soil Chemistry, and Ecosystem Responses to P Dosing

395

and 50 channels. Soluble reactive P was lowest in the control channels and seldom exceeded 5 µg l−1 but often reached >10 µg l−1 in the higher dosed channels. The significant increases in SRP in the water column clearly indicate the ecosystem had exceeded its P assimilative capacity (Richardson and Qian 1999). SRP was quickly removed from the water column and was only elevated beyond 2 m down the channel in the 150 channels. The average for the entire channel was nearly identical in the 30 and 50 channels, but the range was higher in the latter channel, as were individual sites along the channel. The highest TP values were found in the 75 channels with both the 75 and 150 channels showing the highest variations in TP. Surprisingly, the 150 channels had average TP values below the 75 channels. This was not the case in the first year of the experiment when the plant communities were similar (Richardson et al. 1995). However, after a shift in the algal communities and a loss of many of the macrophytes in the 150 channels decreased TP values below those found in the 75 channels were measured. This may have been due to increased uptake by the new green algal plant communities. This reflects an early major shift in the plant communities and in turn different P uptake rates due to the high P loadings. Another important difference between the control channels and the dosed channels was the shift in the amount of SRP, DOP, and PP that made up TP (Fig. 15.6). In the controls channels, PP made up ~50% of the TP down the channels with SRP and DOP each comprising around half of the remaining 10 µg l−1 of TP. However, PP made up at least 60% and had values as high as 78% of the TP in the dosed channels, with PP concentrations ranging as high as 50 µg l−1 in the upper end of the highest two dosing channels. Of concern is not so much the fact that DOP values

Fig. 15.6 The percentage of soluble reactive phosphorus (SRP), dissolved organic phosphorus (DOP), and particulate phosphorus (PP) that comprise total phosphorus in the dosing channels in the Everglades

396

C.J. Richardson et al.

reached upwards of 22% of the TP but rather this is a very difficult form of P to remove from the water column and concentrations reached >10 µg l−1 in the higher dosed channels. The movement of this species of P downstream is of concern since it is more readily available to organisms than once thought (McKelvie 2005). To assess the channels for PO4 dilution effects as well as determine if SRP is being taken up by biota, a dynamic linear model was applied to the SRP and Bromide (Br) concentration data in each channel (Qian et al. 2003). The Br/SRP ratio should gradually increase as water moves down the channel due mainly to biotic uptake of SRP, since the dilution rate of both Br and SRP due to rainfall and backflow from the end of the channel should be the same if there are no other factors such as leakage through the channel walls. It was expected that the ratio would also drop near the end of the channel as the back flux of the outside water would increase. Analysis of Br/SRP ratios in each channel demonstrated that the channel flow and uptake behaved as expected. Bromide was reduced by 25–50% when input entered the head box (location −1 to 0) (Qian and Richardson 1997a), indicating a significant dilution in the head box, which was planned for. If reduction down the channel had been reduced by factors other than dilution, we would have noted a divergence from the predicted dilution rate, which did not occur. The results show that the dosing channels are sealed and well characterized by the conservative Br tracer, and importantly this allowed us to estimate P uptake vs. dilution in the system. Analysis of Br vs. PO4 trends down the higher dosed channels in the first full year of PO4 additions indicated significant uptake from the water column since the PO4 declined in concentration at a much faster rate than the Br tracer (Qualls et al. 1994). Average annual PO4 uptake is shown for selected channels in Table 15.4. The greatest P uptake occurs in the channel segments with the highest concentrations and these are the segments (meter sections) closest to the source of P input. The general trend from this first year water analysis showed a high rate of uptake in the first few meters, with much lower rates further down the channel. Both dosing sites displayed similar total channel uptake rates that corresponded closely with the initial annual SRP loadings of ~0.3, 1.5, 2.5, 3.5, and 8 g m−2 year−1 in the 0, 30, 50, 75, and 150 channels, respectively (see Qian and Richardson 1997a for a full discussion of modeling of SRP (PO4) ) and TP uptake and settling rates, respectively.

Table 15.4 Uptake rates of phosphate (PO4) in selected channels in the first year of dosing (23 November 1992–29 November 1993) Channel name 1–50

2–50

2–75

1–150

2–150

PO4 uptake (g m−2 year−1)

Distance down channel (m) 0–1 2–3 4–5 6–7

1–75

1.5 0.6 0.3 0.5

1.5 0.3 0.2 0.4

3.8 0.7 0.4 0.4

4.1 0.7 0.3 0.6

7.6 3.0 1.2 0.3

Uptake was calculated by the Br/PO4 method outlined in the text

5.5 5.3 0.8 0.9

15 Water Quality, Soil Chemistry, and Ecosystem Responses to P Dosing

397

A series of microcosm studies using water and periphyton from the channel controls was done to see if Phosphate (PO4-P) uptake was significant for the periphyton mat as well as to investigate potential P limitation in the Everglades sloughs (Vaithiyanathan and Richardson 1998). The microcosm experiments showed that PO4-P uptake rates were strongly related to the initial dissolved PO4-P concentration (r = 0.94; p < 0.0001). In all experimental PO4-P treatments, the background levels of PO4-P (2–3 µg l−1) were attained within a few hours. Nutrient analysis of the periphytic algae showed considerable differences in the P content of the algae increasing with increasing concentrations of P addition (Table 15.5). Changes in the C:N:P ratios were primarily due to the increase in P content of the algae. Periphyton and slough macrophytes in the background Everglades exhibited C:N:P ratios indicative of depletion in N and particularly severe P limitations relative to C (Table 15.5). The N:P ratios in periphyton (N:P = 151:1) were higher than in macrophytes (N:P = 56:1) suggesting that the extent of P depletion was even greater for periphyton. Stoichiometric C:P ratios >258 and N:P ratios >23 are indicative of severe P limitations in phytoplankton seston of lakes (Wetzel 2001) and our tested ratios greatly exceed this for our periphyton mat. Verhoeven et al. (1996) suggest that macrophytes with an N:P molar ratios >36 are P limited, which was also verified for Cladium jamaicense along a nutrient gradient study in the northern Everglades by Richardson et al. (1999). As a result, any introduced inorganic phosphate was predicted to be absorbed rapidly by the periphyton, which was demonstrated in our mesocosm experiments, reducing the water column PO4 to near background levels (Vaithiyanathan and Richardson 1998). The experiments also revealed considerable differences in the C:N:P ratios of the algae resulting from P additions (Table 15.5) whereas the water column PO4 concentration showed uniform low values in all the experimental P treatments after few hours. This suggests that the C:N:P ratios of the algae may provide a reliable measure of the P status in the Everglades sloughs whereas the water column PO4 concentration may not truly reflect the P status of the system. A recent radiotracer study in the Everglades suggests that the rapid rate of uptake PO4 from the water column we found in our channels was probably due initially to water column particulates (>0.45 µm) storing 95–99% of the added 32P (Noe et al. 2003). They reported that microbes controlled the short-term cycling

Table 15.5 Nutrient concentration of periphyton mat exposed to experimental PO4-P additions Water column

Periphyton mat

Treatment

Initial PO4-P (µg l−1) Total P (µg g−1) Total N (%)

Total C (%) C:N:P (M)

Control Low P Low med. P High med. P High P

4±1 16 ± 2 32 ± 3 45 ± 6 87 ± 2

25 ± 4 25 ± 2 25 ± 0 25 ± 1 26 ± 2

203 ± 25 325 ± 53 553 ± 97 488 ± 8 721 ± 298

1.38 ± 0.40 1.63 ± 0.17 1.65 ± 0.08 1.65 ± 0.05 1.74 ± 0.18

3,181:151:1 1,987:111:1 1,168:66:1 1,323:75:1 932:53:1

Results presented are mean ± SD from second experiment (9 September 1994) (from Vaithiyanathan and Richardson 1998)

398

C.J. Richardson et al.

and that over the course of their 18 days 32P experiment in the Everglades National Park, 32P moved from the particulates in the water column to periphyton to floc and then to soil. At the end of the experiment, they completed a budget and reported that floc contained 35%, soil 27%, floating periphyton 12%, and surface water only 10% of the added P. Emergent macrophytes displayed little 32P activity especially when compared to aquatic macrophytes (Noe et al. 2003). Importantly, they found that within the calcareous mat 81% of the initial 32P uptake was associated with Ca. These uptake and transfer trends closely follow the 32P cycling patterns found in a northern fen by Richardson and Marshall (1986), although Ca did not play as important a role in the northern fen. To further assess both the role of calcium carbonate (CaCO3) precipitation by the periphyton and the role of P concentrations on CaCO3 precipitation and PO4 coprecipitation, we calculated a saturation index (SI) from the water column Ca (40–85 mg l−1) and carbonate alkalinity (200–280 mg CaCO3 l−1) concentration (Vaithiyanathan et al. 1997). Results showed that the slough water was saturated with CaCO3 (SI 0–1) at the normal range of pH (7.5–8.5) and the temperature (15–35°C) observed in the Everglades sloughs. Calcium carbonate saturation was considerably exceeded (SI > 2.0) at the elevated pH (9.3) and temperature (36°C) conditions observed on the surface of the periphyton mat, resulting in its precipitation from the water column near the periphyton mat (Fig. 15.7a–c). Daytime

Fig. 15.7 Temperature (a), pH (b), and dissolved calcium (c) concentration on the periphyton mat and in the surface, middle, and bottom water column of the Everglades sloughs. Mat: periphyton mat. SW: surface water. MW: mid water column. BW: bottom water

15 Water Quality, Soil Chemistry, and Ecosystem Responses to P Dosing

399

field measurements showed that water column temperature and pH were relatively high on the surface of the periphyton mat compared to the open water surface and mid-water column. Water column temperature and pH values were the lowest near the soil–water interface. In situ calcium measurements were inversely related to pH (r = −0.99; p = 0.001) and temperature (r = −0.92; p = 0.09) (Vaithiyanathan et al. 1997). Calcium concentrations were the lowest on the surface of the periphyton mat (Fig. 15.7c). Low Ca2+ concentrations observed on the periphyton mat reflect the precipitation of calcium carbonate during algal photosynthesis. Laboratory experiments demonstrated the precipitation of calcium carbonate from the slough water with the increase in pH (Vaithiyanathan et al. 1997). Experimental results further showed that CaCO3 precipitation and phosphate coprecipitation from the slough water were strongly inhibited at concentrations >50 µg l−1 of PO4-P ions. Inhibition of CaCO3 precipitation by P was also confirmed in the in situ mesocosm experiments by the strong decrease in the ash content (measure of calcium carbonate content) of the periphyton mat with increase in water column P concentration (Fig. 15.8). Ash content of the algae collected from the periphyton mat and the plexiglass slides in the channels varied from 5 to 35 mg cm−2. Ash content associated with the algae in the Everglades is a reliable measure of its calcium carbonate content (Vymazal and Richardson 1992, 1995; Browder et al. 1994). Ash content of the algae collected from the highest P treatment channel was significantly (p < 0.001) lower compared to the control channels. A strong inverse relationship observed between the water column TP concentration and the algae ash content (r = −0.52; p = 0.00, see Fig. 15.8) suggests a decrease in calcium carbonate precipitation on periphyton with an increase in water column P concentration.

Fig. 15.8 Relationship between periphyton ash content and the water column TP concentration in the experimental channels of the dosing study

400

C.J. Richardson et al.

Decrease in CaCO3 precipitation may be one of the primary factors responsible for the fragmentation of the periphyton mat observed in the mesotrophic areas of the Everglades and the complete disappearance of the periphyton mat in the most eutrophic regions of Everglades and our channels. Six-year mean concentrations of Ca (65 ± 2 mg l−1), Mg (25 ± 1 mg l−1), Na (102 ± 4 mg l−1), and Cl (145 ± 5 mg l−1) were similar among treatment channels without any definitive relationship to P enrichment. These ions primarily showed only a significant difference (p < 0.05) over time (i.e., seasonal differences) (Richardson et al. 1997c). Since Ca is precipitated by periphyton, we expected to observe seasonal differences due to differences in productivity of calcareous periphyton, but none was shown when measuring Ca from the entire water column. The Ca concentration in water is high, however, and the volume of water is large in the channels, so the precipitation of Ca by periphyton may remove insignificant proportions of this ion from the entire water column. Alternatively, the change in these ions as related to P input may require longer P dosing for the effects to become evident. Calcium-to-magnesium ratios were similar to those reported for the WCA-2A gradient study (Qualls and Richardson 1992). Both Na and Cl average 4.6 mM in concentration. This nearly exact correspondence in Na and Cl molar ratios was noted in previous studies along the nutrient enrichment gradient in WCA-2A (Richardson et al. 1991; Qualls and Richardson 1995) and indicates an origin in sea salt deposits. Concentrations for K were uniform throughout all sites and treatments and averaged 7.6 ± 0.7 mg l−1. Sulfate (SO4-S) concentrations were also very consistent at all locations and averaged around 55 ± 55 mg l−1. The dosing channels cation and anion values were similar to pretreatment concentrations, except Ca was 35% higher for some unknown reason after 1992 (Table 15.3 values compared to 6-year Ca means given earlier). Trends in NH4-N concentrations showed significant patterns (p < 0.05) that could be attributed to enhanced uptake with P additions (Table 15.6). The controls at Site 1 had the highest concentrations of NH4-N compared to the P treatment channels, but there was no significant pattern in the other channels associated with increased P enrichment. The same general pattern existed at Site 2, although values were lower. Surprisingly, there was no consistent gradient with distance down the channel in NH4-N with increased P enrichment in the channels, which one might expect since P enrichment and stimulation of productivity has been shown to result in greater uptake of NH4-N (Vaithiyanathan and Richardson 1994). This earlier study in our channels did find a diel pattern of increased NH4-N and NO3-N uptake under increased P enrichment which corresponded directly to increased photosynthesis during the day (Table 15.7). The addition of PO4 resulted in the increased consumption of NH4-N from the water column as shown by the relatively low values of NH4-N in the 150 channel and progressively higher values in the 75, 30, and control (WC, UWC) channels (Table 15.7). The specific uptake rate for NH4-N always exceeded that of NO3-N. Ammonium seems to be the preferred source of inorganic N as suggested by the higher specific uptake rates of NH4-N compared to NO3-N. It has been known for a long time that ammonium suppresses nitrate uptake in many algal species and also in various higher plants. There are several reasons

15 Water Quality, Soil Chemistry, and Ecosystem Responses to P Dosing

401

Table 15.6 Ammonium–nitrogen (mean ± 1 SE) concentration (µg l−1) at 2, 4, and 6 m from each channel head box at the two dosing sites, 1993–1998 Channel

2m

4m

6m

Site 1 UC WC 30 50 75 150

90.82 ± 17.84 124.67 ± 30.66 58.48 ± 8.85 59.14 ± 7.48 60.44 ± 9.91 71.16 ± 10.87

82.33 ± 14.78 74.18 ± 16.99 110.55 ± 22.07 101.20 ± 23.93 50.48 ± 5.40 62.35 ± 13.04 61.41 ± 7.42 60.81 ± 8.08 56.18 ± 9.95 58.49 ± 11.15 76.45 ± 12.21 79.17 ± 15.33

Site 2 UC WC 30 50 75 150

55.09 ± 9.92 52.60 ± 7.24 66.85 ± 24.63 38.42 ± 4.22 36.13 ± 4.20 42.17 ± 4.86

63.31 ± 10.78 50.66 ± 7.91 33.19 ± 2.77 36.36 ± 4.09 29.86 ± 2.38 37.31 ± 4.35

51.52 ± 11.62 54.08 ± 9.71 34.11 ± 4.45 35.50 ± 3.62 33.08 ± 3.51 35.02 ± 4.30

Table 15.7 A diel pattern of NH4-N and NO3-N uptake with increased PO4 dosing in the mesocosms of WCA-2A on 24–25 August 1993 NH4-N (µg l−1) Channels

NO3-N (µg l−1) Channels

Time (h)

150

75

30

WC

UWC

150

75

30

WC

UWC

2 6 10 14 18 22

82 44 34 24 19 38

45 38 57 47 14 19

78 111 100 59 69 84

90 123 110 74 60 71

133 129 90 57 42 197

10 19 17 15 8 9

11 10 9 16 15 7

17 15 20 14 11 9

15 11 17 11 8 14

7 10 9 12 11 15

for this preferential assimilation of ammonium. Neither active nitrate reductase is formed in the presence of NH4+ nor is the NO3− uptake system. And even if active nitrate reductase and NO3− uptake system are present, the addition of NH4+ can lead to a rapid cessation of NO3− utilization (Vymazal 1995). Increased NH4-N uptake during photosynthesis was further evidenced by a strong negative relationship between pH and NH4-N (Vaithiyanathan and Richardson 1994). Thus, midday sampling for NH4-N and NO3-N will probably not reveal uptake patterns or give values representative of the daily change in ion concentrations. Vaithiyanathan and Richardson (1994) also observed a strong negative correlation between pH and dissolved calcium in the water column, which probably reflects enhanced CaCO3 precipitation in the Everglades waters, at higher pH values found during the middle of the day. Pore water pH (6.5 and 7.0) was relatively low compared to the surface water (Table 15.3, Fig. 15.9a) in the control channels. The relatively high pH values in the surface water reflect the loss of CO2 from aquatic photosynthesis. Concentrations

402

C.J. Richardson et al.

Fig. 15.9 Pore water profiles of pH (a), PO4-P (b), NH4-N (c), and Ca (d) in the control channels of the Everglades slough dosing study (from Vaithiyanathan and Richardson 1998) (reproduced from Journal of Environmental Quality, with permission, J. Environ. Qual. 27:1439–1450, 1998)

of PO4-P in the surface water and pore water were relatively uniform (Fig. 15.9b, Table 15.3). Similar values of PO4-P in the surface water and pore water suggest that organic matter mineralization may not be a significant source of surface water PO4-P in oligotrophic sloughs. However, pore water NH4-N concentration in the sloughs was substantially higher than surface water (Fig. 15.9c) and was within the upper range of values reported for oligotrophic wetlands (Moore and Bellamy 1974). The pore water NH4-N profiles suggest that pore water NH4-N flux across the soil–water interface represents a major source of NH4-N for the surface water (Fig. 15.9c). Dissolved calcium concentration was relatively low in the surface water compared to the pore water (Fig. 15.9d) due to the algae-mediated precipitation of calcium carbonate in the surface waters noted earlier. Sediment P chemistry from the dosing channels compared by treatment channels in 1995 (2 years following the start of dosing) and 1997 (4 years following the start of dosing) indicates that the P doses have resulted in an increase in TP concentrations

15 Water Quality, Soil Chemistry, and Ecosystem Responses to P Dosing

403

Fig. 15.10 The amount of total phosphorus found in the sediment (0–15 cm) at the dosing channels in both 1995 and 1997. The mean phosphorus values at each distance down the channel are shown for each channel. See Fig. 15.1 for channel name details

that closely correspond to the amount of P additions (Fig. 15.10). For example, in 1995 the highest sediment P was found in the 150 channels where it averaged nearly 1,000 mg kg−1 of P in the top 10 cm of sediment. By contrast, the control channels averaged around 250 mg kg−1 of P. Control channel values were similar to pretreatment concentrations. The 30, 50, and 75 channels showed values around 400–500 mg kg−1 of P in 1995 but had increased to values closer to 750 mg kg−1 by 1997 (Fig. 15.10). By late 1997, both the 75 and 150 channels displayed sediment P concentrations above 1,000 mg kg−1 and the 150 channels values approached 1,400 mg kg−1 near the discharge point. The sediment TP concentrations closely follow the values reported for TP along the nutrient gradient in northern WCA-2A (Koch and Reddy 1992; Craft and Richardson 1993a,b, 1997; see Chap. 6). This suggests that the phosphate additions in the dosing channels were able to produce sediment TP values that are similar to the long-term enrichment areas found in the Everglades. However, the deeper soil layers in the dosing study do not display P enrichment at depths of 10–25 cm as found along the long-term gradient (Richardson and Vaithiyanathan 1995; Vaithiyanathan and Richardson 1997a). This may suggest that the additions of P to the water column from the pore water may be greater along the gradient than those found at the dosing study after 4 years of dosing. Values and increasing P trends were similar after 6 years of dosing, with P values reaching >1,600 mg kg−1 in the most highly dosed channels (data not shown).

404

15.3.2

C.J. Richardson et al.

Biological Driven Responses to PO4 Dosing

A typical diel oxygen pattern for water in the treatment channels showed a midday peak for all channels at Site 1 (Fig. 15.11). The lowest oxygen concentrations were found just before dawn for all the channels, and highest values were found in midafternoon. The largest daily amplitude was found in the 50 and 75 channels and the lowest in the 150 channels. Oxygen values indicate near oxygen saturation of the water column at midday in the 50 and 75 channels, but they are suppressed somewhat in amplitude in the 150 channels as compared to the controls and other treatments. Box plots showing the daily mean and range of values for each site over a 3-year period from 1995 to 1997 indicate some differences among treatments for Site 1 as compared to Site 2 (Fig. 15.12). We chose this time period to show for comparison since it represented three full years of both wet and dry conditions, which follow the “typical” Everglades wet season/dry season shifts. Site 2 clearly displayed a drop in oxygen at the highest two treatment channels. The daily maximum values at Site 1 showed oxygen saturation for all channels but not in the 75 and 150 channels at Site 2. The daily minimums were also the lowest at Site 2 in the 75 and 150 channels. The largest range for oxygen values was found in the 50 channels. The range of oxygen values in the dosing channels was similar to that reported for stations found along the P gradient in WCA-2A (SFWMD 1999; Fig. 6.15). The SFWMD reported typical oxygen values in interior unenriched sites between 2 and 8 mg l−1. Our P dosing channels displayed similar concentration ranges except for the highest treatments, where the daily minimum values for oxygen

Fig. 15.11 Diel variation in dissolved oxygen in the water column in the experimental channels of the Duke dosing study at Site 1 during June 1995. This site and date were selected as representative of oxygen responses during treatments. Similar trends were found at Site 2 and in other summer months

15 Water Quality, Soil Chemistry, and Ecosystem Responses to P Dosing

405

Fig. 15.12 The daily mean, daily maximum, and daily minimum dissolved oxygen concentrations found in the water column at the Duke phosphorus dosing sites. Plots depict the mean 25th and 75th interquartile range of values found during the 3-year (1995–1997) period at both Sites 1 and 2

were near 1 mg l−1 (Fig. 15.12). However, our mean oxygen concentrations of 4 mg l−1 were nearly double those reported by the SFWMD for the most enriched areas along the gradient. This suggests that other factors such as additional nutrient concentrations of NH4-N, NO3-N, and SO4-S or shading from cattails and sawgrass may be influencing the oxygen concentration at certain sites areas along the gradient since TP water column values are similar at the dosing and gradient enrichment sites where open water exists (see further discussion on this in Chap. 6).

406

C.J. Richardson et al.

Fig. 15.13 The daily mean, maximum, and minimum water column pH found at the Duke dosing site. The data represent both Site 1 and Site 2 during the 3-year (1995–1997) period

Water column pH was fairly uniform over both sites and among the treatments (Fig. 15.13). The daily maximum pH at both sites was near 8.0 with the exception of the two highest treatments at Site 2. The mean was near 7.5 for most sites. The mean values of pH for the dosing sites were similar to those reported for water column chemistry for WCA-2A (SFWMD 1999). Of interest is the fact that the highest pH values were found in the middle range of P treatments and the lowest in the highest two channels, which lacked calcareous periphyton after the first year of dosing. The depletion of free CO2 at higher pH in the enriched channels (Vaithiyanathan and Richardson 1995) might also explain the decline of the submerged U. purpurea (Chap. 16) we noted in higher dosed areas, since this species

15 Water Quality, Soil Chemistry, and Ecosystem Responses to P Dosing

407

apparently can no longer photosynthesize in enriched channels with elevated daytime pH. Supporting this hypothesis is the work of Moeller (1978), who has shown that U. purpurea in New England lakes used only free CO2 and not bicarbonate as a carbon source for photosynthesis. Increased nutrient demand/consumption during photosynthesis also resulted in relatively low dissolved nutrient levels in the water column during the latter part of the day, particularly PO4 (Fig. 15.14). The depletion of PO4 during the day may be so low as to limit primary productivity in Everglades waters. However, increased P additions in the 150 and 75 channels have resulted in significantly higher PO4-P values during the morning hours (Fig. 15.14). It is likely that PO4 utilization by the algae and macrophytes may be proportional to its availability. It is also important to note that during the afternoon when the productivity is at its maximum, there were no major differences in the PO4-P concentration among the channels (Fig. 15.14). Our results clearly indicate that in a P-limited system such as the Everglades, PO4-P concentration of the water samples collected during the latter part of the day may often be misleading with regard to the extent of enrichment of the site. This suggests that further work on monitoring schedules needs to be undertaken in nutrient studies in the Everglades. All the dosing water treatment samples analyzed generally revealed trends in phosphatase activity that increased with distance from the PO4 dispensing header box. All water samples had some level of phosphatase/per unit chlorophyll (APA-c), indicating that this enzyme will be universally present (although in very low concentrations) in Everglades waters with higher PO4 concentrations (Fig. 15.15a). The increase in APA-c ranged by a factor of about 4 between the most and least

Fig. 15.14 Diel variation in PO4-P concentrations in the experimental channels of the dosing site in WCA-2A on 24–25 August 1993

408

C.J. Richardson et al.

eutrophied sites. No single APA-c value denotes the transition from eutrophied to “natural portion” of the water channel because of inherent variability within the system and method of analysis, although APA-c increased rapidly half way down most channels as TP and PO4 decreased in the water column. Values within the range of 500–800 APA-c are within the natural range for this period of July and August (Bush and Richardson 1995). However, storm-resuspended sediment has been shown to alter APA, as would dry spells that concentrate nutrients, flood conditions, or even a short-term surge of nutrients following regional fires (Newman and Reddy 1992). Therefore, all measurements of APA-c should be related to a control, a point at which phosphatase returns to “norms” that indicates a P limitation and maximum values of APA-c. In the dosing channels this is represented by the distance down the control channel where APA regains the level of the open water controls at meter 9 at the end of the channel. (Note: This was verified several times by comparison with open water slough areas away from the dosing facility.) Table 15.8 summarizes the various concentrations of P where “normal” APA-c is or is not regained in Site 2 of the dosing study. Normal APA-c levels appear in only the 30 and 50 channels between 11 and 12 µg l−1 of TP. Phosphatase concentrations did not return to normal levels in either the 75 or 150 channels. Sedimentary APA-c generally showed a similar pattern to APA-c in the water column (Fig. 15.15b). While APA-c in the water column is likely to be affected by short-term fluctuations, the sediment APA-c is likely to reflect concentrations of phosphorus that have been deposited, but remain in the biologically active sediment– water interface region. As changes in phosphate concentration in the sediment–water interface are liable to be less volatile than in the water column, the sediment APA-c probably offers a measure of the equilibrium state of these systems given annual phosphorous loads. The APA-c from the sediment increases down the channel as sediment TP decreases. It is similar to that recorded in the water columns (Fig. 15.15a, b). Phosphatase production by biological communities appears to be inversely related to the phosphate concentrations of both waters and sediments (Fig. 15.15a, b). Phosphatase production is concentrated in the periphyton mat and in the sediment. The amount that is in solution, or attached to planktonic organisms is at least two orders of magnitude less than these other two sources. However, when normalized for biomass, the amount of APA per unit chlorophyll-a was found to be similar in both sediment and water in most channels. A single value for a phosphatase limitation threshold appears difficult to quantify across components, but if a gradient of values was collected with samples from water, soil, and vegetation known to be unenriched and also from enriched areas, the phosphatase threshold of a given locality could be developed. Using this approach a P “limitation range” was identified, indicating that the system is P limited and in equilibrium with “normal” background total P and PO4-P (Richardson et al. 1995, 1999; Richardson and Qian 1999). Does this range of phosphatase values have a particular ecological significance? At some level, it must. Given a somewhat greater degree of latitude, the question then becomes, how much phosphate can the system absorb without being materially altered and do the decreases in phosphatase reflect this? In the 75 and 150 channels,

15 Water Quality, Soil Chemistry, and Ecosystem Responses to P Dosing

409

Fig. 15.15 (a) Total P, PO4-P, and APA-c in the water of the experimental channel dosed with 44 µg l−1 PO4-P in the first 2 years at Site 2 in WCA-2A. (b) Total P and APA-c in the sediment– water interface of the experimental channel dosed with 44 µg l−1 PO4-P at Site 2 in WCA-2A. Measurements taken at increasing distances from the PO4 source (head box). The “9th m” of the 8 m channel is the open water APA and was used as a control value as denoted by *

Table 15.8 The concentration of total P and PO4 where alkaline phosphatase/per unit chlorophyll (APA-c) appears to return to normal levels in unfiltered water samples from Site 2 of the dosing study Channel 30

50

75

150

12 11 #9 #8 Total P (µg l−1) 5 6 #8 #8 PO4-P (µg l−1) The symbol “#” indicates that APA/c did not regain a normal level in this channel (Bush and Richardson 1995)

we saw no return to normal APA-c levels and this had a profound effect on the organisms. In the other channels, we saw a return to normal APA-c levels at 11–12 µg l−1 of TP down the channel.

410

C.J. Richardson et al.

How much of an effect will slightly raised TP and PO4 concentrations have on organisms depends on the niche width of keystone species (sensu Tilman 1982). If the niche is narrow, and we may take it in a P-limited system that some form of P is likely to be a major niche dimension, then a competing species could move in and take over if there is even a small change in P. If, on the other hand, the niche is relatively broad, the threat of competitive replacement is reduced (Tilman 1982; Colinvaux 1992). Undoubtedly, some species will be replaced by small changes in P, because they are already living near the limits of their tolerance; however, it is not clear that because of one or even a small proportion of species levels turnover at the lowest trophic that the system is changed in any fundamental sense. Paleoecological studies repeatedly demonstrate that communities are ephemeral collections of species with no permanent allegiance one to the other (Webb 1987; Bush 1994). Island biogeographic theory (e.g., MacArthur and Wilson 1967) denotes that even in systems that are apparently in equilibrium, the equilibrium is not static but dynamic. Thus some proportion of species turnover should be acceptable, as it is unlikely to produce a cascade effect upon the system. The clear exception to this would be the turnover of a keystone species like U. purpurea, C. jamaicense, or keystone algal species aggregates, e.g., those forming the periphyton mat. If these were to disappear, the ecology could be profoundly affected. More detailed research is needed to quantify the affects of species losses and changes on ecosystem processes and sustainability. A more complete assessment of the P threshold is presented in Chap. 25.

15.3.3

Ecosystem Response

To assess ecosystem level responses, we utilized oxygen profiles as well as indices of productivity and biomass for the slough community. The diel rise and fall of dissolved O2 and CO2 have been used successfully to determine the primary productivity in wetlands (Richardson and Schwegler 1986; Flora et al. 1988), lakes (Verduin 1960; Wetzel 2001), streams (Livingstone 1991), and oceans (Reyes and Merino 1991). The general diel patterns for DO revealed a general increase in the DO at the middle two P dosing treatments and a slight damping of the amplitude of values at the highest treatments (Fig. 15.11). Our results are similar to Flora et al. (1988), who observed a relative increase in DO concentrations in the P-enriched test channels in the Everglades National Park. Increases in pH and DO values have also been measured in lakes after P additions and during algal blooms (Wetzel 2001). However, an assessment of the frequency distribution of DO ranges at each dosing site suggests that a shift in the ranges existed in the treatment channels vs. the control channels (Fig. 15.16). At Site 1, the highest measured ranges of dissolved oxygen were found between 2 and 8 mg l−1 for most channels with the 150 treatment channel showing a slightly truncated DO range as compared to the other channels. In contrast, Site 2 treated channels displayed a higher frequency of lower DO values

15 Water Quality, Soil Chemistry, and Ecosystem Responses to P Dosing

411

Fig. 15.16 Box and whisker plots showing the frequency distribution of < 2 mg l−1 dissolved oxygen (DO) concentrations in the experimental channels of the dosing study at Sites 1 and 2

as shown by a box plot of the DO frequencies (%) for measurements below 2 mg l−1 (Fig. 15.16). Again, Site 1 channels are fairly similar although the middle treatment channels have a higher frequency of DO values below 2 mg l−1. The site analysis also revealed that all the walled channels at Site 2 had an increased frequency of lower DO values. This may be due to a lower amount of circulation in the enclosed channels as compared to the open unwalled control channels. Lower circulation at Site 2 may have been related to the dense populations of Chara and N. odorata (Richardson et al. 1997; see Chap. 16). An analysis of the mean monthly DO values in the water column suggest that changes in the frequency of DO below 1 mg l−1 occur mainly at water column TP concentrations above 20 µg l−1. An assessment of the influence of phosphorus additions on ecosystem community metabolism was based on our diel oxygen curves. This work suggests that water column gross primary productivity (GPP) was comparable to the respiration (CR) in the experimental channels of the TP dosing study (Fig. 15.17). Thus, the P/R ratio was always close to 1 for the channels. High rates of GPP and low rates

412

C.J. Richardson et al.

Fig. 15.17 Gross primary productivity (GPP) and community respiration (CR) measured in g O2 m−2 day−1 in the dosing channels in WCA-2A in 1995 calculated from diel oxygen curves. Values for Sites 1 and 2 represent values averaged over the peak growing seasons in June (Site 1) July and August (Site 2). Line shows subsidy–stress curve for productivity

of net primary productivity have been shown to be typical of Everglades sloughs (Browder et al. 1994). The GPP and CR rates increased with the P additions, reaching a maximum value in the 50 channels. Estimates for the 75 and 150 channels were lower than the 50 channels and the 150 channels were often below the controls. This indicates a declining trend in community metabolic activity for the higher dosing treatments. Moreover, this trend indicates that this ecosystem level attribute follows the subsidy–stress model suggested as a method to assess ecosystem response to limiting nutrients or stress (Odum et al. 1979; see Chap. 25). The GPP and CR estimates in the experimental channels strongly reflect the measured changes in the periphyton and macrophyte communities (Chaps. 18 and 16). Higher community metabolic activity at Site 2 as compared to Site 1 probably reflects the higher seasonal increases in productivity found during the July and August period (higher irradiation and temperature periods) when Site 2 measurements were made as compared to the June measurements at Site 1 (lower irradiation and temperature period). Our productivity estimates in the Everglades sloughs (unenriched) indicate a system of high GPP (5.3–13.5 g O2 m−2day−1) in which most of the production is used by the photosynthesizing organisms and decomposers.

15 Water Quality, Soil Chemistry, and Ecosystem Responses to P Dosing

413

This same P/R balance exists in the P-dosed channels although moderate P additions increased GPP and CR values to nearly 20 g O2 m−2 day−1. These rates are much higher than the rates 1.4–4.5 and 0.6–6.1 g m−2 day−1 reported earlier by Browder et al. (1982) and Swift (1989), respectively. Importantly, the periphyton communities serve as a major contributor to the GPP of the Everglades sloughs. Disappearance of the mat cover would thus be a major and significant change in this community. For example, water column TP above 15 µg l−1 has resulted in an increased loss of the periphyton mat cover and in the decline in Utricularia spp. (Chaps. 16 and 25). To assess the seasonal variations and the yearly community metabolic response from continued P dosing, we compared the changes in productivity and respiration over several years in the treated channels with the control channels (Fig. 15.18). A general trend of increasing community metabolism over controls was found from

Fig. 15.18 Changes in gross primary productivity (GPP) and community respiration (CR) measured in g O2 m−2 day−1 in the P-dosed channels compared to the walled and unwalled control channels over different seasons and years. Primary productivity and community respiration were estimated from the diel dissolved oxygen data

414

C.J. Richardson et al.

Fig. 15.19 The periphyton biomass vs. water column TP concentration in the experimental channels of the dosing study at the end of the 6-year experiment in October of 1998

1993 to 1996 in all the treatment channels except the 150 channels. The highest responses were found in the 30–75 channels. Community respiration was near GPP, but the 1995 values exceeded GPP in the 50 and 75 channels. In addition, the 150 channels displayed both a GPP and CR value below that of the controls. By 1996, after 4 years of P dosing, the highest two channels showed the lowest increase in community metabolism. To appraise the long-term influence of increasing TP concentrations in the water column on production, we measured the amount of periphyton biomass that existed in the dosing channels at the end of the 6 years of elevated P concentrations (Fig. 15.19). Periphyton biomass generally averaged around 400 g dry weight m−2 at 10 µg l−1 TP. Biomass decreased to values near 100 g dry weight m−2 around 30 µg l−1 TP. A decrease in periphyton biomass was found at all channel locations where TP values were consistently above 15 µg l−1 TP.

15.4

Conclusions and Lessons for Restoration

Our research discussed in this chapter focused on factors controlling the biogeochemistry of P in Everglades as well as establishing the relationship among PO4 concentration, TP, water chemistry, and ecosystem responses. Knowing that the Everglades is dominated by alkaline (hard water) chemistry is central to under-

15 Water Quality, Soil Chemistry, and Ecosystem Responses to P Dosing

415

standing what controls nutrient biogeochemistry in the ecosystem. Importantly, the phosphorus dosing mesocosms provided realistic and uniform testing ranges for P threshold determinations during the 6-year experiment in the Everglades The background water column PO4 (5 µg l−1) and TP (10 µg l−1) concentrations in the sloughs selected as research sites were typical of oligotrophic waters in the northern everglades in the 1990s. Water depth in the sloughs in WCA-2A varied seasonally (<10 to >145 cm) in response to the rainfall pattern and water management practices. Variations in water column TP, TDP, TN, TDN, NH4-N, and NO3-N in the sloughs were inversely related to the water depth fluctuations (p < 0.001), suggesting the strong link between hydrology and nutrient regime in the Everglades. The average 6-year PO4 dose (both sites averaged) for each treatment channel was as follows: (30-P = 27 µg l−1 PO4-P, 50-P = 48 µg l−1 PO4-P, 75-P = 61 µg l−1 PO4-P, and 150-P = 126 µg l−1 PO4-P). Our data show that the dosed PO4 treatment levels have created an appropriate number of response sites within the channels to test for “P threshold effects” on organisms at different trophic levels. Six-year PO4 concentrations for the control channels at both sites averaged ~5 µg l−1 PO4-P. Both PO4 uptake and dilution of input water caused the rapid decrease in phosphate (PO4) in the water column. The Br tracer results show that the dosing channels are sealed and well characterized by the conservative tracer, allowing us to estimate PO4 uptake vs. dilution in the system. The higher PO4 values in the water column in the first few meters of the 75 and 150 channels vs. controls indicate that the slough ecosystem at these locations could not utilize all of the added PO4. This created a PO4 concentration in the water column that was higher than background control channels, a clear signal that the ecosystem was overloaded with P at these sites. During the 6 years, no clear pattern was found suggesting a TP or PO4 concentration increase from year to year in the water column in the treated channels, but sediment concentration continued to increase each year. Year-to-year patterns were consistent with the trends observed in control channel total P. Phosphate additions greatly influenced downgradient TP concentrations in the channels. Total P was generally highest in the channels to which we added the highest concentrations of PO4. All treatment channels had significantly different (p < 0.05) TP water column values from the controls. The N/P ratio in treatment channels was much lower than in controls. Total phosphorus was significantly (p < 0.05) and negatively related to water depth in most years. Higher TP values exist under shallow water depths and lower TP values under high water depths. This suggests that water depth must be taken into account when establishing TP values for water quality standards. Of concern was the amount of DOP generated within the dosing system. This form of P is very difficult to remove from the water column and is subject to long distance transport in the ecosystem. Soluble reactive P values also remained high in the system, especially in the higher dosing channels, thus indicating that P loading had exceeded the system’s P assimilative capacity. Importantly, our results clearly indicate that in a P-limited system such as the Everglades, PO4 concentrations in the water collected during the latter part of the day may often be very low compared to early morning or evening concentrations due to high P uptake driven by high midday photosynthetic rates. It also suggests that low nighttime PO4 uptake

416

C.J. Richardson et al.

rates may result in the downstream movement of higher concentrations of P at night. These findings indicate that further work on monitoring and release schedules needs to be undertaken to quantify potential diel effects on P uptake and downstream loss rates. Calcium carbonate precipitated from the water column by algae serves as a matrix, provides structure, and supports the physical integrity of the periphyton mat commonly observed in the oligotrophic Everglades. The mat covers nearly the entire surface of the open water sloughs during the summer season in certain areas of the Everglades. Decrease in calcite precipitation due to P additions may be one of the factors responsible for the fragmentation of the periphyton mat observed in the mesotrophic areas of the Everglades and the complete disappearance of the periphyton mat in the eutrophic areas (Stevenson and Richardson 1995; McCormick and O’Dell 1996). Conspicuous absence of the calcareous periphyton mat in the “tails” downstream from the tree islands in the Everglades (Browder et al. 1994) may also be related to higher P status of the tree islands compared to the surrounding fens (C.J. Richardson, unpublished data). Calcium carbonate precipitation from periphyton photosynthetic activity observed in the Everglades sloughs has also been observed in many hard water systems, and this process is considered to be the dominant mechanism responsible for the self-protection of these waters against eutrophication (Murphy et al. 1983; House 1990; Kleiner 1988). Total inhibition of CaCO3 precipitation by elevated P introductions may enhance P bioavailability in the Everglades and further accelerate the eutrophication of these waters. APA-c is a legitimate tool for assessing water quality P limitations but should not be used in isolation since its values can be quite variable. Separating bacteriagenerated APA-c from that generated by algae or macrophyte roots is another difficulty when trying to determine what the concentrations represent. It is unlikely that the APA-c assay alone can be used to determine a “P threshold” for the Everglades since its maximum concentrations represent a P limitation, and lower values have not been directly calibrated to biotic changes over a range of P concentrations. Our productivity estimates in the Everglades sloughs (unenriched) indicate a system of high GPP (5.3–13.5 g O2 m−2 day−1) in which most of the production is used by the photosynthesizing organisms and decomposers. This same P/R balance exists in the P-dosed channels, although moderate P additions increased GPP and CR values to nearly 20 g O2 m−2 day−1. Thus, the periphyton communities serve as a major contributor to the GPP of the Everglades sloughs. Disappearance of the mat cover would thus be a major and significant change in this community. A trend of increasing community metabolism over controls was found from 1993 to 1996 in all the treatment channels, except the 150 channels. The highest responses were found in the 30–75 channels. Community respiration was near GPP but the 1995 values exceeded GPP in the 50 and 75 channels. In addition, the 150 channels displayed both a GPP and CR value below that of the controls. After 4 years of P dosing, the highest two treatments showed the lowest increase in community metabolism. These productivity trends indicate that the Everglades ecosystem follows a subsidy–stress model and that excessive PO4 additions result in an imbalance in overall community metabolism.

16

Macrophyte Slough Community Response to Experimental Phosphorus Enrichment and Periphyton Removal Curtis J. Richardson, Robert G. Qualls, Jan Vymazal, John G. Zahina, and Panchabi Vaithiyanathan

16.1

Introduction

Natural plant communities generally respond to fertilization with increases in growth rates for many species, but differential plant responses to excess nutrients more often result in species composition change. Some species are better adapted to take advantage of increased availability of nutrient resources and grow faster than others when supplied with abundant nutrients (Tilman 1990; Tilman et al. 1997). In many cases these might be termed “r” adapted species or “weedy” species. Shifts in species abundance in enriched communities often occur as a result of competition for light or space as plant cover increases due to fertilization. Fast growing plants that eventually tend to dominate the fertilized community are generally those species that can dominate sites by suppressing species not able to take advantage of fertilization through greatly increasing their growth, cover, or overall competitiveness. In many wetland macrophyte communities, a different pattern of response to eutrophication appears to have been developed. Macrophyte declines in response to eutrophication have been reported at various sites worldwide (Moss 1976; Phillips 1978; Ostendorp 1989; Vymazal 1995). Eutrophication of a body of water results in dense phytoplankton growth, reduced water clarity because of light adsorption by phytoplankton (shading), and increased concentrations of dissolved organic compounds produced by phytoplankton (Sand-Jensen and Borum 1991). These factors inhibit the growth of most submerged plants. It is likely that dense phytoplankton growth also inhibits growth of rooted submerged macrophytes by generating high pH, thus hampering inorganic carbon uptake and making submerged plants more dependent on the light-driven utilization of HCO3− of for photosynthesis (Maberly and Spence 1983). It has also been shown that shallow lakes that are macrophyte-dominated systems can usually withstand some level of nutrient additions without a shift in major species; however, when a critical level is reached, the stable state of the systems shifts from the clear water, macrophyte-dominated state to a turbid state devoid of macrophytes and dominated by phytoplankton (Scheffer et al. 1993). Turbidity shading does not appear to account for the decline in Phragmites density in Europe, since Phragmites tends to have most of its leaf area

417

418

C.J. Richardson et al.

above the water surface. It is also unlikely that shading could explain the decline in macrophyte populations in the Norfolk broads in England, prompting Phillips (1978) to offer an alternative hypothesis. He found correlative evidence that the decline in macrophyte production was caused by growth of filamentous and epiphytic algae on the macrophytes. He hypothesized that dense phytoplankton growth was prevented until the filamentous and epiphytic algae had suppressed the macrophytes and, in turn, substances that macrophytes secreted to inhibit phytoplankton. A similar inhibitory effect of dense epiphytic communities on rooted macrophytes was found by Sand-Jensen et al. (1985) and Ozimek et al. (1991). Twilley et al. (1985) reported that the accumulation of epiphytic material resulted in >80% attenuation of the incident radiation at the leaf surface. Ozimek et al. (1991) found clusters of the green filamentous alga Cladophora glomerata reduced underwater light by 30–80%, depending on the density of filaments and their compactness and position, i.e., distance from the macrophytes. In contrast, the Everglades is a periphyton-dominated ecosystem with floating periphyton mats, epipelic algal mats (growing on sediments surfaces), epiphytic algae (growing on plant surfaces), and very limited phytoplankton populations (Stevenson and Richardson 1995; see Chap. 18). Experimental results in the Everglades have also shown that heavy periphyton mat cover (95–100%) will result in a significant decrease (99% ± 12) in water column photosynthetically available light radiation (PAR, 400–700 nm at 5 cm water depth) and an 82% ± 13 PAR reduction with just 75% mat cover (Vaithiyanathan and Richardson 1998). The periphyton mat is thus the primary factor responsible for maintaining the reduced light conditions and relatively low DO concentrations characteristic of the benthic environment of the Everglades sloughs. More light is needed to maintain populations of advanced rooted floating and submerged macrophytes as compared to unicellular algae (Sand-Jensen and Borum 1991). Among microalgae, the experimental Ic (photosynthetic light compensation point) values typically vary from 1 to 10 µE m−2 s−1 in a 12 h light/ 12 h dark cycle, with an average value being around 7 µE m−2 s−1 (Geider et al. 1985; Richardson et al. 1983). However, when algae are given sufficient time to adapt, it has been shown that Ic values could be even below 1 µE m−2 s−1 (Geider et al. 1985). Light compensation points for photosynthesis of green shoots of rooted macrophytes are usually between 8 and 50 µE m−2 s−1 (Maberly and Spence 1983; van der Bijl et al. 1989). In this chapter, we will present evidence that may help in explaining the general pattern of response of wetland macrophyte communities to eutrophication, light, and competition as well as assess changes in plant responses to alterations in water depth in open water sloughs. Importantly, we utilized our P dosing experiment to determine the long-term effects of a large range of P enrichments on the slough macrophyte community. Our primary objective was to determine the effect of experimental P enrichment on mortality and vigor of Utricularia purpurea, Eleocharis elongata, Eleocharis cellulosa, and Nymphaea odorata, the main species in our test sloughs. To further aid in the confirmation of macrophyte trends observed in the experimental mesocosm study, we also transplanted these key species

16 Macrophyte Slough Community Response to Experimental Phosphorus

419

in control and highest P-enriched channels to lend insight into the mechanism affecting mortality associated with enrichment and to “calibrate” the mortality and senescence of U. purpurea as a function of PO4 concentration to assess the use of this species as a potential indicator organism. To investigate the effect of short-term reduced algal community competition, we did a removal experiment in which the periphyton mat, the epiphytic algae, or both were removed for 6 months to assess the effects on macrophyte growth of E. cellulosa, N. odorata, and U. purpurea within an unenriched Everglades slough community near the dosing study in WCA-2A. We hypothesize that removal of periphyton mechanically would result in increased macrophyte growth due to increased light availability.

16.2

Slough Plant Communities of the Everglades

Sloughs of the Everglades are located in areas of deeper water and longer hydroperiods than the surrounding sawgrass marshes (Loveless 1959). The vegetation of sloughs is comprised of emergent macrophytes (e.g., Eleocharis spp.) of lower stature than is typical of the surrounding dense sawgrass marsh, floating macrophytes (e.g., N. odorata), and submerged macrophytes (e.g., Utricularia spp.). A macroalga, Chara spp., is another frequent plant that can be considered a macrophyte (large plant). In some seasons, dense, prolific growths of algal periphyton can cover live stems, dead plant stems, and fragments of algal mat from the sediment surface. The cover of macrophytes and periphyton can vary from areas of relatively open water to 100% coverage of the water surface. For more information about the slough vegetation, see Chap. 4. The macrophytes of these sloughs are ecologically vital to these habitats. They contribute to the building of peat. They provide food for animals in the form of live and dead plant material (i.e., the grazing and detritus food chains). They provide cover and habitat structure for fish, insects, and snails (Davis and Ogden 1994a; McCormick et al. 2001). Macrophytes also are likely to be vital in providing substrate for the algae in sloughs (Vymazal and Richardson 1995). The relationship between macrophytes, the algae that grow upon them, and P availability appears to be very important but very complex and poorly understood. Most studies aimed at the influence of increased phosphorus concentration on macrophyte occurrence and production in the Everglades focused on two dominant macrophyte species – sawgrass (Cladium jamaicense) and cattail (Typha domingensis) (Steward and Ornes 1975a,b, 1983; Davis 1991; Koch and Reddy 1992; Sutter 1992; Urban et al. 1993; Newman et al. 1996; Miao and Sklar 1998; Richardson et al. 1999) and not slough species. Regrettably, there is limited information on how P affects the growth and survival of slough macrophytes despite the fact that the Utricularia periphyton complex is considered to be an essential component of the Everglades sloughs (Browder et al. 1994; Vymazal and Richardson 1995). In contrast to cattail and sawgrass, submerged species such as Utricularia spp. or Chara may be adversely affected by P

420

C.J. Richardson et al.

additions. Steward and Ornes (1975b) reported that after 12 weeks of P dosing at a concentration of 10 mg l−1 (0.26 g m−2 week−1) Utricularia, Chara, and associated periphyton disappeared from the treatment plots. However, the results in the literature are not unanimous concerning the P effect on Chara. Craft et al. (1995) observed that after 1 year of adding 4.8 g P m−2 year−1 the decrease in Utricularia biomass was about 90%, while there was a large increase in the Chara biomass. Also, Chiang et al. (2000) reported that Utricularia was replaced by Chara in an experimental plots enriched with phosphorus. Flora et al. (1998) reported a near elimination of Utricularia spp. and an increase in Sagittaria spp. and Panicum spp. biomass as a response to P enrichment (PO4-P concentrations: mean 34.3 µg l−1, range 12–47 µg l−1) in experimental channels in the Shark River Slough area of the Everglades National Park. Flora et al. (1988) and Walker et al. (1989) did observe the disappearance of the submerged macrophyte, Utricularia and associated periphyton in response to continuous dosing with nitrate. However, the decline of Utricularia and periphyton occurred more slowly under nitrate loading as compared to loading with phosphorus. The interpretation of the macrophyte data in the Flora study was hampered by four aspects (1) lack of a walled control channel, (2) lack of replication of channel treatments or sites, (3) the low number of samples taken in each segment, and (4) the fact that only four samples were taken in the control area. A more recent P dosing study suggests that 6-month additions of SRP dosing (30 µg l−1) significantly increased P concentrations in the periphyton mat (30 µg l−1 dose = 1,961 µg g−1 as compared to control concentrations of 149 µg P g−1, while the flocculent detritus layer stored the most P (1.7 g P m−2) at the 30 µg l−1 dose (Noe et al. 2002). No increase in soil or macrophyte P was found. Also no significant differences in P standing stock were observed between SRP doses of 15 µg l−1, 5 µg l−1, and controls. These shorter-term studies have provided important information on P cycling and P storage, but plant response information is incomplete or confounded, as in the Flora study, by a lack of controls on P dosing and problems with nitrogen additions, and only short-term responses are reported in the Noe study. Thus, there is a need for a long-term assessment of the response of macrophytes over a range of PO4 concentrations under varying water levels.

16.3 16.3.1

Methods Plant Analysis

The establishment, operation, experimental design, and P treatment conditions of the dosing study are given in Chaps. 14 and 15. Since each channel represents an enrichment gradient, the entire channel is not represented as an experimental unit. Each meter square quadrant serves as the experimental unit since TP, PO4, and other nutrient concentrations are known for each distance.

16 Macrophyte Slough Community Response to Experimental Phosphorus

421

Macrophytes were counted in March, June, August, and December from August 1992 to August 1998. Plant stem counts on emergent plants were done at each meter in the channels using a m2 plot. Stem densities per species per area of channel were calculated for each plot and sampling period. Plots were located in the center of the channels to avoid edge effects. Submerged plants like Chara and Utricularia were either counted individually within plots or the percent of cover area was estimated. The species studied in the experiment were E. elongata (spikerush), E. cellulosa (spikerush), N. odorata (fragrant water lily), U. purpurea (bladderwort), and Chara spp. (stonewort). Utricularia gibba and U. foliosa (bladderwort) were also present in small numbers and counted individually to give a total number of Utricularia plants. All individuals, or total percent cover were measured in each meter square quadrant. For Eleocharis spp. all stems emerging from the soil were counted, both above and below the water surface. For N. odorata, all leaves were counted, both above and below the water surface. The percent cover by Nymphaea leaves was also estimated because size of individual leaves was variable and percent cover was more closely related to the competition for space. For Utricularia spp., any stem not connected to another was counted as an individual plant. Chara spp. was measured as percent cover, counting plants at and below the water surface. For the transplant study, ten plants from each of four species were transplanted into a control (no P added) and enriched 150 channel at each of the two replicated dosing sites. Thus, the design was two P treatment levels × two sites × four species × ten replicate plants. A second experiment was done with U. purpurea in which transplants were placed into channels with a range of concentrations of PO4 to attempt to show the concentration at which effects were observed. Plants were gathered from outside the channels and five plants were tethered 0.5 m from the input in each of five channels receiving long-term 4-year average input concentrations of 5 (control), 26, 45, 60, and 124 µg l−1 PO4-P. Macrophyte species were collected from an area adjacent to an unwalled control “channel” in mid-May 1996, transplanted in the dosing channels, and analyzed 90 days later in August 1996. They were examined for number of shoots or leaves, length, mortality, greenness or senescence, and presence of periphyton. U. purpurea plants were also gathered from the area adjacent to the outside control channel. Plants were tethered with soft nylon string to a dowel rod located 0.5 m from the water input at the head of the channels. Sufficient slack was allowed so that the plants could rise or sink in the water. Plants of comparable length, 15–25 cm, and green color were selected and a piece of fluorescent nylon string was loosely looped 2.5 cm behind the last whorl on the growing tip of the plant so that extension growth could be measured later. Plants were examined later for mortality, presence of green whorls, and extension growth beyond the labeled end, and the density of periphyton growth on the new Utricularia growth. For experiment 1, ten Utricularia plants were placed in the control and 150 channel at two sites in mid-May, and a census was taken 2 weeks later. For experiment 2, five plants were placed in each of five channels and examined 36 days later in August 1996.

422

C.J. Richardson et al.

Three replicate slough sites were chosen for the periphyton removal experiment in an unenriched area of WCA-2A at the following locations: Rep. (1) 26.15.554 N by 80.23.158 W; Rep. (2) 26.15.517 N by 80.23.192 W; and Rep. (3) 26.15.388 N by 80.23.139 W (Zahina and Richardson 1997). Four 1 m2 plots were established in each slough site, marked with PVC poles. One of the following treatments was randomly assigned to each plot (1) untreated control (UTC) plot was left unchanged; (2) cover removal (C2) plot had initial dead plant material and periphyton mat removed, but no further changes; (3) periphyton removal (PR) plot with initial dead plant material removed, then periphyton removed on a regular basis; and (4) shaded (SH) plot with initial dead plant material removed, periphyton removed on a regular basis, and plot shaded at the water surface with 55% woven nursery shade cloth attached to a floating 1.5 m2 PVC quadrat. The shade cloth was installed in late July when estimated cover at C2 plots approached 50% and was used to mimic the shading effects of the algal mat. The treatment period ran from May to October. Treatment plots were placed in slough areas where test macrophyte species and the algal mat community were well represented. The following three macrophytes were selected and treated in the following manner. Living plant stems of E. cellulosa, a rooted emergent plant of the Cyperaceae family, which has a portion of its photosynthetic stem below the water surface, were counted per plot and measured in situ. Initial dry weight biomass per plot was calculated using stem length measurements and a regression of stem length vs. dry weight biomass calculated from plants harvested from nearby sloughs (Zahina and Richardson 1997). Plant leaves per plot were counted for N. odorata a floating leaved plant. Initial aboveground dry weight biomass (combined leaf and petiole) per plot was calculated using a regression of leaf L × W vs. dry weight biomass from plants harvested from nearby sloughs. Three 25-cm pieces of U. purpurea were collected from surrounding replicate sloughs and introduced into each of the C2, PR, and SH plots. This dominant rootless submersed bladderwort, which lives within the slough water column, is often encrusted with algal mat growth. These plants were introduced when U. purpurea was well within its growing season and when the algal community was at its peak growth rate. Individual plants were tied to the center well pipe with a 0.5 m length of string. Plant lengths and number of whorls were recorded regularly. In November all plant material for these species were harvested for biomass analysis and individual stems lengths were measured for all species. The material was air dried, oven dried at 70°C for 24 h, then weighed. Percent cover (combined algal mat and N. odorata leaves) was visually estimated in UTC and C2 plots, aided by monthly photographs of each plot. Epiphytic periphyton growth measurements on living E. cellulosa stems and N. odorata petioles were made monthly using a Manostat model 15-100-100 calipers (± 0.25 mm). During the treatment period, from May to November, the algal mat was removed from PR and SH plots approximately every 2 weeks. Also, periphyton growth was removed from living E. cellulosa stems, N. odorata petioles, and from U. purpurea in PR and SH plots using a soft artists brush (Grumbacher, model 6106 #8). Pore water wells were established in each of the treatment plots at a depth of 20 cm. Each well consisted of a 1.5 m length of 18.2 cm (2-in.) diameter thick-walled

16 Macrophyte Slough Community Response to Experimental Phosphorus

423

PVC pipe with the bottom end covered by plastic screen. Just before sampling pore water, the well was completely pumped out and, within 3 h, a sample of fresh pore water was drawn out using a vacuum pump and placed directly into an acid-washed plastic vial. During the test period, surface and soil pore water were collected approximately every 2 weeks. Water samples were immediately placed on ice, kept in a dark container, and filtered within 24 h through precombusted GF/C glass fiber filters. These were analyzed for orthophosphate (PO4-P). Once a month the water samples were also analyzed for ammonia (NH4-N) and nitrite (NOx-N) (see Chap. 6 for methods). During the study period, dissolved oxygen, pH, conductivity, temperature, and water depth were monitored monthly. PAR in the water column was measured in August and September. PAR measurements in the water column were made using a LI-COR model LI-192SA Underwater Quantum Sensor (400–700 nm response). All monthly dissolved oxygen measurements were made with an YSI Model 51B oxygen meter (probe diameter = 12.5 mm). On 3 dates during midsummer, dissolved oxygen profiles in the water column were measured using a Microelectrode OM-4 oxygen meter (probe diameter = 2 mm). All pH and temperature measurements were made with an Orion model 250A pH meter with a probe diameter of 3 mm. All conductivity measurements were made with an YSI model 33 S-C-T meter.

16.3.2

Data Analysis

Since each meter square quadrant was the experimental “universe,” the complete population was counted and there is no “standard statistical sampling error” due to subsampling in the population estimate. Since there is a complete population count, error bars are not shown around the data points in the figures. To examine the relationship between total phosphorus and various measured responses, we initially used classification and regression tree (CART) models (Breiman et al. 1984) as implemented in S+ software (Clark and Pregibon 1992; MathSoft 1997). This methodology has several advantages over more familiar linear techniques such as regression analysis and analysis of variance. CART models are not based on assumptions of linearity, additivity, or multiplicative interactions, and they are invariant to monotone transformation of the predictors. They are particularly useful for evaluating complex interrelationships that are not homogenous within the sample space. Predictor variable selection is implicit in the model procedure; many candidate predictors can be specified, but only those resulting in the greatest deviance reduction will be incorporated into the model. Interactions among predictors are not specified a priori; they are determined in the recursive partitioning procedure. However, a disadvantage of CART models is that they do not provide concise mathematical formulae characterizing the relationship between the response and predictor variables. Additionally, CART model predictions are based on mean values of discrete partitions rather than continuous functions of the predictors. However, this approach has proven to be a powerful tool in predicting which biotic attributes

424

C.J. Richardson et al.

change the most along an environmental gradient or in determining the breakpoint where response variables display major differences along the environmental gradient (Amrhein et al. 1999; Qian and Anderson 1999). In this application, we first specified total phosphorus as the response variable with many candidate predictors (e.g., macrophyte species, water depth, year, etc.) to determine the best predictors of P changes. The transplant experiments were statistically analyzed to determine the effects of P concentrations on the major species in terms of their growth rates, mortality, and senescence. The null hypothesis was that there was no association between mortality plus senescence and channel treatment category. This was tested by a chi square analysis. The periphyton removal experiment was statistically analyzed with ANOVA.

16.4 16.4.1

Results and Discussion Water Level and Temporal Effects

The CART modeling allowed us to evaluate which plant species and water variables were most closely associated with total phosphorus concentration differences in the dosing channels. In this application, TP was specified as the response variable with candidate predictors, like water depth, algae, mat cover, etc. Macrophyte responses as a function of month, year, water depth, and TP revealed interesting trends and responses for each species (Fig. 16.1). The biotic attribute with the highest TP level response for all tested attributes was Utricularia spp. (in this initial test we combined all Utricularia spp.). This attribute was followed by water depth as being the second most important variable related to variations in total P concentration. Importantly, TP was the major predictor for U. purpurea and all Utricularia spp. However, for E. elongata and E. cellulosa the major predictor variable was the study year with TP being the second or third best predictor of breakpoints in the data (CART models not shown). Thus, a number of environmental parameters, such as the water depth, temperature, season or complex year-to-year variations, may account for the large temporal variations in the distribution of the Eleocharis spp. observed in the Everglades sloughs, although P increases also had important effects. Similar plant species responses have also been found along the hydrology and P nutrient gradient in northern WCA-2A (Vaithiyanathan and Richardson 1999; King et al. 2004). Specifically, plants populations of Utricularia spp. and Eleocharis spp. were significantly reduced (p < 0.0001) with increased P concentrations along the gradient in northern WCA-2A, although King et al. (2004) clearly showed that plants responded to both elevated P concentrations and changing hydrologic conditions along the gradient. A time sequence plot of the slough macrophytes in the control channels reveals considerably more information concerning seasonal and annual variation in plant populations and density, as well as site differences for some species. The density of

16 Macrophyte Slough Community Response to Experimental Phosphorus

425

Fig. 16.1 A CART model of total phosphorus as a response variable with macrophytes, water depth, algae, etc., being used as the candidate predictors. The data are divided into a succession of nodes or branches based on values or categories of predictive variables. Plant responses are in stems m−2, water depth is reported in cm, TP is measured in µg l−1, and mat cover in % of area

U. purpurea, the most common Utricularia spp. in our Everglades sloughs, varied widely through the course of the experiment even in the control channels (Fig. 16.2). It was most abundant in the spring of 1993 and 1996 and was nearly absent by August 1993 and in the spring of 1997 (Fig. 16.2). It was present in relatively high numbers in Site 1 compared to Site 2, especially during 1993–1994, but population trends were quite similar after 1996 (data not shown). The population trends in the control channels closely followed the changes in water level found at the site (Figs. 15.4 and 16.2). For example, the increase in Utricularia populations in 1996 closely tracked the high water levels found in late 1995. Moreover, the decrease in stem counts in 1997 was during a period of very low water, and the recovery followed a dramatic increase in the water column in the fall of 1997. For the Eleocharis spp., large differences in density were observed between the two experimental dosing study sites. E. cellulosa dominated Site 2 and E. elongata dominated Site 1 (Figs. 16.3 and 16.4). The density of E. cellulosa at Site 2 ranged from 100 to 300 stems m−2 during 1993–1994, but declined substantially to numbers ranging between 50 and 100 stems m−2 thereafter (Fig. 16.3). E. elongata showed a similar distribution pattern as well. It reached its maximum density in 1993 (800 stems m−2) at Site 1 but did not increase above 250 stems m−2 after 1994 (Fig. 16.4). Site 2 had very low populations after 1992 (data not shown). These fluctuations in long-term trends in plant density in the controls (no P additions) demonstrate the importance of other environmental factors or plant life cycles on

426

C.J. Richardson et al.

Fig. 16.2 Temporal variation in the density of Utricularia purpurea in the walled control channel of the experimental dosing study sites in the Everglades sloughs. Note that the low uniform TP concentrations are shown in thick dark lines, the plant stem number per m2 shown as the gray line, and the fluctuations of water level as a thin black line

plant populations in the Everglades. Moreover, these factors must be incorporated into any assessment of changes in plant populations due to nutrient additions. As indicated by the CART analysis Eleocharis spp. populations varied significantly from year to year as shown in Fig. 16.4. Under no P additions, populations seem to respond to the previous years water levels. For example, low water levels in 1993– 1994 showed corresponding higher plant populations, while significant increases in water level in the 1994–1995 years was followed by a significant drop in Eleocharis populations in the control channels. Interestingly, the populations of both species of Eleocharis continued to drop in all the control channels over the 6-year period of the experiment. The same trends were found in the unwalled controls, so it was not due to a wall or shading effect, and in both cases the TP and PO4 concentrations remained constant (Fig. 15.5). The continued decrease in Eleocharis populations in the control channels with no P additions may reduce their value as a P metric, since the population decreased significantly during the study in the control channels where no P was added.

16.4.2

Phosphorus Effects

The response of slough vegetation to water column TP concentrations also varied by species. A comprehensive analysis of the plant density responses to P additions revealed a general decreasing trend in populations of Utricularia spp. at elevated

16 Macrophyte Slough Community Response to Experimental Phosphorus

427

Fig. 16.3 Temporal variation in the density of Eleocharis cellulosa in the walled control channel of the experimental dosing study Sites 1 and 2 in the Everglades sloughs in southern WCA-2A

water column TP concentrations (Fig. 16.5). The highest populations of Utricularia spp. per unit area are found in TP concentrations below 25 µg l−1 PO4-P. However, an actual P threshold is difficult to discern from the graph alone. A robust analysis of dosing channel data to provide a TP threshold is provided in Chap. 25. However, it is clear that populations of all Utricularia spp. are significantly effected by PO4 additions. This also has been reported in earlier P fertilization experiments in the Everglades (Steward and Ornes 1975a; Flora et al. 1988; Vymazal et al. 1994; Craft et al. 1995). A more detailed analysis of U. purpurea was undertaken since this is the dominant species of this plant and it may be a keystone species for sloughs in the Everglades (Fig. 16.6). The decline in density for this species was at a much

428

C.J. Richardson et al.

Fig. 16.4 Temporal variation in the density of Eleocharis elongata in the walled control channels of the experimental dosing study site 1 in the Everglades sloughs. Note that the low uniform TP concentrations are shown in thick dark lines, the plant stem number per m2 shown as the gray line, and the fluctuations of water level in dark thin lines

Fig. 16.5 Changes in plant density per unit area of Utricularia spp. to water column TP concentrations in the experimental channels of both dosing study Sites 1 and 2

~1S µg l–1) that the other two species of Utricularia, which indilower P threshold (= cated that this was the most sensitive of the Utricularia spp. to P additions. The fact that Utricularia is a nonrooted floating plant drawing nutrients from the water column may have influenced the speed and degree of the effects. The interactions of Utricularia

16 Macrophyte Slough Community Response to Experimental Phosphorus

429

Fig. 16.6 Changes in plant density per unit area of Utricularia purpurea to water column TP concentrations in the experimental channels of both dosing study Sites 1 and 2

with the periphyton may also influence their response to PO4. Utricularia growing in control areas was normally covered with growths of calcareous periphyton except at the rapidly growing tips, and it normally floats despite the cover of dense periphyton. Consequently, it is not the presence of periphyton per se which inhibits the growth of Utricularia. Hypotheses that might account for the inhibition of Utricularia are as follows: 1. Direct toxicity of PO4-P at high concentration 2. Stimulation of periphytic algal growth rates so that the growing tips of Utricularia overgrown by periphyton are shaded 3. Change in periphyton to forms that do not allow the Utricularia to float 4. Change in periphyton to forms which have an allelopathic effect on Utricularia 5. Loss of CO2 in water column due to an upward pH shift caused by PO4 additions and algal production Currently, there is no direct evidence to eliminate or support many of these hypotheses. However, hypothesis 1 is very unlikely as there are no studies that have shown any direct toxic effects of P at such low concentrations (≤ 50 µg l−1 PO4-P). Hypothesis 5 has the most scientific evidence in that Moeller (1978) has shown that U. purpurea in New England lakes used only free CO2 and not bicarbonate as a carbon source for photosynthesis. The depletion of free CO2 at higher pH in the enriched channels (Chap. 15) might explain the decline of the submerged U. purpurea, since it can no longer photosynthesize in enriched channels with elevated daytime pH. Some interaction with the periphyton (hypothesis 3) also seems likely in view of the changes in the periphyton community that occurs at the higher P levels.

430

C.J. Richardson et al.

Fig. 16.7 The population response of Eleocharis elongata to water column TP concentration in the experimental channels of dosing study Site 1

The decline of E. elongata is more difficult to explain since it generally has a significant proportion of its stem emerging from the water column, and should have no problem with CO2 fixation. Yet, as reported earlier, a major decline for this species was found in the control channels (Fig. 16.4). We believe significant changes in water levels have affected this species, since its decline follows a significant increase in the water levels at the dosing site. However, it also appears that this species has a low threshold for P additions (Fig. 16.7). The reasons for a decrease in stem density with increased P concentrations in the water column are unknown at this time. It may be related to a change in growth of alga on the surface of the stems. What is clear is that this species significantly declines in population at water TP concentrations above 15 µg l−1. Any changes in Eleocharis spp. would cause a significant ecological change in the slough community since it is the dominant plant and an important substrate for periphyton. In contrast, Chara sp. and N. odorata exhibited an increase in populations and cover with increased TP concentrations in the dosing channels (Figs. 16.8 and 16.9). This suggests that these two species respond positively to increases in PO4 in the water column. In addition, there does not seem to be a clear relationship between PO4 additions or TP in the water column and the plant cover of either species. The response of N. odorata is much different than all the other macrophyte species in that it shows an enormous positive growth and plant cover response over all the PO4 additions tested (Fig. 16.9). Increases in P seem to stimulate the growth and greatly expand the cover of this species in sloughs.

16 Macrophyte Slough Community Response to Experimental Phosphorus

431

Fig. 16.8 The change in percent cover of Chara sp. to increases in water column TP concentration in the experimental channels of the dosing study Sites 1 and 2

Fig. 16.9 The change in percent cover of Nymphaea odorata with increases in water column TP concentration in the experimental channels of dosing study Site 2

432

C.J. Richardson et al.

There appears to be no single hypothesis for explaining the responses of Chara sp., Utricularia spp., and Eleocharis spp. to increased PO4 additions. The CO2 hypothesis is less convincing for the case of the emergent E. elongata and Chara is known to take up bicarbonate so the CO2 limitation would not be a problem for Chara at high pH. The periphyton hypothesis may help in explaining the response of U. purpurea and E. elongata, but Forsberg (1964) demonstrated that PO4 inhibited Chara growth even without the presence of algae. He first noted that Chara seemed to be found in much lower density in eutrophic lakes (> 20 µg l−1 total P) (Forsberg 1964). He later grew Chara in culture and found that 30 µg l−1 PO4 stopped growth and only 8 µg l−1 limited growth. The effect was similar when Forsberg used pure cultures of Chara, thus it is unlikely that the growth responses that we observed in the Everglades was affected by interactions with periphytic algae growing on Chara. Forsberg (1964) also reported that some species of Chara reach optimum growth cover at an elevated PO4-P concentration between 5 and 20 µg l−1. Natural background PO4 water column concentrations in the Everglades are apparently below the optimum for the Chara, i.e., P is limiting growth and additions resulted in a significant increase in its productivity in the channels. It may also be significant that the dense growths of Chara with only slight enrichment were much less calcareous than those of the plants in the control channels.

16.4.3

Transplant Experiments

The survival of U. purpurea and E. elongata was much lower in the highly enriched channel compared to the control channel (Table 16.1). In contrast, while E. cellulosa did not exhibit growth in the enriched channels as did most transplants in the control channel, most plants were healthy and did not die or senesce in the enriched channels. Transplanted, N. odorata grew vigorously in both control and enriched channels (data not shown). All U. purpurea plants survived and grew at least 8 cm in growth extension in the control channels at both sites. In the highly enriched channels all plants died. For statistical analysis, the categories of “growth” and “healthy but no growth” were lumped, and the categories of senescent and dead were Table 16.1 Condition of transplanted species when a census was taken 90 days after transplantation Site 1 Site 2 Control Species

g

h

s

150 channel d

g

h

s

Control d

g

h

s

150 channel d

g

h

s

d

U. purpurea 10 0 0 0 0 0 0 10 10 0 0 0 0 0 0 10 E. elongate 6 2 0 0 0 0 3 7 10 0 0 0 0 3 4 3 E. cellulosa 9 0 0 0 1 7 0 2 9 1 0 0 0 8 0 2 Total numbers less than ten plants indicate that some of the tagged plants could not be found at the end of the experiment (note: all N. odorata plants were in the g class, data not shown) g growth; h healthy plant but no growth; s senescent; d dead

16 Macrophyte Slough Community Response to Experimental Phosphorus

433

lumped. The null hypothesis that there was no association between mortality plus senescence and channel input category tested by a chi square test indicated a highly significant effect of P treatment on mortality plus senescence at both sites. The algal growth on new shoots was sparse in the control channels but dense noncalcareous periphyton grew on the U. purpurea in the enriched channel and on the dowel rod to which the plants were tethered. The dowel was used as an artificial substrate to index periphyton growth during the experiment. An average of 85% of the E. elongata was either dead or senescent in the highly enriched channels at both sites. No plants in the control were dead or even senescent in the control channels; in fact, 87.5% of the plants located after incubation in the control channels exhibited growth. Obviously, the transplanting itself caused little mortality. The chi square test indicated a highly significant effect of PO4 input on mortality plus senescence at both sites. Shoots in the enriched channel had moderate to dense noncalcareous periphyton growth (which could be distinguished from previous calcareous growths when present). Shoots from the control channel had none to light calcareous periphyton growth. Eleocharis cellulosa clearly was not as vigorous in the highly enriched channel as in the control channel as indicated by growth vs. static survival, but mortality was low in the enriched channel. When the categories of “growth” and “healthy” were lumped, there was not a significant effect of P treatment on mortality plus senescence when sites were analyzed alone ( p = 0.15). However, when both sites were lumped, thus increasing sample size (n = 39), there was a significant effect of PO4 additions on mortality plus senescence ( p = 0.04). This significance reflected the fact that mortality plus senescence, while low in both the enriched and unenriched channels, was zero in the unenriched and two in the enriched channels at both sites. The mortality of the E. elongata in the enriched channel was clearly much greater than that of the E. cellulosa while survival (healthy plus growth categories) was similar in the control channels. In experiment 2, in which U. purpurea was transplanted into a series of channels with a range of P inputs, all transplants grew vigorously in the control channel, but all plants died or senesced in all channels receiving P inputs (Table 16.2). A chi square test indicated a highly significant effect (p < 0.001) of PO4 input on mortality plus senescence. This indicated to us early in the dosing experiment that U. purpurea had a P threshold lower than our average lowest treatment Table 16.2 Utricularia purpurea transplant survival as a function of P input concentration in channels receiving a range of P inputs Growth Senescent Dead Channel name and input concentration PO4-P (µg l−1) Unwalled control (5)* 5 0 0 Walled control (5) 5 0 0 30 (26) 0 1 4 50 (45) 0 0 5 75 (60) 0 0 5 150 (124) 0 0 5 Five plants were transplanted into each of five channels at one site and categorized as growing, senescent, or dead. *Phosphate concentrations are shown in parentheses next to channel names

434

C.J. Richardson et al.

(26 µg l−1 PO4-P) we used in the upper reaches of channel 30. However, U. purpurea was growing and surviving at various locations in the lower P enrichment zones down the channels. The results of the transplant studies generally confirm the trends seen in the long-term species density and cover studies in the vegetation of the channels treated with several levels of P input (Figs. 16.5–16.7). U. purpurea transplants died rapidly in the enriched channels, which corresponds to the general absence of Utricularia in the first segments of the enriched channels even when Utricularia occurs frequently in control channels. This also corresponds to observations along a nutrient enrichment gradient that U. purpurea is absent in the most enriched portions of WCA-2A (Vaithiyanathan and Richardson 1999). The very rapid response of U. purpurea to the treatment is probably explained by their inability to photosynthesize under reduced CO2 levels found in the elevated P channels as mentioned earlier. The failure of E. elongata to survive in the most enriched channel at both sites corresponds to the trend noted at Site 1 in the long-term census (Fig. 16.7). While the initial absence or low density of E. elongata at Site 2 prevented us from having the additional weight of data from two sites, the transplant data indicated that the same effect occurred at both sites. This mortality did not occur because of the competition for light and space. At Site 1, the percent cover of other macrophytes and algal mat was zero during the period of incubation. Rapid growth of noncalcareous algae was observed on the leaves of E. elongata in the enriched channel. The coverage did not appear dense enough to heavily shade the E. Elongata leaves as indicated by the visibility of portions of the leaves through the film of periphyton. This observation would be difficult to quantify, however. The greater tolerance of E. cellulosa to the P inputs in the most enriched channel also corresponds to the trends observed at Site 2, where E. cellulosa has persisted at low density in the most enriched channel (Richardson et al. 1997). Again, however, E. cellulosa growth did not appear to benefit from the high nutrient enrichment. N. odorata was not adversely affected and actually increased in size in both controls and enrichment channels. All the growth and mortality effects in the transplant experiments corresponded with trends noted in at least one site in the long-term plant census, but the experimental demonstration of the effect at both sites strengthens the hypothesis that the effects we found are not site specific.

16.4.4

Periphyton Removal Experiment

The algal community showed a clear seasonal pattern similar to that reported by Swift and Nicholas (1987). In the untreated control (UTC), algal mat growth had filled in all the available surface area by late July and then began to decrease considerably in late September (Table 16.3). Percent cover measurements per plot showed that ~80 ± 24% of the water surface in the UTC treatment plots was covered throughout the treatment period in contrast to 7 ± 3% in the periphyton removal

16 Macrophyte Slough Community Response to Experimental Phosphorus

435

Table 16.3 Periphyton mat cover, chemical, and physical measurements from algal removal treatment plots (Zahina and Richardson 1997) UTC PR SH C2 Estimated % cover 80 ± 24 7±3 57 ± 24 46 ± 29 7.2 ± 3.2 6.7 ± 3.1 6.3 ± 2.8 6.9 ± 2.9 Dissolved O2 (mg l−1) pH 7.80 ± 0.3 7.76 ± 0.3 7.77 ± 0.2 7.80 ± 0.2 918 ± 154 915 ± 176 950 ± 155 944 ± 160 Conductivity (µmho cm−1) Temperature (°C) 32.2 ± 2.3 32.6 ± 2.1 32.3 ± 2.0 32.6 ± 2.1 Depth (cm) 36 ± 25 37 ± 25 37 ± 24 36 ± 25 5±6 5±6 4±4 4±7 Slough PO4-P 6±9 9 ± 10 6±6 5±4 Pore water PO4-P 132 ± 208 251 ± 402 52 ± 45 52 ± 46 Slough NH4-N 4,258 ± 1,719 3,879 ± 1,795 3,654 ± 1,981 3,435 ± 1,266 Pore water NH4-N 4±3 5±4 6±7 4±2 Slough NOx-N 7±8 9±9 8±5 8±6 Pore water NOx-N All nutrient measurements are in µg l−1 UTC untreated control; PR plot that had initial dead plant material removed, algal mat and epiphytic algae removed regularly; SH plot that had initial dead plant material removed, algal mat and epiphytic algae removed regularly, and was shaded with 55% weave shade cloth; C2 plot that had only initial dead plant material and algal mat removed Table 16.4 Dissolved oxygen profile of the water column in each treatment plot in WCA-2A of the northern Everglades (Zahina and Richardson 1997) Treatments Surface water Mid-column Sediment–water UTC 13.2 ± 1.9 12.6 ± 0.4 1.5 ± 0.9 PR 8.7 ± 0.5 8.5 ± 0.7 10.9 ± 4.4 SH 10.2 ± 0.7 10.7 ± 1.4 7.0 ± 3.7 C2 16.1 ± 7.1 13.5 ± 2.1 5.0 ± 4.5 All measurements are in mg l−1 (see Table 16.3 for treatment codes)

(PR) plots. When the experiment was installed in May, it was noted that much of the algal community in the treatment plots was present in cylindrical formations produced from thick periphyton growth around Eleocharis stems from the previous year. These would float to the surface and, along with the Utricularia/mat complex, went on to form the basis of much of the floating algal mat community later in the season. Epiphytic periphyton growth on living E. cellulosa stems and on N. odorata petioles was not found until mid-September and attained a maximum thickness of 2.5 ± 0.25 mm by early November. Macrophytes, whether dead or alive, appear to be particularly important to the algal community in that they provide sources of structure. As a result of removal of the floating mat, there was an increase in the algal community at the slough bottom, as indicated by higher dissolved oxygen measurements at the sediment–water interface in the PR treatment (Table 16.4). The relatively anoxic benthic environment reflected in measurements from the UTC treatment is characteristic of unenriched Everglades sloughs (Rader and Richardson 1992) and these data indicate that shading due to heavy mat cover is the primary factor in maintaining it. Algal mats, especially those in association with Utricularia, have been shown to have a high metabolic rate (Brock 1970; Van Meter-Kasanof 1973).

436

C.J. Richardson et al.

To investigate whether the surrounding algal mat can limit nutrients to plants such as Utricularia, which obtains nutrients directly from the water column, we collected water collected from the floating algal mat using two techniques, a syringe with a needle opening of 0.25 mm, which was used to extract water, or collected the water that drained water from algal mat collected from the slough surface. Orthophosphate values of the interstitial water from the algal mat were two to six times those in the surrounding slough water (3–4 vs. 8–23 µg l−1 PO4-P). This was rather surprising in that reduced nutrients levels in a metabolically active mat were expected. The most likely sources of elevated PO4 are decomposition of dead algal and plant material embedded within the mat complex, as well as phosphatase activity (Chap. 15). The combination of the high temperatures (> 35°C) reported from the floating mat (Rader and Richardson 1992), its composition of bacteria, fungi, and algae (Weber 1973), and high metabolic rate, makes the mat an ideal environment for rapid decomposition of plant material. From these results it appears that the decay of dead plant material is important to the living algal mat community in that it can be a substantial local source of nutrients. PAR readings were taken at a 5 cm depth in the water column where there was heavy mat (near 100% cover), fragmented mat (75–90% cover), open water (0% cover), and open water with 55% weave shade cloth. Light reduction was calculated using the following formula %light reduction = 100 − (B × 100)/A, where B is the light reading 5 cm below water surface (µmol s−1 m−2) and A is the light reading above water surface (µmol s−1 m−2). Irradiation data confirmed that the algal mat community could reduce available light to submerged portions of aquatic macrophytes by as much as 98%, compared to a reduction of only 10% in open water (Table 16.5). Since algal mats seasonally cover large areas of the WCAs (Swift and Nicholas 1987; Pan et al. 2000), especially during the summer and fall growing season of many aquatic macrophytes, it is probable that this plays a significant role in limiting macrophyte productivity. Eleocharis production was higher in treatment plots, which had the floating algal mat regularly removed over control plots or plots that were artificially shaded (Table 16.6). The PR treatment showed the greatest increase in dry weight biomass (g m−2), percent increase in biomass, number of stems, and percent increase in number of stems. In each case, the PR treatment increases nearly doubled that of both the UTC and the shaded and periphyton-removed (SH) plots. Over the study period, the increase in biomass for the PR treatment was significant (p < 0.05), compared to increases in the UTC controls and SH treatments. A good correlation (r2 = −0.86) was found between percent gain in biomass and estimated percent mat cover per plot. At Table 16.5 Photosynthetically active radiation (PAR) reduction in the water column 5 cm below the slough water surface in WCA-2A of the northern Everglades (Zahina and Richardson 1997) Site description % Reduction in PAR (400–700 mm) Heavy mat (95–100% cover) Fragmented mat (75–90% cover) Open water (0% cover) Open water covered with 55% weave shade cloth

98.9 ± 1.2 (n = 12) 81.7 ± 13.0 (n = 9) 10.3 ± 4.8 (n = 38) 59.0 ± 5.1 (n = 12)

16 Macrophyte Slough Community Response to Experimental Phosphorus

437

Table 16.6 Changes in the relative amount of Eleocharis cellulosa and Nymphaea odorata biomass per plot over treatment period Initial biomass Final biomass Initial biomass Final biomass Gain in biomass Treatments % E. cellulosa % E. cellulosa % N. odorata % N. odorata per plot (g) UTC 83 ± 12 81 ± 10 PR 79 ± 11 88 ± 8 SH 88 ± 7 90 ± 7 C2 76 ± 9 80 ± 13 See Table 16.3 for treatment codes

17 ± 12 21 ± 11 12 ± 7 24 ± 9

19 ± 10 12 ± 8 10 ± 7 20 ± 13

26 ± 25 39 ± 17 11 ± 28 27 ± 4

10% mat cover biomass increased nearly 75% while at 75% cover biomass increase was only 25%. A strong linear relationship (y = −6.156 + 1.243x, r2 = 0.90) was found between maximum plot depth and mean stem length of E. cellulosa, where no relationship was found between estimated percent mat cover and mean stem length. This suggests that Eleocharis does not elongate its stems above the water surface in response to mat cover and a reduction in PAR below the water surface. It appears that water depth, rather than shading alone, is the primary mechanism that controls stem elongation response in this macrophyte. Eleocharis apparently grows rapidly with natural seasonal decreases in water levels but the species apparently cannot respond to massive increases in water depth as shown earlier (Fig. 16.4). Epiphytic periphyton growth on the plant stems seemed to play a lesser role in affecting Eleocharis production. Epiphytic periphyton appeared at the latter part of the growing season along the plant’s midstem just before annual senescence had begun. This was found to occur in all treatments simultaneously. Although, stem periphyton may influence Eleocharis’ fall season dieback, it appears that senescence is the direct result of a natural seasonal life cycle rather than from algal overgrowth. The UTC treatment showed the largest gain, while the PR treatment resulted in a loss of dry weight biomass for N. odorata (data not shown). However, the change in the biomass for each treatment over the study period was not significant (p < 0.05). A trend of decreased N. odorata biomass in the PR treatment seems to correspond directly with the increase in biomass found in Eleocharis (Table 16.6). One means of characterizing macrophyte interactions and their effects on community structure is to compare the relative amount of biomass of each species in a given area. Little or no change over a period of time, excluding seasonal variations, would represent a stable system, while consistent, larger increases or decreases for a particular species compared to a control indicate that the community structure is changing. The relative amounts of biomass of N. odorata and E. cellulosa in the UTC and SH treatments changed ± 2%, indicating that these plots remained nearly stable over the treatment period with respect to these macrophytes (Table 16.6). However, the PR treatment showed a 9% shift in relative biomass in favor of E. cellulosa. Investigation of this change revealed a strong correlation (r2 = 0.97) between a shift in relative percent biomass from N. odorata to E. cellulosa and estimated percent cover per plot. Considering that N. odorata is unaffected by shading below the water surface and that E. cellulosa biomass increases with increased PAR below the water surface (Table 16.6), we conclude that in the PR

438

C.J. Richardson et al.

treatment, the relative percent biomass of these macrophytes began to change as a result of increased PAR in the water column due to the removal of the floating algal mat. These results suggest that by limiting light below the water surface, the algal mat community can be an important factor in maintaining macrophyte community structure in unenriched Everglades sloughs. The results from the PR and SH treatments indicate that shading does reduce the growth of U. purpurea but not as significantly as we had postulated. Both the length of plant stems produced and the number of whorls of bladders were reduced slightly by shading. After 168 days in the PR treatment, where the algal mat and the epiphytic algae were removed regularly, stem length increased on average 57 cm or 228% as compared to an average of 61 cm (244% increase) in the C2 plots where only initial algal removal occurred. The 55% shading and continual removal experiment (SH) resulted in only a 48 cm expansion (192% increase), the lowest increase in length. The 55% shading (0.28 cm day−1 growth rate) resulted in only a 10–15% reduction in growth overall compared to the PR (0.33 cm day−1 growth rate) and C2 (0.36 cm day−1 growth rate) treatments, respectively. However, during this study it was found that U. purpurea was much less buoyant after the periphyton was removed, similar to other segments in the sloughs, which had no visible attached algal growth. Plants that had little or no attached algae remained in the lower water column, whereas plants without the periphyton removed from those consistently remained at or near the water surface. Utricularia purpurea is usually found in association with algal mats (Rader and Richardson 1992; Swift and Nicholas 1987; Brock 1970; Wagner and Mshigeni 1986) and often is found with algal covering on older portions of its plant body rather than the younger sections (approximately the first 20 cm). This Utricularia/ mat complex behaves as a symbiotic relationship in several ways. U. purpurea appears to be able to benefit from buoyancy gained from the attached floating algae by placing it in an advantaged position at the water surface, rather than in the lower water column where the mat can drastically reduce available PAR (Table 16.5). Increased PO4 levels in the mat complex (noted earlier), as well as cyanophyta nitrogen fixation within the plant’s bladders and mat complex (Wagner and Mshigeni 1986), provide U. purpurea with additional local sources of nutrients in a nutrient-limited environment. This additional nutrient supply may increase Utricularia’s growth, perhaps offsetting any negative effects from algal overgrowth and shading. It appears that a symbiotic relationship exists between Utricularia and the periphyton mat, and removal of the mat structure may result in a loss of Utricularia due to its inability to stay afloat near the surface.

16.5

Conclusions and Lessons for Restoration

It appears that populations of both emergent Eleocharis spp. and submerged Utricularia spp. respond to high and low water levels but in opposite patterns under rapid changes in water level. Utricularia populations directly increase with increasing

16 Macrophyte Slough Community Response to Experimental Phosphorus

439

water levels (Fig. 16.2) and Eleocharis populations decrease, but with a hysteresis effect being evident (Fig. 16.4). Macrophyte responses to TP concentrations were uniform for Utricularia spp. and Eleocharis spp. during the 6-year dosing trials (Figs. 16.5–16.7). A comprehensive analysis of the responses shows that a decrease in most aquatic plant populations is found at higher TP concentrations, especially above the 15 µg l−1 TP zone concentration (see Chap. 25 for quantitative analysis of the P threshold). The exception to this is Nymphaea spp., which shows an increase in population and cover with increased TP concentrations (Fig. 16.9). The results also suggest three patterns of response to PO4 enrichment in the sloughs (1) inhibition – as exemplified by U. purpurea, (2) stimulation – as suggested by Nymphaea response, and (3) an optimum response – as exemplified by Chara. The decrease in abundance of some macrophytes in the experimentally enriched channels superficially resembles the macrophyte “decline” in response to eutrophication observed in many studies (Ostendorp 1989; Phillips 1978; Vaithiyanathan and Richardson 1999). It does not resemble the response of terrestrial communities to fertilization – i.e., the increase in density and competition for space that results in dominance of exploitative species (Tilman 1990). On closer inspection, the responses of a macroalga Chara a submerged vascular plant and an emergent plant would not necessarily require the same mechanism of inhibition. Further research is needed to determine why these responses occurred and whether they share a common mechanism with the “macrophyte declines” observed in other enriched aquatic ecosystems. It is clear from our dosing study that an elevated concentration of phosphorus into the slough waters of the Everglades will result in the loss of several of the key macrophyte species (U. purpurea, E. elongata) from the community and a shift to more nutrient tolerant species like N. odorata and possibly Chara. The impact of these shifts on the other trophic levels is not well known but could be significant for algal, macroinvertebrate, and fish species. The algal removal experiment (periphyton mat and epiphytic) in unenriched sloughs showed enhanced macrophyte productivity and a shift in the algal community (mat to benthic) with increased light availability. This suggests that increased irradiation deep into the water column may lead to the establishment of other macrophyte species uncharacteristic of the oligotrophic sloughs by facilitating seed germination and/or species invasion. Of importance is the finding that several of the key macrophyte species declined over time in apparent response to elevated water levels in WCA-2A. This suggests that the significant raising of water levels under several of the proposed management restoration regimes may result in a loss of key slough species, especially if the lengths of drier periods are reduced in the Everglades, since species like Eleocharis seem to regenerate their populations under lower water conditions.

17

Decomposition of Litter and Peat in the Everglades: The Influence of P Concentrations Robert G. Qualls and Curtis J. Richardson

17.1

Introduction

The accumulation of organic carbon in soil is determined by the balance of net primary production and mineralization of carbon via decomposition. Because the soils of the earth contain about twice as much C in the form of organic matter than is contained in the atmosphere, the factors that control this storage of carbon are important to the potential effects of rising atmospheric CO2 content. A large portion of the soil organic matter of the earth is stored in peatlands such as the Everglades (although boreal peatlands are far more expansive). Not only is carbon mineralized during the process of decomposition, but also the N, P, and other elements bonded to the organic matter are returned for continued cycling in the ecosystem. Consequently, the sequestration of these nutrients in peat often results in nutrient deficiencies in peat ecosystems. This sequestration of P, along with the lack of geological inputs, is likely to be one reason that the Everglades is P limited. The major environmental controls on decomposition are temperature, moisture availability, oxygen availability, pH, and exogenous nutrient availability. Decomposition rates of given types of plant residue are also influenced by intrinsic factors such as lignin content, polyphenol content, and the internal content of N and P in the residues (Swift et al. 1979). The subtropical temperatures of the Everglades lead to very high decomposition rates when oxygen is supplied through drainage (Tate 1980). The pH of the Everglades peat is generally near neutral to slightly alkaline (except in ombrotrophic areas of Loxahatchee peat) and is consequently not a limiting factor. We will concentrate on oxygen availability and exogenous nutrient availability as the most important environmental controls that are subject to human alteration. In this chapter, we review work on decomposition in the Everglades and other wetlands and specifically review in detail our experimental approach to isolating the effect of P enrichment on litter. In addition, we present work on effects of P and N additions on mineralization of peat.

441

442

17.2

R.G. Qualls and C.J. Richardson

Effect of Aerobic vs. Anaerobic Conditions

Because of the dramatic subsidence of Everglades histosols when they are drained, they have a long history of research on the effects of drainage. Rates of soil subsidence in the Everglades Agricultural Area (EAA) are about 3 cm year−1, and 58–73% of the loss is due to microbial decomposition (Volk 1973; Tate 1980). Evolution of CO2 by drained Pahokee muck was measured at 3,300 g C m−2 year−1, far higher than rates of any ecosystem reviewed by Schlesinger (1977). Accompanying these massive rates of carbon loss, about 1,400 kg ha−1 year−1 of N are mineralized, most of which is denitrified (Terry and Tate 1980). How do these rates of decomposition compare with those under undrained (and usually anaerobic) conditions? By comparing rates of net primary productivity (NPP) with rates of accretion of peat since about 1964 in the unenriched area of Water Conservation Area 2A (WCA-2A), Craft and Richardson (1993a) estimated an average of 15–16% of the NPP of 524 g C m−2 year−1 was buried. Thus, about 443 g C m−2 year−1 were lost to decomposition. Comparing this rate to the figure of 3,300 g C m−2 year−1, drainage appears to increase decomposition rate about 7.5-fold. Since decomposition rates are likely to decrease with age of peat, the recent decomposition rate of 443 g C m−2 year−1 is likely to be greater than that which would include older peat. In other terms, 1 year of decomposition of drained peat (at 443 g C m−2 year−1) is equivalent to about 39 years of accretion of undrained peat (about 84 g C m−2 year−1) using Craft and Richardson’s (1993a,b) accretion rate.

17.3

Effect of Exogenous Inorganic Nutrients on Decomposition

The concentration of N and P, necessary for building biomass of the decomposers, is often less than optimal in fresh plant litter. The acceleration of litter decomposition by N additions has long been known (Anderson 1926) but the results in different ecosystems depend on the specific circumstances. Evidence for the acceleration of decomposition rate comes from two types of studies: experimental additions of N or P and correlative studies in which a correlation between decomposition and ambient inorganic N or P concentrations is observed. Among the studies using experimental additions, results have varied. Egglishaw (1972) found that NO3 or NH4 additions accelerated decomposition of cellulose in artificial stream flumes. Howarth and Fisher (1976) found that additions of N and P together stimulated leaf litter decomposition but addition of either N or P separately did not. Experimental additions of PO4 to whole streams in Tennessee increased leaf litter decomposition but additions of NH4 did not (Elwood et al. 1981; Newbould et al. 1983). NO3 additions failed to affect leaf litter decomposition in a study by Triska and Sedell (1976). In the only other wetland study involving experimental additions, Lockaby et al. (1996) found no increase in decomposition rates of lignin or cellulose in

17 Decomposition of Litter and Peat in the Everglades

443

swamp leaf litter in response to either N or P additions. No study of which we are aware, however, has tested a quantitative relationship of decomposition rate across a range of nutrient concentrations. Correlative studies have also tended to suggest acceleration of litter decomposition by exogenous inorganic nutrient availability. In comparing litter decomposition among six hardwater streams, Suberkropp and Chauvet (1995) found that nitrate concentration was the only variable correlated with differences in decomposition. Qualls (1984) found that litter in a stream swamp decomposed much faster at sites with elevated inorganic N and PO4 from agricultural hog farm runoff than at unenriched sites. Comparing three sites along a nutrient enrichment gradient in WCA2A of the Everglades, Davis (1991) observed that litter of Cladium and Typha decomposed faster at a nutrient-enriched site. He attributed this correlation to limitation by either N or P but a number of factors can also vary along a large-scale geographic gradient. Amador and Jones (1993) found that additions of P increased microbial respiration in peat with low P concentrations from the southern Everglades (Amador and Jones 1993). In peat with high P concentrations, however, P additions did not stimulate respiration. Additions of ammonium did not stimulate respiration in the low P content peat but did in peat with a high P content. The stimulation of decomposition rate by inorganic N and P generally thought to be caused by the fact that the C/N or C/P ratios in recently senesced dead plant matter are greater than optimal for building microbial biomass (Swift et al. 1979). Bacteria and fungi, however, can take up exogenous inorganic N and P to supplement the nutrients in litter being decomposed. As a result of this process of immobilization, plant litter in the early stages of decomposition often takes up and stores certain exogenous nutrients. This process of immobilization has been shown to play a role in reducing concentrations of inorganic N on water of swamp streams (Qualls 1984). The initial P concentration in the litter also exerts an influence on the decomposition rate. DeBusk and Reddy (1998) incubated samples of litter and peat gathered from a nutrient enrichment gradient in WCA-2A in the laboratory and found that CO2 production correlated with the initial P content of the substrate, which, in turn, was higher in the nutrient-enriched area of the gradient. Thus, P enrichment might speed decomposition rate of litter in the Everglades both by supplying exogenous P for microbial immobilization and by influencing the initial quality of the substrate. They also found that cattail litter decayed faster than sawgrass litter. We believe that the following hierarchy of effects might explain the varying results of N or P additions on litter decomposition in different ecosystems. First, if the internal N or P content (as expressed by the C/N or C/P ratio) of litter in a given experiment or ecosystem is not high, then additions of exogenous nutrients are unlikely to speed decomposition. If the internal C/N or C/P ratio is high enough to make the uptake of exogenous N or P beneficial, then the ambient inorganic N or P concentrations in the water may be already high enough to supply demand, again causing the addition of more exogenous nutrients to have no effect. Finally, if internal nutrient concentrations are suboptimal and ambient inorganic nutrient concentrations

444

R.G. Qualls and C.J. Richardson

in water are low, then further additions of exogenous nutrients may accelerate decomposition. Then the ratio of N/P in the litter and in the ambient water may determine which element (N or P), when added, may accelerate decomposition. Examining the N/P ratio alone without first showing that the C/P or C/N ratios are suboptimal would be meaningless. For example a plant residue that is very high in both N and P content might provide optimal N and P concentrations initially but the N/P ratio might be extremely high. We will show that litter in the unenriched areas of the Everglades initially contains extraordinarily low P concentrations and P concentrations in the water are also very low. Consequently, we might expect that additions of inorganic P might accelerate decomposition accordingly. In the enriched area of WCA-2A (see Fig. 6.19), soil P levels have been elevated by up to a factor 2.5–3 in peat accumulated since the 1960s (Craft and Richardson 1993a,b; Reddy et al. 1993; DeBusk et al. 1994) in both labile and refractory forms of soil P (Qualls and Richardson 1995). C/N ratios in the peat, however, are similar along this gradient. The potential effects of this P enrichment are of interest because of (1) potential effects on the accretion of peat and (2) potential effects on oxygen demand in the water column caused by decomposition. The question of the effects of P on decomposition and the mechanisms of P removal have also assumed additional importance because the world’s largest constructed wetland for nutrient removal has been constructed on the periphery of the northern Everglades as part of an Everglades restoration program (Guardo et al. 1995; SFWMD 2006). The process of microbial immobilization in litter could play an important role in storing P in peat accreting in these enriched fens. The objectives of the experimental studies summarized here were: 1. To experimentally test whether PO4 enrichment alone affects plant litter decomposition rate 2. More specifically, to test the quantitative relationship between PO4-P concentration and decomposition rate of sawgrass and cattail litter 3. To determine whether increased PO4 concentrations result in increased immobilization of N and micronutrients such as Cu 4. To determine whether the P additions resulted in large increases in the content of soil microbial biomass P and to determine the depths to which any effects would occur 5. To determine effect of N or P enrichment and aerobic vs. anaerobic conditions on C, N, and P mineralization of peat

17.4

Methods

Senesced leaves of Typha domingensis and Cladium jamaicense were placed in litter bags in each of the five experimental channels receiving different input of PO4 additions at each of two sites late in the month of September. These experimental mesocosms were described in an earlier chapter. Bags were placed at a distance of

17 Decomposition of Litter and Peat in the Everglades

445

1 m from the source of nutrient input. Because of uptake and dilution, concentrations of PO4-P were lower than the nominal input concentrations but were measured every 2 weeks at the location of the bags. Concentrations of total P, dissolved organic P, nitrate + nitrite, ammonia, Ca, and K were monitored monthly. Methods of analysis of water chemistry were detailed in Qualls and Richardson (1995). Other methods are detailed in Qualls and Richardson (2000). Bags of Cladium litter were retrieved from the unenriched control and the channel receiving the highest level of PO4 input channels after 32, 128, 245, and 365 days of incubation to observe the trends in decomposition and immobilization of nutrients over time. In addition, three bags (in most cases) of both species from all channels were retrieved after 365 days to observe a more detailed trend in annual decomposition as a function of concentration. Leaves were gently and carefully brushed in a pan of water to remove any periphyton. Examination under a dissecting microscope indicated that no significant amount of leaf material was lost by this procedure. Litter was analyzed for C, N, total P, Ca, K, and Cu content by methods summarized in Qualls and Richardson (2000). After 8 months of P additions, soil cores were taken at the 1 m distance in each of the channels except the ones receiving the lowest dose of added P. A 2 mm slice was shaved off the soil surface to exclude benthic algae from the sample and sliced into 3 cm increments to 24 cm depth. Microbial biomass and other P fractions were measured in these soil increments using the fractionation procedure of Hedley et al. (1982) as detailed in Qualls and Richardson (1995) for other samples taken from the WCA-2A nutrient enrichment gradient.

17.4.1

Mineralization Potential of Peat Along a Nutrient Enrichment Gradient

Enrichment with P may affect not only the decomposition of recently shed litter but also decomposition of older peat and any effects may differ under aerobic vs. anaerobic conditions. Consequently, we evaluated the effects of aerobic vs. anaerobic conditions, and N enrichment and P enrichment on mineralization of peat samples taken from along the WCA-2A nutrient enrichment gradient (see Chap. 15 describing this gradient). We used as closely as possible, the Stanford and Smith (1972) incubation method for “easily mineralizable N” since it has become a widely used standard method but we also measured the loss of carbon, mineralization of P, and the leaching of dissolved organic forms. We also adapted the method to measure anaerobic mineralization. Carbon mineralization over the period was measured by loss of weight and carbon content in the solid soil that we could not account for in the dissolved organic C (DOC) leached. Samples from six plots, representing the extremes of the nutrient enrichment gradient in WCA-2A, from two depths were incubated in the spring of 1991. Samples were taken from three plots in the enriched area and three plots in the unenriched area. Each soil was incubated under the following treatments at 35°C for 26 weeks:

446

R.G. Qualls and C.J. Richardson 1. Aerobic 2. Anaerobic

(a) P amendment (for N mineralization) (b) N amendment (for P mineralization) (a) P amendment (for N mineralization) (b) N amendment (for P mineralization)

Aerobic incubations were done in Falcon Filter Flasks with 5 g equivalent dry mass. After leaching, they were maintained at 0.02 MPa matric water potential by applying regulated vacuum below the 0.2 mm membrane filter. This water potential was chosen because it was close to the optimum for mineralization (Terry and Tate 1980) but was not so wet as to encourage denitrification. Each aerobic sample was leached every 2 weeks with eight cycles of 20 ml of the nutrient solution. Anaerobic samples of 0.5 g equivalent dry masses were incubated in 40 ml of nutrient solution in sealed centrifuge tubes. The nutrient solution was purged with N2 and added inside a bag purged with N2, and tubes were capped and sealed with parafilm. After incubating 2 weeks, the tube was centrifuged, the supernatant was collected, and the soil was washed five times with 0.01 M CaCl2, then new nutrient solution was added anaerobically. The nutrient supplement solutions were meant to remove the limitation of any nutrient except the one under study so that the changes in the substrate quality dictated the rate of decomposition (Stanford and Smith 1972). For the aerobic N mineralization, the solution used by Stanford and Smith was used. In the solution with P amendments, ammonium and nitrate were substituted for PO4. On each set of leachate, we measured PO4, NH4, NO3 + NO2, DON, DOP, and DOC. At the end of the experiment, dry weight, C content, N content, and total P content were measured on initial and final subsamples of the soil.

17.5

Results

The results of the litter decomposition experiment showed that additions of inorganic P clearly accelerated decomposition and resulted in greatly increased amounts of exogenous P immobilized and stored in the litter. At the end of 1 year of incubation, the loss of C by the litter was strongly correlated to the average PO4-P concentration (Fig. 17.1a, b). The slopes of the lines relating fraction of initial C remaining to average PO4-P concentration were similar for both species but the intercept was less for cattail litter. Cattail litter decayed faster than the sawgrass, but the response to PO4 concentration as indicated by the slope of the lines in Fig. 17.1 a, b was similar. The first-order decay constants for the litter based on the predicted regression values of fraction of initial C remaining ranged from 0.46 to 1.11 year−1 for sawgrass litter and 0.59 to 1.30 year−1 for cattail litter over the range of PO4-P concentrations. Only sawgrass litter was sampled over the course of the year, and during that time the sawgrass litter in the channel enriched with the highest level of P lost more C than litter in the unenriched control (Fig. 17.2a). In the channel receiving the highest input of P, the litter accumulated absolute quantities of P that were eight to ten times those present in the initial litter by 250 days (Fig. 17.2b). Between 265 and 365 days,

Fig. 17.1 Percent of initial C remaining after 1 year of decomposition of (a) sawgrass and (b) cattail litter as a function of average PO4-P in the water column in five experimental channels at each of two sites. Each data point represents one litter bag sample (reproduced from Qualls and Richardson 2000, with kind permission from The Soil Science Society of America)

448

R.G. Qualls and C.J. Richardson

Fig. 17.2 Percent of initial C (a), P (b), N (c), and Cu (d) remaining over the course of 1 year of decomposition of sawgrass litter in channels receiving either the highest level of P enrichment or no P enrichment. Lines indicate the averages of data points that represent individual litter bags (reproduced from Qualls and Richardson 2000, with kind permission from The Soil Science Society of America)

17 Decomposition of Litter and Peat in the Everglades

Fig. 17.2 (continued)

449

450

R.G. Qualls and C.J. Richardson

P remained constant or dropped slightly in concentration. In the unenriched channel, P was lost and never accumulated above the quantity present in the initial litter (Fig. 17.2b). Sawgrass litter accumulated more N than was present in the initial litter in both the enriched and unenriched channels after a 30-day period (Fig. 17.2c). Patterns of N accumulation over time were variable but not different between the enriched and unenriched control channels. Copper also accumulated twofold to threefold in absolute quantity over that present in the initial sample (Fig. 17.2d).

Fig. 17.3 Percent of initial P remaining after 1 year of decomposition of sawgrass (a) and cattail (b) litter as a function of average PO4-P in the water column in five experimental channels at each of two sites. Each data point represents one litter bag sample (reproduced from Qualls and Richardson 2000, with kind permission from The Soil Science Society of America)

17 Decomposition of Litter and Peat in the Everglades

451

The fraction of the original mass of element remaining in the litter at 1 year as a function of average PO4-P showed a dramatically increasing immobilization of P up to a concentration of about 15 µg l−1 (Fig. 17.3). However, the percentage of original N, Cu, Ca, and K remaining at 1 year was not related to the average PO4-P concentration in water (Qualls and Richardson 2000). Several factors other than PO4 concentration that might hypothetically affect decomposition could not explain the pattern among channels. The average concentrations of NH4, NO3, Ca, dissolved oxygen, pH, and water temperature were each regressed against the data for proportion of original mass remaining after 1 year, but no significant correlations were found (Qualls and Richardson 2000). All litter bags were inundated with water for the entire year of incubation. The chloroform released-P (proportional to microbial biomass P) content in soil was shown in Fig. 17.4a for the control, the intermediate input, and the highest inputs of P at both sites. In the surface depth increment of the soil of the most enriched channels, the chloroform released-P was approximately nine times higher than the control in the most enriched channel. Chloroform released-P in the surface increment of the channel receiving intermediate P inputs was elevated about three times higher than that in the control. These dramatic differences declined with depth and by the 12–15 cm depth increment, all channels were similar. In contrast, the exchangeable P in the soil was far lower than the microbial biomass P content and was only elevated in the upper increments of the channels receiving the highest inputs (Fig. 17.4b).

17.5.1

Mineralization of Peat

Under aerobic conditions, additions of P resulted in considerably greater C mineralization than N-amended samples (Fig. 17.5a). This difference was observed in both the surface and subsurface soils. In addition, soils from the northern enriched area of WCA-2A mineralized much more C than those from the southern unenriched area when corresponding treatments are compared. Under anaerobic conditions, in the surface soil, P-amended soil from the enriched area mineralized much faster than P-amended soil from the unenriched area (Fig. 17.5b). In the subsurface soil, rates were similar in the soils from unenriched and enriched areas. If we compare aerobic vs. anaerobic C mineralization for comparable amendments and horizons, the aerobic C loss is approximately four times the anaerobic C loss for the P-amended treatments. In the anaerobic incubations, the percentages of carbon mineralized were very low and were not significantly different between the corresponding P-amended (Fig. 17.5) and N-amended (not shown) treatments. In the surface soil, mineralization of N from soils of the enriched area was far greater than mineralization of N from soils of the unenriched area despite the fact that all samples were incubated with P amendments (Fig. 17.5c). In the subsurface soil, however, N mineralization was not greater for soils collected from the enriched area compared to soils collected from the unenriched area. While samples from the surface soil mineralized more N than did the corresponding soils under anaerobic

452

R.G. Qualls and C.J. Richardson

Fig. 17.4 Chloroform released-P (a) and exchangeable PO4-P (b) in the soil of channels receiving inputs of no P, the intermediate P level, or the highest P level. Solid points represent site 1 and hollow points represent site 2. Lines connect the average of the values for the two sites. Points are plotted for the 0–3, 6–9, 12–15, and 21–24 cm depth increments (reproduced from Qualls and Richardson 2000, with kind permission from The Soil Science Society of America)

conditions, in the subsurface soils there was little difference between the corresponding aerobic and anaerobic treatments. In our 6-month experiment to determine potentially mineralized N, the leaching of DON rivaled the mineralization of N from the soil (Fig. 17.5c). Under aerobic conditions, 1.2–1.8% of the soil N was leached as DON while 2.8–8.2% was mineralized. From this data, it is easy to see why such high concentrations of DON were drained from the aerobic soils of the agricultural area. Under aerobic conditions, significantly more DON was leached from soils collected in the enriched area

17 Decomposition of Litter and Peat in the Everglades

453

Fig. 17.5 Percent of C lost (a), mineralization of C (b), N (c), and P (d), during aerobic and anaerobic decomposition of peat from the P-enriched areas of northern WCA-2A and P-unenriched areas of southern WCA-2A in surface (0–5 cm depth) and subsurface (20–25 cm depth) soil over a 6-month period. Error bars represent standard deviation

454

R.G. Qualls and C.J. Richardson

Fig. 17.5 (continued)

than those from the unenriched area (t-test, p < 0.05), the same pattern that was observed in concentrations in pore water in the field. Under anaerobic incubation even more DON was leached, 3.4–6% of the soil N, much more than was mineralized. The high rate of leaching under saturated conditions may partly reflect a more efficient leaching of desorbable organic matter in addition to microbial dissolution of solid matter.

17 Decomposition of Litter and Peat in the Everglades

17.6 17.6.1

455

Discussion Litter Decomposition

Increases in PO4-P concentration appear to increase the decomposition rate of leaf litter in the WCA-2A area of the Everglades, at least over a period of 1 year. This finding is consistent with the study of Davis (1991) who found higher litter decomposition rates at a site in the nutrient-enriched area of WCA-2A compared to a more southerly unenriched site. In this study, we have experimentally isolated the effects of PO4 and demonstrated a quantitative relationship to average PO4 concentration. The increase in decomposition rate due to addition of exogenous nutrient is often regarded as occurring during the initial stages of decomposition when concentrations of the leaf tissue are low (i.e., C/P or C/N ratios are high) (Swift et al. 1979). As litter decomposes, and N and P are concentrated in the litter, a “critical” C/N or C/P ratio is eventually reached, after which net mineralization and release of N and P occur. In this study, much of the acceleration of decomposition may have occurred in the early, more nutrient-limited stages of decomposition. It may be that the early stages of relatively rapid decomposition above the peat surface in a zone which is aerobic before the litter becomes buried in the peat are crucial to the long-term accretion and P content of the peat. In fact, the final concentration of P in the cattail litter after 1 year in the channel receiving the highest P input was 1,274 µg g−1 (± 282 SD) (Table 17.1) which is only slightly less than the P concentration in the 0–5 cm depth of the four most enriched plots (1,400–1,550 µg g−1 soil) along the nutrient enrichment gradient (Qualls and Richardson 1995). This correspondence between the P concentration of the cattail litter in the most enriched channel and that found in the enriched zone of the northern Everglades not only suggests

Table 17.1 The initial concentrations of P, N, Cu, Ca, and the C:P, C:N, and N:P ratios in sawgrass and cattail litter P (µg g−1)

N (%)

Cu (µg g−1) Ca (µg g−1)

Sawgrass Initial 41 (4) Finala 791 (78)

0.299 (0.004) 1.3 (0.2) 2.66 (0.54) 14 (3)

Cattail

0.411 (0.006) 0.9 (0.2)

Initial 112 (5)

Finala 1,274 (112) 2.28 (0.13)

6 (0.2)

2,010 (40) 14,000 (810) 10,900 (220) 25,400 (2,100)

C:P ratio

C:N ratio

N:P ratio

11,800 161 374 17.8

73 34

4,330

117

37

609

20.8

18

a The final concentrations and elemental ratios in litter collected after 1 year in the most P-enriched channel at both sites are also shown to contrast the maximum P concentrations attained (from Qualls and Richardson 2000). The means of litter in the channels receiving the highest P inputs at both sites after 1 year incubation

456

R.G. Qualls and C.J. Richardson

a reasonable correspondence to decay conditions there, but also that the litter approached the critical C/P ratio within a year. The stimulation of submerged litter decomposition by nutrient additions has been demonstrated in several studies, but some others have demonstrated the lack of an effect as reviewed earlier. The initial P content was very low and the initial N/P ratio (by mass) in our cattail litter was 37 and in the sawgrass litter was 73 (Table 17.1). Both of these ratios are relatively high and in the range which might suggest P rather than N limitation of litter decomposition (Vogt et al. 1968). This study in the Everglades mesocosms is the first that we are aware of which has demonstrated quantitative relationship over a range of average nutrient concentrations on litter decomposition.

17.6.2

Immobilization of Macronutrients and Cu by Litter

The immobilization of P by litter in response to increased PO4-P supply may be an important mechanism of removing and storing excess P from the overlying water. It may also help explain the dramatically higher P content and P accretion rate of peat in the enriched areas of WCA-2A (Davis 1991; Craft and Richardson 1993a,b). Craft and Richardson (1993b) found up to 2.5–3 times the P content (µg g−1 soil) in peat deposited since the 1960s in the P-enriched area of WCA-2A compared to the unenriched areas further from the source of nutrient inputs in canal water. In litter decomposed for 1 year, we found up to eight to ten times the absolute mass of P in litter in the most enriched channel compared to the control channel. Since this litter will eventually be incorporated in peat, this uptake of P is likely to be very important in the P budget of the Everglades. We had hypothesized that an increase in microbial activity stimulated by PO4 availability, and accompanied by an increase in P uptake by litter, might also result in a greater uptake of N by the microbial biomass. This might occur even under conditions where N was not limiting to decomposition simply because a higher microbial biomass would result in a higher quantity of N in the biomass. Exogenous N was apparently immobilized and accumulated in litter as shown by the twofold to fourfold increases in N remaining in the litter, but it did not accumulate differently in the P-enriched litter despite the dramatic differences in P in the litter. A greater cycling of C and P through the microbial biomass due to turnover rather than standing stock may explain why the accumulation of P was so much different from that of N. Likewise, the N content of peat was similar in the unenriched and enriched areas (Craft and Richardson 1993a), and indeed we found that N uptake by litter was similar regardless of P treatment. The general patterns of net uptake, and in some cases, subsequent decreases of P, N, Ca, and K were typical of the patterns observed in litter decomposition studies but they differed in magnitude and timing. Brinson (1977), for example, found a net increase in N, P, and Ca in the early phases of decomposition followed by a later net decrease as the weight loss of the litter progressed to the advanced stages of

17 Decomposition of Litter and Peat in the Everglades

457

decomposition, presumably indicating a phase of net mineralization. Brinson also found, as we did, that K leached readily and did not accumulate in litter. Cu was also taken up and accumulated by the decomposing litter to amounts that were generally between 1.5 and 6 times that in the initial litter (Fig. 17.2d) of 1 year. Cu is potentially a limiting nutrient for plant growth in the Everglades peat soils. Cu limits the growth of crop plants grown on drained Everglades peat as it does in many histosols (Coale 1994). This process of uptake of Cu may be one that competes for available Cu with living plants and algae and may be the initial step in long-term sequestration of this potentially limiting nutrient. Humic acids bind Cu strongly and humification of decomposing litter could be an explanation for the uptake by litter, in addition to microbial uptake and immobilization. This is the first report of which we are aware of demonstrating Cu uptake by decomposing litter. The P-enriched areas of the northern Everglades have also been enriched in Cu (Vaithiyanathan and Richardson 1997b) with a gradient in Cu content corresponding to a well-established gradient in soil P concentrations. Since crops in the EAA are fertilized with Cu (Coale 1994), runoff may be a possible source. Thus, immobilization of Cu by litter may aid in reducing concentrations in surface water.

17.6.3

Microbial Biomass P in Soil

Microbial biomass P concentration, as indicated by the chloroform released-P, was greatly elevated in the surface layers of soil. The origin of this P in the microbial biomass might have been (a) immobilization of P in microbes decomposing detritus present in the soil before the P additions began, (b) assimilation of P from P-rich periphyton detritus deposited after P additions began, and (c) assimilation of P from P-enriched macrophyte leaf or root litter deposited after the P additions began. While benthic algal periphyton was observed on the surface, we had shaved off the 2 mm surface layer that appeared to eliminate any noticeable green color. The restriction of the P enrichment of the microbial biomass to the near surface layers might suggest deposition of P-rich detritus from above, but immobilization of P by microbes on older detritus might also be consistent with this observation since the more recent upper layers of soil might be expected to harbor a more active microbial community. It was notable that many macrophytes had a greater density of roots in the peat below the P-enriched layer so it appeared that if there was any translocation of P to the roots from the aboveground parts, it had not cycled to the deeper microbial biomass at that time. In a study of 15 different soils, Brookes et al. (1984) found no relationship between the percentage of P in microbial biomass and the bicarbonate exchangeable PO4 in the soil. In a study of the effect of fertilizer on a cropped soil, Ghoshal and Singh (1995) found a 37% greater microbial biomass P in the fertilized soil than the control. DeBusk and Reddy (1998) found that microbial biomass C in surface peat was greater in the P-enriched area of WCA-2A, suggesting that P enrichment resulted in higher microbial biomass either through effect of P on P-limited

458

R.G. Qualls and C.J. Richardson

microbial growth or by influencing the replacement of sawgrass with the more rapidly decomposing cattail. The dramatic increase in microbial biomass P in our study was greater in magnitude than any comparable study we could locate. Two of the most rapid responses to P enrichment that result in P removal might be adsorption to the soil and microbial immobilization. Comparison of the small elevation of exchangeable P to the massive elevation of microbial biomass P shows the importance of the microbial processes in this peat soil with a limited adsorption capacity (Richardson and Vaithiyanathan 1995) compared to mineral soils. This response might be contrasted to the initial removal of P in Fe- or Al-rich mineral soils by adsorption (Richardson and Marshall 1986; Richardson 1985).

17.6.4

Peat Mineralization

Under aerobic conditions, it seems likely that the decomposition of peat is stimulated by P enrichment. This was likely the result of two separate processes (1) direct effects on microbial decomposition of exogenous PO4 availability under otherwise P-limited conditions and (2) the replacement of sawgrass peat by cattail peat in the enriched area. Under aerobic conditions, the P-amended soils from both depths and locations decomposed much faster than with N amendments. The greater rate of decomposition of peat from the enriched area of WCA-2A compared to the unenriched area in the N-amended treatments could have been caused by the higher internal P concentration of the peat or the fact that peat formed from cattail is lower in lignin and is inherently more labile. We suspect that both factors contribute to this result. The results of the litter decomposition experiment showed that, at least in the first year, cattail litter decomposes faster under any nutrient regime and this effect may extend to cattail peat. In addition, the findings of DeBusk and Reddy (1998) – that CO2 production correlated with the initial P content of the substrate, which in turn, was higher in the nutrient-enriched area of the gradient incubated – suggest that the elevated P content of the peat contributed to the effect. There is other evidence that P enrichment stimulates the decomposition rate of peat from the Everglades since Craft and Richardson (1993b) found that the balance of NPP and peat accretion would require higher rates of decomposition in the P-enriched areas compared to unenriched areas. Our results are also consistent with Amador and Jones (1993) findings that additions of P increased microbial respiration under aerobic conditions in peat with low P concentrations from the southern Everglades. In peat with high P concentrations, however, P additions did not stimulate respiration. In our study, however, even in soils from the enriched area, the P-amended soil decomposed faster than the N-amended soil. In Amador and Jones study, additions of ammonium did not stimulate respiration in the low P content peat but did in peat with a high P content.

17 Decomposition of Litter and Peat in the Everglades

17.7

459

Conclusions and Lessons for Restoration

This research has demonstrated P limitation of decomposition of sawgrass and cattail litter over a range of phosphate concentrations in water. Mesocosms proved useful in isolating the effect of P enrichment on several wetland microbial processes in controlled, replicated, but natural field environments. These studies also demonstrated a dramatically increased uptake of PO4-P by litter in phosphateenriched channels. It also demonstrated uptake of Cu and N by litter but no significant effect by the P enrichment of water on the uptake of Cu, N, Ca, or K. Microbial biomass P concentrations in the soil of the P-enriched channels were elevated by up to a factor of 9, but this elevation of microbial biomass P content was restricted to the upper 12 cm of soil. Responses of the microbial community in litter and the surface layers of the peat soil were rapid, dramatic, and resulted in far greater uptake of P than adsorption as indicated by exchangeable P. In the Everglades, P enrichment has been shown to increase NPP, increase peat accretion, and increase litter decomposition rate – a general acceleration of carbon and nutrient cycling. While litter decomposition rates are increased, large increases in NPP (Davis 1991; Chap. 6) apparently more than compensate for the increased decomposition rate in controlling the rate of peat accretion (Craft and Richardson 1993b). Concern has also been expressed that the increased oxygen demand created by these increases in decomposition may also adversely affect dissolved oxygen concentrations (Amador and Jones 1993) but we have observed no large differences in soil redox potential (Qualls et al. 2001) along the nutrient enrichment gradient in WCA-2A. The remarkable capacity for extremely P-deficient litter and peat to immobilize large quantities of P also works to remove PO4 from water. The general response of the Everglades to P enrichment is that of an extremely P-deficient ecosystem which responds to P additions through increases in several ecosystem level functions.

18

Experimental Assessment of Phosphorus Effects on Algal Assemblages in Dosing Mesocosms Jan Kaštovský, Klára Řeháková, Marek Bastl, Jan Vymazal, and Ryan S. King

18.1

Introduction

Periphyton covered by calcareous precipitations in the Everglades is mostly dominated by cyanobacteria (Cyanoprokaryota, blue-green algae). In some areas of the Everglades, diatoms, desmids, and a few species of filamentous green algae form a significant part of the periphyton. Filamentous green algae (e.g., Spirogyra spp. or Mougeotia spp.) are found mostly in areas with elevated nutrient levels (Van Meter 1965; Wood and Maynard 1974; Swift and Nicholas 1987; Vymazal and Richardson 1995). It has been repeatedly reported that the major factor affecting periphyton growth and species composition is water quality and, especially, phosphorus concentration in the water (Browder et al. 1981; Swift and Nicholas 1987; Rader and Richardson 1992; Vymazal et al. 1994; McCormick and O’Dell 1996; Raschke 1993; McCormick et al. 1996, 1998; Stevenson and Richardson 1994; Pan et al. 2000). The major objectives of this study were (1) to evaluate the effect of phosphorus concentration on species variability and relative abundance of algae and cyanobacteria in natural periphyton communities and (2) to compare periphyton communities growing on natural vs. artificial substrates.

18.2 18.2.1

Material and Methods Study Site and Sampling

The experiments were carried out in the experimental dosing channels in the Everglades WCA-2A (for detailed description, see Chap. 14). Samples of floating periphyton mats were taken by inserting the open end of a 50 ml centrifuge tube (27 mm diameter) through the mat (away from edges) at 2, 4, and 6 m from the head of the channels. The sampling dates were 20 August 1996, 16 December 1996, 26 January 1998, 24 September 1998, 29 October 1998, 23 February 1999, and 7 December 1999. Plexiglas slides were used to compare naturally growing periphyton and periphyton growing on artificial substrate. Slides (7.5 × 2.5 cm2) were suspended 461

462

J. Kaštovský et al.

from PVC poles that were positioned across the channels 2, 4, and 6 m from the head of channels where water was added. The slides were 5 cm below the water surface in the upright position and weighted with a lead sink on the lower side. Slides were left in position for 2 months. All biomass samples were analyzed for dry matter, ash-free dry matter, TP, TN, and TC. The total number of samples was 220 and 237 for mats and slides, respectively. Samples for species composition evaluation of floating mats were taken on 26 January 1998, 24 September 1998, 29 October 1998, 23 February 1999, and 7 December 1999. Samples on Plexiglas were taken on 24 September 1998 and 23 February 1999. Samples were placed into 50 ml centrifuge vials with 1.5% solution of formaldehyde. The calcareous crust was removed by using an 8% solution of citric acid. A total of 72 mat samples and 25 samples from Plexiglas slides were analyzed. The samples were selected from sites that represent as wide as possible range of phosphorus concentrations in the water. Samples were evaluated in an optical microscope, and for individual species the relative abundance within the periphyton complex was given. The evaluation methodology was based on the phytosociological relevés used for higher plants (Braun-Blanquet 1964). The scale of relative abundance estimation (Table 18.1) was modified for algal periphyton and for the needs of statistical analyses (Kaštovský and Komárek 2001). Identification of cyanobacteria (Cyanoprokaryota, blue-green algae) was made using the following literature sources: Gardner (1927), Anagnostidis and Komárek (1988), Komárek (1989), and Komárek and Anagnostidis (1998). For diatoms (Bacillariophyta), identification monographs of Patrick and Riemer (1966) and Kramer and Lange-Bertalot (1988, 1991a,b, 1997) were used. Total phosphorus (TP) concentrations for individual sampling dates were determined as a mean TP concentration during the period of 2 months before the sampling. To evaluate the influence of TP concentrations on periphyton species composition, TP concentrations were grouped into the following categories: <10 µg l−1, 10–15 µg l−1, 15–20 µg l−1, 20–30 µg l−1, and >30 µg l−1. Phosphorus concentrations here represent the arithmetic mean of TP values collected over five sampling dates (see Sect. 18.2.2); thus, P response categories are slightly higher and do not represent the geometric TP data used in the our P threshold analysis (Chap. 25).

Table 18.1 Scale of estimation of relative abundance of periphyton species Symbol Abundance (%) 1 2 3 4 5 6 7

<0.1 0.1–1 1–5 5–20 20–50 50–90 90–100

18 Experimental Assessment of Phosphorus

18.2.2

463

Data Analysis

Constrained ordination technique Canonical Correspondence Analysis (CCA) was used for analyzing the periphyton composition (97 samples) relative to TP concentrations, grouped TP concentrations (five intervals), date of sampling, and sample origin (mats and Plexiglas). Statistical significance of variables was tested by Monte Carlo permutation test. Unconstrained ordination technique Detrended Correspondence Analysis (DCA) was used for analyzing the composition of floating periphyton mats (71 samples). Passive variable TP concentration, number of species in the sample, and Shannon–Weaver diversity index of sample were used. The Canoco for Windows 4.5 package (Ter Braak and Šmilauer 2002) was used for all multivariate vegetation analyses.

18.3

Results and Discussion

A total of 210 species of algae and cyanobacteria were identified in this study (Table 18.2) Cyanobacteria (Cyanoprokaryota, blue-green algae) are clearly the dominant group as far as the biomass and number of species are concerned with the filamentous cyanobacteria being the most important group. The green algae (Chlorophyta) occur in high quantities, especially when TP concentrations are high. Diatoms (Bacillariophyceae) are ubiquitous and the number of species found is the highest of all groups (Table 18.2) but their relative abundance is very low with the exception of a few species such as Amphora coffeaeformis or Mastogloia smithii. Species of other algal groups were found in very low numbers (Table 18.2).

Table 18.2 Number of algal and cyanobacterial species of individual taxonomic groups found during the P-dosing study Group Number of species Cyanoprokaryota–Oscillatoriales Cyanoprokaryota–Chroococcales Cyanoprokaryota–Nostocales prokaryota–Stigonematales Cyano (total Cyanoprokaryota) Bacillariophyceae Conjugatophyceae Chlorophyceae Dinophyceae Xanthophyceae Chrysophyceae Euglenophyceae Total number of species

48 37 17 1 (103) 60 24 18 2 1 1 1 210

464

18.3.1

J. Kaštovský et al.

Effect of TP on Periphyton Composition

Figure 18.1 shows the results of CCA analysis with grouped TP concentrations. The first and the second axes account for 4.6 and 2.2% variability in species data, respectively. The test of significance of all canonical axes indicates that the effect of grouped TP concentration was highly significant (p = 0.001, F = 2.196). In general, the samples taken from TP concentration intervals <10 µg l−1 and 10–15 µg l−1 are very similar, and those two concentration intervals are very different from samples taken from TP concentrations 20–30 µg l−1 and >30 µg l−1 (Fig. 18.1). Samples from

Fig. 18.1 Canonical Correspondence Analysis (CCA) of 97 samples of periphyton and the concentrations of total phosphorus divided into five categories (<10 µg l−1, 10–15 µg l−1, 15–20 µg l−1, 20–30 µg l−1, and >30 µg l−1). For species abbreviations, see Appendix

18 Experimental Assessment of Phosphorus

465

the TP concentration range 15–20 µg l−1 represent a slightly different community, and their character is more similar to that of higher TP concentration ranges. Periphyton community growing at TP concentrations <15 µg l−1 could be considered as nonimpacted by eutrophication, while changes in the periphyton composition start to occur above a 15 µg l−1 TP concentration (Fig. 18.1). The results of CCA analysis with ungrouped TP concentrations indicate that the first and the second axes account for 4.3 and 11% variability in species data, respectively. The test of significance of all canonical axes shows that the effect of TP concentrations was highly significant (p = 0.001, F = 4.284). Individual species respond to increasing TP concentrations differently. 1. Most species either disappear or substantially decrease in abundance. Among those species the most important are: – Members of cyanobacterial genus Aphanothece – Most species of genera Chroococcus, Stigonema, Scytonema, Komvophoron, and Gloeothece interspersa – Diatoms M. smithii, A. coffeaeformis, and Fragilaria ulna – All observed members of the green algae genus Cosmarium However, it is important to note that several species (Scytonema sp., Scytonema sp. 22, Nostoc sp., Nostoc sp. 4, and Phormidium sp. 4) occur only at TP concentrations <10 µg l−1 (Fig. 18.2). 2. The species of the second group exhibit their maximum occurrence between TP concentrations 10 and 30 µg l−1:

Fig. 18.2 Decrease of relative abundance of Scytonema sp. 22 with the increase of TP concentration

466

J. Kaštovský et al.

– Cyanobacteria Gloeothece sp. 9.7 × 7, Chroococcus deltoids, Komvophoron sp. 2 × 1.2, Leptolyngbya sp. 2, sp. 3, sp. 4 × 4, Lyngbya sp. 15, sp. 18, Oscillatoria spp., Phormidium sp. 5 × 1.5, sp. 5 × 6, Pseudanabaena sp. 2, Scytonema sp. 10, sp. 12, sp. 15, sp. 17, and Spirulina subsalsa – All members of the Conjugatophyceae genera Mougeotia (Fig. 18.3) and green algae (Chlorophyceae) Oedogonium, Bulbochaete sp. 50 × 12, and Staurastrum sp. – Diatoms Navicula lanceolata, Navicula rhynchocephala, Navicula sp. 50, Neidium sp., Nitzschia linearis, Nitzschia recta, Cymbella sp. 53, Cymbella cf. cymbiformis, Cymatopleura solea, Diploneis cf. elliptica, Gomphonema olivaceum, and Pinnularia gibba 3. Several species occur across the TP concentration range found in the dosing study. Among those the most abundant are Cyanobacteria Leptolyngbya sp. 1 (Fig. 18.4) and diatoms Cocconeis placentula and Neidum comasii 4. There are only two species among all species found during the study responding positively to a very high increase in TP concentration: cyanobacterium Lyngbya sp. 10.5 and diatom Rhopalodia gibba (Fig. 18.5) Diatoms have frequently been used as indicators of elevated TP concentrations in the Florida Everglades ecosystem (e.g., Swift and Nicholas 1987; McCormick and O’Dell 1996). However, cyanobacteria and green algae could act as indicators of TP concentration in water as well, and their ecological role in the natural periphyton communities is more important. Examples of good indicators of low TP concentration in the water (and, therefore, examples of noneutrophic conditions) include the cyanobacteria Scytonema 12, Scytonema 22, Chroococcus minutus, or Aphanothece variabilis. On the other hand, all species of genera Lyngbya (Cyanobacteria) or Mougeotia and Oedogonium (Chlorophyta) could be considered to be indicators of higher eutrophic levels. Importantly, the findings obtained from

Fig. 18.3 Relative abundance of Mougeotia sp. 11 along the dosing TP gradient

18 Experimental Assessment of Phosphorus

467

Fig. 18.4 Relative abundance of Leptolyngbya sp. 1 along the dosing TP gradient

Fig. 18.5 Relative abundance of Rhopalodia gibba along the dosing TP gradient

this study are in close agreement with the algal results obtained in the natural gradient research in WCA-2A (Chap. 10). Figures 18.6–18.8 give the results of DCA analysis with a projection of passive variables. The first and the second axes account for 11 and 6.8% of variability in species data, respectively. The number of species decreases with increasing TP concentration in water (Figs. 18.6 and 18.7). As a result, the diversity of periphyton communities decreases under eutrophic conditions (Figs. 18.6 and 18.7). Results in Figs. 18.6 and 18.7 indicate that the mean number of species found in samples growing in low TP concentrations was as high as 28, while in samples taken from

468

J. Kaštovský et al.

Fig. 18.6 Detrended Correspondence Analysis (DCA) of 71 samples of floating periphyton mats with projection of isolines of passive variable TP-dosing gradient. For species abbreviations, see Appendix 18.1

high TP concentration locations the mean number of species dropped to only 10. The Shannon–Weaver diversity index (Shannon and Weaver 1949) decreases from 3.3 in low TP sites to 2.1 in high TP sites (Figs. 18.6 and 18.7).

18.3.2

Seasonal Dynamics of Periphyton

Figure 18.9 shows the results of CCA analysis with date of sampling. The first and the second axes show 10.2 and 4.4% variability in species data, respectively. The test of significance of all canonical axes indicates that sampling date was highly significant (p = 0.001, F = 4.235). The results indicate seasonal dynamics for the periphyton community (Fig. 18.9). Samples taken during the winter (December, January, and February) differ from those taken in the fall (September and October). In the winter typical periphyton species are cyanobacteria Phormidium sp. 4, Phormidium sp. 8, Aphanocapsa cf. delicatissima, Leptolyngbya sp. 2, Leptolyngbya sp.

18 Experimental Assessment of Phosphorus

469

Fig. 18.7 Detrended Correspondence Analysis (DCA) of 71 samples of floating periphyton mats with projection of isolines of passive variable algal species number in sample. To compare species number with the TP-dosing gradient, see Fig. 18.6

4, Leptolyngbya sp. 2 × 5, Aphanothece minutissima, C. minutus, Scytonema sp. 10, Lyngbya sp. 12, Lyngbya sp. 15, G. interspersa, and diatoms G. olivaceum, N. linearis, or A. coffeaeformis. On the other hand, samples taken in early fall are dominated by cyanobacteria Pseudanabaena sp. 1, Phormidium sp. 5 × 6, Scytonema sp. 12, Scytonema sp. 22, Leptolyngbya sp. 3, Aphanothece comasii, Lyngbya sp. 6.5, sp. 18, and Conjugatophyceae Mougeotia sp. 4 and Mougeotia sp. 11.

18.3.3

Comparison of Natural and Artificial Substrates

Figure 18.10 gives the results of CCA analysis with the origin of the sample substrate as a factor (natural mats or Plexiglas). The first and the second axes indicate 5.2 and 9% variability in species data, respectively. The test of significance of all canonical axes reveals that the effect of sampling date was highly significant (p = 0.001, F = 5.17). Samples of floating mats differ from samples taken from

470

J. Kaštovský et al.

Fig. 18.8 Detrended Correspondence Analysis (DCA) of 71 samples of floating periphyton mats with projection of isolines of passive variable Shannon–Weaver diversity index. To compare Shannon–Weaver diversity index with TP-dosing gradient, see Fig. 18.6

Plexiglas slides (Fig. 18.10). Species that prefer Plexiglas could be found in floating mats but in substantially lower quantities. The same applies to species that prefer floating mats to Plexiglas. It is true that periphyton growing on Plexiglas gives a real qualitative picture of periphyton species composition at a given site, but quantitatively mats and Plexiglas are not comparable. For example, the dominant species for mats are cyanobacteria Scytonema sp., Phormidium sp. 4, Phormidium sp. 8, A. minutissima, Aphanocapsa cf. delicatissima, Lyngbya sp. 15, Leptolyngbya sp. 2, and G. interspersa. Also, diatoms A. coffeaeformis, M. smithii, F. ulna, G. olivaceum, and C. solea are very common in floating mats. By contrast, Plexiglas is preferred by cyanobacteria Lyngbya sp. 6.5, Lyngbya sp. 10.5, Lyngbya sp. 18, Leptolyngbya sp. 3, Phormidium sp. 5 × 6, A. comasii or A. bacillaris. Green algae Mougeotia sp. 5 and Mougeotia sp. 11 are also often found on Plexiglas.

18 Experimental Assessment of Phosphorus

471

Fig. 18.9 Canonical Correspondence Analysis (CCA) of 97 samples of periphyton and the date of sampling (yyyy/mm/dd). For species abbreviations, see Appendix

The difference between periphyton growing on Plexiglas and periphyton growing in floating mats is also evident from biomass analyses. Results presented in Fig. 18.11 clearly show more intensive growth of periphyton in floating mats as compared to Plexiglas. It is also obvious that 2 months of exposure is not long enough for periphyton to match the naturally growing periphyton biomass. However, it is questionable if higher biomass can actually stay on the Plexiglas since algae sloughing was found on older Plexiglas slides. It is also important to note that the 2-month period is longer than the exposure period usually used in the state of Florida’s Everglades studies, which may well result in underestimates of periphyton communities (SFWMD 2003). Our substrate comparison results also indicate that the content of ash, P, and N differs substantially depending on substrate. Due to higher ash content in samples of naturally growing periphyton in the mat, the biomass P and N concentrations are much lower as compared to periphyton growing on Plexiglas slides (Fig. 18.12).

472

J. Kaštovský et al.

Fig. 18.10 Canonical Correspondence Analysis (CCA) of 97 samples of periphyton and the origin of the sample (Mats sample of floating periphyton mats, Plexi samples of periphyton on Plexiglas slides). For species abbreviations, see Appendix

18.4 Conclusions and Lessons for Restoration A total of 210 species of cyanobacteria and algae were identified during our study in floating mats of periphyton in the experimental P-dosing channels. Most of the species found in the periphyton assemblages were cyanobacteria (103), diatoms (60), and green algae (42). Multidimensional statistical analyses revealed that if TP concentration exceeds 15 µg l−1, the composition of periphyton assemblages begins to undergo both qualitative and quantitative changes. Typical species in unimpacted sites are cyanobacteria of the genera Aphanothece, Chroococcus, Stigonema, Scytonema, Komvophoron, and G. interspersa, diatoms M. smithii, A. coffeaeformis,

18 Experimental Assessment of Phosphorus

473

Fig. 18.11 Comparison of dry matter of floating mats and periphyton growing on Plexiglas slides during the period 1996–1998

and F. ulna, and all observed members of the green algae genus Cosmarium. These species disappear with increasing TP concentrations above 15 µg l−1. Only a few species responded positively to a very high increase in TP concentration: cyanobacterium Lyngbya sp. 10.5 and diatom R. gibba. It has also been found that with increasing TP the number of species is reduced by more than 60% and also the species diversity of the periphyton mats decreases with increasing TP concentrations. The qualitative composition of periphyton growing on Plexiglas is close to that in mats but quantitatively mats and Plexiglas are not similar. These differences have also been confirmed by biomass analyses and thus put into question the use of artificial substrates as indicators of natural community composition or growth potential. Natural mats contain higher amount of ash (higher degree of calcification), and therefore concentrations of P and N in the biomass are lower in periphyton growing on natural substrates. The use of algal community dynamics is thus a very powerful tool in assessing the early responses of Everglades slough communities to increased TP concentrations. The use of natural substrates to characterize the natural communities and shifts in species, biomass, or nutrients is recommended because artificial substrates do not give similar results. Seasonal shifts in algal species are significant and must be taken into account when comparing responses to nutrients. In general, the samples taken from TP concentration intervals <10 µg l−1 and 10–15 µg l−1 are very similar, and those two concentration intervals are very different from samples taken from TP concentrations 20–30 µg l−1 and >30 µg l−1. Samples from the TP concentration ranging from 15 to 20 µg l−1 represent a slightly different community as well, and their character is more similar to that of higher TP concentration ranges. Thus,

474

J. Kaštovský et al.

Fig. 18.12 Comparison of ash, P, and N content in the biomass of floating mats and periphyton growing on Plexiglas slides

18 Experimental Assessment of Phosphorus

475

Appendix 18.1: List of Abbreviations Used in the Figures Abbreviation

Algal species

AMPHCOFF Amphora coffeaeformis APCDE Aphanocapsa cf. delicatissima APTBAC Aphanothece bacillaris APTCO Aphanothece comasii APTMI Aphanothece minutissima APTVA Aphanothece variabilis CRCMI Chroococcus mipitanensis CRCTNX Chroococcus tenax CROMIN Chroococcus minutus CRPOL Chroococcus polyedriformis CYMSO Cymatopleura solea CYMAF Cymbella affinis FRAUL Fragilaria ulna GLTHIN Gloeothece interspersa GOMNOL Gomphonema olivaceum LLY12.1 Leptolyngbya sp. 1 LLY2 Leptolyngbya sp. 2 LLY2 × 2 Leptolyngbya sp. 2 × 2 LLY2 × 5 Leptolyngbya sp. 2 × 5 LEPTSP3 Leptolyngbya sp. 3 LLY4 objects Leptolyngbya sp. 4 LYN6.5 Lyngbya sp. 6.5 LYN10.5 Lyngbya sp. 10.5 LYN12 Lyngbya sp. 12 LYN15 Lyngbya sp. 15 LYN18 Lyngbya sp. 18 MASSMI Mastogloia smithii MOUGSP5 Mougeotia sp. 5 MOUG11 Mougeotia sp. 11 NITZLIN Nitzschia linearis OEDSP5 Oedogonium sp. 5 OEDSP7 Oedogonium sp. 17 OEDSP11 Oedogonium sp. 11 OEDSP20 Oedogonium sp. 20 PHSP5 × 6 Phormidium sp. 5 × 6 PHOSP4 Phormidium sp. 4 PHORSP8 Phormidium sp. 8 PSESP2.1 Pseudanabaena sp. 2 Scytonema sp. SCYTSP SCYSP10 Scytonema sp. 10 SCYTSP12 Scytonema sp. 12 SCYTSP22 Scytonema sp. 22 UGO13 Unidentified green The single numbers denote the width of filaments. The numbers with “×” denotes the width and length of the cell

natural algal communities found at background TP concentrations ≤10 µg l−1 are still similar to slightly elevated TP sites (10–15 µg l−1) after years of SRP additions, while significant changes in periphyton composition start to occur above 15 µg l−1 TP concentration (for a complete P threshold analysis, see Chap. 25).

19

Macroinvertebrate and Fish Responses to Experimental P Additions in Everglades Sloughs Ryan S. King and Curtis J. Richardson

19.1

Introduction

A variety of anthropogenic influences threaten the Everglades. One of the most publicized perturbations to the system has been excessive inputs of phosphorus (P). The Everglades is a P-limited ecosystem, and its biota are adapted for survival under highly oligotrophic conditions (e.g., Browder 1982; Steward and Ornes 1975a,b; Swift and Nicholas 1987; Davis 1991). Indeed, numerous experiments have demonstrated the sensitivity of Everglades biota to P enrichment. In one of the first fertilization experiments, Steward and Ornes (1983) showed that small additions of P resulted in significant increases in the productivity of sawgrass (Cladium jamaicense Crantz.) seedlings, the most abundant macrophyte in the Everglades (Loveless 1959). However, high levels of P actually inhibited sawgrass production (Steward and Ornes 1983). Ensuing field fertilization studies have shown that P enrichment can have profound effects on open-water slough communities, a unique and potentially critical habitat in the Everglades ecosystem (e.g., Loveless 1959). For example, Walker et al. (1989) demonstrated that continuous dosing of high concentrations of inorganic P in sloughs of Everglades National Park nearly eliminated the common submergent macrophyte Utricularia purpurea Walt. and attached floating mats of calcareous periphyton, but stimulated growth of several emergent macrophyte species. In the same study, Flora et al. (1988) and Hall and Rice (1990) found that periphyton accumulation on artificial substrates showed as much as a tenfold increase in biomass as a direct result of P. This biomass increase was concomitant with significant shifts in algal species composition and changes in tissue nutrient content (Hall and Rice 1990). Subsequently, other P-fertilization experiments have been conducted in slough habitats, producing similar responses for both periphyton (Vymazal et al. 1994; Craft et al. 1995; McCormick and O’Dell 1996) and macrophytes (Craft et al. 1995). Despite this clear connection between P enrichment and changes in productivity and species composition of primary producers in the Everglades, remarkably little has been done to assess how these effects manifest themselves at higher trophic levels. To date, only papers resulting from the P-dosing experiment that is presented in this chapter (King 2001; King and Richardson 2003; Qian et al. 2003) have examined responses of invertebrates or fish to P enrichment using experimental

477

478

R.S. King and C.J. Richardson

manipulations. Moreover, relatively few observational studies have been conducted on Everglades invertebrates or fish in relation to nutrient enrichment (Terczak 1980; Urban and Koebel 1992; Rader and Richardson 1994; Jordan 1996; Turner et al. 1999; King 2001; King and Richardson 2002). As part of the P-dosing experiment designed to refine our understanding of how specific concentrations of P affect multiple levels of the Everglades slough ecosystem, we set out to establish dose–response relationships for attributes of macroinvertebrate and fish assemblages to local-scale, long-term experimental P additions. We used flume-style mesocosms with multiple treatments to create experimental P gradients, thus providing a wide range of concentrations to meet with our primary objective. We used the subsidy–stress model (Odum et al. 1979) as a framework to establish several hypotheses regarding macroinvertebrate and fish assemblage responses to experimental P enrichment: 1. Macroinvertebrates and fish are resource limited. Relaxation of P-limitation will result in a positive response in primary production with small doses of P. Macroinvertebrate and fish abundance will mirror responses of primary producers, such that increased in production will increase macroinvertebrate and fish standing stocks. 2. Species turnover of primary producers will result in a stress to specialized macroinvertebrate species (e.g., grazers of specific algae). Opportunistic species will proliferate with P additions, while specialists will be disadvantaged by the loss of sensitive algal taxa, resulting in a subsidy–stress dose–response curve for species richness. The opportunistic small fish assemblage, which has few species, will not exhibit a detectable change in species richness. 3. Shifts in assemblage structure (e.g., increased dominance by one or a few species) will occur for both macroinvertebrates and fish at relatively low doses of P. 4. Succession of macroinvertebrate assemblages will be amplified with P due to greater secondary production and turnover, potentially leading to lesser temporal stability in composition. 5. The direct effect of P on periphyton production and nutrient content will be the most important dimension of organization for macroinvertebrate and fish assemblages in Everglades sloughs. Our goal in this chapter is to emphasize tests of these specific hypotheses. Here, we present a synthesis of the results of approximately 3 years of macroinvertebrate and 1 year of fish studies from the P-dosing experiment.

19.2 19.2.1

Methods Study Area and Experimental Design

The P-dosing experiment was established in the southern interior of WCA-2A, a large, contiguous Everglades landscape. Levees surround the perimeter of WCA-2A, and water-control structures pump water in and out of the fen at irregular intervals

19 Macroinvertebrate and Fish Responses

479

(SFWMD 1992). Concentrations of water-column soluble reactive phosphate (SRP) and unfiltered total phosphorus (UTP) in the vicinity of the study area are typically near 5 and 10 µg l−1, respectively (Vaithiyanathan and Richardson 1998). Sediment and soil TP typically range between 250 and 400 mg kg−1 (Vaithiyanathan and Richardson 1998). Other detailed characteristics of WCA-2A are presented by numerous authors in this book, including geology and hydrology (Chap. 7), soil properties (Chap. 3), water chemistry (Chap. 6), and vegetation (Chap. 4). Two dosing facilities (hereafter called “sites”) were constructed in adjacent sloughs (26°15´N, 80°23´W) in fall of 1992. Each site had five-walled mesocosms, or flumes, 2-m wide × 10-m long with walls approximately 90 cm in height above the slough substrate. Flumes were oriented N–S and separated by 1 m, where permanent boardwalks were built to allow investigators access for sampling. Flumes were not sealed at the south end but were obstructed to prevent large vertebrates from entering. Each flume was randomly assigned one of five SRP treatments: ~5 µg l−1 (control; 0.25 g m−2 year−1), ~22 µg l−1 (1.5 g m−2 year−1), ~39 µg l−1 (2.75 g m−2 year−1), ~57 µg l−1 (3.5 g m−2 year−1), and ~126 µg l−1 (8.2 g m−2 year−1). These concentrations correspond to mean doses and loads for the full 6 years of the study. An additional unwalled control area was established on the west side at both dosing sites to concurrently monitor the potential effects of placing walls around slough habitat. Originally, categorical labels (unwalled control (UC), walled control (WC), 30, 50, 75, and 150, respectively) were assigned to each treatment based on the anticipated SRPdosing concentration, but actual doses were lower. Thus, these names are used at times in the text for convenience but are not intended to reflect the precise dose applied to flumes. SRP was dosed from the northern end of flumes via large mixing tanks. Tanks contained natural slough water mixed with a regulated ratio of SRP concentrate. Flumes were dosed on a continuous schedule except during low- or high-water shutdowns or periodic maintenance to specific flumes. Maintenance shutdowns, particularly during 1997 and 1998, caused cumulative P loads to vary modestly among treatments, so actual loads were calculated for each flume to better estimate cumulative exposure to P. Dosing was applied from 30 November 1992 to 21 September 1998. Greater detail on the design and operation of the P-dosing experiment are presented in Chap. 14. Water depth at the sites varied seasonally and interannually. During the course of study, water depths ranged between near 0 cm and over 100 cm, but typically were 25–75 cm. Depths were generally greatest in late summer to early fall (wet season) and reached their lowest values during spring (dry season). However, in 1998, water levels were quite high (60–90 cm) throughout the dry season. Vegetation of the sloughs was characterized by abundant white water lily (Nymphaea odorata Aiton). Submergent plants were primarily bladderworts, Utricularia fibrosa Walt., Utricularia foliosa L., and U. purpurea Walt. Chara sp., a macroalga, also was locally and seasonally abundant. Spikerushes (Eleocharis cellulosa Torrey and Eleocharis elongata Chapm.) were the dominant emergent plants, with occasional stems of maidencane (Panicum hemitomon Schult.) also observed. A chalkish-colored mat of calcareous periphyton, composed primarily of

480

R.S. King and C.J. Richardson

cyanobacteria and diatoms, often covered much of the slough surface and typified this oligotrophic habitat (Swift and Nicholas 1987). Epiphytic periphyton was also locally abundant, particularly on stems of Eleocharis spp. Richardson et al. (Chap. 16) and Kasˇtovsky´ et al. (Chap. 18) provide a synthesis of vegetation and periphyton responses to P dosing, respectively.

19.2.2

Sampling: Water Chemistry, Periphyton, and Vegetation

Water-column SRP and UTP were measured biweekly throughout the duration of the study. Since uptake and diffusion dramatically affected concentrations down the length of the flumes, surface samples were collected at 0.5-, 1-, 2-, 3-, 4-, 6-, and 8-m stations corresponding to distances from the source of P inputs. Sediment TP (0–10 cm) was measured in 1995, 1997, and 1998 at each station. Other surfacewater measurements (UTN, NH4, NO3, major cations) also were collected but generally were not affected by P dosing. Dissolved oxygen (DO) was measured periodically using multiprobes deployed at 2-m stations in mid-water column for 1-week durations. However, minimum DO concentrations were not sufficiently altered by P dosing to elicit negative effects on macroinvertebrates or fish and will not be considered further here. Synthesis of the dosing study water chemistry results are presented in Chap. 15. Periphyton was collected from a variety of substrates during the study, but primarily from the floating mat and artificial substrates (Plexiglas slides). Because Plexiglas slides best estimated short-term periphyton responses to P and were similar to the substrates used to collect macroinvertebrates in this study, data from these samples were used in subsequent macroinvertebrate analyses. Plexiglas slides were deployed at 2-, 4-, and 6-m stations in flumes for a duration of 2 months. Ash-free dry mass (AFDM), % ash, and total C, N, and P were the primary periphyton analytes. Species composition was examined on some dates earlier in the study, but species data were not generated concurrently with macroinvertebrate or fish data. Periphyton results from the first 4 years of the study are presented in Chap. 18. Macrophyte stem counts and % cover as well as floating periphyton mat % cover were surveyed quarterly using 1-m2 quadrats. Quadrats were centered in the flumes, and macrophytes were surveyed across the entire length of the flumes at all stations. Richardson et al. (Chap. 16) describe the responses of vegetation to P additions.

19.2.3

Sampling: Macroinvertebrates

We initiated the macroinvertebrate component of the P-dosing study in 1996, 4 years after dosing had begun. We sampled macroinvertebrates in May and September 1996, January 1997, and February and September 1998. Sampling was intended to be maintained twice per year, one wet (September) and one dry (January–May)

19 Macroinvertebrate and Fish Responses

481

season event, but dry conditions during summer 1997 forced a shutdown of dosing and prevented a September 1997 collection. Macroinvertebrates were sampled from the dosing flumes using Hester–Dendy artificial substrates. Hester–Dendys (HDs) were used for several reasons. The state of Florida has classified the Everglades as a Class III waterbody and has mandated that water-quality assessments using macroinvertebrates in these areas will be conducted using a composite sample of a minimum of three HD artificial substrate samples (Florida Administrative Code 17-302.560[9]). Additionally, HDs are deployed for a fixed period of time to allow for colonization of the substrate, thus are passive samplers and minimize disturbance to the surrounding natural habitat (e.g., Cairns 1982). Active sampling devices (e.g., dip nets) are destructive to habitat and were inappropriate for use in this long-term experiment. Finally, HDs are standardized for surface area and help reduce the variability associated with sampling from natural substrates (Murkin et al. 1994), which were spatially heterogeneous on the relatively fine scale present in the dosing flumes. We deployed three HDs at each of the 2-, 4-, and 6-m stations from the source of P inputs in all flumes on each date (108 per date). Since three HDs are required to form a composite sample, each station could therefore be considered as an observation. This sampling design was chosen to allow for comparisons among P-dosing treatments as well as observational analyses using water-quality measurements at each station. HDs were deployed for 28 days. HDs were suspended from fixed bars across the width of the flumes and maintained at mid-depth in the water column. Depth of HDs was adjusted on a weekly basis, if necessary, to accommodate for water-level changes. We spaced samplers 0.5 m apart on each bar, centered within the flume. After 28 days, HDs were collected by sliding a 4-l heavy-duty plastic bag under the sampler and pulling the bag to the water surface. This technique alleviated potential dislodging of mobile invertebrates resting on the sampler during retrieval. However, during May 1996, three dosed flumes were nearly completely overgrown (90–100% cover throughout the water column) with Chara sp., a macroalga, which prohibited this collection technique. These HDs had to be untangled from and forced through the Chara mats, undoubtedly agitating and dislodging attached macroinvertebrates. Therefore, HDs collected from these flumes during this event did not meet QA/QC procedures and were subsequently excluded from the final data set. Once collected, samplers were put on ice and returned to the laboratory for processing, where material on the substrates was scraped and rinsed into a sieve, then fixed in 10% buffered formalin for storage. We defined macroinvertebrates as any invertebrate retained in the 0.5-mm sieve, which included Cladocera, Copepoda, and Ostracoda. All macroinvertebrates were sorted from associated periphyton and detritus under a stereomicroscope using 10× magnification. We identified individuals to the lowest possible taxonomic unit. Nomenclature followed Hobbs (1942), Keyser (1975), Sanderson (1982), Pluchino (1984), Thompson (1984), Berner and Pescador (1988), Daigle (1991), Dodson and Frey (1991), Williamson (1991), Daigle (1992), Epler (1995, 1996), Klemm (1995), Pescador et al. (1995), Courtney et al. (1996), and Milligan (1997). Taxa that could not be assigned a species name were identified to morphospecies, a level of taxonomy that differentiated among

482

R.S. King and C.J. Richardson

individuals that were likely distinct species but did not attach specific names (Oliver and Beattie 1996). Multiple specimens of each taxon were retained in a voucher collection and sent to expert taxonomists for independent verification of names. We estimated macroinvertebrate biomass for each HD station-composite after all names had been verified. Because shell mass of snails is generally not considered to constitute biomass, all Gastropoda were separated from other invertebrates. Gastropods were soaked in Bouin’s fixative, a technique used to dissolve the calcareous shells but not affect mass of the flesh (F.G. Thompson, personal communication). Macroinvertebrates were pooled together for each sample, oven-dried in preweighed aluminum pans for 48 h at 60°C, and placed in a desiccator before weighing. Pans were weighed on an analytical balance (± 0.0001 g) to estimate dry mass. Gastropoda were dried and weighed separately from other invertebrates. Many of the smaller invertebrates, particularly Chironomidae and Oligochaeta, had been previously mounted on slides for identification purposes, thus were not usable for direct weighing. These specimens were enumerated into size classes based on length and width, and dry mass subsequently was estimated using a biovolume technique (Smit et al. 1993).

19.2.4

Sampling: Fish

We initiated fish collection in October 1997, and additional collections were conducted in January and May 1998. As was the case with the macroinvertebrates, fish sampling methods had to be nondestructive to habitat in the flumes. This immediately ruled out active sampling devices, such as throw traps, that have been effectively used in Everglades marshes to sample fish (e.g., Loftus and Eklund 1994; Jordan 1996). Instead, we used small fish traps (43 × 25 × 25 cm, 5 cm funnel openings, 2 mm nylon mesh), which were passive samplers and relied on fish activity for collection rather than movement of the sampler itself (e.g., Hubert 1996). These traps functioned similarly to funnel traps and fyke nets commonly employed in wetlands to capture fish (e.g., Brazner 1997). Traps were deployed in pairs at the 2-, 4-, and 6-m distances within each flume (72 traps per date), suspended from bars across the width of the flumes. We aligned the top of each trap with the surface of the water to more effectively capture the primarily surface-oriented small fishes (e.g., mosquitofish, Gambusia holbrooki) found in Everglades slough habitats. We deployed traps for 24 h on each date. Initially, we set out to identify, count, and measure all individuals in the field and release them back into the flumes, but this was deemed almost logistically impossible and statistically limiting due to the relatively poor accuracy of mass estimates made on small fish in the field. Additionally, because flumes could be recolonized by fish from the surrounding sloughs through the opening at the south ends, we assumed that any removals would be offset by recolonization. Thus, we retained fish for laboratory examination. Contents of each trap were emptied into a container where fish were anesthetized,

19 Macroinvertebrate and Fish Responses

483

then fixed in 10% formalin. All individuals were identified, measured for total length (mm), and weighed on an analytical balance (± 0.0001 g). Taxonomic nomenclature followed American Fisheries Society (1990). We also considered the potential that predation by large fish may have an influence on the abundance and composition of small fish in the flumes. We attempted electrofishing as a supplementary technique to the fish traps (Reynolds 1996). We used a pulsed-DC backpack electrofishing unit (Smith-Root, Inc., Vancouver, WA, USA); however, it was not effective due to interference by the heavy vegetation and abundant galvanized pipe comprising the support structure of the dosing facility.

19.2.5

Data Analyses: Macroinvertebrates

The dose–response relationship between P and macroinvertebrates was examined using a suite of assemblage-level attributes. Attributes were selected from 1 or 4 categories (1) standing stocks, (2) taxonomic richness, (3) taxonomic structure, and (4) feeding ecology. Mean or total (total number of taxa) values from each dosing flume were regressed against P load to assess a potential dose–response relationship. Load rather than dosing concentration was used because of slight variation in [PO4] within individual mixing tanks over time and occasional maintenance shutdowns to flumes that caused the loads to vary somewhat among flumes of the same treatment. Therefore, load more accurately represented the actual P treatments and was subsequently used as a continuous predictor variable. Of note, sediment TP concentrations in the flumes during the course of the study spanned a very similar range to those observed along the P gradient in WCA-2A (250–1,800 mg kg−1 at dosing study; 300–1,800 mg kg−1 along P gradient); thus, the treatment with the highest P load was similar to the most enriched areas near canal inflow structures. Load units were standardized to total surface area of the flumes (g m−2 year−1). UC flumes could not be assigned a load value because they received no P additions and thus could not be used in dose–response regressions. These flumes served as qualitative reference points to assess potential wall effects on assemblage attributes. Simple linear and curvilinear relationships were explored between attributes and P load. Several P load durations (immediately prior to and including date of macroinvertebrate collection) were considered, ranging from 28 days (during deployment), 2, 3, 6 months, 1 year, and total (complete study duration). Six-month load was selected since it approximated the aquatic lifespan of most of the longlived taxa present at the dosing facility, and it consistently showed the strongest relationship to attribute values. Data transformations (log and arcsine) were applied when appropriate (Sokal and Rohlf 1995). Multivariate assemblage composition was summarized and related to environmental variables using nonmetric multidimensional scaling (nMDS), an ordination technique based on ranked dissimilarities among samples (Minchin 1987; Clarke 1993; Legendre and Legendre 1998). The objective in the use of nMDS was to

484

R.S. King and C.J. Richardson

recover a multivariate community pattern that could potentially be attributed to an environmental gradient, such as phosphorus concentration. Bray–Curtis dissimilarity was selected as the metric for the macroinvertebrate assemblage data, as it has been shown to be one of the most robust and ecologically interpretable measures for species abundance data (Bray and Curtis 1957; Faith et al. 1987; Clarke 1993; Legendre and Legendre 1998; Legendre and Anderson 1999). Taxon abundances were standardized using a log10(x + 1) transformation to provide greater weight to uncommon taxa (e.g., Faith and Norris 1989; Cao et al. 1998). In the first nMDS analysis, data were ordinated using individual flumes (including UCs) as observational units. This analysis was done separately for each date (n = 12 per date) as well as all dates combined (n = 60), with the former analysis designed to assess the overall effect of P loads on each date, while the latter was intended to track the trajectories of each flume over time and in relation to each other (successional vectors; Legendre and Legendre 1998). Ordinations were also conducted using data at 2-, 4-, or 6-m stations within each flume, again including UCs, for the purpose of relating observational environmental data collected at each station (n = 36 per date) to assemblage composition. Measured environmental variables included macrophyte cover/stem counts (by species), periphyton AFDM, % ash, TC, TN, and TP, water-column SRP and UTP (6-month average), and sediment TP (most recent annual collection). We related these variables to species composition using vector fitting in nMDS space (Faith and Norris 1989; Growns et al. 1992). To complement the nMDS ordinations and cleanly assess the direct and indirect effects of the environment on macroinvertebrate species composition, partial Mantel’s tests were applied in conjunction with path diagrams (Leduc et al. 1992; Sanderson et al. 1995). The Mantel’s test is a multivariate technique that uses distance matrices as variables and allows the investigator to partial out the effect of other variables to obtain an independent estimate of the relative contribution of each variable to macroinvertebrate species composition. Using a priori hypotheses regarding relationships among variables, a path diagram synthesizes the results of the partial Mantel’s tests by depicting the significant paths of relationships (Legendre and Legendre 1998). Phosphorus (SRP, UTP, sediment TP, and periphyton TP), macrophyte (species composition weighted by untransformed cover estimates), periphyton (AFDM and C:N ratio), and macroinvertebrate distance matrices were constructed for this analysis. Individual variables used in P and periphyton distance matrices were standardized to z-scores prior to matrix calculation using Euclidean distance (Legendre and Legendre 1998). Macrophyte and macroinvertebrate matrices used Bray–Curtis dissimilarity as the distance metric.

19.2.6

Data Analyses: Fish

The Everglades small fish assemblage has relatively few common species; thus, the number of attributes for use as response variables to P additions was more limited than for the macroinvertebrates. We calculated individual species abundance and

19 Macroinvertebrate and Fish Responses

485

biomass, total numerical abundance and biomass, and species richness within each flume as a unit. Only six species were abundant enough to make univariate statistical inferences. Consistent with the macroinvertebrate analysis, we used 6-month P load as a predictor in linear and curvilinear regression analyses to estimate potential dose–response relationships. Additionally, we compared the slope and the intercept of length–mass regression equations to assess whether average fish size or condition changed among P treatments. To eliminate potential flume-specific findings and increase sample size, we pooled individuals within each of the six dosing concentration “categories” (i.e., UC, WC, 30, 50, 75, and 150; see Chaps. 14 and 15). Only G. holbrooki had sufficient numbers to justify treatment comparisons (minimum n > 50 per treatment, total n = 693) (Sokal and Rohlf 1995; Anderson and Neumann 1996). We ordinated multivariate species composition using nMDS separately for each date to assess potential gradients in the assemblages that could be attributed to P load or other environmental variables (e.g., vegetation, periphyton). We also examined trajectories of assemblages in each flume over time using successional vectors in nMDS space, as in the macroinvertebrate analysis.

19.3 19.3.1

Results Dose–Response Relationships: Macroinvertebrates

Over 36,000 individuals represented by 123 taxa were collected and identified during the study period (Table 19.1). Amphipoda, Chironomidae (Diptera), Gastropoda, microcrustacea (Cladocera, Copepoda, and Ostracoda), and Oligochaeta were the most numerically dominant taxonomic groups (Table 19.1). Amphipoda, Decapoda, Gastropoda, and Odonata tended to have the largest individuals and thus made the greatest contributions to assemblage biomass. Chironomidae (at least 33 species) had the most species among all higher groups. Many macroinvertebrate assemblage attributes changed significantly as a function of P dosing. The two primary measures of standing stocks, density and biomass, showed a positive response to increasing P load on all collection dates, with significant relationships (p ≤ 0.05) on four of five dates for density and three of five dates for biomass (Fig. 19.1). This trend was particularly apparent during the winter (dry season) collections, January 1997 and February 1998. Water depths were markedly different between these two dates, with February 1998 having much greater depth than January 1997, yet results were remarkably similar (Fig. 19.1). Taxon richness and taxon density were also affected by P. These diversity metrics increased consistently among dates with increased P additions, following a similar pattern to that of total abundance and biomass (Fig. 19.2). Generally, the intermediate P loads resulted in a marked increase in the number of taxa relative to the controls, with diminishing marginal increases with additional P loading.

486

R.S. King and C.J. Richardson

Table 19.1 Dominant invertebrate taxa collected at the P-dosing experiment Taxonomic group No. collected Rank Dominant taxon/taxa Amphipoda Bryozoa Cladocera Cnidaria Coleoptera Copepoda Decapoda

4,304 41 687 6 16 2,426 69

5 15 7 18 16 6 14

Diptera (Chironomidae)

7,239

2

228 454

9 8

5,812

3

Hemiptera Hirudinea Hydracarina Odonata

9 205 166 128

17 10 11 12

Oligochaeta

9,649

1

Ostracoda

5,002

4

4 106

19 13

Diptera (Other) Ephemeroptera Gastropoda

Porifera Trichoptera

Hyalella azteca (Saussure) Plumatella cf. repens (L.) Ceriodaphnia cf. reticulata (Jurine) Hydra sp. Enochrus ochraceus (Melsheimer) Cyclopoida Palaemonetes paludosus (Gibbes) Procambarus cf. alleni (Faxon) Chironomus sp. 2 Chironomus stigmaterus (Say) Cladotanytarsus sp. Endochironomus nigricans (Johannsen) Larsia decolorata (Malloch) Parakiefferiella sp. C Epler Paratanytarsus sp. B Epler Polypedilum halterale group Polypedilum sp. A Epler Tanytarsus sp. G Epler Tanytarsus sp. R Epler Dasyhelea sp. Caenis diminuta Walker Callibaetis floridanus Banks Aphaostracon pachynotus Thompson Physella cubensis (Pfieffer) Planorbella duryi (Weatherby) Planorbella scalaris (Jay) Belostoma sp. Helobdella triserialis (Blanchard) Arrenurus sp. 1 Celithemis eponina (Drury) Ischnura hastata Say Dero digitata (complex) Pristina leidyi Smith Cypridopsis okeechobei Furtos Cytheridella alosa (Tressler) Spongilla cf. cenota (Penney and Racek) Oxyethira sp.

Dose–response relationships for taxon richness and density were significant in January 1997, and February and September 1998. Richness of other major groups (e.g., Chironomidae) generally followed this same pattern, thus no particular group of macroinvertebrates was primarily responsible for the overall increase in number of taxa. Higher taxonomic groups were not negatively affected by P dosing in terms of absolute abundance, but some groups demonstrated greater responses than others. In particular, the primary-consuming Oligochaeta and microcrustacea were most responsive to P. These two groups were in much greater numbers in even the lowest P treatment than in the controls (Fig. 19.3). In contrast,

19 Macroinvertebrate and Fish Responses

487

Fig. 19.1 Dose–response relationships between P load and invertebrate biomass and density for all five collection dates at the P-dosing study. Response variables are means (± 1 SE) per mesocosm

488

R.S. King and C.J. Richardson

Fig. 19.2 Dose–response relationship between P load and taxon richness and taxon density during January 1997. Taxon density values are means (± 1 SE) per treatment, while taxon richness is expressed as the total number of taxa per treatment

Amphipoda (represented exclusively by the abundant Hyalella azteca) revealed potential seasonality in its response to P. Its numbers increased significantly (p ≤ 0.05) and log-linearly with P during January 1997 and February 1998, but did not respond significantly to P dosing during the May and September collection dates. Due to large increases in absolute numbers of some groups, proportional abundance measures subsequently indicated shifts in assemblage structure with P dosing. Particularly, % Gastropoda showed a negative linear or log-linear response to P additions on most dates, suggesting that enrichment had not benefited these periphyton grazers in proportion to the rest of the assemblage (Fig. 19.4). In spite of this, Gastropoda numerical abundance consistently exhibited a positive response to P additions.

19 Macroinvertebrate and Fish Responses

489

Fig. 19.3 Dose–response relationship between P load and densities of microcrustacea and Oligochaeta during January 1997 and February 1998. Response variables are means (± 1 SE) per mesocosm

Of the trophic levels, primary consumers dominated the macroinvertebrate assemblage, generally representing > 90% of the total numbers collected. These periphyton, macrophyte, and detrital consumers showed positive responses to P load on all dates. Secondary and tertiary consumers, or predators, were less responsive to P additions. Their abundance either did not increase above control treatments or showed very subtle increases with increasing P load. However, their proportional abundance was negatively affected on all dates, and significantly during January 1997 and February 1998 (Fig. 19.5). On these dates, % Predators declined in a steep log-linear fashion.

490

R.S. King and C.J. Richardson

Fig. 19.4 Dose–response relationship between P load and % Gastropoda during January 1997 and February 1998. % Gastropoda values are means (± 1 SE) per mesocosm

Rate of change in composition over time was relatively similar among treatments. Examination of successional trajectories suggested that mean vector lengths, a direct measure of the magnitude of change over time, did not vary notably as a function of P load (p > 0.05). However, projection of these vectors in nMDS space revealed that the relative trajectory among dates was affected by P (Fig. 19.6). Assemblages in P-dosed flumes during January 1997 and February 1998 diverged from the paths of the controls, thus reflected increased dissimilarity from control assemblages. This trend was particularly apparent at the greatest loads of P. Moreover, dosed flumes were clearly separated from controls on all dates, complementing the findings from univariate analyses that dosing had not only affected abundance and biomass, but had also manifested itself as a gradient in taxonomic composition. This gradient

19 Macroinvertebrate and Fish Responses

491

Fig. 19.5 Dose–response relationship between P load and % Predators during January 1997 and February 1998. % Predator values are means (± 1 SE) per mesocosm

in composition was directly related to the magnitude of P loading, as highest-dosed flumes were most dissimilar from controls (Fig. 19.6).

19.3.2

Composition–Environment Relationships: Macroinvertebrates

Several observational environmental variables were strongly related to macroinvertebrate assemblage composition. Vector fitting of individual variables into nMDS space showed that all four measures of P (plexislide periphyton TP, water-column

492

R.S. King and C.J. Richardson

Fig. 19.6 Nonmetric multidimensional scaling (nMDS) ordination of invertebrate successional vectors during 1996–1998 at the P-dosing study. Symbols indicate the long-term mean P load per treatment, while numbers indicate the date of collection. For clarity, replicate mesocosms are shown only for each of three of the six P treatments

SRP and UTP, and sediment TP) were significantly (p ≤ 0.05) correlated to the gradient in assemblage composition defined by nMDS Axis 1 (Fig. 19.7). Vectors for these variables were directed away from control sampling stations toward the highest-dosed flumes. Sediment TP, although measured only annually during the study period, had the strongest relationship with composition among the P metrics (r = 0.43–0.77, p ≤ 0.0001 on all dates except September 1996 (p = 0.0260) ). This trend was consistent among all five dates. Vegetation and periphyton variables also were significantly correlated to composition. Three macrophyte species, U. purpurea, U. foliosa, and E. elongata, had vectors directed toward control stations (Fig. 19.7). Thus, the distribution of these macrophytes was negatively related to increasing concentrations of P, a consistent finding throughout the duration of the study. Periphyton AFDM accumulation on plexislides, however, followed a similar trajectory as the P vectors (Fig. 19.7). It was directed toward high-P stations and suggested that it played a role in the observed gradient in macroinvertebrate composition. Periphyton total carbon-tonitrogen (C:N) ratio values decreased as AFDM values increased, typically ranging from 17–22 at control locations to as low as 9–10 at stations of high-P exposure. Consequently, the C:N vector also was related to composition and was directed toward control assemblages. Partial Mantel’s tests, in conjunction with path analysis, were useful in assessing the relative indirect and direct contribution of the environmental variables used in the nMDS analyses on multivariate composition. Although the magnitude of relationships varied slightly among dates, the paths of relationships did not. Periphyton and vegetation both were strongly influenced by P, but the relationship was stronger for

19 Macroinvertebrate and Fish Responses

493

Fig. 19.7 Nonmetric multidimensional scaling (nMDS) ordination of macroinvertebrate assemblage composition at each sampling station from all mesocosms (including unwalled controls) during January 1997. Numbers adjacent to symbols indicate station distance (2, 4, or 6 m) down the flumes. Environmental vectors indicate the direction and magnitude of significant (p ≤ 0.05) correlations to assemblage composition

periphyton (partial Mantel r = 0.181–0.512, all p ≤ 0.05 for periphyton; simple Mantel r = 0.020–0.2725, p ≤ 0.05 only during May 1996 and September 1998 for macrophytes) (Fig. 19.8). Periphyton also was influenced weakly by vegetation structure, even after accounting for the effects of P (partial Mantel r = 0.088–0.195, p ≤ 0.05 only in February 1998). Macroinvertebrate composition appeared to be most influenced by the direct effect of periphyton, in terms of both AFDM and C: N ratio (partial Mantel r = 0.147–0.314, all p ≤ 0.05). Vegetation also accounted for variation in assemblage composition that could not be accounted for by either P or periphyton; this direct effect was not as strong as that for periphyton (partial Mantel r = 0.107–0.228, p ≤ 0.05 during January 1997, and February and September 1998). Surprisingly, a weak, but significant indirect effect of P on composition also was detected (partial Mantel r = 0.0628–0.4910, p ≤ 0.05 in May and September 1996, and January 1997). This residual suggested that the selected measures of periphyton and vegetation could not account for all P-related variation in composition, thus indicating the potential importance of other P-mediated variables in structuring the macroinvertebrate assemblages at the dosing facility.

494

R.S. King and C.J. Richardson

Fig. 19.8 Path diagram synthesizing relationships among P, periphyton, vegetation, and invertebrates based on results of partial Mantel’s tests. Relative strength of relationships is indicated by the magnitude of arrows

19.3.3

Fish Responses to P Additions

We collected a total of 2,785 individuals and identified 15 species, but only six were common (Table 19.2). Rank order of species abundance was consistent with data in the literature on fish assemblages in Everglades sloughs (Table 19.2), suggesting our fish traps were not excessively selective toward particular species. However, in contrast to the macroinvertebrates, our data did not reveal a pattern in abundance, biomass, or number of species that could be attributed to P dosing. Abundance and biomass were highly flume specific and very noisy among treatments. Regressions for individual species abundances and biomass also yielded nonsignificant relationships. Assessment of mass response of the most abundant fish in our study, G. holbrooki, among P-treatment categories also showed no detectable effect. Despite having high statistical power and precision in length and mass estimates, comparison of both slopes and intercepts of length–mass equations among treatments showed that P had not significantly affected condition of this species. Species composition was not affected by P. Rather, temporal variation sorted assemblages among the three dates. October was dominated by Gambusia, Poecilia latipinna, Heterandria formosa, and Fundulus chrysotus, all small fishes. By January, these three species were less common. January brought about a noted increase

19 Macroinvertebrate and Fish Responses

495

Table 19.2 Common and scientific names of fish collected at the P-dosing experiment, shown by rank order of abundance Common name No. collected Scientific name Eastern mosquitofish Sailfin molly Bluefin killifish Least killifish Golden topminnow Flagfish Spotted sunfish Largemouth bass Redear Yellow bullhead Bluegill Lake chubsucker Bluespotted sunfish Seminole killifish Redfin pickerel

1,164 549 226 106 98 98 22 7 5 3 2 2 1 1 1

Gambusia holbrooki Girard Poecilia latipinna (Lesueur) Lucania goodei Jordan Heterandria formosa Agassiz Fundulus chrysotus (Günther) Jordanella floridae Good and Bean Lepomis punctatus (Valenciennes) Micropterus salmoides (Lacépède) Lepomis microlophus (Günther) Ameiurus natalis (Lesueur) Lepomis macrochirus Rafinesque Erimyzon sucetta (Lacépède) Enneacanthus gloriosus (Holbrook) Fundulus seminolis Girard Esox americanus Gmelin

in the abundance of 1+ and adult age-class sunfish, particularly Lepomis punctatus. While primarily invertivores, adult Lepomis were observed actively ambushing individuals of the small fish assemblage at and near the water surface. This species also was captured in most flumes, and so it is unlikely that predation was treatment or flume specific. In May, young-of-the-year (YOY) Micropterus salmoides and Lepomis spp. were common. Additionally in May, greater numbers of Lucania goodei and Jordanella floridae contributed to differences in assemblage structure over time. None of the successional vectors from nMDS indicated that P had a temporal influence on the fish assemblage. Analysis of other environmental variables in relation to fish assemblages indicated that microhabitat structure might be more important than P-status, at least on a local scale. A weak relationship was found between total abundance and biomass of small fishes and the combined abundance of two macrophyte species, E. elongata and E. cellulosa (spikerush). Flumes that had the greatest abundance of fish generally had high stem densities of these plants.

19.4 19.4.1

Discussion Standing Stocks and Trophic Levels

The importance of resource limitation to biotic communities is a source of much contention among ecologists, with many asserting that competition or other constraints are the chief determinants regulating populations (reviewed by Cohen et al. 1990; Pimm et al. 1991), while others suggest that its significance is dependent upon the trophic level an organism belongs to (e.g., Abrams 1993) or the disturbance

496

R.S. King and C.J. Richardson

frequency of the environment (Connell 1975; Schoener 1982). In the dynamic, nonequilibrium (Reice 1994) environment of the Everglades, we hypothesized that interspecific competition and predation would not play a strong role in regulating abundances of macroinvertebrates and that this would allow standing stocks to benefit from nutrient additions. An alternative hypothesis was that productivity would accumulate at the top of the food web due to top-down control of macroinvertebrate standing stocks by invertivorous fish (e.g., Hairston et al. 1960; Oksanen et al. 1981). The log-linear response of macroinvertebrate abundance and biomass to increasing levels of P enrichment, particularly primary consumers, supported the former hypothesis. This finding is also consistent with several nutrient addition studies conducted in other aquatic systems. For example, Hart and Robinson (1990) demonstrated that resources limited stream insects, as P additions to in situ flumes generated significantly greater densities, biomass, and developmental rates of grazing caddisflies in an oligotrophic stream. Similarly, Hershey et al. (1988) and Peterson et al. (1993) showed that whole-stream N and P enrichment of the Kuparuk River in arctic Alaska dramatically altered food webs, consistently increasing secondary production of grazing insects and generally increasing standing stocks. Other mesocosm studies have reached the same conclusion – nutrient and resource limitation can be significant determinants of secondary production and standing stocks in many aquatic systems (e.g., Mundie et al. 1991; Perrin and Richardson 1997). Experimental evidence of resource limitation is less clear in wetland habitats, particularly outside the Everglades. Hann and Goldsborough (1997) used both press and pulse treatments of N and P to small enclosures in the Delta Marsh to conclude that microcrustaceans responded variably over time to enrichment, but responses were tightly coupled to both primary production and taxonomic succession of algae. Murkin et al. (1994) found little to no effect of added nutrients to invertebrates in an enclosure study, but they conceded that nutrient concentrations used were too low to elicit a response in primary productivity. Two other studies conducted in Canadian wetlands, however, demonstrated highly positive responses of invertebrates to fertilization, yielding greater secondary production and standing stocks (Campeau et al. 1994; Gabor et al. 1994). These somewhat equivocal wetland results may be partly due to the inconsistent scales and durations among their studies, most of which were conducted for no more than a few months, as well as the degree of nutrient limitation in their study areas. This study benefited from being long term and continuous, with dosing conducted for over 6 years and macroinvertebrate sampling beginning after nearly 4 years of enrichment already had occurred. We also had replicate sites, each with five P treatments, thus allowed estimates of dose–response relationships rather than just comparing “enriched” vs. “control” treatments. Subsequently, these results provide compelling evidence that the macroinvertebrate community of Everglades sloughs is resource limited, and this limitation is relaxed through the addition of P. Moreover, all levels of P enrichment resulted in a subsidy effect rather than a stress to overall assemblage standing stocks (Odum et al. 1979). Our results conflict with the observational surveys of Terczak (1980), Urban and Koebel (1992), and Turner et al. (1999), all of whom conducted studies in P-enriched

19 Macroinvertebrate and Fish Responses

497

and -unenriched locations in the Water Conservation Areas (WCAs) of the northern Everglades. Terczak (1980) used several samplers to compare macroinvertebrate communities at two enriched and two unenriched sites in the WCAs of the northern Everglades, finding markedly lower abundance at one of the enriched sites while little change at the other. Urban and Koebel (1992) used litter bags in dense stands of vegetation to conclude that samples at a eutrophic site had greater abundance of Oligochaeta and Chironomidae but depressed abundance of other taxa relative to an oligotrophic site in WCA-2A. Alternatively, Turner et al. (1999) found no effect of enrichment on large-invertebrate biomass, although they used throw traps that primarily collected decapods. Aside from differences in collection techniques among these studies, we suspect that differences in vegetation were primarily responsible for their results. The highly enriched areas near canal inflows in WCA-2A are dominated by cattail (Typha domingensis Pers.), while interior WCA-2A is a mosaic of sawgrass stands laced with open-water sloughs (King et al. 2004). Periphyton production and biomass have been shown to be much greater in unenriched sloughs than in stands of cattail in enriched areas (McCormick et al. 1998), mostly from shading caused by high stem densities (Grimshaw et al. 1997). Consequently, dissolved oxygen levels are also much lower in the eutrophic cattail stands sampled by these researchers due to a shift from autotrophic to heterotrophic aquatic production (McCormick et al. 1997). Thus, it is not surprising that authors of these previous studies concluded that P enrichment either had no effect or was detrimental to invertebrate standing stocks in the Everglades. Here, primary production has shifted heavily to emergent macrophytes, thus P and other abiotic variables contributing to the spread of cattail likely acts as a stress rather than a subsidy to the invertebrate community in this eutrophic habitat. Accordingly, King (2001) demonstrated a clear subsidy–stress (sensu Odum et al. 1979) relationship between P and invertebrate biomass along the 10-km P gradient in WCA-2A when area weighted by the coarse-scale vegetation pattern of the landscape. Here, intermediate levels of enrichment with at least small sloughs or patches of open water yielded the greatest biomass. In addition to results reported by King (2001), results from Rader and Richardson (1994) support the notion that vegetation was an important factor governing the findings of the previous studies. Sampling with a sweep net and core sampler in patches of open water rather than dense stands of cattail along a 10-km P gradient in WCA-2A, they found the highest densities of macroinvertebrates at sites with the greatest enrichment of P. This trend mirrored the dose–response curve of macroinvertebrate standing stocks to our experimental P gradient. However, contrary to this study, it is important to note that they did not estimate biomass. Small fish did not respond to P additions as expected. No detectable increase of either standing stocks or mean condition was documented in the fish assemblage. This is somewhat surprising given that macroinvertebrate standing stocks increased with P, but several plausible explanations exist for this apparent disparity. Emigration and immigration to and from surrounding slough habitat may have masked P effects since the flumes had small openings at their southern ends. Predation also may have played a role. While highly speculative since we were

498

R.S. King and C.J. Richardson

unable to accurately estimate presence and abundance of piscivorous fish in the flumes, the small fish assemblage may have been affected by a few large predators, especially on this highly local scale. Consequently, biomass of piscivores, such as largemouth bass (M. salmoides), may have benefited from nutrient additions to the flumes. An interaction between predation and habitat also may have influenced fish abundance. The only statistically detectable effect was a weak correlation between stem densities of Eleocharis spp. and total abundance of small fish. High densities of this emergent macrophyte may have provided refuge from predators in the flumes, a factor that may have been more important than nutrient status. This is partially supported by the work of Jordan (1996) in WCA-1, where he illustrated the importance of high habitat complexity to most of the small fish assemblage of the Everglades. However, Jordan (1996) as well as Turner et al. (1999) found that small fish densities were as much as five times greater in nutrient-rich cattail stands than in any habitat in the oligotrophic interior fen. Turner et al. (1999) hypothesized that increases in productivity cascaded up through the food chain and accumulated as high standing crops of small fishes and consequently, as predicted by food web theory (e.g., Oksanen et al. 1981), resulting in no appreciable increase in invertebrate standing stocks. Interestingly, cattail stands in WCA-2A are unlikely to have few, if any, large predator fish due to the very high stem densities and generally lesser water depth than peripheral canals and airboat trails. Thus, contrary to our slough sites, the small fish assemblage in most large, eutrophic stands of cattail in the Everglades may receive little pressure from predators. This could partially explain the discrepancy between our experimental results and these previous observational studies. Clearly, more experimental data are needed to make inferences regarding mechanisms regulating production, standing stocks, and species composition of the small fish community in the Everglades.

19.4.2

Assemblage Richness and Composition

A subsidy effect also was clearly evident for macroinvertebrate taxonomic richness and density on some dates; however, this dose–response relationship did not completely support our hypothesis. We expected to see a unimodal productivity–diversity relationship (e.g., Tilman 1982; Abrams 1995), largely based on reported sensitivity of many algal taxa, particularly diatoms, to high concentrations of phosphorus (e.g., Swift and Nicholas 1987; Stevenson and Richardson 1995; McCormick and O’Dell 1996; McCormick et al. 1996; Kasˇtovsky´ et al., see Chap. 18). We hypothesized that grazing specialists would be disadvantaged by the reduction or complete exclusion of potentially preferred diatom or other P-sensitive algal taxa (Lamberti and Moore 1984). On the contrary, few of common macroinvertebrate taxa appeared disadvantaged by P additions (although there was tremendous species turnover along the 10-km P gradient in WCA-2A where vegetation changed markedly in response to high levels of P enrichment; King 2001; King and Richardson

19 Macroinvertebrate and Fish Responses

499

2002). This may suggest that few Everglades invertebrate species are truly “specialists.” Indeed, virtually all species documented from the Everglades are well adapted for harsh, often temporary conditions (Gunderson and Loftus 1993), hence have limits of tolerance (sensu Shelford 1913) that far exceed challenges presented by changes in food resources. Additionally, the south Florida landscape was formed very recently (~5,000 YBP) and has very few endemic species. Thus, Everglades macrofauna may best be characterized as an assemblage of rapidly dispersing, opportunistic species rather than a coevolved community uniquely adapted to the Everglades (Jordan 1996). This is not to say that Everglades macroinvertebrates are not sensitive to their environment; the sharp increase in the number of taxa at the lowest P loads above background clearly indicates to the contrary. Nevertheless, this study highly suggests that local-scale P additions to oligotrophic Everglades sloughs will likely increase species richness of macroinvertebrates. Rader and Richardson (1994) also indicated that species richness increased with increasing levels of P enrichment based on observational data they collected in WCA-2A. They found that most taxa that were common to oligotrophic sloughs became more abundant in intermediately and highly enriched locations, while many new taxa were added to the community. These findings parallel these experimental results. However, other authors have suggested that richness has declined due to P enrichment in the Everglades (Terczak 1980; Urban and Koebel 1992). Their work primarily was conducted in heavily vegetated habitats rather than the open-water habitats sampled by Rader and Richardson (1994). Although results were not entirely conclusive, King (2001) indicated that taxon richness changed modestly along the P gradient on a landscape scale, reaching greatest richness in the intermediately enriched transition zone. Because slough-like patches of open water are not common in enriched areas of the Everglades, this rare habitat may have represented an uncharacteristic refugium amongst vast stands of cattail and other invasive macrophyte vegetation in Rader and Richardson’s study. Increases in species richness with P dosing also contributed to increases in taxonomic dissimilarity (β-diversity) relative to control treatments. Shifts in composition closely corresponded to the magnitude of P loads. Dramatic increases in the abundance of ostracods, copepods, cladocerans, oligochaetes, and many chironomids overwhelmed more modest increases of other taxa, such as gastropods, thus changed the proportional structure of taxonomic composition. The apparent decrease in % Gastropoda was of particular interest since these taxa are primarily grazers of algae. However, when calculated as a proportion of total biomass, this negative dose–response relationship was not evident. Because of the relatively small individual mass of most taxa collected on the samplers, a few large specimens (particularly Planorbella duryi and Planorbella scalaris) contributed to noise and subsequently occluded a pattern. While relative biomass is probably a more meaningful measure of structural change in the assemblage, gastropods are a numerically dominant group in Everglades invertebrate assemblages and their relative abundance may be a good indicator of ecological integrity of oligotrophic sloughs. While macroinvertebrate species composition diverged from controls with P additions, successional trajectories generally did not. Small fish trajectories also

500

R.S. King and C.J. Richardson

were not affected by P additions. We hypothesized that P-induced increases in secondary production would result in more rapid emergence or turnover of some species, thus contributing to greater temporal variation in composition. Successional vectors of macroinvertebrates in P-dosed flumes did diverge somewhat from the paths of the controls during January 1997 and February 1998, but the mean rate of change was similar in all flumes. Hann and Goldsborough (1997) reported comparable findings, as the magnitude of temporal variation in cladoceran assemblages in the Delta Marsh differed little among nutrient treated and untreated mesocosms. The highly dynamic abiotic environment of the Everglades, and most wetlands for that matter, is probably the most important dimension governing the direction and magnitude of community succession (e.g., Batzer and Wissinger 1996). Seasonal variation in hydrology, water temperature, dissolved oxygen, and structure of primary producers likely were important forcing factors organizing the Everglades assemblages over time.

19.4.3

Composition–Environment Relationships

A multitude of variables measured within the flumes were correlated to macroinvertebrate assemblage composition. All observed measures of P (water-column SRP and UTP, periphyton TP, and sediment TP) were highly related to a gradient in assemblage composition. The strongest vector among all P measurements was for sediment TP, and this was consistent throughout the duration of the study. This seemed puzzling since HD samplers were suspended within the water column, not the sediment, thus P in the water should have had the greatest influence on biota attached to the samplers. However, SRP was heavily influenced by biological uptake and was often indistinguishable from the controls at low-P loads, especially during midday collections when photosynthetic activity was at its peak. Conversely, water-column UTP had relatively linear relationship with dose up to the 75 treatment but actually declined slightly in the 150 treatment. Sediment TP reached much higher values in the 150 treatments than any other, thus maintained a linear relationship with dosing concentrations from the controls to the highest treatment. Pan et al. (2000) also found a stronger relationship between algal species composition and sediment TP than with UTP in these same experimental flumes and along the P gradient in WCA-2A. Because it is unfiltered and includes particulates, UTP may often contain large amounts of living or dead plant tissue, thus may have been influenced by the coverage of periphyton mat on the water surface. Stevenson and Richardson (1995) and Richardson et al. (2000) both reported significant reductions in floating periphyton cover at the highest dosage of P; perhaps “cleaner” samples were collected in the absence of high mat coverage. Regardless of the mechanism, sediment TP appeared to be a better indicator of actual P loads than did UTP, at least at the highest load. Macrophyte species composition covaried with P doses and also may have played a role in structuring the macroinvertebrate assemblage. Macrophytes had a

19 Macroinvertebrate and Fish Responses

501

significant “direct” effect on macroinvertebrates, as indicated by partial Mantel’s tests. Two macrophytes, U. purpurea and U. foliosa, exhibited high sensitivity to P and thus were highly correlated to a gradient in assemblage composition. These bladderworts are largely carnivorous, relying on microorganisms as their primary source of nutrients. Coincident with the decline of Utricularia was a sharp increase in the abundance of microcrustacea, particularly the ostracods Cytheridella alosa and Cypridopsis okeechobei, and cyclopoid copepods. While the increase in microcrustacea was probably largely due to greater quality and productivity of grazing resources, loss of Utricularia also may have contributed. Microcrustacea often were observed within U. purpurea bladders when they were incidentally retrieved with HD samplers. There is an increasing body of evidence that periphyton forms the base of the food web in wetland habitats (e.g., Neill and Cornwell 1992; Keough et al. 1996; D.A. Wrubleski and N.E. Detenbeck, unpublished manuscript), a contradiction to the prevailing view that macrophyte detritus is most important to secondary production (e.g., Murkin 1989; Batzer and Wissinger 1996). This study indirectly supports this theory. Consistent with our hypothesis, periphyton exhibited a strong direct effect on macroinvertebrate assemblage composition. The highly significant partial Mantel’s coefficients for periphyton, based on biomass and nutrient content on plexislides, suggest that P-induced changes to periphyton had an influence on macroinvertebrate assemblages in the flumes. One of the most notable changes to the periphyton was the substantial decrease in molar total C:N ratios with increasing P concentrations. Ratios often exceeded 20:1 on plexislides collected from control flumes but were as low as 9:1 at high-P loads. High C:N ratios in controls may reflect poor nutritional value, as ratios > 17:1 (organic C:N) are considered unpalatable and inefficient to algivores (e.g., Russell-Hunter 1970; Jones et al. 1998). Although we did not correct for inorganic carbon, % ash on plexislides did not consistently vary with P, suggesting that relative differences in C:N among P treatments would be similar if expressed in terms of organic or total C. Browder (1982), however, suggested that periphyton represented a good source of protein for primary consumers in the Everglades. After correcting for inorganic carbon, she found that periphyton from oligotrophic areas in Everglades National Park had C:N ratios of 4.84–8.0 (but C:N ratios based on total C ranged from 19 to 30). It seems unlikely that periphyton from unenriched locations is an efficient source of energy because the Everglades is characterized as having unusually high standing stocks of periphyton, yet unusually low invertebrate and fish biomass (Goldsborough and Robinson 1996; Turner et al. 1999). What makes this even more mysterious is that periphyton grows very slowly in oligotrophic sloughs (e.g., Swift and Nicholas 1987; Vymazal and Richardson 1995; Browder et al. 1994; see Chap. 18) thus must accumulate over very long time periods. Such high biomass accumulation despite low productivity suggests that it receives relatively little grazing pressure from primary consumers due to one or a combination of factors (Murdoch 1966). These may include low nutritional value, unpalatable species composition (e.g., Jones et al. 1998), chemical deterrents (Lodge et al. 1998), or unavailability due to calcification and structure of the periphyton matrix (e.g.,

502

R.S. King and C.J. Richardson

Turner et al. 1999). Certainly, more research is needed on the interactions among nutrient additions, periphyton, and grazers in the Everglades. However, these data suggest that C:N ratios may be a good indicator of periphyton quality as food to primary consumers, and that P additions may increase the nutritional value of periphyton to macroinvertebrates.

19.5

Conclusions and Lessons for Restoration

Macroinvertebrates are resource limited in Everglades sloughs. Long-term P additions significantly increased macroinvertebrate standing stocks and species richness, but also altered the relative taxonomic and trophic structure of the assemblage. Dramatic responses often occurred at relatively low doses of P but diminished in magnitude with increasingly greater P loads, thus resulting in log-linear dose– response relationships for most attributes of the macroinvertebrate assemblage. Fish, however, did not respond to P additions, possibly due to predation or to localscale heterogeneity in habitat among the P-dosing experiment flumes. P-induced changes to both macrophytes and periphyton influenced macroinvertebrate species composition; however, periphyton productivity and nutrient content had the greatest direct effect.

B. Fertilizer Experiments

20

Plant Community Response to Long-Term N and P Fertilization Jan Vymazal, Christopher B. Craft, and Curtis J. Richardson

20.1

Introduction

The Everglades ecosystem is believed to have evolved under a low-nutrient environment, especially low phosphorus. Sawgrass, the dominant emergent macrophyte species in the Everglades, has low phosphorus requirements (Steward and Ornes 1975a,b). Previous studies concerning the effect of fertilizer additions, especially N and P, on sawgrass communities have been inconclusive. Steward and Ornes (1975b) observed that additions of P at rates of 1.1, 5.6, and 11.2 g m−2 had no effect on growth of sawgrass. Sutter (1992) observed no increase in standing crop of sawgrass receiving 2.9–6 g P m−2 over a 186-day period in a greenhouse experiment. On the other hand, Steward and Ornes (1983) reported a significant increase in shoot and root production 440 days after a one-time P application of 25–400 mg kg−1 soil in the greenhouse. Davis (1989) reported that net annual aboveground production, leaf biomass, and tissue P concentration of both sawgrass and cattail were highest near the Hillsboro canal (the source of nutrient input to WCA-2A) and decreased with distance downstream. However, there was no clear relationship between mean annual surface water N or P concentration and annual productivity of biomass of sawgrass. In contrast, Craft et al. (1995) reported significant increases in sawgrass biomass production with P additions. In all studies, tissue P increased with P additions. In contrast to sawgrass, submerged species such as bladderwort (Utricularia spp.) or the macroalga Chara might be adversely affected by P additions. After 12 weeks of P dosing at a concentration of 10 mg l−1 (0.26 g m−2 week−1 for 22 weeks), Utricularia, Chara, and associated periphyton disappeared from the treated plots (Steward and Ornes 1975b). Flora et al. (1988) also observed the disappearance of the periphyton mat from channels receiving continuous loading of dilute P solutions. On the other hand, Vymazal et al. (1994) pointed out that periphyton communities respond differently to phosphorus additions and must be assessed individually. Craft et al. (1995) observed that after 1 year of adding 4.8 g P m−2 year−1 the decrease in Utricularia biomass was 90%, while there was a large increase in the Chara biomass. After a 6-year monitoring study conducted in WCA-2A, Urban et al. (1993) suggested that both nutrient enrichment and prolonged hydroperiod stimulate

505

506

J. Vymazal et al.

cattail proliferation. They found that, during typical wet years, cattail increased more rapidly than sawgrass at all sites, but nutrient enrichment seemed to accelerate cattail encroachment. During drought years, sawgrass expansion (and cattail decline) occurred at nutrient-enriched sites but not at low-nutrient sites. Newman et al. (1996) reported no difference between cattail and sawgrass growth at low nutrient and low water levels (15 and 30 cm). However, under high-nutrient levels and all water depths, cattail had a significantly higher growth rate than sawgrass. Cattail growth under enriched conditions also was higher than growth at the same water levels under unenriched conditions. Numerous studies have demonstrated that cattail is more aggressive than sawgrass at assimilating N and P to produce biomass (Davis 1991; Koch and Reddy 1992). Miao and Sklar (1998) observed that cattail in WCA-2A had higher N and P concentrations in the biomass and allocated more biomass to leaves, while sawgrass assimilated the nutrients for storage in shoot bases. Recent field studies suggest that cattail can take advantage of highnutrient supplies to increase photosynthesis and growth, enabling it to outcompete sawgrass (Chiang et al. 2000). The growth response of native Everglades vegetation to nitrogen seems to be very low. Steward and Ornes (1983) found that N additions (in the form of nitrate) at rates of 10–50 mg N kg−1 soil had no effect on sawgrass growth. Also, Craft et al. (1995) observed no effect on sawgrass and cattail biomass standing crop from additions of 5.6 to 20.4 g N m−2 year−1 for 2 years. Flora et al. (1988) and Walker et al. (1988) did observe the disappearance of the submerged macrophyte Utricularia and associated periphyton in response to continuous dosing with nitrate. However, the decline of Utricularia and periphyton occurred more slowly under nitrate loading as compared with loading with phosphorus. The limited response of Everglades vegetation to N additions may be attributed to the use of NO _3 as the nitrogen source in experimental studies (Stewart and Ornes 1975b, 1983; Flora et al. 1988; Walker et al. 1988). Nitrate is highly soluble and is readily lost from the wetlands via leaching and denitrification under favorable conditions, e.g., sufficient source of organic carbon, anoxic or anaerobic conditions (Vymazal 1995). Ammonium (NH+4) is a more practical source of N because (1) it is not directly lost from the system via denitrification, (2) it is adsorbed onto cation exchange sites in the soil, reducing leaching losses, and (3) it is the dominant form of inorganic nitrogen in Everglades porewaters and thus readily available to plants. Because ammonia nitrogen is more reduced energetically than nitrate, it is a preferable source of nitrogen for assimilation (Kadlec and Knight 1996). Although ammonia is the preferred nitrogen source, nitrate can also be used by some plant species. Nitrate uptake by wetland plants for plant growth is less favorable than ammonium uptake under most conditions. On the other hand, ammonia could be lost from the system through volatilization. Reddy and Patrick (1984) pointed out that losses of NH3 through volatilization from flooded soils and sediments are insignificant if the pH value is below 7.5, and very often losses are not serious if the pH is below 8.0. At pH of 9.3, the ratio between ammonia (NH3) and ammonium ions (NH+4) is 1:1 and the losses via volatilization are significant. Algal photosynthesis in wetlands as well as photosynthesis by free-floating and submerged macrophytes

20 Plant Community Response to Long-Term N and P Fertilization

507

often creates high pH values >9.0 during the day in the Everglades (Vaithiyanathan and Richardson 1999). Thus, the response of plant communities to long-term N and P additions under controlled experimental conditions is unknown. Moreover, no data exist on the interactive effects of both N and P additions. The objective of the Duke N and P fertilizer study in three Everglades plant communities (sawgrass, mixed sawgrass–cattail, and slough) was to determine the long-term (>6 years) effects of fertilizer additions on (1) biomass production and diversity of macrophytes, (2) nutrient concentrations in the tissues of emergent and submergent macrophytes, (3) changes in plant community species composition, and (4) soil phosphorus chemistry.

20.2 20.2.1

Materials and Methods Experimental Design

The experimental units were established during the summer 1990 in sawgrass, sawgrass–cattail (mixed), and slough communities in WCA-2B of the central Everglades (80°20´00˝ west, 25°13´00˝ north) (Fig. 20.1, Plate 8). Sawgrass plots were established in areas of pure sawgrass. In the mixed plots, sawgrass and cattail were the dominant macrophyte species. Other species occurred only sparsely and included, e.g., Eleocharis cellulosa, royal fern (Osmunda regalis L.), lance-leaf arrowhead (Sagittaria lancifolia L.), alligator weed (Alternanthera philoxeroides (Mart.) Griseb.), fragrant water lily (Nymphaea odorata Ait.), or pennywort (Hydrocotyle umbellata L.). Slough plots were dominated with E. cellulosa, Panicum hemitomon, P. repens, and Utricularia–periphyton assemblage. Those species were present in all treatments, while N. odorata, water hyssop (Bacopa caroliniana (Walt.) Robins), or S. lancifolia were present occasionally. The three plant communities exist along a hydroperiod gradient (Fig. 20.2), with the sawgrass site occupying the drier end and the slough site occupying the wetter end. The mixed site was characterized by greater water level fluctuations than the other two sites because of its proximity to the S-146 water control structure. During wet seasons, the water control structures open and release pulsed water into WCA-2B. The fluctuating hydroperiod at sites nearer the gates represent a more disturbed Everglades hydrology. All three sites experienced nearly year-round surface inundation in the third and fourth years of the study due to higher rainfall and water release schedules (Fig. 20.2). A randomized complete block design (Steel and Torrie 1980) with three replications was used to test the effects of N and P additions on Everglades vegetation. Seven fertilizer treatments and two controls (no fertilizer addition) were used, giving a sum of 27 plots for each vegetation community (Fig. 20.3). Each plot consisted of a 2 × 2 m2 within a larger (4 × 4 m2); fiberglass siding (1 m high × 2 m long) was used to delineate the 2 × 2 m plots. The fiberglass siding was pushed approximately

508

J. Vymazal et al.

Fig. 20.1 Location of study sites in WCA-2B of the Everglades, in southern Florida

20 cm into the soil and low permeability foam was applied to the plot corners to retain the fertilizer. Fertilizer was applied to the 2 × 2 m plots while the larger 4 × 4 m plots served as buffers between treatments. Each replicate contained two control plots. One plot utilized fiberglass siding, and was identical to the fertilized plots, while the second control was a 2 × 2 m plot without fiberglass siding used to determine whether the fiberglass siding affected plant productivity within the 2 × 2 m enclosure. Monitoring wells (15 and 30 cm deep) were established in each 2 × 2 m plot. The wells for collecting soil porewater were made from PVC pipe with the end covered by screen, and they were placed in each plot together with the water level well.

20 Plant Community Response to Long-Term N and P Fertilization

509

Fig. 20.2 Surface water depth in sawgrass, mixed, and slough communities during the period of continuous monitoring in 1990–1994. (Note: Seasonal water patterns are typical of those found throughout the study period)

Fig. 20.3 Experimental design for applying nitrogen and phosphorus to Everglades vegetation. (See Table 20.1 for figure key)

510

J. Vymazal et al.

20.2.2

Fertilizer Application

Nitrogen (as ammonium chloride, 26% N) and P (as superphosphate–monocalcium phosphate, 7.86% P) were applied by hand in amounts shown in Table 20.1 beginning in August 1990 and ending in June 1997. During the first 2 years (1990–1992) nitrogen and phosphorus were applied every 2 months. In 1993 and 1994, the frequency of application was reduced to every 4 months. Previous studies of fertilizer additions to wetland vegetation indicate that these rates are sufficient to produce a response in growth (Broome et al. 1983). The low-P treatment (0.6 g m−2 year−1) was used to compare rates of maximum P accumulation in the proposed study with previous estimates of P accumulation in the enriched portion of WCA-2A (0.21– 0.63 g m−2 year−1, mean 0.40 g m−2 year−1) (Richardson and Craft 1990). A combination of N and P was applied at two levels to assess the interaction between the N and P fertilizers.

20.2.3

Water Analysis

During each visit to the plots, standing water was thoroughly pumped out of the wells. A slight vacuum of about −2 kPa was applied to speed collection of the new water into wells. Samples were drawn by the slight vacuum directly into an acid-washed plastic vial. Surface water samples were collected at mid-depth in the water column. Surface and porewaters were collected from each plot prior to initiation of fertilizer applications. One surface water and two porewater samples (in the slough site only one at 30-cm depth) were collected from each plot, filtered through Gelman type A/E glass fiber filters, and analyzed for PO4-P and NH4-N using a TRAACS 800 autoanalyzer (Method nos. 804-86T and 812-86T; Bran + Luebbe, Inc., Elmford, NY).

Table 20.1 Rates of nitrogen and phosphorus applied to experimental plots P (g m−2 year−1) Treatment N (g m−2 year−1) Control 1 (C1, enclosed) Control 2 (C2, unenclosed) Low P (LP) Medium P (MP) High P (HP) Low N (LN) High N (HN) Medium N + P (MNP) High N + P (HNP)

0 0 0 0 0 5.6 20.4 5.6 20.4

0 0 0.6 1.2 4.8 0 0 1.2 4.8

20 Plant Community Response to Long-Term N and P Fertilization

20.2.4

511

Soil Analysis

Soils were collected from all three communities in 1992–1994 and 1997 using a rectangular box corer (7.5 × 7.5 cm), which minimizes the compression of organic soils. The cores were sliced into 5 cm increments. Field moist soil was analyzed for KCl extractable NH4-N and NO3-N, and bicarbonate extractable PO4-P (Keeney and Nelson 1982; Olsen and Sommers 1982). The remaining soil was dried, ground, and sieved through a 2-mm diameter mesh screen and analyzed for total P and total N. Additional soils (0–5 cm) were also collected from the control, HP, and HNP plots at the sawgrass and slough sites after 4 years of N and P additions. In these samples, P was fractionated into plant available organic and inorganic P, P bound to microbial biomass, Fe- and Al-bound inorganic P, Fe- and Al-occluded inorganic P, Ca-bound inorganic P, and residual P, using the methods of Hedley and Stewart (1982) and Qualls and Richardson (1995).

20.2.5

Vegetation Harvesting and Biomass Analysis

Vegetation was collected outside of the plots (between blocks) prior to initiation of fertilizer treatments to quantify baseline biomass in the three plant communities. Twelve 0.25-m2 quadrants were enumerated and the aboveground standing crop (live material and standing litter) was measured. Aboveground biomass was harvested from the treatment plots in August of 1991–1994 and in October 1997 by clipping one randomly selected 0.25-m2 quadrant in sawgrass and mixed sites and 0.125-m2 quadrants in each slough site plot. In 1997, a 0.25-m2 quadrant was used for slough plots as well. The aboveground vegetation was separated by species and by live and dead (standing litter) material. Plant material was dried at 75°C to a constant weight to estimate dry mass. The dried plants were ground through a 2-mm diameter mesh screen and analyzed for organic C, N, and P. Total P was analyzed as phosphate in a nitric/perchloric acid digestion (Sommers and Nelson 1972) using TRAACS 800 autoanalyzer (Method no. 781-86T, Bran + Lubbe, Inc., Elmsford, NY). Total N and organic C were analyzed by combustion on a CHNS/O analyzer (Perkin-Elmer 2400 CHNS/O, Perkin-Elmer, Norwalk, CT).

20.2.6

Photosynthesis and Leaf Area Index

Between May 1993 and February 1994, photosynthesis on sawgrass and cattail plants was measured in selected treatment plots at the sawgrass and mixed sites. An infrared gas analyzer (Model LI-6200, LI-COR, Inc., Lincoln, NE) was used to measure photosynthesis on fully expanded leaves of similar-aged mature plants. Measurements were taken during the months of May, August, and December 1993 and February 1994, between 9 a.m. and 2 p.m. during the periods of full sunlight

512

J. Vymazal et al.

(>1,000 µmol m−2 s−1). At the sawgrass site, five sawgrass plants were randomly selected from each plot for photosynthesis measurements. At the mixed site, five each of sawgrass and cattail were chosen. Afterwards, the leaves were clipped at the soil surface, and leaf length and leaf area were measured using a leaf area meter (Model LI-3 100, LI-COR, Inc., Lincoln, NE). Leaves were dried, weighed, ground, sieved, and analyzed for tissue N and P content. During the biomass harvest event in August 1993, additional leaf lengths from the clip plots were measured to quantify the relationship between leaf length and leaf area index (LAI). LAI was estimated by the equation LAI = (1/specific leaf weight) × (gbiomass m−2 ground area), where specific leaf weight is the weight of harvested leaves. Ground area photosynthesis was then calculated by multiplying LAI and leaf area photosynthesis. Plant species diversity was determined by enumerating the number of species harvested and the proportion of biomass associated with each species from each quadrant. An index of species diversity (H) was calculated based on the Shannon– Weaver information theory index (Shannon and Weaver 1949).

20.2.7

Statistical Analyses

A randomized complete block design with three replicates was used to test effects of fertilizer additions on standing crop biomass and nutrient content of Everglades vegetation (Steele and Torrie 1980). The data were analyzed using the General Linear Models procedure (SAS 1988). Analysis of variance (ANOVA) was used to compare the enclosed and unenclosed treatment to determine whether the enclosure affected biomass production or nutrient uptake. If no difference was detected, which was the case for all measured parameters, the control treatments were combined for comparison with the fertilizer treatments. Treatment means were separated using the Ryan–Einot–Gabriel–Welsh multiple F-test (SAS 1988).

20.3 20.3.1

Results and Discussion Soil and Water

Phosphorus concentration in the top 5 cm of soil in all treatment plots is presented in Table 20.2. At the sawgrass site, considerably higher values (>1,000 µg g−1) were found only in HP and HNP treatments, and these values were significantly different from the values found at control plots. In addition, P concentration of 935 µg g−1 found at HP plot in 1994 (data not shown) was also significantly higher as compared with the control. Those values are typically found in the most nutrient-enriched areas of the Everglades WCA-2A (Craft and Richardson 1993b; DeBusk et al. 1994;

20 Plant Community Response to Long-Term N and P Fertilization

513

Table 20.2 Concentration of phosphorus (µg g−1) in the top 5 cm of experimental plots 1992 1997 Sawgrass C MN HN LP MP HP MNP HNP

630 609 587 756 879 878 862 1,076*

566 510 628 593 882 1,363* 883 1,463*

788 657 720 840 1,179 1,011 925 1,033

855 864 836 843 894 1,122 775 983

Mixed C MN HN LP MP HP MNP HNP Slough

C 259 234 MN 287 178 HN 255 242 LP 316 278 MP 257 331 1,039* HP 676* MNP 372 356 882* HNP 671* * Significantly different from control treatment within the same sampling year (p = 0.05)

Vaithiyanathan and Richardson 1999). At the mixed site, elevated phosphorus concentrations in the soil (>1,000 µg g−1) were also found in MP, HP, and HNP treatments but the values were not statistically significant. At the slough site, the soil P concentrations were much lower compared with the sawgrass and mixed sites. Elevated concentrations were found only in HP and HNP treatments in 1992, 1994 (data not shown), and 1997. After 4 years of N and P additions, mean phosphorus fractions were evaluated (Table 20.3) in sawgrass and slough control, HP, and HNP plots (Chiang et al. 2000). At the sawgrass site, the HNP plots exhibited significant increases in easily exchangeable inorganic and organic P, Fe- and Al-bound inorganic P, and residual P, compared with the controls. Although not significant, the concentrations of these fractions were higher in the HP plots than in the controls. A much higher amount of P bound to microbial biomass was found at the surface soil, especially in the control plots (348 µg g−1), as compared with treatment plots.

514

J. Vymazal et al.

Table 20.3 Mean soil phosphorus fractions (µg P g−1 soil) at 0–5 cm (n = 3) after 4 years of N and P additions (modified from Chiang et al. 2000) Treatment Site

P fraction

Microbial Pi and Po Easily exchangeable Pi Easily exchangeable Po Fe- and Al-bound P Sonicated Fe- and Al-bound Pi Ca-bound Pi Residual P Total Pb Slough Microbial Pi and Po Easily exchangeable Pi Easily exchangeable Po Fe- and Al-bound Pi Sonicated Fe- and Al-bound P Ca-bound Pi Residual P Total Pb * Significantly different from control (p = 0.05) a Number in parenthesis is percentage of total P b Total P is the summation of the various fractions Sawgrass

Control

High P

High N + P

348 (70)a 39 (8) 22 (4) 13 (3) 19 (4) 47 (10) 5 (1) 493 109 (41) 2 (0.07) 0.7 (0.04) 3 (1) 7 (3) 19 (7) 126 (47) 267

154 (26) 101 (17) 60 (10) 44 (7) 28 (5) 170 (28) 46 (7) 603 171 (35) 32 (7)* 22 (4)* 2 (0.04) 15 (3) 223 (45)* 26 (5) 491*

103 (19) 101 (18)* 69 (12)* 51 (9)* 33 (6) 106 (19) 90 (16)* 553 168 (34) 24 (5)* 18 (4)* 1 (0.02) 19 (4) 170 (34)* 97 (20) 497*

At the slough site, easily exchangeable inorganic and organic P, Ca-bound inorganic P, and total P were significantly higher in the HP and HNP plots than in the controls (Table 20.3). Most of the soil P in the HP plots was bound to Ca (223 µg g−1) and microbial biomass (171 µg g−1), although the microbial fraction was not significantly different from the controls. Total P at the HP plots (491 µg g−1) was nearly twice that of the controls (267 µg g−1). Similar partitioning of soil P was found in the HNP plots, with Ca-bound inorganic P contributing 34% (170 µg g−1) of the total P and with total P concentrations (497 µg g−1) almost twice that of the controls (267 µg g−1) (Chiang et al. 2000). Similar to the sawgrass site, a considerable amount of P was bound to microbial biomass in plots. Controls contributed 34–35% of total soil P but had lower amounts of microbial P than treatment plots. At the P-enriched plots of the slough site, precipitation of P with Ca, microbial immobilization of P, and rapidly cycled easily exchangeable inorganic and organic P were the major P sinks. The large pool of Fe- and Al-bound inorganic P at the sawgrass site may have been somewhat overestimated, as soils extracted with NaOH might result in hydrolysis of some organic P, thus elevating the amounts of inorganic P (Reddy et al. 1998). The large amounts of P stored as residual P, most of which is organic, and Fe- and Al-bound inorganic P indicate that Everglades soil may be an important sink for long-term storage in areas receiving P loadings. The results suggest that P loading results in enrichment of both labile (easily exchangeable inorganic and organic P) and recalcitrant P pools (Fe- and Al-bound inorganic P, residual P) (Chiang et al. 2000). The large amount of P bound to microbial biomass is not surprising, since microbial populations are usually found higher in the surface peat (Qualls and

20 Plant Community Response to Long-Term N and P Fertilization

515

Richardson 1995). The microbial P is probably very labile and cycles rapidly, and it represents a significant pool for P storage. Reddy et al. (1998) observed that in WCAs 1, 2A, and 3, 13–57% of total P in the top 10 cm of the soil was present as microbial biomass and easily exchangeable organic P, as compared with 19–35% and 41–70% of microbial biomass found in our fertilized and unfertilized plots, respectively. The findings from the slough plots are in accordance with those reported by Qualls and Richardson (1995), who observed an increase in Ca-bound inorganic P in enriched surface soils of WCA-2A. It is likely that the increase in Ca-bound inorganic P in the HP and HNP plots was the result of sorption and precipitation with calcium in the calcium carbonate-rich marl soils of the slough (Swift 1984). Nitrogen concentrations in the top 5 cm of soil in experimental treatment plots varied very little during the 1992–1994 and 1997 analyses, and there was no significant difference between control plots and fertilized plots. The mean N concentrations in soil (% dry matter) varied from 1.78 to 2.93 at the sawgrass site, 1.91 to 2.77 at the mixed site, and 1.78 to 2.93 at the slough site. The slough site had the lowest values for soil N concentrations, as was the case with the P concentrations. There was little fluctuation within treatments during the course of the experiment in both surface water and porewater in controls and fertilized treatments, indicating no long-term enrichment of surface water with phosphorus in response to the nutrient additions (data not shown). Craft and Richardson (1994) observed that surface water PO4-P concentrations increased only for several days after fertilizer application. In contrast to PO4-P, NH4-N concentrations were much higher in porewater as compared with surface water.

20.3.2

Tissue Nutrient Concentrations

At the sawgrass site, concentrations of P in live aboveground biomass were significantly higher in the HP and HNP treatments on every sampling with the exception of HP treatment in 1994 (Table 20.4). At a mixed site, significantly higher P concentrations were recorded in the HP treatment 1, 2, and 4 years after nutrient additions were started. In the HNP treatment, significantly higher P concentrations as compared with control were found only after 4 years of fertilization. In general, sawgrass tissue P concentrations were lower than those in cattail. Cattail in the mixed site did not exhibit any significant responses to fertilizer additions during all 7 years, but tissue in the HN and HNP plots was generally greater than in the control plots. Similarly to our findings, Davis (1991) observed that both cattail and sawgrass responded to higher surface water nutrient concentration by increasing tissue P, with leaf tissue P in cattail two times higher than in sawgrass. Newman et al. (1996), Toth (1987, 1988), and Koch and Reddy (1992) also reported that cattail tissue P concentration was more than three times higher than sawgrass under both enriched and unenriched soil conditions. According to Davis (1989), cattail is an opportunistic species that takes advantage of short-term

Table 20.4 Mean phosphorus concentration (µg P g dry mass−1) in aboveground live biomass in fertilized plots at sawgrass, mixed, and slough sites Site/vegetation 1991 1992 1993 1994 1997 Sawgrass Cladium

C MN HN LP MP HP MNP HNP

209 197 218 354 266 541* 193 516*

94 162 170 99 165 329* 126 490*

131 119 100 138 137 327** 238** 599*

169 119 159 119 211 271 160 385**

251 272 241 266 257 359** 300 340**

C MN HN LP MP HP MNP HNP C MN HN LP MP HP MNP HNP

256 263 292 313 302 615* 268 594 580 523 512 413 626 1,026 150 911

205 199 215 257 321 684** 253 371 499 358 355 – 573 709 489 375

254 141 176 486 455 779 207 1,015** 470 478 423 550 621 – 531 997

236 205 144 217 156 509** 171 364 364 219 414 471 551 481 – 362

289 277 253 273 406 402 347 474 532 594 423 659 756 434 440 735

C MN HN LP MP HP MNP HNP C MN HN LP MP HP MNP HNP C

253 244 214 190 211 314 226 259 280 377 311 295 318 339 369 668 144

232 262 276 224 245 380 278 379 207 268 269 228 216 992 257 452 110

299 323 262 196 261 482 311 585** 468 391 334 294 421 893 383 756 174

276 335 296 287 271 503 339 869** 283 324 250 392 372 754** 406 734** 116

257 258 260 232 304 419 321 559** 270 275 247 229 299 262 248 374 174

Mixed Cladium

Typha

Slough Eleocharis

Panicum

Utricularia/ periphyton

MN 217 99 118 77 107 HN 208 97 190 75 118 131 LP 262 317 311 712** 230 MP 373 294 392 588** 1,390* – – – HP 1,151* 305 407 – 160 MNP 577** 849* 892* – 343** HNP 1,078** * Significantly different from all other treatments (p = 0.05); **Significantly different from control treatment within the same sampling year (p = 0.05). Updated from Craft et al. (1995) and Chiang et al. (2000)

20 Plant Community Response to Long-Term N and P Fertilization

517

(annual) increased nutrient supply by assimilating large reserves of nutrients, especially P, and increasing productivity. In contrast, sawgrass responds to nutrient supplies by partitioning the nutrients to storage organs such as the shoot bases (Miao and Sklar 1998). Eleocharis tissue P concentration did not respond to P additions during the first 2 years after initiation of fertilizer additions (Table 20.4). After 3 years (in 1993), P concentration in the live aboveground biomass was significantly higher in the HNP treatment compared with the control plots. This phenomenon was also found in the HNP treatment in 1994 and 1997. Tissue P concentrations in Panicum biomass did not respond significantly to P additions during the first 3 years after fertilizer additions had started (Table 20.4), but concentrations found in the HP and HNP treatments were higher than in control plots. Significantly higher tissue P concentrations were recorded in the HP and HNP plots in 1994 after 4 years of fertilizing. In 1997, however, no significant differences were observed. Like sawgrass, P concentrations in the Utricularia–periphyton complex increased in response to P applications after only 1 year (Table 20.4). After 1 year of fertilization in the HP, MNP, and HNP plots, P concentrations were significantly higher than in control plots. Tissue P in the HNP treatment was significantly higher than the control on every sampling when the Utricularia–periphyton complex was present in the plot. The same pattern applies to the HP treatment, but the Utricularia–periphyton complex disappeared after 2 years of fertilization and has not been recorded since then. In 1994, after 4 years of fertilizer additions, significantly higher tissue P concentrations compared with control plots were recorded in the LP and MP treatments. Nitrogen concentrations in live aboveground biomass of sawgrass (0.31–0.87% dry matter), cattail (0.44–1.30% dry matter), Eleocharis (0.56–1.44% dry matter), and Panicum (0.58–1.48% dry matter) did not respond significantly to nitrogen additions. In addition, nitrogen concentrations varied little during years and among treatments and species as well. The N concentration of Utricularia and associated periphyton was significantly higher in plots receiving the highest rate of P (HP, HNP) than in control treatments when the complex was present in the plots (Table 20.5). The increased N content of the Utricularia–periphyton complex in plots receiving high P loads suggests that sufficient “native” N is available to support periphyton

Table 20.5 Mean nitrogen concentration (% dry mass) in Utricularia/periphyton complex at the slough sites Treatment 1991 1992 1993 1994 1997 C 0.99 0.90 0.91 0.90 MN 1.00 0.87 1.32 0.77 HN 1.11 0.92 0.94 0.76 LP 0.94 1.32 1.04 1.15 MP 1.21 1.23 1.26 0.96 1.96** – – HP 2.04* MNP 1.50 1.21 1.25 – 2.10** 2.35** – HNP 1.77* * Significantly different from control treatment within the same sampling year (p = ** Significantly different from all other treatments (p = 0.05)

1.03 0.79 0.85 0.82 1.06 – 0.97 1.55* 0.05);

518

J. Vymazal et al.

production at P loadings up to 4.8 g P m−2 year−1. Thus, P additions may stimulate N uptake in this macrophyte–algae complex (Craft et al. 1995). In Table 20.6, N:P mass ratios for sawgrass (both in sawgrass and mixed sites), cattail, Eleocharis, Panicum, and Utricularia complex are shown. Koerselman and Meuleman (1996) and Verhoeven et al. (1996) summarized on the basis of survey of a 40 fertilization studies that the N:P mass ratio >16 indicates P-limitation on a community level, while N:P ratio <14 is indicative of N-limitation. The results presented in Table 20.6 suggested that sawgrass is strongly limited by P in control plots both in sawgrass and mixed sites (N:P ratio ranges between 19.7 and 46.8). Single nitrogen additions generally increased this ratio in the MN and HN plots (range between 23.2 and 83.0). Phosphorus additions shifted the N:P ratio in sawgrass live aboveground biomass to less P-limited conditions and in high-P addition plots even to N-limited conditions. The N:P ratio in HP and HNP plots varied between 5.8 and 18.2 (Table 20.6) with an average value of 13.8. Cattail mass N:P ratio in control plots fluctuated little and was close to nonlimiting values (11.1–16.5). Nitrogen and phosphorus additions exhibited expected shifts in nutrient limitation. Single nitrogen additions caused a shift to slight P-limitation (mean N:P ratio 17.2) while single phosphorus additions caused a shift to a slight N-limitation (mean N:P ratio 12.1). Effects of combined N + P additions were not uniform, but HNP plots exhibited an average N:P ratio of 12.9 over the experimental period. Eleocharis exhibited P-limitation in control plots with N:P ratio averaging 27.5. Single nitrogen additions caused little change; the average N:P ratio was 27.0. Only the HP and HNP treatments showed a decrease in N:P ratio over years, and values dropped down to either nonlimiting or even slightly N-limiting conditions in the HNP plots (Table 20.6). Strong P-limiting conditions were found in control plots for Panicum, with N:P ratio values ranging between 21.8 and 53.6 and an average of 38.2. Single nitrogen additions and single P additions (LP and MP) did not cause any shift, with an average N:P ratio values of 36.5 and 34.9, respectively. The NP and HNP additions lowered the value of N:P ratio even to N-limiting conditions in the case of HP addition. Utricularia–periphyton N:P ratio was very high in control plots (52.3–81.8) and increased in plots with single nitrogen additions (46.1–112). Higher P additions decreased the N:P ratio values to nonlimiting conditions, but higher P additions eventually caused the disappearance of Utricularia from the plots.

20.3.3

Macrophyte Biomass

In general, all treatments at the sawgrass site showed an increase in aboveground sawgrass biomass over years (Fig. 20.4). However, significant difference within the sampling year (p = 0.05) in the aboveground biomass (live + standing litter) between control and fertilized plots were observed only at HP and HNP treatments in 1993 (after 3 years of fertilizer application) and 1997 at HNP plot (Fig. 20.4). In 1993, biomass in the control plots averaged 2,167 g m−2 while biomass in HP and HNP treatments amounted to 8,597 and 6,453 g m−2, respectively. In 1997, sawgrass

Table 20.6 N:P ratios in the live aboveground biomass Site/vegetation 1991 1992

1993

1994

1997

Sawgrass sawgrass

C MN HN LP MP HP MNP HNP

38.3 35.5 33.9 21.2 27.4 12.2 35.2 16.9

46.8 42.0 34.1 31.3 32.7 18.2 32.5 11.8

45.8 42.9 83.0 47.8 36.5 17.7 28.2 12.2

25.4 36.1 35.2 32.8 17.5 12.5 23.8 12.2

21.5 30.5 23.2 20.7 18.3 17.5 16.3 17.4

C MN HN LP MP HP MNP HNP C MN HN LP MP HP MNP HNP

29.7 29.3 29.5 26.5 26.8 12.5 29.5 16.7 15.0 15.7 18.8 15.3 14.9 7.1 41.3 10.9

26.8 29.1 27.9 33.5 20.2 9.1 25.1 17.0 16.0 17.3 17.2 – 10.8 10.2 17.2 14.7

19.7 36.2 29.0 12.6 13.0 6.5 20.2 5.8 13.6 18.6 20.0 10.0 9.8 – 10.2 4.7

23.7 29.3 27.8 25.8 25.0 15.1 28.1 16.2 16.5 28.8 13.5 13.0 23.6 12.7 – 24.6

20.8 24.5 24.5 26.0 17.8 15.1 19.6 13.9 11.1 9.3 11.1 9.0 11.0 10.1 12.3 9.8

C MN HN LP MP HP MNP HNP C MN HN LP MP HP MNP HNP C

27.3 29.1 31.8 32.6 30.3 23.2 28.3 30.1 38.2 26.5 43.1 31.9 31.4 28.6 31.4 15.7 68.8

33.2 29.8 27.5 32.6 27.3 21.1 28.1 19.0 53.6 39.9 42.0 43.9 54.2 8.2 45.1 20.4 81.8

25.4 26.0 26.3 38.8 55.1 16.0 24.1 15.4 21.8 29.2 39.5 34.7 35.2 11.5 24.5 14.8 52.3

26.8 25.4 28.0 27.5 28.4 15.9 23.9 10.2 41.7 31.2 42.0 29.1 27.2 13.0 20.9 14.9 77.6

24.9 24.8 21.5 26.3 20.4 16.0 21.8 13.5 35.9 33.5 37.7 31.9 29.4 20.1 34.7 25.1 59.2

MN HN LP MP HP MNP HNP N-limitation is in bold

46.1 53.4 35.9 32.4 17.7 26.0 16.4

87.9 94.8 41.6 41.8 14.1 39.7 24.7

112.0 49.5 33.4 32.1 – 30.7 26.3

100.0 101.0 16.2 16.3 – – –

73.8 72.0 62.6 46.1 – 60.6 45.2

Mixed sawgrass

cattail

Slough Eleocharis

Panicum

Utricularia/ periphyton

520

J. Vymazal et al.

Fig. 20.4 Aboveground biomass (live + standing litter) at the sawgrass site. Asterisks denote significant differences from control treatment within the same sampling year (p = 0.05)

biomass in control and HNP plots averaged 3,073 and 9,940 g m−2. Davis (1989) reported that sawgrass in enriched areas of WCA-2A that are adjacent to our study sites had higher aboveground biomass than plants in unenriched areas, supporting our fertilizer results. Craft et al. (1995) and Chiang et al. (2000) reported that live biomass only was significantly higher in HP and HNP treatments as compared with control plots after 2 years of fertilizer additions in 1992. Moreover, previous field and greenhouse studies also observed no effect of N and P additions on short-term (less than 20 months) aboveground biomass production by sawgrass (Steward and Ornes 1975b; Sutter 1992). A recent short-term field study by Miao and Sklar (1998) also showed that sawgrass biomass was lower in areas of lower water depth and soil P concentrations as compared with areas of higher P and deeper water. Importantly, according to field observations (Chap. 9) long-term (>20 years) higher concentrations of soil P cause decline of sawgrass growth, and sawgrass is often not present in areas of the WCAs with elevated P soil concentrations >1,000 mg kg−1 (Richardson and Qian 1999; Richardson et al. 1999). At the mixed site, the response to fertilization was not so evident and there was no significant increase in total aboveground biomass (live + standing litter) over the whole period of the experiment (Fig. 20.5). Surprisingly, cattail live biomass at the mixed site did not exhibit any significant response to P additions during all 7 years. The fact that cattail did not respond to P additions suggests either that changes in native Everglades macrophyte plant communities occur over a period of longer than 7 years, or that factors in addition to nutrients are important agents of cattail encroachment. For example, changes in water levels and the frequency of fires may also affect the extent to which cattail invades sawgrass-dominated communities

20 Plant Community Response to Long-Term N and P Fertilization

521

Fig. 20.5 Aboveground biomass (live + standing litter) at the mixed site

(Urban et al. 1993). This site was much closer to a flow gate S-144 (see Fig. 20.1), and plots here were subjected to much greater hydrologic changes in water level. Those pulses of water may have overridden fertilizer effects as compared with the sawgrass site, which had little water level fluctuations. At the slough site, the biomass of Utricularia and associated periphyton (Fig. 20.6) was significantly reduced by the highest P applications after only 1 year of P addition. Utricularia was not found in the HP treatment in 1993, 1994, and 1997, while in MNP and HNP treatments Utricularia was absent in 1994. Steward and Ornes (1975b) also observed the disappearance of Utricularia after 12 weeks of P dosing at a rate of 10 mg l−1 (0.26 g m−2 week−1). The authors suggested that Utricularia disappearance might be caused by dense phytoplankton growth and subsequent reduced light penetration. On the other hand, thick periphyton growth may limit the light penetration to macrophytes. It has been demonstrated that epiphytic material absorbs photosynthetically active radiation (PAR) before reaching the leaf surfaces of vascular plants. The accumulation of epiphytic material may result in 80% attenuation of the incident radiation at the leaf surface (Twilley et al. 1985). The development of epiphytic communities on the leaves of vascular plants may reduce net production through several mechanisms other than PAR attenuation, including the reduction of diffusive transport of inorganic carbon, nitrogen, and phosphorus. Another reason for Utricularia disappearance may be the high pH caused by increased photosynthetic activity of periphyton. At higher pH values, carbon dioxide is not present in the water and several species of Utricularia are not able to use bicarbonate as a source of carbon for photosynthesis (L. Adamec, personal communication). The Utricularia–periphyton complex was replaced with Chara in higher P treatments. Chara appeared in the HP, MNP, and HNP plots after 1 year of nutrient addition

522

J. Vymazal et al.

Fig. 20.6 Utricularia/periphyton biomass at the slough site. Asterisks denote significant differences from control treatment within the same sampling year (p = 0.05)

and, additionally, in the MP plots in the fourth year. In the HP treatment, Chara was first recorded in 1992 (362 g m−2) together with Utricularia (16 g m−2). Since 1993, only Chara was recorded in the HP treatment with biomass of 192, 346, and 211 g m−2 in 1993, 1994, and 1997, respectively. There was no significant effect of P or N additions on aboveground biomass of Panicum (hemitomon and repens) and E. cellulosa at the slough site. In general, between 1991 and 1997 biomass of Eleocharis decreased gradually but in all plots including controls. Aboveground biomass (live + standing litter) ranged between 310 and 522 g m−2 in 1991 and from only 92 to 196 g m−2 in 1997. Chiang et al. (2000) reported that sawgrass aboveground live biomass increased with P loading (r2 = 0.76), with biomass accrual leveling-off at cumulative additions of 10–15 g m−2 (Fig. 20.7a). The amount of biomass at the mixed site reached a maximum at 15–20 g P m−2 (Fig. 20.7b). Both sites displayed a steep increase in biomass up to approximately 5 g m−2 of added P. The biomass plateau in the sawgrass and mixed sites could be due to self-shading (especially in the high biomass plots of the sawgrass site), limitation by other nutrients such as potassium (Steward and Ornes 1975a), or stress such as oxygen limitation caused by surface water inundation. At the slough site, the Utricularia–periphyton complex declined in response to cumulative P additions (r2 = 0.75, p < 0.05). Cumulative loads of above 1 g P m−2 resulted in a decrease in biomass (Fig. 20.7c), and levels above 10 g m−2 nearly eliminated the Utricularia–periphyton complex (Chiang et al. 2000).

20 Plant Community Response to Long-Term N and P Fertilization

523

Fig. 20.7 Aboveground live biomass response (g m−2) to cumulative amounts of P addition at (a) sawgrass site, (b) mixed site, and (c) slough site (Utricularia–periphyton complex) (from Chiang et al. 2000)

524

20.3.4

J. Vymazal et al.

Plant Species Diversity

With the exception of the replacement of the Utricularia–periphyton complex by Chara in the P fertilized plots at the slough site, there was no effect of nutrient additions on plant species richness or species diversity in the sawgrass or mixed plots. The loss of Utricularia–periphyton complex also did not affect species number or diversity relative to the control plots because of the concurrent increase by Chara (Chiang et al. 2000).

20.3.5

Phosphorus and Nitrogen Pools

In Table 20.7, phosphorus and nitrogen pools in various compartments of plant biomass are given for control and HNP plots after 7 years of fertilization are given. The results show both for P and N pools the general order belowground > live aboveground > standing litter > old litter, with the belowground pool being by far the highest. However, the data revealed that the old litter pool, which is mostly neglected in harvesting studies, might contribute significantly to the total biomass pool. This is in agreement with generally accepted nutrient distribution in wetlands (Verhoeven

Table 20.7 Mean phosphorus and nitrogen pools in control (C) and high N + P plots after 7 years of N and P additions Site P (g m−2) Sawgrass

ABG-live HNP

N (g m−2) Sawgrass Mixed Slough

OL

Total

NM

3.647 (5,126)

C

NM

0.574 (3,458)

HNP C HNP C HNP C

18.57 5.69 9.85 7.12 7.90 7.46

4.18 3.17 13.22 8.50 NM NM

108.67 59.16 74.34 40.88 39.71 19.94

HNP C

Slough

BG

0.986 1.007 (7,056) 5.321 (2,884) (6,152) 0.230 (952) 0.157 (2,122) 1.306 (3,908) 0.801 0.448 (2,727) 2.536 (1,460) (3,934) 0.397 0.193 (1,184) 0.806 (1,045) (2,229) 0.394 (536) 0.072 (2,719) 3.181 (1,871) 0.221 (656) 0.020 (1,813) 0.333 (989)

C Mixed

SL

HNP

25.13 5.36 10.30 4.16 2.88 1.31

57.14 44.94 40.97 21.10 28.93 11.17

0.277 (1,016) 7.591 (17,108) 0.097 (724)

1.790 (7,706)

0.536 (2,123) 4.321 (10,244) 0.350 (1,323) 1.746 (5,781)

ABG aboveground live biomass, SL standing litter, BG belowground biomass, OL old litter (on the floor, not attached to plants), NM not measured. Biomass data (g DM m−2) are in parentheses

20 Plant Community Response to Long-Term N and P Fertilization

525

1986; Richardson 1991) with peat-litter being by far the largest P and N pool. The results also indicate that sawgrass plants have the largest N and P pools as compared with the two other plant communities (mixed and slough). Both above- and belowground biomass N and P pools were found in the order sawgrass > mixed > slough with the exception of belowground pool in slough which was higher than that at mixed site due to very high P concentration in Nymphaea rhizomes.

20.3.6

Photosynthesis

To estimate productivity rates, net photosynthesis measurements were taken in 1994. High phosphorus additions resulted in increased photosynthesis compared with control plots, but significant differences were observed only at the HNP plots of the sawgrass site (Table 20.8). Chiang et al. (2000) pointed out that mean annual rates of cattail leaf area photosynthesis were much higher than sawgrass, but significant differences in cattail photosynthesis between controls and treatment plots were not observed. Additions of P at the highest rate resulted in significantly higher LAI, biomass, and hence ground area photosynthesis in the HP and HNP plots at both sites (Table 20.8). Biomass harvested in this study represents an integration of primary production over an annual period, while photosynthesis measurements represent a short-term response to hourly and daily conditions. This may explain a significant response to P in terms of biomass, but not photosynthesis. Additions of P increased tissue P content at both sites, while additions of N + P resulted in increased tissue N for cattail and sawgrass at the mixed site (Table 20.8). Contrary

Table 20.8 Mean annual rates of photosynthesis, LAI measured in August 1993 and leaf tissue N and P measured between May 1993 and February 1994 (adapted from Chiang et al. 2000)

Site species

Treatment LAI

C Sawgrass sawgrass HN HP HNP C Mixed sawgrass HN HP HNP C Mixed cattail HN HP HNP

1.88 3.42 16.2* 12.8** 3.12 1.96 5.6* 5.07** 0.19 0.18 1.56* 1.65*

Photosynthesis Photosynthesis (leaf area) (µmol (ground area)a Leaf tissue Leaf tissue P (µg g−1) m−2 s−1) (µmol m−2 s−1) N (%) 13.3 12.3 13.4 15.3* 13.1 13.1 14.7 15.1 16.0 17.5 17.0 18.1

25 42.1 218* 196* 40.9 25.7 82.6* 76.6* 3.0 3.2 26.5* 29.9*

0.83 0.79 0.69 0.83 0.80 0.76 0.83 0.95* 1.07 1.03 1.07 1.27*

266 236 479* 408* 289 252 441* 635* 715 562 l,025* l,137*

Significantly different from control (p = 0.5); **Significantly different from control (p = 0.05) Ground area photosynthesis was calculated by multiplying biomass with LAI

* a

526

J. Vymazal et al.

to most studies (Mooney et al. 1978; Field and Mooney 1986), foliar N concentration was not strongly correlated to photosynthesis for sawgrass or cattail. A regression analysis performed on leaves collected in August showed no significant relationship (p > 0.05) between leaf area photosynthesis and foliar N (r 2 = 0.02) or foliar P concentration (r2 = 0.01) in sawgrass. In contrast, cattail photosynthesis exhibited a significant (p < 0.05) but not strong relationship for both N (r2 = 0.42) and P (r2 = 0.22) (Chaing et al. 2000).

20.4

Conclusions and Lessons for Restoration

Phosphorus additions at a rate of 4.8 g P m−2 year−1 increased soil total P concentrations values above 1,000 µg g−1, i.e., concentrations that are typically found in the most nutrient-enriched areas of the Everglades WCA-2A. Nitrogen additions caused no increase in total N concentration in the top soil layer. Enrichment of soil labile (easily exchangeable inorganic and organic P) and recalcitrant P pools (Feand Al-bound inorganic P and residual P at the sawgrass site, Ca-bound inorganic P at the slough site) was observed in all three plant communities. The highest phosphorus addition of 4.8 g m−2 year−1 (both single and in combination with nitrogen) caused significantly higher tissue P concentrations in live aboveground biomass of all species in all three sites with the exception of cattail where significant increase in tissue P concentration was not observed. However, the species differed at time when significant increase was reached. While sawgrass and Utricularia reached this level after only the first year of fertilization, Eleocharis and Panicum reached this level after 3 and 4 years of fertilization, respectively. The analysis of plant biomass indicates that Cladium, Eleocharis, Panicum, and Utricularia–periphyton complex were P limited in control plots. Nitrogen additions generally increased this limitation while high P additions (4.8 g P m−2 year−1) shifted the N:P status to nonlimiting or slight N-limitation status. Cattail at a mixed site was found at approximately nonlimiting status in control plots and N and P additions shifted the status accordingly. High phosphorus additions also caused significant difference in the aboveground biomass (live + standing litter) in HP and HNP treatments in sawgrass site as compared with control plots. The highest aboveground sawgrass biomass was recorded in 1997 in HNP treatment (9,940 g m−2), which is about seven times higher than the background biomass recorded in 1990 before initiation of fertilizer additions. All fertilizer treatments exhibited increased biomass over the course of the study, but significant increase was reached at different times with high-P treatments reaching this level after 3 years of fertilization. Surprisingly there was no significant increase of either sawgrass or cattail biomass in the fertilized plots as compared with control at the mixed site. This may have been the result of altered hydroperiod conditions. At the slough site, the biomass of Utricularia–periphyton complex was significantly reduced by the highest P applications and Utricularia was not found in the HP treatment in 1993, 1994, and 1997 while in the MNP and HNP treatments Utricularia was

20 Plant Community Response to Long-Term N and P Fertilization

527

absent in 1994. In HP treatment, Utricularia was replaced by macroalga Chara, which also appeared in MNP, HNP, and MP, plots. It appears that Everglades plant communities have a limited ability to utilize added P. In our study, cumulative P loading greater than 10 g m−2 led to leveling-off of aboveground biomass production, as well as a near elimination of the Utricularia– periphyton complex. The encroachment of cattail into sawgrass-dominated communities was not observed. However, the loss of the Utricularia–periphyton complex in the slough and concurrent increase in Chara may have significant implications for the Everglades food web, as the periphyton slough community is an important component of it. The largest pool for both phosphorus and nitrogen is belowground biomass followed by live aboveground biomass, standing litter, and old litter. Additions of P resulted in a significant increase in sawgrass leaf area photosynthesis, increased leaf area of sawgrass and cattail, and increase ground area photosynthesis, especially for sawgrass.

C. Disturbance Experiments

21

The Effects of Disturbance, Phosphorus, and Water Level on Plant Succession in the Everglades Curtis J. Richardson, Alisa Dickson, and Mengchi Ho

21.1

Introduction

The Everglades is a dynamic fen system that has thrived under conditions of low phosphorus, shallow water sheet flow, and annual alterations of wet/dry seasons. It has also adapted to periodic lightening induced fires, droughts, hurricanes, and occasional winter frosts (Davis 1943). The main water inputs into the system before anthropogenic alterations were rainfall and overflow of Lake Okeechobee during flooding events (Davis 1943). As a result, the plant communities within this ecosystem have evolved under low-nutrient conditions with the dominant community being monotypic Cladium jamaicense Crantz stands. Interspersed with the Cladium stands are wet prairies, sloughs, alligator holes, and marsh areas with tree islands that provide a diversity of aquatic habitats (Loveless 1959; Gunderson 1994; Jordan et al. 1997). The Everglades has also experienced alteration of hydrologic patterns and nutrient enrichment (Doren et al. 1997). Historically, the hydroperiod encompassed a 7–12 month time period, during which the water depth was below 50 cm, with dry down periods, where the water table was greater than 30 cm below the soil surface (Wu et al. 1997). In the historical Everglades before human intervention, the total phosphorus concentration ranged from 5 to 10 µg l−1, and phosphorus has been determined to be the limiting nutrient for this ecosystem (Doren et al. 1997; Richardson et al. 1999). The conversion of plant communities in the Everglades is associated with hydrologic alterations, increased nutrients, and anthropogenic disturbances. Jordan et al. (1997) delineated new vegetation classes such as open water, Typha monocultures, and willow tree island communities that have become predominant in disturbed areas of protracted hydroperiods and increased nutrients. David (1996) attributes plant community conversions mainly to altered hydrology. Increased flooding duration increases occurrence of species such as Sagittaria lancifolia, Nymphaea odorata, and Utricularia spp., and drowns tree islands and wet prairie vegetation. Dry conditions on the other hand, lead to severe fires that destroy large areas of the vegetation. Other studies attribute plant community conversions to increased nutrients that favor species such as Typha over Cladium (Doren et al. 1997; Stewart

531

532

C.J. Richardson et al.

et al. 1997). There is no consensus among scientists as to the primary cause of community change, nor are there many quantitative studies that show the effects of hydrologic and nutrient alterations on the Cladium community (Vaithiyanathan and Richardson 1999). The purpose of this study is to quantify the effects of soil disturbance, P additions, and increased water depth on the succession pattern of Everglades plant communities. The hypothesis is that alteration of the conditions to which plant communities are specifically adapted will result in a community different from those historically present in the Everglades. Of special interest is the cause for Typha invasion, because past studies have indicated that increased Typha growth is primarily found in areas of increased nutrients and water levels (Newman et al. 1996). While Typha has been found to be a successful competitor in areas of deeper water and increased nutrients, there is not a definitive agreement as to whether Typha directly competes with and excludes Cladium (Newman et al. 1998). This is of special interest in the Everglades, where Cladium plays an important role for juvenile fish refuge, peat accretion, and an organic phosphorus reservoir (Miao et al. 1997).

21.2 21.2.1

Materials and Methods Plot Establishment

In October of 1990, 12 plots were established in WCA-2B of the northern Everglades (Fig. 21.1). This area was chosen because the nutrient-laden waters from the Everglades Agricultural Areas (EAAs) have not impacted the area, nor has it undergone extensive airboat disturbance (SFWMD 1999). The main water input into the system is rainfall and overflow from WCA-2A during flooding events. There is also some ground water seepage near the dike at the southern end of WCA2A that could contribute to water and nutrient inputs. The site chosen was an old established Cladium stand, ~2 m tall with few cattails (less than 1%). The specific test areas did not contain Typha within the actual plot or buffer zones. This area was selected based on the following criteria: 1. Area exceeded 95% Cladium coverage 2. Existing Cladium was healthy and well established 3. Proximity to seed source of other species, including cattail To test soil disturbance, nutrient, and water level effects on community succession and species invasion, four treatments (replicated three times) were established for a total of 12 treatment plots that were randomly selected with a randomized complete block design (Steel and Torrie 1980). The plots consisted of 3 m × 3 m areas cleared of all vegetation down to the surface of the soil. All plant materials were removed using gas-powered bush trimmers; and cut plant materials were removed from the

21 The Effects of Disturbance, Phosphorus, and Water Level

533

Fig. 21.1 Study site location in Water Conservation Area 2B of the Everglades, Florida. The disturbance study (x-rectangle) is next to the sawgrass (Cladium) and mixed sites

plot. Within the 3 m × 3 m areas, 2 m × 2 m plots were laid out allowing for a 0.5-m buffer zone outside of the actual experimental area. The outer border of each 2 m × 2 m plot was cut 60 cm into the soil, and the plot was covered with two layers of 6-ml black polypropylene plastic sheeting with the edges being forced into the ground at a depth of 60 cm. The purpose of the plastic was to cut off all light sources and separate the plots from surrounding soils to insure that all living plant material died as well as preventing rhizome invasions prior to the start of the experiment in the fall of 1991.

21.2.2

Experimental Design

Vegetation was removed from the plots in the fall of 1990 prior to soil disturbance or P additions. Three plots had only vegetation removed, three had the soil disturbed by tilling the top 15 cm of soil, and three plots were enriched with 0.6 g m−2 year−1 of phosphorus as Na2HPO4 annually in June of each year. The final three plots had the top 45 cm of soil removed and the top 15 cm replaced onto the lower soil surface to maintain the seed bank. To maintain experimental integrity, a 30-cm high PVC wall surrounded the plots. These plots were monitored quarterly from 1992 to 1999. Analysis of variance was conducted using a mixed model to do a repeated measurements test for significant treatment effects, change in growth of plant cover over time, and interaction

534

C.J. Richardson et al.

effects between treatments and time for plant coverage and pore water chemistries (SAS 1988). A Pearson’s correlation coefficient was used to relate plant species and chemical parameters.

21.2.3

Site Monitoring

PVC wells were established in each plot to monitor water levels and pore water chemistry. Two wells at 15- and 30-cm depths were placed in each 2 m × 2 m plot. Water elevation, pH, temperature, conductivity, and oxygen were monitored quarterly along with surface and pore water chemistries. The water samples were filtered through Gelman type A/E glass fiber filters and analyzed for NH 4-N and NO 3-N using a TRAACS 800 autoanalyzer (Method nos. 804-86T, 818-87T, Bran + Luebbe, Inc., Elmford, NY). Phosphorus was analyzed as PO4-P by the ascorbic acid–molybdate blue method (Wetzel and Likens 1990) with a 5-cm cell on a Beckman spectrophotometer. Total P in water was analyzed following persulfate digestion of shaken unfiltered samples (Wetzel and Likens 1990) followed by colorimetric analysis. Atomic absorption spectroscopy using a Perkin-Elmer Model 5100 PC A.A. was utilized to analyze Ca, Na, and Mg. Ion chromatography using a Dionex Ion Model 300 Chromatograph was utilized to analyze Cl− and SO− 24. The plots were subdivided into quadrants and sampled quarterly, in a nondestructive manner, for changes in macrophyte vegetation and periphyton cover. The species for each plot were visually estimated on a percent coverage basis converted from a modified Braun-Blanquet scale (Phillips 1959). In this scale, 0 represents no plant coverage, x represents sparse or <1% coverage, 1 is plentiful but of small cover value < 5%, 2 is numerous covering from 5 to 25%, 3 is any number of individuals covering 25–50% of the area, 4 is 50 75% coverage, and 5 is greater than 75% coverage. These values were converted utilizing the midrange value for each category and averaging over the four quadrants in each plot.

21.3 21.3.1

Results and Discussion Water Depth

Water depths were recorded in each of the treatment plots throughout the course of the study. The average water depths throughout the course of the study were 31.81 cm (±18.86), 28.46 cm (±17.32), 31.55 cm (±17.92), and 67.61 cm (±22.30) for the disturbed, phosphorus, control, and deep-water plots, respectively. The deep-water treatment plots were nearly double the depth of all the other treatments during the study. Typically, water levels dropped around April of each year, with the highest water levels seen in January (Fig. 21.2). The lowest water levels were found from April 1993 to April 1994.

21 The Effects of Disturbance, Phosphorus, and Water Level

535

Fig. 21.2 Water depth averages for the four treatments represent water level treatment trends found throughout the experiment. Control = vegetation removed

21.3.2

Statistical Analysis

The control plots remained 100% Cladium during the period of monitoring. Chara displayed an initial rapid increase in coverage for the vegetation and phosphorus treatments, a slight rise and fall before a rapid increase in the disturbed (tilled) treatments, and a slow establishment in cover for the deep-water treatments (Fig. 21.3). Coverage was highly variable in the first 4 years of the experiment but treatment responses for the latter half of the experiment were more uniform. The highest percent coverage for Chara was in the phosphorus plots followed by the control, disturbed, and then deep-water treatments. The mean percent plant coverage for Chara was highest for the phosphorus treatment plots and could be attributed to open niches created by a decline in other species such as Utricularia and its associated algal mat. A study conducted by Craft et al. (1995) found that in plots with medium and high phosphorus additions, Chara expanded into those plots and replaced the floating algal mats. Chara expansion was also correlated with a decline in Utricularia. The only species that showed significant treatment effects was Chara (p < 0.05). Chara’s growth over time was statistically significant (p < 0.0001). However, there were no interaction effects between species growth and treatments. The deep-water sites had lower mean percent Chara coverage than the other treatment sites (Fig. 21.4). It would seem more intuitive that Chara would have a higher percent coverage in the deep-water sites, as this macroalgal species is dependent upon water environments for survival. Our results suggest that deep water initially can reduce Chara

536

C.J. Richardson et al.

Fig. 21.3 Chara plant coverage for each of the four treatments over the course of the experiment from July 1992 to May 1999. The replicates are averaged

Fig. 21.4 Chara mean percent plant coverage for each of the four treatments. The replicates and years are averaged for the course of the study (p = 0.05)

21 The Effects of Disturbance, Phosphorus, and Water Level

537

Fig. 21.5 Cladium plant coverage for each of the four treatments over the course of the experiment from July 1992 to May 1999. The replicates are averaged

establishment. The increased depth could be related with decreased light source, increased algal mat coverage, or competition from other species. Figure 21.5 shows that for the first half of the experiment (0–3 years), Cladium had low coverage, but it showed a significant increase in plant coverage during study years 4–7. Cladium initially grew in the fertilized treatments and was found 6 months afterwards in the sites with removed vegetation. The tilled treatments suppressed Cladium growth until August of 1995, while the deep-water treatments completely inhibited Cladium recolonization. Cladium showed significant interaction effects between treatment and change in growth over time, p = 0.0026. A significant treatment effect was not found over vegetation removed plots; however, Cladium did exhibit a significant change in growth over time effect, p = 0.0001. Non-walled reference Cladium sites maintained a 100% cover of Cladium throughout the experiment (data not shown). Miao et al. (1997) found Cladium to have enhanced growth rates with nutrient additions. Many succession studies have shown the importance of nutrient limitation as a mechanism for structuring plant communities, and enriched areas can aid in Cladium restoration provided that cattail is removed via fire or mechanical methods (Miao et al. 1997; Richardson et al. 1995, 1999). While Cladium growth is enhanced in phosphorusenriched areas, it is most competitive in low-nutrient environments. For instance, at phosphorus levels greater than 10 µg l−1 in surface water, Typha has a competitive edge over Cladium with an increased leaf production and rapid growth expansion although Typha does exhibit shorter plant longevity than Cladium (Davis 1994).

538

C.J. Richardson et al.

One of the difficulties in Cladium reestablishment is its inherently low seed germination rate (see Chap. 22). Miao et al. (1997) found a maximum seed germination rate of only 28% for Cladium. Cladium most likely recolonized via rhizomes, which explains the time lag before any percent plant cover was observed. Cladium and other plants with diffusive gas transport are also at a disadvantage for colonizing inundated areas when compared to plants with pressurized bulk flow, e.g., Typha (Miao et al. 1997). Busch et al. (1998) found that the average macrophyte cover on marsh study areas appeared to have an inverse relationship with water depth and that Cladium densities were inversely correlated with water depth indicating an affinity to drier environments. Doren et al. (1997) found that at depths greater than 50 cm and inundation periods exceeding 6–10 months, Cladium is partially or totally replaced by wet prairie vegetation. These findings further support the prediction that Cladium plant abundance declines as a function of water depth. This is a result of Cladium’s physiology. Cladium limits water loss from its roots; therefore, it does not have a fully oxidized rhizosphere that would alter the soil chemistry and aid in respiration and nutrient uptake in flooded soils. However, Typha has an extensive oxidized rhizosphere, which enables it to outcompete species such as Cladium in deep-water habitats (SFWMD 1999; Brix 1999). Utricularia showed widely fluctuating percent plant coverage over the course of the 7-year study as well as between treatments (Fig. 21.6). The deep-water treatments were relatively more stable in terms of Utricularia populations. This can be attributed

Fig. 21.6 Utricularia plant coverage for each of the four treatments over the course of the experiment from July 1992 to May 1999. The replicates are averaged

21 The Effects of Disturbance, Phosphorus, and Water Level

539

to these sites having more uniform water depths than the other treatments, which represents more of a slough community that favors Utricularia. A study conducted by Busch et al. (1998) found Utricularia to be positively correlated with algal mat cover as well as increasing water depth and frequency. This supports their hypothesis of a positive feedback system where Utricularia serves as an algal mat substrate as well as a consumer of periphyton-grazing zooplankton (known at the Utricularia– periphyton complex). The increases and decreases in percent Utricularia plant cover for the deep-water treatments correspond to a similar pattern in percent algal mat cover (Fig. 21.6). The vegetation-removed sites generally had a lower percent plant coverage than the other treatments, while the tilled and phosphorus treatments tended to have a similar percent coverage of Utricularia over time (Fig. 21.6). Utricularia exhibited a statistically significant interaction effect between treatment and change in growth over time, p = 0.0033. This species did not indicate a significant treatment effect, however, there was a significant change in growth over time, p = 0.0001. To test the effects of increased hydrology and nutrient enrichment on Typha invasions, we examined our deep-water and phosphorus treatments for Typha plant coverage trends (Fig. 21.7). Typha did not show any significant increases in plant coverage over time, treatment effects, or interaction effects between treatment and time. This indicates that the plot conditions were not conducive to Typha invasions potentially due to low phosphorus treatment concentrations, life history characteristics of Typha, and community interactions. The establishment of Typha from seed germination is rare although it is more likely to happen in periods of dry conditions (Wu et al. 1997). Stewart et al. (1997) found that without light, seed germination of Typha was

Fig. 21.7 Typha plant coverage for each of the four treatments over the course of the experiment from July 1992 to May 1999. The replicates are averaged

540

C.J. Richardson et al.

0–10% while with light, germination increased to 91–100%. A high percentage of Typha germination requires long exposure to light, low oxygen concentrations, and high temperatures (Stewart et al. 1997). High algal mat percent coverage would then inhibit Typha germination as it would reduce light penetration and dissolved oxygen in the water column (Correll 1998). Wu et al. (1997) also concluded that Typha dispersal was predominately spatially dependent. Thus, Typha invasions were potentially reduced in this study due to barrier walls around each plot and a lack of adjacent Typha plants. While floating algal mats and green algae did not indicate significant treatment or interaction effects, they did exhibit significant plant cover growth over time. The floating algal mat plant coverage showed a significant (p = 0.0001) change in growth over time (Fig. 21.8). There was an initial rapid increase in plant coverage and then complete mortality during the first winter. For the remainder of the experiment, the algal mat exhibited a fairly stable pattern of increased coverage in the spring and summers and a decrease in fall and winter. These fluctuations can most likely be attributed to the variations in the hydroperiod and seasonal growth pattern of algae (Vymazal and Richardson 1995). The green algae also showed a significant (p = 0.0001) change in growth over time (Fig. 21.9). There are only a few dates within our study time frame that show an increase in percent coverage of green algae and these events are correlated with a percent coverage decline in the floating algal mat. Algal mats in the Everglades are primarily found on macrophyte stems and floating aquatic species at the water surface and at the fen bottom and are dominated by calcareous cyanobacteria, diatoms, and desmids (Browder et al. 1994). Busch et al. (1998) also found that

Fig. 21.8 Mean algal mat percent plant coverage over time from July 1992 to May 1999. All replicates and treatments are averaged since no treatment effects were found (p = 0.0001)

21 The Effects of Disturbance, Phosphorus, and Water Level

541

Fig. 21.9 Mean green algae percent plant coverage over time from July 1992 to May 1999. All replicates and treatments are averaged since no treatment effects were found (p = 0.0001)

following prolonged periods of inundation, the proportion of diatoms and green algae may increase in the algal mat association. In a study conducted by Doren et al. (1997), declines in algal mats were associated with water quality alterations. Algal mats, composed of pollution-intolerant species, are replaced by eutrophic cyanobacteria and pollution-tolerant green algae associated with increased concentrations of phosphorus or increased nutrient loads in surface waters (SFWMD 1999; Browder et al. 1994).

21.3.3

Water Chemistry

Baseline water chemistry and physical variables along with water depth were also monitored to determine if significant difference existed over time, among treatments or for any interaction between time and treatment (Table 21.1). Significant interaction effects occurred for surface water NH4-N, surface water Mg, pore water NH4-N, pH, and depth. Treatment effects were found for pore water NO3-N, Ca, Na, and Cl with the deep-water treatments having significantly lower concentrations for these chemistries except pore water NO3-N. All chemical and physical parameters showed a significant change over time except for pore water PO4-P. The lack of increase in the PO4 concentrations may be due to the low level of phosphorus additions (0.6 g m−2 year−1) used in this experiment. This P addition level is below the P assimilative capacity calculated for the Everglades (Richardson and Qian 1999). This also suggests that the ecosystem has some capacity to assimilate a low P load without a significant change in the ecosystem, e.g., cattail invasion.

542

C.J. Richardson et al.

Table 21.1 p-Values for surface and pore water chemistry and water quality conditions Parameter Time Treatment Interaction Surface water

Pore water

Water quality

21.4

NH4-N NO3-N PO4-P TP Ca Mg K Na Cl NH4-N NO3-N TP Ca Mg K Na Cl pH Oxygen Temperature Depth

0.0085 0.0001 0.0001 0.0045 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0003 0.0167 0.0001 0.0001 0.0105 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001

0.0212

0.0403

0.0427 0.0182

0.0075

0.0394 0.0376 0.0404 0.0234

0.0001

0.0001

Conclusions and Lessons for Restoration

Plant succession is a natural ecological process, but in areas that operate under anthropogenic controls and not by their typical historic conditions, disturbances can have more of a profound effect on the community and deflect plant communities from the usual pattern of succession. In an area of the Everglades that is not as greatly impacted by nutrient inputs, exotic invasions, or excessive airboat transportation, our experiment demonstrates a shift in plant species and cover dominance due to disturbance. The hydroperiod can be considered the primary controlling mechanism under which the historic Everglades communities evolved, and deep prolonged hydroperiods proved to be important in controlling the reestablishment of Cladium and Typha as well as other species in our study. The results of a recent redox and phosphorus interaction study on Cladium support our finding that deep water inhibits sawgrass germination (Lissner et al. 2003). They report that it would be difficult for C. jamaicense to establish or recruit at sites with low nutrients when flooding results in strongly reduced soils over extended periods, conditions such as those found in our deep water plots. Human management of the Everglades’s hydrology may therefore play an important role in establishing or maintaining oligotrophic Everglades plant communities All the plots were disturbed in some form, which enabled new species to colonize and change the community structure. By contrast, a nearby pure sawgrass stand (control) remained 100% Cladium over the experiment period. This experiment

21 The Effects of Disturbance, Phosphorus, and Water Level

543

was established in 1991, but it was not until 6 years later that Cladium became the dominant species in any individual plot. Cladium communities have low species diversity and turnover rate while maintaining a dense canopy structure that has not been reestablished fully in the course of this study, which suggests a slow rate of recolonization and stand recovery. Chara and Utricularia were the dominant species in all 12 plots and currently still have greater than 50% plant coverage in all plots after 7 years. This would suggest that the area has a higher water level more typical of slough communities than the original monotypic Cladium stands or that it takes longer for Cladium to become established. Due to the direct influences of the periphyton–Utricularia complex on light, water, and nutrient acquisition by macrophytes, as well as indirect interactions with soil formation, physical–chemical changes and trophic webs, algal mats are considered key components of early Everglades marsh plant communities (Busch et al. 1998). The dominance of the slough community plant species, especially the periphyton–Utricularia complex, is associated with an increase in water depth, which suggests that both mat cover and water depth are impeding Cladium’s recolonization of the deep-water areas. Cladium was suppressed in plots for the first few years but most especially in the disturbed treatments while in the deep-water treatments it was completely inhibited over the entire experimental period. Initial colonization began in the phosphorus treatments and then the control sites. This supports the idea that hydrology is the most important factor in controlling Cladium community establishment. It also suggests that phosphorus additions enhance Cladium growth and establishment when conditions for Typha colonization are not optimal. Typha was an initial invader in all treatment sites but was more prevalent in the disturbed and phosphorus treatments, both treatments having higher nutrient levels than the other treatments. However, Typha never had a high percent plant coverage in any individual plot and was not present for the last 3 years. Phosphorus or other nutrient conditions were thus not favorable to the permanent establishment in these plots; however, seed dispersal ability and rhizomatic growth as well as the altered hydrology confound the issue of phosphorus being the primary factor influencing Typha colonization and expansions in these experiments. The Everglades plant communities have natural adaptations to various disturbances. The disturbances in our study paralleled natural disturbances, although the degree and magnitude of our induced disturbances might have been greater. Because the disturbed and deep-water treatments indicated more pronounced community effects, our management recommendations suggest that care must be taken when flow structures or water pumping regimes result in deep-water flooding or soil disturbance and vegetation removal. Soil disturbances are natural occurrences in the Everglades. Holes dug by American alligators provide important aquatic habitat for fauna and flora during times of low hydroperiods. Airboats also disturb the soil and vegetation, and plant communities along the airboat trails change due to introduction of species or pollution in addition to soil disturbance. Managers cannot completely halt airboat travel through the areas, but they can restrict travel to times when the water levels are deep

544

C.J. Richardson et al.

enough that the soil is not scraped or disturbed by the boat. They can also establish regular trails so that new cross-country trails are not consistently being cut. Among the other treatments, there was still a high percentage of species that are more indicative of slough communities. Cladium is stress-resistant but not for extended high hydroperiods. The results of the deep-water treatments and past research indicate that Cladium is either inhibited or outcompeted in areas that have a water depth of 30 cm or greater. In our study, Cladium did not return in the deepwater treatments that were an additional 30 cm lower than the water table for the other treatment sites. Moreover, floods and hurricanes, the natural phenomena that could increase the water table to depths greater than 30 cm, exist in the Everglades only for short time periods. Periods of dry downs are also important to allow Cladium to germinate from seeds as well as enabling them to have a competitive edge over other species. Therefore, managers in control of water levels in the WCAs should allow for periods of dry downs in areas of extensive Cladium stands and try to avoid increasing the water level above 30 cm for any prolonged period of time, although they will probably never be able to mimic the historic hydroperiod. This study did not determine the amount of added P that would cause Typha invasion in the Everglades. However, it does suggest that a loading of 0.6 g m−2 year−1 does not result in the establishment of Typha in former Cladium stands.

D. Germination Experiments

22

Establishment and Seedling Growth of Sawgrass and Cattail from the Everglades Sarah C. Goslee and Curtis J. Richardson

22.1

Introduction

Loveless (1959) described the vegetation of the Everglades as being predominantly sawgrass (Cladium jamaicense), with 65–70% cover. In wetter areas, the vegetation consists of wet prairie communities, and the deepest areas, known as sloughs, contain emergent and aquatic species (Davis 1994; see Chap. 4). Cattail (Typha domingensis) is native to the Everglades, but historical populations were low (Chap. 12). Historically, the Everglades was a low-nutrient system, especially deficient in phosphorus, because it received nutrients mainly from rainwater (Chap. 2). Few plant species are capable of growing in such a wet, infertile habitat. In some areas of the Everglades, particularly the Water Conservation Areas (WCAs), the vegetation is changing. Both sloughs and sawgrass stands are being invaded and replaced by cattails, especially near the water inflows from the Everglades Agricultural Area (EAA), and some sloughs in other areas are being filled in by sawgrass (Richardson et al. 1999). In WCA-2A, sawgrass has decreased from 95% cover to only 61% between 1973 and 1991 (Jensen et al. 1995). This change is correlated with increased nutrient loadings, and also with increased hydroperiod (Urban et al. 1993; Craft et al. 1995). Changes in the frequency of fire and other disturbances may also be a contributing factor. Cattail encroachment is not unique to the Everglades. Wilcox et al. (1985) observed the invasion of cattail into a sedge wetland at Indiana Dunes National Lakeshore. Cattail increased with increased water levels. Colonies became established by seed, and spread vegetatively. Cattail in Indiana was the most successful in stabilized elevated water levels. This could be similar to the process occurring in the Everglades, although there is the added complication of nutrient enrichment. There are many potential consequences of a shift from sawgrass to cattail, both to the Everglades itself and on a larger scale. The native plant community is being altered, and accompanying changes in associated plant species, periphyton, and in the fauna have been reported (SFWMD 1992, 2003, 2004, 2005). In addition, different decomposition characteristics of the two species lead to different rates of peat accretion and different peat characteristics (Chap. 3). Peat accumulates faster in cattail areas because of the higher productivity, but the peat is finer and less

547

548

S.C. Goslee and C.J. Richardson

fibrous than that formed by sawgrass (Davis 1991; Craft and Richardson 1993a). Detritus production is higher in cattail areas, and peat accretion and nutrient accumulation are higher in enriched areas (Reddy et al. 1993; Richardson and Qian 1999). There are also biogeochemical changes associated with the change from sawgrass to cattail. Alterations in nutrient accumulation and cycling occur due to the differences in uptake, productivity, and decomposition between the two species (Davis 1991). Cattail translocates more methane (a greenhouse gas) to the atmosphere than does sawgrass (Chanton et al. 1993). Cattail also has a higher transpiration rate, which could affect the water balance of the Everglades (Koch and Rawlik 1993). This area already loses an extremely high percentage of its water through evapotranspiration. Thus, understanding the complex biology of how cattail and sawgrass germinate, become established, and grow in the early stages of development is key information for restoring native sawgrass to its natural habitat and for sustaining Everglades ecosystem structure and function. The objectives of this study were to test the effects of water level and nutrient availability on the reproductive output, germination, and early growth of sawgrass and cattail. Ultimately, this information would enhance our ability to regenerate sawgrass in the Everglades, and possibly to reduce cattail recruitment. It has been shown for other species that various characteristics of the environment in which a plant is growing can affect its reproductive output. Differences in seed production among plants from the enriched and unenriched areas of the Everglades would affect the ability of those plants to reproduce themselves and to spread to new areas. The main germination experiment was designed to test the effects of water level and nutrient availability on the germination of sawgrass and cattail seeds collected from both enriched and unenriched areas. Two smaller experiments supplemented this study. The first of these experiments tested the effects of chilling, a potential dormancy breaker, on sawgrass germination. The second explored the role of light in stimulating the germination of both species. All germination studies were conducted in the Duke University Greenhouse. Chilling was tested on these two species because chilling is an important dormancy-breaking factor in many species. The Everglades is subtropical, so little if any effect was expected. Cattail has been found to germinate best in high light conditions, but the effects of light on sawgrass are unknown. A study of the growth of sawgrass and cattail seedlings was set up in an attempt to identify the age at which the faster potential growth rate of cattail overcame the initial advantage given to sawgrass by its larger seed size.

22.2

Species Descriptions

Sawgrass (C. jamaicense Crantz) is a member of the Cyperaceae and is found in the coastal southeastern United States (Table 22.1). It is a tall grass-like plant, reaching heights of 3 m. An individual can live several years but will die shortly after flowering. Sexual reproduction is insignificant; sawgrass spreads vegetatively from rhizomes. Steward and Ornes (1975b) examined the autecology of sawgrass, but very little

22 Establishment and Seedling Growth of Sawgrass and Cattail from the Everglades

549

Table 22.1 Biological characteristics, wetland status, and range of cattail and sawgrass (Fernald 1950; Gleason and Cronquist 1963) Typha domingensis Pers.

Cladium jamaicense Crantz

Typhaceae Wetland status: OBL Range: Delaware to South America and inland to California

Cyperaceae Wetland status: OBL Range: SE Virginia to Florida to Texas and Mexico Perennial Culm lifespan: 11–96 weeks Height: to 3 m Photosynthesis: C3

Perennial Culm lifespan: 28–370 weeks Height: 2–3 m Photosynthesis: C3

research has been done on any aspect of its physiology other than nutrient uptake. Conway (1936) published a series of physiological studies on a similar European species, Cladium mariscus. There are two forms of sawgrass, one short and one tall, found in the Everglades (Wood and Tanner 1990). The short form is found in the southern areas of the Everglades, probably where the environment is more stressful. It is not known if the two forms are genetically distinct. The species of cattail most common in the Everglades is T. domingensis Pers. (Typhaceae). It can reach heights of 2–3 m. It has a shorter lifespan than sawgrass and also dies after flowering (Davis 1989). Like sawgrass, it spreads primarily by vegetative reproduction. Cattail is native to the Everglades, but historically it was found only in scattered patches, not in the monoculture it now often forms (Loveless 1959; see Chap. 12). However, Davis (1943) mapped a large area of cattail in the northern part of the Everglades that is today part of the EAA. It was speculated that this large area of cattail and ferns was due to fire effects. One common site for cattail was around alligator holes, which tended to have deeper water and higher nutrient levels than the surrounding fen (Koch and Reddy 1992). Another common site was downstream of some tree islands (C.J. Richardson, personal observation). Sawgrass is adapted to the low-nutrient environment of the Everglades, and has many characteristics found in plants of resource-poor habitats (Chapin 1980, 1991). These include slow growth rate and limited ability to take advantage of increased resource availability (e.g., Davis 1991). It reacts to long-term increases in available nutrients by increasing uptake and productivity, but is insensitive to shorter (yearly) fluctuations. Cattail, however, is an opportunistic species similar to other species of fertile habitats (Chapin 1980), and can react quickly to changes in nutrient supply (Davis 1989). Both species use the C3 photosynthetic pathway (Bender 1971). Increased hydroperiod has also been mentioned as potentially favoring cattail over sawgrass. T. domingensis is a deep-water cattail, growing well even in heavily flooded conditions (Grace 1989). Sawgrass, however, is most productive in stable, shallow water levels (Toth 1987). The combination of deeper water levels and increased nutrients in some areas of the Everglades probably favors cattail more than either alteration alone.

550

22.3

S.C. Goslee and C.J. Richardson

Previous Field Studies

There have been many field studies of the response of sawgrass and cattail to nutrient loading in the Everglades (Toth 1987, 1988; Davis 1989, 1991; Richardson et al. 1991, 1992, 1993, 1994, 1999; Koch and Rawlik 1993; Urban et al. 1993; Craft et al. 1995; Newman et al. 1996; Miao et al. 1997, 2000). However, there has been little or no research in the laboratory or greenhouse on the physiological mechanisms behind these correlations. One mesocosm study (Newman et al. 1996) has addressed the effects of nutrient loading and hydroperiod on the growth of sawgrass and cattail. A field study on photosynthesis and transpiration of sawgrass and cattail completed by Chiang et al. (1994) showed that cattail’s carbon fixation rate was nearly double that of sawgrass. Most of the field studies suggest that both nutrient enrichment and hydroperiod changes are important factors in the decline of sawgrass (Chap. 9). Only a few studies, however, have examined the effects of water depth in the field. Urban et al. (1993) found that cattail increased more rapidly in wet years, and the rate of increase was accelerated by increased nutrients. In dry years, sawgrass was favored. Cattail appears to have less stomatal control than sawgrass, which can maintain constant stomatal conductance over a wide range of vapor pressures (Koch and Rawlik 1993). This suggests that cattail is less able to compensate for dry conditions. Grace (1989) found that T. domingensis is capable of growing in water as deep as 115 cm, although it grows best in somewhat shallower water. Toth (1987, 1988) found that both species grow best in stable shallow water levels, but deep and fluctuating water levels have a more negative effect on sawgrass than cattail. Several studies have examined the effects of only nutrient loading on sawgrass and cattail in the Everglades without considering other factors. Steward and Ornes (1975b) found no significant relationship in sawgrass between either soil phosphorus or water phosphorus concentration and leaf tissue phosphorus concentration. Cattail distribution, however, is correlated with phosphorus concentration in the sediments (Urban et al. 1993). Sawgrass and cattail can tolerate low-nutrient levels, and biomass and productivity increase with long-term fertilization (Davis 1989; Craft et al. 1995; see Chap. 20). Cattail is a more opportunistic species, capable of responding to yearly variation in nutrient levels. Sawgrass is not able to do this; its growth reflects only long-term levels. Thus cattail responds faster and also more strongly to enrichment (Davis 1989, 1991). Chiang et al. (1994) found that phosphorus addition increased photosynthesis rates in both species, but nitrogen addition had no effect. In addition, they found that the presence of cattail did not appear to affect the nutrient uptake and productivity of sawgrass in the site, with a mixed population of both cattail and sawgrass suggesting that increased nutrient levels are not the only factor involved in the change from sawgrass to cattail. All of the studies were concerned with the responses of mature plants. Both species spread almost entirely through vegetative reproduction, so the tillers have the resources of the parent to draw upon and are less vulnerable to environmental stresses than seedlings would be. Moreover, establishment of sawgrass from seeds can be extremely important after a fire (Chap. 9). Still, the clone must originally

22 Establishment and Seedling Growth of Sawgrass and Cattail from the Everglades

551

have become established from seed, so an understanding of the germination and seedling growth of these two species may provide insight into potential restoration and management strategies for both species.

22.4 22.4.1

Methods Seed Production

Cattail and sawgrass seeds were collected from an unenriched area of WCA-2A in July 1995. Seeds from the enriched area were collected in a similar fashion in August 1995. Sawgrass seeds were separated from the fruiting head and hulled by hand. Twenty seed-heads of each species were collected from several sites in the unenriched area; each of these seed-heads was bagged individually. Over the next few days, seeds were separated from the parent stalk and allowed to air dry. Cattail seeds are very small, so an alternative method was used based on the fact that the hairs of the pericarp, which hold the seed, expand on contact with water (Grace and Harrison 1986). A handful of fluff was placed in a blender, covered with water and agitated. The mixture was then poured into a bowl and allowed to settle. The seeds sank, while everything else floated. This process was repeated, and the seeds were removed from the bottom of the bowl by suction and allowed to dry (Sifton 1959; Bonnewell et al. 1983). Cattail seeds sink but sawgrass seeds float, reflecting potentially different means of dispersal. Three sets of 100 were counted out from each set of sawgrass seeds and weighed. For the much smaller cattail seeds, 200 were counted. The remaining seeds were weighed, and the average mass per 100 was used to estimate the total number of seeds present. Of the 19 cattail heads collected in the enriched area, all but one had seeds, while of the 17 collected in the unenriched area, only 6 had enough seeds to collect and weigh (more than about 50 seeds).

22.4.2

Germination Experiments

Seeds from each plant were counted out into four sets of 100. This was not possible in all cases, since some plants had less than 400 seeds. Two of these were placed into 15 ml centrifuge tubes, and two were placed in 5 cm Petri dishes on several layers of moist toweling. Dishes were watered with either distilled water or a solution consisting of 1 ml l−1 of a commercial 7–7–7 (N–P–K) fertilizer. Sawgrass treatments were placed in the greenhouse, and cattail treatments were put in the greenhouse the following day. The water or toweling was changed weekly, and dishes were watered as needed. Cattail sets were examined daily for germination. Germination was defined for the purpose of these experiments as protrusion of the radicle. Sawgrass germination was much slower than cattail germination, so sawgrass sets

552

S.C. Goslee and C.J. Richardson

were examined several times per week rather than daily. Cattail sets were removed from the study if no additional germination occurred for more than a week, or no germination at all for more than a month. Sawgrass sets were removed if there was no additional germination for more than a month. The experiment ended at 6 months.

22.4.3

Chilled Seeds

To simulate wetland conditions, seeds from each species were placed wet in a dark refrigerator. The chilling period was approximately 6 weeks. Cattail seeds germinated even in the refrigerator, especially surprising considering that the parents grew in the subtropical Everglades. Because they germinated under chilling, cattail was dropped from the germination trials. Only distilled water, no fertilizer treatment, was used in the germination trials of the chilled sawgrass seeds, but both tubes and dishes were used.

22.4.4

Dark Germination

The germination of the two species in the absence of light was tested to assess the effect of darkness on germination. Plants were chosen that had seeds left and that had shown fairly high germination rates in the previous trials. Seeds from two sawgrass plants from the enriched and two from the unenriched area were used, and four cattail plants from the enriched area were chosen. No cattail seeds from the unenriched area were available for this experiment. For each plant, six sets of 50 seeds were placed in tubes filled with distilled water. Five of the tubes were sealed with aluminum foil, and the sixth was left uncovered. All tubes were placed in the greenhouse on 22 May 1996. Foil-covered tubes were unwrapped periodically, and the germination rates compared with those of the uncovered tubes. The initial time lapse was too long (total germination had occurred), so the cattail seeds had to be retested beginning on 26 May 1996.

22.4.5

Seedling Growth

Although mature cattail has been found to grow faster than sawgrass, during the germination study it was evident that when immature, sawgrass seedlings grew considerably faster than cattail. This was probably due to the larger initial size and presumably greater stored reserves of sawgrass seeds. To test this, seedlings for this experiment were germinated in tubes of distilled water, from seeds collected in an enriched area of the Everglades. The seeds used were pooled from several plants.

22 Establishment and Seedling Growth of Sawgrass and Cattail from the Everglades

553

Sawgrass seeds were started first, on 28 May 1996, and cattail seeds were placed in water on 3 June 1996 when the first signs of seedling emergence were noted in sawgrass. This procedure ensured that the seedlings of both species were the same age, since cattail was found to germinate much more quickly than sawgrass. One to three days old seedlings were randomly divided into sets of 5. Three growth treatments were used. Some sets of seedlings were placed in tubes of distilled water to determine the amount of growth due to stored resources in the seed. The remaining sets of seedlings were placed in either fertilized or unfertilized Everglades peat. Two large plastic tanks were filled with peat collected in an unenriched area of the Everglades. Both tanks were watered frequently with distilled water to keep the water level just over the top of the peat. Five seedlings from each treatment were harvested weekly. Length of root and shoot was measured, and the number of leaves was counted. Each seedling was then air dried, separated into root and shoot, and weighed.

22.5 22.5.1

Results and Discussion Seed Production

There was little difference in mean mass per seed for sawgrass from the enriched and unenriched control areas, but a larger difference for cattail (Fig. 22.1a, b). Sawgrass seeds are several orders of magnitude larger than cattail seeds. Analysis of variance showed a highly significant (p < 0.0001) species effect, as would be expected from the difference in size of seeds of the two species. The site effect was also significant (p < 0.044), despite the lack of difference in the means of sawgrass seed size. Sawgrass produced more than double the seeds in the control area (Fig. 22.1c). In contrast, cattail produced far fewer seeds (2,000 vs. ≅ 25,000) in the unenriched area than in the enriched area (Fig. 22.1d). The species effect was also significant – cattail produced an order of magnitude more seeds than sawgrass. A comparison of the number of seeds produced by sawgrass in enriched and unenriched sites was not possible since seed collections from the enriched site were done after a major storm and many seeds had fallen from the plant. Although we expected a correlation between seed size and seed number, no such correlation was found for cattail, either by area or pooled. Sawgrass from the unenriched area was found to have a negative correlation between mass per 100 seeds and seed number (r = −0.83). Plants with larger seeds tended to produce fewer of them. Within a species, seed mass is mainly determined by resource competition (Wulff 1995). This was evident in cattail, with smaller seeds in the plants from the unenriched area (Fig. 22.1b). Total seed mass and number were also much lower for cattail from the unenriched area. Plants that are already under nutrient stress, as these cattails probably were, do not have much energy to spare for reproduction. Sawgrass is adapted to low-nutrient environments, so parental location did not

554

S.C. Goslee and C.J. Richardson

Fig. 22.1 Mean and standard deviation for seed mass per 100 seeds (a, b) and seed number (c, d). Sawgrass (Cladium jamaicense) and cattail (Typha domingensis) seeds were harvested from enriched and unenriched areas of WCA-2A

affect seed size. Cattail seeds are small, numerous, and wind-dispersed, which enables them to reach much of the surrounding area. Once the seeds contact the water, they are released from the pappus and sink, to remain in the sediments. Wide dispersal increases the likelihood of cattail seeds being present when a suitable site for germination becomes available. Sawgrass seeds are much larger, and they float. Water transport is a potential dispersal mechanism for sawgrass.

22.5.2

Germination

Seed viability is essential to a plant’s ability to reproduce itself sexually. The goals for this experiment were to compare the germination rates of sawgrass and cattail and to identify the effect of parental environment. In the Everglades the seeds could be dispersed to various types of environments, so treatments were included representing

22 Establishment and Seedling Growth of Sawgrass and Cattail from the Everglades

555

high- and low-nutrient availability and moderate or no flooding. Although flooding inhibits germination of many upland species, some wetland species are stimulated by flooding. The usual requirements of nondormant seeds for germination are water, proper temperature, and oxygen. The optimum conditions for germination must be determined separately for each species. Cattail has been shown to have the highest germination rate (over 90%) in low oxygen, high temperature, and high light conditions similar to those found in exposed mudflats (Sifton 1959; Bonnewell et al. 1983). Generally, 10 days is sufficient for most or all of the cattail seeds to germinate at room temperature (Sifton 1959). It is possible that cattail green leaf extract inhibits seed germination (Rivard and Woodard 1988), but Grace (1983) believes that this is unlikely. Sawgrass germinates best in saturated rather than flooded conditions, but it has a much lower germination rate than cattail (about 30%) and germinates much slower, with only about 16% germination in the first 30 days (Ponzio et al. 1995).

22.5.3

Main Germination Experiment

There were some definite differences in the germination curves of cattail seeds under the various treatments (Fig. 22.2). Seeds collected from the enriched area had a uniformly higher germination rate than seeds from the unenriched area. Of the enriched area seeds, those in tubes germinated slightly faster and reached higher germination rates than those in dishes. Fertilization had little or no effect on seeds

Fig. 22.2 Germination curves for cattail in various treatments. Error bars were omitted to increase clarity. Three letter legend key: E/U collected from enriched/unenriched area, D/F watered with distilled water/fertilizer solution, d/t kept in Petri dish/centrifuge tube

556

S.C. Goslee and C.J. Richardson

from the enriched area but increased the germination rate of seeds from the unenriched area. The seeds in tubes tended to have slightly higher germinations than those in dishes for the same water type. The most immediately obvious characteristic of the sawgrass germination curves (Fig. 22.3) is the lower maximum percent germination and the much longer time needed for both initial and maximum germination when compared to cattail. Within the sawgrass treatments, the seeds from the enriched area germinated better, although the difference was not as large as in the cattail treatments. Within the seeds from the enriched area, those in tubes had the highest germination rate and the steepest slope. Fertilized seeds reached higher maximum germinations than unfertilized seeds. The seeds collected from the unfertilized area showed different trends. Three of the four treatments were similar. The lowest germination was recorded from the fertilized seeds in tubes. This was the treatment that produced the highest germination rate in the seeds collected from the enriched area. Three parameters of the germination curve were examined in more detail: maximum germination, time to initial germination, and time to maximum germination. A summary of the maximum percent germination for all treatments, presented graphically in Fig. 22.4, was analyzed by ANOVA. The species effect was highly significant (p < 0.0001). Cattail germination was higher than sawgrass for nearly every treatment. The site effect was also highly significant (p < 0.0001) – seeds collected from the enriched area had higher germination rates than those from the unenriched area. The container effect was significant (p < 0.02) due to the increased germination in tubes for several treatments. Only two interaction effects were significant: species by site and species by site by container. These were driven by

Fig. 22.3 Germination curves for sawgrass in various treatments. Error bars were omitted to increase clarity. Three letter legend key: E/U collected from enriched/unenriched area, D/F watered with distilled water/fertilizer solution, d/t kept in Petri dish/centrifuge tube

22 Establishment and Seedling Growth of Sawgrass and Cattail from the Everglades

557

Fig. 22.4 Maximum percent germination for cattail and sawgrass seeds in various treatments

the very high percent germination of cattail from the enriched area. Fertilization had no significant effect on maximum percent germination, either alone or in interaction. Differences are also apparent in time to initial germination (Fig. 22.5). Species and site effects were highly significant (p < 0.0001). Cattail always began to germinate first, and seeds from the enriched area always began to germinate before those from the unenriched area. The initial germination results, unlike those for maximum germination, showed that the container had no significant effect but fertilization did. The species by site interaction was highly significant, and the other significant interaction was container by water. Fertilization speeded germination in tubes, but slowed germination in dishes. The third parameter considered was time to maximum germination as defined earlier (Fig. 22.6). The results of this analysis were very similar to the results of the analysis of time to initial germination. As was the case in both previous analyses, species and site had highly significant effects (p < 0.0001). Cattail seeds and seeds collected from the enriched area reached maximum germination faster than other seeds. Container had no effect, and fertilization did. The container by water interaction effect was significant, as was the species by water interaction. Fertilization speeded germination in tubes, but slowed germination in dishes. Research has shown that nutrient additions to the maternal parent plant affect offspring germination rate (Wulff 1995); therefore, we expected seeds of plants from the enriched area to germinate more successfully than those from the unenriched area. This was true for all treatments and all three aspects of germination analyzed.

558

S.C. Goslee and C.J. Richardson

Fig. 22.5 Time to initial germination for cattail and sawgrass seeds in various treatments

Fig. 22.6 Time to maximum germination for cattail and sawgrass seeds in various treatments

22 Establishment and Seedling Growth of Sawgrass and Cattail from the Everglades

559

Presumably, when the maternal parent grew in favorable conditions, the seeds had greater stored reserves to use during germination. This may be why fertilization increased germination in cattail seeds collected from the unenriched area. The nutrients in the water helped to compensate for the lack of stored reserves in the seed itself. In the Petri dishes, fertilization sometimes encouraged mold growth, which may have inhibited the rate of germination, although it did not appear to reduce total percent germination.

22.5.4

Chilled Seeds

The overall response was similar to that previously observed for sawgrass (Fig. 22.3). Seeds collected from the enriched area had much higher germination rates than those from the unenriched area (Fig. 22.7). The only apparent differences in germination curves between the chilled and unchilled seeds were the earlier initial germination of the chilled seeds and the lower maximum germination of the chilled seeds from the unenriched area (data not shown). Chilled seeds also tended to reach maximum germination somewhat faster than unchilled seeds. This is most likely the result of time spent soaking rather than of chilling itself. The water in which the seeds were stored turned brown, suggesting that various compounds were leaching from the seeds. Some of these may be germination inhibitors. Chilling is an unlikely dormancy breaker for species native to the Everglades.

Fig. 22.7 Germination curves for chilled sawgrass seeds. See Fig. 22.2 for letter code

560

22.5.5

S.C. Goslee and C.J. Richardson

Dark Germination

There was no apparent effect of light exposure on sawgrass germination (Fig. 22.8). Cattail showed a significant reduction (t-test, a = 0.05) in germination initially. The maximum percent germination for seeds kept in the dark was nearly as high as for those exposed to light. The seeds exposed to light reached that maximum several days earlier. Consistent with earlier results, cattail germinated faster and had a higher total germination than sawgrass. Cattail germination was slightly inhibited by darkness. This agrees with previous studies. Inhibition by darkness probably helps to ensure that cattail seeds that persist in the seed bank do not germinate until conditions are favorable. Since cattail seeds sink, they might otherwise germinate under deep water and die. No evidence for dark inhibition was seen in sawgrass. Sawgrass seeds float initially, but a light requirement may not provide any advantage, since our tests showed no effect of light on germination.

22.5.6

Seedling Growth

Sawgrass seedlings planted in peat were much larger than cattail seedlings after 40 days and significantly larger after 100 days (Fig. 22.9). Fertilization made no difference in either shoot or root length for sawgrass. Fertilized cattail seedlings had longer shoots and roots, and higher survival rates, than cattail seedlings in

Fig. 22.8 Germination curves for sawgrass and cattail seeds with and without exposure to light. Error bars were omitted to increase clarity

22 Establishment and Seedling Growth of Sawgrass and Cattail from the Everglades

561

Fig. 22.9 Changes in mean root and shoot length over time. The zero line is the soil surface; shoots extend upward and roots extend down

unfertilized peat. Sawgrass seedlings placed in tubes of distilled water lived and grew much longer than similarly treated cattail seedlings. Seedling mass behaved similarly (Fig. 22.10). Sawgrass seedlings reached much higher masses than cattail seedlings. The mass of the sawgrass seed itself declined slightly over time as resources were transferred to the seedling. Fertilization made little difference to sawgrass seedling growth. Fertilized cattail seedlings were slightly larger than unfertilized cattail seedlings. Relative growth rate (RGR) was calculated for each seedling according to the formula (Kveˇt et al. 1971; Hunt 1978) RGR = [ln (W2) − ln (W1)]/(T2 − T1) where W1 is the dry mass initially present at time T1 and W2 is the dry mass at the end of the experiment at time T2.

562

S.C. Goslee and C.J. Richardson

Fig. 22.10 Changes in the mass of sawgrass and cattail seedlings over time

RGRs for sawgrass seedlings started out high at 0.25 g g−1 week−1. There was no initial difference between any treatments. The RGR of the seedlings in tubes dropped sharply over the first few weeks, but then increased. The RGR of the seedlings in peat declined more slowly and appeared to level off around 0.02 g g−1 week−1. The RGR of cattail seedlings tended to increase from the initial low level and then also decline. Cattail seedlings generally had a significantly lower RGR than sawgrass seedlings, even after 80 days. The RGR of the cattail seedlings in tubes was highly variable, but tended to be near or below zero. There was little difference in the root:shoot mass ratio of the fertilized and unfertilized sawgrass seedlings (data not shown). The sawgrass seedlings in tubes eventually reached a higher root:shoot mass ratio than the other two sawgrass treatments. Only the fertilized cattail seedlings reached a large enough size to be able to separate the roots and shoots. These seedlings had a higher root-to-shoot ratio than any of the

22 Establishment and Seedling Growth of Sawgrass and Cattail from the Everglades

563

sawgrass seedlings. An increased root:shoot ratio is a common response to low-nutrient conditions, but was not apparent here except in the most extreme conditions. Seed and seedling size are often correlated (Wulff 1995). This was true at the gross level for sawgrass and cattail. Large seeds, with their greater stored resources, may produce seedlings with a higher tolerance for nutrient deprivation (Wulff 1995), a potential advantage for sawgrass in unenriched areas. This was confirmed by the greater growth of sawgrass seedlings in tubes of distilled water. Another advantage for sawgrass with its floating seeds is that if seeds germinate in the water, they may still have time to reach a suitable substrate.

22.6

Conclusions and Lessons for Restoration

Our results support the view of cattail as an opportunistic species of favorable environments (a ruderal–competitor according to Grime 1977). The high seed production and high germination rate of cattail gives this species a remarkable ability to reproduce itself by seed. When combined with an efficient dispersal mechanism and the ability to persist in the seed bank, the large reproductive output of cattail gives it a high probability of being present when a favorable site becomes available. Cattail is not well adapted for life in a low-nutrient environment. Plants in the unenriched area produced fewer and smaller seeds, and those seeds had lower viability than seeds from the enriched area. External fertilization increased the viability of these seeds somewhat, making it more likely that the seeds could become established in a more favorable environment than that of the parent plant. The main limitation on the expansion of cattail may be the fragility of its young seedlings. While the germination requirements of cattail seeds help to ensure that seedlings develop in the most favorable environments, the small size of cattail seeds means that there are almost no resources initially available to the seedling. They respond to external fertilization, but even in the best conditions, cattail seedlings have low growth rates. Adult cattail plants can grow very quickly and reproduce well vegetatively, so the bottleneck is in initial establishment from seed. Sawgrass is a stress-tolerant species (Grime 1977). Low-nutrient conditions had no effect on seed size. Enrichment did increase rate of germination speed and total viability, but not to the extent seen in cattail. The lower reproductive output of sawgrass is balanced by its greater resource allocation to each seed. Sawgrass seeds are larger, so they contain more stored resources and are better able to become established in low-nutrient environments. The correspondingly large seedlings can become established successfully much faster than the tiny cattail seedlings. Sawgrass seedlings were able to persist for weeks in tubes of distilled water, using only the stored resources from the seed. Unlike cattail, the stress-adapted sawgrass seedlings showed no response to fertilization. Cattail has always been present in the Everglades, but it was restricted to scattered locations, highly disturbed areas, and nutrient-rich locations like alligator holes (Davis and Ogden 1994a). The historically low-nutrient conditions prevented cattail

564

S.C. Goslee and C.J. Richardson

from having high reproductive output and extensive spread. Based on these results, two factors are contributing to the spread of cattail and its invasion of new areas. First, disturbance is providing open areas suitable for germination and seedling growth. Second, the high-nutrient inputs enable the seedlings to thrive and eventually to produce hundreds of thousands of viable seeds, which are then dispersed throughout the Everglades to await favorable conditions. This positive-feedback cycle could threaten the Everglades with ever-increasing areas of cattail. Even if nutrient levels could be instantaneously returned to their former low state, the now enormous seed bank of cattail will allow the continued increase of cattail any time a favorable site becomes available. Finally, cattail appears to be more tolerant of the current altered hydrologic regime than sawgrass and more able to take advantage of high-nutrient conditions. These functional species advantages, coupled with the vast acreage of cattails being cultivated in the nutrient-rich Stormwater Treatment Areas, provide an enormous seed source. Without major changes in water and nutrient management in the Everglades, large expanses of cattail are here to stay, and its removal as a dominant species in enriched areas will be nearly impossible.

Part IV

Modeling Ecosystem Responses to Phosphorus Additions

23

Long-Term Phosphorus Assimilative Capacity (PAC) in the Everglades Curtis J. Richardson and Song S. Qian

23.1

Introduction

Wetlands have been shown to have some capacity to retain and store phosphorus, although it has been clearly demonstrated their efficiency is reduced as P loads increase (Qian and Richardson 1997b; Richardson and Qian 1999). Peatlands have a lower capacity to retain P than mineral-based wetlands (Richardson 1985, 1999). Nutrient storage capacity, often a design feature of constructed wetlands, is defined as the total mass per unit area that can be retained permanently by the system. However, high P-loading levels may result in significant ecosystem changes (e.g., altered community structure and diversity as well as increased productivity) and increases in downstream P and N output concentrations (Richardson et al. 1997a; Richardson and Qian 1999). One key question concerns the ability of wetlands like the Everglades to assimilate and store P without causing significant changes in the flora and fauna within the ecosystem. This chapter presents an updated version of our earlier P assimilative capacity analysis (Qian and Richardson 1997b; Richardson and Qian 1999) and relates our phosphorus assimilative capacity (PAC) concept with current Stormwater Treatment Area (STA) P loadings and release concentrations. Noe et al. (2001) have suggested that the Florida Everglades is extremely sensitive to small increases in P concentrations. However, this does not mean that the Everglades has no P assimilative capacity. Here, we define P assimilative capacity of a wetland as the long-term mass removal capacity per unit area that is transformed and absorbed into the system with no significant ecosystem changes in internal structure or function and no downstream output of the nutrient. This removal capacity is in addition to the nutrients received from rainfall. Thus, two hypotheses could be proposed regarding the additions of low levels of P to the Everglades. First, it could be postulated that continued P additions to the Everglades, even at low load levels (i.e., at concentrations just above background concentrations), will eventually result in significant changes in ecosystem structure and function. Alternatively, it could be postulated that the Everglades has the capacity to assimilate and store P at some elevated P threshold load (i.e., P concentrations above background) that will not result in a significant change in ecosystem structure and function. The first hypothesis suggests that 567

568

C.J. Richardson and S.S. Qian

the ecosystem has no capacity to assimilate P above background levels since any P additions will affect all the trophic levels eventually. The second hypothesis indicates that the ecosystem has a capacity to take up and permanently store some P loadings without significant alterations to the ecosystem. This chapter reviews data and results for a number of studies in the US along with data from a long-term nutrient-gradient study in the Everglades to assess the concept of PAC in the Everglades.

23.2

Methods

Qian and Richardson (1997b) earlier analyzed data from a large number of wetlands throughout the United States to assess the effects of different P loadings on long-term storage rates and effluent concentration patterns. Data from a large number of wetlands throughout the United States were reviewed to assess the effects of different P loadings on long-term storage rates and effluent concentration patterns. The North American Wetlands for Water Quality Treatment Database (NAWDB; Knight et al. 1994) is a United States Environmental Protection Agency effort to collect and summarize the effectiveness of using wetlands, both constructed and natural, as a low-cost alternative for removing pollutants (Knight et al. 1994; Kadlec and Knight 1996). Cross-sectional data sets are appealing because they include a range of responses to nutrient inputs. This means that empirical evidence discovered from these data is likely to have a relatively broad inference base. The NAWDB includes input and output phosphorus concentrations (Pin and Pout, in mg l−1), hydraulic loading rate (qs in cm d−1), treatment area (in ha), and areal input and output P mass loading rates (Lin and Lout, in g m−2 year−1). A piecewise linear model proposed earlier by Reckhow and Qian (1994) was used by the authors to do exploratory analysis and to prepare a nonparametric regression model fit with the data. A Bayesian changepoint detection method (Stephens 1994) was applied to a piecewise linear model of the data, and we developed a procedure for estimating assimilative capacity when site-specific data are available (Richardson 1997). This method was applied to data from the northern Everglades to develop a single-wetland “changepoint” curve. The probability of overloading the Everglades at a given P-loading rate was calculated, and a nonparametrically fitted model of P outflow concentration as a function of P-loading rate was developed (Qian and Richardson 1997b; Qian and Reckhow 1998). To further assess the relationship of P loading to assimilative capacity, we collected and analyzed data on water quality, developed indices for P availability, assessed populations of macrophytes and macroinvertebrates, and measured P storage in soils for 6 years along a P gradient in an area of the northern Everglades (Craft and Richardson 1993b; Walker 1995; Richardson 1997; Rader and Richardson 1994).

23 Long-Term Phosphorus Assimilative Capacity (PAC) in the Everglades

23.3 23.3.1

569

Results PAC Concept

Analysis of the NAWDB (Knight et al. 1994), a cross-sectional data set for 126 natural and constructed wetlands throughout the United States, made evident the existence of wetland PAC (Knight et al. 1994; Reckhow and Qian 1994; Stephens 1994) (Fig. 23.1a). Piecewise linear regression modeling (Reckhow and Qian 1994; Qian and Richardson 1997b) suggests that a significant increase in total phosphorus (P) effluent concentrations will occur from wetlands receiving a mass loading of P much above 1 g m−2 year−1. Analysis of the NAWDB effluent trends reveals that when P loadings to wetlands are kept below 1 g m−2 year−1, P output concentrations remain fairly constant and low (geometric mean of 21 ± 3.3 µg l−1 P) (Fig. 23.1b). Based on an analysis of residual sums of squares, the optimal value for the changepoint (see changepoint zone between regions 1 and 2 on Fig. 23.1a) lies between 0.4 and 1.4 g m−2 year−1, which is compatible with the long-term P accumulation rate reported for wetland ecosystems (Richardson and Marshall 1986; Craft and Richardson 1993b; see Chap. 3). The variations in P effluent output around the changepoint suggest that PAC is a site-specific quantity. Therefore, we postulated that the long-term PAC for wetlands does not exceed 1 g m−2 year−1 P (Richardson and Qian 1999). This assimilative capacity has a 95% credible interval of 1.4–0.4 g m−2 year−1 since this region of the model (region 1 on Fig. 23.1a) maintains stable outflow P concentrations (see expanded scale Fig. 23.1b). The “one gram” level, however, represents an overall North American average against which site-specific assimilative capacity should be estimated when appropriate data are available. Moreover, the “one gram” level should be considered an upper P-loading value for assimilative capacity that sustains natural structure and functions in the wetland. Kadlec (1999) suggested that the central tendency of the cross-sectional data set of the NAWDB cannot be used to draw general conclusions about specific ecosystem behavior. This is true if it is only used as the final model to estimate particular parameters for specific sites. The exact changepoint is a case of such a parameter. We suggest that the Bayesian changepoint method can be used with a piecewise linear model to develop specific parameters for each wetland type. We used this changepoint detecting method utilizing both NAWDB and an independent data set from the Everglades to estimate PAC, i.e., the probability of overloading Everglades ecosystem processes and increasing P effluent above background outputs (Richardson and Qian 1999). The probability analysis suggests a 90% chance of overloading at 1.4 g m−2 year−1, a 40% chance at 1.0 g m−2 year−1, and a 10% chance at 0.7 g m−2 year−1 (see solid line, Fig. 23.1b). These results suggest that the Everglades data support the “one gram P assimilative capacity rule.” Important here is the understanding that the confidence intervals (0.4–1.4 g m−2 year−1) be included in any assessment of the 1 g concept since this defines the probable range of expected capacity for the system that has to be tested with specific ecosystem response variables.

570

C.J. Richardson and S.S. Qian

Fig. 23.1 (a) Input total phosphorus (P)-loading effects on P output concentrations for the North American Wetlands for Water Quality Treatment Database (NAWDB; Knight et al. 1994). Total sites are 126, with data collected over several years, n = 317. In region (1) where loading rate is less than 1 g m−2 year−1, uniform P output concentrations are found (i.e., baseline P output) and output concentration is not a function of the loading rate; while in region (2) the P-loading rate is larger than 1 g m−2 year−1, and output concentrations increase significantly as the loading rate increases. The variation in the output concentrations is large and nonuniform. The changepoint in loading rate is defined as the loading rate value that divides output P concentrations into uniform and nonuniform regions (from Richardson and Qian 1999). Reprinted with permission from Richardson and Qian: “Long-term phosphorus assimilative capacity in freshwater wetlands: A new paradigm for maintaining ecosystem structure and function.” Environmental Science and Technology 33(10):1545–1551. Copyright 1999 American Chemical Society. (b) An expanded scale of the NAWDB better displays the increase in outflow variation that occurs at P loadings between 0 and 5 g m−2 year−1. The curved line indicates the relationship between the P-loading rate and the probability (or risk) of the system being overloaded. The risk–loading relationship was

23 Long-Term Phosphorus Assimilative Capacity (PAC) in the Everglades

571

To further test our hypothesis and separate P storage capacity and PAC, we first calculated P mass loadings and related surface water P and soil P accretion rates to P availability as indexed by measurements of phosphatase activity in water (Wetzel 2001) and N:P ratios in plants (Koerselman and Meuleman 1996). Changes in plant community structure, plant productivity, and macroinvertebrate diversity were then used to estimate change in ecosystem structure and function along a 10-km P gradient in the Everglades. Phosphorus-loading rates ranged from 4.0 g m−2 year−1 in the highly enriched dense cattail zone nearest the Hillsboro discharge canal to background P loadings of 0.3 g m−2 year−1 in the most unenriched sawgrass region 10-km downstream from the inputs (Fig. 23.2). Surface water P concentrations and soil P

Fig. 23.2 The relationship between surface water total phosphorus and mean soil P accumulations along a eutrophication gradient (Transect C) in WCA-2A in the central Everglades of Florida. Estimated phosphorus mass loadings over the past 26 years are shown for set distances from the Hillsboro Canal, and are calculated following a P retention model developed for the Everglades wetland (Walker 1995). The model was run successively to calculate the amount of P stored in each zone, and the remaining P was used as the load to the downstream zone. The degree of enrichment and amount of impact are shown for each zone and are based on the amount of P loadings, elevated water column P concentrations, and changes in macroinvertebrate and plant community structure (Richardson 1997; Rader and Richardson 1994; Craft and Richardson 1997). Vertical dashed lines show the 95% confidence intervals (1.4–0.4 g m−2 year−1) around the 1 g m−2 year−1 P threshold (from Richardson and Qian 1999). Reprinted with permission from Richardson and Qian: “Long-term phosphorus assimilative capacity in freshwater wetlands: A new paradigm for maintaining ecosystem structure and function.” Environmental Science and Technology 33(10):1545–1551. Copyright 1999 American Chemical Society Fig. 23.1 (continued) developed for an Everglades data set, using a Bayesian changepoint estimation method (Stephens 1994; Qian and Richardson 1997b). The dashed line is extended across the NAWDB since the probability of exceeding baseline P output is equal to 1 at loadings above 1.5 g m−2 year−1 (from Richardson and Qian 1999). Reprinted with permission from Richardson and Qian: “Long-term phosphorus assimilative capacity in freshwater wetlands: A new paradigm for maintaining ecosystem structure and function.” Environmental Science and Technology 33(10):1545–1551. Copyright 1999 American Chemical Society

572

C.J. Richardson and S.S. Qian

accretion rates are highly correlated (r = 0.94, p < 0.05), and both decrease exponentially with distance from the input structures (Fig. 23.2). However, the highly enriched area (zone 0–2 km) that received the highest average P loadings (2.7 g m−2 year−1) maintained water column concentrations >100 µg l−1 P (Fig. 23.2). The region of the moderately enriched zone (2–5 km) receiving P loadings above 1 g m−2 year−1 had water column concentrations from 60 to 100 µg l−1 P, well above the average P output concentrations of 21 µg l−1 P displayed by the NAWDB wetlands (Fig. 23.1a, b). It was not until P loadings decreased to below 1 g m−2 year−1 (6–8 km) in the nonimpacted zone that P concentrations reached the mean baseline output concentrations of 21 µg l−1 P reported for the NAWDB (Fig. 23.1a, b). However, the Everglades ecosystem, like many freshwater wetlands, is P limited (Richardson et al. 1999). This causes P concentrations in the water column to be further reduced to a mean background level of 10.4 ± 1.3 µg l−1 in the interior of the Water Conservation Area 2A (WCA-2A) (Richardson et al. 1997a; Qian and Richardson 1997b). Highest P storage capacity in the wetlands is found with increased P loading near the inputs (0–0.7 km) at decreased P removal efficiency (Richardson et al. 1997a; see Chaps. 3 and 6) (Fig. 23.2). These increased loadings above the PAC for the wetland resulted in highest downstream P concentrations and ecosystem changes in the high P-loading zone (Chaps. 2 and 9).

23.3.2

Indices of Ecosystem Response

Two biochemical indices of P availability, phosphatase activity (APA) in the water column and molar N:P ratios in sawgrass (Cladium jamaicense), revealed that P limitations only existed below the 1 g P m−2 year−1 loading threshold (Fig. 23.3a, b). In the region above the 1 g P m−2 year−1 loading, APA enzyme activity per unit of chlorophyll was almost undetectable. By comparison, the region below the 1 g P m−2 year−1 loading displayed significant APA activity in both the filtered and unfiltered samples (Fig. 23.3a). The magnitude of increase in the molar N:P ratios found in sawgrass leaves generally followed the increase in APA activity in the water column (Fig. 23.3a, b). Interestingly, the plant N:P ratios reach their maximum values near the 0.4 g P m−2 year−1 loading rate, but APA activity increased until water column P concentrations reached their lowest concentrations along the gradient (Fig. 23.3b). Both indices of ecosystem P limitations showed their highest values below the 1 g P m−2 year−1 threshold but closer to the 0.4 g P m−2 year−1 loading rate. Detailed measurements of plant community structure, productivity, and macroinvertebrate diversity along this eutrophication gradient caused by nearly 30 years of nutrient-rich agricultural runoff in the northern Everglades revealed a shift in the dominant plant species community from open sloughs and sawgrass (C. jamaicense) to cattail (Typha domingensis) in the area near the discharge point (0–4 km, Fig. 23.4a). The impacted zone was also the area of highest P loadings, water column TP, and maximum P accumulation (Fig. 23.2). In addition, this region

23 Long-Term Phosphorus Assimilative Capacity (PAC) in the Everglades

573

Fig. 23.3 (a) Water column phosphatase activity (APA) per unit of chlorophyll along a nutrient gradient in the northern Everglades (Transect C, Water Conservation Area 2A – WCA-2A). Phosphatase was normalized over the amount of chlorophyll found per milliliter. Both filtered and unfiltered APA are shown to clarify the amount of activity due to varying densities of particles vs. bacteria and algae in the water column. Methods followed the 4-methylumbelliferyl phosphate technique of Wetzel (2001). Vertical lines show the 95% confidence intervals (1.4–0.4 g m−2 year−1) around the 1 g m−2 year−1 P threshold (from Richardson and Qian 1999). (b) The molar N: P ratio for sawgrass leaves along a nutrient gradient in the northern Everglades (Transect C, WCA2A). Methods of analysis for N and P followed Craft and Richardson (1997) and Verhoeven et al. (1996). Vertical dashed lines show the 95% confidence intervals (1.4–0.4 g m−2 year−1) around the 1 g m−2 year−1 P threshold (from Richardson and Qian 1999). Reprinted with permission from Richardson and Qian: “Long-term phosphorus assimilative capacity in freshwater wetlands: A new paradigm for maintaining ecosystem structure and function.” Environmental Science and Technology 33(10):1545–1551. Copyright 1999 American Chemical Society

had the highest loadings of other ions, including nitrogen, calcium, and sodium (Qualls and Richardson 1995; Craft and Richardson 1997; see Chaps. 3 and 6). The lack of change in plant community structure from 1990 to 1996 beyond 5.1 km (Fig. 23.4a) and the low P concentrations in the downstream water column initially

574

C.J. Richardson and S.S. Qian

Fig. 23.4 (a) Plant species composition at six locations downstream from the Hillsboro Canal in the central Everglades. Cattail (Typha domingensis) is the dominant species within the first 3.5 km. A sawgrass (Cladium jamaicense) monoculture has existed for 5 years (1990–1996) at the 5.1-km location and at more distant plots even after 30 years of P loadings. Plant species composition was measured in permanent plots at sampling locations in the central Everglades in June 1990, 1992, 1994, and 1996 to verify potential changes in cover type and assess the changes reported in the impacted and nonimpacted zone since 1973 (Richardson et al. 1997; Jensen et al. 1995). Species composition was determined using the point intercept method and expressed as percent frequency of occurrence along Transect C in WCA-2A (Mueller-Dombois and Ellenberg 1974). Vertical dashed lines show the 95% confidence intervals (1.4–0.4 g m−2 year−1) around the 1 g m−2 year−1 P threshold (from Richardson and Qian 1999). (b) Net annual primary productivity was estimated at six locations in WCA-2A. Data were recalculated from Davis (1989). Vertical dashed lines show the 95% confidence intervals (1.4–0.4 g m−2 year−1) around the 1 g m−2 year−1 P threshold (from Richardson and Qian 1999). Reprinted with permission from Richardson and Qian: “Longterm phosphorus assimilative capacity in freshwater wetlands: A new paradigm for maintaining ecosystem structure and function.” Environmental Science and Technology 33(10):1545–1551. Copyright 1999 American Chemical Society

23 Long-Term Phosphorus Assimilative Capacity (PAC) in the Everglades

575

supported our hypothesis that wetlands can fully assimilate P at loadings of up to 1 g m−2 year−1 without significant change in this ecosystem component. The nonimpacted zone (>5.1–10.5 km) had maintained the dominant native sawgrass communities (>89% frequency) and 18 out of 19 macrophyte and slough species in the first 6 years of our study (Vaithiyanathan and Richardson 1999). However, over the next 6-years cattail frequency increased down the transects to a distance of nearly 9 km, although the frequencies at this distance were less than 5% (see Fig. 9.5 in Chap. 9). These longer-term data suggest that P loading must be maintained closer to the lower boundary of 0.4 g P m−2 year−1 to maintain maximum community species integrity in the Everglades. Sawgrass plant productivity in the impacted zones (0–5.1 km) was nearly double that of the nonimpacted area (Fig. 23.4b; see Chap. 6). Plant frequency, however, was greatly reduced for this more desirable species above a 1 g m−2 year−1 loading (Fig. 23.4a). Macroinvertebrate diversity and number of taxa were highest in the P-enriched areas (P loadings 1.6–4.0 g m−2 year−1) but were reduced to background levels at P inputs (<1.0 g m−2 year−1) (Fig. 23.5) (Rader and Richardson 1994). The relative distribution of taxa within functional feeding groups was similar along the gradient (Rader and Richardson 1994). However, the mean annual density of benthic invertebrates was 7.4 times greater in the most enriched sites compared to nonenriched sites (Rader and Richardson 1994).

Fig. 23.5 Macroinvertebrate Shannon–Wiener diversity and number of taxa were measured seasonally at six locations in WCA-2A from 1990 to 1992 using a D-frame sweep net and benthic cores (Rader and Richardson 1994). Vertical dashed lines show the 95% confidence intervals (1.4–0.4 g m−2 year−1) around the 1 g m−2 year−1 P threshold (from Richardson and Qian 1999). Reprinted with permission from Richardson and Qian: “Long-term phosphorus assimilative capacity in freshwater wetlands: A new paradigm for maintaining ecosystem structure and function.” Environmental Science and Technology 33(10):1545–1551. Copyright 1999 American Chemical Society

576

23.4

C.J. Richardson and S.S. Qian

Discussion

Collectively, our data suggest that the Everglades ecosystem responses and water quality support the P assimilative concept that emerged from our original analysis of the North American Wetland Database. Low P-loading rates and low water column P concentrations exist below the P threshold zone after 30 years of nutrientrich runoff with little or no significant change in ecosystem structure and function, but only at the lower loading boundary of 0.4 g m−2 year−1 after 12 years (Figs. 23.3a, b, 23.4a, b, 23.5, and 9.5). Specifically, the long-term maintenance of the low P-availability region after three decades of P additions has resulted in no significant changes in plant community composition, invertebrate diversity, or invertebrate taxa within functional family groups in the unenriched areas. However, the highest total number of invertebrate taxa, biomass, and density were found in the enriched areas with high plant productivity (Rader and Richardson 1994). It should be noted that WCA-2A is the most highly nutrient-impacted area of the Everglades, and these long-term loading effects may not fully represent community responses in the more pristine areas of WCA-3A or the ENP. In addition it has been shown that the large amount of stored P in the soil in the enriched zone may diffuse into the water column when waters with lower surface P concentrations are introduced as a result of STAs and BMPs (Chap. 6). This will result in an efflux of P into the water column and a downstream movement of P into the less enriched zones. The nutrient-limitation criteria developed by Koerselman and Meuleman (1996) and Verhoeven et al. (1996) suggest that plants with molar N:P ratios >36 are P limited, while those below 31 are N limited (Fig. 23.3b). (Note: The plant molar ratio of 36 equals an equivalent weight-base ratio of 16.) Values above 36 are only found below the 1 g loading zone, a region with elevated APA activity. Both indices suggest that P limitation, a condition found in the most undisturbed area of the Everglades, exists below the P threshold loading zone. Related mesocosm work on responses of the Everglades plant community to varying P dosing concentrations also indicates that annual average P concentrations must be maintained <15 µg l−1 to prevent significant decreases in periphyton mat, macrophyte density, and algal communities (Richardson et al. 2007; see Chap. 25). This water quality concentration is found only near the lower end of the 1 g m−2 year−1 threshold at 0.4 g m−2 year−1 of loading (Fig. 23.3b). We find that once P loadings saturate short-term uptake and storage mechanisms (i.e., in the highly enriched zone), a downstream moving front of elevated P concentrations exists in the wetlands until P loadings are reduced to the PAC of the wetland (Figs. 23.1a, b and 23.2). We thus propose that P loadings into the Everglades be maintained below 1 g m−2 year−1 if long-term storage of P, maintenance of native plant and invertebrate species, and low P effluent concentrations are the goal. Of importance is the fact that the “one gram rule” takes into account the probability that the system will be overloaded with the analysis suggesting a 90% chance of overloading at 1.4 g m−2 year−1, a 40% chance at 1.0 g m−2 year−1, and only a 10% chance at 0.7 g m−2 year−1 (see solid line, Fig. 23.1b). This type of analysis

23 Long-Term Phosphorus Assimilative Capacity (PAC) in the Everglades

577

gives managers the ability to reduce their risk by selecting a loading rate that will not exceed the PAC. Thus, a very conservative loading rate of 0.4 g m−2 year−1 would have <1% probability of exceeding the PAC. This rule may not hold if high continuous loadings of Fe-, Ca-, Al-, or P-binding sediment are added to the water entering the wetland and metal precipitation becomes the main P removal mechanism (Hammer 1989; Cooke et al. 1992).

23.5

Conclusions and Lessons for Restoration

Statistical analysis of a North American Wetland Database (NAWDB) allowed us to develop a mass loading model which was used to separate P assimilative capacity (defined as P absorption with no significant ecosystem change and no elevated P output) from storage capacity (maximum storage) in wetlands. Our analysis indicates that, given ample supplies of other nutrients, average PAC in North American wetlands is near 1 g m−2 year−1 with a confidence interval from 1.4 to 0.4 g m−2 year−1. From this analysis we earlier proposed a “one gram assimilative capacity rule” (Richardson and Qian 1999) for P loadings within natural freshwater wetlands if long-term storage of P, maintenance of community structure and function, and low P effluent concentrations are required. However, the Everglades is a P-limited ecosystem, and biological response data suggest that the lower end of the PAC confidence interval (0.4 g m−2 year−1) be used to assure no significant changes in ecosystem structure and function. Our research supports our hypothesis that the Everglades has the capacity to assimilate and store P at some elevated P threshold load (i.e., with P concentrations slightly above background levels) that will not result in a significant change in ecosystem structure and function. However, it will lose native species, become P saturated in a few years, and export unacceptable amounts of phosphate downstream when phosphorus loading exceeds PAC. But these findings clearly demonstrate that even P-limited wetlands like the Everglades have the capacity to assimilate low levels of P loadings without significant changes in ecosystem structure and function. Importantly, whether P is lost at high concentrations from a wetland depends on the size of the wetland compared to the P loadings. Our studies suggest that all wetlands have the capacity to store some P without significant changes in downstream P concentrations or community changes. If the wetland is of sufficient acreage to decrease downstream P loadings at or below its assimilative capacity, then a high percentage of the P will be retained within the wetland and little or no moving P front in the water column will exist, as the system has reached equilibrium with P additions. This has been shown to be the case for some Everglades areas once loadings are reduced (Chap. 6). However, a P gradient will exist in the Everglades regardless of the concentration of P allowed to enter the Everglades (e.g., 30 or 10 µg l−1) since the planned release of P will be loaded to a small restricted area downstream of the outflow gates or structures, i.e., the load is area dependent and is based on the volume of

578

C.J. Richardson and S.S. Qian

water released times the concentrations of P in the water (Lowe and Keenan 1997). For example, an impacted zone (0–5 km) is found in the region receiving 4.0 to about 1.0 g m−2 year−1 of P loadings, and the nonimpacted area is only found below the PAC (Fig. 23.2). A reduction of P loadings by 50% into a WCA would result in a smaller impact zone provided the volume of water is not increased significantly, which would increase mass P loading even with low concentrations. Importantly, the sizing of the wetlands necessary to remove P loadings to near background levels can be estimated from a model of the relationship of mass P storage inputs to water column P inputs (Qian and Richardson 1997b): A=CiQ/Pli, where A is the total storage area (m2), Ci the P input concentrations (µg l−1), and Pli the P-loading inputs (g m−2 year−1). The equation has been further developed and tested in a paper by Lowe and Keenan (1997). This approach was used to size the STAs that are being used to reduce P loadings into the Everglades (Walker 1995). Of importance is the recent finding that STAs that have received P loadings in excess of 1.0 g m−2 year−1 (1.5–4.2 g m−2 year−1) in the past few years (2003–2005) are releasing annual flow-weighted mean downstream P concentrations that range from 26 (STA-6 in 2003) to 136 µg l−1 of TP (STA-5 in 2003), values far in excess of the proposed standard of 10 µg l−1 of TP (SFWMD 2004, 2005, 2006). Moreover, in 2004 and 2005 all the STAs averaged together each year received a TP loading of 2.0 and 2.2 g m−2 year−1, respectively, and released 41 µg l−1 of TP each year. By contrast, STA-2 received a loading rate of 0.90, 0.90, and 1.90 g m−2 year−1 in 2003, 2004, and 2005, respectively, and correspondingly released 17, 14, 20 µg l−1 of TP. The lowest concentration rate of 12 µg l−1 of TP released during the 3-year period was from STA-6 in 2004, with a loading of 1.0 g m−2 year−1, however the site only received 3.4 MT of loading that year. These data suggest that the “one gram rule” (Richardson and Qian 1999) is applicable to STA design criteria and that sites that are greatly overloaded will continue to release high concentrations of P (see Table 2.2 in Chap. 2). In summary, the total area of current STAs are not sufficient to handle the present loadings and more treatment areas will need to be constructed, a conclusion reached much earlier by Richardson and Craft (1993). To attempt to reach the federally mandated and state-approved 10 µg l−1 of TP, the amount of STA area will need to be greatly expanded from the current 16,500 to 43,000 ha or possibly as high as 67,000 ha if long-term soil P storage rates only average 0.40 g m−2 year−1 as found by Craft and Richardson (1993b). The key to meeting the P water criteria is to provide sufficient filtering areas so that yearly loadings of TP to all STAs remain close to or below 1 g m−2 year−1 even in extremely high loading years.

24

Spatial Distributions of Total Phosphorus and Phosphorus Accretion Rates in Everglades Soils Song S. Qian and Curtis J. Richardson

24.1

Introduction

Historically, the Everglades was a phosphorus-limited ecosystem, with the supply of phosphorus primarily coming from rainfall (Davis 1943; see Chap. 2). After the turn of the twentieth century, the US Army Corps of Engineers built a system of drainage canals and dikes that made possible the subsequent development of much of the northern Everglades for agricultural use (Chap. 8). As a consequence, ecological changes have occurred, especially in the central part of the Everglades south of the agricultural development. The most significant change is the shift of dominant plant species from sawgrass (Cladium jamaicense) to cattail (Typha domingensis). To evaluate the phosphorus effect, several researchers collected soil samples to analyze the soil total phosphorus (STP) concentration and phosphorus accretion rate (PAR; Reddy et al. 1991, 1994a,b; Richardson et al. 1992; Craft and Richardson 1993a,b). The STP spatial distribution can help us to understand the effect of agricultural runoff on the Everglades and to study how the ecosystem responds to the elevated phosphorus concentration in the soil. PAR is an important measure of ecosystem change. As discussed in Richardson and Qian (1999) and Qian and Richardson (1997), the long-term PAR represents the assimilating capacity of a wetland, which is an indicator of ecosystem changes. When the long-term accretion rate exceeds a level of about 1.0 g m−2 year−1, it is believed that the phosphorus loading rate is higher than the assimilative capacity and significant changes might occur when short-term storage capacity is saturated. In this chapter, we develop models of spatial distribution of STP using data from the three Water Conservation Areas (WCAs) in central Everglades collected by the aforementioned authors and also data collected by the Duke University Wetland Center from the northern part of the Everglades National Park (ENP). In addition, we present the spatial distribution of PAR using a relationship between PAR and STP developed using nonparametric regression. We first discuss the data, and then present the statistical methods we used. The results are presented in two parts: one on the relationship between PAR and STP, the other on spatial distributions of STP and PAR.

579

580

24.2

S.S. Qian and C.J. Richardson

Study Area and Data

Our study area consists of three WCAs in the northern and central part of the Everglades and the northern part of the ENP in south Florida. The four parts of our study area are separated by levees and canals, and water is distributed through a network of pump stations and canals. In general, water from the Everglades Agriculture Area flows into the Water Conservation Area 1 (WCA-1) also known as the Loxahatchee National Wildlife Refuge. Outflow from WCA-1 is combined with agriculture runoff in Hillsboro Canal, then diverted into Water Conservation Area 2A (WCA-2A) through three main gates on Hillsboro Canal. From WCA-2A, water flows to Water Conservation Area 3A (WCA-3A) via the Canal L-38 through five gates. And three main gates divert water from WCA-3 to ENP. Water levels in the three WCAs are managed to provide water supply, flood control, recreation, and wildlife preservation (SFWMD 1992). Soil phosphorus data used in this paper were collected by Reddy et al. (1991, 1994a,b) and Richardson et al. (1992). Reddy’s team collected 90 soil samples from WCA-1, 74 from WCA-2A, and 200 from WCA-3 in 1990. They measured and analyzed STP concentration for the top 0–10 and 10–20 cm soil layers (Bruland et al. 2006). The Reddy’s team also collected 12 soil samples from WCA-2A for measuring PAR and STP. In 1991, the Duke University Wetland Center collected 25 soil samples from WCA-2A for measuring STP, and 18 of these samples were also used for measuring PAR. The Duke team also collected six samples from WCA-1 and 24 samples from ENP. All together, we have 419 soil samples of STP, and among them, 68 samples have PAR measurements. Analyses of these data (except the ENP data from Duke) have been published elsewhere (e.g., Reddy et al. 1993; Richardson and Craft 1993; Richardson et al. 1992; Qian 1997a; Bruland et al. 2006). Some of these data were used in modeling studies (Walker 1995; Qian 1997b).

24.3

Predicting PAR Using Soil Total Phosphorus Concentration

PARs are measured using a Cesium (137Cs) dating technique. Rates of peat accretion are determined by measuring 137Cs in soil depth increments. The 137Cs maximum in the soil profile corresponds to the soil surface in approximately 1964, the period of maximum deposition of 137Cs from aboveground thermonuclear weapons tests. From the peat accretion rate and measured total phosphorus (TP) concentration in the same layer, PAR is determined. This procedure is time consuming and expensive. It is therefore desirable to predict PAR based on the top 10 cm STP concentration, since STP is much easier and cheaper to measure. According to studies of the Everglades soils (Craft and Richardson 1993b; Reddy et al. 1993), rooted vegetation

24 Spatial Distributions of Total Phosphorus and Phosphorus Accretion Rates

581

activities keep available phosphorus in the top 5–15 cm of the soil. Therefore, we expect to see a positive correlation between the top 10-cm STP and PAR. However, there is no theoretical support for a functional relationship between STP and PAR. We approach this problem by using a nonparametric regression technique, the locally weighted regression or “loess” (Cleveland 1993), which enables us to pick out the most likely functional relationship based on the data we have. The term “nonparametric” should not be confused with the ordered statistical analysis such as the Wilcoxson test. Let us consider the situation where we want to describe a response variable Y as a function of the predictor variable X, or Y = f (X) + e.

(24.1)

Two extremes of the function f may be defined. At one extreme, we may simply connect all points, moving from left to right along the x-axis. This model has all the roughness of the original data, and it estimates each data point using only that data point (hence no smoothing and no estimation error). At the other extreme, we may fit a linear model to the data using ordinary least squares regression. This model has no roughness, and all data points contribute equally to the prediction at any one point (with error based on the squared deviation between each point and the regression line). A nonparametric regression approach is in between these two extremes, i.e., prediction at a given point is based on data points in a local neighborhood, and the closer to the point to be predicted a data point is, the higher its contribution to the prediction will be. As a consequence, nonparametric regression can be used to describe locally persistent patterns in the data. Details of nonparametric regression can be found in Eubank (1988) for smoothing splines, Härdle (1990) for kernel smoothers, and Hastie and Tibshirani (1990) for a class of models called generalized additive models. Important concepts in developing a nonparametric model are the balance of the roughness of a model and the accuracy of the model prediction (Green and Silverman 1994). The relationship between the top 10-cm STP and PAR is strong (Fig. 24.1). When STP is less than about 300 mg kg−1 (mostly from ENP), the fitted model is almost a flat line with high variability, in other words, an average of about 0.1 g m−2 year−1 can be used as the estimate of PAR. This is expected since the 137Cs method can only measure soil accretion accurately in 1-cm increments. When the soil accretion rate is low, we expect the 137Cs method to have very high relative error. For all the soil samples there are measurements of soil bulk density (based on dry mass). With bulk density it is easy to calculate the volume-based soil TP concentration measured in grams of TP per m3 of soil. However, we think that the measurements of soil volume are not as reliable as the measurements of dry mass, since during the sampling process the soil may be compressed slightly. For this reason, we use the weight-based STP as the predictor of PAR. The solid line in Fig. 24.1 indicates that this relationship between STP and PAR is similar for the WCA-1, WCA-2A, WCA-3A, and ENP. From this nonparametrically fitted relationship, we may further simplify the model by using two joint straight

582

S.S. Qian and C.J. Richardson

Fig. 24.1 Relationship between soil’s total phosphorus concentration (mg kg−1) and total phosphorus accretion rate (g m−2 year−1). Data were collected throughout the study region. The relationship is fitted by using a nonparametric regression model (solid line) and a piecewise linear model (dashed line)

lines (or a piecewise linear model, the dashed line in Fig. 24.1). When STP is less than 326.5 mg kg−1, we use a flat line (at 0.065 g m−2 year−1), and when STP is greater than 326.5, we use a line with a positive slope (the number 326.5 is selected such that the model has the smallest residual sum of squares) log2 (PAR + 1) = 0.091 if STP £ 326.5 mg kg–1 log2 (PAR + 1) = a + b • STP > 326.5 mg kg–1

(24.2)

where a = −0.0810 and b = 5.264 × 10−4. This model can be used to make quick estimate of PAR. (Note that we used log2 transformation on PAR, such that a change of one unit in the transformed scale means a change on the original scale by a factor of 2.) The difference between this piecewise linear model and the nonparametric regression model is small (Fig. 24.1). When accurate prediction is needed, the nonparametric model can be used. In our case, the model was fitted using the “loess” function from the statistical software package S-Plus (MathSoft 1997).

24 Spatial Distributions of Total Phosphorus and Phosphorus Accretion Rates

583

Using S-Plus, predictions can be made once the nonparametric model has been fitted.

24.4

Spatial Distribution of Soil Phosphorus in the Everglades

Since the subregions of this study are isolated from each other, we present the spatial distribution of soil’s phosphorus separately. In studying spatial distributions, kriging (see, e.g., Cressie 1991) is the most commonly used statistical method. However, kriging requires that data to be analyzed are from a stationary process. Ordinary kriging requires that the mean of the data should be the same everywhere in the region. In the Everglades situation, this requirement is not met, since there are sources of phosphorus and the mean STP is higher in regions closer to the sources. When major sources of phosphorus can be identified, it is possible to fit a universal kriging model using the distance to the closest source as a covariate (Qian 1997a) or to use Bayesian kriging. In this study, we choose to use a nonparametric regression method instead of kriging for two reasons. First, it is necessary to identify the correct spatial correlation model for using universal kriging, but there is no guidance for doing so. As discussed in Qian (1997a), using an incorrect correlation model will affect the kriging result significantly. Second, the computational cost for Bayesian kriging, an alternative to the universal kriging, is too high. When comparing contour maps produced by using various spatial interpolation methods (such as various kriging methods and nonparametric regression), there may be no visible difference. However, when the contour maps are used to support further analysis (e.g., estimating the size of the area affected by the runoff), significant differences may arise. By using nonparametric regression models, we can obtain the spatial distribution model without imposing some hard-to-justify assumptions.

24.4.1

Method

The nonparametric regression method we used is the two-dimensional version of loess, which is implemented in S-Plus. Details of loess can be found in Cleveland (1993) and MathSoft (1997). We fit the following model for each of the four subregions in the study area log (STP) = f (Lon, Lat) + z

(24.3)

where Lon represents longitude and Lat is latitude of a sampling site. The fitted model is presented on a 100 × 100 grid for each subregion. With the fitted model we are able to estimate the size of the area in each subregion that has a mean STP within a certain range, presented in terms of percentage of the total area of the subregion. The estimation method is similar to that presented in Qian (1997a).

584

24.4.2

S.S. Qian and C.J. Richardson

Soil Total Phosphorus

In this section, we present spatial distributions of STP in each of the four subregions. Both weight-based (mg kg−1) and volume-based (g m−3) STP are presented (Figs. 24.2–24.5). First, we note that there exists a general north-to-south gradient of STP. Very high STP concentrations (1,700–1,900 mg kg−1) can be found in the southwest corner of WCA-1 and the northeast part of WCA-2A. Near the inlet of WCA-3, the highest STP concentration is between 700 and 750 mg kg−1. While in ENP, the high STP concentration is between 600 and 700 mg kg−1. Note that in the northwest corner of the ENP modeling region STP is between 700 and 800 mg kg−1, which is a result of model extrapolation since no data is available in that corner. This gradient is also reflected in the percentage of area having high STP concentrations in each subregion (Tables 24.1 and 24.2). Water in WCA-1 concentrated around the WCA’s edges because the interior elevation is higher in elevation. Although over 80% of the area has STP less than 500 mg kg−1 (Table 24.1), it does have the highest STP in the entire region. In WCA-2A, the most intensively studied area, high STP occurs near the S-10 inlet structures releasing agricultural runoff, as expected. Comparing the locations of high STP with many published vegetation maps, the areas with STP concentrations larger than 600 mg kg−1 are almost congruent with the areas on the vegetation maps with cattail as the dominant species (Chap. 9). This suggests that high STP concentration may be one of the conditions of cattail expansion in WCA-2A. In both WCA-1 and WCA-2A, there is no significant

Fig. 24.2 Spatial distributions of (a) weight-based total phosphorus concentration (mg kg−1) and (b) volume-based total phosphorus concentration (g m−3) in WCA-1. Soils sampling sites are represented by dots

24 Spatial Distributions of Total Phosphorus and Phosphorus Accretion Rates

585

Fig. 24.3 Spatial distributions of (a) weight-based total phosphorus concentration (mg kg−1) and (b) volume-based total phosphorus concentration (g m−3) in WCA-2A. Soils sampling sites are represented by dots

Fig. 24.4 Spatial distributions of (a) weight-based total phosphorus concentration (mg kg−1) and (b) volume-based total phosphorus concentration (g m−3) in WCA-3. Soils sampling sites are represented by dots

586

S.S. Qian and C.J. Richardson

Fig. 24.5 Spatial distributions of (a) weight-based total phosphorus concentration (mg kg−1) and (b) volume-based total phosphorus concentration (g m−3) in ENP. Soils sampling sites are represented by dots

Table 24.1 Size (in percent of respective total area) of wetlands with given soil total phosphorus (STP, mg kg−1) levels Region < 500 500–750 750–1,000 > 1,000 WCA-1 WCA-2A WCA-3 ENP

81.4 49.4 90.5 94.4

9.2 35.3 9.5 4.7

3.6 7.9 0.0 0.9

5.8 7.4 0.0 0.0

Table 24.2 Size (in percent of respective total area) of wetlands with given soil total phosphorus (STP, g m−3) levels Region < 50 50–75 75–100 > 100 WCA-1 WCA-2A WCA-3 ENP

90.2 78.3 62.7 100.0

4.9 12.7 19.4 0.0

2.2 6.2 17.9 0.9

2.7 2.9 0.0 0.0

difference in the spatial distribution patterns for the weight-based STP and the volume-based STP. However, this difference is apparent in WCA-3. Using the weightbased STP concentrations, we found that STP concentrations decrease rapidly moving away from the inlets (northeast corner). After a few miles from the inlets, the STP concentrations are more or less the same over the rest of the region. However, there is a clear north-to-south gradient of volume-based STP in WCA-3.

24 Spatial Distributions of Total Phosphorus and Phosphorus Accretion Rates

587

This pattern is not to be interpreted simply as the effect of anthropogenic source of phosphorus, since there exists a natural gradient of soil bulk density (Fig. 24.6). The bulk density is high in the northwest part of the region and decreases gradually southward. The spatial patterns of STP in ENP are evaluated only in a small section of the Park south of a major water inlet, where we have 13 soil samples (the rest of ENP samples were taken far away from this cluster). The STP spatial pattern is similar to that in WCA-3, i.e., the weight-based STP concentrations decrease rapidly moving away from the inlet, and the volume-based STP concentrations show a north-to-south gradient that is due largely to the gradient of soil’s bulk density.

Fig. 24.6 Spatial distributions of soil’s bulk density (g m−3) in WCA-3. Soils sampling sites are represented by dots

588

24.4.3

S.S. Qian and C.J. Richardson

Phosphorus Accretion Rate

As we discussed in Sect. 24.1, PAR is an indicator of whether a wetland system is significantly affected by the anthropogenic phosphorus runoff. Richardson et al. (1997a) defined the phosphorus assimilative capacity as the maximum long-term PAR rate an ecosystem can support without causing structural changes. The assimilative rate is believed to be less than 1 g m−2 year−1 (Richardson and Qian 1999; see Chap. 23) for the Everglades wetlands. Spatial distributions of PAR can help us understand the magnitude of anthropogenic influence on the Everglades. Because measuring PAR is expensive and time consuming, we use the relationship developed in this chapter to transform the STP spatial distribution to PAR distributions (Figs. 24.7–24.10).

Fig. 24.7 Spatial distributions of phosphorus accretion rate (PAR, g m−2 year−1) in WCA-1. Soils sampling sites are represented by dots

24 Spatial Distributions of Total Phosphorus and Phosphorus Accretion Rates

589

Fig. 24.8 Spatial distributions of phosphorus accretion rate (PAR, g m−2 year−1) in WCA-2A. Soils sampling sites are represented by dots

With the spatial distribution of mean PAR, we estimate the percentage of area that has a PAR larger than 1.0 g m−2 year−1, as well as the area with PAR between 0.5 and 1.0 g m−2 year−1 (Table 24.3). The latter indicates a potential affected area, since the overall average of measured PARs in WCA-2A is about 0.44 g m−2 year−1 (Craft and Richardson 1993b). In general, high PARs were found in the west edge of WCA-1 and northern part of WCA-2A. Although the percentage of high PAR area in WCA-1 was small, WCA-1 had the highest PAR in the southwest corner (>1 g m−2 year−1). In WCA2A, only about 4% of the total area had PAR larger than 0.5 g m−2 year−1, while over 84% of the area had PARs less than 0.25 g m−2 year−1. It is somewhat surprising to see that only about 15% of the WCA-2A had PAR larger than 0.25 g m−2 year−1 after receiving an annual anthropogenic phosphorus load of nearly 80 metric tons since 1979 (SFWMD 1992). Nearly the entire areas of WCA-3 and ENP have PAR less than 0.25 g m−2 year−1. (Note that the region with PAR larger than 0.25 g m−2 year−1 in ENP is in the northwest corner, where no data were available; therefore, it should not be taken as accurate until more data are available.) Our values for PAR coincide

590

S.S. Qian and C.J. Richardson

Fig. 24.9 Spatial distributions of phosphorus accretion rate (PAR, g m−2 year−1) in WCA-3. Soils sampling sites are represented by dots

closely with the small area of elevated STP in WCA-3 reported by Bruland et al. (2006). We point out that the model predicts the mean PAR at a given STP level. Therefore, predictions are conservative. Variations of the predicted PAR come from two sources (1) the error variance of the PAR–STP relationship (Fig. 24.1, or ε in (24.1) ) and (2) the error variance of the STP spatial distribution model (24.3), which is different for each subregion. Variances of errors in both (24.1) and (24.3) can be estimated by assuming the errors have a normal distribution (e ~ N(0, s 22) and z ~ N(0, s 21) ). With the estimated error variances (s 21, s 22 ) from the fitted models, a Monte Carlo simulation is used to estimate the spatial distribution of the predicted PAR. We first randomly select 100,000 points inside each subregion and using (24.3) to predict the mean STP concentrations. Then, 100,000 random numbers were generated from the distribution of z (N(0, s 22) ). Adding the predicted mean and the random error resulted in

24 Spatial Distributions of Total Phosphorus and Phosphorus Accretion Rates

591

Fig. 24.10 Spatial distributions of phosphorus accretion rate (PAR, g m−2 year−1) in ENP. Soils sampling sites are represented by dots

Table 24.3 Size (in percent of respective total area) of wetlands with given mean phosphorus accretion rate (PAR, g m−2 year−1) levels Region <0.25 0.25–0.5 0.5–1.0 >1.0 WCA-1 WCA-2A WCA-3 ENP

90.37 84.32 99.99 98.87

5.36 11.73 0.01 1.13

3.02 3.95 0.00 0.00

1.25 0.00 0.00 0.00

100,000 predicted STP values. In turn, we have 100,000 mean PAR values. These PAR values were then adjusted based on the error distribution of the PAR–STP relationship (adding 100,000 random numbers from (e ~ N(0, s 21) ). The resulting 100,000 PAR values were used to estimate the percentage of areal coverage for specified PAR intervals for each subregion (Table 24.4). The results in Table 24.3 are conservative estimates (or the lower bounds), and the results in Table 24.4 may be considered as the upper bounds of the affected area. The predicted upper estimates of PAR for WCA-1A, WCA-2A, and the ENP were similar to (Table 24.4) earlier estimates but these analyses show that 33% of WCA-2A were above the 0.25 g m−2 year−1 compared with the lower 15% estimate (Table 24.3). The upper estimates more closely coincide with field measurements of vegetation shifts (Chap. 9).

592

S.S. Qian and C.J. Richardson Table 24.4 Size (in percent of respective total area) of wetlands with given (predicted) phosphorus accretion rate (PAR, g m−2 year−1) levels Region <0.25 0.25–0.50 0.50–1.0 >1.0 WCA-1 WCA-2A WCA-3 ENP

24.5

88.04 66.12 88.70 97.34

8.23 24.03 10.85 2.65

2.73 8.86 0.45 0.01

1.00 0.98 0.00 0.00

Conclusions and Lessons for Restoration

We presented spatial distributions of soil’s total phosphorus and long-term PAR in four separated wetlands in the Everglades. The distribution of PAR is based on the PAR–STP relationship developed using 68 samples from the same study area. The PAR–STP relationship is nonlinear; however, it can be expressed by using a simple piecewise linear model (in logarithm) (equation 24.2). The piecewise linear model is derived from data; however, it reflects the fact that there exists a detection limit of PAR using the 137Cs method. When measuring PAR, soil cores are sliced into many layers with certain thickness (1 cm) for determining the 137Cs profile. When the peat accretion since 1964 is less than this thickness, the maximum of 137Cs should be in the first layer; therefore, this thickness is taken as the amount of peat accretion since 1964. In addition, compression of the soil core may be inevitable, which will result in a larger relative error for small PAR. As a result, the positive association between PAR and STP disappeared when STP is below a certain level (326.5 mg kg−1). In general, STP concentration is higher in regions that are closer to inlets of phosphorus-enriched runoff. The high STP concentration only occurred around the edge of WCA-1 due to the higher elevation in the interior. In WCA-2A, STP concentrations are high in the northeast part of the region and in the west corner near an inlet. The shape of the high STP region is similar to the shape of cattail monoculture area, which indicates the effect of high phosphorus concentration on the ecosystem. In both WCA-3 and ENP, only a small portion of the respective areas was affected by anthropogenic phosphorus; the rest of the region displayed no apparent spatial trend. In the same region, the soil’s bulk density displays a north-to-south decreasing gradient. As a result, volume-based STP also showed the same gradient. Using our model the general trend of PAR distribution follows that of STP. If we consider the background level of PAR to be less than 0.25 g m−2 year−1 (Craft and Richardson 1993b; Reddy et al. 1993), we conclude that anthropogenic sources of phosphorus have no significant influence in WCA-3 and ENP. Additionally, we conclude that small portions of WCA-1 and larger areas in WCA-2A have been affected. The spatial distributions of PAR presented in Figs. 24.7–24.10 and Table 24.3 are mean PARs for respective subregions. If we consider a wetland having a PAR > 0.5 g m−2 year−1 as being significantly affected by agriculture runoff, we can conclude that about 4% of the area of WCA-1 and about 4–10% of the WCA-2A has been significantly affected by anthropogenic source of phosphorus. For WCA-3

24 Spatial Distributions of Total Phosphorus and Phosphorus Accretion Rates

593

and ENP, the significantly affected areas are very small; however, up to 3% of the ENP and 10% of WCA-3 may be slightly affected (having a PAR > 0.25 g m−2 year−1). Again, the number for ENP is less reliable since only 13 samples were used for estimation. However, they were near the input structures and should reflect inputs into the ENP from the north. These findings provide a baseline comparison for future P changes in the Everglades. They indicate that agricultural P additions have primarily affected the northern Everglades (WCA-1A and WCA-2A). PARs have increased as a result of exceeding the long-term assimilative capacity of the ecosystem, and care must be taken in the future to reduce P input loads and concentrations. However, the release of P thorough narrow gates and pump stations, even at low concentrations, will result in higher loading at the edge of the fen. New delivery systems designed to spread water over a larger surface area would help in the reduction of P mass loadings to levels below PAC at specific release sites along the edge of the Everglades. This reduction would result in less downstream movement of P deeper into the fen.

25

An Ecological Basis for Establishment of a Phosphorus Threshold for the Everglades Ecosystem Curtis J. Richardson, Ryan S. King, Song S. Qian, Panchabi Vaithiyanathan, Robert G. Qualls, and Craig A. Stow

25.1

Introduction and Background

Numerous studies have shown that the Everglades fen is a phosphorus-limited ecosystem (Steward and Ornes 1975a,b; Craft and Richardson 1993a; Koch and Reddy 1992; Richardson et al. 1999; Richardson and Qian 1999; Noe et al. 2001). From this it can be hypothesized that increases in phosphorus concentrations in the water column and the soils of the Everglades above the ecosystem’s P assimilative capacity (Chap. 23) will result in significant imbalances in the structure and function of the Everglades ecosystem (Richardson and Qian 1999). In the following chapter we provide information on our experimental results and general organism and ecosystem responses to P additions, as well as provide a statistical basis for determining a P threshold as it relates to ecological imbalance. However, before assessing the experimental P-dosing results, it is necessary to establish quantifiable metrics or indices of trophic level response to P concentrations as well as to determine a priori a scale of acceptable P effects. Various ideas have been proposed to aid in the development of a phosphorus threshold for the Everglades. They range from simply using the background concentration of P found in the water column or soils in the most areas of the Everglades to utilizing technology-based criteria founded on best available P removal technology. It has also been argued that the P standard could be based on the geometric mean or median P concentration found in the water column of a reference or least disturbed area of the Everglades that displays the native balance of plant and animals (SFWMD 2001, 2002, 2003, 2005, 2006). This concept is attractive in that it is based on the recently popular use of a reference area as the standard conditions for comparing wetland functions (Brinson et al. 1995; US EPA 1997b). The selection of the reference area itself is the key to this entire approach since it dictates the baseline environmental conditions that are used for each wetland type by region. This approach requires the selection of numerous reference areas and measurements over time to capture the high natural variation with the system (Brinson et al. 1995). However, this method has not been fully tested or calibrated, especially as it relates to water quality standards in wetlands. Also, the use of interior reference systems alone also does not account for the natural nutrient and hydrologic gradients that exist in wetland from the edge inward (Keddy 2000).

595

596

C.J. Richardson et al.

Importantly, if the reference system is used and no change from the reference P state is allowed then this infers that ecosystems have no capacity to tolerate even minimal stress or disturbance above baseline conditions. This assumption has been seriously challenged with detailed studies in recent years (Botkin 1991; Richardson et al. 1997a; Richardson and Qian 1999). Moreover, the reference approach would need to take into account reference site variability and community differences in water quality as well as seasonal variations, water-depth effects, and the role of fire on nutrient concentrations. This has never been done for the Everglades reference approach. Another approach is the development of biological monitoring (biomonitoring) techniques developed for streams (Karr et al. 1986; Plafkin et al. 1989; Barbour et al. 1996). This is based on selecting biological metrics that can be used to describe structural and functional changes of aquatic communities (algae, macroinvertebrates, fish) in the test stream compared to reference streams in the area. (A metric is defined as a characteristic of the biota that changes in a predictable way with perturbations to the ecosystem.) The selected metrics should have some sound ecological basis (Karr et al. 1986; Barbour et al. 1995). Moreover, it has been shown that a single metric, reflecting a single attribute may fail to indicate the effects of the system stressor (Karr et al. 1986; Barbour et al. 1996). Thus, a multimetric approach that incorporates several aspects of the structure and function of a community must be used to properly assess biotic responses and ecosystem level structure and functions (Kerans and Karr 1994). In general, candidate metrics have focused on measures of richness of species (e.g., number of taxa), compositional measures (e.g., Shannon–Weaver diversity), or trophic measures (e.g., density or % by guild classes) (Barbour et al. 1996). The sensitivity of each metric to distinguish between a reference site and the test site is based on a simple comparison of box and whisker plots (Barbour et al. 1996). Often this approach has focused on macroinvertebrates, with intolerance classes assigned to each taxon and the classes summed. This has resulted in a somewhat subjective assessment procedure. This approach has not been well tested or calibrated for wetlands, nor has the evaluation method for determining sensitivity of metrics been quantified. Moreover, this method has never been used to assess a limiting nutrient for an ecosystem, so the direction of response variables (increase or decrease) is unknown. However, the multimetric approach has some promise for assessing biological responses if properly calibrated. Recently, it has been suggested that periphyton alone can be used to assess environmental conditions of streams and wetlands (Stevenson and Lowe 1986; McCormick and Stevenson 1998; Stevenson et al. 2003). In this approach, the changes in taxonomic composition of sensitive and tolerant species are compared between reference areas and disturbed areas. In addition, periphyton biochemistry, nutrient availability, and biomass production are included as metrics of assessment. This system relies on only one trophic level, and the selection of sensitive species is based on limited wetlands algal response data. The approach of using the most sensitive species has some merit as an early indicator of stressors on the biota, but it has usually been restricted to toxic effects on the organisms and not nutrient

25 An Ecological Basis for Establishment

597

increases to the system (Cairns 1983; Cairns and McCormick 1991). Moreover, the ecological importance of changes in certain algal species is unknown, especially to higher trophic levels. An alternative approach is based on the use of the subsidy–stress model of Odum (Odum et al. 1979) as a testable model to assess functional and structural imbalance in ecosystems (Fig. 25.1). This concept follows Liebig’s concept of limiting factors (Liebig 1840) and Shelford’s “limits of tolerance.” Thus, additions of P – a critical element for all life processes that is in short supply in the Everglades (Davis 1943; Steward and Ornes 1975a; Craft and Richardson 1993a,b; Rader and Richardson 1994; Richardson and Vaithiyanathan 1995; Richardson et al. 1999; Richardson and Qian 1999) – should result in increased plant and animal growth and increased ecosystem level responses. The central questions are at what level are these changes unacceptable and when do they constitute an imbalance of flora and fauna? It is clear that ecosystem change occurs, but not all levels of change equal an imbalance (Botkin 1991). Moreover, change is a normal part of ecosystem responses. Hence, change alone does not equate with imbalance. Nevertheless, an excess of any ion can result in imbalances in ecological communities and ecosystem processes if it causes another ion to become limiting or it results in a toxic effect on organisms within the system. Phosphorus is not toxic to organisms per se, but additional P in an ecosystem may result in increased growth in organisms or unacceptable changes in community structure. Excess levels might also limit growth due to a deficiency of another ion (Liebig 1840; Redfield 1958) or result in cultural eutrophication of the ecosystem (Odum 1971; Wetzel 2001).

Fig. 25.1 Hypothetical phosphorus response curve for the Everglades ecosystem. The curves simulate the output response (as measured by appropriate systems functions like diversity or productivity, etc.) to increasing P additions (Odum et al. 1979; Lemly and Richardson 1997)

598

C.J. Richardson et al.

Shelford (1913) referred to the range of environmental effects on organisms as the “limits of tolerance” for organisms, and he stated that not only was too little of something a limiting factor but also too much might also result in a decreased response. This suggests that individual species had a range of tolerance. Odum et al. (1979) further expanded these concepts and developed the subsidy–stress model for ecosystem responses to material inputs. The subsidy–stress model predicts that the introduction of a nontoxic limiting element like P would result in a subsidy effect (not a decrease) in ecosystem responses like productivity, species diversity, nutrient storage, and cycling rates up to a threshold maximum (Fig. 25.1; Lemly and Richardson 1997). At some P “threshold” concentration or load, the ecosystem would reach a maximum response beyond which a decrease in response would become evident (Chap. 23). It should be noted that the threshold for ecosystems is a zone, not one number; thus, increases in productivity and diversity, as well as changes in community structure, become more evident as the maximum P threshold is reached. Moreover, the variation in ecosystem response becomes greater near the maximum threshold. This suggests that some method of assessing the amount of variation in response variables like the number of taxa or diversity might be an important indicator of imbalance for the system. Of interest is the amount of P that can be added to an ecosystem without significant changes in system responses or imbalance in flora and fauna. The amount of P that can be added above background levels and assimilated into the system without significant changes in ecosystem structure and function or downstream concentrations of P has been defined as the P assimilative capacity (Fig. 25.1; Richardson and Qian 1999; see Chap. 23). Connell’s intermediate disturbance hypothesis also predicts a unimodal shape of the diversity–disturbance relationship at the population/community level (Connell 1978). He states that the unimodal curve is caused by two factors that depress species number at opposite ends of the perturbation gradient. Competitive exclusion is operative as resources are limited (resource depletion), thus reducing the number of coexisting species under undisturbed or small perturbations. At higher inputs, only a few rapid growing populations can withstand the inputs. This model would predict that the maximum diversity and species richness in the Everglades would be found under intermediate intensities, frequencies, and times of disturbance. Significant changes can be expected in ecosystem structure and function once the P assimilative capacity is exceeded (Richardson and Qian 1999; see Chap. 23). This is due to the difference in niche width related to the limited P resource (Tilman 1982). If the niche is narrow in a P-limited system, then some form of P is likely to be a major niche dimension. Thus, a competing species could theoretically move in and take over if there is even a small change in P concentration. Alternatively, the niche width for a species or group of species is relatively broad, and then the threat of competitive replacement is reduced (Tilman 1982; Colinvaux 1992). Undoubtedly, some species will be replaced by small changes in P, because they are already living near the limits of their tolerance. However, it should not be supposed a priori that because of one, or even a small proportion of species turnover that the system is changed in any fundamental sense, especially at the lower trophic levels. By contrast,

25 An Ecological Basis for Establishment

599

an excessive amount of P additions would be reached when the normal operating range of ecosystem responses are significantly affected. The central ecological question is to what degree are increases or decreases in ecosystem responses tolerable? Of equal importance is the selection of biological metrics that can be used to properly assess what constitutes an imbalance in the Everglades ecosystem.

25.2

Indices of Ecological Change or Biological Metrics that can be Used to Assess P Imbalance

Clearly any additions to the Everglades ecosystem that result in a significant loss of ecosystem function and community structure are not acceptable to ecologists or society. Here we define ecological imbalance as a significant alteration in the mean and/or variance of the biological response variable found above vs. below the ecological threshold. In our studies we have analyzed the responses at several trophic levels as well as determined ecosystem level functional indicators. Biological components or metrics that can be used to assess changes in species, community, or ecosystem level responses to additions of P have been a focus of our research effort. Our studies in this phase are based on research along a P gradient in a 6-year mesocosm experiment (Chap. 15) with a constant disturbance intensity of P additions. This experiment was set up to establish a sufficient range of P concentrations that would provide a predictive P response curve that could be used to analyze biotic responses to P additions. We used these responses to help us in selecting metrics that are indicative of changes across all levels of response (e.g., species, community, and ecosystem). Moreover, we used this experimental approach to calibrate appropriate metrics that indicate imbalance at different trophic levels (e.g., algal, macrophyte, macroinvertebrate, community, and ecosystem). The metrics were then compared to responses found along a natural P gradient (Chaps. 6, 9, 10, and 11) in the Everglades. Accordingly, we hypothesized that Everglades communities and ecosystem processes will initially respond positively to P additions up to some “P threshold concentration zone” above which a decrease in response will be measured (Fig. 25.1). It should be noted that our experiments are set up to test this hypothesis as well as the alternate hypothesis that P additions do not result in a positive and significant response initially and that no threshold above background exists for P additions. Potential biotic indicators that we analyzed for responses across several trophic levels are shown in Table 25.1. Nearly 50 metrics were analyzed over five levels of organization and responses are presented in detail in Chaps. 16–19. Metrics are some measure of the biological attributes that represent the elements of structure and function of assemblages and change in some predictable, consistent fashion along a perturbation gradient (Barbour et al. 1995). Here we characterize imbalance as a significant alteration of multiple changepoints across trophic levels. Therefore, change for one biological attribute or metric alone within a trophic level would not constitute an imbalance. This definition corresponds to the following statement

Table 25.1 Potential ecological attributes and metrics of change in Everglades communities Algae level responses

Phosphorus conditions tested −2

Algal density (by substrate, cells cm )

(Previous 2-, 3-, and 6-month water column geometric mean of TP)

Blue-green algae biovolume (by substrate, cm3 per cell) Diatom density (by substrate, cells cm−2) Diatom relative abundance (by substrate, %) Diatom biovolume (by substrate, cm3 per cell) Phosphorus-sensitive algal density (by substrate, cells cm−2) Phosphorus-sensitive algal relative abundance (%) Pollution-sensitive algal relative abundance (%) Pollution-sensitive algal density (cells cm−2 by substrate) Macrophyte level responses

Phosphorus conditions tested

Utricularia spp. (stem densities)

(Previous 2-, 4-, 6-, 8-, 10-, and 12-month water column geometric mean of TP)

Utricularia purpurea (stem densities) Eleocharis cellulosa (stem densities) Eleocharis elongata (stem densities) Nymphaea odorata (areal cover) Typha domingensis (% frequency) Cladium jamaicense (% frequency) Macroinvertebrate level responses

Phosphorus conditions tested

Total abundance (number)

(Previous 1-, 2-, 3-, 6-, and 12-month water column geometric mean of TP)

Total biomass (mg) Temporal variation (CV temporal abundance) Microcrustacea (number) Oligochaeta (number) Predators (%) Gastropoda (%) Sensitive taxa (FDEP) (%) Community level responses

Phosphorus conditions tested

Algal taxa (number)

(Previous 2-, 4-, 6-, and 12-month water column geometric mean of TP)

Calcareous periphyton mat cover (%) Community similarity/dissimilarity (%) Macroinvertebrate taxa (number) Shannon–Wiener diversity (scaled index) Ecosystem level responses

Phosphorus conditions tested −1

(Previous 2- and 3-month water column Diel oxygen concentrations (mg l ) P average except for plant biomass, Productivity/respiration profiles (g m−2 day−1) which used 6–12 month water column Plant biomass (g m−2 at seasonal peak) geometric mean) Periphyton mat cover biomass (g m−2 at seasonal peak) Selected metrics for each indicator are shown in parentheses. The time period that was used for averaging P chemistry and compared to biotic responses is shown in bold. Other time periods tested included 1, 2, and 6 months as well as the cumulative number of months prior to measurement. Final selections were based on biotic uptake response times and lifespan based on literature values for each group

25 An Ecological Basis for Establishment

601

made to the Environmental Regulatory Commission (ERC) by the SFWMD in 1994: “The detection of ecological imbalance not only requires selection of appropriate indicators of ecosystem condition, but also presumes some understanding of the magnitude of the ecosystem change that is deemed acceptable. It is doubtful whether any absolute environmental ‘threshold’ concentration for phosphorus, or any pollutant, exists for aquatic systems…” (SFWMD 1994). Therefore, care has to be taken at each level of response to select and test attributes that represented ecologically significant components of the structure and function of the Everglades. The order shown does not indicate any predetermined importance or prior sensitivity ranking. The responses were measured for either a positive or a negative value from the control conditions over numerous years and seasons. The selection and testing of numerous metrics at several hierarchical levels within the ecosystem were done to provide comprehensive information that could be integrated to provide a sound ecological basis for a P threshold in the Everglades. It has been shown experimentally that “integrating metrics” like functional composition and functional diversity of plants control important ecosystem processes like productivity and nutrient status (Tilman et al. 1997). Tilman’s research and that of Paine (1966) and Vitousek et al. (1987) have shown that factors that change ecosystem composition – such as invasion by novel organisms, nutrient inputs, disturbance frequency, fragmentation, predator decimation, species extinction, and alternative management practices – are likely to strongly affect ecosystem processes. Importantly, the loss or additions of species with certain functional traits can have a great impact, while others have little impact on ecosystem processes (Tilman et al. 1997). Because functional redundancy at lower trophic levels has been recognized in ecosystems for years, focusing on developing metrics for a single species (Cairns 1982, 1983) is not appropriate for determining thresholds. These findings suggest that a P threshold for the Everglades cannot be based on single species responses unless they are keystone species for the ecosystem, a keystone species being one that plays a key role in regulating the composition of the associated community (Paine 1966, 1974, 1980). For example, Paine showed that the removal of Pisaster (starfish) from the intertidal community results in a reduction of species from 15 to 8. Simenstad et al. (1978) found that the removal of the sea otter resulted in a dramatic change in the composition of the entire marine community off the Aleutian Islands. These and other studies suggest that keystone species for an ecosystem are generally found at the higher trophic levels and that lower trophic levels have built in redundancy (Odum 1971; Westman 1985). These studies also suggest that the loss of species at higher trophic levels (carnivores) has a larger proportional ecosystem impact than a loss at the lowest trophic levels. For the Everglades, this would be represented by a loss of the alligator vs. a loss of an algal or bacterial species. However, recent work has shown that removal of some of the macrophytes like Utricularia spp. may have major impacts on the slough mat community in the Everglades (see Chap. 16; Ulanowicz 1995). Redundancy of information was reduced by classifying attributes into categories that reflect relatively distinct components of each assemblage level, then eliminating

602

C.J. Richardson et al.

those that provided little additional information within categories (Barbour et al. 1996). Excessively redundant attributes, such as multiple measures of diversity, were consolidated to a few core attributes within each category. Importantly, we used ecological information to guide this attribute reduction process rather than simple correlation analyses, as correlation does not necessarily connote redundancy (Fore et al. 1996; Karr and Chu 1997). Simply using the correlation approach to eliminate attributes can result in the loss of diagnostic information by excluding attributes that may appear related to others but provide unique insight to the response of the assemblage (Karr and Chu 1997).

25.3

Objectives

The specific objectives of this research were to (1) analyze gradient P research in the Everglades to establish an ecological basis for determining a phosphorus imbalance in the Everglades; (2) investigate in detail the biological and chemical responses from a 6-year phosphorus-dosing study designed to assess ecological responses to P additions; (3) develop a comprehensive trophic level model of population, community, and ecosystem responses to increased P additions; (4) test responses at algal, macrophyte, macroinvertebrate, community, and ecosystem levels to develop a multiple metric model as an indicator of an imbalance of flora and fauna; and (5) integrate experimental research at several scales (lab, mesocosm and flume studies, etc.) with observational studies along a P gradient in the Everglades to develop a hierarchical integrated ecosystem level model of P-induced imbalance including the development and testing of an index of biological integrity (IBI) model for macroinvertebrates.

25.4

Development of an Integrated Hierarchical Model for P Threshold Analysis

To assess these biological indicators and their corresponding metrics, we tested their responses at both control (reference) sites and experimental sites (Fig. 25.2) that had received P additions for the last 6 years (dosing experiment). In addition, we tested sites that had received 30-plus years of N and P as well as ion additions along a north–south gradient subjected to both surface and ground water inputs in Water Conservation Area 2A (WCA-2A) of the northern Everglades (Richardson et al. 1997a; see Chaps. 5–13). Importantly, we developed an integrated approach to combining the data from both the dosing experiment and the gradient studies to advance an ecological criterion for determining the P threshold (Fig. 25.3). Responses at the algal, macrophyte, macroinvertebrate, as well as community and

25 An Ecological Basis for Establishment

603

Fig. 25.2 An integrated model of a long-term database used to determine an ecological-based P threshold for the Everglades. Information from a 6-year Everglades phosphorus-dosing experiment was combined with a 10-year database along a phosphorus gradient in the northern Everglades

ecosystem level responses, were considered in the development of the integrated model (Table 25.1). Final attributes were selected based on ecological importance and statistical analyses (Richardson et al. 2000a; Breiman et al. 1984) at each trophic level in the Everglades. The attributes tested are as follows: Algae–Diatom relative abundance (%), diatom density (cell m−2), diatom biovolume (cm3 per cell); blue-green algae biovolume (cm3 per cell), Macrophytes – Utricularia spp. (stem densities), Utricularia purpurea (stem densities); Macroinvertebrates – biomass (mg), abundance (number), Oligochaeta (number), microcrustacea (number), sensitive species (%), predators (%), Gastropoda (%); and Community – macroinvertebrate taxa (number), Bray–Curtis dissimilarity (BCD; scaled index), calcareous mat cover (%). Selection of these response attributes were based on prior analysis of all attributes (Table 25.1) following accepted ecological criteria used to assess biotic and system responses to perturbations (Green 1979; Karr et al. 1986; Clarke 1993; Cao et al. 1998; Cairns and McCormick 1991; Odum et al. 1979; Barbour et al. 1996; Tilman et al. 1997). The quantitative evaluation of responses of ecosystems, communities, populations, or species to a pollutant is difficult. Arguments exist as to whether or not one should focus on structural (species numbers, taxa and diversity, etc.) or functional properties (productivity, decomposition, etc.) of ecosystems (Cairns and Pratt

604

C.J. Richardson et al.

Fig. 25.3 The location of the Duke University phosphorus-dosing site in Water Conservation Area 2A. Note the dosing site is located in an area of low phosphorus concentrations as denoted by the soil P contours (P in mg kg−1)

1986). We have developed and analyzed properties of both in our dosing study and along a P gradient in WCA-2A. However, a true estimation of the exposure dose is critical to the interpretation of the results. Actual threshold zones by trophic group in our final analysis are based on the averaged weighted doses and the dose– response curves developed over measured geometric mean P concentrations in the channels. The long-term P dose load to the organism is the averaged weighted amount of PO4-P added to the mixing tanks times the volume of water pumped and not the residual amount of PO4-P measured in the water column at point (x) at time (t).

25 An Ecological Basis for Establishment

605

This is especially true for PO4-P dosing, since this limited nutrient is rapidly taken up by organisms (minutes to hours), precipitated, or diluted by the water in the system. Moreover, a comparison showing increases of PO4-P in the water column in treatment channels over the amount of PO4-P found in control sites may be an indication of system saturation; especially if biotic responses are significant. Another concern that must be addressed is the increase of total phosphorus (TP) in the water column, and the ratio of TP to PO4-P. The proposed state standards are based on TP, not on PO4-P, although the former is not as readily available to biota. We have only used TP concentrations as the independent variable in our main response curves, but PO4-P analysis may be a more useful variable. The TP/PO4-P ratio varies by season, at different locations in the Everglades, and in different channels in the dosing study as a consequence of PO4-P additions. This adds additional problems to the data interpretation. It should also be noted that P additions can act directly on the biota or indirectly through alteration of the physical environment as in the case of changing pH levels (Richardson et al. 1995, Chap. 15). Finally, because functional redundancy has been recognized in ecosystems for years, we did not focus on developing bioassays for single species in the mesocosms (Cairns 1982). Our integrated indices were developed over multiple trophic levels (Table 25.1) and used to indicate significant changes in population, community, and ecosystem structure and function.

25.5

Analytical Approach for Data Analysis

To first evaluate which abiotic (e.g., water depth, year) and ecological (e.g., algal, macrophyte, macroinvertebrate) variables were the best indicators of TP concentration, we used Classification and Regression Tree (CART) models, an analysis that recursively partitions observations into groups using the best indicators of TP (Breiman et al. 1984) as implemented in S+ software (Clark and Pregibon 1992; MathSoft 1997). The analysis used data collected over more than 6 years, spanning all seasons and a wide range of water depths. The model selects predictor variables and the splitting points successively based on the reduction of the total deviance. The result is analogous to a dichotomous key where successive choices are made regarding the value of the response variable, based on predictor characteristics (for details see Qian and Anderson 1999). This approach has proven to be a powerful tool in initially predicting which biotic attributes change the most along an environmental gradient or in determining the breakpoint where response variables display major differences along the environmental gradient (Amrhein et al. 1999; Qian and Anderson 1999). The Bayesian hierarchical changepoint model was specifically designed for detecting changes along a gradient (Qian et al. 2003, 2004). Specifically, we assume that the response variable values, y1,…, yn, collected from the n sites along the P gradient of interest, are random samples from the sequence of random variables Y1,…,Yn. The corresponding P concentration values are x1,…,xn, where x1 < x2 < … < xn.

606

C.J. Richardson et al.

Assuming that variables Y1,…,Yn belong to the same family of distributions with parameter θ. The random variables Y1,…,Yn have a changepoint r (1 ≤ r ≤ n) if the parameter value changes at r Y1,..., Yr ~ p (Yi | q1), Yr+1,..., Yn ~ p (Yi | q2). The corresponding P concentration xr is the P threshold or changepoint. The results of the model is presented in the form of a probability distribution of the n P conn centration values x1,…,xn being the changepoint, or p(xi), where ∑ p( xi ) = 1 . i =1

Theoretical background on this type of changepoint analysis can be found in Smith (1975), Raftery and Akman (1986), and Carlin et al. (1992). Qian et al. (2003) provide a complete description of the Bayesian and nonparametric changepoint methods along with the model forms used to analyze the data. We further estimated potential threshold responses in the measurement endpoints to numerical levels of TP for the macroinvertebrates using nonparametric changepoint analysis (nCPA), a technique explicitly designed for detecting threshold responses using ecological data (Qian et al. 2003). nCPA is a derivative of a family of techniques historically used in classification and divisive partitioning of ecological data (e.g., Pielou 1984). This analysis is based on the idea that a structural change in an ecosystem may result in a change in both the mean and the variance of an ecological response variable used to indicate a threshold. When observations are ordered along a stressor gradient, a changepoint is a value that separates the data into the two groups that have the greatest difference in means and/or variances. This can also be thought of as the degree of within-group variance relative to the between-group variance, or deviance (D) (for details see Venables and Ripley 1994; Qian et al. 2003). Analytically, the nCPA examines every point along the stressor gradient and seeks the point that maximizes the reduction in deviance. Thus, each stressor value is a potential changepoint and is associated with a deviance reduction ∆ i = D − ( D≤ i + D> i ), where D is the deviance of the entire data set y1,…,yn, D ≤i is the deviance of the sequence y1,…,yi, and D>i is the deviance of the sequence yi+1, …, yn, where i = 1,…, n. The changepoint r is the i value that maximizes ∆i: r = maxi ∆i. There is one particular value of the predictor y (in this case, TP) that maximizes the reduction in deviance in the response data (in this case, the selected metrics); however, there is uncertainty associated with that value. It is unlikely that any one value of TP is the only value that could represent a changepoint. In reality, depending on the acuteness of the biological change in response to TP, several observations of TP could represent the changepoint, each with varying probabilities. Thus, to assess the risk associated with particular levels of TP, nCPA incorporates estimates of uncertainty in the changepoint (King and Richardson 2003; Qian et al. 2003).

25 An Ecological Basis for Establishment

607

These estimates are calculated using a bootstrap simulation (Efron and Tibshirani 1993). This simulation resamples (with replacement) the original dataset and recalculates the changepoint with each simulation. Bootstrap simulations are repeated 1,000 times. The result is a distribution of changepoints that summarizes the uncertainty among multiple possible changepoints. This uncertainty is expressed as a cumulative probability of a changepoint based on the relative frequency of each changepoint value in the distribution. An additional factor to consider when using nCPA is an estimate of the probability of Type I error. A c2 test statistic (1 df) can be used to evaluate the likelihood that an observed changepoint is real (Qian et al. 2003). However, we only used this statistic to help in assessing the likelihood that changepoints with relatively wide cumulative probability distributions represented real biological changes, as uncertainty around the changepoint was a much more relevant issue (Suter 1993; Germano 1999; Johnson 1999). Analyses were conducted using the custom function “chngp.nonpar” (Qian et al. 2003) in S-Plus 2000 (MathSoft, Inc., Seattle, WA).

25.6

25.6.1

An Analysis of P Thresholds to Determine Ecological Imbalance Dosing Study

To evaluate which abiotic (e.g., water depth, year) and ecological (e.g., algal, macrophyte, macroinvertebrate) variables were the best indicators of TP concentration, we used CART models. The analysis used data collected over more than 6 years, spanning all seasons and a wide range of water depths. The density (stems m−2) of U. purpurea (bladderwort), a floating aquatic macrophyte, was selected as the best macrophyte indicator of TP (Fig. 25.4a). Less than 0.5 stems m−2 of U. purpurea were found at TP concentrations averaging 21 µg l−1 (Fig. 25.4b). By contrast we found up to 63 stems m−2 in channel locations with lower TP concentrations. The second best indicator was the combined population of Utricularia (U. purpurea + U. fibrosa + U. foliosa) closely followed by water depth (data not shown). All three Utricularia spp. – a key plant component of the floating aquatic mat community – demonstrated a consistently pronounced decline with increasing P concentrations (Fig. 25.4c). This was apparently due to their inability to photosynthesize in waters devoid of CO2, a condition found at higher pH levels (Moeller 1978) in P-enriched areas of the Everglades. In fresh waters the total amount of free CO2 available for photosynthesis is variable and highly pH dependent. The pK dissociation relationships of CO2, HCO3-, and CO2indicate that free CO2 is the dominant form in the 3 water column at a pH of 5 and below, while HCO3- is dominant from pH 7 to 9. Above a pH of 9.5, CO2; is the main form of inorganic carbon in the water column 3 (Wetzel 2001). The pH in the Everglades averages 7.5 and ranges from a low of 7.2 at night to a high of >10 during the day in the P treatment channels (Chap. 15).

608

C.J. Richardson et al.

Fig. 25.4 (a) Classification and Regression Tree (CART) model using total phosphorus (TP) as the response variable was modeled against candidate predictor variables (e.g., macrophytes species, water depth, algal species). Utricularia purpurea was selected as the candidate predictor that was most closely associated with P concentrations in the water column of the Everglades. (b) The number of U. purpurea stems m−2 in response to mean TP concentrations found in the water column in August. The TP values are the geometric mean of the water column phosphorus found during the 6-month growth period for the plant species. Data for all seasons and years are combined to give a general response pattern for U. purpurea during the entire experiment (1992– 1998). (c) The number of Utricularia spp. Stems m−2 in response to mean TP concentration in the water column in August of 1993–1998

Next, we used the Bayesian changepoint method to detect variations in response of the population of U. purpurea to TP at different seasons each year (Fig. 25.5a–d). Here each existing data point was assigned a probability of being the changepoint, and the TP concentration associated with the highest probability was selected as the TP threshold. For example, U. purpurea displayed a variation in TP thresholds over different seasons with geometric mean values of 17.2 (August 1995), 22.3 (March 1966),

25 An Ecological Basis for Establishment

609

Fig. 25.5 (a–d) A decrease in the stem counts of Utricularia purpurea per m2 to increasing water column total phosphorus (TP) concentrations in the dosing channels during the summer and spring of 1995–1998. The selected TP threshold was based on Bayesian analyses and is shown as a dark vertical line. Values range from a low of 12.4 µg l−1 TP in April of 1998 to a high of 22.3 µg l−1 TP in March of 1996. The mean and variance of the populations above and below the thresholds have been shown to be significantly different as estimated by a nonparametric deviance analyses (Qian and Anderson 1999). The probability for all the thresholds exceeds 0.97

12.4 (April 1998), and 16.6 µg l−1 TP (August 1998, near the end of the 6-year experiment) (Fig. 25.5d). In all cases the probability of each selected threshold exceeded 0.97. Importantly, the temporal variation among TP thresholds suggested that seasonal and yearly differences must be integrated into the development of a TP threshold for each attribute. To accomplish this, we calculated an overall mean threshold and credible interval (CI) for U. purpurea for the entire experimental period by averaging the geometric mean P thresholds across each time period. The 95% CI is calculated as the interval of P concentrations that includes 95% of the probability mass of the changepoint distribution. The lower bound (xL) is the largest L P concentration such that ∑ i =1 p( xi ) ≤ 0.025 and the upper bound (xU) is the smallest P concentration such that ∑ p( xi ) ≤ 0.025 . Because the distribution is a discrete, i=U the selected 95% CI (xL, xU) may include more than 95% probability mass. In other words, the CI would give a conservative estimate of the range of upper and lower bounds to the TP threshold. The mean P threshold for U. purpurea was 14.8 µg l−1 TP with a 95% CI that ranged from 13.6 to 15.7 µg l−1 TP (Table 25.2). To examine the ecological imbalance across trophic levels, we completed a similar hierarchical n

610

C.J. Richardson et al.

Table 25.2 Results from Bayesian changepoint analysis on all trophic levels from the phosphorusdosing experiment in 1993–1998 (from Richardson et al. 2007) Metric

Maximum probabilities 95% CI

Lower and upper probabilities

Decrease 19.2b (216) Increase (180) 13.0b

0.88

18.4–20.0

0.003–0.000

0.54

9.4–15.8

0.010–0.011

Increase (180) 13.0b

0.77

9.8–19.4

0.004–0.004

Decrease (72) 12.4b

0.67

8.4–26.2

0.016–0.012

Decrease (72) 8.2 Increase (180) 19.9b Increase (108) 18.3b

0.96 0.83 0.99

7.9–8.7 16.5–20.7 17.1–18.3

0.003–0.664 0.000–0.000 0.000–0.000

Decrease (468) Decrease (432)

15.6b

0.94

15.0–15.9

0.002–0.002

14.8b

0.84

13.6–15.7

0.000–0.005

Decrease (108)

14.5b

0.91

12.1–22.2

0.003–0.003

Decrease (72) 23.5b

0.68

14.0–40.3

0.001–0.006

15.3b

0.75

12.0–22.3

0.004–0.007

Observed response (n)a

Mean changepoint

Community Mat cover Bray–Curtis (macroinvertebrates) Macroinvertebrates % Tolerant species % Sensitive species % Predators % Microcrustacea % Oligochaeta Macrophytes (stem counts) Total Utricularia species Utricularia purpurea Algae % Diatom on Eleocharis cellulosa % Diatom on Plexiglas % Diatom on floating mats

Decrease (144)

Changepoints and 95% credible intervals are based on the geometric means of total phosphorus in water a n is the number of sampling periods times 36 sample locations that displayed a significant TP changepoint b Number of P thresholds >10 µg l−1 TP

changepoint analysis over each season and year to identify biological attributes that were consistent metrics of ecological imbalance (Karr and Chu 1997). Ecological imbalance was calculated as a change in the mean and/or variance of the attribute response variable found above vs. below the biological changepoint with the highest maximum probability. At all trophic levels the composition, diversity, or population structure of the attribute were significantly altered above the TP threshold as compared to below the changepoint as shown, for example, in Fig. 25.5a–d. Importantly, the Bayesian hierarchical model takes into account the natural variation in the attributes and thus provides a robust probability estimate of the TP threshold. We then selected metrics at each trophic level with the highest

25 An Ecological Basis for Establishment

611

ecological value and probability threshold to calculate a mean annual TP threshold concentration for each metric (Table 25.2). This multimetric approach is based on the premise that while no one particular attribute of a biotic assemblage will always be a reliable indicator of imbalance, a suite of attributes used in combination will greatly increase our ability to detect impairment (Karr and Chu 1997; Barbour et al. 1995). Collectively, our selected attributes represent taxonomic composition, species richness and diversity, tolerance/intolerance, and trophic and community structure (Table 25.2). The observed responses among trophic levels demonstrated a remarkably similar pattern of response (Table 25.2). First, most metrics displayed a decrease in response with increasing TP, with the exceptions being the BCD index (Legendre and Anderson 1999), % tolerant species, and % microcrustacea species. Second, thresholds were almost all (11 out of 12 metrics or 92%) above 10 µg l−1 TP and below 20 µg l−1 TP, with macroinvertebrate predators displaying the lowest TP changepoints whereas diatoms growing on artificial Plexiglas substrates had the highest TP threshold. The most consistent TP threshold with the highest probability was for Utricularia spp. This metric showed very little variation as noted by the narrow CI (Table 25.2). The 95% intervals were the tightest for the metrics with the highest number of sampling points and varied greatly for the % diatoms on Plexiglas, which had only two sampling dates with detectable changepoints. To assess uncertainty for each of the selected thresholds, we evaluated the maximum probability and CIs for each metric. Providing an estimate of uncertainty is important because no clear consensus definition for imbalance of natural populations of flora and fauna exists in the ecological literature, nor does the “balance of nature” concept still hold as much favor in current ecological theory (Botkin 1991). The weakest maximum probabilities were found for BCD (0.54), % sensitive macroinvertebrate species (0.67), and % diatoms on Plexiglas (0.68) (Table 25.2). Again, this was probably related to the small number of sampling dates that showed a significant TP threshold for these metrics. Probabilities in excess of 0.90 were found for % Oligochaeta, % predators, total Utricularia spp., and % diatoms on Eleocharis stems. However, % predators showed a changepoint on only two sampling dates in the study. Maximum probabilities averaged 0.89, 0.84, 0.78, and 0.71 for the plant macrophytes, macroinvertebrates, algae, and community metrics, respectively. The metrics with both high probabilities and narrow CIs indicate thresholds that accurately reflect a major imbalance in the attributes above the changepoint thresholds for each group as compared to the attribute characteristics below the changepoints (Table 25.2). The probabilities that the threshold is at the lower or the higher extreme of the 95% CI were also tested by developing individual probability ratios of the selected changepoint to the lower and higher CIs. In all cases the probability that the changepoint is ≤ the lower CI or ≥ the upper 95% CI was < 0.01, with most metrics being < 0.005 (Table 25.2). Thus, the probability that the threshold is at the lower or upper 95% CI is close to 0. The overall mean changepoint was 15.6 µg l−1 TP, a value identical to the overall mean of all Utricularia spp. changepoints shown on Table 25.2. The BCD index had the lowest TP threshold (13 µg l−1 TP), lowest maximum probability (0.54), and a wide CI, while mat cover had the highest TP threshold (19.2 µg l−1 TP) with a

612

C.J. Richardson et al.

maximum probability of 0.88 and a narrow CI. The most robust ecological attribute over the entire 6-year period of the experiment was Utricularia spp., with a mean changepoint of 15.6 µg l−1 TP, a very narrow CI of 15.0–15.9, and a maximum probability of 0.94 (Table 25.2). The ecological importance of Utricularia spp. to the Everglades slough mat community (Davis and Ogden 1994; Vymazal et al. 1994) coupled with their highly predictable response to P additions (Vaithiyanathan and Richardson 1999; Table 25.2) suggest they may be both keystone species and highly sensitive indicators of the P threshold for the Everglades. The 95% CI range for all metrics varied from a low of 13 µg l−1 TP to a high of 19 µg l−1 TP. This span of 6 µg l−1 TP represents a measure of the uncertainty of the estimate, which in part may be due to the natural variation found in the TP thresholds across seasons, water depths, and years (Fig. 25.5a–d). Our dosing study results support our hypothesis that the TP threshold for the Everglades is best represented by a TP zone – not a single number – and that it is above 10 µg l−1 TP. Importantly, our results at each trophic level show a similar, well-defined TP threshold with a high probability of maintaining a balanced flora and fauna within the Everglades. Moreover, 92% of all TP thresholds tested are above 10 µg l−1 TP and below 20 µg l−1 TP.

25.6.2

Gradient Study

In regards to our experimental findings, it can be argued that the small spatial scale of our mesocosm study is not representative of the Everglades ecosystem (Rejmánková 2001; Koch and Reddy 1992; Richardson and Qian 1999; McCormick and Odell 1996). To test if our mesocosm results were representative of the TP gradient in the Everglades, we completed TP changepoint analyses for macroinvertebrates along a 10-km nutrient (TP, N, Ca, etc.) gradient (Richardson et al. 1999; Craft and Richardson 1993a; Noe et al. 2001; King 2001). Ten of the metrics evaluated using the observational P-gradient data showed clear responses to TP and were identified as potential metrics. Of these ten candidate metrics, five exhibited consistent responses to TP in the P-dosing experiment: BCD, % tolerant taxa, % sensitive taxa, % Oligochaeta (aquatic worms), and % predators. To provide an even more conservative estimate of the TP thresholds, we compared the median TP values of the dosing and gradient research. Results from correlation analysis among these 5 metrics indicated that no pair was collinear (r < 0.90); thus, each metric was sufficiently unique to retain as core metrics. These 5 metrics were subsequently analyzed individually and as an aggregated Nutrient–IBI using nCPA. Changepoints were detected for all selected metrics and the IBI using the observation P-gradient data (Fig. 25.6). Probabilities of Type I error (p < 0.001) were all quite low, indicating that it was highly likely that changepoints were real and represented a threshold response. The cumulative probability distributions generated from nCPA indicated that a changepoint was ≥ 50% probable between 12.6 and 19.4 µg l−1 TP for individual metrics and at 14.8 µg l−1 TP for the IBI (Fig. 25.6). These changepoints represented

25 An Ecological Basis for Establishment

613

Bray-Curtis Dissimilarity

Cumulative probability distributions P-Gradient Study

% Sensitive Taxa

5% % Tolerant Taxa

50% 95%

P-Dosing Experiment 5%

50% 95%

% Oligochaeta

% Predators

Nutrient Index of Biological Integrity (IBI)

0

10

20

30

40

50

Surface-water TP (µg/L)

Fig. 25.6 Synthesis of results from the P-gradient study and P-dosing study for the identification of a TP criterion protective of biological integrity. Median values from the four dates in the P-dosing experiment were used for the ≥5%, 50%, and 95% cumulative probabilities

biologically significant shifts in assemblage structure and function. Sensitive taxa dropped from a mean of over 21% to only 1.3% above 14.6 µg l−1 TP (Fig. 25.7). Conversely, tolerant taxa increased from only 2.2% to nearly 20% above 17.7 µg l−1 TP (Fig. 25.6). Percent Oligochaeta, a group of aquatic worms, nearly doubled when TP exceeded 13 µg l−1. Mean BCD values (nMDS Axis 1 scores) were highly negative to the left of the 50% probability of a changepoint, while highly positive to the right, indicating a markedly different species assemblage once a cumulative probability of 50% had been exceeded (Fig. 25.6). Elevated TP also resulted in functional changes, reducing the proportion of predators in the assemblages from a mean of 9.2% to only 3.4% at TP levels above 12.6 µg l−1. Finally, mean IBI scores above 14.8 µg l−1 were reduced by one-half when compared to IBI scores below that concentration (Fig. 25.8). Results from the P-dosing experiment mirrored those of the observed P gradient. Changepoints were evident for all metrics and the IBI, and they were repeatable through time. Means and variances of metric values above and below the 50% level of risk were on average lower in the dosing study as compared to the biologically significant changes observed along the P gradient, and highly suggested that the changepoints represented by the P-dosing study were a conservative estimate of the TP threshold (Fig. 25.6). This in part may be due to the fact that the dosing study used the most readily available form of phosphorus (soluble reactive phosphorus, SRP) additions and thus responses represent the most direct effect of P. The gradient

614

C.J. Richardson et al.

Fig. 25.7 Cumulative probability of a changepoint estimated for an individual metric in response to surface water TP. The cumulative probability curve describes the cumulative risk of a change in a response variable (% sensitive taxa, y-axis (right side); depicted by filled circles) associated with a range of stressor values. Cumulative probabilities are calculated using 1,000 bootstrap simulations. Any given location along the curve corresponds to a specific cumulative probability of a changepoint (y-axis (left side) ) at a specific level of TP (x-axis). In this example, there was at least a 5% cumulative probability, or risk, that a detectable change in the mean and/or variance of the % sensitive taxa metric occurred at or below 13.3 µg l−1 TP. In other words, ≥ 5% of the bootstrap simulations resulted in a changepoint that was ≤13.3 µg l−1 TP. Similarly, there was ≥ 50% risk of a changepoint ≤ 14.6 µg l−1 TP, while there was ≥95% probability that a changepoint occurred ≤16.9 µg l−1 TP. Data are from the observed P-gradient study

only received ~50% SRP with the other 50% of P additions being made up of nonreadily available particulate and organic P. The cumulative probability distributions of changepoints indicated that there was a relatively tight range of TP levels likely to result in degradation in biological condition. Both the observational and experimental data revealed that there was a low (typically ≤ 5%) probability that a threshold response occurred ≤ 10 µg l−1 TP for all metrics. There was high (≥ 95%) certainty that the threshold was ≤20 µg l−1 TP for the majority of individual metrics (Fig. 25.6). However, aggregating the individual metrics into the IBI reduced this range of variability (Fig. 25.8). Results indicated ≤ 0.05% probability that a threshold response for the IBI occurred at or below 9 (experimental) and 12.3 (gradient) µg l−1 TP, whereas there was ≥ 95% certainty that a threshold response occurred at ≤ 15 (experimental) and ≤17 (gradient) µg l−1 TP. Although these differences were relatively small, the fact that changepoints from the P-dosing experiment were lower than those in the observed P gradient suggested that changepoints from the experiment were conservative estimates of TP levels that may pose a risk to macroinvertebrate structure and function.

25 An Ecological Basis for Establishment

615

Fig. 25.8 Cumulative probabilities of changepoints for the Nutrient–IBI in response to surface water TP. Results are shown for the observational P-gradient study and the four dates from the P-dosing experiment. Note: The probability of the changepoint can be for selected in terms of acceptable risk as noted in Fig. 24.7

Uncertainty analysis indicated a low (≤ 5%) probability that an IBI threshold occurred at ≤10 µg l−1 TP, while there was ≥ 95% certainty that the threshold was ≤17 µg l−1 TP. The weight-of-evidence produced from the dosing and gradient analyses implies that the P threshold protective for all trophic levels would best be defined as a zone between 12 and 15 µg l−1, and a TP concentration >15 µg l−1 is likely to cause degradation of macroinvertebrate assemblage structure and function, a reflection of biological integrity, in the study area. This finding may assist in the development of a numerical estimate of the associated risk at each trophic level as well as an integrated estimate of risk with the combined hierarchical analysis and

616

C.J. Richardson et al.

the IBI estimate. Both methods give very similar results and provide a robust estimate of the TP threshold for the Everglades, demonstrating that the probability of the threshold being at or below 10 µg l−1 TP is at the lower end of the range for the northern Everglades areas we tested.

25.7 Conclusions and Lessons for Restoration We initially analyzed nearly 50 biological attributes over four levels of ecosystem organization to determine the most responsive ecological attributes to P additions in the water column. The responses were measured for either a positive or a negative imbalance from the control conditions over 6 years of SRP dosing and numerous seasons. The selection and testing of numerous metrics at several hierarchical levels within the ecosystem were done to provide comprehensive information that could be integrated to develop a scientific basis for a P threshold in the Everglades. Our results provide evidence that the P threshold for the Everglades, unlike that for lakes, varies with changing water depths. We further show that the P threshold for the Everglades is highly predictable across trophic levels, but estimates of uncertainty and confidence limits must be utilized to accurately determine the boundaries of biotic P thresholds. Our results further suggest that the TP threshold in the Everglades can be accurately quantified at each trophic level using Bayesian changepoint or nonparametric analysis techniques to (1) estimate numerical values for predictor variables indicating ecological imbalance in both flora and fauna directly from TP values in the water column and (2) provide an estimate of uncertainty by providing 95% CI. Providing an estimate of uncertainty is important since these threshold boundaries are not clearly defined and site variability must be taken into consideration when defining a threshold. To address the imbalance issues, we focused our research on determining the range of “P concentrations” that result in significant changes for 12 of the most robust biological indicators (i.e., metrics) at multiple trophic levels. Importantly we compared and calibrated the results from the dosing study with studies along a 30-year P gradient. In addition, we developed an IBI for the macroinvertebrates so that we could provide a combined TP threshold along with a risk assessment. Our research supports our hypothesis that the Everglades has the capacity to assimilate P slightly above background concentrations and that a P threshold protective for all trophic levels in the northern Everglades would best be defined as a zone between 12 and 15 µg l−1 due to seasonal and water level effects. While a TP concentration of 15 µg l−1 TP is a reasonable estimate of a TP concentration that will maintain a balance in the flora and fauna at the edge of the peatlands, some species, especially in the interior of the southern Everglades and ENP, may require a value closer to 10 µg l−1 TP. Here we define ecological imbalance as a significant increase or decrease in both the mean and/or the variance of response variables as

25 An Ecological Basis for Establishment

617

indicated by changepoints for all trophic levels. Importantly, our Bayesian approach represents a reliable and innovative way of quantifying ecological thresholds along gradients based on estimates of changes in both mean and population variance coupled with a probability analysis. Our method for the first time allows for a quantification of the breakpoints in ecological attributes across multiple trophic levels rather than reliance on indirect estimates of imbalance. The quantification of ecological thresholds from multiple trophic levels rather than reliance on arbitrary estimates of imbalance or single species responses provides for a robust estimate of the TP threshold, and it is applicable to others aquatic ecosystems. Importantly, our results from both the hierarchical analysis and the IBI converge on a similar TP threshold zone and a nonexceedance value of 15 µg l−1. Interestingly, a reading of the current US EPA and State of Florida-approved P criterion for the Everglades indicates that all measured sites must meet a 5-year geometric mean criterion of less than or equal to 10 µg l−1 P in 3 of 5 years, have annual concentrations less than or equal to 11 µg l−1 P across all stations, and have concentrations less than or equal to 15 µg l−1 P annually at all individual stations (SFWMD 2006). An important and not often appreciated fact is that our threshold zone (12-15 µg 1-1) and nonexceedance value of 15 µg l−1 P falls within the current range of values being utilized; moreover, it takes into account the natural, seasonal, and annual variations found within the Everglades ecosystem. Our approach developed here also presents a more powerful approach to quantifying the P threshold for other aquatic ecosystems. Moreover, our analysis clearly shows that the TP threshold is below the TP concentrations that are currently reported for the interior sites in WCA-2A of the Everglades (Chap. 2), and our analysis highlights the importance of understanding the natural decreasing TP nutrient gradient that existed historically from the exterior of the Everglades inward. Of course the final test of the Everglades response is now underway since the Stormwater Treatment Areas (STAs) are in place and are currently discharging TP values that often exceed 40 µg l−1 TP (SFWMD 2003, 2004, 2005, 2006). State of Florida data (see Fig. 2.17) do indicate that the interior sites in the ENP and WCA-3A meet the 5-year geometric mean criterion of less than or equal to 10 µg l−1 P across all sites. However, current inflow P concentrations into WCA1A, WCA-2A, and WCA-3A are still far in excess of the approved P criterion, and interior values for WCA-2A are 17 µg l−1 P (Fig. 2.17). More troubling are the high concentrations of P that are still flowing out of the WCAs and toward the ENP. The northern WCAs release far higher P concentrations than WCA-3A, but all values are well above the US EPA-approved P criterion and EPA standard of 10 µg l−1 P even though farm BMPs have been in place for over a decade and STAs are now in operation. As noted earlier (Chap. 2), if the present trend in P release concentrations continues and new BMPs are not utilized or additional STAs are not built to reduce current P overloading, then the Everglades will continue to receive unacceptable concentrations and loads of P for the foreseeable future. This will have significant consequences for the native biota and ecosystem structure and function.

Part V

Lessons for Restoration of the Everglades

26

An Ecological Approach for Restoration of the Everglades Fen Curtis J. Richardson

26.1

Introduction

One of the monumental environmental achievements in the US over the past 100 years was the consensus reached among local, state, and federal agencies – in addition to the agricultural, conservation, Native American, and development communities – to adopt and endorse the Comprehensive Everglades Restoration Plan (CERP). Everglades restoration is now well underway, but its success still rests to a major degree on our understanding and successful translation of scientific findings into meaningful restoration guidelines. However, restoration is a young science. Knowledge of how to restore wetland communities and ecosystem components with appropriate surface and subsurface hydrologic conditions and nutrient loadings is limited (Zedler 2000). Thus, disagreements on how to proceed will occur, and mistakes will be made, which means that we must be ready to “alter course” as new data on successes and failures become available. As I often say to my students, wetland restoration is not rocket science – it is far more difficult. The scientific information and recommendations in the previous 25 chapters hopefully will advance the scientific understanding of how the Everglades functions. The volume provides new data, synthesizes key findings and insights into current and past conditions in the Everglades, and quantifies the results of multiyear experiments on responses of plant and animal communities to P dosing, soil disturbance, vegetation removal, and hydrologic and nutrient gradients. Each of the research chapters provides a summary of key findings followed by suggested lessons that may help with restoration, especially in terms of community and organism responses. In addition, we highlight potential problem areas or environmental consequences that need to be considered as specific Everglades restoration plans are implemented. Supported by results from the Duke experiments and gradient studies, I now provide in this summary chapter a more comprehensive ecological approach along with science-based restoration recommendations that should dovetail nicely into the current adaptive planning approach used in CERP and recommended by Zedler (2005). Some of these plans and ideas can already be found in current restoration efforts; however, CERP is not a comprehensive, ecologically based restoration approach based on successional principles or peatland hydrodynamics.

621

622

C.J. Richardson

Given current constraints, what might the Everglades of the future look like? Can we restore the Everglades? Our analysis and the earlier findings of other scientists (e.g., Davis and Odgen 1994a; Porter and Porter 2002; Sklar and van der Valk 2002) make evident the fact that even the best plans, perfectly implemented, can never restore the Everglades to its pre-nineteenth-century area, hydrologic conditions, fire successional patterns, or diversity of plant and animal communities. We could not turn back the clock and restore the historic Everglades, even if we did not have to deal with the current massive increase of human populations with attendant urban and agricultural development. Moreover, the anthropogenic impacts on hydrologic flow and nutrient loadings during the last century have challenged our ability even to simply sustain the Everglades in its current condition. But even in the face of these challenges, scientific research has given us hope for revitalizing the remaining Everglades and restoring parts of it to historic conditions, but it will not be easy in a political atmosphere that supports uncontrolled growth of human populations, land development, and – most importantly – unchecked demands for water. Importantly, our research shows that understanding the factors controlling plant community structure and succession in the Everglades are central to any long-term restoration success. This ecologically based approach would allow us to better develop integrated water, fire, and nutrient management plans to maintain a diversity of Everglades plant and animal communities on the landscape as well as restore the critical types of peatland hydrodynamics needed to maintain habitat diversity (Chaps. 2 and 9). For example, in the thousand years before drainage ditches and peat oxidation took place, ombrotrophy may have been found in larger portions of the Everglades, but it is relevant today only for the raised center portion of the Loxahatchee Wildlife Refuge (WCA-1). Water diversions and drainage from deeply cut canals and ditches over the past century have caused massive peat subsidence. Large portions of the Everglades are now nourished by waters that have passed over or through calcareous mineral parent soils and are then released as point sources through canal and gate structures, not overland sheet flow (Chap. 7). Chemistry profiles and gradients found within the Everglades (Chap. 6) make it clear that portions of the Everglades are now in fact nourished by mineral groundwater and would be referred to as “minerogenous” or, in modern terms, “minerotrophic” systems. Historically, the Everglades fen would have had several types of dominant hydrologic systems (Chap. 2). For example, a “limnogenous” peatland that was located along Lake Okeechobee’s southern edge no longer exists because of the effects of the Hoover Dike built in the 1930s. The main hydrologic system where sloughs, tree islands, and aquatic communities existed in the central and southern Everglades would have been classified as “soligenous” peatlands with a minor slope and water flowing generally south. The originally “topogenous” central portion of WCA-1A (formed in a topographical depression) evolved into today’s raised-bog, “ombrotrophic” system (see Fig. 2.1). Thus, it is important to realize that the dominant water sources for various parts of the Everglades have not just simply evolved but, more importantly, have been greatly altered by the vast system of canals and dikes that have been built since the early 1900s. Today most

26 An Ecological Approach for Restoration of the Everglades Fen

623

of the Everglades hydrodynamics would be loosely classified as “soligenous”: however, in reality it is total a managed system and should be reclassified as “managenous” (having managed water flow) and the original flow systems cannot be replaced by bigger pumps and outflow structures (Figs. 2.10, 8.3). Thus, simply allowing water to flow over the land surface is not sufficient to restore and maintain Everglades communities. Future hydrologic management plans must take functional peatland hydrology into account because the different hydrodynamic factors and resulting nutrient status inherent in the various forms of water introductions control plant and animal community across the landscape. Clearly, hydrology is the main ecological driver in the Everglades. Hydrologic flow, hydroperiod, and overall hydropatterns interact with fire, drought, and nutrients to create a template for community development. Because climate, hydrology, fire, nutrients, and invasive species all influence the community structure and patterns that we see evolving in the Everglades today, we need a more integrated view of how these factors interact and control the biota and their communities. Unfortunately, only limited funding is being made available to study the interaction of these factors or how to integrate these findings into a more comprehensive ecological restoration plan. In addition, all original community types that existed in the more natural Everglades system prior to development should be represented in the restoration plans. Importantly, the environmental conditions under which these community types can be successfully reintroduced must also be understood. For example, restoration plans are underway for restoring tree island communities, now almost totally absent from some parts of the Everglades; however, not enough is known about the hydrologic, nutrient, and soil conditions needed to sustain this community type. For this reason, detailed experimental research is underway to try and understand the factors controlling these habitats (see Chap. 2; SFWMD 2006). However, research is not funded well enough to complete this vital work for all communities in the Everglades.

26.2

Target Areas for Community Restoration

Community distributions in the historic Everglades are estimated to have been ~60% sawgrass marsh, 39% slough, and 1% tree islands (Loveless 1959). Davis et al. (1994) estimated the predrainage areas for eight landscape types in the Everglades. Their data are converted into a percent area for the purposes of this discussion (Table 26.1). While the total area of the Everglades has been reduced by about 50%, the relative areas of the different community types have not been equally affected. For instance, three community types have experienced a 100% loss: swamp forest (including pond apple communities, not listed separately), peripheral wet prairie, and cypress stands. Although the sawgrass plains suffered a total loss of 175,000 ha, the relative loss was only 10% based on the current smaller area of the Everglades. This information can be used to develop target areas for the

624

C.J. Richardson

Table 26.1 Target total area based on the predrainage percent area multiplied by the current total area of the Everglades (618,000 ha) (from Davis et al. 1994) Landscape Current Current Predrainage Change in Target total type area (ha) area (%) area (%) area (%) area (ha) Swamp forest 0 0 5 −5 30,900 Sawgrass plains 63,000 10 20 −10 123,600 Slough/tree island/ 271,000 44 27 17 166,860 sawgrass mosaic Sawgrass-dominated 94,000 15 15 0 92,700 mosaic Peripheral wet prairie 0 0 10 −10 61,800 Cypress strand 0 0 1 −1 6,180 Southern marl-forming fens 190,000 31 21 10 129,780 Total area 618,000 100 99 NA 611,820 Note that percent areas have been rounded to the nearest percent. Note: If the target areas are already exceeded, as in the case of southern marl-forming fens, then targets have been reached and would only be reduced to restore an area that is currently extinct (provided hydrogeomorphic conditions are appropriate for the restoration of the extirpated Everglades community type)

different community types using the original percent area for each community type in the predrainage Everglades multiplied by the current total area. These estimates provide a general range, or target, of acceptable acreage for the different community types as shown on Table 26.1. Detailed site work (soils, hydrology, nutrients) would be needed to firmly establish the probability that a particular community could or should be established at any given location. Some community types have already exceeded target expectations. We are not necessarily advocating that these areas should all be reduced; in size however, the option should be considered if a location is conducive to new restoration requirements needed to recreate a lost or significantly reduced community type. For example, we do not advocate the loss of significant amounts of the present slough/tree island/sawgrass complex; however, current planned hydrologic changes in CERP may alter this community in specific areas, and thus some thought must be given as to what other community types can be maintained in these hydrologically altered areas. In addition to these considerations, it is was earlier suggested (Chap. 2) that the pond apple communities that originally occurred along the southern shores of Lake Okeechobee but were eliminated when the land was cleared for agriculture should be reestablished as part of a newly established limnogenous zone south of the lake. Since this community existed on high-nutrient mineral soils, suitable areas for its reestablishment may be found in the nutrient-enriched portions of the EAA. Our analysis (Table 26.1) also indicated that the sawgrass plain, especially dense sawgrass areas, should be doubled. Again, care must be taken to establish and maintain each community type based on historic conditions that allowed for their development, and the ecological requirements of each community type should be matched with planned hydrologic and nutrient conditions at each location within the remaining Everglades. Thus, it may or may not be possible to double this area once the hydrogeological conditions necessary to sustain this community are determined.

26 An Ecological Approach for Restoration of the Everglades Fen

625

Although the slough/tree island/sawgrass mosaic does not show a relative decrease in area, tree islands make up an important community type that should be restored in certain areas of the Everglades, such as WCA-2A (Chap. 8). Tree islands provide important habitats and refugia for various plants and wildlife. More recent studies have demonstrated the importance of tree islands in the Everglades, revealing that they are phosphorus “hot spots” on the landscape. The tree islands act as P reservoirs because roosting birds and predators transfer P there from surrounding low-concentration areas (Sklar and van der Valk 2002). The implications of the storage and release of high P concentrations from the tree islands are not well understood. Additional research is needed to determine what factors are important in establishing and maintaining these communities. Water Conservation Area 2A is the most disturbed area of the former Everglades (Chaps. 9 and 24), and it should be given a high priority for restoration. In a macrophyte survey of 55 sites throughout WCA-2A, C.J. Richardson et al. (unpublished data) found that cattail (>90% cover) was most abundant in plots in the northeastern section in areas close to the canal outflows (Chap. 9). In many of these plots, cattail was the main component of the vegetation. Plots further in west and south of the canal outflow structures had a lower percentage of cattail. The cattail abundance is generally proportional to soil total phosphorus (TP) levels. Areas where sawgrass has been replaced by cattail (see historic vegetation map, Fig. 2.2) should be given priority for sawgrass reestablishment. Furthermore, sloughs and wet prairies have often been replaced by cattail stands, especially in the northern portions of WCA-2A, and hydrologic conditions conducive to these communities need to be established once P loadings are reduced to levels that would allow native species to return. An analysis of the amount of P in the Everglades soils provides a starting point for determining potential restoration areas as well as areas susceptible to future cattail invasions. Qian and Richardson (Chap. 24) examined the total soil phosphorus concentrations in WCA-1, WCA-2A, WCA-3, and Everglades National Park (ENP) to determine the potential for cattail invasion in areas based on the soil TP content. Previous studies indicated that cattail frequency was higher in areas that had TP values over about 500 mg kg−1 (Chap. 9). At TP levels above about 1,000 mg kg−1, cattail frequency increases exponentially and cattail monocultures are present. By combining the information from these two studies, it is possible to identify areas that are conducive to future cattail invasion under current P concentrations (Table 26.2). Table 26.2 The percentage of the Everglades area within each soil phosphorus concentration category (top 10 cm of soil) (from Chap. 24) Location <500 (%) 500–750 (%) 750–1,000 (%) >1,000 (%) WCA-1 81 9 4 6 WCA-2A 49 36 8 7 WCA-3 90 10 0 0 ENPa 94 5 1 0 Note that areas are rounded to the nearest percent. P concentrations are in mg kg−1 a The lower portion of the ENP was not surveyed

626

C.J. Richardson

The soils in the study were classified as follows (1) background levels less than 500 mg kg−1 TP and low potential for cattail invasions, (2) slightly elevated levels of 500–750 mg kg−1, (3) elevated levels of 750–1,000 mg kg−1 TP and high potential for cattail invasion, and (4) highly eutrophic sites of 1,000 mg kg−1 or more that have a great potential for cattail monocultures. Under this classification, about 15% of WCA-2A contains soil TP levels highly conducive to the development of cattail monocultures. In addition, more than half of WCA-2A (51%) is predicted to be susceptible to cattail invasion given the current soil TP levels. Fortunately, in WCA-1, 81% of the area of has low susceptibility to cattail invasions and it is above 90% for both WCA-3 and sampled portions of ENP (Table 26.2). These data suggest that if soil P levels remain constant or P inputs are reduced, only a small percentage of ENP and WCA-3 will be susceptible to further invasion by cattail, assuming that hydrological conditions are carefully monitored and regulated. WCA-1 shows a higher percentage of areas susceptible to cattail invasion than WCA-3 and ENP. WCA-2A is by far the most likely candidate for cattail expansion, and the current cattail population supports this analysis (SFWMD 2000, 2006; see Chap. 9). Thus, when restoration efforts focus on removing cattail, the cattail populations in WCA-2A should be the first main target area.

26.3

Cattail Species Removal and Restoration of Community Types

Cattail (Typha) monocultures in the northern Everglades were not a dominant feature of the historical landscape. However, Davis (1943) documented an area of ~8,000 ha (20,000 acres) that was dominated by ferns and cattail in what is now EAA. The reason for this extensive acreage of cattail is unknown, but it was present prior to significant farming activities in the region (Fig. 2.2 in Chap. 2). It is possible that this community formed as a result of a nutrient release following a massive fire. The only other evidence for abundant cattail in the past occurred shortly after peat formation, around 4,000 YBP (Stone 2000). It may not be feasible to reduce Typha populations to what they were 100 years ago, but a major reduction of Typha from its present abundance is a reasonable goal, especially for WCA-2A. Again, some target levels are needed for each community type based on controlling factors in each area under the planned new water regimes.

26.3.1

The Role of Succession

The successional dynamics of the Everglades are mainly controlled by the interaction of climatic patterns (droughts and rainfall) and human alterations on hydroperiod, which in turn influences fire frequency and the degree of fire intensity as well as the transfer and release of P on the landscape (Chap. 2; Fig. 2.11). Knowledge of

26 An Ecological Approach for Restoration of the Everglades Fen

627

long-term climatic patterns is also important to understand changes within the Everglades. With the impending sea level rise due to global climate change projected to range from 19 to 59 cm or go even higher with significant ice melt by 2090 (IPCC 2007), saltwater becomes even more of a factor as it invades further and further into the southern Everglades and alters freshwater communities (Bartlett et al. 1995). More disturbing are the recent observations that since 1990 sea level has been rising faster than projected by IPCC models and if true the rise is 25% faster (3.3 ± 0.4 mm year−1; 1993–2006) in the last 20 years than the preceding 115 years (Rahmstorf et al. 2007). Although developed for the Big Cypress, the interaction model of hydroperiod and years since severe fire nicely describe the community stages of succession that one might expect in parts of the Everglades under normal hydrologic conditions and with no excessive inputs of nutrients (Duever et al. 1976; see Fig. 26.1). This model was based on earlier observations and studies by Davis (1943), Alexander and Crook (1973), Craighead (1971), and Hofsetter (1973). The major community types present at the origin of the seral stage are dependent on the duration of inundation, with more subtle habitat differences influenced by soil type. Thus, this model only provides a general template for the Everglades, and values for fire frequency and hydroperiod need to be confirmed for each community type. With a light fire or no fire soil peat accumulations (Fig. 2.12) will continue, and communities

Fig. 26.1 Successional patterns and rates of vegetation change in southern Florida plant communities as influenced by years since severe fire and hydroperiod (redrawn from Duever et al. 1976)

628

C.J. Richardson

will succeed from a fen slough habitat towards a more shrub/tree habitat. Fire, however, usually maintains the earlier successional stages. The importance of this hydroperiod/fire control model is that it provides a framework from which to develop water and fire management regimes to maintain each community type. Moreover, as new data come from integrated fire and water studies or observational data from burned areas, adjustments can be made to optimize conditions for each community type in a particular region. Whether appropriate water releases are possible to maintain the necessary hydroperiods and fire frequencies for each vegetation type in the Everglades is not known given the current human and agricultural demands for water and the point source deliveries given by pump stations and release structures. This means that the original soligenous water conditions will be very difficult to reestablish in many areas under current management plans. However, to create a more realistic restoration model for the future, the Everglades phosphorus concentration in the upper soil and floc layer needs to be incorporated into the successional model because it is the limiting nutrient for most of the Everglades. For example, the open water sloughs and wet prairie soils have phosphorus concentrations averaging around 550 mg kg−1, but once soils reach >1,000 mg kg−1, these sites convert to cattail monocultures (Chap. 9). Sawgrass marshes average 600 mg kg−1 in northern WCA-2A (Rivero et al. 2007), but they too convert to monoculture cattail stands at soil concentrations >1,000 mg kg−1. Willow and shrub swamps, hardwood swamps, and tree island sites averaged over 1,000 mg kg−1 of TP in the upper 10 cm, as did the cattail areas in WCA-2A (Rivero et al. 2007). This stored soil P has been shown to diffuse to surface waters and have long-term impacts on increasing water column P concentrations and, in turn, vegetation responses downstream (Chap. 6). Thus, a three-dimensional control model (hydroperiod, time since fire, and soil P concentrations) needs to be developed as a management template for Everglades restoration. Incorporating phosphorus into the model will allow for a more realistic plan for community habitat development and more importantly allow managers to focus first on using limited financial resources to restore those areas that project to have the highest success rate for community restoration and habitat stability. Key to this approach will be the selection of areas that can maintenance optimum hydroperiods, which will control fire regimes and new nutrient releases for each community type. Examples of how a hydroperiod/fire/nutrient model might work are given for a few communities in the next section. The actual requirements given below are only preliminary, and research should be continued to update and complete the three-dimensional control model.

26.3.2

Restoration of Pond Apple Communities

The northern, nutrient-rich section of the EAA along the lake may be a suitable locale for the reestablishment of the original pond apple community that once existed along the higher nutrient shores of Lake Okeechobee. This community was lost when the area was reclaimed for agricultural purposes. Duever et al. (1986)

26 An Ecological Approach for Restoration of the Everglades Fen

629

indicate that pond apple communities occur in areas where the hydroperiod is between about 240 and 360 days year−1 (Fig. 26.1) but where fire has not occurred for at least 10 years. Burning of this community (without a change in hydroperiod) would shift the vegetation back to a floating/emergent aquatic community. Without further burning, the pond apple community would eventually become dominated by hardwoods. Phosphorus concentrations in these relic communities have been shown to be extremely high (>1,000 mg kg−1 of TP); thus, currently enriched farm areas with more mineral soils (shallow peat) may be conducive to reestablishment of this community type. The restoration of this community will require further field research to determine appropriate management regimes for hydrology, nutrients, and fire. One difficulty will be the reestablishment of the limnogenous conditions that originally created this community because the Hoover Dike now prevents lake’s overflow. Research trials are needed to determine the optimum conditions for restoring and maintaining these communities.

26.3.3

Restoration of Wet Prairie

Wet prairies require a hydroperiod that is between that of a sawgrass marsh and a slough (Gunderson and Loftus 1993). Wet prairies burn infrequently; so extensive fires would not be useful in maintaining these communities. The best approach for restoration would be to maintain the appropriate hydroperiod (between 150 and 240 days year−1), maintain low P water concentration inputs (Chap. 25) and soil P concentrations < 500 mg kg−1 of TP, and prevent the invasion of exotics and cattail. Areas best suited for the reestablishment or maintenance of wet prairies include areas of southern WCA-3A and lower areas of WCA-2A and WCA-2B, provided appropriate hydroperiod conditions and nutrient concentrations can be maintained. Research trials are needed to determine the optimum conditions required to restore and maintain these communities.

26.3.4

Restoration of Sloughs

According to Duever et al. (1986), sloughs are inundated year-round and often occupy depressions in peat. Peat depressions are typically formed by fire that burns the top layers of peat (Fig. 2.12). Sloughs may also occupy depressions in mineral soils. Thus, slough restoration may utilize fire to initially create a peat depression. This would be followed up with a strict maintenance of wet season high water levels. Controlled burns and low P water inputs (Chap. 25) with soil P concentration of <500 mg kg−1 of TP may be necessary to maintain these communities, as Duever et al. (1986) have shown that sloughs may be invaded by wet prairie vegetation (emergent aquatics) in the absence of fire (Fig. 26.1) and cattails have invaded many areas with higher soil P (Chap. 9). Specific research is again needed on integrated

630

C.J. Richardson

water, fire, and soil conditions needed to restore and maintain these communities. Importantly, studies of hydrogeologic water sources are needed in most areas to determine the dominant water inputs (i.e., quantify the proportion of water received from rainfall, groundwater, or surface flow), because the sources control the type of vegetation that can be restored in each area.

26.3.5

Restoration of Tree Islands

The restoration of tree islands will likely take more time because paleoecological records show that conditions that result in quick shifts between other community types and tree islands are not common (Willard et al. 2001). Hofmockel, Richardson and Halpin (Chap. 8) and Worth (1988) found that the flooding that resulted from water impoundment led to a loss of tree islands, especially small islands, in the Everglades. Experimental drawdowns in areas where tree islands existed in the past did not result in the reemergence of tree island vegetation (Worth 1988). Current efforts are underway in WCA-1A to test appropriate conditions for tree island development (SFWMD 2006). Importantly, recent studies (e.g., Sklar and van der Valk 2002) have shown that restoration will probably require more than changes in hydrology. Much more research is needed on this topic before any more specific recommendations can be made.

26.3.6

Cattail Removal and Sawgrass Regeneration

In this section we discuss possible ways to remove cattail while at the same time creating suitable conditions for sawgrass establishment. To eradicate Typha populations, it is necessary to have an understanding of its basic physiology and ecology. Linde et al. (1976) recommend that control measures be taken when the plants are at a physiological weak point. They identified a window of opportunity during the growing season when carbohydrate reserves (stored in the rhizome) in Typha latifolia are low. This period was identified as the time prior to the beginning of pollination, when the pistillate spathe leaf emerges and sheds. At study site of Linde et al. (1976) in Wisconsin, this period was a relatively short period, spanning just 2 weeks. In the Everglades, low carbon reserves probably occur more frequently. Typha domingensis typically produces a lesser number of rhizomes than T. latifolia, thus carbohydrate reserves are lower (McNaughton 1966). However, T. domingensis generally produces a greater number of seeds, which would make it easier for this species to recolonize an area after removal. This could be another major problem for the Everglades, because most of the current Stormwater Treatment Areas (STAs) are nearly a monoculture of cattail. This means a greater seed source is being created than currently exists. Crocker (1938) found that cattail seeds survived 5.5 years of dry storage and these seeds had a 70% germination rate

26 An Ecological Approach for Restoration of the Everglades Fen

631

(Chap. 22). This underscores the importance of maintaining conditions that favor sawgrass growth even after cattail eradication not only in the restoration area but also in adjacent areas. Another difference between T. latifolia and T. domingensis found by McNaughton (1966) is that in T. domingensis, reactivation of meristem tissue at the base of shoots from the previous season’s growth is more common. Thus, it is important that eradication programs involve the destruction of shoot meristem tissue to prevent regrowth the following year. The most common Typha spp. removal alternatives include prescribed burning, chemical control, and physical control through a combination of cutting and flooding (Nelson and Dietz 1966; Weller 1975; Sale and Wetzel 1983; Apfelbaum 1985; Motivans and Apfelbaum 1987; Smith and Newman 2001; Kostecke et al. 2004; Pahl et al. 2004). Although chemical control can result in greater than 75% reduction in cattail stems (Nelson and Dietz 1966; Weller 1975), it is typically not an option in parks and protected wetland areas. Mechanical disturbance followed by submergence is the most effective strategy for short-term cattail control and can result in 100% mortality (Nelson and Dietz 1966; Weller 1975; Sale and Wetzel 1983; Pahl et al. 2004). Cattail shoots that are cut and then flooded quickly decay as they become oxygen-deprived and accumulate ethanol due to anaerobic respiration (Sale and Wetzel 1983). In contrast to the relative success of mechanical disturbance, the utility of prescribed burns alone for cattail control is ambiguous and depends on the historical fire regime. Whereas fire in some wetlands has resulted in reduced cattail biomass and the reestablishment of native species (Kostecke et al. 2004; Pahl et al. 2004), fire in other wetlands has increased P availability and contributed to cattail dominance (Smith and Newman 2001). The use of prescribed burns to control cattail expansion should be preceded by site-specific experimentation. Due to the vast acreage of Typha currently present in the Everglades, eradication may be best achieved (due to cost of mechanical removal) through removal by fire and subsequent replanting of Cladium, along with careful monitoring and modification of hydrologic conditions. It should be noted that Cladium did colonize high P soils from seed after a fire in WCA-2A and thus planting may not be required (Chap. 9). The timing of controlled burns should take into account the phenology of Typha discussed earlier as well as climatic considerations. The effects of recent fires in the Everglades may not be representative of what fires in the past were like. Wade et al. (1980) assert that prior to drainage water levels were typically at or above the ground surface except during drought years. However, it is now more common to find organic soil that dries enough to burn, resulting in the damaging muck fires that destroy even fire-adapted vegetation such as sawgrass. In unburned areas, it has been suggested that excessive litter accumulation in sawgrass marshes may result in sawgrass declines (Chap. 2). In contrast, Werner (1975) found that burning of decadent sawgrass stands stimulated recovery. The best regrowth was seen after January and February fires. Current fire suppression policies may have led to a decrease in surface fires, as these are easier to control (Gunderson and Snyder 1994), thus potentially harming sawgrass plains if surface fires are necessary for their maintenance.

632

C.J. Richardson

Wade et al. (1980) suggest that successful sawgrass management depends on the prevailing fire regime. They cite Florida Division of Forestry’s prescription for burning of sawgrass in the WCAs to maintain populations: 1. 2. 3. 4. 5.

A falling water table with surface levels of 7.6–15.2 cm (3–6 in.) An E to S wind at 16.1–24.1 km h−1 (10–15 mph) Relative humidity of less than 60% Stagnation index under 10 (smoke dispersion index) A 2:1 ratio of dead to live fuel (typically achieved by burning every 3–5 years)

Since cattail growth declines in the early dry season, fires during this period may be particularly useful in the eradication of cattail. Furthermore, cattail shoots are sensitive to freezing, so a burn during the cold, dry season may accelerate cattail loss. Using fire to maintain plant communities is not a new concept in south Florida. McCally (1999) points out that the aboriginal tribes that inhabited the south shores of Lake Okeechobee used fire to insure the proliferation of certain plants consumed in their diet. Thus, it is conceivable that fire, in combination with a carefully monitored hydrologic regime, will result in the successful restoration of sawgrass stands without negatively impacting other components of the ecosystem. One issue that still needs to be addressed is under what conditions sawgrass will naturally regenerate on sites that have been occupied by cattail. This will depend in large part on the viability of sawgrass seeds in the seed bank, and whether or not the seeds are too deeply buried. The regeneration of sawgrass following a fire in WCA-2A transect D (Chap. 9) suggests that seeds will sprout under appropriate conditions and vigorously recolonize the peatland even in areas with high soil P concentrations. However, fire will need to be employed ever 3–5 years to keep cattails from reinvading in areas with high residual P soil concentrations.

26.4

Invasive Species

An examination of past vegetative communities reveals that major changes have occurred, particularly during the last 50 years. In addition to the loss of certain community types, there has been invasion by both native (e.g., Typha and Lygodium) and nonnative taxa (e.g., Melaleuca and Schinus). Thus, in some areas of the WCAs, restoration plans will need to first include the eradication of certain invasive taxa that contributed to the displacement or disappearance of various community types. Because extensive programs are underway in south Florida to remove both Melaleuca and Schinus, we recommend a review of these program restoration guidelines. These plans are outlined in Thayer et al. (2000) and South Florida Water Management District reports (SFWMD 2003, 2004, 2005, 2006). The climbing fern Lygodium is an invasive taxon that deserves serious attention because it alters the natural fire ecology. For example, burning mats of the fern may be blown into the tree canopy and start a crown fire. Normally, crown fires are rare in the Everglades (Thayer et al. 2000). This climbing fern is spreading in WCA-2A, WCA-3A, and Big Cypress Park. Biological control measures are currently underway but are limited in scope (Thayer et al. 2000).

26 An Ecological Approach for Restoration of the Everglades Fen

633

Vaithiyanathan and Richardson (1999) found that sloughs in nutrient-enriched areas in WCA-2A had an increase in problem exotics such as water lettuce (Pistia stratiotes) and water hyacinth (Eichhornia crassipes). The native taxa that are normally common in unenriched sloughs, Utricularia purpurea and Eleocharis spp., showed a decline with the increase in exotics. Both water lettuce and water hyacinth are considered weed taxa in other countries. Importantly, a study by J. Zahina (unpublished Duke University Wetland Center report) demonstrated that both Pistia and Eichhornia would not survive in areas of low-nutrient water concentrations after only 2 months of being placed in a low P area. Unfortunately, present management and control practices for these species in south Florida consists mainly of the application of herbicides (Thayer et al. 2000).

26.5

Restoration Plans

In the proposed restoration plans (CERP 1999; SFWMD 2003, 2004, 2005, 2006) the concept of adaptive management is suggested as a means to reduce uncertainty in terms of biotic responses to the newly created hydrologic regimes and concomitant fire and nutrient regimes. However, it is clear that little is known about the dynamic interactions among hydrology, fire, and nutrients in terms of maintaining Everglades plant communities as noted earlier. Moreover, CERP plans, although extensive and based on science, do not address integrated peatland restoration concepts per se. Thus, a series of experiments is badly needed to address interaction effects of hydroperiod, fire, and nutrients to reduce uncertainty concerning the response of fen communities to restoration programs. In brief the steps to restore any areas should include: Step 1. Utilize maps of the current soil nutrient profiles, present vegetation, past vegetation, open water sloughs and channels, and complete detailed elevation maps. Although satellite data are available to map landscape patterns, these data need to be verified via ground truthing. Step 2. Establish target areas for each community type based on the information in Step 1. Step 3. Develop a proper hydrodynamic and hydroperiod regime for each region so that each peatland community type located within the area can be maintained. Step 4. Develop a mechanical removal or a burning-and-hydrologic plan for areas within cattail-dominated areas that will result in the removal of this species and enhance the desired target community. Step 5. Develop a set of experiments that together determine the effects of fire and hydrology on vegetation in nutrient-enriched and -unenriched environments. Although much work has been done on the direct effects of nutrients on vegetation, there has been little experimental work done on the joint effects of fire and hydrology. Additionally, there have not been experiments that address the interactive effects

634

C.J. Richardson

Fig. 26.2 Flowchart of potential successional paths of community development in phosphorusenriched Everglades communities subject to a surface fire under different hydrologic restoration scenarios

that these three variables (hydrology, nutrients, and fire) have on Everglades vegetation and animal communities. An example of a research plan for enriched areas, such as those found in WCA-2A is summarized in Fig. 26.2. These experiments involve a series of controlled burns under different nutrient levels and hydrologic regimes to remove cattail and restore native vegetation and habitats. This design is based on our work with fire management and Everglades communities (Richardson and Huvane 2001; Pahl et al. 2004). The key to this effort is to follow the state guidelines noted earlier (Wade et al. 1980) to prevent deep muck fires.

26.6

Summary of Research Results and Key Restoration Guidelines

To improve the focus of future restoration efforts, some of our key scientific findings and recommendations are summarized in this section and organized under one of three hierarchical levels: landscape, ecosystem, or community. Clearly many of the findings and recommendations span all three levels, but restoration efforts have usually focused on only one or two levels. The species or population levels are not listed separately because restoration for a single species or population is not usually successful unless the community or ecosystem is restored. The specific details of the research supporting these summary findings can be found in earlier chapters as

26 An Ecological Approach for Restoration of the Everglades Fen

635

noted. Our findings and recommendations are based on our work-to-date and thus some of these ideas and plans will no doubt need to be revised or replaced as new research becomes available – adaptive management at work.

26.6.1

Landscape Level

1. The Everglades should be managed as a fen, not a marsh or swamp. The formation of peatlands is totally different from that of marshes and swamps in terms of their hydrogeomorphic development, which has important hydrologic and nutrient management implications for successful restoration of this peatland complex (Chap. 2). 2. Functional peatland hydrology must be taken into account in any management plan. Sources of water provide the basis for maintaining differences in nutrient status and hydrodynamics, factors that control plant and animal community across the landscape (Chap. 2). 3. Phosphorus in rainfall is a major contributor to the overall P budget for the Everglades, overshadowed only in areas where agricultural runoff flows into the Everglades. The accuracy of P measurements in rainfall is not adequate at the present time; thus, estimates of total P loadings to the Everglades are in question. Research to improve estimates is badly needed (Chap. 2). 4. The current area of the STAs is not sufficient to reduce P concentrations going into the Everglades Protection Area to meet the State of Florida and EPAapproved criterion of 10 µg l−1 P or even the proposed threshold range of 1215 µg l−1 P for any one site (Chaps. 2 and 25).

26.6.2

Ecosystem Level

5. The important impact of SRP vs. total P in input, interior, and outflow areas is not often addressed in terms organism responses. Shifts in N:P ratios in the various components of the Everglades also need to be monitored carefully because high ratios of this key index indicate P limitation, a condition representative of the historic Everglades. Changes in this ratio can be used as an index of nutrient alterations (Chap. 2). 6. The key to restoring the Everglades ecosystem and the diversity of plant communities will be determining the hydrologic regimes necessary to maintain the dominant plant species in each community type. Unfortunately, little experimental information exists about the hydrologic conditions necessary to maintain most of the species, especially for those communities of higher plant diversity (Chap. 4). 7. The role of sulfate in the Everglades has been little appreciated, and efforts to reduce inputs levels need to be undertaken. This element has been shown to cause a severe drop in soil and water redox status, kill or restrict plant species

636

C.J. Richardson

once it is reduced to H2S, and greatly accelerate sulfur bacteria activity resulting in increased production of MeHg (Chap. 6). 8. The current hydrologic setting does not mimic the historical natural hydrologic setting. Particularly, the seasonal variations in water depths are no longer coincident with precipitation patterns. Spatially, hydroperiods show changes that probably would not have occurred under natural circumstances, and restoration efforts must be undertaken to restore these patterns (Chap. 7). 9. The hydrology is impacted not only by the magnitude of flow through the water control structures, but also by the distribution of open and closed water control structures. In extensive wetland systems such as the Everglades, the magnitude of the hydraulic gradient is very small; thus, even minute disturbances that affect water elevation can affect the direction and velocity of water flow (Chap. 7). −1 10. The average flow of water was ∼ − 200 m day in the open water areas, and – not surprisingly – the water flow in dense vegetation was nearly zero, with water leaving densely vegetated areas mostly by evapotranspiration. This indicates that large portions of the Everglades did not function as a “river of grass” in the last 1,000 years and thus hydrologic restoration efforts are needed to factor varying flow rates into water release plans (Chap. 7). 11. Most of the hydrologic information that has been collected by the State of Florida, the USGS, and the USACE is water depth data, which is often easier to collect than information on elevation or flow within the fen. However, elevation data are critical to determine water flow directions. It is very important in the Everglades to collect accurate elevation data for equipotential maps, and this should be done for all areas prior to any attempt to restore hydrologic conditions in the Everglades (Chap. 7). 12. Water impoundment has drastically reduced the quantity and quality of upland habitat in the northern Everglades, especially in WCA-2A. Fragmentation of the landscape has primarily affected small islands, leaving the impacted landscape with only large isolated large tree islands by the 1980s. The destruction of small islands with large edge-to-area ratios has negatively impacted wildlife, specifically wading birds that use hydrologic transition zones as foraging habitat. Additionally, loss of small islands has decreased heterogeneity and reduced connectivity of the landscape (Chap. 8).

26.6.3

Community Level

13. Succession in the Everglades is mostly controlled by hydrologic conditions and, in turn, fire frequency and intensity. The impact of excessive nutrients, especially P, is critical in some areas of the Everglades, as is the invasion of exotic species (Chap. 2). 14. The loss of several of the historical plant communities or landscape types in the drained Everglades presents a difficult challenge for future restoration efforts. For example, these losses make it very difficult under the current proposed

26 An Ecological Approach for Restoration of the Everglades Fen

637

hydrologic regimes to restore the swamp forest or maintain large areas of peripheral wet prairies due to inadequate hydroperiod conditions for the dominant species in these community types (Chap. 2). 15. Tree island density and patterns can be sentinels to the effectiveness of our restoration efforts (Chap. 8). 16. The large increase of nonnative species found in the Everglades is of concern. To have more than 10% of the identified taxa as nonnative presents another challenge for restoring the native plant communities. Of specific concern is hemp vine (Mikania scandens), an invasive that is virtually blanketing plant communities of the northern areas of WCA-2A, including cattail-dominated areas (Chaps. 4 and 9). 17. The current expansion of cattail and overall vegetation distribution patterns in WCA-2A and WCA-3A are the direct result of both increased phosphorus inputs and altered hydrologic and fire regimes, especially during the past three decades. The most important nutrient controlling cattail density and distribution is phosphorus, but there are indications that other ions like sodium and sulfate may play some role for many species (Chap. 9). 18. The key to vegetation management for the enriched portions of Everglades will be to reduce P loadings to levels that do not exceed its P assimilative capacity (PAC). It has been suggested that this P threshold may be below 0.5 g m−2 year−1 for TP in the Everglades. In areas where P inputs have been reduced to below 0.5 g m−2 year−1, we have not seen an expansion of cattail or loss of native macrophyte species (Chap. 9). 19. Fire management is another mechanism that may be useful in preventing the spread of cattail into sawgrass communities. Prescribed burning of sawgrass once every 3–5 years during the wet season may be effective in maintaining sawgrass communities against invasion by cattail, even in highly P-enriched soils. However, a great deal of research is needed to effectively understand the role of fire in maintaining Everglades plant communities (Chap. 9). 20. An analysis of the previous 10 years of hydrologic patterns suggested that the most important hydrologic pattern influencing cattail frequency was the number of consecutive days that a site had water levels at or below 5 cm of water depth. This finding, when combined with our experimental work, suggests that low water is first required for cattail invasions and that elevated P in the soil then acts as a catalyst to increase cattail dominance and expansion in these areas (Chap. 9). 21. Sawgrass populations can be maintained in areas where water depths are above 30 cm as long as sufficient drying periods also exist. Our studies clearly indicate that populations of sawgrass can be maintained by careful management of the hydrologic and fire regimes at all of the current soil nutrient regimes found in WCA-2A. Increased water depth >60 cm for extended periods in low P soil areas will result in decreased sawgrass populations and an increase in open water slough areas (Chap. 9). 22. The use of Plexiglas slides to characterize algal communities in the Everglades presents an overestimate of the role of diatoms and also underestimates the presence of the genus Scytonema, especially in short-term monitoring (e.g., 1 or 2 months).

638

C.J. Richardson

A Plexiglas substratum also results in many green filamentous algal species being found along a nutrient gradient in the Everglades as compared to analysis of natural substrates. Thus, the type of substrate must be taken into account when comparing algal communities and the use of biovolume is needed to assess the role of algal groups in community structure (Chap. 10). 23. From a bioassessment perspective, invertebrate assemblages are sensitive to nutrient enrichment, and they respond in predictable ways. Our results suggest that fine levels of taxonomic resolution (i.e., genus-level or species-level data) may be necessary for bioassessment to be accurate. We suggest that fixed counts ≥200 or integrated fixed-area/fixed-count approaches that consistently obtain a minimum of 200 individuals should be considered as minimum subsample sizes for wetland invertebrates (Chap. 11). 24. More paleoecological soil core analyses and broader calibration sets would facilitate our understanding of the extent of human-induced changes to the ecosystem. Laboratory experiments to determine pH and [P] optima of diatom species under controlled situations would contribute to the diatom calibration studies undertaken in the field. With broader coverage and more detailed studies, it may be possible to develop models to predict post-restoration plant communities based on planned restoration activities (Chap. 12). 25. Our findings suggest that managers will have to choose what conditions they want to restore in the Everglades. The water quality and hydrology within the WCAs were more uniform before impoundment than they are now. The dense patterns of calcareous periphyton growth in the P-unenriched slough areas of WCA-2A may be a relatively new phenomenon to the region due to the release of additional calcium from canal cuts in the bedrock. The future of algal communities in the managed Everglades depends on a better understanding of the relationship of proposed water management releases and water quality (Chap. 12). 26. Our 6-year phosphorus-dosing mesocosms provided realistic and uniform testing ranges for determining P thresholds. Importantly, our dosing results clearly indicate that in a P-limited system such as the Everglades SRP concentrations in the water collected during the latter part of the day may often be very low as compared to early morning or evening concentrations due to high P uptake driven by high midday photosynthetic rates. This suggests that low nighttime PO4 uptake rates may result in the downstream movement of higher concentrations of P at night. These findings indicate that further work on monitoring and release schedules needs to be undertaken to quantify potential diel effects on P uptake and downstream loss rates (Chap. 15). 27. Calcium carbonate precipitated from the water column by algae serves as a matrix, provides structure, and supports the physical integrity of the periphyton mat commonly observed in the oligotrophic Everglades. The mat covers nearly the entire surface of the open water sloughs during the summer season in certain areas of the Everglades. This process is considered to be the dominant mechanism responsible for the self-protection of these waters against eutrophication. Decrease in calcite precipitation due to P additions may be one of the factors responsible for the fragmentation of the periphyton mat observed in the

26 An Ecological Approach for Restoration of the Everglades Fen

639

mesotrophic areas of the Everglades and the complete disappearance of the periphyton mat in the eutrophic areas (Chap. 15). 28. A comprehensive analysis of the responses shows that a decrease in most aquatic plant populations is found at higher TP concentrations, especially above a 15 µg l−1 TP concentration. There are exceptions to this, e.g., Nymphaea spp., which shows an increase in population and cover with increased TP concentrations (Chap. 16). 29. The P-dosing results also suggest three patterns of response to PO4 enrichment in the sloughs (1) inhibition – as exemplified by U. purpurea, (2) stimulation – as suggested by Nymphaea response, and (3) an optimum response – as exemplified by Chara (Chap. 16). 30. With P additions, litter decomposition rates are increased, but large increases in net primary productivity apparently more than compensate for the increased decomposition rate in controlling the rate of peat accretion. The remarkable capacity for extremely P-deficient litter and peat to immobilize large quantities of P also works to remove PO4 from water; however, this results in areas with higher peat accretion but significantly elevated soil P concentrations, which alters native Everglades communities (Chap. 17). 31. Long-term P additions not only increased macroinvertebrate standing stocks and species richness significantly, but also altered the relative taxonomic and trophic structure of the assemblage. Dramatic responses often occurred at relatively low doses of P but diminished in magnitude with increasingly greater P loads, thus resulting in log-linear dose–response relationships for most attributes of the macroinvertebrate assemblage (Chap. 19). 32. Plant succession is a natural ecological process, but in areas that operate under anthropogenic controls and not by their typical historic conditions, disturbances can have more of a profound effect on the community and deflect plant communities from the usual pattern of succession. In an area of the Everglades that is not as greatly impacted by nutrient inputs, exotic invasions, or excessive airboat transportation, our disturbance experiment demonstrates a shift in plant species and cover dominance due to alterations in water depth, P additions, and vegetation removal (Chap. 21). 33. Based on our germination studies, two factors are contributing to the spread of cattail and its invasion of new areas. First, disturbance is providing open areas suitable for germination and seedling growth. Second, the high-nutrient inputs enable the seedlings to thrive, and eventually to produce hundreds of thousands of viable seeds, which are then dispersed throughout the Everglades to await favorable conditions. This positive-feedback cycle could threaten the Everglades with ever-increasing areas of cattail. Even if nutrient levels could be instantaneously returned to their former low state, the now enormous seed bank of cattail in the STAs will allow the continued increase of cattail any time a favorable site becomes available. Without major changes in water and nutrient management in the Everglades, large expanses of cattail are here to stay, and its removal as a dominant species in enriched areas will be nearly impossible (Chap. 22).

640

C.J. Richardson

34. Our nationwide analysis indicates that, given ample supplies of other nutrients, average PAC in North American wetlands is near 1 g m−2 year−1 with a confidence interval from 1.4 to 0.4 g m−2 year−1. However, the Everglades is a P-limited ecosystem, and biological response data suggest that the lower end of the PAC confidence interval (0.4 g m−2 year−1) be used to assure no significant changes in ecosystem structure and function (Chap. 23). 35. A P gradient will always exist in the Everglades downstream of inputs regardless of the concentration of P allowed to enter the Everglades (e.g., 30 or 10 µg l−1) because the release of P will be loaded to a small restricted area downstream of the outflow gates or structures; i.e., the load is area dependent and is based on the volume of water released times the concentrations of P in the water (Chap. 23). 36. Our P soil mapping provides a baseline comparison for future P changes in the Everglades. They indicate that P additions due to agriculture have primarily affected the northern Everglades (WCA-1A and WCA-2A), while WCA-3A and the ENP soils show little impact from P additions. Careful planning of new delivery systems that spread water over a larger surface area will help in the reduction of P mass loadings to specific sites to below their PAC. This reduction will result in less downstream movement of P deeper into the fen (Chaps. 6 and 24). 37. Our results suggest that the TP threshold in the Everglades can be accurately quantified at each trophic level using Bayesian changepoint or nonparametric analysis techniques to (1) estimate numerical values for predictor variables indicating ecological imbalance in both flora and fauna directly from TP values in the water column and (2) provide an estimate of uncertainty (Chap. 25). 38. Our research supports our hypothesis that the Everglades has the capacity to assimilate P slightly above background concentrations and that a TP concentration of 10 µg l−1 TP is a very conservative estimate of a TP concentration that will maintain a balance in the flora and fauna in the northern Everglades. However, interior portions of the Everglades, especially in the south may need concentrations nearer the 10 µg l−1 TP value to maintain natural communities. Analysis clearly shows that the P threshold is not a single number but rather a zone that ranges from 12 to 15 µg l−1 TP, because the threshold varies over wet and dry seasons and trophic levels. This threshold zone is within the range of the EPA-approved criterion. In our studies we define the threshold of ecological imbalance as a significant increase or decrease in both the mean and/or the variance of response variables as indicated by changepoints for all trophic levels. Our Bayesian threshold approach presents a powerful approach for quantifying the nutrient threshold in other aquatic ecosystems.

26.7

The Future of the Everglades

One word sums up the future of the Everglades: WATER. To successfully achieve water management goals, future restoration plans must consider the carrying capacity of the Everglades ecosystem while realizing that the value of Everglades is not

26 An Ecological Approach for Restoration of the Everglades Fen

641

just alligators, birds, and flowers – it is also the freshwater supply for south Florida. This will be more evident in the years ahead because south Florida’s current population of 6 million is projected to double over the next 50 years (Stevens 1999). Future demand for water resources is expected to exceed the available regional freshwater resources (USACE 1999). With current population growth the demand for fresh water could increase from the current 4.5 × 106 m3 (1 billion gallons) of water per day to 9.0 × 106 m3 (2 billion gallons) per day by 2050 (USACE 1999). While efforts are underway in CERP to increase the storage of water on the landscape, it is not clear whether in drought years enough water will be made available to sustain key communities in the Everglades. Thus, successful restoration of this ecosystem is contingent upon satisfying the many, often conflicting demands of people, agriculture, business, and biota. By many accounts, the CERP – a state–federal partnership to implement a $10.5 billion restoration plan covering 16 counties over 46,619 km2 (18,000 square miles) of south Florida and restore more natural water flow to 969,000 ha (2.4 million-acres) of marshes, revive habitat for more than 60 threatened and endangered species, and establish a reliable supply of water for millions of Floridians while providing flood control to the south Florida area – is well on the road to success. But how success is defined still varies greatly among the constituent parties. As mentioned in Chap. 2 there is now a great deal of concern about the “future balance of water in south Florida” in terms of what will be available in the future for Everglades restoration vs. agriculture or growing urban demands. In addition, the quality of the water that will be available has not been improved to the level hoped for, and our overall ability to manage such a large and complex set of ecosystems under the environmental extremes that occur quite often in south Florida is very difficult. In addition one can ask whether there will be the political will to save the Everglades when human pressures on water supply reaches a critical threshold. While I have every confidence that scientists will be available to help in developing sound management decisions and will give restoration guidelines, I have less confidence that urban populations will be willing to pay the cost of maintaining the Everglades unless they are educated about the true value of the ecosystem services they receive from this vast fen system. Finally, the outcome of restoration plans is of major global importance because, as it is often said, the Everglades is the sentinel wetland for the world. If we cannot get this restoration right with all our money, engineering technology, environmental laws, and ecological knowledge, then the future of wetlands worldwide is endangered. This statement alone challenges the US and Florida governments to preserve and restore one of the great wonders of the world. If done right, sustaining the ecological integrity of the Everglades will set a precedent for future efforts to sustain large natural wetland regions in the face of exponentially increasing human populations and environmental degradation. The Everglades plan was certainly the model used to help in restoring the Mesopotamian marshes of Iraq (Richardson et al. 2005; Richardson and Hussain 2006). I have every confidence that we can restore and will manage our only US subtropical fen for future generations – provided we use sound science to supply satisfactory

642

C.J. Richardson

water quality and quantity and delivery modes to sustain the ecosystem, and most importantly we communicate its full value to politicians and society. I hope the research findings and management recommendations given in this volume will aid in that endeavor.

Bibliography

Abrams, P.A. 1993. Effect of increased productivity on the abundances of trophic levels. Am. Nat. 141:351–371 Abrams, P.A. 1995. Monotonic or unimodal diversity–productivity gradients: what does competition theory predict? Ecology 76:2019–2027 Abtew, W. and N. Khanal. 1994. Water budget analysis for the Everglades agricultural area drainage basin. Water Resour. Bull. 30:429–439 Abtew, W., J. Obeysekera, M. Irisarry-Oritz, D. Lyons, and A. Reardon. 2003. Evapotranspiration estimation for South Florida. In World Water and Environmental Resources Congress: Proceedings of the Congress: 23–26 June 2003. Philadelphia, Pennsylvania, eds. S.M. Davis and J.C. Ogden. CD-Rom edition. Reston, VA: American Society of Civil Engineers Abtew, W., R.S. Huebner, and V. Ciuca. 2006. Hydrology of the South Florida environment. In 2006 South Florida Environmental Report, SFWMD (South Florida Water Management District), Chapter 5. West Palm Beach, FL: South Florida Water Management District Adorisio, C., C. Bedregal, S. Daroub, J. DeLeon, M. Edwards, C. Garvey, et al. 2006. Phosphorus controls for the Basins Tributary to the Everglades Protection Area. In 2006 South Florida Environmental Report, SFWMD (South Florida Water Management District), Chapter 3. West Palm Beach, FL: South Florida Water Management District Ahn, H. 1997. Detection and treatment of contaminated samples in atmospheric deposition data. In Proceedings of the Conference on Atmospheric Deposition into South Florida. Measuring Net Atmospheric Inputs of Nutrients, eds. G.W. Redfied and N.H. Urban, pp. 51–52. West Palm Beach, FL: South Florida Water Management District Alexander, T.R. and A.G. Crook. 1973. South Florida Ecology Study. Part 1. Recent and LongTerm Vegetation Changes and Patterns in South Florida. Coral Gables, FL: University of Miami. Report NTIS PB 231–939 Allen, T.F.H. and T.B. Starr. 1982. Hierarchy: Perspectives for Ecological Complexity. Chicago, IL: University of Chicago Press Allen, T.F.H., and E.P. Wyleto. 1983. A hierarchical model for the complexity of plant communities. J. Theor. Biol. 101:529–540 Amador, J.A. and R.D. Jones. 1993. Nutrient limitations on microbial respiration in peat soils with different total phosphorus content. Soil Biol. Biochem. 25:793–801 American Fisheries Society. 1990. Common and Scientific Names of Fishes from the United States and Canada, 5th edition. Bethesda, MD: American Fisheries Society Amrhein, J.F., C.A. Stow, and C.M. Wible. 1999. Whole-fish vs. filet PCB concentrations: an analysis using classification and regression tree (CART) models. Environ. Toxicol. Chem. 18:1817–1823 Anagnostidis, K. and J. Komárek. 1988. Modern approach to the classification system of cyanophytes 3-oscillatoriales. Arch. Hydrobiol./Suppl. 71, Algol. Stud. 50/53:327–472 Anderson, J.A. 1926. The influence of available nitrogen on the fermentation of cellulose in the soil. Soil Sci. 21:115–126 643

644

Bibliography

Anderson, R.O. and R.M. Neumann. 1996. Length, weight, and associated structural indices. In Fisheries Techniques, 2nd edition, eds. B.R. Murphy and D.W. Willis, pp. 447–482. Bethesda, MD: American Fisheries Society Anderson, M.P. and W.W. Woessner. 1992. Applied Groundwater Modeling: Simulation of Flow and Advective Transport. San Diego, CA: Academic Anderson, N.J., B. Rippey, and C.E. Gibson. 1993. A comparison of sedimentary and diatominferred phosphorus profiles: implications for defining pre-disturbance nutrient conditions. Hydrobiologia 253:357–366 Angelone, M. and C. Bini. 1992. Trace elements concentrations in soils and plants of Western Europe. In Biogeochemistry of Trace Metals, ed. D.C. Adriano, pp. 19–60. Boca Raton, FL: Lewis Apfelbaum, S.I. 1985. Cattail (Typha spp.) management. Nat. Areas J. 5(3):9–17 Apfelbeck, R.S. 1999. Development of Biocriteria for Wetlands in Montana. Helena, MT: Montana Department of Environmental Quality APHA (American Public Health Association). 1992. Standard Methods for the Evaluation of Water and Wastewater, 18th edition. Washington, DC: American Public Health Association Arrhenius, O. 1921. Species and area. Ecology 9:95–99 Avery, G.N. and L. Loope. 1980. Plants of Everglades National Park: A Preliminary Checklist of Vascular Plants. Homestead, FL: South Florida Research Center, Everglades National Park. Report T-558 Avery, G.B., Jr., C.T. Hackney, and A.N. Clark. 2001. Impact of increased salinity on the geochemistry of tidal wetlands: a model for sea level rise. Presentation at the 16th Biennial Conference of the Estuarine Research Federation (4–8 November 2001), St. Petersburg, FL Bailey, R.C., M.G. Kennedy, M.Z. Dervish, and R.M. Taylor. 1998. Biological assessment of freshwater ecosystems using a reference condition approach – comparing predicted and actual benthic invertebrate communities in Yukon streams. Freshwater Biol. 39(4):765–774 Balogh, S.J., M.L. Meyer, and K.K. Johnson. 1997. Mercury and suspended sediment loadings in the lower Minnesota River. Environ. Sci. Technol. 31:198–202 Balogh, S.J., D.R. Engstrom, J.E. Almendinger, and D.K. Johnson. 1999. History of mercury loading in the upper Mississippi River reconstructed from the sediments of Lake Pepin. Environ. Sci. Technol. 33:3297–3302 Barber, M.L. 2003. The effect of agricultural run-off on mercury biogeochemistry in the Everglades. Ph.D. Dissertation, Duke University Barbour, M.T., J.B. Stribling, and J.R. Karr. 1995. The multimetric approach for establishing biocriteria and measuring biological condition. In Biological Assessment and Criteria: Tools for Water Resource Planning and Decision Making, eds. W.S. Davis and T.P. Simon, pp. 63–80. Boca Raton, FL: Lewis Barbour, M.T., J. Gerritsen, G.E. Griffith, R. Frydenborg, E. McCarron, J.S. White, and M.L. Bastian. 1996. A framework for biological criteria for Florida streams using benthic macroinvertebrates. J. N. Am. Benthol. Soc. 15:185–211 Barbour, M.T., J. Gerritsen, B.D. Snyder, and J.B. Stribling. 1999. Rapid bioassessment protocols for use in streams and wadeable rivers: periphyton, benthic macroinvertebrates, and fish. Washington, DC: US Environmental Protection Agency, Office of Water. EPA 841-0B-99-002 Barnes, T. 2005. Caloosahatchee Estuary conceptual ecological model. Wetlands 25(4):884–897 Bartlett, D.S., K.B. Bartlett, J.M. Hartmann, R.C. Harriss, D.L. Sebacher, R. Pelletier-Travis, D.D. Dow, and D.P. Brannon. 1989. Methane emissions from the Florida Everglades: patterns of variability in a regional wetland ecosystem. Global Biogeochem. Cycles 3:363–374 Bartlett, K.B., R.S. Clymo, C.B. Craft, and C.J. Richardson. 1995. Non-coastal wetlands. In Intergovernmental Panel on Climate Change (IPCC) Working Group II Second Assessment Report, eds. M.G. Oquist and B.H. Svensson, Chapter 6. Paris: International Council of Scientific Unions Bartow, S.M., C.B. Craft, and C.J. Richardson. 1996. Reconstructing historical changes in Everglades plant community composition using pollen distributions in peat. J. Lake Reserv. Manage. 12:313–322

Bibliography

645

Bass, O.L. 1998. Small mammal population dynamics on tree islands. Oral Presentation at Tree Islands of the Everglades Symposium (July 14 and 15), Boca Raton, FL Bates, A.L., W.H. Orem, J.W. Harvey, and E.C. Spiker. 2002. Tracing sources of sulfur in the Florida Everglades. J. Environ. Qual. 31:287–299 Batzer, D.P. and V.H. Resh. 1991. Trophic interactions among a beetle predator, a chironomid grazer, and periphyton in a seasonal wetland. Oikos 60:251–257 Batzer, D.P., and S.A. Wissinger. 1996. Ecology of insect communities in nontidal wetlands. Annu. Rev. Entomol. 41:75–100 Bauer, R.L., K.H. Orvis, and E.G. Edlund. 1991. CalPalyn Pollen Diagram Program. Unpublished computer program, Pollen Laboratory, University of California at Berkeley Bedford, B.L. 1996. The need to define hydrologic equivalence at the landscape scale for freshwater wetland mitigation. Ecol. Appl. 6:57–68 Belanger, T.V., D.J. Scheidt, and J.R. Platko, II. 1989. Effects of nutrient enrichment on the Florida Everglades. J. Lake Reserv. Manage. 5:101–111 Bender, M.M. 1971. Variations in the 13C/12C ratios of plants in relation to the pathway of photosynthetic carbon dioxide fixation. Phytochemistry 10:1239–1244 Benke, A.C., A.D. Huryn, L.A. Smock, and J.B. Wallace. 1999. Length–mass relationships for freshwater macroinvertebrates in North America with particular reference to the southeastern United States. J. N. Am. Benthol. Soc. 18:308–343 Bennetts, R.E., M.W. Collopy, and J.A. Rodgers, Jr. 1994. The snail kite in the Florida Everglades: a food specialist in a changing environment. In Everglades: The Ecosystem and Its Restoration, eds. S.M. Davis and J.C. Ogden, pp. 507–532. Delray Beach, FL: St. Lucie Bennion, H., S. Juggins, and N.J. Anderson. 1996. Predicting epilimnetic phosphorus concentrations using an improved diatom-based transfer function and its application to lake eutrophication management. Environ. Sci. Technol. 30:2004–2007 Berner, L. and M.L. Pescador. 1988. The Mayflies of Florida, revised edition. Gainesville, FL: University of Florida Press van der Bijl, L., K. Sand-Jansen, and A.L. Hjermind. 1989. Photosynthesis and canopy structure of a submerged plant, Potamogeton pectinatus, in a Danish lowland stream. J. Ecol. 77:947–962 Binford, M.W. 1990. Calculation and uncertainty analysis of 210Pb dates for PIRLA project lake sediment cores. J. Paleolimnol. 3:253–267 Birks, H.J.B. and H.H. Birks. 1980. Quaternary Palaeoecology. Baltimore, MD: University Park Press Blake, N.M. 1980. Land into Water – Water into Land. Tallahassee, FL: University Presses of Florida Bloom, N.S. 1992. On the chemical form of mercury in edible fish and marine invertebrate tissue. Can. J. Fish. Aquat. Sci. 49:1010–1017 Bloom, N.S. and E.A. Crecelius. 1983. Determination of mercury in seawater at subnanogram per liter levels. Mar. Chem. 14:49 Bodle, M.J., A.P. Ferriter, and D.D. Thayer. 1994. The biology, distribution and ecological consequences of Melaleuca quinquenervia in the Everglades. In Everglades: The Ecosystem and Its Restoration, eds. S.M. Davis and J.C. Ogden, pp. 341–356. Delray Beach, FL: St. Lucie Bonnewell, V., W.L. Koukkari, and D.C. Pratt. 1983. Light, oxygen and temperature requirements of Typha latifolia seed germination. Can. J. Bot. 61:1330–1336 Botkin, D.B. 1991. Discordant Harmonies, a New Ecology for the Twenty-First Century. Oxford, UK: Oxford University Press Brandt, L.A. 1997. Spatial and temporal patterns of tree islands in the Arthur R. Marshall Loxahatchee National Wildlife Refuge. Dissertation. Gainesville, FL: University of Florida Braun-Blanquet, J. 1964. Pflanzensoziologie, 3rd edition. Berlin Heidelberg New York: Springer Bray, J.R. and J.T. Curtis. 1957. An ordination of the upland forest communities of southern Wisconsin. Ecol. Monogr. 27:325–349

646

Bibliography

Brazner, J.C. 1997. Regional, habitat, and human development influences on coastal wetland and beach fish assemblages in Green Bay, Lake Michigan. J. Great Lakes Res. 23:36–51 Breiman, L., J.H. Friedman, R.A. Olshen, and C.J. Stone. 1984. Classification and Regression Trees. New York, NY: Chapman and Hall Bridgham, S.D. and C.J. Richardson. 1993. Hydrology and nutrient gradients in North Carolina peatlands. Wetlands 13(3):207–218 Brinson, M.M. 1977. Decomposition and nutrient exchange of litter in an alluvial swamp forest. Ecology 58:601–609 Brinson, M.M. and R. Rheinhardt. 1996. The role of reference wetlands in functional assessment and mitigation. Ecol. Appl. 6:69–76 Brinson, M.M., A.E. Lugo, and S. Brown. 1981. Primary productivity, decomposition and consumer activity in freshwater wetlands. Annu. Rev. Ecol. Syst. 12:123–161 Brinson, M.M., F.R. Hauer, L.C. Lee, W.L. Nutter, R.D. Rheinhardt, R.D. Smith, and D.F. Whigham. 1995. A guidebook for application of hydrogeomorphic assessments to riverine wetlands. US Army Corps of Engineers Waterways Experiment Station, Vicksburg, MS. USA. Wetlands Research Program Technical. Report WRP-DE-11 Brix, H. 1999. Genetic diversity, ecophysiology and growth dynamics of reed (Phragmites australis). Aquat. Bot. 64:179–184 Brock, T.D. 1970. Photosynthesis of algal epiphytes of Utricularia in Everglades National Park. Bull. Mar. Sci. 20:952–956 Brookes, P.C., D.S. Powlson, and D.S. Jenkinson. 1982. Measurement of microbial biomass P in the soil. Soil Biol. Biochem. 14:319–329 Brooks, H.K. 1968. The Plio-Pleistocene of Florida with special references to the strata outcropping on the Caloosahatchee River. In Late Cenozoic Stratigraphy of Southern Florida: A Reappraisal, 2nd Annual Field Trip of the Miami Geological Society, ed. R.D. Perkins, pp. 3–64. Miami: Miami Geological Society Broome, S.W., E.D. Seneca, and W.W. Woodhouse, Jr. 1983. The effects of source, rate and placement of nitrogen and phosphorus fertilizers on growth of Spartina alterniflora transplants in North Carolina. Estuaries 6:212–226 Browder, J.A. 1982. Biomass and Primary Production of Microphytes and Macrophytes in Periphyton Habitats of the Southern Everglades. Homestead, FL: National Park Service, South Florida Research Center, Everglades National Park Browder, J.A., S. Black, P. Schroeder, M. Brown, M. Newman, D. Cottrell, et al. 1981. Perspective on the Ecological Causes and Effects of the Variable Algal Composition of Southern Everglades Periphyton, Homestead, FL: South Florida Research Center. Report T-643 Browder, J.A., D. Cotrrell, M. Brown, S. Newman, R. Edards, J. Yuska, et al. 1982. Biomass and Primary Production of Microphytes and Macrophytes in Periphyton Habitats of the Southern Everglades. Homestead, FL: National Park Service, South Florida Research Center, Everglades National Park. Report T-662 Browder, J.A., P.J. Gleason, and D.R. Swift. 1994. Periphyton in the Everglades: spatial variation, environmental correlates and ecological implications. In Everglades: The Ecosystem and Its Restoration, eds. S.M. Davis and J.C. Ogden, pp. 379–418. Delray Beach, FL: St. Lucie Brown, J.G. and A.D. Cohen. 1985. Palynologic and petrographic analyses of peat deposits, Little Salt Spring. Natl Geogr. Res. 1:21–31 Bruland, G.L., S. Grunwald, T.Z. Osborne, K.R. Reddy, and S. Newman. 2006. Spatial distribution of soil properties in Water Conservation Area 3 of the Everglades. Soil Sci. Soc. Am. J. 70:1662–1676 Brush, G.S. 1989. Rates and patterns of estuarine sediment accumulation. Limnol. Oceanogr. 34:1235–1246 Busch, E., W.F. Loftus, and O.L. Bass, Jr. 1998. Long-term hydrologic effects on marsh plant community structure in the southern Everglades. Wetlands 18:230–241 Bush, M.B. 1994. Amazonian speciation: a necessarily complex model. J. Biogeogr. 21:5–17 Bush, M.B and C.J. Richardson. 1995. Phophatase as a biochemical indicator of phosphorus availability. In Effects of Phosphorus and Hydroperiod Alterations on Ecosystem

Bibliography

647

Structure and Function in the Everglades, eds. C.J. Richardson, C.B. Craft, R.G. Qualls, J. Stevenson, P. Vaithiyanathan, M. Bush, and J. Zahina. Chapter 8. Durham, NC: Nicholas School of the Environment, Duke University, Duke Wetland Center Publication 95-05 Bushaw, K.L., R.G. Zepp, M.A. Tarr, D. Schulz-Jander, R.A. Bourbonniere, R.E. Hodson, W.L. Miller, D.A. Bronk, and M.A. Moran. 1996. Photochemical release of biologically available nitrogen from aquatic dissolved organic matter. Nature 381:404–407 Cairns, J., Jr. (ed.). 1982. Artificial Substrates. Ann Arbor, MI: Ann Arbor Science Publishers Cairns, J., Jr. 1983. Are single species tests alone adequate for estimating hazard? Hydrobiologia 137:271–278 Cairns, J., Jr. and P.V. McCormick. 1991. The use of community- and ecosystem-level end points in environmental hazard assessment: a scientific and regulatory evaluation. Environ. Audit. 2(4):239–248 Cairns, J., Jr. and J.R. Pratt. 1986. Editorial on the relation between structural and functional analyses of ecosystems. Environ. Toxicol. Chem. 5:785–786 Campeau, S., H.R. Murkin, and R.D. Titman. 1994. Relative importance of algae and emergent plant litter to freshwater marsh invertebrates. Can. J. Fish. Aquat. Sci. 51:681–692 Cao, Y., D.D. Williams, and N.E. Williams. 1998. How important are rare species in aquatic ecology and bioassessment? Limnol. Oceanogr. 43:1403–1409 Carlin, B.P., A.E. Gelfand, and A.F.M. Smith. 1992. Hierarchical Bayesian analysis of changepoint problems. Appl. Stat. 41:389 Carpenter, S.R., D. Ludwig, and W. Brock. 1999. Management of lakes subject to potentially irreversible change. Ecol. Appl. 9(3):751–771 Carter, M.R., L.A. Burns, T.R. Cavinder, K.R. Dugger, P.L. Fore, D.B. Hicks, H.L. Revells, and T.W. Schmidt. 1973. Ecosystem Analysis of the Big Cypress Swamp and Estuaries. Atlanta, GA: US EPA 904/9-74-002, Region IV CERP (Comprehensive Everglades Restoration Plan). 1999. Central and Southern Florida Project Comprehensive Review Study. Final Integrated Feasibility Report and Programmatic Environmental Impact Statement. April 1999. CD-Rom edition. Jacksonville, FL/West Palm Beach, FL: US Army Corps of Engineers/South Florida Water Management District CERP (Comprehensive Everglades Restoration Plan). 2005. 2005 Report to Congress. (n.p.): US Army Corps of Engineers, http://www.evergladesplan.org/pm/program_docs/cerp_report_ congress_2005.cfm (accessed 1 May 2006) CERP (Comprehensive Everglades Restoration Plan), Welcome to the official website of the Comprehensive Everglades Restoration Plan (CERP), US Army Corps of Engineers, South Florida Water Management District, http://www.evergladespplan.org/index.aspx (accessed 21 March 2007) Chang, S.C. and M.L. Stewart. 1957. Fractionation of soil phosphorus. Soil Sci. 84:133–144 Chanton, J., G.J. Whiting, J.D. Happell, and G. Gerard. 1993. Contrasting rates and diurnal patterns of methane emission from aquatic macrophytes. Aquat. Bot. 46:111–128 Chapin, F.S., III. 1980. The mineral nutrition of wild plants. Annu. Rev. Ecol. Syst. 11:233–260 Chapin, F.S., III. 1991. Integrated responses of plants to stress. BioScience 41:29–36 Chen, Z. and T. Fontaine. 1997. Simulating the impact of rainfall phosphorus inputs on the phosphorus concentrations in the Everglades Protection Area. In Proceedings of the Conference on Atmospheric Deposition into South Florida. Measuring Net Atmospheric Inputs of Nutrients, eds. G.W. Redfied and N.H. Urban, p. 19. West Palm Beach, FL: South Florida Water Management District Chen, M., S.H. Daroub, T.A. Lang, and O.A. Diaz. 2006. Specific conductance and ionic characteristics of farm canals in the Everglades Agricultural Area. J. Environ. Qual. 35:141–150 Chiang, C., C.B. Craft, and C.J. Richardson. 1994. Fertilizer study – the effects of nutrient additions on photosynthesis by cattail (Typha domingensis Pers.) and sawgrass (Cladium jamaicense Crantz) in the Everglades. In Annual Report: Effects of Nutrient Loadings and Hydroperiod Alterations on Control of Cattail Expansion, Community Structure and Nutrient Retention in the Water Conservation Areas of South Florida, eds. C.J. Richardson, C.B. Craft, R.G. Qualls,

648

Bibliography

J. Stevenson, and P. Vaithiyanathan, pp. 49–71. Durham, NC: School of the Environment, Duke University, Duke Wetland Center Publication 94-08 Chiang, C., C.B. Craft, D.W. Rogers, and C.J. Richardson. 2000. Effects of 4 years of nitrogen and phosphorus additions on Everglades plant communities. Aquat. Bot. 68:61–78 Childers, D.L., R.F. Doren, R. Jones, G.B. Noe, M. Rugge, and L.J. Scinto. 2003. Decadal change in vegetation and soil phosphorus pattern across the Everglades landscape. J. Environ. Qual. 32:344–362 Cholnoky, B.J. 1968. Die Ökologie der Diatomeen in Binnengewassern. Lehre: J. Cramer Clark, L.A. and D. Pregibon. 1992. Tree-based models. In Statistical Models, eds. J.M. Chambers and T.J. Hastie, pp. 377–419. Pacific Grove, CA: S. Wadsworth and Brooks Clarke, K.R. 1993. Non-parametric multivariate analyses of changes in community structure. Aust. J. Ecol. 18:117–143 Cleckner, L.B., C.C. Gilmour, J.P. Hurley, and D.P. Krabbenhoft. 1999. Mercury methylation in periphyton of the Florida Everglades. Limnol. Oceanogr. 44(7):1815–1825 Cleveland, W.S. 1993. Visualizing Data. New Jersey: Hobart Coale, F.J. 1994. Sugarcane production in the EAA. In Everglades Agricultural Area (EAA). Water, Soil, Crop and Environmental Management, eds. A.B. Bottcher and F.T. Izuno, pp. 224–237. Gainesville, FL: University Press of Florida Cohen, J.E., F. Briand, and C.M. Newman. 1990. Community Food Webs: Data and Theory. Berlin Heidelberg New York: Springer Colinvaux, P.A. 1992. Ecology, 2nd edition. New York: Wiley Conley, D.J. 1988. Biogenic silica as an estimate of siliceous microfossil abundance in Great Lake sediments. Biogeochemistry 6:161–179 Connell, J.H. 1975. Some mechanisms producing structure in natural communities: a model and evidence from field experiments. In Ecology and Evolution of Communities, eds. M.L. Cody and J.M. Diamond, pp. 460–490. Cambridge, MA: Belknap Connell, J.H. 1978. Diversity in tropical rainforests and coral reefs. Science 199:1302–1310 Conway, V.M. 1936. Studies on the autecology of Cladium mariscus R. Br. I. Structure and development. New Phytol. 35:177–204 Cooke, C.W. 1939. Scenery of Florida, interpreted by a geologist. Fla. Geol. Surv. Bull. 17:1–118 Cooke, J.G., L. Stub, and N. Mora. 1992. Fractionation of phosphorus in the sediment of a wetland after a decade of receiving sewage effluent. J. Environ. Qual. 21:726–732 Cooper, S.R. 1993. A 2,500-year history of anoxia and eutrophication in Chesapeake Bay. Estuaries 16:617–626 Cooper, S.R. 1995. Chesapeake Bay watershed historical land use: impacts on water quality and diatom communities. Ecol. Appl. 5:703–723 Cooper, R.M. and J. Roy. 1991. An Atlas of Surface Water Management Basins in the Everglades: The Water Conservation Areas and Everglades National Park. DRE 300. West Palm Beach, FL: Water Resources Department of Research and Evaluation, South Florida Water Management District Cooper, S.R., J. Huvane, P. Vaithiyanathan, and C.J. Richardson. 1999. Calibration of diatoms along a nutrient gradient in Florida Everglades Water Conservation Area 2A, USA. J. Paleolimnol. 22:413–437 Correll, D.L. 1998. The role of phosphorus in the eutrophication of receiving waters: a review. J. Environ. Qual. 27:261–266 Courtemanch, D.L. 1996. Commentary on the subsampling procedures used for rapid bioassessments. J. N. Am. Benthol. Soc. 15:381–385 Courtney, G.W., H.J. Teskey, R.W. Merritt, and B.A. Foote. 1996. Aquatic Diptera, Part 1. Larvae of aquatic Diptera. In An Introduction to the Aquatic Insects of North America, 3rd edition, eds. R.W. Merritt and K.W. Cummins, pp. 484–514. Dubuque, IA: Kendall-Hunt Cowardin, L.M., V. Carter, F.C. Golet, and E.T. LaRoe. 1979. Classification of wetlands and deepwater habitats of the United States. Washington, DC: US Fish and Wildlife Service, Department of the Interior

Bibliography

649

Craft, C.B. and C.J. Richardson. 1993a. Peat accretion and nitrogen, phosphorus and organic carbon accumulation in nutrient-enriched and unenriched Everglades peatlands. Ecol. Appl. 3:446–458 Craft, C.B. and C.J. Richardson. 1993b. Peat accretion and phosphorus accumulation along a eutrophication gradient in the northern Everglades. Biogeochemistry 22:133–156 Craft, C.B. and C.J. Richardson. 1994. Response of Everglades plant communities to nitrogen and phosphorus additions. In Effect of Nutrient Loadings and Hydroperiod Alterations on Control of Cattail Expansion, Community Structure and Nutrient Retention in the Water Conservation Areas of South Florida, ed. C.J. Richardson, pp. 9–48. Durham: Duke University Wetland Center Publication 94-08 Craft, C.B. and C.J. Richardson. 1997. Relationships between soil nutrients and plant species composition in Everglades peatlands. J. Environ. Qual. 26:224–232 Craft, C.B. and C.J. Richardson. 1998. Recent and long-term organic soil accretion and nutrient accumulation in the Everglades. Soil Sci. Soc. Am. J. 62:834–843 Craft, C.B., J. Vymazal, and C.J. Richardson. 1995. Response of Everglades plant communities to nitrogen and phosphorus additions. Wetlands 15:258–271 Craighead, F.C., Sr. 1971. The Trees of South Florida. Vol. 1: The Natural Environments and Their Succession. Coral Gables, FL: University of Miami Press Craighead, F.C., Sr. 1984. Hammocks of South Florida. In Environments of South Florida, Present and Past II, ed. P.J. Gleason, pp. 53–56. Miami, FL: Miami Geological Society Cressie, N.A.C. 1991. Statistics for Spatial Data. New York: Wiley Crocker, W. 1938. Life span of seeds. Bot. Rev. 4:235–272 CSSS (Canadian Society of Soil Science). 1993. Soil Sampling and Methods of Analysis, ed. M.R. Carter. Boca Raton, FL: Lewis Daigle, J.J. 1991. Florida Damselflies (Zygoptera): A Species Key to the Aquatic Larval Stages. Technical Series Vol. 11, No. 1. Tallahassee, FL: State of Florida Department of Environmental Regulation Daigle, J.J. 1992. Florida Dragonflies (Anisoptera): A Species Key to the Aquatic Larval Stages. Technical Series Vol. 12, No. 1. Tallahassee, FL: State of Florida Department of Environmental Regulation van Dam, H. 1982. On the use of measures of structure and diversity in applied diatom ecology. Nova Hedwigia 73:97–115 Danielson, T.J. 1998. Wetland bioassessment fact sheets. Washington, DC: US Environmental Protection Agency, Office of Wetlands, Oceans, and Watersheds, Wetlands Division. EPA 843-F-98-001 David, P.G. 1996. Changes in plant communities relative to hydrologic conditions in the Florida Everglades. Wetlands 16:15–23 Davis, J.H. 1943. The natural features of southern Florida. Fla. Geol. Surv. Bull. 25:1–311 Davis, J.C. 1986. Statistics and Data Analysis in Geology, 2nd edition. New York: Wiley Davis, S.M. 1989. Sawgrass and cattail production in relation to nutrient supply in the Everglades. In Freshwater Wetlands and Wildlife, CONF-8603101. DOE Symposium Serial No. 61, eds. R.R. Sharitz and J.W. Gibbons, pp. 325–341. Oak Ridge, TN: USDOE Office of Scientific and Technical Information Davis, S.M. 1991. Growth, decomposition and nutrient retention of Cladium jamaicense Crantz and Typha domingensis Pers. in the Florida Everglades. Aquat. Bot. 40:203–224 Davis, S.M. 1994. Phosphorus inputs and vegetation sensitivity in the Everglades. In Everglades: The Ecosystem and Its Restoration, eds. S.M. Davis and J.C. Ogden, pp. 357–378. Delray Beach, FL: St. Lucie Davis, S.M. and J.C. Ogden (eds.). 1994a. Everglades: The Ecosystem and Its Restoration. Delray Beach, FL: St. Lucie Davis, S.M. and J.C. Ogden. 1994b. Toward ecosystem restoration. In Everglades: The Ecosystem and Its Restoration, eds. S.M. Davis and J.C. Ogden, pp. 769–796. Delray Beach, FL: St. Lucie

650

Bibliography

Davis, S.M., L.H. Gunderson, W.A. Park, J.R. Richardson, and J.E. Mattson. 1994. Landscape dimension, composition, and function in a changing Everglades ecosystem. In Everglades: The Ecosystem and Its Restoration, eds. S.M. Davis and J.C. Ogden, pp. 419–444. Delray Beach, FL: St. Lucie DeAngelis, D.L. 1994. Synthesis: spatial and temporal characteristics of the environment. In Everglades: The Ecosystem and Its Restoration, eds. S.M. Davis and J.C. Ogden, pp. 307–320. Delray Beach, FL: St. Lucie DeBusk, W.F. and K.R. Reddy. 1998. Turnover of detrital organic carbon in a nutrient-impacted Everglades marsh. Soil Sci. Soc. Am. J. 62:1460–1468 DeBusk, W.F., K.R. Reddy, M.S. Koch, and Y. Wang. 1994. Spatial distribution of soil nutrients in a northern Everglades marsh: water conservation area 2A. Soil Sci. Soc. Am. J. 58:543–552 Delettre, Y.R. and N. Morvan. 2000. Dispersal of adult aquatic Chironomidae (Diptera) in agricultural landscapes. Freshwater Biol. 44:399–411 Delfino, J.J., T.L. Crisman, J.F. Gottgens, B.E. Rood, and C.D.A. Earle. 1993. Spatial and temporal distribution of mercury in Everglades and Okefenokee wetland sediments. Final Report to South Florida Water Management District, US Geological Survey and Florida Department of Environmental Regulation. Gainesville, FL: University of Florida, Department of Environmental Engineering Sciences Dineen, J.W. 1972. Life in the tenacious Everglades. Central and Southern Flood Control District, West Palm Beach, Florida, In-Depth Report 1:1–10 Dineen, J.W. 1974. Examination of water management alternatives in conservation area 2A. Depth Report, Central and Southern Florida Flood Control District, West Palm Beach, Florida, In-Depth Report 2:1–10 Dixit, S.S. and J.P. Smol. 1994. Diatoms as indicators in the environmental monitoring and assessment program-surface waters (EMAP-SW). Environ. Monit. Assess. 31:275–306 Dixit, S.S., B.F. Cumming, H.J.B. Birks, J.P. Smol, J.C. Kingston, A.J. Uutala, et al. 1993. Diatom assemblages from Adirondack lakes (New York, USA) and the development of inference models for retrospective environmental assessment. J. Paleolimnol. 8:27–47 Dodson, S.I. and D.G. Frey. 1991. Cladocera and other Branchiopoda. In Ecology and Classification of North American Freshwater Invertebrates, eds. J.H. Thorp and A.P. Covich, pp. 723–786. San Diego, CA: Academic Dohrenwend, R.E. 1977. Evapotranspiration patterns in Florida. Fla. Sci. 40:184–192 Doren, R.F., T.V. Armentano, L.D. Whiteaker, and R.D. Jones. 1997. Marsh vegetation patterns and soil phosphorus gradients in the Everglades ecosystem. Aquat. Bot. 56:145–163 Douglas, M.S. 1947. The Everglades: River of Grass. New York: Ballantine Draper, N.R. and H. Smith. 1981. Applied Regression Analysis. New York: Wiley Drouet, F. 1968. Revision of the Classification of the Oscillatoriaceae. Monograph 15. Philadelphia, PA: The Academy of Natural Sciences of Philadelphia Drouet, F. 1981. Revision of the Stigonemataceae with a summary of the classification of the blue-green algae, Nova Hedwigia 66:1–221 Du Rietz, G.E. 1954. Die Mineralbodenwasserzeigergrenze als Grundlage einer natürlichen Zweigliederung der nord-und mitteleuropäischen Moore. Vegetatio 5–6:571–585 Duever, M.J., J.E. Carlson, L.A. Riopelle, L.H. Gunderson, and L.C. Duever. 1976. Ecosystem analyses at Corkscrew Swamp. In Cypress Wetlands for Water Management, Recycling, and Conservation, eds. H.T. Odum, K.C. Ewel, J.W. Ordway, and M.K. Johnston, pp. 707–737. Gainesville, FL: Center for Wetlands Duever, M.J., J.E. Carlson, J.F. Meeder, L.C. Duever, L.H. Gunderson, L.A. Riopelle, et al. 1986. The Big Cypress National Preserve, pp. 32–33. New York: National Audubon Society. Research Report No. 8 Dufrêne, M. and P. Legendre. 1997. Species assemblages and indicator species: the need for a flexible asymmetrical approach. Ecol. Monogr. 67:345–366 DUWC (Duke University Wetland Center). 1995. Comprehensive quality assurance plan for Quality Assurance Section, Florida Department of Environmental Protection. Durham, NC: Duke University Wetland Center (COMPQAP 950381)

Bibliography

651

Eakins, J.D. and R.T. Morrison. 1978. A new procedure for the determination of Lead-210 in lake and marine sediments. Int. J. Appl. Radiat. Isot. 29:531–536 Efron, B. and R.J. Tibshirani. 1993. An Introduction to Bootstrap. London, UK: Chapman and Hall Egglishaw, H.J. 1972. An experimental study of the breakdown of cellulose in fast flowing streams. Mem. 1st. Ital. Idrobiol. 29(Suppl.):405–428 Eisenreich, S.J., P.J. Emmling, and A.M. Beeton. 1977. Atmospheric loading of phosphorus and other chemicals to Lake Michigan. J. Great Lakes Res. 3:291–304 Elwood, J.W., J.D. Newbold, A.F. Trimble, and R.W. Stark. 1981. The limiting role of phosphorus in a woodland stream ecosystem: effect of P enrichment on leaf decomposition and primary producers. Ecology 62:146–158 Enos, P. and R.D. Perkins. 1977. Quaternary Sedimentation in South Florida. Boulder, CO: Geological Society of America Environmental Systems Research Institute, Inc. 1998. Arc Version 7.2.1 Patch 1, Arcview Version 3.1 Epler, J.H. 1995. Identification Manual for the Larval Chironomidae (Diptera) of Florida, revised edition. Tallahassee, FL: Florida Department of Environmental Protection Epler, J.H. 1996. Identification Manual for the Water Beetles of Florida. Tallahassee, FL: Florida Department of Environmental Protection Eubank, R. 1988. Spline Smoothing and Nonparametric Regression. New York: Dekker Faegri, K. and J. Iversen. 1989. Textbook of Pollen Analysis, 4th edition. New York: Wiley Faith, D.P. and Norris, R.H. 1989. Correlation of environmental variables with patterns of distribution and abundance of common and rare freshwater macroinvertebrates. Biol. Conserv. 50:77–98 Faith, D.P., P.R. Minchin, and L. Belbin. 1987. Compositional dissimilarity as a robust measure of ecological distance. Vegetatio 69:57–68 FDEP (Florida Department of Environmental Protection). 1996. Standard operating procedures manual – benthic macroinvertebrate sampling and habitat assessment methods: 1. Freshwater streams and rivers. Tallahassee, FL: Florida Department of Environmental Protection FDEP (Florida Department of Environmental Protection). 2003. Appendix 2A-3: Annual Summary Statistics for Phosphorus Concentrations at Everglades Protection Area Monitoring Stations. In 2003 Everglades Consolidated Report. West Palm Beach, FL: South Florida Water Management District Feng, K.E. and Molz, F.J. 1997. A 2-D, diffusion-based, wetland flow model. J. Hydrol. 196:230–250 Fennema, R.J., C.J. Neidrauer, R.A. Johnson, T.K. MacVicar, and W.A. Perkins. 1994. A computer model to simulate natural Everglades hydrology. In Everglades: The Ecosystem and Its Restoration, eds. S.M. Davis and J.C. Ogden, pp. 249–290. Delray Beach, FL: St. Lucie Fernald, M.L. 1950. Gray’s Manual of Botany, 8th edition. New York: American Book Company Field, C.B. and H.A. Mooney. 1986. The nitrogen/phosphorus relationship in wild plants. In On the Economy of Plant Form and Function, ed. T.J. Givinsh, pp. 25–55. Cambridge, UK: Cambridge University Press Fisher, R. 1953. Dispersion on a sphere. Proc. R. Soc. Lond. A 217:295–305 Fisher, N.I. 1993. Statistical Analysis of Circular Data. Cambridge: Cambridge University Press Fisher, M.M. and K.R. Reddy. 2001. Phosphorus flux from wetland soils affected by long-term nutrient loading. J. Environ. Qual. 30:261–271 Flint, R.F. 1971. Glacial and Quaternary Geology. New York: Wiley Flora, M.D., D.R. Walker, D.J. Scheidt, R.G. Rice, and D.H. Landers. 1988. The Response of the Everglades Marsh to Increased Nitrogen and Phosphorus Loading: Part I. Nutrient Dosing, Water Chemistry, and Periphyton Productivity. Homestead, FL: National Park Service, South Florida Research Center, Everglades National Park Fore, L.S., J.R. Karr, and R.W. Wisseman. 1996. Assessing invertebrate responses to human activities: evaluating alternative approaches. J. N. Am. Benthol. Soc. 15:212–231

652

Bibliography

Forsberg, C. 1964. Phosphorus, a maximum factor in the growth of Characeae. Nature 201:517–518 Foster, D.R., M. Fluet, and E.R. Boose. 1999. Human or natural disturbance: landscape-scale dynamics of the tropical forests of Puerto Rico. Ecol. Appl. 9:555–572 Franklin, J. 1993. Preserving biodiversity: species, ecosystems or landscapes? Ecol. Appl. 3:202–205 Frazier, W.J. and D.R. Schwimmer. 1987. Regional Stratigraphy of North America. New York: Plenum Frederick, P.C. and M.G. Spalding. 1994. Factors affecting reproductive success of wading birds (Ciconiiformes) in the Everglades ecosystem. In Everglades: The Ecosystem and Its Restoration, eds. S.M. Davis and J.C. Ogden, pp. 659–691. Delray Beach, FL: St. Lucie Funkhauser, J.W. and W.R. Evitt. 1959. Preparation techniques for acid insoluble microfossils. Micropaleontology 5:369–375 Gabor, T.S., H.R. Murkin, M.P. Stainton, J.A. Boughen, and R.D. Titman. 1994. Nutrient additions to wetlands in the Interlake region of Manitoba, Canada: effects of a single pulse addition in spring. Hydrobiologia 279/280:497–510 Gambrell, R.P. 1994. Trace and toxic metals in wetlands. J. Environ. Qual. 23:883–891 Gardner, N.L. 1927. New Myxophyceae from Puerto Rico. Mem. N.Y. Bot. Garden 7:1–144 Gawlik, D.E. and D.A. Rocque. 1998. Avian communities in bayheads, willowheads and sawgrass marshes of the central Everglades. Wilson Bull. 110:45–55 Geider, R.J., B.A. Osborne, and J.A. Raven. 1985. Light dependence of growth and photosynthesis in Phaeodactylum tricornutum (Bacillariophyceae). J. Phycol. 21:609–619 Geller, A. 1986. Comparisons of mechanisms enhancing biodegradability of refractory lake water constituents. Limnol. Oceanogr. 31:755–764 Germano, J.D. 1999. Ecology, statistics, and the art of misdiagnosis: the need for a paradigm shift. Environ. Rev. 7:167–190 Gernes, M.C. and J.C. Helgen. 1999. Indexes of Biotic Integrity (IBI) for Wetlands: Vegetation and Invertebrate IBIs. Minneapolis, MN: Minnesota Pollution Control Agency, Environmental Outcomes Division. Report to USEPA, #CD995525-01 Ghoshal, N. and K.P. Singh. 1995. Effects of farmyard manure and inorganic fertilizer on the dynamics of soil microbial biomass in a tropical dryland agroecosystem. Biol. Fert. Soils 19:231–238 Gilmour, C.C. 2003. Status report on the effect of water quantity and quality on methlymercury production. In 2003 Everglades Consolidated Report, Appendix 2B-2. West Palm Beach, FL: South Florida Water Management District Gilmour, C.C. and E.A. Henry. 1991. Mercury methylation in aquatic systems affected by acid deposition. Environ. Pollut. 71:131–169 Gilmour, C.C., E.A. Henry, and R. Mitchell. 1992. Sulfate stimulation of mercury methylation in freshwater sediments. Environ. Sci. Technol. 26:2281–2287 Gilmour, C.C., G.S. Riedel, M.C. Ederlington, J.T. Bell, J.M. Beniot, G.A. Gill, and M.C. Stordal. 1998. Methylmercury concentrations and production rates across a trophic gradient in the Northern Everglades. Biogeochemistry 40:326–346 Gilpin, M.E. and I.A. Hanski. 1991. Metapopulation Dynamics: Empirical and Theoretical Investigations. London: Academic Gleason, P.J. 1972. The origin, sedimentation, and stratigraphy of a calcite mud located in the southern freshwater Everglades. Ph.D. Dissertation, Pennsylvania State University Gleason, P.J. (ed.). 1974a. Environments of South Florida: Present and Past. Miami, FL: Miami Geological Society Gleason, P.J. 1974b. Chemical Quality of Water in Conservation Area 2A and Associated Canals. West Palm Beach, FL: Resource Planning Department, Central and Southern Florida Flood Control District. Technical Bulletin 74-1 Gleason, P.J. (ed.). 1984. Environments of South Florida: Present and Past II. Memoir 2. Coral Gables, FL: Miami Geological Society Gleason, H.A. and A. Cronquist. 1963. Manual of Vascular Plants of Northeastern United States and Adjacent Canada. New York: D. Van Nostrand

Bibliography

653

Gleason, P.J. and W. Spackman, Jr. 1974. Calcareous periphyton and water chemistry in the Everglades. In Environments of South Florida: Present and Past, Memoir 2, ed. P.J. Gleason, pp. 146–181. Miami, FL: Miami Geological Society Gleason, P.J. and P. Stone. 1994. Age, origin and landscape evolution of the Everglades peatlands. In Everglades: The Ecosystem and Its Restoration, eds. S.M. Davis and J.C. Ogden, pp. 149–197. Delray Beach, FL: St. Lucie Gleason, P.J., A.D. Cohen, W.G. Smith, H.K. Brooks, P.A. Stone, R.L. Goodrick, and W. Spackman, Jr. 1984. The environmental significance of Holocene sediments from the Everglades and saline tidal plain. In Environments of South Florida: Present and Past II, ed. P.J. Gleason, pp. 297–351. Coral Gables, FL: Miami Geological Society Goldsborough, L.G. and G.G.C. Robinson. 1996. Pattern in wetlands. In Algal Ecology, eds. R.J. Stevenson, M.L. Bothwell, and R.L. Lowe, pp. 77–117. San Diego, CA: Academic Goodrick, R.L. 1984. The wet prairies of the northern Everglades. In Environments of South Florida, Present and Past II, ed. P.J. Gleason, pp. 185–190. Miami, FL: Miami Geological Society Grace, J.B. 1983. Autotoxic inhibition of seed germination by Typha latifolia: an evaluation. Oecologia 59:366–369 Grace, J.B. 1989. Effects of water depth on Typha latifolia and Typha domingensis. Am. J. Bot. 76:762–768 Green, R.H. 1979. Sampling Design and Statistical Methods for Environmental Biologists. New York, NY: Wiley Green, P.J. and B.W. Silverman. 1994. Nonparametric Regression and Generalized Linear Models: A Roughness Penalty Approach. London: Chapman and Hall Greeson, P.E. 1977. Diel oxygen curve method for estimating primary productivity and community metabolism. In Techniques of Water Resources Investigations of USGS, pp. 269–279. Washington, DC: US Geological Survey Grieb, T.M., C.T. Driscoll, S.P. Gloss, C.L. Schofield, G.L. Bowie, and D.B. Porcella. 1990. Factors affecting mercury accumulation in fish in the upper Michigan peninsula. Environ. Toxicol. Chem. 9:919–930 Grime, J.P. 1977. Evidence for the existence of three primary strategies in plants and its relevance to ecological and evolutionary theory. Am. Nat. 111:1169–1194 Grimshaw, H.J., R.G. Wetzel, M. Brandenburg, K. Segerblom, L.J. Wenkert, G.A. Marsh, W. Charnetzky, J.E. Haky, and C. Carraher. 1997. Shading of periphyton communities by wetland emergent macrophytes: decoupling of algal photosynthesis from microbial nutrient retention. Arch. Hydrobiol. 139:17–27 Growns, J.E., J.A. Davis, F. Cheal, L.G. Schmidt, R.S. Rosich, and S.J. Bradley. 1992. Multivariate pattern analysis of wetland invertebrate communities and environmental variables in Western Australia. Aust. J. Ecol. 17:275–288 Grunwald, M. 2006. The Swamp: The Everglades, Florida, and the Politics of Paradise. New York: Simon and Schuster Guardo, M., L. Fink, T.D. Fontaine, S. Newman, M. Chimney, R. Bearzotti, and G. Goforth. 1995. Large-scale constructed wetlands for nutrient removal from stormwater runoff: an Everglades restoration project. Environ. Manage. 19:879–889 Guentzel, J.L., W.M. Landing, G.A. Gill, and C.D. Pollman. 2001. Processes influencing rainfall deposition of mercury in Florida. Environ. Sci. Technol. 35:863–873 Gunderson, L.H. 1990. Historical hydropatterns in vegetation communities of Everglades National Park. In Freshwater Wetlands and Wildlife, eds. R.R. Sharitz and J.W. Gibbons, pp. 1099– 1111. Oak Ridge, TN: US Department of Energy Gunderson, L.H. 1994. Vegetation of the Everglades: determinants of community composition. In Everglades: The Ecosystem and Its Restoration, eds. S.M. Davis and J.C. Ogden, pp. 323–340. Delray Beach, FL: St. Lucie Gunderson, L.H. and W.T. Loftus. 1993. The Everglades. In Biodiversity of the Southeastern United States: Lowland Terrestrial Communities, eds. W.H. Martin, S.G. Boyce, and A.C.E. Echtemacht, pp. 199–255. New York: Wiley

654

Bibliography

Gunderson, L.H. and J.R. Snyder. 1994. Fire patterns in the southern Everglades. In Everglades: The Ecosystem and Its Restoration, eds. S.M. Davis and J.C. Ogden, pp. 291–306. Delray Beach, FL: St. Lucie Gunderson, L.H., C.S. Holling, and G.D. Peterson. 2002. Surprises and sustainability: cycles of renewal in the Everglades. In Panarchy: Understanding Transformations in Human and Natural Systems, eds. L.H. Gunderson and C.S. Holling, pp. 315–332. Washington, DC: Island Hairston, N.G., F.E. Smith, and L.B. Slobodkin. 1960. Community structure, population control, and competition. Am. Nat. 94:421–425 Hall, G.B. and R.G. Rice. 1990. Response of the Everglades Marsh to Increased Nitrogen and Phosphorus Loading, Part III: Periphyton Community Dynamics. Homestead, FL: National Park Service, South Florida Research Center, Everglades National Park Hammer, D.A. (ed.). 1989. Constructed Wetlands for Wastewater Treatment: Municipal, Industrial, and Agricultural. Chelsea, MI: Lewis Hann, B.J. and L.G. Goldsborough. 1997. Responses of a prairie wetland to press and pulse additions of inorganic nitrogen and phosphorus: invertebrate community structure and interactions. Arch. Hydrobiol. 140:169–194 Happell, J.D., G.J. Whiting, W.S. Showers, and J.P. Chanton. 1993. A study contrasting methane emission from marshlands with and without active methanotrophic bacteria. J. Geophys. Res. 98:307–314 Härdle, W. 1990. Smoothing Techniques: With Implementation in S. Berlin Heidelberg New York: Springer Harrison, S. and L. Fahrig. 1995. Landscape pattern and population conservation, In Mosaic Landscapes and Ecological Processes, eds. L. Hannson, L. Fahrig, and G. Merriam, pp. 293–308. London: Chapman and Hall Hart, D.D. and C.T. Robinson. 1990. Resource limitation in a stream community: phosphorus enrichment effects on periphyton and grazers. Ecology 71:1494–1502 Harvey, J.W., J.T. Newlin, and S.L. Krupa. 2006. Modeling decadal timescale interactions between surface water and ground water in the central Everglades, Florida, USA. J. Hydrol. 320:400–420 Hastie, T.J. and R.J. Tibshirani. 1990. Generalized Additive Models. London: Chapman and Hall Hawkins, C.P. and R.H. Norris. 2000. Performance of different landscape classifications for aquatic bioassessments: introduction to the series. J. N. Am. Benthol. Soc. 19:367–369 Hawkins, C.P., R.H. Norris, J.N. Hogue and J.W. Feminella. 2000. Development and evaluation of predictive models for measuring the biological integrity of streams. Ecol. Appl. 10:1456–1477 Hedley, M.J. and J.W.B. Stewart. 1982. Method to measure microbial phosphate in soils. Soil Biol. Biochem. 14:377–385 Hedley, M.J., J.W.B. Stewart and B.S. Chauhan. 1982. Changes in inorganic soil phosphorus fractions induced by cultivation practices and by laboratory incubations. Soil Sci. Soc. Am. J. 46:970–976 Hemond, H.F. 1980. Biogeochemistry of Thoreau’s Bog, Concord, Massachusetts. Ecol. Monogr. 50:507–526 Hemond, H.F. 1983. The nitrogen budget of Thoreau’s Bog. Ecology 64:99–109 Hendry, C.D., P.L. Brezonik, and E.S. Edgerton. 1981. Atmospheric deposition of nitrogen and phosphorus in Florida. In Atmospheric Pollutants in Natural Waters, ed. S.J. Eisenreich, pp. 199–215. Ann Arbor, MI: Ann Arbor Science Publishers Hershey, A.E., A.L. Hiltner, M.A.J. Hullar, M.C. Miller, J.R. Vestal, M.A. Lock, S. Rundle, and B.J. Peterson. 1988. Nutrient influence on a stream grazer: Orthocladius microcommunities respond to nutrient input. Ecology 69:1383–1892 Hobbs, H.H. 1942. The crayfishes of Florida. Fla. Univ. Publ. Biol. Sci. Ser. 3:1–179 Hoffman, W., G.T. Bancroft, and R.J. Sawicki. 1994. Foraging habitat of wading birds in the Water Conservation Areas of the Everglades. In Everglades: The Ecosystem and Its Restoration, eds. S.M. Davis and J.C. Ogden, pp. 585–614. Delray Beach, FL: St. Lucie

Bibliography

655

Hoffmeister, J.E. and H.G. Multer. 1968. Geology and origin of the Florida keys. Geol. Soc. Am. Bull. 79:1487–1502 Holling, C.S. 1973. Resilience and stability of ecological systems. Annu. Rev. Ecol. Syst. 4:1–23 Holling, C.S., L.H. Gunderson, and C.J. Walters. 1994. The structure and dynamics of the everglades system: guidelines for ecosystem restoration. In Everglades: The Ecosystem and Its Restoration, eds. S.M. Davis and J.C. Ogden, pp. 741–756. Delray Beach, FL: St. Lucie Holmgren, G.G.S., M.W. Meyer, R.L. Chaney, and R.L. Daniels. 1993. Cadmium, lead, zinc, copper and nickel in agricultural soils of the United States of America. J. Environ. Qual. 22:335–348 van Horn, S. 1996. Hydrometeorologic Monitoring Network Metadata Report, WRE #344. West Palm Beach, FL: South Florida Water Management District House, W.A. 1990. The prediction of phosphate coprecipitation with calcite in freshwaters. Water Resour. 24:1017–1023 Howarth, R.W. and S.G. Fisher. 1976. Carbon, nitrogen, and phosphorus dynamics during leaf decay in nutrient-enriched microecosystems. Freshwater Biol. 6:221–228 Hubert, W.A. 1996. Passive capture techniques. In Fisheries Techniques, 2nd edition, eds. B.R. Murphy and D.W. Willis, pp. 157–182. Bethesda, MD: American Fisheries Society Hunt, R. 1978. Plant Growth Analysis. The Institute of Biology’s Studies in Biology No. 96. London: Edward Arnold Hurlbert, S.A. 1971. The nonconcept of species diversity: a critique and alternative parameters. Ecology 52:577–586 Hustedt, F. 1927–1930. Die Kieselalgen Deutschlands, Österreichs und der Schweiz (3 vols.), In Dr. L. Rabenhorst’s Kryptogamen-Flora von Deutschland, Österreich und der Schweiz. Band 7, Die Kieselalgen. Akademische Verlagsgesellschaft Hvorslev, M.J. 1951. Time lag and soil permeability in ground water observations. US Army Corps Eng. Waterway Exp. Station Bull. 36 IPCC (Intergovernmental Panel on Climate Change). 2007. Climate Change 2007: Summary for Policymakers. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Geneva: IPCC Secretariat IUPAC (International Union of Pure and Applied Chemistry). 1978. Nomenclature, symbols, and their usage in spectrochemical analysis-II. Data interpretation. Spectrochim. Acta 33:241–245 Ives, J.C. 1856. Memoir to Accompany a Military Map of the Peninsula of Florida South of Tampa Bay. New York: [US] War Department Jackson, T.A. 1988. The mercury problem in recently formed reservoirs of northern Manitoba (Canada): effects of impoundments and other factors on the production of methyl mercury by microorganisms. Can. J. Fish. Aquat. Sci. 45:97–121 Jensen, J.E. 1998. Calibration of Everglades modern pollen with vegetation and nutrient levels in Water Conservation Area 2A. M.Sc. Thesis, Duke University Jensen, J.R., K. Rutchey, M.S. Koch, and S. Narumalani. 1995. Inland wetland change detection in Everglades water conservation area 2A using a time series of normalized remotely sensed data. Photogramm. Eng. Remote Sens. 61:199–209 Jensen, J.E., S.R. Cooper, and C.J. Richardson. 1999. Development of a calibration model of modern pollen along a nutrient gradient in Everglades Water Conservation Area-2A, USA. Wetlands 19:675–688 Johnson, D.H. 1999. The insignificance of statistical significance testing. J. Wildlife Manage. 63:763–772 Jones, L.A. 1948. Soils, geology, and water control in the Everglades region. Univ. Fla. Agri. Exp. Station Bull. 442. Gainesville, FL Jones, J.I., B. Moss, and J.O. Young. 1998. Interactions between periphyton, nonmolluscan invertebrates, and fish in standing freshwaters. In The Structuring Role of Submerged Macrophytes in Lakes, eds. E. Jeppesen, Ma. Søndergaard, Mo. Søndergaard, and K. Christoffersenp, pp. 69–90. Berlin Heidelberg New York: Springer Jordan, C.L. 1984. Florida’s weather and climate implications for water. In Water Resources Atlas of Florida, eds. E.A. Fernald and D.J. Patton, pp. 18–35. Tallahassee: Florida State University

656

Bibliography

Jordan, C.F. 1996. Spatial Ecology of Decapods and Fishes in a Northern Everglades Wetland Mosaic. Gainesville, FL: University of Florida Jordan, C.F., Jelks, H.L., and Kitchens, W.M. 1997. Habitat structure and plant community composition in a northern Everglades wetland landscape. Wetlands 17:275–283 Kadlec, R.H. 1999. The limits of phosphorus removal in wetlands. Wetlands Ecol. Manage. 7:165–175 Kadlec, R.H. and R.L. Knight. 1996. Treatment Wetlands. Boca Raton, FL: CRC/Lewis Kang, W.-J., J.H. Trefry, T.A. Nelson, and H.R. Wanless. 2000. Direct atmospheric inputs versus runoff fluxes of mercury to the lower Everglades and Florida Bay. Environ. Sci. Technol. 34:4058–4063 Kapp, R.O. 1969. How to Know Pollen and Spores. Dubuque: Brown and Company Karr, J.R. 1981. Assessment of biotic integrity using fish communities. Fisheries 6:21–27 Karr, J.R. and E.W. Chu. 1997. Biological Monitoring and Assessment: Using Multimetric Indexes Effectively. Seattle, WA: University of Washington. EPA 235-R97-001 Karr, J.R., K.D. Fausch, P.L. Angermeier, P.R. Yant, and I.J. Schlosser. 1986. Assessing Biological Integrity in Running Waters: A Method and Its Rationale. Special Publication 5. Urbana, IL: Illinois Natural History Survey Kaštovský, J. and J. Komárek. 2001. Phototrophic microvegetation of thermal springs in Karlovy Vary, Czech Republic. In Algae and Extreme Environments, eds. J. Elster, J. Seckbach, W.F. Wincent, and O. Lhotsky, pp. 107–119. Nova Hedwigia Beih. 123 Keddy, P.A. 2000. Wetland Ecology: Principles and Conservation. Cambridge, UK: Cambridge University Press Keeney, D.R. and D.W. Nelson. 1982. Nitrogen – inorganic forms. In Methods of Soil Analysis, eds. A.L. Page, R.H. Miller, and D.R. Keeney, pp. 643–698. Madison, WI: American Society of Agronomy Keiber, J.R., X. Zhou, and K. Mopper. 1990. Formation of carbonyl compounds from UV-induced photodegradation of humic compounds in the sea. Limnol. Oceanogr. 35:1503–1515 Keith and Schnars, P.A. 1993. G.P.S. Geodetic Survey in the Everglades, WCA-2A and WCA-3A, prepared for Peeples, Earl and Blank, P.A. Keough, J.R., M.E. Sierszen, and C.A. Hagley. 1996. Analysis of a Lake Superior coastal food web with stable isotope techniques. Limnol. Oceanogr. 41:136–141 Keyser, D. 1975. Ostracodes of the mangroves of South Florida, their ecology and biology. Bull. Am. Paleontol. 65:489–499 King, R.S. 2001. Dimensions of invertebrate assemblage organization across a phosphorus-limited Everglades landscape. Ph.D. Dissertation, Duke University King, R.S. and C.J. Richardson. 2002. Evaluating subsampling approaches and macroinvertebrate taxonomic resolution for wetland bioassessment. J. N. Am. Benthol. Soc. 21:150–171 King, R.S. and C.J. Richardson. 2003. Integrating bioassessment and ecological risk assessment: an approach to develop numerical water-quality criteria. Environ. Manage. 31:795–809 King, R.S. and D.A. Wrubleski. 1998. Spatial and diel availability of flying insects as potential duckling food in prairie wetlands. Wetlands 18(1):100–114 King, R.S., K.T. Nunnery, and C.J. Richardson. 2000. Macroinvertebrate assemblage response to highway crossings in forested wetlands: implications for biological assessment. Wetlands Ecol. Manage. 8(4):243–256 King, R.S., C.J. Richardson, D.L. Urban, and E.A. Romanowicz. 2004. Spatial dependency of vegetation–environment linkages in an anthropogenically influenced wetland ecosystem. Ecosystems 7:74–97 Klein, H. and C.R. Causaras. 1982. Biscayne aquifer, southeast Florida, and contiguous surficial aquifer to the north, In Principal Aquifers in Florida, ed. B.J. Franks, pp. 1–4. Tallahassee, FL: US Geological Survey. Water-Resources Investigations. Open-File Report Kleiner, J. 1988. Coprecipitation of phosphate with calcite in lake water: a laboratory experimental modeling phosphorus removal with calcite in lake Constance. Water Resour. 22:1259–1265 Klemm, D.J. 1995. Identification Guide to the Freshwater Leeches (Annelida: Hirudinea) of Florida and other Southern States. Tallahassee, FL: Florida Department of Environmental Protection

Bibliography

657

Kloor, K. 2000. Everglades restoration plan hits rough waters. Science 288:1166–1167 Knight, R.L., R.H. Kadlec, S.G. Reed, R.W. Ruble, J.D. Waterman, and D.S. Brown. 1994. North American Wetlands for Water Quality Treatment Database, EPA/600/C-94/002 Koch, M.S. and P.S. Rawlik. 1993. Transpiration and stomatal conductance of two wetland macrophytes (Cladium jamaicense and Typha domingensis) in the subtropical Everglades. Am. J. Bot. 80:1146–1154 Koch, M.S. and K.R. Reddy. 1992. Distribution of soil and plant nutrients along a trophic gradient in the Florida Everglades. Soil Sci. Soc. Am. J. 56:1492–1499 Koch-Rose, M.S., K.R. Reddy, and J.P. Chanton. 1994. Factors controlling seasonal nutrient profiles in a subtropical peatland of the Florida Everglades. J. Environ. Qual. 23:526–533 Koerselman, W. and A.F.M. Meuleman. 1996. The vegetation N:P ratio: a new tool to detect the nature of nutrient limitation. J. Appl. Ecol. 33:1441–1450 Komárek, J. 1989. Studies on the cyanophytes of Cuba 7–9. Folia Geobot. Phytotax. Praha 24:131–206 Komárek, J. and K. Anagnostidis. 1998. Cyanoprokaryota, Vol. 1: Chroococcales. In Süsswasserflora von Mitteleuropa 19/1, eds. H. Ettl, G. Gärtner, H. Heynig, and D. Mollenhauer, Jena: Gustav Fischer Koroleff, F. 1983. Simultaneous oxidation of nitrogen and phosphorus compounds by persulfate. In Methods of Seawater Analysis, 2nd edition, eds. K. Grasshoff, K. Kremling, and M. Ehrhardt, pp. 168–169. Weinheimer, Germany: Verlag Chemie Kostecke, R.M., L.M. Smith, and H.M. Hands. 2004. Vegetation response to cattail management at Cheyenne Bottoms, Kansas. J. Aquat. Plant Manage. 42:39–45 Krabbenhoft, D.P., W.H. Orem, G.R. Aiken, and C.C. Gilmour. 2006a. Unraveling the Complexities of Mercury Methylation in the Everglades. Florida Greater Everglades Ecosystem Restoration Conference – 2006 “Planning, Policy and Science” Geer Conference abstract. Lake Buena Vista, FL, 5–9 June 2006 Krabbenhoft, D.P., W.H. Orem, G.R. Aiken, and C.C. Kendall. 2006b. Aquatic cycling of mercury in the Everglades. USGS fact sheet http://sofia.usgs.gov/projects/evergl_merc/ Krammer, K. and H. Lange-Bertalot. 1987–1991. Bacillariophyceae (4 vols.). In Süßwasserflora von Mitteleuropa 2/1–4, eds. H. Ettl, G. Gärtner, H. Heynig, and D. Mollenhauer. Jena: Gustav Fischer Krammer, K. and H. Lange-Bertalot. 1988. Bacillariophyceae, Vol. 2: Bacillariaceae. In Süßwasserflora von Mitteleuropa 2/2, eds. H. Ettl, G. Gärtner, H. Heynig, and D. Mollenhauer. Jena: Gustav Fischer Krammer, K. and H. Lange-Bertalot. 1991a. Bacillariophyceae, Vol. 3: Centrales, Fragilariaceae, Eunotiaceae. In Süßwasserflora von Mitteleuropa 2/3, eds. H. Ettl, G. Gärtner, H. Heynig, and D. Mollenhauer. Jena: Gustav Fischer Krammer, K. and H. Lange-Bertalot. 1991b. Bacillariophyceae, Vol. 4: Achnanthaceae. In Süßwasserflora von Mitteleuropa 2/4, eds. H. Ettl, G. Gärtner, H. Heynig, and D. Mollenhauer. Jena: Gustav Fischer Krammer, K. and H. Lange-Bertalot. 1997. Bacillariophyceae, Vol. 1: Naviculaceae. In Süßwasserflora von Mitteleuropa 2/2, 2nd edition, eds. H. Ettl, G. Gärtner, H. Heynig, and D. Mollenhauer. Jena: Gustav Fischer Kuhn, N.L., I.A. Mendelssohn, K.L. McKee, H. Brix, and S.L. Miao. 2002. Root phosphatase activity in Cladium jamaicense and Typha domingensis grown under ambient and elevated phosphorus levels. Wetlands 22:794–800 Kushlan, J.A. 1974. Observations on the role of the American alligator (Alligator Mississippiensis) in the southern Florida wetlands. Copeia 1974:993–996 Kushlan, J.A. 1987. External threats and internal management: the hydrologic regulation of the Everglades, FL, USA. Environ. Manage. 11:109–119 Kushlan, J.A. 1989. Wetlands and wildlife: the Everglades perspective. In Freshwater Wetlands and Wildlife, CONF-8603101. DOE Symposium Serial No. 61, eds. R.R. Sharitz and J.W. Gibbons, pp. 773–790. Oak Ridge, TN: USDOE Office of Scientific and Technical Information Kushlan, J.A., S.A. Voorhees, W.F. Loftus, and P.C. Frohring. 1986. Length, mass, and calorific relationships of Everglades animals. Fla. Sci. 49:65–79

658

Bibliography

Kusler, J. and Niering, W. 1998. Wetland assessment: have we lost our way? Natl Wetlands Newslett. 20:8–14 Květ, J., J.P. Ondok, J. Nečas, and P.G. Jarvis. 1971. Methods of growth analysis. In Plant Photosynthetic Production: Manual of Methods, eds. Z. Šesták, J. Čatský, and P.G. Jarvis, pp. 343–391. The Hague, The Netherlands: Dr. W. Junk N.V. Publishers Lamberti, G.A. 1996. The role of periphyton in benthic food webs. In Algal Ecology, eds. R.J. Stevenson, M.L. Bothwell, and R.L. Lowe, pp. 533–572. San Diego, CA: Academic Lamberti, G.A. and J.W. Moore. 1984. Aquatic insects as primary consumers. In The Ecology of Aquatic Insects, eds. V.H. Resh and D.M. Rosenberg, pp. 164–195. New York: Praeger Lange, T.R., H.E. Royals, and L.L. Connor. 1993. Influence of water chemistry on mercury concentration in largemouth bass from Florida lakes. Trans. Am. Fish. Soc. 122:74–84 Larsen, P. 1994. Everglades water budget – principles and overview. Technical Advisory Committee presentation, Governor’s Commission for a Sustainable South Florida (October 19), Miami, FL Larsen, D.P. and A.T. Herlihy. 1998. The dilemma of sampling streams for macroinvertebrate richness. J. N. Am. Benthol. Soc. 17:359–366 Leduc, A., P. Drapeau, Y. Bergeron, and P. Legendre. 1992. Study of spatial components of forest cover using partial Mantel tests and path analysis. J. Veg. Sci. 3:69–78 Legendre, P. and M.J. Anderson. 1999. Distance-based redundancy analysis: testing multispecies responses in multifactorial ecological experiments. Ecol. Monogr. 69:1–24 Legendre, P. and L. Legendre. 1998. Numerical Ecology, 2nd edition. Amsterdam, The Netherlands: Elsevier Lemly, A.D. and R.S. King. 2000. An insect-bacteria bioindicator for assessing detrimental nutrient enrichment in wetlands. Wetlands 20:91–100 Lemly, A.D. and C.J. Richardson. 1997. Guidelines for risk assessment in wetlands. Environ. Monit. Assess. 47:117–134 Levinspeil, O. 1972. Chemical Reaction Engineering. New York: Wiley Liebig, J. 1840. Chemistry in Its Application to Agriculture and Physiology. London, UK: Taylor and Walton Lietz, A.C. 2000. Analysis of Water-Quality Trends at Two Discharge Stations – One Within Big Cypress National Preserve and One Near Biscayne Bay – Southern Florida, 1966–1994. Tallahassee, FL: US Geological Survey. Water-Resources Investigations. Report 00–4099 Light, S.S. and J.W. Dineen. 1994. Water control in the Everglades: a historical perspective. In Everglades: The Ecosystem and Its Restoration, eds. S.M. Davis and J.C. Ogden, pp. 47–84. Delray Beach, FL: St. Lucie Likens, G.E. and F.H. Bormann. 1995. Biochemistry of a Forested Ecosystem, 2nd edition. Berlin Heidelberg New York: Springer Linde, A.F., T. Janisch, and D. Smith. 1976. Cattail – The Significance of Its Growth, Phenology and Carbohydrate Storage to Its Control and Management. Madison, WI: Department of Natural Resources. Technical Bulletin 94 Line, J.M., C.J.F. ter Braak, and H.J.B. Birks. 1994. WACALIB version 3.3 – a computer program to reconstruct environmental variables from fossil assemblages by weighted averaging and to derive sample-specific errors of prediction. J. Paleolimnol. 10:147–152 Lissner, J.I., A. Mendelssohn, B. Lorenzen, H. Brix, K.L. McKee, and S. Miao. 2003. Interactive effects of redox intensity and phosphate availability on growth and nutrient relations of Cladium jamaicense. Am. J. Bot. 90:736–748 Livingstone, D.R. 1991. Organic xenobiotic metabolism in marine invertebrates. In Advances in Comparative and Environmental Physiology, ed. R. Gilles, pp. 45–185. Berlin Heidelberg New York: Springer Lockaby, B.G., R.S. Wheat, and R.G. Clawson. 1996. Hydroperiod influence on litter conversion to soil organic matter in a floodplain forest. Soil Sci. Soc. Am. J. 60:1989–1993 Lodge, T.E. 2005. The Everglades Handbook: Understanding the Ecosystem, 2nd edition. Boca Raton, FL: CRC

Bibliography

659

Lodge, D.M., G. Cronin, E. van Donk, and A.J. Foelich. 1998. Impact of herbivory on plant standing crop: comparisons among biomes, between vascular and nonvascular plants, and among freshwater herbivore taxa. In The Structuring Role of Submerged Macrophytes in Lakes, eds. E. Jeppesen, Ma. Søndergaard, Mo. Søndergaard, and K. Christoffersen, pp. 149–174. Berlin Heidelberg New York: Springer Loftus, W.F. and A.M. Eklund. 1994. Long-term dynamics of an Everglades small-fish assemblage. In Everglades: The Ecosystem and Its Restoration, eds. S.M. Davis and J.C. Ogden, pp. 461–483. Boca Raton, FL: St. Lucie Long, R.W. and O. Lakela. 1971. A Flora of Tropical Florida. A Manual of the Seed Plants and Ferns of Southern Peninsular Florida. Coral Gables, FL: University of Miami Press Lorenzen, B., H. Brix, I.A. Mendelssohn, K.L. McKee, and S.L. Miao. 2001. Growth, biomass allocation and nutrient use efficiency in Cladium jamaicense and Typha domingensis as affected by phosphorus and oxygen availability. Aquat. Bot. 70(2):117–133 Loveless, C.M. 1959. A study of the vegetation in the Florida Everglades. Ecology 40:1–9 Lowe, E.F. and L.W. Keenan. 1997. Managing phosphorus based cultural eutrophication in wetlands: a conceptual approach. Ecol. Eng. 9:109–118 Ludlam, S.D., S. Feeney, and M.S.V. Douglas. 1996. Changes in the importance of lotic and littoral diatoms in a high arctic lake over the last 191 yrs. J. Paleolimnol. 16:187–204 Lutz, J.R. 1977. Water Quality and Nutrient Loading of the Major Inflows from the Everglades Agricultural Area to the Conservation Areas, Southeast Florida. Technical Publication 77–6. West Palm Beach, FL: South Florida Water Management District Maberly, S.C. and D.H.N. Spence. 1983. Photosynthetic inorganic carbon use by freshwater plants. J. Ecol. 71:705–724 MacArthur, R.H. and E.O. Wilson. 1967. The Theory of Island Biogeography. Princeton, NJ: Princeton University Press MacVicar, T.K. and S.T. Lin. 1984. Historical rainfall activity in central and southern FL: Average, return period estimates and selected extremes. In Environments of South Florida: Present and Past II. Memoir 2, ed. P.J. Gleason, pp. 477–509. Coral Gables, FL: Miami Geological Society Manly, B.F.J. 1997. Randomization, Bootstrap, and Monte Carlo Methods in Biology, 2nd edition. London: Chapman and Hall Mantel, N. 1967. The detection of disease clustering and a generalized regression approach. Cancer Res. 27:209–220 MathSoft (1997), S-Plus 4 User’s Guide. Seattle, WA: MathSoft Maxted, J.R., M.T. Barbour, J. Gerritsen, V. Poretti, N. Primrose, A. Silvia, et al. 2000. Assessment framework for mid-Atlantic coastal plain streams using benthic macroinvertebrates. J. N. Am. Benthol. Soc. 19:128–144 May, R.M. 1975. Patterns of species abundance and diversity. In Ecology and Evolution of Communities, ed. M.L. Cody and J.M. Diamond, pp. 81–120. Cambridge, MA: Harvard University Press McAndrews, J.H., A.A. Berti, and G. Norris. 1973. Key to the Quaternary Pollen and Spores of the Great Lakes Region, Royal Ontario Museum: Life Sciences Misc. Publications McCally, D. 1999. The Everglades: An Environmental History (The Florida History and Culture). Gainesville, FL: The University of Florida Press McCormick, P.V. and M.B. O’Dell. 1996. Quantifying periphyton responses to phosphorus in the Florida Everglades: a synoptic-experimental approach. J. N. Am. Benthol. Soc. 15:450–468 McCormick, P.V., P.S. Rawlik, K. Lurding, E.P. Smith, and F.H. Sklar. 1996. Periphyton–water quality relationships along a nutrient gradient in the northern Everglades. J. N. Am. Benthol. Soc. 15:433–449 McCormick, P.V., M.J. Chimney, and D.R. Swift. 1997. Diel oxygen profiles and water column community metabolism in the Florida Everglades, USA. Arch. Hydrobiol. 140:117–129 McCormick, P.V., R.B.E. Shuford, III, J.G. Backus, and W.C. Kennedy. 1998. Spatial and seasonal patterns of periphyton biomass and productivity in the northern Everglades, Florida, USA. Hydrobiologia 362:185–208

660

Bibliography

McCormick, P.V., S. Newman, S. Miao, R. Reddy, D. Gawlik, C. Fitz, et al. 1999. Ecological needs of the Everglades. In 1999 Everglades Interim Report, ed. G. Redfield, Chapter 3. West Palm Beach, FL: South Florida Water Management District McCormick, P.V., S. Newman, G. Payne, S. Miao, and T. Fontaine. 2000. Ecological effects of P enrichment. In 2000 Everglades Consolidated Report, Chapter 3. West Palm Beach, FL: South Florida Water Management District McCormick, P.V., S. Newman, S.L. Miao, D.E. Gawlik, D. Marley, K.R. Reddy, and T.D. Fontaine. 2001. Effects of anthropogenic phosphorus inputs on the Everglades. In The Everglades, Florida Bay, and Coral Reefs of the Florida Keys: An Ecosystem Sourcebook, eds. J.W. Porter and K.G. Porter, pp. 83–126. Boca Raton, FL: CRC McDowell, L.L., J.C. Stephens, and E.H. Stewart. 1969. Radiocarbon chronology of the Florida Everglades peat. Soil Sci. Soc. Am. Proc. 33:743–745 McElhinny, M.W. 1973. Paleomagnetism and Plate Tectonics. Cambridge: Cambridge University Press McGarigal, K. and B.J. Marks. 1995. Fragstats: Spatial pattern analysis program for quantifying landscape structure. Portland, OR: USDA Forest Service, Pacific Northwest Research Station. General Technical Report PNW-GTR-351 McKelvie, I.D. 2005. Separation, preconcentration and speciation of organic phosphorus in environmental samples. In Organic Phosphorus in the Environment, eds. B.L. Turner, E. Frossard, E.T.H. Lindau, and D.S. Baldwin, pp. 1–20. Cambridge, MA: CAB International McKnight, D., E.M. Thurman, and R.L. Wershaw. 1985. Biogeochemistry of aquatic humic substances in Thoreau’s bog, Concord, Massachusetts. Ecology 66:1339–1352 McNaughton, S.J. 1966. Ecotype function in the Typha community-type. Ecol. Monogr. 36:297–325 McPherson, B.F., G.Y. Hendrix, H. Klein, and H.M. Tysus. 1976. The Environment of South Florida: A Summary Report. U.S.G.S. Professional Paper 1011. Washington, DC: US Geological Survey Meffe, G.K. and C.R. Carroll. 1994. Principles of Conservation Biology, 2nd edition. Sunderland, MA: Sinauer. 1997; Groome, M.J., G.K. Meffe, and R.L. Carroll. 2006. 3rd edition Mermutt, A.R., J.C. Jain, L. Song, R. Kerrich, L. Kozak, and S. Jana. 1996. Trace element concentrations of selected soils and fertilizers in Saskatchewan, Canada. J. Environ. Qual. 25:845–853 van Meter-Kasanof, N.N. 1973. Ecology of the micro-algae of the Florida Everglades. Part 1. Environment and some aspects of freshwater periphyton, 1959–1963. Nova Hedwigia 24:619–664 Meyer, E. 1989. The relationship between body length parameters and dry mass in running water invertebrates. Arch. Hydrobiol. 117:191–203 Miao, S.L. and F.H. Sklar. 1998. Biomass and nutrient allocation of sawgrass and cattail along a nutrient gradient in the Florida Everglades. Wetlands Ecol. Manage. 5:245–263 Miao, S.L., R.E. Borer, and F.H. Sklar. 1997. Sawgrass seedling response to transplanting and nutrient additions. Restor. Ecol. 5:162–168 Miao, S.L., S. Newman, and F.H. Sklar. 2000. Effects of habitat nutrients and seed sources on growth and expansion of Typha domingensis. Aquat. Bot. 68(4):297–311 Mierle, G. and R. Ingram. 1991. The role of humic substances in the mobilization of mercury from watersheds. Water Air Soil Pollut. 56:349–357 Mihuc, T.B. 1997. The functional trophic role of lotic primary consumers: generalist vs. specialist strategies. Freshwater Biol. 37:455–462 Mikkelsen, D.S. and Brandon, D.M. 1975. Zinc deficiency in California rice. Calif. Agric. 29:8–9 Miller, J.A. 1997. Hydrogeology of Florida. In The Geology of Florida, eds. A.F. Randazzo and D.S. Jones, pp. 69–88. Tallahassee, FL: University of Florida Press Milligan, M.R. 1997. Identification Manual for the Aquatic Oligochaeta of Florida, Vol. I: Freshwater Oligochaetes. Tallahassee, FL: Florida Department of Environmental Protection

Bibliography

661

Minchin, P.R. 1987. An evaluation of the relative robustness of techniques for ecological ordination. Vegetatio 69:89–107 Miskimmin, B.M., J.W.M. Rudd, and C.A. Kelly. 1992. Influence of dissolved organic carbon, pH and microbial respiration rates on mercury methylation and demethylation in lake water. Can. J. Fish. Aquat. Sci. 49:17–22 Mitsch, W.J. and J.G. Gosselink. 2000. Wetlands, 3rd edition. New York: Wiley Moeller, R.E. 1978. Carbon uptake by the submerged hydrophyte Utricularia purpurea. Aquat. Bot. 55:209–216 Mooney, H.A., P.J. Ferrar, and R.O. Slayter. 1978. Photosynthetic capacity and carbon allocation patterns in diverse growth forms of Eucalyptus. Oecologia 36:103–111 Moore, P.D. and D.J. Bellamy. 1974. Peatlands. Berlin Heidelberg New York: Springer Moore, P.D., J.A. Webb, and M.E. Collinson. 1991. Pollen Analysis, 2nd edition. Oxford, UK: Blackwell Scientific Publications Mopper, K., X. Zhou, R.J. Kieber, D.J. Kieber, R.J. Sikorski, and R.D. Jones. 1991. Photochemical degradation of dissolved organic carbon and its impact on the oceanic carbon cycle. Nature 352:60–62 Moss, B. 1976. The effects of fertilization and fish community structure and biomass of aquatic macrophytes and epiphytic algal populations. An ecosystem experiment. J. Ecol. 64:313–342 Moss, D., J.F. Wright, M.T. Furse, and R.T. Clarke. 1999. A comparison of alternative techniques for prediction of the fauna of running-water sites in Great Britain. Freshwater Biol. 41:167–181 Mueller-Dombois, D. and H. Ellenberg. 1974. Aims and Methods of Vegetation Ecology. New York: Wiley Mundie, J.H., K.S. Simpson, and C.J. Perrin. 1991. Responses of stream periphyton and benthic insects to increases in dissolved inorganic phosphorus in a mesocosm. Can. J. Fish. Aquat. Sci. 48:2061–2072 Murdoch, W.W. 1966. Community structure, population control and competition – a critique. Am. Nat. 100:219–226 Murkin, H.R. 1989. The basis for food chains in prairie wetlands. In Northern Prairie Wetlands, ed. A.G. van der Valk, pp. 316–338. Ames, IA: Iowa State University Press Murkin, H.R., D.A. Wrubleski, and F.A. Reid. 1994a. Sampling invertebrates in aquatic and terrestrial habitats. In Research and Management Techniques for Wildlife and Habitats, ed. T.A. Boukhout, pp. 349–369. Bethesda, MD: The Wildlife Society Murkin, H.R., J.B. Pollard, M.P. Stainton, J.A. Boughen, and R.D. Titman. 1994b. Nutrient additions to wetlands in the Interlake region of Manitoba, Canada: effects of periodic additions throughout the growing season. Hydrobiologia 279/280:483–495 Murphy, T.P., K.J. Hall, and I. Yesaki. 1983. Co-precipitation of phosphate with calcite in a naturally eutrophic lakes. Limnol. Oceanogr. 28:58–69 Nealon, D. 1984. Groundwater Quality Study of the Water Conservation Areas. DRE-180. West Palm Beach, FL: South Florida Water Management District Nedoma, J., J. Vrba, T. Hanzl, and L. Nedbalová. 2001. Quantification of pelagic filamentous microorganisms in aquatic environments using the live-intercept method. FEMS Microbiol. Ecol. 38:81–85 Neill, C. and J.C. Cornwell. 1992. Stable carbon, nitrogen, and sulfur isotopes in a prairie marsh food web. Wetlands 12:217–224 Nelson, N.F. and R.H. Dietz. 1966. Cattail control methods in Utah. Salt Lake City, UT: Utah Department of Fish and Game. Publication 66-2 Nepf, H.M. 1999. Drag, turbulence, and diffusion in flow through emergent vegetation. Water Resour. Res. 35:479–489 Newbold, J.D., J.W. Elwood, M.S. Schulze, R.W. Stark, and J.C. Barmeier. 1983. Continuous ammonium enrichment of a woodland stream: uptake kinetics, leaf decomposition and nitrification. Freshwater Biol. 13:193–204 Newman, F.I. 1966. A method of estimating the total length of root in a sample. J. Appl. Ecol. 3:139–145

662

Bibliography

Newman, S. and K.R. Reddy. 1992. Sediment resuspension effects on alkaline phosphatase activity. Hydrobiologia 245:75–86 Newman, S., J.B. Grace, and J.W. Koebel. 1996. The effects of nutrients and hydroperiod on mixtures of Typha domingensis, Cladium jamaicense, and Eleocharis interstincta: implications for Everglades restoration. Ecol. Appl. 6:774–783 Newman, S., J. Schuette, J.B. Grace, K. Rutchey, T. Fontaine, K.R. Reddy, and M. Pietrucha. 1998. Factors influencing cattail abundance in the northern Everglades. Aquat. Bot. 60:265–280 NOAA (National Oceanic and Atmospheric Administration), NNDC Climate Data Online, http:// www.ncdc.noaa.gov/onlineprod/drought/xmgr.html (accessed 21 March 2007) Noe, G.B., D.L. Childers, and R.D. Jones. 2001. Phosphorus biogeochemistry and the impact of phosphorus enrichment: why is the Everglades so unique? Ecosystems 4:603–624 Noe, G.B., D.L. Childers, A.L. Edwards, E. Gairser, K. Jayachandran, D. Lee, et al. 2002. Short-term changes in an oligotrophic Everglades wetland ecosystem receiving experimental nutrient enrichment. Biogeochemistry 59:239–267 Noe, G.B., L.J. Scinto, J. Taylor, D.L. Childers, and R.D. Jones. 2003. Phosphorus cycling and partitioning in an oligotrophic Everglades wetland ecosystem: a radioisotope tracing study. Freshwater Biol. 48:1993–2008 Obeysekera, J. and K. Rutchey. 1997. Selection of scale for Everglades landscape models. Landscape Ecol. 12(1):7–18 Odum, E.P. 1971. Fundamentals of Ecology, 3rd edition. Philadelphia, PA: W.B. Saunders Odum, E.P. 1984. The mesocosm. BioScience 34:558 Odum, E.P., J.T. Finn, and E.H. Franz. 1979. Perturbation theory and the subsidy–stress gradient. BioScience 29:349–352 Oksanen, J. 1996. Is the humped–shaped relationship between species richness and biomass an artefact due to plot size? J. Ecol. 84:293–295 Oksanen, L., S.D. Fretwell, J. Arruda, and P. Niemela. 1981. Exploitation ecosystems in gradients of primary productivity. Am. Nat. 118:240–261 Oliver, I. and A.J. Beattie. 1996. Invertebrate morphospecies as surrogates for species: a case study. Conserv. Biol. 10:99–109 Olsen, S.R. and L.E. Sommers. 1982. Phosphorus. In Methods of Soil Analysis, eds. A.L. Page, R.H. Miller, and D.R. Keeney, pp. 403–430. Madison, WI: American Society of Agronomy Olsson, U. 1986. Radiometric dating. In Handbook of Holocene Palaeoecology and Palaeohydrology, ed. B.E. Berglund, pp. 273–312. New York: Wiley O’Neill, R.V., D.L. DeAngelis, J.B. Waide, and T.F.H. Allen. 1986. A Hierarchical Concept of the Ecosystem. Princeton, NJ: Princeton University Press Ornes, W.H. and K.K. Steward. 1973. Effect of phosphorus and potassium on phytoplankton populations in field enclosures. Washington, DC: US Department of the Interior. South Florida Environmental Project Ecological Report No. DI-SFEP-74-07 Ostendorp, W. 1989. Die-back of reeds in Europe – a critical review of literature. Aquat. Bot. 35:5–26 Ozimek, T., E. Pieczynska, and A. Hankiewicz. 1991. Effects of filamentous algae on submerged macrophyte growth: a laboratory experiment. Aquat. Bot. 41:309–315 Pahl, J.W., C.J. Richardson, and P.V. Sundareshwar. 2004. Fire and disturbance as mechanisms to limit the expansion of Typha domingensis in the northern Everglades. Presentation at the Seventh Intecol International Wetlands Conference (July 29), Utrecht, Netherlands Paine, R.T. 1966. Food web complexity and species diversity. Am. Nat. 100:65–75 Paine, R.T. 1974. Intertidal community structure: experimental studies on the relationship between a dominant competitor and its principal predator. Oecologia 15:93–120 Paine, R.T. 1980. Food webs: linkage interaction strength and community infrastructure. J. Anim. Ecol. 49:667–685 Palmer, M.W. 1990. The estimation of species richness by extrapolation. Ecology 71:1195–1198 Palmer, M.A., C.M. Swan, K. Nelson, P. Silver, and R. Alvestad. 2000. Streambed landscapes: evidence that stream invertebrates respond to the type and spatial arrangement of patches. Landscape Ecol. 15:563–576

Bibliography

663

Pan, Y., R.J. Stevenson, P. Vaithiyanathan, J. Slate, and C.J. Richardson. 2000. Changes in algal assemblages along observed and experimental phosphorus gradients in a subtropical wetland, USA. Freshwater Biol. 44:339–353 Parker, G.G. and C.W. Cooke. 1944. Late Cenozoic Geology of Southern Florida, with a Discussion of the Ground Water. Bulletin 27. Tallahassee, FL: Florida Geological Survey Parker, G.G., G.E. Ferguson, and S.K. Love. 1955. Water Resources of Southeastern Florida. Washington, DC: US Geological Survey. Water-Supply Paper 1255 Patrick, R. and C.W. Reimer. 1966 and 1975. The Diatoms of the United States. Vols. 1 and 2. Monographs of the Academy of Natural Sciences of Philadelphia, No. 13 Payne, G.G., K.C. Weaver, and S.K. Xue. 2006. Status of phosphorus and nitrogen in the Everglades protection area. In 2006 South Florida Environmental Report, Chapter 2c. West Palm Beach, FL: South Florida Water Management District Perkins, R.D. 1977. Depositional framework of Pleistocene rocks in South Florida. In Quaternary Sedimentation in South Florida, eds. P. Enos and R.D. Perkins, pp. 131–198. Boulder, CO: Geological Society of America Perrin, C.J. and J.S. Richardson. 1997. N and P limitation of benthos abundance in the Nechako River, British Columbia. Can. J. Fish. Aquat. Sci. 54:2574–2583 Pescador, M.L., A.K. Rasmussen, and S.C. Harris. 1995. Identification Manual for the Caddisfly (Trichoptera) Larvae of Florida. Tallahassee, FL: Florida Department of Environmental Protection Peters, N.E. and Reese, R.S. 1995. Spatial and temporal variations of weekly atmospheric deposition in a small area on the shore of Lake Okeechobee, Florida. Atmos. Environ. 29:179–187 Peterson, B.J., L. Deegan, J. Helfrich, J.E. Hobbie, M. Hullar, B. Moller, T.E. Ford, A. Hershey, A. Hiltner, G. Kipphut, A. Lock, D.M. Fiebig, V. McKinley, M.C. Miller, R. Vestal, R. Ventullo, and G. Volk. 1993. Biological responses of a tundra river to fertilization. Ecology 74:653–672 Pettersson, K. and M. Jansson. 1978. Determination of phosphatase activity in lake water – a study of methods. Int. Ver. Theor. Ange. Limnol. 20:1226–1230 Petuch, E.J. 1986. The Pliocene reefs of Miami: their geomorphological significance in the evolution of the Atlantic Coastal Ridge, Southeastern Florida, USA. J. Coastal Res. 2:391–408 Phillips, E.A. 1959. Methods of Vegetation Study. New York: Holt, Rinehart and Winston Phillips, G.L. 1978. A mechanism to account for the decline in macrophytes in progressively eutrophic freshwater wetlands. Aquat. Bot. 4:103–126 Phillips, J.D. 1999. Divergence, convergence, and self-organization in landscapes. Ann. Assoc. Am. Geogr. 89:466–88 Pielou, E.C. 1984. The Interpretation of Ecological Data: A Primer on Classification and Ordination. New York: Wiley Pimm, S.L., J.H. Lawton, and J.E. Cohen. 1991. Food web patterns and their consequences. Nature 350:669–674 Plafkin, J.L., M.T. Barbour, K.D. Porter, S.K. Gross, and R.M. Hughes. 1989. Rapid Bioassessment Protocols for Use in Streams and Rivers: Benthic Macroinvertebrates and Fish. EPA/440/489-001. Washington, DC: Office of Water, US Environmental Protection Agency Pluchino, E.S. 1984. Guide to the Common Water Mite Genera of Florida. Technical Series Vol. 7, No. 1. Tallahassee, FL: Florida Department of Environmental Regulation Ponzio, K.J., S.J. Miller, and M.A. Lee. 1995. Germination of sawgrass, Cladium jamaicense Crantz, under varying hydrologic conditions. Aquat. Bot. 51:115–120 Porter, J.W. and K.F. Porter. 2002. The Everglades, Florida Bay, and Coral Reefs of the Florida Keys: An Ecosystem Sourcebook. Boca Raton, FL: CRC Preston, F.W. 1948. The commonness, and rarity, of species. Ecology 29:254–283 Puri, H.S. and R.O. Vernon. 1964. Summary of the Geology of Florida and Guidebook to the Classic Exposures. Tallahassee, FL: Florida Geological Survey. Special Publication No. 5 Qian, S.S. 1997a. Estimating the area affected by phosphorus runoff in an Everglades wetland: a comparison of universal kriging and Bayesian kriging (with discussion). Environ. Ecol. Stat. 4:1–29

664

Bibliography

Qian, S.S. 1997b. An illustration of model structure identification. J. Am. Water Resour. Assoc. 33:811–824 Qian, S.S. and C.W. Anderson. 1999. Exploring factors controlling the variability of pesticide concentrations in the Willamatte River basin using tree-based models. Environ. Sci. Technol. 33:3332–3340 Qian, S.S. and K.H. Reckhow. 1998. Modeling phosphorus trapping in wetlands using nonparametric Bayesian regression. Water Resour. Res. 34:1745–1754 Qian, S.S. and C.J. Richardson. 1997a. Testing the feasibility of using first order kinetics for modeling phosphorus in an Everglades wetland (WCA-2A): a dynamic linear modeling approach. In Effects of Phosphorus and Hydroperiod Alterations on Ecosystem Structure and Function in the Everglades, ed. C.J. Richardson, Chapter 20. Durham, NC: Nicholas School of the Environment, Duke University Qian, S.S. and C.J. Richardson. 1997b. Estimating the long-term phosphorus accretion rate in the Everglades: a Bayesian approach with risk assessment. Water Resour. Res. 33:1681–1688 Qian, S.S., R.S. King, and C.J. Richardson. 2003. Two statistical methods for the detection of environmental thresholds. Ecol. Model. 166:87–97 Qian, S.S., Y. Pan, and R.S. King. 2004. Soil total phosphorus threshold in the Everglades: a Bayesian changepoint analysis for multinomial response data. Ecol. Indicators 4:29–37 Qualls, R.G. 1984. The role of leaf litter nitrogen immobilization in the nitrogen budget of a swamp stream. J. Environ. Qual. 13:640–644 Qualls, R.G. and B.L. Haines. 1991. Geochemistry of dissolved organic nutrients in water percolating through a forest ecosystem. Soil Sci. Soc. Am. J. 52:1112–1123 Qualls, R.G. and C.J. Richardson. 1992. Gradient study – water chemistry along the gradient: comparison of 1990 and 1991 periods. In Effects of Nutrient Loadings and Hydroperiod Alterations on Control of Cattail Expansion, Community Structure and Nutrient Retention in the Water Conservation Areas of South Florida, ed. C.J. Richardson, pp. 196–285. Annual Report. Durham, NC: School of the Environment, Duke University Wetland Center publication 92-11 Qualls, R.G. and C.J. Richardson. 1995. Forms of soil phosphorus along a nutrient enrichment gradient in the northern Everglades. Soil Sci. 160:183–198 Qualls, R.G. and C.J. Richardson. 2000. Phosphorus enrichment affects litter decomposition, immobilization, and soil microbial phosphorus in wetland mesocosms. Soil Sci. Soc. Am. J. 64:799–808 Qualls, R.G. and C.J. Richardson. 2003. Factors controlling concentration, export, and decomposition of dissolved organic nutrients in the Everglades of Florida. Biogeochemistry 62(2):197–229 Qualls, R.G., M.H. Dorfman, and J.D. Johnson. 1989. Evaluation of the efficiency of ultraviolet disinfection systems. Water Res. 23:317–325 Qualls, R.G., C.J. Richardson, R. Johnson, and J. Zahina. 1994. Dosing study – response of Everglades slough communities to increased concentrations of PO4: operation of experimental field mesocosms. In Annual Report: Effects of Nutrient Loadings and Hydroperiod Alterations on Control of Cattail Expansion, Community Structure and Nutrient Retention in the Water Conservation Areas of South Florida, eds. C.J. Richardson, B.C. Craft, R.G. Qualls, R.J. Stevenson, and P. Vaithiyanathan. Chapter 3. Durham, NC: School of the Environment, Duke University Qualls, R.G., C.J. Richardson, and L.J. Sherwood. 2001. Soil reduction-oxidation potential along a nutrient enrichment gradient in the Everglades. Wetlands 21:403–411 Rader, R.B. and C.J. Richardson. 1992. The effects of nutrient enrichment on algae and macroinvertebrates in the Everglades: a review. Wetlands 12:121–135 Rader, R.B. and C.J. Richardson. 1994. Response of macroinvertebrates and small fish to nutrient enrichment in the Everglades. Wetlands 14:134–146 Raftery, A.E. and V.E. Akman. 1986. Bayesian analysis of a poison process with a change-point. Biometrika 73:85–89

Bibliography

665

Rahmstorf, S.A., Cazenave, J.A. Church, J.E. Hansen, R.F. Keeling, D.E. Parker, and R.C.J. Somerville. 2007. Recent climate observations compared to projections. Science 316:709 Raschke, R.L. 1993. Diatom (Bacillariophyta) community response to phosphorus in the Everglades National Park, USA. Phycologia 32:48–58 Reckhow, K.H. and S.S. Qian. 1994. Modeling phosphorus trapping in wetlands using generalized additive models. Water Resour. Res. 30:3105–3114 Reddy, K.R. and W.H. Patrick, Jr. 1984. Nitrogen transformations and loss in flooded soils and sediments. CRC Crit. Rev. Environ. Contr. 13:273–309 Reddy, K.R. and P.S.C. Rao. 1983. Nitrogen and phosphorus fluxes from a flooded organic soil. Soil Sci. 136:300–307 Reddy, K.R., W.F. DeBusk, Y. Wang, R. DeLaune, and M. Koch. 1991. Physico-chemical properties of soils in the Water Conservation Area 2 of the Everglades. West Palm Beach, FL: Final report submitted to South Florida Water Management District Reddy, K.R., R.D. DeLaune, W.F. DeBusk, and M.S. Koch. 1993. Long-term nutrient accumulation rates in the Everglades. Soil Sci. Soc. Am. J. 57:1147–1155 Reddy, K.R., W.F. DeBusk, Y. Wang, and S. Newman. 1994a. Physico-chemical properties of soils in the Water Conservation Area 1 of the Everglades. West Palm Beach, FL: Final report submitted to South Florida Water Management District Reddy, K.R., W.F. DeBusk, Y. Wang, and S. Newman. 1994b. Physico-chemical properties of soils in the Water Conservation Area 3 of the Everglades. West Palm Beach, FL: Final report submitted to South Florida Water Management District Reddy, K.R., Y. Wang, W.F. DeBusk, M.M. Fisher, and S. Newman. 1998. Forms of soil phosphorus in selected hydrologic units of the Florida Everglades. Soil Sci. Soc. Am. J. 62:1134–1147 Redfield, A.C. 1958. The biological control of chemical factors in the environment. Am. Sci. 46:205–21 Redfield, G.W. and N.H. Urban, eds. 1997. Proceedings of the Conference on Atmospheric Deposition into South Florida. Measuring Net Atmospheric Inputs of Nutrients. West Palm Beach, FL: South Florida Water Management District Reice, S.R. 1994. Nonequilibrium determinants of biological community structure. Am. Sci. 82:424–435 Rejmankova, E. 2001. Effect of experimental phosphorus enrichment on oligotrophic tropical marshes in Belize, Central America. Plant Soil 236:22–53 Reyes, E. and M. Merino. 1991. Diel dissolved oxygen dynamics and eutrophication in a shallow, well-mixed Tropical lagoon (Cancun, Mexico). Estuaries 14:372–381 Reynolds, J.B. 1996. Electrofishing. In Fisheries Techniques, 2nd edition, eds. B.R. Murphy and D.W. Willis, pp. 221–251. Bethesda, MD: American Fisheries Society Reynoldson, T.B., R.C. Bailey, K.E. Day, and R.H. Norris. 1995. Biological guidelines for freshwater sediment based on BEnthic Assessment of SedimenT (the BEAST) using a multivariate approach for predicting biological state. Aust. J. Ecol. 20:198–219 Reynoldson, T.B., R.H. Norris, V.H. Resh, K.E. Day, and D.M. Rosenberg. 1997. The reference condition: a comparison of multimetric and multivariate approaches to assess water-quality impairment using benthic macroinvertebrates. J. N. Am. Benthol. Soc. 16:833–852 Richardson, C.J. 1985. Mechanisms controlling phosphorus retention capacity in freshwater wetlands. Science 228:1424–1427 Richardson, C.J. 1991. Biogeochemical cycles: regional. In Wetlands and Shallow Continental Water Bodies, ed. B.C. Patten, pp. 259–279. The Hague, The Netherlands: SPB Academic Richardson, C.J. 1995. Wetlands Ecology. In: Encyclopedia of Environmental Biology, ed. W.A. Nierenberg, Vol. 3, pp. 535–550. San Diego, CA: Academic Richardson, C.J., ed. 1997. Effects of Phosphorus and Hydroperiod Alterations on Ecosystem Structure and Function in the Everglades. Duke Wetland Center Publication 97-05. Durham, NC: Nicholas School of the Environment, Duke University Richardson, C.J. 1999. The role of wetlands in storage, release, and cycling of phosphorus on the landscape: a 25-year retrospective. In Phosphorus Biogeochemistry in Sub-Tropical Ecosystems, eds. K.R. Reddy, G.A. O’Connor, and C.L. Schelske, pp. 47–68. Boca Raton, FL: CRC/Lewis

666

Bibliography

Richardson, C.J. 2000. Freshwater Wetlands. In North American Terrestrial Vegetation, eds. M.G. Barbour and W.D. Billings, pp. 448–499. New York: Cambridge University Press Richardson, C.J. and Craft, C.B. 1990. Phase One: A Preliminary Assessment of Nitrogen and Phosphorus Accumulation and Surface Water Quality in Water Conservation Areas 2A and 3A of South Florida. Duke Wetland Center Publication 90-01. Durham, NC: School of Forestry and Environmental Studies, Duke University Richardson, C.J. and C.B. Craft. 1993. Effective phosphorus retention in wetlands: fact or fiction. In Constructed Wetlands for Water Quality Improvement, ed. G.A. Moshiri, pp. 271–282. Boca Raton, FL: CRC/Lewis Richardson, C.J. and N.A. Hussain. 2006. Restoring the Garden of Eden: an ecological assessment of the Marshes of Iraq. BioScience 56:477–489 Richardson, C.J. and J.K. Huvane. 2001. Everglades Restoration: A Primer. Final Report to the EAA Environmental Protection District, March 2001. Durham, NC: Nicholas School of the Environment and Earth Sciences, Duke University Richardson, C.J. and P.E. Marshall. 1986. Processes controlling movement, storage and export of phosphorus in a fen peatland. Ecol. Monogr. 56:279–302 Richardson, C.J. and S.S. Qian. 1999. Long-term phosphorus assimilative capacity in freshwater wetlands: a new paradigm for sustaining ecosystem structure and function. Environ. Sci. Technol. 33:1545–1551 Richardson, C.J. and B.R. Schwegler. 1986. Algal bioassay and gross productivity experiments using sewage effluent in a Michigan wetland. Water Res. Bull. 22:111–120 Richardson, C.J. and P. Vaithiyanathan. 1995. Phosphorus sorption characteristics of Everglades soils along a eutrophication gradient. Soil Sci. Soc. Am. J. 59:1782–1788 Richardson, C.J. and P. Vaithiyanathan. 1997. A preliminary analysis of three years of atmospheric nutrient deposition in WCA-2A. In Proceedings of the Conference on Atmospheric Deposition into South Florida. Measuring Net Atmospheric Inputs of Nutrients, eds. G.W. Redfield and N.H. Urban, pp. 41–42. West Palm Beach, FL: South Florida Water Management District Richardson, K., J. Beardall, and J.A. Raven. 1983. Adaptation of unicellular algae to irradiance: an analysis of strategies. New Phytol. 93:157–191 Richardson, C.J., C.B. Craft, R.G. Qualls, R.B. Rader, and R.R. Johnson. 1991. Annual Report: Effects of Nutrient Loadings and Hydroperiod Alterations on Control of Cattail Expansion, Community Structure and Nutrient Retention in the Water Conservation Areas of South Florida. Duke University Wetland Center publication 91-08. Durham, NC: School of the Environment, Duke University Richardson, C.J., C.B. Craft, R.R. Johnson, R.G. Qualls, R.B. Rader, L. Sutter, and J. Vymazal. 1992. Effects of Nutrient Loadings and Hydroperiod Alterations on Control of Cattail Expansion, Community Structure and Nutrient Retention in the Water Conservation Areas of South Florida. Wetland Center Publication 92-11. Durham, NC: Duke University Richardson, C.J., C.B. Craft, R.R. Johnson, R.G. Qualls, R.B. Rader, L. Sutter, and J. Vymazal. 1993. Annual Report: Effects of Nutrient Loadings and Hydroperiod Alterations on Control of Cattail Expansion, Community Structure and Nutrient Retention in the Water Conservation Areas of South Florida. Duke Wetland Center Publication 92-11. Durham, NC: School of the Environment, Duke University Richardson, C.J., C.B. Craft, R.G. Qualls, R.J. Stevenson, and P. Vaithiyanathan. 1994. Annual Report: Effects of Nutrient Loadings and Hydroperiod Alterations on Control of Cattail Expansion, Community Structure and Nutrient Retention in the Water Conservation Areas of South Florida. Duke Wetland Center Publication 94–08. Durham, NC: School of the Environment, Duke University Richardson, C.J., C.B. Craft, R.G. Qualls, R.J. Stevenson, P. Vaithiyanathan, M. Bush, and J. Zahina. 1995. Annual Report: Effects of Phosphorus and Hydroperiod Alterations on Ecosystem Structure and Function in the Everglades. Duke Wetland Center Publication 95–05. Durham, NC: Nicholas School of the Environment, Duke University Richardson, C.J., S.S. Qian, C.B. Craft, and R.G. Qualls. 1997a. Predictive models for phosphorus retention in wetlands. Wetlands Ecol. Manage. 4:159–175

Bibliography

667

Richardson, C.J., P. Vaithiyanathan, E.A. Romanowicz, and C.B. Craft. 1997b. Macrophyte community responses in the Everglades with an emphasis on cattail (Typha domingensis) and sawgrass (Cladium jamaicense) interactions along a gradient of long-term nutrient additions, altered hydroperiod and fire. In Effects of Phosphorus and Hydroperiod Alterations on Ecosystem Structure and Function in the Everglades, ed. C.J. Richardson, Chapter 14. Durham, NC: Nicholas School of the Environment, Duke University Richardson, C.J., P. Vaithiyanathan, R.G. Qualls, and C. Stow. 1997c. Dosing study chemistry analysis: four-year response (1992–1996) of Everglades sloughs to increased concentrations of PO4: operation of experimental field mesocosms and water quality analysis. In Effects of Phosphorus and Hydroperiod Alterations on Ecosystem Structure and Function in the Everglades, ed. C.J. Richardson, Chapter 15. Durham, NC: Nicholas School of the Environment, Duke University Richardson, C.J., R.G. Qualls, and P. Vaithiyanathan. 1997d. Dosing study – changes in macrophyte community composition and calcareous mat cover over four years of P additions to Everglades mesocosms. In Effects of Phosphorus and Hydroperiod Alterations on Ecosystem Structure and Function in the Everglades, ed. C.J. Richardson, Chapter 17. Durham, NC: Nicholas School of the Environment, Duke University Richardson, C.J., G.M. Ferrell, and P. Vaithiyanathan. 1999. Nutrient effects on stand structure, resorption efficiency, and secondary compounds in Everglades sawgrass. Ecology 80:2182–2192 Richardson, C.J., P. Vaithiyanathan, R.J. Stevenson, R.S. King, C.A. Stow, R.G. Qualls, and S.S. Qian. 2000a. The Ecological Basis for a Phosphorus (P) Threshold in the Everglades: Directions for Sustaining Ecosystem Structure and Function. Duke Wetland Center Publication 2000-02. Durham, NC: Nicholas School of the Environment, Duke University Richardson, C.J., P. Vaithiyanathan, J. Vymazal, J. Komárková, K. Kubecková, J. Kaštovský, and P. Znachor. 2000b. Everglades Monitoring, Recovery and Restoration. Phase II Year-End Report. Durham, NC: Duke University Wetland Center Richardson, C.J., R.S. King, S.S. Qian, P. Vaithiyanathan, C.A. Stow, and R.G. Qualls. 2003. A Scientific Basis for Determining Phosphorus Imbalance Effects in the Everglades. Final Report to Environmental Regulation Commission, State of Florida. Duke University Wetland Center Publication 2003-3. Durham, NC: Nicholas School of the Environment and Earth Sciences, Duke University Richardson, C.J., P. Reiss, N.A. Hussain, A.J. Alwash, and D.J. Pool. 2005. The restoration potential of the Mesopotamian marshes of Iraq. Science 307:1307–1311 Richardson, C.J., R.S. King, S.S. Qian, P. Vaithiyanathan, and R.G. Qualls. 2007. Estimating ecological thresholds for phosphorus in the Everglades. Environ. Sci. Technol. 41(28):8084–8091 Ritter, A. and R. Muñoz-Carpena. 2006. Dynamic factor modeling of ground and surface water levels in an agricultural area adjacent to Everglades National Park. J. Hydrol. 317:340–354 Rivard, P.G. and P.M. Woodard. 1989. Light, ash, and pH effects on the germination and seedling growth of Typha latifolia (cattail). Can. J. Bot. 67:2783–2787 Rivero, R.G., S. Grunwald, T.Z. Osborne, K.R. Reddy, and S. Newman. 2007. Characterization of the spatial distribution of soil properties in Water Conservation Area 2A, Everglades, Florida. Soil Sci. 172(2):149–166 Robertson, W.B. 1953. A Survey of the Effects of Fire in Everglades National Park. Washington, DC: US Department of Interior, National Park Service Robertson, W.B., Jr. and P.C. Frederick. 1994. The faunal chapters: contexts, synthesis, and departures. In Everglades: The Ecosystem and Its Restoration, eds. S.M. Davis and J.C. Ogden, pp. 709–737. Delray Beach, FL: St. Lucie Romanowicz, E.A. and C.J. Richardson. 1997. Hydrologic investigation of Water Conservation Area 2A. In Effects of Phosphorus and Hydroperiod Alterations on Ecosystem Structure and Function in the Everglades, ed. C.J. Richardson, pp. 12.1–12.29. Durham, NC: Nicholas School of the Environment, Duke University Romanowicz, E.A., C.J. Richardson, and P. Vaithiyanathan. 1996. Evidence for differential ground-water flow inputs and multiple surface-water flow domains in the northern Everglades (Water Conservation Area 2-A), Florida. EOS, Trans. Am. Geophys. Union Suppl. 77:140

668

Bibliography

Rood, B.E., J.F. Gottgens, J.J. Delfino, C.D. Earle, and T.L. Crisman. 1995. Mercury accumulation trends in Florida Everglades and Savannas Marsh flooded soils. Water Air Soil Pollut. 80:981–990 Rosenberg, D.M. and Resh, V.H. (eds.). 1993. Freshwater Biomonitoring and Benthic Macroinvertebrates. New York: Chapman and Hall Rosendahl, P.C. and P.W. Rose. 1981. Freshwater flow rates and distribution within the Everglades marsh. In Proceedings of the National Symposium on Freshwater Inflow to Estuaries, eds. R.D. Cross and D.L. Williams. San Antonio, TX, 9–11 September 1980. Vol. II. Slidell, LA: US Fish and Wildlife Service Rosenzweig, M.L. 1971. Paradox of enrichment: destabilization of exploitation ecosystems in ecological time. Science 171:385–387 Rosenzweig, M.L. and Z. Abramsky. 1993. How are diversity and productivity related? In Species Diversity in Ecological Communities: Historical and Geographical Perspectives, eds. R.E. Ricklefs and D. Schluter, pp. 52–65. Chicago, IL: University of Chicago Press Russell-Hunter, W.D. 1970. Aquatic Productivity: An Introduction to Some Basic Aspects of Biological Oceanography and Limnology. London, UK: MacMillan Rutchey, K. and L. Vilcheck. 1994. Development of an Everglades vegetation map using a SPOT image and the global positioning system. Photogramm. Eng. Remote Sens. 60:767–775 Rutchey, K. and L. Vilcheck. 1999. Air photo interpretation and satellite imagery analysis techniques for mapping cattail coverage in a northern Everglades impoundment. Photogramm. Eng. Remote Sens. 65(2):185–191 Rydin, H. and J.K. Jeglum. 2006. The Biology of Peatlands. New York: Oxford University Press Sale, P.J.M. and R.G. Wetzel. 1983. Growth and metabolism of Typha species in relation to cutting treatments. Aquat. Bot. 15:321–334 Sample, B.E., R.J. Cooper, R.D. Greer and R.C. Whitmore. 1993. Estimation of insect biomass by length and width. Am. Midl. Nat. 129:234–240 Sanchez, S.A. 1990. Soil-Testing and Fertilization Recommendations for Crop Production on Organic Soils in Florida. Gainesville, FL: University of Florida, Agricultural Experiment Station, Institute of Food and Agricultural Sciences. Bulletin 876 Sanderson, M.W. 1982. Aquatic and semiaquatic heteroptera. In Aquatic Insects and Oligochaetes of North and South Carolina, eds. A.R. Brigham, W.U. Brigham, and A. Gnilka, pp. 6.1–6.94. Mahomet, IL: Midwest Aquatic Enterprises Sanderson, R.A., S.P. Rushton, A.J. Cherrill, and P. Byrne. 1995. Soil, vegetation and space: an analysis of their effects on the invertebrate communities of a moorland in north–east England. J. App. Ecol. 32:506–518 Sand-Jensen, K. and J. Borum. 1991. Interactions among phytoplankton, periphyton and macrophytes in temperate freshwaters and estuaries. Aquat. Bot. 41:137–175 Sand-Jensen, K., N.P. Revsbach, and B.B. Jørgensen. 1985. Microprofiles of oxygen in epiphyte communities on submerged macrophytes. Mar. Biol. 89:55–66 Sanudo-Wilhelmy, S. and G.A. Gill. 1999. Impact of the clean air act on the levels of toxic metals in urban estuaries: the Hudson River estuary revisited. Environ. Sci. Technol. 33:3477–3491 SAS. 1988. Statistical Analysis System. SAS User’s Guide. Release 6.03. Cary, NC: SAS Institute Scheffer, M., S.H. Hosper, M.L. Meijer, B. Moss, and E. Jeppesen. 1993. Alternative equilibria in shallow lakes. Trends Ecol. Evol. 8:275–279 Schelske, C.L., A. Peplow, M. Brenner, and C.N. Spencer. 1994. Low-background gamma counting: applications for 210Pb dating of sediments. J. Paleolimnol. 10:115–128 Schlesinger, W.H. 1977. Carbon balance in terrestrial detritus. Annu. Rev. Ecol. Syst. 8:51–81 Schlesinger, W.H. 1978. Community structure, dynamics and nutrient cycling in the Okefenokee cypress swamp forest. Ecol. Monogr. 48:43–65 Schlesinger, W.H. 1997. Biogeochemistry: An Analysis of Global Change. 2nd edition. New York: Academic Schmidt, W. 1997. Geomorphology and Physiography of Florida, In The Geology of Florida, eds. A.F. Randazzo and D.S. Jones, pp. 1–12. Tallahassee, FL: University of Florida Press

Bibliography

669

Schoener, T.W. 1982. The controversy over interspecific competition. Am. Sci. 70:586–595 Schomer, N.S. and R.D. Drew. 1982. An Ecological Characterization of the Lower Everglades, Florida Bay, and the Florida Keys. FWS/OBS-82/58.1. Washington, DC: US Fish and Wildlife Service, Office of Biological Services Schroeder, L.D., D.L. Sjoquist, and P.E. Stephan. 1986. Understanding Regression Analysis: An Introductory Guide. Newberry Park, CA: Sage University Schueneman, T.J. and Sanchez, C.A. 1994. Vegetable production in the EAA. In Everglades Agricultural Area (EAA), eds. A.B. Bottcher and F.T. Izuno, pp. 238–277. Gainesville, FL: University Press of Florida Scott, T.M. 1997. Miocene to Holocene history of Florida. In The Geology of Florida, eds. A.F. Randazzo and D.S. Jones, pp. 57–68. Tallahassee, FL: University of Florida Press Scott, J.T., R.D. Doyle, and C.T. Filstrup. 2005. Periphyton nutrient limitation and nitrogen fixation potential along a wetland nutrient depletion gradient. Wetlands 25:439–448 SCS (Soil Conservation Service). 1978. Soil Survey of Palm Beach County Area of Florida. Washington, DC: US Government Printing Office Secretary of the Army. 1949. Comprehensive Report on Central and Southern Florida for Flood Control and Other Purposes. Washington, DC: US Government Printing Office. House Document 643 SFWMD (South Florida Water Management District). 1990. Surface Water Improvement and Management Plan for the Everglades. West Palm Beach, FL: South Florida Water Management District SFWMD (South Florida Water Management District). 1992. Surface Water Improvement and Management Plan for the Everglades. Supporting Information Document. West Palm Beach, FL: South Florida Water Management District SFWMD (South Florida Water Management District). 1994. Effects of Phosphorus and Hydrology on the Everglades. Unpublished handout from staff presentation to ERC. West Palm Beach, FL: South Florida Water Management District SFWMD (South Florida Water Management District). 1995. Users Guide to REMO, Remote Access to the South Florida Water Management District’s Water Quality and Hydrometeorological Databases. West Palm Beach, FL: South Florida Water Management District SFWMD (South Florida Water Management District). 1998. Natural Systems Model. Version 4.5 Documentation. West Palm Beach, FL: South Florida Water Management District, Hydrologic Systems Modeling Division SFWMD (South Florida Water Management District). 1999. 1999 Everglades Interim Report. West Palm Beach, FL: South Florida Water Management District SFWMD (South Florida Water Management District). 2000. 2000 Everglades Consolidated Report. West Palm Beach, FL: South Florida Water Management District SFWMD (South Florida Water Management District). 2001. 2001 Everglades Consolidated Report. West Palm Beach, FL: South Florida Water Management District SFWMD (South Florida Water Management District). 2002. 2002 Everglades Consolidated Report. West Palm Beach, FL: South Florida Water Management District SFWMD (South Florida Water Management District) 2003. 2003 Everglades Consolidated Report. West Palm Beach, FL: South Florida Water Management District SFWMD (South Florida Water Management District). 2004. 2004 Everglades Consolidated Report. West Palm Beach, FL: South Florida Water Management District SFWMD (South Florida Water Management District). 2005. 2005 South Florida Environmental Report. West Palm Beach, FL: South Florida Water Management District SFWMD (South Florida Water Management District). 2006. 2006 South Florida Environmental Report. West Palm Beach, FL: South Florida Water Management District SFWMD (South Florida Water Management District), Lake Okeechobee WSE Schedule, http:// www.sfwmd.gov/org/pld/hsm/reg_app/lok_reg/index.html (accessed 21 March 2007) Shannon, C.E. and W. Weaver. 1949. The Mathematical Theory of Communication. Urbana, IL: University of Illinois Press

670

Bibliography

Sharitz, R.R. and D.P. Batzer. 1999. An introduction to freshwater wetlands in North America and their invertebrates. In Invertebrates in Freshwater Wetlands of North America: Ecology and Management, eds. D.P. Batzer, R.B. Rader, and S.A. Wissinger, pp. 1–22. New York: Wiley Sharma, M. and R. Agarwal. 2003. Maximum likelihood method for parameter estimation in non-linear models with below detection data. Environ. Ecol. Stat. 10:445–454 Shelford, V.E. 1913. Animal Communities in Temperate America. Chicago, IL: University of Chicago Press Shuman, M.S. 1990. Carboxyl acidity of aquatic organic matter: possible systematic errors introduced by XAD extraction. In Organic Acids in Aquatic Ecosystems: Report to the Dahlem Workshop on Organic Acids in Aquatic Ecosystems, eds. E.M. Perdue and E.T. Gjessing, pp. 97–108. New York: Wiley Sifton, H.B. 1959. The germination of light-sensitive seeds of Typha latifolia L. Can. J. Bot. 37:719–739 Silveira, J E. 1996. Landscape dynamics in the Everglades: vegetation pattern and disturbance in Water Conservation Area 1. Ph.D. Dissertation, University of Florida Sjors, H. 1948. Mire vegetation in Bergslagen Sweden. Acta Phytogeographica Suecica 21:1–299 Sklar, F.H. and A.G. van der Valk (eds.). 2002. Tree Islands of the Everglades. Dordrecht, The Netherlands: Kluwer Sklar, F.H., C. McVoy, R. Van Zee, D.E. Gawlik, K. Tarboton, D. Rudnick, and S. Miao. 2002. The effects of altered hydrology on the ecology of the Everglades. In The Everglades, Florida Bay, and Coral Reefs of the Florida Keys: An Ecosystem Sourcebook, eds. J.W. Porter and K.G. Porter, pp. 40–73. Boca Raton, FL: CRC Slate, J.E. 1998. Inference of present and historical environmental conditions in the Everglades with diatoms and other siliceous microfossils. Ph.D. Dissertation, University of Louisville Slate, J.E. and R.J. Stevenson. 2000. Recent and abrupt environmental change in the Florida Everglades indicated from siliceous microfossils. Wetlands 20:346–356 Small, J.K. 1929. From Eden to Sahara: Florida’s Tragedy. Lancaster, PA: The Science Press Smit, H., E. dudok van Heel, and S. Wiersma. 1993. Biovolume as a tool in biomass determination of Oligochaeta and Chironomidae. Freshwater Biol. 29:37–46 Smith, S.M. and S. Newman. 2001. Growth of southern cattail (Typha domingensis Pers.) seedlings in response to fire-related soil transformations in the northern Florida Everglades. Wetlands 21(3):363–369 Smith, M.J., W.R. Kay, D.H.D. Edward, P.J. Papas, K.S.J. Richardson, M. Simpson, D.J. Cale, P.H.J. Horwitz, J.A. Davis, F.H. Yung, R.H. Norris, and S.A. Halse. 1999. AusRivAS: using macroinvertebrates to assess ecological condition of rivers in Western Australia. Freshwater Biol. 41:269–282 Smol, J.P. 1992. Paleolimnology: an important tool for effective ecosystem management. J. Aquat. Ecosyst. Health 1:49–58 Snyder, G.H. and J.M. Davidson. 1994. Everglades agriculture: past, present, and future. In Everglades: The Ecosystem and Its Restoration, eds. S.M. Davis and J.C. Ogden, pp. 85–116. Delray Beach, FL: St. Lucie Sokal, R.R. and F.J. Rohlf. 1995. Biometry, 3rd edition. New York: W.H. Freeman Sommers, L.E. and D.W. Nelson. 1972. Determination of total phosphorus in soils: a rapid perchloric acid digestion procedure. Soil Sci. Soc. Am. Proc. 36(6):902–904 Stanford, G. and S.J. Smith. 1972. Nitrogen mineralization potentials of soils. Soil Sci. Soc. Am. Proc. 36:465–472 Steel, R.G.D. and J.H. Torrie. 1980. Principles and Procedures of Statistics: A Biometrical Approach. New York: McGraw-Hill Steinman, A.D., K.E. Havens, H.J. Carrick, and R. Van Zee. 2002. The past, present, and future hydrology and ecology of Lake Okeechobee and its watersheds. In The Everglades, Florida Bay, and Coral Reefs of the Florida Keys: An Ecosystem Sourcebook, eds. J.W. Porter and K.G. Porter, pp. 19–37. Boca Raton, FL: CRC

Bibliography

671

Stephens, D.A. 1994. Bayesian retrospective multiple-changepoint identification. Appl. Stat. 43:159–178 Stephens, J.C. and L. Johnson. 1951. Subsidence of organic soils in the upper Everglades region of Florida. Soil Crop Sci. Soc. Fla. Proc. 11:191–237 Sterner, R.W. and J.J. Elser. 2002. Ecological Stoichiometry: The Biology of Elements from Molecules to the Biosphere. Princeton, NJ: Princeton University Press Stevens, W.K. 1999. Putting Things Right in the Everglades. The New York Times, April 13, p. D1 Stevenson, R.J. and L.R. Lowe. 1986. Sampling and interpretation of algal patterns for water quality assessment. In Rational for Sampling and Interpretation of Ecological Data in the Assessment of Freshwater Ecosystems, ed. B.G. Isom, pp. 118–149. Philadelphia, PA: American Society for Testing and Materials Stevenson, R.J. and C.J. Richardson. 1994. Dosing study: response of Everglades periphyton communities to increased concentrations of PO4 in experimental mesocosms. In Annual Report: Effects of Nutrient Loadings and Hydroperiod Alterations on Control of Cattail Expansions, Community Structure and Nutrient Retention in the Water Conservation Areas of South Florida, eds. C.J. Richardson, C.B. Craft, R.G. Qualls, R.J. Stevenson, and P. Vaithiyanathan, pp. 121–190. Durham, NC: School of the Environment, Duke University Stevenson, R.J. and C.J. Richardson. 1995. Response of Everglades algal communities to increased concentrations of PO4 in experimental mesocosms. In Annual Report: Effects of Phosphorus and Hydroperiod Alterations on Ecosystem Structure and Function in the Everglades, eds. C.J. Richardson, C.B. Craft, R.G. Qualls, R.J. Stevenson, P. Vaithiyanathan, M. Bush, and J. Zahina, pp. 107–184. Durham, NC: Nicholas School of the Environment, Duke University Stevenson, A.C., S. Juggins, H.J.B. Birks, D.S. Anderson, N.J. Anderson, R.W. Battarbee, F. Berge, R.B. Davis, R.J. Flower, E.Y. Haworth, V.J. Jones, J.C. Kingston, A.M. Kreiser, J.M. Line, M.A.R. Munro, and I. Renberg. 1991. The Surface Waters Acidification Project Palaeolimnology Programme: Modern Diatom/Lake-Water Chemistry Data-Set. London: ENSIS Publishing Steward, K.K. and W.H. Ornes. 1975a. Assessing a marsh environment for wastewater renovation. J. Water Pollut. Contr. Fed. 47:1880–1891 Steward, K.K. and W.H. Ornes. 1975b. The autecology of sawgrass in the Florida Everglades. Ecology 56:162–171 Steward, K.K. and W.H. Ornes. 1983. Mineral nutrition of sawgrass (Cladium jamaicense Crantz) in relation to nutrient supply. Aquat. Bot. 16:349–359 Stewart, H., S.L. Miao, M. Colbert, and C.E. Carraher, Jr. 1997. Seed germination of two cattail (Typha) species as a function of Everglades nutrient levels. Wetlands 17:116–122 Stober, Q.J., R.D. Jones, and D.J. Scheidt. 1995. Ultra-trace level mercury in the Everglades ecosystem: a multimedia canal pilot study. Water Air Soil Pollut. 80:991–1001 Stockmarr, J. 1971. Tablets with spores used in absolute pollen analysis. Pollen Spores 8:615–621 Stoermer, E.F., N.A. Andresen, and C.L. Schelske. 1992. Diatom succession in the recent sediments of Lake Okeechobee, Florida, USA. Diatom Res. 7:367–386 Stone, P.A. 2000. Prehistoric hydrologic shifts in the Everglades and implications to restoration. Poster presentation at the Greater Everglades Ecosystem Restoration (G.E.E.R.) Science Conference (December 11–15), Naples, FL Strober, M.C., R.D. Jones, and D.J. Scheidt. 1995. Ultra-trace level mercury in the Everglades ecosystem: a multimedia canal pilot study. Water Air Soil Pollut. 80:991–1001 Suberkropp, K. and E. Chauvet. 1995. Regulation of leaf breakdown by fungi in streams: influences of water chemistry. Ecology 76:1433–1445 Suter, G.W. 1993. Ecological Risk Assessment. Chelsea, MI: Lewis Suter, G.W. 1996. Abuse of hypothesis testing statistics in ecological risk assessment. Hum. Ecol. Risk Assess. 2:331–347 Sutter, L.A. 1992. Effects of phosphorus concentrations and water depth on sawgrass (Cladium jamaicense Crantz) growth and nutrient uptake. M.Sc. Thesis, Duke University

672

Bibliography

Swift, D.R. 1981. Preliminary Investigations of Periphyton and Water Quality in the Everglades Water Conservation Areas. Technical Publication 81–05. West Palm Beach, FL: South Florida Water Management District Swift, D.R. 1984. Periphyton and water quality relationships in the Everglades Water Conservation Areas. In Environments of South Florida Present and Past II, ed. P.J. Gleason, pp. 97–117. Coral Gables, FL: Miami Geological Society Swift, D.R. 1989. Water quality section. In Vegetation Water Quality. Estuarine Salinity and Productivity in C-111 Basin. West Palm Beach, FL: South Florida Water Management District Swift, D.R. and R.B. Nicholas. 1987. Periphyton and Water Quality Relationships in the Everglades Water Conservation Areas: 1978–1982. Technical Publication 87–02. West Palm Beach, FL: South Florida Water Management District. Swift, M.J., O.W. Heal, and J.M. Anderson. 1979. Decomposition in Terrestrial Ecosystems. Studies in Ecology, Vol. 5. Berkeley, CA: University of California Press Tate, R.L., III. 1980. Microbial oxidation of histosols. Adv. Microb. Ecol. 4:169–210 Technicon Industrial Systems. 1988. Phosphorus, Total. TRAACS 800 Method, Bran & Luebbe Industrial Method No. 787-86T Ter Braak, C.J.F. 1990. CANOCO – Version 3.10. Unpublished computer program, Agricultural Mathematics Group, 6700 AC Wageningen Ter Braak, C.J.F. 1995. Ordination. In Data Analysis in Community and Landscape Ecology, eds. R.H.J. Jongman, C.J.F. ter Braak, and O.F.R. Van Tongeren, pp. 91–173. Cambridge, UK: Cambridge University Press Ter Braak, C.J.F. and P. Šmilauer. 2002. CANOCO Reference Manual and CanDraw for Windows Users Guide: Software for Canonical Ordination (Version 4.5). Ithaca, NY: Microcomputer Power Terczak, E.F. 1980. Aquatic Macrofauna of the Water Conservation Areas, September 1979–1980. West Palm Beach, FL: South Florida Water Management District TerraServer USA, http://terraserver.microsoft.com. Last referenced 25 May 2006 Terry, R.E. and R.L. Tate, III. 1980. Nitrogen mineralization in Florida histosols. Soil Sci. Soc. Am. J. 44:747–750 Thayer, D., A. Ferriter, M. Bodie, K. Langeland, K. Serbesoff, D. Jones, and B. Doren. 2000. Exotic plants in the Everglades. P9. 14.1–14.48. In 2000 Everglades Consolidated Report. West Palm Beach, FL: South Florida Water Management District Thompson, F.G. 1984. The Freshwater Snails of Florida: A Manual for Identification. Gainesville, FL: University of Florida Press Thurman, E. 1985. Organic Geochemistry of Natural Waters. Dordrecht, The Netherlands: Nijhoff/Junk Tilman, D. 1982. Resource Competition and Community Structure. Princeton, NJ: Princeton University Press Tilman, D. 1990. Competition and Plant Communities. Princeton, NJ: Princeton University Press Tilman, D., J. Knops, D. Wedin, P. Reich, M. Ritchie, and E. Siemann. 1997. The influence of functional diversity and composition on ecosystem processes. Science 277:1300–1302 Tobe, J.D., K.C. Burks, R.W. Cantrell, M.A. Garland, M.E. Sweeley, D.W. Hall, et al. 1998. Florida Wetland Plants. An Identification Manual. Tallahassee, FL: Florida Department of Environmental Protection Toth, L.A. 1987. Effects of Hydrologic Regimes on Lifetime Production and Nutrient Dynamics of Sawgrass. Technical Publication 87-6. West Palm Beach, FL: South Florida Water Management District Toth, L.A. 1988. Effects of Hydrologic Regimes on Lifetime Production and Nutrient Dynamics of Sawgrass. Technical Publication 88-6. West Palm Beach, FL: South Florida Water Management District Triska, F.J. and J.R. Sedell. 1976. Decomposition of our species of leaf litter in response to nitrate manipulation. Ecology 57:783–792

Bibliography

673

Tsukuda, S., M. Sugiyama. Y. Harita, and K. Nishimura. 2006. Atmospheric phosphorus deposition in Ashiu, central Japan – sources apportionment for the estimations of true input to a terrestrial ecosystem. Biogeochemistry 77:117–138 Tufte, E.R. The Visual Display of Quantitative Information. Cheshire, CT: Graphics Press Turner, A.M., J.C. Trexler, C.F. Jordan, S.J. Slack, P. Geddes, J.H. Chick, and W.F. Loftus. 1999a. Targeting ecosystem features for conservation: Standing crops in the Florida Everglades. Conserv. Biol. 13:898–911 Turner, S.J., J.E. Hewitt, M.R. Wilkinson, D.J. Morrisey, S.F. Thrush, V.J. Cummings, and G. Funnell. 1999b. Seagrass patches and landscapes: the influence of wind-wave dynamics and hierarchical arrangements of spatial structure on macrofaunal seagrass communities. Estuaries 22:1016–1032 Twilley, R.R., W.M. Kemp, K.W. Staver, J.C. Stevenson, and W.R. Boynton. 1985. Nutrient enrichment of estuarine submersed vascular plant communities. 1. Algal growth and effects on production of plants and associated communities. Mar. Ecol. Prog. Ser. 23:179–191 Ulanowicz, R.E. 1995. Utricularia’s secret: the advantage of positive feedback in oligotrophic environments. Ecol. Model. 79:49–57 Urban, D.L. 2000. Using model analysis to design monitoring programs for landscape management and impact assessment. Ecol. Appl. 10:1820–1832 Urban, N.H. and J.W. Koebel. 1992. Colonization of Macroinvertebrates in Sawgrass and Cattail Litter during Two Years of Decomposition in the Florida Everglades. Unpublished report. West Palm Beach, FL: South Florida Water Management District Urban, D.L., R.V. O’Neill, and H.H. Shugart, Jr. 1987. Landscape ecology: a hierarchical perspective can help scientists understand spatial patterns. BioScience 37:119–127 Urban, N.H., S.M. Davis, and N.G. Aumen. 1993. Fluctuations in sawgrass and cattail density in Everglades Water Conservation Area 2A under varying nutrient, hydrologic and fire regimes. Aquat. Bot. 46:203–223 USACE (US Army Corps of Engineers). 1987. Corps of Engineers Wetlands Delineation Manual. Vicksburg, MS: US Army Engineer Waterways Experiment Station. Technical Report Y-87-1 USACE (US Army Corps of Engineers). 1994. Central and Southern Florida Project Reconnaissance Report, Comprehensive Review Study. Jacksonville, FL: US Army Corps of Engineers USACE (US Army Corps of Engineers). 1999. Central and Southern Florida Project Comprehensive Review Study: Final Integrated Feasibility Report and Programmatic Environmental Impact Statement. Jacksonville, FL: US Army Corps of Engineers and South Florida Water Management District USACE (US Army Corps of Engineers), Jacksonville District and South Florida Water Management District. 2006. Central and Southern Florida Project. Final Supplemental Environmental Impact Statement, Interim. Operational Plan (IOP) for Protection of the Cape Sable Seaside. Sparrow (December 2006), Jacksonville FL USACE (US Army Corps of Engineers), Jacksonville District Website, Everglades Division Modified Water Deliveries, http://www.saj.usace.army.mil/dp/mwdenp-c111/index.htm (accessed 21 March 2007) US EPA (United States Environmental Protection Agency). 1983. Methods for Chemical Analysis of Water and Wastes. Cincinnati, OH: US EPA US EPA (United States Environmental Protection Agency). 1989. Risk Assessment Guidance for Superfund. Vol. I. Human Health Evaluation Manual (Part A). Interim Final. Washington, DC: US Environmental Protection Agency, Office of Emergency and Remedial Response. EPA/540/I-89/002 US EPA (United States Environmental Protection Agency). 1997a. Wetlands: biological assessment methods and criteria development workshop. Proceedings, 18–20 September 1996, Boulder, Colorado (Washington, DC): US EPA, Office of Water. EPA 843–3–97–001 US EPA (United States Environmental Protection Agency). 1997b. Field and Laboratory Methods for Macroinvertebrate and Habitat Assessment of Low Gradient, Nontidal Streams. Wheeling, WV: Mid-Atlantic Coastal Streams Workgroup, Environmental Services Division, Region 3

674

Bibliography

US EPA (United States Environmental Protection Agency). 1998. National Strategy for the Development of Regional Nutrient Criteria. Washington, DC: US EPA Office of Water. Publication 822-R-98-002 Vaithiyanathan, P. and C.J. Richardson. 1994. Dosing study – diel variations in pH, dissolved oxygen, and inorganic nutrients in an Everglades slough: effects of phosphorus additions. In Annual Report: Effects of Nutrient Loadings and Hydroperiod Alterations on Control of Cattail Expansion, Community Structure and Nutrient Retention in the Water Conservation Areas of South Florida, eds. C.J. Richardson, C.B. Craft, R.G. Qualls, R.J. Stevenson, and P. Vaithiyanathan. Chapter 4. Durham, NC: School of the Environment, Duke University Vaithiyanathan, P. and C.J. Richardson. 1995. Dosing study – diel variations in pH and dissolved oxygen in an Everglades slough: effects of phosphorus additions. In Annual Report: Effects of Phosphorus and Hydroperiod Alterations on Ecosystem Structure and Function in the Everglades, eds. C.J. Richardson, C.B. Craft, R.G. Qualls, R.J. Stevenson, P. Vaithiyanathan, M. Bush, and J. Zahina, pp. 58–69. Durham, NC: Nicholas School of the Environment, Duke University Vaithiyanathan, P. and C.J. Richardson. 1997a. Nutrient profiles in the Everglades: Examination along the phosphorus gradient. Sci. Total Environ. 205:81–95 Vaithiyanathan, P. and C.J. Richardson. 1997b. Accumulation of copper and zinc along a eutrophication gradient in the Florida Everglades. In Effects of Phosphorus and Hydroperiod Alterations on Ecosystem Structure and Function in the Everglades, ed. C.J. Richardson, Chapter 11. Durham, NC: Nicholas School of the Environment, Duke University Vaithiyanathan, P. and C.J. Richardson. 1998. Biogeochemical characteristics of the Everglades sloughs. J. Environ. Qual. 27:1439–1450 Vaithiyanathan, P. and C.J. Richardson. 1999. Macrophyte species changes in the Everglades: examination along a eutrophication gradient. J. Environ. Qual. 28:1347–1358 Vaithiyanathan, P., C.J. Richardson, R.G. Kavanaugh, C.B. Craft, and T. Barkay. 1996. The relationships of eutrophication to the distribution of mercury and to the potential for methylmercury production in the peat soils of the Everglades. Environ. Sci. Technol. 30:2591–2597 Vaithiyanathan, P., T. Minto, and C.J. Richardson. 1997a. Calcium carbonate precipitation in the Everglades sloughs: influence of water column phosphorus concentration. In Effects of Phosphorus and Hydroperiod Alterations on Ecosystem Structure and Function in the Everglades, ed. C.J. Richardson, Chapter 6. Durham, NC: Nicholas School of the Environment, Duke University Vaithiyanathan, P., J. Zahina, S.R. Cooper, and C.J. Richardson. 1997b. Examination of the seed bank along a eutrophication gradient in the Northern Everglades. In Annual Report: Effects of Phosphorus and Hydroperiod Alterations on Ecosystem Structure and Function in the Everglades, ed. C.J. Richardson, Chapter 10. Durham, NC: Nicholas School of the Environment, Duke University van der Valk, A.G. and T.R. Rosburg. 1997. Seed bank composition along a phosphorus gradient in the northern Florida Everglades. Wetlands 17:228–236 Van Meter, N. 1965. Some quantitative and qualitative aspects of periphyton in the Everglades. M.Sc. Thesis, University of Miami Venables, W.N. and B.D. Ripley. 1994. Modern Applied Statistics with S-Plus. Berlin Heidelberg New York: Springer Verduin, J. 1960. Phytoplankton communities of western Lake Erie and the CO2 and O2 changes associated with them. Limnol. Oceanogr. 5:57–83 Verhoeven, J.T.A. 1986. Nutrient dynamics in minerotrophic peat mires. Aquat. Bot. 25:117–137 Verhoeven, J.T.A., W. Koerselman, and A.F.M. Meuleman. 1996. Nitrogen- or phosphorus-limited growth in herbaceous, wet vegetation: relations with atmospheric inputs and management regimes. Trends Ecol. Evol. 11:494–497 Vignoles, C.B. 1823. Observations upon the Floridas. New York: E. Bliss and E. White Vinson, M.R. and C.P. Hawkins. 1996. Effects of sampling area and subsampling procedure on comparisons of taxa richness among streams. J. N. Am. Benthol. Soc. 15:392–399

Bibliography

675

Vitousek, P.M., L.R. Walker, L.D. Whiteaker, D. Mueller-Dombois, and P.A. Matson, P.A. 1987. Biological invasion by Myrica faya alters ecosystem development in Hawaii. Science 238:802–804 Vogt, K.A., C.C. Grier, and D.J. Vogt. 1968. Production, turnover, and nutrient dynamics of above- and belowground detritus of world forests. Adv. Ecol. Res. 15:303–377 Voinov, A., C. Fitz, and R. Costanza. 1997. Landscape model provides Everglades management tool. GIS World 10:48–50 Volk, B.G. 1973. Everglades histosol subsidence. 1. CO2 evolution as affected by soil type, temperature, and moisture. Proc. Soil Crop Sci. Soc. Fla. 32:132–135 Von Post, L. and E. Granlund. 1926. Sodra Sveriges torvtillgangar I. Sver. Geol. Unders. C 335:1–127 Vymazal, J. 1995. Algae and Element Cycling in Wetlands. Boca Raton, FL: CRC/Lewis Vymazal, J. and C.J. Richardson. 1992. Dosing study – determination of phosphorus dosing threshold concentrations altering Everglade slough communities: baseline sampling of macrophytes and periphyton. In Annual Report: Effects of Nutrient Loadings and Hydroperiod Alterations on Control of Cattail Expansion, Community Structure and Nutrient Retention in the Water Conservation Areas of South Florida, eds. C.J. Richardson, C.B. Craft, R.R. Johnson, R.G. Qualls, R.B. Rader, L. Sutter, and J. Vymazal, Chapter 4. Durham, NC: Duke University Vymazal, J. and C.J. Richardson. 1995. Species composition, biomass, and nutrient content of periphyton in the Florida Everglades. J. Phycol. 31:343–354 Vymazal, J., C.B. Craft, and C.J. Richardson. 1994. Periphyton response to nitrogen and phosphorus additions in Florida Everglades. Algol. Stud. 73:75–97 Vymazal, J., J. Komárková, K. Kubecková, J. Kaštovský, and P. Znachor. 2000. Species Composition of Periphyton Growing on Natural and Artificial Substrata along the C-Transect of the Everglades WCA-2A. Part 1: Comparison of Species Abundance. Prague, Czech Republic/Ceské Budejovice, Czech Republic: Ecology and Use of Wetlands/Institute of Hydrobotany Vymazal, J., J. Komárková, K. Kubecková, J. Kaštovský, and M. Bastl. 2001a. Species Composition of Periphyton Growing on Artificial Substrata along the C-Transect of the Everglades WCA-2A. Part 2: Periphyton Biovolume–Biomass. Prague, Czech Republic/Ceské Budejovice, Czech Republic: Ecology and Use of Wetlands/Institute of Hydrobotany Vymazal, J., J. Kaštovský, V. Baxová, and M. Bastl. 2001b. Species Composition of Periphyton Growing on Natural and Artificial Substrata along the C-Transect of the Everglades WCA-2A. Part 3: Comparison of Periphyton Growing on Natural and Artificial Substrata. Prague, Czech Republic/Ceské Budejovice, Czech Republic: Ecology and Use of Wetlands/Institute of Hydrobotany Wade, D.D., J.J. Ewel, and R.H. Hofstetter. 1980. Fire in South Florida Ecosystems. Asheville, NC: Southeastern Forest Experiment Station. Forest Service General Technical Report SE-17 Wagner, G.M. and K.E. Mshigeni. 1986. The Utricularia–Cyanophyte association and its nitrogen-fixing capacity. Hydrobiologia 141:255–261 Waide, R.B., M.R. Willig, C.F. Steiner, G. Mittelbach, L. Gough, S.I. Dodson, G.P. Juday, and R. Parmenter. 1999. The relationship between productivity and species richness. Annu. Rev. Ecol. Syst. 30:257–300 Walbridge, M.R. 1991. Phosphorus availability in acid organic soils of the lower North Carolina coastal plain. Ecology 72:2083–2100 Walker, W.W., Jr. 1995. Design basis for Everglades stormwater treatment areas, Water Resour. Bull. 31:671–685 Walker, W.W., Jr. 2000. Estimation of a Phosphorus TMDL for Lake Okeechobee. West Palm Beach, FL. Final Report for FDEP and US Department of the Interior Walker, W.W., Jr. and R.H. Kadlec. 2006. Dynamic Model for Stormwater Treatment Areas. Model Version 2. West Palm Beach, FL. Report for US Department of the Interior and US Army Corps of Engineers Walker, D.R., M.D. Flora, R.D. Rice, and D.J. Scheidt. 1989. Response of the Everglades Marsh to Increased Nitrogen and Phosphorus Loading: Part II: Macrophyte Community Structure

676

Bibliography

and Chemical Composition. Homestead, FL: National Park Service, South Florida Research Center, Everglades National Park Walters, C., L. Gunderson, L., and C.S. Hollings. 1992. Experimental policies for water management in the Everglades. Ecol. Appl. 2:189–202 Ware, F.J., H. Royals, and T. Lange. 1990. Mercury contamination in Florida largemouth bass. Proc. Annu. Conf. Southeastern Assoc. Fish Wildlife Agencies 44:5–12 Washington, H.G. 1984. Diversity, biotic and similarity indices: a review with special relevance to aquatic ecosystems. Water Res. 18:653–694 Watt, A.S. 1947. Pattern and process in the plant community. J. Ecol. 35:1–22 Watts, W.A. 1975. A late quaternary record of vegetation from Lake Annie, south-central Florida. Geology 3:344–346 Weaver, K., G. Payne, and S. Xue. 2007. Status of Water Quality in the Everglades Protection Area. In The 2007 South Florida Environmental Report, Vol. I: The South Florida Environment. Chapter 3A. https://my.sfwmd.gov/pls/portal/docs/page/pg_grp_sfwmd_sfer/portlet_sfer/ tab2236041/2007_volume1/chapters/v1_ch_3a.pdf (20 June 2007) Webb, T. 1987. Is vegetation in equilibrium with climate? How to interpret late Quarternary pollen data. Vegetatio 67:75–91 Weber, C.I. 1973. Biological Field and Laboratory Methods for Measuring the Quality of Surface Waters and Effluents. EPA-670/4-73-001. Cincinnati, OH: US Environmental Protection Agency Weber, J.H. 1993. Review of possible paths for abiotic methylation of mercury (II) in the aquatic environment. Chemosphere 11:2063–2077 Weiher, E. 1999. The combined effects of scale and productivity on species richness. J. Ecol. 87:1005–1011 Weins, J.A., N.C. Stenseth, V. Van Horne, and R.A. Ims. 1993. Ecological mechanisms and landscape ecology. Oikos 66:369–380 Weller, M.W. 1975. Studies of cattail in relation to management for marsh wildlife. Iowa State J. Res. 49:383–412 Werner, H.W. 1975. The Effects of Fire on Sawgrass in Shark Slough. Homestead, FL: USDI National Park Service Westman, W.E. 1985. Ecology, Impact Assessment, and Environmental Planning. New York, NY: Wiley Wetzel, R.G. 2001. Limnology. Lakes and River Ecosystems. New York: Academic Wetzel, R.G. and Likens, G.E. 1990. Limnological Analyses. Berlin Heidelberg New York: Springer Wetzel, R.G., P.G. Hatcher, and T.S. Bianchi. 1995. Natural photolysis by ultraviolet irradiance of recalcitrant dissolved organic matter to simple substrates for microbial metabolism. Limnol. Oceanogr. 40:1369–1380 Wetzel, P.R., A.G. van der Valk, S. Newman, D.E. Gawlik, T. Troxler Gann, C.A. CoronadoMolina, D.L. Childers and F.H. Sklar. 2005. Maintaining tree islands in the Florida Everglades: nutrient redistribution is the key. Front. Ecol. Environ. 3:370–376 Whiting, G.J., J. Chanton, D. Bartlett, and J. Happell. 1991. Methane flux, net primary productivity and biomass relationships in a sub-tropical grassland community. J. Geophys. Res. 96:13067–13071 Wieder, R.K., M. Novak, W.R. Schell, and T. Rhodes. 1994. Rates of peat accumulation over the past 200 years in five Sphagnum dominated peatlands. J. Paleolimnol. 12:35–47 Wiggins, G.B., R.J. Mackay, and I.M. Smith. 1980. Evolutionary and ecological strategies of animals in annual temporary pools. Arch. Hydrobiol./Suppl. 58:96–206 Wilcox, D.A., S.I. Apfelbaum, and R.D. Hiebert. 1985. Cattail invasion of sedge meadows following hydrologic disturbance in the Cowles Bog wetland complex, Indiana. Wetlands 4:115:128 Willard, D.A. 1997. Pollen Census Data from Southern Florida: Sites along a Nutrient Gradient in Water Conservation Area 2A. Open-File Report 97–497. Reston, VA: US Geological Survey Willard, D.A. and L.M. Weimer. 1997. Palynological Census Data from Surface Samples in South Florida. Open-File Report 97-867. Reston, VA: US Geological Survey

Bibliography

677

Willard, D.A., L. Brewster-Wingard, C. Fellman, and S.E. Ishman. 1997. Paleontological Data from Mud Creek, Southern Florida. Open-File Report 97–736. Reston, VA: US Geological Survey Willard, D.A., L.M. Weimer, and W.L. Riegel. 2001. Pollen assemblages as paleoenvironmental proxies in the Florida Everglades. Rev. Palaeobot. Palynol. 113(4):213–235 Willard, D.A., C.W. Holmes, M.S. Korvela, D. Mason, J.B. Murray, W.H. Orem, and T. Towles. 2002. Paleoecological insights on fixed tree island development in the Florida Everglades: I. Environmental controls. In 2002 Tree Islands of the Everglades, eds. Sklar, F.H. and A.G. van der Valk, Dordrecht, The Netherlands: Kluwer Williams, D.D. and B.W. Feltmate. 1992. Aquatic Insects. Wallingford, Oxon, UK: C.A.B. International Williams, S.J., K. Dodd, and K.K. Gohn. 1990. Coasts in crisis. Circular No. 1075. (n.p.): US Geological Survey Williamson, C.E. 1991. Copepoda. In Ecology and Classification of North American Freshwater Invertebrates, eds. J.H. Thorp and A.P. Covich, pp. 787–822. San Diego, CA: Academic Willoughby, H.L. 1898. Across the Everglades: A Canoe Journey of Exploration. Philadelphia, PA: J.B. Lippincott Wilson, S.U. 1974. Metabolism and biology of a blue-green algal mat. M.Sc. Thesis, University of Miami Wilson, C.A., R.J. Mitchell, J.J. Hendricks, and L.R. Boring. 1999. Patterns and controls of ecosystem function in longleaf pine–wiregrass savannas. II. Nitrogen dynamics. Can. J. Forest Res. 29:752–760 Winfrey, M.R. and J.W.M. Rudd. 1990. Environmental factors affecting the formation of methylmercury in low pH lakes. Environ. Toxicol. Chem. 9:853–869 Winter, T.C. and M.K. Woo. 1990. Hydrology of lakes and wetlands. In Surface Water Hydrology, Vol. 0–1, Geology of North America, eds. M.G. Wolman and H.C. Riggs, pp. 159–187 and Plate 2. Boulder, CO: Geological Society of America Wissinger, S.A. 1999. Ecology of wetland invertebrates: synthesis and applications for conservation and management. In Invertebrates of Freshwater Wetlands of North America: Ecology and Management, eds. D.P. Batzer, S.A. Wissinger, and R.B. Rader. pp. 1043–1086. New York: Wiley Wood, E.J.F. and N.G. Maynard. 1974. Ecology of the micro-algae of the Florida Everglades. In Environments of South Florida, Present and Past, ed. P.J. Gleason, pp. 123–145. Miami, FL: Miami Geological Society Wood, J.M. and G.W. Tanner. 1990. Graminoid community composition and structure within four Everglades management areas. Wetlands 10(2):127–150 Worth, D. 1983. Preliminary Environmental Responses to Marsh Dewatering and Reduction in Water Regulation Schedule in Water Conservation Area-2A. Technical Publication 83–6. West Palm Beach, FL: South Florida Water Management District Worth, D. 1988. Environmental Response of WCA-2A to Reduction in Regulation Schedule and Marsh Drawdown. District Technical Publication 88–2. West Palm Beach, FL: South Florida Water Management Wright, H.E., D.A. Livingstone, and E.J. Cushing. 1965. Coring devices for lake sediments. In Handbook of Paleontological Techniques, eds. B. Kummel and D. Raup, pp. 494–520. San Francisco, CA: Freeman Wright, H.E., D.H. Mann, and P.H. Glaser. 1984. Piston corers for peat and lake sediments. Ecology 65:657–659 Wrubleski, D.A. and N.E. Detenbeck. Stable carbon, nitrogen, and sulfur isotope determinations of trophic relationships in prairie pothole wetlands. Unpublished manuscript Wu, Y., F.H. Sklar, and K. Rutchey. 1997. Analysis and simulations of fragmentation patterns in the Everglades. Ecol. Appl. 7:268–276 Wulff, R.D. 1995. Environmental maternal effects of seed quantity and germination. In Seed Development and Germination, eds. J. Kigel and G. Galili, pp. 491–509. New York: Dekker

678

Bibliography

Zahina, J.G. and C.J. Richardson. 1997. A survey of vascular plants in Water Conservation Area 2 of the northern Everglades. In Effects of Phosphorus and Hydroperiod Alterations on Ecosystem Structure and Function in the Everglades, ed. C.J. Richardson, Chapter 7. Durham, NC: Nicholas School of the Environment, Duke University Zamora-Muñoz, C. and J. Alba-Tercedor. 1996. Bioassessment of organically polluted Spanish rivers, using a biotic index and multivariate methods. J. N. Am. Benthol. Soc. 15:332–352 Zedler, J.B. 2000. Progress in wetland restoration ecology. Trends in Ecology and Evolution 15:402–407 Zedler, J.B. 2005. Ecological restoration: guidance from theory. San Francisco Estuary and Watershed Science 3(2): Article 4. http://repositories.cdlib.org/jmie/sfews/vol3/iss2/art4

Index

A 32 P experiment, 398 Abiotic DOC mineralization, 366 Aboriginal tribes, 631 Accretion carbon, 344 nitrogen, 344 phosphorus, 344 Accumulation rates BSi, 344 Ca, 333, 344 Cd, 141 Cu, 140 Hg, 144, 145 Na, 333 P, 50, 144, 150, 330 Pb, 140, 144 Achnanthes minutissima var. scoticai, 329, 337 as indicator of flow patterns, 348 Adams-Onis Treaty, 322 Adaptive management, 2, 10, 44, 47, 65, 633, 635 Adaptive management plan, 47, 633 Aerobic incubations, 446 Agricultural runoff, 33, 592 nutrients, 51 Agriculture, 3, 9, 16, 28, 80, 161 algal response, 348 land area, 195 water storage reservoirs, 45 Airboat disturbance, 543 Aleutian Islands, 601 Algae fresh biomass, 266 lifeforms, 266 natural and artificial substrates, 469 number of species, 463 photosynthetic activity, 500

sampling methods, 262 taxonomic groups, 266 Algal assemblages, 461 Algal indicators Aphanothece variabilis, 466 Chroococcus minutus, 466 Lyngba, 466 Mougeotia, 466 Oedogonium, 466 Scytonema, 466 Algal mats, 81 cover growth, 540 interstitial water, 436 Algal responses, 261 Algivores, 303 Alismataceae, 26 Alligator mississippiensis, 543, 601 Alligator weed. See Alternanthera philoxeroides Allochthonous, 278 Alternanthera philoxeroides, 90, 507 Aluminum-bound organic phosphorus, 513 Ambrosia artemisiifolia, 26, 91 American alligator. See Alligator mississippiensis Ammodramus maritimus mirabilis, 45 Ammonia, 106, 423, 506 volatilization, 506 Ammonium, 366, 506 Ammonium chloride, 510 Ammonium-nitrogen concentrations, 400 Amphipoda, 485 Amphora coffeaeformis, 463, 465 Amphora lineolata, 329, 334, 337 Amphora veneta, 329 Anaerobic mineralization, 445 Anaerobic respiration, 631 Anastasia Formation, 15, 170 679

680 Annona glabra, 192 communities, 624 restoration, 628 Anoxic benthic environment, 435 Anthropogenic disturbance, 4, 15, 23, 44, 70, 141, 172, 225, 343, 348, 361, 551, 587 paleoecological evidence, 24, 348 Anthropogenic impacts, 622 Anthropogenic influence, 588 APA activity, 408, 410, 417, 572, 576 Aphanothece, 465 Aquifer storage ASR wells, 44 Arrowhead. See Sagittaria Artificial substrates algae, 478 Assemblage composition determinants macroinvertebrates, 314 Assimilative capacity phosphorus, 569 Atlantic Coastal Ridge, 168 Attributes assemblage-level, 483 Australian melaleuca tree. See Melaleuca Australian pine. See Casuarina Autochthonous, 278

B Bacillariophyceae, 270, 463 Bacopa caroliniana, 507 Balance of nature, 611 Bayesian changepoint detection method, 568, 606, 608 Bayesian Hierarchical Changepoint Model, 605, 610 Bayesian kriging, 583 Bayhead/swamp forests, 80 Beach-elder. See Iva imbricate Beardius truncatus, 311, 314 Bedrock CaCO3, 124, 160, 347 configuration, 16 depressions, 171 hydraulic conductivity, 176 limestone, 170 map, 16 permeability, 15 Below detection limit (BDL) water quality data, 108 Belowground biomass macrophytes, 157

Index Best Management Practices, 617 P reduction program, 54 Big Cypress National Preserve, 627 Bioassessment, 278, 638 multimetric approach, 316 multivariate approach, 316 Biogenic silica, 322 Biological attributes, 599 algae-diatom relative abundance, 603 blue-green algae biovolume, 603 Bray-Curtis dissimilarity, 603 calcareous mat cover, 603 diatom biovolume, 603 diatom density, 603 Gastropoda percent, 603 macroinvertebrates abundance, 603 biomass, 603 taxa, 603 Oligochaeta number, 603 predators percent, 603 sensitive species percent, 603 trophic level, 603 Utricularia purpurea stem densities, 603 Biological metrics, 596 ecological change, 599 Biological monitoring, 596 Biomass, 437 response cumulative P additions, 522 Biomonitoring, 596 Biovolume evaluation algae, 265 Bladderwort. See Utricularia Blue-green algae, 461 Bluefin killifish. See Lucania goodei BMP. See Best Management Practices Board of Drainage Commisioners, 323 Bog, 31 Bootstrap simulation, 225, 252, 286, 328, 607, 614 Bottom-up control, 316 Br/SRP ratios, 394 Brachysira vitrea, 329, 334, 337 Braun-Blanquet scale, 462, 534 Braun-Blanquet cover classes, 282–283, 536 Bray-Curtis dissimilarity, 286, 288, 603

Index Brazilian pepper. See Schinus terebinthifolius Bromide, 107 sodium bromide, 387, 394 Bulbochaete, 466 Bulk density, 135, 587 as function of peat depth, 64 Bulrush. See Scirpus Buttonbush. See Cephalanthus occidentalis

C C:N ratio, 443, 501 C:N:P ratios, 397 C:P ratios, 397, 443 C3 photosynthetic pathway Cattail and sawgrass, 549 Cadmium, 140, 141 Caecidotea, 303 Calcite Mud, 20 Calcium carbonate, 20, 22, 161, 169, 346, 388, 417, 515, 638 Calcium-bound inorganic phosphorus, 514 Calcium-to-magnesium ratios, 399 Caloosahatchee Canal, 39 Canal L-38, 580 Canal systems, 197 Canal-and-levee effects, 256 Canals P gradient, 315 water Quality, 28 Canonical Correspondence Analysis periphyton composition, 463 Cape Sable Seaside Sparrow. See Ammodramus maritimus mirabilis Carbon budgets, 352 cycling, 351 litter loss, 446 microbial biomass, 359 Carbon dating. See Radiometric dating Carbon dioxide, 82, 521 gaseous, 358 Carbon mineralization, 445, 451 Carbon-14, 66 Carbonate alkalinity, 398 Carboxylic acids, 368 Carnivores, 601 CART models. See Classification and regression tree models Casuarina litorea, 28, 329 Cattail control mechanical disturbance, 631 Cattail invasion potential, 625

681 Cattail removal, 630 Cattail. See Typha domingensis Cellulose, 442 Central and South Florida Project for Flood Control and Other Purposes, 40, 195, 196 Central and Southern Florida Flood Control District, 323 Centroids, 223 Cephalanthus occidentalis, 26 CERP. See Comprehensive Everglades Restoration Plan Cesium-137, 66 activity, 326 from thermonuclear weapons tests, 580 Changepoints, 569, 599, 612 cumulative probability distributions, 614 Channel flow, 389 Channels designation P dosing, 376 Chara, 130, 230, 236, 385, 411, 535 P treatments fertilizer study, 420, 481, 505, 521, 542 optimum growth cover PO4-P concentration, 432 percent coverage, 535 populations, 430 Chemical control of Typha, 630 Chenopodiaceae/Amaranthaceae, 329 Chenopodium, 26, 238 Chilled seeds, 552, 558 Chironomidae, 285, 301, 485 Chloride molar ratios, 400 pore water, 175 Chloride gradient, 122, 124, 236 Chlorophyceae biomass, 269 Chlorophyta, 463, 466 growth, 540 Chroococcales, 266 Chroococcus, 465 Chroococcus deltoids, 466 Chrysobalanus icaco, 192 CI. See Credible interval Cladium jamaicense, 78, 154, 329, 535 aboveground biomass, 154, 526 biomass, 518 carbon fixation rate, 550 establishment, 547 fire management, 631 frequency, 231 relationship with soil TP, 238

682 Cladium jamaicense (cont.) germination curves, 556 germination experiments, 551 germination rate, 554 lifespan, 549 litter C:N ratio, 455 C:P ratio, 455 Cu concentrations, 455 decay, 446 N:P ratio, 455 P accumulation, 450 P concentrations, 455 litter bags, 444 N:P ratios, 229, 455, 518, 572 nitrogen concentrations, 517 P effects, 217 photosynthesis, 511, 455 physiology, 538 reestablishment, 538 regeneration, 630 Relative Growth Rate, 562 seed production, 551, 553 seedling growth, 547 species description, 548 tissue nutrient concentrations, 515 Cladium mariscus, 7, 89, 549 Cladocera, 481, 485 Cladophora glomerata, 418 Cladotanytarsus, 311 Class III waterbody Everglades classification, 481 Classification and Regression Tree Models, 423, 605 Clean Water Act Section 101a, 278 Climate, 36, 171, 215 phosphorus inputs, 49, 51 precipitation, 37 Climate change, 627 Climbing fern, 632 Clone sawgrass, 550 Cluster analysis pollen assemblages, 338 Cocoplum. See Chrysobalanus icaco Coleoptera, 303 Community level restoration guidelines, 636 Community metabolism, 415 Community structure, 559 Competitive exclusion, 558

Index Composition vegetation fire, 257 Compositional patterns fine-scale, 256 Comprehensive Everglades Restoration Plan, 1, 9, 39, 621, 641 goals, 2, 57, 194, 633 website, 44 Concept of limiting factors, 597 Conjugatophyceae, 466 biomass, 269 Cosmarium, 269 Mougeotia, 269 Pleurotaenium, 269 Spirogyra, 269 Staurastrum, 269 Connell’s intermediate disturbance hypothesis, 598 Constant Rate of Supply Model, 326, 329 Constructed wetlands, 569 Control reference sites, 535, 602 Controlled burns, 629 cattail removal succession, 634 Copepoda, 481, 485 Copper, 140, 141, 450 as limiting nutrient, 457 Co-precipitation P and CaCO3, 388 Cosmarium, 465 Credible interval, 569, 609 CRS Model. See Constant Rate of Supply Model Cyanobacteria, 461 Cyanophyta nitrogen fixation, 438 Cyanoprokaryota, 266, 270, 461 Oscillatoriales, 270 Chroococcales, 270 Nostocales, 271 total biomass, 271 Cyclopoid copepods, 501 Cyclotella meneghiniana, 334 Cymatopleura solea, 466 Cymbella, 327, 334, 466 Cymbella cymbiformis, 466 Cyperaceae, 548 pollen, 341 Cypress, 80 Cypridopsis okeechobei, 501 Cytheridella alosa, 501

Index D Dahoon holly. See Ilex cassine Dark germination, 552, 560 Dark inhibition, 560 Database ACCESS, 5 Davis, J.H. The Natural Features of Southern Florida, 3 Decapoda, 301 Decomposition, 156, 332, 352, 358 effect of aerobic vs. anaerobic conditions, 442 effects of exogenous inorganic nutrients, 367, 442 litter and peat, 441, 445, 455 Deep water, 84 Delta Marsh, 496 Denitrification, 446, 506 Depth calculations soft-bottom, 222 Detrended Correspondence Analysis algal analysis, 328, 467 periphyton mat composition, 463 Diatoms, 28, 334, 463 analysis, 321 as indicators of high soil P, 329 as indicators of low soil P, 329 assemblages, 347 cluster analysis, 334 Cymatopleura solea, 466 Cymbella, 466 Cymbella cymbiformis, 466 decline in diversity, 347 Detrended Correspondence Analyses, 328 Diploneis elliptica, 466 identification, 264 indicator species, 328 indicator taxa, 327 methods, 326 native taxa, 348 Navicula lanceolata, 466 Navicula rhynchocephala, 466 Nedium, 466 Nitzschia linearis, 466 Nitzschia recta, 466 oligotrophic (low-nutrient) sites, 272 species abundance, 337 Diel oxygen, 389, 413, 600 Diel pattern, 404 NH4-N and NO3-N uptake, 400 Diel variation PO4-P, 408 Dike systems, 197, 198

683 Diploneis elliptica, 466 Diptera, 485 Dissolved Organic Carbon, 352, 361, 369, 446, 607 biodegradation, 367 humic fractions, 367 hydrophilic acid fraction, 367 methods, 107, 352 pore water, 362 Dissolved Organic Matter, 351, 355 controlling processes, 369 Dissolved Organic Nitrogen, 357, 454 budgets, 357, 362 Dissolved Organic Phosphorus dosing channels, 396 Dissolved Oxygen, 411 community effects, 129 dosing, 404 frequency, 411 sediment-water interface, 435 State of Florida criteria, 129 Disturbance, 639 germination, 563 seedling growth, 563 Disturbance experiment deep-water treatments, 537 experimental design, 533 fertilized treatments, 537 plot establishment, 532 plot location, 532 randomized complete block design, 536 site monitoring, 533 tilled treatments, 537 vegetation removed plots, 537 water chemistry, 541 water depth, 534 Diversity (Shannon’s H’), 336 DO. See Dissolved Oxygen DOC. See Dissolved Organic Carbon Dog-fennel. See Eupatorium DOM. See Dissolved Organic Matter DON. See Dissolved Organic Nitrogen Dormancy, 559 Dose response curves, 604 Dosing, 385–418 Ca, 399 Cl, 399 Mg, 399 Na, 399 operations, 387 TP loadings, 391 sampling, 388

684 Dosing channels ENP, 420, 477 oxygen, 406 SRP concentrations, 394 TP concentrations, 394 Dosing experiment P threshold, 602 Dosing flumes, 391, 406 algal assemblages sampling, 461 applied PO4 concentrations, 391 background water quality, 391 control channel TP values, 391 diel oxygen pattern, 404 DOP, 106, 362, 394 ecosystem response, 411 fish data analysis, 484 fish sampling, 482 litter decomposition, 455 macroinvertebrate data analysis, 483 macroinvertebrate sampling, 480 P/R, 413 phosphatase activity, 408 PP, 394 productivity, 413 respiration, 413 SRP, 394 vegetation, 478 Chara, 478 Eleocharis cellulosa, 478 Eleocharis elongata, 478 Nymphaea, 478 Panicum hemitomon, 478 Utricularia foliosa, 478 Utricularia purpurea, 478 water depth, 478 water level changes, 393 Dosing mesocosms. See Dosing flumes Douglas, Marjory Stoneman, 1, 13 Drainage blocked, 14 effects, 16, 23, 34, 170, 347, 442 canals, 195, 262 carbon, 359 indicators of changed hydroperiod, 23, 227 nutrients, 63, 70 Drouet classification algae, 272 Drought, 23, 37, 46, 226, 245, 391 Dry downs, 544 Dry levee, 87 Duck-potato. See Sagittaria lancifolia

Index E EAA. See Everglades Agricultureal Area Ecological imbalance, 599, 610, 616 Ecological thresholds multiple trophic levels, 617 Ecosystem function, 32, 164, 200, 599 Ecosystem level restoration guidelines, 635 Ecosystem services, 641 Eichhornia crassipes, 84, 90, 633 Eleocharis, 230 N:P mass ratio, 518 P-limitation, 518 tissue P concentrations, 517 water depth controls, 437 Eleocharis cellulosa, 418, 422, 425, 507, 522 biomass per plot, 422 density, 425 mortality, 433 Eleocharis elongata, 418, 425 density, 425 stem density, 430 survival, 432 Enallagma civile, 311 Environmental factors plant communitites, 246 Environmental Regulatory Commission, 601 Epiphytes effects of dense communities on macrophytes, 418 Epiphytic habitats, 87 Epiphytic periphyton, 422, 437 growth measurements, 422 Epiphyton vegetatively attached periphyton, 283 Equipotential surface, 179, 183 ERC. See Environmental Regulatory Commission Euclidean distances, 225 Eudocimus albus, 209 Eunotia, 28, 334 acidophilous species, 347 Eunotia arcus, 347 Eunotia camelus, 347 Eupatorium, 26 Eutrophic lakes, 432 Eutrophic sites, 272 Eutrophication, 26, 32, 69, 322, 417 gradient, 134, 572 mercury, 147 periphyton community, 261, 417, 465 Evaporation rates, 172 Evapotranspiration, 38, 171, 174, 364

Index Everglades area, 59 as P-limited ecosystem, 595 classification, 29 classification as Class III waterbody, 481 comparison with other North American peatlands, 70 fen restoration, 621–642 formation, 15 geologic settings, 167 historic formation, 15 historical background, 322 hydrology gradients, 167 maps, 6, 17, 99, 169, 198, 282 plant communities, 74 Everglades Agricultural Area, 16, 195 carbon budget, 352 DON budgets, 364 phosphorus loadings, 54 Everglades National Park, 6, 17, 67, 117, 323 14 C, 66 cattail invasion potential, 626 N:P ratios, 54 soil C, N, P, 64 spatial patterns of STP, 586 Everglades peatland complex formation, 14 Everglades Protection Area, 55 rainfall P inputs, 49 Everglades snail kite. See Rostrhamus sociabilis plumbeus Exchangeable inorganic phosphorus, 514 Experimental drawdowns, 630 Experimental sites location, 7, 279, 602 Exposure dose P, 604

F Feeding groups macroinvertebrates, 575 Fen, 1, 31, 56, 72, 398 classification, 29, 83 definition, 13 Ferns, 26 Fertilizer application N and P, 510 Fertilizer study, 505–527 aboveground biomass (Live + standing litter), 518 biomass analysis, 511 cattail response to P, 520

685 hydroperiod, 507 leaf area index, 511 phosphorus and nitrogen pools, 524 photosynthesis, 511 randomized complete block design, 507 soil analysis, 511 water analysis, 510 Fire, 4, 7, 18, 33, 35, 67, 159, 216, 229, 234, 244, 257, 627, 628, 629 Fire management, 165, 632, 634, 637 succession, 627 Fire suppression, 277, 631 Fish abundance and Eleocharis stem densities, 495, 498 biomass, 305 subsidy-stress relationships, 305 data analysis, 484 density, 314, 316 subsidy-stress relationships, 305 hypotheses P-enrichment, 478 mercury concentrations, 142 response to P additions, 477, 494 sampling, 286, 482 Flats, 78, 85 biomass, 84 Floating mats, 469 biomass, 474 removal, 435 Florida Project Comprehensive Review Study, 43 problems, 43 Flow gates S-10A, C, D, E, 173 S-144, 173, 521 Flumes, 376, 385 construction, 378 design and materials, 379 battery platform, 379 channels, 380 mixing tanks, 379 pump platform, 379 solar array, 380 location, 377 physical layout of design, 377 systems operation, 380 electrical, 382 plumbing, 380 Formation of the Everglades historic Everglades, 15 Fort Thompson Formation, 15, 66, 170 Fossils indicators, 328

686 Fragilaria, 329, 334, 337 Fragilaria ulna, 465 Fragrant waterlily. See Nymphaea odorata Freeze, 245 Frustulia rhomboides, 334 Functional redundancy, 601, 605 Future water demand, 641

G Gambusia holbrooki, 316, 482, 495 Gastropoda, 285, 485, 499 Geologic history, 168 Geologic history of Everglades peatland. See Peatland Germination, 551, 554 curves, 555 inhibitors, 559 Germination experiments Cladium jamaicense, 551 methods chilled seeds, 552 dark germination, 552 seedling growth, 552 Typha domingensis, 551 Germination studies, 639 Gleason, P.J., 16 Gloeothece interspersa, 465, 466 Gomphonema olivaceum, 466 Gomphonema parvulum, 329, 334, 337 GPP. See Gross Primary Productivity Gradient studies, 97–101 enrichment gradients methods, 104 sampling locations, 104 site establishment, 104 Gradient studies GPS locations, 98–101 objectives, 98 Gradients course-scale, 249 DOC concentrations, 363 fine-scale, 249 hydrology, 220, 595 nutrient, 116, 121, 262, 595 soil analysis, 220 Graham, Bob, 323 Great Blue Heron, 209 Great Egret, 209 Green algae. See Chlorophyta Gross Primary Productivity sloughs, 411, 413, 414

Index Groundwater flow, 39, 43, 622 WCA-2A, 175

H Hardwood swamps TP, 628 Hardwood upland hammocks, 76, 80, 218 Heavy metal, 104, 112, 140 Hemiptera, 303 Herbert Hoover Dike, 41, 622, 629 Hester-Dendy artificial substrates, 481 Heterandria Formosa, 316 Hierarchical changepoint analysis, 610 Hierarchy theory, 280–281 Hillsboro Canal, 117, 195, 343, 580 Hirudinea, 303 Historic Everglades original area, 215 surface sheet flow, 194 Histosols, 19, 442, 457 Human population, 59, 195, 197, 323, 622 Humic acids copper, 457 Hurricanes, 38, 53 Hydraulic conductivity, 176 Hydraulic gradient, 171, 183 Hydraulic loading rate, 568 Hydrocotyle umbellata, 507 Hydrogeology water sources, 630 Hydrologic budget WCA-2A, 172 Hydrologic equivalence, 32 Hydrologic pattern, 637 Hydrology future plans, 43 shifts on the landscape, 39 Hydropattern, 33, 177, 215 Hydroperiod, 7, 20, 26, 56, 67, 187, 227, 243, 627, 628 definition, 33 macroinvertebrate response, 243, 342, 507, 540 model, 627 restoration, 629–633 vegetation response, 243, 255, 342, 507, 540 Hypothesis assemblage responses cattail invasions 218 macroinvertebrates and fish, 311, 478, 496, 501 mercury methylation, 147

Index periphyton change, 237, 312 restoration, 35 vegetation patterns, 255 Utricularia, 539 Utriculatia decline, 406, 429 hydrology, 255 intermediate disturbance, 598 microbial activity PO4, 456 P assimilative capacity, 567 P threshold, 599 subsidy-stress, 279 Utricularia inhibition, 429

I IBI. See Index of Biological Integrity Ice age Wisconsin, 170 Ilex cassine, 192 Imbalance definition, 599, 610 flora and fauna, 375, 597, 609 indicator. 611 Impact zone, 578 Impacted zone indicators, 311 taxa, 311 Anopheles, 311 Caecidotea, 311 Scirtes, 311 Uranotaenia sapphirina, 311 Inceptisols, 19, 20 Index of Biological Integrity, 316, 613 Indiana Dunes National Lakeshore, 547 Indicator species, 242, 328, 607 Indicator Species Analysis, 288, 298, 311 Indicator taxa Cladotanytarsus, 315 Cypretta brevisaepta, 315 Parakiefferiella, 315 Paraponyx, 315 Tanytarsus, 315 Indices ecosystem response, 572 metrics, 277–278 trophic level response, 595 Initial germination, 556 INSPAN. See Indicator Species Analysis Interim Operation Plan, 45 Intermediate disturbance hypothesis, 598 Inundation, 197 Invasive species, 90, 232, 632

687 Invasive vegetation, 310 Mikania, 310 Sarcostemma, 310 Typha, 310 Invertebrates benthic, 575 taxa list, 288 Iraq, 641 Island Biogeographic Theory, 410 Iva imbricata, 26

J Jordanella floridae, 495

K Kernel smoothers, 581 Keystone species, 410, 427, 601 Kissimmee Basin, 41 Komvophoron, 465 Kriging, 583 Kriging model, 583

L Labile phosphorus, 134 Lake Okeechobee, 39, 41, 624 P concentration, 52 water level, 41 water supply/environmental schedule, 41 Land development, 622 Landscape P gradient, 52 Landscape evolution of Everglades peatland. See Peatland Landscape fragmentation, 636 Landscape level restoration guidelines, 635 Largemouth bass. See Micropterus salmoides Law water Resources Development Act, 45 Lead, 140, 163 Lead-210, 51, 66 Leaf area index, 511 Least killifish. See Heterandria Formosa Lemna, 257 Lepomis punctatus, 495 Leptolyngbya, 263 Liebig, Justis von concept of limiting factors, 597 Lignin, 442 Liguus, 209

688 Limits of tolerance, 279, 499, 597 Limnogenous, 31, 47, 622, 629 Litter, 156, 158, 367, 443, 524 biomass, 156, 520 copper accumulation, 457 decomposition, 358, 455, 639 microbial immobilization, 446 Livingstone piston coring device, 325 Location of research sites, 6 Lower East Coast, 41 Loxahatchee Channel, 171 Loxahatchee National Wildlife Refuge. See Water Conservation Area 1 Lucania goodei, 495 Lycopodium spike, 327 Lygodium, 632 Lyngbya, 263

M Macroinvertebrate assemblages biomass, 301, 485 density, 485 Macroinvertebrates assemblage-environment relationships, 299 bioassessment, 278 biomass estimation, 285 biomass response to P, 287, 311 composition-environment relationships, 491, 493 data analysis, 286, 483 density, 286 diversity, 287, 572, 575 dose-response relationships, 485 gradient, 283 hypotheses P-enrichment, 478 indicator species, 311 metrics, 612 predators, 489 response to P additions, 477 sampling, 283, 480 sampling design, 281 sensitive taxa, 613 species accumulation curves, 301 species diversity, 312 species list, 289–298 species richness, 287, 499 study area, 284 subsamples, 300 fixed counts, 300 fixed-area, 300 successional vectors, 500

Index taxon density, 485 taxon richness, 485, 498 taxonomic dissimilarity, 499 taxonomic resolution, 284 tolerant taxa, 613 Macronutrients litter immobilization copper, 456 Macrophyte survey oxygen study, 129 Macrophytes, 418 biomass, 518 declines, 417 direct effects on macroinvertebrates, 500–501 gradient responses, 277 N:P ratios, 397 phosphorus effects, 426 plant species listing, 88–93 responses season, 424 temperature, 424 TP level, 424 water depth, 424 species distribution, 248 in WCA-2A, 230 species increase with P enrichment, 240 Magnesium, 262 Magnolia virginiana, 192 Maidencane. See Panicum hemitomon Managed water flow, 31, 47, 623 Managenous. See Peatland Managenous water flow. See Managed water flow Mantel test, 252 Marl, 15, 78, 81, 170, 218 calcium carbonate, 20, 515 Marsh, 83, 85 definition, 14 Mass spectrometry C-14, 18 Mastogloia smithii, 329, 334, 337, 463, 465 Mat cover biomass, 437 metric, 610 pH, 399 temperature, 399 Matric water potential, 446 Maximum germination, 556 time, 557 Mean total N, 120

Index Melaleuca quinquernervia nonnative taxa, 28, 33, 216, 632 Meotrophic, 32 Mercury, 112, 142 demethylation, 147 deposition, 146 methylation, 147 Mesocosm, 385 construction, 378 experiment, 376 flow rate, 381 P treatment levels, 376 phosphorus experiments, 375 systems operation, 380 electrical, 382 plumbing, 380 Mesopotamian marshes, 641 Metal accumulation rates soil, 141 Metaphyton cover, 305 floating periphyton mats, 283 Methane emission, 360 loss, 358 production, 352 Method detection limits (MDL), 106 Metric multidimensional scaling, 224 Bray-Curtis dissimilarity, 224, 286, 298, 484, 603 partial Mantel tests, 224 Metrics, 316, 596, 599 definition, 596 density, 596 feeding ecology, 316 guild classes, 596 number of taxa, 596 richness of species, 596 taxonomic richness, 316 taxonomic structure, 316 tolerance/intolerance, 316 trophic level response, 595 Miami Canal, 45, 140, 195, 212 Miami Formation, 15 Miami limestone, 66, 170 Miccosukee Tribe, 44 Microalga Chara, 527 Microbial C biomass, 359 Microbial decomposition, 442 Microbial methylation, 147 Microbial P biomass, 360, 457, 514, 515 soil, 457

689 Microbial phosphorus, 514, 515 Microbial respiration, 443, 458 Microcosm studies P uptake periphyton, 395 Microcrustacea, 485 Micropterus salmoides, 495 Mikania scandens, 85, 88, 230, 310 biomass, 156 Mineralization potential enrichment gradient, 445 Minerotrophic system, 31, 622 Minerogenous system, 31, 622 Minimum detection limit (MDL), 108 Mire definition, 13 water, 31 Model Bayesian model, 605 changepoint, 605 Classification and Regression Tree, 423, 605 Constant Rate of Supply, 326, 329 Hierarchical model, 605 Integrated hierarchical model P threshold, 602 Interaction of hydroperiod and fire, 627 Kriging, 583 Natural System, 41, 171, 174 Phosphorus Accretion Rate, 582 piecewise linear model, 569 subsidy-stress model, 279–280, 478, 597 three dimensional control Everglades, 628 Modified water delivery plan, 45 Monitoring wells fertilizer study, 508 Monocalcium phosphate, 510 Monte Carlo simulation, 463, 590 Mosquitoes, 316 Coquillettidia perturbans, 316 larvae, 316 Mansonia titillans, 316 Uranotaenia sapphirina, 316 Mosquitofish. See Gambusia holbrooki Mougeotia diatom, 466 MSL. See Sea level Multidimensional scaling, 288 Mycteria Americana, 45, 209 Myrica cerifera, 28, 192, 329

690 N N:P ratio, 54, 397, 455, 571, 572 P limitations, 14, 518 Native seed bank, 33, 215, 324, 533, 560, 632 Natural System Model, 41, 174 Navicula confervacea, 329, 337 Navicula lanceolata, 466 Navicula rhynchocephala, 466 NAWDB. See North American Wetlands for Water Quality Treatment Database nCPA. See Nonparametric Changepoint Analysis Nedium, 466 Net Primary Productivity, 351, 412, 442 belowground, 358 WCA-2A, 357 Never Glades, 13 Niche width, 410, 598 Nitrate reductase, 400 Nitrogen, 450, 506 accumulation, 67, 69 and phosphorus fertilization experimental design, 507 nitrogen concentrations, 62, 121, 256, 262, 515, 517 study location, 507 as function of peat depth, 64 budgets, 524 elevated, 257 soils, 14, 62, 138 standing stock, 156, 158 Nitrogen fixation, 4, 438 Nitzschia, 334 Nitzschia amphibia, 329, 332 Nitzschia linearis, 466 Nitzschia palea, 329, 337 Nitzschia recta, 466 nMDS. See Nonmetric multidimensional scaling Nonmetric multidimensional scaling, 483 Nonmetric multidimensional scaling ordination, 252 Nonparametric Changepoint Analysis (nCPA), 606 Nonparametric regression, 581 North American Wetlands for Water Quality Treatment Database, 568 North New River Canal, 195 Nostoc, 465 NPP. See Net Primary Productivity Nuphar lutea, 7, 79, 93, 349 Nutrient gradient, 228 Nutrient limitation, 351, 518, 537 criteria, 576

Index Nutrient storage, 154, 158 Nutrient storage capacity, 4, 567 Nutrient-IBI, 612 Nutrients, 355 anthropogenic inputs, 53, 343 deprivation, 563 runoff, 51 Nymphaea, 7, 230 Nymphaea odorata, 7, 329, 418, 422, 507 pollen, 342 populations, 430 rhizomes, 524 shading, 437 Nymphaeaceae pollen, 339

O Odonata, 303, 485 Oecetis, 311 Oedogonium, 266, 269, 466 Old litter, 156, 524 Oligocaeta, 485 Oligotrophy, 32, 129, 402 Ombrotrophy, 31, 622 Opportunistic species, 563 Ordination, 288, 314 invertebrate species composition, 307 landscape-scale dimension, 314 local-scale dimension, 314 Organic anions ionic balance contribution, 368 Organic carbon accumulation, 67, 441 as function of peat depth, 64 Organic phosphorus, 514 Organic pollution Chironomus stigmaterus indicators, 311 Goeldichironomus holoprasinus indicators, 311 Orthophosphate (SRP), 50, 105, 262, 385 doses, 386 dosing biological driven responses, 404 interstitial water, 436 residual, 386 Oscillatoria, 263, 466 Oscillatoriales, 266 Osmunda regalis, 507 Ostracoda, 481, 485, 501 Oxyethira, 311

Index Oxygen concentrations, 129, 133, 391, 404, 435 methods, 389 plant germination, 555

P P criterion, 617 P efflux rates, 152 P enrichment shift in species dominance, 347 P fractions soil gradient, 445 P gradient edge effect, 577 P limitation biochemical indices, 572 P loading fertilizer study, 522 gradient study, 571 rainfall, 49 P threshold model, 606 P threshold load, 567 P/R ratio, 411 PAC. See Phosphorus Assimilative Capacity Pa-hay-okee grassy lake, 13 Palaemonetes paludosus, 301, 311 Paleoecology analysis, 321 anthropogenic disturbance, 24 calibration, 324 core analysis, 638 core collection, 325 study site WCA-2A, 323 timeline, 29 Pamlico Terrace, 170 Pan evaporation, 171, 364 Panicum hemitomon, 420, 478, 507, 522 N:P mass ratio, 518 P-limitation, 518 tissue P concentrations, 517 Pappus cattail, 553 PAR. See Phosphorus Accretion Rate Parakiefferiella, 311 Paratanytarsus, 311 Partial Mantel’s test, 492 P-assimilative capacity. See Phosphorus Assimilative Capacity Path diagram relationships, 492

691 Pattern and process, 197 Peat, 60–62 accretion, 65, 352, 580 increased water effects, 70 nutrient loading, 70 accretion rates, 329 carbon, 451 classification, 62 core sites, 19 core stratigraphies, 21, 22 Everglades peat, 60 formation 5000 YBP, 18 Gandy peat, 60, 192 geochemistry, 330 Loxahatchee peat, 60 mineralization, 458 nitrogen, 451 nutrient accumulation, 70 Okeechobee muck, 60 radiocarbon dating, 16, 18 soil depths, 18, 20, 21 thickness, 65 types, 60 Peat batteries, 85 Peat mineralization anaerobic conditions, 451 Peatland definition, 13 formation, 14 geologic history, 15, 168 limnogenous peatland, 31, 622 managenous, 31, 623 minerotrophic peatland, 31 ombrogenous peatland, 31 ombrotrophic peatland, 31 restoration concepts, 633 soligenous peatland, 31, 623 topogenous peatland, 31, 622 Peat-litter P and N pool, 524 Peltandra virginica, 75, 85, 89, 236 Pennywort. See Hydrocotyle umbellata P-enrichment gradient WCA-2A, 116, 228 WCA-3A, 120 Percent cover algal mat, 422 Periphyton, 81, 261 AFDM, 82, 492 ash content, 471 biomass, 82, 415 biomass analysis, 471 biomass N and P concentrations, 471 C:N ratio, 303, 492

692 Periphyton (cont.) calcareous periphyton, 82, 346, 638 composition Effect of TP, 464 diversity, 467 food quality for primary consumers, 502 mat cover, 418 N:P ratios, 399 noncalcareous, 433 removal, 417 removal experiment, 422, 434 seasonal dynamics, 468 Periphyton mats ash content, 398 N and K, 256 Persea borbonia, 80, 86, 92, 192 pH calcium carbonate, 346 depletion of free CO2, 429, 607 in water column, 398, 406, 434 pore water, 402 Phormidium, 465 Phosphatase limitation threshold, 410 production, 410 unit chlorophyll (APA-c), 408 Phosphatase activity, 363, 408, 572 Phosphate coprecipitation, 398 diel pattern, 406 uptake, 396 rates, 397 response inhibition, 439 optimum response, 439 stimulation, 439 Phosphorus, 115 accretion rates, 579, 582, 588 percentage of area, 589 accumulation in soil, 67, 69 adsorption, 458 affected area, 592 as function of peat depth, 64 budgets, 524 content in live leaf and root tissue, 43 criterion, 53 dilution, 389 doses, 478 dosing facilities, 376 dosing mesocosms, 638 dosing operations, 387 fractionation methods, 109 hot spots, 36, 625 imbalance, 599

Index impacted zones, 281 litter immobilization, 451, 456 loads, 478 methods, 113 microbial immobilization, 458 plant succession, 531 rainfall, 48 reference zones, 281 sodium phosphate, 387 soil efflux, 113, 150 spatial distribution, 579, 588, 591 speciation, 105, 106, 386 standing stock, 156, 158 threshold, 595–617 determination by technology-based criteria, 595 transition zones, 281 uptake channels, 389 Phosphorus Accretion Rate, 579 background level, 592 modeling nonparametric regression model, 582 piecewise linear model, 582 Phosphorus Assimilative Capacity, 567–578, 595, 637 Phosphorus concentration geometric mean, 108, 595 median, 595 Phosphorus dosing. See also Dosing flumes ecosystem responses, 385 experimental setup, 376 macrophyte community response, 417 orthophosphorus, 385 particulate phosphorus, 396, 398 plant analysis methods, 420 soil chemistry, 385 Soluble Reactive Phosphorus, 385, 396 total phosphorus, 384 water quality, 384 Phosphorus loadings, 54, 571, 579 WCAs, 228, 570 Phosphorus threshold, 595 gradient study, 612 zone, 615, 616 Phosphorus threshold concentration zone, 599 Photochemistry mineralization experiment, 366 Photosynthesis, 511, 525 foliar N concentration, 526 foliar P concentration, 526 measurements, 524 Photosynthetic C3 pathway cattail and sawgrass, 549

Index Photosynthetic light compensation, 418 Photosynthetically available radiation (PAR), 418 light reduction, 436 measurement, 423 readings, 436 Physical control of Typha, 630 Physiography, 168 Phytosociological relevés, 462 Pickerel-weed. See Pontederia cordata Piecewise linear model, 569 Piezometer, 177 Pigweed. See Chenopodium Pine. See Pinus Pinnularia gibba, 466 Pinus, 26 pollen, 338 Pisaster, 601 Pistia stratiotes, 232, 633 pK dissociation relationships CO2, 607 CO3=, 607 HCO3-, 607 Planorbella duryi, 499 Planorbella scalaris, 499 Plant biomass, 84, 154 compartments, 353–355, 363, 524 P gradient, 154 Plant communities, 17, 74, 246, 419, 507, 627 controlling factors, 33, 237 response to N and P fertilization, 232, 505 response to disturbance, 531 structure, 215, 572 Plant density responses to P additions, 426 Plant frequency gradient, 231 Plant gradient survey methods, 219 Plant litter organic nutrients, 355 Plant species composition along nutrient gradients in WCA-2A, 230 Plant species composition in WCA-3A, 234 Plant succession, 627 disturbance, 531, 639 phosphorus, 531 water level, 531 Plants species diversity method, 512 Plexiglas algae, 273, 462, 469 mat comparisons, 470 Plexiglas slides, 262, 480, 501, 637 periphyton, 491

693 P-limitation ecosystem, 595 Eleocharis, 518 PO4. See Phosphorus; speciation Pollen absolute diagram, 339 analysis, 322 as P indicator, 329 assemblages, 348 diversity, 349 diversity changes, 340 indicator taxa, 327 influx, 339 methods, 327 profiles, 23, 25, 27 spore taxa, 337 Polygonum, 84, 93, 160, 237, 329 Pond apple. See Annona glabra Pond apple forests, 80 Ponds, 79 Pontederia cordata, 75, 80, 90, 232, 236 Pore water, 237, 423 Ca, 402, 541 Na, 541 NH4-N, 402, 541 NO3-N, 541 pH, 401, 541 PO4, 402 Potassium, 256, 262, 400, 457, 522 Precipitation rainfall, 37, 171, 225, 226, 398 atmosphic P deposition, 50 nutrients, 47, 48 orthophosphate, 50 total P concentration, 50 Prescribed burning, 259, 630, 631, 637 Primary consumers, 489 Procambarus fallax, 301, 311 Production macrophytes, 157, 363, 437, 520, 524, 525 Productivity Everglades sloughs, 413 plant gradient, 572, 574 subsidy-stress model, 418 Project overview central questions, 4 research program, 4 Pseudanabaena, 263, 466

Q Quality Assurance Plan (QAP), 105

694 R r adapted species, 417 Radiocarbon dating, 18 Radiometric dating Pb-210, C-14, Cs-137, 18, 21, 22, 580 Ragweed. See Ambrosia Rainfall. See Precipitation Raised-bog, 622 Recalcitrant phosphorus, 514 Recovery wells, 44 Red bay. See Persea borbonia Redox soil, 257, 332, 358, 360, 459, 542 Reference area, 105, 223, 595 Reference system, 595 Reference zone, 246, 248, 253, 282, 311 Refugia, 625 Relative Growth Rate (RGR), 561 Research program, 5 Reservoirs water area, 212 Residence times WCA-2A, 175, 188, 387 Resource limitation, 495, 496 Restoration, 621–642 community restoration target areas, 623 ecological approach, 621–642 lessons algal responses, 274–275 carbon cycling, 368–370 decomposition of litter and peat, 459 Dissolved Organic Matter export, 368–370 ecosystem responses to P dosing, 415–418 effects of disturbance, P, and water level on plant succession, 542–544 enrichment gradients, 151–165 environmental and human factors, 56–58 establishment and seedling growth of Typha domingensis and Cladium jamaicense, 563–564 historical changes, 350–351 hydrology gradients, 187 macroinvertebrate responses, 317–319 macroinvertebrate and fish response to P, 502 macrophyte community response, 258–260 macrophyte slough community response to experimental P, 438

Index management decisions for tree islands, 212 P dosing and soil chemistry, 415–418 P dosing and water quality, 415–418 P effects on algal assemblages, 472 P threshold, 616–617 phosphorus accretion rates, 592–593 phosphorus Assimilative Capacity, 577–578 plant community response to N and P fertilization, 526–527 soil characteristics, 72–73 spatial distribution of TP, 592–593 vegetation and algae, 87 plans, 633 pond apple communities, 628 sloughs, 629 successional dynamics, 626 tree islands, 630 wet prairie, 629 Restoration guidelines, 634–640 Restoration model, 628 Restudy. See Florida Project Comprehensive Review Study RGR. See Relative Growth Rate Risk uncertainty analysis, 606 River of Grass, 1, 13, 188, 636 Root:shoot mass ratio, 562 Rostrhamus sociabilis plumbeus, 45, 209 Royal fern. See Osmunda regalis, Ruderal-competitor, 563

S Sagittaria, 75, 78, 89, 160 P response, 422 pollen, 340 Sagittaria lancifolia, 26, 230, 232. 236, 237, 329, 420, 507 Salix, 76, 80, 85, 93, 232, 237 TP, 232, 628 Salix caroliniana, 236, 257 Saltwater, 627 Salvinia minima, 89, 232, 257 Sampling centroids vegetation, 223 Sangamon interglacial stage, 170 Sapric peat, 351 Sarcostemma, 231 Saturation index Ca, 398 Save Our Everglades, 323

Index Sawgrass. See Cladium jamaicense Schinus terebinthifolius nonnative taxa, 28, 33, 216, 632 Schoenoplectus tabernaemontani, 90, 232 Scirpus, 232 Scytonema, 465 Sea level, 14, 59, 168, 170 Sea level rise, 627 Sea otter, 601 Secondary production, 500 Sediment P chemistry, 403 TP, 303, 492, 500 Sediment APA-c, 410 Seed bank, 631 Seed germination rate, 538 Seed mass, 552 number, 552 Seed production, 552 Seed viability, 554 Seedling growth, 552, 560 Seedling mass, 561 Seedling relative growth rate, 561 Sentinel wetland, 641 Seral stage, 627 SFWMD. See South Florida Water Management District Shannon-Weaver diversity, 596 algae, 463, 468 diatoms, 327 pollen, 327 Shark River Bedrock Slough, 19, 45, 79, 171, 420 Shelford’s Law of Tolerance, 279, 499, 497 Shrub swamps TP, 628 Silica, 128, 238, 332 Silver Bluff Terrace, 170 Slough surface orthophosphate, 129, 394, 436 Sloughs, 78, 84 dosing experiment, 385 fertilizer experiment, 507 plant communities, 7, 26, 75, 217, 230, 249, 385, 419 restoration, 629 Small, J.K. From Eden to Sahara, 24 Smoothing splines, 581 Sodium, 124 accumulation, 333 gradient, 122, 262 molar ratios, 400 porewater, 237 vegetation, 257

695 Soil accumulation rates, 329, 342 BSi, C, Ca, N, Na, P, 330 bulk density, 62, 135, 137, 153, 353, 581, 587 natural gradient, 586 characteristics, 59 chloroform released P, 451 dating methods, 326 disturbance, 532 exchangeable P, 451 methods, 60 microbial biomass P, 451 nitrogen concentrations, 62, 515 organic carbon, 62 phosphorus, 62, 628 phosphorus concentrations, 62, 512 respiration, 352 total phosphorus pore water gradient, 229 spatial distribution, 579 threshold, 242 volume-based STP, 584 weight-based STP, 584 Soil gradient survey methods, 219 Soil phosphorus fractionation, 109, 133, 517 historical concentrations, 134 Kriged mapping results, 137 spatial distribution model, 583 Soil physical properties, 110 bulk density, 110 spatial extent, 110 Soil subsidence rate, 442 Soil Total Phosphorus (STP). See Soil; total phosphorus Soligenous. See Peatland Soluble Reactive Phosphorus, 53, 105, 116, 118, 385, 396, 613 loadings, 396 South Florida, 168 South Florida Water Management District, vii website, 3, 41 Spatial distribution patterns volume-based STP, 586 weight-based STP, 586 Species composition point intercept method, 220 Species turnover, 410 Spirogyra, 266, 269, 273, 461 Spirulina subsalsa, 270, 466 S-Plus, 607 Spotted sunfish. See Lepomis punctatis

696 SRP. See Soluble reactive phosphorus St. Lucie Canal, 39, 40, 42 Standing litter, 524 Starfish. See Pisaster STAs. See Stormwater Treatment Areas Staurastrum, 466 Stigonema, 263, 465, 472 Stomatal conductance cattail and sawgrass, 550 Stormwater Treatment Areas, 55, 567, 630 design criteria, 578 P discharges, 56, 617 P loadings, 56, 578 reductions, 56 seed source Typha domingensis, 564 STP. See Soil; total phosphorus Stress response, 279, 301 Stress-tolerant species, 563 Subsampling effects macroinvertebrates, 299 Subsidy effect, 598 Subsidy responses, 301 Subsidy-stress gradient, 279 Subsidy-stress model, 279–280, 478, 597 Substrata, natural algae, 273 Subsurface soil N mineralization, 451 Subtropical fen, 72, 641 Succession, 627 controlling factors, 33, 35, 57, 216, 225, 273 model, 627 restoration, 531, 537, 626 vectors, 453, 495 Sugarcane fields DOC, 365 production, 195 Sugarcane production, 195 Sulfate dosing concentrations, 400 gradients, 125–127 Sulfur bacteria, 636 Superphosphate, 510 Surface soil N mineralization, 451 Surface water Dissolved Oxygen, 109 irradiation reduction PAR, 436 Mg, 541 NH4-N, 541 Surface water flow, 185 resistance, 185

Index Surface water quality Water Conservation Area 2A, 115 Water Conservation Area 3A, 115 Swamp bay. See Magnolia virginiana Swamplands Act of 1850, 195 Sweep net, 497

T Tamiami Basin, 171 Taxonomic groups, 87 Temperature daily, 36 mat, 399 Thelypteris, 26, 236 Three dimensional control model Everglades, 628 Threshold probability, 609 Throw traps, 497 Tillers cattail and sawgrass, 550 Tolypothrix, 263 Top-down control, 316, 496 Top-down regulation, 312 Topogenous. See Peatland Total P loads EPA, 55 Hillsboro Canal, 118 TP threshold nonexceedance value, 617 TP/PO4-P ratio, 605 Trace metals analysis, 112, 142 Transition zone indicators Macroinvertebrates, 311 Transitional levee Habitat, 86 Transplant experiments Dosing, 432 Transplant study Dosing, 421 Utricularia purpurea, 421 Tree islands, 80, 86, 191–214, 625 Classes, 202 Comparison between WCA-1 and WCA-2, 211 Edge-to-edge ratio, 208 Habitat edge, 206 Imagery analysis, 201 Landscape pattern, 203 Loss, 204 Nearest neighbor, 206

Index Patch density, 206 Perimeter reduction, 211 Restoration, 630 Total numbers, 204 Total perimeter, 206 TP, 628 Tree snails. See Liguus Trichoptera, 303 Trophic levels, 495 Type 1 error, 607 Typha domingensis, 154, 549, 607 Aboveground biomass, 154 Carbohydrate reserves, 630 Carbon fixation rate, 550 Control Mechanical disturbance, 631 Establishment, 547 Frequency, 230, 575 Relationship with soil TP, 238 Germination, 540, 554 Germination experiments, 551 Germination rate, 554 Invasion potential, 625 Invasions, 539 Leaf area, 524 Lifespan, 549 Litter C:N ratio, 455 C:P ratio, 455 Cu concentrations, 455 decay, 446 N accumulation, 450 N:P ratio, 455 P concentrations, 455 litter bags, 444 N:P ratios, 229, 518 P effects, 217 photosynthesis, 511, 525 pollen, 337 Relative Growth Rate, 562 seed production, 551 removal, 626, 630 chemical control, 630 physical control, 630 prescribed burning, 630 restoration of community types, 626 seed bank, 639 seed germination rate, 630 seedling growth, 547 species descriptions, 548 tissue nutrient concentrations, 515 transpiration rate, 548 Typhaceae, 90, 549

697 U Uncertainty risk, 606 analysis, 611, 615 US Army Corps of Engineers, 1, 39, 197, 579 Central and Southern Florida Project for Flood Control and Other Purposes, 195, 323 Comprehensive Plan (1948), 39 website, 39, 45 US Environmental Protection Agency, 42, 617 5 X rule nondetects, 108 P criterion, 617 Utricularia, 75, 85, 92, 230 decline, 420, 535 disappearance, 506, 518, 521, 538 inhibition, 429 hypothesis, 429 N:P mass ratio, 518 populations water levels, 425, 538 successional pattern, 538 Utricularia foliosa, 501 Utricularia purpurea, 93, 242, 418, 501 as keystone species, 427 decline, 349, 406, 420 density, 425, 429 growth rate, 438 indicator species, 242, 607 P threshold, 433 shading effects, 438 stem lengths, 438 survival, 432 Utricularia spp. keystone species, 612 Utricularia-periphyton complex, 517 nitrogen content, 517 phosphorus concentrations, 517 UV radiation abiotic mineralization of DOC, 366

V Vegetation, 257 communities, 73 loss of wetland types, 73 historic map, 17 Historical changes, 321 Pattern, 318 Responses to hydrologic indicators, 242 Vegetation composition coarse-scale ordination, 251 Vegetation gradient partial Mantel tests, 251

698 Vegetation studies, 217 Vegetation-environment linkages, 223, 249 allogenic, 253 autogenic, 253 Vegetative reproduction, 549, 550

W Water Conservation Area 1, 9, 192, 580 peat accretion, 67, 580, 582, 622 cattail invasion potential, 626 DON budgets, 364 Total Soil Phosphorus, 584 Water Conservation Area 2A, 83, 120, 125, 154, 197 area, 217 cattail invasion potential, 626 DON budgets, 364 enriched carbon budget, 352 high water levels, 227 plant communities, 83 biomass, 83 soil P, C, and N gradients, 138 sulfate, 126 Total Soil Phosphorus, 584 unenriched carbon budget, 352 water control structures, 172 Water Conservation Area 2B fertilizer study, 532 Water Conservation Area 3A, 83, 118, 120, 124, 125, 158, 236, 580 area, 218 cattail invasion potential, 626 DON budgets, 364 nutrient storage, 158 plant biomass, 158 plant communities, 83 biomass, 83 soil P, C, and N gradients, 138 Total Soil Phosphorus, 584 Water Conservation Areas, 579 inflow P concentrations, 52 interior P concentrations, 52 N:P ratios, 54 nitrogen gradient, 54 outflow P concentrations, 52 spatial P distributions, 579 Water demand, 622, 641 agriculture, 641 Everglades, 641 urban, 641 Water depth, 236, 243, 534, 607 recurrence interval calculations, 223 recurrence intervals, 180–182

Index Water flow, 39, 170, 636 direction, 183 historic conditions, 41 Water hyacinth. See Eichhornia crassipes Water hyssop. See Bacopa caroliniana Water impoundment, 636 Water irradiation reduction, 436 Water lettuce. See Pistia stratiotes Water level dosing, 393 effects, 242, 259, 346, 347, 425, 479, 535 Lake Okeechobee, 41 sloughs, 79 WCA-2A, 199, 227 WCA-2B, 509, 535 WCA-3A, 236 Water quality historic, 348 pH. 347 statistical analysis of data, 108 surface, 104, 262 Water Resources Development Act (2000), 45 Water sampling Dissolved Organic Phosphorus, 106 Methods, 105 Particulate P, 106 Water velocity profiles, 185 Wax myrtle. See Myrica cerifera WCAs. See Water Conservation Areas West Palm Beach Canal, 195 Wet prairies, 75, 78 restoration, 629 vegetation, 538 Wetland bioassessment, 278 Wetland loss types Everglades, 73, 74 White ibis. See Eudocimus albus Willoughby, H.L., 39 Willow. See Salix Wood stork. See Mycteria Americana

Z Zinc, 140, 141 β-diversity χ2 test, 607

Ecological Studies Current List of Titles in Ecological Studies Series: Volume 185 Ecology and Conservation of Neotropical Montane Oak Forests M. Kappelle

Volume 194 Clusia: A Woody Neotropical Genus of Remarkable Plasticity and Diversity U.E. Luttge

Volume 186 Biological Invasions in New Zealand R.B. Allen

Volume 195 The Ecology of Browsing and Grazing I.J. Gordon

Volume 187 Managed Ecosystems and CO2: Case Studies, Processes, and Perspectives J. Nosberger

Volume 196 Western North American Juniperus Communities: A Dynamic Vegetation Type O. Van Auken

Volume 188 Boreal Peatland Ecosystems R.K. Wieder

Volume 197 Ecology of Baltic Coastal Waters U. Schiewer

Volume 189 Ecology of Harmful Algae E. Graneil

Volume 198 Gradients in a Tropical Mountain Ecosystem of Ecuador E. Beck

Volume 190 Wetlands and Natural Resource Management J.T.A. Verhoeven Volume 191 Wetlands: Functioning, Biodiversity Conservation, and Restoration R. Bobbink Volume 192 Geological Approaches to Coral Reef Ecology R.B. Aronson Volume 193 Biological Invasions W. Nentwig

Volume 199 Hydrological and Biological Responses to Forest Practices: The Alsea Watershed Study J. Stednick Volume 200 Arid Dune Ecosystems: The Nizzana Sands in the Negev Desert S. Breckle Volume 201 The Everglades Experiments: Lessons for Ecosystem Restoration C.J. Richardson