Carbon Dioxide, Populations, and Communities
Carbon Dioxide, Populations, and Communities
This is a volume in the PH...
75 downloads
1118 Views
23MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
Carbon Dioxide, Populations, and Communities
Carbon Dioxide, Populations, and Communities
This is a volume in the PHYSIOLOGICAL ECOLOGY series Edited by Harold A. Mooney
Carbon Dioxide, Populations, and Cornrnunities Edited by
Christian Korner Botanisches lnstitut der Universitat Basel Basel, Switzerland
Fakhri A. Bazzaz Department of Organismic and Evolutionary Biology Harvard University Cambridge, Massachusetts
Academic Press San Diego
New York
Boston
London
Sydney
Tokyo
Toronto
This book is printed on acid-free paper. Copyright 9 1996 by ACADEMIC PRESS, INC. All Rights Reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher.
Academic Press, Inc. A Division of Harcourt Brace & Company 525 B Street, Suite 1900, San Diego, California 92101-4495
by Academic Press Limited 24-28 Oval Road, London NW1 7DX Library of Congress Cataloging-in-Publication Data Carbon dioxide, populations, and communities / edited by Christian KOrner, Fakhri A. Bazzaz. p. cm.-- (Physiology ecology series) Includes bibliographical references (p. ) and index. ISBN 0-12-420870-3 (alk. paper) 1. Plants, Effect of atmospheric carbon dioxide on. 2. Atmospheric carbon dioxide--Environmental aspects. 3. Plant communities. 4. Plant ecophysiology. I. KSrner, Christian. II. Bazzaz, F. A. (Fakhri) A.) III. Series: Physiological ecology. QK753.C3C38 1996 581.5'222--dc20 96-33959 CIP
PRINTED IN THE UNITED STATES OF AMERICA 96 97 98 99 00 01 EB 9 8 7 6 5
4
3
2
1
Contents
Contributors Preface xix
xiii
Part Population-Level
I Responses
I. II. III. IV.
Introduction 3 The Genetic Bases for Evolutionary Responses to Climate Change Thermal Sensitivity and Evolutionary Responses to Climate Change Summary 11 References 12
I. II. III. IV.
Introduction 13 Experimental Methods Results and Discussion Conclusions 20 References 21
References
30
15 17
4 7
I. Plant Responses to Environmental Change: Theory and Review of Previous Work 31 II. An Experiment to Test Genotypic Responses to Increased CO2 38 III. Results from the Experiment and Discussion 39 IV. Outlook 47 V. Summary 48 References 49
I. II. III. IV. V. VI.
Introduction 51 54 Genetic Variability in C O 2 Responses Effects of Elevated CO2 on the Selection Process 61 What Characters Will Be Selected? 68 Possible Effects of Evolutionary Changes on Ecosystem Processes Summary 74 References 75
Part I I Commtmity-Level
Responses
I. II. III. IV. V.
Introduction 85 Current Vegetation Changes in Western Europe 86 Plant Functional Types and Response to Elevated CO2 Feedbacks 89 Summary 91 References 91
I. II. III. IV. V.
Introduction 93 Methods 94 Results 96 Discussion 97 Summary 98 References 99
88
73
I. II. III. IV.
Introduction 10l Responses at the Level of the Individual 102 Responses at the Level of the Plant Community Conclusions and Recommendations 118 References 119
I. II. III. IV. V.
Introduction 123 The Experimental Designs 124 Species Responses within Community Ecosystem Responses 127 Discussion and Conclusion 131 References 136
103
126
Community Microcosms I. II. III. IV. V. VI.
Introduction 139 The Jasper Ridge CO2 Experiment Methods 142 Analysis 144 Results 145 Discussion 151 References 155
140
I. Introduction 159 II. Design of CO2 and Plant Diversity Treatments in Calcareous Grassland Communities 160 III. Response of Calcareous Grassland Communities to Manipulations of CO2 and Plant Diversity 164 IV. Discussion 166 V. Summary 173 References 174
I. II. III. IV.
Introduction 177 CO2 and Vegetation Change Conclusions 188 Summary 189 References 190
I. II. III. IV. V.
Introduction 197 Site Description and Methodology 198 The Response of Primary Producers 200 Other Trophic Levels 203 Conclusions 204 References 205
I. II. III. IV. V.
Introduction 209 Experimental Setup and Methods Results 216 Discussion 221 Summary 226 References 227
I. II. III. IV. V. VI.
Introduction 231 The Role of Fire in Plant Communities 232 CO2 Effects on Vegetation and Fire 235 Predicting High CO2 Effects on Future Fire Cycles Research Priorities 243 Summary 244 References 245
181
211
241
Part I I I Interactions
Organismic
I. Introduction 253 II. The Symbiotic N z Fixation: A Highly Flexible Way to Assimilate Nitrogen 254 III. The Link between Plant Growth, Nitrogen Assimilation, and N 2 Fixation 255 IV. The Link between Elevated CO2 and N Availability in the Soil 255 V. The Response of Symbiotic N 2 Fixation to Elevated CO2 in the Field: A Response to Both Increased Legume N Demand and Increased Strength of the Ecosystem N Sink? 258 VI. Model and Conclusion 259 VII. Summary 260 References 261
I. II. III. IV.
Introduction 265 Responses of Symbiotic Fungi to a COz-Enriched Environment Community and Ecosystem Level Consequences 268 Summary 271 References 272
I. Introduction 273 II. Survey of Differential Species Responses within Species Mixtures Exposed to Elevated CO2 275 III. Conclusions 282 References 284
266
I. II. III. IV.
Introduction 287 Materials and Methods Results and Discussion Summary 297 References 298
I. II. III. IV. V. VI. VII. VIII.
288 289
Introduction 301 Theory 301 An Experimental Test 303 Temperature, Phenology, and Competition 311 Species Competitive Abilities 313 Competition and Internal Plant Nutrient Status 314 Longer Term Implications 314 Summary 315 References 316
I. Introduction 319 II. Shade Tolerance of Spruce Seedlings and Some Co-occurring Grasses at Ambient and Elevated CO2 Concentration in Air 320 III. Interaction between Spruce Seedlings and at Ambient and Elevated CO2 327 IV. Conclusions 329 References 330
I. II. III. IV. V.
Introduction 333 Developmental Patterns and Competitive Ability 334 Predicting the Effect of Elevated CO2 on Development 336 Effect of Developmental Patterns on CO2 Responsiveness 341 Summary 344 References 345
I. II. III. IV. V. VI.
Introduction 347 Direct Effects of CO2 349 Interactive Effects: CO2 and Resource Availability 353 Indirect Effects: CO2 and Climate 354 Tritrophic Interactions 355 Predicting Insect Outbreaks: The Functional Attribute Approach 356 VII. Conclusions and Recommendations 357 References 359
I. II. III. IV. V.
Introduction 363 Ruminant Digestion 364 Impact of Elevated CO2 on Forage Quality 366 Impact of Elevated CO2 on Cattle Production in Tallgrass Prairie Conclusions 369 References 370
Part IV Theory, Modeling, Concepts
I. II. III. IV. V. VI.
The Need for Functional Types 375 Methodological Aspects 376 Interspecific Variation 380 Differences between Species at Low N Levels Ecological Aspects 390 Summary 391 References 406
I. Introduction 413 II. Light Interception in Mixed Species Stands
389
414
367
xii
III. Effect of CO2 Elevation on the Canopy Development in Two Annuals 418 IV. Nitrogen Allocation and Optimal Leaf Area Index under Elevated CO2 420 V. Summary and Conclusion 426 References 428
I. II. III. IV. V.
Introduction 431 Extrapolation 432 Reductionism from Below 434 Two Approaches to Research on Elevated CO2 Prediction and Uncertainty 439 References 440
I. II. III. IV. V. VI. VII. VIII. IX.
Index
Introduction 443 Why Study Variance? 444 Genotypic Responses in Populations 445 Responses of Plant Communities 446 Plant-Plant Interactions 449 Plant-Microbe Interactions 451 Plant-Animal Interactions 451 Theory, Modeling, Concepts 452 Ecosystem and Global Consequences 453 References 455 457
436
Contributors
on
D. D. Aekerly (413), Department of Organismic and Evolutionary Biology, Harvard University, Cambridge, Massachusetts 02138 John A. Arnone III (101), Institute of Botany, University of Basel, CH-4056 Basel, Switzerland Lisa M. Auen (363), Department of Agronomy and Department of Animal Science and Industry, Kansas State University, Manhattan, Kansas 66506 F. A. Bazzaz (413, 443), Department of Organismic and Evolutionary Biology, Harvard University, Cambridge, Massachusetts 02138 A. Birrer (31), Institut for Umweltwissenschaften, Universit~it ZOrich, CH8057 Zfirich, Switzerland L. O. Bj6rn (197), Department of Plant Physiology, University of Lund, Lund S-22100, Sweden Herbert Blum (253, 287), Institute of Plant Sciences, Swiss Federal Institute of Technology, 8092 Zfirich, Switzerland T. V. Callaghan (197), Department of Animal and Plant Sciences, The Centre for Arctic Ecology, The University of Sheffield, Sheffield $10 5BR, United Kingdom Bruce D. Campbell (301,375), AgResearch, Grasslands Research Centre, Palmerston North, New Zealand Nona R. Chiariello (139) ,Jasper Ridge Biological Preserve, Stanford University, Stanford, California 94305 Robert C. Cochran (363), Department of Animal Science and Industry, Kansas State University, Manhattan, Kansas 66506 Peter S. Curtis (13), Department of Plant Biology, The Ohio State University, Columbus, Ohio 43210 Paolo De Angelis (209), Department of Forest Environment and Resources, University of Tuscia, 1-01100 Viterbo, Italy Shivcharn Dhillion 1 (123), Centre d'Ecologie Fonctionnelle et Evolutive, Centre de la Recherche Scientifique, F-34033 Montpellier, France Christopher B. Field (139, 443), Department of Plant Biology, Carnegie Institution of Washington, Stanford, California 94305 1 Present Address: Department of Biology and Nature Conservation, Agricultural University of Norway (NLH), .3ts N-1432, Norway.
•
,.~
Beret Fischer (253), Institute of Plant Sciences, Swiss Federal Institute of Technology, 8092 ZOrich, Switzerland Marco Frehner (253), Institute of Plant Sciences, Swiss Federal Institute of Technology, 8092 Z/irich, Switzerland C. Gehrke (197), Department of Plant Ecology, University of Lund, Lund S-223 62, Sweden; and Abisko Naturvetenskapliga Station, Abisko S98107, Sweden Jan Gloser (319), Department of Plant Physiology, Faculty of Science, Masaryk University, 61137 Brno, Czech Republic J. P. Grime (85), NERC Unit of Comparative Plant Ecology, Department of Animal and Plant Sciences, The University of Sheffield, Sheffield $10 2TN, United Kingdom Jean-Louis Gtdllerm (123), Centre d'Ecologie Fonctionnelle et Evolutive, Centre de la Recherche Scientifique, F-34033 Montpellier, France D. Gwyrm-Jones (197), Department of Animal and Plant Sciences, The Centre for Arctic Ecology, The University of Sheffield, Sheffield S10 5BR, United Kingdom Alan L. Hart (301), AgResearch, Grasslands Research Centre, Palmerston North, New Zealand Ueli A. Hartwig (253, 287), Institute of Plant Sciences, Swiss Federal Institute of Technology, 8092 Z~rich, Switzerland Thomas Hebeisen (253, 287), Institute of Plant Sciences, Swiss Federal Institute of Technology, 8092 Z~rich, Switzerland George R. Hendrey (253, 287), Department of Applied Science, Brookhaven National Laboratory, Upton, Long Island, New York 11973 T. Hirose (413), Biological Institute, Faculty of Science, Tohoku University, Aoba, Sendai 980-77, Japan M. Jasiefiski (51), Department of Organismic and Evolutionary Biology, Harvard University, Cambridge, Massachusetts 02138 U. Johanson (197), Department of Plant Physiology, University of Lund, Lund S-22100, Sweden Hyrum B.Johnson (177), United States Department of Agriculture, Agricultural Research Service, Grassland, Soil & Water Research Laboratory, Temple, Texas 76502 T. Herin Jones (93), NERC Centre for Population Biology, Imperial College at Silwood Park, Ascot, Berks SL5 7PY, United Kingdom Susan Kalisz2 (13), Kellogg Biological Station, Michigan State University, Hickory Corners, Michigan 49060 Joel G. Kingsolver (3), Department of Zoology, University of Washington, Seattle, Washington 98195 Present Address: Department of Biological Sciences, University of Pittsburgh, Pittsburgh, Pennsylvania 15260.
2
Dawn Jenkins Klus (13), Kellogg Biological Station, Michigan State University, Hickory Corners, Michigan 49060 Ch. K6rner (159, 443), Institute of Botany, University of Basel, 4056 Basel, Switzerland Elena Kuzminsky (209), Department of Forest Environment and Resources, University of Tuscia, 1-01100 Viterbo, Italy C. Lavigne (31), Institut for Umweltwissenschaften, Universit~it Ziirich, CH8057 ZCtrich, Switzerland Sharon P. Lawler ~ (93), NERC Centre for Population Biology, Imperial College at Silwood Park, Ascot, Berks SL5 7PY, United Kingdom John H. Lawton (93), NERC Centre for Population Biology, Imperial College at Silwood Park, Ascot, Berks SL5 7PY, United Kingdom Paul W. Leadley (159), Botanical Institute, University of Basel, 4056 Basel, Switzerland J. A. Lee (197), Department of Animal and Plant Sciences, The University of Sheffield, Sheffield S10 2UQ, United Kingdom Richard L. Lindroth (347), Department of Entomology, University of Wisconsin, Madison, Wisconsin 53706 Andreas L/iseher (253, 287), Institute of Plant Sciences, Swiss Federal Institute of Technology, 8092 Ziirich, Switzerland Giorgio Matteueci (209), Department of Forest Environment and Resources, University of Tuscia, 1-01100 Viterbo, Italy Herman S. Mayeux (177), United States Department of Agriculture, Agricultural Research Service, Grassland, Soil & Water Research Laboratory, Temple, Texas 76502 Shahid Naeem 4 (93), NERC Centre for Population Biology, Imperial College at Silwood Park, Ascot, Berks SL5 7PY, United Kingdom Marie-Laure Navas 5 (123), Centre d'Ecologie Fonctionnelle et Evolutive, Centre de la Recherche Scientifique, F-34033 Montpellier, France Josef N6sberger (253, 287), Institute of Plant Sciences, Swiss Federal Institute of Technology, 8092 Z/irich, Switzerland Clenton E. Owensby (363), Department of Agronomy, Kansas State University, Manhattan, Kansas 66506 H. Wayne Polley (177), United States Department of Agriculture, Agricultural Research Service, Grassland, Soil & Water Research Laboratory, Temple, Texas 76502 3Present Address: Department of Entomology, Universityof California, Davis, Davis, California 95616. 4Present Address: Department of Ecology, Evolution, and Behavior, Universityof Minnesota, St. Paul, Minnesota 55108. 5Present Address: CEFE and Biologie et Pathologie V6g6tale, ENSA-M,F-34060 Montpellier, France.
• Hendrik Poorter (375), Department of Plant Ecology and Evolutionary Biology, Utrecht University, 3508TB Utrecht, The Netherlands Catherine Potvin (23), Department of Biology, McGill University, Montr6al, Qu6bec H3A 1B1, Canada E. G. Reekie (333), Biology Department, Acadia University, Wolfville, Nova Scotia BOP 1X0, Canada Heather L. Reynolds (273), W. K. Kellogg Biological Station, Michigan State University, Hickory Corners, Michigan 49060 Catherine R0umet (375), Centre d'Ecologie Fonctionelle et Evolutive, Centre National de la Recherche Scientifique-Centre Louis Emberger, 34033 Montpellier, France Jacques Roy (123), Centre d'Ecologie Fonctionnelle et Evolutive, Centre de la Recherche Scientifique, F-34033 Montpellier, France Rowan F. Sage (231), Department of Botany, University of Toronto, Toronto, Ontario M5S 3B2, Canada Ian R. Sanders (265), Institute of Botany, University of Basel, CH-4056 Basel, Switzerland Giuseppe E. Scaraseia-Mugnozza (209), Department of Forest Environment and Resources, University of Tuscia, 1-01100 Viterbo, Italy B. Schmid (31), Institut for Umweltwissenschaften, Universit~it Zfirich, CH8057 ZOrich, Switzerland M. Sonesson (197), Department of Plant Ecology, University of Lund, Lund S-223 62, Sweden; and Abisko Naturvetenskapliga Station, Abisko S-98107, Sweden S. C. Thomas (51), Department of Organismic and Evolutionary Biology, Harvard University, Cambridge, Massachusetts 02138 LindseyJ. Thompson (93), NERC Centre for Population Biology, Imperial College at Silwood Park, Ascot, Berks SL5 7PY, United Kingdom Charles R. Tisehler (177), United States Department of Agriculture, Agricultural Research Service, Grassland, Soil & Water Research Laboratory, Temple, Texas 76502 StephenJ. Tonsor 6 (13), Kellogg Biological Station, Michigan State University, Hickory Corners, Michigan 49060 Denise Tousignant 7 (23), Department of Biology, McGill University, Montr6al, Qu6bec H3A 1B1, Canada Chris Van Kessel (253), Department of Soil Science, University of Saskatchewan, Saskatoon, Saskatchewan S7N 0W0, Canada Present Address: Department of Biological Sciences, Universityof Pittsburgh, Pittsburgh, Pennsylvania 15260. 7 Present Address: Ministere des ressources naturelles, Peponiese St. Modeste, St-Modeste, Quebec J0L 3W0, Canada.
6
Jacob Weiner 8 (431), Biological Laboratories, Harvard University, Cambridge, Massachusetts 02138 Richard M. Woodfin (93), NERC Centre for Population Biology, Imperial College at Silwood Park, Ascot, Berks SL5 7PY, United Kingdom SUvia Zanetti (253, 287), Institute of Plant Sciences, Swiss Federal Institute of Technology, 8092 ZCtrich, Switzerland
8 Present Address: Department of Biology, Swarthmore College, Swarthmore, Pennsylvania 19081.
This Page Intentionally Left Blank
Preface
In many aspects of biological science, variance is considered a nuisance. Variance makes experimental design difficult, requires much replication and detailed statistical analysis, and may greatly increase the need for expensive research facilities beyond the reach of many investigators. The results of some critical experiments may be statistically insignificant because of low replication. Consequently, many researchers in plant sciences, including those investigating CO2, have attempted to eliminate variance by cloning plants or by working with a small set of individuals and assuming that they represent the species. On the other hand, some researchers consider assemblages of plants and treat whole plant communities as a single "big leaf." In many of these cases, these smooth results and their mean values eliminate variance among individuals in the populations, which is the material for natural selection and evolution. Variance in response among individuals and species may be the most important aspect to consider in predicting the vegetation response to the continued enrichment of the atmosphere with CO2. Because of the speed with which CO2 in the atmosphere is increasing (~1.8 ppm/year), there is little time for the evolution of a new set of genotypes specifically adapted to this new global level of CO2 in the atmosphere. Therefore, it is the current genetic structure of populations that will determine the response of plants to elevated levels of CO2, especially in long-lived woody or clonal plants. Research on the response of a variety of plant species over the past two decades has convincingly shown that species do differ in their response to CO2. These differences can be very large, even among co-occurring species of a community. It is also well established that genotypes within a population differ in their response to CO2 enrichment. Thus, in a changing CO2 environment there may be winner (positively responding) and loser (less responding) individuals or species. This differential response may be underway already. Furthermore, this differential response may determine the genetic structure of future populations, may change the dominance relationship in communities, and thus may alter the importance of some plant functional groups, which may have feedbacks on the functioning of ecosystems. It may also influence biological diversity of some ecosystems. We must remember, however, that under natural conditions, selection by one xix
factor may not be the case, as factors of the physical and the biological environment interact in a variety of ways and may collectively influence the direction and the strength of selection. For example, CO2 enrichment may increase the water-use efficiency of a species or a genotype, providing access to more soil moisture to its otherwise less efficient neighbor, which in turn may offer the preferred food for a butterfly that may be an important pollinator for a third species. Scientists concerned about the global increase in CO2 met in the hills of the SwissJura Mountains in August 1994 and discussed the topics in this volume. They emphasized the need for a better understanding of the differences among individuals and of interactions among themselves a n d with other organisms. They sought to identify critical questions in this research area. This meeting represents one of the major steps in the Global Change and Terrestrial Ecosystems (GCTE) research plan, and its outcome is intended to complement that of the preceding workshop on ecosystem response to elevated CO2, the results of which appeared last year in this series. We gratefully acknowledge the support of the Swiss National Science Foundation, the Swiss Academy of Sciences, the United States Electric Power Research Institute (EPRI), and the United States National Science Foundation. The Leuenberg Conference Center near Basel provided a charming environment for this workshop. CH. KORNER
F. A. BAZZAZ
I Population-Level Responses
This Page Intentionally Left Blank
1 Physiological Sensitivity and Evolutionary Responses to Climate Change
The global changes in C O 2 and climate expected to occur during the coming decades are but one of many types of environmental changes resulting from human activities during the past century. The ecological and evolutionary consequences of pollution, pesticides, heavy metals, and other environmental insults during the past 40 years have been well documented. It is natural to ask whether and how the lessons we have learned from such studies may be used to anticipate the evolutionary consequences of future climate change for populations and species. In this chapter I will argue that the evolutionary consequences of climate change may differ importantly from those documented by field studies of pesticides and many other environmental toxins for a rather simple reason: pesticides represent an abrupt step-change in the environment, whereas climate change represents a progressive, directional alteration of environmental conditions. To illustrate this point, I will suggest and defend two conjectures about the evolutionary consequences of climate change: 1. The evolutionary responses of populations and species to climate change will involve polygenic, not monogenic, genetical responses. 2. Species whose individuals have broad physiological tolerances to climatic conditions will be less able to adapt evolutionarily to rapid and
3
Copyright 9 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.
4
climate change than those species whose individuals have intermediate tolerances. For the sake of discussion, I will argue these points with greater confidence than is perhaps warranted. My goal, however, is to show how lessons learned from the evolution of pesticide resistance and from quantitative genetic models of physiological sensitivity may provide some useful guidelines for evolutionary studies of climate change.
Most discussions of the biological consequences of climate change have downplayed any potential role for evolution, arguing that the predicted rapid rate of climate change will preclude evolutionary responses: species will either adjust ecologically or become extinct. It is certainly true that the relatively long generation times and small effective population sizes of many trees and large vertebrates make evolutionary responses to future climate change, which may occur on time scales of one to several decades, ineffective. However, the long history of studies of the evolution of resistance to pesticides and heavy metals clearly demonstrates that evolutionary responses to environmental changes can be rapid indeed in species of interest to humans. It is more useful to ask in what cases rapid evolution is likely to be important. First, rapid evolution is most likely in populations with large population sizes (e.g., >105-106), with short generation times (e.g., <1 year), and with high intrinsic rates of increase (e.g., rm > 0.5/generation). Thus, evolutionary responses to climate change may be quite likely in pest species. Second, rapid evolution is more likely when migration and dispersal are geographically restricted. Thus, evolutionary adaptation to climate change may be of importance in nature reserves and on other habitat islands, where the diluting effects of gene flow are reduced and where range shifts into new areas may be constrained.! geographically. Studies of the evolution of insecticide resistance in insect pests provide some useful insights into the possible genetic bases for evolutionary responses to climate change. A key question is whether evolutionary adaptation to climate change is more likely to involve a few major gene loci of large effect (monogenic), or many gene loci each of small effect (polygenic). Although the genetic basis for insecticide resistance has been intensively studied since the 1950s (Crow, 1957), the early studies appeared to be conflicting. Only in the last decade has a clear pattern in this literature been suggested: most laboratory studies, in which artificial selection is used
1.
5
to select for resistance from an initially susceptible population, indicate a polygenic basis for resistance. Conversely, many field studies, in which resistant genotypes are sampled from a population that has experienced frequent heavy insecticide doses, indicate a monogenic basis for resistance (Roush and McKenzie, 1987). The example of diazinon resistance in sheep blowflies is instructive (McKenzie and Batterman, 1994). Diazinon was widely used in Australia to control sheep blowflies, which rapidly evolved resistance to this insecticide. Genetic studies showed that this resistance was the result of the same allelic substitution at a single locus in different populations; and this resistant genotype was maintained near fixation during two decades of routine application of diazinon in the field. Subsequently, four populations of resistant blowflies were established in the laboratory and subjected to artificial selection for further increases in resistance. Within eight generations, more highly resistant strains had been developed, and genetic analyses showed that the increased resistance was due to multiple loci on at least four chromosomes. Thus, while two decades of field selection resulted only in monogenic evolutionary changes, laboratory selection led quickly to polygenic responses (McKenzie and Batterman, 1994). What causes these differing evolutionary responses both in the laboratory and in the field? The best current explanation is that the evolutionary response may depend on the intensity of selection (Roush and McKenzie, 1987; McKenzie and Batterman, 1994). Consider a laboratory population undergoing artificial selection for increased resistance to an insecticide (Fig. 1, top). During artificial selection, the intensity of selection (i.e., the fraction of the population killed by selection) must be chosen such that surviving individuals form a population sufficiently large to avoid substantial inbreeding and subsequent drift. Typically the insecticide dosage is chosen to achieve a selection intensity less than 80-90%, even for relatively large laboratory populations (103-104). As a result, selection occurs within the range of existing genetic variation in resistance in the population, which is often the result of many loci. The situation in the field is quite different. Here frequent and heavy doses are traditionally used to eliminate the entire pest population, and selection in this case could exceed 99% or more (Fig. 1, bottom). Even in large pest populations, only the rare individual with a mutant allele of large effects on resistance would be likely to survive. Because these rare alleles would only appear initially in heterozygotes, such alleles would need to be partially or fully dominant to be expressed. Such resistant alleles frequently have deleterious pleiotropic effects in the absence of insecticides, hence their initial rarity. Thus the very high insecticide doses, and hence the high selection intensity, typical of the field situation may have selected for resistance caused by one or a few alleles of larger effect.
6
Figure I (Top) The distribution on the left (BEFORE) represents the distribution ot susceptible phenotypes within a population of insects before selection. When an insecticide dosage (Selection intensity, dashed line) is applied, only those individuals with resistances greater than the applied dosage (to the right of the dashed line) will survive. After repeated generations of selection, the distribution of phenotypes in the population will shift to the right as a result of evolution (AFFER). These selective conditions, typical of artificial selection studies in the laboratory, may preferentially select for a polygenic response. (Bottom) The distribution on the left (BEFORE) represents the distribution of susceptible phenotypes within a population of insects before selection. When a high insecticide dosage (Selection intensity, dashed line) is applied, only those individuals possessing mutations with large effects on resistance to the insecticide will survive. After repeated generations of selection favoring rare mutants of large effect on insecticide resistance, the distribution of phenotypes in the population will shift far to the right as a result of evolution (AFTER). These selective conditions, typical of insecticide applications in the field, may preferentially select for a monogenic response. Adapted from McKenzie and Batterman (1994), Fig. 1. T a b a s h n i k (1995) a r g u e s t h a t this d i c h o t o m y b e t w e e n m o n o g e n i c field r e s i s t a n c e a n d p o l y g e n i c l a b o r a t o r y r e s i s t a n c e is t o o simplistic. H e s u g g e s t s t h a t t h e e v i d e n c e t h a t r e s i s t a n c e typically h a s a m o n o g e n i c basis in t h e field is e q u i v o c a l a n d t h a t s o m e o f t h e m o s t t h o r o u g h s t u d i e s o f r e s i s t a n c e d o n o t fit n e a t l y i n t o e i t h e r t h e m o n o g e n i c o r p o l y g e n i c e x t r e m e s . F o r o u r p u r p o s e s , h o w e v e r , t h e p o i n t is t h a t m o r e g r a d u a l s e l e c t i o n is m o r e likely to s e l e c t o n e x i s t i n g q u a n t i t a t i v e v a r i a t i o n in r e s i s t a n c e in a p o p u l a t i o n a n d t h a t t h e c o n d i t i o n s o f s e l e c t i o n c a n i n d e e d i n f l u e n c e t h e g e n e t i c basis of resistance.
1.
7
What might this suggest for evolutionary responses to climate change? One characteristic of global climate change is that it will not occur in an abrupt step, but as a gradual, directional increase in CO2 and temperature. To those organisms most likely to respond evolutionarily--those with relatively short generation times and large population sizes--such gradual changes are likely to generate selection intensifies per generation more similiar to laboratory artificial selection regimes than to the extremes seen in field insecticide applications. Thus we might predict that the evolutionary responses to climate changes will involve primarily polygenic, not monogenic, control. One implication of this suggestion is that experimental studies exploring evolutionary responses to climate change need to consider the type of environmental change and the selection intensity imposed on the study population. In particular, studies that utilize abrupt step-changes in CO2 or temperature may affect the type of genetic responsesmpolygenic or m o n o g e n i c - - t h a t occur, in ways that do not accurately reflect the rates of change predicted from global climate change (see Tousignant and Potvin, Chapter 3). Again, the key issue is whether or not the imposed environmental change falls within the range of the existing genetic variation in the population. The notion that step-changes and progressive directional changes in the environment may yield qualitatively different responses has not received much attention. To further explore the importance of these differences, I will briefly discuss some recent models that address the question how does the physiological tolerance of individuals in a population affect the population's evolutionary response to selection?
One useful way to characterize the effects of temperature or other physical factors on the performance or fitness of an individual organism is in terms of a which maps environmental conditions onto physiological or ecological performance (Fig. 2). (For convenience I will focus on temperature effects throughout this secdon, but the basic ideas should apply to most other physical factors.) Performance initially increases with temperature, reaches some "optimal" temperature for performance, then declines rapidly as it approaches upper lethal levels. Frequently one can characterize the performance curve in terms of three parameters: optimal temperature (Z), the temperature at which performance is maximum; performance breadth (~r), the breath or width of the performance curve; and the maximum performance (Rmax), the level of performance at
Rmllx
-
Z ENVIRONMENTALTEMPERATURE Thermal performance curves illustrating the relationship between an individual's performance (assumed directly proportional to fitness) and the environmental temperature it experiences. For each individual, there is an optimal temperature (Z) at which performance is maximized (Rmax).Performance curves are given for two individuals with identical optimal temperature (Z) and maximal performance (Rmax),but that differ in thermal performance breadth (tr). Adapted from Huey and Kingsolver (1989).
the optimal t e m p e r a t u r e (Fig. 2). We shall assume that the measure of p e r f o r m a n c e chosen is directly related to fitness. O n e natural question is how do the values of these parameters affect the ecological and evolutionary response to climate change (e.g., climate warming) ? Suppose we consider two individuals with identical Z and Rmax, but that differ in p e r f o r m a n c e breadth t r m t h a t is, one is a thermal " specialist" (small tr), the other a thermal "generalist" (large tr) (Fig. 2). Suppose the environmental t e m p e r a t u r e 0 is initially the same as Z, but then 0 increases somewhat. Obviously the reduction in p e r f o r m a n c e (and hence the reduction in fitness) of the thermal generalist is less (Fig. 2). Similarly, a population of thermal generalists will suffer a smaller decline in m e a n fitness than a population of thermal specialists in the face of a small increase in environmental temperature. Clearly, thermal generalists are at an ecological advantage in the face of climate warming. But how does thermal p e r f o r m a n c e breadth affect the evolutionary response of a population to sustained, directional climate warming? We have recently e x a m i n e d this question (Huey and Kingsolver, 1993), modifying a quantitative genetic m o d e l developed by Lynch and Lande (1993) (see Fig. 3). The p e r f o r m a n c e curve identifies the optimal t e m p e r a t u r e (Z) and thermal p e r f o r m a n c e breadth (tr) of each individual (Fig. 2). Suppose that optimal t e m p e r a t u r e Z is now a polygenic trait with some constant phenotypic and genetic variation in the population, but that all individuals
1.
9
Diagram illustrating the effect of thermal performance breadth on a population's evolutionary response to climate warming. Here f(Z) is the frequency distribution of phenotypic trait Z, the optimal temperature for performance. In each panel, the solid line represents the change in environmental temperature (0), and the dashed line represents the change in the population mean value of Z with time. As time proceeds, a lag develops between the environmental optimum and the population mean phenotype. For populations with large thermal performance breadths (top), this lag will be greater than for populations with small performance breadths (bottom). From Huey and Kingsolver (1993), Fig. 5.
in the p o p u l a t i o n (with fixed, constant, effective p o p u l a t i o n size) have identical p e r f o r m a n c e b r e a d t h (t r) a n d m a x i m u m p e r f o r m a n c e (Rma~). Initially the e n v i r o n m e n t a l t e m p e r a t u r e 0 is at t h e m e a n o p t i m a l t e m p e r a ture Z o f the p o p u l a t i o n ; 0 t h e n increases at a c o n s t a n t m e a n rate, b u t with s o m e stochastic ( r a n d o m ) variation. Given this situation, t h e m e a n o p t i m a l t e m p e r a t u r e Z o f t h e p o p u l a t i o n will evolve toward increasingly h i g h e r values, b u t will lag b e h i n d the e n v i r o n m e n t a l t e m p e r a t u r e (Fig. 3). m
0 If the rate of climate warming is too rapid, the population's lag will become too great, its mean fitness will approach zero, and extinction will occur. Hence one can identify a critical rate of climate change above which population extinction will quickly occur (Fig. 3). Using this model, we can address how performance breadth affects the critical rate of climate change that a population can sustain. Consider the simplest case in which the genetic variation in optimal temperature Z in the population is constant with rime and independent of performance breadth. The model then predicts that the critical rate of climate change will initially increase with increasing performance breadth, quickly reach a maximal value, and then decline with increasing performance breadth (Fig. 4). Thus the model predicts that populations with intermediate performance breadths will be able to sustain the highest rates of climate c h a n g e - that populations of thermal generalists are more likely to become extinct in the face of rapid climate change. Stochastic variation in climate decreases the critical rate of change and increases the performance breadth at which the rate is maximal, but does not alter the qualitative result (Fig. 4). Interestingly, these results do not depend on the existence of tradeoffs between specialists and generalists (Huey and Kingsolver, 1993). What produces this apparently paradoxical result? The key once again is the importance of the intensity of selection. For a population of thermal
~0
"-2".
0.4
o. r
0
2
.,'>
~
~,~"
3 Theoretical predictions of the critical rate of climate change (in ~ above which population extinction will occur, as a function of thermal performance breadth (in ~ and the amount of stochastic climatic variation. Parameter values are for the case in which genetic variance is independent of time and of performance breadth. Adapted from Huey and Kingsolver (1993), Fig. 8.
1.
11
specialists, the change in climate each generation substantially reduces the population's mean fitness, generating strong selection on the population. Given available genetic variation, this intense selection will result in a substantial evolutionary response to selection and will reduce the lag of the population behind the environment (Fig. 4). For a population of thermal generalists, however, the change in climate each generation has a smaller effect on the population's mean fitness, and thus generates weaker selection. Because of the smaller evolutionary response to this weaker selection, the lag of the population behind the environment may become large: if the lag becomes too great, extinction will occur. Thus thermal generalists may be at an ecological advantage, but at an evolutionary disadvantage, relative to populations of individuals with intermediate thermal breadths (see Huey and Kingsolver, 1993, for further details). This result is sensitive to assumptions about the determinants of the genetic variation in the population. For example, if genetic variation in optimal temperature results largely from a balance of mutation, selection, and drift, one might expect populations of thermal generalists to possess greater genetic variance in Z than populations of thermal specialists; in this case, thermal specialists no longer can sustain higher critical rates of climate change than generalists (Huey and Kingsolver, 1993). However, Burger and Lynch (1995) have recently extended these models to allow for dynamic evolutionary changes in population size and genetic variance over time in the population. Their simulations suggest that, even when genetic variance is determined by a mutation-selection-drift balance and varies over time, the mean time to extinction will be greatest for populations with intermediate performance breadths. These models, while clearly simplistic, highlight some of the complexities in predicting the evolutionary responses of populations to progressive climate change. I believe the main lesson from the models is this: the evolutionary responses of populations to climate change cannot be predicted solely based on ecological information, but must include some understanding of the genetics of physiological performance traits, about which we know very little. Is optimal temperature a polygenic trait? Do populations of intermediate performance breadth possess less genetic variation for performance traits than populations of generalists? Will specialists and generalists respond in different ways evolutionarily to abrupt step-changes versus progressive changes in climate? These questions have hardly been asked, much less answered, by physiological ecologists; but the answers may be key to predicting the evolutionary responses to climate change.
Evolutionary responses to climate change are most likely to be of importance in abundant, short-lived species such as pests and in populations
12
restricted to habitat islands including nature reserves. There are compelling, empirical, and theoretical reasons to believe that evolutionary responses to abrupt step-changes in the environment may differ qualitatively from responses to gradual, progressive changes, because of the consequences of selection intensity for evolutionary response. Studies of the genetic bases for pesticide resistance over the past 40 years suggest that the evolutionary responses to climate change will most likely involve polygenic responses. Quantitative genetic models suggest that the performance breadth of physiological traits will play an important and complex role in evolutionary adaptation to climate change and indicate that the determinants of genetic variation in physiological performance are key to understanding the evolutionary responses of populations to progressive climate change.
I thank Ray Huey and Bruce Tabashnik for useful discussion of the issues considered here, an anonymous reviewer for helpful comments on the earlier version, and Ray Huey for help with Fig. 2. Research was supported in part by NSF Grants BSR-890130 and IBN-9220748 to the author.
Burger, R., and Lynch, M. (1995). Evolution and extinction in a changing environment: A quantitative genetic analysis. 49, 151-163. Crow, J. F. (1957). Genetics of insect resistance to chemicals. 2, 227-246. Huey, R. B., and Kingsolver, J. G. (1989). Evolution of thermal sensitivity of ectotherm performance. 4, 131-135. Huey, R. B., and Kingsolver, J. G. (1993). Evolution of resistance to high temperature in ectotherms. 142, $21-$46. Lynch, M., and Lande, R. (1993). Evolution and extinction in response to environmental change. "Biotic Interactions and Global Change" (P. M. Kareiva, J. G. Kingsolver, and R. B. Huey, eds.), pp. 234-250. Sinauer, Sunderland, MA. McKenzie, J. A., and Batterman, P. (1994). The genetic, molecular and phenotypic consequences of selection for insecticide resistance. 9, 166-169. Roush, R. T., and McKenzie, J. A. (1987). Ecological genetics of insecticide and acaricide resistance. 32, 361-380. Tabashnik, B. E. (1995). Genetic basis of insecticide resistance: Dichotomy or continuum. 10, 164-165.
2 Intraspecific Variation in C02 Responses in Raphanus raphanistrum and Plantago lanceolaCa: Assessing the Potential for Evolutionary Change with Rising Atmospheric CO2
Elevated atmospheric C O 2 has well-documented effects on plant physiology and growth that may alter important ecosystem processes, particularly carbon storage and nutrient cycling (Mooney 1991). Differential sensitivity among species to elevated CO2 may also alter interspecific interactions (e.g., competition, symbiosis, herbivory), resulting in changes in community structure (Bazzaz, 1990). Predictions of the ecological consequences of elevated atmospheric CO2 have almost always been based on mean physiological and growth responses from a small, and often arbitrary, subsample of the species under consideration. Global change analyses have also generally assumed both a uniform response to CO2 across populations and CO2 response levels that are stable over time. However, mean responses estimated from a small number of individuals are probably not representative of the range of responses found among distinct populations of a species. The range of responses determines the evolutionary potential of a population within a community. This is a restatement ofFisher's (1930) fundamental theorem:
13
Copyright 9 1996 by Academic Press, Inc. All fights of reproduction in any form reserved.
14
In the CAM species for example, both populations and genotypes within populations differed significantly in the relative proportion of CO2 uptake in the daytime versus that at night (Kalisz and Teeri, 1986). When such differences in CO2 uptake result in differences in growth, survivorship, a n d / o r fecundity, the most fit genotypes will come to predominate in future generations. The result will be a shift in the mean response of the population as a result of natural selection. If a species' mean physiological or growth response to elevated CO2 evolves over time, either positively or negatively, then predictions regarding possible community and ecosystem level responses to rising CO2 may need to be revised. The first step in understanding how communities and ecosystems evolve under elevated COz requires characterization of the variation within populations that constitute the community. A primary requirement for COz to act as a selective agent is the presence of heritable variation in CO2 responses of fitness-related traits among individuals in a population. Work with other anthropogenic environmental perturbations such as heavy metal, ozone, and herbicide pollution has shown that rapid evolutionary responses are possible within natural populations given even low levels of genetic variation in tolerance to these stresses (Bradshaw, 1991). However, not all populations or species possess the requisite genetic variation. Clearly, where heritable variation is absent, no adaptive change is possible. The nature and magnitude of genetic variation in CO2 responses are poorly known at present. Variation among two cultivars in grain production at elevated CO2 was shown by Ziska and Teramura (1992), Wulff and Alexander (1985) found differences among four maternal families of in germination and early growth at twice ambient CO2, and Garbutt and Bazzaz (1984) documented population-level differences in reproduction at high COz in Fajer (1992), however, found no significant differences in the magnitude of growth or leaf secondary chemistry responses to elevated CO2 among five clones, whereas Curtis (1994) found significant variation in the magnitude of changes in lifetime reproduction due to elevated CO2 among five paternal families of Making accurate predictions of responses to selection generally requires accurate narrow-sense estimates of heritabilities (Falconer, 1981). The narrow-sense heritability requires in turn an estimate of the additive genetic variance within the population. Additive genetic variance estimates are most often based on the correlations between paternal half-sibs or between pollen parent and offspring correlations. Although these approaches give the best estimates of the additive genetic variances, accurate estimates require many hundreds or thousands of offspring. This endeavor is extremely expensive when artificial elevation of atmospheric CO2 is involved.
2.
15
Because so little is known about the genetic basis for phenotypic variation in response to elevated CO2, more conservative initial approaches are necessary. A lower cost, initial estimate of the potential for evolutionary change can be gained from an examination of the variance among maternal fullsib families or from a limited number of paternal half-sib families. After significant variation at one of these levels is established, a more complete genetic analysis can be conducted. To improve our understanding of the evolutionary potential in CO2 responses, we characterized 12 additional paternal families in maternal families and conducted a separate study of 18 (Klus unpublished). Here we integrate results from these studies, presenting evidence for intraspecific, family-level variation in physiological, growth, and reproductive responses to twice ambient CO2. In the experiments, the goal was to obtain a preliminary test for the presence of additive genetic variance for leaf area and number of flowers, traits known to influence fitness in the field. In the experiments, the goal differed. Under normal atmospheric conditions populations of P. have already been shown to have heritable variation for a broad variety of traits likely to be affected by CO2 level or to affect fitness (Tonsor and Goodnight, 1996; Teramura and Strain, 1979; Wolff and Van Delden, 1989). Therefore, the goal in the experiments was to gain a mechanistic understanding of the range of responses to elevated CO2 within two populations and to gain a preliminary estimate for evolutionary potential in mechanistically important traits.
Raphanus raphanistrum Two separate experiments were conducted, in 1992 and in 1993. Seeds for parental genotypes were obtained from the same location in Deer Isle, Maine, in the year preceding each study. For the 1992 study, 5 individuals from separate maternal plants were randomly selected to serve as pollen donors and 20 randomly selected individuals, also from separate maternal plants, served as pollen recipients. Each pollen donor (male parent) was crossed with 4 recipients (female parents) in a nested design (females within males), yielding 5 paternal families. For the 1993 experiment, 12 separate males were crossed with each of 3 separate females in a standard diallel breeding design, yielding 12 paternal families. With these designs, significant paternal effects are interpreted as evidence for additive genetic variance among paternal genotypes. Experiments were conducted at the University of Michigan Biological Station, Pellston, Michigan, in large, open-top field chambers. Half of the
16
chambers were maintained at approximately 700/zl liter -1 CO2 and the remainder at ambient CO2 (--350/zl liter -1 CO2). Seeds were planted in 2.5-liter pots filled with a mixture of locally derived sand and topsoil and thinned to one plant per pot after germination. All pots were watered daily as needed. In 1992 there were no nutrient amendments whereas in 1993 each pot received three 23-mg additions of NHaNO~ over a 3-week period. Germination occurred within 4 - 6 days followed by growth of a basal rosette, with bolting and flowering beginning --30 days after planting. Each flower was hand pollinated 1-2 days after opening. Days to bolting, total leaf area at bolting (measured nondestructively), total n u m b e r of flowers produced, and total seed set (1992 only) were recorded for each plant. In 1992, 32 replicates per paternal family were randomly divided among 4 chambers for a total of 160 pots per CO2 treatment (5 paternal families • 4 maternal families • 4 replicates per female • 2 chambers per CO2 treatment). In 1993, 36 replicates per paternal family were divided among 6 chambers for a total of 216 pots per CO2 treatment (12 paternal families • 3 maternal families • 2 replicates per female • 3 chambers per CO2 treatment). Data were analyzed with a fixed-effects model ANOVA where CO2 and father were main effects and mother was either fully crossed (1993) or nested within fathers (1992). Response of progeny from individual fathers was evaluated with separate two-way ANOVAs where CO2 and m o t h e r were main effects.
Plantago lanceolata Seeds were collected from 12 open-pollinated plants (= maternal families) in each of two populations in southern Michigan. The experiment was conducted at the Kellogg Biological Station, Hickory Corners, Michigan, in small, open-top field chambers using the design of Curtis and Teeri (1992). Half of the chambers were maintained at approximately 700/zl liter -1 CO2 and the remainder at ambient CO2. Seeds were germinated at the CO2 level in which they were to be grown. Six seedlings from each maternal family were divided among four chambers at each CO2 treatment. Seedlings were individually planted in 2.5-liter pots filled with a mixture of locally derived sand and topsoil and placed in the chambers in early June 1992. All pots were watered daily as needed but received no additional nutrients. After 3 weeks' growth, all chambers were covered with a neutral density shade cloth to alleviate symptoms of high light stress and remained so shaded throughout the experiment. Net COz assimilation was measured on intact, newly expanded leaves after ca. 100 days' growth. Measurements were made in the laboratory with an ADC LCA2 photosynthesis system under saturating light intensifies using the same CO2 level under which the plants were grown. Plants were harvested after 127 days' growth. Sample size for each family was <-6 and 6 of
17
2.
the 24 families were excluded from analysis because two or fewer individuals had germinated in a particular COz environment. Data were analyzed with a fixed-effects model ANOVA with CO2 and population as main effects and family nested within population. In this analysis, significant CO2 treatment effects indicate a consistent difference between CO2 levels across populations and families. Significant population and family level effects indicate that the populations and families respond differently in terms of assimilation, which can contribute to an evolutionary response to elevated CO2. Family level differences in this design include genetic and maternal environmental effects. Thus, they indicate only the potential for a narrow-sense heritability.
We found significant genetic variation in life history and reproductive characteristics among families in both years, but markedly different responses to CO2 in 1993 compared to 1992 (Fig. 1). In 1992 there was no effect of CO2 on leaf area at bolting but there was a strong positive COz effect on flower production (Fig. 1A, P < 0.001). This stimula-
80 L_
~
~ ~
.-I
~
~
60
o
40
.1~
20
Ooe -000 I
0
I
I
_
J
I
1
I
I
I
I
I
I
1
2
3
4
5
6
7
8
I
|
I,
I
I
I
I
I
I
I
9 10 11 12 13 14 15 16 17
Paternal family Genotypic variation in life history and reproductive responses of grown at elevated (solid symbols) or ambient (open symbols) CO2. Experiments were conducted in two years, 1992 (A) and 1993 (B), using seeds derived from a c o m m o n parental population. Vertical bars indicated one SE, n - 6 (1992) or n = 8 (1993). *P < 0.06, **P < 0.01.
18
tion of reproduction was not experienced uniformly across genotypes, however, ranging from an 8% increase in family 4 (not significantly different from zero) to a 50% increase in family 5 (P < 0.01). In the 1993 experiment, growth at elevated CO2 resulted in significantly less leaf area at bolting (9% decrease overall, P < 0.01), but this effect again was unevenly distributed across families (Fig. 1B). Nine families showed no significant CO2 effect on leaf area while three families had significant reductions, with a maximum of 23% less leaf area at bolting in family 8 under elevated CO2. Unlike results from 1992, flower production showed a pattern similar to that of leaf area at bolting, with a significant decrease overall at high CO2 (P < 0.001) but with considerable variation among families in the magnitude of this response. Most families showed no significant CO2 effect on numbers of flowers while families 7 and 9 had 40 and 38% reductions, respectively. These results are consistent with previous work documenting heritable variation in fitness characters in (Mazer, 1987). They do not, however, provide a clear picture of how rising CO2 might act on those characters. In 1992, an exceptionally cool and cloudy summer in northern lower Michigan, elevated CO2 increased reproductive output and led to a change in rank order of seed production among five genotypes (Curtis 1994). The weather in 1993 was considerably warmer and sunnier than that in 1992, and while leaf area growth prior to bolting reflected the improved growth conditions, flower production was equivalent to (ambient CO2, overall) or lower than (elevated CO2, overall) that recorded in 1992. Although CO2 effects on flower production varied across genotypes, we found no evidence in 1993 for changes in rank order of genotypes in ambient compared to elevated CO2. also exhibited among family variation in responses to elevated CO2 (Fig. 2). The average stimulation of photosynthesis at elevated CO2 was 50% over ambient, which is comparable to that reported for other herbaceous species (Drake and Leadley, 1991). It is important to note that this average results from 11 families that show no significant stimulation and 8 families that show up to 75% greater assimilation at elevated levels compared to ambient CO2 levels. This is the first time that intraspecific variation in this response has been described. Plant size, as measured by whole-plant biomass at the end of the growing season, also varied significantly among these families. Fifteen families showed small or nonsignificant CO2 effects, two families showed significant positive responses, and one family showed a strong negative response. Plant size is well correlated with reproductive output under field conditions (Tonsor and Goodnight, 1996). These size differences among families are therefore likely to affect plant fitness. Klus (unpublished data) found that some families acclimated photosynthetically
2.
19
35 30
~
~'10
0
I
I
I
I
I
I
I
I
I
I
I
1
I
1
1
I
1
I
I
I
1
I
I
I
1
I
I
35 .o
ca.
15
oe , -
10 0
0
I
I
I
I
I
I
0
I
I
I
Genotypic variation in photosynthetic and growth responses of grown at elevated (solid symbols) and ambient (open symbols) CO2.Assimilation was measured in September after 100 days' growth, and plants were harvested after 127 days' growth. Seeds were collected from plants growing in either a lakeshore population (families 2-15) or a field population (families 31-44). Vertical bars indicated one SE, n = 3-6. *1 = P < 0.1, *P < 0.05, **P < 0.01.
resulting in indistinguishable assimilation rates u n d e r to elevated CO2, a m b i e n t a n d elevated CO2 by the e n d of the season, w h e r e a s o t h e r families showed n o sign of such acclimation. T h e s e differences in the p h o t o s y n t h e t i c r e s p o n s e to CO2 could h e l p explain differences in g r o w t h r e s p o n s e to CO2 a m o n g families. With biomass as well as with assimilation, the average r e s p o n s e of the species was positive, b u t this average r e s u l t e d f r o m a mix o f families, s o m e of which r e s p o n d e d negatively, s o m e n o t at all, a n d s o m e positively. has b e e n shown to r e s p o n d rapidly to b o t h n a t u r a l a n d artificial selection o n a r a n g e of ecologically i m p o r t a n t c h a r a c t e r s (Wu a n d Antonovics, 1976; Wolff a n d Van D e l d e n , 1989). T h u s for o u r p o p u l a t i o n s of P. the a m o n g family variation could serve as the basis for the evolution of i n c r e a s e d c a r b o n gain in h i g h CO2 e n v i r o n m e n t s . P r e d i c t i o n s of the of evolution u n d e r i n c r e a s e d a t m o s p h e r i c c a r b o n gain r e q u i r e m o r e d e t a i l e d quantitative g e n e t i c studies.
Clearly, shifts in the proportional representation of these or families within their respective populations would change our perceptions of the species' performance with a community. The extent to which families responding to elevated CO2 will increase in abundance in natural communities depends on two factors. The first factor is the extent to which family level differences in assimilation, allocation, and growth reflect genetic or maternal environmental components. Tonsor and Goodnight (1996) found that maternal family-based heritability estimates of assimilation rate under ambient CO2 conditions in approximately double the significant narrow-sense estimates for this species, but they saw no maternal effects on biomass. This means that there was both a nuclear genetic and a maternal component to the inheritance of assimilation rate, while size was affected only by nuclear genes. In Tonsor and Goodnight's study of nine traits, only those traits associated with physiology and leaf anatomy showed significant maternal effects on family means, suggesting an important maternal genetic role via cytoplasmic inheritance for those traits. If the significant maternal family variance we found in P. has the same genetic basis as was seen in Tonsor and Goodnight's study, both assimilation rate and biomass can be expected to evolve under elevated CO2. The second factor influencing the extent to which responding families will increase in abundance in natural communities is the degree to which the variation measured is maintained under field conditions. This degree is influenced by genotype-environment interactions, where competition is a part of the normal field environment. We have little basis for judgment here, and the only solution is field studies that combine measures of community-level and population genetic-level changes.
Our data demonstrate substantial among family intraspecific variation in response to elevated CO2. This variation is a necessary precondition to any evolutionary response to rising CO2 and could lead to large shifts in average population-level responses under elevated CO2 conditions. Evolutionary changes within populations could thus alter both ecosystem-level processes such as carbon storage and community-level interactions among species. It is important, therefore, to understand the magnitude and distribution of this variation as we refine our predictions of the ecological consequences of global climate change. We remain a long way from accurately predicting rates of evolutionary change within a community context. Although this chapter documents the best studies to date of the evolutionary potential of herbaceous plants under elevated CO2, much larger and more detailed studies of evolvability under
2.
21
field c o n d i t i o n s are necessary. A clearer u n d e r s t a n d i n g o f the physiological, allocational, a n d life-historical causal m e c h a n i s m s for intraspecific variation in r e s p o n s e to elevated CO2 are n e e d e d . T h e s e can c o n t r i b u t e to a m o r e effective use o f f u n d s d e v o t e d to the necessarily large-scale studies o f evolvability u n d e r elevated CO2. O n c e the m o s t a p p r o p r i a t e traits for study have b e e n clearly identified, c o m p a r i s o n s o f g e n o t y p e s using b o t h quantitative genetics a n d its r e c e n t hybrid with m o l e c u l a r genetics, quantitative t r a i t locus (QTL) identification, can be u s e d to describe the g e n e t i c a r c h i t e c t u r e u n d e r l y i n g this variation. U n d e r s t a n d i n g the g e n e t i c a r c h i t e c t u r e is a crucial step in assessing the evolvability o f the traits. Studies that i n c l u d e s i m u l t a n e o u s m e a s u r e s o f the c h a n g i n g g e n e t i c c o m p o s i t i o n o f species a n d the relative a b u n d a n c e s o f species within the c o m m u n i t y are especially i m p o r t a n t in p r e d i c t i n g w i t h i n - c o m m u n i t y adaptation versus successional c o m m u n i t y r e p l a c e m e n t as r e s p o n s e s to c h a n g e s in a t m o s p h e r i c chemistry. T h e i n t e g r a t i o n o f physiological, allocational, a n d g e n e t i c studies c o n d u c t e d in n a t u r a l p l a n t c o m m u n i t i e s c o u l d contribute substantially b o t h to o u r ability to p r e d i c t whole-ecosystem c a r b o n storage a n d to the i n t e g r a t i o n o f subdisciplines within the field o f ecology. I n t e g r a t e d studies o f species- a n d community-level r e s p o n s e s to elevated CO2 r e p r e s e n t b o t h a f o r m i d a b l e c h a l l e n g e a n d an exciting o p p o r t u n i t y for the unification o f ecology.
We thank Jon Ervin, Pam Woodruff, Nina Consolatti, Andrea Case, and Katherine Syverson for their assistance in the field and Allison Snow and Jim Teeri for critical scientific and material assistance. This research was supported with funds from the Ohio State University, an NSF predoctoral fellowship to D.J.K., NSF Grant CIR-9113598 to the Kellogg Biological Station, Michigan State University, the University of Michigan Biological Station, and the University of Michigan Project for the Integrated Study of Global Change.
Bazzaz, F. A. (1990). The response of natural ecosystems to the rising global C O 2 levels. 21, 167-196. Bradshaw, A. D. (1991). Genostasis and the limits to evolution. B 333, 289-305. Curtis, P. S., and Teeri, J. A. (1992). Seasonal responses of leaf gas exchange to elevated carbon dioxide in 22, 1320-1325. Curtis, P. S., Snow, A. A., and Miller, A. S. (1994). Genotype-specific effects of elevated CO2 on fecundity in wild radish 97, 100-105. Drake, B. G., and Leadley, P. W. (1991). Canopy photosynthesis of crops and native plant communities exposed to long-term elevated COs. 14, 853-860.
22 Fajer, E. D., Bowers, M. D., and Bazzaz, F. A. (1992). The effect of nutrients and enriched CO2 environments on production of carbon-based allelochemicals in A test of the carbon/nutrient balance hypothesis. 140, 707-723. Falconer, D. S. (1981). "Introduction to Quantitative Genetic Analysis." Longman, New York. Fisher, R. A. (1930). "The Genetical Theory of Natural Selection." Clarendon, Oxford. Garbutt, D., and Bazzaz, F. A. (1984). The effects of elevated CO2 on plants. III. Flower, fruit and seed production and abortion. 98, 433-446. Kalisz, S., and Teeri, J. A. (1986). Population-level variation in photosynthetic metabolism and growth in 67, 20-26. Klus, D.J., Kalisz, S., Tonsor, S.J., Curtis, P. S., and Teeri, J. A. Intraspecific variable responses to elevated atmospheric CO2: Resource partitioning in above- and below-ground tissues in L., unpublished. Mazer, S.J. (1987). The quantitative genetics of life history and ftness components in L. (Brassicaceae): Ecological and evolutionary consequences of seed-weight variation. 130, 891-914. Mooney, H. A., Drake, B. G., Luxmoore, R. J., Oechel, W. C., and Pitelka, L. F. (1991). Predicting ecosystem responses to elevated CO2 concentrations. 41, 96-104. Teramura, A. H., and Strain, B. R. (1979). Localized populational differences in the photosynthetic response to temperature and irradiance in 57, 25592563. Tonsor, S. J., and Goodnight, C.J. (1996). Testing the effect of mating structure on the partitioning of phenotypic variance in in press. Wolff, K., and Van Delden, W. (1989). Genetic analysis of ecologically relevant morphological variability in Response and correlated response to bidirectional 62, 153-160. selection for leaf angle. Wu, L., and Antonovics, J. (1976). Experimental ecological genetics in II. Lead tolerance in and from a roadside. 57, 205-208. Wulff, R. D., and Alexander, H. M. (1985). Intraspecific variation in the response to COs enrichment in seeds and seedlings of 66, 458-460. Ziska, L. H., and Teramura, A. H. (1992). Intraspecific variation in the response of rice (Oryza to increased COz- photosynthetic, biomass and reproductive characteristics. 84, 269-276.
3 Selective Responses to Global Change: Experimental Results on Brassicajuncea (L.) Czern.
The climate change forecasted for the next century is predicted to occur at an unprecedented rate (Huntley, 1991). Projections are an increase of ---4~ in global mean temperature over some 50 years. The rates of plant migration necessary to track such changes in climate are higher than the maximum rates of migration estimated from the Quarternary climate change (Huntley, 1991). Therefore, it is unlikely that plants will successfully migrate in response to global change, hence the threat of massive extinction. Holt (1990) and Bradshaw and McNeilly (1991) suggested that evolutionary changes resulting in genotypes adapted to the new climatic environment might prevent the extinction of some species. Bradshaw and McNeilly (1991) also suggested that, in the absence of genetic changes, adaptive plasticity could provide an alternative strategy. Yet the potential for evolutionary responses to global change has received little attention to date. The experiment discussed in this chapter was designed to assess the evolutionary consequences of global change. We selected a wild mustard species, (L.) Czern., over 7 generations, for growth under conditions simulating global change. Throughout this chapter we refer to the long-term growth environments as Selection treatments. Specifically, we want to quantify the relative importance of genetic and plastic responses to Selection. was chosen as a model species because it is self-incompatible, which prevents excessive inbreeding (Williams and Hill, 1986), and it has a short life cycle, which, under our experimental conditions, lasted 50 days. Copyright 9 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.
Two replicate lines of 70 plants were selected u n d e r conditions simulating either the present atmosphere ( " C o n t r o l " treatment) or, by gradually increasing CO2 and temperature, global change ("Predicted" treatment). In the Control treament, conditions were maintained at the initial temperature (22~176 and CO2 conditions ( 3 7 0 / x l . liter -1) during the 7 generations of Selection. In contrast, conditions in the Predicted treatment changed gradually from one generation to the next (Table I). During Selection, daily m e a n temperature increased by 3.6~ while CO2 approximately doubled. Degree-days were used to model the desired changes in temperature over time, by combining increased mean temperatures and more frequent heat stress. This design accommodates the prediction of an increase in length and frequency of heat waves with the greenhouse effect (Mearns 1984). To " m i m i c " Natural Selection, we selected for high fecundity in both treatments. After the plants were harvested and the siliques dried, the 18 most productive plants within each selection line, based on total silique mass, were selected and their seeds used to sow the next generation. Consequently, plants from the Control and Predicted treaments were subjected to identical fecundity selection and inbreeding. Overall, inbreeding ranged between 2.3 and 3.9% and the possible significance of inbreeding and inbreeding depression for our results is discussed elsewhere (Potvin and Tousignant, 1996). The choice of total silique biomass as a fitness estimate was validated by calculating genetic correlations a m o n g fitness-related traits at the start of Selection, based on Rose and Charlesworth (1981). There were positive genetic correlations between, on one hand, total silique mass and, on the other hand, total seed n u m b e r and total seed mass (ranging from 0.812 to 0.995). To identify evolutionary responses after Selection, we carried out a reciprocal transplant experiment. Plants selected u n d e r Predicted conditions
Generation number
CO2 level (/xl 9liter-1)
0 1 2 3 4 5 6 7
370 410 450 490 530 570 610 650
Day/night temp. (~ 22.0/16.0 22.0/16.0 22.4/16.4 22.8/16.8 23.2/17.2 23.6/17.6 24.0/18.0 24.4/18.4
No. of days of heat stress (32~176
Daily mean temp.
0 2 3 4 5 6 7 8
20.0 20.4 21.0 21.6 22.1 22.6 23.1 23.6
3.
on
25
were grown in both Control and Predicted conditions and vice versa. The chambers were p r o g r a m m e d to either one of the conditions prevailing at the 7th generation. This approach has been used in the field and in artificial selection experiments to test for genetic adaptation (Jain and Bradshaw, 1966; Maxon Smith, 1977; Agren and Schemske, 1993). The underlying assumption is that a plant adapted to its environment will perform better in that environment than in a new one. In other words, given genetic adaptation, home transplants should do better than foreign ones. At first harvest we examined growth during the prereproductive life stage. The analysis of variance (ANOVA) indicated a significant main effect of Selection and a significant Selection by Environment interaction (Table II). Overall, the vegetative biomass was higher for plants selected in the Predicted regime. examination of the results indicated that the vegetative biomass of plants of selected and grown in the Predicted environment was significantly higher than that of plants grown under any other (Fig. 1A). At the end of the reproductive phase, vegetative biomass was significantly lower in the Predicted plants and the adaptive response was lost (Fig. 1B). The response of silique biomass to the Predicted growth conditions was highly significant but their was no effect of the Selection treatment. Overall, silique biomass was 3.45 g under Control conditions and 0.95 g under Predicted ones. Direct exposure of Control plants to Predicted conditions (CP/CC) reduced silique biomass from 3.2 to 1.0 g (Fig. 1C). Silique biomass of the Predicted plants likewise was reduced by growth in the Predicted environment with no sign of genetic adaptation. The analysis of the reciprocal transplant answered our first main question. Vegetative biomass responded to the stimulated global change mainly through genetic changes. Between 15 and 18% of the total variation in vegetative biomass was due to the effect of Selection. Furthermore, in the prereproductive growth period, these genetic changes were adaptive. Conversely, the response of silique biomass was plastic and maladaptive. The reproductive output of an annual plant without a seed bank such as provides a direct estimate of fitness (Primack and Kang, 1989). Consequently, the second part of this chapter will further examine the responses of silique biomass to the gradual changes in environment imposed in the Predicted treatment. Despite a fecundity selection, silique biomass decreased significantly under both treatments during the 7 generations of Selection (Fig. 2). This reduction in silique biomass was not accompanied by a reduction in vegetative biomass. The decrease in total silique biomass was significantly faster for the Predicted selection lines, overall -0.443 g 9 generation -1, than for the Control lines, overall -0.334 g 9 generation -1 (Table III). A closer look at the data suggests that whereas silique biomass decreased continuously in the Predicted environment, it
26
6
6
I
-r-
I
-T-
I
(A) Vegetative biomass of at the prereproductive (Day 21) and (B) late reproductive (Day 49) harvests, and (C) pod biomass at maturity (Day 49). Vegetative biomass was the mass of stems, leaves, and roots. Plants of both selection lines were pooled and sample size is indicated in parentheses. Significant differences among groups using Tukey's HSD, P < 0.05, are indicated by different letters. Differences in sample sizes were due to postgermination mortality. CC: "Control" lines grown under "Control" conditions; CP: "Control" lines grown under "Predicted" conditions; PP: "Predicted" lines grown under "Predicted" conditions; PC: "Predicted" lines grown under "Control" conditions. (Modified from Potvin and Tousignant, 1996.)
3.
on
27
0 0
(A) Mean pod biomass of over seven generations of selection in the "Control" or "Predicted" conditions and (B) mean divergence of the "Predicted" lines from the average value of "Control" ones. Data were not available for the first generation. In (A) each value represents the mean of 70 plants. CI: "Control" line 1; C2: "Control" line 2; PI: "Predicted" line 1; P2: "Predicted" line 2.
stabilized at g e n e r a t i o n s 3 - 4 for the C o n t r o l e n v i r o n m e n t . This e x p o n e n t i a l d e c r e a s e in silique biomass can be best c a p t u r e d using n o n l i n e a r r a t h e r t h a n linear regression, as shown by the h i g h e s t r 2 values (Table III). T h e e x p o n e n t i a l e q u a t i o n p r e d i c t e d asymptotic r e p r o d u c t i v e biomass of 2.02 a n d 2.08 g for the two C o n t r o l selection lines. T h e r e was n o clear a m e l i o r a t i o n in curve fitting w h e t h e r linear or nonlinear r e g r e s s i o n was u s e d to analyze the silique biomass of P r e d i c t e d plants. F r o m the o b s e r v e d c h a n g e in silique biomass t h r o u g h time, we p r e d i c t e d the o u t c o m e of Selection in the P r e d i c t e d e n v i r o n m e n t . For b o t h selection lines, we u s e d linear r e g r e s s i o n to p r e d i c t the cumulative selection differential, AS, r e d u c i n g slique biomass to zero. S is d e f i n e d for e a c h g e n e r a t i o n as the d i f f e r e n c e existing b e t w e e n the m e a n p a r e n t a l value for a trait a n d the m e a n r e s p o n s e of the offsprings for that trait (Falconer, 1989). A s e c o n d series of r e g r e s s i o n was u s e d to p r e d i c t the n u m b e r of g e n e r a t i o n s n e e d e d to r e a c h such AS value. T h e cumulative selection differential r e q u i r e d to r e d u c e re-
28
Sources (A) Treat. Env. Line(Treat.) Treat.XEnv. Env. XLi(Tre.)
Sums of squares
0.726 0.241 0.161 0.884 0.169
1 1 2 1 2
9.26 3.07 1.04 11.28 1.08
0.0032 0.08 0.36 0.0012 0.35
8.360 1.226 0.176 0.673 0.240
1 1 2 1 2
13.44 1.97 0.14 1.08 0.19
0.0005 0.17 0.87 0.30 0.83
0.389 114.1 0.498 0.647 2.316
1 1 2 1 2
0.83 243.3 0.53 1.38 2.47
0.37 0.0001 0.60 0.24 0.09
(8) Treat. Env. Line(Treat.) Treat. XEnv. Env. •
(c) Treat. Env. Line(Treat.) Treat. XEnv. Env. XLi(Tre.)
productive growth to zero was 8.7 and 9.4 for the two selection lines. Assuming a constant rate of increase in ASthrough time, complete reproductive failure would occur in the Predicted environment within 9.8-10.1 generations. Therefore, the Predicted environment would lead to local extinction of B. with an additional 3 generations of Selection.
Selection lines "Current" 1 "Current" 2 "Predicted" 1 "Predicted" 2
L i n e a r regression
N o n l i n e a r regression
3.74-0.311 g e n r~ = 0.721 a 3.883-0.334 gen r~ = 0.626 a 4.093-0.434 gen r~ = 0.813 b 4.281-0.446 gen r~ = 0.882 b
2.045 + e x p ( - 0 . 3 2 9 * g e n ) r2 -- 0.933 2.084 + e x p ( - 0 . 3 3 7 * g e n ) r~ = 0.907 1.696 + e x p ( - 0 . 3 0 6 * g e n ) r~ --- 0.847 1.81 + e x p ( - 0 . 2 9 2 * g e n ) r~ = 0.866
For the linear regressions, slopes followed by a different letter are statistically different with P < 0.05. a
3.
on
was apparently unable to adjust to the simulated global change either plastically or evolutionarily. Could the model of Kingsolver (Chapter 1) explain this absence of evolutionary response? In our experiment, the CO2 concentrations increased by 40 /~1 9 liter -1 between two successive generations, while the mean daily temperature rose by 0.4-0.6~ The environmental changes were clearly well within the natural variation in CO2 and in temperature encountered by The maximum high temperature, 32/26~ was also within the temperature range encountered in the field (Polowick and Sawhney, 1988). Our experimental situation might thus have been analogous to that of the "thermal generalists" (see Kingsolver, Chapter 1). Each generation was sown with seeds from 18 mothers, which represents 25% of the original population. Kingsolver's model suggests that this intensity of selection might not have been strong enough to sustain the rate of changes in environmental conditions. A lag between the population tolerance and the environmental conditions was thus created. As this lag increased, we observed a decrease in silique biomass, hence fitness, that could lead to the future extinction of the population. The evolutionary model of Kingsolver thus provides an appropriate theoretical framework to examine our experimental results. In conclusion, we would like to examine our results in a broader evolutionary perspective. We wish to make three main claims: first, the directional and gradual changes in the environment did exert a selective pressure on Second, evolutionary responses to such gradual change in CO2 concentration and climate are possible and were evidenced by the genetic changes in vegetative biomass. Finally, the magnitude and direction of those changes will clearly be trait-specific. Vegetative biomass responded mainly by genetic changes whereas silique biomass showed only plastic responses. We believe that our understanding of the evolutionary consequences of global change cannot be predicted from an abrupt change in environmental conditions. Our experimental results support the idea that ecological information is not sufficient to predict the evolutionary responses to climate change (see Kingsolver, Chapter 1). More specifically, for B. any inferences of the evolutionary consequences of global change drawn from the response of vegetative biomass alone would have been misleading. A better understanding of the evolutionary consequences of global change might come from the analysis of the heritabilities and genetic correlation structure of fitness-related traits.
During the 18 months of Selection, we had continuous support from M. Romer and C. Cooney from the McGill Phytotron. We thank M. Beaudet and D. Cantin for their assistance during
seemingly endless seed counting and weighing. Discussions with Drs. S. Arnold, Y. Carri/~re, and M. Johnson helped to clarify the evolutionary aspects of this work R. G. Shaw has been most helpful in disentangling problems in the quantitative genetic approach. This research was funded by a "Nouveaux chercheurs" research grant, Fonds FCAR, Quebec, to C.P.D.T. acknowledges a NSERC postgraduate fellowship.
Agren, J., and Schemske, D. W. (1993). The cost of defense against herbivores: An experimental study of trichome production in 141, 338-350. Bradshaw, A. D., and McNeilly, T. (1991). Evolutionary response to global climatic change. 67 (Suppl. 1), 5-14. Falconer, D. S. (1989). "Introduction to Quantitative Genetics." Wiley, New York. Holt, R. D. (1990). The microevolutionary consequences of climate change. 5, 311-315. Huntley, B. (1991). How plants respond to climatic change: Migration rates, individualism, and the consequences for plant communities. 67(Suppl. 1), 15-22. Jain, S. K., and Bradshaw, A. D. (1966). Evolutionary divergence among adjacent plant populations. 21, 407-441. Maxon Smith, J. W. (1977). Selection for response to CO2-enrichment in glasshouse lettuce. 17, 15-22. Mearns L. O., Katz R. W., and Schneider S. H. (1984). Extreme high-temperature events: Changes in their probabilities with changes in mean temperature. 23, 1601-1613. Polowick, P. L., and Sawhney, V. K. (1988). High temperature induced male and female sterility in Canola. L.). 62, 83-86. Potvin, C., and Tousignant, D. Evolutionary consequences of simulated global change: Genetic adaptation or adaptive phenotypic plasticity. in press. Primack, R., and Kang, H. (1989). Measuring fitness and natural selection in wild plant populations. 20, 367-396. Rose, M. R., and Charlesworth, B. (1981). Genetics of life history in I. Sib analysis of adult females. 97, 173-186. Williams P. H., and Hill, C. B. (1986). Rapid-cycling populations of 232, 13851389.
Genetic Variation in the Response of Plant Populations to Elevated in a NutrientPoor, Calcareous Grassland
Results from models and experiments on the effects of rising C O 2 differ in that the models often predict large effects on ecosystems (Woodward, 1992; World Conservation Monitoring Centre, 1992), whereas only small effects at the ecosystem level have been observed in experiments (e.g., Billings 1984; Grulke 1990; Ktrner and Arnone, 1992, but see Curtis 1989). In particular, though expected from the response of single species grown under high CO2, shifts in species composition within a community are rare and appear to be highly dependent on characteristics of the ecosystem considered, such as the availability of nutrients and light and the intensity of competition (Bazzaz, 1990). In a study of a natural tallgrass prairie, Owensby (1993) found essentially no change in the species composition and basal cover of the major warm-season dominants in response to CO2 treatments over a 3-year period. The time scale of the experiments might, however, often be too short for shifts in species composition to be observed (cf. Milchunas and Lauenroth, 1995). Ktrner and Arnone (1992), for example, found no significant response of the total biomass of an artificial ecosystem to elevated CO2 but pointed out that the slight observed biomass increase under elevated CO2 and the harmful levels of starch in leaves might alter the dominance relationships in this community in the long term. 31
Copyright 9 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.
32
We suggest that both in models and in experiments there are implicit misconceptions about the nature of ecosystems because they are thought to contain assemblages of species or of individuals of species that have only mean properties. Variation within species is assumed to be negligible, interactions among species are rarely taken into account, and evolution is not allowed. A general feature of hierarchically organized systems is that lower level processes express more variation than higher level processes, because constraints are more stringent at the higher level (e.g., total biomass versus component-species biomass in an ecosystem; see O'Neill 1986; Odum, 1989). However, even though the lower level dynamics in an ecosystem such as individual and population processes might not greatly influence p h e n o m e n a observed at the ecosystem level in the short term, they cannot be neglected to understand the mechanisms that could lead to ecosystem-level modifications in the longer term. In this chapter, we shall focus on the dynamics of genetic variation within species and discuss its consequences for predicting the response of an ecosystem to an increase in CO2. Our aim is to demonstrate that lower level dynamics must be considered if CO2 research is to yield reliable predictions.
A. Genetic Variation and Plasticity Variation in natural populations is caused by several factors, including plastic response to variations in environmental conditions, genetic differences among individuals, and developmental noise. Phenotypic plasticity is environmentally induced phenotypic change that occurs within an organism's lifetime (Bradshaw, 1965; Stearns, 1989). Evidence suggests much of this environmentally induced variation is selectively advantageous (see, e.g., Marshall 1986) so that plasticity as a trait itself could be sensitive to selection (de Jong, 1990a,b). Moreover, genetic variation for plasticity in fitness-related traits has been observed in many plants (e.g., Scheiner and Goodnight, 1984; Silander, 1985; Schmid and Weiner, 1993; Stratton, 1994; Cheplick, 1995). Because of their open organization and their inability to move when conditions change, plants are thought to adjust to a change in their environm e n t largely by phenotypic plasticity, whereas mobile animals usually respond by behavior (e.g., migration) (Schmid, 1992). To illustrate the importance of understanding genetic variation, phenotypic plasticity, and genetic variation in phenotypic plasticity within a species to predict the effect of increasing CO2, it is convenient to start with a theoretical example. The range and mean value of conditions under which an individual can grow vary among species and even within species among genotypes. Two parameters can describe the fitness response or tolerance curve (Lynch and Gabriel, 1987) of a genotype to a given environmental factor: (1) the value of the environmental factor at which the fitness reaches the maximum
4.
33
for the genotype and (2) the range of values of this factor over which the genotype has positive fitness, i.e., over which it can grow, survive, and reproduce. T h e first p a r a m e t e r may be called the " n i c h e position" a n d the second one the " n i c h e width" of the genotype (Roughgarden, 1974). A genotype (or species) with a wide niche is called a "generalist," whereas a genotype (or species) with a narrow niche is called a "specialist." For an individual genotype the phenotypic expression of a quantitative character across an environmental factor is called a " r e a c t i o n n o r m " (Woltereck, 1909). It can be r e p r e s e n t e d by a functional response of the genotype to the environmental factor (de Jong, 1995). Now consider a set of genotypes with the same niche positions but different niche widths across an e n v i r o n m e n t a l gradient (Fig. 1). U n d e r the assumption that the fitness response (shown in a g r a p h of the niche) is related to the response of some resource-use character such as rooting d e p t h (shown in a reaction n o r m ) in each genotype, the generalist genotypes in the set are expected to have a m o r e plastic response in the resource-use character than the specialist genotypes. T h a t is, character values for the resource-use character, but not for fitness, should differ m o r e a m o n g contrasting environments in the generalists than in the specialists. Figure 1 presents the reaction n o r m s of different genotypes and the resulting fitness responses across an arbitrary e n v i r o n m e n t a l gradient. An ideally plastic genotype would maintain a high fitness in every e n v i r o n m e n t by adjusting the resource-use character along an optimal reaction n o r m (see also Schmid, 1992). However, genetic, physiological, and mechanical constraints likely exist, possibly in the form of trade-offs between traits, which restrict plasticity. Consequently, reaction n o r m s will usually diverge from the line of ideal plasticity and therefore the fitness response of any genotype will be limited to a certain range of e n v i r o n m e n t a l values beyond
Reactionnorms for a hypothetical resource-use character across an arbitrary environmental gradient and resulting niche responses or tolerance curves of three genotypes with similar niche position but different niche width. The ideally plastic reaction norm is indicated by the thin, dashed line with slope 1 in the left graph; the genotype with the steepest reaction norm is almost ideally plastic; and the one with the horizontal reaction norm is not plastic at all.
34 which conditions become too extreme for growth, survival, and reproduction. Returning to the generalist-specialist comparison, the generalist genotypes are assumed to have higher fitness than the specialist genotypes over most environmental conditions except those that are optimal for the specialist. This is because the ability to perform a broader variety of tasks through plasticity in resource-use characters may be possible only at the cost of reduced maximum performance. The view of a cost of plasticity has, however, been challenged by results from some experiments in which individuals that performed better in an "optimal" environment also performed better in other environments (Huey and Hertz, 1984). We shall assume for the purpose of the demonstration that all the genotypes have the same fitness under optimal conditions, and different assumptions would not change the conclusions drawn from this example. Let us now consider how a generalist and a specialist genotype will respond in a short-term experiment during which replicate individuals of the genotypes are grown at either ambient or elevated COp. Assuming that plants are adapted to the COp concentration they experience today or did some decades earlier, i.e., that their niche positions are fixed, and given, by definition, that generalist genotypes exhibit steeper reaction norms in traits that relate to CO2 uptake, these generalist genotypes should perform better under high COp than specialist genotypes.
B. Response at the Species Level As with genotypes, a generalist species has a broad fitness response and a specialist species has a narrow fitness response across an environmental gradient, such as some resource axis (Roughgarden, 1974). Specialist species always consist of a small n u m b e r of specialist genotypes with a similar niche position. Generalist species, however, may be either polymorphic and consist of specialist genotypes differing in niche position or monomorphic and consist of generalist genotypes with similar niche positions, or generalist species may represent some intermediate condition (Fig. 2). To stress the relevance of studying genetic variation in the response to increasing CO2, we shall now compare the results expected from shortterm experiments (in which plants are grown under elevated CO2) and the long-term evolution of specialist and generalist species differing in genetic variation a n d / o r plasticity. For this purpose it is convenient to characterize a species by the variation among its constituent genotypes in niche position and niche width. Six possible combinations are described in Fig. 3. For each of them, we present first the plasticity of the genotypes in a resource-use character (here related to CO2 uptake) and the distribution of niche positions (optimal COp concentration), then the performance of the species in a short-term experiment in which it is grown in monoculture under ambient and elevated COp, and, finally, its expected perfor-
4.
35
Environmental gradient Figure 2 Performance of two hypothetical generalist species (a, b) and one hypothetical specialist species (c) along an arbitrary environment gradient. Thick lines indicate the niches of the species; thin lines indicate niches of genotypes. The generalist species (a) is composed of a few similar generalist genotypes, generalist species (b) is composed of specialist genotypes with different niche positions, and the specialist species (c) is composed of a few similar specialist genotypes.
mance if it were allowed to evolve over a longer time period with continuously increasing CO2 concentrations. The continuous increase shall be slow enough for shifts in the frequency distribution of genotypes to keep pace, and the shifts shall come about only by selection and recombination without the appearance of new favorable mutations. The six species depicted in Fig. 3 represent the following cases: (A) low genetic variation for optimum CO2, no plasticity, no genetic variation in plasticity; (B) low genetic variation for optimum C 0 2 , high plasticity, no genetic variation in plasticity; (C) no genetic variation for optimum CO2, some genotypes with high plasticity, genetic variation in plasticity; (D) high genetic variation for optimum CO2, no plasticity, no genetic variation in plasticity; (E) high genetic variation for optimum CO2, intermediate plasticity, no genetic variation in plasticity; and (F) high genetic variation for optimum CO2, some genotypes with high plasticity, genetic variation in plasticity. If selection-recombination cycles are not allowed in the short term, the genetic composition of a population cannot change and only species containing generalist genotypes with high plasticity in CO2-resource use will be able to adjust to the new conditions and maintain a high performance (Fig. 3B and, to a lesser extent, Fig. 3E). This will be the case in shortterm experiments with suddenly increased CO2. With a slow and continuous increase in CO2, however, it is the species containing different specialist genotypes, i.e., the species with a large genetic variation within populations in CO2-resource use, that will maintain a high performance by an evolutionary adjustment of its mean response (Figs. 3 D - F ) . Obviously, many simplifications must be made in the theoretical example. In reality, there are other niche dimensions or characters in addition to C O 2 o r CO2-optimum for growth/reproduction, respectively, that will be
36
Reaction norms for a hypothetical COz-use character of six species differing in genetic variation and plasticity and their expected performance in ambient CO2 (amb.), in elevated CO2 in a short-term experiment (high ST), and in elevated CO2 if the increase is slow enough for evolution to happen (high LT). The numbers at the top of the bars in the
4.
37
modified with future climatic change, and interactions among the different characters might prove important. In particular, the existence of genetic variation in response to CO2 might not be relevant if it is negatively correlated with other traits (e.g., if there were a negative correlation between optimum temperature and optimum CO2 concentration for photosynthesis). Interactions among species also are to be taken into consideration. CO2 might, for example, change the relationship between higher plants and mycorrhiza (see Sanders, Chapter 17). Three predictions can, however, be derived from the theoretical example. First, the best "phytometers" for bioindication in global change should be specialists with low plasticity and low genetic variation in resource-use characters (Fig. 3A). Because of this low plasticity and low genetic variation, specialist species should indeed be the first ones to be outcompeted when environmental conditions change. An interesting problem that remains to be investigated is whether a specialist with respect to one environmental factor is also a specialist with respect to other environmental factors. In the first case it would be possible to use indicator species for nutrient-poor grasslands such as (described below) also to detect changes in atmospheric composition. Second, in short-term experiments on the effect of CO2on plants, generalist species consisting of similar generalist genotypes, i.e., with high plasticity (Fig. 3B), should do better than generalist species consisting of different specialist genotypes (Fig. 3D) because the first can adjust faster to the new conditions than the second. Consequently, it is impossible to predict the effect of elevated CO2 on the species composition of a community from short-term experiments, unless it is known whether the differences in the specific responses are due to phenotypic plasticity or genetic variation. Third, if the conditions change slowly enough for the genetically variable species to evolve, we may in the long term expect a plastic species to be outcompeted by the better adapted genotypes of a genetically variable species. Under this scenario, environmental change will inevitably lead to a reduction in genetic and species diversity in an ecosystem. These points stress the necessity to consider the dynamics of the genetic composition of the species in order to predict the effect of environmental changes in communities of species (Billington and Pelham, 1991; Bradshaw and McNeilly, 1991).
right set of graphs give the ranking of the species if their performance is compared under the same conditions (i.e., amb., high ST, or high LT). As in Fig. 1, the dashed line represents an ideally plastic genotype. The left mark on the x-axis represents the ambient level of CO2 and the right mark represents an elevated level of CO2.The width of the shaded area visualizes the amount of genetic variation for optimum CO2 in the species; see text for interpretation.
38
Therefore the questions we wished to address experimentally were the following: 1. How do generalists and specialists phenotypically respond to elevated CO2? 2. Is there genetic variation within species in the response to elevated CO2? 3. What proportion of the variation in phenotypic characters has genetic causes (heritability) ? In the following sections we present results obtained for a model pair of a specialist and a generalist species. (specialist) and (generalist) are two closely related grassland perennials. Known genotypes of these two species were grown in a CO2 field experiment in a nutrient-poor, calcareous grassland south of Basel, Switzerland (see Leadley and K6rner, Chapter 11).
In a research program a field experiment was set up to investigate the effects of varying plant diversity on ecosystem responses to elevated CO2 (K6rner, 1995; Leadley and K6rner, Chapter 11; see also Weber and Schmid, 1995). The aim was to provide ecologists of different expertise with properly replicated and maintained installations in which the influence of elevated CO2, competition, and three biodiversity treatments on plant growth and ecosystem processes could be tested. As a compromise between the conflicting demands of realism and precision (see Levins, 1968), the experiment was carried out under natural field conditions using species and homogenized soil from the site but with controlled plantings in three biodiversity treatments (5-, 12-, 31-species plots). The ambient and elevated CO2 treatments were applied using screen-aided CO2 enrichment (SACE, see K6rner 1995; Leadley and K6rner, Chapter 11). A number of test plants were grown in a competition-free microenvironment in tubes within the planted plots. For as many species as possible, defined genetic material was propagated in the glasshouse and similar-sized seedlings or cuttings were transplanted to the field plots according to a predefined random arrangement. The species according to our a priori information from the floristic literature (e.g., Hess 1972; Landolt, 1977) represent an ideal generalist-specialist pair, and this classification could be supported in nutrient-gradient experiments with added competition (Birrer and Schmid, unpublished data). occurs over a broad range of environ-
39
4.
ments from nutrient rich to relatively nutrient poor, from dry to wet, from grassland to disturbed sites, and even to road margins within forests; P. is restricted to nutrient-poor, dry grasslands and has received legal status as indicator species for this community type in Switzerland. Twenty-eight genotypes of each species were collected at random from selected patches at the field site. To assess genetic variation and treatment-by-genotype interactions, each genotype was grown in competition-flee tubes and in all the treatment combinations that contained transplanted communities (3 diversity levels • 2 CO2 levels; Table I). Measurements of fitness-related characters were taken at the start of the experiment in the spring and repeated in 2-3 months. The analysis presented was carried out on the second set of measurements.
A. Species Responses No main effects of C O 2 o n the investigated characters could be detected (for n u m b e r of leaves and cumulative leaf length see Table II). The performance of the plants as indicated by the measured characters declined with increasing biodiversity from 5-species plots to 12-species plots to 31-species plots. However, because each of these three biodiversity levels was represented by a particular (replicated) mixture, this effect could have been due to the presence of a particular strong competitor in the m e d i u m and an even stronger one in the high diversity plots (see Naeem 1994). Indeed, one certain effect of high diversity is the increased chance of always having a particular species with high performance "at h a n d " in any particular environment. Large microgeographic variation was reflected in highly significant plot effects (Table II). This indicated the importance of using proper replication at the plot level because the effects of CO2 and biodiversity had to be compared with the variation among plots to obtain. significance tests (see Hurlbert, 1984).
CO 2
level
Ambient Elevated (---600 ppm)
Diversity treatment
Competitionfree tubes
5 species
12 species
31 species
12 plots 12 plots
4 plots 4 plots
4 plots 4 plots
4 plots 4 plots
40
Table II Results from Analyses of Variance of the Measured Quantitative Characters a N u m b e r of leaves (In) Source of variation
SS c
Cumulative leaf length (In)
pd
SS
P
(A) Without competition Block CO2 Plot Species CO2 • species
1 1 21 1 1
Block CO2 Biodiversity CO2 • biodiv. Plot Species CO~ • species Biodiversity • species CO2 • biodiv. • species
1 1 2 2 17 1 1 2 2
3.838 0.040 7.261 9.862 0.957
<0.001 >0.1 <0.001 <0.001 <0.05
1 1 21 1 1
3.724 0.078 19.352 15.218 0.587
<0.1 >0.1 <0.001 <0.001 >0.1
(B) With competition 28.907 <0.001 1.616 >0.1 1.595 >0.1 1.872 >0.1 9.372 <0.001 7.092 <0.001 0.005 >0.1 0.442 >0.1 1.229 <0.05
1 1 2 2 17 1 1 2 2
18.972 1.510 10.753 8.638 8.705 60.605 0.138 0.094 1.982
<0.001 >0.1 <0.01 <0.05 <0.05 <0.001 >0.1 >0.1 <0.1
Only the treatment and species effects are shown, the analysis of genetic variation within species is presented in Table 3. o degrees of freedom. cSS, sum of squares. dp, error probability. a
In contrast to the absence of main effects of CO2 there were significant interactions between this factor and biodiversity (cumulative leaf length), and among these two factors and species (number of leaves) (Table IIB). Thus, whereas the different levels of biodiversity could not be distinguished at ambient CO2, they had a differential effect on cumulative leaf length, presumably a good indicator of plant fitness, under elevated CO2: it was highest in the low-biodiversity treatment at elevated CO2 and lowest in the high-biodiversity treatment at elevated CO2 (Fig. 4A). The result demonstrates that the effect of diversity under elevated CO2 cannot be predicted from studying the effect at ambient CO2. The CO2-by-biodiversity interaction possibly reflected increased competition due to a positive response to elevated CO2 of some other species in the medium-biodiversity and even more so in the high-biodiversity treatments. Again it may be argued that in a larger set of species there is a greater probability that one species benefits more than proportionately from increasing CO2 concentrations but, again, this would neither be trivial nor would it invalidate the cautionary statement concerning predictions about biodiversity effects under elevated CO2. It
4. Variation in the Response of Plant Populations
41
-
~
loo
.~ ~_, g E~
4oo
~-"--~-'~-....~
\ \ \
P. grandiflora ~,.
Z
|
\ \
Cumulative leaf length (a) and leaf number (b) of the two Prunella species grown without competition or with competition at different levels of biodiversity and in ambient and elevated C02. Bar in lower left of upper graph in both (a) and (b) indicates one standard error of differences among means.
could also be argued that competition was greater under high diversity because there is reduced opportunity for escape by plasdc niche shifts. Prunella vulgaris responded negadvely to elevated C02 whereas P. grandiflora was not affected. Thus, the generalist species surprisingly was at
42
disadvantage compared with the specialist species at elevated CO2 in the competition-free tubes (significant 2-way interaction CO2-byspecies for number of leaves in Table IIA), as well as in medium- and highbiodiversity plots (significant 3-way interaction CO2-by-biodiversity-byspecies for number of leaves in Table IIB; see Fig. 4B). This is evidence against a high plasticity with regard to CO2 in and indicates either that this species is a generalist because of high genetic variability or that, with regard to CO2 as resource axis, the generalist-specialist distinction does not hold for the two species.
B. Genotypic Responses Significant differences among genotypes in a given environment or in response to environmental changes indicate the existence of genetic variation (broad-sense heritability). The amount of this genetic variation is the most important characteristic of a species in predicting its evolutionary potential for adaptation to environmental change because the expected shift in the mean of a character is directly proportional to its heritability (Bradshaw, 1984). We estimated the broad-sense heritabilities (H 2) of the investigated quantitative characters under ambient CO2, under elevated CO2, and over both levels pooled for plants growing both in the competition-free tubes (Table IliA) and in the diversity treatments (Table IIIB). The estimates for ambient and elevated CO2 were calculated from mixed-model analyses of variance as the quotient of the variance component due to differences among genotypes (s2) and the sum of this component and the residual variance (s2es) within plots (Falconer, 1989): 2 H(1,2)
=
s~ + S~Res
The pooled heritability was calculated as the quotient of the variance component due to differences among genotypes (s2) and the sum of this component, the component for the genotype-by-CO2 interaction (sZxco2), and the residual variance (S2Res)within plots: H~ / =
s~ s~ + s G2x C O
2
+ S2R~s"
A significant genotype-by-CO2 interaction indicates that there is heritability in the response tO elevated CO2 and that the slopes of the responses of the genotypes (reaction norms) are significantly nonparallel. An estimate of the heritability of this response to CO2, i.e., of the heritability of plasticity, can be obtained in an analogous way (see Scheiner and Lyman, 1989):
43
(ml))
(m2))
(m3))
(m4))
(0.05) b (0.11) (0.00)
(A) Without competition Ramet length (ln) Number of leaves (In) Cumulative leaf length (ln)
0.58 (0.40) (0.28)
0.52 (0.15) (0.28)
0.51 0.25 0.31
Ramet length (ln) Number of leaves (ln) Cumulative leaf length (ln)
0.57 (0.00) (0.00)
0.51 (0.25) (0.37)
(0.10) (0.00) (0.00)
0.45 (0.17) (0.27)
(B) With competition Ramet length (ln) Number of leaves (ln) Cumulative leaf length (ln)
0.38
0.31
0.33
(0.00)
(0.09)
(o.oo)
o.11
(o.oo)
0.26
0.14
0.20
(0.00)
Ramet length (ln) Number of leaves (ln) Cumulative leaf length (ln)
(0.00) (0.00) (0.09)
0.17 (0.05) 0.28
(0.04) (0.06) 0.15
(0.04) (0.00) (0.01)
a The response-heritability was calculated using the mean square of the CO2-by-genotype interaction from an analysis of variance. bValues in parentheses are not significant (P > 0.05).
S2XCO2 H~4) = -
s~ + s 2G•
2 + SLs
Significant genotype-by-CO2 interactions may occur in the following general cases (Fig. 5): (a) if a p p a r e n t genetic variation in a new e n v i r o n m e n t (elevated CO2) is increased c o m p a r e d to the present one (ambient CO2), (b) if a p p a r e n t genetic variation in a new e n v i r o n m e n t is decreased compared to the present one, or (c) if levels of variation in a new and the present e n v i r o n m e n t are similar but the rankings of the genotypes differ between environments. Case (a) might be expected if the new e n v i r o n m e n t (in this case elevated CO2) had never before been experienced and therefore stabilizing or directional selection could not have r e d u c e d heritability whereas it did so in the present e n v i r o n m e n t (Falconer, 1989). The opposite case (b) seems unlikely, because it would imply that the population is more precisely p r e a d a p t e d to the new e n v i r o n m e n t than to its present. Case (c) would be expected if some of the genotypes were better p r e a d a p t e d to the new e n v i r o n m e n t whereas other genotypes were better adapted to the
44
Three types of significant variations in slopes of reaction norms. Type (a) would be expected if selection had reduced apparent genetic variation in the present environment (but obviously not in the new one; for further explanation see text). The diagrams are drawn without main effects of C02, that is, the average slopes are not significantly different from zero.
present environment. This has often been observed in reciprocal transplant experiments involving two environments and has been referred to as a home-versus-away effect (Schmid, 1992). With regard to the two CO2 environments, however, this explanation can be ruled out because no genotype could already have experienced the elevated CO2 environment. Therefore, a crossing of reaction norms must be taken as an expression of unselected genetic variation. In the present study, some of the observed variation in morphological characters of the two species clearly had genetic causes. The highest heritabilities were found in the generalist and the lowest in the specialist (see Table III). This suggests a higher genetic variation in P. and is consistent with the view that this generalist, with wide character distributions, consists of many specialist genotypes, each with a narrower distribution than the entire species, and not of a few plastic genotypes. This also indicates a strong potential for the evolution of this generalist species with increasing CO2 and therefore agrees with the findings of others (Parsons, 1992) that u n d e r a global-change scenario, generalists, in this case should replace specialists, i.e., in the longer term. This long-term prediction is in contrast to the short-term advantage of observed here. The reaction norms of individual genotypes of the three species across the two CO2 treatments appear to represent a r a n d o m scatter of lines when plotted (Fig. 6). In particular, there was no consistent response to increased CO2 and no evidence for larger apparent genetic variation in the new than in the present environment (see Table III), as would have been expected if stabilizing or directional selection had reduced the heritability in the present environment (see discussion of Fig. 5). However, the reaction
4.
Reaction norms of replicated genotypes of (a) and (b) grown in competition-free tubes. The means of the lines are significantly different among each other in and the slopes of the lines are significantly different among each other in that is, this species expresses significant genetic variation in the response to CO2 (see also Table 3). Bars indicate standard errors of differences among means.
n o r m s of the genotypes did significantly differ in slope for r a m e t length in (see Table IIIA and Fig. 6). These differences in slopes, which c o r r e s p o n d to type (c) in Fig. 5, must be i n t e r p r e t e d as significant genetic variation in the response to increased CO2, but presumably with little adaptive value in the past. Nevertheless, because the genetic variation occurs in characters that potentially affect plant fitness, it can be expected that, everything else not considered, with the rising CO2 the genotypes " p r e a d a p t e d " to high CO2 will b e c o m e d o m i n a n t over and eventually replace the genotypes best a d a p t e d to the past and present atmospheric conditions. T h e r e d u c e d slope-heritabilities u n d e r competition (see Table IIIB) indicate, however, that the genetic variation that is a p p a r e n t without competition may be masked in the natural vegetation (see Bazzaz and Sultan, 1986).
C. Consequences for Predicting the Response of Ecosystems O u r results are consistent with results from a small n u m b e r of cases r e p o r t e d in the literature that also have shown that within single populations
46 of plant species variation in the response to elevated CO2 can be significant. Garbutt and Bazzaz (1984) found variation in the response to elevated CO2 among four populations of with regard to flowering dynamics, total number of flowers, final biomass, and total number of flowers per unit dry mass. At the genotypic level, the increase in the number of flowers and seeds under elevated CO2 was found to be genotype dependent in (Curtis 1994). Wulff and Alexander (1985) reported significant differences in growth and germination among progenies of five maternal lines of grown at ambient and doubled CO2 concentrations. Polley and co-workers (1995) found differential modifications of water-use efficiency among wheat cultivars induced by elevated CO2. The fact that some genotypes show a positive, some others a neutral, and yet others a negative response to increased CO2 has several important implications for research on the effect of rising CO2 and global change. Results from any study working with replicates of a single or a few genotypes per species may not be extrapolated to predict community-level responses, not even at the local scale of a single field site, because depending on the sampled genotypes the response of each studied species may vary from positive to negative! Two possible solutions to evaluating a species response are short-term experiments in which the reaction norm of a representative n u m b e r of genotypes is measured separately (this study) and long-term experiments which allow changes in the genetic composition of the species to be observed. Because of the existence of genetic variation in response to elevated CO2, anticipated global change to a large extent may be absorbed by evolution within species rather than lead to substitutions of species. Indeed, when the response lines of the two species are compared (Fig. 6), there is much overlap between the two species. It is known that for many species most genetic variation exists within populations, with small but significant components added among populations, races, etc. (Baur and Schmid, 1996). However, we lack information on whether this progression of decreasing size of variance components is continued at the abovespecies level, and we do not know how quantitative genetic variation is partitioned into components among and within functional groups of species (cf. comparative approach of Bell, 1989). Therefore, we cannot predict at which level the greatest changes will occur in response to rising CO2. Another implication of these results is that characters measured on individual plants and genotypes are more sensitive indicators of possible environmental impacts on ecosystems than characters measured at higher levels of organization because the response of a higher level is the result of many lower level responses that may differ in direction. Quite generally, the fact that studying single genotypes can be useful for indication of environmental change does not mean that such studies can also be useful for prediction.
4.
47
Osenberg (1994) have recently shown that in environmental monitoring, individual-based parameters perform better than chemical, physical, or population-based parameters.
It is becoming increasingly clear that genetic variation within species may play an important role in the response of plant communities and ecosystems to global change. This genetic variation makes it possible for species to evolve, that is, to adjust to environmental change by shifting their mean characteristics. Indirect evidence for the possibility of withinspecies evolution has been obtained by Dippery (1995) with Plants grown under very low CO2 were unable to set seed because of bud abortion. Although the CO2 concentration used in the experiment was one of the lowest that this species might have experienced, this suggests that it is now composed of genotypes that have become adapted to present levels of CO2 Further examples of genetic variation or evolution in response to increasing CO2 are presented (see Curtis Chapter 2; Kingsolver, Chapter 1; and Tousignant and Potvin, Chapter 3). The possibility of evolutionary changes in the characteristics of species under the selective pressure of increasing CO2 should therefore be considered in models that try to predict effects of global change on ecosystems. This will be necessary in particular when considering organisms with short generation times. Several questions remain to be answered before such evolution can correctly be predicted. First, as shown by the variable responses of different genotypes to elevated CO2, the expression of genetic variation can vary across environments. This cautions against premature conclusions, because the responses could be modified by other genotype-by-environment interactions, e.g., different weather conditions, competition, and pathogens. A second caveat is that negative genetic correlations can exist among characters. For example, the significant heritability of leaf number would be irrelevant if it were correlated with "compensatory" genetic variation in leaf size, resulting in zero heritability at the level of whole-plant fitness. This calls for further analyses of a large set of characters and the study of their correlations to better understand which traits are relevant to estimate individual fitness. In particular, the longer-term survival of the planted genotypes in the different experimental treatments needs to be assessed. Finally, in order to predict the effect of an increase of CO2 concentration on the evolution of a given species, the intensity of the selection pressure exerted by this increase will have to be compared with the effects of other evolutionary forces, whether selective (effect of increasing temperature, UV radiation, nitrogen depositions, frequency of extreme events) or not
48 ( r a n d o m genetic drift, for example), and take into account the existence of potential constraints such as photosynthetic pathways (C~, C4, CAM).
We propose that the response of ecosystems to global change, in particular increasing CO2, will d e p e n d mainly on the extent to which plant species adjust to new conditions by plasticity or evolutionary change, which d e p e n d s on genetic variation. We show that the possibility of extrapolating results of short-term experiments to long-term prediction d e p e n d s on the relative i m p o r t a n c e of these two processes. Moreover, if genetic variation exists within species in the response to elevated CO2, it will not be possible to predict evolutionary changes from heritability estimates m a d e u n d e r present atmospheric conditions. In a field e x p e r i m e n t we grew genotypes of two closely related grassland perennials without competition and with competition at three levels of biodiversity u n d e r a m b i e n t and elevated levels of CO2. T h e generalist P. genetically more variable but not more plastic than the specialist and was at a relative disadvantage u n d e r elevated CO2. Thus, the hypothesis that specialists should be replaced by generalists when the e n v i r o n m e n t changes was not supported in the short term in this exemplary case study. Long-term predictions about the response to environmental change will, however, be difficult according to our result because showed significant genetic variation in its response to increased CO2, indicating that its long-term response cannot be predicted from heritability estimates m a d e u n d e r present conditions. F u r t h e r m o r e , an effect of biodiversity on the growth of the two species was only visible at elevated CO2, indicating that biodiversity may b e c o m e more i m p o r t a n t as the envir o n m e n t changes. T h e existence of large genetic variation within populations in ecologically relevant characters demonstrates that species cannot be considered as constants in ecosystem studies and global change models. Characters m e a s u r e d at the level of the individual may therefore be more promising indicators of ecosystem change than higher level characters.
We thank C. K6rner, P. Leadley, I. Sanders, T Steinger, and their many co-workersfor excellent collaboration in this experiment. We also thank F. A. Bazzaz, C. K6rner, D. Matthies, A. McLellan, D. Prati, T. Steinger, and an anonymous referee for critical comments on the manuscript. This research was supported by the SwissPriority Programme (SPP) Environment of the Swiss National Science Foundation (Grant 5001-035229 to B. Schmid and B. Baur and by the Swiss Federal Office for Education and Science (Grant 93.0271 to B. Schmid and C. K6rner).
4.
49
Baur, B., and Schmid, B. (1996). Spatial and temporal patterns of genetic diversity within species. "Biodiversity: A Biology of Numbers and Difference" (K. J. Gaston, ed.). Blackwell Sci., Oxford, in press. Bazzaz, F. A. (1990). The response of natural ecosystems to the rising global CO2 levels. 21, 167-196. Bazzaz, F. A., and Sultan, S. E. (1986). Ecological variation and the maintenance of plant diversity. "Differentiation Patterns in Higher Plants" (K. Urbanska, ed.), pp. 69-93. Academic Press, London. Bell, G. (1989). A comparative method. 133. 553-571. Billings, W. D., Peterson, K. M., Luken, J. O., and Mortensen D. A. (1984). Interaction of increasing atmospheric carbon dioxide and soil nitrogen on the carbon balance of tundra microcosms. 65, 26-29. Billington, H. L., and Pelham, J. (1991). Genetic variation in the date of budburst in Scottish birch populations: Implications for climate change. 5, 403-409. Bradshaw, A. D. (1965). Evolutionary significance of phenotypic plasticity in plants. 13, 115-155. Bradshaw, A. D. (1984). The importance of evolutionary ideas in ecology--and vice versa. "Evolutionary Ecology" (B. Shorrocks, ed.), pp. 1-25. Blackwell Sci., Oxford. Bradshaw, A. D., and McNeilly, T. (1991). Evolutionary response to global climatic change. 67 (Suppl. 1), 5-14. Cheplick, G. P. (1995). Genotypic variation and plasticity of clonal growth in relation to nutrient availability in 83, 459-468. Curtis, P. S., Drake, B. G., Leadley, P. W., Arp, W. J., and Whigham, D. F. (1989). Growth and senescence in plant communities exposed to elevated CO2 concentrations on an estuarine marsh. 78, 20-26. Curtis, P. S., Snow, A. A., and Miller, A. S. (1994). Genotype specific effects of elevated CO2 on fecundity in wild radish 97, 100-105. deJong, G. (1990a). Genotype by environment interaction and the genetic covariance between environments: Multilocus genetics. 81, 171-177. de Jong, G. (1990b). Quantitative genetics of reaction norms. J. 3, 447-468. de Jong, G. (1995). Phenotypic plasticity as a product of selection in a variable environment. 145, 493-512. Dippery, J. K., Tissue, D. T., Thomas, R. B., and Strain, B. R. (1995). Effects of low and elevated CO2 on C3 and C4 annuals. I. Growth and biomass allocation. 101, 13-20. Falconer, D. S. (1989). "Introduction to Quantitative Genetics," 3rd ed. Longman, New York. Garbutt, K., and Bazzaz, F. A. (1984). The effects of elevated COz on plants. 98, 433-446. Grulke, N. E., Riechers, G. H., Oechel, W. C., and Laeger, C. (1990). Carbon balance in tussok tundra under ambient and elevated atmospheric CO2. 83, 485-494. Hess, H. E., Landolt, E., and Hirzel, H. (1972). "Flora der Schweiz," Vol. 3. Birkh~iuser Verlag, Basel. Huey, R. B., and Hertz, P. E. (1984). Is a jack-of-all-temperatures a master of none? 38, 441-444. Hurlbert, S. H. (1984). Pseudoreplication and the design of ecological field experiments. 54, 187-211. K6rner, C. (1995). Biodiversity and CO2: Global change is under way. 4, 234-243. K6rner, C., and Arnone, J. A. (1992). Responses to elevated carbon dioxide in artificial tropical ecosystems. 257, 1672-1675.
50
Landolt, E. (1977). Okologische Zeigerwerte zur Schweizer Flora. 64, 1-208. Levins, R. (1968). "Evolution in Changing Environments." Princeton Univ. Press, Princeton, NJ. Lynch, M., and Gabriel, W. (1987). Environmental tolerance. 129, 283-303. Marshall, D. L., Levin, D. A., and Fowler, N. L. (1986). Plasticity in yield components in response to stress in and (Leguminosae). 127, 508-521. Milchunas, D. G., and Lauenroth, W. K. (1995). Inertia in plant community structure: State changes after cessation of nutrient-enrichment stress. 5, 452-458. Naeem, S., Thompson, L., Lawler, S. P., Lawton, J. H., and Woodfin, R. M. (1994). Declining biodiversity can alter the performance of ecosystems. 368, 734-737. Odum, E. P. (1989). "Ecology and Our Endangered Life-Support Systems." Sinauer, Sunderland, MA. O'Neill, R. V., DeAngelis, D. L., Waide, J. B., and Allen, T. F. H. (1986). "A Hierarchical Concept of Ecosystems." Princeton Univ. Press, Princeton, NJ. Osenberg, C. W., Schmitt, R.J., Holbrook, S. J., Abu-Saba, K. E., and Flegal, A. R. (1994). Detection of environmental impacts: Natural variability, effect size, and power analysis. 4, 16-30. Owensby, C. E., Coyne, P. I., Ham, J. M., Auen, L. M., and Knapp, A. K. (1993). Biomass production in a tallgrass prairie ecosystem exposed to ambient and elevated CO2. 2, 55-70. Parsons, P. A. (1992). Biodiversity and climatic change. "Conservation of Biodiversity for Sustainable Development" (O. T. Sandlund, K. Hindar, and A. D. H. Brown, eds.), pp. 155-167. Scandinavian Univ. Press, Oslo. Polley, H. W., Johnson, H. B., and Mayeux, H. S. (1995). Nitrogen and water requirements of C3 plants grown glacial to present carbon dioxide concentrations. 9, 86-96. Roughgarden,J. (1974). Niche width: Biogeographic patterns among lizard populations. 108, 429-442. Scheiner, S. M., and Goodnight, C.J. (1984). The comparison of phenotypic plasticity and genetic variation in populations of the grass 38, 845-855. Scheiner, S. M., and Lyman, F. (1989). The genetics of phenotypic plasticity. I. Heritability. 2, 95-107. Schmid, B. (1992). Phenotypic variation in plants. 6, 45-60. Schmid, B., and Weiner, J. (1993). Plastic relationships between reproductive and vegetative mass in 47, 61-74. Silander, J. A. (1985). The genetic basis of the ecological amplitude of II. Variance and correlations analysis. 39, 1034-1052. 39, Stearns, S. C. (1989). The evolutionary significance of phenotypic plasticity. 436-445. Stratton, D. A. (1994). Genotype-by-environment interactions for fitness of show fine-scale selective heterogeneity. 48, 1607-1618. Weber, M., and Schmid, B. (1995). Reductionism, holism, and integrated approaches in biodiversity research. 20, 49-60. Woltereck, R. (1909). Weitere experimentelle Untersuchungen •ber Artver~inderung, speziell ftber das Wesen quantitativer Artunterschiede bei Daphniden. 19, 110-192. Woodward, F. I. (1992). A review of the effects of climate on vegetation: Ranges, competition, and composition. "Global Warming and Biological Diversity" (R. L. Peters, and T. E. Lovejoy, eds.), pp. 105-123. Yale Univ. Press, New Haven, CT. World Conservation Monitoring Centre (1992). "Global Diversity: Status of the Earth's Living Resources." Chapman & Hall, London. Wulff, R. D., and Alexander H. M. (1985). Intraspecific variation in the response to CO2 enrichment in seeds and seedlings of 66, 485-460.
5 Genetic Variability and the Nature of Micro evolutionary Responses to Elevated CO2
There are four important questions one might pose regarding natural selection and increasing atmospheric CO2. First, is there genetic variability within plant populations on which rising CO2 might act? Second, how will rising CO2 affect the component processes of natural selection (namely, the expression of phenotypic variation in fitness, the degree to which this variation is heritable, and genetic covariance of other traits with fitness)? Third, what phenotypic traits will be favored under elevated CO2 conditions? And fourth, how will evolutionary changes affect plant community composition and ecosystem processes? Empirical studies addressing potential evolutionary implications of rising CO2 have primarily addressed the first question (Strain and Cure, 1985; Wulff and Alexander, 1985; Curtis 1994); the third and fourth questions have received some cursory attention in recent reviews (McLauglin and Norby, 1991; Roose, 1991; Strain, 1991; but see Geber and Dawson, 1993). This chapter addresses primarily the second question (which has received very little attention) and focuses on the direct effects of CO2 rather than indirect effects mediated by climate change. Our thesis is that an understanding of the impact of elevated CO2 on the component processes of selection is of central importance in predicting which traits are likely to be affected, and how selection of these traits may ultimately affect population, community, and ecosystemlevel processes. 51
Copyright 9 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.
52
In examining the implications of rising CO2 to natural selection, a useful comparison may be made to the relatively well-studied case of heavy metal tolerance (Antonovics 1971; Baker, 1987; Shaw, 1991). Drawing on this work, Bradshaw and McNeilly (1991) present a simple conceptual model of the effects of global change on selection processes. This model has a fixed selection criterion: individuals displaying phenotypes above this threshold survive and reproduce, whereas other individuals do not (Fig. 1: Model 1). Bradshaw and McNeilly's primary intent in presenting this model is to point out the importance of genetic variation in determining the potential future success of a species under profound environmental change. A more genetically variable population will generally have a greater proportion of individuals falling above the selection criterion cutoff, and would thus be expected to be more successful under altered conditions. Selection for metal tolerance generally involves a direct and nearly typological response (tolerant versus intolerant genotypes). In contrast, selec-
Three possible modes of directional selection under anthropogenic environmental change.
5.
53
tion pressures under elevated C O 2 a r e likely to be indirect and nontypological. Selection under elevated CO2 is likely to be indirect in that the immediate effect on most genotypes in most species is enhanced vegetative growth mediated by higher carbon uptake (cf. Bowes, 1991). However, at the population level such overall growth enhancements are likely to enhance local competitive interactions (see also Bazzaz and McConnaughay, 1992). This would result in a greater importance of resources other than CO2 in limiting plant growth. Selection pressures under this scenario would thus be highly contingent on local-neighborhood conditions and local supply rates of other plant resources. These considerations suggest some alternative conceptual approaches to the "mode of selection" problem, such as those illustrated as Models 2 and 3 in Fig. 1. To include the importance of resource competition, one might assume that some constant proportion of individuals displaying the highest fitnesses in a given population successfully reproduce (Fig. 1: Model 2). Alternatively, to incorporate the idea that enhanced density dependence would be local, one might assume that a constant number of the most "fit" individuals successfully reproduce in subpopulations of varying size (Fig. 1: Model 3). Wallace (1968) originally used the terms "hard" versus "soft" selection to distinguish between Models 1 and 2, respectively. These terms have since been co-opted, and are now frequently used to refer to preversus postdispersal selection in multiple-deme models (e.g., Wade, 1985; Holsinger and Pacala, 1990). Model 3 corresponds in important respects to the idea of "sieve selection," used to describe selection under conditions of strong local density dependence (Antonovics 1973 cited in Antonovics, 1978; see also Wilson and Levin, 1986). One might assume that because selective pressures under elevated CO2 are indirect, they would also tend to be weak. On the other hand, selection for heavy-metal tolerance has been celebrated as a microevolutionary model in part because selection pressures are often very strong. In this regard it is useful to examine previous work on selection against metal tolerance in noncontaminated areas. In many respects selection against tolerance may provide a closer parallel to selection under elevated CO2, because selection is dependent more on differential growth and reproduction than on survivorship, and the intensity of selection is highly density dependent (Cook 1972; Hickey and McNeilly, 1975; see also Namkoong 1993). Under high-density a n d / o r high-productivity situations, metal-tolerant genotypes are at a strong disadvantage. These genotypes can generally persist only under conditions of low competitive pressure. Observations of the incidence of metal tolerance along environmental gradients also suggest that density-dependent selection against metal tolerance may be quite strong (McNeilly, 1968; MacNair 1993). Along similar lines, Davies and Snaydon (1976) compared selection pressures on grass species
54
in the Park Grass Experiment, involving long-term fertilizer and liming treatments in plots maintained over more than 50 years. In general, greater coefficients of selection were found under conditions of higher overall plant growth. In summary, although not directly addressing the case of CO2, previous studies clearly indicate that indirect selection pressures mediated by competitive interactions can be very strong in natural plant populations. We have argued that the mode of selection under elevated CO2 will generally differ from selection for heavy metal tolerance and similar forms of anthropogenic change involving typological, noncontingent responses. Two other important general properties of a selective regime are its directionality and its spatial extent. With regard to the latter, selection for metal tolerance is generally local in scale, with a high probability of introduction of new genetic information from plants outside of the affected area (e.g., McNeilly, 1968; MacNair 1993). In contrast, selection under elevated CO2 is truly global in scale, affecting all extant plant populations. With the exception of "temporal migration" from seed banks, there will be no introduction of nonselected genetic material as plants evolve under rising CO2. Like selection for metal tolerance, elevated CO2 is likely to primarily result in directional selection, as opposed to stabilizing or disruptive selection. The global ambient CO2 concentration averaged over an annual interval is monotonically increasing (Houghton 1992). This progressive, gradual nature of CO2 increase does contrast markedly with selection for metal tolerance. Models of evolutionary responses to rising CO2 will thus need to consider genetic variation in responses across a range of CO2 levels. Additionally, there is a well-documented correlation between seasonal variation in CO2 and average CO2 concentrations, especially at high latitudes (e.g., Tans 1990). The selective regime under rising CO2 could thus involve some increase in variance associated with an increase in the mean. However, the amplitude of this variation is small relative to overall rise in CO2 levels. Moreover, winter peaks in CO2 are unlikely to be of importance since most plants are photosynthetically dormant during this period (thus driving the seasonal oscillations in atmospheric CO2). Higher winter peaks in CO2 may, however, be of some selective consequence in species showing pronounced photosynthetic activity during this period, such as winter annuals or winter-photosynthesizing conifers.
A. Conceptual and Mensurational Issues Natural selection is contingent on heritable variation. This raises the question of how we can best detect and measure genetic variability in C02
5.
55
responsiveness. In answering this question, it is first important to note that "CO2 responsiveness" at the whole plant level is not a trait expressed by an individual, but is rather an aspect of a plant's " n o r m of reaction" (i.e., the expression ofa phenotype across environmental states) (Schmalhausen, 1949). A second important point is that we are principally concerned with genetic variability in fitness-related traits, such as lifetime reproductive output. Much existing work on whole-plant responses to elevated CO2 focuses on vegetative aspects of plant growth and development, and thus may not be of direct relevance. Genetic variability in CO2 response may be detected rather simply as an interaction term in analysis of variance (e.g., Curtis 1994). However, such interactions are generally scale dependent, and also do not serve to measure the degree of genetic variability. Heritability coefficients quantify the proportion of variation in phenotypes that may be accounted for by genetic variation (Falconer, 1989): "narrow-sense" heritabilities measure the additive genetic c o m p o n e n t of variance, whereas "broad-sense" heritabilities measure the total variance accounted for by all genetic effects (i.e., additive, epistatic, and dominance components). It is sometimes of interest to quantify heritabilities for traits that summarize responses across environments. For this purpose clonal propagules or siblings may be grown in each environment, and heritabilities are calculated for the relative or absolute differences in mean performance across environments. Although this approach has been c o m m o n in studies of herbicide and metal tolerance, such measures are compromised by problems of both interpretation and analysis. For example, heritability values for metal resistance measured in this way commonly reflect variability in control populations rather than in experimental populations (MacNair, 1991). It is important to stress that for most purposes the heritability of responsiveness per se is not of direct interest. Rather, we wish to know the heritability of each trait in each environmental state, and the genetic covariances between traits measured in these different states (Falconer, 1952; Via and Lande, 1985; Via, 1987; see also Roose, 1991). Plant growth responses to elevated CO2 have most often been quantified in terms of response ratios, the conventional measure being the ratio of whole-plant biomass at 700/350 ppm CO2 (cf. Kimball, 1983; Cure and Acock, 1986; Poorter, 1993; Ceulemans and Mousseau, 1994). Similar data have been reported for intraspecific variation (Wulff and Alexander, 1985; Curtis 1994). From a comparative perspective, a fundamental problem with this measure is that it will generally vary through plant ontogeny. This ontogenetic variation arises inevitably from the sigmoidal nature of plant growth. To understand why this is so, consider two plants, one at ambient and one at elevated CO2, and both showing logistic growth. The biomass ratio (/3) as a function of time (t) would be:
56
where the constants Khi and/~o are asymptotic biomass values at elevated and ambient CO2, respectively, rhi and ~io are the corresponding initial relative growth rate parameters, and No is initial biomass. If rhi -- ~io then /3 will remain constant through ontogeny, being determined primarily by the ratio of/<,i to / ~io, the function will always be " h u m p e d " in form, displaying a local maxima at an intermediate value of t. Physiologically, enhanced carbon gain at elevated CO2 is directly linked to early growth rate, providing a strong expectation that rhi > ~io.Therefore, we can generally anticipate that differences in plant size (either relative or absolute) between CO2 levels will increase early in plant ontogeny. In contrast, the physiological basis for predicting effects on asymptotic maximal size is much less clear. Thus, as growth approaches an asymptote under both environmental conditions, the ratio may often decrease and approach a constant value. In order for the "response ratio" to remain constant throughout plant ontogeny, plant growth would have to be linear, which it is not (Causton and Venus, 1981). More generally, these patterns indicate that no single summary statistic, such as a response ratio, can adequately summarize effects of CO2 on plant "growth." Plant growth curves generally must be specified by at least two parameters (cf. Causton and Venus, 1981). It therefore follows that growth responses must also be quantified by at least two parameters, such as those describing effects on early exponential growth, and effects on asymptotic size. Empirical studies that have conducted detailed growth analyses under varying CO2 levels have generally found such " h u m p e d " patterns of response through ontogeny (e.g., Rogers 1984; Tolley and Strain, 1984; Bazzaz 1989; Coleman and Bazzaz, 1992; Tremmel and Patterson, 1994; Walters 1993; see also Fig. 2). Moreover, these analyses indicate that peak size differences are reached at different times for different species. Therefore, no single time frame during development can be used for comparison. Any effort to understand variation within or among species in CO2 growth responses should take these basic ontogenetic patterns into account. Most experiments on CO2 growth responses report only final harvest measurements, however. Also, the majority of studies focus overwhelmingly on early growth. For example, the median growth interval of studies reviewed by Poorter (1993) is 43 days. From a microevolutionary standpoint, some researchers have argued that growth trajectories should generally be treated as "infinite-dimensional" characters, in which the phenotype of an individual is treated as a function (of arbitrary form) rather than as a finite set of measurements (Kirkpatrick and Lofsvold, 1989, 1992). The variance-covariance matrix of a trait mea-
57
5. 5 A 4 3 2
1
"~-'-----~_
_
_
._.o
1
c~
.,,.~
0
0
5 B
3
:
:
:
:
:
:
:
:
:
:
:
:
I
-,
.=
-.
:
.-
:
-.
4 2
3 2 o x=
:
1
~
o
1 0
"
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
:
5 4 3 2 1 0
.
.
.
.
20
.
.
.
.
.
40
.
.
.
.
.
.
60
.
.
.
.
.
80
.
.
100
Time since germination (d)
120
20
40
60
80
100
120
Time since germination (d)
Growth curves (left panels) and response ratios (right panels) for three genotypes of u n d e r a m b i e n t a n d elevated CO2 conditions. In the first set of panels, o p e n symbols indicate plants at a m b i e n t CO2 (350 p p m ) , closed symbols plants at elevated CO2 (700 p p m ) .
sured at several points in time yields eigenvalues and eigenfunctions, and their magnitude and shape, respectively, indicate which evolutionary modifications in growth trajectories are more or less likely. For several reasons, this general approach is well suited to describing and modeling genetic variation in plant growth responses to CO2. In particular, a growth equation (such as the logistic) need not be assumed a priori, and the evolutionary response of the entire growth trajectory can be modeled (cf. Kirkpatrick 1990). Moreover, the infinite-dimensional formalism can be similarly extended to treat the evolution of norms of reaction (Gomulkiewicz and Kirkpatrick, 1992), such as response to CO2 levels treated as a continuous function. Models of this sort yield insights into evolutionary dynamics of traits (see below for an example of such an analysis) and would also be of particular use in projecting evolutionary changes under conditions of
58
gradually increasing C O 2. The primary disadvantage of this approach is the need for very large data sets to parameterize models (i.e., data on the continuous growth of plants of known ancestry at a range of CO2 levels). It should also be emphasized that current models of the evolution of growth trajectories have not incorporated density- or frequency-dependent processes. Although reproductive responses are of primary importance from an evolutionary perspective, the vast majority of existing comparative data on whole-plant responses to CO2 focus on vegetative growth. Reproductive output is often closely correlated with vegetative biomass and other measures of plant size in herbaceous plants (e.g., Samson and Werk, 1986; Weiner, 1988; Kawano 1989; Hartnett, 1990). However, there are a variety of experimental results suggesting effects of elevated CO2 on reproduction that are disdnct from vegetative growth enhancement (see, e.g., Ziska and Teramura, 1992; Enoch and Honour, 1993; Ackerly and Bazzaz, 1995; Bazzaz 1995a). Reekie (1994) provide evidence that long-day plants generally display accelerated reproduction at elevated CO2, whereas short-day plants exhibit delayed reproduction. Either accelerated or delayed reproduction may result in altered patterns of reproductive allometry under elevated CO2. There may also be changes in seed quality under elevated CO2 (Garbutt and Bazzaz, 1984; Farnsworth and Bazzaz, 1995). Further work on these issues is called for, particularly multiplegeneration experiments. However, it is clear that differential vegetative growth may not be a safe basis by which to infer evolutionarily relevant variation in responses to elevated CO2. In summary, studies of C O 2 effects on plant growth have generally taken what is inherently a multidimensional problem, and reduced it to a single dimension (i.e., the "response ratio"). Future work aimed at understanding either intra- or interspecific variation in CO2 effects on plant growth should account for the variation of effects through plant ontogeny, and also across the range of CO2 levels to be encountered by plant populations. Additionally, there is evidence that reproductive responses to elevated CO2 may often diverge from vegetative responses. Attempts to understand the evolutionary implications of rising CO2 should therefore directly examine proximate fitness measures such as lifetime reproductive output. B. Empirical Data on Variation within and among Species The preceding discussion points to the general importance of growth analyses and reproductive data in studies of evolutionary responses to CO2. In what follows, we will make extensive use of data from a recent experiment characterizing norms of reaction of Medic. genotypes to CO2 and conspecific density. Detailed methods and results of the study will be presented in full elsewhere (S. C. Thomas, M. Jasiefiski, and F. A.
5.
59
Bazzaz, manuscript in preparation; see also Bazzaz 1995). In brief, the study incorporated a fully crossed design using 8 genotypes grown individually in 2-liter pots and in dense, monospecific stands, under elevated (700 ppm) and ambient (350 ppm) CO2 conditions. The experiment was conducted in a controlled-environment facility at Harvard University, and each CO2 level was represented by three spatially interspersed blocks. Each block had 4 replicate individual and 12 replicate high-density plants for each genotype. Available soil volumes and nutrient input levels were matched to those of eastern Massachusetts. The genotypes were selfed progeny of cloned plants that had previously been systematically sampled from a natural population. Growth analyses were based on allometric estimates of biomass derived from destructive harvests of plants grown under similar conditions. The importance of growth analysis in understanding genetic variation in whole-plant responses to CO2 is well illustrated by the data shown in Fig. 2. (Note that each point represents an average of 24-36 replicate plants.) Genotype A shows a pronounced early growth enhancement and a positive, though smaller, enhancement in final size. The biomass response ratio thus declines from near 2 to 1.2. In contrast, genotype C shows a very strong early growth enhancement, but considerably smaller final biomass at elevated than at ambient CO2 levels. The biomass response ratio thus declines from near 3 to less than 0.6. (Losses in estimated biomass late in ontogeny reflect leaf drop during fruit development in this species .) The general " h u m p e d " pattern of response shown in these three genotypes was seen in all genotypes at both density conditions. Given the range of variation in biomass response ratios seen within a single species, much of the variation in CO2 responses reported among species could actually represent intraspecific variation that is either genetic or ontogenetic in nature. This point is dramatically illustrated by the range of variation in CO2 response found among genotypes sampled from an Illinois old field (Fig. 3). The overall distribution of growth enhancement ratios, pooling all genotypes and sequential measurements, is statistically indistinguishable (G test: P > 0.05) from the distribution of response ratios across 155 species reported by Poorter (1993). In order to partition variation in CO2 responses into within and between species components, the minimal data requirement would be growth analyses of genotypes within some sample of plant species. Ideally, such data should also encompass a range of environments for each species, so as to allow an evaluation of the relative importance of"environmental" variation in CO2 responsiveness. Unfortunately, no such data set yet exists. Existing data do give some indication that genetic variation in CO2 responsiveness is actually large, relative to variation among species. Table I lists the conventional "response ratios" for genotypes in four studies investigating genetic
60
F i g u r e 3 Distributions of CO2 growth " e n h a n c e m e n t ratios" (biomass at 700 p p m / b i o m a s s at 350 ppm) for 8 genotypes of pooling 7 weekly measurements and 2 density treatments (top), and 155 plant species compiled by Poorter (1993) (bottom). The 2 distributions do not differ significantly (G test for homogeneity: P > 0.050).
variation in CO2 responses in annual plants. Table II gives the result of a mixed model analysis of variance that partitions variance in response ratio within and among species, and among environmental states. For each
Species
Env.
Genotypes
Range of response
Reference
2 nutr. 2 nutr. 1 only
4 6 5
0.47-2.37 0.25-1.37 0.93-1.39
Wulff and Alexander, 1985 Fajer 1992 Curtis 1994
2 dens.
8
0.30-1.40
Thomas manuscript in preparation
a Responses are calculated as the ratio of plant performance at 700/350 ppm for the most proximate fitness characteristic measured. The ranges reported are pooled across environments in studies for which measurements were made under different environmental treatments.
61
5.
Effect
d.f.
SS
MS
F
P
Species Environment Genotype Error Total
2 1 16 21 40
0.120788 0.377143 3.80075 2.31536 6.83487
0.060394 0.377143 0.237547 0.110255
0.25424 3.4206 2.1545
0.7786 0.0785 0.0502
For this analysis environmental treatments were classified as "good" (high nutrient or low density) vs "poor" (low nutrient or high density).
species the most proximate fitness measure was used (i.e., either final total reproductive or vegetative biomass). Also, in three of the four studies multiple environmental states were measured. For the purposes of the analysis, these were scored as "favorable" (high nutrients or low density) or "unfavorable" (low nutrients or high density). This preliminary analysis suggests that intraspecific variation exceeds interspecific variation: 56% of total variance in response ratio is explained by genotype, approximately 6% is explained by environmental state, and less than 2% is explained by species. The genotype term approaches significance at the P < 0.05 level, while the environment term is also marginally significant. Existing data thus suggest that intraspecific variation in responses to elevated CO2 is very substantial, perhaps even greater in magnitude to interspecific variation. However, we emphasize that this preliminary analysis does not take into account the important issue of ontogenetic changes in growth enhancement. Additionally, models that incorporate CO2 effects in the context of local resource competition would be necessary to rigorously evaluate the relative importance of genetic change versus change in species composition under rising CO2.
A. A Quantitative Genetic Framework Quantitative genetic models represent a potentially powerful tool for understanding the evolutionary dynamics under global change (e.g., Lynch and Lande, 1993). A central assumption of these models is that phenotypic traits of interest are determined by many genes of small effect (MitchellOlds and Rutledge, 1986; Falconer, 1989). However, there are important cases of evolutionary responses to anthropogenic disturbance in which
62
single locus changes have very large effects (such as copper tolerance in (MacNair, 1977, 1991). It is within the range of possibility that single genes could also have large effects on CO2 responses. For example, Musgrave (1986) examined CO2 responses in pea hybrids differing in the presence vs. absence of the cyanide-resistant respiratory pathway. Presence of the pathway was associated with a very low growth response to CO2, the explanation offered being that carbohydrate production was respired in hybrids possessing the pathway. Many metric traits display a right-skewed distribution of gene effects, with some degree of "major gene" influence as well as many genes of small effect (Hill and Caballero, 1992). It seems likely that CO2 responses would have a similar distribution of gene effects; however, studies addressing the genetic basis of CO2 responses in natural plant populations are entirely lacking. In the simplest quantitative genetic models of selection, the rate of change in fitness is a product of the heritability of fitness, and the relative variation in fitness [Eq. (2a)]. The response of some trait correlated with fitness may be estimated by substituting the product of fitness heritability and the covariance of the trait with fitness in this expression [Eq. (2b)]. The "Chicago school" multivariate quantitative genetic models of selection essentially extend this expression to predict selection on a set of traits, incorporating a genetic variance-covariance matrix in the place of the heritability term (Lande, 1979, 1982; Arnold and Wade, 1984) h2 h2
(2a)
2 2.
(2b)
Here R or R' is the selection rate on fitness or on a correlated trait, respectively; h2 is narrow sense heritability, V~ is variance in fitness, W is mean absolute fitness before selection, and ~ is the covariance of a given 2 is simply the square of the trait with fitness. Note that the term coefficient ofvariance of fitness (cf. Thomas and Bazzaz, 1993). The product h2 2 is sometimes called the "opportunity for selection" (denoted I), and has been used as a measure ofevolvability (Crow, 1958; Houle, 1992). Elevated CO2 may potentially affect the selection process by systematically altering some or all of the variables in these expressions, namely, the relative variance in fimess within local populations, the heritability of fitness-related traits, and the genetic covariance of particular traits of interest with fitness. In the subsequent sections, these parameters are addressed in turn.
B. Phenotypic Variability Rising C O 2 may affect the selection process simply by altering the degree of phenotypic variability in fitness related traits. One mechanism by which this could occur is the acceleration of size differences due to enhanced
63
5.
overall growth. In the exponential phase of plant growth, small differences in growth rates among individuals result in exponential increases in size differences. Over time an even-aged set of plants that displays a normal distribution of seedling sizes will show increasing relative variation (Koyama and Kira, 1956; Uchmafiski, 1985). The addidon of any resource is expected to accelerate this process. By this reasoning, one would expect greater phenotypic variability in size in plant populations under elevated CO2. However, the sigmoidal nature of plant growth modifies this expectation. Plants grown at higher resource states may reach asymptotic sizes earlier, which could potentially result in decreased variability in asymptotic size or reproductive output under high resource conditions. A second mechanism by which elevated CO2 could influence phenotypic variability in plant size and reproductive output is by accelerating competitive interactions for other plant resources. In general, size variability increases through stand ontogeny in even-aged plant monocultures (e.g., Thomas and Weiner, 1989), and monocultures growing at higher densities display greater variability in plant size (Weiner and Thomas, 1986). Addition of nutrients and other resources often results in greater variation in size and reproductive output (Weiner, 1985; Rice, 1990). Elevated CO2 might similarly result in increased size variability. Morse and Bazzaz (1994) specifically addressed this issue in experiments with stands of two annual plants. Their results provide some evidence that elevated CO2 may accelerate size hierarchy formation and self-thinning. However, from an evolutionary perspective, it is of greatest interest to examine variability in reproductive output, rather than in size. Table III
C.V. of final total seed mass
Opportunity for selection (I)
Broad-sense heritability (H 2)
Response to selection (R)
350 p p m
0.557
Individually grown 0.31
0.052 ,
0.016
700 p p m
0.400
0.16
0.236
0.038
350 p p m
1.442
H i g h density 2.08
0.050
0.104
700 p p m
1.752
3.07
0.104
0.320
Statistical significance of differences were tested using a boot-strapping approach (cf. Thomas and Bazzaz, 1993), with 2000 iterations for each test. * Indicates pairwise comparison is significant at P > 0.05. Based on Bazzaz 1995. a
64
3,1.
presents data from the experiment. The data are consistent with the hypothesis that enhanced competition results in higher variability in fitness under elevated CO2: a significant difference in the coefficient of variation of final seed mass was found at high density, though not for individually grown plants. Not all species show such a pronounced response, however (S. C. Thomas, M. Jasiefiski, and F. A. Bazzaz, unpublished data).
C. Heritability The heritability of fimess-related traits is expected to be close to zero, as selection will operate to remove genetic variance for such traits (Fisher, 1930). Nonzero heritabilities may commonly be maintained by negative genetic correlations among a set of fitness-related traits. However, if traits are expressed in an evolutionarily novel environment in which selection has had no opportunity to act, then heritabilities may be much higher even in the absence of "trade-offs" among fitness components (e.g., Service and Rose, 1985). This raises the central issue of whether or to what degree elevated atmospheric CO2 (i.e., in the 350-700 ppm range) constitutes an evolutionarily novel environment. From a very long-term paleoecological perspective, CO2 levels were considerably higher in the geological past (B6ger, 1980; Spicer and Corfield, 1992). However, the past 160,000 years have been a period of relatively low atmospheric CO2 levels (Barnola 1987). Life spans of terrestrial plants vary in general from less than 1 year to several hundred years (Harper and White, 1974). Because the majority of plant species are iteroparous, generation times are generally much shorter. If changes in CO2 concentrations have directly or indirectly resulted in even modest selective effects (e.g., selection coefficients of order 0.00010.01), then there has almost certainly been sufficient time for preindustrial low CO2 levels to have eroded genetic variance related to earlier evolutionary processes driven by high CO2. The idea that rising CO2 constitutes an evolutionarily novel environment also depends on the nature of the selective regime generated. As noted above, the selective impact of rising CO2 may generally be expressed indirectly, particularly by exacerbating competitive interactions for other resources (as well as through the indirect effects of CO2-forced climatic change). In a review that briefly addresses this issue, Roose (1991) suggests that "these secondary effects of increased CO2 do not create novel environments, but rather environments which already occur elsewhere" (p. 122). There are several strong arguments against this view. First, almost any "novel" selective pressure is novel only locally, not globally. To take the paradigmatic example, mining activities expose heavy-metal-rich soils of a sort that generally already occur elsewhere; yet this does not alter the fact that local populations colonizing mine-spoils may have high genetic variance in fitness components due to a lack of previous heavy-metal expo-
5.
65
sure. Second, competitive regimes generated by altered C O 2 levels may indeed be qualitatively novel. For example, if plant monocultures are able to sustain a higher leaf area index under elevated CO2, then this could generate a qualitatively novel light environment under the canopy of a given species. Along these lines, one study detected substantial effects of CO2 on red-far red ratios of light transmitted through a canopy (Arnone and K6rner, 1993). Rising CO2 could also act as a novel evolutionary environment through its effects on plant development. Two general classes of developmental effects may be of importance in this regard. First, CO2 has known biochemical interactions in plants that are not mediated by the carboxylase activity of rubisco. For example, CO2 is directly involved in the regulation of ethylene biosynthesis (e.g., Horton, 1985; Cheverry 1988). Also, CO2 binds to rubisco to create the active form of the enzyme, and also regulates the activity of rubisco activase (for a review see Bowes, 1991). Second, elevated CO2 may have important novel effects on plant development that arise from increased carbon accumulation. One notable example is changes in nonstructural carbohydrate chemistry. For example, wheat shows a qualitatively different pattern of fructan accumulation under elevated CO2, with large amounts synthesized very early in ontogeny (Smart 1994). Completely novel carbohydrates are synthesized by certain conifer species at elevated CO2 (H. Lee, personal communication). Such quantitative and qualitative changes in carbohydrate chemistry may have important and novel effects on plant morphogenesis. One possible example of such an effect is dramatic changes in leaf form seen in (Fig. 4). Under elevated CO2 plants produced exaggerated "sun-leaf" morphologies, with significant changes in leaf length and degree of dentition. An alternative perspective on fitness heritabilities under rising CO2 derives from studies of genotype-specific performance under varying competitive regimes. A variety of studies indicate that competition in plant populations is often asymmetric with respect to size: large individuals usurp resources at the expense of small individuals (Lomnicki, 1988; Weiner, 1990). This phenomenon has the potential to greatly amplify small differences in size among individuals in a population, particularly in even-aged stands. Differences in size early in plant ontogeny may generally be largely due to microenvironmental heterogeneity. The amplification of early size differences may therefore act to enhance environmental variation in plant performance. This reasoning leads to a prediction of lower heritabilities for size under conditions of enhanced competition, such as at high density (Thomas and Bazzaz, 1993). Higher resource levels would also be expected to result in earlier and more intense competitive interactions. By this reasoning, one might expect reduced heritabilities for fitness-related traits under elevated CO2.
66
o 350 ppm 9700 ppm
1.4
o
9
1.2
~
O-'"""
-"~ o
9
1.0
.."
9
1 " ~ ~
~-~176 ~
0.8 ~~
0.6
~176176 o*~
0
0.2
014
016
018
110
1;2
1;4
1J6
1;8
i
Log leaf area (cm 2 ) 4 Effects of elevated CO2 on leaf shape in Altered patterns of plant development under rising CO2 may result in qualitatively novel phenotypes. From Thomas and Bazzaz, 1995. Figure
To summarize these arguments, there is strong reason to believe that increasing CO2 levels could act as an evolutionarily novel environment, perhaps most importantly through effects on plant developmental processes. This is expected to be associated with increased genetic variability in fitness-related traits at elevated CO2. However, accelerated competitive interactions may under some conditions act to increase the importance of small-scale environmental sources of variation in plant performance. This effect may act to reduce genetic variability in fitness-related traits at elevated CO2. Do trait heritabilities actually respond to elevated CO2? Data from the experiment indicate a consistent pattern of higher heritabilities at elevated CO2 than at ambient CO2 for final seed mass (Table III). This pattern is most pronounced for the individually grown plants, although high-density plants show a nonsignificant trend in this direction. Heritabilities at either CO2 level are also generally lower at high density. The overall pattern is consistent with the idea that accelerated competitive interactions may enhance the role of "environmental noise" in determining plant performance (cf. Thomas and Bazzaz, 1993). There is thus some evidence
5.
67
that both "evolutionary novelty" and "environmental noise" may play important roles in determining heritabilities of fitness-related traits under elevated CO2. D. Genetic Correlation Structure
Responses of a given trait to selection depend on the covariance of that trait with fitness. More generally, the response to selection of any set of phenotypic characteristics will depend on the overall genetic variancecovariance structure (Lande, 1982; Falconer, 1989). An ultimate task, however, is to establish a connection between quantitative-genetic estimates of variation and covariation of traits, developmental processes, and functional relationships of traits (Riska, 1989). Only then will we be able to provide mechanistic explanations of the evolution of suites of traits in novel environments (Chapin 1993). Such limitations notwithstanding, quantitative genetics of covariances among traits continues to be a basic framework for evolutionary considerations. The verdict is still out, however, as to the feasibility of using phenotypic correlations among traits in lieu of, more difficult to obtain, genetic correlations (Cheverud, 1988; Willis 1991; Roff, 1995). Although genetic variance-covariance structure is often treated as constant in quantitative genetic models, it is not fixed. There is a substantial literature that examines changes in genetic correlation structure with environmental conditions (e.g., Giesel 1982; Itoh and Yamada, 1990; Wilkinson 1990). (Although genetic covariances are directly used in calculating selection effects, comparative analyses are often conducted with genetic correlations.) In plants, changes in genetic correlation structure have been best documented with respect to changes in local density (Geber, 1990; Mazer and Schick, 1991; Thomas and Bazzaz, 1993; but see Shaw and Platenkamp, 1993; see also Young 1994). A body of theoretical work exists regarding expected changes in genetic variance-covariance structure under selection and drift (Crow and Kimura, 1970, pp. 236-239; Avery and Hill, 1977, 1979; Turelli, 1988). However, less attention has been given to possible effects of environmental changes, novel or otherwise. The statement has even been made that " . . . theory cannot predict whether the environmental changes that select for new phenotypes will change environmental or genetic covariances" (Turelli, 1988, p. 1344). One possible basis for prediction may, however, stem from observations that functionally or developmentally related traits generally show high (positive or negative) genetic correlations (e.g., Clark, 1987; Cowley and Atchley, 1990). One might predict in a novel environment of any sort that functionally related genetic correlation structure would tend to weaken. Similarly, environmental changes altering developmental processes could result in an overall lowering of absolute values of genetic
correlations. Another possibility is that changes in genetic variance in fitness could affect relationships between fitness correlates and other traits simply by altering overall phenotypic variance. Specifically, given some underlying genetic relationship, an increase in variance in one trait could result in increased genetic covariance with another trait. The sign of such covariance, however, is hard to predict, especially in traits whose phenotypic expression depends on the allocation of a single resource (van Noordwijk and deJong, 1986; de Jong and van Noordwijk, 1992). The empirical data on allow for a preliminary analysis of effects of elevated CO2 on genetic correlation structure (Fig. 5). In spite of the relatively small sample size (in terms of numbers of genotypes), both statistically significant genetic correlations and statistically significant differences between CO2 treatments are detected. At both ambient and elevated CO2 there are high positive genetic correlations between final size metrics: namely, height, leaf area, and biomass. Also in both treatments there is a high negative genetic correlation of these characteristics with seed size (mass). However, while at ambient CO2 initial relative growth rate shows a strong positive genetic correlation with final size characteristics, this correlation is very weak at elevated CO2. Genetic correlations of any of the traits examined to the most proximate fitness measure (final fruit mass) do not attain statistical significance (P < 0.05) for either COz level. However, there is a suggestive negative genetic correlation of final plant height with fruit mass that is considerably higher at elevated CO2. Overall, genetic correlations appear to be somewhat weaker under elevated CO2 conditions. These observations suggest surprisingly large effects of CO2 on genetic correlation structure; however, we emphasize that in this preliminary analysis we have not performed a conservative pooled test for the entire character matrix (cf. Shaw, 1991b). Long-term responses to selection may often be determined by fitness differentials, rather than genetic correlation structures, unless genetic correlations are very high (e.g., Via, 1987; Zeng, 1988). This point is of some interest from the perspective of evolutionary responses to a relatively sudden environmental change such as rising CO2. Although fitness maximization may prevail over the long term, it is likely that effects of genetic correlation structure will be especially pronounced over the first several generations of selection. This points to the importance of further studies aimed at elucidating genetic correlation structure and its consequences of evolution under rising CO2.
It has commonly been assumed that traits directly involved in physiologi cal responses to C02 may respond evolutionarily to rising C02. Thus, a priori
5.
69
Effects of elevated CO2 on genetic correlation structure in Data are for individually grown plants. Black bars indicate significant positive genetic correlations; white bars indicate significant negative correlations. Hatched bars indicate nonsignificant positive correlations; stippled bars indicate nonsignificant negative correlations. Genetic correlations were calculated as Pearson product-moment correlations between genotypic means. Bar width is proportional to the absolute value of the correlation.
predictions regarding expected evolutionary responses may be derived from considerations of optimal allocation patterns. For example, one might e x p e c t e v o l u t i o n a r y r e s p o n s e s involving c h a n g e s in n i t r o g e n a l l o c a t i o n f r o m r u b i s c o to e i t h e r d a r k cycle e n z y m e s i n v o l v e d in R U B P o r p h o s p h a t e
70
8.
regeneration (Bowes, 1991,1993; Stitt, 1991 ) or to light-harvesting molecules such as chlorophyll. However, there is presently a complete absence of studies addressing genetic correlations of such physiological traits with reproductive output under elevated CO2. Similarly, it is not clear that evolutionary responses ofstomatal density will match developmental responses of individual plants (see Beerling and Chaloner, 1993). Reported long-term trends in stomatal density (e.g., Woodward, 1987; Pefiuelas and Matamala, 1990) may or may not involve genetic change. It also cannot be assumed that either net photosynthesis orvegetative growth responses to CO2 will necessarily increase as plants evolve under rising CO2. For example, there is a negative genetic correlation offruit production with final plant height observed in the population under elevated CO2 (Fig. 5). The predicted short-term selective response would thus be for plants of smaller final stature. If rising CO2 levels generally act to exacerbate competitive interactions within and among plant species (Bazzaz and McConnaughay, 1992), one might expect CO2 to favor traits that are also favored under conditions of high density or high productivity. Such traits may include early germination, rapid early growth, delayed reproduction, and increased stem allocation, among others (e.g., Thomas and Bazzaz, 1993). A second a priori hypothesis is that selection under elevated CO2 will often involve amelioration of "nonadaptive" plastic responses. Consider, for example, species that show strong developmental effects, such as altered reproductive timing (cL Reekie 1994). Genots, pes that retain a flowering schedule more closely matched to their current seasonal pattern might be strongly favored under rising CO2. Similarly, forms of that do not develop exaggerated sun leaves under elevated CO2 (see Fig. 4) may be differentially favored. Future studies should be cognizant of possible indirect selective pressures brought about by enhanced competitive interactions and also by developmental effects of rising CO2. The potential for evolutionary change in plant traits can also be explored through an analysis of their covariance structure during ontogeny (Kirkpatrick 1994). In the case of per-plant leaf area measured in at six points in time, the eigenfunction associated with the dominant eigenvalue did not change sign in high-density populations, at either CO2 level (Fig. 6; M. Jasiefiski, S. C. Thomas, and F. A. Bazzaz, manuscript in preparation). This means that selection for an increase (or decrease) in leaf area at one age has the same effect at every point during ontogeny. A similar effect was found in the case of plants grown individually, although much smaller fraction of total variance in growth trajectories was concentrated in the first eigenvalue (i. e., 60% of variance, rather than over 90%, as in high-density plants). In the low-density plants, the second eigenfunctions changed signs (Fig. 6), indicating possible trade-offs in responses to selection for early versus late leaf area; e.g. selection for an increase in leaf area
71
5.
700 ppm CO2
350 ppm CO2
1st eigenfunction
High density
o.eoj
/
~
o.1o
0.05 ~
0.05
30
40
50
60
70
80
30
40
50
60
70
80
1st eigenfunction 0.15
0.15
0.10
0.10
0.05
0.05
Low
density
2nd eigenfunction 0.2
0.2
0.1
0.1
0.0 ~ , J 3 0
-0.1[
40
50
'
'
0.0
-0.1 Time since germination (d)
Figure 6 Effectsof elevated CO2 and density on the covariance structure of leaf area during growth in Line thickness of the eigenfunction corresponds to the fraction of variance associated with its eigenvalue. In high-densityplants, second eigenvalues accounted for less than 5% of total variance and are not shown. From M. Jasiefiski, S. C. Thomas, and F. A. Bazzaz, manuscript in preparation. of young plants will lead to decreases in leaf area of older plants, and vice versa. Density of plants had a stronger direct effect than CO2 on the potential for evolutionary modifications in growth trajectories: selection on leaf area in plants grown without competition had m o r e possible outcomes than selection in high-density populations. This result points again to density as an i m p o r t a n t proxy for CO2 effects. A recent selection e x p e r i m e n t on (see Tousignant and Potvin, Chapter 3) provides virtually the only direct experimental evidence
regarding traits favored under rising C O 2. Following eight generations of truncation selection for reproductive output (under a regime of increasing CO2, temperature, and temperature variation), selected lines exhibited increased biomass, but reduced allocation (in both absolute and relative terms) to reproduction. One possible explanation for these results is that plants under elevated CO2 were effectively under a selective regime of increased aboveground competition. A variety of life history models predict increased size and decreased reproductive allocation under such conditions (e.g., Gadgil and Bossert, 1970). However, the inclusion of the temperature increases and heat shock episodes in the experimental design make these experimental results difficult to interpret unambiguously. Inbreeding effects could also have resulted in the observed reproductive declines. One earlier study attempted to select lettuce for high growth response to elevated CO2 (Maxon-Smith, 1977). No significant response was detected after eight generations of selection; however, the author suggests that the experiment may have been compromised by poor control of CO2 and temperature levels. Some intriguing physiological responses to selection under very low CO2 conditions have been documented in a series of experiments with doubled haploid lines of (Delgado and Medrano, 1991; Delgado, 1992a,b, 1993). Lines selected for survival at CO2 levels near the CO2 compensation point (60 ppm) were found to display increased vegetative growth under both field and laboratory conditions. However, this was not associated with either increased leaf-level photosynthesis or decreased photorespiration. Rather, selected lines showed increased rates of leaf production and lower respiration rates (Delgado 1992b, 1993). Additionally, leaves of selected lines possessed smaller mesophyll cells than did controls, which may have decreased mesophyll limitation of CO2 diffusion under low CO2 conditions. The implications of these findings to selection under high CO2 are not entirely clear. However, it is of some interest that screening of genotypes under low CO2 levels has previously been used as a strategy to select for high photosynthetic rates and growth potential in crop species (see also Cannell 1969; Menz 1969). One potentially important source of direct evidence regarding traits favored under rising CO2 is a comparison between populations of plants growing near naturally high CO2 environments with populations under ambient levels. Unfortunately, no systematic study of this nature has yet been attempted. However, one obscure and very early study addressing the ecology of crater plants on the island of Java contains some suggestive observations on possible effects (Von Faber, 1925, 1927). Von Faber's attention was initially drawn to CO2 vents by the presence of extremely green leaves of plants in the immediate vicinity of the vents. He later documented that these plants did in fact show exceptionally high chloro-
5.
phyll content. Von Faber also speculated that high C O 2 levels could compensate for very low light levels, facilitating the evolution of extreme shade plants in such environments (Schimper, 1935, p. 146). In summary, at present no strong a priori theory exists allowing for predictions of what traits will be favored under rising CO2. Two speculative hypotheses are advanced: first, selection may act largely through enhanced local competition, thus favoring suites of traits enhancing fitness under high density or high productivity environments; and second, in cases where CO2 has pronounced effects on plant developmental processes, selection may operate to ameliorate such effects. Definitive empirical studies aimed at identifying traits favored under rising CO2 are presently lacking. A variety of approaches, including selection experiments, quantitative genetic investigations, and comparative studies of populations exposed to naturally high CO2 levels are necessary to make progress in this area.
There are well-documented examples in which genetic change in populations can have profound impacts on ecosystem function. In the case of a heavily contaminated mine spoil, ecosystem properties such as carbon and nitrogen flux may be entirely contingent on patterns ofintraspecific genetic variability (Antonovics 1971; Shaw, 1991). If no tolerant genotypes were available to colonize heavily contaminated areas, then no carbon fixation by any autotrophic organism would exist. The evolutionary implications of a resource increase are, however, less clear at the ecosystem level. Some bounds on potential evolutionary effects on ecosystem function might be derived from selection experiments which monitor both ecosystem properties and genetic composition. The data from the experiment provides some preliminary data on this issue. Using quantitative genetic projections [Eq. (2)], we calculated a < 1% per generation increase in net primary productivity in either high or low density populations (Bazzaz 1995). This result was primarily due to a low correlation between reproductive and vegetative growth enhancement, in spite of relatively high values for the total response to selection. Since data regarding traits selectively favored under rising CO2 are very scant, predictions regarding consequences of evolutionary change to ecosystem processes are largely premature. However, such lack of information does not mean that evolutionary processes could not potentially play an important role in determining future trends in ecosystem function. For example, enhanced root turnover or root exudate production (for reviews see Stulen and den Hertog, 1993; Rogers 1994) may not result in
74
enhanced plant performance. Physiological characteristics that provide the basis for these responses could thus be rapidly selected out of plant populations, depending on genetic variability and fitness consequences. If plants do indeed generally exhibit greatly enhanced levels of belowground carbon release under elevated COs, the evolutionary implications for soil microbes may be very large. In this case, short generation times could facilitate very rapid evolutionary responses, perhaps favoring the ability to rapidly utilize newly available fluxes of belowground carbon. It would be remarkable indeed if a doubling of the primary substrate for the world's most abundant enzyme did not have profound evolutionary implications for the world's autotrophic organisms. The few existing studies documenting genetic variation in plant responses to elevated COs have consistently found that not all genotypes respond positively (cf. Wulff and Alexander, 1985; Curtis 1994; Bazzaz 1995; Wayne and Bazzaz, 1995). In this regard, we may confidently predict, both on the basis of general theory and existing data, that selection under elevated COs will involve some loss of genetic variability in plant populations. This may or may not have long-term repercussions to community composition or ecosystem function. However, because COs-related selection affects all plant populations throughout the globe, such a threat to global genetic resources should be examined in earnest. Even considering massive anthropogenic changes in land use patterns and atmospheric pollutants, globally rising COs could well be the single most profound selective agent affecting our planet's autotrophic organisms.
Recent empirical work has documented substantial intraspecific genetic variation in plant growth responses to elevated CO2. thus raising the issue of selective responses to elevated CO2 in plant populations. In contrast to the well-studied case of selection for heavy-metal tolerance, selection under rising CO2 is likely to be density dependent and contingent on local availability of other plant resources. The component processes of natural selection, namely, the expression of phenotypic variation in fitness, the degree to which this variation is heritable, and genetic covariance of other traits with fitness, may each respond in predictable ways to rising COs conditions. Each of these parameters is examined, with specific reference to reaction norm experiment investigating the responses of genotypes of to population density and COs. Phenotypic variation within plant populations may be enhanced by elevated COs: for example, variation in size-related traits may increase due to accelerated divergence in growth a n d / o r accelerated competitive interactions. Heritabilities of fitness-related
5.
traits m a y o f t e n b e e x p e c t e d to b e h i g h e r , b e c a u s e e l e v a t e d C O 2 c o n s t i t u t e s a n o v e l e n v i r o n m e n t a l state n o t p r e s e n t in t h e r e c e n t selective h i s t o r i e s o f m o s t p l a n t species. G e n e t i c c o r r e l a t i o n s t r u c t u r e m a y also b e a l t e r e d , p e r h a p s d u e to a d i s r u p t i o n o f f u n c t i o n a l i n t e g r a t i o n o f p l a n t d e v e l o p m e n t . E x i s t i n g d a t a s u g g e s t t h a t r i s i n g CO2 levels will h a v e p r o f o u n d selective c o n s e q u e n c e s o n p l a n t p o p u l a t i o n s . H o w e v e r , t h e r e is very little d a t a availa b l e to s u g g e s t w h a t p h e n o t y p i c c h a r a c t e r i s t i c s will u l t i m a t e l y b e s e l e c t e d . T h e c o n s e q u e n c e s o f s u c h s e l e c t i o n o n c o m m u n i t y a n d e c o s y s t e m level p r o c e s s e s a r e also p r e s e n t l y u n c l e a r , a l t h o u g h s e l e c t i o n u n d e r r i s i n g CO2 will n e c e s s a r i l y r e s u l t in s o m e loss o f g e n e t i c v a r i a t i o n a m o n g p l a n t p o p u l a tions.
We thank F. A. Bazzaz for collaboration on the experiment reported here, and for inspiration and discussions of the ideas presented. This work was supported by a grant from the U.S. Department of Energy (DEFGO2 ER 60257).
Ackerly, D. D., and Bazzaz, F. A. (1995). Plant growth and reproduction along C O 2 gradients: Non-linear responses and implications for community change. 1, 199-208. Antonovics, J. (1978). The population genetics of mixtures. "Plant Relations in Pastures" (J. R. Wilson, ed.), pp. 233-252. CSIRO, Melbourne. Antonovics, J., Bradshaw, A., and Turner, R. G. (1971). Heavy-metal tolerance in plants. Res. 7, 1-85. Arnold, S.J., and Wade, M.J. (1984). On the measurement of natural and sexual selection: Theory. 38, 709-719. Arnone, J. A., and K6rner, C. (1993). Influence of elevated CO2 on canopy development and red:far-red ratios in two-storied stands of 94, 510-515. Avery, P.J., and Hill, W. G. (1977). Variability in genetic parameters among small populations. 29, 193-213. Avery, P.J., and Hill, W. G. (1979). Variance in quantitative traits due to linked dominant genes and variance in heterozygosity in small populations. 91, 817-844. Baker, A.J. (1987). Metal tolerance. 106 (Suppl.), 93-111. Barnola, J. M., Raynaud, D., Korotkevich, Y. S., and Lorius, C. (1987). Vostok ice core provides 160,000-year record of atmospheric CO2 329, 408-414. Bazzaz, F. A., and McConnaughay, K. D. M. (1992). Plant-plant interactions in elevated CO2 environments. 40, 547-563. Bazzaz, F. A., Garbutt, K., Reekie, E. G., and Williams, W. E. (1989). Using growth analysis to interpret competition between a C3 and a C4 annual under ambient and elevated CO2. 79, 223-235. Bazzaz, F. A., Bassow, S. L., Berntson, G. M., and Thomas, S. C. (1996). Elevated CO2 and terrestrial vegetation: Implications for and beyond the global carbon budget. "Global
76
Change in Terrestrial Ecosystems" (B. Walker and W. Steffen, eds.), pp. 43-76. Cambridge University Press, Cambridge. Bazzaz, F. A., Jasiefiski, M., Thomas, S. C., and Wayne, P. (1995). Microevolutionary responses to elevated CO2 environments in experimental populations of plants: Parallel results from two model systems. 92, 8161-8165. Beerling, D. J., and Chaloner, W. G. (1993). Evolutionary responses of stomatal density to global COs change. 48, 343-353. B6ger, P. (1980). The Os/COscycle: Development and atmospheric consequences. "Biochemical and Photosynthetic Aspects of Energy Production" (A. San Pietro, ed.), pp. 175-190. Academic Press, New York. Bowes, G. (1991 ). Growth at elevated CO2: Photosynthetic responses mediated through rubisco. 14. 795-806. Bowes, G. (1993). Facing the inevitable: Plants and increasing atmospheric COs 44, 309-332. Bradshaw, A. D., and McNeilly, T. (1991). Evolutionary response to global climate change. 67 (Suppl. 1), 5-14. Cannell, R. Q., Brun, W. A., and Moss, D. N. (1969). A search for high net photosynthetic rate among soybean genotypes. Sc/. 9, 840-841. Causton, D. R., and Venus, J. C. (1981). "The Biometry of Plant Growth." Arnold, London. Ceulemans, R., and Mousseau, M. (1994). Effects of elevated atmospheric COs on woody plants. 127, 425-446. Chapin, F. S. I., Autumn, K., and Pugnaire, F. (1993). Evolution of suites of traits in response to environmental stress. 142, $78-$92. Cheverry, J. L., Sy, M. O., Puliqueen J., and Marcellin, P. (1988). Regulation by CO2 of 1aminocyclopropane-l-carboxylic acid conversion to ethylene in climacteric fruits. 72, 535-540. Cheverud, J. M. (1988). A comparison of genetic and phenotypic correlations. 42, 958-968. Clark, A. G. (1987). Genetic correlations: The quantitative genetics of evolutionary constraints. "Genetic Constraints on Adaptive Evolution" (V. Loeschcke, ed.), pp. 25-45. SpringerVerlag, Berlin. Coleman, J. S., and Bazzaz, F. A. (1992). Effects of COs and temperature on growth and resource use of co-occurring C3 and C4 annuals. 73, 1244-1259. Cook, S. C., Lefebvre, C., and McNeilly, T. (1972). Competition between metal tolerant and 26, 366-372. normal plant populations on normal soil. Cowley, D. E., and Atchley, W. R. (1990). Development and quantitative genetics of correlation structure among body parts of 135, 242-268. Crow, J. F. (1958). Some possibilities for measuring selection intensities in man. 30, 1-13. Crow, J. F., and Kimura, M. (1970). "Introduction to Population Genetics Theory." Burgess, Minneapolis, MN. Cure, J. D., and Acock, B. (1986). Crop response to carbon dioxide doubling: A literature survey. 38, 127-145. Curtis, P. S., Snow, A. A., and Miller, A. S. (1994). Genotype-specific effects of elevated CO2 on fecundity in wild radish 97, 100-105. Davies, M. S., and Snagdon, R. W. Rapid population differentiation in a mosaic environment. III. Measures of selection pressures. 36, 59-66. de Jong, G., and van Noordwijk, A.J. (1992). Acquisition and allocation of resources: Genetic (co)variances, selection, and life histories. 139, 749-770. Delgado, E., and Medrano, H. (1991). Field performance and leaf characteristics of L. genotypes selected by low COs survival. 25, 313-322.
5.
77
Delgado, E., Parry, M. A. J., Vadell, J., Lawlor, D. W., Keys, A.J., and Medrano, H. (1992a). Effects of water stress on photosynthesis, leaf characteristics, and productivity of field-grown L. genotypes selected for survival at low CO2 43, 1001-1008. Delgado, E., Azc6n-Bieto, J., Aranda, X., Palaz6n, J., and Medrano, H. (1992b). Leaf photosynthesis and respiration of high CO2-grown tobacco genotypes selected for survival in a low CO2 atmosphere. 98, 949-954. Delgado, E., Parry, M. A.J., Lawlor, D. W., Keys, A.J., and Medrano, H. (1993). Photosynthesis, ribulose-l,5-bisphosphate carboxylase, and leaf characteristics of L. genotypes selection by survival at low CO2 concentrations. J. 44, 1-7. Enoch, H. Z., and Honour, S.J. (1993). Significance of increasing ambient CO2 for plant growth and survival, and interactions with air pollution. "Interacting Stresses on Plants in a Changing Climate" (M. B. Jackson and C. R. Black, eds.), pp. 51-76. SpringerVerlag, Berlin. Fajer, E. D., Bowers, M. D., and Bazzaz, F. A. (1992). The effect of nutrients and enriched CO2 environments on production of carbon-based allelochemicals in A test of the carbon/nutrient balance hypothesis. 140, 707-723. Falconer, D. S. (1952). The problem of environment and selection. 86, 293-298. Falconer, D. S. (1989). "Introduction to Quantitative Genetics," 3rd ed. Longman, London. Farnsworth, E. J., and Bazzaz, F. A. (1995). Inter- and intra-generic differences in growth, reproduction, and fitness of nine herbaceous annual species grown in elevated CO2 environments. 104, 454-466. Fisher, R. A. (1930). "The Genetical Theory of Natural Selection." Clarendon, Oxford. Gadgil, M., and Bossert, W. H. (1970). Life historical consequences of natural selection. 104, 1-24. Garbutt, tC, and Bazzaz, F. A. (1984). The effects of elevated CO2 on plants. III. Flower, fruit and seed production and abortion. 98, 433-446. Geber, M. A. (1990). The cost of meristem limitation in Negative genetic 44, 799-819. correlations between fecundity and growth. Geber, M. A., and Dawson, T. E. (1993). Evolutionary responses of plants to global change. "Biotic Interactions and Global Change" (P. M. Kareiva, J. G. Kingsolver, and R. B. Huey, eds.), pp. 179-197. Sinauer, Sunderland, MA. Giesel, J. T., Murphy, P. A., and Manlove, M. N. (1982). The influence of temperature on genetic interrelationships of life history traits in a population of What tangled data sets we weave. 119, 464-479. Gomulkiewicz, R., and Kirkpatrick, M. (1992). Quantitative genetics and the evolution of reaction norms. 46, 390-411. Harper, J., and White, J. (1974). The demography of plants. 5, 419-463. Hartnett, D. C. (1990). Size-dependent allocation to sexual and vegetative reproduction in four clonal composites. 84, 254-259. Hickey, D. A., and McNeilly, T. (1975). Competition between metal tolerant and normal plant populations: A field experiment on normal soil. 29, 458-464. Hill, W. G., and Caballero, A. (1992). Artificial selection experiments. 23, 287-310. Holsinger, K. E., and Pacala, S. W. (1990). Multiple-niche polymorphisms in plant populations. 135, 301-309. Horton, R. F. (1985). Carbon dioxide flux and ethylene production in leaves. "Ethylene and Plant Development" (J. A. Roberts and G. A. Tucker, eds.), pp. 37-46. Butterworths, London. Houghton, J. T., Callander, B. A., and Varney, S. K. (1992). "Climate Change 1992: The Supplementary Report to the IPCC Scientific Assessment." Cambridge Univ. Press, Cambridge,UK. Houle, D. (1992). Comparing evolvability and variability of quantitative traits. 130, 195-204.
Itoh, Y., and Yamada, Y. (1990). Relationships between genotype x environment interaction and genetic correlation of the same trait measured in different environments. 80, 11-16. Kawano, S., Hayashi, S., Aria, H., Yamamoto, M., Takasu, H., and Oritani, T. (1989). Regulatory mechanisms of reproductive effort in plants. III. Plasticity in reproductive energy allocation and propagule output of two grass species, cv. Akihikarai and cultivated at varying densities and nitrogen levels, and the evolutionary-ecological implications. 4, 75-99. Kimball, B. A. (1983). Carbon dioxide and agricultural yield: An assemblage and analysis of 430 prior observations. 75, 779-788. Kirkpatrick, M., and Lofsvold, D. (1989). The evolution of growth trajectories and other complex quantitative characters. 31, 778-783. Kirkpatrick, M., and Lofsvold, D. (1992). Measuring selection and constraint in the evolution of growth. 46, 954-971. Kirkpatrick, M., Lofsvold, D., and Bulmer, M. (1990). Analysis of the inheritance, selection, and evolution of growth trajectories. 124, 979-993. Kirkpatrick, M., Hill, W. G., and Thompson, R. (1994). Estimating the covariance structure of traits during growth and aging, illustrated with lactation in dairy cattle. 64, 57-69. Koyama, H., and Kira, T. (1956). Intraspecific competition among higher plants. VIII. Frequency distribution of individual plant weight as affected by the interaction between plants. D 7, 73-94. Lande, R. (1979). Quantitative genetic analysis of multivariate evolution, applied to brain: Body size allometry. 33, 402-416. Lande, R. (1982). A quantitative genetic theory of life history evolution. 63, 607-615. Eomnicki, A. (1988). "Population Ecology of Individuals." Princeton Univ. Press, Princeton, NJ. Lynch, M., and Lande, R. (1993). Evolution and extinction in response to environmental change. "Biotic Interactions and Global Change" (P. M. Kareiva, J. G. Kingsolver, and R. B. Huey, eds.), pp. 234-250. Sinauer, Sunderland, MA. MacNair, M. R. (1977). Major genes for copper tolerance in 268, 428-430. MacNair, M. R. (1991). The genetics of metal tolerance in natural populations. "Heavy Metal Tolerance in Plants: Evolutionary Aspects" (A. J. Shaw, ed.), pp. 235-253. CRC Press, Boca Raton, FL. MacNair, M. R., Smith, S. E., and Coumbes, Q.J. (1993). Heritability and distribution of variation in degree of copper tolerance in at Copperopolis, California. 71, 445-455. Maxon-Smith, J. W. (1977). Selection for response to CO2-enrichment in glasshouse lettuce. 17, 15-22. Mazer, S. J., and Schick, C. T. (1991). Constancy of population parameters for life-history and floral traits in L. II. Effects of planting density on phenotype and heritability estimates. 45, 1888-1907. McLauglin, S. B., and Norby, R.J. (1991). Atmospheric pollution and terrestrial vegetation: Evidence of changes, linkages, and significance to selection processes. "Ecological Genetics and Air Pollution" (G. E. Taylor, L. F. Pitelka, and M. T. Clegg, eds.), pp. 61-102. Springer-Verlag, Berlin. McNeilly, T. (1968). Evolution in closely adjacent plant populations. III. tenuis on a small copper mine. 23, 99-108. Menz, K. M., Moss, D. N., Cannell, R. Q~, and Brun, W. A. (1969). Screening for photosynthetic efficiency. 9, 692-694.
5.
79
Mitchell-Olds, T., and Rutledge,J.J. (1986). Quantitative genetics in natural plant populations: A review of the theory. 127, 379-402. Morse, S. R., and Bazzaz, F. A. (1994). Elevated COz and temperature alter recruitment and size hierarchies in C3 and C4 annuals. 75, 966-975. Musgrave, M. E., Strain, B. R., and Siedow, J. N. (1986). Response of two pea hybrids to CO2 enrichment: A test of the energy overflow hypothesis for alternative respiration. 83, 8157-8161. Namkoong, G., Bishir, J., and Roberds, J. H. (1993). Evolutionary effects of density-dependent selection in plants. 62, 57-62. Noordwijk van, A.J., and de Jong, G. (1986). Acquisition and allocation of resources: Their influence on variation in life history tactics. 128, 137-142. Pefiuelas, J., and Matamala, R. (1990). Changes in N and S leaf content, stomatal density and specific leaf area of 14 plant species during the last three centuries of CO2 increase. 41, 1119-1124. Poorter, H. (1993). Interspecific variation in the growth response of plants to an elevated ambient CO2 concentration. 104/105, 77-97. Reekie, J. Y. C., Hicklenton, P. R., and Reekie, E. G. (1994). Effects of elevated CO2 on time 72, 533-538. of flowering in four short-day and four long-day species. Rice, K.J. (1990). Reproductive hierarchies in Effects of variation in plant density and rainfall distribution. 71, 1316-1323. Riska, B. (1989). Composite traits, selection response, and evolution. 43, 1172-1191. Roff, D. A. (1995). The estimation of genetic correlations from phenotypic correlations: A test of Cheverud's conjecture. 74, 481-490. Rogers, H. H., Cure, J. D., Thomas, J. F., and Smith, J. M. (1984). Influence of elevated CO2 24, 361-366. on growth of soybean plants. Rogers, H. H., Runion, G. B., and Krupa, S. V. (1994). Plant responses to atmospheric CO2 enrichment with emphasis on roots and the rhizosphere. 83, 155-189. Roose, M. L. (1991). Genetics of response to atmospheric pollutants. "Ecological Genetics and Air Pollution" (G. E. Taylor, L. F. Pitelka, and M. T. Clegg, eds.), pp. 111-126. Springer-Verlag, Berlin. Samson, D. A., and Werk, K. S. (1986). Size-dependent effects in the analysis of reproductive effort in plants. 127, 667-680. Schimper, A. F. W. (1935). "Pflanzengeographie auf physiologischer Grundlage," 3rd ed. Fischer, Jena. Schmalhausen, I. I. (1949). "Factors of Evolution: The Theory of Stabilizing Selection." Univ. Chicago Press, Chicago. Service, P. M., and Rose, M. R. (1985). Genetic covariation among life-history components: The effect of novel environments. 39, 943-945. Shaw, A.J., ed. (1991a). "Heavy Metal Tolerance in Plants: Evolutionary Aspects." CRC Press, Boca Raton, FL. Shaw, R. G. (1991b). The comparison of quantitative genetic parameters between populations. 45, 143-151. Shaw, R. G., and Platenkamp, G. A.J. (1993). Quantitative genetics of response to competitors in 47, 801-812. Smart, D. R., Chatterton, N. J., and Bugbee, B. (1994). The influence of elevated CO2 on nonstructural carbohydrate distribution and fructan accumulation in wheat canopies. 17, 435-442. Spicer, R. A., and Corfield, R. M. (1992). A review of terrestrial and marine climates in the Cretaceous with implications for modelling the 'Greenhouse Earth.' 129, 169-180.
Stitt, M. (1991). Rising CO2 levels and their potential significance for carbon flow in photosynthetic cells. 14, 741-762.
80
Strain, B. R. (1991). Possible genetic effects of continually increasing atmospheric CO2. "Ecological Genetics and Air Pollution" (G. E. Taylor, L. F. Pitelka, and M. T. Clegg, eds.), pp. 177-202. Springer-Verlag, Berlin. Strain, B. R., and Cure, J. D. (1985). "Direct Effects of Increasing Carbon Dioxide on Vegetation." United States Dept. of Energy, Carbon Dioxide Research Division, DOE/ER 0238, Office of Energy Research, Washington, DC. Stulen, I., and den Hertog, J. (1993). Root growth and functioning under atmospheric CO2 enrichment. 104/105, 99-115. Tans, P. P., Fung, I. Y., and Takahashi, T. (1990). Observational constraints on the global atmospheric CO2 budget. 247, 1431-1438. Thomas, S. C., and Bazzaz, F. A. (1993). The genetic component in plant size hierarchies: Norms of reaction to density in a species. 63, 231-249. Thomas, S. C., and Bazzaz, F. A. (1996). Elevated CO~ and leaf shape: Are dandelions getting toothier? 83, 106-111. Thomas, S. C., and Weiner, J. (1989). Growth, death and size distribution change in an population. 77, 524-536. Tolley, L. C., and Strain, B. R. (1984). Effects of CO2enrichment on growth of and seedlings under different irradiance levels. 14, 343-350. Tremmel, D. C., and Patterson, D. T. (1994). Effects of elevated CO~ and temperature on development in soybean and five weed. 74, 43-50. Turelli, M. (1988). Phenotypic evolution, constant covariances, and the maintenance of additive variance. 42, 1342-1347. Uchmafiski, J. (1985). Differentiation and frequency distributions of body weights in plants and animals. 310, 1-75. Via, S. (1987). Genetic constraints on the evolution of phenotypic plasticity. "Genetic Constraints on Adaptive Evolution" (V. Loeschcke, ed.), pp. 47-71. Springer-Verlag, Berlin. Via, S., and Lande, R. (1985). Genotype-environment interaction and the evolution of phenotypic plasticity. 39, 505-522. Von Faber, F. C. (1925). Untersuchungen fiber die Physiologie der javanischen SolfatarenPhflanzen. 18/19, 89. Von Faber, F. C. (1927). Die Kraterpflanzen Javas in physiologisch-tkologischer Beziehung. no. 1. Wade, M.J. (1985). Soft selection, hard selection, kin selection, and group selection. 125, 61-73. Wallace, B. (1968). Polymorphism, population size, and genetic load. "Population Biology and Evolution" (R. C. Lewontin, ed.), pp. 87-108. Columbia Univ. Press, New York. Walters, M. B., Kruger, E. L., and Reich, P. B. (1993). Growth, biomass distribution and CO2 exchange of northern hardwood seedlings in high and low light: Relationships with successional status and shade tolerance. 94, 7-16. Wayne, P. M., and Bazzaz, F. A. (1995). Seedling density modifies the growth response of yellow birch maternal families to elevated carbon dioxide. 1,315-324. Weiner, J. (1985). Size hierarchies in experimental populations of annual plants. 66, 743-752. Weiner, J. (1988). The influence of competition on plant reproduction. "Reproductive Ecology of Plants: Patterns and Strategies" (J. Lovett-Doust and L. Lovett-Doust, eds.), pp. 228-245. Oxford Univ. Press, New York. Weiner, J. (1990). Asymmetric competition in plant populations. 5, 360-364. Weiner, J., and Thomas, S. C. (1986). Size variability and competition in plant monocultures. 47, 211-222. Wilkinson, G. S., Fowler, K., and Partridge, L. (1990). Resistance of genetic correlation structure to directional selection in 44, 1990-2003.
5.
01
Willis, J. H., Coyne, J. A., and Kirkpatrick, M. (1991). Can one predict the evolution of quantitative characters without genetics? 45, 441-444. Wilson, J. B., and Levin, D. A. (1986). Some genetic consequences of skewed fecundity distributions in plants. 73, 113-121. Woodward, F. I. (1987). Stomatal numbers are sensitive to increases in COzfrom pre-industrial levels. 327, 617-618. Wulff, R. D., and Alexander, H. M. (1985). Intraspecific variation in the response to COz enrichment in seeds and seedlings of 66, 458-460. Young, H.J., Stanton, M. L., Ellstrand, N. C., and Clegg, J. M. (1994). Temporal and spatial variation in heritability and genetic correlations among floral traits in wild radish. 73, 298-308. Zeng, Z.-B. (1988). Long-term correlated response, interpopulation covariation, and interspecific allometry. 42, 363-374. Ziska, L. H., and Teramura, A. H. (1992). Intraspecific variation in the response of rice (Oryza to increased COzmPhotosynthetic, biomass and reproductive characteristics. 84, 269-276.
This Page Intentionally Left Blank
II Community-Level Responses
This Page Intentionally Left Blank
The Changing Vegetation of Europe: What Is tile Role of Elevated Carbon Dioxide?
Although there is no doubt that atmospheric carbon dioxide concentrations are rising rapidly (Keeling 1982) and will continue to do so (IPCC, 1992), there is considerable uncertainty with regard to the consequences for plant communities and ecosystems (Korner, 1993). The difficulties in predicting the impact of elevated CO2 first become evident in the laboratory or growth chamber and multiply as we move outdoors and begin to consider large-scale processes operating over extended periods of time. From laboratory studies we know that plant species differ in responsiveness to elevated CO2 (Hunt 1991; Poorter, 1993) and we may be certain that patterns of response detected under controlled conditions will be subject, in more natural habitats, to the modifying effects of other environmental factors, some of which (temperature, rainfall, UV-B) are themselves implicated in global environmental change. At this point in the analysis it is tempting to conclude that the task of predicting the ecological impacts of rising CO2 falls almost exclusively in the domain of plant physiology (e.g., Schulze and Mooney, 1993). Little doubt remains that physiological insights are needed for a mechanistic and predictive understanding of vegetation responses to elevated CO2. However, in this chapter I shall argue that the most urgent requirement is to place CO2 research in the context of other global and regional changes in vegetation driven by more powerful forces. and
85
Copyright 9 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.
At the present time, the most potent forces for change acting on vegetation are the effects of land use. These arise from the direct effects of human activity (habitat modification by agriculture, forestry, industry, human settlements, overgrazing) and indirect effects (eutrophication through groundwater and atmospheric pollutants, and phytotoxicity resulting from aerial and soil contamination). Reviewed on a world scale, the most consistent effect of these phenomena is the inexorable replacement of mature, often species-rich ecosystems by early successional states in which the vegetation is composed of recently established, fast-growing clonal herbs and ephemeral species. This process has two important implications for studies which seek to predict the impacts of rising CO2. The first is the notion that vegetation is already experiencing such radical processes of change that impacts of CO2 are perhaps most appropriately analyzed as a fine-tuning of the rates and trajectories of changes which are already well advanced and are driven by land use. The second implication arises from the faster growth rates and reduction in the average life span in the constituent species of vegetation arising from modern forms of intensive or disruptive land use (Grime 1988). Later in this chapter we will examine evidence that these expanding species are more responsive to elevated CO2. Moreover, the higher rates of population turnover characteristic of the vegetation of disturbed and intensively exploited landscapes create conditions in which the plant cover is likely to respond more quickly to selection driven by elevated CO2 either by permitting more invasions and extinctions or by allowing rapid genetic changes within component populations. Hence land use is likely to be an essential factor in any calculations of the direction and rate of vegetation responses to elevated CO2. In order to explore further the interaction between changing land use and CO2 responses, let us now look at recent evidence of floristic change in Western Europe.
In a comparative discriminant analysis of the functional traits of increasing and decreasing species in the vascular plant floras of the British Isles, The Netherlands, and West Germany (Thompson, 1994), three main conclusions have been drawn: 1. A key measure discriminating between increasing and decreasing species is S radius. S radius is a measure of proximity to the stress-tolerant corner of a CSR strategic triangle of plant functional types (Grime, 1974; Grime 1988); a large S radius is correlated with low growth rate, low rates of tissue turnover, and low mineral nutrient requirements. In densely
6.
populated England, the Netherlands, and to a lesser extent, West Germany, decreasing species are stress-tolerators (large S radius), whereas increasing species are fast-growing and typical of eutrophic, disturbed habitats. 2. In the sparsely populated northern and western regions of the British Isles differences between "winners" and "losers" are very slight. 3. Surprisingly, regenerative attributes (seed weight, seed persistence in soil, wind dispersal) are very poor predictors of success and failure in the modern northwest European landscape. The same sources are used in Fig. I to plot the mean Sradius of increasing and decreasing species against human population density in seven European countries. Thompson's (1994) interpretation of these results accords with an earlier hypothesis of Hodgson (1986a,b) in suggesting that in the densely populated countries of Western Europe, land use is increasingly polarizing the flora into two parts. The successful, fast-growing part is tolerant of human activities and is ecologically attuned to intensively managed grassland, arable fields, road verges, gardens, spoil, and urban waste-
Figure 1 Relationship between mean S radius of increasing and decreasing species and h u m a n population density in seven European countries. S radius of the two groups is not significantly different in Scotland, N. Ireland, or Wales. The two groups are significantly different in Republic of Ireland (P = 0.049), England, western Germany, and The Netherlands (all P < 0.001).
88
j. t'. ~ m e
land. Because these habitats are common, and because soils, seeds, and plant fragments are moved freely between them by human agencies, these plants are highly mobile, rapidly colonizing new sites as they become available. Species of this type will have little difficulty in migrating in response to land use change. In contrast, the slow-growing, stress-tolerant part of the flora, typical of unimproved grassland, lowland heath and old woodland, is increasingly excluded from the wider landscape. How will rising COs influence these vegetation changes resulting from land use? Has responsiveness to COs concentration already played a significant role in the promotion of fast-growing, resource-demanding species in intensively developed landscapes? To address these questions it is necessary to consider our present understanding of how different types of plants respond to elevated COs.
Over the period from 1987 to 1994 a comprehensive screening program (Integrated Screening Programme, ISP) was conducted at Sheffield (with links to other centers) to compile standardized information on the laboratory characteristics of a large number of common herbaceous vascular plants of the British Isles. As part of the ISP a series of CO2 screening experiments were carried out at Horticulture Research International's site at Littlehampton (Hunt 1991, 1993, and 1995). These experiments covered 36 different species; many of them were later restudied to provide confirmation. Strong responses to elevated CO2 were recorded when plants were held at 18~ In some species there was a 27% increase in biomass after only 8 weeks' growth. However, such levels of response mainly occurred in robust, fast-growing perennials (e.g., of the kind which often dominate vegetation in productive and undisturbed habitats, such as river banks and recently abandoned farmland and gardens. These plants exhibit sustained vegetative growth and develop large roots, shoots, and storage organs prompting the hypothesis that responsiveness to elevated COs may be related to possession of rapidly expanding carbon sinks. This interpretation is supported by the observation (Hunt 1995) that the benefit of elevated COs to the fast-growing ephemeral, is not sustained beyond the early vegetative phase of development. A consistent feature of the ISP results has been the low level of response by slowgrowing evergreen species of unproductive habitats (i.e., high S radius plants). These results suggest that the capacity to respond strongly to elevated COs is prominent among the species that are currently expanding in abundance in the more intensively developed landscapes of heavily populated countries of Western Europe.
89
6.
The need for caution in extrapolating from the responses of individual potted plants in growth chambers to the dynamics of communities of plants in the field is self-evident. The majority of laboratory studies (including the ISP) are designed to standardize conditions and facilitate comparison and interpretation. Realism is sacrificed to the extent that the complexities of season, weather, soil microbiology, decomposition, nutrient dynamics, interspecific competition, and plant-animal interactions are excluded. Some insights into the complications that can arise when attention is turned to "real" systems are evident in Fig. 2 (A,B), which presents results from an experiment (Diaz 1993) in which early-successional plant communities were allowed to reassemble from natural seed banks and soils removed to
Shoot biomass n
s
Cover
~
~ n s
Dominance Carbohydrates Nitrogen SOIL 9
Microbial N ..
-40
-30
-20
-10
0
10
20
30
MicrobialC 40
% change under doubling CO2 Figure 2 (A) Responses of and soil microflora grown in microcosms to a doubling of atmospheric CO2 (700 ppm) as compared to controls at 350 ppm. Vegetation was allowed to develop for 84 days by natural recruitment from seed banks in soils removed from a tall herb community in Derbyshire and placed in microcosms (6 replicates per treatment) in cabinets without nutrient addition. Shoot biomass was measured as milligram dry weight, cover as n u m b e r of touches in a point-quadrat analysis, dominance as biomass of R. total community biomass, carbohydrates (starch + glucose + sucrose) as milligram/gram fresh weight and nitrogen as milligram/gram dry weight of fully expanded young leaves, microbial C and N as milligram/gram dry soil; ns, nonsignificant; *, P < 0.05; **, P < 0.01 (ANOVA). (B) Effects of atmospheric doubling of CO2 concentration (ppm) and fertilizer addition on foliar N content of grown in microcosm for 60 days. Deionized water (control) and full-strength Rorison solution (fertilized) were added throughout the experiment as 100 ml per microcosms every 4 days. Bars designated by the same letter are not different at P < 0.05 (ANOVA).
90
j.P. Gr/me 25 o l
2 0 ow
15 o l
Q ~
10~
z
5 --
0
~ 350 ppm
700 ppm
Continued
laboratory microcosms providing ambient and elevated
CO
2 concentra-
tions.
In the example shown in Fig. 2A the potentially responsive showed leaf stunting when exposed to elevated CO2. These symptoms coincided with carbohydrate accumulation and reduced levels of foliar nitrogen. From soil analyses, circumstantial evidence showed that deficiency had been induced by export of carbohydrate from roots to soil and subsequent sequestration of nitrogen in a massively expanded soil microflora. This hypothesis was supported by the results of a second experiment (Fig. 2B) in which the leaf nitrogen content of plants grown at elevated CO2 did not increase when additional mineral nutrients were added to the soil. These results are a useful reminder of the feedbacks that can occur in response to elevated CO2. Species responses within ecosystems are likely to be determined by factors such as soil microbiology, which are additional to those (e.g., sink strength) that emerge as key factors in simple laboratory assays.
6.
91
This chapter assembles circumstantial evidence that the types of herbaceous plants which are currently expanding in a b u n d a n c e in heavily populated countries of Western Europe are also strongly responsive to elevated concentrations of carbon dioxide. This evidence suggests that responses to carbon dioxide may be contributing to the recent success of these species. An attractive feature of this hypothesis is the observation that the responsive species occupy habitats in which other essential resources are relatively a b u n d a n t and are unlikely, therefore, to limit the stimulatory effect of carbon dioxide enrichment. There are several reasons why we should treat this interpretation with caution. The data comparing responses to elevated CO2 are based on simplified laboratory conditions and refer mainly to herbaceous plants grown for short periods of time with nonlimiting supplies of moisture and mineral nutrients. It may be especially i m p o r t a n t to recognize that on natural soils responses to CO2 may be limited by n u t r i e n t stress arising from microbial sequestration of mineral nutrients. In attempts to u n d e r s t a n d the interactions between land use and elevated CO2 it may be essential to differentiate between short and long cycles of secondary succession. Where vegetation destruction and mineral n u t r i e n t release occurs frequently, it seems likely that selection associated with elevated CO2 may act in parallel with eutrophication in p r o m o t i n g fast-growing clonal herbs. However, it can be argued that where the intervals between major disturbance events are longer (coppiced woodlands, plantations, long-rotation grasslands, b u r n e d heathlands), elevated CO2 could accelerate succession by favoring plants that develop m o r e slowly but provide m o r e substantial sinks for carbon and mineral nutrients. These hypotheses require experiments of sufficient scale and duration to allow rigorous tests of the effects of CO2 e n r i c h m e n t on successional processes.
This chapter draws on information collected with colleagues at UCPE as part of the Integrated Screening Programme and the Terrestrial Initiative in Global Environmental Research, both of which are supported by the Natural Environment Research Council.
Diaz, S., Grime,J. P., Harris,J., and McPherson, E. (1993). Evidence of a feedback mechanism limiting plant response to elevated carbon dioxide. 364, 616-617.
Grime, J. P. (1974). Vegetation classification by reference to strategies. 250, 26-31. Grime, J. P., Hodgson, J. G., and Hunt, R. (1988). "Comparative Plant Ecology: A Functional Approach to Common British Plants." Unwin Hyman, London. Hodgson, J. G. (1986a). Commonness and rarity in plants with special reference to the Sheffield flora. I. The identity, distribution, and habitat characteristics of the common and rare species. 36, 199-252. Hodgson, J. G. (1986b). Commonness and rarity in plants with special reference to the Sheffield flora. II. The relative importance of climate, soils, and land use. 36, 254-274. Hunt, R., Hand, D. W., Hannah, M. A., and Neal, A, M. (1991). Response to COz enrichment in 27 herbaceous species. 5, 410-421. Hunt, R., Hand, D. W., Hannah, M. A., and Neal, A. M. (1993). Further responses to CO2 enrichment in British herbaceous species. 7, 661-668. Hunt, R., Hand, D. W., Hannah, M. A., and Neal, A. M. (1995). Temporal and nutritional influences on the CO2 response in selected British grasses. 76, 207-216. International Panel on Climate Change Report (1992). "Climate Change 1992: The Supplementary Report to the IPCC Scientific Assessment." Cambridge Univ. Press, Cambridge, UK. Keeling, C. D., Bacastow, R. B., and Whorf, T. P. (1982). Measurements of the concentration of carbon dioxide at Mauna Loa Observatory, Hawaii. "Carbon Dioxide Review" (W. C. Clark, ed.). Korner, C. H. (1993). CO2 fertilization: The great uncertainty in future vegetation development. "Vegetation Dynamics and Global Change" (A. M. Solomon and H. H. Shugart, eds.), pp 53-70. Chapman and Hall, London. Poorter, H. (1993). Interspecific variation in the growth response of plants to an elevated ambient CO2 concentration. 104/105, 77-97. Schulze, E. D., and Mooney, H. A., eds. (1993). "Design and Execution of Experiments on CO2 Enrichment." Commission of the European Communities, Brussels. Thompson, K. (1994). Predicting the fate of temperate species in response to human disturbance and global change. "NATO Advanced Research Workshop on Biodiversity, Temperate Ecosystems, and Global Change" (T. J. B. Boyle and C. E. B. Boyle, eds.), pp. 6176. Springer Verlag, Berlin.
7 Changing Community Composition and Elevated CO2
Anthropogenic forcing of biogeochemical cycles, such as the global carbon cycle, can potentially alter community composition (e.g., Bazzaz and Carlson, 1984; Bazzaz and Fajer, 1992; Melillo 1990; Gates, 1994). Likewise, alterations in the species composition of communities can alter ecosystem processes (Tilman and Downing, 1994; Naeem 1994a, 1995; Schulze and Mooney, 1993). Together these interactions describe a potential feedback between changing levels of biotic diversity and changing levels of CO2. This potential feedback between biotic diversity and the carbon biogeochemical cycle may be significant on a global scale. Ecosystems worldwide are simultaneously experiencing both anthropogenic alterations in diversity (e.g., Wilson and Peter, 1988; Soul~, 1991; Ehrlich and Wilson, 1991; Groombridge, 1992; Sisk 1994) and anthropogenic increases in atmospheric CO2. Such feedbacks between biotic factors and biogeochemical cycles are important for modeling global change (Lashof, 1989; Schneider, 1992). Empirical evidence for the feedback between diversity and carbon cycling is limited. We present, however, evidence from two mesocosm experiments using a controlled environmental facility referred to as the "Ecotron" 93
Copyright 9 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.
94 (Lawton 1993) which support the possibility of interactions between community composition and C02 flux through an ecosystem (biomass accumulation) and the possible feedback between them. In the first experiment, henceforth the "biodiversity experiment," community composition was manipulated and ecosystem biomass accumulation was measured as a response variable. In the second, henceforth the "elevated C02 experiment," C02 levels were manipulated and community composition and biomass accumulation were measured as response variables. We discuss the implications of these results for global-change research on elevated C02.
A. The Biodiversity Experiment The Ecotron is described in both Lawton (1993) and Thompson (1993). The materials, methods, and results concerning the association between biodiversity and ecosystem functioning (the biodiversity experiment) can be found in Naeem (1994a,b, 1995). Briefly, in this experiment plant and animal species composition within model terrestrial communities were experimentally manipulated to produce 3 levels of diversity: low (4 replicates), intermediate (4 replicates), and high diversity (6 replicates) mesocosms (see Table I for a list of species). The Ecotron consists of 16 chambers which are divided into 2 banks of 8 chambers each. Because each bank is serviced by separate environmental regulating machinery, random assignment of treatments to chambers was statistically blocked by bank to test for block effects. Diversity was manipulated in all trophic levels (Table I). CO2 flux was measured continuously throughout the experiment after Day 120 (total experiment duration was 210 days), but only data collected over 48-hour periods, on a biweekly basis, when chambers were closed and undisturbed by the activities of researchers, were used for statistical analyses. B. The Elevated CO2 Experiment This experiment conducted in the Ecotron consisted of one bank of chambers with eight replicate mesocosms exposed to ambient levels of CO2 and eight replicates of the second bank exposed to ambient +200 ppm CO2. CO2 flux was measured as above, but data used for statistical analyses were collected weekly. C. Differences between the Two Experiments Although the edaphic, daily temperature and humidity conditions were the same for both experiments, the elevated CO2 experiment differed from the biodiversity experiment in several significant ways. First, more light
Biodiversity Species
Low
Int.
High
C02
Plants
u
Animals Mollusk
Earthworm W o o d louse Collembola
cf. cf. Insect herbivores
Parasitoids
a Note similarities in community composition between intermediate (Int.) diversity treatment and the current COs experiment.
96 (18% full sunlight) was provided in the elevated C O 2 experiment (10% full sunlight was provided for the biodiversity experiment). Second, though trophic complexity and number of species were similar between the elevated CO2 experiment and the intermediate-diversity treatment of the biodiversity experiment, some species substitutions were made (Table I). Third, unlike the biodiversity experiment, the Ecotron does not permit random assignment of atmospheric treatments, such as ambient versus elevated CO2, to replicate mesocosms. This results in a pseudoreplicated design which necessitates a follow-up experiment in which treatments to banks (enhanced and ambient CO2) are reversed (currently running, but not reported here). Fourth, injection of augmented CO2 into the atmospheric stream servicing the elevated CO2 communities occasionally resulted in spurious levels of recorded CO2 exhausts in five of the eight chambers during the period of measurement reported here (the problem has since been rectified). We therefore estimated COz flux using only the chambers (chambers 3, 5, and 6) whose CO2 flux readings were not suspect. Finally, periodic harvests were conducted for standing biomass estimation. This practice periodically reduced net CO2 flux.
The biodiversity experiment showed that manipulating community composition alters biomass accumulation, as measured by CO2 flux (Fig. 1). In our model system, higher diversity assemblages sequestered more carbon, a result that is robust for other combinations of plant species (Naeem 1994a, 1995). Although communities rarely differed within intervals, over the duration of the experiment a repeated measures analysis of variance (RMANOVA) showed that higher diversity communities sequestered more carbon than lower diversity communities (RMANOVA; among group = 2, 8; F = 4.5; P < 0.05; interaction = 22, 88; F = 2.7; P < 0.001). Results from the CO2 experiment show that overall levels of biomass accumulation (as measured by COz flux) are statistically different between ambient and elevated CO2 communities of intermediate diversity (RMANOVA; = 26, 1; F = 5.6; P < 0.001, Fig. 1). Note that this statistic was obtained by the conservative method of comparing weekly mean CO2 fluxes of elevated CO2 communities with the weekly mean COz fluxes of ambient CO2 communities. Periodic harvests, estimates of community composition, and additional analyses reveal that a complex number of ecological changes over the course of the experiment are associated with these results. These results are still in a preliminary form and will be reported in detail elsewhere.
7.
97
Figure 1 CO~ flux (mol m -2 d -x) in the biodiversity and elevated experiments. Time is measured in weeks from start of experiment. For the purposes of comparison, this figure shows only a partial set (that which corresponds temporally to the shorter sequence in the biodiversity experiment) of the data from the elevated CO2 experiment. Arrows indicate harvest dates ("winters") in which most of the vegetation is removed and the system is allowed to recover. The top graph (a) plots results from the elevated COs experiment. The lower graph (b) plots results from the biodiversity experiment. Note that negative values indicate sequestered carbon whereas positive values indicate greater microbial, invertebrate, and plant respiration than photosynthetic activity. Note also the different scales and note the greater negative values for the elevated CO2 experiment are partly the result of additional light provided in this experiment. Error bars represent one SE. AMB, ambient CO2; ELEV, elevated CO2; HI, high diversity; INT, intermediate diversity; LOW, low diversity.
Results from these experiments suggest that declining diversity within an ecosystem can decrease biomass accumulation and, conversely, that elevated CO2 can change the biomass accumulation of an ecosystem. Interpreted more broadly, our results suggest that ecosystem response to elevated CO2 is a function of both diversity and CO2 levels.
Though biomass accumulation may be trivially a function of the species found within an ecosystem, the response of biomass accumulation to random declines in diversity is not a trivial problem. For example, neither theory nor experiments in intercropping provide steadfast rules for how diversity and yield (biomass accumulation) might be associated in even simple agroecosystems (Vandermeer, 1989; Swift and Anderson, 1993). Our biodiversity experiment suggested that if a decline in plant diversity is associated with decreasing interception of light by the canopy, then CO2 sequestration by an ecosystem may decline. Our elevated CO2 experiment, however, suggests that changes in ecosystem biomass accumulation generated by elevated COz could compensate for such a loss of carbon sequestration. Although we cannot readily extrapolate from these simple experimental systems to larger more complex naturally occurring systems, these experiments point to an often neglected possibility that understanding how ecosystems will respond to elevated COz will be a function of how diversity changes over the next few decades. Our results suggest that manipulating both CO2 and community composition may improve our understanding of global change. Most research on the ecological consequences of elevated CO2 has been conducted using, on average, 550-700 ppm CO2, or levels likely to occur 50-60 years from now (Houghton 1990) and this research has rarely manipulated community composition. By the time these 50-60 years pass, changing COz, in addition to many other globally changing factors (e.g., N fertilization and habitat fragmentation) (Vitousek, 1994), may have already changed community composition. Indeed, some authors (e.g., K6rner, Chapters 11 and 28; Polley 1994, and Chapter 12) have argued that some of these effects have already occurred. Even without the effects of elevated CO2, the community composition of most ecosystems is likely to be substantially altered in the near future (e.g., Wilson and Peter, 1988; Soul6, 1991; Ehrlich and Wilson, 1991; Groombridge, 1992; Sisk 1994; Lawton and May, 1995). Understanding the interactions and feedbacks between ecosystem processes and community composition and how human impacts contribute to these processes will prove useful for predicting and understanding the effects of elevated CO2 on global change.
Current research on the ecological consequences of elevated C O 2 supports two direct interactions between communities and atmospheric CO2. First, altering levels of CO2 can change the relative abundance of species in communities. Second, altering the species composition of communities can change the ecosystem's ability to absorb CO2 (accumulate biomass).
7.
99
T h e s e two d i r e c t i n t e r a c t i o n s c o n s t i t u t e a f e e d b a c k b e t w e e n c h a n g i n g diversity a n d c h a n g i n g CO2 levels. Results f r o m two e x p e r i m e n t s c o n d u c t e d in a c o n t r o l l e d e n v i r o n m e n t a l facility ( t h e E c o t r o n ) s u p p o r t t h e e c o l o g i c a l bases f o r t h e s e i n t e r a c t i o n s . S i n c e virtually all e c o s y s t e m s a r e c u r r e n t l y b e i n g s i m u l t a n e o u s l y e x p o s e d to b o t h a n t h r o p o g e n i c a l l y i n d u c e d d e c l i n e s in diversity a n d i n c r e a s e d CO2, s t u d i e s t h a t m a n i p u l a t e b o t h CO2 a n d diversity as e x p e r i m e n t a l f a c t o r s will p r o v i d e m o r e p o w e r f u l i n s i g h t s i n t o g l o b a l c h a n g e t h a n single f a c t o r s t u d i e s c a n p r o v i d e a l o n e .
Bazzaz, F. A., and Carlson, R. W. (1984). The response of plants to elevated C O 2 . I. Competition among an assemblage of annuals at different levels of soil moisture. 62, 196-198. Bazzaz, F. A., and Fajer, E. D. (1992). Plant life in a CO2-rich world. 266, 68-74. Ehrlich, P. R., and Wilson, E. O. (1991). Biodiversity studies: Science and policy. 253, 758-762. Gates, D. M. (1994). "Climate Change and Its Biological Consequences." Sinauer, Sunderland, MA. Groombridge, B. (1992). "Global Biodiversity: Status of the Earth's Living Resources." A report compiled by the World Conservation Monitoring Centre. Chapman & Hall, London. Houghton, J. T., Jenkins, G. J., and Ephraums, J. J. (1990). "Climate Change: The IPPC Scientific Assessment." Cambridge Univ. Press, Cambridge, UK. Lashof, D. A. (1989). The dynamic greenhouse: Feedback processes that may influence future concentrations of atmospheric trace gases and climatic change. 11, 7-31. Lawton, J. H., and May, R. M. (1995). "Extinction Rates." Oxford Univ. Press, Oxford, UK. Lawton, J. H., Naeem, S., Woodfin, R. M., Brown, V. K., Gange, A., Godfray, H. C.J., Heads, P. A., Lawler, S., Magda, D., Thomas, C. D., Thompson, L. J., and Young, S. (1993). The Ecotron: A controlled environmental facility for the investigation of population and ecosystem processes. 341, 181-194. Melillo, J. M., Callaghan, T. V., Woodward, F. I., Salati, E., and Sinha, S. K. (1990). Effects on ecosystems. "IPCC, Climate Change, The IPCC Scientific Assessment" pp. 282-310. Cambridge Univ. Press, Cambridge, UK. Naeem, S., Thompson, L.J., Lawler, S. P., Lawton,J. H., and Woodfin, R. M. (1994a). Declining biodiversity can alter the performance of ecosystems. 368, 734-737. Naeem, S., Thompson, L. J., Lawler, S. P., Lawton, J. H., and Woodfin, R. M. (1994b). Biodiversity loss in model ecosystems: A reply to Andr6 371, 565. Naeem, S., Thompson, L.J., Lawler, S. P., Lawton, J. H., and Woodfin, R. M. (1995). Empirical evidence that declining species diversity may alter the performance of terrestrial ecosystems. 347, 249-262. Polley, H. W., Johnson, H. B., and Mayeux, H. S. (1994). Increasing CO2: Comparative responses of the C4 grass and grassland invader 75, 976-988. Schneider, S. H. (1992). The climate response to greenhouse gases. 22, 1-32. Schulze, E. D., and Mooney, H. A. (1993). "Biodiversity and Ecosystem Function." SpringerVerlag, New York. Sisk, T. D., Lauder, A. E., Switky. K. B., and Ehrlich, P. R. (1994). Identifying extinction 44, 592-604. threats. Soul6, M. E. (1991). Conservation: Tactics for a constant crisis. 253, 744-750.
1
Swift, M.J., and Anderson, J. M. (1993). Biodiversity and ecosystem function in agricultural systems. "Biodiversity and Ecosystem Function" (E. D. Schulze and H. A. Mooney, eds.), pp. 15-41. Springer-Verlag, New York. Thompson, L.J., Thomas, C. D., Radley, J. M., Williamson, S., and Lawton, J. H. (1993). The effect of earthworms and snails in a simple plant community. 95, 171-178. Tilman, D., and Downing, J. A. (1994). Biodiversity and stability in grasslands. 367, 363-365. Vandermeer, J. (1989). "The Ecology of Intercropping." Cambridge Univ. Press, Cambridge, UK. Vitousek, e. (1994). Beyond global warming: Ecology and global change. 75,1861-1902. Wilson, E. O., and Peter, F. M. (1988). "Biodiversity." National Academy of Science, Washington, DC.
Predicting Responses of Tropical Plant Communities to Elevated C02: Lessons from Experiments with Model Ecosystems
Tropical ecosystems and plant communities contain an enormous proportion of the world's known species and represent about 42% of the world's biomass carbon reserves (Brown and Lugo, 1982; Olson 1983). Furthermore, these ecosystems are expected to be among the most responsive to the direct effects of rising atmospheric carbon dioxide concentrations (Long, 1991; Hogan 1991; Lugo, 1992). Despite the inevitable importance of tropical ecosystems to global species conservation and to the world's C balance, no experimental data exist on the response of native tropical plant communities to elevated atmospheric CO2! Our knowledge to date is based on a total of seven actual experiments conducted using moist tropical plant species either grown under conditions of nonlimiting nutrient supply as individuals (Oberbauer 1985; Reekie and Bazzaz, 1989; Ziska 1991) or in competitive arrays in model communities (Reekie and Bazzaz, 1989), or in model plant communities with varying degrees of nutrient limitation (Ktrner and Arnone, 1992; Arnone and Ktrner, 1993; Arnone and Ktrner, 1995; and Arnone 1995). In order to improve our chances at accurately predicting the responses of the tropical biome, it is essential to more accurately represent the predominantly low to moderately low soil fertilities found in these regions (e.g., Whittaker, 1975; Sfmchez, 1976). For example, 63% of the soils in the moist tropics is represented by Oxisols and Ultisols (Vitousek and Sanford, 1986). Perhaps a 1 01
Copyright 9 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.
more critical deficit is the complete lack of data on the responses to elevated CO2 of dry tropical vegetation and communities (savannas and scrub vegetation) and even tropical seasonal forests. These terrestrial systems may be among the most responsive to rising atmospheric CO2 because high CO2 reduces leaf stomatal conductance and can reduce plant water use, at least in the short term (e.g., Strain and Cure, 1985; Jackson 1994; see Chiarello and Field, Chapter 10). Despite these deficits in our knowledge and although studies conducted in native tropical systems are desperately needed, we have gained substantial insights into the range of possible responses of native tropical plant communities to elevated CO2 as a result of the diverse data sets generated by these seven experimental studies. Perhaps the most significant finding is that nutrient supply plays a pivotal role in determining the magnitude and speed of responses of various species in a community to elevated CO2. The objective of this chapter is to synthesize our current understanding of how, and through which mechanisms, tropical plant communities will likely respond to rising atmospheric CO2. I particularly emphasize results from the studies with multispecies plant assemblages in which intra- and interspecific competition for above- and belowground resources was incorporated into the experimental design. The reason for this focus stems from the recognition that responses at the community level often cannot be reliably predicted from responses seen in individually grown plants (e.g., Bazzaz and Carlson, 1984; Zangerl and Bazzaz, 1984; Bazzaz and McConnaughay, 1992; K6rner, 1995). However, I attempt to integrate results from pot and model community studies in this synthesis. I would also like to refer the reader to earlier syntheses on the responses of multispecies model communities to elevated CO2 (Bazzaz 1985; K6rner, 1995; Bazzaz, 1990), as well as to works on the potential response of native vegetation (Strain and Bazzaz, 1983; Bazzaz, 1990; Hogan 1991) to high CO2. In addition, K6rner (1993) reviewed the utility of model ecosystems in CO2 research.
Hogan (1991) reviewed the results of the only three experiments to be conducted on individually grown (i.e., potted) tropical plants, and then speculated on the potential responses which might occur at the levels of the plant population, community, and ecosystem. These studies by Oberbauer (1985), Reekie and Bazzaz (1989), and Ziska (1991) have shown that a variety of species from the moist tropics respond both physiologically and in terms of growth to elevated atmospheric CO2. How-
8.
ever, the results from these studies paint a picture that is far from consistent. The responses of the six C~, one C4, and two CAM species observed by Ziska were the most consistent with results typically reported for a wide range of agricultural plants and other temperate species growing under ample nutrient supply (cf. Strain and Cure, 1985). Oberbauer reported a substantial increase in biomass accumulation under elevated CO2 in seedlings of both species tested, but actually measured decreases in leaf-level photosynthesis. Reekie and Bazzaz observed no CO2 effect on biomass or photosynthesis of individually grown plants but did measure decreases in stomatal conductance. Hogan and Ziska attributed the discrepancies among these studies to the negative effects of pot size on sink strength and the potential to respond to elevated CO2 (Ziska used relatively large pots, whereas Oberbauer and Reekie and Bazzaz used relatively small pots) (cf. Thomas and Strain, 1991). Berntson (1993) refuted this claim by showing that the effects of small pot size can be eliminated by increasing nutrient additions. Certainly both pot size and the actual amount of nutrients available to the plant can affect plant growth. Furthermore, all three of these experiments were conducted under unnaturally high nutrient conditions indicating that the species responses to elevated C O 2 observed may be quite untypical. Undoubtedly, the interpretation of C02 responses of potted individuals is confounded by extremes in pot size and by relatively large additions of fertilizer.
A. Species Shifts Although still artificial, model plant communities allow us to assess the responses of species to elevated atmospheric CO2 in an environment which, in most cases, more closely reflects natural competitive conditions than those represented by isolated plants (K6rner 1993). The four experiments with model communities of moist tropical plant species completed to date demonstrate some significant common responses to high CO2 as well as some important differences. All of these studies began with the establishment of replicate plant communities containing either several to many species planted in relatively homogeneous substrates (but all containing some amount of native soil) or, in Arnone and K6rner (1993), a twostoried monoculture in which the responses of overstory and understory plants were compared (Table I). Aside from the number of species used, experimental conditions differed dramatically, especially between the Atnone and K6rner studies and that of Reekie and Bazzaz (1989). Ground area, substrate depth, and average volume of substrate available per individual plant were in some cases orders of magnitude different. Planting density
Substrate
Species Reekie and Bazzaz (1989) 5 species
Functional groups
Moist tropical
Tree Tree Tree Tree Tree KOrner and Arnone (1992) "Experiment I" 15 species Moist tropical
Pioneer tree Tree Shrub Shrub Climbing vine Climbing vine Climbing vine G r o u n d vine G r o u n d vine Herb. m o n c o t Herb. m o n c o t Herb. m o n c o t Herb. m o n c o t Herb. m o n c o t
Planting density (m -s)
400
Ground area (m -s)
0.075
0.30 0.45 0.75 0.30 0.30 0.30 0.30 0.90 0.45 0.30 0.30 0.60 0.30 0.60 0.45
soil: turface (1:1) Fertilizer 4 • ?
11
Vol. (liters)
8
CO2 levels (/~11 -~)
300 525 700
111
340 610
94
LAI
NPP equivalent ( g m - S y r -1)
6.65
sand: vermiculite (1:1) + compost-soil layer Fertilizer 20 g m -s Osmocote NPK and micro. Tot. N equiv, of 400 kg N ha -~
1300
Shift in Biomass spp. abund. response at harvest ~ at harvest (A % rel. to (% of c o n t r o l ) c o n t r o l [COs])
n.s.
n.s.
n.s.
12.3 b 11.9 b 11.4 b
1875 1840 1940
control -2 +3 at 525, 700 +40 +100 +122 +115 -40 -45 +55 +60 -30 +15
6.9 6.6
P = 0.09 2780 3377
P = 0.10 control +11 at 610 +10 +1 +20 +13 +22 -17 + 4 0 (*) +5 +15 +47 (*) +17 +23 - 9 (*) +5 +8
Peter's NPK m Very high: "nonlimiting"
80 80 80 80 80
6.6
Type
Depth (cm)
Effective trt. period (d)
significant
33.0 35.4 at 525, 700 +1.8 +1.6 +4.1 +5.3 -8.0 -17.7 +10.6 +8.5 -8.5 +2.3 P = 0.10
n.s.
13.4 at 610 +0.06 - 5 . 2 3 (*) -0.19 +0.16 +0.74 +0.16 +1.99 (*) -0.08 -0.15 +0.69 (*) +0.12 +2.76 -0.88 -0.15 0.00
Number of communities (n)
Arnone and K6mer (1993) "Experiment II" 1 species Moist tropical (two-storied canopies) Over-/understory
5.7
6.65
same as K6rner & Arnone (1992) Tot. N equiv, of ca. 360 kg N ha -t
3.0/2.7
20
1300
340 610 340
21
Arnone and K6rner (1995) "Experiment HI"
Moist tropical
Tree Tree G r o u n d vine dcot. G r o u n d vine mcot. Herb. m o n c o t Herb. m o n c o t Herb. moncot
a Shifts in species contribution
n.s.
11.6
6.65
0.30 0.45 0.90 0.45 0.30 0.60 0.30
to t o t a l c o m m u n i t y
bIncludes foliage which extended
sand and tropical soil inoculum top layer (1 cm) and dried ground plant material Fertilizer 129 g m -~ NPK and micro. Total N equiv, of 118 kg N ha -~ yr -~
aboveground
beyond the ground
25
1700
340 610
530
4.0 3.9
b i o m a s s ; * * P < 0.01, * P < 0.05, (*) P < 0.15.
area of the tubs (most probably).
cIncludes leaf litter and standing necromass, but excludes root litter.
P = 0.02 375 c
P = 0.02 +18"
n.s. (% understory)
47O ~
1.75/ 0.11 1.73/ 0.08
610
7 species
n.s. 1.86 1.81
+ 1 8 " / + 1 6 n.s.
n.s.
n.s.
n,s. (sig. by spp. groups) 20.4
-18 +35 +136 0 +9 -40 -18
-4.4 +6.4 +0.9 0 +2.9 -4.1 -1.7
815 ~ 910 c
106 was also lower (6.6-11.6 versus 400 individuals per square meter) as was the a m o u n t of nutrients supplied/available in the Arnone and K6rner experiments. The following is a relative ranking of the nutrient availabilities of the four studies: Reekie and Bazzaz (1989)~>)K6rner and Arnone, 1992 > Arnone and K6rner 1993 ~> Arnone and K6rner, 1995. Furthermore, the starting size of individuals used in each experiment varied. Reekie and Bazzaz began with very small seedlings, Arnone and K6rner (1993 and 1995) with larger (30-70 cm tall) but still relatively small stamred individuals, and K6rner and Arnone (1992) with larger (up to 1 m tall) individuals in order to create a highly structured stand from the start which enabled the analysis of the CO2 effects on various life-forms occupying different positions along the vertical light gradient. Finally, the duration of these experiments varied considerably ranging from 21 to 530 days (Table I), and Reekie and Bazzaz were able to include four replicate communities per CO2 treatment, instead of the two used in the Arnone and K6rner studies. Despite the significant increases in community biomass accumulation observed under both ambient and elevated CO2 over the course of all of these experiments, significant CO2-induced shifts in species dominance (or in composition of overstory versus understory plants) were reported for only two of them (Table I). Reekie and Bazzaz (1989) found highly significant and substantial shifts in the contribution of individual tree species to community aboveground biomass with increasing CO2 concentration in stands planted with equal densities of each species. For example, increased in abundance while decreased at elevated CO2. Perhaps most striking is that these shifts occurred even though CO2 level had no effect on overall community aboveground biomass or on leaf area index (LAI). The authors showed that the success of a species was positively related to its mean canopy height measured at harvest. In much more nutrient-poor systems, Arnone and K6rner (1995, Experiment III) reported significant changes in the abundance of groups of species under elevated CO2 but no significant shifts in any single species (Table I). The magnitude of the CO2-induced shifts Arnone and K6rner observed over 530 days were considerably less than those observed by Reekie and Bazzaz (1989). The contribution of a pioneer species, to community biomass was reduced over the course of the experiment and this shift was slightly (n.s.) enhanced under elevated CO2 (Fig. 1). In contrast, the slower-growing and the understory monocot increased in abundance in all communities, with a trend toward even greater increases u n d e r elevated CO2. No significant changes in biomass accumulation or LAI were observed u n d e r elevated CO2. Although no significant overall species shifts were observed in the relatively fast-growing nutrient-rich systems of Experiment I (K6rner and Arnone, 1992), showed a substantial 5% mean decrease in its share of community aboveground
8.
107
Figure 1 Mean changes in the contribution of the seven species of moist tropical plants to communitybiomass (including coarse root biomass) over 530 daysin Experiment III in systems maintained at ambient and elevated atmospheric CO2 concentrations.
biomass (P = 0.11) u n d e r elevated C O 2 measured at harvest, while two other and species showed marginally significant increases in their share of community aboveground biomass (P = 0.14 and 0.08, respectively, Table I). The success of at elevated CO2 u n d e r one set of competitive experimental conditions and its relative failure at elevated CO2 u n d e r another set of conditions does, however, point to the difficulty of predicting species-specific responses to elevated CO2 based on results from model systems alone. Thus, it appears that the most reliable prediction is that some level of shifts in species dominance will take place in native moist tropical communities in a CO2-rich world, and shifts may occur more rapidly in nutrient-rich systems containing very young individuals than in nutrient-poorer systems or in those containing older plants. However, we are unable to predict with any degree of certainty which species will win and which will lose.
B. Individually Grown Versus Competitively Grown Plants How well do the responses of individually grown plants extrapolate to their responses when grown u n d e r competition? Generally, existing data support the notion of a poor correspondence between responses in individually grown plants and competitively grown plants. Reekie and Bazzaz (1989) report correlations between some autecological morphological traits and responses to elevated CO2 in their tropical plants growing u n d e r competitive conditions. However, they found no correlations for other morphological and physiological traits. Perhaps their most interesting finding was that mean canopy height and shape (leaf area profiles) were strongly influ-
108
enced by competition and C O 2 level. Bazzaz and McConnaughay (1992) use data from an experiment by Williams (1988), with serpentine grassland species, to illustrate the pronounced mismatch between projections of community species composition derived from CO2 responses of isolated individuals and their actual success in heterospecific stands. Likewise, the experiments of Arnone and K6rner with model communities of tropical plant species indicate a relatively poor correspondence between prognostications made based on single-plant experiments (Hogan 1991) and responses of plants in multispecies stands. For example, Arnone and K6rner found no uniform increase across all species in leaf area per plant under competitive conditions in any of their three experiments, as is implied by Hogan (1991) and predicted from models (e.g., Oikawa, 1990). Moreover, in none of the three community experiments did Arnone and K6rner find the often predicted CO2-induced increase of LAI (Eamus andJarvis, 1989; Nijs 1989). Even under nonlimiting nutrient conditions, Reekie and Bazzaz (1989) found no effect of CO2 concentration on leaf area per plant of either individuals or plants grown in model communities. When nutrients were supplied at near natural levels, LAI even tended to decrease under elevated CO2 (K6rner and Arnone, 1992; Arnone and K6rner, 1993, 1995). This was associated with higher leaf mortality in a number of species at high CO2 (Fig. 2). These leaf area responses are remarkably consistent over a wide range of nutrient conditions, species mixes, plant life-forms, overall LAIs, and durations of experiments, further exemplifying the danger of simple extrapolation from singleplant responses obtained under unlimited nutrient supply to responses under competition and more realistic nutrient regimes. C. Effects of Plant Morphology on Competitive Performance: Biomass Allocation Patterns and Plant Life-form
In tropical plant species growing competitively in model communities in Experiment III (Arnone and K6rner, 1995), the effects of elevated CO2 on patterns of biomass allocation were similar to those observed in a very large number of other species grown as individual plants (e.g., Strain and Cure, 1985; Rogers 1994; Poorter, 1993). These patterns include greater (initial) enhancement of root growth than shoot growth under elevated CO2 (Fig. 3) and consequently increased allocation to root biomass. These changes are commonly associated with greater root:shoot ratios, lower leaf weight ratios (LWR), and leaf area ratios (LAR) (e.g., Norby 1992) but may not be so pronounced when mineral nutrients are abundant (Oberbauer 1985; Reekie and Bazzaz, 1989; Ziska 1991). For instance, Ziska and coworkers (1991) only found increases in root:shoot ratio in two of nine species, and lower LWRs in half of the C3 species studied.
8.
-
+t I-
E
v m
[~ Arab0~ I
_.1
Time courses of leaf area index (LAI) development in Experiments I, II, and III for communities maintained at ambient (open symbols) and elevated (filled symbols) atmospheric CO2 concentrations (mean _+ SE of two communities per CO2 level).
LWR appears to explain the relative competitive success of the seven species in Experiment III (Arnone and K6rner, 1995). They found that the greater the LWR of a species the larger the positive shifts observed in species dominance within the plant community over the 530-d experiment (Fig. 4). Since no CO2-induced changes were observed in specific leaf area (SLA) for any of the species (Arnone 1995), this relationship also holds true for LAR. This effect was independent of CO2 treatment, however, and the relationship appears to reverse itself if the very high
-2'o o Relative effects of growth under elevated CO2 on LAI, biomass accumulation of various organs, and on production of aboveground necromass in Experiments I, II, and III (difference between mean ambient and elevated values expressed as a percentage of the ambient mean; a, senescing leaves; **, P < 0.01; *, P < 0.05; (*), P < 0.15).
110
o
0-
ffl
Ct
Relationship between leaf weight ratios (LWR, including coarse roots) measured for each of the seven species at the start of the 530-day Experiment III and their competitive outcomes as measured by the change in a species' contribution to total Community biomass over the course of the experiment. Each point represents one species in one of the four communities. Each curve represents a second-order polynomial fitted either to all points (longer curve, r z = 0.52) or to all but the four data points from (shorter curve, r 2 = 0.48). Curves fitted separately to ambient and elevated CO2 points were not different, so the curve shown was fitted to the entire set of points (save Species key: Ct = Ce = Ep = He = E1 = Fib = Fip =
a g r o u n d - c r e e p i n g vine, were i n c l u d e d in the analysis. R e e k i e a n d Bazzaz (1989) u s e d p r i n c i p l e c o m p o n e n t analysis a n d stepwise r e g r e s s i o n to evaluate which m o r p h o l o g i c a l a n d physiological characteristics o f individually g r o w n plants o f five tropical p l a n t species w o u l d best e x p l a i n t h e i r success w h e n g r o w i n g in c o m p e t i t i v e arrays. T h e two m e a s u r e s w h i c h e x p l a i n e d almost 75% o f the variation in c o m p e t i t i v e success were m o r p h o l o g i c a l traits: m e a n c a n o p y h e i g h t a n d leaf a r e a ratio. N e t leaf-level p h o t o s y n t h e s i s e x p l a i n e d less t h a n 9% o f the variability in the data. I n d e e d even in p o t t e d tropical plants, O b e r b a u e r (1985) f o u n d t h a t b i o m a s s a c c u m u l a t i o n was g r e a t e r in plants g r o w n u n d e r elevated CO2 even t h o u g h leaf-level p h o t o s y n t h e s i s was lower in these plants! All o f these o b s e r v a t i o n s g e n e r a l l y suggest t h a t CO2-induced alterations in p l a n t m o r p h o l o g y a n d p l a n t d e v e l o p m e n t may be m o r e useful p r e d i c t o r s o f species' c o m p e t i t i v e success t h a n c h a n g e s in t h e i r p h o t o s y n t h e t i c p e r f o r m a n c e . Biomass allocation p a t t e r n s are closely tied to p l a n t life-form in t h a t c e r t a i n life-forms e x h i b i t various d e g r e e s o f m o r p h o l o g i c a l plasticity in r e s p o n s e to e n v i r o n m e n t a l stimuli a n d to c o m p e t i t i o n . Thus, m o r p h o l o g i -
8.
111
cal constraints can profoundly affect the type and magnitude of response to atmospheric CO2 level (Table I). Once one begins to think in this vein one must consider constraints imposed on a species' resource (light and nutrients) capturing ability conferred on it by both its lateral and vertical space occupancy in the canopy and in the soil. Indeed, Reekie and Bazzaz (1989) showed the importance of differences in canopy height and architecture in determining the success of tropical tree seedlings u n d e r elevated CO2. In Experiment III, Arnone and K6rner also found that plant lifeform could help explain its response in model communities. For example, herbaceous monocots such as were able to proliferate rapidly in lateral directions but could not grow significantly in height, whereas tree species such as and were not able to grow laterally very quickly but could grow in height (Arnone and K6rner, 1995; see also Hunt 1991). the pioneer with its short-lived leaves and rapid growth in open stands, gained an early advantage in all communities. In contrast, the slower growing tree species with its long-lived leaves and its continuous occupation of the soil with its roots allowed it to gradually increase in dominance in all communities, but more so u n d e r elevated CO2. Within the herbaceous monocots, the success of (20% of start biomass and 38% of harvest biomass at ambient CO2, and 41% at elevated CO2) and the relative failure of (39% at the start, 12% and 8%, respectively at harvest) was largely due to the relative slowness of leaf and lateral filler proliferation in more rapid proliferation of new tillers in Furthermore, root systems of were more extensive than those of
D. The Vertical Dimension: Interactive Effects of LAI, Light, and Light Quality Plants growing in competitive arrays shade themselves and each other and create gradients of decreasing light availability from the top to the bottom of the canopy. As LAI increases, light transmittance in the stand decreases. These changes are known to have variable effects on different plant species depending on their position in the community. Under conditions of elevated atmospheric CO2 it has often been hypothesized that LAIs should increase (e.g., Nijs et al., 1989; Eamus and Jarvis, 1989). One reason that this could occur is through an increased production and retention of leaves of understory plants growing in deep shade. Greater leaf retention u n d e r very low photon flux densities (PFDs) u n d e r elevated CO2 could result from improved leaf carbon balance (Pearcy and Bj6rkmann, 1983) afforded by CO2-induced reductions in light compensation point and by increases in q u a n t u m use efficiency (e.g., Ehleringer and Bj6rkmann, 1977). In all three of the Arnone and K6rner experiments and in the study of Reekie and Bazzaz (1989) no evidence in support of this hypothesis was
observed (Table I). In fact, as a consequence of the unchanged LAIs under elevated CO2 (Fig. 2; Reekie and Bazzaz, 1989), no differences in light transmittance within stands has been observed. Accordingly, we found no CO2-induced shifts in vertical leaf area distribution for any of the 15 species growing in the communities in Experiment I (Fig. 5). Thus, LAI appears to be relatively insensitive to atmospheric CO2 concentration. In Experiment II, Arnone and K6rner (1993) specifically examined the responses to elevated CO2 ofunderstory plants growing in two-storied monospecific stands of the extremely fast-growing tropical species Both understory and overstory plants in all communities increased significantly in size over the 21-d CO2 treatment. However, they observed no differences in LAI of either the overstory or understory plant canopies between ambient and elevated CO2 communities (Fig. 3). They also found no enhancement of biomass accumulation in understory plants in high CO2 communities, as was hypothesized, but did observe significantly greater (17%) height growth and internode length of understory plants growing in communities maintained at elevated CO2 (Fig. 3). These growth responses are typical of shade avoidance reactions in response to reductions in the red:far-red ratio (e.g., Smith, 1982), and suggested that elevated CO2 may alter properties of leaves such that they absorb more red light a n d / o r reflect more far-red light (e.g., Ballar6 1987). Indeed, Arnone and K6rner measured significantly lower R:FR ratios beneath overstory leaves produced under elevated CO2 than under leaves produced under ambient CO2. Although these results do not rule out other possible direct effects of elevated CO2 on understory plant behavior, they do suggest that CO2-induced alterations in light quality within a community could eventually have a pronounced effect on plant-plant interactions and competitive outcomes. For example, recruitment patterns and species composition of later successional stages in tree-fall gaps may be affected by shifts in R:FR ratios and variable inherent sensitivities to reductions in R:FR ratio of gapcolonizing species.
250]
Meanleaf area profiles of each of the 15 species growing in model communities in Experiment I after 94 days of exposure to ambient (open symbols) and elevated (filled symbols) atmospheric CO2 concentrations.
8.
E. Photosynthetic Performance as a Determinant of Species' Competitive Success
M. Gruber (unpublished data) measured whole-shoot CO2 exchange of the most dominant species in Experiment III in order to evaluate species' contributions to ecosystem CO2 flux and to test how well shootlevel physiology would correspond to species' competitive performance. Measurements were made in the last third of the experiment with an openIRGA system by enclosing individual shoots from each species in each community in transparent polyethylene bags. During this phase of the experiment relatively pronounced "successional" shifts in species dominance were underway in all communities, however no obvious CO2 treatment effects were seen. Net shoot CO2 flux on each individual was measured at the growth CO2 concentration for about 16 hours, which included the latter half of the photoperiod and most of the dark period. In Experiment III, Gruber found greater shoot assimilation rates under elevated CO2 in all of the four species evaluated (Table II). In and these increases were marginally significant (P < 0.15) and amounted to 53% and 123% of the rates measured at ambient CO2, respectively. Higher assimilation rates under elevated CO2 in may have
Shoot dark respiration
S h o o t assimilation (/~mol CO2 m -2 leaf s -1) Species
Amb. CO 2 2.57 1.41 0.99 0.77
_+ _+ + _+
Elev. CO2
0.39 0.14 0.08 0.20
2.85 2.16 1.04 1.71
+ + _+ _+
0.16 0.27 0.08 0.17
(/~mol CO2 m -2 leaf s -1) Diff. Diff. (%) +11 + 5 3 (*) +6 + 1 2 3 (*)
A m b . COz 0.69 0.24 0.12 0.09
+ +_ +_ _+
0.10 0.04 0.03 0.02
Elev. CO2 0.76 0.27 0.09 0.04
+_ +_+ _+
0.01 0.02 0.02 0.04
(%) +10 +12 -25 -55
a.f ANOVA
CO 2
1
Results
Resid.
2
Species
3
Sp. • C O 2 Residual
3 6
n.s.
n.s.
(*)
Levels of statistical significance: ** = P < 0.01; * = P < 0.05; (*) = P < 0.15. a All parameters were measured at growth CO2 concentrations. Mean net assimilation rates measured on individual leaves on Day 305 were 72% greater (P = 0.03) under elevated CO2 (13.6 _+ 0.2 /~mol CO2 m -s leaf s-1) than under ambient COs (7.9 _+ 1.1 /.~mol COs m -2 leaf s-l). The values in the table are means _+ SE with n = 2 communities per COs level.
114 contributed to its relative greater gains in these communities than in communities maintained at ambient CO2 (Table I). on the other hand, suffered marginally greater losses under elevated CO2 despite its greater shoot assimilation rates and its reduced dark respiration rates. Likewise, the 25% lower (n.s.) shoot dark respiration rates measured in may have conferred a substantial advantage to this species over time and facilitated its relative greater expansion in elevated CO2 communities. Species such as which had similar shoot CO2 assimilation and dark respiration under both ambient and elevated CO2, but surprisingly 72% greater leaf-level photosynthesis (see legend of Table II), tended to lose a greater share in community biomass by the end of the experiment. This again illustrates the unreliability of using physiological traits alone to predict competitive outcomes. Thus it can be argued that physiological performance with respect to shoot C O 2 balance may have contributed to competitive outcome in some species but may have been relatively unimportant determinants of success in other species. The same held true for the connections between leaf stomatal conductance and species competitive outcome in Experiment III. In the latter third of this experiment, A. Kocyan (unpublished data) measured substantial decreases in leaf diffusive conductance under elevated CO2 in the five most dominant species, but only small reductions (5%) in whole ecosystem evapotranspiration. It is unclear how lower water consumption by any of the species would confer any particular competitive advantage since water was not limiting in this experiment. No differences in leaf stomatal conductance between ambient and elevated CO2 were found in Experiment I, and this corresponded well with the absence of differences in ecosystem evapotranspirational water losses. Due to the 75 % greater LAI in this experiment (--- 7), relative to Experiment III (~- 4), air humidity was higher and may have precluded stomatal responses to CO2.
F. Belowground Interactions among Plants' and Species' Competitive Success Under elevated atmospheric CO2 levels, competition for nutrients (e.g., phosphorus) should increase as a result of greater root growth, greater allocation of carbon to root systems, and maintenance of larger fine root populations of many or all plant species within a community. Thus the importance of interactions among roots and between roots and soil organisms would also be expected to increase in a CO2-rich world, but these have not been quantified. However, qualitative observations on the timing and extent of root growth and proliferation of various species in model communities confirm the major role they play in determining competitive outcome regardless of CO2 effects. For example, Bazzaz (1990) attributed the relatively successful outcome of higher atmospheric CO2 concentrations
8.
115
(see Reekie and Bazzaz, 1989) to its greater allocation to roots and ultimately to its rapid occupation of the soil volume early in the experiment. Arnone and K6rner (1995) observed a similar p h e n o m e n o n in their Experiment III under relatively nutrient-poor conditions. Early in the experiment (first 100 d), the pioneer species quickly occupied the entire top several centimeters of the soil with its fine roots and grew equally rapidly in height under both ambient and elevated CO2. As the experiment progressed, root systems and shoots of the understory herbaceous monocots (especially and the relatively slow growing trees began to compete strongly with its growth. By the end of the experiment (530 days) in all communities had lost tremendous share in its contribution to community leaf area and biomass, especially. Belowground competitive pressure from a species with leaves unable to shade those of appeared to be the main cause for the severe suppression of growth in all communities, and this effect tended to increase at high CO2. In another experiment embedded in Experiment III (J. Arnone, unpublished data), abilities of species to exploit nutrient-rich soil microsites in an otherwise "nutrient-limited" system were investigated. A pronounced increase in proliferation of fine roots (-<2 mm) in nutrient patches (created using fertilizer sticks inserted vertically over the entire soil depth) was observed under both CO2 regimes. Surprisingly, however, no overall increase in proliferation was observed under elevated CO2, nor were any species-specific fine root responses evident. Apparently, fine root proliferation in this system was not carbon limited.
G. Insect Herbivory and Plant Species Dominance When other biological factors which are known to influence plant community structure are incorporated in analyses of vegetation responses to elevated CO2, the need for field experimentation becomes more obvious. For example, Coley (1983) showed that insect herbivores can dramatically affect plant growth and ecology in tropical ecosystems, and further demonstrated differential species responses to herbivory as affected by resource availability and investment in defense compounds (Coley 1985). In all but two experiments (Thompson and Drake, 1994; Arnone 1995) reported so far designed to evaluate the effects of elevated CO2 on insect herbivory, feeding responses have been measured either by providing larvae in petri dishes with leaf material produced under ambient and elevated CO2, or by placing larvae on individual plants (cf. Lincoln 1993). Predictions based on these studies indicate that larvae will eat more leaf tissue on high CO2-grown plants in order to compensate for lower leaf nutritional quality (i.e., lower C:N ratios) and that larval fitness will decline (e.g., Fajer 1991). Although these studies have provided useful infor-
m a t i o n a b o u t the r e s p o n s e s of insect herbivores to c h a n g e s in leaf tissue quality, they are i n h e r e n t l y i n a d e q u a t e for p r e d i c t i n g how herbivory may be affected in c o m p l e x c o m m u n i t i e s w h e r e h e r b i v o r e s n o t only r e s p o n d to variations in leaf n u t r i t i o n a l quality, b u t w h e r e they can c h o o s e a m o n g species a n d positions in the c a n o p y in which to feed. Results f r o m an e x p e r i m e n t c o n d u c t e d at the e n d of E x p e r i m e n t III (Arnone 1995) illustrate these points. A r n o n e a n d co-workers i n t r o d u c e d e q u i v a l e n t p o p u l a t i o n s of f o u r t h instar a lepidopteran generalist, to c o m p l e x m o d e l ecosystems c o n t a i n i n g seven species of moist tropical plants m a i n t a i n e d u n d e r low m i n e r a l n u t r i e n t supply. Larvae were allowed to feed freely for 14 days by which time they h a d r e a c h e d the seventh instar. Prior to larval i n t r o d u c t i o n s , p l a n t c o m m u n i t i e s h a d b e e n c o n t i n u o u s l y e x p o s e d to e i t h e r a m b i e n t or elevated CO2 for 1.5 years. In c o n t r a s t to E x p e r i m e n t I w h e r e n u t r i e n t supply was relatively h i g h ( K 6 r n e r a n d A r n o n e , 1992, Table III), n o m a j o r shifts in leaf n u t r i t i o n a l quality (e.g., leaf n i t r o g e n c o n c e n t r a t i o n , Fig. 6) were observed in this low fertility system b e t w e e n CO2 t r e a t m e n t s for any of the seven species. F u r t h e r m o r e , n o c o r r e l a t i o n s were observed b e t w e e n leaf quality a n d leaf biomass cons u m p t i o n by larvae (e.g., Fig. 6). In situations w h e r e significant r e d u c t i o n s in l e a f N c o n c e n t r a t i o n s o c c u r (Table III, E x p e r i m e n t I), h e r b i v o r e f e e d i n g
Species
Amb. CO2
Elev. CO2
Diff. (A%)
20.6 +- 1.2 12.4 + 0.1 15.6 + 0.4 18.6 + 0.5 22.0 + 0.7 18.9 _+ 0.6 14.6 + 0.3 20.7 + 0.8 16.3 _+ 0.6 16.6 +_ 0.6 16.1 +_ 0.7 18.8 _+ 0.6 17.7 _+ 0.8 12.7 -+ 0.3 12.8 _+ 0.4
16.0 _+ 0.6 10.9 + 0.2 12.8 + 0.2 17.1 _+ 0.4 18.9 +_ 0.8 16.8 _+ 0.7 11.7 _+ 0.3 19.1 + 0.3 15.1 _+ 0.4 15.1 _+ 0.4 14.9 _+ 0.5 18,3 _+ 0.5 15.0 _+ 0.7 11.3 +_ 0.2 13.0 +_ 0.4
-4.6 -1.5 -2.8 -1.5 -3.1 -2.1 -2.9 -1.6 -1.5 -1.5 -1.2 -0.5 -2.7 -1.4 +0.2
8. 0.4
0.3" J... r
0.2-
t~
0.1"13
E tO
0.0 0.4
. . . .
. . . .
E
. 0~ ..Q
_1
|
|
20 Leaf nitrogen concentration (mg N g-l) Figure 6 Species' leaf biomass consumption by late-instar larvae after 14 days of feeding as a function of leaf nitrogen concentration for plants growing in communities which had been exposed for 515 days to either ambient (open symbols) or elevated (filled symbols) atmospheric CO2 concentrations in Experiment III. Key: (a) through (d): men 9 (e) and (f): dicots. LeafN concentrations shown as insets represent the mean _ SE of two communities per CO2 level.
may still be strongly determined by this factor. Total leaf area (IAI = 4) and biomass of all plant communities were similar when caterpillars were introduced. However, leaf biomass of some species was slightly g r e a t e r - and for other species slightly less (e.g., communities exposed to elevated CO2. Larvae showed the strongest preference for leaves, the plant species that was least abundant in all communities (and which lost in relative abundance under high CO2), and fed relatively little on plants species which were more abundant. Interspecific differences in leaf quality most certainly influence feeding patterns of these generalists. Low concentrations of defense compounds in leaves may have accounted at least in part for the overwhelming preference shown for them by the larvae (cf. Coley 1985). These results indicate leaf tissue quality is not necessarily affected by elevated CO2 when nutrient supply is very restricted. However, on nutrientrich sites, and where "choice" of food is limited (e.g., only two species of graminoids), leaf nutritional quality may still be a major determinant of herbivory (Thompson and Drake, 1994). Herbivores appear to become
more dependent on less-preferred plant species in cases where elevated CO2 results in reduced availability of leaves of a favored plant species. This greater dependency may eventually affect insect populations adversely. Thus, once again we see a more complex and quite different response to an elevation of atmospheric carbon dioxide when results from leaf/ individual-plant studies are compared with those obtained in relatively complex systems. In native systems, insect herbivores may either amplify or suppress changes in plant species' abundance caused directly by variable responses to elevated CO2.
Perhaps the most significant lesson from this handful of experimental studies with tropical plants is that responses of individually grown plants to high CO2 do not scale well (with consistent predictability) to those obtained when individuals are grown under competitive conditions. For example, predictions of dramatic increases in biomass accumulation and leaf area production at the level of the community, especially in naturally "nutrient-limited" systems, is consistently contradicted by evidence from model and communities (e.g., see also K6rner, 1995). More astonishing is that shifts in species' abundance have occurred in model systems under elevated CO2 even in the absence of significant overall stimulation of biomass accumulation. Sufficient evidence now exists to state with some degree of confidence that the speed and extent to which shifts occur under elevated CO~ is linked to the age of the plants and to the availability of other growth-limiting resourcesmespecially nutrients. Furthermore, results from Experiments I-III, where Arnone and K6rner attempted to link community-level changes with those occurring at the level of the entire ecosystem, indicate that in nature innumerable positive and negative feedbacks occurring at all levels of system complexity will influence the structure and composition of plant communities (e.g., effects of herbivory, changes in light quality within the stand, and increased rhizodeposition on nitrogen availability) (cf. Diaz 1993; Arnone, 1995; Zak 1993). For the most part, the consequences of such complicated interactions among biotic and abiotic factors and elevated atmospheric CO2 cannot be predicted with any degree of satisfaction for any native system. At present, we are incapable of making sound predictions of the nature and extent of shifts in native communities. Working in native vegetation may help us overcome some of the major limitations of working in model systems-effects on the actual specific competitive outcomes and species' shifts created by the particular choice of species, life-forms, and their relative abundances and arrangements used in artificial communities. However, it
8.
119
will also introduce greater experimental noise and may make the detection of CO2 effects more difficult. Therefore, specifically designed and coordinated experiments in both native and model systems will yield the most fruitful results. The ecologically important c o m m o n responses observed across all of these model systems, in spite of the strong effects which resulted from a given community design and growth conditions, attests to their robustness. Thus the best place to begin perhaps is to test whether and to what extent the c o m m o n responses to elevated COz found in all model systems thus far occur in native systems in the field. For now, all we can say with any degree of certainty is that elevated atmospheric COz concentrations will lead to changes in species dominance in tropical plant communities, just as it will in other geographic regions. Eventually this will alter patterns of plant succession and vegetation cover.
I am grateful to my colleagues at the Botanisches Institut of the University of Basel who generously assisted me with many aspects of the studies discussed in this chapter. I especially thank Christian K6rner for his stimulating discussions and his comments on an earlier version of the manuscript. The canton of Basel Stadt provided funding for this research.
Arnone, J. A. III. (1995). Soil nitrogen availability in an alpine grassland under elevated atmospheric COz. Submitted for publication. Arnone,J. A., III, and K6rner, Ch. (1993). Influence of elevated CO2 on canopy development and red:far-red ratios in two-storied stands of 94, 510-515. Arnone,J. A., III, and K6rner, Ch. (1995). Soil and biomass carbon pools in model communities of tropical plants under elevated CO2. 104, 61-71. Arnone, J. A., III, Zaller, J. G., Ziegler, C., Zandt, H., and K6rner, C. (1995). Leaf quality and insect herbivory in model tropical plant communities after long-term exposure to elevated atmospheric CO2. 104, 72-78. Ballar6, C. L., Sgmchez, R. A., Scopel, A. L., Casal, J. J., and Ghersa, C. M. (1987). Early detection of neighbour plants by phytochrome perception of spectral changes in reflected sunlight. 10, 551-557. Bazzaz, F. A. (1990). The response of natural ecosystems to the rising global COs levels. 21, 167-196. Bazzaz, F. A., and Carlson, R. W. (1984). The response of plants to elevated CO2. I. Competition among an assemblage of annuals at two levels of soil moisture. 62, 196-198. Bazzaz, F. A., Garbutt, K., and Williams, W. E. (1985). Effect of increased atmospheric carbon dioxide concentration on plant communities. "Direct Effects of Increasing Carbon Dioxide on Vegetation. Carbon Dioxide Research, State of the Art" (B. R. Strain and J. D. Cure, eds.), pp. 155-170. U. S. Dept. of Energy, Washington, D.C., Publ. No. ER-0238. Bazzaz, F. A., and McConnaughay, K. D. M. (1992). Plant-plant interactions in elevated CO2 environments. 40, 547-563.
120 Berntson, G. M., McConnaughay, K. D. M., and Bazzaz, F. A. (1993). Elevated CO2 alters deployment of roots in "small" growth containers. 94, 558-564. Brown, S., and Lugo, A. E. (1982). The storage and production of organic matter in tropical forests and their role in the global carbon cycle. 14, 161-187. Coley, P. D. (1983). Herbivory and defense characteristics of tree species in a lowland tropical forest. 53, 209-233. Coley, P. D., Bryant, J. P., and Chapin, F. S., III. (1985). Resource availability and plant antiherbivore defense. 230, 895-899. Diaz, S., Grime,J. P., Harris,J., and McPherson, E. (1993). Evidence of a feedback mechanism limiting plant responses to elevated carbon dioxide. 364, 616-617. Dixon, R. K., Brown, S., Houghton, R. A., Solomon, A. M., Trexler, M. C., and Wisniewski, J. (1994). Carbon pools and flux of global forest ecosystems. 263, 185-190. Eamus, D., and Jarvis, P. G. (1989). The direct effects of increase in the global atmospheric COs concentration on natural and commercial temperate trees and forests. 19, 1-55. Ehleringer, J., and Bj6rkmann, O. (1977). Quantum yields for COs uptake in Cs and C4 plants. 59, 86-90. Fajer, E. D., Bowers, M. D., and Bazzaz, F. A. (1991). The effects of enriched COs atmospheres on the buckeye butterfly, 72, 751-754. Hogan, K. P., Smith, A. P., and Ziska, L. H. (1991). Potential effects of elevated COs and changes in temperature on tropical plants. 14, 763-778. Hunt, R., Hand, D. W., Hannah, M. A., and Neal, A. M. (1991). Responses to COs enrichment in 27 herbaceous species. 5, 410-420. Jackson, R. B., Sala, O. E., Field, C. B., and Mooney, H. A. (1994). CO2 alters water use, carbon gain, and yield for the dominant species in a natural grassland. 257-262. K6rner, Ch. (1995). The response of complex multispecies systems to elevated COs. "Global Change and Terrestrial Ecosystems" (B. H. Walker and W. L. Steffen, eds.), in press. Cambridge Univ. Press, Cambridge, UK. K6rner, Ch., and Arnone, J. A., III. (1992). Responses to elevated carbon dioxide in artificial tropical ecosystems. 257, 1672-1675. K6rner, Ch., Arnone, J. A., III, and Hilti, W. (1993). The utility of enclosed artificial ecosystems in COs research. "Design and Execution of Experiments on CO2 Enrichment" (E.-D. Schulze and H. A. Mooney, eds.), pp. 185-198. Ecosystem Research Report 6, Commission of the European Communities, Brussels. Lincoln, D. E., Fajer, E. D., and Johnson, R. H. (1993). Plant-insect herbivore interactions in elevated COs environments. TREE 8, 64-68. Long, S. P. (1991). Modification of the response of photosynthetic productivity to rising temperature by atmospheric COs concentrations: Has its importance been underestimated? 14, 729-739. [Opinion] Lugo, A. E. (1992). The search for carbon sinks in the tropics. 64, 3-9. McConnaughay, K. D. M., Berntson, G. M., and Bazzaz, F. A. (1993). Limitations to CO2induced growth enhancement in pot studies. 94, 550-557. Nijs, I., Impens, I., and Behaeghe, T. (1989). Leaf and canopy responses of to long-term elevated atmospheric carbon-dioxide concentration. 177, 312-320. Norby, R. J., Gunderson, C. A., Wullschleger, S. D., O'Neill, E. G., and McCracken, M. K. (1992). Productivity and compensatory responses of yellow-poplar trees in elevated COs. 357, 322-324. Oberbauer, S. F., Strain, B. R., and Fetcher, N. (1985). Effect of COs enrichment on physiology and growth of seedlings of two tropical tree species. 65, 352-356. Oikawa, T. (1990). Modelling primary production of plant communities. 27, 63-80.
8.
121
Olson, J. s., Watts, J. A., and Allison, L.J. (1983). "Carbon in Live Vegetation of Major Ecosystems, ORNL/EIS-109. Oak Ridge National Laboratory, Oak Ridge, TN. Pearcy, R. W., and Bj6rkmann, O. (1983). Physiological effects. "CO2 and Plants: The Response of Plants to Rising Levels of Carbon Dioxide" (E. R. Lemon, ed.), pp. 65-78. Westview Press, Boulder, Co. Poorter, H. (1993). Interspecific variation in the growth response of plants to an elevated ambient CO2 concentration. 104/105, 77-97. Reekie, E. G., and Bazzaz, F. A. (1989). Growth and photosynthetic response of nine tropical species with long-term exposure to elevated carbon dioxide. 79, 212-222. Rogers, H. H., Runion, G. B., and Krupa, S. V. (1994). Plant responses to atmospheric CO2 enrichment with emphasis on roots and the rhizosphere. 83,155-189 S~inchez, P. A. (1976). Properties and Management of Soils in the Tropics." Wiley, New York. Smith, H. (1982). Light quality, photoreception and plant strategy. 33, 481-518. Strain, B. R., and Bazzaz, F. A. (1983). Terrestrial communities. "CO2 and Plants: The Response of Plants to Rising Levels of Carbon Dioxide" (E. R. Lemon, ed.), pp. 45-60. Westview Press, Boulder, Co. Strain, B. R., and Cure, J. D. (1985). "Direct Effects of Increasing Carbon Dioxide on Vegetation. Carbon Dioxide Research, State of the Art," Publication No. ER-0238. U.S. Dept. of Energy, Washington, DC. Thomas, R. B., and Strain, B. R. (1991). Root restriction as a factor in photosynthetic acclimation of cotton seedlings grown in elevated carbon dioxide. 96, 627-634. Thompson, G. B, and Drake, B. G. (1994). Insects and fungi on a C3 sedge and a C4 grass exposed to elevated CO2 concentrations in open-top chambers in the field. 17, 1161-1167. Vitousek, P. M., and Sanford, R. L.,Jr. (1986). Nutrient cycling in moist tropical forests. 17, 137-167. Whittaker, R. H. (1975). Communities and Ecosystems." MacMillan, New York. Williams, W. E., Garbutt, K., and Bazzaz, F. A. (1988.) The response of plants to elevated COz. V. Performance of an assemblage of serpentine grassland herbs. 28, 123-130. Zak, D. R., Pregitzer, K. S., Curtis, P. S., Teeri, J. A., Fogel, R., and Randlett, D. L. (1993). Elevated atmospheric CO2 and feedback between carbon and nitrogen cycles. 151, 105-117. Zangerl, A. R., and Bazzaz, F. A. (1984). Response of plants to elevated CO2. II. Competitive interactions among annual plants under varying light and nutrients. 62, 412-417. Ziska, L. H., Hogan, K. P., Smith, A. P., and Drake, B. G. (1991). Growth and photosynthetic response of nine tropical species with long-term exposure to elevated carbon dioxide. 86, 383-389.
This Page Intentionally Left Blank
Responses to Elevated C02 in Mediterranean Old-Field Microcosms: Species, Community, and Ecosystem Components
Despite voluminous literature on the effect of CO2 on plants cultivated in pots, great uncertainty exists concerning the eventual response of ecosystems and of species and communities within ecosystems (Bazzaz, 1990; Mooney 1991; Strain and Thomas, 1992; K6rner, 1993). The difficulty to extrapolate from short-term laboratory experiments on individual species to the long-term response of ecosystems arises from many sources. These include the interactions between CO2 and other environmental factors which are inadequately reproduced in controlled environments, the interactions between ecosystem components (e.g., plant-plant, plant-soil biota, plant-predators), and the too short time and space scales of the experiments with regard primarily to the dynamics of soil organic matter, the evolutionary response of populations, and the dynamics of vegetation. Although the free-air CO2 enrichment system is technically the most suitable to address the problem at the ecosystem level (Schulze and Mooney, 1994), its very high running cost has limited its use to a few agricultural systems. However, whole ecosystem enclosures such as open-top chambers or greenhouses (K6rner 1994; Vourlitis and Oechel, 1994; Jenkins and Wright, 1994) provide an adequate alternative as far as realistic atmospheric, competitive, and soil conditions can be maintained. Our current knowledge of reponses to CO2 in an ecosystem context strictly rely on such experimental facilities. 123
Copyright 9 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.
Extensive studies have been conducted and reported for three natural ecosystems: the Alaskan tundra (Oechel 1994), a coastal salt marsh (Drake and Leadley, 1991), and a tallgrass prairie (Owensby 1993, and also Chapter 24 in this volume). Experiments are also conducted on alpine grassland (Diemer, 1994), temperate grassland (Leadley and K6rner, see Chapter 11), Mediterranean grasslands (Jackson 1994' and also Chapter 10 in this volume) and macchia (see Scarascia, Chapter 14). Here we present results of studies on Mediterranean old-field ecosystems using, on a larger scale, the excavated microcosms technique pioneered by Billings (1984). In these old-field ecosystems, the species diversity is high (about 25 species/m 2, with very few C 4 species), the fertility of the soil is low and drought usually occurs at the end of spring and during summer. This set of characteristics differs from the ones of the three ecosystems extensively studied so far. It allows us to address in particular the following points: the differential response of plant species, and the relationship between CO2 response and availability of nutrients and water. This chapter is a synthesis of the results of three recent successive experiments, with characteristics imposed by the availability of CO2 enrichment facilities and by the hypotheses to be tested. We concisely extracted the main results of each experiment, looking at their consistency and complementarity. The differential response of the species, the response of ecosystem processes, and the interactions between ecosystem response and the diversity of the plant community are the topics addressed. We finally put into perspective the different time scales of the mechanisms involved at the different levels and their relative potential consequences.
A. Experiment A: Intact Microcosms in Growth Chambers
Sixteen 0.5 • 0.5 • 0.2 m soil monoliths with intact vegetation were excavated from an olive tree orchard abandoned 20 years ago, located 15 km west of Montpellier. The community was an open Mediterranean grassland dominated by Asteraceae, Fabaceae, and Poaceae. The soil was brown calcareous with gravel and numerous stones. Its pH and C / N were 7.6 and 9, respectively. Four microcosms were transferred to each of four growth chambers, two run at 350/~mol mo1-1 of CO2 and two at 700. Microcosms were rotated within and between growth chambers every 2 weeks and 2 months, respectively. Mean weekly day and night temperatures were set to match the values recorded at a nearby meteorological station. Light intensity and photoperiod were also set to mimic seasonal changes. Water was added such that water potential of the plants in the growth chamber was in the
9.
125
range of the water potential of the plants The experiment lasted from mid-February to early July 1992 when all annual species had completed their growth cycle. See Navas (1995) for additional information.
B. Experiment B: Intact Microcosms in Greenhouses; Interaction with Nutrients and Temperature Thirty-two 0.71 X 0.71 X 0.28 m monoliths with intact vegetation were excavated from the periphery of Camp Redon, the experimental field at CEFE-CNRS in the outskirts of Montpellier. The type of vegetation was similar to the previously m e n t i o n e d one with 25% of species in common, including some of the d o m i n a n t species. The soil was also of a similar type (with pH = 7.3 and C / N = 9) but deeper and with fewer stones. The microcosms were transferred to two climate-controlled greenhouses providing outside conditions of temperature and vapor pressure deficit and with either 350 or 700/~mol mo1-1 CO2. The CO2 treatment was crossed with a fertilization treatment: 0 or 100 kg ha -~ of N, P, and K was applied at the beginning of the experiment. Water was added along the same principle as in Experiment A but fertilized microcosms n e e d e d more water than nonfertilized ones during the second half of the experiment due to higher biomass production. Four microcosms were used per treatment but represented pseudoreplicates because they were grown in a single greenhouse. The experiment lasted from 3 April 1993 to the end of the cycle of annual species (early July). C. Experiment C: Planted Microcosms in Greenhouses; Interaction with Biodiversity Fifty-two 0.71 X 0.71 X 0.28 m monoliths were excavated from the center of Camp Redon at the end of s u m m e r 1993. The site was ploughed 4 years before and left fallow. As in Experiments A and B, the vegetation was dominated by annual grasses, composites, and legumes. The soil was similar to previous type (with pH 8.2 and C / N = 10.1) but with even fewer stones. The microcosms were transferred to four greenhouses of which two were at 350/~mol mo1-1 CO2 and two at 700. Air temperature and vapor pressure deficit in all greenhouses tracked outside conditions. Microcosms were rotated within and between greenhouses in the fall and in the spring. Vegetation of the monoliths was removed and the soil surface was tilled to 5 cm. Artificial communities were established by transplanting 1-week-old seedlings at a density c o m m o n in old fields (684 plants/m2). The soil was watered near field capacity during a 2-month establishment period and then kept between 60 and 80% of field capacity. In May, July, and August a d r o u g h t was imposed with the soil kept between 40 and 60% of field capacity.
Nine plant diversity treatments were designed to test the potential role of functional groups (annual grasses, annual composites, annual legumes, and perennial grasses) and analogue species within these functional groups. Five out of the nine diversity treatments were applied both u n d e r ambient and elevated COs: bare soil; one grass; six grasses; three grasses mixed with three composites; and a mixture of two grasses, two composites, and two legumes. All species in these treatments were annuals. Each diversity treatm e n t at each COs level was installed in four microcosms. However, due to the duration of planting, these four microcosms constitute a date of plantation treatment rather than true replicates. We will report here primarily first-year results of the most complex treatment (two grasses, two composites, and two legumes) as well as some results on the impact of removing from this treatment the legumes and the legumes plus the composites.
A. Biomass Production Among all species present in Experiment A, 57 species had at least five individuals at each COs level and were then examined. As many species responded negatively as positively but the COs effect was significant for only five species: and had a positive response and a negative one. Looking for trends across all species, as many annual as perennial species were either stimulated or inhibited and the proportion of species stimulated under elevated COs was 60% in the composites, 40% in the legumes, and 30% in the grasses. A similar diverse response was found in Experiment B. Interestingly, for the 19 species present in all CO2 and fertilization treatments, a highly significant positive correlation was found between the COs and fertilization responses. In particular, four species (Galium and were extremely responsive to both treatments. In Experiment C, the two composites did not respond whereas in each of the grass and legume families, one species responded positively and respectively) and the other negatively and respectively). B. Growth and Phenology The n u m b e r of leaves and branches at the census dates were increased by 20 to 30% at elevated COs for the two legumes studied, and in both Experiments A and C. The composites did not respond whereas some of the grasses (e.g., showed
9.
127
a tendency to have a reduced number of leaves and tillers. For all species, the change in leaf number resulted from a change in branch or tiller number. Among the 17 species for which fruit set time was recorded in Experiment A, it occurred significantly earlier (by 1-4 weeks) for 3 species and significantly later (by 1-3 weeks) for 4 species. In Experiment B, among the 5 species that were analyzed for the time of flower emergence, 3 had a delayed emergence by 1-3 weeks (Fig. 1) while 2 were only marginally affected. Fertilization also delayed flower emergence, but elevated CO2 and fertilization, when applied together, had the opposite effect (flower emergence was accelerated by about 1 week). In Experiment C, the number of flowers at a given time also suggests a species-specific effect of CO2 on reproductive phenology, but no trends at the family or biological type can be drawn from any of the three experiments.
C. Reproduction In Experiment A, 19 species were analyzed for reproductive parameters. Although the effects of elevated CO2 were significant in only a few cases (see Table I for the species for which all parameters were measured), some trends appear at the family level: the number of fruits and seeds per plant were reduced for the grasses and increased for the legumes. The ratio of reproductive to vegetative biomasses was reduced for 13 of the 19 species (significantly for and and increased for 6 species (4 composites and 2 species; significantly for
D. General Trends Besides a species-specific, positive or negative effect of CO2, some general trends at the family level emerge (Table II). The most notable is the opposite response of the grasses and the legumes (negative for the grasses and positive for the legumes) with regard to the number of fruits (and consequently of seeds). This appears to be the consequence of the effect of CO2 on the number of tillers or branches. and for which the number of ramifications are increased by CO2, are legumes, but they are also species with an indeterminate type of growth. Which one of these two characters, atmospheric nitrogen fixation and indeterminate growth, makes them respond positively to CO2 is unknown.
A. CO2 Exchanges The diurnal C O 2 n e t exchanges (plant photosynthesis minus plant and soil respiration) were stimulated by CO2 on a short-term basis, by generally
CO2 effect on the floral phenology of three species (V.2 vegetative phase; FI.1 flower bud stage; F1.2 open flower stage; F1.3 senescent flowers; Fr. fruiting stage).
about 50%, at the beginning of the experiment and along the season. This stimulation, however, was not maintained (it was not detectable 1 month after the start of the CO2 treatment, Experiment A) and on the long-term the diurnal CO2 exchanges were not different at ambient and elevated CO2 (Fig. 2a). The downward adjustment of photosynthesis, calculated from short-term reversed CO2 levels, was between 20 and 40% (similar results in Experiments A and B). In the fertilized microcosms at peak biomass, the long-term stimulation by CO2 of the diurnal CO2 net exchanges was larger than the short-term one (48 versus 30%) due, at least partly, to an
9.
Species
Fruit number Seed number Seed number Reproductive biom./ per plant per fruit per plant vegetativebiomass 0 0
0
0
0
*, P < 0.05.
increase in LAI at high C O 2 (Experiment B). Nocturnal C O 2 exchanges (plant plus soil respiration) were consistently doubled u n d e r elevated CO2 (Experiments A and B). When above- and belowground components of the CO2 exchanges were dissociated (Experiment C), net canopy photosynthesis was stimulated by 20%, canopy night respiration was reduced by 40%, and belowground respiration was increased u n d e r elevated CO2. B.
StandingBiomass
Aboveground biomass at the end of the season (July) was not significantly increased by elevated CO2 in any of the three experiments (Fig. 2b). However, when fertilization was added (Experiment B), the CO2 effect resulted in a significant 40% increase. Root biomass at the end of the season was not affected by CO2 treatment in Experiment A. In Experiment C, root mass and length were increased by 33 and 24%, respectively. For the whole community, therefore, root-specific length was then not increased by CO2. Litter could be analyzed separately only in Experiment A: it was doubled u n d e r elevated CO2. C. Water and Nutrient Fluxes Total water consumption during the n o n d r o u g h t period was slightly less at high CO2. When drought was imposed, the soil desiccation rate was
Family Grassses Legumes Composites
Vegetativebiomass Numberof modules 0 +
+ 0
Reproductive biomass
Fruitand seed number
0 0 0
+ 0
(a)Effect of CO2 on the CO2 diurnal exchanges of Mediterranean old-field ecosystems at the end of April. Light conditions during measurements were 800, 700, and 1200/zmol m -2 s -1 in Experiments A, B, and C, respectively. (b) Effect of CO2 on aboveground biomass in the same ecosystems at the end of the season (July). The water conditions mimicked outside ones in Experiments A and B, but were nonlimiting in Experiment C except for a d r o u g h t period in May.
marginally different between the CO2 treatments (Experiment C). Plant xylem water potential was generally less negative (by 0.2-0.5 MPa) at elevated than at ambient CO2 (Experiments A, B, C). Nitrogen mineralization rate was doubled under high CO2, but the amount of nutrients removed from the soil and stored in the end-of-season aboveground (dead) biomass was reduced under high CO2 for all of the 12 mineral ions analyzed, the reduction was larger for the oligoelements compared to N, P, or K. Because biomass was nearly identical in the two CO2 treatments, this reduction was the consequence of a decrease in mineral concentration of the plant tissues.
9.
131
For tissue, both carbon content and C / N ratio were increased (Experim e n t C).
D. Soil Microbiology Preliminary results show a 70% increase in microbial biomass-C (chloroform fumigation-extraction) u n d e r elevated CO2. There is also a significant increase in hyphal lengths as well as in arbuscularomycorrhizal infection. The soil enzymes investigated (dehydrogenase, cellulase, phosphatase, and xylanase) also show an increased activity under elevated CO2. When screening a large variety of substrates (e.g., carbohydrates, carboxylic acids polymers, amides, amines, amino acids), the soil bacterial community from the elevated CO2 treatment utilized a higher n u m b e r of substrates (Experiment C, April sampling). E. Interaction between Community Type and Ecosystem Response Three plant diversity treatments (grasses, grasses + composites, grasses + composites + legumes), all with a similar total n u m b e r of species (six), allow us to address this topic. The cycle of the annual composites is about 1 month longer than the cycle of the annual grasses and thus is more affected by the summer drought. The hypothesis underlying the comparison of the first two treatments was that elevated CO2 will alleviate the drought stress and have a larger effect on the grasses + composites treatm e n t than on the grasses treatment. Since pot and agronomic experiments have shown that legumes respond more to CO2 than non-nitrogen-fixing species, the hypothesis was that the treatment including legumes will respond more to CO2 than will the other treatments. No significant difference in the response to CO2 was found among the three treatments for the parameters analyzed so far, including photosynthesis and biomass production. However, preliminary results of the soil parameters (microfloral) suggest interactions could take place in relation to the presence of leguminous species.
A. Effect of CO2 on Individual Species Biomass Production The C O 2 growth response of the individual species present in our microcosms was characterized by a high interspecific variability. A negative response was almost as frequent as a positive one and a large intrapopulation variability made most differences not significant. A large interspecific variability was also found in the laboratory experiments on individually potted plants (e.g., Kimball, 1983) and some types of species (fast-growing species, legumes) have been shown to be, on average, more responsive than others
(Poorter, 1993; Hunt 1994). However, the percentage of the overall variability explained by the specific response of the plant categories identified is low (see Poorter Chapter 25), and part of this categorization (the legumes that respond more than the nonlegumes) is generally not validated by data from natural community experiments. Negative responses to CO2 have also been found in laboratory-grown plants, but they generally represent a low percentage of the species (< 10%). The much higher percentage of species responding negatively found in our microcosms could result from the low fertility level of our natural soil compared to the one used in pot experiments, if the CO2 response is decreased at low nutrient level. Such a decrease is found, for single-grown plants, in the literature review of Poorter and co-workers (see Chapter 25), but the opposite is found in the larger review of Idso and Idso (1994). In plant communities, Bazzaz (1990) suggested that interspecific competition changes the effect of CO2. The more responsive species strongly outcompete the other species to the point that the latter perform more poorly at elevated COz than at ambient COz. Changes in competitive relationships with CO2 concentration have been observed but only in laboratory experiments (Bazzaz and Garbutt, 1988). B. Effect of CO2 on Phenology and Reproduction The vegetative and reproductive phenology of some species were strongly affected by CO2 and an interaction with soil fertility was shown. In the literature, phenology have been shown to be either delayed, advanced, or not affected by CO2, depending on the particular species and environmental conditions (Rawson, 1992) and these effects are not always correlated with the effects on growth (Marc and Gifford, 1984; Reekie and Bazzaz, 1991; Ellis 1995). Reekie (1994) found the time of flowering was delayed in four short-day species and advanced in four long-day species at elevated CO2. No trend was found in Experiment A between the time of fruiting and its modification by CO2. In Experiment B, the general trend for a delayed reproductive phenology concerned short-day species (spring flowering species) and is in agreement with the results of Reekie (1994). However, the reverse trend observed when fertilization was applied (Experiment B) confirms the difficulty in comparing phenological responses to CO2 in studies that are not done under similar experimental conditions (Reekie 1994). Effect of CO2 on reproductive output was extremely variable between species and within populations. However, when the numbers of fruits and seeds were considered, the effect of CO2 tended to be negative for the grasses and positive for the legumes. (This effect appeared to be mediated by a change in the number of fillers or branches, the two legumes analyzed having an indeterminate type of growth.) So far, the limited information
9.
133
available about the effects of C O 2 o n reproduction suggest they are rather unpredictable but can have a large impact on the demography of natural population (Bazzaz, 1990). Results on cereals (Ford and Thorne, 1967; Gifford, 1977; Baker 1990) show an effect of CO2 on tillering contrary to the one we observed in wild grasses. However, the effects of CO2 on modular growth at vegetative and reproductive stages appear worthy of further analysis. C. Intraspecific Variability of the Response to CO~ In all experiments, the variability at the species level was very high and resulted in very few species significantly affected by CO2 despite a consequent average response for many of them. The variability between experiments was also very large: none of the species whose biomass was significantly affected by CO2 were common to the three experiments, but among the 10 common species, opposite trends of the response were found among experiments. For some species, too few individuals were found in either one or all treatments, but in general this variability suggests strong interactions between C O 2 t r e a t m e n t s and other factors. One of these factors might be local soil fertility since we showed a strong interaction between CO2 and nutrient treatments for species biomass production and phenology. This local fertility might be partly dependent on the competition with neighboring plants. We also found, in Experiment C, a large variability between the four "replicates" of a given treatment, sown over a 2-week period each, suggesting a strong interaction between the CO2 effects and the relative phenology of the species. Within and between experiment variability for a given species, response can also result from a differential CO2 response of the genotypes and populations (cf. Tousignant and Potvin, Chapter 3), which is expected to be higher in natural populations than in selected lines often used in agronomic experiments. This high variability certainly dictates specific experimental designs, but it also likely characterizes the actual response of the plants to CO2, which is strongly dependent on interactive factors. Instead of being judged only as a serious experimental drawback for working with natural systems, this variability ought to be seen as the reality we need to analyze and one which involves too many interactions for its functioning to be currently predicted from simplified systems. D. Effect on Ecosystem Processes
The responses of the ecosystem processes to elevated CO2 are summarized in Fig. 3. Photosynthesis and standing biomass were marginally stimulated by CO2 and often not significantly. Changes in water regimes (seasonal ones in Experiment A or between Experiments B and C) were not associated with changes in the response to CO2 (photosynthesis in Experiment A or
134
Figure 3 Summaryof the effect of elevated CO2 on the major processes in Mediterranean old-field ecosystems characterized by a low soil fertility and a high plant species diversity. biomass in E x p e r i m e n t B versus C). A large stimulation was observed only when fertilizer was added. O n the contrary, soil parameters (respiration, mineralization rate, microbial biomass, mycorrhizal infection, enzymic activity) were significantly increased by CO2 u n d e r the natural low soil fertility. These results, in line with those of K6rner and Arnone (1992), Diaz (1993), and Oechel (1994) support the hypothesis that low fertility
9.
ecosystems--the major part of the earth land area--will have a limited plant growth response to CO2 (Mooney 1991). This empirical evidence totally contradicts the conclusion of Idso and Idso (1994), based on 10 years of pot experiments by the international community, that "the percentage growth responses of natural ecosystems to CO2 enrichment could well be greater than those of managed agricultural systems." The increased soil activity under elevated CO2 is probably the consequence of an increased input of carbon through larger exudates a n d / o r increased root turnover (O'Neill, 1994). Feedbacks to the plant response primarily through changes in nutrient availability have not been established (Curtis 1994a). A change in litter quality and in its decomposition could alter the nutrient availability for the plant and could serve as a feedback mechanism. Studies by Gorissen (1995) using grass roots and by Kemp (1994) using leaf litter show contrary effects.
E. Plant, Population, Community and Ecosystem Responses: What to Look at and at Which Time Scale? The complexity of the response t o C O 2 is due to its indirect effect on plant growth and development, its interaction with the other environmental factors, and its interactions with other ecosystem components. This was well known a decade ago (Lemon, 1983; Strain and Cure, 1985) and this complexity has certainly been documented since. Up to now, most research has been conducted on individual, young plants probably with the thought that the direct effect of CO2 on photosynthesis and subsequently growth would be the dominant one and that most other parameters (e.g., seed production) would change proportionately. A conclusion of our work on natural communities and of recent publications (e.g., Reekie 1994) is that it is not so and that indirect effects on specific characters (e.g., branching or phenology) dominate the response in some species. More studies, in conditions as close as possible as ones, are needed before establishing how general this is, but it suggests that the prediction of the species' response to CO2 will be more difficult than expected and will not be based on a single (growth-related) criteria. The focus of research needs to move from the first month of growth to the whole cycle of the plant. However, evidence of within-population genetic variation in response to CO2 (Curtis 1994b) and of selective effects of CO2 (see Tousignant and Potvin, Chapter 3) strongly invite researchers to extend studies to several generations. Similarly, indirect impacts of CO2 at the community or at the ecosystem levels occur at the time scale of the season (e.g., relative dominance of C3 versus C4 species (Owensby 1993), increased soil moisture (Jackson 1994)), but longer term effects are unknown and several antagonistic hypotheses can be formulated. Long-term, ecosystem scale experiments on representative earth biota are needed to answer our
questions at the ecosystem level. They need to include the studies of mechanisms that are traditionally associated with ecosystem function (fluxes of energy and matter) but also mechanisms that take place at the levels of the species and of the community.
We thank the technical personnel from the physiological ecology group and from the experimental field section of CEFE as well as from the plant biology and pathology unit from the Ecole Nationale Suptrieure Agronomique (Montpellier) whose dedicated work was determinant to the success of the experiments. This research was funded by CNRS (Programme Environnement, Comit6 6cosyst~mes), the Minist~re de l'Environnement (Comit6 EGPN), the European Union (DGXII, Environment programme, contract EV5V-CT94-0428) through grants toJ. Roy and the Minist~re de l'Agriculture (Direction gtntrale de l'Enseignement et de la Recherche) through a grant to M.-L. Navas.
Baker, J. T., Allen, L. H., Jr., and Boote, K.J. (1990). Growth and yield responses of rice to carbon dioxide concentration. J. 115, 313-320. Bazzaz, F. A. (1990). The response of natural ecosystems to the rising global CO2 levels. 21, 167-196. Bazzaz, F. A., and Garbutt, K. (1988). The response of annuals in competitive neigbourhoods: 69, 937-946. Effect of elevated CO2. Billings, W. D., Peterson, K. M., Luken, J. O., and Mortensen, D. A. (1984). Interaction of increasing atmospheric carbon dioxide and soil nitrogen on the carbon balance of tundra microcosms. 65, 26-29. Curtis, P. S., O'Neill, E. G., Teeri, J. A., Zak, D. R., and Pregitzer, K. S. (1994a). Belowground responses to rising atmospheric COs: Implications for plants, soil biota, and ecosystem processes. 165, 1-6. Curtis, P. S., Snow, A. A., and Miller, A. S. (1994b). Genotype-specific effects of elevated COs on fecundity in wild radish 97, 100-105. Diaz, S., Grime,J. P., Harris, J., and McPherson, E. (1993). Evidence of a feedback mechanism limiting plant response to elevated carbon dioxide. 364, 616-617. Diemer, M. W. (1994). Mid-season gas exchange of an alpine grassland under elevated COs. 98, 429-435. Drake, B. G., and Leadley, P. W. (1991). Canopy photosynthesis of crops and native plant communities exposed to long-term elevated COs. 14, 853-860. Ellis, R. H., Craufurd, P. Q., Summerfield, R.J., and Roberts, E. H. (1995). Linear relations between carbon dioxide concentration and rate of development towards flowering in sorghum, cowpea and soybean. 75, 193-198. Ford, M. A., and Thorne, G. N. (1967). Effect of COs concentration on growth of sugar-beet, barley, kale, and maize. 31, 629-644. Gifford, R. M. (1977). Growth pattern, COs exchange and dry weight distribution in wheat growing under differing photosynthetic environments. 4, 99-110. Gorissen, A., Van Ginkel, J. H., Keurentjes, J. J. B., and Van Veen, J. A. (1995). Grass root decomposition is retarded when grass has been grown under elevated COs. 27, 117-120.
9.
137
Hunt, R., Hand, D. W., Hannah, M. A., and Neal, A. M. (1993). Further responses to CO2 enrichment in British herbaceous species. 7, 661-668. Idso, K. E., and Idso, S. B. (1994). Plant responses to atmospheric CO~ enrichment in the face of environmental constraints: A review of the past 10 years' research. 69, 153-203. Jackson, R. B., Sala, O. E., Field, C. B., and Mooney, H. A. (1994). CO2 alters water use, carbon gain, and yield for the dominant species in a natural grassland. 257-262. Jenkins, A., and Wright, R. F. (1994). The 'Climex' project. Raising CO2 and temperature to whole catchment ecosystems. "Design and Execution of Experiments on CO2 Enrichment" (E.-D. Schulze and H. A. Mooney, eds.), pp. 211-219. Commission of European Communities, DGXII, Brussels. Kemp, P. R., Waldecker, D. G., Owensby, C. E., Reynolds, J. F., and Virginia, R. A. (1994). Effects of elevated CO2 and nitrogen fertilization pretreatments on decomposition of tallgrass prairie leaf litter. 165, 115-127. Kimball, B. A. (1983). Carbon dioxide and agricultural yield: An assemblage and analysis of 430 prior observations. J. 75, 779-788. K6rner, Ch. (1993). CO2 fertilization: The great uncertainty in future vegetation development. "Vegetation Dynamics and Global Change" (A. M. Solomon and H. H. Shugart, eds.), pp. 53-70. Chapman & Hall, New York. K6rner, Ch., and Arnone, J. A., III (1992). Responses to elevated carbon dioxide in artificial tropical ecosystems. 257, 1672-1675. K6rner, Ch., Arnone, J. A., III, and Hilti, W. (1994). The utility of enclosed artificial ecosystems in CO2 research. "Design and Execution of Experiments on CO2 Enrichment" (E.-D. Schulze and H. A. Mooney, eds.), pp. 185-198. Commission of European Communities, DGXII, Brussels. Lemon, E. R. (1983). and Plants. The response of plants to rising levels of atmospheric carbon dioxide." Westview Press, Boulder. CO. Marc, J., and Gifford, R. M. (1984). Floral initiation in wheat, sunflower, and sorghum under carbon dioxide enrichment. 62, 9-14. Mooney, H. A., Drake, B. G., Luxmoore, R. J., Oechel, W. C., and Pitelka, L. F. (1991). Predicting ecosystem responses to elevated CO2 concentrations. 41, 96-104. Navas, M. L., Guillerm, J. L., Fabreguettes, J., and Roy, J. (1995). Community structure, co2 fluxes, and production of mediterranean old field microcosms under elevated CO~. 1, 325-335. Oechel, W. C., Cowles, S., Grulke, N., Hastings, S.J., Lawrence, B., Prudhomme, T., Riechers, G., Strain, B., Tissue, D., and Vourlitis, G. (1994). Transient nature of CO2 fertilisation in Artic tundra. 371, 500-503. O'Neill, E. G. (1994). Responses of soil biota to elevated atmospheric carbon dioxide. 165, 55-65. Owensby, C. E., Coyne, P. I., Ham, J. M., Auen, L. M., and Knapp, A. K. (1993). Biomass production in a tallgrass prairie ecosystem exposed to ambient and elevated CO2. 3, 644-653. Poorter, H. (1993). Interspecific variation in the growth response of plant to an elevated ambient CO2 concentration. 104/105, 77-97. Rawson, H. M. (1992). Plant responses to temperature under conditions of elevated CO2. 40, 473-490. Reekie, E. G., and Bazzaz, F. A. (1991). Effect of elevated CO~ on phenology and growth in four annual species grown as individuals and in competition. 69, 2475-2481. Reekie, J. Y. c., Hicklenton, P. R., and Reekie, E. G. (1994). Effect of elevated CO2 on time of flowering in four short-day and four long-day species. 72, 533-538. Schulze, E.-D., and Mooney, H. A. (1994). Comparative view on design and execution of experiments at elevated CO2. "Design and Execution of Experiments on CO2 Enrich-
138 ment" (E.-D. Schulze and H. A. Mooney, eds.), pp. 407-413. Commission of European Communities, DGXII, Brussels. Strain, B. R., and Cure, J. D. (1985). "Direct Effects of Increasing Carbon Dioxide on Vegetation." U.S. Department of Energy, Washington, DC. Strain, B. R., and Thomas, R. B. (1992). Field measurements of CO2 enhancement and climate change in natural vegetation. 64, 45-60. Vourlitis, G., and Oechel, W. C. (1994). Microcosms in natural experiments. "Design and Execution of Experiments on CO2 Enrichment" (E.-D. Schulze and H. A. Mooney, eds.), pp. 199-210. Commission of European Communities, DGXII, Brussels.
1 Annual Grassland Responses to Elevated C02 in Multiyear Community Microcosms
Responses to elevated CO2 can involve components that are difficult to resolve in natural communities. Spatial variation in plant community structure, disturbance regime, and belowground properties can introduce so much noise that community responses to elevated CO2, not to mention species responses, can be detected only with highly replicated experiments. Even then, detecting CO2 responses against the inherent variability may require intensive measurements, which are often inconsistent with preserving an intact community. Microcosm experiments can serve as more uniform, accessible analogs if they are functionally similar to the natural community. Functional similarity is probably easiest to achieve in the short term. Establishing long-term microcosms that approximate natural processes is likely to be possible only with certain ecosystems. Here we report on microcosm communities established as a 4-year experiment starting in fall 1992 as part of a study of the response of California annual grassland to elevated CO2 (Field 1996a). We examine species and community aboveground production and plant density during the microcosms' second year, the 1993-1994 growing season. We focus on (1) microcosm responses to elevated CO2, (2) microcosm responses to the addition of nutrients or exotic species, and (3) interactions among elevated CO2, nutrient addition, and presence of exotics. The first of these can be compared with responses measured in open-top chambers in natural 139
Copyright 9 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.
140 grassland. We use the results to address the general issue of whether microcosms are good analogs for assessing long-term field responses to elevated CO2, and, if so, to what extent they provide access to questions that cannot be answered in the field.
Both microcosm and field components of the Jasper Ridge CO2 experiment are located within the Jasper Ridge Biological Preserve of Stanford University. The microcosm facility is 1.5 km west of the experimental field sites, which are ridge-top grasslands in the core of the preserve. Together the microcosms and field studies are part of a long-term study of the roles of resource availability, species characteristics, and community composition in controlling ecosystem responses to increasing CO2 (Field 1996a). The field studies take advantage of a sharp, natural boundary between shallow, nutrient-poor, serpentine-derived soils supporting native forbs and grasses, and more fertile sandstone soils dominated by Eurasian annual grasses. In the microcosm experiments presented here, we focus on serpentine communities grown on serpentine soil, but we also consider their susceptibility to invasion by introduced, sandstone species found in the adjacent grassland. Combining low-to-modest productivity, high species diversity, dominance by annual species, a high frequency of soil disturbance, and high variation in annual precipitation, California annual grasslands constitute a good model system for probing how resources and species characteristics shape an ecosystem's response to increasing CO2 (Field 1996a).These features make it feasible to have experimental units small enough to be highly replicated and still encompass much of the functional diversity of the ecosystem. For example, the serpentine field chambers are 0.65 m in diameter and typically contain 1000-3000 individual higher plants with 10-15 species per 0.01 m 2. The major axes of functional diversity in the microcosm experiments that we present here are phenology, which varies considerably among the annual forbs and is correlated with roofing depth and other properties that affect ecosystem processes (Gulmon 1983), and grasses versus forbs, which respond differently to nutrient manipulations in the field (Hobbs 1988). The species that fit these categories are dominant in the serpentine grassland and lack a significant seed bank, offering the possibility o f detecting changes in species or functional composition in multiyear experiments. Annual legumes and perennial forbs and grasses
10.
141
are also present in the system but are less dominant a n d / o r have significant seed dormancy. In designing the microcosm experiments, we aimed for a community that was representative of the grassland as a whole, both in terms of major species and functional groups, while at the same time allowing comparisons with the CO2 fumigation area. A second goal was to have each species and functional group represented sufficiently to detect treatment effects on each species (or functional group) and the impact of their responses for the whole community. The principal functional groups we included were early-flowering, native annual forbs; late-flowering, native annual forbs; mid-season, native annual grasses; and mid-season, introduced annual grasses (Table I). Although most serpentine species in these categories are widely distributed throughout the grassland, several major species have distributions that involve patchiness above the scale of the field site. for example, is an early annual abundant in the field chambers for the CO2 experiment, but it is rare or absent over large areas of grassland. It was not included in the microcosms. One of the late annuals, is very c o m m o n in most of the grassland but is replaced
SS 625 1250
o
1250 1900
.4 total
1250
1940
1260
4000
2640
75
50
150
100
8065
8425
a S, serpentine species only; ss, serpentine plus sandstone species. Nomenclature as in Hickman (1993).
by a second late annual, in most of the field CO2 plots. For the microcosms, we included both species of late annuals.
The microcosm facility consists of 20 enclosures that contain two or three community microcosms 0.4 m in diameter and either 24 or 28 smaller microcosms 0.2 m in diameter, all with a 0.95-m-deep soil column (Field 1996a). Soil for the large microcosms consists of 0.8 m subsoil/ crushed rock from a serpentine quarry and 0.15 m serpentine topsoil. Before filling the pots, we watered the topsoil to pregerminate resident seed, air-dried the soil outdoors, and then shredded it. The small microcosms are used for a variety of monoculture and community experiments. A 2.5-mtall open-top chamber of polyethylene (0.15 m m thick in Year 1, 0.10 m m thick in Year 2) sits above each enclosure. Each open-top chamber was topped with a polystyrene grill (Year 1) or polyethylene mesh (Year 2) to evenly distribute rainfall over the enclosure. Treatments consisted of a balanced three-factor design involving two CO2 treatments (ambient versus ambient plus 35 Pa CO2), two nutrient treatments (ambient versus surface application of slow-release fertilizer (14% N, 14% P, 14% K, equivalent to 20 g N, 20 g P, and 20 g K m -2, Osmocote), and two communities (serpentine species versus serpentineplus-sandstone species). The 20 enclosures were set up as a 4 • 5 array, with each 2 • 5 half of the array air-fed by a large blower. To avoid extreme differences in plant height among adjacent microcosms within an enclosure, nutrient treatments were assigned at the level of the enclosure in a checkerboard pattern. The two community treatments (serpentine versus serpentine-plus-sandstone species) were distributed to provide four replicate enclosures (two with ambient nutrients, two with added nutrients) of all three possible combinations of two microcosms, and four replicate enclosures of the two combinations of three microcosms with both community treatments present, for a total of five combinations. Within enclosures, community treatments were positioned randomly. We set targets for initial species abundances (Table I) based on four sources of information: (1) species versus area relationships for four sites in the serpentirie grassland, (2) previous censuses of 900 serpentine plots of 0.01 m 2, (3) censuses of an additional 30 0.01-m 2 plots from the first year of open-top chamber experiments in the field, and (4) previous experiments growing serpentine communities in similarly deep pots. We selected a standard community of five native species that would also be appropriate for 1-year companion experiments in the smaller pots (0.2 m diameter). To test the hypothesis that increased CO2 might increase the invasibility
10.
143
of the community, we added three introduced species to form the serpentine-plus-sandstone treatment. In order to meet our goals of preserving community structure while at the same time achieving sufficient biomass of each species, we focused on the most d o m i n a n t species, which show a roughly inverse relationship between individual biomass and density. The most abundant early-annual forbs had average densities of about 1000-5000 m -z, and individual plant biomass at maturity averaged 2 - 6 mg. The late-flowering annual forbs had much lower density (100-500 m-2), with much higher individual biomass (0.05-1 g). Limiting our n u m b e r of species to five native annuals (_+three introduced annuals) helped define a community of species that typically co-occur and produce comparable aboveground biomass per m 2. Where possible, we included two species in each functional group in order to test whether members of a single functional group responded similarly to elevated CO2 (Table I). We chose two of the four very c o m m o n species of native, early-annual forbs, both of the two dominant species of native, late-annual forbs, and the only species of native annual grass present. Target densities were determined from studies of field plots that included the two late-flowering native annual forbs and the two early-flowering forbs. For all species, target seedling densities were set to the mean adult densities in the field. Because of the reduced species n u m b e r relative to the field, this yielded total densities about two-thirds the field average but sufficient for full production (unpublished data). Seeds for all experiments were collected near the field site. For the introduced annual grasses, we included the two species, and that were present in the C O 2 fumigation plots on serpentine soil. We also included an introduced species that increased height and seed production in 1 year (Jackson 1994) but not in another year (Jackson 1995) in the field fumigations on sandstone soil. not present in any serpentine plots. In previous studies, adding nutrients to serpentine grassland enhanced invasion by and but not (Hobbs 1988). When introduced grasses were included in the microcosms, the target densities of the native species were reduced by one-third (Table I). Target densities for and were set at the mean of field plots containing these species. We set the target density lower than the other grasses, reflecting typically greater biomass. Only the species used in the community microcosms were used for experiments in the smaller microcosms surrounding them. Seeds of all species were planted November 24 and 25, 1992, into microcosms positioned in their enclosures with soil levels 30 m m below the pot rim. The largest seeds were placed on the soil surface in the serpentine-plus-sandstone treatment, 10 m m of soil was added above
144 them, then all remaining seeds were added except then another 10 m m of soil, and finally seeds were sprinkled on the surface. The microcosms received natural rainfall inputs, with two supplemental waterings (each equal to 20 m m precipitation) during late spring. During the two periods of peak flowering by forbs (early April and late August), we removed the open-top chambers for 1 week to provide access for pollinators. After the end of the first growing season, we placed a chimney of aluminum window screen on top of each microcosm to maximize self-seeding within microcosms and to minimize seed dispersal. Except for removal of the open-top chambers for pollinator access, treatments were continuous for 2 years. Prior to germination in Year 2 (October 1993), we repeated the Osmocote application for the added-nutrient treatment. After germination, the monoculture experiments in smaller microcosms had weed seedlings of the grass species, indicating that we were not completely successful in preventing seed dispersal between microcosms. A 0.15-m-diameter ring was placed on the soil surface as a long-term census plot in the center of each community microcosm with ambient nutrients. We counted individuals per species during peak flowering. Densities of plants and fillers were so high in the microcosms with added nutrients that counts of individuals were not possible. In late May 1994, following flowering of the early species, we harvested aboveground biomass from one-eighth the area of each microcosm (a 0.013-m 2 wedge-shaped section). We sorted the material into groups: individual species, abscised leaves not identifiable to species, and litter from the previous year. T h e n we dried and weighed each fraction. In September during full flowering of we measured biomass indirectly in the ambient-nutrient communities by measuring the length of every stem on every plant of and using field-harvested plants to establish a length versus dry weight relationship.
We analyzed biomass and density responses separately because harvests and censuses were based on different samples, and censusing was limited to the ambient-nutrient treatments. Based on the harvests from all communities, we evaluated the effects of nutrient addition, community composition, elevated CO2, and interaction effects on aboveground production by each species and the community using ANOVAs (Systat, Evanston, IL). We used the censuses of ambient-nutrient communities to evaluate density responses of species as a function of community composition, elevated CO2, and their interactions. One species of late-flowering, native annual forb, was lost from all but one community microcosm through low
10.
145
rates of germination, survival, and set seed during Year 1. It was omitted from the analyses. Accidental seed dispersal among microcosms (within an enclosure) potentially confounded the effects of treatments on dynamics of introduced grass species. To separate species dynamics into treatment effects versus effects of invasion from neighboring microcosms, we used an index of the population of potential invaders as a covariate in the ANOVAs. The index we used for each species was its background population, calculated as the number of seeds of that species planted in other microcosms of the same enclosure during Year 1. For the communities with only serpentine species, this index included the background population both in small monoculture microcosms and in community microcosms that included introduced grasses (the serpentine-plus-sandstone treatment).
A. Production Totals
In the ambient-nutrient microcosms with only serpentine species, aboveground community production in May 1994 averaged 78.2 g m -2 under ambient CO2 and 78.9 g m -2 under elevated CO2. For the serpentine-plussandstone community, the same treatments averaged 61.8 and 75.0 g m -2. Values from the field experiments were quite similar, averaging 65-74 g m -2 in the three treatments (no chamber, ambient CO2 chamber, and high CO2 chamber) (Fig. 1). Aboveground production in the added-nutrient serpentine communities (which received applications of 20 g N, P, and K m -2 for 2 years) averaged 512-666 g m -z, compared with 657 g m -2 for natural serpentine grassland after two annual applications of Osmocote, 1988). Addedeach equivalent to 31 g N, P, and K m -2 (Hobbs nutrient communities that included sandstone species averaged 827 and 950 g m -2 for ambient and elevated CO2 treatments. Aboveground litter in the ambient-nutrient microcosms averaged 7.817.3 g m -z, about one-ninth the values in the field for the same year, and one-sixth the field values the previous year (their second year of CO2 fumigation) (Fig. 1). Litter values for the added-nutrient microcosms were intermediate between the field and ambient-nutrient microcosms. By the May 1994 census, plant density in the ambient-nutrient microcosms had increased by 76 and 28% above the November 1992 seeding densities for the serpentine-only and the serpentine-plus-sandstone communities (Fig. 2). The resulting densities of 14,250 m -2 in the serpentine-only community and 10,800 m -2 in the serpentine-plus-sandstone community bracket the value (12,000 m - z ) for serpentine field plots, which have densities of and intermediate between the two microcosm communities.
146
(Top) Aboveground production (g m -2) in ambient-nutrient and added-nutrient community microscosms measured in their second year (May 1994) as compared with measurements from field CO2 experiments during April 1994 and with nearby plots studied by Hobbs (1988). (Bottom) Aboveground litter from the previous year's production (gm -2) in community microcosms measured in their second year (May 1994) as compared to measurements from field CO2 experiments during their second year (April 1993) and third year (April 1994). Treatments in the field COz experiments include ([--]) no chamber, (El) open-top chamber with blower and no COs addition, (11) chamber with blower at ambient plus 35 Pa CO2. (Hobbs 1988.)
The late-annual was nearly a b s e n t f r o m the a d d e d - n u t r i e n t c o m m u n i t i e s b u t c o n s t i t u t e d 6 - 1 2 % o f the a b o v e g r o u n d b i o m a s s o f a m b i e n t - n u t r i e n t c o m m u n i t i e s in the May 1994 harvest (Fig. 3). B e t w e e n p r o d u c t i o n in t h e a m b i e n t - n u t r i e n t microMay a n d S e p t e m b e r , cosms i n c r e a s e d by u p to eighffold. This late p r o d u c t i o n i n c r e a s e d total
10.
147
Initial seeding density by species for the two community treatments (serpentine species only (s) vs serpentine-plus-sandstone species (ss)) and mature plant density in the microscosms during the second year at ambient CO2 (c) vs elevated CO2 (C). All microcosms shown here were grown with ambient nutrients (f).
aboveground production for the growing season to 93-121 g m -2, of which 28-36% was late annual biomass. B. Nutrient Effects The added-nutrient treatment affected the aboveground production of every species ( P = 0.000-0.0186), but grasses and forbs had opposite responses (Table II; Fig. 3). All four grass species had higher aboveground production with the nutrient addition, while all three forb species had higher production without it. Community production and aboveground litter also increased with added nutrients (P < 0.0001). C. Community Effects The initial community composition (serpentine versus serpentine-plussandstone species) affected five of the seven annual species in terms of aboveground production, plant density, or both (Table II; Figs. 2, 3). For all five species, production a n d / o r density was higher in the community with the higher initial seeding density of that species (P = 0.000-0.0853).
148
Aboveground biomass by species (g m-2) for the community microcosms in the second year of growth (1994). Treatments include ambient nutrients (f) vs added nutrients (F), serpentine-plus-sandstone community (ss) vs serpentine only (s), and ambient COs (c) vs elevated (C). production was too small to appear.
T h e native species fared better in the serpentine-only communities, and the i n t r o d u c e d species were m o r e successful in the serpentine-plus-sandstone communities, indicating that they had not been lost from the serpentineplus-sandstone t r e a t m e n t but had not fully invaded the serpentine-only community. In May, c o m b i n e d production for the c o m m u n i t y as a whole (including unresponsive species) was higher in the serpentine-plussandstone communities than the serpentine-only communities (P = 0.0399), driven by a c o m m u n i t y effect in the added-nutrient t r e a t m e n t ( P = 0.0273 for nutrient • c o m m u n i t y interaction) (Table II; Fig. 1). Total plant density was 2 4 - 2 8 % higher in the serpentine-only communities than the serpentine-plus-sandstone (P = 0.0017) (Table II; Fig. 2). T h e r e was no effect of c o m m u n i t y on aboveground litter.
D. CO~ Effects In the microcosms, elevated CO2 increased the a m o u n t of aboveground litter present in May (P < 0.0001) (Fig. 1; Table II) but did not affect aboveground production m e a s u r e d at the same time (P = 0.34 for CO2 effect). In the ambient-nutrient t r e a t m e n t in September, elevated CO2 increased the density of ( P = 0.0149) (Table II) and r e d u c e d individual biomass (P < 0.0001). These effects did not balance,
Nutrients • Nutrients
Biomass
Community
F > f
P = 0.0186
Biomass Density
F > f
P = 0.0036 --
ss > s
P = 0.0853
Biomass
F>f
P<0.0001
ss>s
P<
ss > s
P < 0.0001
CO2
community
N u t r i e n t s • CO2
Community
X CO2
Density
Density
~
Biomass
F>f
P<0.0001
Density
~
Biomass Density
f>
F
Biomass Density
P = 0.0013 ~
P = 0.0001
0.0001
s>ss
P<0.0001
s > ss
P = 0.0017
s > ss
P = 0.0455
s > ss
P = 0.0025
P < 0.0001
0.0001 --
Biomass
P = 0.0127 P = 0.0926
0.0003
Density B i o m a s s (Sept.)
C>
c
P=
0.0845
D e n s i t y (Sept.)
C>
c
P=
0.0149
C>
c
P=
0.0699
C>
c
P=
0.0961
m
Community B i o m a s s (May)
F > f
Density (May
P < 0.0001 ~
ss > s
P = 0.0399
s > ss
P = 0.0017
B i o m a s s (total) Litter Biomass
F > f
P < 0.0001
P = 0.0273
C > c
P < 0.0001
P < 0.0001
" B e c a u s e f e r t i l i z e d t r e a t m e n t s w e r e n o t c e n s u s e d , we d i d n o t test f o r effects o n d e n s i t y d u e to n u t r i e n t a d d i t i o n o r its i n t e r a c t i o n s ( i n d i c a t e d by m ) . Effects o f c o m m u n i t y a n d CO2 o n d e n s i t y w e r e d e t e r m i n e d o n l y f o r t h e a m b i e n t n u t r i e n t t r e a t m e n t s . All v a l u e s a r e f r o m M a y 1994 h a r v e s t s a n d c e n s u s e s e x c e p t f o r a d d i t i o n a l m e a s u r e m e n t s o f at m a t u r i t y in S e p t e m b e r . T h e " C o m m u n i t y " 0.1 a r e s h o w n .
b i o m a s s r e s p o n s e is b a s e d o n M a y v a l u e s o f e a r l y a n n u a l s plus S e p t e m b e r v a l u e s f o r
All P v a l u e s less t h a n
150
and total aboveground production was higher under elevated CO2 than ambient CO2 (P = 0.0845). For the ambient-nutrient microcosms, the combined total of early annual production measured in May plus late annual (i.e., production measured in September increased under elevated CO2 (P = 0.0961) (Table II; Fig. 3). E. Nutrient x Community Effects Two grass species, one native and one introduced showed nutrient • community interactions (Table II; Fig. 3). Both responded more to nutrient addition in the serpentine-only community than in the serpentine-plus-sandstone community. Aboveground community production showed the opposite interaction effect--community production responded to nutrient addition more in the serpentine-plus-sandstone community than in the serpentine community ( P - 0.0273). F. Nutrient x CO, Effects Litter production increased under elevated CO2 in the added-nutrient treatment but not in the ambient-nutrient treatment (P < 0.0001) (Fig. 1; Table II). G. Community x CO2 Effects Two species exhibited interaction effects. Under ambient-nutrient, ambient CO2 conditions, was nearly twice as dense in the serpentineonly treatment as in the serpentine-plus-sandstone treatment, but under elevated CO2 the two communities had similar density of 0.0127 for interaction effect) (Fig. 2; Table II). aboveground biomass in May showed a qualitatively similar pattern (P = 0.0926) (Fig. 3). H. Nutrient x Community x CO2 Effects No three-way interaction effects were observed. I. Community Responses and Invasion Although invasion of the serpentine-only microcosms by introduced grasses occurred, source population size did not explain a significant fraction of the variance in density or aboveground production of any of the three introduced grass species in the serpentine-only treatment (P > 0.3 for all). Invasion patterns reflected nutrient effects and interspecific differences among invaders (Figs. 2, 3). In the added-nutrient treatments, dominated aboveground biomass production in the serpentine-plussandstone communities. It also established in the serpentine-only communities with added nutrients, where it produced one-fourth to one-half as much aboveground biomass as the native grass was much less successful invading the ambient-nutrient communities, suggesting that soil
10.
151
fertility, not source population size, was the major determinant of invasion. The added-nutrient treatment had a similar but much smaller effect on invasion by abundance was very low under both nutrient, treatmentsmless than 1% of production. Among the serpentineplus-sandstone microcosms, 6 of 12 added-nutrient communities retained while 11 of 12 ambient-nutrient communities lost it. Among the serpentine-only microcosms, invaded one added-nutrient microcosm but no ambient-nutrient microcosms.
The microcosms were very similar to nearby field plots in terms of overall production, plant density, and sensitivity to nutrients and elevated CO2. Peak aboveground biomass in ambient-nutrient microcosms in May was generally within _+10% of the values from the field and never different by more than 15 g m -2. After 1 year of self-seeding, plant density in ambientnutrient microcosms converged on field values, with treatments ranging from 18% above to 5% below average field densities. As in previous field experiments (Hobbs 1988), moderate nutrient additions increased aboveground production more than fivefold and virtually eliminated forbs from the community. Overall, species that responded to the treatments sorted along functional axes (Fig. 4), strongly paralleling field patterns. Responses to elevated CO2 in the field and in ambient-nutrient microcosms were similar not only in magnitude but also in mechanism. Total aboveground production in the microcosms increased with elevated CO2, but only as a consequence of increased late-season growth in the late-annual guild. This pattern matches the results from field plots (unpublished data) and provides clear confirmation that the late-season CO2 response in the field is i n d e p e n d e n t of unique features of the soil depth distribution or bias in the placement of the field chambers. In the added-nutrient microcosms, we did not observe increased production with elevated CO2 (Figs. 1, 2) even though added-nutrient monocultures did respond to elevated CO2 (unpublished data). One explanation is that late annuals were virtually eliminated from the added-nutrient microcosms, so they lacked the functional group necessary for a community response. A second possibility is that production responded during the first year of the experiment, resulting in more litter under elevated CO2 (Fig.l; Table II), which may have limited the production response during Year 2. Of the two species that declined or disappeared from the microcosms, one is consistent with field observations, and the other (Calycadenia) is not. The poor establishment of and differential persistence of and are difficult to explain. Greater seed dormancy
Lasthenia
Summary of responses along functional axes to community (ss vs s), nutrient (f vs F), and CO2 (c vs C) treatments. The inequalities indicate the density a n d / o r production response of species within a functional dimension, e.g., grass species had higher production u n d e r a d d e d nutrients than ambient nutrients (F > f), whereas the converse was true of forbs (f > F). Details of the responses are given in Table II.
in (Gulmon, 1992) could account for low germination in Year 1 but the almost complete absence of seedlings in Year 2 argues against this explanation. Whatever the cause, failed to establish both in communities and in companion monoculture experiments. The similarities between the performance of ecosystems in the microcosms and in the field deemphasize the role of some potentially important controls on primary production. They provide strong evidence, for example, that soil structure on the spatial scale of centimeters and larger, developed over long time periods, does not play a dominant role in this ecosystem. This is consistent with the conclusion that gopher disturbance churns the field soils at approximately 5-year intervals (Hobbs and Hobbs, 1987). The field and microcosm communities also differ in plant species diversity, with the microcosms containing many but not all of the major species and many fewer species total. At least in the short term, field-level production does not depend on field-level biodiversity. This is consistent with the hypothesis that production saturates relatively quickly with plant biodiversity, once the major functional groups are represented (Vitousek and
153 Hooper, 1993). It does not, of course, rule out the possibility that the higher diversity might be important for coping with severe stress (Tilman and Downing, 1994) or that diversity at other trophic levels may be important in regulating plant production(Naeem 1994). Aboveground litter was substantially lower in the microcosms than in the field, probably reflecting a n u m b e r of factors. First, the microcosms were measured in their second year, eliminating possible contributions of litter older than 1 year. At Jasper Ridge, losses in grassland litter mass are approximately 35% in the first year and another 15% the second (J. des Rosiers, unpublished data), indicating an important contribution of older litter to the standing pool. Second, biomass in the field plots is up to 10% perennials, which are absent from the microcosms. Litter from perennials is coarser than litter from annuals, and it is more likely to be held off the ground, where it is less subject to attack by decomposers. The resulting aboveground litter pool in the field is roughly one to two times the aboveground production (Fig. 1). In the microcosms it was 10-20% of aboveground production. The microcosms provide excellent access to a range of questions involving ecosystem responses to modified resources or species composition. Although it is conceivable to design field experiments with a factorial combination of elevated CO2, altered soil nutrients a n d / o r water, and altered composition of the community of plants and potentially invading plants, the work and financial resources required to establish such an experiment would be difficult to justify for most ecosystems, especially in the context of the high spatial variability in the field. The free-air CO2 exchange experiments in Switzerland demonstrate the feasibility of one approach to establishing a large n u m b e r of treatments in the field (see A LOscher Chapter 19). Microcosms allow equally large or even larger numbers of treatments. Microcosms have the added disadvantage of chamber effects (Curtis 1989), but the added advantage of relatively easy access to the entire rooting volume. In the results discussed here, the microcosms provide confirmation of some patterns observed in the field and initial access to some patterns where the field provides no information. The central confirmation concerns the mode of increased growth in response to increased CO2. It is increasingly clear that the CO2 responses of individual species grown in pots may be poor predictors of the responses of complete communities, with the general pattern that production is less CO2-sensitive in communities than in individual plants (Ackerly and Bazzaz, 1995). So few data are available that it is not yet possible to assess whether this conclusion applies as well at the ecosystem scale. At Jasper Ridge, ecosystem production is sensitive to CO2, even though greenhouse studies on the individual species indicate small CO2 responses (Williams 1988) and microcosm experiments with
154 monoculmres reveal C02 responses of production only under increased nutrient availability (unpublished data). The response of production to increased CO2 in the ambient-nutrient microcosms is driven through responses of summer production in This increase in summer production under elevated CO2 does not appear to be dominated by direct CO2 effects. is, in fact, not strongly directly sensitive to CO2, based on biomass accumulation from germination until the point when ecosystem-scale, aboveground biomass reaches it peak (unpublished data). Although it is very difficult to prove that does not become CO2-sensitive late in its life, the simplest explanation is that responds to increased soil moisture and not directly to increased COz. Thus far, CO2 responses mediated through effects of decreased stomatal conductance on soil moisture have been observed only in water-limited grasslands and only at the ecosystem scale (Field 1995). In the results reported here, this appears as stimulated growth of late-season annual, after the date of maximum aboveground biomass. At Jasper Ridge, manifestations of CO2 effects on stomatal conductance appear in soil moisture (Fredeen 1996), microbial biomass (Hungate 1996a), soil nitrogen dynamics (Hungate 1996a), and ecosystem carbon balance (Hungate et al., 1996b). Ecosystem-scale experiments on tallgrass prairie yield a qualitatively similar pattern, with increased COs leading, in dry years only, to decreased water stress and increased duration of season for active plant growth (Owensby 1993). If the CO2 sensitivity in ambient-nutrient microcosms is driven largely through effects on the water balance, why are the late annuals the only plant functional group stimulated? It appears that the CO2 effects on soil moisture are largely limited to the end of the rainy season (Fredeen 1996), when the late annuals are the only plants still active. With the Mediterranean-type climate at Jasper Ridge, soil moisture increases with the onset of the rainy season in October or November and typically remains high until March, at which point the plant canopy is well-enough developed for differences in stomatal conductance to influence ecosystem evapotranspiration (Fredeen 1995) and soil moisture. Early and mid-season annuals fail to respond to the increased late-season soil moisture because their phenology, regulated largely by photoperiod (Mooney unpublished), makes them incapable of using additional resources available only late in the growing season. Late-season soil moisture increased under elevated COz in field plots on sandstone but the effect was much less clear near the surface on serpentine plots (Fredeen 1996). This contrast probably reflects the more important role of soil evaporation in the plots with incomplete canopy development. Using measurements to a greater depth, Field (unpublished)
10.
155
c o n f i r m e d that elevated C O 2 c a n lead to increased soil moisture in microcosms. At Jasper Ridge, the initial species change in response to increased CO2 is an increase in the a b u n d a n c e of late-season annuals. But moisture balance also plays a role in the relative d o m i n a n c e of different early-season annuals (Hobbs and Mooney, 1991) and in the invasibility of grassland by shrubs (Williams 1987). We speculate that over an e x t e n d e d period, species changes driven by increased CO2 may be m u c h m o r e extensive than changes in the functional groups of annuals. They may lead to major changes in ecosystem type, from grassland to shrubland or shrubland to forest, a hypothesis p r o p o s e d (Polley 1994), but also challenged (Archer 1995), as an explanation for the r e c e n t expansion of shrubland in arid regions of the southern U n i t e d States. In the microcosm e x p e r i m e n t s described here, elevated CO2 did not stimulate c o m m u n i t y invasion from s u r r o u n d i n g populations of Eurasian grasses d o m i n a n t on sandstone sites. Fertilization, however, did stimulate invasion. We interpret this as indicating that while early-season soil moisture may be correlated with invasion in the field (Hobbs and Mooney, 1991), stored or late-season soil moisture may not be effective in facilitating invasion by these early-maturing annuals. T h e d o m i n a n c e of Eurasian grasses on sandstone sites attests to their ability to thrive in the Jasper Ridge climate, given appropriate soil conditions. For some Eurasian grasses such as and appropriate conditions include serpentine soils in years with high rainfall and in areas where natural soil disturbance at least temporarily increases local n u t r i e n t availability. T h e inability of to persist in the serpentine microcosms with a m b i e n t n u t r i e n t is consistent with its absence from serpentine grassland on Jasper Ridge, even areas that have been successfully invaded by other Eurasian grasses.
The Jasper Ridge C O 2 experiment is supported by grants from the U.S. National Science Foundation to the Carnegie Institution of Washington, Stanford University, and the University of California at Berkeley. The U.S. Department of Energy contributed support in the project's first year. Thanks to Terry Chapin, Beth Holland, and Hal Mooney for major roles in planning the microcosm experiments. Many people contributed to the microcosm component in the Jasper Ridge CO2 experiment, including Sue Thayer, Barbara Mortimer, Julie des Rosiers, Howard Whitted, GeeskeJoel, Pep Canadell, Bruce Hungate, Julie Whitbeck,James Gorham, Anne Blanche Adams, Hailin Zhong, Heather Reynolds, and Barbara Lilley. This is Carnegie Institution of Washington, Department of Plant Biology Publication 1284.
Ackerly, D. D., and Bazzaz, F. A. (1995). Nonlinear responses of plants to incrementally rising atmospheric CO2. 1, 199-207.
156 Archer, S., Schimel, D. S., and Holland, E. A. (1995). Mechanisms of shrubland expansion: Land use, climate or CO2. 29, 91-99. Curtis, P. S., Drake, B. G., and Whigham, D. F. (1989). Nitrogen and carbon dynamics in C3 and C4 estuarine marsh plants grown under elevated CO2 in situ. 78, 297301. Field, C. B., Jackson, R. B., and Mooney, H. A. (1995). Stomatal responses to increased CO2: Implications from the plant to the global scale. 18, 1214-1225. Field, C. B., Chapin, F. S., III, Chiariello, N. R., Holland, E. A., and Mooney, H. A. (1996a). The Jasper Ridge CO2 experiment: Design and Motivation. "Carbon Dioxide and Terrestrial Ecosystems" (G. W. Koch and H. A. Mooney, eds.), pp. 121-145. Academic Press, San Diego. Fredeen, A. L., Koch, G. W., and Field, C. B. (1995). Effects of atmospheric COs enrichment on ecosystem CO2 exchange in a nutrient and water limited grassland. 22, 215-219. Fredeen, A. L., Randerson,J. T., Holbrook, N. M., and Field, C. B. (1996). Elevated atmospheric COs increases late-season water availability in a water-limited grassland ecosystem. Submitted for publication, Gulmon, S. L. (1992). Patterns of seed germination in California serpentine grassland species. 89, 27-41. Gulmon, S. L., Chiariello, N. R., Mooney, H. A., and Chu, C. C. (1983). Phenology and resource use in three co-occurring grassland annuals. 58, 33-42. Hickman, J. C. (1993). "The Jepson Manual--Higher Plants of California." Univ. of California Press, Berkeley. Hobbs, R.J., and Hobbs, V.J. (1987). Gophers and grassland: A model of vegetation response to patchy soil disturbance. 69, 141-146. Hobbs, R.J., and Mooney, H. A. (1991). Effects of rainfall variability and gopher disturbance on serpentine annual grassland dynamics. 72, 59-68. Hobbs, R. J., Gulmon, S. L., Hobbs, V. J., and Mooney, H. A. (1988). Effects of fertilizer addition and subsequent gopher disturbance on a serpentine annual grassland community. 75, 291-295. Hungate, B. A., Chapin, F. S., III, Zhong, H. L., Holland, E. A., and Field, C. B. (1996a). Stimulation of grassland nitrogen cycling under CO2 enrichment. in press, Hungate, B. A., Jackson, R. B., Chapin, F. S., III, and Field, C. B. (1996b). Carbon cycling in annual grasslands under carbon dioxide enrichment. Submitted for publication. Jackson, R. B., Sala, O. E., Field, C. B., and Mooney, H. A. (1994). COs alters water use carbon gain, and yield in a natural grassland. 98, 257--262. Jackson, R. B., Luo, Y., Cardon, Z. G., Sala, O. E., Field, C. B., and Mooney, H. A. (1995). Photosynthesis, growth, and density for the dominant species in a COw,enriched grassland. 22, 220-224. Naeem, S., Thompson, L.J., Lawler, S. P., Lawton,J. H., and Woodfin, R. M. (1994). Declining biodiversity can alter the performance of ecosystems. 368, 734-737. Owensby, C. E., Coyne, P. I., Ham, J. M., Auen, L. M., and Knapp, A. K. (1993). Biomass production in a tallgrass prairie ecosystem exposed to ambient and elevated COs. 3, 644-653. Polley, H. W., Johnson, H. B., and Mayeux, H. S. (1994). Increasing COs: Comparative responses of the C4 grass and grassland invader 75, 976-988. Tilman, D., and Downing, J. A. (1994). Biodiversity and stability in grasslands. 367, 363-365. Vitousek, P. M., and Hooper, D. U. (1993). Biological diversity and terrestrial ecosystem biogeochemistry. "Biodiversity and Ecosystem Function" (E.-D. Schulze and H. A. Mooney, eds.), pp. 3-14. Springer-Verlag, Berlin.
10.
157
Williams, K., Hobbs, R. J., and Mooney, H. A. (1987). Invasion of an annual grassland in Northern California by ssp. 72, 461-465. Williams, W. E., Garbutt, K., and Bazzaz, F. A. (1988). The response of plants to elevated CO2. V. Performance of an assemblage of serpentine grassland herbs. 28, 123-130.
This Page Intentionally Left Blank
11 Effects of Elevated C02 on Plant Species Dominance in a Highly Diverse Calcareous Grassland
Earth's ecosystems will face two serious problems in the future that may threaten their stability: change in global climate and loss of species diversity (e.g., Peters and Lovejoy, 1992). There has been considerable discussion of the role of plant diversity in maintaining ecosystem function. Some researchers believe that any loss of species will negatively affect ecosystems whereas others believe that there is a great deal of functional redundancy in species-rich ecosystems (Baskin, 1994). There have, however, been few tests of the role of species in maintaining ecosystem function (but see Ewel 1991; Naeem 1994a; Tilman and Downing, 1994). Much of the work on the effects of climate change on plants and ecosystems has focused on increased atmospheric CO2 concentration. The role of CO2 in altering community structure has been intensively studied in controlled environments. These experiments suggest that the CO2 response of a species grown individually is not a good predictor of its response in the presence of competition, but that initial relative growth rate and allocation patterns may be important predictors (Bazzaz, 1990). There have been, however, relatively few tests of the effects of elevated CO2 on ecosystem processes (Drake and Leadley, 1991; Owensby 1993; Oechel 1994; Sch~ippi and K6rner, 1996) or community composition (Curtis 1989; Owensby 1993; see also Chiariello and Field, Chapter 10) in intact ecosystems. 159
Copyright 9 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.
In a long-term experiment in which we manipulate plant species diversity and atmospheric CO2 concentration, we are attempting to determine the extent to which elevated CO2 induces changes in community structure and affects ecosystem function, and how changes in plant diversity affect ecosystem function. Calcareous grasslands were chosen for this study, because they are some of the most species-rich temperate ecosystems and they are disappearing due to altered land use and m a n a g e m e n t practices. We have taken a hierarchical approach to studying these communities and report here on the first-year responses of the whole community, functional groups, and individual species in the field. Schmid (Chapter 4) report on the response of genotypes of two of the species in this study. The field study was combined with a study of model grassland ecosystems in controlled environments, and here we report the response of individual species to elevated CO2 in these model ecosystems. We use these results as a basis for discussing difficulties in detecting and predicting the changes in community structure caused by CO2. We also discuss the usefulness and limitations of plant species manipulations as a means of gaining additional insight into the potential effects of changes in plant community structure on ecosystem function.
A. A Long-Term Field Study The field study site is an unfertilized, calcareous meadow 20 km south of Basel, Switzerland, in the foothills of the Jura Mountains near the village of Nenzlingen. The elevation is 550 m, the slope is ---20~ and the exposure is SW. The soils are classified as a transition Rendzina, with a well-developed, rock-free, loamy topsoil and a rapid transition at 10-15 cm to a rocky subsoil. The shallow depth of the topsoil, moderate summer precipitation, and southern exposure of this site make it prone to drought between June and August. This unfertilized meadow was used for nonintensive cattle grazing for as long as records are available, but since the beginning of 1993 it has been mown twice a year to maintain the high species diversity typical of these calcareous grasslands. A detailed comparison of grazed and mown calcareous grasslands in this region has shown that biomass and productivity are similar, plant species composition of abundant species does not differ substantially (e.g., is the dominant species), plant diversity is very high under both types of management, and plant diversity is slightly higher in grazed systems because some species are not found in mown meadows (e.g., (Zoller, 1954; M. Schl~ipfer, personal communication).
11.
161
on
The experimental design consists of (1) C O 2 t r e a t m e n t s on plots of "natural" calcareous meadow, i.e., undisturbed treatments, and (2) factorial combinations of CO2 treatments and differing plant species n u m b e r in artificially established plant communities at the same field site, i.e., diversity treatments. The experimental design and the n u m b e r of replicates per treatment are summarized in Table I. The n u m b e r of replicates in the undisturbed treatments (eight) is double that of diversity treatments (four) because of relatively high spatial heterogeneity in the undisturbed community compared to the diversity plots. Plots for the diversity treatments were prepared by removing all of the topsoil for all of the plots in a block, homogenizing the soil, returning the soil to the plots, and then covering the plots for 1 month with clear plastic in m i d s u m m e r 1993 to kill resprouting plants. A ring of pots installed around the border of the plots provided an opportunity to grow selected species in the absence of competition (see Schmid Chapter 4 for further details). Most of the species were raised from seed, starting in late spring 1993. and were vegetatively propagated from material collected at the field site. Plantlets were then planted into these plots in September 1993 using a hexagonal grid with a 3.85-cm spacing between plants. The species used in the three diversity treatments are listed in Table II. Selection of the species for the diversity treatments was determined by (1) knowledge of the floristic composition of calcareous grasslands in the region, (2) an extensive survey of the plant biomass and tiller n u m b e r at the study site in May 1993, and (3) the need to include species that were of special significance to individual research projects. For example, is the dominant plant species in this and other calcareous grasslands of the Jura and, therefore, was planted at high densities in all three diversity treatments. Four species,
CO 2 treatment
C--control (no chamber and no added CO2) Amambient (chamber, but no added CO~) E--elevated (chamber with added CO2)
Undisturbed (natural grassland community)
Diversity treatments (artificially established communities) Low Medium High 4
4
4
4
4
4
4
4
4
Diversity treatment Functional group
Low (5 species)
Medium (12 species)
High (31 species)
Graminoids
Legumes
Forbs (nonlegume)
and were planted at much higher densities than they naturally occur because a large n u m b e r of replicates were needed to study genotypic variation in these species. We also know from the May 1993 survey that the total n u m b e r of vascular plant species in the study area (1200 m 2) was --~110, the total n u m b e r found in 48 plots of 0.04 m 2 was 74, and the average n u m b e r in a 0.04 m 2 plot was 25, so the n u m b e r of species in the center of the high diversity plots (31 species in 0.55 m 2) is roughly the n u m b e r that would be found in a similar sized area in the undisturbed grassland (C. Hufschmid, personal communication).
11.
on
163
C O 2 exposure commenced at the end of March 1994. Hexagonal opentop chambers 1.2 m across by 1.2 m high were used to maintain CO2 concentrations of 600/xl liter -1. Their steel frames are covered with clear plastic foil (Melinex, ICI Inc.). A novel aspect of this chamber is that it is open at the bottom (a gap of 7 cm), which may have ameliorated some of the effects that open-top chambers have on climate and provided better turbulent mixing of CO2. We found that this turbulent mixing of CO2 works well, so in 1995 we reduced the height of the chambers to 0.5 m with little change in the control of CO2 concentrations spatially or temporally. This open-bottom design also allows some animals, especially snails and slugs, to move freely into and out of the chamber. Air with or without added CO2 enters the chamber through perforations in a ring of 10-cm-diameter pipe m o u n t e d at the base of the chamber. The rate of exchange of air in the chambers was ~-2 changes min -1. Carbon dioxide concentrations were monitored with a CO2 IRGA in air samples drawn continuously from the center of every elevated CO2 chamber. Measurements of CO2 concentrations at other positions indicated that CO2 concentrations were more variable and slightly lower near the edge of the chamber than at the center of the chamber. CO2 concentrations in the elevated treatments were maintained with one mass flow controller per chamber using custom developed electronics and software. Measurements of microclimate in the chambers indicated that air temperatures were ---2~ warmer on sunny days, relative humidity was proportionally lower, light intensity was 10% lower, and wind speeds were more constant than outside the chamber. In addition, because the upslope wall of the chamber intercepted rainfall, the soils in the upslope areas of the plots received less rainfall and were drier than the downslope areas of the plots. Water was applied several times during the season to compensate for the lack of rainfall in the upslope part of the plots. The problem of lack of rain in the upper part of the chambers was nearly eliminated in 1995, because we reduced the height of the chambers to 0.5 m. We clipped vegetation to 5 cm in all plots in June and October 1994. Data presented below are plant dry weights above 5 cm from the June 1994 mowing. We have focused primarily on data from the diversity treatments. Data from the diversity plots are from the central 0.55 m 2 a r e a of the plot only, because the tubes for competition-free growth surrounded this area (see above). Some species from the diversity plots do not appear in the analysis or were analyzed at the genus level, because (1) the data for these species appear elsewhere (see Schmid Chapter 4), (2) the plants were not tall enough to have significant biomass above 5 cm, or (3) the species were too difficult to separate in the dried samples of leaf parts (there are no data on a species basis for some of the grasses and and species where pooled in the high diversity treatment). Data from the undis-
turbed treatments are for the whole plot. A randomized complete block analysis was used for whole plot level measures of biomass. For the analysis of subplot factors, i.e., functional group and species biomass, the subplot response was tested against the subplot by main factor interaction term and subplot by main factor interactions were tested against the residual error. All analyses were done using ANOVAs in JMP (SAS Institute, Inc.).
B. A Companion Study of Grassland Communities in Controlled Environments In conjunction with the field study, we also studied the C O 2 response of model communities composed of six plant species used in the field study, and The plants were grown in rectangular, 20 liter containers in naturally lit, controlled environment chambers. The containers (26 cm wide • 36 cm long • 26 cm deep) were filled with 5 cm of field soil on top of a 50-50 mixture of field soil and calcareous gravel. Plants were planted on 11 and 12 Jan 1994 using a hexagonal grid with 3.75 cm between plants. Daytime CO2 concentrations for the three CO2 treatments were 330, 500, and 660/~1 liter -1 averaged over the duration of the experiment. Other environmental conditions were 18~ day/10~ night for the first 35 days and then 24~ 18~ for the remainder of the experiment, a 16-hr photoperiod, and ~ 4 0 0 700/~mol m -2 sec -1 light (artificial lights supplemented natural light on cloudy days). We report aboveground dry weights per species from a final harvest that took place between 16 and 20 May 1994. Statistical analysis was similar to that used for the species level analysis in the field experiment.
A. The Field Study Community Aboveground biomass in the undisturbed plots was 242 + 24 g m -2 (mean _+ SE, n = 8) at ambient CO2 and 256 + 11 g m -2 at elevated CO2, and this difference was not statistically significant. Figure 1 shows that the aboveground biomass increased in the diversity treatments at elevated CO2 ( P = 0.10, marginally significant) and with increasing n u m b e r of plant species (P = 0.001). The CO2 response was greatest in the high diversity treatment (27%) and this was the only diversity level in which the CO2 effect was significant (P = 0.04) based on linear contrasts. The increase in biomass due to diversity treatment was greater between the m e d i u m and high diversity (32%) than between the low and m e d i u m diversity (15%). There was substantially less biomass in the undisturbed plots (i.e., natural vegetation) than in the diversity plots (i.e., planted
11.
on
165
Figure I Response of artificially established calcareous grassland communities to factorial combinations of CO2 and diversity treatments. The CO2 treatments are A, ambient CO2, and E, elevated CO2 (control treatments are not shown for the sake of clarity). Diversity treatments are low (5 species), medium (12 species), and high (31 species). (A) Aboveground biomass from the June 1994 mowing (means _+ SE, n = 4). (B) Fraction of biomass by functional groupings of graminoids, nonleguminous forbs, and legumes.
c o m m u n i t i e s in d i s t u r b e d soil). Control plots (i.e., no c h a m b e r ) h a d 16% g r e a t e r biomass t h a n a m b i e n t plots (i.e., c h a m b e r , but no CO2 a d d e d ) in the u n d i s t u r b e d t r e a t m e n t , but this difference was n o t significant. Aboveg r o u n d biomass was also h i g h e r in control plots t h a n in a m b i e n t plots in the low (25%) a n d high (14%) diversity treatments, b u t was slightly lower in the m e d i u m ( - 3 % ) diversity t r e a t m e n t (there was a marginally significant increase across all diversity levels, P 0.09). We believe that the t r e n d toward h i g h e r biomass in the control plots was primarily d u e to the lack of rainfall in the u p s l o p e area of the c h a m b e r e d plots. Relatively little c h a n g e o c c u r r e d in c o m m u n i t y c o m p o s i t i o n in r e s p o n s e to CO2 at the level of f u n c t i o n a l g r o u p s (i.e.,
166 graminoids, legumes, and nonleguminous forbs) in the diversity treatments, except in the high diversity treatment where there was a significant increase in the fraction of graminoids at elevated CO2 (Fig. 1B, P = 0.02). This change was due to an increase in graminoid biomass at elevated CO2 that was accompanied by no change in forb or legume biomass. There were no statistically significant changes in community composition at the functional group level due to elevated CO2 in the undisturbed treatment (data not shown). Figure 2 shows the biomass response of individual species in the high diversity plots. In this diversity treatment, substantial differences were found in the biomass of species within each of the functional groups (P < 0.001 in all three functional groups), but species x CO2 interactions were not significant (P = 0.43 for grasses, P = 0.36 for forbs, and P = 0.96 for legumes). That is, species were different in size overall, but we have no evidence that, within a functional group, species differed in their response to elevated CO2. Similar responses were observed in the low and medium diversity plots. There is evidence from the undisturbed plots that species within a functional group did respond differentially. For example, most grass species did not respond to elevated CO2, but significantly increased its leaf length, percentage cover, and biomass (Ch. R6tzel, personal communication). Genotype level responses in the field study are discussed in Chapter 4 by Schmid
B. The Controlled Environment Study We also found substantial differences in species responses to elevated CO2 in the model grassland ecosystems grown in controlled environments. Figure 3 shows the aboveground biomass of the individual species. A highly significant species • CO2 interaction occurred in this study (P = 0.002). (P = 0.06) and (P = 0.01) responded positively to CO2, and had nonsignificant responses to COz, and responded very negatively to CO2 (P = 0.02).
A. Potential Consequences of Changes in Plant Community Composition at Elevated CO2 Will elevated CO2 alter plant community composition and, if so, will the functioning of ecosystems be altered by these changes in plant community composition? These are two of the most significant scaling problems that are faced by ecosystem researchers and modelers when trying to predict
11.
on
167
v~.~......_...----
Figure 2 Aboveground biomass (above 5 cm) of individual species in the high diversity treatment from the June 1994 mowing: (A) graminoids, (B) nonleguminous forbs, and (C) legumes. Lines at the bottom of panel B are 1, (solid line); 2, sp. (dotted line)" 3, (solid line), and 4, (dotted line). Biomass for and were combined, as was the biomass of and
E
~ o ,IQ
7s
, ~
50
"0 cO
L
25
o ,,o
Abovegroundbiomass (cut at ground level) of the individual species in the model ecosystem as a function of CO2 concentration. Lines are least squares, quadratic fits to whole container means of aboveground dry weight for each species (n = 3 per species per CO2 level). Arrows along the x-axis indicate the three CO2 treatment concentrations.
the long-term response of ecosystems to elevated C O 2. If community composition is relatively insensitive to changes in CO2, or if ecosystem function is relatively independent of changes in community composition, then there is an increased likelihood that ecosystem level models that make no species distinctions (e.g., Running and Nemani, 1991; Raich 1991) can successfully predict the response of ecosystems to elevated CO2. Additionally, short-term experiments with elevated CO2 may give results that are similar to experiments that are of sufficient duration to allow community composition and associated ecosystem processes to "stabilize" [by this definition no long-term experiment exists, because the longest running experiment with intact ecosystems is 7 years in a brackish marsh of the Chesapeake Bay (Curtis 1989; Drake and Leadley, 1991) ]. Certainly, caution must be exercised in interpreting our results, because both the field and the controlled environment experiments must be considered short-term experiments given that these calcareous grasslands are dominated by perennials. Relatively short-term laboratory experiments show large differences in the response of calcareous grassland species to elevated CO2. We found statistically significant interspecific variation in our model grassland ecosystem (Fig. 3). Part of this variation arose from the large negative response of to elevated CO2. Sanders (see Chapter 17) suggests that this negative response might be attributable to a shift in the mycorrhizal association from mutualistic to parasitic in this species. The species X CO2 interaction in the model ecosystem study did not hinge entirely on the negative response of since the species X CO2 interaction was marginally significant (P = 0.07) with excluded from the analysis.
11.
on
169
Two other controlled environment studies of individual species in pots by Ferris and Taylor (1993) and H u n t (1991) have also shown that there is substantial variation in response of calcareous grassland species to elevated CO2, although the statistical significance of this variation was not tested. Despite the large variation in species response in short-term, controlled environment studies, we found little interspecific variation in response to elevated CO2 in the field. There was no significant interspecific variation within functional groups in any of the diversity treatments and there was no significant variation among functional groups except in the high diversity treatment. The exceptions to this lack of variation may be important though, because they occurred in treatments with high plant species diversity. That is, was a large positive responder in a group of nonresponsive graminoids in the undisturbed treatment and graminoids as a group were "winners" in the high diversity treatment. Clearly, it is too early to say whether these trends are representative of the long-term response of this system. In a tallgrass prairie, Owensby (1993) found substantial year-to-year variation in CO2 response of individual species that appeared to d e p e n d on the severity of drought. In spite of this year-to-year variability, elevated CO2 has been shown to cause directional changes in community composition in both tallgrass prairie (Owensby 1993) and saltmarsh communities (Curtis 1989). Differences between the artificially established and undisturbed grassland communities are difficult to attribute to any single factor. Plant productivity was higher when plants were grown in disturbed soils, i.e., in the diversity treatments in the field and in the model ecosystem, which may have been due to increased soil fertility (although no fertilizer was added to any of the treatments). However, there are no clear patterns of CO2 response that correspond to disturbed versus undisturbed soils. There was no response of aboveground biomass or functional groups to elevated CO2 in the low diversity, medium diversity, or the undisturbed treatments (Fig. 1), but there were significant biomass responses to elevated CO2 in the high diversity treatment (Fig. 1) and model ecosystem (data not shown) that were accompanied by shifts in community composition (Fig. 1B and Fig. 3). Species n u m b e r does not appear to explain the differences either, since the model ecosystem (six species) and low diversity treatment (five species) had few species, but dissimilar responses. The two treatments with the highest species number, i.e., the undisturbed and high diversity treatment, also had dissimilar responses. Responses of individual species were similar when compared across treatments. For example, was a nonresponsive species in the undisturbed treatments (data not shown), the diversity treatments (T. Steinger,
170
personal communication), and in the model ecosystem (Fig. 3). and also responded similarly in the model ecosystems (Fig. 3) when compared to the diversity treatments (see Schmid Chapter 4). We have no direct evidence that the functioning of this grassland ecosystem has been altered by CO2-related changes in plant community composition. However, we do find that the aboveground biomass of this ecosystem and its response to elevated CO2 is dependent on community composition; that is, biomass increased with increasing diversity level and only the high diversity treatment had a large, positive response to elevated CO2 (Fig. 1). We also find that other measures of ecosystem function such as net ecosystem CO2 and water vapor flux respond similarly to the diversity treatments (R. Stocker, personal communication), which suggests that this ecosystem is potentially sensitive to CO2-induced changes in plant community composition. The degree to which changes in community composition could alter ecosystem function might, however, be limited by ecosystem level constraints, especially nutrient availability. But even these ecosystem level constraints might be profoundly affected by shifts in species dominance. For example, the %N in aboveground biomass of calcareous grassland species ranges from ---1-3% with legumes tending to have the highest %N and grasses the lowest %N (Morecroft 1994; and unpublished data from our study). Thus, a shift to graminoid species (e.g., Fig. 1B, high diversity) might increase the nutrient use efficiency of the plant community or decrease rates of decomposition. Changes in N cycling due to changes in species composition might be far more profound (see review by Hobbie, 1992) than typically assumed in ecosystem level models based on CO2induced changes in plant tissue composition within species. B. How Useful Are Functional Groups in CO2 Research? Functional groupings in C O 2 research may help predict (1) the response of individual species to elevated CO2 and (2) the response of ecosystems to CO2-induced changes in community composition. To fulfill the first role, it is necessary to identify groupings of species that explain a large fraction of the total variation in species-specific response to elevated CO2. To fulfill the second role, the response of functional groups to elevated CO2 must be predictable and ecosystem function must be regulated by community composition as defined by functional groups. In this second case, it is not essential that the response of individual species is predictable. Predicting the response of individual species to elevated CO2 is essential if we are to develop appropriate conservation strategies for protecting important species in a high CO2 world. However, no functional grouping yet examined explains a substantial fraction of the variation in speciesspecific response to elevated CO2. Hunt (1991) have suggested that
11.
on
171
Grime's (1977) growth strategy classifications are predictors of C O 2 r e s p o n s e . Their analysis shows that the correlation between growth strategy and CO2 responsiveness is statistically significant, but we feel that the low explanatory power of this relationship ( R 2 - 0.24) will limit its usefulness. Poorter (1993) suggested several other functional groupings; however, the best grouping, fast versus slow growing species, explained only 21% of the variance dry weight response to elevated CO2. We have used a coarse taxonomic grouping for functional classifications, which also corresponds in a general way to morphological groupings. Close examination of this functional grouping suggests that the relationship between taxonomic groups and CO2 response is weak. Within the forb classification, for example, responds negatively in the field (see Schmid Chapter 4) and in the model ecosystems (Fig. 3), shows no response in the field (see Schmid Chapter 4) or in model ecosystems (Fig. 3), and responded positively (Fig. 3). Physiologically based classifications, such as the Ca versus C4 dichotomy, do not appear to work much better and Poorter (1993) found that the differences in CO2 response between these groups are inconsistent and smaller than might be expected. The C3 versus C4 dichotomy explained CO2 responses of brackish marsh species (Curtis 1989) and rangeland species (see Polley Chapter 12), but it does not work in some artificial communities incontrolled environments (Bazzaz, 1990) or in tallgrass prairie (Owensby 1993). Our feeling is that CO2 responses are highly species-specific and that functional groupings may never explain more than a small fraction of this variation. This experiment addresses the two main problems associated with aggregating species into functional groups for the purpose of predicting changes in ecosystem function in response to changes in community structure. The response of functional groups to CO2 in the diversity experiment (Fig. 2) addresses the problem of whether the response of functional groups to elevated CO2 is predictable. One of the important a pr/or/predictions that we might have made is that legumes should "win" as a group at high CO2 (K6rner, 1993b; Overdieck, 1986; LOscher Chapter 19), but they did not in the field experiment (Fig. 1B). It may be that our experiment has not been running long enough to detect an increase in legumes, but our best guess is that legumes may "lose" at elevated CO2, because legumes did not respond positively, but graminoids did (Fig. 1B). The addition of species within a functional group in the diversity treatments (Table II) addresses the question of whether ecosystem function is affected by changing species composition within a functional group. Adding species within a functional group in the field study altered an important ecosystem property, aboveground biomass (Fig. la). In this case, we would have to say that it is species composition and not functional groups that define ecosystem
172
function. One might argue that we did not define our functional groups well, but this will be a problem in all studies of functional groups because there is considerable disagreement among ecologists over the definition of functional groups (cf. K6rner, 1993a). Based on this analysis, the outlook for developing generalities at the functional group level looks a bit bleak, but other research programs are actively addressing the relationships between functional groupings and elevated CO2 response (e.g., see Roy Chapter 9), and things may improve with more information.
C. How Should Responses to "Diversity" Manipulations Be Interpreted? This and other biodiversity studies should be interpreted with extreme caution because diversity effects (which are often strictly defined as changes in the n u m b e r and evenness of distribution of species) are usually confounded with species effects (defined as a dependence on the identity of the species irrespective of diversity level). The only way to avoid confounding species with diversity effects is to grow a given set of species in monospecific stands and in factorial combinations at each diversity level (e.g., factorial experiment in pots by Naeem 1994b), preferably in model ecosystems (e.g., see Roy Chapter 9) or natural ecosystems. We have not used a factorial design and neither have most other diversity studies (e.g., Naeem 1994a; Tilman and Downing, 1994; Ewel 1991), because of the large n u m b e r of replications involved in a full factorial design u n d e r realistic competitive conditions. Teasing out species effects in nonfactorial experiments will rest on the qualitative judgments involved in designing and interpreting them. We also believe that species effects are part of the larger question about how biodiversity affects ecosystem function. Specifically, if ecosystem function responds to changes in the n u m b e r and identity of species, as it did in our study and most other studies, then we can say that ecosystem function is sensitive to plant community composition, even if we cannot quantitatively attribute the proportions of change that were due to changes in species n u m b e r versus species identity. It certainly would be more accurate if nonfactorial studies were referred to as "community composition" rather than "diversity" experiments, but the "diversity" label is useful and accurate if interpreted broadly. There are also a n u m b e r of other problems associated with diversity experiments related to the methods that are used to establish stands of differing diversity. The first problem is one of initial conditions. It is difficult to start an experiment with plants of the same size (one might even ask if this is desirable). We had ---50% lower initial biomass per unit land area in low diversity compared with medium and high diversity because all five of the species in the low diversity were in the lowest range of weights per plant at planting. Naeem (1994a) planted seeds of all species at the same date in their experiment, but this approach has several important
11.
on
173
drawbacks: (1) differences in the timing of germination may lead to inequities in competitive ability, (2) seedlings do not behave like mature plants, and (3) the system will be rapidly expanding. The second problem with diversity studies is related to the difference in behavior of expanding versus nonexpanding systems. Most planted systems are expanding systems whereas many natural ecosystems are nonexpanding. This will have a tremendous impact on, for example, competitive interactions and ecosystem carbon sequestering (K6rner, 1996). A third problem arises when natural ecosystems are treated to alter their diversity. For example, Tilman and Downing (1994) used fertilizer treatments to alter diversity in grasslands (diversity decreased with increasing fertility) and then measured the resistance and resilience to drought stress. Responses to diversity in their study may have been confounded with interactions between fertilizer treatment and drought stress, because the species favored by high nutrients may also be the species that are the most sensitive to drought (Givinish, 1994). Tilman and Downing (1994) did not account for shifts in individual species in their study, but Tilman (1994) explained that they did eliminate some of the possible confounding effects by accounting for shifts in C~ versus C4 functional groups, such as changes in root:shoot ratios, in their statistical models. There are several ways in which these problems with diversity experiments can be diminished. In planted communities, it will help to wait for communities to reach "equilibrium" so that the problems associated with differing initial conditions and expanding systems are reduced. Most reports of diversity responses are from expanding systems (Naeem 1994a,b; Ewel 1991; and our study). There is also the potential of synthesizing the results from many diversity experiments to circumvent the species effect problem, because each research group used different species and, therefore, the combination of studies effectively makes a factorial design. For example, if we compare the results from five studies of diversity in grassland ecosystems, namely Tilman and Downing (1994), Naeem (1994a,b), Roy (see Chapter 9) and our study (Fig. 1), we find that measures of ecosystem function, resistance, or resilience were sensitive to the number of species in every study.
1. Elevated C O 2 did not detectably increase aboveground plant biomass in natural or artificial calcareous grassland communities in the first year of exposure. The exceptions are the high-diversity plots (on homogenized soil) in which aboveground biomass increased 27% in response to elevated CO2.
174
2. Variation exists in the response of individual species or groups of species to elevated CO2 in this and other experiments with calcareous grassland species, and this variation may lead to changes in species dominance. 3. A b o v e g r o u n d p l a n t b i o m a s s is a f f e c t e d by p l a n t c o m m u n i t y c o m p o s i t i o n . T h i s m a y b e d u e to c h a n g e s in p l a n t s p e c i e s n u m b e r , b u t m a y also b e d u e to o t h e r f a c t o r s s u c h as s p e c i e s i d e n t i t y a n d d i f f e r e n c e s in i n i t i a l c o n d i t i o n s . C o n s i d e r a b l e c a u t i o n m u s t b e e x e r c i s e d in i n t e r p r e t i n g t h e r e s u l t s f r o m this a n d o t h e r diversity s t u d i e s .
This research was funded by a grant from the Swiss National Science Foundation as part of the Swiss Priority Program on Environment (NF Grant SPPU 5001-035214). We gratefully thank Reto Stocker, Pascal Niklaus, Christina R6tzel, and many others for making the field research project possible. We also thank Jiirg St6cklin for his contribution to the model ecosystem study.
Baskin, Y. (1994). Ecologists dare to ask: How much does diversity matter? 264, 202-203. Bazzaz, F. A. (1990). The response of natural ecosystems to the rising global CO2 levels. 21, 167-196. Curtis, P. S., Drake, B. G., Leadley, P. W., Arp, W. J., and Whigham., D. F. (1989). Growth and senescence in plant communities exposed to elevated CO2 concentrations on an estuarine marsh. 78, 20-26. Drake, B. G., and P. W. Leadley. (1991 ). Canopy photosynthesis of C3 and C4 plant communities exposed to long-term elevated CO2 treatment. 14, 853-860. Ewel, J. J., Mazzarino, M.J., and Berish, C. W. (1991). Tropical soil fertility changes under monocultures and successional communities of different structure. 1, 289-302. Ferris, R., and Taylor, G. (1993). Contrasting effects of elevated CO2 on the root and shoot growth of four native herbs commonly found in chalk grassland. 125, 855-866. Givinish, T.J. (1994). Does diversity beget stability? 371, 113-114. Grime, J. P. (1977). Evidence for the existence of three primary growth strategies in plants and its relevance to ecological and evolutionary theory. 111, 1169-1194. Hobbie, S. E. (1992). Effects of plant species on nutrient cycling. 7, 336-339. Hunt, R., Hand, D. W., Hannah, M. A., and Neal, A. M. (1991). Response to CO2 enrichment in 27 herbaceous species. 5, 410-421. K6rner, Ch. (1993a). Scaling from species to vegetation: The usefulness of functional groups. "Biodiversity and Ecosystem Function" (E.-D. Sculze, and H. A. Mooney, eds.), pp. 117-140. Springer-Verlag, Berlin. K6rner, Ch. (1993b). CO2 fertilization: The great uncertainty in future vegetation development. "Vegetation Dynamics and Global Change" (A. M. Solomon, and H. H. Shugart, eds.), pp. 53-70. Chapman & Hall, New York.
11.
on
175
K6rner, Ch. (1996). The response of complex, multispecies systems to elevated CO2. "Global Change and Terrestrial Ecosystems" (B. H. Walker and W. L. Steffen, eds.), in press. Cambridge Univ. Press, Cambridge, UK. Morecroft, M. D., Sellers, E. K., and Lee, J. A. (1994). An experimental investigation into the effects of atmospheric nitrogen deposition on two semi-natural grasslands. J. 82, 475-483. Naeem, S., Thompson, L.J., Lawler, S. P., Lawton,J. H., and Woodfin, R. M. (1994a). Declining biodiversity can alter the performance of ecosystems. 368, 734-737. Naeem, S., Thompson, L. J., Lawler, S. P., Lawton, J. H., and Woodfin, R. M. (1994b). Biodiversity in model ecosystems (response to letter from Andr~ 1994). 565. Oechel, W. C., Cowles, S., Grulke, N., Hastings, S.J., Lawrence, B., Prudhomme, T., Riechers, G., Strain, B., Tissue, D., and Vourlitis, G. (1994). Transient nature of CO2 fertilization in arctic tundra. 371, 500-503. Overdieck, D. (1986). Long-term effects of an increased CO2 concentration on terrestrial plants in model ecosystems: Morphology and reproduction of L. and 30, 323-332. Owensby, C. E., Coyne, P. I., Ham, J. M., and Auen, L. M. (1993). Biomass production in a tallgrass prairie ecosystem exposed to ambient and elevated levels of CO2. 3, 644-653. Peters, R. L., and Lovejoy, T. E., eds. (1992). "Global Warming and Biological Diversity." Yale Univ. Press, New Haven, CT. Poorter, H. (1993). Interspecific variation in the growth response of plants to elevated ambient CO2 concentration. 104/105, 77-97. Raich, J. w., Rastetter, E. B., Melillo, J. M., Kicklighter, D. W., Steudler, P. A., Peterson, B.J., Grace, A. L., Moore, B., III, and V6r6smarty, C.J. (1991). Primary net productivity in South America: Application of a global model. 1, 399-429. Running, S. W., and Nemani, R. R. (1991). Regional hydrologic and carbon balance responses of forests resulting from climate change. 19, 349-368. Sch~ippi, B., and K6rner, Ch. (1996). Growth responses of an alpine grassland to elevated CO2. 105, 43-52. Tilman, D., and Downing, J. A. (1994). Biodiversity and stability in grasslands. 367, 363-365. Tilman, D., Downing, J. A., and Wedin, D. A. (1994). Does diversity beget stability? (Response to letter from Givinish, 1994.) 371, 114. Zoller, H. (1954). Die typen der Weisen des Schwiezer Juras. Beitr. Geobot. Landsaufn. Schweiz, Heft 33, Bern.
This Page Intentionally Left Blank
1 Are Some of the Recent Changes in Grassland Communities a Response to Rising CO2 Concentrations?
The abundance and density of C3 trees and shrubs on predominantly C4 grasslands in many parts of the world (Africa, Australia, North America, South America) have increased dramatically during approximately the last 125 years (reviews by Mayeux 1991; Archer, 1994). This rapid change in vegetation is evident from historical accounts and photographic records (Hastings and Turner, 1965) and repeated aerial photography (Archer 1988; Knight 1994) and censuses of permanent plots or similar areas (Glendening, 1952; Buffington and Herbel, 1965; Hennessy 1983), analyses of the stable carbon isotope composition of soil organic matter (Tieszen and Archer, 1990; Steuter 1990; McPherson 1993), and palynological evidence (Davis and Turner, 1986). Some woody species (e.g., mesquite) apparently expanded within their historical ranges from refugia, like drainages or rocky outcrops that had been occupied for centuries, or increased in stature and visibility from suppressed populations (Bogusch, 1952; Johnston, 1963). For other species (e.g., creosotebush), recent changes are an acceleration of an increase in density and range extension initiated centuries before (Hunziker 1977; Johnson and Mayeux, 1992). Production of grasses and other herbaceous species in many of these systems declines following woody ingress (Glendening, 1952; Heitschmidt and Dowhower, 1991; 177
178
Hobbs and Mooney, 1986), thus reducing the value of grasslands and savannas for livestock grazing and other uses. The shift in growth or life form composition also alters ecosystem-level processes, including surfaceatmosphere transfers of matter and energy, that potentially influence rates and patterns of carbon sequestration (McPherson 1993; H. B.Johnson, unpublished data) and local, regional, or even global climates (Schlesinger 1990). By reducing grass production, for example, woody plants may lessen the frequency and intensity of fires and the accompanying return of carbon and nitrogen to the atmosphere (Medina, 1982; Hobbs 1991). By increasing spatial heterogeneity of soil water and nitrogen (Schlesinger 1990) or in other ways altering the hydrology of grasslands (Joffre and Rambal, 1993), woody ingress can change surface albedo, evapotranspiration, runoff, and fluxes of trace gases to the atmosphere. Causes of the increase in abundance of woody plants have been extensively discussed because of the obvious importance of the change to economics and to longer term environmental concerns (e.g., Hastings and Turner, 1965; Grover and Musick, 1990; Bahre and Shelton, 1993; Archer, 1994). There is, however, relatively little experimental evidence documenting the cause or causes of woody invasion. Much of the evidence implicating specific causal factors in vegetation change is based on anecdotal accounts and is correlative and descriptive in nature. Expansion of woody plants on grasslands has traditionally been attributed to rather site-specific changes, most of which increased recruitment or reduced the often size-dependent mortality of woody seedlings. These explanations include suppression of fire which often selectively kills woody seedlings (Bragg and Hulbert, 1976; Wright 1976), changes in the populations of rodents (Brown and Heske, 1990) which may both consume and disperse shrub seed and kill woody seedlings (Glendening and Paulsen, 1955; Gibbens 1992), changes in temperature or precipitation (Hastings and Turner, 1965; Neilson, 1986), removal of browsers (Belsky, 1984), and the effects of livestock introduced during the last 125 years (Bahre and Shelton, 1993; Archer 1994). Effects of livestock include the ingestion and dissemination of seed of some woody species (Glendening and Paulsen, 1955; Brown and Archer, 1987; Archer, 1989) and overgrazing that reduced grass production and the frequency of fires, reduced interference from grasses for light, water, and other resources (Bush and Van Auken, 1990), and may have increased spatial heterogeneity of soil nitrogen and water (Schlesinger 1990). Each of these explanations for vegetation change is demonstrably important to some areas or for some species, but none of the traditionally advanced explanations for shrub encroachment satisfactorily explains the success of the diverse group of woody invaders across the diverse climatic and physiographic regions involved worldwide. Included are woody species
12.
179
(e.g., spp., spp., spp.) which differ greatly in potential growth rates, life histories, and tolerances to fire, drought, and shading, and grasslands on which precipitation differs markedly in seasonality, intensity, and predictability. Frequent fire alone probably was not sufficient to prevent invasion by for example, for it may not kill plants beyond the seedling stage (Wright 1976), but fire kills even large individuals of some species and effectively limits or prevents invasion by these plants (Burkhardt and Tisdale, 1976). Conversely, seed dispersal by livestock greatly increased the rate and extent of grassland invasion by (Brown and Archer, 1987; Archer, 1989) and perhaps other hard-seeded species, but cattle do not consume seed of or and thus cannot have directly influenced dissemination of these plants. Traditionally cited explanations for woody ingress sometimes also fail to explain the changes observed even in individual species. York and DickPeddie (1969), for example, concluded that heavy grazing increased the cover of and other woody species in desert grassland in New Mexico, but increases rapidly regardless of grazing regime (Glendening, 1952; Smith and Schmutz, 1975; Hennessy 1983). It is, of course, difficult to completely exclude prior effects of livestock and other factors in most studies, but similar increases in abundance have been noted in other woody species when grazing was absent (Rice and Westoby, 1978; Williams 1987). Widespread encroachment of woody species on grasslands began shortly after atmospheric CO2 concentration rose above its preindustrial level of 270-280 ppm. The global nature of the increase in CO2, near synchrony of changes in CO2 and vegetation, and multiple benefits of higher CO2 to C~ plant growth suggest that the historical rise in CO2 concentration contributed to woody ingress. Higher CO2 concentration, for example, decreases photorespiration (Sharkey, 1988; Johnson 1993a) and increases the quantum yield (Ehleringer and Bj6rkman, 1977; Long and Drake, 1991) and temperature optimum of C~ photosynthesis (Long, 1991). Rising CO2 increases rates or amounts of N 2 fixation by woody legumesymbioses (Norby, 1987; Thomas 1991; Polley 1994) and often increases the nitrogen use efficiency and growth of woody (Norby 1986; Polley 1994) and other C~ plants (Polley 1995a). Further, tolerance of C~ plants to heat, drought, salinity, and other stresses is improved by higher CO2 concentration (Idso, 1989). These and other changes may contribute to shift the competitive balance in favor of taller plants (Reekie and Bazzaz, 1989), possibly including C~ shrubs over C3 grasses, and of taller C3 over co-occurring C4 plants at elevated CO2 (e.g., Bazzaz and Carlson, 1984; Arp 1993). The relative and, sometimes, absolute effect of a per-unit increase in CO2 concentration on C~ physiology
(Johnson 1993a; Polley 1993b) and growth (Polley 1992; Polley 1994) may be greater over the subambient levels indicative of the past than at future levels, suggesting that the historical rise in CO2 concentration, albeit small (70-80 ppm), also generally favored greater production, taller vegetation, and taller C~ over Ca competitors (Fig. 1). Clearly, there are instances in which competitive success at higher CO2 concentrations is explained by factors other than or in addition to photosynthetic pathway (Bazzaz 1989). Results from CO2-enrichment studies on C4-dominated tallgrass prairie in Kansas, however, generally support the suggestion that taller Ca plants may increase on grassland as CO2 concentration rises. Owensby (1993), for example, reported that basal cover of relatively tall and deep-roofing Ca forbs on the prairie increased following a doubling of the current CO2 concentration. Frequency (Nie 1992) and basal cover (Owensby 1993) of the shorter Ca grass declined in chambered plots exposed both to the current and elevated CO2 concentration, though both frequency and basal cover declined less rapidly in plots fumigated with high CO2 (frequency declined less rapidly at high CO2 only when soils were well watered). Increased widths of annular rings of some trees (LaMarche 1984; Hari and Arovaara, 1988) and greater C accumulation in temperate forests
400
o
0
I
_
Abovegroundbiomass of C3and C 4 species that developed from the seed bank of a Texas savanna soil exposed for 13 weeks to a gradient in daytime CO2 concentration from near 350 to 150 ppm. Figure reprinted fromJohnson (1993a). Reprinted with permission of Kluwer Academic Publishers.
12.
(Kauppi 1992) in recent decades, suggest that the historical increase in CO2 has stimulated growth of at least some woody vegetation. Woody plants also usually grow more rapidly when exposed to atmospheres enriched in CO2, even when soil nutrients or water is limiting (Tolley and Strain, 1985; Norby 1986; Conroy 1990; Norby and O'Neill, 1991; Conroy 1992). Further, initial benefits of elevated CO2 to growth of woody plants may be multiplied over time when environmental conditions are favorable (Idso 1991). Resolving the role of rising atmospheric CO2 concentration in the changing structure of grasslands and, more generally, in vegetation change is not straightforward. The dynamics of any species is the resultant of complex interactions among a number of factors, including its life history, physiology, and morphology, that vary in relative importance with the biotic and abiotic environments. Despite the difficulty and perhaps futility of attempting to assign a relative weight to livestock, fire suppression, climate change, rising CO2, and other factors that contributed to vegetation change on grasslands and savannas, it remains important from a management perspective to identify when each of the potential contributors to change may be most influential. Here we briefly discuss what we perceive as the major obstacles to progress in identifying the relative contribution or importance of the historical increase in atmospheric CO2 concentration to vegetation change on grasslands and savannas. Drawing largely from the vast and rapidly expanding literature documenting effects of elevated CO2 levels on plants and ecosystems, we suggest mechanisms by which and conditions under which rising CO2 may have contributed significantly to the observed increase in abundance of woody plants in many parts of the world.
Perhaps the greatest difficulty in experimentally defining how rising C O 2 influenced vegetation on grasslands results from inequities between the processes typically measured when CO2 is varied and those most relevant to questions of vegetation change. Conjecture that CO2 has contributed to a shift from mostly C4 grasses to woody vegetation or from generally shorter to taller plants is based largely on physiological data and on observations that growth generally responds more positively to CO2 in C~ than C4 plants (Poorter, 1993). The question of woody ingress, however, inherently is one of population biology or community ecology. Relevant processes at these levels include the life histories and morphologies, as well as physiologies, of species as influenced by biotic interactions and the abiotic environment. Most of the evidence that rising CO2 concentration shifts the competitive balance in favor of C~ over co-occurring C4 species comes from experiments
with herbaceous plants in which competing species were planted nearly simultaneously and were of similar growth form or potential height and rooting depth (Bazzaz and Carlson, 1984; Arp 1993; Johnson 1993a). Competitive success in most of these experiments therefore was highly correlated with the response of growth rate or biomass production to CO2. These data yield information on the likely responses of vegetation in highly disturbed habitats to CO2, but have limited relevance to questions of change in established vegetation. Very little of the CO2 work explicitly addressed the quesdon of vegetation change in perennial systems in which invasive seedlings emerge in mature vegetation, a situation where the competitive success of seedlings will depend on their tolerance to resource depletion by neighboring adults (Goldberg, 1990), or competing plants differ in growth form. Of practical necessity, most CO2 studies have been of limited duration and, with few exceptions (e.g., Bazzaz 1992; Morse and Bazzaz, 1994), have not considered possible feedbacks between growth responses of plants to CO2 and the timing or magnitude of reproduction or other demographic parameters. Can known physiological and whole-plant responses to rising CO2 concentration be applied to predict the dynamics of grassland communities and responses of the diverse group of species involved? Linkages between physiology and populations or communities are inherently difficult to establish because relevant processes often differ among levels. When directed or guided by observations at community and population levels, however, physiological studies often may provide insight into those processes or periods during a plant's life cycle that are most critical to vegetation change. Vitousek (1993) similarly argued that recent success in applying physiological data to ecosystem-scale questions has resulted when physiological explanations were constrained by observations at higher levels. Progress in identifying the contribution of rising CO2 to the observed shift in grassland vegetation is further limited by the diverse biologies and life histories of woody invaders and by the wide range of environmental conditions that characterize invaded grasslands. When single species are considered or the spatial and temporal scales of observation are reduced to levels typical for most experimental evaluations, variation in the composition of vegetation increases and biotic and edaphic characteristics, local disturbances, past history, and annual weather variations strongly influence vegetation dynamics (Prentice, 1986). Not surprisingly, attempts to identify causal factors for vegetation change at these scales, even for single species, often yield contradictory results. General trends therefore are best sought at higher spatial and longer temporal scales. A. Interactions with Water Availability
It may prove useful to first approach the question of woody encroachment on grasslands at higher, perhaps regional, scales by identifying possible
12.
183
interactions of CO2 with mechanisms known to control the potential balance of trees or shrubs and grasses. At regional scales, climatic variables related to water availability and water balance appear most influential in controlling the potential distributions of vegetation types (Stephenson, 1990; Belsky, 1995) and their productivities (Sala 1988). Stephenson (1990), for example, demonstrated that actual evapotranspiration decreases and water deficit (evaporative d e m a n d not met by available water) increases along regional gradients in southern North America dominated first by forests, then by grasslands, and finally by shrublands (Fig. 2). Shrubs dominate the driest areas along the gradient either because they tolerate extreme soil and atmospheric drought or access deeply placed soil water not generally available to grasses. Because seedlings generally are more susceptible to drought than established individuals, however, woody establishment in these relatively arid environments is strongly correlated with rainfall and often occurs in pulses following large precipitation events (Turner, 1990). As water availability increases along this moisture gradient, production of grasses also increases. Woody seedlings can encounter severe interference from grasses for water and other resources in upper soil layers (Harrington, 1991). In some grasslands and savannas, therefore, the relative abundance of woody plants increases with increasing precipitation (Williams and Hobbs, 1989; Medina and Silva, 1990) and with soil or biotic factors that reduce water loss from surface soils and encourage accumulation in deeper layers (Knoop and Walker, 1985). In still more mesic areas, woody plants dominate at equilibrium apparently because they are superior competitors to shorter grasses for light (e.g., Tilman, 1988; Smith and Huston, 1989).
Mean annual evapotranspiration and annual deficit for major plant formations in North America (reproduced from Stephenson, 1990, with permission from the Universityof Chicago). Deficit is evaporative demand not met by available water. The diagonal connecting actual evapotranspiration of 1500 mm and deficit of 1500 mm depicts correlated changes in water balance and vegetation across an east-to-westtransect of increasing aridityin the southern United States.
A t m o s p h e r i c C O 2 directly affects the c o u p l i n g b e t w e e n climatic w a t e r b a l a n c e a n d v e g e t a t i o n by altering the efficiency with which plants use water, p r o d u c t i o n p e r u n i t o f t r a n s p i r a t i o n (Fig. 3). At the leaf level, water use efficiency (WUE) o f t e n increases by a b o u t the same relative a m o u n t as d o e s CO2 c o n c e n t r a t i o n in b o t h C3 a n d C4 species (Morison, 1993; Polley 1993a). S o m e o f this p o t e n t i a l increase in W U E may n o t have b e e n realized by plants in n a t u r e if h i g h e r W U E was d e r i v e d largely f r o m lower transpiration. T h e decline in t r a n s p i r a t i o n w o u l d have b e e n partially offset by an a c c o m p a n y i n g increase in leaf t e m p e r a t u r e a n d a d e c r e a s e in atmospheric humidity. For C~ plants, however, m u c h o f the CO2-mediated increase in W U E d u r i n g the last 200 years a p p a r e n t l y was realized as h i g h e r p h o t o s y n t h e s i s a n d g r e a t e r biomass p r o d u c t i o n (Polley 1993b; Polley 1994). Because o f the s t r o n g c o r r e l a t i o n that exists b e t w e e n water availability a n d p r i m a r y productivity in grasslands a n d deserts ( W e b b
(I)
~_
o zi5 W c:~ ~ v
oo
a. < I
I
I
CO 2 (ppm) Apparent water use efficiency, biomass produced divided by water lost to evapotranspiradon, of two cultivars of (wheat; Seri M82 and Yaqui 54) and individuals of the woody legume (honey mesquite). The two species were grown in separate experiments across daytime gradients in CO2 concentration from near 350 to 200 ppm. Wheat was well-watered (n = 20) or droughted by withholding additional water for the last 50 days of the 100-day experiment (n = 20). droughted by adding water only after soil moisture had declined to 65% of volumetric content at field capacity for 9 of the 14 months of the experiment (n = 6). The line is a linear regression of water use efficiency on CO2. Data are replotted from Polley (1994) and (1995a).
12.
05
1983; Sala 1988), the productivity of many grasslands should have increased substantially as atmospheric CO2 rose. To the extent that rising CO2 increases WUE and plant productivity, it also will increase potential leaf area and the potential for plant competition for light (Tilman, 1988; Smith and Huston, 1989). These changes, in turn, should favor taller growth forms like trees and shrubs at the expense of grasses, at least in relatively mesic grasslands. By decreasing the amount of water required to sustain growth, the CO2mediated increase in plant WUE may also have increased survival of woody seedlings in more arid environments. Analyses of the carbon (C) isotope composition ofaridland plants have shown that low values of leaf discrimination against 1~C (A), indicative of higher water use efficiency, are associated with greater plant longevity (Ehleringer and Cooper, 1988; Smedley 1991) and with greater survival during drought (Ehleringer, 1993). Projections based on changes in plant WUE also are compatible with observed effects of CO2 on the C3/C4 composition of annual communities grown at different soil moisture levels (Bazzaz and Carlson, 1984). Characteristics like osmotic adjustment and the roots: leaves ratio that impact plant water balance and affect niche overlap among competing species may also have been altered as CO2 rose (Miao 1992). Rising CO2 may also indirectly favor woody encroachment by decreasing transpiration rates of existing vegetation and thereby increasing soil water availability in grasslands and savannas (Polley 1995b). Shallow-roofing grasses are superior competitors to woody plants for water from upper soil depths. The abundance of deep-roofing trees and shrubs on some grasslands therefore depends on the amount of water that reaches soil below most grass roots (Knoop and Walker, 1985). When higher WUE derives largely from a decrease in transpiration per unit leaf area, as it often does in C4 grasses, rising CO2 will reduce the rate of soil water depletion provided the decline in transpiration per unit leaf is not offset by an increase in leaf area or leaf temperature. This change may in turn increase soil water content and favor greater percolation to depths where woody roots predominate. Higher leaf temperatures (Kirkham 1991) and, in years with below normal precipitation, greater leaf area (Owensby 1993) partly offset water savings expected on C4-dominated prairie from a decrease in conductance and transpiration per unit leaf area at elevated CO2 concentration. Still, Kirkham (1991) found that as a result of a 7-15% decline in evapotranspiration, soil moisture levels in prairie to a depth of 2 m were consistently higher at elevated than at the current CO2 concentration, even during periods of relatively severe drought (Owensby 1993). B. Effects on Seedling Establishment and Growth
Population-level data for woody invaders and larger-scale studies of rates and patterns of woody ingress or expansion can be used to identify crucial
] 86 periods or processes during the life cycles of woody plants that may be most sensitive to CO2 or to other factors. A seemingly trivial but important assumption implicit in attempts to identify factors that caused vegetation change is that woody populations were near equilibrium with grasses or were declining relative to grasses near the beginning of the 19th century. If woody populations were instead expanding, historical changes may simply have accelerated a vegetation change already predisposed under existing climatic conditions. Definitive data to judge the stability of woody/ grass vegetation prior to the industrial revolution obviously are few and it must be recognized that a true equilibrium in vegetation composition is never attained (Johnson and Mayeux, 1992; Brown and Gersmehl, 1985). The assumption that woody populations were in equilibrium near the beginning of the 19th century appears reasonable for some species, but only approximately true for others. Archer (1989), for example, used a model of woody cluster development and age-size relationships to predict that most plants in relatively mesic upland grassland/savanna in south Texas appeared since the late 1800s. Model results are consistent with other evaluations suggesting that the woody legume has only recently moved in great numbers onto grassland from refugia occupied perhaps for centuries (Bogusch, 1952). Abundance of creosotebush, like mesquite, has increased dramatically in former desert grassland during approximately the last 125 years (Buffington and Herbel, 1965). The recent increase in creosotebush, however, may be an acceleration of a process that proceeded for millenia. Creosotebush first appeared in quantity in the macrofossil record from the southwestern United States about 11,000 years ago. Johnson and Mayeux (1992) calculated that a subsequent expansion rate of about 16,000 h a / year was necessary to account for the distribution of creosotebush in the Chihuahuan, Sonoran, and Mojave Deserts. Johnson and Mayeux (1992) also discuss macrofossil evidence that blackbrush began to expand on these desert grasslands prior to European settlement. Similarly, Davis and Turner (1986) concluded from palynological data that expansion in Arizona began at least 2000 years before livestock were introduced. It appears then that changes during the historical period likely were not necessary to cause the ingress or expansion of some woody species. Changes apparently were, however, necessary to cause woody ingress or expansion onto grassland at the accelerated rates observed recently. Given the relatively long life spans of many of the invasive species (Archer, 1994), the apparently limited changes in abundance of some woody species in the decades to centuries preceeding widespread and concentrated human intervention on grasslands and savannas implies that prior to the industrial revolution, seedling recruitment and subsequent survival to reproductive maturity did not greatly exceed that required to replace existing woody plants. Models, for example, suggest that factors such as fire slow,
12.
187
but do not prevent, the eventual dominance over grasses by some woody species provided that over their life span established trees contribute more than one reproductive individual to the next generation (Hochberg 1994). Seed production of mature individuals or seedling establishment and survival must have been very low for those woody invaders that were not expanding greatly prior to the 19th century. By enhancing seedling establishment and survival a n d / o r seed dispersal, recent changes likely also increased the rates at which previously expanding woody populations grew. Given the diverse biologies and life histories of the woody plants involved, including species that differ greatly in growth rates and susceptibility to fire and other agents of mortality, the implication is that some influential mechanism or more likely a combination of mechanisms prevented or limited seedling establishment and survival of woody plants on grasslands. Not surprisingly then, most of the factors that traditionally are believed to cause or accelerate woody expansion or ingress either reduced the often size-dependent mortality of woody seedlings or enhanced woody establishment and growth by relieving water stress or by otherwise reducing interference from resident grasses or increasing the availability of soil resources. Grassland fires (Wright 1976) and herbivory by small mammals (Gibbens 1992), for instance, selectively kill small seedlings of the woody legume, Suppression of fire and reductions in small mammal populations during the last two centuries thus likely contributed to a marked increase in seedling survival of this species. Conversely, establishment rates of mesquite can be increased by grazing or mowing (Bush and Van Auken, 1990) or after severe drought in some grasslands (Hennessy 1983), apparently because these changes reduce interference from taller grasses for light or other resources. Rising CO2 may similarly have enhanced woody recruitment and reduced seedling mortality by reducing the amount of resources that plants required in order to grow until largely uncoupled from interference with neighboring grasses and the primary causes of mortality. By increasing plant water and nitrogen use efficiencies and, in woody legumes, rates and amount of Nz-fixation, for example, rising CO2 can reduce the amount of water and nitrogen that woody plants required from the soil. Increasing CO2 may also greatly increase growth rates of individually grown or spaced plants of woody invaders and other C~ plants (Polley 1994; Johnson 1993b), changes that should facilitate woody ingress following disturbance by reducing the period during which woody seedlings are most susceptible to mortality and to interference from neighboring grasses. Some woody seedlings minimize interference for water with neighboring grasses by rapidly extending their roots below those of most grasses (Brown and Archer, 1990; Bragg 1993), a process that would have been abetted by an increase in growth rate as CO2 rose. There are situations, however, where effects of CO2 on woody plants likely
were minimal. Seedling tolerance of shade potentially was improved by an increase in the quantum yield (Ehleringer and Bj6rkman, 1977; Long and Drake, 1991) or decrease in the light compensation point of photosynthesis as CO2 rose (Hand 1993; Polley 1993b), but existing information suggests these changes had little influence on the growth of shaded plants (K6rner and Arnone, 1992; see Gloser, Chapter 21). Limited data also suggest that rising CO2 concentration sometimes is of little benefit to the growth of C~ shrubs confined to the same roofing volume as grasses (Polley 1994).
Available evidence suggests that effects of the historical increase in CO2 likely were greatest in grasslands or for woody species in which water and, to a lesser extent, nitrogen limited woody establishment and early growth. Effects of CO2 also may have been pronounced in situations where rising CO2 significantly increased growth rates of woody seedlings. It is likely, therefore, that rising CO2 interacted positively with disturbances like grazing and prolonged drought that reduced the cover and growth of grasses and their ability to compete with woody seedlings. Woody plants that grew more quickly were susceptible for a shorter time to mortality from fire and herbivory and more rapidly attained reproductive maturity. Even a small reduction in the age of first reproduction can greatly increase the reproductive output of an individual (Cole, 1954). Woody plants that grew faster also more quickly became large enough to significantly impact ecosystem dynamics in ways that can feed back to enhance woody ingress. Large woody plants, for example, shade neighboring grasses or shelter small mammals that feed on herbaceous plants (Hobbs and Mooney, 1986) and, in some systems, thereby reduce grass production and the frequency or local intensity of fire. Large shrubs, by concentrating and cycling nutrients and changing hydrological properties beneath their canopies, may increase spatial heterogeneity of water and nitrogen (Schlesinger 1990). These changes may feed back to reduce the production and establishment of grasses and enhance shrub regeneration. Large trees or shrubs also serve as foci for bird-dispersed seed of other woody species (Archer 1988). Rising CO2 likely was less influential in vegetation change on grasslands where the growth of woody seedlings was strongly limited by light because of the highly size-asymmetrical nature of competition for this resource (Weiner, 1990; Bazzaz and McConnaughay, 1992). Although it appears reasonable that a change in CO2 would directly and indirectly affect the productivity and composition of grasslands and savannas, it is not obvious on what time scale this should occur. Atmospheric
12.
189
slowly during the 19th and early 20th centuries (Friedli 1986), and thus probably played little role in the initial phase of woody expansion (Archer, 1994). Further, there can be a lag of decades to centuries between shifts in atmospheric conditions and vegetation change (Davis, 1986) that probably would have limited how quickly effects of the recent rise in CO2 became evident. The typical lag between cause and effect often results from the limited rates at which propagules disperse and the inertial resistance of established vegetation to changemthe ability of perennial vegetation that established during favorable periods to persist under unfavorable conditions for decades without reproducing. The limit to change imposed by vegetative inertia may partly have been offset because of the highly asymmetric nature of competition between grasses and trees or shrubs that greatly favors the taller growth form when woody plants overtop grasses. Similarly, limitations of propagule dispersal were partly relieved for those species with seed that are disseminated by livestock. Still, effects of the near 30% increase in CO2 during the last two centuries likely are not fully evident. Vegetation almost certainly is not in equilibrium with atmospheric CO2, which presently is changing at an accelerated rate. CO 2 rose
The widespread and rapid increase in abundance and density of a diverse group of woody species on grasslands during the last two centuries likely is the result of several, often interrelated and reinforcing changes (Smeins, 1983; Belsky, 1990) including fire suppression, overgrazing, changes in rodent and browser populations, climate change, and the 30% increase in atmospheric CO2 concentration. Rising CO2 concentration, by increasing plant water use efficiency, increased potential production of vegetation in frequently water-limited grasslands and may, by reducing transpiration, have increased soil water availability and percolation to depths where woody roots predominate. In the absence of severe and repeated disturbance, these changes should in time have favored taller plants, such as trees or shrubs, at the expense of shorter grasses, consistent with patterns observed along natural gradients of precipitation or soil resource availability (Tilman, 1988). Given the lag of decades to centuries that often occurs between changes in atmospheric driving variables like CO2 concentration and shifts in vegetation, it is highly unlikely that the potential response of vegetation to the historical increase in CO2 concentration has been fully realized or that rising CO2 alone was sufficient to account for the rates and patterns of woody ingress observed in some areas during approximately the last 125 years.
190 Effects of rising CO2 concentration that were initiated during the past century or more should become increasingly evident with time and as CO2 continues to increase during the foreseeable future. Because so little of the CO2 work to date has addressed questions of vegetation change in perennial systems, experimentation in appropriate ecosystems is badly needed. To refine our understanding of the role of rising COs in woody ingress on grasslands and in vegetation change generally, it also will be necessary to understand interactions between disturbances and the potential productivity of grasslands on the species and growth form composition of these communities. Progress in recent theoretical (Tilman, 1988) and empirical studies is evident (Wilson and Tilman, 1991). To predict likely effects of atmospheric changes on individual species, it also will be necessary to understand how COs concentration affects environmental conditions under which species can grow and reproduce. Each stage in a plant's life cycle, from germination and establishment through vegetative growth and reproduction, potentially can be influenced directly or indirectly by atmospheric COs concentration. Relevant processes or periods must be identified. Rates of woody expansion on grassland are highly sensitive to the n u m b e r of reproductive plants that are recruited into the population from mature individuals (Hochberg 1994), suggesting that factors that altered seedling establishment and survival were critical to woody expansion. It is of particular importance in predicting vegetation change, then, to understand the interaction of CO2 with seedling requirements for resources such as water that are likely to limit seedling establishment, survival, and growth. We suggest that rising COs likely was most influential in those grasslands in which water or N limited woody recruitment and in which rising COs significantly increased growth rates and the fecundity of woody plants. Ultimately, the ability to predict effects of rising C O s concentration on the dynamics of woody or other species requires an improved understanding of relationships between physiological and growth responses to CO2 and the persistence of plant populations (Austin, 1992).
We thank A. B. Frank, H. W. Hunt, and G. R. McPherson who provided helpful comments on earlier versions of the manuscript.
Archer, S. (1989). Have southern Texas savannas been converted to woodlands in recent history? 134, 545-561.
12.
191
Archer, S. (1994). Woody plant enchroachment into southwestern grasslands and savannas: Rates, patterns, and proximate causes. "Ecological Implications of Livestock Herbivory in the West" (M. Vavra, W. A. Laycock, and R. D. Pieper, eds.), pp. 13-68. Society for Range Management, Denver, CO. Archer, S., Scifres, C., and Bassham, C. R. (1988). Autogenic succession in a subtropical savanna: Conversion of grassland to thorn woodland. 58, 111-127. Arp, W.J., Drake, B. G., Pockman, W. T., Curtis. P. S., and Whigham, D. F. (1993). Interactions between C3 and C4 salt marsh plant species during four years of exposure to elevated atmospheric CO2. 104/105, 133-143. Austin, M. P. (1992). Modelling the environmental niche of plants: Implications for plant community response to elevated CO2 levels. 40, 615-630. Bahre, C.J., and Shelton, M. L. (1993). Historic vegetation change, mesquite increases, and climate in southeastern Arizona. 20, 489-504. Bazzaz, F. A., and Carlson, R. W. (1984). The response of plants to elevated CO2. I. Competition among an assemblage of annuals at two levels of soil moisture. 62, 196-198. Bazzaz, F. A., and McConnaughay, K. D. M. (1992). Plant-plant interactions in elevated CO2 environments. 40, 547-564. Bazzaz, F. A., Garbutt, K., and Reekie, E. (1989). Using growth analysis to interpret competition between a C~ and a C4 annual under ambient and elevated CO2. 79, 223-233. Bazzaz, F. A., Ackerly, D. D., Woodward, R. I., and Rochefort, L. (1992). CO2 enrichment and dependence of reproduction on density in an annual plant and a simulation of its population dynamics. 80, 643-651. Belsky, A.J. (1984). The role of small browsing mammals in preventing woodland regeneration in the Serengeti National Park, Tanzania. 22, 271-279. Belsky, A.J. (1990). Tree/grass ratios in East African savannas: A comparison of existing models. 17, 483-489. Belsky, A.J. (1995). Spatial and temporal landscape patterns in arid and semi-arid African savannas. "Mosaic Landscapes and Ecological Processes" (L. Hansson, L. Fahrig, and G. Merriam, eds.), pp. 31-56. Chapman & Hall, London. Bogusch, E. R. (1952). Brush invasion in the Rio Grande plain of Texas. 4, 85-91. Bragg, T. B., and Hulbert, L. C. (1976). Woody plant invasion of unburned Kansas bluestem 29, 19-24. prairie. J. Bragg, W. K., Knapp, A. K., and Briggs, J. M. (1993). Comparative water relations of seedling and adult Quercus species during gallery forest expansion in tallgrass prairie. 56, 29-41. Brown, J. R., and Archer, S. (1987). Woody plant seed dispersal and gap formation in a North American subtropical savanna woodland: The role of domestic herbivores. 73, 73-80. Brown, J. R., and Archer, S. (1990). Water relations of a perennial grass and seedling vs adult woody plants in a subtropical savanna, Texas. 57, 366-374. Brown, D. A., and Gersmehl, P.J. (1985). Migration models for grasses in the American midcontinent. 75, 383-394. Brown, J. H., and Heske, E.J. (1990). Control of a desert-grassland transition by a keystone rodent guild. 250, 1705-1707. Buffington, L. C., and Herbel, C. H. (1965). Vegetational changes on a semidesert grassland range from 1858 to 1963. 35, 139-164. Burkhardt, J. W., and Tisdale, E. W. (1976). Causes of juniper invasion in southwestern Idaho. 57, 472-484. Bush, J. K., and Van Auken, O. W. (1990). Growth and survival of seedlings associated with shade and herbaceous competition. 151, 234-239. Cole, L. C. (1954). The population consequences of life history phenomena. 29, 103-137.
192 Conroy, J. P., Milham. P.J., Mazur, M., and Barlow, E. W. R. (1990). Growth, dry weight partitioning, and wood properties of D. Don after 2 years of CO2 enrichment. 13, 329-337. Conroy, J. P., Milham, P.J., and Barlow, E. W. R. (1992). Effect of nitrogen and phosphorus availability on the growth response of to high CO2. 15, 843-847. Davis, M. B. (1986). Climatic instability, time lags, and community disequilibrium. "Community Ecology" (J. Diamond and T.J. Case, eds.), pp. 269-284. Harper & Row, New York. Davis, O. K., and Turner, R. M. (1986). Palynological evidence for the historic expansion of juniper and desert shrubs in Arizona, U.S.A. Rev. 49, 177-193. Ehleringer, J. R. (1993) Variation in leaf carbon isotope discrimination in Implications for growth, competition, and drought survival. 95, 340-346. Ehleringer, J., and Bj6rkman, O. (1977). Quantum yields for COz uptake in C3 and C, plants. 59, 86-90. Ehleringer, J. R., and Cooper, T. A. (1988). Correlations between carbon isotope ratio and microhabitat in desert plants. 76, 562-566. Friedli, H., L6tscher, H., Oeschger, H., Siegenthaler, U., and Stauffer, B. (1986). Ice core record of the 13C/lZC ratio of atmospheric CO 2 in the past two centuries. 324, 237-238. Gibbens, R. P., Beck, R. F., McNeely, R. P., and Herbel, C. H. (1992). Recent rates of mesquite establishment in the northern Chihuahuan Desert. 45, 585-588. Glendening, G. E. (1952). Some quantitative data on the increase of mesquite and cactus on a desert grassland range in southern Arizona. 33, 319-328. Glendening, G. E., and Paulsen, H. A., Jr. (1955). "Reproduction and Establishment of Velvet Mesquite as Related to Invasion of Semidesert Grasslands." United States Department of Agriculture, Forest Service Technical Bulletin 1127. Goldberg, D. E. (1990). Components of resource competition in plant communities. "Perspectives on Plant Competition" (J. B. Grace, and D. Tilman, eds.), pp. 27-49. Academic Press, San Diego. Grover, H. D., and Musick, H. B. (1990). Shrubland encroachment in southern New Mexico, U.S.A.: An analysis of desertification processes in the American southwest. 17, 305-330. Hand, D. W., Warren Wilson, J., and Acock, B. (1993). Effects of light and CO2 on net photosynthetic rates of stands of aubergine and 71, 209-216. Hari, P., and Arovaara, H. (1988). Detecting CO2 induced enhancement in the radial increment of trees. Evidence from northern timber line. J. For. Res. 3, 67-74. Harrington, G. N. (1991). Effects of soil moisture on shrub seedling survival in a semi-arid grassland. 72, 1138-1149. Hastings,J. R., and Turner, R. M. (1965). Univ. of Arizona Press, Tucson. Heitschmidt, R. K., and Dowhower, S. L. (1991). Herbage response following control of honey mesquite within single tree lysimeters. J. 44, 144-149. Hennessy, J. T., Gibbens, R. P., Tromble, J. M., and Cardenas, M. (1983). Vegetation changes from 1935 to 1980 in mesquite dunelands and former grasslands of southern New Mexico. 36, 370-374. Hobbs, N. T., Schimel, D. S., Owensby, C. E., and Ojima, D. S. (1991). Fire and grazing in the tallgrass prairie: Contingent effects on nitrogen budgets. 72, 1374-1382. Hobbs, R. J., and Mooney, H. A. (1986). Community changes following shrub invasion of grassland. 70, 508-513. Hochberg, M. E., Menaut, J. c., and Gignoux, J. (1994). The influences of tree biology and fire in the spatial structure of the West African savannah. 82, 217-226. Hunziker, J. H., Palacois, R. A., Poggio, L., Naranjo, C. A., and Yang, T. W. (1977). Geographic distribution, morphology, hybridization, cytogenetics and evolution. "Creosotebush:
12.
193
The Biology and Chemistry of in New World Deserts" (T. J. Mabry, J. H. Hunziker, and D. R. DiFeo, eds.), pp. 10-47. Dowden, Hutchinson and Ross, Stroudsburg, PA. Idso, S. B. (1989). "Carbon Dioxide and Global Change: Earth in Transition." IBR Press, Tempe, AZ. Idso, S. B., Kimball, B. A., and Allen, S. G. (1991). CO2 enrichment of sour orange trees: 2.5 years into a long-term experiment. 14, 351-353. Joffre, R., and Rambal, S. (1993). How tree cover influences the water balance of Mediterranean rangelands. 74, 570-582. Johnson, H. B., and Mayeux, H. S. (1992). Viewpoint: A view on species additions and deletions and the balance of nature. 45, 322-333. Johnson, H. B., Polley, H. W., and Mayeux, H. S. (1993a). Increasing CO2 and plant-plant interactions: Effects on natural vegetation. 104/105, 157-170. Johnson, H. B., Polley, H. W., and Mayeux, H. S. (1993b). Elevated CO2 amplifies expression of genetic variability in 74, 296. Johnston, M. C. (1963). Past and present grasslands of southern Texas and northeastern Mexico. 44, 456-466. Kauppi, P. E., Mielikainen, K., and Kuusela, K. (1992). Biomass and carbon budget of European forests, 1971 to 1990. 256, 70-74. Kirkham, M. B., He, H., Bolger, T. P., Lawlor, D. J., and Kanemasu, E. T. (1991). Leaf photosynthesis and water use of big bluestem under elevated carbon dioxide. 31, 1589-1594. Knight, C. L., Briggs, J. M., and Nellis, M. D. (1994). Expansion of gallery forest on Konza Prairie Research Natural Area, Kansas, USA. 9, 117-125. Knoop, W. T., and Walker, B. H. (1985). Interactions of woody and herbaceous vegetation in a southern African savanna. 73, 235-253. K6rner, C., and Arnone, J. A., III. (1992). Responses to elevated carbon dioxide in artificial tropical ecosystems. 257, 1672-1675. LaMarche, V. C., Jr., Graybill, D. A., Fritts, H. C., and Rose, M. R. (1984). Increasing atmospheric carbon dioxide: Tree ring evidence for growth enhancement in natural vegetation. 225, 1019-1021. Long, S. P. (1991). Modification of the response of photosynthetic productivity to rising temperature by atmospheric CO2 concentrations: Has its importance been underestimated? 14, 729-739. Long, S. P., and Drake, B. G. (1991). Effects of the long-term elevation of CO2 concentration in the field on the quantum yield of photosynthesis of the C3 sedge, 96, 221-226. Mayeux, H. S., Johnson, H. B., and Polley, H. W. (1991). Global change and vegetation dynamics. "Noxious Range Weeds" (L. F. James, J. o. Evans, M. H. Ralphs, and R. D. Child, eds.), pp. 62-74. Westview Press, Boulder, CO. McPherson G. R., Boutton, T. W., and Midwood, A.J. (1993). Stable carbon isotope analysis of soil organic matter illustrates vegetation change at the grassland/woodland boundary in southeastern Arizona, USA. 93, 95-101. Medina, E. (1982). Nitrogen balance in the grasslands of central Venezuela. 67, 305-314. Medina, E., and Silva, J. F. (1990). Savannas of northern South America: A steady state regulated by water-fire interactions on a background of low nutrient availability.J. 17, 403-413. Miao, S. L, Wayne, P. M., and Bazzaz, F. A. (1992). Elevated CO2 differentially alters the responses of coocurring birch and maple seedlings to a moisture gradient. 90, 300-304. Morison, J. I. L. (1993). Response of plants to CO2 under water limited conditions. 104/105, 193-209.
194 Morse, S. R., and Bazzaz, F. A. (1994). Elevated CO2 and temperature alter recruitment and size hierarchies in C3 and C4 annuals. 75, 966-975. Neilson, R. P. (1986). High-resolution climatic analysis and southwest biogeography. 232, 27-34. Nie, D., Kirkham, M. B., Ballou, L. K., Lawlor, D.J., and Kanemasu, E. T. (1992). Changes in prairie vegetation under elevated carbon dioxide levels and two soil moisture regimes. 3, 673-678. Norby, R.J. (1987). Nodulation and nitrogenase activity in nitrogen-fixing woody plants stimulated by CO2 enrichment of the atmosphere. 71, 77-82. Norby, R.J., and O'Neill, E. G. (1991). Leaf area compensation and nutrient interactions in CO2-enriched seedlings of yellow-poplar (Liri0dendr0n 117, 515-528. Norby, R.J., O'Neill, E. G., and Luxmoore, R.J. (1986). Effects of atmospheric CO2 enrichment on the growth and mineral nutrition of seedlings in nutrient-poor soil. 82, 83-89. Owensby, C. E., Coyne, P. I., Ham, J. M., Auen, L. M., and Knapp, A. K. (1993). Biomass production in a tallgrass prairie ecosystem exposed to ambient and elevated COs. 3, 644-653. Polley, H. W., Johnson, H. B., and Mayeux, H. S. (1992). Growth and gas exchange of oats and wild mustard at subambient COs concentrations. 153, 453-461. Polley, H. W., Johnson, H. B., Marino, B. D., and Mayeux, H. S. (1993a). Increase in C3 plant water-use efficiency and biomass over glacial to present COs concentrations. 361, 61-64. Polley, H. W., Johnson, H. B., Mayeux, H. S., and Malone, S. R. (1993b). Physiology and growth of wheat across a subambient carbon dioxide gradient. 71, 347-356. Polley, H. W., Johnson, H. B., and Mayeux, H. S. (1994). Increasing COs: Comparative responses of the C4 grass and grassland invader 75, 976-988. Polley, H. W., Johnson, H. B., and Mayeux, H. S. (1995a). Nitrogen and water requirements of C3 plants grown at glacial to present carbon dioxide concentrations. 9, 86-96. Polley, H. W., Mayeux, H. S.,Johnson, H. B., and Tischler, C. R. (1995b). Viewpoint: Implications of rising atmospheric COs concentration for soil water availability and shrub/grass ratios on grasslands and savannas. Submitted for publication. Poorter, H. (1993). Interspecific variation in the growth response of plants to an elevated ambient CO2 concentration. 104/105, 77-97. Prentice, I. C. (1986). Vegetation responses to past climatic variation. 67, 131-141. Reekie, E. G., and Bazzaz, F. A. (1989). Competition and patterns of resource use among seedlings of five tropical trees grown at ambient and elevated COs. 79, 212-222. Rice, B., and Westoby, M. (1978). Vegetative responses of some Great Basin shrub communities protected against jack-rabbits or domestic stock. 31, 28-34. Sala, O. E., Parton, W. J., Joyce, L. A., and Lauenroth, W. K. (1988). Primary production of the central grassland region of the United States. 69, 40-45. Schlesinger, W. H., Reynolds, J. F., Cunningham, G. L., Huenneke, L. F., Jarrell, W. M., Virginia, R. A., and Whitford, W. G. (1990). Biological feedbacks in global desertification. 247, 1043-1048. Sharkey, T. D. (1988). Estimating the rate of photorespiration in leaves. 73, 147-152. Smedley, M. P., Dawson, T. E., Comstock, J. P., Donovan, L. A., Sherrill, D. E., Cook, C. S., Ehleringer,J. R. (1991). Seasonal carbon isotopic discrimination in a grassland community. 85, 314-320. Smeins, F. E. (1983). Origin of the brush problemmA geographical and ecological perspective of contemporary distributions. "ProceedingsmBrush Management Symposium," pp. 5-16. Texas Tech Univ. Press, Lubbock, TX.
12.
195
Smith, D. A., and Schmutz, E. M. (1975). Vegetative changes on protected versus grazed desert grassland ranges in Arizona. J. 28, 453-458. Smith, T., and Huston, M. (1989). A theory of the spatial and temporal dynamics of plant communities. 83, 49-69. Stephenson, N. L. (1990). Climatic control of vegetation distribution: The role of the water balance. 135, 649-670. Steuter, A. A., Jasch, B., Ihnen, J., and Tieszen, L. L. (1990). Woodland/grassland boundary changes in the middle Niobrara Valley of Nebraska identified by 613Cvalues of soil organic matter. 124, 301-308. Thomas, R. B., Richter, D. D., Ye, H., Heine, P. R., and Strain, B. R. (1991). Nitrogen dynamics and growth of seedlings of an N-fixing tree (Gliricidia (Jacq.) Walp.) exposed to elevated atmospheric carbon dioxide. 88, 415-421. Tieszen, L. L., and Archer, S. (1990). Isotopic assessment of vegetation changes in grassland and woodland systems. "Plant Biology of the Basin and Range" (C. B. Osmond, L. F. Pitelka, and G. M. Hidy, eds.), Vol. 80, pp. 293-321. Springer-Verlag, Heidelberg. Tilman, D. (1988). "Plant Strategies and the Dynamics and Structure of Plant Communities," Vol. 26. Princeton Univ. Press, Princeton, NJ. Tolley, L. C., and Strain, B. R. (1985). Effects of CO2 enrichment and water stress on gas exchange of and seedlings grown under different irradiance levels. 65, 166-172. Turner, R. M. (1990). Long-term vegetation change at a fully protected Sonoran desert site. 71, 464-477. Vitousek, P. M. (1993). Global dynamics and ecosystem processes: Scaling up or scaling down? "Scaling Physiological Processes Leaf to Globe" (J. R. Ehleringer, and C. B. Field, eds.), pp. 169-177. Academic Press, San Diego. Webb, W., Szarek, S., Lauenroth, W., Kinerson, R., and Smith, M. (1983). Primary productivity and water use in native forest, grassland, and desert ecosystems. 59, 1239-1247. Weiner, J. (1990). Asymmetric competition in plant populations. 5, 360-364. Williams, K., and Hobbs, R.J. (1989). Control of shrub establishment by springtime soil water availability in an annual grassland. 81, 62-66. Williams, K., Hobbs, R.J., Hamburg, S. P. (1987). Invasion of an annual grassland in northern California by ssp. 72, 461-465. Wilson, S. D., and Tilman, D. (1991). Interactive effects of fertilization and disturbance on community structure and resource availability in an old-field plant community. 88, 61-71. Wright, H. A., Bunting, S. C., and Neuenschwander, L. F. (1976). Effect of fire on honey mesquite. 29, 467-471. York, J. c., and Dick-Peddie, W. A. (1969). Vegetation changes in southern New Mexico during the past hundred years. "Arid Lands in Perspective" (W. G. McGinnies and B.J. Goldman, eds.), pp.157-166. Univ. of Arizona Press, Tucson.
This Page Intentionally Left Blank
13 Effects o f E n h a n c e d [W-B Radiation and Elevated Concentrations o f CO2 on a Subarctic H e a t h l a n d
Furore increases in the levels of surface UV-B radiation (280-320 nm) reaching polar regions as a result of stratospheric ozone depletion have been predicted (Farman, Gardiner, and Shanklin, 1985; Frederick and Snell, 1988; Proffitt 1990; Hoffman and Deshler, 1991; Kerr, 1993; Gleason 1993). Current knowledge on the biological implications of these changes for both fauna and flora of these regions is limited. This chapter concentrates mainly on the direct effects of UV-B radiation on the flora of the Northern polar region, and effects on other life forms are also briefly outlined. Studies of the effects of UV-B on vegetation have been restricted to temperate species and crops, and most of these studies have been laboratory rather than field based. Generally following exposure, most of these species display a reduction in plant growth. Such reductions can be caused by inhibition of photosynthesis, DNA damage, or changes in the amounts of regulatory compounds (Bornman and Teramura, 1993; United Nations Environment Program (UNEP), 1994). Effects are generally cumulative in that the damage accumulates as the period of exposure progresses (Sullivan and Teramura, 1992; Johanson 1995a; Johanson 1995b). Concomitant with changes in UV-B radiation, the atmospheric concentrations of CO2 have been predicted to rise from present (365 ml liter -1) to 197
Copyright 9 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.
198 elevated (600 ml liter -1) concentrations by the year 2050 (Watson 1990). Plant responses to elevated CO2 are species-specific, although the majority of growth responses are positive. Investigations to date have looked at the effects of elevated CO2 from the single plant to the community level (e.g., Krupa and Kickert, 1989; Eamus and Jarvis, 1989; Drake and Leadley, 1991; Newton, 1991; Baker and Allen, 1994). Only one study site to date has been established to look at the effects of elevated CO2 on a polar ecosystem (Toolik Lake, Alaska, 68 ~ 38' N, 149 ~ 34' W). The early part of the study looked at tussocks exposed to different air temperatures and different concentrations of atmospheric CO2 (Tissue and Oechell, 1987). A stimulation of new filler production was observed in at elevated CO2, whereas the growth of mature tillers were found to be unaffected by the treatment. However, subsequent experiments looked at carbon flux and sequestration in the ecosystem. Results suggested that the system had reached a complete status of homeostasis following a 3-year exposure to elevated concentrations of carbon dioxide (Oechell 1993). Little is known about the way plants respond to simultaneous UV-B and CO2 exposure. Hypothetically in some species the positive effect of CO2 may ameliorate the negative effects of UV-B. Most studies have shown reductions in growth at enhanced UV-B irrespective of CO2 concentration (Teramura, Sullivan, and Lewis, 1990; Ziska and Teramura, 1992; Sullivan and Teramura, 1994). However, such studies looking at the interactive effects of UV-B radiation and elevated CO2 have concentrated on temperate plant species grown individually in pots (Rozema, Lenssen, and van de Staaij, 1990; Ziska and Teramura, 1992; Stewart and Hoddinot, 1993; van de Staaij, Lenssen, Stroetenga, and Rozema, 1993; Sullivan and Teramura, 1994). Field experiments are therefore required to quantify the responses of naturally occurring plant species where responses may be different. Against this background, during May 1993 field experiments were initiated to study the simultaneous effects of UV-B radiation and elevated CO2 on arctic vegetation. The experiments were established on a subarctic heathland ecosystem at Abisko, Sweden (see below). The principal aim of this chapter is to discuss the results of the first 2 years of experimentation and outline future initiatives.
The experiments were conducted on plots situated at the Abisko Scientific Research Station, Abisko, Sweden (N 68 ~ 21'; 18~ 49'E). The research site was based in an open canopy birch forest community Ehrh. ssp. corresponding to the
13.
on
199
variant" described by Sonesson and Lundberg (1974). This site contained scattered birch trees 3 m high with a field layer dominated by deciduous and evergreen ericaceous dwarf shrubs. The major species included (Hagerup.), (L.), (L.), and (L.). The main grass species found in the heathland were (L.) and (L.), which were sparsely distributed within the field layer. Major cryptogams included (Hedw.), (Hedw.), (L.), and (L.) which inhabited the understory (see Johanson 1995a, for general composition of the community). The understory heath vegetation was exposed to enhanced levels of UVB (representing a 15% ozone depletion) using aluminium frames housing fluorescent lamps following the methods ofJohanson (1995a). Each of the 16 frames (2.5 m • 1.3 m • 1.5 m high) housed six fluorescent lamps (Q-PANEL UVB-313, Cleveland, OH). The middle 70 cm of the two center lamps were covered by aluminum foil to ensure uniform radiation over the plots. In the frames used for enhanced UV-B treatment, each lamp was covered by a preburned cellulose diacetate filter (0.13 mm, Courtaulds, Derby, UK) housed on a UV-transmitting plexiglass plate (R6hm 2458, R6hm GmbH, Damstdt, Germany). This was done to ensure absorption of UV-C radiation (<280 nm) but allow transmission of the UV-B radiation. Lamps in control plots were blocked by 5-mm-thick glass to absorb any short wavelength UV radiation. Each of the plots received natural UV-B radiation, which allowed us to compare natural and enhanced levels of UV-B radiation. Daily irradiation was centered around noon, using timers which controlled three lamps at once allowing a stepwise increase and decrease in daily UV-B. Irradiation times were changed every second week to match seasonal changes in UV-B radiation. The model of Bj6rn and Murphy (1985) was used to calculate changes in UV-BBE radiation as a result of a 15% ozone depletion. In addition to enhanced UV-B radiation the heathland was simultaneously exposed to elevated (600 ml liter -1 + 7%) or present (360 ml liter -1 -+ 2%) atmospheric concentrations of carbon dioxide (CO2). There were four combinations of treatments each replicated four times (enhanced/natural levels of UV-B radiation versus elevated/present atmospheric concentrations of carbon dioxide). For this purpose 16 open-top chambers (0.85 m diameter and 0.50 m height, 0.73 m 2 area and 0.37 m 3 volume) were placed centrally under each of the 16 frames. The chambers were constructed from UV-transmitting plexiglass (R6hm 2458, R6hm GmbH, Damstdt, Germany) to minimize effects on natural radiation. Atmospheric air was blown into each chamber from mixing boxes housing fans (Model TD350/125, Soler and Palau, Spain). Air entered each chamber
200
through a plenum with numerous holes ensuring approximately five air changes per minute. In CO2-enriched chambers pure CO2 was trickled into the air supply via flow controllers (Platon Ltd, UK). The CO2 concentration within the chambers was continuously monitored at a vegetation height of 10 cm using an IRGA (Series 2000, ADC, UK), relay multiplexer (model AM416), and data logger (model CR10, Campbell Scientific, U.S.A.). Sensors were placed within the chambers to measure photosynthetically active radiation (PAR sensor, Skye Instruments, UK), air temperature (thermocouple's), and relative humidity (Series 2000, Skye Instruments, UK) linked to a datalogger (Delta Logger, Delta-T Services, UK). The experimental site was established during August 1992 and exposure to the respective treatments initiated during snow melt, May 1993. The treatments were maintained throughout the spring and summer and were stopped in early September when autumn began. The same sequence of events will be adopted during each subsequent growing season of this 6year study.
A. Dwarf Shrubs (Major Canopy Species) Exposure to elevated C O 2 during the first growing season affected the phenology of deciduous dwarf shrubs, while no effects of UV-B could be observed (Gwynn-Jones unpublished). exposed to elevated COz showed later shoot development and earlier leaf senescence signifying a reduced growing season. However, such profound CO2-induced changes in development were not observed during the subsequent season. Similarly no effects of either CO2 or UV-B could be observed on the vegetative phenology of the evergreen dwarf shrub during the first two growing seasons. A reduced peak season, photosynthetic capacity was observed in V. during the first season of exposure to elevated CO2. This may have been due to end-product inhibition of photosynthesis via accumulation of starch. Similar responses were suggested by Tissue and Oechell (1987) after only 3 weeks of exposing to elevated CO2 in the Alaskan tundra. Such down regulation of the photosynthetic machinery may be anticipated in the nitrogen-limited soils of the subarctic. During the first season of exposure the growth of V. increased at elevated CO2 regardless of changes in photosynthesis and phenological development. However, no CO2 or UV-B effects were observed on the growth of the evergreen dwarf shrub The only response apparent during a second season of exposure was an interaction effect of UV-B and CO2 on the growth of V. which was reduced during
13.
on
201
simultaneous exposure. This may have resulted from an interaction between effects of CO2 and UV-B on photosynthesis. During the second season of exposure increases in f o w e r and berry production of the deciduous species V. were observed u n d e r enhanced UV-B (Gwynn-Jones unpublished ). No effects on flowering were expected during the first season because flower bud formation in dwarf shrubs are generally determined by the conditions of the previous season (Emanuelson and Callaghan, 1994). Stimulation of flowering in the second season may have been caused by the effects of I.W-B on the hormone control of flower bud formation. The only CO2 treatment effect apparent was a reduction in the proportion of flowers producing berries in V. The cause of this is uncertain, but perhaps may be linked to the interaction between nitrogen and carbon assimilation. Although elevated COz should have ensured adequate carbohydrate supply to satisfy sink capacity (Farrar and Williams, 1991), competing sinks for nitrogen within the plant may have been affected in such a way as to interfere with berry production. Exposure to enhanced levels of UV-B radiation during the first season of exposure had no effects on the leafflavonoid (UV-B protecting pigments) content of both deciduous and evergreen dwarf shrub species. Data are not available for the seasonal changes, but in a related study investigating the long-term effect (>3 years) of UV-B radiation per se on the same heathland vegetation, Johanson (unpublished) has found increases in flavonoids. The effects of UV-B x CO2 on flavonoid accumulation will be investigated in future seasons. Little is known about the energy cost of producing and maintaining protective pigments in the plant. Increased investment in these compounds at enhanced UV-B may be at the expense of growth, although such loss of energy to photoprotection may not be as great at elevated CO2 because plant maintenance costs are often reduced (Reuveni and Gale, 1985; Bunce, 1990; Bunce and Caulfield, 1990). Studies to date have concentrated mainly on the effects of UV-B and CO2 on aboveground growth and development of the dwarf shrubs, because studies of the rhizosphere in the current system have presented difficulties. However, some of the aboveground responses observed may be partially mediated through changes occurring in the rhizosphere. A review by Rogers, Brett Runion, and Krupa (1994) extensively discussed the effects of elevated CO2 belowground processes, but our knowledge on UV-B effects are limited. The heathland surface soil horizon is dominated by the roots of the ericaceous dwarf shrub which may be indirectly and directly influenced by such perturbations. Individual studies on decomposition (Gehrke 1995), soil water relations (Gwynn-Jones and Pantis, unpublished) and carbon exudation from roots (Norby 1987) all suggest that there
202
could be major effects of these perturbations on the rhizosphere. Any process that influences the availability of soil nutrients may be of particular importance within the heathland.
B. Grasses (Minor Canopy Species) Grass species contribute only a small proportion of total plant cover within the heathland, hence experiments on the effects of UV-B and CO~ on grasses have been carried out in controlled environments. The predominant grass species within the heathland are and Experiments on these species have addressed the individual rather than the interactive effects of these two perturbations. Of particular interest in these species is the nonlinear damage response observed with increasing levels of UV-B radiation. A 40% decrease in dry weight was observed following 60 d exposure to enhanced UV-B radiation representing a 15% ozone depletion (cf. natural level) while no effects were observed at a higher UV-B dosage (representing 25% ozone depletion) (Gwynn-Jones andJohanson, unpublished). Damage caused by UV-B radiation could be overcome at higher levels via stimulation of tillering. Such tillering could have been stimulated at the higher UV-B level by the direct effect of auxin which may control apical dominance. shows a characteristic increase in growth following exposure to elevated concentrations of CO2 (Parsons, unpublished). Given the responses observed individually at enhanced UV-B and elevated CO2 it is difficult to make predictions as to the response of this species to simultaneous exposure. From the positive responses observed, it could be hypothesized that combined effects of these two variables may increase the significance of grasses in the heathland. However, there are no apparent changes in the importance of grasses within our main field site, which remain at low densities following two full seasons of exposure.
C. Cryptogams (Understory Species) showed increased shoot growth at elevated CO2 but no effects of UV-B were apparent after 2 years' exposure. Results suggest that the effect was more p r o n o u n c e d in previous year shoots (c + 1) than in those developed during the current year (c). It could be hypothesized that the influence of CO2 on dwarf shrubs may be partially responsible for changes in the growth of these mosses (see Section III A in this chapter). Mosses such as initiate their highest growth rates when deciduous canopy species are leafless--at the beginning and end of each growing season (Karlsson, 1987). Changes in V. phenology may have been partly responsible for the increased shoot growth of due to increased exposure to photosynthetically active radiation. This would emphasize the importance of studying plant responses to environmental per-
13.
on
203
turbation within natural communities as opposed to short-term, singlespecies experiments.
A. Decomposition Experiments at Abisko to date have concentrated on I.W-B effects on decomposition although material has been collected to study the decomposition of tissue previously exposed to simultaneous UV-B and CO2 exposure. The decomposition rate of leaf litter was found to be reduced overall under enhanced I_W-B levels due to both direct and indirect influences (Gehrke 1995). Direct impacts were due to a reduction in the n u m b e r of active microorganisms at enhanced UV-B which was reflected by a reduced microbial respiration. Indirect effects of UV-B were expressed as changes in leaf litter quality, where increases in phenolic (e.g., tannin) substances were apparent following field exposure of species to enhanced I_W-B. Such phenolic compounds complexing with proteins may cause decreased digestibility to microorganisms (Richards, 1987) thus slowing the rate of decomposition. Leaf tissue quality is also influenced by exposure to elevated CO2 as the C : N ratio is commonly increased (see review by Woodward, 1992). However, such marked effects on tissue quality may not necessarily influence the rate of decomposition as this will be dependent on the populations and activities of both microflora and fauna within the ecosystem. It could be predicted that the combination of enhanced UV-B and elevated CO2 may result in reductions in the rate of litter decomposition. Such a response over sequential growing season may result in reduced soil fertility, lower primary production, and greater storage of soil carbon.
B. Herbivory Changes in leaf quality as a result of plant exposure to elevated CO2 and UV-B may also influence the degree of insect herbivory in the subarctic heathland. Bazzaz and Fajer (1992) showed that the Buckeye butterfly was adversely affected by high CO2 as caterpillars would grow more slowly, feeding on plantain grown at elevated CO2. Exposure to I.W-B radiation may also influence leaf tissue quality (Teramura, 1983; Hatcher and Paul, 1994) and hence the success of herbivores. Indeed, Hatcher and Paul (1994) found increases in the levels of leaf phenolics in pea plants exposed to enhanced levels of I.W-B radiation under laboratory conditions. Feeding the leaves of these plants to larva of the moth L. had no deleterious effects on growth rates as the level of nitrogen had also increased following exposure to I.W-B radiation.
204 Further studies are needed to fully understand such relationships and these should be performed in the field and not under artificial laboratory conditions. A field-based project was commenced during the summer of 1995 to look at the direct and indirect effects of UV-B and COz on the moth which feeds predominantly on birches and deciduous dwarf shrubs.
Results from our experiments based in a subarctic heathland suggest that vegetation responses to UV-B and CO2 are species-specific. The deciduous dwarf shrub V. was found to be most sensitive to the perturbations showing both CO2 and UV-B responses during the period of exposure. Responses to elevated CO2 included changes in photosynthesis, phenology, and growth, which were only observed during the first season of exposure. However UV-B responses were apparent during the second season of exposure where the flowering and berry yield of this species was stimulated at e n h a n c e d UV-B. The other dwarf shrub species present within the heathland appeared to be unresponsive to these environmental perturbations in aspects of physiology, demography, and growth during the two seasons of exposure. Compared to many laboratory investigations, the dose simulating a 15% reduction in the ozone layer is fairly modest. This, combined with the fact that all the plants studied are long-lived perennials, suggests that small, potentially cumulative damage may be occurring. O u r evidence for field responses of cryptogam to UV-B and CO2 is limited, although we have observed some stimulation of growth in the moss H. at elevated COz during the second season of exposure. Further field studies are required to understand the long- and short-term sensitivity of the understory species (including lichens and other moss species) to such environmental perturbation. An understanding of the effects of global climate change on seminatural ecosystems must be u n d e r p i n n e d by realistic and long-term experimentation. The experiment described in this chapter is one attempt to do so. O u r ecosystem approach is diagramatically illustrated in Fig. 1. It shows the breadth of our approach and the interdependence of trophic levels addressed. Responses observed in the first 2 years are fairly modest. It is probable that the ecological interest will increase with time. This should allow, for example, a thorough investigation not only of cumulative plant-specific responses to perturbations, but also a full evaluation of the effects on other ecosystem components and, in particular soil processes.
13.
on
~ / structureI Figure
1 Diagram showing the relationship between the ecosystem as a whole and the environmental perturbations addressed.
We are grateful to the CEC for financial support and to the Abisko Naturvetenskapliga Station (Abisko, N. Sweden) for allowing these experiments to be conducted and providing excellent technical and administrative support.
Baker, J. T., and Allen, L. H., Jr. (1994). Assessment of the impact of rising carbon dioxide and other climate changes on vegetation. 83, 223-235. Bazzaz, F. A., and Fajer, E. D. (1992). Plant life in a CO2-rich world. 1, 18-24. Bj6rn, L. O., and Murphy, T.M. (1985). Computer calculations of solar ultraviolet radiation at ground level. 23, 555-561. Bornman, J.F., and Teramura, A. H. (1993). Effects of ultraviolet-B radiation on terrestrial plants. "Environmental Photobiology" (]. Young, ed.), pp. 427-471. Plenum, New York. Bunce, J. A. (1990). Short- and long-term inhibition of respiratory carbon dioxide efflux by elevated carbon dioxide. 65, 637-642. Bunce,J. A., and Caulfield, F. (1990). Reduced respiratory carbon dioxide efflux during growth at elevated carbon dioxide in three herbaceous perennial species. 67, 325-330. Drake, B. G., and Leadley, P. W. (1991). Canopy photosynthesis of crops and native plant communities exposed to long-term elevated CO2. 14, 853-860. Eamus, D., and Jarvis, P. G. (1989). The direct effects of increases in the global atmospheric CO2 concentrations on natural and commercial temperate trees and forests. 19, 2-55. Emanuelsson, U., and Callaghan, T. V. (1994). Population structure and process of tundra plants and vegetation. "The Population Structure of Vegetation" (J. White, ed.), pp. 399-439. Junk of Bodstricht press. Farman, J. C., Gardiner, B. G., and Shanklin, J. D. (1985). Large losses of total ozone in Antarctica reveal seasonal C1Ox/NOx interaction. 315, 207-210.
206 Farrar, J. F., and Williams, M. L. (1991). The effects of increased atmospheric carbon dioxide and temperature on carbon partitioning, source-sink relations, and respiration: Commissioned review. 14(8), 819-831. Frederick, J. E., and Snell, H. E. (1988).Ultraviolet radiation levels during the Antarctic spring. 241, 438-440. Gehrke, C., Johanson, U., Callaghan, T., Chadwick, D. and Robinson, C. H. (1995). The impact of enhanced ultraviolet-B radiation on litter quality and decomposition processes in leaves from the sub-arctic. 72, 213-222. Gleason, J. F., Bhartia, P. K., Herman, J. R., McPeters, R., Newman, P., Stolarski, R. S., Flynn, L., Labow, G., Larko, D., Seftor, C., Wellemeyer, C., Komhyr, W. D., Miller, A., and Planet, W. (1993). Record low global ozone in 1992. 290, 523-526. Hatcher, P. E., and Paul, N. P. (1994). The effects of elevated UV-B radiation on herbivore of pea by Autographa gamma. 71 (3), 227-233. Hoffman, D.J., and Deshler, T. (1991). Evidence from balloon measurements for chemical depletion of stratospheric ozone in the Arctic winter of 1989-1990, 349, 300-305. Johanson, U., Gehrke, C., Bj6m, L. O., Callaghan, T. V., and Sonesson, M. (1995a). The effects of enhanced UV-B radiation on a sub-arctic heath ecosystem. 24, 106-111. Johanson, U., Gehrke, C., Bj6rn, L. O., and Callaghan, T. V. (1995b). The effects of enhanced UV-B radiation on the growth of dwarf shrubs in a sub-arctic heathland. 9(5), 713-719. Karlsson, P. S. (1987). Niche differentiation with respect to light utilization among coexisting dwarf shrubs in a sub-arctic woodland. 8, 35-39. Kerr, R. A. (1993). The ozone hole reaches a new low. 262, 501. Krupa, S., and Kickert, R. N. (1989). The greenhouse effect: Impacts of ultraviolet-B (UVB), carbon dioxide (CO2), and ozone (03) on vegetation. 61, 263-293. Newton, P. C. D. (1991). Direct effects of increasing carbon dioxide on pasture plants and communities. 34, 1-24. Norby, R., O'Neill, E. G., Hood, W. G., and Luxmoore, R.J. (1987). Carbon allocation, root exudation, and mycorrhizal colonization of seedlings grown under CO2 enrichment. 3, 203-210. Oechell, W. C., Hastings, S.J., Vourlitis, G., Jenkins, M., Riechers, G., and Grulke, N. (1993). Recent change of Arctic tundra ecosystems from a net carbon dioxide sink to a source. 261 (6412), 520-523. Proffitt, M. H., Margitan, J. J., Kelly, K. K., Loewenstein, M., Podolske, J. R., Jones, and Chan, K. R. (1990). Ozone loss in the Arctic polar vortex inferred from high-altitude aircraft measurement. 347, 31-36. Reuveni,J., and Gale,J. (1985). The effect of high levels of carbon dioxide on dark respiration and growth of plants. 8, 623-628. Richards, B. N. (1987). "The Microbiology of Terrestrial Ecosystems." Longman, New York. Rogers, H. H., Brett Runion, G., and Krupa, S. V. (1994). Plant responses to atmospheric CO2 enrichment with emphasis on roots and the rhizosphere. 83, 155-189. Rozema, J., Lenssen, G. M., and van de Staaij, J. W. M. (1990). The combined effects of increased atmospheric COs and I.W-B radiation on some agricultural and salt marsh species. "The Greenhouse Effect and Primary Productivity in European Agroecosystems" (J. Goudriaan., H. van Keulen, and H. H. van Laar, eds.), pp. 68-71. Pudoc Wageningen. Sonesson, M., and Lundberg, B. (1974). Late quaternary forest development in Tornetr/isk area, northern Sweden. I. Structure of modern forest ecosystems. 25, 121-133. van de Staaij,J. W. M., Lenssen, G. M., Stroetenga, M., and Rozema, J. (1993). The combined effect of elevated COs levels and UV-B radiation on growth characteristics ofElymus 104/105, 433-439. Stewart, J. D., and Hoddinot, J. (1993). Photosynthetic acclimation to elevated atmospheric 88, 493-500. carbon dioxide and UV irradiation in
13.
on
207
Sullivan, J. H., and Teramura, A. H. (1992). The effects of ultraviolet-B radiation on loblolly pine. II. Field-grown seedlings. 6, 115-120. Sullivan, J. H., and Teramura, A. H. (1994). The effects of UV-B radiation on loblolly pine. III. Interaction with CO2 enhancement. 17, 311-317. Teramura, A. H. (1983). Effects of UV-B radiation on the growth and yield of crop plants. 58, 415-427. Teramura, A. H., Sullivan, J. H., and Lewis (1990). Interaction of elevated UV-B radiation and COs on productivity and photosynthetic characteristics in wheat, rice and soybean. 94, 470-475. Tissue, D. T., and Oechell, W. C. (1987). Response of to elevated COs and temperature in the Alaskan tussock tundra. 68, 401-410. United Nations Environment Program (UNEP) (1994). Effects of increased solar ultraviolet radiation on terrestrial plant. "Environmental Effects of Ozone Depletion--1994 Assessment," pp. 49-65. Nairobi, Kenya. Watson, R. T., Rodhe, H., Oeschger, H., and Siegenthaler, U. (1990). Greenhouse gases and aerosols. "Climate Change--IPCC Scientific Assessment" (J. T. Houghton, G.J.Jenkins, and J. J. Ephramus, eds.), pp. 1-40. Cambridge Univ. Press, Cambridge, UK. Woodward, F. B. I. (1992). Predicting plant responses to global environmental change. New 122, 239-251. Ziska, L. H., and Teramura, A. H. (1992). CO2enhancement of growth and photosynthesis in rice (Oryza Modification by increased ultraviolet-B radiation. 99, 473-481.
This Page Intentionally Left Blank
1 Carbon Metabolism and Plant Growth under Elevated C02 in a Natural Quercus ilex L. "Macchia" Stand
Mediterranean-type woodland communities represent slightly more than 10% of the total forest surface of the world (Walter, 1985), and they make up the natural vegetation of some of the most populated and economically active areas of the globe. At the same time, the Mediterranean biome represents the intermediate vegetation between the desert zone and the temperate forests. It is, therefore, crucial to be able to anticipate the possible effects of environmental changes on these plant communities given their essential role on protecting lands that are under a strong pressure by man and climate. In Mediterranean-type ecosystems, the two main factors limiting primary productivity are water and nutrient availability (Specht, 1973; Debano and Conrad, 1978). Additionally, the frequency of disturbances is high due to the occurrence of wildfires during the summer dry season, harvesting of biomass, and animal grazing. Many plant species of the Mediterranean regions are evergreen sclerophyll shrubs and trees adapted to low water and nutrient availability, and also able to rapidly recover after disturbance by resprouting from protected buds (Naveh, 1974). L. (holm oak) is the dominant tree species of most mature communities over large areas of the Mediterranean basin (Romane and Terradas, 1992). This species avoids the damaging effects of summer 209
Copyright 9 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.
210
water deficits through mechanisms that maintain positive turgor during periods of reduced water availability (Rhizopoulou 1989; Rhizopoulou and Mitrakos, 1990; Terradas and Say6, 1992), by means of stomatal closure (Romane and Terradas, 1992), and through morphological adaptations that improve the efficiency of the plant hydraulic system (Rambal, 1993). Predictions on the effects of increasing CO2 concentrations on natural communities of trees and herbs have been traditionally inferred from short-term studies conducted on plants raised under controlled conditions (Cure and Acock, 1986; Sengupta and Sharma, 1993; Ceulemans and Mousseau, 1994). Hence, a large wealth of information has been accumulated regarding the physiology of photosynthesis and transpiration (Sfitt, 1991; Eamus, 1991), stomatal activity (Idso, 1991; Bunce, 1993), leaf anatomy (Radoglou andJarvis, 1990), and biomass growth and distribution into components (Bazzaz, 1993; Luo 1994). However, according to Cipollini (1993) concern has arisen over the validity of extrapolations from short-term, small-scale experiments. Only a few long-term experiments at the community and ecosystem level have been conducted worldwide (Oechel and Riechers, 1986; Drake 1989) although a number of CO2 exposure experiments in natural conditions were recently initiated, in particular within the European ECOCRAFI' research network on "The likely impact of rising [CO2] and temperature on European forests." Critical questions that should be solved by experiments in natural habitats are (i) which plant functional type will be successful in a future double CO2 world, particularly in a drier environment; (ii) whether photosynthesis undergoes down-regulation adjustments over long-term exposure (Sage 1989); and in case of no major acclimating effects, (iii) how the excess of organic carbon is utilized by plants, particularly woody species, trees, and shrubs. Evidence exists about the increase of nonstructural carbohydrates in plants under elevated CO2, particularly in leaves (Wullschleger 1992), but also about the increment of some tree biomass components as woody stem and roots (K6rner and Arnone, 1992). The objective of our research has been, therefore, to examine the impact of long-term exposure to elevated CO2 concentration in a natural Mediterranean community dominated by (high "macchia"). Emphasis is placed on measurements of carbon metabolism and light energy utilization by the leaves in order to assess the physiological responses that subtend growth rather than just trying to measure short-term biomass increments at the tree level alone.
14.
A. The Experimental Site This study capitalizes on a long-term research conducted over the past 5 years aimed at understanding the interactions between structure, functions, and microclimate in a woody plant community, representing the natural vegetation of the Mediterranean, coastal sand dunes. The site is located near Montalto di Castro (Viterbo), along the Thyrrenian coast, 100 km northwest of Rome, at the E.N.E.L. (National Agency of Electric Energy) reservation (Lat. 42 ~ 22' N, 11 ~ 32' E). The vegetation is a Mediterranean evergreen " m a c c h i a " ecosystem, 4 - 6 m tall, dominated by trees, with a dense shrub layer made up of L., and L. Toward the seashore, this Br. B1. community is substituted by a dense belt of Caneva; toward the interior, beyond the last sand dune, the vegetation becomes a 15 m tall, deciduous forest composed of spp., and spp. trees (Fig. 1). Within the intermediate belt of high total aboveground biomass is 35 Mg ha -1 of dry matter, vegetation covers about 80% of the g r o u n d surface, and leaf area index (LAI) ranges from 3 to 4. H o l m oak
Vegetationdistribution along a profile from the coast to the interior at the Montalto di Castro experimental site.
212
trees represent 34% of total n u m b e r of woody plants, 53% of aboveground biomass, and 62% of LAI of the forest community; total aboveground productivity is around 2.5 Mg ha -1 year -1 with the oak contributing far more than one-half (Matteucci, 1991). The climate in this area is typically temperate-Mediterranean, with a mean annual temperature of 15~ the m a x i m u m temperature in s u m m e r can be greater than 35~ and the m i n i m u m winter temperature can be less than -5~ The total annual rainfall is around 610 mm; its distribution during the year typically peaks in February and in late September; consequently, the dry season lasts from May until early September (Fig. 2). B. The Experimental Setup Within this " m a c c h i a " stand, six open-top chambers (OTCs) have been installed to test the effect of atmospheric CO2 e n r i c h m e n t on clumps of natural, Mediterranean vegetation, starting from early spring 1992. The OTCs are made of an aluminium frame, coated with a transparent sheet of PVC "Cristal," 0.4 m m thick. The size of the OTCs is 4 m in diameter and 6 m in height. The airflow rate of 12,000 m 3 hr -1 inside the OTCs changes the air three to four times per minute in order to maintain the microclimate inside the chambers similar to outside. The air temperature inside the OTCs (Fig. 3) was, on average, 1.1~ higher than outside as measured continuously over a 2-year period with a CR-10 data logger (Campbell-USA). On the contrary, relative humidity was, on average, not significantly affected. The light environment inside the OTCs was affected
Climatediagram of the experimental site, according to Walter and Lieth scheme (Walter, 1985); thin and thick line are, respectively, monthly mean temperature (1 division, 10~ and monthly mean precipitation (1 division, 20 mm).
14.
[ <>o u t s i d e
E
1
10.0
2.0 ~:
1'2 Figure 3 Air temperature variation during the experimental period and differences between the temperatures inside and outside the OTCs. The temperature is calculated as the mean over a 10-day period.
by the PVC cover in relation to the solar elevation angle; at solar angles > 5 0 ~ the r e d u c t i o n of PPFD was b e t w e e n 0 a n d 10%, while at lower solar elevations a 30% o f r e d u c t i o n was observed. T h e CO2 c o n c e n t r a t i o n of the air inside the OTCs is either a m b i e n t or a m b i e n t plus 3 5 0 / x m o l mo1-1. In each O T C the woody vegetation c l u m p (about 30 years of age) is m a d e up, o n the average, by trees a n d by four a n d seven shrubs. T h e m e a n stem d i a m e t e r of is 7 _+ 0.8 (-+SE) cm and canopy height reaches 3.7_+0.35 m. C o r r e s p o n d i n g values for 4 _+ 0.7 a n d 2.4 -+ 0.11, a n d for 3 -+ 0.13 a n d 1.7 _+ 0.19. C. M e t h o d s
1. D u r i n g the third year of C O 2 e x p o s u r e (1994), the r e s p o n s e of n e t assimilation to the CO2 c o n c e n t r a t i o n s (A/Ci curves) was m e a s u r e d in the spring season, w h e n the water supply did n o t limit photosynthesis. C a r b o n dioxide a n d water vapor e x c h a n g e s of leaves
214 (two samples X two OTCs) were measured in the field on each of the three species with the Compact Minicuvette System (CMS, Heinz Walz GmbH, Germany), a portable, temperature and water vapor controlled, open-path, gas-exchange system. The rate of leaf photosynthesis and transpiration, as well as conductance and intercellular concentration of CO2 were calculated according to the equations of von Caemmerer and Farquhar (1981). Air temperature and air vapor pressure deficit inside the cuvette were maintained constant (25~ and VPG ---13 hPa) during measurements of steady state A/Ci curves (Table I). The cuvette was m o u n t e d on a tripod to reach sunlit branchlets; an artificial light source was utilized (HQI-Osram) to obtain about 1200/.tmol m -2 sec -~ of incident PPFD in the PAR region; measurements were taken at steady state only, about every 40 min. The response of A to Ci was fitted by nonlinear regression methods (SY STAT 5.0), with a nonrectangular hyperbola (Eq. 5 of the Table 4.1 reported by Thornley, 1976).
2. Fluorescence emission from the leaves was measured with a modulated fluorometer (PAM 101, Heinz Walz GmbH, Germany) on dark-adapted leaf samples detached from the plants of the three woody species enclosed in the OTCs (three samples x two OTCs). Fluorescence measurements were made over a 2-year period (from 1992 through 1993) in different seasons and at different times of the day to derive m i n i m u m daily values of the photochemical efficiency of PSII (Butler, 1978; Demmig-Adams 1989). For each measurement, leaf disks were collected from the upper part of the plants and were dark adapted in an aluminium container for 15 rain. With this time length the relaxation of the fast c o m p o n e n t of nonphotochemical quenching is reported to occur (Krause and Weis, 1991). Photochemical efficiency of PSII was then estimated by where is the maximum fluorescence intensity emitted from a leaf disk on application of a saturating light pulse,
Species
Treatment
Ta a
m.s.d.
PPFD
m.s.d.
VPG
m.s.d.
Ambient Elevated
25.3 25.0
0.08 0.07
1100 1113
26.0 92.8
14.2 13.4
0.48 0.36
Ambient Elevated
25.0 25.1
0.07 0.12
1137 1063
21.5 24.3
12.3 12.1
0.68 0.43
Ambient Elevated
25.1 25.0
0.01 0.01
1341 1354
8.7 11.5
12.7 14.6
0.12 0.05
"Ta, cuvette air temperature (~ PPFD, incident photon flux density (/zmol quanta m -~ s-l); VPG, vapour pressure gradient (hPa); m.s.d., mean values of the standard deviations of the A/Ci curves.
14.
215
while F0 is the m i n i m u m fluorescence intensity emitted in response to a negligible level of actinic light. As reported in the literature, this ratio is a measure of photoinhibition of photosynthesis (Ogren, 1991). During natural daily courses, generally decreases with minimum values at midday; this reduction can be related to two broad processes: an increase of nonradiative thermal deactivation and an increase of damage and repair of PSII reaction centers (Demmig-Adams and Adams, 1992; Long 1994). 3. End-products of carbon metabolism, pigments concentration, and nitrogen content were analyzed in the leaves of the three woody species in the same period as the gas-exchange measurements. Total nonstructural carbohydrates (TNC) were analyzed from leaf disks (two samples x two OTCs) collected in the morning at 9:00 AM after about 3 hr of daylight. The material was sampled from the u p p e r part of the canopy of trees and of and shrubs, included in the OTCs. Leaf disks were frozen directly in the field u n d e r liquid N2 and later stored at -80~ Sugars were extracted from disks dried for 2 min in a microwave oven and boiled for 30 min in distilled water (Wong, 1990; K6rner and Miglietta, 1994). Soluble sugars were then analyzed spectrophotometrically using the Boehringer M a n n h e i m Biochemicals kit 716260 (Germany). Starch was determined according to H u b e r and Israel (1982) and Rufty and H u b e r (1983). Chlorophyll extraction from the leaf disks (four samples x two OTCs) was carried out by dimethylformamide (DMF) whereas chlorophylls a and b content was d e t e r m i n e d spectrophotometrically, on the resulting solution, according to Moran (1982). Nitrogen content of leaves (two samples x two OTCs) was determined by Kjeldhal digestion followed by distillation in vapor of ammonia and titration. 4. Anatomical observations were carried out, in the s u m m e r 1994, on mature leaves from the southern part of the upper crown collected from one plant per species in each OTC. Two transverse sections per leaf were analyzed microscopically at five different point locations (four samples X two OTCs) for thickness of the epidermis, the palisade layer, and the spongy mesophyll. 5. The Mediterranean " m a c c h i a " species enclosed in the OTCs are considered slow-growing plants. In fact, Bruno (1977) observed a mean annual increment of the basal area of trees, comparable in size to our plants, of about 1.4 c m 2 y r -] (<2-mm-diameter increment). Height i n c r e m e n t of new shoots occurs in early spring and frequently in late s u m m e r or a u t u m n (Giovannini 1992). Therefore, shoot length and shoot structure were measured two times per year in the d o m i n a n t tree species Ten representative branches per tree were
Figure 4 Simplified scheme of a representative branch. The numbers indicate successive length growth increments.
marked in shoots per old shoots measured:
each OTC (10 samples X 2 0 T C s , Fig. 4). All of the lateral branch were counted and numbered. In addition, 10 1-yearper branch were marked and the following parameters were shoot length and diameter, leaf number, and leaf area.
Water availability represents a major environmental constraint in the Mediterranean environment. Together with seasonal variations of air temperature and solar radiation, water stress determines the cyclic pattern of vegetative activity typical of the region. At our experimental site, predawn plant water potential (PWP) varied between - 0 . 3 and - 0 . 5 MPa in winter and spring to - 3 . 0 and - 4 . 0 MPa in summer during the 3 years since the beginning of the study in 1992 (Fig. 5). The oscillation of PWP and
14.
217
Seasonal trend of plant water stress measured with a pressure chamber; the data points are the mean values of the predawn leaf water potential (PWP) of trees growing inside the OTCs at both ambient and elevated [COz], or outside (control) (n = 4).
environmental factors interacts with elevated C O 2 concentration in (Scarascia-Mugnozza 1996), and possibly also in the other species. Net CO2 uptake was markedly increased by elevated CO2 under optimal soil moisture conditions. A lack of downward regulation of A under elevated CO2 in and was evident after 3 years of exposure (A/Ci curves, Table II). Photosynthetic rates (A) were increased in elevated CO2 by 122% in 98% in and only 32% in (Table III). In contrast, showed the strongest reponse in transpiration (E) and stomatal conductance (Gs) to elevated CO2 ( - 4 3 % ) , whereas and were less affected ( - 1 1 % , n.s.). The A/E ratio, the instantaneous transpiration efficiency, increased by 130-150%, in all three species. The intercellular CO2 concentration, Ci, increased significantly in all three species whereas the ratio of internal versus external CO2 concentration, Ci/Ca, did not change significantly. Finally, carboxylation efficiency (CE) significantly decreased in but did not change in the other two woody species. Elevated CO2 caused TNC to increase by 18% in ( P = 0.24), by 51% in (P = 0.26), and by 22% in (P = 0.28, Fig. 6). However, substantial differences in TNC concentration were already evident among the three species at ambient CO2; had the highest TNC concentration which was about twice that measured in both
Table II
A E
G a s - E x c h a n g e P a r a m e t e r s , C a l c u l a t e d from N C i C u r v e s P e r f o r m e d in the Field
Ambient
Elevated
Ambient
Elevated
Ambient
Elevated
11.02
24.50
16.81
22.26
11.99
25.03
2.125
Gs
152.5
ITE
5.15
Ci Ci/Ca CE
204 0.584 0.07308
1.900
3.750
135.0
2.150
305.0
12.93
182.5
4.49
435 0.612 0.07953
3.025 233.8
11.19
244 0.698 0.09015
482 0.679 0.06452
4.17 249 0.712 0.07181
2.600
Units /xmol CO2 m-2 sec-1 m m o l H20 m-2 see-1
187.5
m m o l Hz0 m-2 see-1 9.65 /xmol C O J m m o l H20 491 /xmol mo1-1 0.692 Adimensional 0.06405 /zmol C O J /zmol mo1-1
Data in the table, for each treatment, are the means of values measured at the respective growing [COs].
and Most of the TNC was composed of sucrose in all the species and treatments. Glucose, fructose, and starch were less abundant. The response of photochemical efficiency of photosystem II to CO2 treatment again showed different trends in the three species and over the 2 study years (Fig. 7; Table IV). Whereas showed consistently lower, but not always significant, values under elevated CO2, no clear effects were found in the two other species. Furthermore, the effect of elevated CO2 was amplified by summer drought. At the peak of summer d r o u g h t in August 1992 and July 1993, the reduction of in was about 20% greater than that measured at other times. However, no
test A E Gs ITE Ci Ci/Ca CE
+ 122% -11% -11% +151% + 113% +5% +9%
P < 0.01 n.s. n.s. P < 0.01 P < 0.01 n.s. n.s.
test +32% -43% -40% +149% +97% -3% -28%
P < P < P= P < P < n.s. P <
0.01 0.01 0.06 0.05 0.05 0.05
test + 98% -14% -20% +131% +97% -3% -11%
P < P < n.s. P < P < n.s. n.s.
0.01 0.05 0.01 0.05
The percentages represent the difference between measurements conducted at 710 and 350 ppm of [CO2] on plants grown at elevated and ambient COs concentrations, respectively, relative to the ambient value. The probability of the null hypothesis (zero difference between the two populations) was tested with a t test for two populations with unequal variance (n = 4). a
14.
219
Figure 6
Total nonstructural carbohydrates (T.N.C.) content of leaves of trees and and shrubs, included in the OTCs, as percentage of dry weight. The material was sampled from the u p p e r part of the canopy in the morning at 9:00 AM after about 3 hours of daylight. Error bars are standard errors of means (n = 4).
Minimum daily values of PSII photochemical efficiency (Fv/Fm) measured, in different periods of the years, on leaves of the three woody species enclosed in the OTCs at ambient [CO2]. Error bars are standard errors (n = 6).
220
Giuseppe E. Scarascia-Mugnozza et al.
Q. ilex
t test
P. lentis,
t test
P. angus,
t test
J u n 92 A u g 92
-6.5% -27.5%
P < 0.01 P < 0.01
-7.4% m
P < 0.10
-7.8%
P < 0.05
Nov 92 F e b 93
-1.8% -4.9%
n.s. n.s.
+4.8% --
P < 0.01
+0.1% m
n.s.
Minimum
Fv/F=
A p r 93
-2.8%
P < 0.05
+7.5%
P < 0.10
+0.8%
n.s.
J u n 93 J u l 93
-4.9% -19%
n.s. P < 0.01
~ -7.2%
n.s.
-0.1%
n.s.
a The probability of the null hypothesis (zero difference between the two populations) was tested with a t test for two populations with unequal variance (n = 6).
significant effects were observed in the two shrub species during summer 1993. In the third year, 1994, climate was more humid and the seasonal change in water stress was reduced compared to previous summers. Fv/T'm u n d e r elevated CO2 increased by 22% (P < 0.01) during the summer period in decreased in by ~10% (P < 0.05), but showed no change in Elevated CO2 resulted in a decrease in nitrogen concentration (per dry weight) at the leaf level in ( - 1 7 % , P = 0.07), an insignificant decrease in P = 0.41), and an insignificant increase in (+10%, P = 0.43). These variations among species disappeared when nitrogen concentration was expressed on a projected leaf area basis (Fig. 8). All species showed clear increases in photosynthetic nitrogen use efficiency with the maximum increase observed in (+122%). It should be noted that in high CO2 showed no reduction in rubisco activity (data not shown), no decrease in protein concentration, and no change in the activation state. However, a statistically significant 35% reduction in total chlorophyll concentration was found (Fig. 9, Table V). A similar decrease (25%) was observed in whereas total chlorophyll concentration did not change in Effects on the chlorophyll a / b ratio were not consistent, increasing at elevated CO2 in Q. decreasing significantly in and remaining u n c h a n g e d in Leaf anatomy was differentially affected by CO2 treatments in the three species. leaves did not increase the overall thickness of their lamina, but the thickness of the upper epidermis significantly increased (10%, P < 0.05). On the other hand, and leaves became thicker by 22% (P < 0.05) and 24% (P < 0.05), respectively, with all leaf tissues contributing similarly (Kuzminsky and De Angelis, 1995).
14.
Total nitrogen content of leaves, expressed on a leaf area basis, of the three woody species enclosed in the OTCs at ambient and elevated [CO2]. Error bars are standard errors (n= 4).
A detailed analysis of the effects of elevated C O 2 o n shoot growth was c o n d u c t e d only for but we are presently doing so for the o t h e r two species. T h e average size of all the 1-year-old shoots p r o d u c e d by the 10 representative b r a n c h e s per tree increased in response to CO2 e n r i c h m e n t . Shoot length, shoot diameter, and n u m b e r of leaves per shoot increased significantly by 200, 38, and 36%, respectively (Table VI). Mean total leaf area per branch, on the other hand, was slightly but not significantly, reduced. T h e same was true for the total n u m b e r of shoots per branch. T h e overall result of these changes led to a slight reduction in the total leaf area per b r a n c h and to a large increase in cumulative shoot length a n d sapwood area per branch. However, these variations were not statistically significant. O n the other hand, cumulative shoot volume per branch, as well as sapwood and volume per unit leaf area, were significantly greater u n d e r elevated COz. T h e increase of total shoot volume per b r a n c h u n d e r elevated COz was largely due to the a u t u m n flush.
T h e physiological adjustments to e n v i r o n m e n t a l change, like the increase of CO2 concentration, are often expected to be in a direction that improves fitness. But, as r e p o r t e d by G u n d e r s o n a n d Wullschleger (1994), " w h e t h e r or not photosynthetic acclimation to elevated CO2 results in improved
Total c h l o r o p h y l l c o n t e n t (a) o n a leaf area basis a n d chla/chlb ratio (b) o f the study plants, at a m b i e n t a n d elevated [CO2]. E r r o r bars are s t a n d a r d e r r o r s (n = 8).
test Chl Chl Chl Chl
a b a+ b a/b
-33.2% -41.6% -35.4% +15.7%
P P P P
< < < <
0.01 0.01 0.01 0.01
test -27.1% -18.5% -25.0% -10.9%
P P P P
< < < <
0.01 0.05 0.01 0.01
test -1.3% +0.6% -0.9% -1.6%
n.s. n.s. n.s. n.s.
a The probability of the null hypothesis (zero difference between the two populations) was tested with a t test for two populations with unequal variance (n = 8).
14.
Branch average No. o f n e w s h o o t s No. o f n e w leaves M e a n l e a f size ( c m 2) L e a f a r e a ( c m 2) L e n g t h (cm) S a p w o o d a r e a ( c m 2) V o l u m e ( c m 3) S a p . / l e a f a r e a ( c m Z / m 2) V o l . / l e a f a r e a ( c m 3 / m 2) Shoot average L e n g t h (cm) Diameter (mm) No. o f leaves L e a f a r e a ( c m 2)
Ambient
Elevated
Diff. %
t test
30.9 152 3.0 404.2 79.9 0.729 1.91 17.05 41.95
26.4 173 2.0 329.5 132.8 1.175 6.15 36.15 189.55
- 15% + 14% -34% - 18% +66% +61% +222% + 112% +352%
n.s. n.s. P < n.s. n.s. n.s. P < P < P <
3.1 0.85 4.7 13.89
9.5 1.17 6.4 12.33
+206% + 38% +36% - 11%
P < 0.01 P < 0.01 P < 0.01 n.s.
0.01
0.05 0.01 0.01
a Data refer to the annual production of new shoots on a per branch basis. The probability of the null hypothesis (zero difference between the two populations) was tested with a t test for two populations with unequal variance (n = 20).
fitness remains to be seen." Recent experiments on plant responses to elevated CO2 concentrations lead, in fact, to contrasting conclusions. Gaudill6re and Mousseau (1989) and Jarvis (1989) detected a progressive decrease over time of the photosynthetic rate of young trees, growing in containers under elevated [CO2]. Contrary to these results, Wullschleger (1992) and Gunderson (1993) were able to stimulate net photosynthetic rate of yellow-poplar and white oak seedlings, rooted in the ground, by 12-144% over a 3-year study period utilizing a CO2-enriched atmosphere by 150 and 300 /~mol mo1-1. Similar conclusions were also drawn from experiments on L. trees, growing in the field, by Idso and Kimball (1991). Different processes are reported to be involved in the acclimation of photosynthesis in trees: at the biochemical level, changes in the activity of rubisco a n d / o r in the light harvesting complex were found; at the leaf and canopy levels, changes in leaf size, thickness, and longevity and changes in canopy architecture were described; also at the stomatal level, changes in density and size were found, although most data are from herbariums rather than from CO2 enrichment experiments (for reviews see Ceulemans and Mousseau, 1994; Gunderson and Wullschleger, 1994). In this study, lack of down-regulation of photosynthesis was shown by A / Ci response curves of oak trees and shrubs of On the other hand, had a different response to elevated CO2, with a marked
224 downward adjustment of carboxylation efficiency. Differences among the first two woody species and also evident for such parameters as stomatal conductance and transpiration. Obviously, these divergent responses produced a similar result when net assimilation was expressed in relationship to transpiration, the A / E ratio. Instantaneous transpiration efficiency (ITE), as well as the Ci/Ca ratio, did not vary significantly among species even though ITE significantly increased in response to elevated CO2, as has been frequently reported (cf. review by Eamus, 1991). The similarity of ITE for the three woody species is in agreement with results of 13C analysis conducted by Valentini (1992) on the same species and site. The marked increase of the water use efficiency of plants exposed to CO2 enrichment is a well-known p h e n o m e n o n (Eamus, 1991) and is expected to significantly improve the water balance of plant communities in a double CO2 scenario. However, it is still debated whether transpiration of these communities will definitely be reduced because of a decrease of stomatal conductance or will remain unchanged because of a parallel increase of foliage production (Mooney and Koch, 1994). No definite information on changes of trees leaf area was obtained during the course of our experiment. However, if the preliminary information on leaf area per branch is confirmed in the years to come, a marked improvement of stand water balance is to be expected. The three Mediterranean woody species under study have distinct strategies to cope with water limitation, as shown by the analysis of deuterium discrimination (Valentini 1992). depends more on rainwater than and shows a behavior that can be generally defined as "water spender" (Table II). will probably respond to elevated CO2 by increased carbon uptake whereas would reduce its stomatal conductance and, therefore, its water consumption. a shrub species that can also reach a tree size, has a behavior intermediate between the two species mentioned previously, but possibly closer to the These differential responses would favor driving succession from "macchia" toward more stable, high forest communities dominated by Chaves and Pereira (1992) and Guehl (1994) have underlined the need for more studies on the interactions between water stress and elevated CO2. As suggested by these authors, a possible interaction of elevated CO2, water stress, and high temperature could result in an increased susceptibility to photoinhibition. Our results seem to confirm this hypothesis; u n d e r elevated CO2 conditions, daily minimum reached lower values than in ambient CO2, during the hot and dry summer months. This trend was, however, more evident in than in the other woody species, possibly as a result of the dominant, fully sun-exposed position of the oak trees within the "macchia" canopy.
14.
225
Photoinhibition is related to a variety of different processes (Long 1994). It consists in an overall reduction of the quantum use efficiency and can be interpreted as an acclimation response to elevated CO2, as indicated by the down-regulation of carboxylation efficiency in The reduction of total chlorophyll concentration with elevated CO2, a p h e n o m e n o n also observed by others (e.g., Wullschleger 1992; Houpis 1988; Oberbauer 1985), may be somehow related to those two acclimation responses. Evidence for this is suggested by the fact that reductions in chlorophyll concentration occurred only in and the species exhibiting the other acclimatory responses, but not in the species in which no acclimation in either carboxylation efficiency or in photoinhibition were observed. Reductions in the total chlorophyll concentration observed in and were also the result of different degrees of reduction in chlorophyll a and chlorphyll b. The resultant increases in Chl a / b ratio in and decreases in suggest differential photosynthetic response mechanisms between species. The more p r o n o u n c e d reduction of Chl b concentration in may be related to an adjustment of the antenna complex to match the reduced photochemical efficiency of photosystem II. The relatively greater decrease Chl a concentration in P. points to an adjustment of the n u m b e r of "photosynthetic units" in relation to down-regulation of carboxylation efficiency. Our data also suggest that specific species level shifts in chlorophylls u n d e r elevated CO2 may correspond to their respective ecologies. The excess carbon produced by greater photosynthesis under elevated CO2 can be accumulated as carbohydrates in leaves and other organs as a result of either reduced sink activity (e.g., Stitt, 1991), or limited phloem loading capacity (K6rner 1995). An increase in TNC concentration in plants grown under CO2 enrichment has been observed in many species (cf. Strain and Cure, 1985) including herbaceous plants (e.g., Poorter 1992), tree saplings (e.g., Wullschleger 1992), and in tropical species (K6rner and Arnone, 1992). K6rner and Miglietta (1994) report TNC increases of 40-50% in a herbaceous community growing near natural CO2 vents, compared to concentrations measured in the same species further away from the vents, and almost a 100% increase in leafTNC concentration in the deciduous oak species However, these investigators observed no differences in TNC concentration in leaves of the climax tree species growing close to or far from the CO2 vent. This finding agrees well with our results in as well as with the general trend toward relatively small ( + 15- + 50%) CO2-induced differences in leaf TNC concentrations in the other two woody species in our study. The occurrence of the greatest increase in leaf TNC concentration in the species which did
not acclimate either in terms of carboxylation or as photoinhibition (P. corresponds well with the hypothesis. Thus it seems that some Mediterranean evergreen woody plants may undergo only minor changes in tissue quality when exposed to CO2 enrichment, whereas others (perhaps fewer species) may show stronger decreases in leaf tissue quality (i.e., lower leaf N concentration and greater TNC concentrations). A possible ecological consequence of increased leaf TNC concentration could be compensatory feeding by herbivores on the plants of the Mediterranean plant community (cf. Lincoln 1993). The interaction between carbon metabolism and biomass partitioning in woody plants is also regulated by nutrient availability (Luo 1994). In the nutrient-limited Mediterranean environment, a CO2 increase may in part relieve this limitation by a greater efficiency of N utilization, especially in tree species, characterized by a large proportion of their biomass allocated to components with low nutrient cost. This aspect will deserve much more attention in future works. Finally, the effect of CO2 exposure was analyzed in terms of shoots and branch growth in trees. Contrary to the conclusion drawn by K6rner and Miglietta (1994) for oak trees naturally exposed to CO2-enriched air, we found a significant increase in shoot volume (length and diameter) of in our study. Even more relevant was the increase in shoot volume and sapwood per unit of leaf area. These results are similar to many observations conducted in potted and planted seedlings and saplings (Radoglou and Jarvis, 1990; Eamus and Jarvis, 1989; Bazzaz 1990). In our study, however, other important compartments of tree biomass such as the root component also need to be investigated in detail before we are able to fully understand the effects of CO2 on C storage by trees and forests.
Given the relevance of the Mediterranean woodland communities from an economical and environmental perspective, it is critical to be able to predict possible effects of global change on these ecosystems and to eventually adopt adequate mitigation strategies. In Mediterranean-type ecosystems, the two main factors limiting primary productivity are water and nutrient availability. These limitations are likely to moderate the responses of plants and communities to global change, particularly to increases in atmospheric CO2 and biosphere warming. After a 3-year period, the community responses to elevated CO2 seem to indicate a wide range of acclimation processes by tree and shrub species in relation to their different ecological strategies. The arboreal species Q. would be favored because of its ability to increase carbon uptake under
14.
e l e v a t e d C O 2. T h i s w o u l d f a v o r t h e s u c c e s s i o n f r o m " m a c c h i a " t o w a r d m o r e s t a b l e , h i g h f o r e s t c o m m u n i t i e s d o m i n a t e d by trees. H o w e v e r , in a w a r m e r , d o u b l e - C O 2 f u t u r e w o r l d , t h e e v e r g r e e n f o r e s t c o m m u n i t i e s , d o m i n a t e d by o r r e l a t e d t r e e s p e c i e s , m a y also p r e s e n t a d v a n t a g e s to d e c i d u o u s f o r e s t e c o s y s t e m s o n sites w h e r e n u t r i e n t s a r e r e l a t i v e l y l i m i t i n g . T h i s r e s p o n s e c o u l d b e h y p o t h e s i z e d b e a r i n g in m i n d : (i) t h e m o r e o p p o r t u n i s t i c b e h a v i o r o f e v e r g r e e n t r e e s w h i c h a r e also a b l e to u t i l i z e t h e m i l d w i n t e r s f o r c a r b o n a s s i m i l a t i o n , a n d (ii) t h e i r g r e a t e r n u t r i e n t u t i l i z a t i o n e f f i c i e n c y o f b i o m a s s p r o d u c t i o n . T h i s t r e n d c o u l d also b e s t r e n g t h e n e d by a p o s s i b l e r e d u c t i o n in p r e c i p i t a t i o n at M e d i t e r r a n e a n l a t i t u d e s a n d a c o n s e q u e n t i n c r e a s e o f t h e e c o s y s t e m w a t e r stress.
The skillful technical assistance of Tullio Oro, Matilde Tamantini, and Sandro Federici and the support of National Electric Board (E.N.E.L.) and of Autostrade Company are gratefully acknowledged. Research was performed within the EC Environment Programme (1991-1994) as Contract EVSV-CT92-0127 (coordinated by Professor P. Jarvis, University of Edinburgh, Scotland).
Bazzaz, F. A. (1993). Use of plant growth analysis in global change studies: Modules, individuals, and population. "Design and Execution of Experiments on CO2 Enrichment" (E. D. Schulze and H. A. Mooney, eds.), pp. 53-68. Commission of the European Communities EUR 15110 EN. Bazzaz, F. A., Coleman, J. S., and Morse, S. R. (1990). Growth responses of seven major cooccurring tree species of the northeastern United States to elevated CO2. 20, 1479-1484. Bruno, F., Gratani, L., and Manes, F. (1977). Primi dati sulla biomassa e produttivit~t della lecceta di Castelporziano (Roma): Biomassa e produzione di 35/ 36, 109-118. Bunce, J. A. (1993). Effects of doubled atmospheric carbon dioxide concentration on the responses of assimilation and conductance to humidity. 16, 189-197. Butler, W. L. (1978). Energy distribution in the photochemical apparatus of photosynthesis. 29, 345-378. Ceulemans, R., and Mousseau, M. (1994). Tansley review no. 71: Effects of elevated atmospheric COz on woody plants. 127, 425-446. Chaves, M. M., and Pereira, J. S. (1992). Water stress, CO2, and climate change. 253(43), 1131-1139. Cipollini, M. L., Drake, B. G., and Whigham, D. (1993). Effects of elevated CO2 on growth and carbon/nutrient balance in the deciduous woody shrub (L.) Blume (Lauraceae). 96, 339-346. Cure,J. D., and Acock, B. (1986). Crop responses to CO2 doubling: A literature survey. 38, 127-145.
228 Debano, L. F., and Conrad, C. E. (1978). The effect of fire on nutrients in a chaparral ecosystem. 59, 489-497. Demmig-Adams, B., and Adams, W. W. III (1992). Photoprotection and other responses of plants to high light stress. 43, 599-626. Demmig-Adams, B., Adams, W. W. III, Winter, K., Meyer, A., Schreiber, U., Pereira, J. S., Krfiger, A., Czygan, F. C., and Lange, O. L. (1989). Photochemical efficiency of photosystem II, photon yield of 02 evolution, photosynthetic capacity, and carotenoid composition during the midday depression of net CO2 uptake in growing in Portugal. 177, 377-387. Drake, B. G., Ledley, P. W., Arp, w.J., Nassiry, D., and Curtis, P. S. (1989). An open top chamber for field studies of elevated atmospheric CO2 concentration on salt marsh vegetation. 3, 363-371. Eamus, D. (1991). The interaction of rising CO2 and temperatures with water use efficiency. 14, 843-852. Eamus, D., and Jarvis, P. G. (1989). The direct effects of increase in the global atmospheric CO2 concentration on natural and commercial temperate trees and forests. 19, 2-55. Gaudill~re, J. P., and Mousseau, M. (1989). Short term effect of CO2 enrichment on leaf development and gas exchange of young poplars cv I 214). 10(1), 95-105. Giovannini, G., Perulli, D., Piussi, P., and Salbitano, F. (1992). Ecology of vegetative regeneration after coppicing in macchia stands in central Italy. 99/100, 331-343. Guehl, J. M., Picon, C., Aussenac, G., and Gross, P. (1994). Interactive effects of elevated CO2 and soil drought on growth and transpiration efficiency and its determinants in two European forest tree species. 14, 707-724. Gunderson, C. A., and Wullschleger, S. D. (1994). Photosynthetic acclimation in trees to rising atmospheric CO2: A broader perspective. 39, 369-388. Gunderson, C. A., Norby, R.J., and Wullschleger, S. D. (1993). Foliar gas exchange responses of two deciduous hardwoods during 3 years of growth in elevated CO2: No loss of photosynthetic enhancement. 16, 797-807. Houpis, J. L.J., Surano, g. A., Cowles, S., and Shinn,J. H. (1988). Chlorophyll and carotenoid concentrations in two varieties of seedlings subjected to long-term elevated carbon dioxide. 4, 187-193. Huber, S. C., and Israel, D. W. (1982). Biochemical basis for partitioning of photosynthetically fixed carbon between starch and sucrose in soybean ( Merr.) leaves. 69, 691-696. Idso, S. B. (1991). A general relationship between CO2-induced increase in net photosynthesis and concomitant reductions in stomatal conductance. 31(4), 381-383. Idso, S. B., and Kimball, B. A. (1991). Downward regulation of photosynthesis and growth at high CO2 levels. No evidence for either phenomenon in three-year study of sour orange trees. 96, 990-992. Jarvis, P. G. (1989). Atmospheric carbon dioxide and forest. 324, 369-392. K6rner, Ch., and Arnone, J. A., III (1992). Response to elevated carbon dioxide in artificial tropical ecosystems. 257, 1672-1675. K6rner, Ch., and Miglietta, F. (1994). Long-term effects of naturally elevated CO2 on Mediterranean grassland and forest trees. 99, 343-351. K6rner, Ch., Pel~ez-Riedl, S., and van Bel, A . J . E . 1995. CO2 responsiveness of plants: A possible link to phloem loading. 18, 595-600. Krause, G. H., and Weis, E. (1991). Chlorophyll fluorescence and photosynthesis: The basics. 42, 313-349.
14.
229
Kuzminsky, E., and De Angelis, P. (1995). Effects of an elevated atmospheric [CO2] on leaf anatomy of three Mediterranean woody plants. 27(3), 20-22. Lincoln, D. A., Fajer, E. D., and Johnson, R. H. (1993). Plant-insect herbivore interactions in elevated CO2 environments. 8, 64-68. Long, S. P., Humpries, S., and Falkowski, P. G. (1994). Photoinhibition of photosynthesis in nature. 45, 633-662. Luo Y., Field, C. B., and Mooney, H. A. (1994). Predicting responses of photosynthesis and root fraction to elevated CO2: Interactions among carbon, nitrogen, and growth. 17, 1195-1204. Matteucci, G. (1991). "Struttura, Fenologia e Fisiologia del Carbonio Nelle Fasce di Vegetazione Dunale di Montalto di Castro (VT)." Thesis of Laurea degree, University of Tuscia, Faculty of Agriculture, Viterbo, Italy. Mooney, H. A., and Koch, G. W. (1994). The impact of rising COs concentrations on the terrestrial biosphere. 23(1), 74-76. Moran, R. (1982). Formulae for determination of chlorophylls pigments extracted with dimethylformamide. 69, 1376-1381. Naveh, Z. (1974). Effects of fire in the Mediterranean region. "Fire and Ecosystems" (T. T. Kozwloski and C. E. Ahlgren, eds.), pp. 401-481. Academic Press, New York. Oberbauer, S. F., Strain, B. R., and Fetcher, N. (1985). Effect of COzenrichment on seedling physiology and growth of two tropical tree species. 65, 352-356. Oechel, W. C., and Riechers, G. H. (1986). Impacts of increasing COs on natural vegetation, particularly the tundra. "Proceedings of Climate-Vegetation Workshop, 27-29 January 1986," pp., 36-42. NASA/GSFC, Greenbelt, MD. Ogren, E. (1991). Prediction of photoinhibition of photosynthesis from measurements of fluorescence quenching components. 184, 539-544. Poorter, H., Gifford, R. M., Kriedemann, P. E., Wong, S. C. (1992). A quantitative analysis of dark respiration and carbon content as factors in the growth response of plants to elevated COs. 40, 501-513. Radoglou, K. M., and Jarvis, P. G. (1990). Effects of COs enrichment on four poplar clones. I. Growth and leaf anatomy. 65, 617-626. Rambal, S. (1993). The differential role of mechanisms for drought resistance in a Mediterranen evergreen shrub: A simulation approach. 16, 35-44. Rhizopoulou, S., and Mitrakos, K. (1990). Water relations of evergreen sclerophylls. I. Seasonal changes in the water relations of eleven species from the same environment. 65, 171-178. Rhizopoulou, S., Angelopulos, K., and Mitrakos, K. (1989). Seasonal variations of accumulated ions, soluble sugars and solute potential in the expressed sap from leaves of evergreen sclerophyll species. 10, 311-319. Romane, F., and Terradas, J. (1992). "Quercus L. Ecosystems: Function, Dynamics, and Management. Advances in Vegetation Science 13." Kluwer Academic, Dordecht/Boston/London. Rufty, T. W., Jr., and Huber, S. C. (1983). Changes in starch formation and activities of sucrose phosphate synthase and cytoplasmic fructose-l,6-bisphosphate in response to source-sink alterations. 72, 474-480. Sage, R. F., Sharkey, T. D., and Seemann, J. R. (1989). Acclimation of photosynthesis to elevated COs in five C3 species. 89, 590-596. Scarascia-Mugnozza, G., De Angelis, P., Matteucci, G., and Valentini, R. (1996). Long-term exposure to elevated [CO2] in a natural L. community: Net photosynthesis and photochemical efficiency of PSII at different levels of water stress. in press. Sengupta, U. K., and Sharma, A. (1993). Carbon dioxide effects on photosynthesis and plant growth. "Photosynthesis: Photoreactions to Plant Productivity" (Y. P. Abrol, P. Mohanty, and Govindjee, eds.), pp. 479-508. Kluwer Academic, The Netherlands.
230 Specht, R. L. (1973). Structure and functional response of ecosystems in the Mediterranean climate of Australia. "Mediterranean type ecosystems. Origin and structure" (F. DiCastri and H. A. Mooney, eds.), pp. 113-120. Chapman & Hall, London. Stitt, M. (1991). Rising CO2 levels and their potential significance for carbon flow in photosynthetic cells. 14, 741-762. Strain, B. R., and Cure, J. D. (1985). "Direct Effects of Increasing Carbon Dioxide on Vegetation. Carbon Dioxide Research, State of the Art," Publication No. ER-0238. U.S. Department of Energy, Washington, DC. Terradas, J., and Sav6, R. (1992). The influence of summer and winter water relationships on the distribution of 99/100, 137-145. Thornley, J. H. M. (1976). "Mathematical Models in Plant Physiology." Academic Press, London. Valentini, R., Scarascia Mugnozza, G. E., and Ehleringer, J. R. (1992). Hydrogen and carbon 6, isotope ratios of selected species of a Mediterranean macchia ecosystem. 627-631. von Caemmerer, S., and Farquhar, G. D. (1981). Some relationships between the biochemistry of photosynthesis and the gas exchange of leaves. 153, 376-387. Vu, J. V. C., Allen, L. H., Jr., and Bowes, G. (1989). Leaf ultrastructure, carbohydrates, and protein of soybeans grown under CO2 enrichment. 29, 141-147. Walter, H. (1985). "Vegetation of the Earth and Ecological Systems of the GeoBiosphere," 3rd English ed. Springer-Verlag, Berlin. Wong, S. C. (1990). Elevated atmospheric partial pressure of CO2 and plant growth. Res. 23, 171-180. Wullschleger, S. D., Norby, R.J., and Hendrix, D. L. (1992). Carbon exchange rates, chlorophyll content and carbohydrate status of two forest tree species exposed to carbon dioxide enrichment. 10, 21-31.
1 Modification of Fire Disturbance by Elevated C02
Ecological disturbances are episodic events that affect the structure and function of plant communities, usually by killing individuals and opening patches for colonization (Sousa, 1984). Disturbance not only has an immediate impact on species composition and material flow in communities, but by altering site characteristics, it influences community dynamics long after the event. All ecological communities are influenced in some way by disturbance, and in most, multiple disturbances have important roles. To understand plant communities, disturbance regimes need to be characterized. In turn, to understand effects of rising atmospheric CO2 on plant communities, it is important to identify interactions between CO2 variation and specific disturbance regimes. Through effects on plant productivity and water use, atmospheric CO2 enrichment may affect disturbance frequency, magnitude, and resiliency of the vegetation to disruptive events. However, CO2 effects on disturbance are considered infrequently in global change discussions. For example, interactions between rising CO2 and fire are not addressed in recent treatments of global change and fire regimes (Crutzen and Goldammer, 1993; Torn and Fried, 1992), and have been the subject of only a few, brief, speculative summaries (Strain and Bazzaz, 1983; Oechel and Strain, 1985; Ryan 1991; Mayeux 1994). As a consequence, CO2 effects on disturbance regimes may be an important overlooked issue in the global change debate. 231
Copyright 9 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.
In this chapter, I will address possible interactions between atmospheric CO2 enrichment and fire. Of the major disturbances affecting communities, fire may be the most sensitive to CO2 variation because warm, dry conditions favoring fire may also promote high CO2 responsiveness in plants (Hogan 1991; Mooney 1991; Sage 1995). Fire is also widespread, having a defining role in a broad range of plant communities (Kozlowski and Ahlgren, 1974; Sousa, 1984). Thus, modification of fire regimes may be one of the more significant ways rising CO2 could affect future vegetation.
With the possible exceptions of rainforests, tundra, and deserts, fire is a major disturbance in all vegetation regions of the earth, occurring with a return time of less than 300 years (Aber and Melillo, 1991; Trabaud 1993; Goldammer, 1993). In drier regions, fire frequency can be high (return time < 50 years) and communities are dominated by fire-adapted species such as grasses, stem-sprouting shrubs, and pines (Trabaud, 1987). In moist forests, where conditions favorable to fire are largely restricted to infrequent drought episodes, fires occur infrequently (return time >100 years) and are typically intense, setting the community back to an early successional stage (Kilgore, 1981; Sousa, 1984). The primary mechanism by which fire influences composition and threedimensional structure of plant communities is through flame-induced mortality. Grasslands, savannas, chaparrals, and pine woodlands are prominent examples of natural communities maintained by fire mortality of hardwood vegetation. Fire prevention in these communities permits establishment of dense thickets within a few decades of burning (Christensen, 1981). Through selective mortality, fires of intermediate frequency and intensity can enhance species diversity and community heterogeneity; however, extreme fire frequency can reduce diversity by favoring relatively few species well adapted to fire disturbance (Sousa, 1984; Whisenant, 1990). Even within a fire-maintained community, the timing and frequency of burning has important effects on which fire-adapted species persist, often for reasons not directly related to fire mortality. In the tallgrass prairie of North America, for example, frequent burning reduces soil fertility by volatilizing nitrogen, thus favoring nitrogen-use-efficient C4 grasses such as big bluestem over C3 grasses and forbs with lower nitrogen use efficiency (Seastedt 1991; Wedin, 1995). Timing of fire within a growing season has important consequences in mixed communities of C3 and C4 vegetation. In the American tallgrass prairie, spring burning favors summer-active C4 plants by damaging spring-active C3 vegetation. In c o n -
15.
233
trast, summer burning suppresses C 4 vegetation in favor of summer-dormant C~ species (Howe, 1994). While climatic conditions act as a course controller of fire intervals and determine the timing of individual fire events, specific vegetation types can have overriding influence in nonextreme climates by either enhancing or attenuating fire cycles. Fire cycles are enhanced by invasion of rapidly growing, fire-tolerant species which produce fine, highly flammable litter that readily carries fire, even in mesic habitats (D'Antonio and Vitousek, 1992). As a consequence, fire frequency rises substantially and favors dominance of species adapted to frequent fire. Formation of a tallgrass "prairie peninsula" in Illinois and Indiana, for example, occurred as a result of a grass-enhanced fire cycle that allowed grassland to extend hundreds of kilometers into regions where climate favors mesic, hardwood forest (Howe, 1994). Widespread grassland establishment has recently occurred in Hawaii, Central and South America, western North America, and Australia as a result of acceleration of fire cycles caused by invasion of exotic grasses (D'Antonio and Vitousek, 1992). The post-19th century conversion of the shrub-dominated steppes of western North America into annual grasslands dominated by cheatgrass, is one of the spectacular examples of how fire-cycle enhancement can radically alter community characteristics (Fig. 1; Young and Evans, 1978; D'Antonio and Vitousek, 1992). Prior to European invasion, the western steppes of North America were dominated by mixtures of brush and perennial C3 bunchgrasses (e.g., spp., Agroand Billings, 1994). Heavy cattle and sheep grazing between 1860 and 1920 converted vast areas of this region into dense sagebrush scrub (Billings, 1994). Beginning in the late 19th century, an annual C~ grass native to central Asia was introduced into western North America with imported cattle and grain. Spreading from railheads, colonized vast areas of the intermountain West within 30 years, largely as a result oflifestock movement and continued overgrazing (Mack, 1986). Once established, filled spaces between shrubs with a fine, highly flammable litter that readily carried fire, particularly in wetter than average years when productivity peaked (Billings, 1990, 1994; D'Antonio and Vitousek, 1992). Fire-return time fell from decades to less than 10 years, and shrub-dominated landscapes were replaced by annual grasslands (Fig. 1). Even with human attempts to control fire, infestation has shortened fire-return time to less than establishment times for the woody vegetation, perpetuating annual grass dominance at the expense of native vegetation (Young and Evans, 1978). Conversely, establishment of woody vegetation can break fire cycles by preventing rapid accumulation of fine fuels in the understory. In the semideserts of the American Southwest, overgrazing of perennial grasslands has
234
Intermountain steppe vegetation dominated by sagebrush top), infested with growing between sagebrush (middle), and completely dominated by monoculture (bottom). Photographs taken August 8, 1995, along Highway 395, near Alturas, California (41.5~ 120.5~
15.
235
allowed woody shrubs such as mesquite spp.) and creosote bush to dominate large portions of the region (Wright 1976; Archer 1988; Schlesinger 1990). As these shrubs matured, vegetation-free gaps formed between the woody plants. Establishment of annual grasses is often poor in these gaps and, as a result, continuous fuel lanes have not readily formed, thereby interrupting fire cycles and allowing shrub dominance to persist. In wetter regions capable of supporting mesic forest, such as the eastcentral United States, human interruption of fire cycles allows woody vegetation to replace fire-adapted grasses within a few decades (Christensen, 1981). Suppression of grass cover by the overstory removes the source of highly flammable fine fuel, and the litter that does accumulate is less likely to burn because it is shaded and therefore dries slowly (Wilson and Agnew, 1992). Greater moisture content of the litter also allows for rapid decomposition and little long-term accumulation (Christensen, 1987). As a consequence, establishment of hardwood trees or shrubs can prevent fuel accumulation and thereby lengthen fire intervals from a few years to decades or centuries, allowing a range of fire intolerant species to become established and dominate the community.
The importance of a disturbance in a community primarily depends on four factorsmareal extent, frequency, intensity, and the sensitivity of the vegetation to the particular disturbance event (Sousa, 1984). Excluding human action, fire frequency and intensity depend on conditions controlled by both climate and vegetation. Primary climatic variables are precipitation, temperature, humidity, wind, and frequency of ignition by lighting. Vegetation characteristics affecting fire are (1) fuel load and spatial distribution; (2) fuel composition, size, and energy content; and (3) fuel water content (Pyne, 1984). Through its effect on photosynthesis and transpiration, atmospheric CO2 enrichment can modify most of the vegetationdependent parameters affecting fire disturbance, particularly fuel load and horizontal continuity, fuel chemistry, and moisture content (Fig. 2). While enhancement of photosynthesis by elevated CO2 varies substantially between species and growth conditions, most field-grown C3 plants respond to a doubling of atmospheric CO2 during growth with a sustained 20-100% enhancement of photosynthesis (Drake and Leadley, 1991; Gunderson and Wullschlager, 1994; Sage, 1994). Warmer temperature stimulates the CO2 responsiveness of photosynthesis, reflecting inhibitory CO2 effects on photorespiration and probable temperature stimulation of sink strength (Hogan 1991; Farrar and Williams, 1991; Sage 1995). In the C~ annual
~ - a n6+~_hape ~ Crowning ]~"
Figure 2 Relationshipsbetween major variables contributing to fire disturbance, modified from Pyne (1984). Hypothesizedresponse of each fire variable to a doubling of atmospheric CO2 content is indicated by a "+" or a "-," with + + + indicating a relative high response (20-40%), + + a moderate response (10-20%), + a low response (1-10%), and- -a moderate inhibitory response.
for example, doubling C O 2 increases net photosynthesis at 24~ by 25%, while at 34~ doubling CO2 increases photosynthesis by about 50% (Sage 1995). The maximum stimulation of biomass productivity in most C3 plants following a doubling of atmospheric CO2 is also between 20 and 100%, with high values occurring in warmer conditions (Cure, 1985; Bazzaz, 1990; Hogan 1991; Hunt 1991). To achieve high productivity responses, plants require nutrient-enriched soils, warmer temperatures, and high light intensity (Bazzaz, 1990). Younger plants, and plants in isolation or low density respond more to elevated CO2 than older plants in crowded conditions (Bazzaz, 1990; Dippery 1995). Plant communities in cold, low nutrient, or dense configurations (for example, reconstructed rainforest, arctic tundra, and British grassland) tend to have low responsiveness to CO2 enrichment (K6rner and Arnone, 1992; Diaz 1993; Oechel 1994). Fire-adapted communities are probably highly responsive to increasing CO2, particularly during the first few seasons after fire. By removing canopy vegetation and accumulated litter, fire increases light availability, temporarily enhances soil nutrient levels, and initially reduces competition for water and nutrients (Blank 1994). Juvenile vegetation and early successional species tend to exhibit high response to a doubling of CO2 with productivity of fire-adapted annual grasses being particularly sensitive (Smith 1987; Bazzaz, 1990; Hunt 1991). For example, the annual grasses and exhibited biomass increases of 40-90% and 41%, respectively, following a doubling of CO2 (Smith 1987; Jackson 1994).
15.
A. Fuel Production and Composition An i m p o r t a n t effect of elevated C O 2 o n fire regimes would likely result from the early stimulation of biomass productivity and the subsequent increase in the rate of fuel accumulation (Fig. 3A). With m o r e rapid fuel production, fuel continuity is established m o r e quickly, allowing m i n i m u m fuel thresholds to be reached sooner after the previous b u r n (Fig. 3B). Thus, within a given fire cycle, proportionally m o r e time would be spent above a m i n i m u m fuel threshold, leading to greater fire frequency if ignition probability remains unchanged. A 50% e n h a n c e m e n t in the density of for example, is estimated to cause a 1.5- to 2-fold increase in fire frequency (Whisenant, 1990). Greater biomass p r o d u c t i o n would also increase fuel loading at any point in time, so in addition to fire danger arising earlier within a fuel cycle, once a fire did occur, flame intensity, fire temperature, and rate of spread would be greater (Figs. 3B, 4). The combination of greater frequency and higher intensity is responsible for substantial modification of vegetation patterns (Christensen, 1985), and can lead to large reductions in biodiversity when both become high (Zedler 1983; Billings, 1990; Whisenant, 1990). Higher fire intensity increases the probability that fire will scorch the overstory or seedbank, and higher fire frequency may kill plants before establishment, allowing relatively few species to d o m i n a t e a site. Elevated CO2 may also increase the energy content of fuel by raising the concentration of r e d u c e d organic compounds. Accumulation of resins, oils, and fats increase heat of combustion because these substances have greater
-d 121 L_ C~
Possible effect of doubling atmospheric CO~ level on the relationship between (A) fuel load (relative to a hypothetical maximum) and the time since last fire and (B) rate of fire spread in a grassland as a function of time since last fire. Relationships modified from Pyne (1984). In panel A, a generalized fuel accumulation scenario is assumed for vegetation with high postfire CO2 responsiveness. In panel B, differences between the curves reflect stimulatory CO2 effects on fuel loading, fuel continuity, and fuel energy content. Arrows in panel B indicate minimum fuel thresholds for fire spread.
Possible effect of doubling atmospheric C O 2 levels on temperature at the soil surface during a fire event. Adopted from Wright (1976) for a perennial grassland. Differences between the lower two curves reflect differences in fuel loading, assuming a 50% e n h a n c e m e n t in standing biomass as a result of doubling CO2. Possible effects of high starch accumulation are shown in the top curve.
energy content per unit mass (Pyne, 1984). Little information exists on the effects of elevated CO2 on tissue flammability; for example, changes in wood chemistry in response to CO2 enrichment are not well known. A few studies have documented enhanced secondary compound synthesis in leaves following CO2 enrichment. Sharkey (1991) reported a doubling of isoprene production in red oak leaves following growth at doubleambient CO2. Phenolic content often increases in hardwood species grown in elevated CO2 (Cipollini 1993; Julkunen-Tiitto 1993; Roth and Lindroth, 1994). For example, in paper birch leaves, tannin concentration increased from 21 to 26% of dry weight (Roth and Lindroth, 1994). However, terpenoid contents appear little affected by rising CO2 (Lincoln, 1993; Roth and Lindroth, 1994). The primary compositional change in leaves following CO2 enrichment is the accumulation of nonstructural carbohydrates, particularly starch (Allen 1988; Farrar and Williams, 1991; Wong, 1990). On average, leaf starch levels appear to increase from less than 10% of dry weight to between 10 and 20% of leaf dry weight. In the extreme, nonstructural carbohydrate and starch levels increase from below 10% to over 40% of leaf dry weight in older leaves grown at high CO2, as has been documented for cotton leaves (Wong, 1990; Thomas and Strain, 1991). High, nonstructural carbohydrate accumulation increases flammability of tissue because it readily undergoes pyrolysis and increases carbon:mineral content in fine fuel. Changes of mineral content
15.
239
between 0.1 and 5% of dry weight have significant effects on combustion intensity and rate of fire spread (Philpot, 1970; Pyne, 1984). Reductions in mineral content of 20% have been measured for a variety of C~ species following CO2 enrichment (Overdieck, 1993), although cases of extreme carbohydrate accumulation could reduce mineral content by as much as 40%. For fuels starting with a 1% mineral content, a 20-40% reduction in mineral content can result in a 5-10% increase in the rate of fuel volatilization (Philpot, 1970). An additional, and possibly more important, effect of elevated carbohydrate and phenolic content in biomass is that litter may decompose more slowly (Lambers, 1993), contributing to more rapid fuel accumulation. Most fire damage occurs as a result of ignition events some distance away (>1 km), emphasizing the importance of areal spread to fire disturbance (Christensen, 1985). CO2-enrichment effects on fuel loading, horizontal continuity, and energy content increase the surface area impacted by an individual fire, because (1) the rate of spread is directly dependent on these variables; (2) fewer fuel gaps would be present to stop or slow the fire; (3) greater flame intensity would increase the ability of fire to traverse large fuel gaps such as rivers, roads, and fire lines; and (4) higher intensity would better enable a fire to persist during unfavorable periods such as high humidity or cool nights (Rothermel, 1983). B. CO2 Effects on Fuel Moisture Content
Fuel moisture affects all components of a fire model (Fig. 2), largely because water in the fuel must be distilled before ignition and pyrolysis become possible (Pyne, 1984). Increasing moisture content raises ignition temperature because heat must be expended in the distillation process (Fig. 5). Higher fuel moisture lengthens the time required for water distillation, and thus the time an ignition event such as a spark must persist before pyrolysis can begin (Wright and Bailey, 1982). Energy consumed in water distillation also reduces energy released during combustion, so that fire intensity, temperature, and rate of spread are reduced as moisture content rises. Fuels with greater than 30% moisture content are generally considered inflammable (Fig. 5; Pyne, 1984). Moisture content of living biomass is typically above 60%, and is therefore considered inflammable until dried by the heat of an approaching fire. Increasing CO2 reduces stomatal conductance and in doing so reduces leaf transpiration at constant temperature and humidity (Morison, 1993; Sage, 1994). Consequences of this response for fuel moisture are difficult to generalize from published CO2 responses because stomatal responses to elevated CO2 are highly variable between species and interact with drought, humidity, light, and temperature (Wong, 1993; Sage and Reid, 1994; Santrucek and Sage, 1996). Furthermore, reductions in transpiration caused by
240
? E :~
+"
/
+
..,
"
""
-
Fuel moisture content Figure 5 Generalized relationship between (A) heat input required for fuel ignition and fuel moisture content and (B) reaction (or combustion) intensity and fuel moisture content (adapted from Pyne, 1984). In panel A, the horizontal arrow and dashed lines indicate a possible effect of elevated CO2 on fuel moisture; the exact magnitude of the CO2 effect will vary with conditions. In panel B, differences between the two curves reflect changes in fuel load, fuel chemistry, and fire intensity as a result of a doubling of atmospheric CO2.
stomatal closure can be attenuated by leaf warming and humidity decline (Morison, 1993). Despite these uncertainties, water content of live biomass u n d e r d r o u g h t or high evaporative d e m a n d is usually increased by exposure to elevated CO2. In the annual grass growing in open-top fumigation chambers in the California coast range, a doubling of atmospheric CO2 reduced stomatal conductance up to 50% relative to ambient CO2 controls (Jackson 1994). Transpiration also declined by as much as 50%, which was associated with 0.2-0.4 MPa higher midday water potential of leaves over much of the growing season. Soil water content was also enhanced, being on average 34% greater in the higher CO2 treatment at the end of the growing season (Jackson 1994). Sionit (1981) reported that a tripling of CO2 alleviated d r o u g h t intensity in wheat, with the reduction in relative water content being 10 percentage points less in plants grown in high relative to ambient CO2 levels. Similarly, soybean, aster, and numerous woody species have been observed to exhibit 0.20.5 MPa greater water potentials during drought (Rogers 1984; Wray and Strain, 1986; Tyree and Alexander, 1993). A 0.5 MPa change in water potential corresponds to a 5-10% change in relative water content above the turgor loss point (Turner, 1987). Water potential in C4 plants also responds to elevated CO2. In the C4 prairie grass double ambient CO2 treatments exhibited 0.1-0.4 MPa greater water potential on 11 of 13 midday sample dates t h r o u g h o u t a growing season (Owensby 1993).
15.
241
Increases in stem and leaf water status reduce flammability of live biomass because more heat and time are required to dry the tissue; however, fire danger is largely a function of dead fuel characteristics (Wright and Bailey, 1982). Effects of rising CO2 on moisture content of dead vegetation likely will be insignificant except where soil moisture can be easily absorbed by surface litter. Moisture content of litter above ground level is largely dependent on relative humidity (Pyne, 1984), which could fall if canopy transpiration rates are reduced by elevated CO2 (Morison, 1993). The more important consequences of CO2-enrichment effects on plant water status would likely be less rapid tissue senescence during drought, and a subsequent maintenance of high live : dead fuel ratios. In fire-prone areas, drought-induced senescence of grass vegetation often creates the finder dry fuel that promotes high fire danger (Christensen, 1981). By slowing exhaustion of soil water reserves, elevated CO2 may delay death of the herbaceous cover and the onset of the fire season. Moreover, elevated CO2 increases the ability of plants to tolerate a given level of soil drought by promoting osmotic adjustment and root growth, and possibly reduced xylem embolism (Rogers 1992; Morse 1993; Tyree and Alexander, 1993). Although experiments are needed to determine exact relationships between elevated CO2 and the timing of drought-induced senescence, past work indicates that onset of severe water stress can be delayed 1-2 weeks by doubling of atmospheric CO2 levels (Wray and Strain, 1986). The significance of delays in drought-induced senescence would probably be most important in areas lacking a prolonged dry season. In seasonally dry regions where fire danger can last 3-5 months, a 1- to 2-week delay in canopy death would have little consequence and would probably be overwhelmed by CO2 effects on fuel accumulation (Fig. 6A). In regions with no regular dry season, 2-week delays in canopy senescence by CO2 enrichment could substantially narrow the period of high fire danger during drought intervals lasting 3-6 weeks (Fig. 6B).
The consequence of C O 2 enrichment for particular fire regimes will vary widely between regions and will be strongly influenced by local climatic conditions. While CO2 effects on fuel loading and moisture content could offset each other and minimize fire-CO2 interactions in some regions, other areas may experience radical alteration of vegetation dynamics if CO2 enrichment causes a switch between contrasting fire-fuel regimes. Plant communities most susceptible to shifts in fire/fuel regimes may be those at the extreme ends of their respective moisture gradients. At the arid extreme, elevated CO2 will likely stimulate production of fire-promoting annuals
oO
L_ 9- -
"--
A
B
(9
~.- Canopy
,
!
T
t
/
.'
-
~
, ...........................
Hypotheticaleffects of C O 2 doubling on the probability of fire disturbance in (A) an annual grassland in a seasonally dry climate and (B) a grassland in a climate with episodic spring and summer precipitation. In panel B, "acceleration of leaf senescence" indicates the period when drought-induced canopy dieback may accelerate, with the left arrow indicating this period for current conditions, and the right arrow indicating this period for high-CO2 conditions.
such as or Accelerated accumulation of fuel could substantially increase fire frequency and intensity, further p r o m o t i n g conversion of shrub and woodland to a n n u a l grassland. Increases in atmospheric CO2 level since 1900 may have already contributed to a n n u a l grassland formation in the i n t e r m o u n t a i n steppes of western N o r t h America (Mayeux 1994), and may currently be allowing and other a n n u a l grasses spp., and spp.) to invade increasingly harsh sites t h r o u g h o u t the arid West. Recent reports note that and B. rubrum have spread into warmer, drier regions of the Nevada desert in the past 40 years (Young and Tipton, 1990; H u n t e r , 1990), a period which c o r r e s p o n d e d to approximately 50/~mol mo1-1 increase in atmospheric CO2. Because h i g h e r CO2 stimulates water use efficiency a n d d r o u g h t tolerance in C~ plants, f u r t h e r expansion of annual grasses into harsh sites may be expected for the next century. O n e c o n s e q u e n c e of B. invasion is a dramatic reduction in the native flora, particularly the desert e p h e m e r a l s which p r o d u c e spectacular spring blooms. Billings (1990) d o c u m e n t e d biotic i m p o v e r i s h m e n t occurring as a result of invasion in western Nevada, a n d Young (1987) described how invasion of into the Desert Q u e e n Valley of Nevada between 1976 a n d 1986 has suppressed Indian rice grass (Oryz0psis f o r m e r c o d o m i n a n t , and has nearly eliminated the once c o m m o n desert
15.
243
candle If atmospheric C O 2 enrichment does allow fire-promoting annuals to invade increasingly xeric locations, then future vegetation change may involve a large expansion of annual grasslands in the floristically diverse semideserts of the American Southwest and northern Mexico, with a consequent reduction in the biodiversity of this region. In contrast, at the mesic end of a fire regime, maintenance of live vegetation during drought intervals may significantly reduce fire probability. Coupled with faster establishment of woody vegetation in response to rising atmospheric CO2, disruption of grass fire/fuel cycles could more readily occur, allowing for expansion of woody species into grassland and savanna vegetation. In central North America, interaction between elevated CO2 and fire disturbance may promote westward expansion of hardwood vegetation into the tallgrass prairie region, while in the West, further expansion of annual grasslands may occur in the steppe, chaparral, and southwestern desert vegetation zones.
The combination of global warming and reduced temperate zone precipitation is widely predicted to enhance future fire danger. Because of direct CO2 effects on fuel production, fire models focusing on climate interactions only may substantially underestimate future fire dynamics, particularly in warm, dry climates where CO2 effects are predicted to be pronounced (Mooney 1991). As discussed here, interactions between elevated CO2 and fuel production, fuel continuity, and live:dead fuel ratios could be important modifiers of fire impacts that need to be experimentally evaluated. Because fire disturbance is a highly stochastic, regionally variable phenomenon, predicting future fire-CO2 interactions will require a large data base collected from many regionally focused studies. Realistically, financial resources will limit the number of new projects emphasizing fireCO2 interactions; however, acquisition of relevant data could be obtained through minor modifications of existing research projects addressing plant responses to rising CO2. For example, most current high CO2 research focuses on information pertinent to production assessments, as shown in the left column of Table I. With minor additional effort and cost, data useful for fire assessments could be acquired during regular sampling intervals (Table I, right column). Instead of simply determining total biomass produced, biomass could be sorted into fuel categories (live/dead, size, position in the canopy), while moisture status could be assessed in terms of relative water content in addition to water potential. Sampled biomass could be analyzed for energy content, mineral content, and combustion intensity in addition to the usual carbohydrate and nitrogen analy-
244
Typical measurements in a high CO2 study
Measurements needed to evaluate a CO2 effect on fire intensity
Photosynthesis rate Respiration Growth, yield Canopy area Vegetation density
Fuel accumulation rate Fuel load Dead:live biomass Fuel size and compaction Fuel breaks
Leaf conductance Transpiration rate Water potential
Transpiration rate Moisture content of litter Moisture content of live biomass Moisture content of soil
Vapor pressure deficit Leaf, air temperature
Humidity at litter layer Temperature of litter layer
Allocation Carbon:nitrogen Secondary compound content
Three-dimensional fuel distribution Carbon, resin, mineral content Fuel energy content
sis. Once collected, this data could be entered into fire models to produce comprehensive evaluations of the probable significance of CO2-fire interactions in specific vegetation types. Vegetation plots produced in high COs could be burned near the termination of experiments, allowing for actual assessments of fire behavior, mortality patterns, and subsequent recovery. From this information, long-term surveys could be designed to detect interactions between rising CO2 and fire disturbance as they begin to occur, rather than when they are fully manifested across a landscape.
I have extrapolated from known responses of plants to rising CO2 to speculate on some possible consequences of elevated CO2 on fire disturbance in natural ecosystems. The following appears clear. First, fire is one of the major disturbances for communities growing in all but the wettest and driest locations on earth. Second, the warm, dry conditions promoting fire may also promote high CO2 responsivess of growth, water balance, and live:dead fuel ratios. Third, by promoting vegetation change, rising COs may alter fire/fuel regimes and create a self-perpetuating system based on altered fire probabilities. The most important COs effects may occur where elevated COs promotes a switch between contrasting fire/fuel cycles (for
15.
245
e x a m p l e , if a n n u a l g r a s s e s a r e a b l e to b e c o m e e s t a b l i s h e d in t h o r n s c r u b o r if w o o d y v e g e t a t i o n c a n b e c o m e e s t a b l i s h e d in C 4 - d o m i n a t e d g r a s s l a n d ) . Finally, in m o s t r e g i o n s , t h e e n h a n c e m e n t o f f u e l l o a d a p p e a r s to b e t h e m o r e critical c o n s e q u e n c e o f e l e v a t e d CO2 a n d fire d i s t u r b a n c e . A m a j o r issue is t h e d e g r e e to w h i c h r e s p o n s e s in f u e l m o i s t u r e c o n t e n t a n d l i v e : d e a d f u e l r a t i o s offset CO2 effects o n f u e l a c c u m u l a t i o n a n d tissue flammability. Future research should focus on these relationships, because t h e y m a y b e critical in d e t e r m i n i n g d i s t r i b u t i o n o f c o m m u n i t i e s at t h e wetter, more productive edge of fire-adapted vegetation zones.
I thank Dr. Wayne Polley, Dr. Stan Smith, and Dr. James Young for helpful reviews of the manuscript.
Aber, J. D., and Melillo, J. M. (1991). "Terrestrial Ecosystems." Saunders, Philadelphia. Allen, L. H. Jr., Vu, J. c. v., Valle, R. R., Boote, K.J., and Jones, P. H. D. (1988). Nocturnal carbohydrates and nitrogen of soybean grown under carbon dioxide enrichment. 28, 84-94. Archer, S., Scifres, C., Bassham, C. R., and Maggio, R. (1988). Autogenic succession in a subtropical savanna: Conversion of grassland to thorn woodland. 52, 111-127. Bazzaz, F. A. (1990). The response of natural ecosystems to the rising CO2 levels. 21, 167-196. Billings, W. D. (1990). a biotic cause of ecosystem impoverishment in the Great Basin. "The Earth in Transition: Patterns and Processes of Biotic Impoverishment" (G. M. Woodwell, ed.), pp. 301-322. Cambridge Univ. Press, Cambridge, UK. Billings, W. D. (1994). Ecological impacts of cheatgrass and resultant fire on ecosystems in the western Great Basin. "Ecology and Management of Annual Rangelands" (S. B. Monsen, and S. G. Kitchen, eds.), pp. 22-30. United States Forest Service General Technical Report INT-GTR-313. Intermountain Research Station, Ogden, UT. Blank, R. B., Allen, F., and Young, J. A. (1994). Extractable soil anions in soils following 58, 564-570. wildfire in a sagebrush-grass community. Christensen, N. L. (1981). Fire regimes in southeastern ecosystems. "Fire Regimes and Ecosystem Properties" (H. A. Mooney, T. M. Bonnicksen, N. L. Christensen, J. E. Lotan, and W. A. Reiners, eds.), pp. 112-136. General Technical Report WO-26. United States Department of Agriculture, Intermountain Research Station, Ogden, UT. Christensen, N. L. (1985). Shrubland fire regimes and their evolutionary consequences. "The Ecology of Natural Disturbance and Patch Dynamics" (S. T. A. Pickett, and P. S. White, eds.), pp. 85-100. Academic Press, Orlando. Christensen, N. L. (1987). The biogeochemical consequences of fire and their effects on the vegetation of the coastal plain of the southeastern United States. "The Role of Fire in Ecological Systems" (L. Trabaud, ed.), pp 1-21. SPB Academic Publishing, The Hague, The Netherlands.
246
Cipollini, M. L., Drake, B. G., and Whigham, D. (1993). Effects of elvated CO2 on growth and carbon/nutrient balance in the deciduous woody shrub (L.) Blume (Lauraceae). 96, 339-346. Crutzen, P.J., and Goldammer, J. G., eds. (1993). "Fire in the Environment: The Ecological Atmospheric, and Climatic Important of Vegetation Fires." Wiley, Chichester. Cure, J. D. (1985). Carbon dioxide doubling responses: A crop survey. "Direct Effects of Increasing Carbon Dioxide on Vegetation" (B. R. Strain andJ. D. Cure, eds.), pp. 99-116. U.S. DOE/ER-0238, National Technical Information Service, Springfield, VA. D'Antonio, C. M., and Vitousek, P. M. (1992). Biological invasions by exotic grasses, the grass fire cycle, and global change. 23, 63-87. Diaz, S., Grime,J. P., Harris,J., and McPherson, E. (1993). Evidence of a feedback mechanism limiting plant response to elevated carbon dioxide. 364, 616-617. Dippery, J. K., Tissue, D. T., Thomas, R. B., and Strain, B. R. (1995). Effects of low and elevated CO2 on C3 and C4 annuals I. Growth and biomass allocation. 101, 13-20. Drake, B. G., and Leadley, P. W. (1991). Canopy photosynthesis of crops and native plant communities exposed to long-term elevated CO2. 14, 853-860. Farrar, J. F., and Williams, M. L. (1991). The effects of increased atmospheric carbon dioxide and temperature on carbon partitioning, source-sink relations and respiration. 14, 819-830. Goldammer, J. G. (1993). Historical biogeography of fire: Tropical and subtropical. "Fire in the Environment" (P.J. Crutzen andJ. G. Goldammer, eds.), pp. 297-314. Wiley, Chichester. Gunderson, C. A., and Wullschlager, S. D. (1994). Photosynthetic acclimation in trees to rising CO2: A broader perspective. 39, 369-388. Hogan, K. P., Smith, A. P., and Ziska, L. H. (1991). Potential effects of elevated CO2 and changes in tmeperature on tropical plants. 14, 763-778. Howe, H. F. (1994). Managing species diversity in tallgrass prairies: Assumptions and implications. 8, 691-704. Hunt, R., Hand, D. W., Hannah, M. A., and Neal, A. M. (1991). Response to CO2 enrichment in 27 herbaceous species. 5, 410-421. Hunter, R. (1990). Recent increases in populations on the Nevada test site. "Symposium on Cheatgrass Invasion, Shrub Die-off, and Other Aspects of Shrub Biology and Management" (E. D. McArthur, E. M. Romney, S. D. Smith, and P. T. Tueller, eds.), pp. 22-25. United States Forest Service General Technical Report INT-276. Intermountain Research Station, Ogden, UT. Jackson, R. B., Sala, O. E., Field, C. B., and Mooney, H. A. (1994). COp alters water use, carbon gain, and yield for the dominant species in a natural grassland. 98, 257262. Julkunen-Tiitto, R., Tahvanainen, J., and Silvola, J. (1993). Increased COp and nutrient status changes affect phytomass and the production of plant defensive secondary chemicals in 95, 495-498. Kilgore, B. M. (1981). Fire in ecosystem distribution and structure: Western forests and scrublands. "Fire Regimes and Ecosystem Properties" (H. A. Mooney, T. M. Bonnicksen, N. L. Christensen, J. E. Lotan, and W. A. Reiners, eds.), pp. 58-89. General Technical Report WO-26. United States Department of Agriculture, Intermountain Research Station, Ogden, UT. K6rner, C., and Arnone, J. A., III (1992). Regulation of the vesicular-arbuscular mycorrhizal symbiosis. 43, 1672-1675. Kozlowski, T. T., and Ahlgren, C. E. (1974). "Fire and Ecosystems." Academic Press, New York. Lambers, H. (1993). Rising COp, secondary plant metabolism, plant-herbivore interactions, and litter decomposition. 104/105, 263-271.
15.
247
Lincoln, D. E. (1993). The influence of plant carbon dioxide and nutrient supply on susceptibility to insect herbivores. 104/105, 273-280. Mack, R. N. (1986). Alien plant invasion into the intermountain west: A case history. "Ecology of Biological Invasions of North America and Hawaii" (H. A. Mooney and J. A. Drake, eds.), pp. 191-213. Springer-Verlag, New York. Mayeux, H. S.,Johnson, H. B., and Polley, H. W. (1994). Potential interactions between global change and intermountain annual grasslands. "Ecology and Management of Annual Rangelands" (S. B. Monsen and S. G. Kitchen, eds.), pp. 95-100. United States Forest Service General Technical Report INT-GTR-313. Intermountain Research Station, Ogden, UT. Mooney, H. A., Drake, B. G., Luxmore, R.J., Oechel, W. C., and Pitelka, L. F. (1991). Predicting ecosystem responses to elevated C O 2 concentrations. 41, 96-104. Morison, J. I. L. (1993). Response of plants to COz under water-limited conditions. 104/105, 193-209. Morse, S. R., Wayne, P., Miao, S. L., and Bazzaz, F. A. (1993). Elevated CO2 and drought alter tissue water relations of birch Marsh.) seedlings. 95, 599602. Oechel, W. C., and Strain, B. R. (1985). Native species responses to increased atmospheric carbon dioxide concentration. "Direct Effects of Increasing Carbon Dioxide on Vegetation" (B. R. Strain and J. D. Cure, eds.), pp. 117-154. U. S. DOE/ER-0238, National Technical Information Service, Springfield, VA. Oechel, W. C., Cowles, S., Grulke, N., Hastings, S.J., Lawrence, B., Prudhomme, T., Riechers, G., Strain, B., Tissue, D., and Vourlitis, G. (1994). Transient nature of CO2 fertilization in Arctic tundra. 371, 500-503. Overdieck, D. (1993). Elevated CO2 and the mineral content of herbaceous and woody plants. 104/105, 403-411. Owensby, C. E., Coyne, P. I., Ham, J. M., Auen, L. M., and Knapp, A. K. (1993). Biomass production in a tallgrass prairie ecosystem exposed to ambient and elevated CO2. 3, 644-653. Philpot, C. W. (1970). Influence of mineral content on the pyrolysis of plant materials. For. 16, 461-471. Pyne, S.J. (1984). "Introduction to Wildland Fire." Wiley, New York. Rogers, H. H., Sionit, N., Cure, J. D., Smith, J. M., and Bingham, G. E. (1984). Influence of elevated carbon dioxide on water relations of soybeans. 74, 233-238. Rogers, H. H., Peterson, C. M., McCrimmon, J. N., and Cure, J. D. (1992). Response of plant 15, 749-752. roots to elevated atmospheric carbon dioxide. Roth, S. K., and Lindroth, R. L. (1994). Effects of CO2-mediated changes in paper birch and white pine chemistry on gypsy moth performance. 98, 133-138. Rothermel, R. C. (1983). "How to Predict the Spread and Intensity of Forest and Range Fires." General Technical Report INT-143. United States Department of Agriculture, Intermountain Forest and Range Experimental Station, Ogden, UT. Ryan, K. C. (1991). Vegetation and wildland fire: Implications of global climate change. 17, 169-178. Sage, R. F. (1994). Acclimation of photosynthesis to elevated COz: The gas exchange perspective. 39, 351-368. Sage, R. F., and Reid, C. D. (1994). Photosynthetic response mechanisms to environmental change in C3 plants. "Plant-Environment Interactions" (R. E. Wilkinson, ed.), pp. 413-499. Dekker, New York. Sage, R. F., Santrucek, J., and Grise, D.J. (1995). Temperature effects on the photosynthetic response of C3 plants to long-term CO2 enrichment. 121, 67-77. Santrucek, J., and Sage, R. F. (1996). Acclimation of stomatal conductance to a CO2-enriched 23, atmosphere and elevated temperature in in press.
248 Schlesinger, W. H., Reynolds, J. F., Cunningham, G. L., Huenneke, L. F., Jarrell, W. M., Virginia, R. A., and Whifford, W. G. (1990). Biological feedbacks in global desertification. 247, 1043-1048. Seastedt, T. R., Briggs, J. M., and Gibson, D.J. (1991). Controls on nitrogen limitation in tallgrass prairie. 87, 72-79. Sharkey, T. D., Loreto, F., and Delwiche, C. F. (1991). High carbon dioxide and sun/shade effects on isoprene emission from oak and aspen tree leaves. 14, 333-338. Sionit, N., Strain, B. R., Hellmers, H., and Kramer, P.J. (1981). Effects of atmospheric COs concentration and water stress on water relations of wheat. 142, 191-196. Smith, S. D., Strain, B. R., and Sharkey, T. D. (1987). Effects of COs enrichment on four Great Basin grasses. 1, 139-143. Sousa, W. P. (1984). The role of disturbance in natural communities. 15, 353-391. Strain, B. R., and Bazzaz F. A. (1983). Terrestrial plant communities. "COs and Plants" (E. R. Lemon, ed.), pp. 177-213. American Association for the Advancement of Science. Westview Press, Boulder, CO. Thomas, R. B., and Strain, B. R. (1991). Root restriction as a factor in photosynthetic acclima96, 627-634. tion of cotton seedlings grown in elevated carbon dioxide. Torn, M. S., and Fried, J. S. (1992). Predicting the impacts of global warming on wildland fire. 21, 257-274. Trabaud, L. (1987). Fire and survival traits of plants. "The Role of Fire in Ecological Systems" (L. Trabaud, ed.), pp. 65-89. SPB Academic Publishing, The Hague, The Netherlands. Thrabaud L. V., Christensen, N. L., and Gill, A. M. (1993). Historical biogeography of fire in temperate and Mediterranean ecosystems "Fire in the Environment" (P. J. Crutzen and J. G. Goldammer, eds.), pp. 277-295. Wiley, Chichester. Turner, N. C. (1987). The use of the pressure chamber in studies of plant water status. "Proceedings of International Conference on Measurement of Soil and Plant Water Status" (R. J. Hanks and R. W. Brown, eds.), pp. 13-24. Utah State Univ. Press, Logan, UT. Tyree, M. T., and Alexander, J. D. (1993). Plant water relations and the effects of elevated COs: A review and suggestions for future research. 104/105, 47-62. Wedin, D. A. (1995). Species, nitrogen, and grassland dynamics: The constraints of stuff. "Linking Species and Ecosystems" (C. Jones and J. H. Lawton, eds.), pp. 253-262. Chapman & Hall, New York. Whisenant, S. G. (1990). Changing fire frequencies on Idaho's Snake River plains: Ecological and management implications. "Symposium on Cheatgrass Invasion, Shrub Die-off, and Other Aspects of Shrub Biology and Management" (E. D. McArthur, E. M. Romney, S. D. Smith, and P. T. Tueller, eds.), pp. 4-10. United States Forest Service General Technical Report INT-276. Intermountain Research Station, Ogden, UT. Wilson, J. B., and Agnew, A. D. Q. (1992). Positive-feedback switches in plant communities. 23, 263-306. Wong, S. C. (1990). Elevated atmosphere partial pressure of COs and plant growth. Res. 23, 171-180. Wong, S. C. (1993). Interaction between elevated atmospheric concentration of COs and humidity on plant growth: Comparison between cotton and radish. 104/105, 211-221. Wray, S. M., and Strain, B. R. (1986). Response of two old field perennials to interactions of COs enrichment and drought stress. 73, 1486-1491. Wright, H. A., and Bailey, A. B. (1982). "Fire Ecology." Wiley-Interscience, New York. Wright, H. A., Bunting, S. C., and Neuenschwander, L. F. (1976). Effect of fire on honey mesquite, jr. 29, 467-471. Young, J. A., and Evans, R. A. (1978). Population dynamics after wildfires in sagebrush grasslands. J. 31, 283-289.
15.
249
Young, J. A., and Tipton, F. (1990). Invasion of cheatgrass into arid environments of the Lahontan Basin. "Symposium on Cheatgrass Invasion, Shrub Die-off, and Other Aspects of Shrub Biology and Management" (E. D. McArthur, E. M. Romney, S. D. Smith, and P. T. Tueller, eds.), pp. 37-46. United States Forest Service General Technical Report INT-276. Intermountain Research Station, Ogden, UT. Zedler, P. H., Clayton, R. G., and McMaster, G. S. (1983). Vegetation change in response to extreme events: The effect of a short interval between fires in California chaparral and coastal shrub. 64, 809-818. Young, J. A., Evans, R. A., and Swanson, R. A. (1987). Snuff the candles in the desert. 13, 3-4.
This Page Intentionally Left Blank
III Organismic Interactions Chapters 16 and 17
Chapters 18 through 22
Chapters 23 and 24
This Page Intentionally Left Blank
1 Symbiotic Nitrogen Fixation: One Key to Understand the Response of Temperate Grassland Ecosystems to Elevated C02?
The potential positive effect of elevated C O 2 o n plant growth has long been known. Striking differences exist in the responses between species, between plants growing in different environments, or between field- and pot-grown plants (see articles in this volume). Attempts to explain plant response to elevated CO2 and processes that take place in the shoot zone are already investigated intensively, although open questions are left behind. In contrast to the shoot, processes in the rhizosphere are much less investigated. There is now increasing awareness that belowground processes have to be included when attempts are made to explain plant responses to elevated CO2. Through plants, elevated CO2 may affect physical, chemical, and biological characteristics of the soil and, in turn, the plant response to elevated CO2. However, only vague information about such processes is available; as a result many answers that may explain plant responses to elevated CO2 may still be hidden below ground. It is generally accepted that nutrient supply has a significant impact on the plant CO2 response (Baker 1990; Comins and McMurtrie, 1993; Cure 1988; Goudriaan and Ruiten, 1983). Especially under fertile soil conditions nitrogen is mostly the main limiting factor for plant growth. Copyright 9 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.
254 As a result, nitrogen nutrition is to a great extent determinant for the plant biomass response to elevated CO2. In this context, the symbiotic interaction between legume plants and bacteria offers a unique possibility to give the host plant access to the unlimited nitrogen source from the atmosphere. Therefore, it is likely that legumes can account for their own CO2-induced, increased N demand through symbiotic N2 fixation, regardless of the nitrogen availability in the soil (Hartwig and N6sberger, 1994). Due to the intimate relationships between the C and N cycles in ecosystems, symbiotic N2 fixation could even account for a possible CO2-induced, increased N sink of the whole ecosystem (Gifford, 1992). A model is proposed that builds on a physiological relationship between symbiotic N 2 fixation and the nitrogen sink of the whole ecosystem and thus highlights the role of symbiotic N2 fixation for C sequestration into the biosphere.
Symbiotically fixed nitrogen is the most significant natural N source for many grassland ecosystems. This endosymbiosis is the result of a successful root infection by host-specific, soil-born, rhizobial bacteria. As a visible evidence of this symbiosis, typically root nodules are formed within which strictly procaryotic genes for nitrogen fixation are expressed. Assuming an adequate infection of a host legume with an effective strain, N2 fixation can be regulated by a variable oxygen diffusion resistance in the nodules (Criswell 1976; Weisz and Sinclair, 1987; Minchin 1986; Witty 1984; Hartwig 1987); carbohydrate status of plants is 1990, usually not a primary regulating factor for N 2 fixation (Hartwig 1994; Weisbach 1996). Rate of nodulation (Streeter, 1988) as well as nodule morphology (Dakora and Atkins, 1990; Parsons and Day, 1990; James 1991) are also involved in the regulation of N 2 fixation. Such mechanisms allow the plant to tune Nz-fixation activity quickly and accurately to respond to varying plant nitrogen demands caused by changing environmental conditions such as nitrate fertilization (Streeter, 1988), drought (Durand 1987), defoliation (Hartwig 1987), or temperature (Kessler 1990). This way either nitrogen deficiency or toxicity can be prevented (Hartwig and Ntsberger, 1994; Hartwig 1994). There are strong indications that a nitrogen feedback regulation is involved in such an adaptation (Parsons 1993; Heim 1993; Oti-Boateng and Silsbury, 1993; Oti-Boateng 1994). It is noteworthy that the ecological range of the legume-rhizobia symbiosis is narrower than that of the plant itself. This was reported for low temperature (Kessler 1990; Bordeleau and Prtvost, 1994), low phos-
16.
phorus (Cadisch 1993), and for low potassium (Sangakkara unpublished data). When legumes are exposed to extreme conditions, the symbiosis either becomes less effective or even ceases completely.
The driving force for nutrient assimilation is growth (Ingestad, 1982) or increase in biomass. Thus, if elevated CO2 should lead to more biomass production, nutrient, in particular nitrogen uptake and assimilation, has to be increased adequately (Fig. 1) in order to maintain a concentration sufficient to keep supporting plant metabolism. Model predictions showed that a more efficient rubisco enzyme requires less N which was subsequently reflected in lower concentrations of N in leaves of plants grown under elevated CO2 concentrations (Andrews and Lorimer, 1987; Morell 1992; Badger, 1992). Nevertheless, one cannot expect a strong increase in biomass production without a more or less appropriate increase in nitrogen assimilation either from the available soil N pool or from the atmosphere (Comins and McMurtrie, 1993). It is obvious that an effectively nodulated legume hardly faces the problem of insufficient nitrogen assimilation. As a result, under elevated CO2 with or without increased mineral N assimilation from the soil, nitrogen fixation is expected to increase in concert with the increased biomass production (Fig. 1). Such a powerful system for adequate N acquisition may be crucial in order to take full advantage of the extra carbon under elevated CO2, presuming other environmental conditions such as light, water supply, temperature, or other mineral nutrients (e.g., P, K) are not limiting growth. This presumption is indeed supported by the fact that C3 plants, capable to symbiotically fix nitrogen, usually show a stronger CO2 response compared to non-Nz-fixing species (Poorter, 1993).
As atmospheric C O 2 increases, the amount and the C : N ratio of plant litter from both roots and shoots as well as the amount and the C : N ratio of the root exudates will increase. Since root exudates contain considerable amounts of easily digestible, C-rich substances like sugars, organic acids, and amino acids (Rovira, 1969), growth of soil microbial biomass will be stimulated (Merckx 1987; Van de Geijn and Van Veen, 1993; Zak 1993, Norby 1987; Lekkerkerk 1990; Eamus and Jarvis,
J Biomass ~
x7
nonlegumes
"xeO;;~ . I ~
Schematic description of the various environmental conditions, processes, and nitrogen pools in a grassland ecosystem that might influence nitrogen cycling. As indicated by the symbolized valves, the arrows indicate either augmenting or diminishing the processes.
1989). Although total N mineralization may be accelerated by increased soil microbial biomass and activity, mineral N will get depleted at the same time because of the increased demand for mineral nitrogen by vigorously growing microbial populations. This implies that the growth of microbial biomass becomes limited by available nutrients like nitrogen rather than by carbon (Jingguo and Bakken, 1989; Merckx 1987; Helal and Sauerbeck, 1986). As a result, nitrogen availability may limit microbial sequestration of plant-released carbon in a similar way as it limits C sequestration into phytomass (Merckx 1987; Van Veen 1989, 1991;
16.
257
Liljeroth 1990). Such a concept was even extended by Diaz (1993) who proposed a feedback regulation of the plant CO2 response in such a way that the increased release of carbon substrates into the rhizosphere leads to sequestration of soil minerals into an expanded microflora which leads to a more p r o n o u n c e d nutritional limitation of plant growth. It could be expected that the rhizosphere microbial populations would grow vigorously in this C-rich environment and thus aggressively compete with the plant for available soil nitrogen. It is also conceivable that " b o o m and bust" of soil microbial populations along the trophic chain (bacteria, fungi, protozoa) lead to periods of higher and lower nutrient availability in the soil; a dynamic that could be significantly altered by a sudden increase in the C : N ratio of their primary resources. Thus, microbial populations might result in temporal immobilization of N. The described process, however, has only a limited capacity to immobilize nitrogen in space as well as in time and will not account for a significant long-term N immobilization. It is more likely that processes associated with a more vigorous turnover of soil microbial populations might account temporarily for reduced mineral N availability in the soil. Under elevated CO2 and sufficient mineral nitrogen, the turnover rate of root-derived, organic matter might well be stimulated. As a result, decomposition of more resilient soil organic matter is inhibited (Liljerot 1990). Soil microbes preferentially feed on easily digestible, C-rich root exudates and dead fine roots as opposed to resilient, N-rich soil organic matter and thus reduce N mineralization from soil organic matter (Jingguo and Bakken, 1989). At the same time soil microbes use soil mineral nitrogen to meet their own N demand. These processes could be magnified by the competition between fast-growing soil microbes using easily digestible substrates versus microbes that decompose soil organic matter or similarly, microbes that synthesize resilient molecules (e.g., chitins, melanins) versus microbes that synthesize more easily digestible compounds. This could also lead to a shift in the build-up of more resilient molecules versus easily digestible ones. As a result, the size of the soil organic matter pool would increase. Subsequently, mineral N that was originally assimilated by soil microbes in competition with plants will be incorporated into the increasing a m o u n t of a resilient soil organic matter pool and thus be excluded from the nutrient cycle. Although increased soil microbial populations are not considered to assimilate all the available mineral nitrogen u n d e r elevated CO2, increased microbial populations could lead to a higher denitrification potential since many soil microbes can also utilize nitrate as a terminal electron acceptor. Increased metabolism in the soil (root and microbial respiration) would increase oxygen consumption which could lead to more anaerobic conditions in the soil. The improved water use efficiency of plants u n d e r elevated
258
CO2 (Goudriaan and Unsworth, 1990; Morison, 1985) could lead to longer periods of increased soil moisture content which then can lead to anaerobic soil spots and to a greater proportion of water-filled (anaerobic) soil pores following a precipitation event. As a result, higher rates of denitrification from facultative anaerobic microorganisms (using nitrate as the terminal electon acceptor) could lead to considerable nitrogen loss under elevated CO2 and the availability of mineral N in the soil would decrease. Nevertheless, all the above described processes remain speculative. However, it is clear that the C : N ratio in plants increases as the atmospheric CO2 concentration increases. Consequently, the C : N ratio of plant residues introduced into the soil will increase. Due to the intimate relationship between C and N cycles, this will ultimately lead to an intensified immobilization of soil mineral nitrogen. As a result, the availability of soil mineral nitrogen is likely to decrease under elevated CO2. In summary, there are many indications which lead one to expect that strength a n d / o r size of the nitrogen sink in an ecosystem increases, at least temporarily, under elevated CO2. This means at the same time that, if this increased N sink cannot be satisfied, additional net C sequestration might get curbed more and more, finally reaching a new equilibrium.
Various environmental effects (including elevated C02) affect plant growth and thus nitrogen demand in nonlegumes and legumes (Fig. 1). As a result, N assimilation from symbiotic N 2 fixation of a legume-rhizobia symbiosis will be tuned to meet the increased legume N demand under elevated CO2 (Masterson and Sherwood, 1978; Hartwig 1990; Ryle and Powell, 1992; Williams and Phillips, 1980; Finn and Brun, 1982). In the presence of mineral nitrogen in the soil, nitrogen fixation is reduced in proportion to the amount of the available soil nitrogen. Under field conditions, nitrogen is always available to a legume. Thus, the extent of competition for soil available nitrogen by associated nonlegumes is one factor that determines the proportion of nitrogen derived from N2 fixation (Harderson 1988; Ta and Farris, 1987; Nesheim and Boller, 1991; Seresinhe 1994; Zanetti unpublished data). As if competed for nitrogen by associated nonlegumes, the proportion of nitrogen derived from symbiotic N2 fixation will also increase if soil nitrogen availability is reduced by any other mechanism (leaching, denitrification, low mineralization, nitrogen immobilization). At this point, a direct physiological link between symbiotic nitrogen fixation and, through the mineral N availability
259
16.
in the soil, the nitrogen sink of the whole ecosystem is closed (Fig. 1). Symbiotic nitrogen fixation may be able to account for the increased N sink of the whole ecosystem and thus buffer a possible imbalance between C and N cycles as it may occur under elevated CO2 (Gifford, 1992). Hence, symbiotic N 2 fixation would enable increased C sequestration into the entire ecosystem following an elevation of the atmospheric CO2 concentration. Data from a model grassland ecosystem on a fertile soil established in the Swiss FACE project (for more background information on this experiment see Lascher Chapter 19) support this concept. Under elevated CO2, the clover biomass production and the proportion of clover in the sward increased (Hebeisen unpublished data; Lascher Chapter 19). It appears that the entire additional amount of nitrogen assimilated in clover under elevated CO2 came through symbiotic N2 fixation (Table I; Zanetti 1995; Zanetti unpublished data). Compared to ambient CO2, clover plants grown under elevated CO2 fixed even more nitrogen than would be expected from the increased biomass production; therefore, more additional nitrogen was sequestered into the ecosystem as required to meet increased nitrogen demand (Table I) for increased clover biomass production. These first observations would be the result of an increased N sink both in the legume and in the ecosystem as a whole and thus be consistent with the proposed model.
Although performance of symbiotic N 2 fixation is N sink regulated at the plant level, a physiological connection of symbiotic N2 fixation with the ecosystem as a whole is proposed. At the ecosystem level, fertilizer N, demand of N by legumes and associated nonlegumes, nitrogen mineralization and immobilization, leaching, atmospheric N deposition and nitrogen
Atmospheric CO2 (Pa) 35 (ambient) 60 (elevated) P
Total N in aboveground biomass (g N m -z)
Total N fixed (g N m -2)
41 45 n.s.
15 22 <0.05
,
Average values from the 5-month experimental period, 1993, in white clover monocultures in the frequent cutting treatments [from Zanetti unpublished field (FACE) data] are shown.
260 loss through dentrification, all contribute to the availability of mineral N in the soil and thus to the demand for symbiotically fixed nitrogen (Fig. 1). Consequently, all these parameters will affect the performance of Nz fixation. At elevated CO2, one can consider that an increase in plant biomass along with an increased C" N ratio of plant litter and exudates introduced into the soil occurs, and thus accept the concept that the N sink of the whole ecosystem increases under those circumstances. As a result, N2 fixation increases. This will occur both through an increase in the legume proportion in the sward and by increasing Nz-fixing activity in each individual legume. Hence, symbiotic N2 fixation has to be viewed not only as a key factor of the legume's response to elevated CO2 (C sequestration into legume phytomass) but also as an important factor for C sequestration into the entire ecosystem; symbiotic N2 fixation might be crucial in assessing the fate of extra carbon in terrestrial ecosystems.
The symbiosis opens the unlimited access to atmospheric nitrogen as a source of nitrogen for the host plant. Symbiotic N2 fixation is considered to be tuned to the plant's demand for symbiotically fixed nitrogen. The role of symbiotic nitrogen fixation in response to an elevated atmospheric CO2 concentration in grassland ecosystems has never been assessed. However, the response of plants to elevated CO2 is highly dependent on the availability of nutrients, especially on nitrogen. It is generally accepted that as atmospheric CO2 concentration increases, the C : N ratio of plant residues and exudates increases. The availability of soil nitrogen, therefore, decreases presumably due to temporal N immobilization. Both a COz-stimulated, increased plant growth (thus requiring more nitrogen) and an increased N demand of other species and processes in the soil under elevated COz will result in a larger N sink of the whole ecosystem. One possibility for maintenance of the balance between the C and the N cycles in the system would be increased N import into the ecosystem through symbiotic N2 fixation. Preliminary data on the performance of symbiotic N2 fixation under elevated CO2 in the field (Swiss grassland FACE experiment) support such a concept: although the soil did not provide more nitrogen to the plants under elevated CO2, white clover showed a pronounced biomass response and increased its proportion in the sward at the expense of the associated grasses. Moreover, symbiotic nitrogen fixation in each individual clover plant increased strongly. This way, clover appeared to satisfy an increased N sink of the whole ecosystem through increased N2 fixation. It is proposed that under elevated CO2,
16. Symbiotic Nitrogen Fixation
symbiotic N2 fixation enables the sequestration of additional carbon into the entire ecosystem.
S. Koller, J. Nagy, and K. F. Lewin built the COs control and monitoring system. We thank Anni Dflrsteler, K. R~egg, P. Schl/issel, and P. Jager for technical assistance. The FACE experiment is mainly supported by the Swiss National Energy Fund. Further support was provided by the Swiss National Science Foundation, the Swiss Department of Energy, and the Swiss Federal Institute of Technology.
Andrews, T. J., and Lorimer, G. H. (1987). Rubisco: Structure, mechanisms, and prospects for improvement. "The Biochemistry of Plants, A Comprehensive Treatise" (M. D. Hatch and N. K. Boardman, eds.), Vol. 10, pp. 132-219. Academic Press, San Diego. Badger, M. (1992). Manipulating agricultural plants for a future high CO2 environment. 40, 421-429. Baker, J. T., Allen, L. H., Boote, g.J.,Jr.,Jones, P., andJones,J. W. (1990). Rice photosynthesis and evapotranspiration in subambient, ambient, and superambient carbon dioxide concen82, 834-840. trations. Bordeleau, L. M., and Pr~vost, D. (1994). Nodulation and nitrogen fixation in extreme environments. 161, 115-125. Cadisch, G., Sylvester-Bradley, R., Boller, B. C., and N6sberger, J. (1993). Effects of phosphorus and potassium on N2 fixation (15N-dilution) of field-grown and C. 31, 329-340. Comins, H. N., and McMurtrie, R. E. (1993). Long-term response of nutrient-limite forests to COs enrichment; equilibrium behavior of plant-soil models. 3, 666-681. Criswell, J. G., Havelka, U. D., Quebedeaux, B., and Hardy, R. W. F. (1976). Adaptation of nitrogen fixation by intact soybean nodules to altered rhizosphere pO2. 58, 622-625. Cure, J. D., Israel, D. W., and Rufty, T. W., Jr. (1988). Nitrogen stress effects on growth and seed yield of non-nodulated soybean exposed to elevated carbon dioxide. 28, 671-677. Dakora, F., and Atkins, C. B. (1990). Effect of pO2 on growth and nodule functioning of 93, 948-955. symbiotic cowpea L. Walp.). Diaz, S., Grime,J. P., Harris, J., and McPherson, E. (1993). Evidence of a feedback mechanism limiting plant response to elevated carbon dioxide. 364, 616-617. Durand, J. L., Sheehy, J. E., and Minchin, F. R. (1987). Nitrogenase activity, photosynthesis, and nodule water potential in soybean plants experiencing water deprivation. 38, 311-321. Eamus, D., and Jarvis, P. G. (1989). The direct effects of increases in the global atmospheric concentration of COs on natural and temperate trees and forests. 19, 1-55. Finn, G. A., and Brun, W. A. (1982). Effect of atmospheric COs enrichment on growth, nonstructural carbohydrate content, and root nodule activity in soybean. 69, 327-331. Gifford, R. M. (1992). Interaction of carbon dioxide with growth-limiting environmental factors in vegetation productivity: Implications for the global carbon cycle. "Advances
262 in Bioclimatology" (R. L. Desjardius, R. M. Gifford, T. Nilson, and E. A. N. Greenwood), Vol. 1, pp. 26-58. Goudriaan, J., and de Ruiter, H. E. (1983). Plant growth in response to CO2 enrichment, at two levels of nitrogen and phosphorus supply. I. Dry matter, leaf area, and development. 31, 157-169. Goudriaan, J., and Unsworth, M. H. (1990). Implications of increasing carbon dioxide and climate change for agricultural productivity and water resources. "Impact of Carbon Dioxide, Trace Gases and Climate Change on Global Agriculture," pp. 1-11. ASA Special Publication. Hardarson, G., Danso, S. K. A., and Zapata, F. (1988). Dinitrogen fixation measurements in alfalfa-ryegrass swards using nitrogen-15 and influence of the reference crop. 28, 101-105. Hartwig, U. A., and N6sberger, J. (1994). What triggers the regulation of nitrogenase activity in forage legume nodules after defoliation? 161, 109-114. Hartwig, U. A., Boller, B., and N6sberger, J. (1987). Oxygen supply limits nitrogenase activity of clover nodules after defoliation. 59, 285-291. Hartwig, U. A., Boller, B. C., Baur-H6ch, B., and N6sberger, J. (1990). The influence of carbohydrate reserves on the response of nodulated white clover to defoliation. 65, 97-105. Hartwig, U. A., Heim, I., L~scher, A., and N6sberger, J. (1994). The nitrogen-sink is involved in the regulation of nitrogenase activity in white clover after defoliation. 92, 375-382. Heim, I., Hartwig, U. A., and N6sberger, J. (1993). Current nitrogen fixation is involved in the regulation of nitrogenase activity in white clover L). 103, 1009-1014. Helal, H. M., and Sauerbeck, D. (1986). Effect of plant roots on carbon metabolism of soil microbial biomass. Sc/. 149, 181-188. Ingestad, T. (1982) Relative addition rate and external concentration; driving variables used in plant nutrition research. 5, 443-453. James, E. K., Sprent, J. I., Minchin, F. R., and Brewin, N.J. (1991). Intercellular location of glycoprotein in soybean nodules--Effect of altered rhizosphere oxygen concentration. 14, 467-476. Jingguo, W., and Bakken, L. R. (1989). Nitrogen mineralization in rhizosphere and nonrhizosphere soil, effect of the spatial distribution of N-rich and N-poor plant residues. "Nitrogen in Organic Wastes Applied to Soils" (]. A. Hansen and K. Henrikson, eds.), pp. 81-97. Academic Press, London. Kessler, W., Boller, B. C., and N6sberger, J. (1990). Distinct influence of root and shoot temperature on nitrogen fixation by white clover. 65, 341-346. Lekkerkerk, L. J. A., Van de Geijn, S. C., and Van Veen, J. A. (1990). Effects of elevated atmospheric CO~-levels on the carbon economy of a soil planted with wheat. "Soils and the Greenhouse Effect" (A. F. Bouwman, ed.), pp. 423-429. Wiley, Chichester, UK. Liljeroth, E., Van Veen, J. A., and Miller, H.J. (1990). Assimilate translocation to the rhizosphere of two wheat lines and subsequent utilization by rhizosphere microorganisms at two soil nitrogen concentration. 22, 1015-1021. Lfischer, A., Hebeisen, T., Zanetti, S., Hartwig, U. A., Blum, H., Hendrey, G. R., and N6sberger, J . (1996). Differences between legumes and nonlegumes of permanent grassland in their responses to free-air carbon dioxide enrichment. "Carbon Dioxide, Populations, and Communities" (Ch. K6rner and F. Bazzaz, eds.), pp. 287-300. Academic Press, San Diego. Masterson, C. L., and Sherwood, M. T. (1978). Some effects of increased atmospheric carbon dioxide on white clover (Trif01ium and pea 49, 421-426. Merckx, R., Dijkstra, A., den Hartog, A., and Van Veen, J. A. (1987). Production of rootderived material and associated microbial growth in soil at different nutrient levels. 5, 126-132.
16.
263
Minchin, F. R., Minguez, M. I., Sheehy, J. E., Witty, J. F., and Skit, L. (1986). Relationships between nitrate and oxygen supply in symbiotic nitrogen fixation by white clover. 37, 1103-1113. Morell, M. I~, Paul, K., Kane, H. J., and Andrews, T.J. (1992). Rubisco: maladapted or misunderstood? 40, 431-441. Morison, J. I. L. (1985). Sensitivity of stomata and water use efficiency to high CO2. 8, 467-474. Nesheim, L., and Boller, B. C. (1991). Nitrogen fixation by white clover when competing 133, 47-56. with grasses at moderately low temperatures. Norby, R.J., O'Neil, E. G., Hood, W. G., and Luxmore, R.J. (1987). Carbon allocation, root exudation, and mycorrhizal colonization of seedlings grown under CO2 enrichment. 3, 203-210. Oti-Boateng, C., and Silsbury, J. H. (1993). The effects of exogenous amino acid on acetylene reduction activity of L. cv. Fiord. 71, 71-74. Oti-Boateng, C., Wallace, W., and Silsbury, J. H. (1994). The effect of the accumulation of carbohydrate and organic nitrogen on nitrogen fixation (acetylene reduction) of faba bean cv. Fiord. 73, 143-149. Parsons, R., and Day, D. A. (1990). Mechanism of soybean nodule adaptation to different oxygen pressures. 13, 501-512. Parsons, R., Stanforth, A., Raven, J. A., and Sprent, J. I. (1993). Nodule growth and activity may be regulated by a feedback mechanism involving phloem nitrogen. 16, 125-136. Poorter, H. (1993). Interspecific variation in the growth response of plants to an elevated ambient CO2 concentration. 104/105, 77-97. Rovira, A. D. (1969). Plant root exudates. 35, 35-57. Ryle, G.J.A., and Powell, C. E. (1992). The influence of elevated CO2 and temperature on biomass production of continuously defoliated white clover. 15, 593-599. Seresinhe, T., Hartwig, U. A., Kessler, W., and N6sberger,J. (1994). Symbiotic nitrogen fixation of white clover in a mixed sward is not limited by height of repeated cutting. J. 172, 279-288. Streeter, J. (1988). Inhibition of legume nodule formation and N2 fixation by nitrate. Sc/. 7, 1-23. Ta, T. C., and Faris, M. A. (1987). Effects of alfalfa proportions and clipping frequencies on timothy-alfalfa mixtures. II. Nitrogen fixation and transfer. 79, 820-824. Van de Geijn, S. C., and Van Veen, J. A. (1993). Implications of increased carbon dioxide levels for carbon input and turnover in soils. 104/105, 283-292. Van Veen, J. A., Merckx, R., and Van de Geijn, S. C. (1989). Plant and soil-related controls of the flow of carbon from roots through the soil biomass. Plant Soil 115, 43-52. Van Veen, J. A., Liljeroth, E., Lekkerkerk, L.J.A., and Van de Geijn, S. C. (1991). Carbon fluxes in plant-soil systems at elevated atmospheric CO2 levels. 1, 175-181. Weisbach, C., Hartwig, U. A., Heim, I., and N6sberger, J. (1996). Whole nodule carbon metabolites are not involved in the regulation of the oxygen permeability and nitrogenase activity in white clover nodules. 110, 539-545. Weisz, P. R., and Sinclair, T. R. (1987). Regulation of soybean nitrogen fixation in response to rhizosphere oxygen. 84, 900-905. Williams, L. E., and Phillips, D. A. (1980). Effect of irradiance on development of apparent nitrogen fixation and photosynthesis in soybean. 66, 968-972. Witty, J. F., Minchin, F. R., Sheehy,J. E., and Minguez, M. I. (1984). Acetylene-induced changes in the oxygen diffusion resistance and nitrogenase activity of legume root nodules. 53, 13-20. Zak, D. R., Pregitzer, K. S., Curtis, P. S., Teeri, J. A., Fogel, R., and Randlett, D. (1993). Elevated atmospheric CO2 and feedback between carbon and nitrogen cycles. 151, 105-117.
264 Zanetti, S., Hartwig, U. A., Hebeisen, T., Lfischer, A., Hendrey, G. R., Blum, H., and N6sberger, J. (1995). Effect of elevated atmospheric CO2 on performance of symbiotic nitrogen fixation in white clover in the field (Swiss FACE experiment) and possible ecological implication. "Nitrogen Fixation: Fundamentals and Applications" (I. A. Tikhonovich, V. I. Romanow, N. A. Provorov, and W. E. Newton, eds.), p. 615. Kluwer Academic, Dordrecht, The Netherlands.
17 Plant-Fungal Interactions in a CO2-Rich World
Much attention is currently focused on the effects of elevated C O 2 o n the functioning of ecosystems and the way in which plant communities may be modified due to the differential effects of elevated atmospheric CO2 on the growth of the component species (see Leadley and K6rner, Chapter 11). Interactions among the biotic components of communities are important in determining community structure and some of these, particularly competition and herbivory, have received a great amount of attention. Symbiotic interactions, of either a mutualistic or a parasitic nature have, until recently, been largely ignored as determinants of community structure (Minchella and Scott, 1991; Dobson and Crawley, 1994). Studies on the effects of fungal pathogens on plant populations (Carlsson 1990; Carlsson and Elmqvist, 1992; Schmid, 1994) and the strong selective pressures which fungi may exert on plants (Clay, 1991) have lead to a wider acceptance of the role of pathogenic interactions in plant community structure. Similarly, mutualistic interactions between plants and arbuscular mycorrhizal fungi are increasingly being seen as important factors both in reproductive fitness in annual plant populations (Carey 1992; Koide 1994) and in determining community structure (Grime 1987; Sanders and Koide, 1994). Moreover, the abundance of the arbuscular mycorrhizal symbiosis in most ecosystems and the physiological nature of the interaction, as well as the movement of a large proportion of photosynCopyright 9 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.
thate from the host plant to the fungus and the transfer by the fungus of phosphate and other minerals from the soil to plant tissues, suggests an important role of this symbiosis in the functioning of natural ecosystems. Only those plant-fungal interactions where the fungal partner is completely dependent on host plant assimilate are considered in this chapter. Arbuscular mycorrhizal fungi and many fungal pathogens of plants are obligate symbionts and are directly dependent on the photosynthates from their hosts. It is likely, therefore, that any change to the carbon budget of the host, as is likely to be induced by a changing atmospheric CO2 concentration, will affect the environment of the symbiotic fungi, even though they are not directly exposed to a CO2-enriched environment. Effects of this environmental change on the fungi may subsequently lead to changes in the interaction between the two partners of the symbiosis and such changes could have far-reaching effects at the community level and possibly at the ecosystem level. Little experimental evidence currently exists regarding the ways in which the outcome of plant-fungal interactions may be modified by atmospheric CO2 enrichment. In this chapter I suggest some possible outcomes and consequences of atmospheric CO2 enrichment at both the community and the ecosystem levels and provide some experimental evidence to support some of these scenarios. No clear answers are provided to these questions and the main aim is to alert global change researchers to the fact that community and ecosystem modification, as a result of global change, cannot be predicted without knowledge of potential responses of symbiotic interactions to these environmental changes. Hopefully this chapter may stimulate more interest in this important but neglected area.
A prerequisite to considering the modifications that may occur to plantfungal interactions as a result of elevated atmospheric CO2 is firstly to know how the fungal partners may respond to this environmental change. If an increased CO2 concentration results in a greater photosynthetic rate, then it would seem likely that the fungus would have a potentially greater supply of photosynthate. If photosynthate is usually limiting, then greater fungal growth would be expected. Very little evidence exists as to how symbiotic fungi respond to elevated levels of atmospheric CO2, although numerous reports infer that responses will, and do, occur. Although an increase in ectomycorrhizal root density in has been suggested (Norby 1987), the plants were inoculated with the natural ectomycorrhizal flora. Some of these fungi may
17.
not have been completely dependent on plant assimilates. A recent study by Ineichen (1995) does, however, demonstrate that seedlings inoculated with the ectomycorrhizal fungus form three times as many mycorrhizal root clusters at 600 /xmol mo1-1 CO2 concentration than at ambient CO2. Moreover, significantly greater amounts of extraradical mycelium were produced at 600/xmol mo1-1 CO2 compared to the ambient level. The model system of Ineichen (1995) was an axenic culture system and the fungus only had the possibility of obtaining assimilate from the seedlings. The responses of arbuscular mycorrhizal fungi (probably the commonest symbiotic fungi) to elevated CO2 have only been studied for a few plant species although the consequences of the mycorrhizal symbiosis to the functioning of ecosystems has been considered (see section III in this chapter). In (a C4 grass) mycorrhizal fungal colonization increased by 46% when the plants were grown at elevated CO2 (Monz 1994). In the same study, mycorrhizal colonization was unchanged by elevated CO2 in (a C3 grass). Photosynthetic rates in B. are known to increase at elevated CO2 (Morgan 1994) which should result in higher carbohydrate productivity. In several studies, changes in mycorrhizal colonization, under ambient CO2 concentrations, have been positively correlated with soluble carbohydrate concentrations in the root cortical cells (Schwab 1991). If excess soluble carbohydrates from increased photosynthetic rates were directed to the root then this could account for the increased levels of mycorrhizal colonization in this species. However, conflicting studies show no correlation between soluble carbohydrate levels in root cortical cells and the extent of mycorrhizal colonization (Schwab 1991). In studies carried out in experimental ecosystems, composing six abundant plant species of limestone grassland (for a description of the experiment see Leadley and K6rner, Chapter 11), the percentage of root length in the total plant community that was occupied by mycorrhizal fungi was not significantly affected by elevated CO2 concentrations of either 500 or 650/xmol mo1-1, as compared to the ambient CO2 concentration of 350 /xmol mo1-1. In this investigation, total root and shoot mass were also not significantly changed by the CO2 treatments (Leadley and K6rner, Chapter 11). However, in the same experiment, Leadley and K6rner observed a change in the structure of the community which occurred as a result of a detrimental growth effect on at elevated levels of CO2. Mycorrhizal fungal responses were also observed in (Fig. 1) with a significant increase in the proportion of the root length of this species being occupied by mycorrhizal fungi. In contrast, a coexisting species, exhibited a significant decrease in mycorrhizal colonization (Fig. 1), although there was no significant effect of CO2 treatment
.N,_
Figure 1 Percentage root length colonized by mycorrhizal fungi in and grown at three concentrations of CO2. The CO2 X species interaction is significant at P - 0.02, as determined by ANOVA. Different letters above bars indicate a significant difference (P -< 0.05) according to the LSD test.
on plant biomass in this species. This finding indicates the potential for differential mycorrhizal responses in the c o m p o n e n t species of a community. This study only incorporates measurements of the structures of the mycorrhizal fungi in the plant. Responses of the extraradical hyphae of these fungi to elevated CO2 are presently unknown. The studies of changes in mycorrhizal colonization at elevated CO2 in and and those by Monz (1994) on and are the first steps to understanding whether there could be a potential effect of CO2 enrichment on the mycorrhizal symbiosis. Now that changes in the fungal colonization have been observed in a few plants, more detailed studies are required to elucidate the cause (i.e., what may cause mycorrhizal colonization to change, and what the result of this change is on the functioning of the symbiosis, host nutrition, and growth). The response of fungal pathogens of plants to elevated CO2 concentrations is poorly understood. Powdery mildew, caused by the obligate biotroph of grasses was significantly more severe at 700/zmol mol 1 CO2 concentration than at ambient CO2 (Thompson 1993). The disease severity was attributed to an increase in the plant water content that occurred as a host response to the elevated CO2 treatment.
The speculative outcomes of changes in the interactions between symbiotic fungi and plants in a CO2-enriched environment are summarized in Fig. 2. In the mycorrhizal symbiosis interactions leading to either a more mutualistic, a less mutualistic, or even a parasitic association are conceivable (Fig.
17. Plant-Fungal Interactions
=-[
Possibleconsequences of elevated COz on mutualistic and parasitic plant-fungal interactions. Interactions that tend to the left become more mutualistic and those that tend to the right become more pathogenic under the influence of elevated C02.
2). Our studies indicate that the latter of these may be affected in at elevated COz as the colonization by mycorrhizal fungi increases while growth of the host plant decreases; the mycorrhizal fungi may be acting as a greater sink for photosynthates, causing the detrimental growth effect. Interestingly, in other studies mycorrhizal fungi were proposed as possible sinks for excess assimilated carbon at elevated COz. Diaz (1993) suggested that nonmycorrhizal plants increased carbon exudation into the soil which results in an expansion of the soil microflora. Diaz (1993) also observed a detrimental effect of CO2 on the growth of nonmycorrhizal plants and attributed this response to an increase in nutrient sequestration by the expanded soil microflora. Experimental investigations by Jakobsen and Rosendahl (1990) suggested that more carbon is invested outside of the roots when in the mycorrhizal state than in the nonmycorrhizal state. In ectomycorrhizal symbioses, mycorrhizas alter both the quality and quantity of carbon allocated belowground (Rygiewicz and Andersen, 1994). Another possible interpretation of the results of Diaz (1993) is that external mycorrhizal hyphae originating in the roots of the coexisting plant
species could be exploiting the soil more efficiently (Fig. 2), thereby limiting nutrient availability to nonmycorrhizal plants. To get a clearer picture of the significance of these findings to ecosystem functioning requires more in-depth studies on the responses of fungal structures to elevated CO2 (e.g., arbuscules involved in nutrient exchange, vesicles for storage of reserves, and extraradical hyphae involved in nutrient acquisition as well as the responses of the plants to changes in their efficiency of nutrient acquisition). It is essential to know how the movement of fixed carbon belowground and the acquisition of nutrients in the ecosystem are affected by changes in symbiotic interactions. In addition, it is essential to understand exactly what effect changes in total root length colonized by fungi might mean for the functioning of the mycorrhizal symbiosis. Does an increase in root length colonization at elevated CO2 lead to greater or lesser plant biomass productivity and how does this affect the phosphorus nutrition of the host plant? Irrespective of whether mycorrhizal responses to CO2 enrichment will affect ecosystem functioning, the fact that arbuscular mycorrhizal fungi respond differently to CO2 enrichment in the few different host species that have been tested leads to the suggestion that such responses could significantly affect the population biology of some species and cause changes to plant community structure. The impact of this is, however, dependent on whether the mycorrhizal symbiosis responds differently in many plant hosts. One predicted outcome of the response of mycorrhizal fungi in would be a reduction of this species in grassland communities which would lead to a change in community structure. Such changes in community structure would also be likely as a result of the detrimental growth ofnonmycorrhizal plants at elevated CO2 concentrations as observed by Diaz (1993). Similarly, if mycorrhizal fungi increased the ability of some plant species to acquire nutrients, as suggested above, the dominance in plants within communities would change. The response of fungal pathogens to elevated CO2 can cause greater disease severity (Thompson 1993). If, however, only the plant, and not the fungus, responds positively to CO2 enrichment, then the fungus may become less virulent; the outcome of such a response being that the interaction between plant and fungus would be more neutral (Fig. 2). In the case of other plant-fungal interactions, the effects of the fungi are not always clearly mutualistic or pathogenic. The fungal endophyte spp. can prevent flower development in grasses but may also cause increased growth of the host (Marks and Clay, 1990). Marks and Clay (1990) demonstrated that fungal-induced host plant growth was unaffected by CO2 enrichment, although there was a main effect of CO2 treatment on plant growth. The response of the fungus in this situation is not known but is essential to the full understanding of the outcome of this interaction at elevated
17.
Three outcomes of C O 2 e n r i c h m e n t on this interaction are possible (Fig. 2). If the fungus grows relatively better than the plant at elevated CO2 concentrations then it could potentially prevent m o r e flowers from developing. In this situation the fungus would have a m o r e pathogenic effect as it would reduce reproductive fitness and limit genetic variability in the host population. If both partners respond positively to elevated CO2 concentrations then the o u t c o m e of the relationship will remain u n c h a n g e d and the same percentage of flowers will develop as at ambient CO2 concentrations. A third o u t c o m e would occur if the plant can outgrow the fungus. A greater percentage of flowers may reach maturity before the fungus can prevent their development. In this situation the interaction is then less pathogenic but is unlikely to be more mutualistic. C O 2.
The strong influences of p l a n t - f u n g a l interactions on plant populations, communities, and the functioning of ecosystems could be modified in a CO2-enriched environment. Modifications of p l a n t - f u n g a l interactions, either b e c o m i n g m o r e mutualistic, m o r e pathogenic, or b e c o m i n g neutral, could occur. Little information currently exists on the extent and ways in which these interactions may be modified. Evidence from our studies on the mycorrhizal symbiosis suggest that positive responses in the growth of the fungus at elevated CO2 do not result in a m o r e mutualistic association in the case of one plant species but the causes and effects of this change in the fungal colonization are not e n o u g h to conclude that it results in a less mutualistic association. The same responses were not observed in a n o t h e r coexisting plant species in the c o m m u n i t y indicating that such responses are host species specific. Such effects would be likely to modify the structure of plant communities. Similarly, changes in the o u t c o m e of plantpathogenic fungus interactions at elevated CO2 concentrations could also have strong effects on plant population biology. The outcomes of such changes would alter reproductive fitness and genetic variability in plant populations and should therefore be investigated in m u c h m o r e detail if successful predictions are to be made in the future regarding vegetation responses to a CO2-rich world.
This work was supported by a grant from the Swiss National Science Foundation under the Priority Programme Environment. I thank Professor A. Wiemken, Drs.J. Arn6ne, P. Leadley, and K. Groppe, and an anonymous reviewer for constructive criticisms of the manuscript.
Carey, P. D., Fitter, A. H., and Watkinson, A. R. (1992). A field study using the fungicide benomyl to investigate the effect of mycorrhizal fungi on plant fitness. 90, 550-555. Carlsson, U., and Elmqvist, T. (1992). Epidemiology of anther-smut disease and numeric regulation of populations of 90, 509-517. Carlsson, U., Elmqvist, T., Wennstr6m, A., and Ericson, L. (1990). Infection by pathogens and population age of host plants. J. 78, 1094-1105. Clay, IL (1991). Parasitic castration of plants by fungi. 6, 162-166. Diaz, S., Grime,J. P., Harris, J., and McPherson, E. (1993). Evidence of a feedback mechanism limiting plant response to elevated carbon dioxide. 364, 616-617. Dobson, A., and Crawley, M. (1994). Pathogens and the structure of plant communities. 9, 393-398. Grime, J. P., Mackey, J. M., Hillier S. M., and Read, D.J. (1987). Floristic diversity in a model system using experimental microcosms. 328, 420-422. Ineichen, K., Wiemken, V., and Wiemken, A. (1995). Shoots, roots and ectomycorrhiza formation of pine seedlings at elevated atmospheric carbon dioxide. 18, 703-707. Jakobsen, I., and Rosendahl, L. (1990). Carbon flow into soil and external hyphae from roots of cucumber plants. 115, 77-83. Koide, R. T., Shumway, D. L., Mabon, S. A. (1994). Mycorrhizal fungi and the reproduction of field populations of Medic. (Malvaceae). 126, 123-130. Marks, S., and Clay, K. (1990). Effects of CO2 enrichment, nutrient addition, and fungal endophyte-infection on the growth of two grasses. 84, 207-214. Minchella, D.J., and Scott, M. E. (1991). Parasitism: A cryptic determinant of animal community structure. 6, 250-254. Monz, C. A., Hunt, H. W., Reeves, F. B., and Elliot, E. T. (1994). The response of mycorrhizal colonization to elevated CO2 and climate change in and 165, 75-80. Morgan, J. A., Knight, W. G., Dudley, L. M., and Hunt, H. W. (1994). Enhanced root system C-sink activity, water relations and aspects of nutrient acquisition in mycotrophic subjected to COz enrichment. 165, 139-146. Norby, R. J., O'Neill, E. G., Hood, W. G., and Luxmoore, R.J. (1987). Carbon allocation, root exudation and mycorrhizal colonization of seedlings grown under CO2 enrichment. 3, 203-210. Rygiewicz, P. T., and Andersen, C. P. (1994). Mycorrhizae alter quality and quantity of carbon allocated below ground. 369, 58-60. Sanders, I. R., and Koide, R. T. (1994). Nutrient acquisition and community structure in cooccurring mycotrophic and nonmycotrophic old-field annuals. 8, 77-84. Schmid, B. (1994). Effects of genetic diversity in experimental stands of Evidence for the potential role of pathogens as selective agents in plant populations. 82, 165-175. Schwab, S. M., Menge, J. A., and Tinker, P. B. (1991). Regulation of nutrient transfer between host and fungus in vesicular-arbuscular mycorrhizas. 117, 387-398. Thompson, G. B., Brown, J. K. M., and Woodward, F. I. (1993). The effects of host carbon dioxide, nitrogen and water supply on the infection of wheat by powdery mildew and aphids. 16, 687-694.
18 Effects of Elevated CO2 on Plants Grown in Competition
Plants and other photosynthetic organisms are the primary entry point of atmospheric CO2 into the terrestrial biosphere. Understanding vegetation responses to elevated CO2 is, therefore, critical for predicting how terrestrial ecosystems will change in response to rising atmospheric CO2. Much attention has been focused at the level of individual plant species responses to elevated CO2. Under optimal growth conditions common responses include enhanced photosynthetic rate, water use efficiency (WUE), and growth (Oechel and Strain, 1985; Bazzaz, 1990; K6rner, 1993; Poorter, 1993; Poorter Chapter 25). Photosynthesis and growth responses are clearly influenced by photosynthetic pathway (C3 v e r s u s C a ) , and enzymatic differences make C~ species inherently more responsive to increases in CO2 concentration than C4 species. Less is known about CO2 responses of species grown in mixtures. Inherent differences in photosynthetic pathway should be important in understanding responses of species grown in mixtures (Fig. la). However, interspecific interactions may modify such responses. Bazzaz and McConnaughay (1992) note that, due to its gaseous nature, CO2 itself should rarely be a contested resource, except perhaps in dense stands of vegetation where significant drawdown may occur. However, positive growth responses to addition of one resource, such as CO2, require that other resources do not limit growth (Field 1992) (Fig. lb). Others (Oechel and Strain, 1985; Bazzaz, 1990; 273
Copyright 9 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.
(a)
Figure I Predictedresponses of species to elevated CO2under conditions of (a) nonlimiting resources, (b) severe resource limitation (c) limiting soil resources, and (d) abundant soil resources.
Bazzaz and McConnaughay, 1992) have suggested that this may explain why productivity of nutrient-rich salt-marsh communities significantly increases in response to elevated CO2 (Drake, 1992) but nitrogen-limited tundra communities do not show sustained increases in productivity (Oechel 1994). Even when total community productivity does not change, responses of particular component plant species and subsequent changes in community composition with elevated CO2 may occur (Bazzaz and Garbutt, 1988). At present no theoretical framework exists for predicting how elevated CO2 will modify the competitive balance among species. I suggest that species differences in ability to compete for potentially limiting resources can have large effects on species responses to elevated CO2. Depending on resource levels, differences in competitive ability should be a critical determinant of whether any gain in photosynthate or WUE under elevated CO2 can be converted into improved growth and performance. When soil nutrients or water are limiting, superior competitors for belowground resources should be favored by elevated CO2 (Fig. lc). Examples are species with high allocation to root (Tilman, 1988), high specific root length (Elberse and Berendse, 1993), or with mycorrhizal or Nz-fixing symbionts (L~scher Chapter 19). Soil nutrient limitation may be
18.
intensified by elevated C O 2 because of potential increases in the flow of labile carbon to soil through root exudation, followed by stimulation of free-living soil microbes and consequent nutrient sequestration (Diaz 1993; although see Lamborg 1984; Zak 1993). If so, mycorrhizal and Nz-fixing species may be especially favored, if these microbial symbionts serve as a sink for excess photosynthate, preventing its flow to free-living soil microbes and directing the beneficial effects of increased photosynthate on microbial growth back to the host plant as increased mineral nutrient supply (Goudrian and de Ruiter, 1983; Strain and Bazzaz, 1983; Diaz 1993). Conversely, when belowground resources are plentiful and light is limiting, superior competitors for light should be favored by elevated CO2 (Fig. l d). Examples are species with high allocation to stems and leaves, and high specific leaf area (SLA) (Tilman, 1988; Elberse and Berendse, 1993). In the following sections I review the literature on responses to elevated CO2 of species grown in mixtures, evaluating the importance of differences in ability to compete for limiting resources versus differences in photosynthetic pathway in determining differential species responses. When focusing on competitive interactions it is essential to group studies into those conducted under moderate to high soil resource conditions (potentially lightlimited systems) and those conducted under limiting soil resource conditions, because this will determine whether species with high belowground or aboveground allocation are superior competitors. In addition, as Diaz (1993) have pointed out, a further distinction can be made between studies depending on the character of the soil microbial communitym artificial soil mixes are likely to have depleted microbe communities relative to natural or partially natural soils. I consider mixtures of C3 and C4 species first and then mixtures composed solely of C3 species (mixtures of C4 species have received little attention thus far).
A. Cn v e r s u s
C4
Species
Studies conducted in greenhouses or growth chambers under well-fertilized, well-watered conditions have consistently shown that performance of C4 species declines with increasing CO2 whereas that of C~ species increases with CO2 when grown in mixtures (Carter and Peterson, 1983; Patterson 1984; Wray and Strain, 1987; Bazzaz and Garbutt, 1988). Total mixture biomass was unchanged by elevated CO2 in these studies, and gains in biomass of C3 species were counterbalanced by declines in biomass of C4 species, suggesting that competition
276
for a limiting resource(s) occurred. Because plants were well watered and fertilized in these studies, the limiting resource was probably light. Is there evidence that elevated CO2 enhanced performance of the superior competitor (for light)? In the Wray and Strain (1987) study, the C3 species was indeed the better competitor at ambient COs, and its inherently greater size and leaf area per plant suggest that it would have had an advantage in competition for light. Similar patterns of competitive ability and inherent size were found for C3 species versus C4 species in the Bazzaz and Garbutt (1988) study. In the other studies, however, there were either no differences in competitive ability between C~ and C4 species at ambient COs (Patterson 1984), or the C3 species was actually a poorer competitor relative to the C4 species at ambient COs (Carter and Peterson, 1983). These latter studies suggest that the greater response of C~ species to elevated COs due to photosynthetic pathway may give them a competitive edge over C4 species. A possible mechanism for this is that enhanced photosynthesis of C3 species under elevated COs leads to increases in growth early on, before resources become limiting. Such increases in size may then enhance the ability of C3 species to acquire resources in competition with C4 species. Thus, early growth responses of seedlings to elevated CO2 may strongly influence competitive balance (Bazzaz 1989). In contrast to the studies described above, at least two studies have found increases in total mixture aboveground biomass and total aboveground biomass of C~ species with elevated COs, with no changes in total biomass of C4 species (Zangerl and Bazzaz, 1984; except in the lowest nutrient treatment), even at low soil moisture (Bazzaz and Carlson, 1984). Such responses suggest that in these studies resources were not limiting, and competitive interactions between C3 and C4 species were minimal. In such situations, photosynthetic pathway appears to be a good predictor of differential species responses. At least one growth-chamber study found that elevated COs enhanced competitive ability of a C4 species more than that of a C3, even though the C3 species was the better competitor at ambient COs due to its greater seed size, and had a greater increase in rate of photosynthesis per unit leaf with elevated CO2 (Bazzaz 1989). The Ca species showed an earlier growth response to elevated COs than the C~ species, and Bazzaz hypothesized that this difference in timing of COs response was responsible for the relatively greater enhancement in competitive performance of the C4 species. Studies conducted in a highly productive, nutrient-rich, salt-marsh community showed increases in total mixture biomass as well as increases in biomass and percentage composition of C~ species in response to elevated C02, with little if any
18.
277
responses of C4 species (Curtis 1989, 1990; Drake, 1992; Drake and Leadley, 1991). Although light is expected to be limiting in nutrient-rich, high-productivity systems (Tilman, 1988), these responses again suggest minimal competitive interactions and point to the importance of photosynthetic pathway in predicting responses of C~ and C4 species when resources are not limiting. In contrast, studies conducted in a productive but drought-susceptible tallgrass prairie system found that elevated CO2 favored dominant C4 grasses over a subdominant C3 grass (Owensby 1993; Mo 1992). Positive growth responses of C4 dominants were due to improved water relations (Owensby 1993). Light competition from the taller C4 grasses was implicated in the C~ grass's inability to respond to elevated CO2, although nutrient competition from the C4 grasses may have also been a factor (Owensby 1993). Phenology of the C~ grass was actually earlier than that of the C4 dominants, but low N mineralization rates prevented the C3 grass from responding to elevated CO2 earlier in the growing season. Percentage composition and basal cover (although not aboveground biomass or leaf area) of subdominant C~ forbs in the system did increase in response to elevated CO2, and Owensby (1993) hypothesized that differences in roofing depth may have reduced competition between C~ forbs and C4 dominants for soil nutrients and water. Zangerl and Bazzaz (1984) reported that increases in C3 biomass with elevated CO2 were offset by decreases in C4 biomass, resulting in no significant differences in aboveground biomass of a C3-C4 mixture grown in unfertilized sandy soil in a glasshouse. Competition for soil nutrients may have been a factor, because with fertilization the reductions in C4 biomass with elevated CO2 disappeared; however, no data on mechanisms of competition were presented in this study. No studies were found that involved mixtures of C~ and C4 species grown in lower fertility, natural or partially natural soil. This is an important area for research, however, since differences in mycorrhizal infection o r N 2 fixation are likely to have the greatest effects on species responses to elevated CO2 at lower soil fertilities. Results from laboratory experiments with ectomycorrhizae (Norby 1987), VAM mycorrhizae (Monz 1994), and N2 fixers (L~scher Chapter 19) indicate that growth ofsymbionts are stimulated under elevated CO2. However, this area has received relatively little attention thus far, and more research is needed (Stulen and den Hertog, 1993). Mycorrhizae could prove important in understanding crop-weed responses to elevated CO2, because many crop-weed combinations occur between C~ and Ca plants, yet many C3 weeds are not normally mycorrhizal, whereas many crop species are (Francis and Read, 1994). Furthermore, C3
278
and C 4 prairie grasses have also been shown to differ in their degree of dependence on mycorrhizae; dominant C4 grasses such as Vitman are obligately mycorrhizal whereas C~ grasses such as L. are facultatively mycorrhizal (Hetrick 1988, 1990). To summarize (Table 1), out of 14 studies examining CO2 responses of C~-C4 mixtures, 7 were conducted in moderate-to-rich, artificial soil; 6 in moderate-to-rich, natural or partially natural soil; 1 in lower fertility, artificial soil; and 0 in lower fertility, natural soil. Of the 13 studies conducted in moderate-to rich, artificial or partially natural soil, 6 observed positive growth responses of C3 species, with no response of C4 species (Zangerl and Bazzaz, 1984; Bazzaz and Carlson, 1984; Curtis 1989, 1990; Drake and Leadley, 1991; Drake, 1992). In these studies competitive interactions appeared to be minimal and results to reflect inherent differences in C3 and C4 physiology. Of the remaining 7 studies, in which competitive interactions appeared to play a role, 4 provided support for the idea that superior competitive ability for light was important in predicting CO2 response at high soil fertility (Wray and Strain, 1987; Bazzaz and Garbutt, 1988; Mo 1992; Owensby 1993). Of the other 3 studies, early timing of CO2 response was implicated in 1 study (Bazzaz 1989), and I suggested that this could be a factor in explaining competitive outcome in the remaining 2 studies as well (Carter and Peterson, 1983; Patterson 1984). Competition for soil nutrients appeared likely in the single study conducted in lower fertility soil (Zangerl and Bazzaz, 1984), however no information on mechanisms of competition were presented or could be inferred from this study. B. Cn versus C3 Species Differential responses to elevated CO2 have been observed among tropical trees (Reekie and Bazzaz, 1989), temperate forest saplings (Williams 1986), and annual forbs (Reekie and Bazzaz, 1991), all grown in glass chambers. These studies were conducted either in fertilized and watered soil mixes (Reekie and Bazzaz, 1989), or in unfertilized loam-peat-perlite mixes (Williams 1986; Reekie and Bazzaz, 1991). Only Williams (1986) reported whether total mixture biomass changed with elevated CO2; it did not. In support of the expectation that responses to elevated CO2 will be determined primarily by competition for light when soil resources are abundant, Reekie and Bazzaz (1989) found that changes in mean canopy height under elevated CO2 explained 60% of the variation in competitive performance. However, it is unclear why mean canopy height responded differently to elevated CO2 in different species, and in particular, why mean canopy height was reduced by elevated CO2 in the species with the greatest mean canopy height (and the greatest
O u t c o m e of competition Superior competitors wins
C3 wins C~ vs C4
Fertile soil Artificial soil
References
0% (n = 2)
100%
Carter and Peterson, 1983; Patterson 1984; Wray and Strain, 1987; Bazzaz and Garbutt, 1988; Bazzaz 1989 Mo 1992; Owensby 1993
100% (n = 1)
No data
Zangerl and Bazzaz, 1984
80% ( n -
5)
40% Light competitor wins
Natural soil Infertile soil Artificial soil
Natural soil C3 vs C3
Fertile soil Artificial soil
}
Soil resource competitor wins No data
No studies NA a
Natural soil
NA
Infertile soil Artificial soil Natural soil
NA NA
Light competitor wins
33% (n = 3) 0% ( n =
Soil resource competitor wins
2),? (n=
0% (n = 1) 100% (n = 2)
Studies in which competitive interactions appeared to be minimal were excluded. a NA, Not applicable.
1)
Reekie and Bazzaz, 1989; Reekie and Bazzaz, 1990; Williams 1986 K6rner and Amone, 1992; Diaz 1993; Jackson submitted for publication Williams 1988 Diaz 1993; Chiariello and Field, Chapter 10
280
abundance) at ambient CO2. No mechanisms for observed changes in species performance were suggested by the other studies. One case of equally positive responses to elevated CO2 has been reported for pairs of congeneric temperate forest tree seedlings grown in a glasshouse, resulting in doublings of total pair biomass compared with ambient CO2 (Rochefort and Bazzaz, 1992). Seedlings were fertilized, watered as necessary, and harvested after 3 months of growth, so it is likely that competition for soil resources and light were minimal and the seedlings' positive responses were a simple consequence of their C~ photosynthetic pathway. Unlike C3 species of productive salt-marsh systems (Curtis 1990), the dominant species in a productive tall-herb system grown in natural topsoil in growth cabinets were not able to convert additional photosynthate into increased growth under elevated CO2 (Diaz 1993). Rather, elevated CO2 resulted in no changes in total mixture biomass, accumulations of nonstructural carbohydrate in leaves, leaf discoloration, and increased belowground carbon (C), even if supplementary mineral nutrients were given (Diaz 1993). Diaz (1993) also measured significant increases in soil microbial C and nitrogen (N) and suggested that increased C effiux from plant roots to soil had tipped the balance of plant-microbe competition for soil nutrients in favor of microbes. The species were nonmycorrhizal, and as discussed below, Diaz suggested that mycorrhizal species might fare better under elevated CO2. Similar observations were made in model productive humid tropical ecosystems situated in a greenhouse (K6rner and Arnone, 1992) and it would be interesting to know whether the species examined were also nonmycorrhizal or, if they were mycorrhizal, whether they were likely to have been infected with mycorrhizal fungi. The difference in response to elevated CO2 between salt marsh systems and productive terrestrial systems may be a consequence of the continuous replenishment of nutrients to salt marshes with the tides, which would likely mitigate any plant-microbe competition. In contrast, Jackson (submitted for publication) found increases in photosynthesis and density of naturally growing a moderatefertility grassland species, under elevated CO2. Average VAM mycorrhizal infection for this grassland system has been measured at 70% (Whitbeck, 1994), and average mycorrhizal colonization of at 47% (Koide and Mooney, 1987) and 51-75% (Hopkins, 1987). These studies suggest that with elevated CO2, soil resources may be limiting to growth responses even in moderate-to-rich soil, perhaps because of increased plant-microbe competition. Species with mycorrhizae may be able to avoid competition for soil nutrients to some extent, and so realize potential growth responses to elevated CO2.
18.
281
No biomass responses to elevated CO2 were found in a mixture of six early-season serpentine annual species grown under well-watered conditions in a sandy soil mix in the greenhouse, despite the fact that two of these species did respond positively when grown individually at a lower density (Williams 1988). Biomass of species grown in mixture was significantly reduced compared to that of individually grown species, indicating that competition did occur. Although the soil mix used was probably richer than the natural soil of these species, it may still have been infertile enough to constrain growth responses to elevated CO2, particularly when plants were grown in competition. Consistent with this, percent nitrogen significantly decreased with elevated CO2 in three of the competitively grown species, and there were nonsignificant trends in the same direction for the other three species. No changes in plant community biomass were observed under elevated CO2 in an field study conducted in serpentine grassland, although a deep-rooted, late-season annual did significantly increase biomass in response to elevated CO2, apparently due to improved water relations (see Chiariello and Field, Chapter 10). Due to senescence of early-season annuals, plant density in serpentine grassland is much lower later in the season, and so reduced competition for water or soil nutrients may also have contributed to the late-season annual's positive growth response to elevated CO2. Consistent with the study by Williams (1988), Chiariello and Field reported no changes in biomass of early-season annuals. Water limitation may have been important in constraining field responses of early-season annuals, although such species were well watered in the Williams study yet still did not respond to elevated CO2 in mixture (however, as discussed above, this may have been due to nutrient limitation). Serpentine species are typically 76-100% infected with VAM mycorrhizae in the field (Hopkins, 1987), yet it may be that soil fertility is low enough in this notoriously low-fertility soil (Proctor and Woodell, 1975; Turitzin, 1982) to constrain CO2 responses even of mycorrhizal species. Furthermore, the ubiquity of mycorrhizal associations in this system decreases opportunity for differential competitive ability for soil nutrients. Diaz (1993) also did not find changes in total aboveground biomass with elevated CO2 for a low fertility acidic grassland mixture grown in natural topsoil in growth cabinets. However, a nonsignificant shift in species composition occurred in favor of a mycorrhizal species. This suggests that when species are infected, mycorrhizae may be important in understanding shifts in species composition in response to elevated CO2, even when soil fertility is too low to support a community-level response (Diaz 1993).
282 To summarize (Table I), out of 10 studies examining C O 2 responses of C~-C3 mixtures, 4 were conducted in moderate-to-rich, artificial soil, 3 in moderate-to-rich, natural or partially natural soil, 1 in lower fertility, artificial soil, and 2 in lower fertility, natural or partially natural soil. Of the 3 studies conducted in moderate-to-rich, artificial soil in which competitive interactions appeared to play a role, 1 provided support for the idea that superior competitive ability for light was important in predicting CO2 response at high soil fertility (Reekie and Bazzaz, 1989). No mechanisms for observed changes in species performance were suggested by the other 2 studies (Williams 1986; Reekie and Bazzaz, 1991). Of the 3 studies conducted in moderate-to-rich, natural or partially natural soil, plantmicrobe competition may have been a factor in constraining plant responses in 2 studies (K6rner and Arnone, 1992; Diaz 1993). In the third study (Jackson submitted for publication) a mycorrhizal species showed positive growth responses to elevated CO2, in line with Diaz suggestion that mycorrhizae may be important to plant response to elevated CO2. The importance of light competition could not be inferred from this study, however. Of the 3 studies conducted in lower fertility soil, severe soil resource limitation appeared to constrain responses in 2 studies (Williams 1988; see Chiariello and Field, Chapter 10), except for a species with access to deep water supplies (see Chiariello and Field, Chapter 10). In the remaining study conducted in partially natural soil, plant-microbe competition was again inferred (Diaz 1993), but appeared to be alleviated in mycorrhizal species.
Differential responses to elevated C O 2 of species grown in mixtures have been observed in the majority of studies examining this question. When competitive interactions are minimal, photosynthetic pathway is a useful predictor of performance, and C3 species typically increase performance relative to C4 species under elevated CO2. When competitive interactions are strong and soil fertility is high, elevated CO2 may favor species that are superior competitors for light, regardless of their photosynthetic pathway (Wray and Strain, 1987; Bazzaz and Garbutt, 1988). However, there are some studies which suggest that the timing of CO2 response can be of greater importance than photosynthetic pathway or competitive ability at ambient CO2 (Patterson 1984; Carter and Peterson, 1983; Bazzaz 1989). Only one study examined responses of C~-C4 mixtures grown at low soil fertility to elevated CO2, finding that performance of C3 species was favored (Zangerl and Bazzaz, 1984). The study, however, was not designed to ad-
18.
283
dress changes in competitive ability with elevated C O 2. A potentially critical but unexplored issue is how differences in mycorrhizal infection or N2 fixation could affect the outcome of competition between C3 and C4 species at different soil fertilities. This issue could be very important in understanding and manipulating C4 crop-C3 weed interactions, as well as crop-weed interactions in general. In the absence of differences in photosynthetic pathway, I expected that differences in ability to compete for limiting resources would be even more important in determining responses of interspecifically grown species to elevated CO2. This expectation was unconfirmed, however, because most of the relevant studies were not designed to examine changes in competitive ability or in mechanisms of competition with elevated CO2. Some studies of C~ mixtures exposed to elevated CO2 have supported the idea that soil resource limitation may constrain responses to elevated CO2 (Diaz 1993; Williams 1988). Such resource limitation may occur even in higher fertility soil, presumably because of increases in C flow to soil, and consequent increases in plant-microbe competition (Diaz 1993). Species differences in mycorrhizal infection (Diaz 1993) or in N2 fixation (see Lascher Chapter 19) are therefore potentially critical predictors of differential species responses to elevated CO2, although thus far this expectation has little support. More studies in which physiological and morphological traits are measured and related to performance in mixture are needed to determine whether there are consistent subgroupings of species that respond in predictable ways to elevated CO2. In order to determine whether differences in competitive ability for resources are important in determining responses to elevated CO2, appropriate competition designs should be included (Underwood, 1986), and attention should be paid to whether soil resources or light are limiting to growth. Traits that may be particularly important to competitive ability in an elevated C O 2 environment are photosynthetic pathway, mycorrhizae, and N2 fixation. Traits that have not been discussed thus far are relative growth rate (RGR) and plasticity, or flexibility of physiology or morphology. Very little attention has been paid to these two traits in the context of differential responses of species grown in mixtures to elevated CO2, yet they are potentially important. Potentially fast-growing species (crop species, ruderals) may have a greater sink strength for carbon, and in experiments involving species grown individually under nonlimiting resource conditions fast-growers have shown greater responses to elevated CO2 than slower growing species (Poorter, 1993; Poorter Chapter 25). In the face of changes in resource availability such as increases in CO2, species with high trait plasticity may be better able than less plastic species to shift physiology or morphology so as to maintain balanced resource acquisition.
284
Finally, studies that include as many species as possible (Poorter, 1993) and that are conducted in natural soil have a higher likelihood of finding functional groupings relevant to understanding responses of natural communities to elevated CO2. For this reason it may be more productive to focus on measuring traits and performance of species grown in relatively complex mixtures, and comparing results between communities from different habitats, rather than to focus on simple, species-poor mixtures, except in agricultural research.
This research was supported by National Science Foundation Grant BSR90-20135. The manuscript benefited greatly from comments by F. S. Chapin III, C. D'Antonio, Ch. K6rner, and an anonymous reviewer.
Bazzaz, F. A. (1990). The response of natural ecosystems to the rising global CO2 levels. 21, 167-196. Bazzaz, F. A., and Carlson, R. W. (1984). The response of plants to elevated CO2 I. Competition among an assemblage of annuals at two levels of soil moisture. 62, 196-198. Bazzaz, F. A., and Garbutt, K. (1988). The response of annuals in competitive neighborhoods: Effects of elevated CO2. 79, 223-235. Bazzaz, F. A., and McConnaughay, K. D. M. (1992). Plant-plant interactions in elevated CO2 environments. 40, 547-563. Bazzaz, F. A., Garbutt, K., Reekie, E. G., and Williams, W. E. (1989). Using growth analysis to interpret competition between a C3 and a C4 annual under ambient and elevated CO2. 79, 223--235. Carter, D. R., and Peterson, K. M. (1983). Effects of a CO2-enriched atmosphere on the growth and competitive interaction of a C3 and a C4 grass. 58, 188-193. Curtis, P. S., Drake, B. G., Leadley, P. W., Arp, W., and Whigham, D. (1989). Growth and senescence of plant communities exposed to elevated COz concentrations on an estuarine marsh. 78, 20-26. Curtis, P. S., Balduman, L. M., Drake, B. G., and Whigham, D. F. (1990). Elevated atmospheric CO2: Effects on belowground processes in C3 and C4 estuarine marsh communities. 71 (5), 2001- 2006. Diaz, S., Grime,J. P., Harris, J., and McPherson, E. (1993). Evidence of a feedback mechanism limiting plant response to elevated carbon dioxide. 364, 616-617. Drake, B. G. (1992). A field study of the effects of elevated CO~ on ecosystem processes in a Chesapeake Bay wetland. 40, 579-595. Drake, B. G., and Leadley, P. W. (1991). Canopy photosynthesis of crops and native plant communities exposed to long-term elevated CO2. 14, 853-860. Elberse, W. Th., and Berendse, F. (1993). A comparative study of the growth and morphology of eight grass species from habitatswith 7,223-229. Field, C. B., Chapin, F. S., III, Matson, P. A., and Mooney, H. A. 1992. Responses of terrestrial ecosystems to the changing atmosphere: A resource-based approach. 23, 201-235.
18.
285
Francis, R., and Read, D.J. (1994). The contributions ofmycorrhizal fungi to the determination of plant community structure. 159, 11-25. Goudrian, J., and de Ruiter, H. E. (1983). Plant growth in response to CO2-enrichment, at two levels of nitrogen and phosphorus supply. 1. Dry matter, leaf area, and development. 31, 157-169. Hetrick, B. A. D., Gerschefske Kitt, D., and Wilson, G. W. T. (1988). Mycorrhizal dependence and growth habit of warm-season and cool-season tallgrass prairie plants. 66, 1376-1380. Hetrick, B. A. D., Wilson, G. W. T., and Todd, T. C. (1990). Differential responses of C~ and C4 grasses to mycorrhizal symbiosis, phosphorus fertilization, and soil microorganisms. 68, 461-467. Hopkins, N. A. (1987). Mycorrhizae in a California serpentine grassland community. 65, 484-487. Jackson, R. B., Luo, Y., Cardon, Z. G., Sala, O. E., Field, C. B., and Mooney, H. A. (1995). Photosynthesis, growth, and density for the dominant species in a CO2-enriched grassland. 22, 221-225. Koide, R. T. and Mooney, H. A. (1987). Spatial variation in inoculum potential of vesiculararbuscular mycorrhizal fungi caused by formation of gopher mounds. 107, 173-182. K6rner, C. (1993). CO2 fertilization: The great uncertainty in future vegetation development. "Vegetation Dynamics and Global Change" (A. M. Solomon and H. H. Shugart, eds.), pp. 53-70. Chapman & Hall, New York. K6rner, C., and Arnone, J. A., III. (1992). Responses to elevated carbon dioxide in artificial tropical ecosystems. 257, 1672-1675. Lamborg, M. R., Hardy, R. W. F., and Paul, E. A. (1984). Microbial effects. "CO2 and Plants: The Response of Plants to Rising Levels of Atmospheric CO2" (E. R. Lemon, ed.), pp. 131-176. American Association for the Advancement of Science Selected Symposium, Washington, DC. Mo, G., Kirkham, M. B., He, H., Ballou, L. K., Caldwell, F. W., and Kanemasu, E. T. (1992). Root and shoot weight in a tallgrass prairie under elevated carbon dioxide. 32(3), 193-201. Monz, C. A., Hunt, H. W., Reeves, F. B., and Elliot, E. T. (1994). The response of mycorrhizal colonization to elevated CO2 and climate change in and 165, 75-80. Norby, R.J., O'Neill, E. G., Hodd, W. G., and Luxmoore, R.J. (1987). Carbon allocation root exudation and mycorrhizal colonization. 3, 203-210. Oechel, W. C., and Strain, B. R. (1985). Native species responses to increased atmospheric carbon dioxide concentration. "Direct Effects of Increasing Carbon Dioxide on Vegetation" (B. R. Strain and J. D. Cure, eds.), pp. 119-154. U.S. DOE/ER-0238, National Technical Information Service, Springfield, VA. Oechel, W. C., Cowles, S., Grulke, N., Hastings, S.J., Lawrence, B., Prudhomme, T., Riechers, G., Tissue, D., and Vourlitis, G. (1994). Transient nature of CO2 fertilization in Arctic tundra. 371, 500-503. Owensby, C. E., Coyne, P. I., Ham, J. M., Auen, L. M., and Knapp, A. K. (1993). Biomass production in a tallgrass prairie ecosystem exposed to ambient and elevated COz. 3(4), 644-653. Patterson, D. T., Flint, E. P., and Beyers,J. L. (1984). Effects of CO2 enrichment on competition between a C, weed and a C~ crop. 32, 101-105. Poorter, H. (1993). Interspecific variation in the growth response of plants to an elevated ambient CO2 concentration. 104/105, 77-97. Proctor, J., and Woodell, S. R.J. (1975). The ecology of serpentine soils. 9, 255-366.
286 Reekie, E. G., and Bazzaz, F. A. (1989). Competition and patterns of resource use among seedlings of five tropical trees grown at ambient and elevated CO2. 79, 212-222. Reekie, E. G., and Bazzaz, F. A. (1991). Phenology and growth in four annual species grown in ambient and elevated COz. 69, 2475-2481. Rochefort, L., and Bazzaz, F. A. (1992). Growth response to elevated CO2 in seedlings of four 22, 1583-1587. co-occurring birch species. Strain, B. R., and Bazzaz, F. A. (1983). Terrestrial plant communities. "CO2 and Plants: The Response of Plants to Rising Levels of Carbon Dioxide" (E. R. Lemon, ed.), pp. 177-222. Westview, Boulder, Co. Stulen, I., and den Hertog, J. (1993). Root growth and functioning under atmospheric COz enrichment. 104/105, 99-115. Tilman, D. (1988). "Plant Strategies and the Dynamics and Structure of Plant Communities,"Monographs in population biology 26. Princeton Univ. Press, Princeton, New Jersey. Turitzin, S. N. (1982). Nutrient limitations to plant growth in a California serpentine grassland. 107, 95-99. Underwood, T. (1986). The analysis of competition by field experiments. "Community Ecology: Pattern and Process" (J. Kikkawa and D.J. Anderson, eds.), pp. 240-268. Blackwell Scientific Publications, Melbourne. Whitbeck, J. (1994). Effects of above- and below-ground resource distribution on the ecology of vesicular-arbuscular mycorrhizas. Ph.D. Dissertation, Stanford, CA. Williams, W. E., Garbutt, K., Bazzaz, F. A., and Vitousek, P. M. (1986). The response of plants to elevated C O 2 IV. Two deciduous-forest tree communities. 69, 454-459. Williams, W. E., Garbutt, K., and Bazzaz, F. A. (1988). The response of plants to elevated COzmV. Performance of an assemblage of serpentine grassland herbs. 28(2), 123-130. Wray, S. M., and Strain, B. R. (1987). Competition in old-field perennials under CO2 enrichment. 68(4), 1116-1120. Zak, D. R., Pregitzer, K. S., Curtis, P. S., Teeri, J. A., Fogel, R., and Randlett, D. L. (1993). Elevated atmospheric CO2 and feedback between carbon and nitrogen cycles. 151, 105-117. Zangerl, A. R., and Bazzaz, F. A. (1984). The response of plants to elevated CO2 II. Competitive 62, 412-417. interactions among annual plants under varying light and nutrients.
1 Differences between Legumes and Nonlegumes of Permanent Grassland in Their Responses to Free-Air Carbon Dioxide Enrichment: Its Effect on Competition in a Multispecies Mixture
Plant growth is generally stimulated by an increase in atmospheric C O 2 concentration, with the average increase for a doubling of the actual CO2 concentration being approximately 30% (Kimball, 1983; Newton, 1991; Poorter, 1993). Clear differences between the growth responses of C3 and C4 plants to elevated CO2 are also reported; it was demonstrated that these differences may have a strong effect on competition between plants (Bazzaz and Carlson, 1984; Smith 1987; Wray and Strain, 1987a,b; Johnson 1993). C3 species are of particular importance for temperate grasslands and within this group different responses to elevated CO2 are reported. Crop plants and fast-growing wild species increased more in weight than slow-growing wild species. C3 species capable of symbiotic N2fixation showed greater increases in weight as compared to other C~ plants (Hunt 1991, 1993; Poorter, 1993). Most of the results mentioned above were obtained from short-term experiments in controlled environments and with unlimited supply of nu287
Copyright 9 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.
trients. It is known, however, that nutrient availability (Goudriaan and de Ruiter, 1983; Conroy 1992; Conroy and Hocking, 1993; Rogers 1993) and competition (Williams 1988) may alter CO2 effects on plant growth. Thus, to predict CO2-induced changes in the species composition of plant communities, information is needed on the response of plant species to CO2 enrichment under competition and field conditions (Bazzaz, 1990; Lawlor and Mitchell, 1991; Newton, 1991; Woodward 1991; McConnaughay 1993). For long-term predictions of the development of the floristic composition of permanent grassland in a CO2-enriched environment, not only are the interspecific differences in the responses to CO2 important but also the intraspecific variability in the responses to CO2 (Patterson and Flint, 1990; Bradshaw and McNeilly, 1991; Woodward 1991). If some genotypes show stronger responses to CO2 than others, then they may have a competitive advantage at elevated CO2. As a result, evolutionary adaptation of plant populations to future changes in atmospheric CO2 concentration could result. Little is known about intraspecific variation in the response to CO2. Differences between rice cultivars in their responses to CO2 were found by Ziska and Teramura (1992). Genotypic variation within native populations in response to CO2 is reported for germination in (Wulff and Alexander, 1985) and for lifetime fecundity in (Curtis 1994). However, Fajer (1992) found no genotypic differences in as far as growth and leaf biochemistry responses to CO2 were concerned. The aim of the present experiment was to test, in the field, inter- and intraspecific differences in the responses to CO2 of seven perennial species typical of permanent temperate grasslands. We assumed that the differences between legumes and nonlegumes in their responses to CO2 can have significant effects on the species proportion of plant communities. This hypothesis was tested within a multispecies mixture in which all seven species were in competition with each other.
The effect of an increase in atmospheric C O 2 partial pressure, from ambient (35 Pa) to 60 Pa, on the harvestable biomass (cutting height 4.5 cm) of native grassland species was investigated in a field experiment at the field station of the Institute of Plant Sciences at Eschikon near Zurich (Switzerland) in 1993. The experiment was on a fertile, eutric cambisol with potassium and phosphorus contents considered adequate for intensive grassland. The FACE system (Free-Air Carbon Dioxide Enrichment) (Lewin 1992) was used to increase the atmospheric CO2 concentration. The
19.
289
experiment consisted of three blocks, each containing a fumigated (60 Pa COz) area and a control area with ambient CO2 concentration. Accuracy of COz control for the three fumigated areas was within 60 Pa _+ 6 Pa for 87-90% of the time (1-minute means). The seven plant species were grouped into two functional types according to their ability to fix nitrogen symbiotically: (1) nonlegumes: (L.) J. et C. Presl, Huds., (L.) P. B., and L.; and (2) legumes: L. and L. Seven different genotypes of each of these plant species were sampled at two sites in native grasslands near Zurich. The genotypes were vegetatively propagated in the greenhouse for 6 months. At the end of May 1993, when the grass plants had three to five tillers, plants from all species were transplanted to field plots representing two competition treatments. In the first treatment, individual plants of all seven species and genotypes were placed in artificial gaps (initial diameter 8 cm) in an established (cv. Bastion) sward. Thus each individual plant grew in competition with The distance between the gaps was 35 x 35 cm. In the second competition treatment, individuals of the seven species and genotypes were planted in a mixture (planting distance 9 x 9 cm) on bare soil. These two competition treatments were established in all three fumigated areas and in the three control areas of the FACE experiment. A total n u m b e r of 588 individuals (7 species x 7 genotypes x 2 competition treatments x 2 COz partial pressures x 3 blocks) were planted. The plants were exposed to the COz treatments from the end of May to November 1993 and were harvested three times. Annual fertilizer application was 5.5 g P m -z, 23.9 g K m -z, and 1.4 g N m -2. The experimental design was a split-split-plot with COz as the main plot factor, the competition treatment as the s u b p l o t factor, and the species and their genotypes as sub-subplot factors. Statistical analyses were carried out using the statistical analysis package SAS (SAS Institute, Cary, NC) on natural logarithm transformed harvestable dry weight of the second and third cuts for each individual plant. The biomass of the first cut was not analyzed, because planting effects were probably combined with treatment effects. Logarithm transformation was chosen to consider relative differences in plant biomass.
A. Variability between Legumes and Nonlegumes in the
Response to C02 The response to COz of the two functional groups and of the seven plant species is presented in Table I. Both functional groups and all species
35 Pa C O 2 60 Pa COp SE (In dry weight, mg per plant)
Species A. In
Relative COp effect (%)
sward 6.833 7.343 6.854 7.118 6.411 6.912
7.251 7.558 6.917 7.497 6.655 7.176
0.144 0.093 0.122 0.112 0.172 0.0575
+ 52 + 24 +7 + 46 + 28 + 30
10.068 9.868 9.968
11.094 10.908 11.001
0.115 0.131 0.087
+ 179 + 183 + 181
Nonlegumes
8.601 8.615 8.023 7.771 6.628 7.928
8.390 8.807 7.721 7.692 7.286 7.979
0.150 0.093 0.127 0.112 0.167 0.0580
- 19 + 21 -26 - 8 +93 +5
Legumes
9.377 7.758 8.568
10.064 8.268 9.166
0.119 0.346 0.1644
+99 + 67 + 82
Nonlegumes
Legmnes B. In multispecies mixture
Biomass is expressed as the natural logarithm (n = 21).
showed positive biomass responses to elevated C O 2 when they were grown in gaps in sward. Harvestable biomass of nonlegumes at elevated CO2 increased by 30%, whereas the biomass of the legumes increased by 181%, an increase six times greater than that of the nonlegumes. Within each of the two functional groups, there were no significant interspecific differences in the responses of harvestable biomass to CO2. The strong response of the legumes to CO2 was also evident in the multispecies mixture (Table 1B). At elevated CO2, the legumes produced 82% more yield than at ambient CO2. The nonlegumes, however, produced only 5% more harvestable biomass, and their responsiveness to CO2 was not as strong as when they were grown in gaps in the sward. The three grass species, 19%), ( - 2 6 % ) , and ( - 8 % ) , even showed negative responses to an increase in CO2 when they were grown in the multispecies competition treatment. These results could underestimate the
19.
291
response t o C O 2 of species that significantly increase their root-to-shoot ratio at elevated CO2. The large differences in the responses to CO2 between N2-fixing and Nznonfixing perennial grassland species found in both competition treatments confirm the results of various authors and their co-workersm Overdieck (1984), Hardacre (1986), Nijs (1989a,b), and Newton (1994)mall of whom used and The differences between legumes and nonlegumes observed in our experiment were, however, larger and more consistent over time than differences reported in these other studies. This may be related to the period chosen for fumigation and to the sward structure. Because planting was done at the end of May the grasses did not develop reproductive organs during the experiment. The response of the grasses may therefore be stronger in the second year when the highly productive growth phase in spring can occur. The legumes, however, grow most during warm periods (Mitchell, 1956) when growth is especially stimulated by CO2 (Newton 1994). Consequently, the timing of the experiment may have been more advantageous to the legumes than to the grasses. In another FACE experiment adjacent to the one described here, established swards of and monocultures showed mean CO2 responses of +6% for the grass and + 19% for the legume under different management treatments (Hebeisen unpublished data). To manipulate the availability of resources other than CO2, these management treatments included defoliation frequency (four or six times per year) and nitrogen fertilization (100 or 420 kg N ha -a year-a). We assume that the smaller responses to CO2 of these established swards, compared to those of plants grown in the gaps in our study, are related to the sward structure. This would be consistent with the findings of Overdieck (1984) who observed stronger responses to CO2 during the establishment of the sward rather than later when the sward was denser. Thus it appears that plants in gaps developed and grew exponentially for a relatively long period of time and did not experience significant limitation by other resources than CO2. Hence, increases in relative growth rate due to CO2 are larger for developing plants in gaps than for plants in dense swards where light may rapidly become a limiting factor. These results further demonstrate that sward structure influences the extent of the CO2 effect on growth but not the qualitative differences of the CO2 effects on legumes and nonlegumes. The remarkable growth response of the legumes, as compared to the nonlegumes, to elevated CO2 was observed in both experiments in swards of different structure and age as well as under different management treatments which affected the supply of resources (compare Hartwig Chapter 16). The responses of developing plants in the gaps in the L. sward may represent the phase of development from young plants
292
to adult plants. This could influence the species composition of the plant community. In the seven-species mixture of our experiment, the proportion of legumes and in the harvested biomass increased from 43% at ambient CO2 to 63% at 60 Pa CO2. The established swards (second year after sowing) of the adjacent experiment appeared to represent the CO2 effect on well-established plants in a dense sward under different management treatments. In these established swards comparable results with a significant increase in the proportion of in the mixture with were observed. It is therefore suggested that the larger response of legumes as compared to nonlegumes is indicative of different phases of the plant's life cycle and of differently managed grassland on fertile soils. The negative CO2 responses of (-19%), (-26%), and T. ( - 8 % ) (Table 1), when grown in the multispecies mixture, are considered to be related to a CO2-induced increase in competition, mainly of the legumes. The harvestable biomass of these three species in the mixture decreased from 292 g m -2 (39% of total harvestable biomass) at ambient CO2 to 202 g m -2 ( 2 0 % of total harvestable biomass). A similar, negative CO2 response was also observed in the adjacent experiment with established swards where showed a 6% decrease in dry weight when grown in competition with These results indicate that elevated atmospheric CO2 may have significant effects on the competitive interactions in temperate grassland. Species that showed the strongest response to CO2 increased their competitive abilities at elevated CO2 as compared to species with a weak response to CO2. In the multispecies mixture, the proportion of nonlegumes decreased from 49% at ambient to 36% at 60 Pa CO2. Comparable CO2-induced effects on competition were also demonstrated in controlled environments, in experiments with competition between C3 and C4 plants (Bazzaz and Carlson, 1984; Wray and Strain, 1987a,b; Johnson 1993). Based on our results, it is concluded that species with a weak response to CO2, mainly grasses, will be less competitive in a CO2-rich environment in the future. This will lead to changes in the proportion of species in plant communities on fertile grassland.
B. The Availability of Mineral Nitrogen in the Soil and Symbiotic Nitrogen Fixation as Possible Explanations of Interspecific Differences in Responses to CO2 The stronger C O 2 response of the legumes grown in the field as compared to nonlegumes contrasts with our results from growth chamber experiments using a sand/nutrient-solution system. Under these conditions, the yield responses of and to CO2 were similar (data not shown), probably because of the nonlimiting availability of nutrients. It has been repeatedly shown that the availability of nutrients strongly influences the
19.
293
plant's response t o C O 2 (Goudriaan and de Ruiter, 1983; Conroy 1992; Conroy and Hocking, 1993; Rogers 1993). In the adjacent field experiment, the relatively weak response of the established swards to CO2 was accompanied by a significant reduction in the nitrogen concentration of the harvested material (from 28.3 mg 9g-1 to 23.1 mg 9g-1 (p < 0.001) at elevated CO2), which resulted in a decrease in the yield of nitrogen. however, showed no significant reduction in nitrogen concentration (47.7 m g . g -1 compared to 45.9 m g . g -1 at 60 Pa CO2) but showed a significant increase in symbiotic nitrogen fixation (Zanetti unpublished data; Hartwig Chapter 16). The difference in nitrogen assimilation between the two species may have caused the different responses to CO2. If mineral nitrogen is the limiting nutrient for nonlegumes, then an increase in CO2 concentration will have only minor effects on growth. It has been shown that plants grown at elevated CO2 may release more energy-rich but nutrient-poor root exudates with a high C / N rado into the rhizosphere (Norby 1987; Lekkerkerk 1990). Growth of the soil microflora may be stimulated which may lead to temporal immobilization of mineral nitrogen (Van Veen 1991; Diaz 1993; Van de Geijn and Van Veen, 1993). Such relationships indicate that less mineral nitrogen may have been available to the plants at elevated CO2 which may have limited the growth ofnonlegumes (Hartwig Chapter 16). Because the legume's growth does not depend solely on the availability of mineral soil nitrogen, symbiotic nitrogen fixation appears to have allowed to respond more dramatically than the nonlegumes leading to increases in dominance of the Nz-fixing species at elevated CO2. Phosphorus nutrition is also known to influence the plant's response to CO2 (Goudriaan and de Ruiter, 1983; Conroy 1992; Rogers 1993). In our experiment, however, soil fertility tests showed that phosphorus and potassium contents in the soil were adequate for intensive grassland. species are known to respond more strongly to an insufficient supply of phosphorus and potassium than grasses (Mengel and Steffens, 1985). Insufficient supplies of these two elements not only reduce the legumes' growth but also their ability to fix nitrogen symbiotically (Cadisch 1993). Assumably, a limited supply of phosphorus and potassium would primarily weaken the response of the species to CO2 but not of the grasses. This could explain the different responses of the legumes to CO2 observed by Leadley and K6rner (Chapter 11) in a nutrient-poor chalk grassland.
C. Intraspecific Variability in Responses to CO2 The intraspecific variability in yield and in response to CO2 of seven genotypes of the species is summarized in Table II and Figs. 1 and 2. Highly
294
Source
MS In DW
P
Main plots Blocks (B) CO2 Error A (B X COz)
2 1 2
5.975 8.888 0.586
ns ns
Subplots Competition CO2 x competition Error B (B X competition (CO2))
1 1 4
47.910 1.654 2.974
<0.05 ns
Sub-subplots Species CO2 X species Species X competition CO2 X species x competition Genotype (species) CO2 X genotype (species) Competition X genotype (species) CO2 x competition X genotype (species) Error Total
5 5 5 5 35 35 35 35 314 485 c
108.589 2.287 13.964 0.675 3.198 0.502 0.628 0.340 0.406
<0.0001 <0.005 <0.0001 ns <0.0001 ns <0.05 ns
L., (L.) J. and C. Presl, Huds., (L.) P.B., L., and L. was excluded from this analysis because it was not possible to distinguish between the different genotypes (due to its stoloniferous growth) in the multispecies mixture. bConsisting of these six species and c 18 missing values
significant genotypic variability in harvestable biomass (independent of the CO2 and competition treatments) was found [Table II: see "genotype(species), P < 0.001"]. When planted in the gaps in the sward, the genotype's yields ranged from 1.1 to 3.6 g per plant for and from 15.3 to 76.3 g for (Fig. 2; Antilog). This demonstrates the intraspecific heterogeneity of the tested plant material. In contrast, no intraspecific variability in response to CO2 was observed [Table II: see "CO2 X genotype(species), ns"; Fig. 1]. Thus a fast evolutionary adaptation of these species in their responses to elevated CO2 seems unlikely because of the lack of variability (Gartside and McNeilly, 1974; Bradshaw and Mortimer, 1986; Bradshaw and McNeilly, 1991). The interspecific differences in responses to CO2 should, therefore, persist, and thus the long-term proportion of species in grasslands should change in a CO2rich environment in the future.
19.
9. r~
8
-
a
6-
Effect of atmospheric C O 2 concentration on harvested biomass of seven genotypes (@) of the two perennial grassland species and Biomass is expressed as the natural logarithm (bars = SE; n = 6).
No genetic variability in the response t o CO 2 was found by Fajer (1992) who investigated growth and leaf biochemistry in six clones of P. Schmid (Chapter 4) also found no genetic variation in the response to increased CO2 of and when grown in the field with competition. Without competition, however,
9~ c"
8
._~ ~
7-
a
6-
-
I
I
I
1
Effect of competitive e n v i r o n m e n t sward or multispecies mixture) on the harvested biomass of seven genotypes (@) of the two perennial grassland species and Biomass is expressed as the natural logarithm (bars = SE; n = 6).
296 significant genetic variation was detected in and Genetic variation in the response of germination and growth to CO2 were found in (Wulff and Alexander, 1985). Responses of seed production of rice (Ziska and Teramura, 1992) and (Curtis 1994; Curtis Chapter 2) varied between genotypes. Significant adaptation of biomass under high CO2 was found at early vegetative stages in (see Tousignant and Potvin, Chapter 3), but this was lost with time. Wulff and Alexander (1985) demonstrated that genetic variability in response to CO2 is sometimes detectable in a relatively small set of plant material (five maternal families of one species). Nevertheless, the test of intraspecific variability in grassland species has to be extended to a bigger set of species and genotypes as well as to longer experimental periods to obtain more general conclusions. However, the limited set of species and genotypes of this study demonstrated that interspecific variability in response to CO2 was far greater than the intraspecific variability.
D. Intraspecific Variability in Responses to Competitive Environments Based on the observed interspecific differences and the lack of genetic variability in responses to CO2, we suggest that floristic composition in fertile grasslands will change. These CO2-induced changes in the proportion of species, as also observed in our multispecies mixture, will alter the competitive environment in the plant communities. The long-term changes in floristic composition in a CO2-enriched atmosphere may be influenced by the evolutionary adaptation of the species to the new competitive environment. There is evidence that evolutionary adaptation of plant populations to competitive environments does occur (Turkington and Harper, 1979; Turkington, 1989; LOscher and Jacquard, 1991; LOscher 1992). 9Such adaptation would be particularly important for plant species with a weak response to CO2 and exposure to increased competitive pressure associated with CO2 elevation. Such an adaptation could help to lessen the indirect negative effects of the CO2 increase. However, evolutionary adaptation depends on sufficient genetic variability within the plant species. Genetic variability in the response of the plant's growth to the competitive environment was detected in the tested plant material [Table II: see "competition • genotype(species), P < 0.05"; Fig. 2). This indicates that evolutionary adaptation of the species to the CO2-induced changes in the competitive environment seems plausible. Through such an adaptation these species with a weak response to CO2 may partly avoid the indirect negative effects of elevated CO2. This could act as a buffer against further changes in species composition within grasslands.
19.
In a flee-air carbon dioxide enrichment experiment on a fertile soil, inter- and intraspecific differences in the growth responses to CO2 of seven perennial species typical of permanent, temperate grasslands were investigated. The effect of interspecific differences in the response to CO2 on competitive interactions was tested within a multispecies mixture in which all seven species were in competition with each other. Plant species differed widely in their biomass responses to elevated CO2 (60 Pa). CO2 fumigation stimulated growth in the first year by 30% in nonlegumes and by 181% in legumes when they were grown in gaps in an established sward. Large interspecific differences in the responses to CO2 were also evident in the multispecies mixture. In this mixture the legumes had a strong positive response to elevated CO2 (+ 82% biomass), whereas three grass species responded negatively to elevated CO2. Therefore, significant changes in the proportions of the species occurred in the multispecies mixture. Our data indicate that, under elevated CO2, the availability of mineral nitrogen in the soil may have limited the growth of nonlegumes, and symbiotic nitrogen fixation of the legumes may explain their strong response to CO2. In contrast to interspecific differences in responses to CO2, no intraspecific variability in these responses was detected. Hence it is unlikely that a rapid evolutionary change in the species' responses to COz will occur at the population level. Therefore, interspecific differences in responses to CO2 may persist, and species with a weak response to COz will be less competitive in a CO2-rich environment. It is thus suggested that the longterm proportion of species in grassland communities may change as the atmospheric CO2 concentration increases. Consequently, the competitive environment would be altered in grassland communities. Intraspecific variability in the plants' responses to the competitive environments sward or multispecies mixture) was detected in our experiment, and it is, therefore, suggested that this variability may be the basis for subsequent evolutionary adaptation of the populations to the predicted CO2-induced changes in competitive environments. Such a process may aid species with a weak response to CO2 to partly avoid the indirect negative effects of elevated CO2. This could act as a buffer against further changes in the composition of species within grassland communities.
S. Koller, J. Nagy, and K. F. Lewin built the CO2 control and monitoring system. We thank K. Rfiegg, P. Schlfissel, and P. Jager for technical assistance. This research was supported by
the Swiss National Energy Research Fund, the Swiss National Science Foundation, the Swiss Department of Energy, and the Swiss Federal Institute of Technology. We are grateful to M. Schoenberg for checking the English.
Bazzaz, F. A. (1990). The response of natural ecosystems to the rising global CO2 levels. 21, 167-196. Bazzaz, F. A., and Carlson, R. W. (1984). The response of plants to elevated CO2 I. Competition among an assemblage of annuals at two levels of soil moisture. 62, 196-198. Bradshaw, A. D., and McNeilly, T. (1991). Evolutionary response to global climatic change. 67, 5-14. Bradshaw, T., and Mortimer, M. (1986). Evolution in communities. "Community Ecology: Pattern and Process" (J. Kikkawa and D.J. Anderson, eds.), pp. 309-341. Blackwell, Oxford. Cadisch, G., Sylvester-Bradley, R., Boller, B. C., and N6sberger,J. (1993). Effects of phosphorus and potassium on N2 fixation (15N-dilution) of field-grown and C. 31, 329-340. Conroy, J., and Hocking, P. (1993). Nitrogen nutrition of C3 plants at elevated atmospheric CO2 concentrations. 89, 570-576. Conroy, J. P., Milham, P.J., and Barlow, E. W. R. (1992). Effect of nitrogen and phosphorus availability on the growth response of to high CO2. 15, 843-847. Curtis, P. S., Snow, A. A., and Miller, A. S. (1994). Genotype-specific effects of elevated CO2 on fecundity in wild radish 97, 100-105. Diaz, S., Grime,J. P., Harris, J., and McPherson, E. (1993). Evidence of a feedback mechanism limiting plant response to elevated carbon dioxide. 364, 616-617. Fajer, E. D., Bowers, M. D., and Bazzaz, F. A. (1992). The effect of nutrients and enriched CO2 environments on production of carbon-based allelochemicals in test of the carbon nutrient balance hypothesis. 140, 707-723. Gartside, D. W., and McNeilly, T. (1974). The potential for evolution of heavy metal tolerance in plants. II. Copper tolerance in normal populations of different plant species. 33, 303-308. Goudriaan, J., and de Ruiter, H. E. (1983). Plant growth in response to CO2 enrichment, at two levels of nitrogen and phosphorus supply. I. Dry matter, leaf area and development. 31, 157-169. Hardacre, A. K., Laing, W. A., and Christeller, J. T. (1986). The response of simulated swards of perennial ryegrass and white clover to enriched atmospheric CO2: Interaction with nitrogen and photosynthetic photon flux density. N. Z.J. 29, 567-573. Hunt, R., Hand, D. W., Hannah, M. A., and Neal, A. M. (1991). Response to CO2 enrichment in 27 herbaceous species. 5, 410-421. Hunt, R., Hand, D. W., Hannah, M. A., and Neal, A. M. (1993). Further responses to COz enrichment in British herbaceous species. 7, 661-668. Johnson, H. B., Polley, H. W., and Mayeux, H. S. (1993). Increasing CO2 and plant-plant interactions--Effects on natural vegetation. 104, 157-170. Kimball, B. A. (1983). Carbon dioxide and agricultural yield: An assemblage and analyses of 430 prior observations. 75, 779-788. Lawlor, D. W., and Mitchell, R. A. C. (1991). The effects of increasing COz on crop photosynthesis and productivity: A review of field studies. 14, 807-818.
19.
299
Lekkerkerk, L. J. A., Van de Geijn, S. C., and Van Veen, J. A. (1990). Effects of elevated atmospheric CO2-1evels on the carbon economy of a soil planted with wheat. "Soils and the Greenhouse Effect" (A. F. Bouwman, ed.), pp. 423-429. Wiley, New York. Lewin, K. F., Hendrey, G. R., and Kolber, Z. (1992). Brookhaven National Laboratory FreeAir Carbon Dioxide Enrichment Facility. 11, 135-141. Lfischer, A., and Jacquard, P. (1991). Coevolution between interspecific plant competitors? 6, 355-358. LCtscher, A., Connolly, J., and Jacquard, P. (1992). Neighbour specificity between and from a natural pasture. 91, 404-409. McConnaughay, K. D. M., Berntson, G. M., and Bazzaz, F. A. (1993). Plant responses to carbon dioxide. 361, 24. Mengel, I~, and Steffens, D. (1985). Potassium uptake of rye-grass and red clover (Trif01ium as related to root parameters. 1, 53-58. Mitchell, K.J. (1956). Growth of pasture species under controlled environment. I. Growth at various levels of constant temperature. 38A, 203-216. Newton, P. C. D. (1991). Direct effects of increasing carbon dioxide on pasture plants and communities. N. Z.J. 34, 1-24. Newton, P. C. D., Clark, H., Bell, C. C., Glasgow, E. M., and Campbell, B. D. (1994). Effects of elevated CO2 and simulated seasonal changes in temperature on the species composition and growth rates of pasture turves. 73, 53-59. Nijs, I., Impens, I., and Behaeghe, T. (1989a). Effects of different CO2 environments on the photosynthesis-yield relationship and the carbon and water balance of a white clover L. cv. Blanca) sward. 40, 353-359. Nijs, I., Impens, I., and Behaeghe, T. (1989b). Leaf and canopy responses of to long-term elevated atmospheric carbon-dioxide concentration. 177, 312-320. Norby, R.J., O'Neil, E. G., Hood, W. G., and Luxmore, R.J. (1987). Carbon allocation, root exudation and mycorrhizal colonization of seedlings grown under COz enrichment. 3, 203-210. Overdieck, D., Bossemeyer, D., and Lieth, H. (1984). Long-term effects of an increased CO2 concentration level on terrestrial plants in model-ecosystems. 3, 344-352. Patterson, D. T., and Flint, E. P. (1990). Implications of increasing carbon dioxide and climate change for plant communities and competition in natural and managed ecosystems. "Impact of Carbon Dioxide, Trace Gases, and Climate Change on Global Agriculture" (B. A. Kimball, ed.), pp. 83-110. Am. Soc. Agron., Spec. Publ. 53, Madison, WI. Poorter, H. (1993). Interspecific variation in the growth response of plants to an elevated ambient CO2 concentration. 104, 77-97. Rogers, G. S., Payne, L., Milham, P., and Conroy, J. (1993). Nitrogen and phosphorus requirements of cotton and wheat under changing atmospheric CO2 concentrations. 156, 231-234. Smith, S. D., Strain, B. R., and Sharkey, D. (1987). Effects of CO2 enrichment on four Great Basin grasses. 1, 139-143. Turkington, R. (1989). The growth, distribution and neighbour relationships of in a permanent pasture. V. The coevolution of competitors. 77, 717-733. Turkington R., and HarperJ. L. (1979). The growth, distribution and neighbour relationships of in a permanent pasture: IV. Fine-scale biotic differentiation. 67, 245-254. Van de Geijn, S. C., and Van Veen, J. A. (1993). Implications of increased carbon dioxide levels for carbon input and turnover in soils. 104, 283-292. Van Veen, J. A., Liljeroth, E., Lekkerkerk, L.J.A., and Van de Geijn, S. C. (1991). Carbon fluxes in plant soil systems at elevated atmospheric CO2 levels. 1, 175-181. Williams, W. E., Garbutt, K., and Bazzaz, F. A. (1988). The response of plants to elevated CO2. V. Performance of an assemblage of serpentine grassland herbs. 28, 123-130.
300
Woodward, F. I., Thompson, G. B., and McKee, I. F. (1991). The effects of elevated concentrations of carbon dioxide on individual plants, populations, communities, and ecosystems. 67, 23-38. Wray, S. M., and Strain, B. R. (1987a). Competition in old-field perennials under COz enrichment. 68, 1116-1120. Wray, S. M., and Strain, B. R. (1987b). Interaction of age and competition under CO2 enrichment. 1, 145-149. Wulff, R. D., and Alexander, H. M. (1985). Intraspecific variation in the response to CO2 enrichment in seeds and seedlings of 66, 458-460. Ziska, L. H., and Teramura, A. H. (1992). Intraspecific variation in the response of rice (Oryza sativa) to increased CO2mPhotosynthetic, biomass and reproductive characteristics. 84, 269-276.
2 Competition between Grasses and Trifolium repens with Elevated Atmospheric CO2
Although there is ready agreement among scientists that competition must be considered when predicting ecosystem responses to CO2 (e.g., Bazzaz, 1990; Woodward 1991; Woodward, 1992; Bazzaz and Fajer, 1992; Bazzaz and McConnaughay, 1992), it is still not clear how generalizations about competitive processes might be achieved. In order to develop general predictions of ecosystem change in response to rising atmospheric CO2, mechanistic understanding is needed of (i) how CO2 may alter competition for resources; (ii) how competition processes are influenced by the functional characteristics of the organisms making up the community; and (iii) how other environmental variables will alter the effect of CO2 on competitive interactions.
Increases in atmospheric CO2 concentration represent an increase in resource availability; therefore, we might expect that ecological theory on the effect of resource availability on competition could assist the development of predictions on the effects of increases in atmospheric CO2 concentration on community and ecosystem processes. It has been suggested (Grime, 1977, 1979; Campbell and Grime, 1992) that some plants are more 301
Copyright 9 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.
302
aggressive competitors than others because of evolutionary specializations determined by prevailing conditions of resource availability and disturbance in particular habitats. This theory suggests that plants suited to conditions of high productivity and low disturbance are strongly specialized for competition both above- and belowground, with tall stature, extensive lateral spread, and high growth rates permitting active foraging for resources (Campbell, Grime, and Mackey, 1991 ). In contrast, plants suited to frequent disturbance have high growth rates but rapidly shunt resources in favor of reproductive organs and seeds, resulting in a restricted capacity to maintain vegetative organs for resource capture and thus a lower competitive ability. Similarly, plants suited to resource-limited conditions have low growth rates, store nutrients in long-lived structures, and are strongly defended against herbivory, thus having a low capacity to compete for resources. From this theory, it might be predicted that plants from productive habitats will be the most rapid to increase competitive effects on vegetation in response to an increase in resource availability and therefore that the competitive effects of fast-growing plants from productive habitats should increase more with elevated CO2 than the competitive effects of plants with lesser competitive ability. In addition, literature on the biochemistry and CO2 responsiveness of C3 and C4 species (e.g., Pearcy 1981; Zangerl and Bazzaz, 1984; Bazzaz and Garbutt, 1988; Bazzaz 1989) indicates that C~ species can be more responsive to CO2 than C4 species. This constitutes another dimension for consideration when predicting effects of CO2 on competitive interactions. The expression of competitive ability is also strongly determined by prevailing environmental conditions such as temperature. Seasonal changes in temperature are an important factor in the changing relative proportions of species throughout a growing season, especially where species of differing phenology are present in the community. The resulting switches of community composition and competitive interactions are important in determining the effects of competition on community structure. As temperature also is likely to strongly influence the response of species to CO2 (Campbell 1993), it is important to consider the combined effects of phenology and competitive ability when considering how competition processes between different species may change in response to increased atmospheric CO2. In herbaceous grassland communities, competition from grasses is an important factor limiting the growth and performance of the legume species L. (Harris, 1987, 1990). Changes in competition arising from increases in atmospheric CO2 may alter the growth and distribution of populations (Overdiek 1984; Overdiek and Reining, 1986; Hardacre 1986; Newton 1994) but it is unclear how the specific
20.
303
interactions between and various grass species will determine the overall response of grassland ecosystems. From the above theory, we might predict that, in medium to high fertility conditions, fast-growing C~ species from fertile habitats would develop greater short-term competitive effects on at high CO2 than would C4 species or species of lesser competitive ability from infertile habitats. We also might predict that increases in temperature will shift the competitive success in favor of at elevated CO2. To arrive at a more mechanistic understanding of competition, it is also useful to focus attention on the capture and utilization of specific resources. Phosphorus nutrition has been implicated as particularly important in relation to competition between grasses and Phosphorus response has been used to explain differences in growth between the legumes T. and (Hart and Jessop, 1984, Hart and Collier, 1994), and is considered to be an essential component of the processes determining competition between and grasses (Jackman and Mouat, 1972). Evidence also exists that phosphorus supply affects leaf area, photosynthesis, and nitrogen fixation in (Hart and Greer, 1988; Hart, 1989; Hart and Collier, 1994). In view of the potential importance of phosphorus nutrition in grass-clover competition, perturbations in phosphorus nutrition brought about by the presence of c o m p a n i o n / competitive grasses were examined as a first step in using resource capture parameters to develop a mechanistic interpretation of CO2 effects on grassclover competition. This chapter describes an experiment designed to test the predicted effects of elevated CO2 on competition from grasses and the phosphorus status of under strictly controlled conditions. To test the predictions, nine different grass species were chosen which varied in morphology, seasonal growth, photosynthesis type, and plant strategy Grime, 1977). We used single, growing stolon tips of (which had been clonally propagated from a single genotype) as biosensors of the competitive environment. The resulting growth, morphology, and chemical composition of the developing plants provided a test of how the competitive ability of several different grass species may be altered by elevated CO2 and different temperatures.
A. Methods
Characteristics of the nine grass species examined in the experiment are described in Table I. Seeds of the nine grass species were sown into 4.5liter pots of Opiki peat loam soil mixed with river sand. Some pots were
304
Species
Photosynthesis Established type strategy a
Description L.
L.
Rhizomatous, prostrate, densely tillered, perennial grass, warm season growth, moderate fertility sites Tufted perennial grass, moderate fertility sites, tolerant of low moisture Free-seeding, annual subtropical grass
C3
CSR
C~
CR
C4
R/SR
Robust perennial grass, summerautumn growth Tufted, moderately densely tillered, perennial grass, cool season growth with summer dip, moderate to high fertility Upright, coarse-leaved, tufted subtropical perennial grass, summer growth, moderate to high fertility, moist sites Stoloniferous and rhizomatous, dense creeping grass, warm season growth, moderate to high fertility sites Short, perennial grass, low fertility, dry sites, appears to operate as an annual in very dry sites
Cs
CR
C3
CR
C4
CR
C4
CSR
C3
R/SR
C4
S/CSR
L. Scop. Schreb. L.
Poir.
Chiov. (Zotov) Connor and Edgar comb.nov (Poir.) Robyns and Tourn.
Tufted coarse-leaved, perennial grass, summer growth, moderate to low fertility, dry sites
a Following the CSR systemof classifyingplant types (Grime R, ruderal.
1988). C, competitor; S, stress tolerator;
also left unplanted. The pots were placed in four different controlled environment rooms at the National Climate Laboratory, Palmerston North, New Zealand, all with a photosynthetic photon flux density of 700/xmol m -2 s -1 and a 12 h photoperiod. Individual rooms were maintained with a d a y / n i g h t temperature regime of either 28/23~ or 18/13~ and a CO2 concentration of either 350 /xl/liter or 700 /xl/liter, to give a factorial combination of temperature and CO2 treatments. There were eight replicate pots of each treatment. The grasses were established under clipping for 70 days. Immediately after the fourth clipping to a height of 20 mm, a single, rooted stolon of a single genotype of L. cv. Grasslands Huia was planted at the edge of each pot and positioned to grow across
20.
305
the grass sward. Roots on the stolon were inoculated with prior to planting and were observed to be nodulated at the final harvest. The pots growing at 28/23~ were harvested after 35 days and those growing at 18/13~ were harvested after 42 days. One youngest mature leaf was removed from each plant and each leaf split longitudinally. Tissue from one-half was killed in liquid nitrogen and stored at -70~ for inorganic phosphorus analysis. Tissue from the remaining half of each leaf was dried at 80~ and weighed. Inorganic phosphorus (Pi) was extracted from the frozen portion of leaf using ice-cold perchloric acid (Irving and Bouma, 1984) and the Pi concentration in the extract determined by a manual modification of the method of Asher (1980). Total phosphorus (Px) was determined on the dried portion of leaf using the method of Haslemore and Roughan (1976), with a 2-ml volume of the initial solution of sulphuric acid/hydrogen peroxide. For the estimation of organic nitrogen, 0.3 ml aliquots were taken from sulphuric acid/hydrogen peroxide solutions used for PT determinations. Digestion mixture (0.35 ml) (Williams and Twine, 1967) was added to each aliquot and the solution digested at 315~ The solution was then diluted to 25 ml and the nitrogen content of the aliquot determined by the method of Reay (1985). The grass shoot biomass was harvested to ground level in each pot, dried, and weighed. The remaining shoot material was removed from each pot, and the number of growing points recorded before the plant was dried at 80~ and weighed. The total leaf material was calculated by adding leaf dry mass from the sample for chemical analysis and the remaining portion. Data were analyzed by analysis of variance, with significance calculated using the F test and least significant difference (LSD) test. The dry mass and dry mass proportion data were logarithm-transformed prior to analysis of variance. B. Dry Mass Results A clear hierarchy in competitive suppression of was evident within the range of grasses when grown under current CO2 concentrations (Fig. 1). The greatest competitive suppression of dry mass at 18/ 13~ was observed with and At the higher temperature regime of 28/23~ the subtropical grasses and the most aggressive grasses, followed by and Elevations in COz caused marked changes both in the growth of isolated plants and in the competitive effect of the grass species on T. (species • CO2 X temperature effect significant, P < 0.03). When
"~ 1.o E a
The aboveground dry mass of growing with different temperate (C3) and subtropical grasses (C4) at CO2 concentrations of either 350/zl/liter (open bars) or 700/zl/ liter (solid bars) and controlled day/night temperatures of either (a) 18/13~ or (b) 28/ 23~ The C3 species are Ac, Dg, Fa, Lp, and Rc, The C4 species are Ds, Pc, Pd, and Sa, The vertical bar is LSD (P < 0.05).
plants were grown without any companion species, growth was stimulated by CO2 at both temperatures. The stimulation of growth by CO2 was less at 18/13~ than at 28/23~ (Fig. 2a), but this represented about a 20-25% increase in dry mass in both cases (Fig. 2b). When companion species were present, however, there was a marked change in the effect of elevated CO2 (Fig. 2). In general, increases in CO2 at 18/13~ increased competitive suppression of so that growth was decreased at elevated CO2. In contrast, elevated CO2 at 28/23~ decreased competitive suppression of
20.
0
e.-o
o
~
Ds
E 121
,
Dry mass difference at 18/13~ (g)
,
... o
1:1
1.5
0 m
m
1.0
E
The (a) difference between dry mass at 700/zl/liter CO2 and dry mass at 350/xl/ liter CO2 and (b) ratio of dry mass at 700/zl/liter CO2 to dry mass at 350 ~l/liter CO2 of T. plants growing with different grass c o m p a n i o n species at different temperatures. The species are as listed in Fig. 1. The solid circles represent C3 species and the o p e n circles represent C4 species.
Significant variation existed in the specific effects of the different grasses. At t e m p e r a t u r e s of 18/13~ elevated CO2 c o n c e n t r a t i o n depressed the growth of with all species except and In contrast, at 28/23~ the growth of was e n h a n c e d by elevated CO2 with all grasses except and D. Two grass species and h a d approximately similar competitive effects on in response to CO2 irrespec-
308
tive of the temperature. A group of others had markedly less competitive effects on in response to CO2 at 28/23~ compared to 18/ 13~ The effect of CO2 on absolute difference in dry mass (Fig. 2a) showed similar trends to the effect on the percentage stimulation of dry mass (Fig. 2b). The dry mass of Cs grasses was increased by CO2 at both temperatures but the effect was proportionally less at 28/23~ than at 18/13~ (species x CO2 X temperature effect significant, P < 0.01) (Fig. 3)-. The subtropical (Ca) grasses showed no positive responses to CO2 at either temperature. The species are ranked in Figs. 3a and 3b in order of the degree of competitive suppression of It is clear by comparing Figs. 1 and 3 that there was no strong trend to higher dry mass with higher competitive effect, indicating that shoot structure and competitive processes belowground are also likely to be important determinants of the competition processes. When the dry mass is expressed as a proportion of the total dry mass in each pot (Fig. 4), it is evident that the effect of elevated CO2 differed depending on the prevailing temperature (species x CO2 x temperature effect significant, P < 0.01). Increases in CO2 generally reduced the contribution of to the aboveground biomass at 18/13~ except for the subtropical (Ca) grasses and In contrast, at 28/23~ increased CO2 generally increased the contribution of to the biomass, except with the very aggressive Ca grasses and The greatest increases in the proportion of at this temperature were observed with the Cs grasses and R~ and with the subtropical (Ca) grasses. The proportion of was generally considerably less at 18/13~ than at 28/23~ and the changes in proportions were less at the lower temperature. C. Shoot Morphology Results
The above changes in dry mass were also accompanied by changes in plant morphology, expressed in terms of the number of growing points developed on each plant (species X CO2 X temperature effect significant, P < 0.003). The number of growing points per plant was unaltered by CO2 when plants were growing in isolation (Table II). However, when plants were in competition with grasses, growing point numbers were generally reduced by elevated CO2 at 18/13~ but increased by elevated C O 2 at 28/23~ consistent with changes in dry mass (Fig. 1). D. Nutrient Status Results The phosphorus and nitrogen contents of the youngest mature leaves of provide a further assay of the competitive environment created
20.
E 121
1.0
E 121
1.0
The aboveground dry mass of different temperate (C3) and subtropical (C 4) grasses at COs concentrations of either 350 /xl/liter (open bars) or 700 /zl/liter (solid bars) and controlled d a y / n i g h t temperatures of either (a) 18/13~ or (b) 28/23~ The species are as listed in Fig. 1. The vertical bar is LSD (P < 0.05).
by the grasses at different temperatures and C O 2 concentrations. The results for a subset of the species (Table III) show that boththe average inorganic and total phosphorus contents were higher at 35:0/xl/liter COs and temperarares o f 28/23~ ( P < 0.0!1), and there were significant species~.-x CO2 • temperature ~interactions ~(p < 0.0.07). However; the differences among species ' d i d not correspond closely with~,the observed differences in T. drymass: produced consistently lower inorganic and total
offl
G..
t~
o t~ m
O
2
I1.
Figure 4 The proportion of total aboveground dry mass contributed by when growing with different temperate (C3) and subtropical grasses (C4) at CO2 concentrations of either 350/zl/liter (open bars) or 700/zl/liter (solid bars) and controlled day/night temperatures of either (a) 18/13~ or (b) 28/23~ The species are as listed in Fig. 1. The vertical bar is LSD (P < 0.05).
phosphorus concentrations in the leaves of especially at the highest temperature and CO2 concentration. The ratio of inorganic to total phosphorus provides an index of the internal phosphorus status in tissue. Higher values of this ratio (Pi/PT) indicate that phosphorus is less limiting to plant growth. The values in Table III indicate that increases in CO2 r e d u ~ d phosphorus stress at both temperatures in the absence of grass, but did not substantially alter phosphorus stress in the presence of grass (species • CO2 effect significant,
20.
18/13~
Companion species Nil
28/23~
350/zl/liter CO2
700/zl/liter COz
350/zl/liter CO2
700/zl/liter CO2
8 5 3 4 3 4 7 6 7 6
7 2 1 3 1 1 3 1 2 1
34 5 7 9 3 3 5 2 6 6
34 8 6 10 7 10 7 6 10 8
LSD (P < 0.05) = 3.
P < 0.0002). On average, phosphorus stress was greater at 18/13~ than at 28/23~ (P < 0.0001). The ratio (Pi/PT) was generally lowest in A. and but, despite these being similar to the value observed in isolated plants at 350/zl/liter, there was no increase in internal phosphorus status with an increase in CO2 concentration. When was growing alone there was a strong decline in leaf nitrogen content, with an increase in CO2 at 28/23~ but little change at 18/13~ (Table III). In contrast, the presence of most grasses increased the leaf nitrogen content at both temperatures (species • CO2 • temperature effect significant, P < 0.008). Changes in both phosphorus and nitrogen contents did not provide a simple explanation for the changes in dry mass observed in Fig. 1.
This experiment was designed to test some specific predictions about the changes in short-term competition processes that result from increases in CO2 concentration. The plants inserted into the grass swards in this e x p e r i m e n t constitute an effective biosensor and changes in the dry mass of these plants are taken here as an integrating measure of the units of the various resources (photons, CO2, water, and nutrients) unavailable to as a result of the presence of the grasses (Welden and Slausen, 1986; Campbell and Grime, 1992). Two main points are evident from the results. Firstly, the effect of CO2 was strongly d e t e r m i n e d by the prevailing conditions of temperature. T.
18/13~ 350/~1/1 CO2
Species (a) Inorganic P (/~mol/g DM) Nil
LSD (P < 0.05) = 17 Mean (b) Total P (/~mol/g DM) Nil
LSD (P < 0.05) = 42 Mean (c) Inorganic P / T o t a l P Nil
LSD (P < 0.05) = 0.09 Mean (d) Nitrogen ( m g / g DM) Nil
LSD (P < 0105), = 19 Mean
28/23~ 700/~1/1 CO2
350/~1/1 CO2
700 ~1/1 CO~
30 25 36 27 32 27
24 14 16 19 28 31
42 36 43 47 43 66
32 24 63 43 42 33
29
22
46
39
130 122 121 126 118 107
69 81 77 94 107 146
129 118 102 108 84 202
79 86 164 127 124 104
121
96
124
113
0.22 0.20 0.31 0.21 0.28 0.25
0.35 0.17 0.21 0.19 0.24 0.21
0.29 0.30 0.46 0.46 0.50 0.33
0.40 0.28 0.37 0.33 0.31 0.33
0.25
0.23
0.39
0.34
51 52 43 50 43 51
54 41 45 55 61 57
48
52
72 41 35 64 ~60 ~:62
~
41 54 68 63 58 60
~, ,:
'
"o
.
,
,
: 3
,,.,
, :
,
.
57
56
- " ' ~
" ' i " ~
20.
313
has a temperature optimum for growth of about 20-25~ whereas most temperate grasses have a temperature optimum for growth of around 18-20~ and subtropical grasses have an optimum of about 30-35~ (Mitchell, 1956). In this experiment, increased CO2 concentrations at temperatures of 18/13~ generally increased competitive suppression of T. by the Ca and C4 grasses (except whereas at 28/23~ increases in CO2 generally decreased competitive suppression by grasses. This suggests that phenology and the temperature responses of different species will be strong determinants of the response of competition processes and community structure to CO2. A seasonal switching in competition between species with high and low temperate optima may act to reduce the competitive dominance exerted by any one component of the community in response to elevated CO2. However, it is suggested that the greatest change in competition in response to CO2 is likely to occur during warmer periods, favoring those species competing most effectively at that time of year. As confirmed in several previous studies (e.g., Bazzaz 1989; Bazzaz and McConnaughay, 1992), the C4 grass species generally showed lower responses to CO2. There is already a strong seasonal switching from temperate (C3) grass dominance in spring to greater clover dominance in summer in warm temperate grasslands (Harris, 1987). Changes in competitive interactions with future increases in atmospheric CO2 concentration may therefore have the effect of accentuating this existing seasonal trend.
Secondly, it is confirmed that the most aggressive C3 species showed the greatest increases in competitive effects in response to CO2. The Ca species and the C4 grasses were less competitive at 700/~l/liter CO2. Reference to Fig. 1 and Table I confirms that (with the notable exception of the most aggressive competitors against with elevated CO2 were those grasses with more competitive attributes (CR Grime 1988), whereas the weaker competitors were generally grasses classified as closer to the R or S corners of the triangle (Grime, 1977). The results are therefore generally consistent with predictions of plant strategy theory (Grime, 1977, 1979). This general plant functional-type scheme may therefore have value as a first filter in predicting effects of CO2 on vegetation processes. However, the unpredictable response of indicates further investigation is required into the responses of species of intermediate competitive ability.
In addition to changes in growth, changes were also evident in the mineral nutrient status of at the different temperatures and CO2 concentrations. Internal phosphorus stress was generally greater at 18/ 13~ than at 28/23~ Although an increase in CO2 generally decreased dry mass at 18/13~ but increased dry mass at 28/23~ there was no accompanying trend in internal phosphorus stress. This suggests that internal phosphorus status can vary within quite wide limits independently of dry mass. It is interesting that internal phosphorus stress was reduced by increasing CO2 at both temperatures when plants were growing in isolation, but was rather unaffected by CO2 in the presence of grasses. This indicates that here, increases in COzwere generally unable to relieve internal phosphorus stress in the presence of competition from grasses. The effects of the different grass companion species on internal phosphorus status were all generally similar, but consistently caused the greatest internal phosphorus stress under all conditions. is believed to be a strong competitor for phosphorus (Jackman and Mouat, 1972). When plants were grown in isolation there was a decrease in leaf nitrogen status at high CO2, consistent with that observed in previous studies. However, the presence of the different grass species altered this response, so that there was generally little change in nitrogen content with increases in CO2. This suggests complex effects on nitrogen nutrition of the legume in the presence of competition from different grasses, possibly resulting from the different light environment and demand for investment in the carbon-fixing protein rubisco. Part of this effect may also be exerted through the effect of the different grasses on the ability of nitrogenase activity to respond to CO2 (Crush and Campbell, 1993). Effects of CO2 on root function and rhizosphere processes also warrant further study (Rogers 1994).
Elevations in atmospheric CO2 therefore resulted in changes in competition processes expressed both in short-term changes in dry matter accumulation rates of and in fundamental changes in internal nutrient status and growing point numbers. These changes might be expected to alter the abilities of the plants to respond to subsequent seasonal switches in temperature. Thus, in predicting future effects of elevated atmospheric CO2 concentration, it is important to bear in mind that the response of the plants in intact communities is likely to be the product both of short-
20.
315
term prevailing environmental changes in temperature and longer-term (weeks to months long) constraints or enhancement of shoot numbers induced during previous growing periods.
In order to test some predictions on mechanisms of C O 2 effects on competition processes, single stolon tips of were inserted into nine different synthetic grass swards as biosensors of competition processes. The differences that developed in the dry mass, morphology, leaf phosphorus, and leaf nitrogen contents as the plants grew through these swards indicated that the prevailing conditions of temperature and CO2 concentration were important determinants of the rate of development of competition effects. Increases in CO2 appeared to favor grasses at cool temperatures but favored at higher temperatures. The greatest suppression of was observed with competitive grasses growing at 18/13~ and elevation of CO2 concentration at this temperature resulted in a further suppression of in favor of grass. Competitive suppression of T. was significantly reduced when temperature was increased to 28/ 23~ and elevated CO2 increased the growth of at this temperature. The response to CO2 by most severely restricted by competition from and In contrast, competition from and subtropical Ca grasses was considerably reduced at elevated CO2. The results confirmed the prediction that highly competitive C3 grasses are likely to increase most in competitive effects with increases in CO2. Changes in internal phosphorus and nitrogen status, and in growing point numbers of indicate that significant changes in plant structure and nutritional status accompanied changes in dry mass due to CO2. Internal phosphorus stress was decreased by elevated CO2 in the absence of grass but was increased by elevated CO2 in the presence of grass, suggesting that increases in the aboveground CO2 resource has consequences for the utilization and physiological status of resources of belowground origin. This indicates that longer term effects of these shortterm competitive effects on might be expected to carry through into subsequent seasons in intact communities. The findings also indicate that phenological switches in community composition associated with seasonal changes in temperature are also likely to determine the response of competition processes to atmospheric CO2 concentration, and consequent effects on community composition. The presence of different species clearly alters the effect of CO2 on Further research is needed to determine how the functional characteristics of plants can alter effects of CO2 on competition processes both above- and belowground.
Thanks to Derryn Williamson, William Laing, Elaine Glasgow, Chris Hunt, and Paul Newton for technical assistance and Harry Clark and Phil Grime for helpful discussions during the preparation of the manuscript. This research was funded by the Foundation for Research, Science, and Technology. The project is part of core research in the Global Change and Terrestrial Ecosystem (GCTE) project of the International Geosphere-Biosphere Programme.
Asher, L. E. (1980). An automated method for determination of orthophosphate in the presence of labile polyphosphates. 44, 173-175. Bazzaz, F. A. (1990). The response of natural ecosystems to the rising global CO2 levels. 21, 167-196. Bazzaz, F. A., and Fajer, E. D. (1992). Plant life in a CO2-rich world. 266, 68-74. Bazzaz, F. A., and Garbutt, K. (1988). The response of annuals in competitive neighbourhoods: Effects of elevated CO2. 69, 937-946. Bazzaz, F. A., and McConnaughay, K. D. M. (1992). Plant-plant interactions in elevated CO2 environments. 40, 547-563. Bazzaz, F. A., Garbutt, K., Reekie, E. G., and Williams, W. E. (1989). Using growth analysis to interpret competition between a C3 and a C4 annual under ambient and elevated CO2. 79, 223-235. Campbell, B. D., and Grime,J. P. (1992). An experimental test of plant strategy theory. 73, 15-29. Campbell, B. D., Grime, J. P., and Mackey, J. M. L. (1991). A trade-off between scale and precision in resource foraging. 87, 532-538. Campbell, B. D., Laing, W. A., and Newton, P. C. D. (1993). Variation in the response of pasture plants to carbon dioxide. "Proceedings of the XVII International Grassland Congress," pp. 1125-1126. Crush, J. R., and Campbell, B. D. (1993). Effect of different grass species on nitrogen fixation by white clover under conditions of elevated carbon dioxide and temperature. "Proceedings of the XVII International Grassland Congress," pp. 1130-1131. Grime, J. P. (1977). Evidence for the existence of three primary strategies in plants and its relevance to ecological and evolutionary theory. 111, 1169-1194. Grime, J. P. (1979). "Plant Strategies and Vegetation Processes." Wiley, Chichester. Grime, J. P., Hodgson, J. G., and Hunt, R. (1988). "Comparative Plant Ecology: A Functional Approach to Common British Species." Unwin Hyman, London. Hardacre, A. K., Laing, W. A., and Christeller, J. T. (1986). The response of simulated swards of perennial ryegrass and white clover to enriched atmospheric CO2: Interaction with nitrogen and photosynthetic photon flux density. N. Z.J. 29, 567-573. Harris, W. (1987). Population dynamics and competition. "White Clover" (M. J. Baker and W. M. Williams, eds), pp. 203-298. CAB International, Wallingford. Harris, W. (1990). Pasture as an ecosystem. "Pastures: Their Ecology and Management" (R. H. M. Langer, ed.), pp. 75-131. Oxford Univ. Press, Oxford. Hart, A. L. (1989). Nodule phosphorus and nodule activity in white clover. 32, 145-149. Hart, A. L., and Collier, W. A. (1994). The effects of phosphorus and form of nitrogen supply on leaf cell size and nutrient content in and 49, 96-104.
20. Hart, A. L., and Greer, D. H. (1988). Photosynthesis and carbon export in white clover plants grown at various levels of phosphorus supply. 73, 46-51. Hart, A. L., and Jessop, D. (1984). Leaf phosphorus fractionation and growth responses to phosphorus of the forage legumes and 61, 435-440. Haslemore, R. M., and Roughan, P. G. (1976). Rapid chemical analysis of some plant constituents. 27, 1171-1178. Irving, G. C.J., and Bouma, D. (1984). Phosphorus compounds measured in a rapid and simple leaf test for the assessment of the phosphorus status of subterranean clover. 24, 213-218. Jackman, R. H., and Mouat, M. C. H. (1972). Competition between grass and clover for phosphate. I. Effect of browntop Sibth.) on white clover L.) growth and nitrogen fixation. 15, 653-666. Mitchell, K.J. (1956). The influence of light and temperature on the growth of pasture species. "Proceedings of the 7th International Grassland Congress," pp. 58-69. Newton, P. C. D., Clark, H., Bell, C. C., Glasgow, E. M., and Campbell, B. D. (1994). Effects of elevated CO2 and simulated seasonal changes in temperature on the species composition and growth rates of pasture turves. 73, 53-59. Overdiek, D., and Reining, F. (1986). Effect of atmospheric CO2 enrichment on perennial ryegrass L.) and white clover (Trif01ium L.) competing in managed model-ecosystems. 7, 357-366. Overdiek, D., Bossemeyer, D., and Lieth, H. (1984). Long-term effects of an increased CO2 concentration level on terrestrial plants in model-ecosystems. 3, 344-352. Pearcy, R. W., Tumosa, N., and Williams, K. (1981). Relationships between growth, photosynthesis and competitive interactions for a C3 and a C4 plant. 48, 371-376. Reay, P. F. (1985). An improved determination of ammonia in Kjeldahl digests and acidic solutions with a buffered Berthelot reaction. 176, 275-278. Rogers, H. H., Runion, G. B., and Krupa, S. V. (1994). Plant responses to atmospheric COz enrichment with emphasis on roots and the rhizosphere. 83, 155-189. Welden, C. W., and Slausen, W. L. (1986). The intensity of competition versus its importance: An overlooked distinction and some implications. 61, 23-44. Williams, C. H., and Twine, J. R. (1967). "Determination of Nitrogen, Sulphur, Phosphorus, Potassium, Sodium, Calcium and Magnesium in Plant Material by Automatic Analysis." Commonwealth Scientific and Industrial Research Organisation, Division of Plant Industry Technical Paper No. 24, pp. 3-16. Woodward, F. I. (1992). Predicting plant responses to global environmental change. New 122, 239-251. Woodward, F. I., Thompson, G. B., and McKee, I. F. (1991). The effects of elevated concentrations of carbon dioxide on individual plants, populations, communities and ecosystems. 67, 23-38. Zangerl, A. A., and Bazzaz, F. A. (1984). The response of plants to elevated CO2. II. Competitive 62, 412-417. interactions among annual plants under varying light and nutrients.
This Page Intentionally Left Blank
21 Impact of Elevated C02 Concentration on Interactions between Seedlings of Norway Spruce (Picea abies) and a Perennial Grass
Calamagrostis epigejos
Spontaneous regeneration of tree stands from seeds is of crucial importance for the sustainable development of natural forest ecosystems and man-made plantations. Successful growth of young tree seedlings in a forest understory or in clearings is a very delicate process, sensitive to various environmental conditions, and its change with increasing CO2 concentration in the atmosphere seems to be highly probable. Undisturbed and completely closed coniferous forests of central Europe usually have very scarce vegetation in the understory. Daily integrals of penetrating photosynthetic active photons are not seldom below 1 mol m -z, and it is too low for growth of both herbs and tree seedlings. Their successful development is possible only after a thinning of the tree canopy. The thinning may be done intentionally (by selective cutting of trees) or due to some other causal factors (insect defoliation, disease, abiotic stresses including anthropogenic pollutants). Early growth and development of tree seedlings in more favorable radiation environment are influenced by competitive interactions with herbaceous vegetation dominated mainly by grasses. In contrast to the majority of grass species, the tree seedlings are inherently slow growers without and
319
Copyright 9 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.
320 tiller formation or other forms of horizontal spreading, which makes their competitive ability very weak. It is well known that they have only little chance to survive in sites with ample supply of resources, e.g., in completely open forest clearings, which are quickly colonized by dense stands of herbaceous vegetation. Successful growth of tree seedlings is thus possible only in a rather narrow range of environmental conditions, unfavorable enough for the fast growth of potential herbaceous competitors (mainly grasses), but still not critical for their own growth. It is obvious that the balance between tree seedlings and population of grasses under such conditions depends on differences in their stress tolerance. The most favorable habitats for forest regeneration can be found in the understory of semiopen tree stands, where incident radiation is the most limiting resource for plant growth. This may indicate that shade tolerance of tree seedlings successfully growing under such conditions must be a little bit higher than in co-occurring and potentially fast-growing herbaceous vegetation, but proper comparative measurements have never been realized. Shade tolerance is not a fixed property of a plant species, but it can be changed by various biochemical, physiological, and morphogenetical acclimatory mechanisms (Bj6rkman, 1981; Pearcy and Sims, 1994). Interspecific differences in the capability to acclimate to low radiation are considerable, and may be modified by various others factors, such as temperature, light quality, and nutrient availability (Smith, 1981; Thompson 1992; Boardman, 1977). It is highly probable that the shade tolerance may also be altered by CO2 concentration in air. Elevated CO2 usually increases quantum yield of photosynthesis and decreases light compensation point of net CO2 uptake (Long and Drake, 1991; Gloser and Bart~ik, 1994), which are essential prerequisites for increased shade tolerance of plants. In this chapter I will first present results of some experiments focused on determination of shade tolerance of spruce seedlings and grasses of the genus at ambient and elevated CO2 concentration in air. Possible consequences of elevated CO2 for changes in competitive interactions between spruce seedlings and the grasses are then discussed.
Norway spruce L. Karst) is undoubtedly the most important forest conifer in central Europe. Its physiological study has been prompted recently, particularly in the context of several ecological projects searching for causes of large-scale forest decline (for review see Klein and Perkins,
21.
321
1988; Schulze, 1989). Spruce plantations and natural stands occur frequently on acid soils with a low amount of mineral nutrients. Among the few herbaceous species, which can be found in the understory, perennial grasses prevail. The most widespread and abundant is probably (L.) Drejer (syn. L.). This low and inherently slow-growing species (Poorter 1990) is a rather weak competitor even after canopy opening. Grasses of the genus and C. are also extremely widespread in spruce woodlands. In contrast to they have very high competitive potential. All belong to the inherently fast-growing sun plants, but their ability to persist in shaded habitats is remarkable. In improved light conditions they expand rapidly by vegetative spreading and form dense mats with peak aboveground standing crop above 1 kg m -2 (Pysek, 1991). The best performance in changing radiation environment and the strongest competitive ability from all of the above-mentioned species was shown by in our previous field observations and experiments (Gloser and Gloser, 1996). Therefore, we have used this species in most of our recent work. For comparative assessment of shade tolerance, seedlings of and Norway spruce were grown in plastic containers with inorganic substrate (Perlite) and Hoagland nutrient solution using two growth chambers with different C O 2 concentrations in the air (350 cm ~m -3 and 700 cm ~m-S). The plants in each chamber were grown at two different levels of photosynthetic photon flux (PPF). One-half of the plants was exposed to 200/xmol m -2 s -1 PPF (which was equivalent to about 11.5 mol m -2 day -1 at 16 h photoperiod), the second half was grown at 26/xmol m -2 s -1 PPF (about 1.5 mol m -2 d a y - l ) . Relative growth rate (RGR) of the plants and some other growthanalytical parameters were estimated by destructive sampling. Gas exchange parameters (photosynthetic uptake of CO2 by leaves, respiratory CO2 output from all organs) were measured using an open-type system with differential infrared gas analyzer (for more details see Gloser and Bart~k, 1994; Gloser and Gloser, 1996). The two levels of radiation applied in our experiments were chosen intentionally, in order to simulate two interesting radiation environments observed in the field. As was found in several localities, the daily photon flux (during summer months) of about 1.5 mol m -2 day -1 is usually very close to the lower limit for growth of spruce seedlings and grasses. At values above 10 mol m -2 d a y -1, the grass stands in the forest understory become too dense for successful establishment and growth of spruce seedlings. ,
A. "Acclimation to Low Light atAmbient CO2 Concentration in Air AcClimation to extremelylow, incident radiation had very: gimilar features in both species: decreased rate of metabolic processes (net photosynthesis,
dark respiration), highly increased values of specific leaf area, and much smaller changes in biomass allocation (Table I). The overall capture and utilization of incident radiation, integrally manifested in growth rates, was much better in than in spruce seedlings. The analysis of results indicated that the key characteristics responsible for the interspecific differences in RGR at both light treatments were net photosynthetic rate (PN), leaf area ratio (LAR), and specific leaf area (SLA). Relative changes in these parameters under the influence of low PPF were very similar in both species. Acclimatory changes in specific respiration rate were much more pronounced in than in spruce. The overall shape of light curves of PN (measured in leaf blades only!) was quite different for plants cultivated at different light regimesm maximum values of PN (at light saturation) and saturating values of PPF were always much smaller in plants from low light treatment (Fig. 1). The
Growth regime:
CO2 Radiation
(cm 3 m -a) (mol m -2 day -1 )
P max app R shoot R root RGR EAR LWR SLA P max app R shoot R root RGR LAR LWR SLA
350
700
1.5
11.5
1.5
11.5
(/zmol m -2 s -1) (mol mo1-1) (/zmol kg -1 s -1) (/zmol kg -1 s -1) (mg g-1 day-l) (m 2 kg -1) (kg kg -1) (m 2 kg -1)
6.7 0.058 8.0 10.1 7.5 9.42 0.65 14.5
10.8 0.059 11.5 16.5 30.2 6.77 0.67 10.1
10.5 0.070 8.2 11.5 8.8 8.28 0.60 13.8
15.4 0.071 12.0 17.4 37.1 5.41 0.57 9.5
(/zmol m -2 s -1) (mol mo1-1) (/.~mol kg -1 s -1) (/zmol kg -1 S-1) (mg g-1 day-l) (m 2 kg -1 (kg kg -1) (m 2 kg -1)
8.5 0.055 9.5 8.1 28.2 29.81 0.55 54.2
15.1 0.057 18.2 23.0 95.1 19.28 0.51 37.8
12.8 0.067 11.0 9.0 35.1 27.04 0.52 52.0
21.3 0.069 21.0 25.4 114.8 14.76 0.46 32.1
Temperature during cultivation was 25/20~ photoperiod, 16 h. P max: maximum net photosynthetic CO2 uptake by leaves (measured at light saturation, 20~ and growth CO~ concentration); 9 app: apparent quantum yield of photosynthesis; R: dark respiration rate, measured as CO~ output at 20~ RGR: maximum relative growth rate; LAR: leaf area ratio (projected leaf area per whole-plant dry mass); LWR: leaf weight ratio (leaf dry mass to whole plant dry mass); SLA: specific leaf area (projected leaf area per leaf dry mass unit).
21. Impact of Elevated C02 Concentration
~
w bft.
w -r I---
,
:-:2
.-
-
-
Figure 1 Relationship between the net photosynthetic rate (expressed per unit leaf area, top graphs, and per leaf dry mass unit, bottom graphs) and photosynthetic p h o t o n flux for plants grown in two different radiation regimes (1.5 mol photons m -2 day -] and 11..6 mol p h o t o n s m -2 day-l). The plants were grown and m e a s u r e d at ambient CO2 concentration in air (350 cm 3 m -3, o p e n circles) or at elevated CO2 concentration in air (700 cm 3 m -3, solid circles).
interspecific differences were less pronounced. It should be mentioned that there were even smaller differences between the two species in their PN (per unit leaf area), when compared at pertinent growth light intensifies (26 or 200/zmol m -2 s -1 PPF). In spite of striking differences in shape of the light curves between plants from high and low light environments, apparent q u a n t u m yield of photosynthesis (derived from initial slope of light curves) was very similar in all cases. This is in accordance with the finding of Evans (1987) and some other authors, as reviewed by Boardman (1977) and Bj6rkman (1981),
that adaptation to shade is usually not connected with changes in quantum yield. Comparisons based on instantaneous PN calculated per unit leaf area did not include all operational costs of the leaves (night respiration) and costs of leaf construction. Night respiration can be incorporated by calculating PN on a 24-h cycle. Leaf construction costs can be indirectly included expressing PN per leaf mass unit, as suggested by Givnish (1988). In this case, the better performance of becomes especially apparent (lower graphs in Fig. 1). When expressed per dry mass unit, the PN values of were much higher than PN of spruce, even at very low PPF during measurement. The leaves of were much thinner and, consequently, less costly in terms of energy investment into growth of unit leaf area than leaves of It is generally accepted that carbon economy is of crucial importance for shade-tolerant species. Whole-plant daily carbon gain is undoubtedly a more appropriate characteristic in this respect than any of the photosynthetic or respiratory parameters alone. It is usually difficult to measure diurnal CO2 exchange of whole intact plants when grown in a nonsterile substrate because of high evolution of CO2 from associated microorganisms in the root compartment. In our case, the daily carbon gain was calculated from separate gasometric measurements of shoots and roots. The shoots were covered during measurements by a special foil enclosure, in order to preserve the spatial orientation of leaves intact. Root respiration was measured subsequently (on the second day) in detached and gently washed roots. As can be seen from Table II, the interspecific differences in whole-plant carbon gain were caused mainly by photosynthetic carbon u p t a k e m differences in the daily integral of respiratory carbon loss were much smaller. Daily dry mass increments (calculated from carbon gain) were quite close to values of RGR estimated by direct harvest method. All the characteristics given in Table II were expressed in a standardized form (per 1 g of whole-plant biomass) for the sake of better comparability. Another particularly useful criterion for the quantification of shade tolerance may be the value of photosynthetic photon flux at which the diurnal whole-plant carbon gain is zero--light compensation point of the wholeplant carbon g a i n (CPCG). The m o r e shade-tolerant plants should have positive carbon balance at lower photon flux than the less tolerant plants. Direct experimental estimation of CPCG is rather difficult and time consuming, iItiseems: m u c h m o r e feasible a n d ve~satile tocalculate~ ~alues of CPCG from imeasuremen tsof. ph0tosyn thetic carbon ~uptake :and~integral carbon losses ;in respiration, (Fig, 2). Fully acclimated::plants of~:.G always had~islightly lower CPCG:~than :swuee seedlings.
21.
Plant species: CO2 concentration in air:
350
700
350
700
Photosynthetic C-uptake Respiratory C-loss Net whole-plant C-gain Daily dry mass increment (mg g-l) CPCG (mol m -2 d a y -1 PPF)
7.96 4.95 3.01 7.92 1.03
9.29 5.48 3.81 9.29 0.98
16.58 4.61 11.97 30.50 0.82
20.25 5.24 15.01 37.52 0.77
The data were normalized to refer to whole-plant mass of 1 g. Data on daily dry mass increments of plants (which are in this case equivalent to RGR) were calculated from carbon gain. Calculation of the CPCG (light compensation point of the daily whole-plant carbon gain) is shown in Fig. 2.
c~ Picea
Calamagrostis /
-Z~OA I O m
m <.__. 6 o ~2 _1 <
example of calculation of the light compensation points of daily whole,plant carbon~gain (CPCG) for plants and grown at extremely low daily photon flux (1 5 mol m -2 da -1) and at two different CO2 concentrations in air. Daily photosynthetic carbon uptake, measured a t two levels of PPF, is depicted by circles;daily respiratory' losses of carbon (RA and P~ for ambient and elevated C02 ~concentrations respectively) are shown. as ;ih6rizontal lines. Based on data presented 'in 'Table II~
326 B. Changes in Shade Tolerance at Elevated CO2 Elevated C O 2 had a very p r o n o u n c e d effect on growth rate and on photosynthetic characteristics of leaves of both species grown in high and low radiation regimes. E n h a n c e m e n t of light-saturated PN in elevated CO2 was very similar (about 40%) in leaves of both species grown at higher irradiance. The plants grown at lower light even responded slightly more sensitively. As can be seen from Fig. 1, the effect of CO2 concentration during measurement of PN declined with decreasing PPF, but it was significant even at the lowest PPF applied (10/~mol m -~ s-i). Apparent quantum yield of photosynthesis increased in leaves exposed to elevated CO2 concentration by about 20% in both species from both light treatments. Some slight increase in specific dark respiration rates in plants grown at elevated CO2 were observed. A positive effect of elevated CO2 concentration on net photosynthesis at low irradiance was reported in many other species, both herbaceous (Long and Drake, 1991; Nijs 1989; Campbell 1990), and in conifers (Grulke 1993; Higginbotham 1985). The positive effect of CO2 at low light is best expressed in increased values of quantum yield of photosynthesis, which is evidently a consequence of inhibited photorespiration. The light compensation point should be lowered in this case, but changes in mitochondrial respiration may interfere. There was no significant decrease in light compensation point in leaves of our plants at elevated CO2. Similar results were reported by Grulke (1993). In any case, higher quantum yield at elevated CO2 may substantially increase carbon gain and relative growth rate of shaded plants, as can be d o c u m e n t e d by our results (Tables I and II). Relative values of responses of spruce seedlings and to elevated CO2 were unexpectedly similar in most parameters measured, in spite of substantial differences in morphology and growth potential of the two species. The inherently slower growing plants are usually considered as less responsive to elevated CO2 (Hunt 1991; Bowler and Press, 1993). Final evaluation of the impact of elevated CO2 concentration in air on the whole-day carbon balance is summarized in Table II. The plants grown at elevated CO2 concentration increased their daily carbon gain by about 26.6% and 25.4% (Calamagr0stis). Significant decrease of CPCG was found only in (by about 6%). Importantly, no acclimation to elevated CO2 concentration in air was detectable in our plants even after 80 days of cultivation. This was tested in some additional measurements in a reversed CO2 regimen: plants grown at 350 cm 3 m -3 were measured at 700 cm ~ m -s, and vice versa (see also Gloser and Bartfik, 1994). All of the data presented in Tables I and II was collected with plants grown and measured at the same CO2 concentration.
21.
327
The analytical data on responses of individual plants of juvenile spruce and perennial grass to irradiance and CO2 concentration can be utilized most efficiently after their scaling up to population or to community level, as r e c o m m e n d e d by Bazzaz (1993). There is not enough field data yet to create a detailed plant-environment model. However, some speculation on behavior of a simple, two-population community may be done. The growth form and growth strategy of conifers and grasses are completely different, which makes their interaction very peculiar and variable in time. Perennial grasses surviving in the understory of a dense forest must possess all important features of stress-tolerant plants Grime, 1979). After a thinning of the tree canopy, the potentially fastgrowing grasses (like our switch to competitive strategy, exploiting resources rapidly. Spruce seedlings during the first few years after germination can also be classified as typical stress tolerators surviving in a low-resource environment. Inherently slow growth of spruce seedlings is caused primarily by investment of carbon and energy into long-lived leaves and supporting axial structures. Thanks to these morphogenetical features they are able to overgrow and, eventually, to repress the grasses over some years. As we can deduce from this very brief outline, long-term competitive success and dominance of grasses in forest clearings is possible only on the condition that the co-occuring tree seedlings are driven by their competitors to extinction, and not only inhibited in growth. In this respect, knowledge of the critical physiological points (like the CPCG) is very important. The tree seedlings have very little chance of survival in fully developed stands of or because the leaf area index of these grasses is regularly very high (8-12) and PPF below the leaf canopy is diminished to about 0.5% of the above-stand PPF (Gloser, unpublished). The daily integrals of PPF are in this case much lower than the CPCG of spruce seedlings, as was found in our experiments. In addition to aboveground competition for light, interactions of spruce and grasses in belowground compartments should be considered. Relative importance of the two types of competition was tested in experiments with plant mixtures. Seedlings of Norway spruce and were grown in plastic containers filled with homogenized soil transported from a field locality. Four plants of each species were planted into each container in the " m o n o c u l t u r e " variant, and 4 + 4 plants in the " m i x t u r e " variant. The containers of the " m i x t u r e " variant were divided into two subvariants.
In the first o n e the interspecific c o m p e t i t i o n for light was e x c l u d e d by b e n d i n g aside the grass shoots (by m e a n s o f nylon threads). In this case light h a d free access to the spruce seedlings p l a n t e d in the central part of the containers. In the second subvariant the grass shoots were kept intact. Cultivation of all variants was r e p e a t e d in a m b i e n t a n d at elevated CO2 (350 cm ~ m -~ a n d 700 cm ~ m -3, respectively). Results of the e x p e r i m e n t (Table III) c o n f i r m e d o u r hypothesis, that growth o f spruce seedlings in the p r e s e n c e of is inhibited mainly by c o m p e t i t i o n for light in the shoot c o m p a r t m e n t (shading by taller grass leaves). T h e spruce seedlings did n o t r e s p o n d to shading by e l o n g a t i o n of shoots. T h e i r h e i g h t after nearly 3 m o n t h s of cultivation was in all variants in the r a n g e f r o m 70 to 100 mm. Spruce seedlings h a d n o detectable effect o n the growth of m u c h taller ( 4 0 0 - 5 0 0 m m ) plants of as was f o u n d by c o m p a r i s o n of grass growth rate with a n d without the p r e s e n c e o f spruce seedlings. After 80 days of cultivation at elevated CO2, the biomass o f grass plants was m o r e than two times h i g h e r than at a m b i e n t CO2 concentration. T h e n u m b e r of new fillers a n d the leaf area were also nearly d o u b l e d at elevated CO2. T h e growth of spruce seedlings cultivated in m i x t u r e with was always significantly inhibited. It was n o t possible to find any beneficial effect of elevated CO2 o n total dry mass i n c r e m e n t s of spruce seedlings growing in mixture. As was shown already, elevated CO2 may normally increase growth rate of plants in shade. In o u r e x p e r i m e n t the beneficial effect of CO2 o n spruce seedlings was c o u n t e r a c t e d by g r e a t e r light i n t e r c e p t i o n within the m o r e stimulated grass
Dry mass (mg)
Leaf area (cm2)
61 + 7 112 +_ 15 1635 + 132 3640 + 310
4.2 _+ 0.5 6.3 _+ 0.8 325.1 _+ 30.2 518.5 -+ 45.8
Without shoot interaction C350 C700 C350 C700 9 With shoot interaction
,2
C350 C700 C350 C700
30_+3 32 _+5 1339 _+ i21 2980 + 270
Mean valuesfrom 16 plants +_1 S.E.
"
'
2.8 -+ 0.3 . . . . . . 2.5 + 0,3 ~: ~2'-~78.4~-I~'_30.0 +~ ~ .... :!4681iii,__442.5 ~ . ,, 9
21.
329
shoots. The amount of leaf area of given in Table 3, was equivalent to LAI about 7.3 in C700 treatment and 4.4 in C350 treatment. The spruce seedlings in C700 treatment (with shoot interaction) had already starved by the last weeks of the experiment.
Any estimation of potential impact of elevated COz on plant-plant interactions is a complex task, impossible to be done without some simplification (Bazzaz and McConnaughay, 1992; Woodward 1991). The most straightforward approach to the study of interactions between tree seedlings and grass populations at elevated CO2 would be long-term field experiments with CO2 enrichment over representative segments of forest floor. To my best knowledge, no such field studies of this phenomena have been properly carried out so far. A more simplified approach using artificially planted tree-grass mixtures grown in controlled conditions and extrapolation from analytical studies of singly grown plants was used in our work. Consequently, the presented results are only the first approximation to the problem. It is also quite clear that some other variables, such as nutrient and water availability or symbiotic and parasitic biota, may alter the interactions. Nevertheless, if we admit that competition for light and shade tolerance of competing plants has the most important role in the interactions, and that the whole-plant carbon gain is a suitable integral measure of shade tolerance, it is possible to conclude the following: (a) Elevated CO2 concentration in air will enhance the potential growth rate of spruce seedlings even in the forest understory with very low values of incident radiation. The increase in growth rate of these seedlings will be of a similar magnitude as a growth stimulation of the competing grass Thanks to the inherently higher growth rate of C. the increase in its biomass and leaf area will be much higher than in spruce seedlings. This, together with inherently fast growth of grass shoots and enhanced tillering rate, will substantially strengthen the competitive ability of a grass population, at least during the first few years of spruce seedling growth. (b) As was derived from our experiments in simplified conditions, the light compensation point of whole-plant carbon gain (CPCG) of spruce seedlings was not much lower at elevated COz. It is highly probable that a similar response could also be found in field-grown plants. We may thus predict that the forest zone with beneficial conditions for the growth of tree seedlings will not be shifted simply to a deeper shade. The narrowing of this zone and thus impaired spontaneous reforestation on large areas of coniferous forests is the most probable scenario for the near future.
This work was supported by Grant 501/94/0493 from the Grant Agency of the Czech Republic.
Bazzaz, F. A. (1993). Scaling in biological systems: Population and community perspectives. "Scaling Physiological Processes Leaf to Globe" (J. R. Ehleringer and C. B. Field, eds.), pp. 233-254. Academic Press, San Diego. Bazzaz, F. A., and McConnaughay, K. D. M. (1992). Plant-plant interactions in elevated CO2 environments. 40, 547-563. Bj6rkman, O. (1981). Responses to different quantum flux densities. "Encyclopedia of Plant Physiology" (O. L. Lange, P. S. Nobel, C. B. Osmond, and H. Ziegler, eds.), New Ser., Vol. 12A, pp. 57-107. Springer-Verlag, Berlin. Boardman, N. K. (1977). Comparative photosynthesis of sun and shade plants. 28, 355-377. Bowler, J. M., and Press, M. C. (1993). Growth responses of two contrasting upland grass species to elevated CO2 and nitrogen concentration. 124, 515-522. Campbell, W. J., Allen, L. H., and Bowes, G. (1990). Response of soybean canopy photosynthesis to CO2 concentration, light, and temperature. 41, 427-433. Evans, J. R. (1987). The dependence of quantum yield on wavelengh and growth irradiance. 14, 69-79. Givnish, T.J. (1988). Adaptation to sun and shade: a whole-plant perspective. 15, 63-92. Gloser, J., and Bart~ik, M. (1994). Net photosynthesis, growth rate and biomass allocation in a rhizomatous grass grown at elevated COz concentration. 30, 143-150. Gloser, V., and Gloser,J. (1996). Acclimation capability of and to changes in radiation environment. 32, 203-212. Grime, J. P. (1979). "Plant Strategies and Vegetation Processes." Wiley, Chichester. Grulke, N. E., Hom, J. L., and Roberts, S. W. (1993). Physiological adjustment of two full-sib families of ponderosa pine to elevated CO2. 12, 391-401. Higginbotham, K. O., Mayo, J. M., Hirondelle, S. L., and Krystofiak, D. K. (1985). Physiological ecology of lodgepole pine in an enriched COz environment. Res. 15, 417-421. Hunt, R., Hand, D. W., Hannah, M. A., and Neal, A. M. (1991). Responses to CO2 enrichment in 27 herbaceous species. 5, 410-421. Klein, R. M., and Perkins, T. D. (1988). Primary and secondary causes and consequences of contemporary forest decline. 54, 1-43. Long, S., and Drake, B. G. (1991). Effect of the long-term elevation of CO2 concentration in the field on the quantum yield of photosynthesis of the C3 sedge, 96, 221-226. Nijs, I., Impens, I., and Beaheghe, T. (1989). Leaf and canopy responses of to long-term elevated atmospheric carbon-dioxide concentration. 177, 312-320 Pearcy, R. W., and Sims, D. A. (1994). Photosynthetic acclimation to changing light environments: scaling from the leaf to the whole plant. "Exploitation of Environmental Heterogeneity by Plants" (M. M. Caldwell and R. W. Pearcy, eds.), pp. 145-174. Academic Press, San Diego.
21.
331
Poorter, H., Remkes, C., and Lambers, H. (1990). Carbon and nitrogen economy of 24 wild species differing in relative growth rate. 94, 621-627. Pysek, P. (1991). Biomass production and size structure of populations in different habitats. 63, 9-20. Schulze, E.-D. (1989). Air pollution and forest decline in a spruce forest. 244, 776-783. Smith, H. (1981). Adaptation to shade. "Physiological Processes Limiting Plant Productivity" (C. B. Johnson, ed.), pp. 159-173. Butterworths, London. Thompson, W. A., Kriedemann, P. E., and Craig, I. E. (1992). Photosynthetic response to light and nutrients in sun-tolerant and shade-tolerant rainforest trees. I. Growth, leaf anatomy and nutrient content. 19, 1-18. Woodward, F. I., Thompson, G. B., and McKee, I. F. (1991). The effect of elevated concentrations of carbon dioxide on individual plants, populations, communities and ecosystems. 67 (Suppl.), 23-38.
This Page Intentionally Left Blank
2 The Effect of Elevated CO2 on Developmental Processes and Its Implications for Plant-Plant Interactions
As clearly demonstrated by other chapters in this book, elevated C O 2 has the potential to substantially alter competitive processes and change community composition (Arnone, Chapter 8; Campbell and Hart, Chapter 20; Chiariello and Field, Chapter 10; Gloser, Chapter 21; Leadley and K6rner, Chapter 11; Polley Chapter 12; Reynolds, Chapter 18; Roy Chapter 9; Scarascia-Mugnozza Chapter 14; see also Bazzaz, 1990). Such changes in community composition could have important ecological consequences; for example, changes in community composition have the potential to influence whole-ecosystem response to CO2. Ben Bolker and colleagues recently developed a model that examines how variation in species response to CO2 may influence growth of northern temperate forests (Bolker 1995). They modeled response to a doubling of CO2 assuming either (1) species differed in the extent to which elevated CO2 increased growth using actual enhancement data for individual species from greenhouse studies, or (2) all species responded in the same manner to elevated CO2 (i.e., the mean response of species in the greenhouse study). Due to increases in the importance of responsive species, incorporating species variation substantially increased the response of the forest as a whole. The species diversity effect is 260% larger than the mean effect after 100 years. Changes in species composition with elevated CO2 will also have direct economic consequences. In forest ecosystems and Communities
333
Copyright 9 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.
species composition may alter the balance between economically unimportant and more desirable species. Similarly, changes in the species composition of pastures could have important consequences for herbage quality. Clearly, the ability to predict how community composition will change with elevated CO2 should be one of the prime objectives of current research. Unfortunately, our ability to predict how community composition will change with elevated CO2 is limited. Experiments with laboratory microcosms and field experiments with small-statured communities provide information for a small subset of communities under specific environmental conditions. To make accurate predictions for a wide range of conditions we will need to develop mechanistic models of plant competition under elevated CO2 (Dahlman, 1985; Reynolds and Acock, 1985). Because the development of such models is still in its infancy, I wish to discuss the type of information and data such models will need to incorporate if we are to accurately predict the results of competition. In particular, I wish to illustrate that the effect of elevated CO2 on developmental processes may be an overlooked, but important, issue. To do so, I will review some recent studies that examine the effect of elevated CO2 on development. The focus will be on work that I have been involved with, not because these studies are necessarily the best examples, but because they are the studies with which I am most familiar.
To date, attempts to predict the effect of elevated C O 2 o n competition have focused largely on the effects on growth of the individual species (e.g., Bolker 1995). It is well established that there are marked differences among species in the extent to which growth is enhanced by elevated CO2 (Cure, 1985; Hunt 1991; Poorter, 1993). It has also been experimentally demonstrated that differences in responsiveness can lead to marked changes in competitive balance when species are grown in mixtures (e.g., Carter and Peterson, 1983). Obviously, a knowledge of the relative responsiveness of species to elevated CO2 will be crucial to attempts to model the effect of C O 2 o n community composition. Elevated CO2, however, is also known to affect a wide range of plant processes including many aspects of development such as leaf morphology and anatomy, biomass allocation, time of flowering, rate of senescence, branching patterns, and stem elongation (Strain and Cure, 1985; Eamus and Jarvis, 1989; Bazzaz, 1990). Rate of development has been shown to have very marked effects on competitive outcome. Ross and Harper (1972), for example, have shown that even minor differences in time of emergence among seedlings can have dramatic consequences for the ability of plants
22.
on
335
to compete with their neighbors. Therefore, it has been predicted that CO2-induced changes in rate of development are likely to affect community composition (St. O m e r and Horvath, 1983). Developmental changes also have the potential to alter the ability of species to compete for resources through changes in morphology. For example, CO2-induced changes in stem elongation and branching patterns will affect canopy structure. Canopy structure has been shown to have a marked influence on the ability of species to compete for light (e.g., Gaudet and Keddy, 1988). To illustrate the importance of CO2-induced changes in development for competitive outcome, I will describe a study that examines the effect of C O 2 o n competition among seedlings of five tropical trees: and (Reekie and Bazzaz, 1989). Plants were grown both as individuals and in competitive arrays at 350, 525, or 700 p p m CO2. When grown in competitive arrays, marked differences existed among CO2 levels in community composition. decreased in importance with CO2 while and increased in importance (Fig. 1). Given that there were no differences among species in growth response to CO2 when grown as individuals, the marked changes in community composition are surprising. One explanation for the change in community composition is possible effects of CO2 on the availability of other resources. Decreases in stomatal conductance with elevated CO2, for example, might increase
Figure I Effectof atmospheric CO2on compositionof a simple, artificial communitycontaining seedlings of five tropical trees. Percentage composition values are based on total aboveground biomassat the end of the experiment. From Reekie and Bazzaz (1989), with permission from Springer-Verlag.
water availablity, whereas increases in growth with C O 2 might decrease light availability. Therefore, if these species differ in their ability to compete for either of these resources, this could explain the shift in community composition even though the species responded similarly to CO2 when grown as individuals. However, the shift in species composition was not correlated with changes in either light or water availability but with changes in canopy structure. The height at which leaves were displayed decreased with CO2 in and increased in and to a lesser extent in the remaining species, increasing the ability of the latter species to compete for light (Fig. 2). The changes in canopy structure with CO2 appeared to be due to a combination of developmental changes including height growth, branching patterns, and leaf senescence.
The fact that elevated C O 2 c a n alter the ability of plants to compete through its effects on development is disturbing, because it will make it substantially more difficult to predict how community composition may be altered by CO2. Not only must we have information on how overall growth of individual species is affected by CO2, we must have knowledge of its effect on developmental patterns. Particularly disturbing is that these developmental effects, although often very marked, are not necessarily predictable. For example, we conducted a study that examined the effect of elevated COz on time to flowering in a group of forbs from an annual grassland in Texas (Reekie and Bazzaz, 1991). Plants were grown both as individuals and in competitive arrays and resource availability was varied by growing plants in different pot sizes. Depending on species and environmental conditions, elevated COz delayed, hastened, or had no effect on time to flowering (Table I). Variable effects of CO2 on time of flowering as well as on other aspects of development have been observed in a number of different studies besides this one (Hesketh and Hellmers, 1973; Paez 1980; St. Omer and Horvath, 1983; Garbutt and Bazzaz, 1984; Marc and Gifford, 1984). Before we can even hope to predict how elevated CO2 will influence competition from a knowledge of its effects on individual species (Reynolds and Acock, 1985), we will have to first be able to predict how CO2 will influence development under specific circumstances for individual species. In the same way we have attempted to place species in response categories based on the effect of CO2 on growth (e.g., Cure, 1985; Hunt 1991; Poorter, 1993; Grime, Chapter 6; Roy C h a p t e r 9 ) , we must also do this for the CO2 effect on development. To illustrate this I will describe
22.
on 35O
700
~" 1.0 v
0.2
Effect of atmospheric CO2 on leaf area profile of seedlings of five tropical trees. Each division on the horizontal axis represents 0.01 m 2. Leaf area was summed over 0.1-m intervals for presentation. The designations C, M, P, S' and T represent ~ The numbers given above each species represent m e a n canopy height (m) for that treatment. Plants were grown as individuals but plants grown competitiVely showed a similar response. From Reekie and Bazzaz (1989), with permission from Springer-Verlag.
338
Species
Pot size ( m l )
Competition
Effect on phenology
500 1000
No Yes
Hastens flowering Hastens flowering
500 1000
No Yes
Hastens flowering No effect
150 500 1000 1000
No No No Yes
Hastens flowering Hastens flowering No effect No effect
150 500 1000 1000
No No No Yes
Delays flowering Delays flowering No effect No effect
two recent experiments that attempt to find some pattern in the effect of CO2 on plant development. One possible explanation for the highly variable effects of elevated CO2 on development comes from the plant development literature. Physiologists investigating the mechanism of the photoperiodic response have noted that extremely high CO2 levels ( ~ 1-5%) can modify the response of plants to daylength. Purohit and T r e g u n n a (1974) observed that u n d e r short-day conditions, elevated CO2 delayed or inhibited flowering in the short-day plants, and and induced flowering in the normally long-day plant Hicklenton and Jolliffe (1980) confirmed that elevated CO2 delays flowering in u n d e r short-day conditions and induced flowering in this normally short-day plant u n d e r long-day conditions. In other words elevated CO2 appears to reverse the photoperiodic response, converting long-day plants into short-day plants and vice versa. It is true that the CO2 levels used in these studies are many times higher than the levels we will ever see in the atmosphere. It is also true that some of these same studies have demonstrated that more moderate CO2 levels have no effect on the photoperiodic response. However, these physiological studies only expose plants to elevated CO2 for relatively short periods of time (i.e., a few hours to 1-2 weeks). More moderate CO2 levels may have similar effects when plants are exposed to these levels over their entire lifetime. We addressed this question in the following study. We wished to determine whether the variation that has been observed in the response of flowering to moderate increases in CO2 can be attributed
22.
on
339
to differences among species in their photoperiodic response (Reekie 1994). Four short-day species and and four long-day species and were grown at 350 and 1000 ppm under identical experimental conditions except for daylength. Short-day plants were grown in an 9-h photoperiod while long-day plants received a 17-h photoperiod. However, both short- and long-day plants received the same total amount of light. Long-day plants received 9 h of 300/zmol m -2 s1 and 8 h of 8/zmol m 2 s-1. Short-day plants received 1 h of 300/zmol m 2 s-1 and 8 h of 308/zmol m 2 s-1. Elevated CO2 had different effects on flowering phenology in the two groups of species (Fig. 3); it uniformly hastened flowering in the long-day species and delayed flowering in the short-day species. Note that these results do not necessarily mean that short- and long-day plants will always respond to CO2 in this fashion. An experiment with demonstrated that if it was grown for a prolonged period in short days before being switched to an inductive long photoperiod, the difference between CO2 levels disappeared entirely. The effect of CO2 on phenology may depend on the entire sequence of photoperiods received. Previous developmental studies suggest that extremely high levels of CO2 affect flowering by either directly or indirectly interacting with phytochrome, the pigment involved in sensing daylength (Purohit and Tregunna, 1974; Hicklenton and Jolliffe, 1980). The above study suggests this interaction also occurs at more moderate COs levels if plants are exposed to these levels over their entire life. If there is indeed an interaction between moderate levels of CO2 and the effects of phytochrome on development, this interaction has far-reaching implications for understanding the effect of increasing atmospheric COs concentration on plants. Not only would it help us understand and predict how changes in COs may affect flowering, it may also help explain some of the many other effects of elevated COs. Phytochrome is directly involved in controlling a wide range of developmental processes in plants, including flowering, circadian rhythms, stem elongation, leaf morphology, leaf anatomy, seed germination, and branching patterns (Salisbury and Ross, Note that many of these processes have also been shown to be affected by elevated CO2 (Strain and Cure, 1985; Eamus and Jarvis, 1989; Bazzaz, 1990). If elevated COs does interfere with the effects of phytochrome, the highly variable effects of COs on development might be explainable in terms of interactions with other factors known to influence phytochrome. The next study I will describe (Reekie, Hicklenton, and Reekie, unpublished data) explores this possibility. In as in many species, daylength through its interaction with phytochrome has a profound effect on plant morphology affecting height, branching patterns, and specific leaf area. We wished to
340
Figure 3 Effect of atmospheric CO2 on time to flowering of four long- and four short-day species. Open bars represent plants grown at 350 ppm and hatched bars represent plants grown at 1000 ppm CO2. The level of significance for differences between CO2 levels is given above the bars for each species. Based on data from Reekie (1994).
determine whether differences in daylength would affect the response of this species to elevated CO2. In this experiment all plants received 300/zmol m 2 s-1 over an 8-h photoperiod. Long-day conditions were simulated in one-half the plants~ by interrupting the n i g h t with a period of low light:~(8~mol m,? S -1) f o r a p e r i o d of ~h, Similar n i g h t interruption ~eatiments~ have b e e n used,to simutate long,day conditions~in a v , ~ i e ~ oflphotoperiodiC' experiments (Salisbu~: ::and Ross,i! 1991):il T h e leVel o f CO~: in,teracted istrongly ~ t h daylength ~( T a b l e II )i,~ fac~: itihad~0ppoSite effects o n ~bran:c~ng in ~ e ~ o treatments, decreasing:nurnber :~ofib r a n c h e s ~ n short days ~a n d increa~rtg, number~ o f b r ~ c h e s : in long ~days.:;similarly,~ elevated GO~ h a d no' effect ;onl height~or d a y s to: flowering,~ irr,long ::days iwhile lilt
22.
Short days Variable No. of branches Height (cm) Days to flower Biomass (g)
341
on
Long days
Low COz
High CO2
Low CO2
High CO2
12.2 6.9 87 1.81
9.0 8.4 72 3.14
6.6 12.6 65 2.20
7.2 12.6 65 2.92
There were significant (P < 0.05) daylength and CO~ effects and a significant CO2 X daylength interaction for each of the variables examined.
increased height and decreased days to flowering in short days. The CO2 • photoperiod interaction was also carried over into growth measurements. Although elevated CO2 enhanced growth in both treatments, the effect was much greater in short days than in long days. The enhancement ratio (biomass of high-CO2 plants to biomass of l o w - C O 2 plants) was 1.73 for short-day plants compared to 1.32 for long-day plants. This study illustrates one of the problems involved in trying to generalize regarding the effect of CO2 on plants; the effects often vary with environment. However, if CO2 does interact with the effects of phytochrome, this will at least provide us with a framework with which we may be able to understand and eventually predict these contrasting effects. For example, due to changes in daylength with season an interaction between phytochrome and CO2 could help explain seasonal variation in the effect of CO2 on growth. It is commonly observed that the beneficial effects of elevated CO2 decline as the season progresses (e.g., Bazzaz 1989). Although there are probably several explanations for this effect including resource limitation and changes in plant responsiveness with age (McConnaughay 1993), the above experiment suggests that changes in photoperiod with the season could be involved. Similarly, it has been suggested that the responsiveness of plants to CO2 may vary depending on whether they grow in the open or in the shade of other plants (Arnone and K6rner, 1993). Increases in the red:far-red light ratio in shade have a direct effect on phytochrome (Morgan and Smith, 1981) which could in turn help explain changes in CO2 responsiveness.
The above study illustrates that differences in developmental patterns triggered by photoperiod can have marked effects on the ability of plants
to respond to elevated CO2; however, it does not explain why the response to CO2 differs. Although there are probably a number of possible explanations for the greater responsiveness of plants grown in short days, one explanadon suggested by the morphological data is the capacity of the plants to utilize the extra photosynthate which is produced at elevated CO2. Plants grown in short days produced more branches than plants in long days (Table II). Increasing the number of branches affects the number of actively growing meristems, and therefore will increase the size of the sink for the photosynthate produced by the leaves. It has been suggested that artificially manipulating source-to-sink ratio by removing leaves (sources) or meristems (sinks), or by limiting growth of the belowground sink through pot size limitation, can significantly affect the responsiveness of plants to elevated CO2 (e.g., Arp, 1991). The next study I will describe addresses the quesdon of how developmental differences in source-to-sink ratio may affect response to CO2. Within the genus there is a wide range of cultivated species and variedes differing markedly in their developmental patterns and as a consequence, in the size and location of carbon sinks. Broccoli and cauliflower are both annuals that produce a very large inflorescence sink. Marrow stem kale and Chinese broccoli (B. are also annuals, but the major carbon sink is in the stem. Turnip is a biennial with a large root sink in its first year of growth. White mustard and rape are annuals that lack any large carbon sink in the vegetative state. However, once pollination occurs the developing seeds would be significant carbon sink. To examine the influence of these different developmental patterns on response to elevated CO2 seeds of each of these, cultivars were germinated and grown at either 350 or 1000 ppm CO2 (Reekie, MacDougal, and Hicklenton, unpublished data). Plants were harvested at two different stages: (1) as young seedlings before the development of any marked differences in allocation patterns among the culdvars, and (2) at time of flowering in broccoli and cauliflower. At this time there were marked differences in allocation patterns among all the cultivars but white mustard and rape were still vegetative. As young seedlings, all cultivars responded very positively to elevated CO2 (Fig. 4). The long-term growth response of the different cultivars, on the other hand, was independent of whether the sink was located in the root, stem, or flower, but was very dependent on sink size (Fig. 4). Those cultivars with no obvious carbon sink (i.e., white mustard and rape) showed no significant growth enhancement by the end of the experiment. It might be argued that the cultivars chosen for this experiment are only caricatures of real plants in that differences in development have been greatly exaggerated by ardficial selection. However, this experiment does illustrate that
22.
on
343
Effect of atmospheric CO2 on growth of seven cultivars with contrasting patterns of development at two developmental stages (see text for description). An asterisk indicates that CO2 had a significant effect on growth at the 0.05 level. Reekie, MacDougal, and Hicklenton, unpublished data.
differences in developmental patterns can have very marked effects o n C O 2 responsiveness. General reviews of the response of a variety of unrelated species to elevated CO2 also support the idea that differences in developmental patterns are extremely important (e.g., Cure, 1985).
344
In this chapter I have attempted to make several points: (1) elevated C O 2 affects a wide range of developmental processes; (2) these developmental changes have the potential to substantially alter competitive outcome; (3) the effects of elevated CO2 on development are highly variable depending on species and environmental conditions; (4) evidence exists that the effect of CO2 on development is due to an interaction with the effects of phytochrome on developmental processes; (5) the variable effects of CO2 may be related to the influence of other factors on phytochrome such as photoperiod and red:far-red ratio; and (6) differences in developmental patterns have the potential to substantially alter the growth response of plants to elevated CO2. These conclusions have important consequences regarding our ability to predict the effect of elevated CO2 on plant-plant interactions. To date, attempts to predict the effect of elevated CO2 on competitive outcome have focused on the effect of CO2 on growth of the individual species in the community. Because developmental patterns can influence the effect of CO2 on growth, this knowledge could be used to help assign species to functional groups. Due to the large n u m b e r of species in a natural community it will not be possible to obtain the necessary data on the response of individual species to elevated CO2. Instead we must depend on our ability to generalize concerning the response of different functional groups. However, developmental patterns are often very plastic depending on a range of environmental conditions which will make it difficult to make simple generalizations regarding response to elevated CO2. The fact that CO2 itself affects development also suggests that predicting the growth response to CO2 will not be staightforward. Further, these developmental effects have the potential to directly influence competitive outcome. Unfortunately, the effects of COz on development are even more variable than its effects on growth. In the same m a n n e r that we attempt to assign species to functional groups based on the response of their growth to elevated CO2 we must also assign species to functional groups based on the effect of CO2 on development. We have made some progress in this area since it appears that daylength response categories (i.e., short- versus long-day plants) may be important, but because of interactions with other environmental parameters such as photoperiod and red:far-red ratio, predictions are unlikely to be straightforward. In conclusion, the ability to predict the effect of CO2 on growth and development will be crucial to any attempt to model the effect of CO2 on p l a n t - p l a n t interactions. However, much more work is required before accurate predictions will be possible.
22.
on
Arnone, J., and K6rner, Ch. (1993). Influence of elevated C O 2 o n canopy development and red:far red ratios in two-storied stands of 94, 510-515. Arp, W.J. (1991). Effects of source-sink relations on photosynthetic acclimation to elevated CO2. 14, 869-875. Bazzaz, F. A. (1990). The response of natural ecosystems to the rising global CO2 levels. 21, 167-196. Bazzaz, F. A., Garbutt, K., Reekie, E. G., and Williams, W. E. (1989). Using growth analysis to interpret competition between a C3 and C4 annual under ambient and elevated CO2. 79, 223-235. Bolker, B. M., Pacala, S. W., Bazzaz, F. A., Canham, C., and Levin, S. A. (1995). Species diversity and ecosystem response to carbon dioxide fertilization: Conclusions from a temperate forest model. 1, 373-381. Carter, D. R., and Peterson, K. M. (1983). Effects of a CO2 enriched atmosphere on the growth and competitive interaction of a C3 and a C4 grass. 58, 188-193. Cure, J. D. (1985). Carbon dioxide doubling responses: A crop survey. "Direct Effects of Increasing Carbon Dioxide on Vegetation" (B. R. Strain and J. D. Cure, eds.), pp. 99-116. United States Department of Energy, National Technical Information Service. Springfield, VA. Dahlman, R. C. (1985). Modelling needs for predicting responses to CO2 enrichment: Plants, communities and ecosystems. 29, 77-106. Eamus, D., and Jarvis, P. G. (1989). The direct effects of increase in the global atmospheric CO2 concentration on natural and commercial temperate trees and forests. 19, 1-55. Garbutt, K., and Bazzaz, F. A. (1984). The effects of elevated COz on plants. III. Flower, fruit 98, 433-446. and seed production and abortion. Gaudet, C. L., and Keddy, P. A. (1988). A comparative approach to predicting competitive ability from plant traits. 334, 242-243. Hesketh, J. D., and Hellmers, H. (1973). Floral initiation in four plant species growing in CO2 enriched air. 11, 51-53. Hicklenton, P. R., and Jolliffe, P. A. (1980). Carbon dioxide and flowering in Choisy. 66, 13-17. Hunt, R., Hand, D. W., Hannah, M. A., and Neal, A. M. (1991). Response to CO2 enrichment in 27 herbaceous species. 5, 410-421. Marc, J., and Gifford, R. M. (1984). Floral initiation in wheat, sunflower, and sorghum under carbon dioxide enrichment. 62, 9-14. McConnaughay, K. D. M., Bernston, G. M., and Bazzaz, F. A. (1993). Limitations to CO2induced growth enhancement in pot studies. 94, 550-557. Morgan, D. C., and Smith, H. (1981). Non-photosynthetic responses to light quality. "Encyclopedia of Plant Physiology," Vol. 12A, pp. 109-134. Springer-Verlag, Berlin. Paez, A., Hellmers, H., and Strain, B. R. (1980). Carbon dioxide effects on apical dominance in 50, 43-46. Poorter, H. (1993). Interspecific variation in the growth response of plants to an elevated ambient CO~ concentration. 104/105, 77-97. Purohit, A. N., and Tregunna, E. B. (1974). Effects of carbon dioxide on and in short days. 52, 1283-1291. Reekie, E. G., and Bazzaz, F. A. (1989). Competition and pattern of resource use among seedlings of five tropical trees grown at ambient and elevated COz. 79, 212-222. Reekie, E. G., and Bazzaz, F. A. (1991). Phenology and growth in four annual species grown in ambient and elevated CO2. 69, 2475-2481.
346 Reekie, j. Y. c., Hicklenton, P. R., and Reekie, E. G. (1994). Effects of elevated CO2 on time of flowering in four short-day and four long-day species. 72, 533-538. Reynolds,J. F., and Acock, B. (1985). Modeling approaches for evaluating vegetation responses to carbon dioxide concentration. "Direct Effects of Increasing Carbon Dioxide on Vegetation" (B. R. Strain and J. D. Cure, eds.), pp. 33-52. United States Department of Energy, National Technical Information Service. Springfield, VA. Ross, M. A., and Harper,J. L. (1972). Occupation of biological space during seedling establishment. 60, 77-88. Salisbury, F. B., and Ross, C. (1991). "Plant Physiology," 4th Ed. Wadsworth, Belmont, CA. St. Omer, L., and Horvath, S. M. (1983). Elevated carbon dioxide concentrations and wholeplant senescence. 64, 1311-1314. Strain, B. R., and Cure,J. D. (1985). "Direct Effects of Increasing Carbon Dioxide on Vegetation." United States Department of Energy, National Technical Information Service. Springfield, VA.
23 Consequences of Elevated Atmospheric CO2 for Forest Insects
Nearly half of the macroscopic species in the world consist of green plants and the insects that feed on them (Strong 1984). Thus, few other interactions can compare with respect to the sheer number ofpairwise species associations likely to be affected by rising concentrations of atmospheric carbon dioxide. The impact of enriched atmospheric CO2 on interactions between forest trees and insects is of particular importance. Forests cover about one-third of the earth's land surface, conduct approximately two-thirds of global photosynthesis, and provide "room and board" for a large fraction of the world's phytophagous insects. These insects can in turn affect forest productivity, species composition, energy flow, and nutrient cycling (Mattson and Addy, 1975; Schowalter 1986; Veblen 1991). Although the importance of insects to forest dynamics is well established, the roles of particular factors affecting insect population regulation and community interactions remain the subject of debate. Widely accepted generalizations have proven elusive. For example, with respect to insect outbreaks per se, Price (1990) stated that "we still have much to learn, generalities are hard to achieve, and little really definitive understanding of insect herbivore population dynamics is available." Given the complex and interactive nature of the factors influencing insect populations (Fig. 1), and the tremendous diversity of insect and plant species, it is not 347
Copyright 9 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.
I ~tm~ her'c~176I~"
; t
Altered plant: physiology phenology tissue chemistry survival ~ ~ .
Complexof interacting factors influencing insect performance and population dynamics.
surprising that powerful and generally applicable theories have failed to emerge. Consequently, predictions about how changes in the global atmosphere may alter future insect population or community dynamics may be presumptuous indeed.
23.
349
Nevertheless, we now know enough about how particular factors (e.g., host quality) influence insect performance, and how these factors may change under conditions of elevated CO2 concentrations, to hazard some projections. Comprehensive assessment of the major factors affecting insect fitness that are likely to be influenced by CO2 (Fig. 1) is beyond the scope of this chapter. Here I will focus on the direct, indirect, and interactive effects of CO2 on insect performance as mediated specifically by changes in host chemical composition.
Chemical composition is likely the single most important determinant of foliar quality to herbivorous insects (Schultz, 1988; Ehrlich and Murphy, 1988). Both nutritional components (e.g., carbohydrates, protein) and secondary metabolites (e.g., alkaloids, tannins) influence insect performance. Changes in host chemical quality may contribute to the onset of insect outbreaks (Mattson and Haack, 1987; Price 1990). A. Plant Chemistry Carbon-nutrient balance theory (Bryant 1983; Bazzaz 1987; Tuomi 1988) posits that changes in the relative availability of carbon and mineral nutrients alter plant carbohydrate production, which in turn influences accumulation of carbon-based secondary and storage compounds. Because increasing atmospheric CO2 concentrations alter the availability of carbon to plants, shifts in tissue chemical composition are likely to occur (i.e., C : N ratios should increase). Only recently have researchers begun to address the effects of enriched CO2 on tree chemistry. Even this relatively small number of studies, however, reveals patterns in plant responses (Table I). Foliar nitrogen concentrations generally decline (82% of species studied). Soluble sugars show variable responses (decreases to increases) among species, and among sugar types within species. Starch concentrations, however, usually increase (80% of species studied). Responses of these primary metabolites mirror those typically observed with herbaceous plants (Lincoln 1993; also see Owensby Chapter 24). With respect to secondary metabolites, responses vary among chemical classes and plant species (Table I). Simple phenolic compounds (monomeric phenols, flavonoids, phenolic glycosides) have been quantified in only four woody species and responses span the range from decreases to increases. Hydrolyzable tannins have been measured in only two species; gallotannin concentrations were unaffected by CO2 treatment whereas ellagitannin concentrations increased in but decreased in
Species
Nitrogen
nc
Quercus rubra
nc
Carbohydrates
Allelochemicals
Reference b
Sugars: nc Starch: nc
C o n d e n s e d tannin: + Gallotannin: nc Ellagitannin: +
Sugars: + Starch: +
Volatile terpenes: nc Sesquiterpene lactone: nc Coumarin: nc Flavonoid: nc
Sugars: nc Starch: +
C o n d e n s e d tannin: +
4
Sugars: + / n c Starch: +
C o n d e n s e d tannin: +
5
Sugars: +
Simple phenols: nc Flavonoids: n c / + Phenolic glycosides: n c / + C o n d e n s e d tannin: +
4
Sugars: + / Starch: +
C o n d e n s e d tannin: +
Starch: +
C o n d e n s e d tannin: + Volatile oils: nc
7
Sugars: + / n c Starch: nc
C o n d e n s e d tannin: nc Monote r penes ; nc
8
Starch: +
Monoterpenes:-/nc
9
Sugars: nc Starch: +
Phenolic glycosides: + C o n d e n s e d tannin: nc
1
Sugars: nc Starch: +
C o n d e n s e d tannin: nc Gallotannin: nc Ellagitannin: -
1
Sugars: nc
Phenolic glycoisdes: C o n d e n s e d tannin: +
10
1
2,3
Entries limited to studies that measured both primary and secondary metabolites. a _, decrease; nc, no change; +, increase. b (1) Lindroth 1993; (2) Johnson and Lincoln, 1990; (3) Johnson and Lincoln, 1991; (4) B. Traw and F. Bazzaz, personal communication; (5) Lindroth 1995; (6) Lavola and Julkunen-Tiitto, 1994; (7) W. Foley, personal communication; (8) Roth and Lindroth, 1995; (9) Williams 1994; (10) Julkunen-Tiitto 1993.
Condensed tannins exhibit the most consistent pattern of response of all allelochemicals investigated to date, increasing in 70% of species studied. Terpenoid concentrations, surprisingly, have not proven to be responsive to plant CO2 environment.
23.
351
Several inferences, albeit preliminary, may be drawn from these studies. First, the secondary chemistry of woody plants is more responsive to atmospheric CO2 concentrations than is that of herbaceous plants, for which virtually no evidence of allelochemical modification exists. Second, changes in levels of secondary compounds are more likely to occur for "static" metabolites (e.g., condensed tannins) due to accumulation of stable endproducts, than for "dynamic" metabolites (e.g., phenolic glycosides), which may be subject to transport or metabolic turnover (Reichardt 1991). Finally, current evidence suggests that the shikimic acid pathway, leading to production of phenolics, is more responsive to CO2 environment than is the mevalonate pathway, leading to production of terpenoids. B. Altered Biochemistry or Accelerated Development?
Enriched CO2 environments may accelerate plant phenological development, and common developmental changes in foliar chemistry include concomitant decreases in nitrogen levels and increases in tannin levels (Slansky and Scriber, 1985; Waterman and Mole, 1989). Thus the question arises as to whether phytochemical changes are due to CO2-mediated alteration of biochemical processes, or are simply a consequence of accelerated maturation. Because nearly all studies to date have sampled foliage only once during an experiment, this issue cannot yet be resolved. Preliminary data from studies with and suggest that for trees grown under elevated CO2, tannins accumulate in excess of levels acquired simply due to leaf developmental processes (Lindroth 1995; Roth and Lindroth, unpublished data). Moreover, the generally good fit of overall plant C : N ratios to predictions of carbon-nutrient balance theory (Lindroth, 1996; Kinney 1996) suggests that atmospheric CO2 levels effect changes in biosynthetic processes distinct from those of developmental physiology. C. Insect Performance CO2-mediated changes in tree chemistry can alter the performance of phytophagous insects (Table II). Although magnitudes of effects are generally small, in some cases they can be quite striking. Again, several patterns emerge even from the relatively few studies conducted to date. For example, insect consumption increases in most (71%) cases. The magnitude of the feeding response typically exceeds that of all other performance parameters; increases of 100% have been noted in several studies (Lindroth, 1996). Effects on growth are more variable, showing no change in half of the studies and a slight decline (typically 10-15%) (Lindroth, 1996) in nearly half of the studies. How is it that insects can eat more, yet grow less than or equal to insects fed ambient CO2 foliage? Insects fed high CO2 foliage typically exhibit reduced food processing efficiencies. Efficiencies of conver-
Lindroth
Growth Insect species
H o s t species
Consumption
Growth
efficiency
Reference b
nc
-
1
1
Antheraea
1
Lymantria
saccharum
nc
nc
nc
2
nc
-
-
3
nc
+
nc
2
+
nc
nc
3
-
-
nc
2
rubra
2
2 nc
-
2
+
nc
4
+
nc
4
+
nc
tridentata
tridentata
decrease; nc, no change; +, increase. Lindroth et 1994; (2) Lindroth et Lincoln, 1990; (5) Williams and Lincoln, 1994.
-
5
a _,
b (1)
1993; (3) R o t h
and Lindroth, 1994; (4) Johnson and
sion of ingested food into larval biomass generally decline by 20-60% (Lindroth, 1996). Decrements are due more to reductions in the efficiency with which absorbed nutrients are converted to biomass than to decreases in the efficiency with which food is digested. Reductions in growth efficiencies may result from exposure of larvae to higher doses of allelochemicals, due to both higher consumption rates and higher concentrations of such compounds in the food consumed. Only one study has assessed efficiencies of use of a specific nutrient by larvae feeding on foliage from trees grown under enriched CO2. Williams et al. (1994) found that reduced foliar nitrogen concentrations led to reduced nitrogen consumption in pine sawflies. Interestingly, however, nitrogen accumulation rates were similar for larvae in different CO2 treatments because insects fed high CO2 foliage (low N) had improved nitrogen utilization efficiencies. Thus, some insects may not experience negative consequences of feeding on high CO2 foliage due to inherent capacities for adjustments in digestive physiology. Given that tree responses to elevated CO2 vary among species, it is not surprising that insect responses to high CO2 foliage also vary among tree
23.
species. Under natural conditions this may lead to shifts in insect host preferences. The majority of phytophagous insects are specialists, for whom host shifts would present a major evolutionary hurdle. But some of the most serious forest pests, such as the gypsy moth, are generalists. Shifts in relative host preferences among these species are likely to occur.
A. Plant Chemistry Responses of plants to elevated concentrations of atmospheric C O 2 a r e influenced by the availability of other resources (e.g., nutrients, water, light) required for maintenance and growth. Carbon-nutrient balance theory and the more general growth-differentiation balance theory (Herms and Mattson, 1992) suggest that levels of carbon-based secondary or storage compounds will increase under environmental conditions in which growth is inhibited more than photosynthesis. Examples include conditions of low mineral nutrient availability and moderate drought. Thus, the impacts of enriched CO2 on tree chemistry and tree-insect interactions are likely to be modulated by resource availability. Soil fertility (particularly N) limits forest productivity in eastern North America (Pastor and Post, 1988) and interacts with atmospheric CO2 to alter plant growth (Bazzaz, 1990; Field 1992). Thus mineral nutrient availability is also likely to interact with CO2 to affect the foliar chemistry of woody plants. Studies with (Johnson and Lincoln, 1991), (Julkunen-Tiitto 1993) and (Kinney 1996) showed that decrements in leaf nitrogen concentrations caused by enriched CO2 atmospheres are smaller under conditions of low nutrient availability. No such interaction was found in or Q. rubra (Kinney 1996). Similarly, CO2 and nutrient availability interact to influence foliar starch concentrations in (Johnson and Lincoln, 1991) and (Kinney 1996), but not in A. or Q. rubra (Kinney 1996). With respect to secondary metabolites, simple phenolic concentrations in (JulkunenTiitto 1993) and flavonoid and condensed tannin concentrations in (Lavola and Julkunen-Tiitto, 1994) are influenced more strongly by soil nutrient availability under elevated atmospheric CO2 than under ambient CO2. The reverse is true, however, for condensed tannins in (Julkunen-Tiitto 1993). No interaction between CO2 and nutrient availability occurred for levels of tannins in or Q. rubra (Kinney 1996). Soil water availability also limits forest productivity in eastern North America (Pastor and Post, 1988) and also interacts with CO2 to affect plant
growth (Bazzaz, 1990; Field 1992). To date, no published study has addressed the interaction between CO2 and drought stress on plant chemistry. A recent study by Roth and Lindroth (1995; unpublished data), however, did so; we found independent but no interactive effects of CO2 and drought stress on tannin concentrations in and B. Insect Performance
Given that C O 2 and resource availability can interact to alter phytochemical composition of woody plants, performance of phytophagous insects is also likely to be affected. Only two studies, however, have addressed the issue. Johnson and Lincoln (1991) reported that for CO2 and nutrient availability interacted to affect grasshopper growth but not consumption. Both insect growth rate and efficiency were more strongly affected by host nutrient availability under ambient CO2 conditions than under elevated CO2 conditions. Kinney (1996) found that for gypsy moth (Lymantria larvae fed consumption rates were strongly affected by nutrient availability at elevated, but not ambient, CO2 concentrations. CO2 • nutrient interactions did not affect growth rates of insects fed or any performance parameter of insects fed These results show that compensatory feeding responses of insects fed high-CO2 foliage may be modulated by mineral resource availability to host plants, and that such dynamics will differ among host plant species.
Enriched C O 2 atmospheres may influence plant-insect interactions not only with other environmental components, but also by modifying environmental factors that in turn affect the quality and quantity of vegetation (Fig. 1). One of the most important of such factors is climate. The importance of specific weather variables to insect fitness is well established. Moreover, climate has long been accorded a pivotal role in the population dynamics of forest insects, particularly in triggering outbreaks of eruptive species. Whether climatic anomalies actually outbreaks (the "climatic release" hypothesis), however, has been difficult to ascertain because many other factors (e.g., plant host quality) are correlated with weather changes (Martinat, 1987; Price 1990). Indeed, indirect effects of climate, mediated through changes in plant chemistry, may be as, or more, important than direct effects (Martinat, 1987; Mattson and Haack, 1987). Minor changes in climate conditions may have amplified effects when transduced through trophic interactions. For
23.
355
example, a temperature increase of only 1~ tripled the potential population growth rate of geometrid moths feeding on mountain birch (Ayres and MacLean, 1987). Larval performance was influenced directly by temperature (more rapid development) and by temperature-induced acceleration in leaf maturation (decline in food quality). Of particular importance in considerations of CO2-induced changes in weather and trophic interactions is the interaction between temperature and tree phenological development. Many insect species, including about half of forest pest species, have evolved life histories in which the period of maximum nutritional demand is temporally linked to the early season growing period, when developing plant tissue is most available (Ayres, 1993). This tissue is generally characterized by high levels of nitrogen and low levels of fiber and secondary metabolites. Consumption of mature leaf tissue can impose severe constraints on insect performance. Thus, many insect species have a narrow window of opportunity within which to complete the feeding phase of development (Hunter and Lechowicz, 1992; Lindroth, unpublished data). General circulation models predict average air temperatures to increase by 2-4~ under anticipated atmospheric CO2 concentrations (Houghton 1990). Increasing temperatures may affect the developmental race between insects and host trees if the temperature sensitivity of the interacting species differ (Ayres, 1993). If insect development is accelerated more rapidly than leaf maturation, the potential for population outbreaks of eruptive species will increase. Alternatively, if leaf development is accelerated more rapidly than insect development, the temporal window of high food quality will become even narrower and insect fitness will be impaired.
Parasitoids, predators, and pathogens play critical roles in the population dynamics of insects. That plant quality can significantly influence interactions between insects and such natural enemies is now well established. Given that enriched CO2 atmospheres modify plant nutritional quality sufficiently to alter insect consumption, development, and growth rates, the question arises as to whether the performance of parasitoids, predators, and pathogens may change as well. To date, no published studies exist on the subject. Two recent experiments by my research group, however, show that effects on third trophic levels are minor, suggesting that secondary consumers may be buffered from CO2-induced changes in the chemical composition of tree foliage.
356
Roth and Lindroth (1995) evaluated performance of a hymenopteran parasitoid, reared on larvae of the gypsy moth fed foliage from four different species grown under ambient or enriched CO2. Of several performance parameters measured, the only significant effects included a slight increase in mortality and very small reduction in adult female size for parasitoids in high CO2 treatments. COz treatment did not alter the effect of parasitoids on gypsy moth performance (no COz x parasitoid interactions) with the exception that parasitization increased food consumption by gypsy moth larvae, and this increase was greater for insects fed high CO2 foliage than for those fed ambient CO2 foliage. We also addressed the effects of tree CO2 environment on performance of a dipteran parasitoid of the gypsy moth. This parasitoid develops within the gut of its larval host, and its performance is influenced by tannins consumed by the host (Bourchier, 1991). Development times (oviposition to pupation) of parasitoids were prolonged 20%, and puparia weights increased slightly, in larval hosts reared on high CO2 foliage (Roth and Lindroth, unpublished data). Consistent with delayed parasitoid development, the time elapsed from oviposition to gypsy moth death increased in larvae fed elevated COz foliage. Increases in atmospheric CO2 are more likely to affect shifts in the performance of natural enemies whose efficacy is strongly affected by foliar chemical composition. A likely candidate here is the gypsy moth nuclear polyhedrosis virus (NPV). This virus is the most important biological factor contributing to declines of outbreak populations of the moth, and pathogenicity varies with inter- and intraspecific variation in host plant quality (Keating and Yendol, 1988; Hunter and Schultz, 1993). Because viral pathogenicity declines when the virus is ingested with foliage containing high levels of tannins, virulence may decrease under enriched CO2 conditions. We are now investigating that possibility. Atmospheric CO2 concentrations may also affect interactions between insects and natural enemies via changes in climate. Predator and parasitoid attack rates can be altered by temperature and humidity (Martinat, 1987; Mattson and Haack, 1987). Virulence of the disease agents may be strongly affected by weather. For example, fungal pathogens germinate best under conditions of high humidity. Thus a fungal pathogen of the gypsy moth, caused dramatic reductions in gypsy moth populations throughout much of the northeastern United States during the especially wet spring and summer of 1992.
Of primary ecological interest and importance in assessments of the effects of enriched CO2 on plant-insect interactions is how insect popula-
23.
tions will be affected. The factors currently regulating forest insect population densities are so complex and interrelated as to largely preclude accurate predictions of outbreaks by eruptive species. To hazard predictions of outbreaks under environments altered by global change is, therefore, problematical at best. Landsberg and Stafford Smith (1992) presented a scheme for assessing the potential for outbreaks of particular insect species in response to global change. The framework analyzed insect populations "in terms of functional attributes that are both important in population regulation and responsive to global change." These "functional attributes" are primary characteristics of the host plants, insects, or natural enemies that play important roles in population dynamics of pest species. For all its shortcomings (as recognized by Landsberg and Stafford Smith), the scheme is useful in providing a starting point for organization of the many factors likely to influence population outbreaks in a globally changed world. For example, consider the case of the larch bud moth (Zeiraphera in subalpine larch forests. These insects are univoltine, and population densities are largely regulated by delayed negative feedback responses of host trees (reductions in foliar nitrogen, increases in resins and fiber). Lansberg and Stafford Smith (1992) suggested that population outbreaks are unlikely to increase due to global atmospheric changes. "Defensive" responses to herbivory are unlikely to be affected, whereas increased climatic variability will likely reduce population increases and warmer climate may promote regulation by predaceous mites and mirids. Obviously, the accuracy of any derived predictions will be determined by the completeness of understanding of the many factors affecting population dynamics of a particular species. A case in point is the gypsy moth (Table III). Development of both larvae and their food (young leaves) is sensitive to temperature. If insect development is accelerated more rapidly than leaf development the potential for outbreaks increases, but whether this is the case we do not know. Warmer temperatures may reduce mortality due to over-winter cold, but increase mortality due to natural enemies. Consumption of high-CO2 foliage will likely reduce insect growth and fecundity, but may also reduce susceptibility to pathogens and parasitoids. If so many critical gaps exist in our understanding of the population dynamics of a species as well studied as the gypsy moth, we are a long way from making accurate predictions for most insect species.
Elevated atmospheric C O 2 concentrations increase the C : N ratios of tree foliage; levels of nitrogen decline whereas those of carbon-based storage compounds and some secondary metabolites rise. Increases in allelochemi-
358
Plant characteristics Cue for growth Tissue response to CO2 Tissue response to drought Herbivore characteristics Cue for activity Response to warmer temperature Food resource Response to high-CO2foliage
Temperature sensitive Increased C: N ratio Enhanced quality(?) Temperature sensitive Accelerated development, reduced over-winter mortality Young leaves Prolonged development, reduced growth(?), reduced fecundity(?)
Enemy characteristics Response to climate
Pathogens benefitted by increased temperature and humidity Response to hosts on high-CO2 foliage High C:N ratios reduce pathogen and parasitoid efficacy(?)
cal contents occur most consistently with metabolic end-products such as condensed tannins. Chemical responses vary considerably, however, among metabolic pathways (e.g., shikimic acid versus mevalonate) and among tree species. Goals of future research must include assessments of the relative responsiveness of the major pathways of secondary product biosynthesis and identification of patterns in interspecific variation in responses vis fi vis life history traits (e.g., growth rates, leaf longevity). Nothing is currently known about intraspecific genetic variation with respect to chemical responses to CO2. Given that genetic variation is the raw material for evolutionary adaptations, such information is required for understanding long-term ecosystem responses to enriched CO2. Finally, initial research has shown that CO2-mediated changes in tree chemistry are influenced by environmental factors such as resource availability and, possibly, climatic factors. Such variables complicate predictions of community and ecosystem responses to CO2, but must be taken into account. Changes in tree chemical composition resulting from enriched CO2 or C O 2 X environment interactions alter the performance of tree-feeding insects. In general, consumption rates increase, development rates are prolonged, growth rates decline or remain the same, and food processing efficiencies decrease. Responses vary, however, among insects of the same species feeding on different tree species, and among different insect species feeding on the same host species. Research is needed to identify both unique and general insect responses. For example, will various insectfeeding guilds (e.g., foliage-chewers, phloem-feeders, root-feeders) be dif-
23.
359
ferentially affected? If the effects of C O 2 o n plant chemistry accumulate over a growing season, will summer-feeding insects be affected more than spring-feeding insects? A gaping hole in our understanding exists at the population level. For example, will reduced growth and prolonged developm e n t alter mortality and natality rates such that populations decline? From the perspective of tree defoliation, to what degree may population declines offset increased consumption rates of individual insects? Nothing is known about potential intraspecific variation in insect response to CO2-induced changes in host quality. Because phytochemistry has played a pivotal role in the evolution of insect host range and host utilization abilities, further evolutionary adaptations are likely to be elicited by changes in atmospheric CO2 concentrations. Plant chemistry markedly affects the most fundamental of ecosystem dynamics, energy flow and nutrient cycling. Moreover, atmospheric CO2 directly, indirectly, and interactively influences plant chemical composition. Given what we know about chemical responses of individual trees and feeding responses of individual insects to CO2, there is little reason not to expect differences in responses among forest communities as well. Could it be, for instance, that spruce-fir or pine forests (secondary chemistry dominated by products of the mevalonate pathway) may be less chemically responsive than aspen-birch or oak forests (secondary chemistry dominated by products of the shikimic acid pathway) ? Consideration of such questions will require long-term studies at increasingly higher levels of ecological complexity, coupled with development of mechanistic models that span multiple levels of biological organization.
Research funds were provided by NSF Grants BSR-8918586 and DEB-9306981, and USDA Grant 87-CRCR-2581.This work contributes to the Core Research Programme of the Global Change in Terrestrial Environments (GCTE) Core Project of the International Geosphere Biosphere Programme (IGBP).
Ayres, M. P. (1993). Plant defense, herbivory, and climate change. "Biotic Interactions and Global Change" (J. G. Kingsolver, P. M. Kareiva, and R. B. Huey, eds.), pp. 75-94. Sinauer, Sunderland, MA. Ayres, M. P., and MacLean, S. F., Jr. (1987). Development of birch leaves and the growth energetics of (Geometridae). 68, 558-568. Bazzaz, F. A. (1990). The response of natural ecosystemsto the rising global CO2 levels. 21, 167-196.
360 Bazzaz, F. A., Chiariello, N. R., Coley, e. D., and Pitelka, L. F. (1987). Allocating resources to reproduction and defense. 37, 58-67. Bourchier, R. S. (1991 ). Growth and development of (Meigan) (Diptera: Tachinidae) parasitizing gypsy moth larvae feeding on tannin diets. 123,10471055. Bryant, J. e., Chapin, F. S., III, and Klein, D. R. (1983). Carbon/nutrient balance of boreal plants in relation to vertebrate herbivory. 40, 357-368. Ehrlich, P. R., and Murphy, D. D. (1988). Plant chemistry and host range in insect herbivores. 69, 908-909. Field, C. B., Chapin, F. S., III, Matson, P. A., and Mooney, H. A. (1992). Responses of terrestrial ecosystems to the changing atmosphere: A resource-based approach. 23, 201-235. Herms, D. A., and Mattson, W.J. (1992). The dilemma of plants: To grow or defend. Q. Rev. 67, 283-335. Houghton, J. T., Jenkins, G.J., and Ephraums, J.J. (1990). "Climate Change." Cambridge Univ. Press, New York. Hunter, A. F., and Lechowicz, M.J. (1992). Foliage quality changes during canopy development of some northern hardwood trees. 89, 316-323. Hunter, M. D., and Schultz, J. c. (1993). Induced plant defenses breached? Phytochemical induction protects an herbivore from disease. 94, 195-203. Johnson, R. H., and Lincoln, D. E. (1990). Sagebrush and grasshopper responses to atmospheric carbon dioxide concentration. 84, 103-110. Johnson, R. H., and Lincoln, D. E. (1991). Sagebrush carbon allocation patterns and grasshopper nutrition: The influence of CO2 enrichment and soil mineral limitation. 87, 127-134. Julkunen-Tiitto, R., Tahvanainen, J., and Silvola, J. (1993). Increased CO~ and nutrient status changes affect phytomass and the production of plant defensive secondary chemicals in 95, 495-498. Keating, S. T., and Yendol, W. G. (1988). Influence of selected host plants on gypsy moth (Lepidoptera; Lymantriidae) larval mortality caused by a baculovirus. 16, 459-462. Kinney, K. K., Lindroth, R. L.,Jung, S. M., and Nordheim, E. V. (1996). Effects of atmospheric CO~ and soil NOg availability on deciduous trees: phytochemistry and insect performance. in press. Landsberg, J., and Stafford Smith, M. (1992). A functional scheme for predicting the outbreak potential of herbivorous insects under global atmospheric change. 40, 565-577. Lavola, A., andJulkunen-Tiitto, R. (1994). The effect of elevated carbon dioxide and fertiliztion on primary and secodary metabolites in birch, (Roth). 99, 315-321. Lincoln, D. E., Fajer, E. D., and Johnson, R. H. (1993). Plant-insect herbivore interactions in elevated CO2 environments. 8, 64-68. Lindroth, R. L. (1996). CO~-mediated changes in tree chemistry and tree-Lepidoptera interactions. "Carbon Dioxide and Terrestrial Ecosystems" (G. W. Koch and H. A. Mooney, eds.), pp. 105-120. Academic Press, San Diego. Lindroth, R. L., Kinney, K. K., and Platz, C. L. (1993). Responses of deciduous trees to elevated atmospheric CO2: Productivity, phytochemistry, and insect performance. 74, 763-777. Lindroth, R. L., Arteel, G. E., and Kinney, K. K. (1995). Responses of three saturniid species to paper birch grown under enriched CO2 atmospheres. 9, 306-311. Martinat, P.J. (1987). The role of climatic variation and weather in forest insect outbreaks. "Insect Outbreaks" (P. Barbosa and J. C. Schultz, eds.), pp. 241-268. Academic Press, New York.
23.
361
Mattson, W.J., and Addy, N. D. (1975). Phytophagous insects as regulators of forest primary production. 190, 515-522. Mattson, W. J., and Haack, R. A. (1987). The role of drought in outbreaks of plant-eating insects. 37, 110-118. Pastor, J., and Post, W. M. (1988). Response of northern forests to CO2-induced climate change. 334, 55-58. Price, P. W., Cobb, N., Craig, T. P., Fernandes, G. W., Itami, J. K., Mopper, S., and Preszier, R.W. (1990). Insect herbivore population dynamics on trees and shrubs: New approaches relevant to latent and eruptive species and life table development. In "Insect-Plant Interactions" (E. A. Bernays, ed.), Vol. 2, pp. 1-38. CRC Press, Boston. Reichardt, P. B., Chapin, F. S., III, Bryant, J. P., Mattes, B. R., and Clausen, T. P. (1991). Carbon/nutrient balance as a predictor of plant defense in Alaskan balsam poplar: Potential importance of metabolite turnover. 88, 401-406. Roth, S. K., and Lindroth, R. L. (1994). Effects of CO2-mediated changes in paper birch and white pine chemistry on gypsy moth performance. 98, 133-138. Roth, S. K., and Lindroth, R. L. (i995). Elevated atmospheric CO2: Effects on phytochemistry, insect performance, and insect-parasitoid interactions. 1, 173-182. Schowalter, T. D., Hargrove, W. W., and Crossley, D. A., Jr. (1986). Herbivory in forested ecosystems. 31, 177-196. Schultz, J. C. (1988). Many factors influence the evolution of herbivore diets, but plant chemistry is central. 69, 896-897. Slansky, F., Jr., and Scriber, J. M. (1985). Food consumption and utilization. "Comprehensive Insect Physiology, Biochemistry, and Pharmacology," Vol. 4, "Regulation: Digestion, Nutrition, Excretion" (G, A. Kerkut and L. I. Gilbert, eds.), pp. 87-163. Pergamon, New York. Strong, D. R., Jr., Lawton, j. H., and Southwood, R. (1984). "Insects on Plants: Community Patterns and Mechanisms." Blackwell, Oxford. Tuomi, J., Niemal~t, P., Chapin, F, S. i., Bryant, J. P., and Siren, S. (1988). Defensive responses of trees in relation to their carbon/nutrient balance. "Mechanisms of Woody Plant Defenses against Insects. Search for Pattern" (W. J. Mattson, J. Levieux, and C. BernardDagan, eds.), pp. 57-72. Springer-Verlag, New York. Veblen, T. T., Hadley, K. S., Reid, M. S., and Rebertus, A.J. (1991). The response of subalpine 72, 213-231. forests to spruce beetle outbreak in Colorado. Waterman, P. G., and Mole, S. (1989). Extrinsic factors influencing production of secondary metabolites in plants. "Insect-Plant Interactions" (E. A. Bernays, ed.), Vol. 1, pp. 107-134. CRC Press, Boca Raton, FL. Williams, R. S., Lincoln, D. E,, and Thomas, R. B. (1994). Loblolly pine grown under elevated CO2 affects early instar pine sawfly performance. 98, 64-71.
This Page Intentionally Left Blank
24 Effects of Elevated Carbon Dioxide on Forage Quality for Ruminants
Carbon dioxide ( C 0 2 ) levels in our atmosphere have increased by 30% over the past 200 years, with much of that increase coming during the past 100 years (Boden 1990). Expectations are that by the middle of the 21st century atmospheric CO2 will be double current levels. Much conjecture has occurred as to the impact of that increase in atmospheric CO2, primarily centering around climate change, but with a recent concerted effort on the effects of elevated CO2 on terrestrial ecosystem processes. Although there have been several research reports concerning the effects of elevated CO2 on insect diet effects (Butler 1986; Lincoln 1986; Fajer, 1989; Fajer 1989), there has been relatively little work on ruminant responses. Worldwide, natural ecosystems provide the majority of food resources for ruminants (Semple, 1970), with rangelands supplying 95% of the food needs for wild ruminants (Holochek 1989). Therefore, the potential impact of CO2 enrichment on forage quality and dietary conversion efficiencies is of paramount importance. Because nutrient resources in natural systems are fixed to a great degree, the impact of increased carbon fixation with elevated CO2 may lead to dietary deficiencies of essential nutrients for herbivores (Owensby 1993a,b). Although C : N ratios of different plant species have been variable with CO2 enrichment, the prevalent response, when productivity has been increased by 363
Copyright 9 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.
CO2 enrichment, has been an increase in the C:N ratio (Newton, 1991). Lower nitrogen concentration in forages results in reduced digestibility and conversion efficiency of ingested forage to ruminant growth or reproduction (Huston and Pinchak, 1991). Because ruminant digestion is microbial (Hume and Warner, 1980), reduced forage quality would lower both the amount of forage digested and the rate of digestion (Huston and Pinchak, 1991). Ruminants retain forages in the rumen for extended periods which allows efficient nutrient extraction (Bell, 1971). Based on reported changes in leaf chemistry and insect responses to forages grown under elevated CO2, we present the following hypotheses: 9 Reduced nitrogen concentration and increased fiber in forages exposed to elevated CO2 will reduce ruminant intake and assimilation. 9 Growth and reproduction of ruminants will be reduced under elevated CO2. 9 Wild ruminants will be more affected by reduced forage quality than domestic livestock, because dietary supplementation can alleviate some of the deficiencies for domestic livestock. Next, we review ruminant digestion processes, the impact of forage quality on forage intake, digestibility, animal productivity, the impact of elevated CO2 on forage quality, and present the results of research on tallgrass prairie.
Ruminants (cattle, sheep, goats, etc.) have evolved a capacious pregastric fermentation structure with four compartments (Fig. 1) where a symbiotic relationship exists with microbes that have an ability to break down complex
A schematic of the four-compartmented stomach of the ruminant.
24.
365
structural polysaccharides (cellulose and hemicellulose) to compounds that can be absorbed by the animal. That adaptation allows for use of forages produced on the vast rangeland acreage throughout the world. Half the meat and almost all of the milk are produced by ruminants. The fermentation of forage occurs primarily in the reticulo-rumen portion of the gastrointestinal tract of ruminants. From the reticulo-rumen part of the stomach, food passes into the omasum, whose function is not fully understood, but some absorption of fermentation products occurs there, and then into the abomasum which is similar to the simple stomach of monogastric animals. The r u m e n develops rapidly and at maturity represents 80% of the stomach capacity (Campbell and Lasley, 1969). The entire stomach occupies ~ of the abdominal cavity (Church, 1969). Ruminants only chew the forage enough to mix it with saliva and to form a bolus. Further chewing occurs when the animal regurgitates the bolus during rumination. In order to reduce the particle size of the ingesta, ruminants regurgitate it, swallow regurgitated fluids, remasticate the solids accompanied by reinsalivation, and reswallow the bolus. Rumination has an impact on the digestibility of forages, because rumination increases digestibility by reducing particle size which, in turn, affects average time feed remains in the rumen. Ruminant digestion is a complex interaction among the diet, the microbial population, and the animal. Within the rumen, flow characteristics can be partitioned into a liquid phase, particles large enough to pass from the r u m e n and particles that are trapped in the r u m e n because they are too large to pass through the omasal orifice. The ingesta is in fairly large pieces and floats on top of the r u m e n fluid. Those larger pieces are involved in rumination and are broken down into smaller pieces. After exposure to the microbial population and rumination which reduce particle size, the ingesta passes into the remainder of the digestive tract. The r u m e n microbes break down almost all soluble carbohydrates in the forage and degrade 40-80% of the ingested protein. The products of r u m e n fermentation are volatile fatty acids (acetic, propionic, valeric, and butyric), lactic acid, carbon dioxide, methane, microbial protein, and microbial polysaccharides. The volatile fatty acids are absorbed through the r u m e n wall into the blood stream for use in metabolism. Lactic acid, microbial protein, and microbial polysaccharides are passed into the remainder of the digestive tract where further digestion and absorption occurs. Methane and carbon dioxide are expelled orally by a process called The effect of the ruminant digestive system is to hold the structural components of the ingesta in the r u m e n for a relatively long period. Bell (1971) stated that the r u m i n a n t digestion strategy is to maximize efficiency of the use of protein at the expense of the superabundant energy supply. When cell wall constituents (complex structural polysaccharides) are high in the diet, the ingesta is held longer in the r u m e n to ensure protein
extraction. Retention time in the rumen for cattle for most forages is 50-60 hours. Herbaceous dicots generally have lower retention times than grasses. The rate of passage of ingesta through the gastrointestinal tract increases with low fiber, low lignin diets. Intake of roughages by ruminants is regulated by rumen capacity. The more quickly the rumen empties the greater the intake will be. Digestibility of forages determines rate of passage. The more digestible a forage is (low cell wall content, low fiber and lignin content), the greater the rate of passage, with the result being increased intake. The very nature of ruminant digestion makes it impossible for the ruminant to compensate for a lower quality diet by consuming more forage. Another consequence of global climate change may be an increased air temperature. Higher daytime air temperature also will reduce forage intake. Dwyer (1961) reported a negative linear relationship between grazing time and average daytime temperature and indicated that there was not an increase in nighttime grazing to compensate for reduced daytime grazing. Nutrient deficiencies, such as low protein in the forage, will reduce digestibility and therefore rate of passage of forages through the rumen due to lowered microbial activity. Microbes need high amounts of protein to sustain fermentation. The major impact of lowered forage quality is reduced intake and consequently lowered productivity. Secondary productivities (rates of energy storage at different trophic levels) in an ecosystem are described by the following diagram: Ingestion = Assimilation + Egestion
Respiration + Productivity
Growth + Reproduction
In ruminants, as the nitrogen concentration of the diet is reduced and fiber component concentrations are increased, the amount of the ingesta assimilated is reduced and egesta is increased. That material is fed directly into the decomposer food chain. Any ecosystem perturbation that reduces forage quality will reduce the assimilation component from which productivity as growth a n d / o r reproduction is derived.
Plants exposed to elevated C O 2 have consistently had reduced tissue N concentration (Bazzaz, 1990; Newton, 1991). The reduced N concentration has developed very early in vegetative growth and has persisted to maturity,
24.
367
particularly with Ca species. Conventional wisdom implicated nitrogen dilution with increased carbon acquisition (Field 1992), thereby increasing nitrogen use efficiency (NUE). Numerous studies have shown reductions in N concentration for C3 and C4 species under elevated CO2 across a wide range of nitrogen availabilities (Hocking and Meyer, 1985; Larigauderie 1988; Coleman 1991). However, reduced N concentration has been reported in natural ecosystems for species with no increased biomass production under elevated CO2 (Curtis 1989; Owensby 1993b). The cause may be due to reductions in enzymes associated with photosynthesis (Bowes, 1991; Long, 1991; van Oosteen 1992; Wong, 1979). Sage (1987) estimated that RuBP and PEPC carboxylases in a C4 and RuBP in a Ca represented as much as 25% of the total leaf N, which could explain decreased N concentration in biomass under elevated CO2. Another possible explanation for a portion of the reduced N concentration of leaves exposed to elevated CO2 may lie in reduced chlorophyll content. Cave ( 1981 ) found reduced chlorophyll in leaves with CO2 enrichment. Further reductions in forage quality may come from morphologic changes associated with elevated CO2. Thomas and Harvey (1983) reported that leaves of plants under elevated CO2 can have more waxes and extra layers of epidermal cells which may further reduce forage quality for ruminants. Cuticle development appears to be a major deterrent to ruminal digestion. Brazle (1979) and Cummins and Dobson (1972) reported that forage cuticle reduced microbial degradation of ingested forages.
We collected diet samples from ambient and elevated C O 2 (2• ambient) plots using three esophageally fistulated sheep at 2-week intervals throughout the growing season in 1989 [see Owensby (1993a) for CO2enrichment procedures] by grazing one-half of 4.5-m-diameter plots. The diet samples were frozen immediately, freeze-dried, and acid detergent fiber (ADF) (Van Soest, 1967) and N concentration (Linder and Harley, 1942; Technicon Industrial Systems, 1977) determined. Forage quality tests individually indicate relative nutritive value among treatments, but fail to integrate the impact of a treatment across different measured parameters. For example, a reduction in nitrogen concentration or an increased fiber concentration indicate a reduction in forage quality, but they do not indicate the impact on the efficiency of utilization of the energy in the diet or the impact on ruminant production. We used the
368
ADF and N values from the ambient and elevated C O 2 diet samples to estimate the growth response of yearling steers grazing tallgrass prairie. Chemical composition affects the energetic value of plant materials when used as feeds for livestock (Blaxter, 1962). Therefore, we estimated the magnitude of the impact on livestock gain that could result from the changes in chemical composition observed in response to enhanced CO2 concentration. Weight gain projections were calculated for beef cattle that were assumed to be between 12 and 24 months of age and experiencing relatively rapid growth while grazing tallgrass prairie during the late spring and summer periods. The initial step in the simulation process involved estimation of organic matter digestion (OMD) from the ADF concentration (Minson, 1982). Accuracy of OMD predictions using this approach has been quite good when compared with OMD values determined directly in cattle consuming tallgrass prairie forage (Sunvold and Cochran, 1991). Subsequently, digestible energy (DE) concentration was estimated from the OMD concentration and then the metabolizable energy (ME) concentration was predicted from the DE concentration (Minson, 1990). The efficiency of ME use for maintenance (K,,) and the efficiency of ME use for production (i.e., gain; K1+p) were estimated using the associated protein values and ME concentrations (expressed as a percentage of gross energy) as described by Blaxter (1989). Metabolizable energy values and their associated efficiencies were then used to estimate the gain that might be realized by a rapidly growing, yearling steer consuming a given amount of forage. Although weight would obviously be changing over the course of the grazing period, to simplify the simulations an average weight of 250 kg was used in all calculations and forage intake was assumed to be the same for cattle consuming forage from CO2-enriched and ambient CO2 environments (intake would likely be lower for cattle consuming forage produced in an elevated CO2 environment). Amount of forage intake was assumed to decrease as season progressed based on measurements from previous studies at our location. The amount of net energy needed for maintenance was estimated by dividing the K~ into the sum of an estimate of fasting heat production (FHP) and activity (Agricultural Research Council, 1980). Once maintenance was accounted for, we estimated the amount of gain that could be supported from the remaining ME. This was determined by multiplying the ME available for gain by the KI+p. Finally, the calorific values of weight gain presented by Blaxter (1962) were used to convert megacalories of gross energy in the gain into weight gain values. Estimated gain for steers consuming forage produced under elevated CO2 in 1989 was lower than that produced under ambient CO2 summed over the 150-day growth period (2X CO2, 80.6 kg; 1 x CO2, 99.6 kg), with the greatest reduction in gain coming in the early season (Fig. 2).
24.
0.2--
Estimatedsteer gain (kg/day) derived from acid detergent fiber and crude protein of diet samples collected on the indicated dates in 1989 by esophageallyfistulated sheep from tallgrass prairie exposed to 2 • ambient and ambient atmospheric CO2. Means within a date with a common letter do not differ (LSD, P < 0.10).
Since N and fiber concentrations in the diet of ruminants impact forage digestibility and utilization efficiency, the reported reduced N and increased fiber concentrations in plants grown u n d e r elevated CO2 will likely impact r u m i n a n t productivity negatively. Data reporting reduced productivity or increased consumption for insects consuming diets of plants grown u n d e r elevated CO2 support that conclusion. Contrary to the results from insect studies, where intake increased as diet quality decreased, r u m i n a n t intake declines as forage quality decreases. Therefore, there cannot be a compensatory intake response to maintain productivity levels comparable to current levels. For domestic livestock, diets can be supplemented to compensate for reduced forage quality, but with wild ruminants, or for ruminants in developing countries, diet supplementation is not an option. The result will be reduced growth and reproduction. Further, changes in climate may impact foraging by ruminants. High daytime air temperatures currently reduce total grazing time for cattle with little or no compensatory nighttime grazing. A future high-CO2 world seems destined to reduce individual animal performance. For domestic livestock enterprises, increased stocking rates can occur because of the reduced intake of lower quality forage, and dietary supplementation may be used to maintain current production levels, but that will increase cost of production. Wild r u m i n a n t diet quality will be affected, and it is likely that they will have reduced growth and reproduction.
Agricultural Research Council (1980). "The Nutrient Requirements of Ruminant Livestock." C. A. B. International, Wallingford, UK. Bazzaz, F. A. (1990). The response of natural ecosystems to the rising global CO2 levels. 21, 167-196. Bell, R. H. V. (1971). A grazing ecosystem in the Serengeti. 225, 86-93. Blaxter, K. L. (1962). "The Energy Metabolism of Ruminants." Hutchinson, London. Blaxter, K. L. (1989). "Energy Metabolism in Animals and Man." Cambridge Univ. Press, Cambridge, UK. Boden, T. A., Kanciruk, P., and Farrell, M. P. (1990). "Trends '90: A Compendium of Data on Global Climate Change." Carbon Dioxide Information Analysis Center, Oak Ridge National Laboratory, Oak Ridge, TN. Bowes, G. (1991 ). Growth at elevated CO2: Photosynthetic responses mediated through rubisco. 14, 795-806. Brazle, F. K., Harbers, L. H., and Owensby, C. E. (1979). Structural inhibitors of big and little bluestem digestion observed by scanning electron microscopy. 48, 1457-1463. Butler, G. D., Kimball, B. A., and Murray, J. R. (1986). Populations of (Homoptera: Aleyrodae) on cotton grown in open-top chambers enriched with CO2. 15, 61-63. Campbell, J. R., and Lasley, J. F. (1969). "The Sciences of Animals That Serve Mankind." McGraw-Hill, New York. Cave, G., Tolley, L. C., and Strain, B. R. (1981). Effect of carbon dioxide enrichment on chlorophyll content, starch content, and starch grain structure in subterraneum leaves. 51, 171-174. Church, D. C. (1969). "Digestive Physiology and Nutrition of Ruminants," Vol. I. Oregon State Univ. Press, Corvallis. Coleman, J. S., Rochefort, L., Bazzaz, F. A., and Woodward, F. I. (1991). Atmospheric CO2, plant nitrogen status and the susceptibility of plants to an acute increase in temperature. 14, 667-674. Cummins, D. G., and Dobson, J. W., Jr. (1972). Digestibility of bloom and bloomless sorghum leaves as determined by a modified technique. 64, 682-683. Curtis, P. S., Drake, B. G., and Whigham, D. F. (1989). Nitrogen and carbon dynamics in C3 and C4 estuarine marsh plants grown under elevated CO2 78, 297-301. Dwyer, D. D. (1961). "Activities and Grazing Preferences of Cows with Calves in Northern Osage County, Oklahoma." Okla. Agr. Expt. Sta. Bull. B-558. Fajer, E. D. (1989). The effects of enriched COz atmospheres on plant-insect herbivore interactions: Growth responses of larvae of the specialist Nymphalidae). 81, 514-520. Fajer, E. D., Bowers, M. D., Bazzaz, F. A. (1989). The effects of enriched carbon dioxide atmospheres on plant-insect herbivore interactions. 243, 1198-1200. Field, B. F., Chapin, F. S., III, Matson, P. A., and Mooney, H. A. (1992). Responses of terrestrial ecosystems to the changing atmosphere: A resource-based approach. 23, 201-235. Hocking, P. J., and Meyer, C. P. (1985). Responses of noogoora burr Bertol.) to nitrogen supply and carbon dioxide enrichment. 55, 835-844. Holochek, J. L., Peiper, R. D., and Herbel, C. H. (1989). "Range Management: Principles and Practices." Prentice-Hall, Englewood Cliffs, NJ. Hume, I. D., and Warner, A. C. I. (1980). Evolution of microbial digestion in mammals. "Digestive Physiology and Metabolism in Ruminants" (Y. Ruckebusch and P. Thivend, eds.), pp. 665-684. AVI, Westpost, CT.
24.
371
Huston, J. E., and Pinchak, W. E. (1991). Range animal nutrition. "Grazing Management: An Ecological Perspective" (R. K. Heitschmidt and J. W. Stuth, eds.), pp. 27-64. Timber Press, Portland, OR. Larigauderie, A., Hilbert, D. W., and Oechel, W. C. (1988). Effect of CO2 enrichment and nitrogen availability on resource acquisition and resource allocation in a grass, 77, 544-549. Lincoln, D. E., Couvet, D., and Sionit, N. (1986). Response of an insect herbivore to host plants grown in carbon dioxide enriched atmospheres. 69, 556-560. Linder, R. C., and Harley, C. P. (1942). A rapid method for the determination of nitrogen in plant tissue. 96, 565-566. Long, S. P. (1991). Modification of the response of photosynthetic productivity to rising temperature by atmospheric CO2 concentrations: Has its importance been underestimated? 14, 729-739. Minson, D.J. (1982). Effect of chemical composition on feed digestibility and metabolizable energy. 52, 592-612. Minson, D.J. (1990). "Forage in Ruminant Nutrition," pp. 93-95. Academic Press, San Diego. Newton, P. C. D. (1991). Direct effects of increasing carbon dioxide on pasture plants and communities. N. Z.J. 34, 1-24. Owensby, C. E., Coyne, P. I., Ham, J. M., Auen, L. M., and Knapp, A. K. (1993a). Biomass production in a tallgrass prairie ecosystem exposed to ambient and elevated levels of CO2. 3, 644-653. Owensby, C. E., Coyne, P. I., and Auen, L. M. (1993b). Nitrogen and phosphorus dynamics of a tallgrass prairie ecosystem exposed to elevated carbon dioxide. 16, 843-850 Sage, R. F., Pearcy, R. W., and Seemann, J. R. (1987). The nitrogen use efficiency of C~ and C4 plants. 85, 355-359. Semple, A. T. (1970). "Grassland Improvement." CRC Press, Cleveland. Sunvold, G. D., and Cochran, R. C. (1991). Technical note: Evaluation of acid detergent lignin, alkaline peroxide lignin, acid insoluble ash, and indigestible acid detergent fiber as internal markers for prediction of alfalfa, bromegrass, and prairie hay digestibility by beef steers. J. 69, 4951-4955. Technicon Industrial Systems (1977). Individual/simultaneous determination of nitrogen a n d / o r phosphorus in BD acid digests. Thomas, J. F., and Harvey, C. N. (1983). Leaf anatomy of four species grown under continuous CO2 enrichment. 144, 303-309. van Oosten, J. J., Afif, D., and Dizengremel, P. (1992). Long-term effects of a COz-enriched atmosphere on enzymes of the primary carbon metabolism of spruce tree. 30(5), 541-547. Van Soest, P.J. (1967). Development of a comprehensive system of feed analyses and its application to forages. 26, 119-128. Wong, S. C. (1979). Elevated atmospheric partial pressure of COzand plant growth. I. Interactions of nitrogen nutrition and photosynthetic capacity in C3 and C4 plants. 44, 68-74.
This Page Intentionally Left Blank
IV Theory, Modeling, Concepts
This Page Intentionally Left Blank
25 Interspecific Variation in the Growth Response of Plants to Elevated CO : A Search for Functional Types
The current increase in the atmospheric C O 2 concentration poses upon us the challenge to predict how the growth and functioning of plants will be affected. Given the wide variety of species in nature, and the very different conditions u n d e r which they may grow, such an understanding is difficult to achieve by focusing on the response of just a few particular species. Rather, some kind of generalizations across species have to be made, with which we can structure our insight into the patterns and processes by which plants respond to a change in climate. One such approach is to analyze whether species that show c o m m o n physiological a n d / o r morphological traits, or species that are phylogenetically related, differ in response compared to other, less related groups of species (cf. K6rner, 1993a). It is our belief that such a classification of organisms in a limited n u m b e r of groups of functionally similar species, so-called functional types, could provide a framework to further the understanding of the CO2 fertilization effect. Therefore, the main objective of this chapter is to quantitatively analyze whether such functional types with regard to CO2 enrichment can be discerned, and how well different groups can be separated. Two other aspects will receive attention in this chapter. First, it is imperative to arrive at an understanding of how the growth e n h a n c e m e n t is and
375
Copyright 9 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.
376
achieved. Logically, the process of photosynthesis has attracted a lot of attention (e.g., Bowes, 1991; Sage, 1994). However, although the increase in photosynthesis forms the most important primary response of plants to high CO2 concentrations, the actual growth enhancement is determined by an array of secondary changes as well, like those in respiration, allocation, morphology, and chemical composition (cf. Poorter 1992). To our knowledge, up to now only one full analysis has been made of the carbon budget of high-CO2-grown plants (Den Hertog 1993). It is difficult to infer any generalization from this one experiment. As a second-best approach to gain insight into the integrated response of plants one can use a detailed growth analysis (cf. Lambers and Poorter, 1992). We will quantitatively analyze how growth parameters such as net assimilation rate (NAR) and leaf area ratio (IJkR) are affected by elevated CO2. The second aspect concerns the growth conditions. The analysis of the growth response of individually grown plants will be restricted to plants that have experienced conditions which impose relatively little limitation to growth, particularly with respect to nutrients and water. Especially in cases where one would like to extrapolate possible differences between functional groups of species from controlled to natural conditions, it is imperative to know how the growth response is modulated by limiting factors. Given the predominance of nutrient limitation in the field, we will concentrate on how nutrient availability alters the growth response to high CO2.
To arrive at a framework of different functional types, data of a large number of species grown in a variety of conditions are required. Although some larger scale comparative experiments have been carried out (e.g., Hunt 1991, 1993; Campbell 1993; Poorter 1993), none of these experiments is large enough to enable an extended analysis, generally due to a limited subset of species chosen. Therefore, the large amount of information required for such an analysis has never been collected in one, structured experiment. An alternative approach has been adopted by Poorter (1993), in which the literature up to 1991 was compiled and analyzed a posteriori for differences in response to elevated CO2 between groups of species. Since that time many new experiments have been published. In this chapter we update the review of Poorter (1993), arriving at a data set of 500 observations on a total of 250 species, published in 180 articles. We calculate the average response of some groups of species and compare the emerging patterns with research papers in which the response of various groups of species has been investigated explicitly.
25.
377
A. Protocol The protocol followed is similar to that of Poorter (1993). Briefly, only those literature data were analyzed which met certain criteria. Experiments were considered where plants were grown individually, in the absence of competition and where both shoots and roots were included in the analysis, for plants that were not reported to be flowering or fruiting. For each species in each report, the ratio of the weight at high CO2 (between 60 and 80 Pa) and control CO2 levels (between 30 and 40 Pa) was calculated. This ratio, which we will call weight ratio throughout the rest of this chapter, is log-normally distributed, so we In-transformed the data before calculating averages and applying the appropriate statistical tests. However, for the convenience of the reader, we back-transformed averages and ranges for groups of species or observations, before including them in figures and tables. To safeguard against outlying data, which could be caused by various reasons (ranging from unobserved problems during the growth of the plants up to our processing of the data), we removed the 5% largest and the 5% smallest observations of each category of interest before analyzing the data statistically. In this way we obtain better estimates of the main core of the data set, unaffected by a small a m o u n t of outlying measurements (Barnett and Lewis, 1978).
B. Methodological Aspects of Weight Ratios Although weight ratios are a simple and appropriate way to scale the response of plants to high CO2, some biological and statistical aspects deserve attention. First, weight ratios are time-dependent. Typically for C~ herbs, they increase on start of the CO2 treatment, reach the highest level after 10-20 days, and may decrease slightly or more strongly thereafter, depending among others on the process of flowering and fruit set (for references see Poorter, 1993). In our screening we use data for plants that generally did realize their first response, but which did not yet enter the reproductive phase. However, one can never be sure that responses differ between experiments just because of differences in the duration of those experiments. A partial solution to this problem would be to express the growth stimulation as the increase in average RGR over the investigated period of time. However, the vast majority of the experiments do not include a proper growth analysis. Alternatively, it is possible from the weight ratio at a given time and the duration of the experiment, to calculate what the difference in average RGR between treatments has been (see Appendix 1). In this way an expression for the growth stimulation is obtained, which is corrected for the duration of the experiment, assuming that the difference between treatments develops steadily over time. In the present analysis, we calculated both the weight ratio and the RGR stimulation, and considered
378 differences between groups of species only to be consistent, if they are significantly affected for both parameters. Second, up to now, it may not have been fully appreciated to what extent variability between plants, as well as the total n u m b e r of plants harvested per treatment, affects the weight ratio. Generally, plant weights in a population of individually grown plants show a log-normal distribution. Variability then can best be expressed as standard deviation in the In-transformed individual dry weights (SDlnw). Typically, in populations of plants used in a growth analysis, this standard deviation ranges from small (<0.15) to very large (>0.6) (H. Poorter and E. Garnier, unpublished results). The larger the variability, the less precise the weight estimate. As both the estimate of average dry weight of the high-CO2- and the control-COz-grown plants are subject to imprecision, the ratio of these two values will show even wider confidence limits. The second parameter that determines the actual precision of the weight ratio estimate (that is, how closely the estimated weight ratio approximates the true weight ratio) is the n u m b e r of plants that is harvested per treatment. What will the variability in weight ratio be in the average experiment? Assuming (a) that the true weight ratio equals 1.45, a value that is close to the average we observed for the total group of C~ species, (b) that the variability in plant weight is the median value observed in the unpublished compilation of H. Poorter and E. Garnier (0.3), and (c) that the n u m b e r of plants harvested per CO2 treatment is the median of the experiments compiled here (8), what will be the distribution curve of the observed weight ratios? With help of a simple simulation model the distribution curve was found to be that in Fig. 1A. As expected, the top of the distribution is at the value of the true weight ratio. More interestingly, it can be seen that there is large variation around this value, with weight ratios that in extreme cases were found to be less than 1 or more than 2. No objective criteria exist by which we can define which weight ratios are acceptable and which are not. However, given a true weight ratio of 1.45, we feel that, from a biological perspective, the observed ratio should not be less than 1.25 or higher than 1.65. Otherwise, the estimate will be too imprecise to compare the response of different species, because differences between species will be relatively small anyway. In 33% of the simulated experiments this goal was not reached, and values deviating more than 0.2 were found. It is rather difficult to reduce the uncertainty around the estimated weight ratios. Figure 1B shows what fraction of the experiments can be categorized as "unsatisfactory," based on the criterion described above. In cases with low variability in the plant population of interest (SDlnw < 0.2) and eight or more replicates per harvest the chance to arrive at an "unsatisfactory" value is modest (Fig. 1B). However, at high variability (SDlnw >
25.
379
(A) Distribution of the observed weight ratios, when harvests of randomly drawn plants are simulated from a high CO2 and control CO2 population. The true weight ratio of the two populations was set to 1.45, the SDInwof both populations to 0.3, and the sample size in both cases to 8. (B) The chance that the observed weight ratio deviates more than 0.2 units from the true weight ratio, as depending on the SDlnwof the plant populations. The relationship is given for three different sample sizes. The three broken vertical lines indicate the 25th, the 50th, and the 75th percentile of the SOlnw values as compiled by H. Poorter and E. Gamier (unpublished). (The xth percentile indicates that x% of the observations in the data set are lower than that value.)
0.45) the risk to arrive at aberrant estimates is high, even when large numbers of plants are harvested per treatment. Variability in response ratios within a single species has been noticed more often (e.g., Potvin and Strain, 1985; Hunt et al., 1993; Poorter, 1993). Although biological reasons could contribute to this variability (different
380 experimental conditions, genotypic variation, developmental stage at harvest), we think that the simple explanation given above may, unfortunately, explain such variation to a large extent. The first conclusion we infer from this simulation is that most of the weight ratios, as listed in Appendix 2, are probably too imprecise to infer conclusions about differences in response between species or treatments. Rather, it is a combination of a large n u m b e r of experiments that will yield a more precise estimate of groups of species. Secondly we conclude that, due to this large variability, we will never be able to explain more than a modest fraction of the total variation by categorizing species into functional groups. Thirdly, we r e c o m m e n d that in cases where the final weight ratio is of interest, at least 10 plants per treatment are harvested, and in populations with high variability even more. Concurrently, it may be worthwhile to consider controlling the variability in the experimental population in one way or another (cf. Poorter, 1989).
C. Limitations to the Approach There are two complications in a literature analysis like this. First, the articles published may show a certain bias compared to the results obtained. For example, if weight ratios below 1.0 have been found for Cs species, or values clearly above 1.0 for C4 species, investigators may have considered this abnormal and refrained from publication. Second, there is a danger in compiling a wide variety of experiments, in that the functional grouping may be confounded with a major difference in growth conditions between species. This would bias the results and could invalidate the conclusions to a certain extent. However, the large n u m b e r of studies involved (approx. 185, with a total of 500 observations) makes large systematic differences in growth conditions across groups less likely and therefore reduces the methodological problems mentioned above. An important advantage of the combination of such a large a m o u n t of studies is that, due to the variation in conditions used, the results provide a better generalization of the expected effects of high CO2 than a single large experiment carried out u n d e r one condition only. However, as the approach we follow bears the risk of overinterpretation, the general trends observed in this metaanalysis will be compared with results obtained in specific experiments, as far as they are available. \
A. Between Species with Different Photosynthetic Pathways All the experiments listed in Poorter (1993) and those of the CO2 literature which have appeared subsequently have been compiled in Appendix 2. Mean values for the weight ratios of the functional groups of C~, C4,
25.
and CAM species, as well as parameters that characterize their frequency distribution, are shown in Fig. 2A. There is a large distinction between species with different photosynthetic pathways. In terms of net biomass accumulation, C3 species respond stronger (47%) than C4 (10%) and CAM species (19%). The lower response of the latter becomes evident when the RGR differences are considered (Fig. 2B). Responses of CAM plants are
Figure 2 (A) Box plots of the observed weight ratios of all the C3 (n = 226), C4 (n = 24) and CAM (n = 6) species listed in Appendix 2, at the left side, and of the categories Crop species (n = 23), Wild herbaceous species (n = 108) and Woody species (n = 95), at the right side. (B) Idem for the absolute RGR stimulation of the same categories. The middle line in each box indicates the geometric mean of the observed distribution, the bottom and top parts of the box the 25th and 75th percentiles, and the bottom and top " e r r o r bars" the 5th and 95th percentiles of the observed distribution. Signs at the top side of the graphs indicate whether differences between means of adjacent groups are significant (ns, nonsignificant; +, 0.05 < P < 0.10; *P < 0.05; **P < 0.01; ***P < 0.001).
382
analyzed after much longer durations of growth. No time trends for the growth stimulation of CAM species are available, so it is not clear whether the stimulation is transient, or steady but small. It seems to be c o m m o n knowledge that C4 plants do not show a growth response to high atmospheric CO2 concentrations. From the compiled data we derive that there is a large variability in the weight ratio, with values in specific experiments ranging from 0.8 to 1.5. This might be caused by the same chance effect as the distribution in Fig. 1A. However, there are two reasons to think of a more biological explanation. First, although the effect on C4 plants is smaller than for C~ plants, the average weight ratio for the C4 plants differs significantly from 1.0 (P < 0.05). Furthermore, there seems to be systematic variation between species. The two most frequently measured C4 species, and differ significantly and almost systematically from each other, with showing a stronger response (1.26) than Zea (1.09). Several explanations have been put forward to account for this difference in behavior. Poorter (1993) suggested that, apart from secondary responses, a small but consistent increase in the rate of photosynthesis may play a (differential) role in the response of C4 plants. T r e m m e l and Patterson (1994) mentioned that the first leaves of some C4 species show the normal C3 type of photosynthesis, and that this may cause such species to be responsive to high CO2, at least in the short term. The latter explanation fits with the fact that for they found a strong time dependency in the weight ratio, with young plants responding more strongly. More insight into the long-term growth response of plants can be achieved by the technique of growth analysis. I n this approach, relative growth rate is factorized into two components: net assimilation rate (NAR, the increase in biomass per unit of leaf area and time) and leaf area ratio (I2kR, leaf area : plant weight). As explained in Appendix 1, it is the absolute difference in RGR that is of interest when comparing variation in the CO2 response of species. However, if one would like to analyze which of the growth parameters is responsible for the increase in RGR, it is more appropriate to express RGR and its components in a relative way, normalizing the values at high CO2 with respect to their value at control levels. Compiling growth analysis experiments of-~60 C~ species, a modest increase in RGR is found of on average 7% (Fig. 3A). The increase in RGR is correlated with a 22% increase in NAR, due to a large extent to an increase in photosynthesis. However, this increase is counterbalanced by a decrease in LAR (on average 11%). The LAR can be further factorized into the specific leaf area (SLA, leaf area: leaf weight) and the leaf weight ratio (LWR, leaf weight:total plant weight). As can be seen in Fig. 3A, the decrease in LAR is caused completely by a decrease in the SLA of the plants, which is to a large extent explained by the accumulation of nonstructural carbohydrates
25. 1.6
1.3
"--
1.0
>
0 0
0 0
e-
.-~ 7-
1.3
(A) Box plots of the observed relative changes in the growth parameters RGR, NAR (net assimilation rate), LAR (leaf area ratio), SLA (specific leaf area) and LWR (leaf weight ratio), as observed for 63 different C~ species. (B) Idem for 8 C4 species. For further information see Fig. 2. Data are from Badger (1992), Baxter (1994), Bowler and Press (1993), Callaway (1994), Carter and Peterson (1983), Chu (1992), Coleman and Bazzaz (1992), Collins (1976), DeLucia (1994), Den Hertog (1993), Hocking and Meyer (1991), Hurd and Thornley (1974), Musgrave and Strain (1988), Pettersson and McDonald (1992), Poorter (1993), Roumet (1993), Rozema (1993), Ryle (1992a,b), Van de Staaij (1993), Wong (1992), Wong (1993), and Wyse (1980) as cited in this reference list; and data from Bazzaz (1989), Bhattacharya (1985), Bunce (1990), Cure (1987, 1988),Jansen (1986),Jolliffe and Ehret (1985), Marks and Clay (1990), Mauney (1978), Neales and Nicholls (1978), Overdieck (1988), Patterson (1986), Patterson and Flint (1980, 1982), Patterson (1988), Peet (1986), Poorter (1988), Rogers (1984), Sionit (1983), Sionit (1982), Thomas (1991), and Wong (1990) as cited in the reference list from Poorter (1993).
384
(e.g., Wong, 1990). Although there is some variation in LWR values, the mean response of these relatively well-nourished plants is nil (cf. Stulen and Den Hertog, 1993). It should be noted that in quite a number of cases the first harvest of the growth analysis started after the onset of CO2 enrichment. As most of the growth response to elevated CO2 occurs in the first 10-20 days, changes in RGR and NAR may have been underestimated to some extent. A limited number of growth analyses have been carried out with C4 species. Generally, the responses are much smaller (Fig. 3B). Although the number of species in this group that is investigated (n = 8) does not allow much of a generalization, it is remarkable to see that the results seem to mirror those of the C~, with decreases in RGR and NAR, and increases in SLA.
B. Within C3 Species Sink strength has been put forward as a main modulator of the response of C3 species. As such, within the group of C3 species three main categories can be discerned. First, crop species, which have been selected to grow vigorously and thus may have a large sink strength during development. Second, the wild herbaceous species, where human selection pressure to increase sink strength generally has been much smaller. And third, woody species, which are generally slow-growing and morphologically and physiologically quite different from the first two groups anyway. On average, the crop species are the strongest in responding, at least in the vegetative growth period, with a mean dry weight increase of 58% (Fig. 2A, The response of wild herbaceous species and trees is smaller, as far as their growth stimulation is concerned (42 and 44%, respectively). However, given the larger time scale of the experiments with trees than with herbs, there is a distinction between the average stimulation in RGR of herbaceous and woody species (Fig. 2B). Generally, growth stimulation by high CO2 in herbaceous species is only temporary. Because few time trends have been analyzed for trees, it remains speculative whether they will be stimulated over a longer time period (but see Bazzaz 1993). In a comparative growth analysis of tree seedlings and a grass species, Gloser (Chapter 21) found the tree seedlings to be stimulated throughout the season in contrast with the grass species. Unfortunately, our insight into the time dependency of the growth stimulation is still fragmentary.
C. Within Wild Herbaceous C3 Species The most clear-cut difference we found in our analysis was when wild herbaceous C~ species were categorized in fast-growing, intermediate, and slow-growing species (Fig. 4). There is a general trend that fast-growing species respond much stronger than slow-growing species (60% versus
25. 3.0
~'~ n,'
Figure 4 (A) Box plots of the observed weight ratios of C3 species of the categories inherently slow-growing (n = 42), intermediate (n = 36), and fast-growing (n = 30) wild herbaceous C3 species, to the left side, and for evergreen (n = 40) and deciduous (n = 43) woody species, to the right side. (B) Idem for the absolute RGR stimulation of those categories. For further information see Fig. 2.
27%). Using these categories, 18% of the total variation in the weight ratios of Appendix 2 could be explained, and 22% in the variation in RGR stimulation. These values in themselves are not particularly high. However, given the large variability that can be expected for weight ratios (see Section IIB), this fraction of explained variance is probably at the high side of what could be achieved anyway. The grouping was carried out on the basis of a number ofmsubjectivem criteria (see Poorter, 1993). Therefore, it is good to compare these results with literature. In Fig. 5 we compiled the available data on experiments where different plant species were grown at elevated CO2 and where RGR
386
40 9
I~
~
20
rv'
mS""~ ~
0
8
1
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
RGR c (rag g-1 day-l) Relationship between the absolute RGR stimulation due to high CO2 and the observed RGR at control levels of CO2. Each line represents a linear regression through literature data of different plant species grown in the same experiment. Continuous lines indicate experiments with herbaceous species, dotted lines are shown for experiments with woody plants. Data for the herbaceous plants are of Baxter (1994; 1), Bowler and Press (1993; 2), Campbell (1993; 3) and Gamier (unpublished; 4), Gobin (unpublished; 5), Mortensen (1985; 6); Musgrave and Strain (1988; 7), Overdieck (1988; 8), Poorter (1993; 9), Roumet (unpublished; 10), Tremmel and Patterson (1993; 11), and Wong (1993; 12). Data for woody species are from Lindroth (1993; a), Mortensen (1994a; c), (1994b; d), Tolley and Strain (1984; e), Wiggins (unpublished; f) and Wong (1992; g). In the experiment indicated with a *, a herb and a tree were compared Chapter 21). In those cases where the first harvest of the growth analysis occurred after the start of the CO2 enrichment, RGR stimulation was calculated according to Appendix 1.
values were d e t e r m i n e d . A l t h o u g h n o t e a c h of these r e p o r t s gave positive, significant relationships b e t w e e n the growth rate of a species a n d its RGR stimulation at h i g h CO2, the average t r e n d is a positive one. A striking result of c o m b i n i n g the data f r o m all the e x p e r i m e n t s is the e m e r g e n c e o f a clear r e l a t i o n s h i p b e t w e e n the RGR of plants at c o n t r o l levels a n d their r e s p o n s e to CO2 (Fig. 5). I n d e p e n d e n t of w h e t h e r RGR was low b e c a u s e the species u n d e r investigation were tree seedlings ( d o t t e d lines in Fig. 5) or b e c a u s e h e r b a c e o u s plants were grown over p r o l o n g e d p e r i o d s o f time a n d t h e r e f o r e showed low RGR's (like the e x p e r i m e n t s n u m b e r e d 1 a n d 4), the growth stimulation does n o t e x c e e d the 10 m g g-1 day-1. For the fast-growing h e r b a c e o u s species m e a s u r e d over relatively s h o r t p e r i o d s stimulation t u r n e d o u t to be at least twice as high. T h e large r e s p o n s e o f c r o p species in the vegetative stage, as discussed in Section IIIB, fits into this picture, as they will generally be i n t e r m e d i a t e to fast-growing as well.
387
25.
An alternative hypothesis to explain the pattern of variation in wild C~ species has been put forward by H u n t (1991, 1993). In a combined analysis of three experiments conducted in 3 consecutive years, with a total n u m b e r of 36 species, they found that the response of species was best characterized by their position in the strategy triangle of Grime (1979). The further the species were away from the C-point of the triangle (the "competitive" strategy), the less they did respond. We have several reasons why we think the relationship as observed by H u n t is not a general one. First, we were not able to reproduce this relationship when we tried to do so for 37 New Zealand grassland species (data not shown). Second, the negative relationship as given in H u n t (1993, Fig. 2) seems to be caused mainly by two outlying values, one of which is the average of an extremely high and a below-average weight ratio. Given the relatively high values in the standard deviations of In-transformed dry weights for at least some species (0.4 and 0.6 as derived from Fig. 2 of H u n t 1991), we suggest that these outliers are most likely caused by variability in their plant material (cf. Section IIB). The third and most important reason to question the relationship is a methodological one. Stated somewhat simplified, the C-radius theory expects that competitors (sensu Grime, 1979) will respond strongly to high CO2, whereas species with a ruderal or stress-tolerating strategy will not. Given that there were very few ruderals in the experiment of H u n t it is likely that the negative relationship is determined to a large extent by the competitors and stress tolerators only. In our opinion, it would be more appropriate to test for the differences between the groups of competitors, stress tolerators, and ruderals explicitly. When we use the full information of our data base, for all the species that have been classified by Grime (1988), we find that stress tolerators do respond significantly less than competitors and ruderals, and that no significant differences can be detected between the two latter groups (Table I). Therefore, the differences reduce to a contrast between the slow-growing species (stress tolerators) and fast-growing ones (ruderals and competitors). Such a division seems to make more sense, given the differences in physiology
Strategy Stress tolerators Competitors Ruderals
WH/Wc
RGRH-RGRc (mg g-1 day-l)
n of species
1.23 1.56 1.64
5.1 13.6 17.2
12 6 8
Plants categorized into the middle part of the triangle (CSR, CS, SR, or CR) were not included.
and morphology in the vegetative stage between stress tolerators on one side, and ruderals and competitors on the other side (cf. Poorter and Remkes, 1990; Poorter 1990; Garnier, 1992). However, if the analysis would be extended to the reproductive phase, it is possible that ruderals and competitors would respond in quite different ways. If ruderals start to flower and stop increasing in vegetative biomass when they reach a certain size, the difference that shows up early in life may disappear later. We made some other a posteriori analyses within the group of wild herbaceous species. No consistent differences were observed between monocotyledonous and dicotyledonous species (Fig. 6). Species that are capable
l
, i
I
I I ..............i..............• ............................[ ............i ..............
(A) Box plots of the observed weight ratios of monocotyledonous (n = 56) and dicotyledonous (n = 74) herbaceous C3-species, to the left side, and those species that potentially can fix N2 in symbiosis (n = 20) and those that cannot (n = 112), to the right side. (B) Idem for the absolute RGR stimulation of those categories. For further information see Fig. 2.
25.
389
of symbiotically fixing N2 did respond somewhat more strongly than the other herbs, though not significantly. However, given the high N availability in most of these experiments, nodulation is likely to have been inhibited, making the comparison somewhat obsolete. Direct evidence comparing the two groups of species is scarce. In the data set of Campbell (1993), with 12 leguminous species/cultivars and 20 nonleguminous ones, no significant difference between the groups was found either, although the tendency was similar as in Fig. 6. See also LOscher (Chapter 19).
D. Within Woody Species A large amount of data has been published in recent years about the response of woody species. These experiments were mainly on tree seedlings, and the median duration of these experiments is 110 days only (see Appendix 2). Thus, most research has focused on only a very limited time span of the total life cycle of these species (cf. K6rner, 1993b). Within the group of tree species, we could envisage a difference between deciduous and evergreen species, similar to the difference between fastand slow-growing species observed above. Although deciduous species do have a slightly higher RGR stimulation, this difference does not show up in the weight ratio (Fig. 4, the difference is not consistent, some direct experiments are required to obtain more insight into this problem. For a more detailed analysis of the physiological performance of tree species see Ceulemans and Mousseau (1994).
The above literature survey comprises experiments in which the availability of nutrients will have been relatively high compared to what plants generally encounter in nature. To what extent will this affect the weight ratios? Evidence is conflicting, with some reports showing weight ratios to decrease with diminishing nutrient availability, and others where the weight ratio is constant or even increases. We could imagine that this may partly coincide with the severity of the applied nutrient stress. I n some reports the low N treatment hardly caused any growth reduction (e.g., Larigauderie 1994), whereas in others the low N plants weigh only a small fraction of the control. To allow for this variation, we plotted the weight ratio (weight of high CO2 relative to those of control levels) against the relative reduction in size of the plants at control levels of CO2 due to the low nutrient treatment, for both trees and herbaceous C~ species (Fig. 7). The observed relationship shows a lot of scatter, as expected (Section IIB). A way to correct for this is to consider the data of each species and experiment separately. We did so by calculating regression lines through all the points
o
9
0.0
9
o
o
o
o
o
0',
9
;.0
Wc,Iow NU / Wc,high NU Figure 7 Weight ratios of plants grown at high and low nutrient availability. The x-axis values indicate the severity of the nutrient stress, calculated as the weight of the plants at low nutrient availability divided by the total plant weight at the highest nutrient availability. The circles pertain to observations on herbaceous Ca species, the squares to observations on woody species. The continuous and broken line indicate the average response of herbaceous and woody species, respectively. Values are from Bassow (1994), Bazzaz and Miao (1993), Bowler and Press (1993), Coleman (1993), E1 Kohen and Mousseau (1994), E1 Kohen (1992), Griffin (1993), Hocking and Meyer (1985), Johnsen (1993), Kerstiens and Hawes (1994), Larigauderie (1994), McConnaughay (1993), Norby and O'Neill (1991), Radoglou and Jarvis (1992), Silvola and Alholm (1992, 1993), Wilkins (1994), Wong (1990), and Wong (1992), as listed in the References, and from Cure (1988), Goudriaan and De Ruiter (1983), Larigauderie (1988), Marks and Clay (1990), Oberbauer (1986), and Patterson and Flint (1982), as cited in the references from Poorter (1993).
o f Fig. 7 that p e r t a i n e d to o n e species in o n e e x p e r i m e n t . T h e m e a n slope o f all these regression lines is positive (P < 0.01), indicating that the average g r o w t h r e s p o n s e of plants to h i g h CO2 d e c r e a s e s w h e n n u t r i e n t availability decreases. This applies to woody a n d h e r b a c e o u s C3 species to the same extent, as can be seen f r o m the regression lines in Fig. 7. For b o t h g r o u p s of species the r e s p o n s e is almost nil at e x t r e m e l y limiting conditions. It s h o u l d be realized t h a t the r e s p o n s e s h e r e are analyzed in a relative way. It is clear t h a t the growth stimulation at low N levels is m u c h lower anyway, w h e n the absolute r e s p o n s e s are c o n s i d e r e d ( K t r n e r , 1993b).
As m e n t i o n e d above, the data listed in A p p e n d i x 2 have b e e n o b t a i n e d o n individually grown plants, at relatively h i g h availabilities of water a n d
25.
391
nutrients. To what extent this information can be extrapolated to the natural field situation, with its myriad of interactions and feedbacks between a wide range of organisms is unknown. As far as such extrapolations are permitted, there are two points that deserve attention. First, one should realize that a high weight ratio does not imply that the high-CO2 plants have increased in "functional size" to the same extent. This is due to the fact that the higher dry weight is caused partly by an increased accumulation ofnonstructural carbohydrate in the leaves. Thus, in a competitive situation, where light interception, and thus total leaf area, is of more importance than leaf weight, differences between types of species will be smaller than expected on the basis of the analysis of Section III. Second, responses of biomass to CO2 in natural environments will generally be less than in these compiled experiments, due to a lower nutrient availability (Field 1992; Section IV; but see Gifford, 1994). Let us make a simple contrast of inherently slow-growing species growing in low-nutrient (low-resource) environments and inherently fast-growing species in nutrient-rich (highresource) environments (cf. Grime, 1979; Poorter and Remkes, 1990). Given the low nutrient availabilities in the habitats of the inherently slowgrowing species, and the very modest response of those species anyway (Fig. 4), we do not expect much of a change in biomass in such environments. A somewhat larger response can be anticipated of the faster-growing species growing in their nutrient-richer habitat (say, an increase in biomass of ---25%). We doubt whether this by itself will have a large impact on the vegetation. Indeed, in the experiments in natural vegetations under way, generally very small responses in biomass are observed (K6rner Chapter 28). A somewhat different question is whether the floristic composition of the vegetation will alter. Up to now, we have assumed that all species within a functional group will respond more or less similarly. However, if there is considerable variation between species within the functional types discerned, this may still affect the floristic composition of the vegetation, even if the total a m o u n t of biomass does not change that much. Also at this point the available evidence is scarce (Bazzaz, 1990). Intended more as a provocative statement than a conclusion we suggest that, given these considerations, the effect of an increase in the atmospheric CO2 concentration on the growth of plants in natural vegetations is relatively small, and that any changes in the ecosystem are more likely to occur via an increased water use efficiency or via changes in the chemical composition of the plants, rather than via a biomass response.
In this chapter we discussed aspects of the growth response of plants to elevated CO2. A simple way to express the stimulation is to calculate the
392 ratio between the biomass of high-CO2-and control-CO2-grown plants, the weight ratio. A major problem with weight ratios is that their variability is high, even when relatively large numbers of plants are harvested. Another weak point of this ratio is that it does not include a time effect. Alternatively, the absolute stimulation in relative growth rate may be used. The amount of information required for this analysis cannot be obtained from a single experiment, and therefore results of 250 plant species from a large number of experiments have been compiled in order to calculate the average response of groups of plants. Basically, the conclusions are similar to those of Poorter (1993). C3 plants were found to be more responsive (47% increase in weight) than those with the C4 and CAM photosynthetic pathway. However, on average C4 species also respond significantly to elevated CO2 (10%). There is insufficient insight into the causes of this stimulation. Also within the C~ group of species differences were found between functionally different types of species. On average, potentially fast-growing wild species and crop species show relatively strong growth responses to high CO2 (58 and 60% increase in weight, respectively), whereas the response of inherently slow-growing species is only half of that (27%). Woody species have an intermediate response when final weights are considered (44%), but are far less responsive when the longer time span in those experiments is taken into account. When the growth response of plants to high CO2 is analyzed at a low nutrient availability, on average a lower weight ratio is found, with no growth stimulation at all at severely limiting conditions. As far as these results of plants grown individually can be extrapolated to natural vegetations, this is one of the reasons that we expect the growth response in the field to be small.
We thank Eric Garnier,Jan Gloser, Kevin Griffin, Marilyn Ball, and Olivier Gobin for providing us with as yet unpublished data. Hans Lambers and Adrie van der Werf made helpful suggestions on an earlier draft of the manuscript.
Assume that plants have a weight W1 at time tl and that they grow with a constant RGR over the time interval tl to t2. Assume that plants are separated into two groups at time h, high-CO2-grown plants (H) and control plants (C). The weight of the control plants at time t2 is then given by
393
25. Wc --
W1 9 e RGRc (t2-tl)
(1)
and that of the high CO2 plants by W H --
W 1 " e RGRH (t'2-tl).
(2)
The weight ratio at time t2 is then given by WH
Wc
W1 9 eRGRH (t2-t 1)
=
W~ 9e~e~ (~-,,/
(3)
and thus - - e(RGRH -RGRc) (t~-tl).
(4)
Wc Consequently, the absolute average stimulation of RGR over time can be calculated as
w~ ln~ Wc A RGR = . t2-tl
(5)
Note that it is the absolute difference in RGR over a certain time period that is of interest, rather than the relative response, as RGR is a relative p a r a m e t e r itself already. From Eq. 4 one can derive that it is an absolute difference in RGR that causes a certain relative difference in plant weight over a given time span. Equation (4) can be rewritten as WH
Wc
"- e
(k-l) r (t2-tl)
where r = RGRc and = RGRH. By definition, potentially fast-growing species have higher r than slow-growing species. If high CO2 would have the same relative effect k on the RGR of both fast- and slow-growing species, then (k - 1)r would be higher for the fast-growing species, and, consequently, the weight ratio of these species. In physiological terms, this could be the case if the leaf a r e a : p l a n t weight ratio is higher for the fast- than for the slow-growing species, and photosynthesis is stimulated in all species to the same extent.
A compilation of the ratio of total weight of plants grown at a high (60-80 Pa) and at a control concentration (30-40 Pa) of CO2. Data are of various literature sources and comprise 503 observations on a total of 256
394
species. Final yield was taken when plants remained vegetative. In other cases plant weight before flowering or fruiting was used. In those cases where mean relative growth rates were given, these values were used for the calculation of the weight ratio, as they summarize data of more than one harvest. For each species and reference, the n u m b e r of days that the experiment lasted (n days), and the total n u m b e r of plants harvested per treatment on which the ratio is based (n plants), are given. Previous no. indicates the n u m b e r of references in Poorter (1993) for the same species. These references are not repeated here, but those data are also used to calculate the mean weight ratio per species. Mean values per species and per category are backtransformed values of averaged log-transformed ratios, to correct for the intrinsically skewed nature of ratios. For the C~ wild species it is indicated whether they are potentially slow-growing (s), intermediate (i), or fast-growing (f). For each category, the median value of the weight ratio, the duration of the experiment, and the n u m b e r of plants harvested per treatment are given. n n days plants
Species
Mean Previous weight no. ratio
Reference
A. C3 crop species
1.43
20
10
1.51
35
5
1.93 1.80 2.08
18 32 30
12 12 4
Gobin
(unpublished)
Tremmel and Patterson (1993) Rogers (1992) Rufty (1994) Thomas (1993)
1
1.08
0 1 1 1 13
1.43 1.56 1.83 2.10 1.71
1.82 1.64
3.91 1.42 1.79 2.13 2.23 2.29 1.65
20 28 28 28 20 28 31
10 10 10 10 10 10 ?
1.75 2.01 2.75
35 35 28
3 3 10
Gobin (unpublished) Campbell (1993) Campbell (1993) Campbell (1993) Gobin (unpublished) Campbell (1993) Stanghellini and Bunce (1993) Ziska and Bunce (1993) Ziska and Bunce (1994) Campbell (1993)
2.04 1.82
1.38 1.93
395
25.
n
Species
n
days plants 1.51
9
9
1.19
113
20
1.20
113
10
1.33
113
20
1.62
113
10
1.30
29
6
1.61
60
3
1.34 1.41 1.41 1.58
28 40 40 28
4 14 14 ?
1.06
45
6
Reference Sicher et
(1994)
Ziska and Teramura (1992b) Ziska and Teramura (1992a) Ziska and Teramura (1992b) Ziska and Teramura (1992a) Radoglou and Jarvis (1992)
Retuerto and Woodward (1993) Billes (1993) Nicolas (1993) Nicolas (1993) Rozema (1993) Radoglou and Jarvis (1993)
Mean weight ratio
0
1.51
2
1.37
0
1.30
4 5 0
1.36 1.64 1.61 1.47
1.46 1.23 1.45 1.80
u Median values (n = 97)
Previous no.
1.58
28
8
1.24 1.33
28 50
6 5
1.55
35
5
1.68 1.35
40 150
? 5
B. C3 wild species Dippery (1995) Coleman and Bazzaz (1992) Tremmel and Patterson (1993) Coleman (1993) Downton and Grant (1994)
1.43
1.35 1.71
1.59 2.00
58 79
6 7
Bowler and Press (1993) Baxter (1994)
1.57 1.82
396
n
Species
days plants 2.20
28
Previous no.
n
10
Reference Campbell
Mean weight ratio
(1993) 1
1.10
1 0
1.40 1.38
1
1.21
Ferris and Taylor (1993)
0
0.96
1
1.50
1
1.35
1 1
1.13 1.15
0 1
1.21 1.00
0 2
1.27 1.46
1 0 1
3.60 1.03 1.66
1.00 1.92 1.21
49 28 52
8 10 8
Hunt Campbell Hunt
(1993) (1993) (1993)
0.96
100
5
1.14 2.53 1.67
52 20 63
8 10 10
Hunt (1993) Gobin (unpublished) Johnson and Lincoln (1991)
1.12 1.19 1.21 1.00
100 25 20 52
16 11 10 8
Lenssen (1993) Lenssen (1993) Gobin (unpublished) Hunt (1993)
1.27 1.19 1.73
49 20 144
8 10 4
Hunt (1993) Gobin (unpublished) Gamier et a/. (unpublished)
1.03 1.34 1.37
20 20 180
10 10 4
Gobin (unpublished) Gobin (unpublished) Garnier (unpublished)
1.49
20
10
Gobin
(unpublished)
1.16 1.49 2.92 2.57 1.04 1.72 1.41
1.81 2.16
28
10
Campbell
(1993)
?
Gloser and Bartok (1994)
1.69 2.16
397
25.
Species
n
n
days
plants
Previous no.
Reference
1.37
35
5
1.90 1.18
21 49
10 8
Tremmel and Patterson (1993) Gobin (unpublished) Hunt (1993)
1.51
21
10
Gobin
(unpublished)
Mean weight ratio
1
1.30
1 1 1
1.13 1.10 1.48 1.90 1.18
1.51 1.59
1.25
52
8
Hunt
(1993)
2.14 1.26
1.74
28
10
Campbell
(1993)
1.74 1.67
2.22
28
10
Campbell
(1993)
2.22
1.27 1.56 1.63 1.74
52 28 56 35
8 10 3 3
Hunt (1993) Campbell (1993) Ziska and Bunce (1994) Ziska and Bunce (1993)
2.31
1.72 1.28 1.30 1.16 1.25
49
8
1.06 1.29 1.31 1.36 1.40 1.64 1.64 1.67
71 29 65 34 23 65 65 35
18 15 12 15 17 12 14 5
Hunt
(1993)
Lenssen (1993) Lenssen (1993) Lenssen (1993) Lenssen (1993) Lenssen (1993) Lenssen (1993) Van de Staaij (1993) Tremmel and Patterson (1993)
1.25 1.33 1.37
1.67
n
Species
days plants 1.12
52
Mean Previous weight no. ratio
n
8
Reference Hunt
Campbell
(1993)
1.11
1
1.19
0
1.30
2 1 0 0 1
1.44 1.00 0.42 1.23 1.00
1.30
28
10
1.24 0.42 1.23
58 189 52
4 6 8
1.76 2.67
28 20
10 10
Campbell Gobin
(1993) (unpublished)
1 0
1.68 2.67
1.63
20
10
Gobin
(unpublished)
0
1.63
1
1.02
1
O.78
1 0
1.46 1.03
0
1.61
0
1.26
Hunt Baxter Hunt
Hunt
(1993)
1
(1995) (1994) (1993)
1.03
49
8
(1993)
1.57 1.66 1.26
28 28 20
10 10 10
Campbell Campbell Gobin
1.61 1.00 1.75 2.11 1.93
20 49 100 28 28
10 8 5 10 10
Gobin (unpublished) Hunt (1993) Ferris and Taylor (1993) Campbell (1993) Campbell (1993)
(1993) (1993) (unpublished)
(wild) 1.61 1.55
1.93 1.57
1.27
146
4
1.31
146
4
Garnier (unpublished) Garnier (unpublished)
1.27 1.31 1.17
spec?
0.92
399
25.
n
Species
Mean Previous weight no. ratio
n
days plants
1.22 1.00
62 49
6 8
1.94
28
10
Reference
2
1.36
Bowler and Press (1993) Hunt (1993)
0 0 1
1.22 1.00 1.32
Campbell
1
1.67
1 0 1 2
2.72 2.09 1.02 1.28
3 0 0 1 0 1 1
1.48 0.90 1.30 1.23 2.09 1.03 1.48
0
1.65
1
0
1.31 2.08
0
1.92
1 1
1.46
(1993)
2.09
28
10
Campbell
(1993)
1.38
28
10
Campbell
(1993)
0.90 1.30 1.50 2.09
100 105 28 28
5 6 10 10
Ferris and Taylor (1993) Baxter (1994) Campbell (1993) Campbell (1993)
1.65
62
15
Lenssen (1993)
2.08
28
10
Campbell
1.92
100
5
(1993)
Ferris and Taylor (1993)
1.70
2.30
28
10
Campbell
(1993)
0
2.30
1.76
28
10
Campbell
(1993)
0
1.76
1.70
28
10
Campbell
(1993)
0
1.70
1.77 1.83 2.16 2.24 2.31 2.66 0.86
28 42 28 28 28 28 28
10 8 10 10 10 10 10
Campbell (1993) Ryle (1992a) Campbell (1993) Campbell (1993) Campbell (1993) Campbell (1993) Campbell (1993)
1.77 2.00
1.24
400
Previous no.
Mean weight ratio
10
Van der Eerden (1993) Gobin (unpublished)
0
1.61
21
10
Gobin
0
1.32
1.16 1.73
52 28
8 10
Hunt Campbell
(1993) (1993)
4 0
1.45 1.73
1.44
38
8
1.76
365
20
Samuelson and Seiler (1992)
0
1.76
1 1
1.40 1.21
0
1.25
0
1.38
1
1.60
0
2.20
n
n
days
plants
1.78
40
?
1.61
21
1.32
Species
Median values (n = 174)
Reference
(unpublished)
C. C3 woody species
1.11 1.20 1.47 1.07 1.20 1.42 1.58 1.03 1.44 1.66 2.67 4.68 2.20
70 210 900 60 900 210 70 35 60 51 100 140 100
4 6 6 10 6 6 5 5 7 3 10 6 5
1.02 1.03 1.32 1.94
210 900 79 90
6 6 4 10
2.15
90
10
Bassow (1994) Bazzaz and Miao (1993) Bazzaz (1993) Bazzaz (1990) Bazzaz (1993) Bazzaz and Miao (1993) Bunce (1992) Reid and Strain (1994) Lindroth (1993) Bunce (1992) Bazzaz (1990) Tschaplinski (1995) Wiggins (unpublished)
Bazzaz and Miao (1993) Bazzaz (1993) Bassow (1994) Rochefort and Bazzaz (1992) Rochefort and Bazzaz (1992)
1.44 1.73 4.13 1.28
2.15 0.90
401
25.
n
n
days
plants
1.24 1.57
60 90
10 10
1.18 1.58
37 122
16 4
1.67
38
3
3.07 0.99 1.21 1.21 1.44 2.44
34 900 66 210 79 90
20 6 9 6 4 10
1.01 1.20 1.26
540 150 150
4 16 24
1.27
180
18
1.66 1.20 2.01 1.30
270 63 63 100
10 30 30 5
Species
Previous no.
Reference Bazzaz (1990) Rochefort and Bazzaz (1992) Mortensen (1994b) Silvola and Ahlholm (1995) Pettersson and McDonald (1992) Pettersson (1993) Bazzaz (1993) Miao (1992) Bazzaz and Miao (1993) Bassow (1994) Rochefort and Bazzaz (1992) Rouhier (1994) E1 Kohen (1993) E1 Kohen and Mousseau (1994) E1 Kohen (1992)
Mean weight ratio
0
1.39
0
1.76
0
1.39
1
1.23
1.14 Kaushal (1989) Cruz (1993) Cruz (1993) Wiggins (unpublished)
1.66 1.55 1.30
1.61 2.11 3.44 1.57 3.90
42
4
Conroy
(1992)
2.03 1.45 1.32 2.56 1.19
402
Species
n
n
days
plants
Previous no.
Reference
1.50 1.83 1.62
70 60 150
6 10 22
Reid and Strain (1994) Bazzaz (1990) E1 Kohen (1993)
1.38 1.51
210 900
6 6
Bazzaz and Miao (1993) Bazzaz (1993)
1.25 1.38
81 82
54 54
1.42
128
6
1.16
850
10
Mortensen (1994a) Mortensen (1994a) Tschaplinski Norby
(1995) (1992)
Mean weight ratio
1
2.74
0
1.66
0 1 0
1.62 1.10 1.44 1.26 1.25 1.38 0.90 1.52 1.42 2.35 1.32
1.12 1.61
59 126
5 10
Bunce (1992) Berryman
(1993)
1.12 1.61 0.92
1.32
90
4
Downton and Grant (1994)
1.46 1.17 1.79 1.30
1.07 1.14 1.23 1.35 1.35 1.44
111 49 111 110 401 112
90 16 90 90 16 20
1.47 1.30 1.50
115 155 112
90 48 20
1.65
305
80
Mortensen (1994a) Mortensen (1994b) Mortensen (1994a) Mortensen (1994a) Mortensen (1994a) Yakimchuk and Hoddinott (1994) Mortensen (1994a) Johnsen (1993) Yakimchuk and Hoddinott (1994) Samuelson and Seiler (1994)
1.22
1.46
1.40
1.70
403
25.
n
n
days
plants
1.75
152
40
1.16 1.24 1.43
115 390 300
90 45 10
1.82
112
20
1.02
112
90
1.04 1.29 1.67 1.26 1.48
112 270 120 60 166
90 10 12 16 8
1.76
407
3
1.20 1.03 1.09 1.21 1.28 1.40 1.59 1.79
100 112 112 112 100 120 177 172
10 90 90 90 5 48 5 10
1.82
166
8
1.63
126
6
Species
Previous no.
Reference Samuelson and Seiler (1993) Mortensen (1994a) Lee (1994) Stewart and Hoddinott (1993) Yakimchuk and Hoddinott (1994) Mortensen (1994a) Mortensen (1994a) Kaushal (1989) Guehl (1994) Callaway (1994) Griffin (unpublished) Johnson (1995) Bazzaz (1990) Mortensen (1994a) Mortensen (1994a) Mortensen (1994a) Griffin (1993) Lewis (1994) Thomas (1994) Larigauderie (1994) Griffin (unpublished) Tschaplinski
(1995)
spec. x
spec.
1.49
158
5
1.26
70
34
Curtis and Teeri (1992)
1.34 0.64 0.65 1.23 1.48
152 195 195 195 60
24 5 5 5 7
Zak (1993) Ceulemans (1994) Ceulemans (1994) Ceulemans (1994) Lindroth (1993)
Curtis
(1995)
Mean weight ratio
0
1.20
0
1.62
0 2 0 0 0 0
1.02 1.38 1.04 1.29 1.67 1.49
1.35 1.20 1.17
1.45
1.15 1.13 2.13 1.65 1.58 1.30
1.06
1.48
n
n
days
plants
1.41
180
16
1.81 1.33 1.03 1.06 1.36 2.00 2.81
540 60 450 116 450 90 90
3 10 6 90 6 9 10
Species
2.37 0.89 1.17 1.25 1.46 2.21 3.72
120 64 198 210 900 60 520
12 5 14 6 6 7 13
Previous no.
Reference Kerstiens and Hawes (1994) Wilkins (1994) Bazzaz (1990) Gorissen (1995) Mortensen (1994) Gorissen (1995) Condon (1992) Condon (1992)
Guehl (1994) Bunce (1992) Vivin (1995) Bazzaz and Miao (1993) Bazzaz (1993) Lindroth (1993) SeegmOller and Rennenberg (1994)
Mean weight ratio
0
1.59
0 1
1.33 1.11
0 0 1
2.00 2.81 1.43
1.20 1.78 2.37 0.89 1.17 1.59
1.17 1.07 1.32
x
1.46
120
4
3.27
120
3
Silvola and Ahlholm (1992) Silvola and Ahlholm (1993)
1.46 3.27 1.19 2.64 1.21
Median values (n = 174):
1.40
100
10
1.40
112
8
Bazzaz
(1990)
1.40
D. C4 species 1.26 1.02 0.90
28
6
Dippery
(1995)
1.26
405
25.
Species
n
n
days
plants
1.14
35
5
1.27
50
5
1.39
40
?
1.37 0.94
76 28
8 10
Previous no.
Reference
Mean weight ratio
Tremmel and Patterson (1993) Coleman and Bazzaz (1992) Coleman (1993)
Morgan
(1994)
Campbell
(1993)
1
0.63
1
1.14
1
1.30
1
1.06
1
1.23
5
1.37
3 1
1.11 1.45
1
1.22
0.95
28
10
Campbell
(1993)
0
0.95
1.15
28
10
Campbell
(1993)
0
1.15
1
1.21
2 0
1.11 0.88
0.88
28
10
Campbell
(1993)
1.73 1.52 1.09
35
5
Tremmel and Patterson
1.10
(1993)
Median values (n = 49)
0.80 0.90 0.95 0.83
58 58 57 28
28 14 35 10
0.87 1.14
?
?
30
8
Lenssen (1993) Lenssen (1993) Lenssen (1993) Campbell (1993)
Rozema (1993)
0.88
0.83 0.94 1.09
n n days plants
Species
Mean Previous weight no. ratio
Reference
E. CAM species
Median values (n=9) All species Median values (n = 503)
1.10
126
4
Nobel
(1994)
1.34 1.44
161 84
5 5
Cui (1993) Cui and Nobel (1994)
1.15
161
5
1.42
49
8
1
1.36
1 1
1.22 1.08
1 1
0.90 1.14
1
1.22
Badger, M. (1992). Manipulating agricultural plants for a future high CO~ environment. 40, 421-429. Barnett, V., and Lewis, T. (1978). "Outliers in Statistical Data." John Wiley, Chichester. Bassow, S. L., McConnaughay, K. D. M., and Bazzaz, F. A. (1994). The response of temperate tree seedlings grown in elevated CO2 to extreme temperature events. App. 4, 593-603. Baxter, R., Ashenden, T. W., Sparks, T. H., and Farrar, J. F. (1994). Effects of elevated carbon dioxide on 3 montane grass species. I. Growth and dry matter partitioning. 45, 305-315. Bazzaz, F. A. (1990). The response of natural ecosystems to the rising global CO2 levels. 21, 167-196. Bazzaz, F. A., and Miao, S. L. (1993). Successional status, seed size, and responses of tree seedlings to CO2, light, and nutrients. 74, 104-112. Bazzaz, F. A., Coleman,J. S., and Morse, S. R. (1990). Growth responses of major co-occurring tree species of the northeastern United States to elevated COz. 20,1479-1484. Bazzaz, F. A., Miao, S. L., and Wayne, P. M. (1993). CO2-induced growth enhancements of co-occurring tree species decline at different rates. 96, 478-482. Berryman, C. A., Eamus, D., and Duff, G. A. (1993). The influence of CO2 enrichment on growth, nutrient content, and biomass allocation of 41, 11-21.
25.
407
Billes, G., Rouhier, H., and Bottner, P. (1993). Modifications of the carbon and nitrogen allocations in the plant L.) soil system in response to increased atmospheric CO2 concentration. 157, 215-225. Bowes, G. (1991). Growth at elevated CO2: Photosynthetic responses mediated through rubisco. 14, 795-806. Bowler, J. M., and Press, M. C. (1993). Growth responses of 2 contrasting upland grass species to elevated CO2 and nitrogen concentration. 124, 515-522. Bunce, J. A. (1992). Stomatal conductance, photosynthesis and respiration of temperate deciduous tree seedlings grown outdoors at an elevated concentration of carbon dioxide. 15, 541-549. Callaway, R. M., DeLucia, E. H., Thomas, E. M., and Schlesinger, W. H. (1994). Compensatory responses of CO2 exchange and biomass allocation and their effects on the relative growth rate of ponderosa pine in different CO~ and temperature regimes. 98, 159-166. Campbell, B. D., Laing, W. A., and Newton, P. C. D. (1993). Variation in the response of pasture plants to carbon dioxide. "Proceedings of the XVII International Grassland Congress," pp. 1125-1126. Carter, D. R., and Peterson, K. M. (1983). Effects of a CO2-enriched atmosphere on the growth and competitive interaction of a C3 and C4 grass. 58, 188-193. Ceulemans, R., and Mousseau, M. (1994). Effects of elevated atmospheric CO2 on woody plants. 127, 425-446. Ceulemans, R., Perez-Leroux, A., and Shao, B. Y. (1994). Physiology, growth and development of young poplar plants under elevated CO~. "Vegetation Modelling and Climate Change Effects" (F. Veroustrate and R. Ceulemans, eds.), pp. 81-98. SPB Academic Publishing, The Hague, The Netherlands. Chu, C. C., Coleman,J. S., and Mooney, H. A. (1992). Control ofbiomass partitioning between roots and shoots: Atmospheric CO2 enrichment and the acquisition and allocation of carbon and nitrogen in wild radish. 89, 580-587. Coleman, J. S., and Bazzaz, F. A. (1992). Effects of CO2 and temperature on growth and resource use of co-occurring C3 and C4 annuals. 73, 1244-1259. Coleman, J. S., McConnaughay, K. D. M., and Bazzaz, F. A. (1993). Elevated CO2 and plant nitrogen-use--Is reduced tissue nitrogen concentration size-dependent? 93, 195-200. Collins, W. B. (1976). Effect of carbon dioxide enrichment on growth of the potato plant. 11, 467-469. Condon, M. A., Sasek, T. W., and Strain, B. R. (1992). Allocation patterns in two tropical vines in response to increased atmospheric CO2. 6, 680-685. Conroy, J. P., Milhat, P.J., and Barlow, E. W. R. (1992). Effect of nitrogen and phosphorus availability on the growth response of to high CO2. 15, 843-847. Cruz, C., Lips, S. H., and Martins-Lougao, M. A. (1993). The effect of nitrogen source on photosynthesis of carob at high CO2 concentrations. 89, 552-556. Cui, M., and Nobel, P. S. (1994). Gas-exchange and growth responses to elevated CO~ and light levels in the CAM species 17, 935-944. Cui, M., Miller, P. M., and Nobel, P. S. (1993). COz exchange and growth of the crassulacean acid metabolism plant under elevated CO2 in open-top chambers. 103, 519-524. Curtis, P. S., and Teeri, J. A. (1992). Seasonal responses of leaf gas-exchange to elevated carbon dioxide in 22, 1320-1325. Curtis, P. S., Vogel, C. S., Pregitzer, K. S., Zak, D. R., and Teeri, J. A. (1995). Interacting effects of soil fertility and atmospheric CO2 on leaf area growth and carbon gain physiology x (Dode) Guinier. 129, 253-263. in
408 DeLucia, E. H., Callaway, R. M., and Schlesinger, W. H. (1994). Offsetting changes in biomass allocation and photosynthesis in ponderosa pine in response to climate change. 14, 669-677. Den Hertog, J., Stulen, I., and Lambers, H. (1993). Assimilation, respiration and allocation of carbon in affected by atmospheric CO2 levels. 104/105, 369-378. Dippery, J. K., Tissue, D. T., Thomas, R. B., and Strain, B. R. (1995). Effects of low and elevated CO2 on C3 and C4 annuals. I. Growth and biomass allocation. 101, 13-20. Downton, W. J. S., and Grant, W. J. R. (1994). Photosynthetic and growth responses of variegated ornamental species to elevated COs. 21, 273-279. E1 Kohen, A., and Mousseau, M. (1994). Interactive effects of elevated COs and mineral nutrition on growth and COs exchange of sweet chestnut (Castanea 14, 679-690. E1 Kohen, A., Rouhier, H., and Mousseau, M. (1992). Changes in dry weight and nitrogen partitioning induced by elevated COs depend on soil nutrient availability in sweet chestnut Mill.). 49, 83-90. E1Kohen, A., Venet, L., and Mousseau, M. (1993). Growth and photosynthesis of two deciduous forest species at elevated carbon dioxide. 7, 480-486. Ferris, R., and Taylor, G. (1993). Contrasting effects of elevated COs on the root and shoot growth of 4 native herbs commonly found in chalk grassland. 125, 855-866. Field, C. B., Chapin, F. S., Matson, P. A., and Mooney, H. A. (1992). Responses of terrestrial ecosystems to the changing atmosphere: A resource-based approach. 23, 201-235. Gamier, E. (1992). Growth analysis of congeneric annual and perennial grass species. 80, 665-675. Gifford, R. M. (1994). The global carbon cycle: A viewpoint on the missing sink. 21, Gloser, J., and Bartak, M. (1994). Net photosynthesis, growth rate and biomass allocation in a rhizomatous grass grown at elevated COs concentration. 30, 143-150. Gorissen, A., Kuikman, P.J., and Van de Beek, H. (1995) Carbon allocation and water use in juvenile Douglas fir under elevated COs. 129, 253-263. Griffin, K. L., Thomas, R. B., and Strain, B. R. (1993). Effects of nitrogen supply and elevated 95, carbon dioxide on construction cost in leaves of (L.) seedlings. 575-58O. Grime, J. P. (1979). "Plant Strategies and Vegetation Processes." Wiley, Chichester. Grime, J. P., Hodgson, J. G., and Hunt, R. (1988). "Comparative Plant Ecology: A Functional Approach to Common British Species." Unwin Hyman, London. Guehl, J. M., Picon, C., Aussenac, G., and Gross, P. (1994). Interactive effects of elevated COs and soil drought on growth and transpiration efficiency and its determinates in two European forest tree species. 14, 707-724. Hocking, P.J., and Meyer, C. P. (1985). Responses of noogoora burr Bertol.) to nitrogen supply and carbon dioxide enrichment. 55, 835-844. Hocking, P.J., and Meyer, C. P. (1991). Effects of COs enrichment and nitrogen stress on growth, and partitioning of dry matter and nitrogen in wheat and maize. 18, 339-356. Hunt, R., Hand, D., Hannah, M. A., and Neal, A. M. (1991). Response to COs enrichment in 27 herbaceous species. 5, 410-421. Hunt, R., Hand, D. W., Hannah, M. A,, and Neal, A.M. (1993). Further response to COs enrichment in British herbaceous species. 7, 661-668. 9 Hurd, R. G., and Thornley, J. H . M. (1974). An analysis of the growth of young tomato plants in water culture at different light integrals and COs concentrations. I: Physiological aspects. . . . . .... 38, 375-388.
25.
409
Johnson, D. W., Ball, T., and Walker, R. F. (1995). Effects of elevated CO2 and nitrogen on nutrient uptake in ponderosa pine seedlings. 168/169, 535-545. Johnson, K. H. (1993). Growth and ecophysiological responses of Black Spruce seedlings to elevated CO2 under varied water and nutrient additions. 23, 1033-1042. Johnson, R. H., and Lincoln, D. E. (1991). Sage brush carbon allocation patterns and grasshopper nutrition: The influence of CO2 enrichment and soil mineral nutrition. 87, 127-134. Kaushal, P., Guehl,J. M., and Aussenac, G. (1989). Differential growth response to atmospheric carbon dioxide enrichment in seedlings of and ssp. Laricio var. Corsicana. J. For. Res. 19, 1351-1358. Kerstiens, G., and Hawes, C. V. (1994). Response of growth and carbon allocation to elevated CO2 in young cherry (Pmnus L.) saplings in relation to root environment. 128, 607-614. K6rner, C. (1993a). Scaling from species to vegetation: The usefulness of functional groups. "Biodiversity and Ecosystem Functioning" (E. D. Schulze, and H. A. Mooney, eds.), pp 118-140. Springer-Verlag, Berlin. K6rner, C. (1993b). CO2 fertilization: The great uncertainty in future vegetation development. "Vegetation Dynamics and Global Change" (A. M. Solomon and H. H. Shugart, eds.), pp. 53-70. Chapman & Hall, New York. Lambers, H., and Poorter, H. (1992). Inherent variation in growth rate between higher plants: A search for physiological causes and ecological consequences. 23, 188-261. Larigauderie, A., Reynolds, J. F., and Strain, B. R. (1994). Root response to CO2 enrichment and nitrogen supply in loblolly pine. 165, 21-32. Lee, H. S.J., Muray, M., Evans, L., Pettersson, R., Leith, I., Barton, C. V. N., and Jarvis, P. G. (1994). Effects of elevated CO2 on Sitka spruce seedlings. 104/105, 458-459. Lenssen, G. M. (1993). "Response of C3 and C4 Species from Dutch Salt Marshes to Atmospheric CO2 Enrichment." Thesis, Free University, Amsterdam. Lenssen, G. M., Lamers, J., Stroetenga, M., and Rozema, J. (1993). Interactive effects of atmospheric COz enrichment, salinity and flooding on growth of C3 and C4 salt marsh species. 104/105, 379-388. Lewis, J. D., Thomas, R. B., and Strain, B. R. (1994). Effect of elevated CO2 on mycorrhizal colonization of loblolly pine (Pinus L.) seedlings. 165, 81-88. Lindroth, R. L., Kinney, K. K., and Platz, C. L. (1993). Responses of deciduous trees to elevated atmospheric COz~Productivity, phytochemistry, and insect performance. 74, 763-777. McConnaughay, K. D. M., Berntson, G. M., and Bazzaz, F. A. (1993). Limitations to CO2 induced growth enhancement in pot studies. 94, 550-557. Miao, S. L., Wayne, P. M., and Bazzaz, F. A. (1992). Elevated CO2 differentially alters the responses of co-occurring birch and maple seedlings to a moisture gradient. 90, 300-304. Morgan, J. A., Knight, W. G., Dudley, L. M., and Hunt, H. W. (1994). Enhanced root system C-sink activity, water relations and aspects of nutrient acquisition in mycotrophic subjected to CO2 enrichment. 165, 139-146. Mortensen, L. M. (1985). Nitrogen oxides produced during CO2 enrichment. II. Effects of different tomato and lettuce cultivars. 101, 411-415. Mortensen, L. M. (1994a). The influence of carbon dioxide or ozone concentration on growth and assimilate partitioning in seedlings of nine conifers. 44, 157-163. Mortensen, L. M. (1994b). Effects of carbon dioxide concentration on assimilate partitioning, photosynthesis and transpiration of Roth. and (L.) Karst. seedlings at two temperatures. 44, 164-169.
410 Musgrave, M. E., and Strain, B. R. (1988). Response of two wheat cultivars to CO2 enrichment under subambient oxygen conditions. 87, 346-350. Nicolas, M. E., Munns, R., Samarakoon, A. B., and Gifford, R. M. (1993). Elevated CO2 improves the growth of wheat under salinity. 20, 349-360. Nobel, P. S., Cui, M. Y., Miller, P. M., and Luo, Y. Q. (1994). Influences of soil volume and an elevated CO2 level on growth and CO2 exchange for the crassulacean acid metabolism plant 90, 173-180. Norby, R.J., and O'Neill, E. G. (1991). Leaf area compensation and nutrient interactions in CO2-enriched seedlings of yellow poplar (Liriodendron L.). 117, 515-528. Norby, R. J., Gunderson, C. A., Wullschleger, S. D., O'Neill, E. G., and McCracken, M. K. (1992). Productivity and compensatory responses of yellow poplar trees in elevated CO~. 357, 322-324. Overdieck, D., Reid, C., and Strain, B. R. (1988). The effects of preindustrial and future CO2 concentration on growth, dry matter production and the C/N relationship in plants at low nutrient supply: (cowpea), (okra) and (radish). 62, 119-134. Pettersson, R., and McDonald, A.J.S. (1992). Effects of elevated carbon dioxide concentration on photosynthesis and growth of small birch plants Roth.) at optimal nutrition. 15, 911-919. Pettersson, R., McDonald, A.J.s., and Stadenberg, I. (1993). Response of small birch plants Roth) to elevated CO2 and nitrogen supply. Poorter, H. (1989). Plant growth analysis: Towards a synthesis of the classical and the functional approach. 75, 237-244. Poorter, H. (1993). Interspecific variation in the growth response of plants to an elevated CO2 concentration. 104/105, 77-97. Poorter, H., and Remkes, C. (1990). Leaf area ratio and net assimilation rate of 24 wild species 83, 553-559. differing in relative growth rate. Poorter, H., Remkes, C., and Lambers, H. (1990). Carbon and nitrogen economy of 24 wild species differing in relative growth rate. 94, 621-627. Poorter, H., Gifford, R. M., Kriedemann, P. E., and Wong, S. C. (1992). A quantitative analysis of dark respiration and carbon content as factors in the growth response of plants to elevated CO2. 40, 501-513. Potvin, C., and Strain, B. R. (1985). Effects of CO2 enrichment and temperature on growth of two C4 weeds, and 63, 1495-1499. Radoglou, K. M., and Jarvis, P. G. (1992). The effects of CO2 enrichment and nutrient supply on growth, morphology, and anatomy of L. seedlings. 70, 245-256. Radoglou, K. M., and Jarvis, P. G. (1993). Effects of atmospheric CO2 enrichment on early growth of a plant with large cotyledons. 16, 93-98. Reid, C. D., and Strain, B. R. (1994). Effects of CO2 enrichment on whole-plant carbon budget of seedlings of and 98, 31-39. Retuerto, R., and Woodward, F. I. (1993). The influence of increased CO2 and water supply on growth, biomass allocation, and water-use efficiency of L. grown under different wind speeds. 94, 415-427. Rochefort, L., and Bazzaz, F. A. (1992). Growth response to elevated CO2 in seedlings of 4 co-occurring birch species. 22, 1583-1587. Rogers, H. H., Peterson, C. M., McCrimmon, J. N., and Cure,J. D. (1992). Response of plant roots to elevated atmospheric carbon dioxide. 15, 749-752. Rouhier, H., Billes, G., E1Kohen, A., Mousseau, M., and Bottner, P. (1994). Effect of elevated CO~ on carbon and nitrogen distribution within a tree (Castanea Mill) soil system. 162, 281-292.
25.
411
Roumet, C., Bel, M. P., Jardon, F., Salager, J. L., and Roy, J. (1993). Diversite de la response de graminees ~tune augmentation de la teneur en CO2 atmospherique. 18, 35-44. Rozema, J. (1993). Plant responses to atmospheric carbon dioxide enrichment: Interactions with some soil and atmospheric conditions. 104/105, 173-190. Rufty, T. W., Thomas, R. B., Cure, J. D., and Cure, W. W. (1994). Growth response of cotton 91, 503-509. to CO2 enrichment in differing light environments. Ryle, G.J.A., Powell, C. E., and Davidson, I. A. (1992a). Growth of white clover, dependent on Nz fixation, in elevated CO2 and temperature. 70, 221-228. Ryle, G.J.A., Powell, C. E., and Tewson, V. (1992b). Effect of elevated CO2 on the photosynthesis, respiration and growth of perennial ryegrass. J. 43, 811-818. Sage, R. F. (1994). Acclimation of photosynthesis to increasing atmospheric COz: The gas exchange perspective. 39, 351-358. Samuelson, L. J., and Seiler, J. R. (1992). Fraser fir seedling gas exchange and growth in response to elevated CO2. 32, 351-356. Samuelson, L.J., and Seiler, J. R. (1993). Interactive role of elevated CO2, nutrient limitations, and water stress in the growth responses of Red Spruce seedlings. For. Sc/. 39, 348-358. Samuelson, L.J., and Seiler, J. R. (1994). Red Spruce seedling gas exchange in response to elevated CO~, water stress and soil fertility treatments. 24, 954-959. SeegmCdler, S., and Rennenberg, H. (1994). Interactive effects of mycorrhization and elevated carbon dioxide on growth of young pedunculate oak L.) trees. 167, 325-329. Sicher, R. C., Kremer, D. F., and Rodermel, S. R. (1994). Photosynthetic acclimation to elevated COz occurs in transformed tobacco with decreased ribulose-l,5-bisphosphate carboxylase/ oxygenase content. 104, 409-415. x grown at Silvola, J., and Ahlholm, U. (1992). Photosynthesis on willows different CO2 concentrations and fertilization levels. 91, 208-213. Silvola,J., and Ahlholm, U. (1993). Effects of CO2 concentration and nutrient status on growth, growth rhythm and biomass partitioning in a willow, 67, 227-234. Silvola, J., and Ahlholm, U. (1995). Combined effects of CO2 concentration and nutrient status on the biomass production and nutrient uptake of birch seedlings (Betula 168/169, 547-553. Stanghellini, C., and Bunce, J. A. (!993). Response of photosynthesis and conductance to light, CO2, temperature and humidity in tomato plants acclimated to ambient and elevated 29, 487-497. CO2. Stewart, J. D., and Hoddinott, J. (1993). Photosynthetic acclimation to elevated atmospheric carbon dioxide and UV irradiation in 88, 493-500. Stulen, I., and Den Hertog, J. (1993). Root growth and functioning under atmospheric COz enrichment. 104/105, 99-115. Thomas, R. B., Reid, C. D., Ybema, R., and Strain, B. R. (1993). Growth and maintenance components of leaf respiration of cotton grown in elevated carbon dioxide partial pressure. 16, 539-546. Thomas, R. B., Lewis, J. D., and Strain, B. R. (1994). Effects of leaf nutrient status and photosynthetic capacity in loblolly pine L.) seedlings grown in elevated atmospheric CO~. 14, 947-960. Tolley, L. C., and Strain, B. R. (1984). Effects of CO2 enrichment on growth of and seedlings under different irradiance levels. 14, 343-350. Tremmel, D. C., and Patterson, D. T. (1993). Responses of soybean and 5 weeds to CO~ enrichment under 2 temperature regimes. 73, 1249-1260. Tschaplinski, T.J., Stewart, D. B., Hanson, P.J., and Norby, R.J. (1995). Interactions between drought and elevated CO2 on growth and gas exchange of seedlings of three deciduous tree species. 129, 63-71.
412 Van der Eerden. L., Dueck, T., and Ptrez-Soba, M. (1993). Influence of air pollution on carbon dioxide effects on plants. "Climate Change: Crops and Terrestrial Ecosystems" (S. C. van de Geijn, J. Goudriaan, and F. Berendse, eds.), pp 59-70. CABO-DLO, Wageningen. Van de Staaij,J. W. M., Lenssen, G. M., Stroetenga, M., and Rozema, J. (1993). The combined effects of elevated CO2 levels and UV-B radiation on growth characteristics ofElymus (= 104/105, 433-439. Vivin, P., Gross, P., Aussenac, G., and Guehl, J. M. (1995). Whole plant CO2 exchange, carbon partitioning and growth in seedlings exposed to elevated COs. 33, 201-211. Wilkins, D., Van Oosten, JJ, & Besford, R. T. (1994) Effects of elevated COs on growth and chloroplast proteins in 14, 769-779. Wong, S. C. (1990). Elevated COs and plant growth. II. Nonstructural carbohydrate content and its effect on growth parameters. 23, 171-180. Wong, S. C. (1993). Interaction between elevated atmospheric concentration of COs and humidity on plant growth: Comparison between cotton and radish. 104/105, 211-221. Wong, S. C., Kriedemann, P. E., and Farquhar, G. D. (1992). COs X nitrogen interaction on seedling growth of four species of 40, 457-472. Wyse, R. (1980). Growth of sugarbeet seedlings in various atmospheres of oxygen and carbon dioxide. 20, 456-458. Yakimchuk, R., and Hoddinott, J. (1994). The influence of ultraviolet-B light and carbon dioxide enrichment on the growth and physiology of seedlings of 3 conifer species. J. For. Res. 24, 1-8. Zak, D. R., Pregitzer, K. S., Curtis, P. S., Teeri, J. A., Fogel, R., and Randlett, D. L. (1993). Elevated atmospheric COs and feedback between carbon and nitrogen cycles. 151, 105-117. Ziska, L. H., and Bunce, J. A. (1993). Inhibition of whole plant respiration by elevated COs as modified by growth temperature. 87, 459-466. Ziska, L. H., and Bunce, J. A. (1994). Increasing growth temperature reduces the stimulatory effect of elevated CO2 on photosynthesis or biomass in 2 perennial species. 91, 183-190. Ziska, L. H., and Teramura, A. H. (1992a). Intraspecific variation in the response of rice to increased CO2--Photosynthetic, biomass and reproductive characteristics. 84, 269-276. Ziska, L. H., and Teramura, A. H. (1992b). COs enhancement of growth and photosynthesis in rice (Oryza Modification by increased ultraviolet-B radiation. 99, 473-481.
2 CO2 Elevation and Canopy Development in Stands of Herbaceous Plants
Atmospheric C O 2 concentrations have strong direct effects on plant physiology and growth, due to changes in photosynthetic carbon assimilation and allocation of assimilate among various components of growth. Studies of plants grown individually under a variety of environmental conditions have contributed much to our understanding of such changes, and have formed the basis for considerable speculation about potential responses of plant communities in a changing climate. However, a variety of experimental studies demonstrate that plant responses to rising CO2 concentrations may be altered considerably when plants are growing in monospecific stands or mixed species communities (Bazzaz and McConnaughay, 1992). One of the direct consequences of growth in dense stands is the increased intensity of competition for light. Under such conditions, the effects of changing CO2 concentrations (or levels of other resources) may be critically mediated .by alterations in canopy structure and the costs and benefits of leaf placement in different strata of the community. For example, Reekie and Bazzaz (1989) studied seedlings of five tropical tree species growing in experimental stands at different CO2 concentrations, and found that shifts in competitive success were best explained by alterations in canopy architecture and the deployment of leaves at different canopy heights. Changes in the location of canopy leaves among neighboring individuals will have significant effects on competitive interactions and may influence and Communities
413
Copyright 9 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.
414
the biological diversity of natural communities. Additionally, at the ecosystem level, models of plant productivity are critically dependent on the distribution of leaf area, and consequent light interception, across communities and the potential changes in these parameters are fundamental to an understanding of the natural carbon cycle (Field, 1991). Understanding how species interact in elevated CO2 environments and other global change conditions is critical to evaluate the potential impact of global change on populations and communities. In this chapter, we address three questions regarding canopy structure in elevated CO2 communities: (i) What is the relative efficiency of light capture by dominant and subordinate members of a plant community? (ii) Is the efficiency of light capture in experimental stands of competing species modified by atmospheric CO2 concentrations? and (iii) Based on a theoretical model of nitrogen allocation and photosynthesis in relation to the attenuation of light through a canopy, is the optimal leaf area index of a plant stand modified by atmospheric CO2 concentrations?
Natural plant communities are composed of a variety of species with a broad range of plant heights and abundances. The variation in plant height, and in the placement of leaf area, has important implications for the capture of light (photosynthetic photon flux density, PPFD), the most important energy source for plant growth. Leaves in the upper layer of the canopy overshade the leaves in the lower layer and thus a gradient of light climate develops within a canopy (Monsi and Saeki, 1953). When a plant community is composed of a range of species with different plant heights, a hierarchy of individuals and species can be established in the canopy with respect to the availability of PPFD (Weaver and Clements, 1929; Keddy and Shipley, 1989; Weiner, 1990). Species that are able to grow tall can place their leaves in upper layers of the canopy. Such species intercept greater fractions of available PPFD and dominate in the stand. Shorter, subordinate species occupy lower layers in the canopy and receive smaller fractions of the available PPFD, nevertheless coexisting with taller species. Thus dominant and subordinate species in a plant community can be differentiated along this light gradient (Grime, 1987). What mechanisms are involved in the coexistence of these species in a plant community? Hirose and Werger (1995) presented a model, based on leaf area distribution and the cost and efficiency of light capture, to analyze canopy structure in multispecies communities and applied it to an herbaceous plant community on a floating fen.
26.
415
A. T h e M o d e l
Attenuation of PPFD through the canopy is assumed to follow Beer's law (Monsi and Saeki, 1953): I = I0 e x p ( - / ~ ,
(1)
where Fis the cumulative leaf area index (LAI) from the top of the canopy, I0 and I are PPFD on a horizontal level above the canopy and within the canopy at depth F, respectively, and K is the coefficient of light extinction. PPFD intercepted by the leaves of species i in the jth layer in the canopy (~bij) is given by (Aj~/[~;A~),
(2)
where -A/j is the a m o u n t of PPFD absorbed in layer j and Aj~ is the leaf area of species/in layer j. Since = A/JAFj and from Eq. (1), A / j = -K/0 exp(-KFj), Eq. (2) is rewritten as ~bij= K/0 e x p ( - / ~ j )
9Afj,
(3)
where Fj is the cumulative LAI at layer j. Thus ~bijcan be determined from K and the distribution of leaf area of each species in the canopy. Total PPFD absorbed by species i (~) is given by = ]Lj~bii.
(4)
Because photons are intercepted by leaves, a positive correlation is expected between photon absorption and leaf area. A power equation is fitted to the relationship between total photon absorption (~) and leaf area (A): b,
(5)
where 9 and A are defined for each species in a stand and a and b are positive constants. To achieve high photon absorption (~), plants need not only to develop a large leaf area, but also have to place their leaves at relatively higher positions in the canopy, or in more open locations to avoid direct shading from above; such plants may invest a large fraction of biomass in supporting tissues such as stems, petioles, or heavier leaf venation. Then we expect a positive correlation between 9 and the aboveground biomass (M) as well: d,
(6)
where cand dare positive constants. Dividing both sides of Eq. (5) by A gives (I)area --" ~ / A
= aA b-1.
(7)
(I)area is the photon flux absorbed per unit leaf area which is defined for each species. Likewise, dividing both sides of Eq. (6) by M gives
416
(8)
(I)mass :
(I)mass is the photon flux absorbed per unit aboveground biomass and defined for each species. IfMis considered as the investment cost to absorb photons, is the benefit gained for that investment. T h e n (I)mass as a ratio of the benefit to the cost indicates an efficiency of aboveground biomass use to absorb photons. The following relationship holds between (I)mass and ~area: (IDmass --" ALAR"
(I)area ,
(9)
where ALAR (aboveground leaf area ratio) is the ratio of leaf area per unit a m o u n t of aboveground dry mass (not total dry mass as in conventional growth analysis). ALAR is further analyzed as: AIAR = A L M R . SLA,
(10)
where ALMR (aboveground leaf mass ratio) is the ratio of leaf dry mass to the aboveground dry mass and SLA (specific leaf area) is the leaf area per unit leaf dry mass. Hirose and Werger (1995) hypothesized that tall dominant species have higher ~area than shorter subordinate species because the former develops their foliage in the upper layers of the canopy. They further hypothesized that (I)mass of the tall dominant species is not necessarily higher than that of short subordinate species because the former should invest more biomass to supporting tissues.
B. Partitioning of Photons among Species in a Plant Community The above hypothesis was tested in a tall herbaceous plant community (Thelypterido-Phragmitetum) on a floating fen at the time of peak standing crop (Hirose and Werger, 1995). There were 11 species coexisting in an area of 0.5 X 1 m. Total aboveground standing dry mass was 407 g / m 2. Three taller dominant species, and accounted for 94.6% in dry mass and 8 subordinate species for the rest of 5.4%. The green LAI was 3.42, of which the 3 tall dominant species accounted for 93.4% and 8 subordinate species for the rest of 6.6%. Ofincident PPFD, 77.5% was intercepted by the photosynthetic tissue of the canopy and 8.0% by the dead leaves. The rest (14.5 %) reached ground level. The 3 dominant species intercepted 75% of the incident PPFD and the 8 subordinate together 2.5%. Power equations fitted to the relationships between photon absorption (@) and leaf area (A) over 11 species and between photon absorption (@) and total aboveground biomass (M) were: @ = 0.40 A H9 (r ~ = 0.988) and @ = 0.74 34TM (r 2 = 0.968). Photon absorption increased more than proportionately with increasing leaf area (b > 1) and less than proportionately with increasing biomass (d < 1, though not significant). Dividing both sides of these two equations by A and M, respectively, gave:
26. (~area -- 0 . 4 0
417 A~
(I)mass --" 0.74 M -~176
(I)area increased significantly with increasing A (P < 0.01; Fig. la), while (I)mass t e n d e d to decrease with M (not significant; P > 0.1; Fig. lb). The a m o u n t of PPFD absorbed per unit leaf area ((I)area) was large in tall, d o m i n a n t species high in the canopy, whereas the a m o u n t of photons absorbed per unit aboveground mass ((I)mass) was not higher in those species. (I)mass is a p r o d u c t of ALAR and (I)area (Eq. 9). There was a trade-off relationship between ALAR and (I)area , leading to relatively constant values of ~ma~s. Similar tradeoffs were r e p o r t e d by Givnish (1982), who showed that taller species, which invest m o r e resources in supporting tissues to remain mechanically stable, display lower proportional allocation to foliage. To obtain a high p h o t o n flux per unit leaf area, plants should place their leaves at higher positions in the canopy. Such plants have to invest a large fraction ofbiomass in supporting tissues. Tall species a p p e a r e d to have an advantage over subordinate species in receiving a large fraction of incident PPFD, whereas subordinate species have an advantage in efficiently using their biomass to capture PPFD. Both ALMR and SLA contributed to larger ALAR (see Eq. 10) of subordinate species, though the contribution of SLA was m u c h larger. Light use efficiency (photosynthetic return per unit a m o u n t of absorbed PPFD) d e p e n d s both on species and on the level of PPFD
1.0
0.1 0.01
01
10 Leaf area
100
0.1
1
10
100
Total mass
1 Relationship (a) between photon absorption per unit leaf area (ordinate) and leaf area (abscissa) and (b) between photon absorption per unit aboveground dry mass (ordinate) and aboveground dry mass (abscissa) for 11 species in the canopy of the Thelypterido-Phragmitetum, Both axes are in relative values (total lead area, 100; total aboveground dry mass, 100, and total photon absorption, 100). (Redrawn from Hirose and Werger, 1995, with permission). Figure
418
absorbed (Bj6rkman, 1982). If higher light use efficiencies at lower levels of PPFD are taken into account, the efficiency of biomass use for dry mass production is even higher in subordinate species. However, we may hypothesize that species maintaining a certain level of (I)m~sis a necessary condition for their coexistence in a plant community.
Canopy consequences of CO2 elevation are difficult to predict, because component species respond to CO2 elevation differently from each other and competition may change their response to CO2 elevation (Bazzaz, 1990). If taller dominant species expand relatively more leaf area under elevated CO2 conditions, (I)area of shorter subordinate species would be reduced. Consequently, CI)massof the latter species would decrease and would not maintain themselves in the canopy unless their ALAR would increase with COs elevation to counterbalance the decrease in (I)area. Using two co-occurring annual species, and we carried out an experiment to see if COs elevation alters canopy development (Hirose unpublished). Both species were from disturbed environments in the American Midwest. These species were selected because they both produce few lateral branches, while having different distinct patterns of leaf display along the primary axis; leaves of increase in size at upper nodes, whereas those of decrease prior to production of a terminal inflorescence. We established mono- and mixed stands of these species at two growth stages (39 and 53 days since planting) under ambient (350/~1/1) and elevated (700/xl/1) COs conditions in the glasshouse with natural light conditions. Temperature was maintained at 25~ day and 20~ night. Lighting was supplemented by halogen bulbs. developed leaf area faster but stopped development earlier than did and no significant difference was found in leaf area between the two species at the later growth stage (Fig. 2a). Neither COs nor interference had significant effects on leaf area development. Averaged over the two COs levels, stand LAI increased from 0.76 to 2.29 in the stand, from 1.15 to 2.16 in the stand, and from 1.02 to 2.25 in the mixed stand. distributed its leaves to lower layers, whereas developed its leaves mainly in upper layers. Although and had similar leaf areas and plant heights in the mixed stand, intercepted a much larger fraction of available photons than due to the difference in leaf placement (Fig. 2b). CO2 elevation had little effect on photon absorption. Averaged over the COs levels, 57, 69, and 68% of incident PPFD were intercepted on Day 39, and 83, 89, and 86% were
26.
419
(a) Leaf area per plant, (b) photon absorption per plant ( in relative units), (c) photon absorption per unit leaf area, and (d) aboveground dry mass in response to CO2 for mono- and mixed stands of and on Days 39 and 53 after emergence. Error bars show _+1 SE of the mean (for details of experimental methods and calculation of photon absorption see Hirose manuscript in preparation).
i n t e r c e p t e d o n Day 53 by and monospecific stands a n d the m i x e d stand, respectively. P h o t o n a b s o r p t i o n p e r unit leaf area ( ~ .... ) d e c r e a s e d with growth (Fig. 2c), reflecting increasing m u t u a l shading particularly w h e n a m o n g leaves. h a d h i g h e r (I~area than the two species were grown in mixture. T h e r e was a significant effect of CO2 elevation o n the a b o v e g r o u n d dry mass (Fig. 2d). O n average, biomass increased with CO2 elevation by 6 - 2 9 % at the y o u n g e r stage a n d by 2 9 - 3 6 % at the later growth stage. Because t h e r e was n o significant difference in leaf area d e v e l o p m e n t a n d light i n t e r c e p t i o n u n d e r two CO2 conditions (Figs. 2 a - 2 c ) , this increase in biomass should be ascribed to h i g h e r n e t assimilation rates (biomass production p e r unit leaf area) u n d e r the elevated CO2 c o n c e n t r a t i o n . This result is s o m e w h a t different f r o m earlier studies. T h e biomass increase in the
420 present study was comparable to 33-37% increase in yield due to doubling CO2 averaged from diverse crops and wild species (Kimball, 1983; Cure and Acock, 1986; Poorter, 1993; Poorter Chapter 25). However, earlier studies showed that the increase in biomass was attributable to increased leaf area as well as increased photosynthetic activity of leaves. Coleman and Bazzaz (1992), studying the effect of COs and temperature on growth of two annuals, and suggested that the effects of CO2 elevation on growth were primarily due to changes in leaf area production and loss, and to a lesser degree to effects on photosynthesis, nitrogen, and water use efficiency. In competition, however, exhibited a relative enhancement in performance at elevated CO2, apparently due to increased net assimilation rates late in growth (Bazzaz 1989). K6rner and Arnone (1992) also found no significant difference in leaf area development in artificial tropical ecosystems established in glasshouses under ambient and elevated COs conditions. These results suggest that enhancement of net assimilation rate may be a particularly important component of COs responses in competitive conditions, where plant-plant interference and self-shading may constrain the deployment of leaf area. However, this does not contradict the importance of leaf display (e.g., Reekie and Bazzaz, 1989), because net assimilation rates depend on both the amount of light intercepted (i.e., the relative positions of leaves in the canopy), and the efficiency of light utilization in photosynthesis. In the last section, we approach this problem from a different perspective using an optimality model for stand level carbon gain to assess the potential COs responses of leaf area index in dense stands.
The effects of elevated C O 2 o n carbon gain in a dense stand depend on the changes in leaf photosynthetic parameters throughout the canopy. At the leaf level, the quantum yield of photosynthesis under light-limiting conditions is higher in plants grown in elevated COs atmospheres (Long and Drake, 1991). The increase in quantum yield lowers the light compensation point of instantaneous and diurnal carbon assimilation in elevated COs. Based on this leaf level response, an increase in leaf area index might be expected in closed stands, because additional leaves with positive carbon balance could be maintained in the shaded, lower canopy layers. However, this prediction does not take into account the distribution of leaf nitrogen and photosynthetic potential through the canopy. In particular, COsinduced changes in rates of photosynthesis a n d respiration will alter the relative value of leaves in different canopy layers. Here we investigate a
26.
421
model that incorporates potential changes in leaf photosynthetic parameters due to elevated CO2 in relation to the allocation of nitrogen through the canopy. The model determines the joint optimum of the number of leaf layers (leaf area index) and the distribution of nitrogen among those layers that maximizes stand level carbon gain, given a fixed supply of nitrogen for the entire stand. The model was parameterized using data from our experiment on and (see Section III above), supplemented with information obtained from the literature as necessary. Preliminary analysis of the model presented below suggests that optimal leaf area index is only slightly higher in elevated CO2, though total carbon gain is greatly enhanced. Furthermore, the predictions of the model appear to be more sensitive to changes in respiration rates, relative to leaf nitrogen and CO2 concentration than to the small changes in quantum yield observed in elevated CO2 atmospheres. Carbon gain at the whole plant or stand level depends on the number of leaves or leaf layers, and the distribution of incident light availability and photosynthetic potential (i.e., leaf nitrogen) among these leaves. Daily leaf-level nitrogen use efficiency of photosynthesis (PNUE, diurnal carbon gain per unit leaf nitrogen) is higher in high-light environments, but it declines as leaf nitrogen increases. For a plant with a given number of leaves in a range of light environments, total carbon gain is maximized when the marginal returns in carbon gain relative to nitrogen concentration in any particular leaf are equal for all leaves (Field, 1983). As a result, carbon gain is maximized when leaf nitrogen concentrations are higher in leaves in high light conditions; observed nitrogen allocation patterns closely parallel these predicted optima, though the mechanisms underlying these distributions are still not understood (Field, 1983; Hirose and Werger, 1987; Traw and Ackerly, 1995; see Chen 1993 for an alternative approach to nitrogen distribution patterns). Previous analyses of optimal nitrogen allocation have assumed that the number and size of leaves are fixed; if these are allowed to vary, the optimal nitrogen distribution and maximum attainable carbon gain will change correspondingly, and a joint optimum may be found that maximizes shoot level carbon gain in relation to all of these parameters. Formally, the joint optimum is the set of values of leaf nitrogen concentration (N, per unit area), leaf number (L), and individual leaf area (m) that maximize canopy level carbon gain and satisfy two constraints:
dN~
= m,c~
~m~N~ = NT, i=1
(11)
(12)
422
where c~ is a constant and is the total nitrogen in all leaves. In contrast to Field (1983), Eq. (11) includes leaf size on the right-hand side because we are considering leaf nitrogen concentration per unit area (not per leaf) and the reallocation of a fixed a m o u n t of nitrogen will alter Nconcentration in two leaves in proportion to their relative sizes. When both leaf size and n u m b e r are allowed to vary, the optimal (and somewhat trivial) solution for a plant is to produce a single large leaf in the uppermost canopy position with a leaf nitrogen concentration that maximizes leaf level PNUE. If leaf n u m b e r (L) is held constant, the model predicts that optimal leaf size will be larger in response to low ambient light levels or increased nitrogen availability (Ackerly, 1993). The result relative to light availability is of interest as it suggests that the decrease in leaf size in high light may enhance nitrogen use efficiency, in addition to improved regulation of leaf energy balance and water loss. Here, we investigate the question of optimal leaf area index by specifying a total nitrogen supply per unit of ground area (g/mS), setting a fixed leaf size of 0.1 m s, and solving for values of c~ and L that maximize carbon gain; leaf area index is then calculated as Following Hirose and Werger (1987), we utilized a nonrectangular hyperbola to characterize the light response of photosynthesis. The nitrogen dependence of lightsaturated photosynthetic capacity and dark respiration rates were based on data obtained for in ambient and elevated COs concentrations (Hirose unpublished data). The nitrogen dependence of quantum yield (0) and the curvature parameter (0) were based on an earlier study of (Hirose and Werger, 1987). For elevated COs, the intercept of the regression of quantum yield on leaf nitrogen concentration was increased by 0.01 u n d e r elevated COs (cf. Long and Drake, 1991). These equations are summarized in Table I. Midday PPFD levels at the top of the canopy were set at 1800/zmol m -s s -1 and the daily course of PPFD was modeled following a sine square curve (Hirose and Werger, 1987). Light attenuation in successive layers of the canopy followed Eq. 1, with k = 0.7 for both ambient and elevated COs concentrations (Hirose unpublished data). Leaf-level daily carbon assimilation was calculated for light environments corresponding to successive leaf layers u n d e r ambient or elevated COs concentrations and for a range of leaf nitrogen concentrations from 0.6 to 4 g N/mS; values of calculated numerically for use in Eq. (11). We analyzed this model to determine optimal leaf area index and nitrogen allocation, in relation to the total nitrogen supply in the canopy. This comparison eliminates differences between ambient and elevated COs stands due to differences in overall growth rates (cf. Coleman 1994), and standardizes the comparisons relative to the total amount of nitrogen available for deployment in the entire canopy. Optimal leaf area index and
26.
Regression on leaf nitrogen concentration ( g / m 2) Parameters (/zmol m -2 sec -1) Pm~ Pm~ Rd P~ 4) 4) 0
CO2 level (tzl/liter)
Slope
Intercept
350 700 350 700 350 700 Both
15.44 21.37 1.381 1.333 0.0188 0.0188 -0.251
-- 1.436 - 3.852 --0.141 0.0514 0.0211 0.0311 1.1
Values for Pmaxand R~ were obtained from measurements of (Hirose unpublished data) and values for quantum yield (~b) and the curvature parameter (0) were obtained from Hirose and Werger (1987).
stand-level carbon gain both increase sharply with increases in total foliar nitrogen, while mean leaf nitrogen concentrations show a less dramatic response (Figs. 3a-3c). As total nitrogen availability increases, optimal LAI is also slightly higher in elevated CO2 environments, and leaf nitrogen concentrations are correspondingly lower (cf. Hilbert 1991). Stand level carbon gain is consistently enhanced by about 45% under elevated CO2 at all levels of nitrogen availability. [Oikawa (1986) obtained a similar result in a model of forest productivity, predicting an increase in optimal leaf area index from about 6.3 to 6.6, in response to a doubling of CO2 concentration, while overall productivity increased by over 50%.] The optimal nitrogen allocation gradient, measured as the difference between the upper and lowermost leaves, increases with total nitrogen supply, due to increases in the predicted nitrogen levels of the upper leaves (Fig. 3d). The gradient is less steep in elevated CO2, based on comparisons at similar LAI or total nitrogen availability. Based on the photosynthetic parameters used in the model, the daily light compensation points for the lower leaves in the canopy (with predicted leafN concentrations of 1 g / m 2) are approximately 6 and 5 mol m -2 day -1 for ambient and elevated CO2, respectively. Based on a light extinction coefficient of 0.7, these values correspond to leaf area indices of 2.64 and 2.95 for the two environments. Thus on the basis of light compensation points alone, a higher LA/ is predicted for elevated CO2. However, a greater LA/was also predicted for lower nitrogen supply levels, where the lowermost leaves will be well above their light
424 c 1.8 -1
~-, 1.4
* [
t
0.8.
A- 0.7. ~.~ 0.6. 9
;~,~ 0 " 8 ~ .
~
e
HighCO2
I ~0"3"~,~
0.6
0.2 1
2
!
2
3
b
3
4.0
1.9
r
0.5.
~
1.8 3.0
..,
Top layer
1.7
o o
1.6
9
"3 1.5
~
2.0
,
"3 1.0 "- 1.4 N
Bottom layer
1.3
~ 1
2
3
Total stand leaf nitrogen (g/m2 ground area)
0
2 3 Total stand leaf nitrogen (g/m2 ground area)
(a) Predicted optimal leaf area index (LAI) that maximized canopy level carbon gain, in relation to total canopy nitrogen supply (g/m 2 of ground area) and atmospheric CO2 concentrations (ambient, 350 /~l/liter; elevated, 700 /~l/liter. (b) Predicted canopy level carbon gain at optimal LAI and optimal distribution of nitrogen among canopy layers. Across all nitrogen levels the enhancement due to elevated CO2 is about 45%. (c) Mean leaf nitrogen concentration at optimal LAI. (d) Nitrogen allocation gradient in relation to nitrogen supply, shown as the leaf nitrogen concentrations of the uppermost and lowermost leaf layers in the canopy.
compensation points. Thus the enhancement of LAI under elevated CO2 predicted by the model is due to the maximization of nitrogen use efficiency throughout the canopy, not simply to a change in the light compensation point of the lowest layer. The predictions of this model, regarding optimal leaf area index, were tested by plotting LAI versus total leaf nitrogen for the experimental stands of and grown under two CO2 concentrations, and harvested at two different ages (as described in Section III above). Leaf area
26.
425
index was positively correlated with total nitrogen, except for at elevated CO2, and the observed values were slightly higher than those predicted by the model (Fig. 4). Most importantly, though, there were no significant differences in this relationship between CO2 concentrations. What are the possible explanations for this discrepancy? First, the predictions of the model are very sensitive to the relationship between the photosynthetic light response curves and leaf nitrogen concentration, and to the shifts in that relationship u n d e r elevated CO2. In particular, shifts in the nitrogen dependence of respiration rates may strongly influence the results, as the relative costs of nitrogen in different leaf layers is strongly affected. The response of respiration to elevated CO2 continues to be one of the most poorly understood components of plant carbon balance (Amthor, 1996). For this analysis we were not able to obtain all of the parameters for the light response curves directly from plants in the experimental stands, so it is impossible to say whether the functions we used accurately reflect the changes observed in these stands. We emphasize that this is a preliminary analysis of this model, as there are a large n u m b e r of variables that require additional empirical validation, and further research on physiological responses is critical to achieve more confident predictions regarding stand level responses. Additionally, it is very important to recognize that any tests of optimality models d e p e n d on the validity of the optimality criterion. In this case, we assume that a stand of plants will be structured to maximize stand level carbon gain. Any criterion of this sort is an implicit evolutionary statement
oc 9o
9
0
0
1
2
3
0
1
2
3
4
Total stand leaf nitrogen (g/m2 ground area) Relationship between leaf area index (LAI) and total canopy nitrogen content in (a) and (b) (Hirose unpublished data). Open symbols, 350/zl/liter; solid symbols, 700/~l/liter.
426
with several important components, i.e., (i) maximization of short-term carbon gain is correlated with lifetime fitness; (ii) sufficient genetic variation in the characters underlying carbon gain (e.g., the processes regulating nitrogen allocation and leaf area index) has been available to selection; (iii) selection has been an important factor shaping plant responses under the conditions we are studying (this is particularly problematic when considering responses to changing environmental conditions); and (iv) we have identified the most important causal pathway linking these characters with overall performance, such that the optimality criterion we identify is the strongest selective factor. In addition to these conditions, optimality models applied to stand level properties assume either that the individuals in the stand exhibit optimal behavior, which is manifest as optimal properties at the stand, level as well, or that there has been selection on properties at the stand level (e.g., due to genetic relatedness of neighbors). The latter condition will certainly not be met in mixed-species stands; in this case, game theory models of individual behavior, and the consequences for population properties (e.g., Iwasa 1984), will be a more appropriate theoretical approach to problems such as the optimal leaf area index under different environmental conditions. However, even if these such conditions are not all met, the analysis of this model provides important insights into the sensitivity of carbon gain to different aspects of leaf function and stand structure.
In this chapter we first showed a model to analyze the canopy structure of plant communities with many species (Section II). The model assumed that PPFD attenuates exponentially through the canopy and that PPFD is intercepted by constituent species in proportion to their leaf area. It was applied to an herbaceous plant community, Thelypterido-Phragmitetum, which contained 11 species with different plant heights. Tall species in the canopy received higher PPFD averaged over leaf area ((I)area),whereas PPFD absorbed per unit aboveground biomass (~m~s) of tall species was not higher than that of the subordinate species. This is because tall species invested larger fractions of their biomass to supporting tissues. It was suggested that species maintaining a certain level of (I)mass is a necessary condition for their coexistence in the plant community. To examine if CO2 elevation alters canopy development and competitive interactions, we established mono- and mixed-species stands of Abutilon and at two growth stages under ambient and elevated CO2 conditions (Section III). distributed its leaves to lower layers, while developed its leaves mainly in upper layers. Although and
26.
427
had similar leaf areas in the mixed stand, occupying upper layers in the canopy absorbed a much larger fraction of available photons than At the later growth stage, reduced (I)mass and increased (])mass when they were mixed, indicating a competitive advantage in the latter species. However, the present experiment did not indicate that CO2 elevation significantly changed the competitive relationships of these species. Analysis of a model of optimal leaf area index and nitrogen allocation, in relation to stand level carbon gain, indicated that elevated CO2 levels should lead to only slight enhancements in LAI, compared to stands in ambient CO2 with similar total foliar nitrogen. However, carbon gain of these stands is greatly enhanced due to the large e n h a n c e m e n t in assimilation rates and nitrogen use efficiency. In contrast to these predictions, the stands of and exhibited no effect of CO2 concentration on leaf area index, relative to canopy nitrogen levels. Currently, more detailed studies of photosynthetic responses to light and nitrogen u n d e r ambient and elevated CO2 are needed to improve models of canopy level responses. In particular, predicted canopy responses are quite sensitive to CO2 effects on respiration rates. Several studies suggest that respiration declines in elevated CO2, in parallel with a decline in leaf nitrogen concentrations. However, the effect of CO2 concentration on the relationship of respiration to leaf nitrogen concentrations has received little attention (Amthor, 1996), and is critical in this context. These studies suggest several critical areas of future research on canopy structure and function in relation to global change. First, in order to scale from leaf level processes to stand level carbon gain, we need a much better understanding of the combined effects of light, leaf nitrogen, and atmospheric CO2 on leaf level photosynthesis. Much research has, understandably, focused on CO2 response of photosynthesis (i.e., A-C~ curves) u n d e r nonlimiting light conditions. As we have demonstrated above, the responses of leaf photosynthesis in the shaded, lower layers of the canopy are the critical determinants of potential changes in canopy structure. These shaded leaves also contain less nitrogen, so the d e p e n d e n c e of light curve parameters must be established across a range of leaf nitrogen concentrations. Secondly, if, as we suggest, elevated CO2 enhances biomass accumulation, but has little effect on leaf area index, where does the biomass go? For determinate annuals (such as it is possible that the entire life cycle will be completed more rapidly, leading to phenological shifts in the importance of different species during the growing season (cf. Reekie and Bazzaz, 1991). For indeterminate species, such as forest communities, there may be increased biomass sequestration belowground a n d / o r more rapid turnover of the canopy. The latter possibility would imply a reduction in leaf lifespan and a more rapid turnover of nitrogen through the plant
428
canopy. Will nitrogen be recycled more efficiently within the plant canopy, increasing long-term nitrogen use efficiency, or will it be taken up and returned to the soil more rapidly, leading to short-term depletion of available nitrogen in the soil? The effects of elevated CO2 on canopy functioning must be considered in a broader context in relation to the functioning of the entire ecosystem. Finally, in large, closed-canopy stands (such as forests) changes in the distribution of leaf area may alter the light availability in different strata of the community. If growth of dominant species is enhanced, and leaf area index in the upper canopy increases, understory light levels may be reduced. In a single-species stand this might simply result in the loss of a layer of leaves, or mortality of suppressed individuals. In mixed-species stands, however, a decline in understory light levels could eliminate less shade-tolerant species and reduce biological diversity of the community. Alternatively, increased C O 2 concentrations may allow these species to survive under lower light levels, mitigating these effects. Resolution of these questions will require detailed studies of the effects of atmospheric CO2 concentrations on leaf and whole-plant carbon balance under low light levels. Further study of these issues is particularly important, as the CO2 responses of productive plant communities with high leaf area indices and high biological diversity play a particularly important role in the global carbon budget.
We thank Christian K6rner for comments and discussion that improved the manuscript. Special thanks are also due to Brian Traw and Reiko Yambe for help with experiments and nitrogen analyses. This study was partly supported by a grant-in-aid of the Japan Ministry of Education, Science and Culture to T.H. (06454007) and by US-NSF grants to D.D.A. and F.A.B.
Ackerly, D. D. (1993). "Phenotypic Plasticity and the Scale of Environmental Heterogeneity: Studies of Tropical Pioneer Trees in Variable Light Environments." Ph. D. Thesis, Harvard University, Cambridge, MA. Amthor,J. S. (1996). Plant respiratory responses to elevated CO2 partial pressure. "Advances in Carbon Dioxide Effects Research" (L. H. Allen, Jr., M. B. Kirkham, D. M. Olszyk, and C. Whitman, eds.), in press. American Society of Agronomy, Madison, WI. Bazzaz, F. A. (1990). Response of natural ecosystems to the rising CO2 levels. 21, 167-196. Bazzaz, F. A., and McConnaughay, K. D. M. (1992). Plant-plant interactions in elevated CO2 environments. 40, 547-563. Bazzaz, F. A., Garbutt K., Reekie, E. G., and Williams, W. E. (1989). Using growth analysis to interpret competition between a C3 and a C4 annual under ambient and elevated CO2. 79, 223-235.
26.
429
Bj6rkman, O. (1982). Responses to different quantum flux densities. "Physiological Plant Ecology. I. Encyclopedia of Plant Physiology (O. L. Lange, P. S. Nobel, C. B. Osmond, and H. Ziegler, eds.), Volume 12A, pp. 57-107. Springer-Verlag, Berlin. Chen, J.-L., Reynolds, J. F., Harley, P. C., and Tenhunen, J. D. (1993). Coordination theory of leaf nitrogen distribution in a canopy. 93, 63-69. Coleman, J. S., and Bazzaz, F. A. (1992). Effects of CO2 and temperature on growth and resource use of co-occurring C3 and C4 annuals. 73, 1244-1259. Coleman, J. S., McConnaughay, K. D. M., and Ackerly, D. D. (1994). Interpreting phenotypic variation in plants. 9, 187-191. Cure, J. D., and Acock, B. (1986). Crop responses to carbon dioxide doubling: A literature survey. 38, 127-145. Field, C. B. (1983). Allocating leaf nitrogen for the maximization of carbon gain: Leaf age as a control on the allocation program. 56, 341-347. Field, C. B. (1991). Ecological scaling of carbon gain to stress and resource availability. "Response of Plants to Multiple Stresses" (H. A. Mooney, W. E. Winner, and E.J. Pell, eds.), pp. 35-65. Academic Press, San Diego. Givnish, T.J. (1982). On the adaptive significance of leaf height in forest herbs. 120, 353-381. Grime, J. P. (1987). Dominant and subordinate components of plant communities: Implications for succession, stability, and diversity. "Colonization, Succession and Stability" (A. J. Gray, M.J. Crawley, and P.J. Edwards, eds.), pp. 413-428. Blackwell Sci. Oxford. Hilbert, D. W., Larigauderie, A., and Reynolds, J. F. (1991). The influence of carbon dioxide and daily photon-flux density on optimal leaf nitrogen concentration and root: shoot ratio. 68, 365-376. Hirose, T., and Werger, M.J.A. (1987). Maximizing daily canopy photosynthesis with respect to the leaf nitrogen allocation pattern in the canopy. 72, 520-526. Hirose, T., and Werger, M.J.A. (1995). Canopy structure and photon flux partitioning among species in a herbaceous plant community. 76, 466-474. Iwasa, Y., Cohen, D., and Leon, J. A. (1984). Tree height and crown shape, as results of competitive games. J. 112, 279-297. Keddy, P. A., and Shipley, B. (1989). Competitive hierarchies in herbaceous plant communities. 54, 539-550. Kimball, B. A. (1983). Carbon dioxide and agricultural yield: An assemblage and analysis of 430 prior observations. 75, 779-789. K6rner, Ch., and Arnone, J. A., III (1992). Responses to elevated carbon dioxide in artificial tropical ecosystems. 257, 1672-1675. Long, S. P., Drake, B. G. (1991). Effect of the long-term elevation of CO2 concentration in the field on the quantum yield of photosynthesis of the C3 sedge, 96, 221-226. Monsi, M., and Saeki, T. (1953). l]ber den Lichtfaktor in den Pflanzengesellschaften und seine Bedeutung f~r die Stoffproduktion. 14, 22-52. Oikawa, T. (1986). Simulation of forest carbon dynamics based on a dry-matter production model. III. Effects of increasing CO2 upon a tropical rainforest ecosystem. 99, 419-430. Poorter, H. (1993). Interspecific variation in the growth response of plants to an elevated and ambient CO2 concentration. 104/105, 77-97. Reekie, E. G., and Bazzaz, F. A. (1989). Competition and patterns of resource use among seedlings of five tropical trees grown at ambient and elevated CO2. 79, 212222. Reekie, E. G., and Bazzaz, F. A. (1991). Phenology and growth in four annual species grown in ambient and elevated CO2. 69, 2475-2481.
430 Traw, M. B., and Ackerly, D. D. (1995). Leaf age, light levels and nitrogen allocation in five species of rain forestpioneer trees. 82, 1137-1143. Weaver, J. E., and Clements, F. E. (1929). "Plant Ecology." McGraw-Hill, New York. Weiner, J. (1990). Asymmetric competition in plant populations. 5, 360-364. Woodrow, I. E. (1994). Optimal acclimation of the C3 photosynthetic system under enhanced C02. 39, 401-412.
27 Problems in Predicting the Ecological Effects of Elevated CO
The rising level of atmospheric CO2 is a major global anthropogenic change that we can track and predict with great confidence. We know that atmospheric CO2 has increased, and we can make relatively good predictions of the levels we can expect to see in the near future. But the ecological effects of this rising COz are not as easy to predict, and this is exactly what scientists are being asked to do by policy makers. Indeed, from the policy makers' point of view, the need for predictions of global change is the for research on elevated CO2. How good is our ability to make reasonable predictions, and how can we best improve such predictions? Most would agree that the answer to the first question is " n o t very good" at this point in time. This makes the second question even more important. The difficulty in making predictions of the ecological effects of rising COz levels stems from two basic problems. First, ecology is a young science that does not have a body of widely accepted theory applicable to the questions of global change. There is no short-term solution to this problem, and the longer term solution is to promote the development of the science of ecology. As I will argue later in this chapter, research on the effects of elevated CO2 can perhaps make a significant contribution to this longer-term goal. The second major problem in making reliable predictions about the ecological effects of elevated CO2 is that while scientists are being asked 431
Copyright 9 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.
to make predictions at the community, ecosystem, and biosphere level, most of the available information exists at lower levels (K6rner, 1993). Figure 1 shows the traditional hierarchy of the levels of organization in biology, although the point here would apply to alternative hierarchical schemes (e.g., O'Neill 1986) as well. The predictions that are most needed concerning the effects of elevated CO2 are at the top three levels, but most of our information about the effects of CO2 are from experiments conducted at lower levels. For example, there are numerous experiments looking at the effects of elevated CO2 on leaf-level photosynthesis, whole growth, and development of individual plants (see reviews by Bazzaz [1990], Mooney [ 1991], and Woodward [ 1991] ), but because of costs and logistical constraints, we are just beginning to see longer-term experiments on populations and communities in the field. One solution that has been proposed to this problem of making predictions at higher levels of organization is scaling up from the lower levels (Ehleringer and Field, 1993). But what exactly does "scaling up" mean? I find two very different meanings of this term in recent literature: (1) Extrapolation within one level of organization; or (2) Actual prediction of higher level phenomena using information from a lower level. I will discuss them both in turn.
Extrapolation usually means simply extending a quantitative relationship beyond the range of the data on which the relationship is based. Extrapolation is certainly possible and reasonable in many cases. For example, if one
1 Levelsof organization in biology.
27.
433
could accurately measure NPP in many randomly placed 1-m 2 quadrats within a grassland, one can extrapolate to get a good estimate of NPP over a larger a r e a . T h e issues here are purely practical, not scientific. Extrapolating a specific quantitative relationship entails m u c h more risk, as can be seen clearly in a simple example from population growth. Figure 2 shows a simple computer-generated logistic growth curve with r a n d o m normal noise added. If we have data from only one part of the curve, we would not be able to extrapolate very successfully, because we would not have any information on the overall shape of the relationship If we have information only from the beginning of the curve, we would be inclined to conclude that growth is exponential (geometric growth). If we have information on the central part of the curve, growth would appear to be approximately linear (arithmetic growth). If we have data on the right-hand part of the curve, we would likely conclude that the growth rate is continuously decreasing, such as in a simple saturating function. Extrapolating any of these trends to the other regions of the curve would lead to major errors. One needs either data over the whole range of the relationship, or huge sample sizes to provide the statistical power to see subtle changes in the derivatives over smaller ranges. As another general example of this problem, one can point to the development and use of systems analysis models in ecosystem ecology (Patten, 1983). These "black box" models are often developed to predict some specific ecosystem processes, and calibrated using empirical data. Such models are often pretty good at predicting new combinations of variables within the range of the data used to calibrate them, but these same models are often very poor at making predictions outside the range of the data with which they are calibrated. The reason is similar to the
%~oo
oo
0
O0
0 0
o
o~
oO
0
o
80 Logistic growth curve with random normal variation.
434 example of the logistic growth curve: the characterization of quantitative relationships is usually good enough for interpolating within the range of the calibrating data, but this characterization is not good enough to make predictions far beyond the range of the data. The same issue arises in statistical models. In conclusion, extrapolation can be dangerous, but it is certainly possible and reasonable in some cases. It is important to note, however, that extrapolation does not usually involve a change in the level of organization in question. Rather, it usually refers to questions of scale within one level of organization.
I call the second meaning of scaling up which I found in the literature "reductionism from below," meaning the actual prediction of higher level p h e n o m e n a from lower level information. I call it "from below" because reductionism is usually from above, i.e. it starts with the higher level phenomenon. In scaling up our starting point is information at the lower level. It is my contention here that this type of scaling up generally fails. For example, there is no basis for assuming that responses of a system to an environmental factor at a higher level of organization will be similar to responses at a lower level, and this has been documented in elevated CO2 research (Reynolds and Acock, 1985; Reynolds 1993). Short-term regulatory responses of leaves to elevated CO2 (i.e., an increase in the rate of photosynthesis) does not predict whole-plant biomass accumulation or acclimatory responses (Mooney and Koch, 1994). On the contrary, acclimatory responses often damp out regulatory response (Bazzaz, 1990). Similarly, the performance of plants grown singly at elevated CO2 may not be a good predictor of their performance when competing. (Bazzaz and Garbutt, 1988; Bazzaz and McConnaughay, 1992). Evolutionary responses to elevated CO2, which we have a strong basis to expect (see Chapters 1-5), present the most difficult problems for prediction. There is no basis for assuming that the plastic response of an organism to an environmental factor will be similar or even in the same general direction as evolutionary responses to that same factor. For example, a plant may respond to shade by etiolating, but natural selection may favor slower growth and shorter stature in the shade (as we see in understory herbs). Similarly, Woodward (1987) presented evidence that plants respond to increasing COs by decreasing the number of stomata (although this conclusion has been disputed [K6rner, 1988]). Even if we assume that plants do develop fewer stomata in a COs-enriched environment, I see no reason why we should expect evolutionary responses to be in
27.
435
the same direction. Simply put, evolutionary responses may damp out effects of elevated CO2 (as populations evolve in an environment of elevated CO2), or evolutionary responses may amplify short-term effects, as competitive relationships are altered and species evolve in different ways. There is no way to predict the long-term outcome at this point. The problems of scaling up are not limited to evolutionary change. For example, CO2 is a resource, and there has been significant progress in theories of resource utilization and limitation (Tilman, 1986; Bloom 1985; Chapter 28). These theories could provide some reasonable predictions concerning the effects of elevated CO2 as a resource. Evidence is accumulating, however, that developmental effects of CO2 on plants (e.g., Reekie and Bazzaz, 1991; Loehle, 1995) may be more important than resource-mediated effects. Since elevated CO2 is a novel environment, developmental effects will not be predictable. Several species seem to show increased reproductive output in high CO2 environments, but other species, e.g., show the opposite response (Bazzaz 1995). There is no way to predict with any confidence the response of reproductive output or allocation to elevated CO2 in any species until we do the appropriate experiments. As K6rner (1993) has pointed out, there is no way that we can now, or will be able in the foreseeable future, to predict ecosystem processes from the physiological properties of organisms. To make the philosophical point that p h e n o m e n a at any level of organization are ultimately reducible to and driven by p h e n o m e n a at lower levels, does not mean that we are anywhere near being able to do this. Numerous ecologists (e.g., Allen and Starr, 1982; O'Neill 1986) have argued that pure reductionism in ecology will usually fail. If such reductionism were possible we should all be molecular biologists, or physical chemists, not ecologists. Levin (1993) suggested there may be laws for scaling up, but even if this is so, we are at present very far from discovering and applying them. The paradigm for scaling up in biology has been the use of biochemistry in medicine: antibiotics are molecules that kill bacteria in a test tube, then scale up to cure disease at whole-body level, and then scale up to control epidemics at the population level. However, this type of successful scaling up has proven to be the exception, not the rule. The success of this example of scaling up may be because it takes place mostly within individuals, and the individual is the product of natural selection. Scaling up to supraorganismal levels, such as the community or ecosystem, operates u n d e r no such constraints. Confidence in our ability to scale up depends on the available data and our theoretical understanding of the relationships between the levels of organization over which we are scaling (O'Neill 1986). The important argument against scaling up is not philosophical, but depends on the
available data and state of the art. If, in the scores of experiments that have been done on the effect of elevated CO2, we did observe simple transparent reductionism, it would be quite reasonable to apply this in predicting CO2 effects. For example, if increased CO2 almost always resulted in increased photosynthesis at the leaf level, and increased biomass accumulation at the whole plant, population, and community levels, we would have a strong basis for applying this prediction generally. But CO2 research has not yielded such general and simple patterns. As Bazzaz (1990) has pointed out, competitive outcomes will be modified by CO2 and by the interaction of CO2 with other environmental factors as different species behave differently in a high COz world, and their response will depend on the identity of the competing species. We cannot predict the behavior of a system from a lower level without either (1) evidence for such simple patterns across several levels, or (2) a well-developed theory that spans the levels in question. The data we have does not support (1), and (2) would require a level of ecological theory far beyond what we have available today or in the foreseeable future. The effects of CO2 on terrestrial ecosystem will ultimately be reducible to physiology and interactions among individuals and their environments. But when, as in ecology, we do not have a very good understanding of the processes in question, scaling up from a lower level of organizations is much less reliable than predictions based on data from the same level as the phenomena to be predicted.
In light of this, we can distinguish two basic types of studies on the effects of elevated CO2 on plants. (1) Reductionist experiments which study the mechanisms of CO2 effects; and (2) Holistic experiments which look at CO2 effects on whole systems that are as similar as possible to those about which we are trying to make predictions. These two classes of experiments represent two legitimate, but in many cases fundamentally different, scientific goals. Scientific understanding is ultimately based on reductionism and mechanism, but the best currently accessible predictions of many p h e n o m e n a often come from nonmechanistic, holistic "calculation tools" Loehle, 1983; see also Peters, 1991). To argue that one of these two scientific goals is better or more important than the other misses the p o i n t - - t h e y represent different goals, although this is not to say that they do not: interact. Both can have integrity and
27.
437
scientific validity (and both can be done poorly). I am suggesting that, over the short term, there is often a trade-off between these two goals. As an example of mechanistic research on the effects of CO2 I refer to an e x p e r i m e n t that has been proposed (Bazzaz 1995; K6rner, pers. comm.) on the effect of elevated CO2 on the process of self-thinning (density-dependent mortality) in plant populations. Such an experiment could provide valuable data on the interaction between resource levels and density-dependent mortality. It might even provide insights into the mechanisms of density-dependent mortality beyond questions concerning CO2. Such an experiment, if done reasonably well, would be very worthwhile scientifically. But I believe such an experiment would be practically useless in the near term in helping us to predict the effects of elevated CO2 on terrestrial ecosystems that we will be seeing in the coming decades. Similarly, an e x p e r i m e n t in which we enrich a whole-plant community with CO2 for as long as possible will probably be m u c h more valuable for predicting what will h a p p e n in the coming years, but it will probably not be very useful in showing us the mechanisms by which these changes occur. I call this latter type of e x p e r i m e n t "brute-force empiricism." Much of m o d e r n medicine is based on such brute-force empiricism. We know a treatment works, but we often do not know the mechanism. Would one be willing to take a drug based purely on the data from studies and chemical theories? No, the principles of public health require that clinical trials be performed. If clinical trials are not possible, we want experiments on animals similar to humans. Similarly, if we're going to predict effects of elevated CO2 on terrestrial ecosystems, the best type of data will be from experiments, natural or planned, which are as close to the thing we're trying to predict as is possible. But the clinical trial of a new medicine whose mechanism of action is not known will probably not provide useful information on the mechanism. It will merely tell us if the medicine works in a specific population. The only reliable predictions possible for complex p h e n o m e n a of which we have very limited understanding come from brute-force empiricism and, when necessary, extrapolation. To make a reliable prediction in such a case, one should study the p h e n o m e n o n itself, or a system as similar to it as possible. In young sciences such as ecology and environmental science, data are more trustworthy than theory in making predictions (Peters, 1991; Weiner, 1995). How are we able to predict the effects of specific treatments other than elevated CO2 on terrestrial plant communities? For example, we know from experience that increasing the nutrients in many nutrient-poor plant communities will result in increased biomass and a reduction in species diversity. But we know this from e x p e r i e n c e m t h e e x p e r i m e n t has been conducted many times. The theories that we have at this point to explain
438 why this occurs are still after-the-fact explanations; they are not really the basis for our prediction that when we add phosphorous to an oligotrophic lake, we will get a huge increase in algal growth and a concurrent decrease in algal diversity. Similarly if we are asked to make a prediction of the effects of building a highway on local populations of plant and animals, the most useful type of information would be the effects of other roadbuilding projects on other communities, not deductions from ecological theories. Although such experiments have been done many times, the experiment of elevated CO2 is being done for the first time. The challenges presented by global change are before us. Predicting and analyzing the structure and function of ecological systems on large spatial and long temporal scales are research challenges of rare potential but daunting difficulty. The potential derives from both need and The difficulty reflects the diversity and non linearity of ecological responses. (Ehleringer and Field, 1993; emphasis mine) The dichotomy I have described fits Ehleringer and Field's eloquent diagnosis (Figure 3). Fundamental mechanistic research on effects of elevated CO2 can and should be justified on its own terms, and it will eventually contribute to our understanding and prediction of global change, but it cannot be justified in terms of predicting global change in the near term. But if the goal is obtaining the best prediction of change in terrestrial systems my claim is that an imperfect experiment at the level of organization we want to predict will be better than a perfectly designed experiment at a much lower level. According to this argument, the following studies are most likely to yield reasonable predictions in the near future: (1) The study of naturally occurring communities of high C02, e.g., volcanic vents in Italy (Miglietta and Raschi, 1993; Miglietta 1993; K6rner and Miglietta, 1994),Java (von Faber, 1925), and California (Koch, 1993). According to the arguments advanced above, despite the limitations of such studies (e.g., possible confounding factors such as other contaminating gases, limited replication, etc.), they probably represent the best available information we have for predicting effects of elevated CO2 on communities and ecosystems. This is because such studies are perhaps the only ones that are at the appropriate level temporally. I believe the potential value of comparative studies on naturally-occurring high CO2 communities, in comparison with experimental studies, has been greatly underestimated by researchers. (2) Whole community experiments with elevated CO2, as realistic and long term as possible (e.g., open-top chambers, FACE experiments) (3) Microcosm versions of (2).
27.
Figure 3
Two basic types of research on effects of elevated CO2.
(4) Paleological evidence of community changes correlated with changes in CO2. If it can be established that CO2 levels were much higher in the Cretaceous, paleological data on terrestrial plant communities could be of value. Even information on terrestrial systems during periods of lower CO2 over the past 100,000 years may be useful via extrapolation.
Some of the controversies concerning predictions of global change may result from the two very different uses of the word "prediction" in science. A hypothesis is a prediction: a claim about the behavior of the world based on a theory or model. Hypotheses are one of our most important research tools. Many of the most exciting and important hypotheses in science are controversial. But the word "prediction" is also used to describe consensus expectations from the scientific and engineering communities on which extra-scientific decisions can and should be based. Although the rising levels of CO2 are predictions in this latter sense, even the best of our predictions about the ecological effects of rising CO2 are only hypotheses, part of the research program in global change, but not yet firm bases for policy decisions. It has been said many times that as science progresses we answer questions, but we often raise more questions than we answer. Perhaps one of the most important things we have learned from CO2 research over the past decade is that things are not as simple as we might have hoped, that we do not know very much, that it will not be easy to make predictions about global effects. Ecology as a science is not yet developed enough to produce the predictions we are being asked to make. We must resist the temptation to confuse the importance of an issue with our ability to understand it. Questions concerning the ecological effects of anthropogenic environmental changes such as elevated CO2 are perhaps among the most
important scientific questions facing the world today, but it does not follow that we have or soon will have the means to answer them. Policy makers want predictions, and they give scientists grants to produce such predictions. I do not think ecology will be well served if we claim to understand more than we do. Rather we are obliged to communicate to policy makers the concept of uncertainty with which they seem so uncomfortable. But if we can use the opportunity presented by global changes such as elevated CO2 to do the fundamental ecological research needed to develop a scientific understanding of the processes involved, we can bring scientific opportunity and practical need together in a way that will further both our science and our public responsibility.
I thank S. Bassow, F. A. Bazzaz, M. Jasienski, C. K6rner, C. Loehle, S. C. Thomas, P. Voss, and an anonymous reviewer for comments on an earlier version of this chapter. Special thanks to F. A. Bazzaz for hosting my visit to Harvard. This work was supported by a Bullard Fellowship from Harvard Forest.
Allen, T. F. H., and Starr, T. B. (1982). "Hierarchy: Perspectives for Ecological Complexity." Univ. of Chicago Press, Chicago. Bazzaz, F. A. (1990). The response of natural ecosystems to the rising global CO2 levels. 21, 167-196. Bazzaz, F. A., and Garbutt, K. (1988). The response of annuals in competitive neighborhoods: Effects of elevated CO2. 69, 937-946. Bazzaz, F. A., and McConnaughay, K. D. M. (1992). Plant-plant interactions in elevated COs environments. 40, 547-563. Bazzaz, F. A., Bassow, S. L., Berntson, G. M., and Thomas, S. C. (1996). Elevated COs and terrestrial vegetation: Implications for and beyond the global carbon budget. "Global Change and Terrestrial Ecosystems" (B. Walker, ed.), pp. 43-76. Cambridge Univ. Press, Cambridge, UK. Bloom, A.J., Chapin, F. S., and Mooney, H. A. (1985). Resource limitation in plantsmAn economic analogy. 16, 363-392. Ehleringer, J. R., and Field, C. B. (eds.) (1993). "Scaling Physiological Processes: Leaf to Globe." Academic Press, San Diego. Koch, G. W. (1993). The use of natural situations of COs enrichment in studies of vegetation responses to increasing atmospheric COs. "Design and Execution of Experiments on COs Enrichment" (E.-D. Schulze and H. A. Mooney, eds.), pp. 381-391. Ecosystem Research Report 6, Commission of the European Community, Brussels. K6mer, C. (1988). Does global increase of COs alter stomatal density? 181, 253-257. K6rner, C. (1993). COs fertilization: The great uncertainty in future vegetation development. "Vegetation Dynamics and Global Change" (A. M. Solomon and H. H. Shugart, eds.), pp. 53-70. Chapman & Hall, London.
27.
441
K6rner, C., and Miglietta, F. (1994). Long term effects of naturally elevated C O 2 o n mediterranean grassland and forest trees. 99, 343-351. Levin, S. A. (1993). Concepts of scale at the local level. "Scaling Physiological Processes: Leaf to Globe." (J. R. Ehleringer and C. B. Field, eds.), pp. 7-19. Academic Press, San Diego. Loehle, C. (1983). Evaluation of theories and calculation tools in ecology. 19, 230-247. Loehle, C. (1995). Anomalous responses of plants to CO2 enrichment. 73, 181-187. Miglietta, F., and Raschi, A. (1993). Studying the effect of elevated CO2 in the open in a naturally enriched environment in central Italy. 105, 391-400. Miglietta, F., Raschi, A., Bettarini, I., Resti, R., and Selvi, F. (1993). Natural COz springs in Italy: A resource for examining long-term response of vegetation to rising atmospheric CO2 concentrations. 16, 873-878. Mooney, H. A., and Koch, G. W. (1994). The impact of rising COz concentrations on the terrestrial biosphere. 23, 74-76. Mooney, H. A., Drake, B. G., Luxmoore, R. J., Oechel, W. C., and Pitelka, L. F. (1991). Predicting ecosystem responses to elevated CO2 concentrations. 41, 96-104. O'Neill, R. V., DeAngelis, D. L., Wade, J. B., and Allen, T. F. H. (1986). "A Hierarchical Concept of Ecosystems." Princeton Univ. Press, Princeton, NJ. Patten, B. C. (ed.) (1983). "Systems Analysis and Simulation in Ecology." Academic Press, New York. Peters, R. H. (1991). "A Critique for Ecology." Cambridge Univ. Press, Cambridge, UK. Reekie, E. G., and Bazzaz, F. A. (1991). Phenology and growth in four annual species grown in ambient and elevated CO2. 69, 2475-2481. Reynolds, J. F., and Acock, B. (1985). Predicting the response of plants to increasing carbon 29, 107-129. dioxide: A critique of plant growth models. Reynolds, J. F., Hilbert, D. W., and Kemp, P. R. (1993). Scaling ecophysiology from the plant to the ecosystem: A conceptual framework. "Scaling Physiological Processes: Leaf to Globe" (J. R. Ehleringer and C. B. Field, eds.), pp. 127-140. Academic Press, San Diego. Tilman, D. G. (1986). Resources, competition and the dynamics of plant communities. "Plant Ecology" (M. J. Crawley, ed.) pp. 51-75. Blackwell Sci., Oxford. von Faber, F. C. (1925). Untersuchungen fiber die Physiologie der javanischen SolfatarenPflanzen. 18/19, 89. Weiner, J. (1995). On the practice of ecology. 83, 153-158. Woodward, F. I. (1987). Stomatal numbers are sensitive to increases in CO2 from pre-industrial levels. 327, 617-618. Woodward, F. I., Thompson, G. B., and McKee, I. F. (1991). The effects of elevated concentrations of carbon dioxide on individual plants, populations, communities, and ecosystems. 67 (Suppl. 1), 23-38.
This Page Intentionally Left Blank
2 The Significance of Biological Variation, Organism Interactions, and Life Histories in CO2 Research
The diversity of structures, functions, and responses to external factors among individuals of a population and between species and assemblages of species provides the raw material for natural selection, including the basis for adaptation to changing environmental conditions. Biological diversity also sustains the richness of ecosystem functions. Because of this intrinsic variability of living systems the responses to any sort of external forcing will also vary. This volume explores the interaction between one aspect of global change, increased atmospheric CO2, and the responses of communities of organisms exposed to it. In contrast to most of the preceding studies with increased CO2, the studies here emphasize ways that the responses of populations and communities can amplify, suppress, or redirect the responses of individual organisms. Increasing atmospheric CO2 is an external driver likely to affect the whole biosphere. CO2, in addition to solar energy and water, is a basic resource for life on earth. Its atmospheric concentration has increased by 27% since the beginning of the industrial revolution and is likely to double within less than 100 years. Although atmospheric CO2 has fluctuated in the past, the recent rapid increase, caused primarily by fossil fuel consumption and anthropogenic deforestation, is unprecedented. Understanding the responses of plants and plant communities to altered carbon supply is a central theme of global change research. This volume focuses on the varia443
Copyright 9 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.
444 tion of responses in populations and communities to the continuing increase of atmospheric COz, and it sets the stage for assessing the implications of these responses for other aspects of biosphere function.
This book appears at an important point in the development of our understanding of plant and ecosystem responses to elevated CO2. Until recently, CO2 enrichment was considered to be almost always beneficial to plants. Based mainly on data from agricultural plants grown with abundant nutrients, water, and light, it was assumed that natural communities will uniformly and strongly increase growth and carbon storage under elevated CO2. Results of a number of recent CO2 enrichment studies in less productive, more natural settings, where competition for space and other resources was allowed, reveal a richer suite of responses. Stimulation of plant growth by increased CO2 is often very modest and less important for community and ecosystem function than a number of other effects, some of which are positive and some of which are negative (Bazzaz 1995; K6rner, 1996; Field 1995). Most past studies of plant responses to increased atmospheric CO2 have been designed to maximize the resolution for detecting CO2 effects by minimizing biological and environmental variability. Although this approach has been invaluable for characterizing the CO2 sensitivity of particular genotypes of a few species and for exploring direct and indirect mechanisms underlying CO2 effects, this approach is not sufficient as a basis for understanding at the community or ecosystem scale. This book brings variance to the center stage. The future of our biosphere under changed atmospheric conditions will largely depend on how selection will act on this variance and how it will translate into new assemblages of genotypes and species. We must understand the variance of plant responses to elevated CO2 in a variable environmental matrix and the mechanisms behind them in order to correctly predict the furore of a CO2-enriched biosphere, both in terms of biomass production and biological diversity. This recognition motivates the focus on population and community studies. In general, the critical information for understanding the implications of CO2 effects on the composition of species and communities is only beginning to accumulate. The chapters in this volume synthesize much of the currently available information, considering both the nature of the responses and their consequences for other processes. Though providing a still incomplete picture, these studies present a compelling case for major roles of COz responses in genotypes within species and species within communities. These roles are clearly important enough that population
28.
445
and community processes should receive a high priority in the continuing initiative to understand and predict CO2 responses at the ecosystem and global scales.
This volume deals largely with experiments. Experimentalists are using a number of model systems designed to produce unequivocal results under the conditions of the experiment, but their application to the real world depends to a large measure on scaling in space a n d time beyond the experimental conditions. Mathematical models, including both conceptual and simulation models, provide powerful but dangerous tools for the scaling needs. Model predictions become most suspect when pushed far outside the range of validation experiments, for example when used to project responses to gradually increasing CO2 on the basis of initial responses to a step-change or when response functions from fast-growing agricultural plants are used to predict responses in late successional, natural communities. Kingsolver (Chapter 1) discusses compelling empirical and theoretical reasons to suppose that evolutionary responses to abrupt step-changes in the environment will be qualitatively different from responses to gradual progressive changes. Under natural conditions, selection is likely to involve polygenetic responses rather than the monogenetic ones usually tested in laboratories. Several research groups have now documented substantial intraspecific variation in response to CO2 (cf. Ackerly and Bazzaz, 1995; Bazzaz 1995, Curtis Chapter 2). In most wild plant species, especially woody ones, life cycles are too slow to study selective processes under elevated CO2. Yet, some extremely shortlived species may serve as useful model organisms. Tousignant and Potvin (Chapter 3) used a fast-growing wild mustard (Brassicajuncea) and observed its biomass production over seven generations during which CO2 was increased by 40 ppm and temperature was increased by 0.5 K for each successive generation. Gradual selection for faster vegetative growth occurred and was verified as genotypic with reciprocal transplant experiments. By the end of the reproductive phase, however, this response disappeared and fruit biomass fell from 3.5 g in controls to only 1 g in the final climate (630 ppm CO2). This late response was fully plastic (phenotypic) with no sign of genetic adaptation. CO2 responses of plants can have genetic as well as phenotypic dimensions that manifest themselves throughout the life cyclema challenge for scientists working with trees or slow-growing and late-reproducing perennials.
446 What are the mechanisms that make a genotype a winner under increased CO2? Curtis (Chapter 2) present data for that bear a clear message: Genotypes with a pronounced and positive photosynthetic gain under elevated CO2 are not consistent "winners" in terms of biomass. In fact, 11 of the 19 seed families that showed increased CO2 uptake showed no biomass response, while some of the seed families with little or no photosynthetic response to CO2 had quite large biomass responses. Hence, the marked intraspecific variation in the CO2 response of growth is not tightly linked to CO2 responses of carbon fixation per unit leaf area. With morphological features, such as genetically determined plant size, were much better predictors (Curtis Chapter 2). Habitat breadth in current environments is not necessarily a good predictor of responses to elevated CO2 under natural conditions. Schmid (Chapter 4) present results of a study in which various genotypes of a specialist and a generalist species of were exposed to CO2-enriched atmospheres in their natural field environment. In contrast to the initial hypothesis that niche breadth would confer greater CO2 sensitivity, the specialist profited much more from elevated CO2 and exhibited a wider spectrum of genetic variation in CO2 responsiveness. Surprisingly, genetic variance of biomass increased under elevated CO2, underlining the potential importance of intraspecific differences in response. Thomas and Jasienski (Chapter 5) highlight the role of plant density and plant-plant interactions. They suggest that because rapidly rising CO2 is a relatively novel selective agent, its strength may be difficult to predict. If selection by CO2 enrichment is sufficiently strong, it may not be effectively counterbalanced by selection by other environmental factors such as frost and drought. This biased selection for one or few traits could reduce overall fitness in the long term.
The responses of plant communities, their pollinators, dispersers, symbionts, and pathogens to elements of the global change are critical regulators of biological diversity, ecosystem function, and the provision by ecosystems of goods and services valued by humans. But, relevance to ecosystem and global questions is not the only motivation for focus on community and population processes in a CO2-enriched world. In addition, external forcing from global change provides an invaluable probe for mechanistic studies at the community and population levels. Community responses to increased CO2 may involve several levels of interaction. Increases in the population of one species and losses in others can fundamentally change interspecific interactions at all levels of complexity. For example, an important pollinator of one
28.
447
species may become rare as a result of decreases in the abundance of a preferred food plant for larvae, or root competition between two species could allow entry of a third species largely occupying a different soil horizon. Some generalizations are now emerging from work with wild plants under elevated CO2; for example, Grime (Chapter 6) suggests that, on fertile land, slow-growing stress tolerators are likely to become losers, whereas fastgrowing species will increase. On less fertile sites chances will be more equal. Based on observations on the British Isles, Grime noted, with surprise, that regenerative attributes are poor predictors of CO2 sensitivity of species. Screening tests indicate that fast-growing herbaceous species currently expanding in heavily populated countries of Europe are also most responsive to CO2 enrichment (cf. Poorter Chapter 25). This observation is consistent with higher CO2 responsiveness of early versus late successional species (Bazzaz and Miao, 1993; Sch~ippi and K6rner, 1996), which may explain their current success and facilitate increased dominance in the future. Naeem (Chapter 7) use assemblages of annual plants of varying species number to investigate the role of biodiversity in modulating CO2 effects on communities.They conclude that the higher diversity assemblages produce more biomass per unit land area and that this effect is enhanced under elevated CO2. Manipulating both CO2 and community composition will certainly improve our understanding of vegetation responses to global change. A key factor in such experiments will be the availability of resources other than CO2. The literature is replete with evidence suggesting multiway interactions between species, communities, and resources in regulating ecosystem CO2 responses. Most of the research to date has been done with fast-growing, small plants, largely annuals. Arnone (Chapter 8) presents a review of our current understanding of the responses of longer lived tropical plants to elevated CO2. The central theme of this work is that responses of individually grown plants to high CO2 do not scale simply to results of experiments with individuals grown under competitive conditions or complete ecosystems. When mineral nutrients severely constrain plant growth, speciesspecific differences in foraging success for these resources determine the overall successional process and the ultimate community structure. Results like these highlight the need for effective integration of individual, community, and ecosystem-scale processes under realistic climate change scenarios. Of course, actual competitive outcomes may depend on the species present, their life forms, and their relative abundances and arrangements. Given the heterogeneity in the field, carefully designed and coordinated experiments in both natural and model systems will provide critical extensions of field results (K6rner, 1995b; Lawton, 1995). A conclusion supported by both Naeem and Arnone (Chapters 7 and 8, respectively) is that CO2 enrichment eventually will alter patterns
448 of plant succession and vegetation cover. Whether such changes will lead to alterations of ecosystem processes is one of the key questions for the future. Roy (Chapter 9) point to this scaling problem and present results of a study with species-rich microcosms containing Mediterranean grassland species. Of 57 species, half responded negatively and half positively to CO2 enrichment. But, in only 5 of these species were responses significantly positive. There was no significant difference between annuals and perennials or between major plant families such as Asteraceae (60% of all species stimulated), legumes (40% stimulated) and grasses (30% stimulated). One interesting response was that the CO2 responses of grasses and legumes when mixed were negatively correlated (when grasses were stimulated, legumes were suppressed, and vice versa). This observation does not match predictions derived from studies with isolated, well-fertilized plants (see review by Poorter Chapter 25). The changes in community composition did not translate into a significant effect of elevated CO2 on total biomass, though root mass tended to increase. The study clearly documents that competitive interactions can lead to negative responses of some wild plant species to elevated CO2 (Hunt 1993; Ackerly and Bazzaz, 1995). In the long term, changes at community level may be more important than short-term biomass responses (Bolker 1995). This idea is further supported by results from microcosm experiments with Mediterranean grassland in California (Chiariello and Field, Chapter 10). In this experiment, increased CO2 led to decreased water consumption by some species, allowing other species to profit. Under elevated CO2, the Jasper Ridge assemblage of late season annuals profited from savings in soil moisture by species that are active earlier in the season. Higher soil moisture may lead to fundamental changes of community structure including the invasion of shrubby species or even forest trees. Natural grassland community responses to elevated CO2 are presented by Leadley and K6rner (Chapter 11). They report no significant increase in aboveground plant biomass in either natural communities or in communities with manipulated biodiversity. Similar to the observations by Roy (Chapter 9), legumes did not profit from CO2 enrichment nor did species with high potential growth rates, in contrast to the conclusions of Grime and Poorter (Chapters 6 and 25, respectively). Surprisingly, the slow-growing sedge was the only species out of 30 that showed a clear and strong positive response to CO2 enrichment. It seems that the conventional assumptions about functional group responses to elevated CO2 need to be reassessed in a more natural context. Polley (Chapter 12) use documented natural patterns of plant succession, combined with theoretical considerations, to evaluate the possibility that vegetation change due to elevated CO2 may already be underway. They suggest that rising CO2 was most influential in grasslands, in which
28.
449
water or nitrogen previously limited woody plant recruitment, and in which rising CO2 is now significantly increasing growth rate and fecundity of woody plants. These authors join others in this volume in emphasizing the importance of life history in predicting long-term responses of vegetation to elevated CO2. Predictions become even more challenging when interactions between elevated CO2 and other elements of global change come into play. GwynnJones (Chapter 13) present data from a field experiment with natural subarctic vegetation under both elevated CO2 and enhanced UV-B radiation. In these experiments, all species behave individualistically. The deciduous dwarf shrub was very sensitive to perturbations, showing both CO2 and UV-B responses in the first season. Responses were, however, less pronounced in the second season. Other dwarf shrub species were rather unresponsive to either environmental perturbation. CO2 fertilization may lead to greater plant biomass, litter production, or both. Under dry conditions, this may lead to increased fire frequency (Crutzen and Goldammer, 1993). CO2 effects may be most important where elevated COz promotes a switch between contrasting fire/fuel cycles. COz effects on fire frequency, mediated through enhancement of fuel load, may be a major controller of species composition (Sage, Chapter 15). Independent of changes in fire frequency, Mediterranean shrublands and forests may be particularly sensitive to COz enrichment through amelioration of drought stress. Scarascia (Chapter 14) report that the Mediterranean arboreal species was favored over typical macchia shrubs, perhaps as a consequence of COz-related improvements in soil moisture status. Thus, CO2 effects may facilitate transitions from shrub to forest communities unless, of course, a drier climate counteracts the positive effect of CO2 on soil moisture (Field 1995).
Population and community responses to elevated C O 2 may depend on a number of factors, including species characteristics, intraspecific variation, plant density, other biotic factors (herbivores, pathogens, pollinators), soil resources, climate, and local history. Carefully designed experiments are required to unravel these multidimensional interactions. Species within a mixed community show a range of responses to elevated CO2 in the majority of studies to date (Reynolds, Chapter 18). One of the best-studied interactions is between C3 and C4 species. Even with C3-C4 interactions, where the contrast in photosynthetic physiologies suggests a consistent outcome of competition, actual interactions can depend strongly on soil fertility, moisture, and species characteristics. Responses of other functional groupings
450
are even more challenging to predict under natural growth conditions. For this reason Reynolds (Chapter 18) concluded that it may be more productive to focus on understanding performance of species grown in relatively complex mixtures, than to extrapolate from experiments on simple, species-poor mixtures, except in agricultural research where such communities predominate. Examples of complex, multidimensional interactions are presented by L~scher (Chapter 19) and Campbell and Hart (Chapter 20). L~scher report large differences in the CO2 responses of seven perennial species, typical of permanent temperate grasslands, grown within a matrix of On these fertile soils, legumes showed a strong positive response to elevated CO2 while three grass species responded negatively. Surprisingly, no intraspecific variability in these responses was detected. In contrast to the pattern observed by L~scher et al. (Chapter 19) in Swiss lowlands, Campbell and Hart (Chapter 20) report no stimulation of legume biomass in swards of grazed turf under the cooler temperatures in New Zealand. At lower temperatures, increased CO2 tended to favor grasses. But competitive suppression of under increased CO2 was significantly reduced when temperature was increased to 28/23~ The competitive outcome between legumes and grasses under elevated CO2 is also sensitive to the grass species used. Under elevated CO2, and restricted most severely, whereas was a weak competitor. The effect of increased CO2 on a given target species depends on the community context. When multiple life forms interact, morphology can become a key determinant of interaction, and predictions of CO2 responses must integrate beyond the CO2 responses of photosynthesis and growth in individual plants and single-species communities. Gloser (Chapter 21) assessed competitive outcome under elevated CO2 in an artificially planted tree-grass mixture. Although the predicted reduction of the light compensation point and increase of shade tolerance of spruce under elevated CO2 did not occur, the competitive success of a was clearly enhanced under CO2 fertilization, reducing early tree-seedling success and tree recruitment. Differences in morphology can influence CO2 responses of plants more than physiological differences (K6rner, 1993a). CO2 enrichment can also influence morphology within a single life form, as illustrated for tropical tree species by Reekie and Bazzaz (1987). Developmental processes represent a second major challenge to our ability to predict plant-plant interactions under elevated CO2 (Reekie, Chapter 22). The fact that CO2 can influence development directly, severely constrains attempts at prediction, particularly in cases where not all phases of the life cycle can be studied, as in trees.
28.
Effects of increased C O 2 o n plants and plant communities are very likely to lead to secondary (and tertiary) effects on microbes, including symbionts, parasites, and decomposers. One of the most important plant-microbe interactions is the symbiosis between nitrogen-fixing bacteria and legumes. Since nitrogen is a limiting resource for growth in many natural ecosystems, any facilitation of nitrogen fixation as a result of CO2 fertilization could have far-reaching effects on community structure and ecosystem function. Hartwig (Chapter 16) found that CO2 fertilization in the field enhances growth and nitrogen fixation by white clover and increases its proportion in a grass sward at the expense of associated grasses. This confirms results of earlier growth chamber observations (Overdieck, 1986; Nijs 1989) but contrasts with the results of Campbell and Hart (Chapter 20). Although nitrogen fixing symbionts are restricted to few plant families, plant-fungal mutualisms exist in the vast majority of species. Increased carbon availability may stimulate nutrient uptake and delivery by mycorrhizas, yet increased mycorrhizal biomass could also lead to increased carbon demands and perhaps diminishing returns on nutrients. Most studies to date indicate no or modest effects of increased CO2 on mycorrhizal abundance (e.g., Whitbeck, 1994). Recent evidence from work with grassland microcosms suggests that mycorrhizal fungi may not necessarily become more mutualistic under elevated CO2, but may even become parasitic (Sanders, Chapter 17). The responses appear to be species-specific, enhancing the possibility that effects of CO2 enrichment on plant-mycorrhiza interactions will modify the structure of plant communities.
Effects of C O 2 enrichment on plant tissue composition suggest secondary effects not only on microbes but also on animal consumers. Feedbacks from differential consumption can affect species composition and abundance (Lindroth, Chapter 23; Lincoln et al., 1993). In general, insects fed material grown under increased CO2 increase consumption, probably as a consequence of the lower protein concentration in CO2-fertilized leaf tissue. Responses may, however, vary among insects. Differential responses may even be found between males and females of the same species (Traw and Bazzaz, unpublished). There are a number of gaping holes in our understanding at the population level. For example, will reduced growth and prolonged development of insects alter mortality and natality rates such that populations decline? From the perspective of tree defoliation, to what degree will decreases in herbivore populations offset increased
452
consumption rates by individual insects? Lindroth concludes that little is known about potential intraspecific variation in insect responses to CO2induced changes in host quality. The same is true for the response of herbivores when provided with a choice of different species (cf. Arnone, Chapter 8). Considering such questions will require long-term studies in ecosystems with realistic levels of complexity. Large ruminants may respond quite differently from insects (Owensby Chapter 24). Unlike insects, where intake typically increases as diet quality decreases, ruminant intake usually declines with decreasing forage quality. With ruminants, the response of consumption amplifies the decrease in growth per unit of consumption, probably resulting in reduced growth and reproduction. According to Owensby a future high CO2 world seems destined to reduce individual animal performance.
Modeling the future of our biosphere requires generalization. Variation within species and populations is currently handled incompletely and in only some of the models designed to predict CO2 responses at the landscape or global scale. One attractive approach to generalization is tosearch for functional types that may serve as substitutes for taxonomic units but respond consistently to increased CO2 under natural conditions. Poorter (Chapter 25) present a review of plant CO2 responses in which they attempt to distill "functional group-specific" CO2 responses. In general, they confirm the greater CO2 sensitivity of C~ plants but provide evidence for substantial responses in C4 and CAM plants. On average, potentially fast-growing wild species and crop species show relatively strong growth responses to CO2 (when grown under optimal conditions and in isolation) whereas inherently slow-growing species show much smaller responses. When nutrient availability is low, plant growth is, on average, unresponsive to high CO2. Poorter conclude that this is one of the reasons we expect growth responses of most natural vegetation in the field to be small. Under competitive conditions the differential responses of species may drastically change, and the overall productivity response may be smaller than for individually grown plants (Bazzaz and McConnaughay, 1992; Baz1995). Aboveground plant parts interact primarily by affecting the light climate of their neighbors. Hirose (Chapter 26) examine interactions between canopy structure and CO2 enrichment. Theoretical considerations (Long, 1991) suggest that elevated CO2 should lead to moderate increases in leaf area index. This is observed in some studies, especially with agricultural species. In other examples, leaf area index is unaffected by increased CO2.
28.
453
Experimental stands of and exhibited no effect of C O 2 on leaf area index (Hirose Chapter 26), similar to reports for other plant canopies (Arnone, Chapter 8; H~ittenschwiler and K6rner, 1996a). Some natural plant communities may adjust individual allometry in such a way that leaf area per unit plant mass is reduced rather than increased. In cases where leaf area index does increase in response to elevated CO2, the result could be a loss of suppressed individuals, elimination of less shadetolerant species from the understory, and decreased biological diversity. Alternatively, increased CO2 may enhance survival under lower light, mitigating shading effects (Long, 1991; Bazzaz and Miao, 1993; H~ittenschwiler and K6rner, 1996b).
The diversity and multidimensionality of possible population and community responses to increased CO2 pose major challenges for the extension of these responses to the scale of ecosystem function, the global carbon cycle, or impacts on ecosystem goods and services valued by humans. To date, most of the models, both conceptual and quantitative, designed to explore questions at these scales either completely ignore population and community responses, or they treat only a few aspects. This approach is a reasonable starting point, but the field is now at the portal of an era where further progress in understanding CO2 responses is critically dependent on an effective integration across fields and approaches, including physiology, population and community studies, and ecosystem studies. Early predictions of ecosystem-scale responses to increased CO2 were typically extrapolated from studies on individual plants (e.g., Strain and Bazzaz, 1983). More recent assessments include limited whole-ecosystem data (e.g., Mooney 1991; K6rner, 1996; Koch and Mooney, 1996) and an acknowledgment of the likely importance of population and community processes (e.g., Bazzaz, 1990; K6rner, 1995a). These assessments lack, however, a framework for translating an appreciation that these processes are important into concrete predictions about outcomes. As a consequence of this lack of a framework and a shortage of data, most of the quantitative models used to explore ecosystem and global CO2 responses work as if the vegetation consisted of only a single species, or perhaps a single plant, a single decomposer, and a single herbivore (e.g., Melillo 1993; Ojima 1993). Early models predicting species variation in responses to global change did not include CO2 effects, and simulated responses to only altered temperature and precipitation (e.g., Pastor and Post, 1988).
454 A number of recently developed simulation models predict changes in the equilibrium distribution of potential natural vegetation, in response to altered climate and increased CO2 (e.g., Nielson, 1993; Prentice 1993). These models, which redistribute vegetation types on the basis of moisture and carbon balance, still treat each major biome as if it were a single species. Dynamic versions of these models, now under active development, replace biomes with a number of functional types, intended to compete more or less realistically and to co-occur, where appropriate. Yet, the challenge of defining broadly useful functional groups is substantial, and the information necessary to associate functional groups with CO2 responses is incomplete (K6rner, 1993b). Modeling approaches based on the hypothesis that biomes stay in place or that responses of vegetation types can be characterized as responses of single aggregate species are a reasonable starting point, given the clear need for synthesis and the fact that global changes on the time scale of decades to a century are fast relative to the life span of many trees. Yet, an accumulating body of evidence, including each of the chapters in this volume, highlights issues where population a n d / o r community aspects of ecosystem-scale CO2 responses are critically important, on time scales that are potentially quite rapid. Broadly successful models at the ecosystem and larger scales will require a combination of increased data and improved tools for generalization. The community has, thus far, not converged on a single approach for providing these. Studies on natural ecosystems, microcosms, and synthetic canopies can all provide essential information. The challenge is deciding how to use the information from each level and avoiding the temptation to focus on a subset of well-understood or easily modeled processes (Weiner, Chapter 27). The chapters in this volume highlight a number of mechanisms that will be central to the development of integrated understanding of plant and ecosystem responses to increased atmospheric CO2 in future integrative models, but more important points were made here than we can summarize. Yet, many of the conclusions form a relatively compact suite. (1) Intraspecific variation in CO2 responses is often comparable to or even greater than interspecific differences. Understanding intraspecific variation and its consequences is a critical prerequisite for developing mechanistic predictive tools. (2) CO2 responses of plants grown in communities can rarely be extrapolated directly from responses of individual plants grown alone. (3) Morphological and developmental traits are major determinants of plant CO2 responsiveness. As a consequence, they may exert strong influences on both community structure and ecosystem function. (4) Success under increased CO2 often reflects responses to indirect effects, for example, altered soil moisture, herbivory, or symbiotic associations.
28.
(5) Ecosystem scale consequences of increased C O 2 c a n be amplified or suppressed by population and community responses. For example, a change in fire frequency or a transition from a grassland to a shrubland can have large effects on production, carbon storage, or watershed yield. (6) Population and community processes are typically the dominant controls on the ability of ecosystems to provide goods and services valued by humans. A narrow focus on production or carbon storage is likely to miss many of the most consequential ecosystem responses to increased CO2. (7) Ecosystem and global scale models designed to provide useful predictions of the consequences of global CO2 enrichment for policy must place a priority on improving the characterization of population and community processes.
Ackerly, D. D., and Bazzaz, F. A. (1995). Plant growth and reproduction along C O 2 gradients: 1, Non-linear responses and implications for community change. 199-207. Bazzaz, F. A. (1990). The response of natural ecosystems to the rising global CO2 levels. 21, 167-196. Bazzaz, F. A., and McConnaughay, K. D. M. (1992). Plant-plant interactions in elevated CO2 environments. 40, 547-563. Bazzaz, F. A., and Miao, S. L. (1993). Successional status, seed size, and responses of tree seedlings to CO2, light, and nutrients. 74, 104-112. Bazzaz, F. A., Jasienski, M., Thomas, S., and Wayne, P. (1995). Microevolutionary responses in experimental populations of plants to CO2-enriched environments: Parallel results from two model systems. 92, 8161-8165. Bolker, B. M., Pacala, S. W., Bazzaz, F. A., Canham, C. D., and Levin, S. A. (1995). Species diversity and ecosystem response to carbon dioxide fertilization: Conclusions from a temperate forest model. 1, 373-381. Crutzen, P.J., and Goldammer, J. G. (eds.) (1993). "Fire in the Environment. The Ecological, Atmospheric, and Climatic Importance of Vegetation Fires." Wiley, Chichester. Field, C. B., Jackson, R. B., and Mooney, H. A. (1995). Stomatal responses to increased COz: Implications from the plant to the global scale. 18, 1214-1226. H~ttenschwiler, S., and K6rner, Ch. (1996a). System-level adjustments to elevated COz in model spruce ecosystems. in press. H~ttenschwiler, S., and K6rner, Ch. (1996b). Effects of elevated CO2 and increased nitrogen deposition on photosynthesis and growth ofunderstorey plants in spruce model ecosystems. in press. Hunt, R., Hand, D. W., Hannah, M. A., and Neal, A. M. (1993). Further responses to CO2 enrichment in British herbaceous species. 7, 661-668. Koch, G. W., and Mooney, H. A. (eds.) (1996). "Carbon Dioxide and Terrestrial Ecosystems," Physiological Ecology Series. Academic Press, San Diego. K6rner, Ch. (1993a). CO2 fertilization: The great uncertainty in future vegetation development. "Vegetation Dynamics and Global Change (A. M. Solomon and H. H. Shugart, eds.), pp. 53-70. Chapman & Hall, New York/London. K6rner, Ch. (1993b). Scaling from species to vegetation: The usefulness of functional groups. "Biodiversity and Ecosystem Function" (E. D. Schulze & H. A. Mooney, eds.), pp. 117-140. Ecological Studies 99. Springer-Verlag, Berlin/Heidelberg/New York.
456 K6rner, Ch. (1995a). Biodiversity and COs: Global change is under way. 4, 234-243. K6rner, Ch. (1995b). Towards a better experimental basis for upscaling plant responses to elevated COs and climate warming. 18, 1101-1110. K6rner, Ch. (1996). The response of complex multispecies systems to elevated COs. "Global Change and Terrestrial Ecosystems" (B. H. Walker and W. L. Steffen, eds.), pp. 20-42. Cambridge Univ. Press, Cambridge, UK. Lawton, J. H. (1995). Ecological experiments with model systems. 269, 328-331. Lincoln, D. E., Fajer, E. D., and Johnson, R. H. (1993). Plant-insect herbivore interactions in elevated COs environments. 8, 64-68. Long, S. P. (1991). Modification of the response of photosynthetic productivity to rising temperature by atmospheric COs concentrations: Has its importance been underestimated? 14, 729-739. Melillo, J. M., Kicklighter, D. W., McGuire, A. D., Moore, B., III, Vorosmarty, C. J., and Grace, A. L. (1993). Global climate change and terrestrial net primary production. 363, 234-240. Mooney, H. A., Drake, B. G., Luxmoore, R. J., Oechel, W. C., and Pitelka, L. F. (1991). Predicting ecosystem responses to elevated COs concentrations. 41, 96-104. Nielson, R. P. (1993). A mapped ecotone response to climatic change: Some conceptual and modelling approaches. 3, 385-395. Nijs, I., Impens, I., and Behaeghe, T. (1989). Effects of long-term elevated atmospheric COs concentration on and canopies in the course of a terminal drought stress period. 67, 2720-2725. Ojima, D. S., Parton, W. J., Schimel, D. S., Scurlack, J. M. O. (1993). Modeling the effects of climatic and COs changes on grassland storage of soil C. 70, 643-657. Overdieck, D. (1986). Long-term effects of an increased COs concentration on terrestrial plants in model ecosystems. Morphology and reproduction of L. and 30, 323-332. Pastor, J., and Post, W. M. (1988). Responses of northern forests to COrinduced climate change. 334, 55-58. Prentice, I. C., Sykes, M. T., and Cramer, W. (1993). A simulation model for the transient effects of climate change on forest landscapes. 65, 51-70. Reekie, E. G., and Bazzaz, F. A. (1987). Reproductive effort in plants. 1. Carbon allocation to reproduction. 129, 876-896. Sch~ippi, B., and K6rner, Ch. (1996). Growth responses of an alpine grassland to elevated COs. 105, 43-52. Strain, B. R., and Bazzaz, F. A. (1983). Terrestrial plant communities. and Plants: The Response of Plants to Rising Levels of Carbon Dioxide" (E. Lemon, ed.), pp. 177-222. American Association for the Advancement of Science, Washington, DC. Whitbeck, J. L. (1994). "Effects of Above- and Below-Ground Resource Distribution on the Ecology of Vesicular-Arbuscular Mycorrhizas." Ph.D. thesis, Stanford University, Stanford, CA.
Index
canopy development model, 418-420 canopy development model, 418-420 Animals biodiversity effects, 451-452 ruminants, forage plant quality, 363-369 carbon dioxide impact, 366-367 cattle production impact, 367-369 future research directions, 369 ruminant digestion, 364-366 Bacteria, nitrogen fixation, 253-261 field studies, 258-259 model mechanics, 259-260 nitrogen availability, 255-258 plant growth effects, 255 symbiotic process, 254-255 Biodiversity community composition changes, 93-99 discussion, 97-98 methods, 94-96 overview, 93-94 results, 96 summary, 98-99 grassland plant species dominance, 159-174 community responses, 164-166 discussion, 166-173 aboveground biomass changes, 173-174 community composition consequences, 166-170 diversity manipulation interpretation, 172-173 functional groups, 170-172
response variation, 174 experimental design, 160-164 Mediterranean old-field microcosms, 124-126 significance, 443-455 community responses, 446-449 ecosystem consequences, 453-455 genotypic population responses, 445-446 global consequences, 453-455 modeling, 452-453 plant-animal interactions, 451-452 plant-microbe interactions, 451 plant-plant interactions, 449-450 theory, 452-453 variance study, 444-445 Biomass allocation patterns calcareous grassland plants, species dominance, 173-174 competitive performance role, 108-111 grasses verses 305-308 Mediterranean old-field microcosms, 126, 129, 131-132 global change responses, 23-30 competition, 319-329 interactions, 327-329 low light acclimation, 321-325, 329 shade tolerance, 320-326 CAM species, interspecific growth response variation, 380-384 Canopy development, herbaceous plant stands, 413-428 418-420
Canopy development, herbaceous plant stands (continued) 418-420 future research directions, 427-428 leaf area index optimization, 420-426 light interception, 414-418 model, 415-416 photon partitioning, 416-418 nitrogen allocation, 420-426 Cattle, Ruminant animals C4 plants, competition C3 versus C4 species, 275-278, 282-283 grasses verses 301-315 dry mass results, 305-308 long term implications, 314-315 methods, 303-305 nutrient status, 308-311,314 phenology, 311-313 shoot morphology results, 308 species competitive abilities, 313 temperature effects, 311-313 theory, 301-303 grassland legumes versus nonlegume responses, 287-297 intraspecific response variability, 293-296 methods, 288-289 nitrogen availability effects, 292-293 response variability, 289-292 Chemistry, Plant chemistry Climate, change responses 23-30 evolutionary responses, 7-11 genetics, 4-7 grasses verses 311-313 overview, 3-4 thermal sensitivity, 7-11 Communities, composition changes calcareous grasslands, 166-170 discussion, 97-98 methods, 94-96 overview, 93-94 results, 96 summary, 98-99 fire disturbances, 231-245 carbon dioxide effects, 235-241 cycle prediction, 241-243 fire role, 232-235 research priorities, 243-244
Competition developmental process implications, 333-344 carbon dioxide responsiveness, 341-343 competitive ability, 334-336 developmental patterns, 334-336 effects prediction, 336-341 plant-plant interactions differential responses, 275-283 C3 versus C~ species, 278-282 C~ versus C4 species, 275-278, 282-283 functional groupings, 284 grasses verses 301-315 dry mass results, 305-308 long term implications, 314-315 methods, 303-305 nutrient status, 308-311,314 phenology, 311-313 shoot morphology results, 308 species competitive abilities, 313 temperature effects, 311-313 theory, 301-303 grassland legumes versus nonlegume responses, 287-297 intraspecific response variability, 293-296 methods, 288-289 nitrogen availability effects, 292-293 response variability, 289-292 overview, 273-275 versus 319,-329 low light acclimation, 321-325, 329 shade tolerance, 320-326 species interactions, 327-329 species dominance, calcareous grassland plants, 159-174 aboveground biomass changes, 173-174 community composition consequences, 166-170 community responses, 164-166 diversity manipulation interpretation, 172-173 experimental design, 160-164 functional groups, 170-172 response variation, 174 tropical ecosystems belowground interactions, 114-115 leaf area index, 111-112
photosynthetic performance role, 113-114 plant morphology effects, 108-111 Cryptogams, ultraviolet-B radiation effects, sub-arctic heathlands, 202-204 Decomposition, ultraviolet-B radiation effects, sub-arctic heathlands, 203 Developmental processes, competitive success implications, 333-344 carbon dioxide responsiveness, 341-343 competitive ability, 334-336 developmental patterns, 334-336 effects prediction, 336-341 Disturbances, Ecology Diversity, Biodiversity Dominance, Species dominance Ecology fire disturbances, 231-245 carbon dioxide effects, 235-241 fuel composition, 237-239 fuel moisture content, 239-241 fire cycle prediction, 241-243 fire role, 232-235 research priorities, 243-244 functional groups, interspecific growth response variation, 390-391 prediction problems, 431-440 extrapolation, 432-434 holistic experiments, 436-439 reductionism experiments, 436-439 theory, 434-436 uncertainty, 439-440 Europe, vegetation changes, 85-91 carbon dioxide response, 88 current changes, 86-88 feedbacks, 89-90 functional plant types, 88 future prospects, 91 Evolution change potential, 13-21 experimental methods, 15-17 future research directions, 21 overview, 13-15 results, 17-20 climate change responses 23-30 genetics, 4-7 overview, 3-4 thermal sensitivity, 7-11
grassland plant population responses, 31-48 experimental design, 38-39 outlook, 47-48 results, 39-47 theory, 31-38 microevolutionary responses, 51-75 conceptual issues, 54-58 ecosystem process effects, 73-74 empirical data, 58-61 genetic correlation structure, 67-68 genetic variability, 54-61 heritability, 64-67 overview, 51-54 phenotypic variability, 62-64 quantitative genetic framework, 61-62 selection characteristics, 68-73 selection process effects, 61-68 Fire ecology, 231-245 carbon dioxide effects, 235-241 fuel composition, 237-239 fuel moisture content, 239-241 fire cycle prediction, 241-243 fire role, 232-235 research priorities, 243-244 Forage plants, quality, 363-369 carbon dioxide impact, 366-367 cattle production impact, 367-369 future research directions, 369 ruminant digestion, 364-366 Functional groups European vegetation changes, 88 interspecific growth response variation, 375-392 C3 species differences, 384-389 ecological aspects, 390-391 low nitrogen level differences, 389-390 methodology, 376-380 limitations, 380 protocol, 377 weight ratio aspects, 377-380 need, 375-376 photosynthetic pathway differences, 380-384 woody species differences, 389 plant-plant competition, 284 research usefulness, 170-172, 375-376 Fungi, community interactions, 265-271 consequences, 268-271 symbiotic responses, 266-268
Genetic variability, Evolution Grassland communities community responses, 139-155 analysis, 144-145 discussion, 151-155 aboveground litter, 153 carbon dioxide response, 153-155 nutrient effects, 151 seed dormancy, 151-152 soils, 152 experimental design, 140-142 methods, 142-144 results, 145-151 carbon dioxide effects, 148-150 community effects, 147-151 invasion, 150-151 nutrient effects, 147, 150 production totals, 145-147 competition grasses verses 301-315 dry mass results, 305-308 long term implications, 314-315 methods, 303-305 nutrient status, 308-311,314 phenology, 311-313 shoot morphology results, 308 species competitive abilities, 313 temperature effects, 311-313 theory, 301-303 legumes versus nonlegume growth responses, 287-297 intraspecific response variability, 293-296 methods, 288-289 nitrogen availability effects, 292-293 response variability, 289-292 versus 319-329 low light acclimation, 321-325, 329 shade tolerance, 320-326 species interactions, 327-329 species dominance, 159-174 aboveground biomass changes, 173-174 community composition consequences, 166-170 community responses, 164-166 diversity manipulation interpretation, 172-173 experimental design, 160-164
functional groups, 170-172 response variation, 174 genetic variation, 31-48 experimental design, 38-39 outlook, 47-48 results, 39-47 theory, 31-38 nitrogen fixation, 253-261 field studies, 258-259 model mechanics, 259-260 nitrogen availability, 255-258 plant growth effects, 255 symbiotic process, 254-255 ultraviolet-B radiation effects, sub-arctic heathlands, 202 woody plant invasion, 177-190 carbon dioxide influence, 181-190 seedling establishment, 185-189 vegetation change effects, 189-190 water availability interactions, 182-185, 188-189 overview, 177-181 Growth response grassland community competition, legume versus nonlegume growth responses, 287-297 intraspecific response variability, 293-296 methods, 288-289 nitrogen availability effects, 292-293 response variability, 289-292 interspecific variation, 375-392 Cs species differences, 384-389 ecological aspects, 390-391 functional types need, 375-376 low nitrogen level differences, 389-390 methodology, 376-380 limitations, 380 protocol, 377 weight ratio aspects, 377-380 photosynthetic pathway differences, 380-384 woody species differences, 389 Heathland, Sub-arctic heathland Herbaceous plants, canopy development, 413-428 418-420 418-420 future research directions, 427-428 leaf area index optimization, 420-426 light interception, 414-418
model, 415-416 photon partitioning, 416-418 nitrogen allocation, 420-426 Herbivory tropical plant communities, insect responses, 115-118 ultraviolet-B radiation effects, sub-arctic heathlands, 203-204 Heritability, Evolution Holistic experiments, ecological prediction problems, 436-439 Insects evolutionary climate change responses, genetics, 4-7 forest communities, 347-359 direct effects, 349-353 accelerated development, 351 altered biochemistry, 351 insect performance, 351-353 plant chemistry, 349-351 indirect effects, 354-355 interactive effects, 353-354 outbreak prediction, 356-357 recommendations, 357-359 resource availability, 353-354 tritropic interactions, 355-356 plant species dominance interactions, 115-118 ultraviolet-B radiation effects, sub-arctic heathlands, 203-204 Invasion, Competition grassland communities community responses, 150-151 woody plants, 177-190 carbon dioxide influence seedling establishment, 185-189 vegetation change effects, 189-190 water availability interactions, 182-185, 188-189 Leaf area index carbon metabolism, 221-222 competitive success role, tropical plants, 111-112 herbaceous plant canopy development, 420-426 Legumes competition grasses verses 301-315 dry mass results, 305-308
long term implications, 314-315 methods, 303-305 nutrient status, 308-311, 314 phenology, 311-313 shoot morphology results, 308 species competitive abilities, 313 temperature effects, 311-313 theory, 301-303 grassland legumes versus nonlegume responses, 287-297 intraspecific response variability, 293-296 methods, 288-289 nitrogen availability effects, 292-293 response variability, 289-292 nitrogen fixation, 253-261 field studies, 258-259 model mechanics, 259-260 nitrogen availability, 255-258 plant growth effects, 255 symbiotic process, 254-255 Light, Photosynthetic pathways; Shade tolerance competitive success role, tropical plants, 111-112 herbaceous plant canopy development, 414-418 model, 415-416 photon partitioning, 416-418 versus competition, low light acclimation, 321-325, 329 L., photochemical efficiency, 218-219 Livestock, Ruminant animals Mediterranean communities old-field microcosms, 123-136 community responses, 124-127 experimental designs, 124-126 species responses, 126-127 ecosystem responses, 127-131 carbon dioxide exchange, 127-129 community type interactions, 131 ecosystem processes, 133-135 nutrient fluxes, 129-131 soil microbiology, 131 standing biomass, 129 water fluxes, 129-131 individual species biomass production, 131-132 intraspecific response variability, 133
Mediterranean communities old-field microcosms (c0ntinued) phenology changes, 126-127, 132-133 reproduction changes, 127, 132-133 time scale considerations, 135-136 woodlands, 209-226 discussion, 223-226 experimental methods, 211-216 anatomical observations, 215 biochemical analyses, 215 fluorescence measurements, 214-215 gas exchange measurements, 213-214 growth measurements, 215-216 setup, 212-213 site selection, 211-212 results, 216-223 carbon dioxide uptake, 217-218 leaf anatomy, 221-222 nitrogen concentration, 220 photochemical efficiency, 218-219 shoot growth, 222-223 total nonstructural carbohydrates, 218-219 water availability, 216-217 Microbes, Bacteria; Fungi Mosses, ultraviolet-B radiation effects, subarctic heathlands, 202-203 Natural selection, Evolution Nitrogen concentration effects, Mediterranean woodlands, 220 fixation, grassland ecosystems field studies, 258-259 interspecific variability, 292-293 model mechanics, 259-260 nitrogen availability, 255-258 overview, 253-254 plant growth effects, 255 summary, 260-261 symbiotic process, 254-255 herbaceous plant canopy development, 420-426 interspecific growth response variation, low level differences, 389-390 Nutrients competitive success role annual grassland microcosms, 147, 150-151 grasses verses 308311,314
Mediterranean old-field microcosms, 129-131 tropical plant communities, 118 forest insect community effects, 349-351 Phenology competitive responses, grasses verses 311-313 Mediterranean old-field microcosms, 126-127, 132-133 Photosynthetic pathways, Light competition C3 versus C3 species, 278-282 C3 versus C4 species, 275-278, 282-283 grasses verses 301-315 dry mass results, 305-308 long term implications, 314-315 methods, 303-305 nutrient status, 308-311,314 phenology, 311-313 shoot morphology results, 308 species competitive abilities, 313 temperature effects, 311-313 theory, 301-303 grassland legumes versus nonlegume responses, 287-297 intraspecific response variability, 293-296 methods, 288-289 nitrogen availability effects, 292-293 response variability, 289-292 tropical ecosystems, community responses, 113-114 interspecific growth response variation, 380-384 competition, 319-329 interactions, 327-329 low light acclimation, 321-325, 329 shade tolerance, 320-326 carbon dioxide elevation response, 13-21 experimental methods, 15-17 future research directions, 21 overview, 13-15 results, 17-20 Plant chemistry, processes forest insects consequences, 349-351 Mediterranean woodlands community analysis, 215, 218-219 carbon dioxide responses experimental design, 38-39
results, 39-47 genotypic responses, 42-45 prediction consequences, 45-47 species responses, 39-42 L., carbon metabolism, 209-226 discussion, 223-226 experimental methods, 211-216 anatomical observations, 215 biochemical analyses, 215 fluorescence measurements, 214-215 gas exchange measurements, 213-214 growth measurements, 215-216 setup, 212-213 site selection, 211-212 results, 216-223 carbon dioxide uptake, 217-218 leaf anatomy, 221-222 nitrogen concentration, 220 photochemical efficiency, 218-219 shoot growth, 222-223 total nonstructural carbohydrates, 218-219 water availability, 216-217 Radiation,
Light; Ultraviolet-B radiation carbon dioxide elevation response, 13-21 experimental methods, 15-17 future research directions, 21 overview, 13-15 results, 17-20 Reductionism, ecological prediction problems experiments, 436-439 theory, 434-436 Root growth, competitive success role, tropical plants, 114-115 Ruminant animals, forage plant quality, 363-369 carbon dioxide impact, 366-367 cattle production impact, 367-369 future research directions, 369 ruminant digestion, 364-366 Seeds dormancy, annual grassland community responses, 151-152 establishment, woody plant invasion, 185-189
Shade tolerance,
Light
versus interactions, 319-329 species interactions, 327-329 tolerance mechanisms, 320-326 low light acclimation, 321-325, 329 tolerance changes, 326 Shoot growth grasses verses 308 L., carbon metabolism, 222-223 Shrubs, Woody plants Soil annual grassland community responses, 152 Mediterranean old-field microcosms, microbiology, 131 Species dominance calcareous grassland plants, 159-174 community responses, 164-166 discussion, 166-173 aboveground biomass changes, 173-174 community composition consequences, 166-170 diversity manipulation interpretation, 172-173 functional groups, 170-172 response variation, 174 experimental design, 160-164 insect effects, 115-118 tropical plant community responses, 103-107 Sub-arctic heathland, ultraviolet-B radiation effects, 197-205 decomposition, 203 herbivory, 203-204 methodology, 198-200 primary producer response cryptogams, 202-204 dwarf shrubs, 200-202 grasses, 202 species-specific responses, 204-205 site description, 198-200 Symbiotic interactions nitrogen fixation, grassland ecosystems, 253-261 field studies, 258-259 model mechanics, 259-260 nitrogen availability, 255-258 plant growth effects, 255 symbiotic process, 254-255
Symbiotic interactions (c0ntinued) plant-fungal interactions, 265-271 consequences, 268-271 symbiotic responses, 266-268 Temperature competitive effects, grasses verses 311-313 evolutionary climate change responses, 7-11 Trees, Woody plants grassland plant competition, 301-315 long term implications, 3 1 4 - 3 1 5 methods, 303-305 phenology, 311-313 results dry mass, 305-308 nutrient status, 308-311,314 shoot morphology, 308 species competitive abilities, 313 temperature effects, 311-313 theory, 301-303 Tritropic interactions, forest insects, 355-356 Tropical ecosystems, plant community responses, 101-119 community responses, 103-111 biomass allocation patterns, 108-111 competitive performance, 108-111 individually grown versus competitively grown plants, 107-108 plant morphology effects, 108-111 species shifts, 103-107 competitive success belowground interactions, 114-115 leaf area index, 111-112 photosynthetic performance role, 113-114 plant morphology effects, 108-111 individual responses, 102-103 insect herbivory responses, 115-118 leaf area index effects, 111-112 light effects, 111 - 112 recommendations, 118-119 Ultraviolet-B radiation, sub-arctic heathland, 197-205 decomposition, 203 herbivory, 203-204 methodology, 198-200 primary producer response
cryptogams, 202-204 dwarf shrubs, 200-202 grasses, 202 species-specific responses, 204-205 site description, 198-200 Water grassland communities, woody plant invasion, 182-185, 188-189 Mediterranean communities old-field microcosms, 129-131 woodlands, 216-217 Woody plants forest insect consequences, 347-359 direct effects, 349-353 accelerated development, 351 altered biochemistry, 351 insect performance, 351-353 plant chemistry, 349-351 indirect effects, 354-355 interactive effects, 353-354 outbreak prediction, 356-357 recommendations, 357-359 resource availability, 353-354 tritropic interactions, 355-356 functional groups, interspecific growth response variation, 389 grassland community invasion, 177-190 carbon dioxide influence, 181-190 seedling establishment, 185-189 vegetation change effects, 189-190 water availability interactions, 182-185, 188-189 overview, 177-181 Mediterranean ecosystems, 209-226 discussion, 223-226 experimental methods, 211-216 anatomical observations, 215 biochemical analyses, 215 fluorescence measurements, 214-215 gas exchange measurements, 213-214 growth measurements, 215-216 setup, 212-213 site selection, 211-212 results, 216-223 carbon dioxide uptake, 217-218 leaf anatomy, 221-222 nitrogen concentration, 220 photochemical efficiency, 218-219 shoot growth, 222-223
total nonstructural carbohydrates, 218-219 water availability, 216-217 versus competition, 319-329
low light acclimation, 321-325, 329 shade tolerance, 320-326 species interactions, 327-329 ultraviolet-B radiation effects, sub-arctic heathlands, 200-202
Physiological Ecology Series Editor Harold A. Mooney
Fakhri A. Bazzaz Robert W. Pearcy
Editorial Board F. Stuart Chapin James R. Ehleringer Martyn M. Caldwell E.-D. Schulze
T. T. KOZLOWSKI. Growth and Development of Trees, Volumes I and II, 1971 D. HILLEL. Soil and Water: Physical Principles and Processes, 1971 V. B. YOUNGER and C. M. McKELL (Eds.). The Biology and Utilization of Grasses, 1972 J. B. MUDD and T. T. KOZLOWSKI (Eds.). Responses of Plants to Air Pollution, 1975 R. DAUBENMIRE. Plant Geography, 1978 J. LEVITT. Responses of Plants to Environmental Stresses, Second Edition Volume I: Chilling, Freezing, and High Temperature Stresses, 1980 Volume II: Water, Radiation, Salt, and Other Stresses, 1980 J. A. LARSEN (Ed.). The Boreal Ecosystem, 1980 S. A. GAUTHREAUX, JR. (Ed.). Animal Migration, Orientation, and Navigation, 1981 F. J. VERNBERG and W. B. VERNBERG (Eds.). Functional Adaptations of Marine Organisms, 1981 R. D. DURBIN (Ed.). Toxins in Plant Disease, 1981 C. P. LYMAN, J. S. WILLIS, A. MALAN, and L. C. H. WANG. Hibernation and Torpor in Mammals and Birds, 1982 T. T. KOZLOWSKI (Ed.). Flooding and Plant Growth, 1984 E. I. RICE. Allelopathy, Second Edition, 1984
M. L. CODY (Ed.). Habitat Selection in Birds, 1985 R.J. HAYNES, I~ C. CAMERON, K. M. GOH, and R. R. SHERLOCK (Eds.). Mineral Nitrogen in the Plant-Soil System, 1986 T. T. KOZLOWSKI, P.J. KRAMER, and S. G. PALI_ARDY. The Physiological Ecology of Woody Plants, 1991 H. A. MOONEY, W. E. WINNER, and E.J. PELL (Eds.). Response of Plants to Multiple Stresses, 1991 F. S. CHAPIN III, R. L. JEFFERIES, J. F. REYNOLDS, G. R. SHAVER, and J. SVOBODA (Eds.). Arctic Ecosystems in a Changing Climate: An Ecophysiological Perspective, 1991 T. D. SHARKEY, E. A. HOLLAND, and H. A. MOONEY (Eds.). Trace Gas Emissions by Plants, 1991 U. SEELIGER (Ed.). Coastal Plant Communities of Latin America, 1992 JAMES R. EHLERINGER and CHRISTOPHER B. FIELD (Eds.). Scaling Physiological Processes: Leaf to Globe, 1993 JAMES R. EHLERINGER, ANTHONY E. HALL, and GRAHAM D. FARQUHAR (Eds.). Stable Isotopes and Plant Carbon-Water Relations, 1993 E.-D. SCHULZE (Ed.). Flux Control in Biological Systems, 1993 MARTYN M. CALDWELL and ROBERT W. PEARCY (Eds.). Exploitation of Environmental Heterogeneity by Plants: Ecophysiological Processes Above- and Belowground, 1994 WILLIAM K. SMITH and THOMAS M. HINCKLEY (Eds.). Resource Physiology of Conifers: Acquisition, Allocation, and Utilization, 1995 WILLIAM K. SMITH and THOMAS M. HINCKLEY (Eds.). Ecophysiology of Coniferous Forests, 1995 MARGARET D. LOWMAN and NALINI M. NADKARNI (Eds.). Forest Canopies, 1995 BARBARA L. GARTNER (Ed.). Plant Stems: Physiology and Functional Morphology, 1995 GEORGE W. KOCH and HAROLD A. MOONEY (Eds.). Carbon Dioxide and Terrestrial Ecosystems, 1996 CHRISTIAN KORNER and FAKHRI A. BAZZAZ (Eds.). Carbon Dioxide, Populations, and Communities, 1996
This Page Intentionally Left Blank