The Ecology and Silviculture of Oaks

The Ecology and Silviculture of Oaks The Ecology and Silviculture of Oaks Paul S. Johnson and Stephen R. Shifley US ...

Author: Paul S. Johnson | Stephen R. Shifley | Rogers Rogers

97 downloads 1199 Views 8MB 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

DOWNLOAD PDF

The Ecology and Silviculture of Oaks

The Ecology and Silviculture of Oaks

Paul S. Johnson and Stephen R. Shifley US Department of Agriculture Forest Service North Central Research Station Columbia, Missouri USA and

Robert Rogers College of Natural Resources University of Wisconsin/Stevens Point Stevens Point, Wisconsin USA

CABI Publishing

CABI Publishing is a division of CAB International CABI Publishing CAB International Wallingford Oxon OX10 8DE UK Tel: +44 (0)1491 832111 Fax: +44 (0)1491 833508 Email: [email protected] Web site: www.cabi-publishing.org

CABI Publishing 10 E 40th Street Suite 3203 New York, NY 10016 USA Tel: +1 212 481 7018 Fax: +1 212 686 7993 Email: [email protected]

© CAB International 2002. All rights reserved. No part of this publication may be reproduced in any form or by any means, electronically, mechanically, by photocopying, recording or otherwise, without the prior permission of the copyright owners. A catalogue record for this book is available from the British Library, London, UK. Library of Congress Cataloging-in-Publication Data Johnson, Paul S. The ecology and silviculture of oaks / Paul S. Johnson and Stephen R. Shiﬂey, and Robert Rogers. p. cm. Includes bibliographical references. ISBN 0-85199-570-5 (alk. paper) 1. Oak--United States. 2. Oak--Ecology--United States. I. Shiﬂey, Stephen R. II. Rogers, Robert, 1941 - III. Title. SD 397.O12 J64 2001 634.97210973--dc21 2001035883 ISBN 0 85199 570 5

Typeset in Melior by Columns Design Ltd, Reading Printed and bound in the UK by Biddles Ltd, Guildford and King’s Lynn

Contents

Preface Acknowledgements Introduction Conﬂicting Environmental Philosophies Silviculture: a Consilient Discipline References

xi xiii 1 1 5 7

Part I. Ecology 1 Oak-dominated Ecosystems Introduction The Taxonomy of Oaks The Geographic Distribution of US Oaks Species ranges and groupings Distribution of oaks by hierarchically classiﬁed ecoregions Eastern Oak Forests The Northern Hardwood Region The Central Hardwood Region The Southern Pine–Hardwood Region The Forest–Prairie Transition Region Western Oak Forests The Southwestern Desert–Steppe Region The Paciﬁc Mediterranean–Marine Region References

8 8 9 10 10 14 20 20 26 33 36 40 40 43 48

2 Regeneration Ecology I: Flowering, Fruiting and Reproduction Characteristics Introduction Flowering Male ﬂowers Female ﬂowers Factors Affecting Acorn Production Weather Premature abscission Variation in acorn production

54 54 55 55 58 61 61 62 64

v

vi

Contents

Acorn Predation and Dispersal Insects: destroyers of acorns Rodents: predation and dispersal Birds: predation and dispersal Oak Seedling Establishment Germination and initial establishment Early growth Seedling Sprouts Shoot dieback and root : shoot ratio Occurrence of shoot dieback Stump Sprouts and Related Growth Forms Deﬁnitions and origins Frequency of sprouting Sprout growth and mortality References

69 70 77 81 84 84 86 92 92 94 98 98 100 101 106

3 Regeneration Ecology II: Population Dynamics Introduction Regeneration Strategy Reproductive mechanisms: seeding and sprouting Accumulation of oak reproduction Fluctuation in population density Regeneration Potential Regeneration mode Modelling theory and objectives Stand-level regeneration models: purpose, problems and limitations References

117 117 118 118 121 142 147 148 155 157 158

Part II. Site Productivity and Stand Developmment 4 Site Productivity Introduction Measures of Site Productivity Relation of Site Productivity to Ecological Classiﬁcation Productivity and Related Self-sustaining Properties of Oak Forests Effects of timber harvesting on site productivity Modifying site productivity through fertilization Methods of Evaluating Site Quality Site index Site evaluation alternatives to site index References

168 168 169 171 172 173 175 176 176 185 189

5 Development of Natural Stands Introduction Forest Canopy Layers Disturbance Disturbance type Disturbance size and frequency of occurrence Development of Even-aged Stands The stand initiation stage

194 194 194 195 196 196 199 201

Contents

The stem exclusion stage The understorey reinitiation stage The complex stage Development of Uneven-aged Stands Disturbance–Recovery Cycles References

vii

203 207 215 216 217 224

6 Self-thinning and Stand Density Introduction Self-thinning Reineke’s model The 3/2 rule Stand Density and Stocking Terminology Maximum and minimum growing space Stand density diagrams References

227 227 227 227 229 235 235 237 241 251

Part III. Silviculture, Growth and Yield 7 Even-aged Silvicultural Methods Introduction Natural Regeneration Methods The clearcutting method The shelterwood method The seed tree method Artiﬁcial Regeneration Methods Oak nursery stock Oak plantation establishment Enrichment planting Intermediate Cuttings Deﬁnitions and theory Application Special Problems: Reducing Insect and Disease Impacts Gypsy moth Oak decline Oak wilt Economic, Environmental and Social Considerations The clearcutting method The shelterwood and seed tree methods References

254 254 254 255 255 274 277 278 278 280 284 294 294 295 305 305 314 316 320 320 322 322

8 Uneven-aged Silvicultural Methods Introduction The Single-tree Selection Method Principles of application Specifying the distribution Applicability to oak forests Group Selection Method Economic, Environmental and Social Considerations References

335 335 337 337 342 352 365 372 374

viii

Contents

9 Silvicultural Methods for Multi-resource Management Introduction Oak Savannas Extent and characteristics Disturbance processes Managing oak savannas Managing Stands for Acorn Production Assessing and predicting acorn crops Effects of tree size and stand characteristics Guidelines for sustaining acorn production Old-growth Oak Forests Extent and characteristics Silvicultural options Forests in transition to old growth Old-growth forests at the landscape scale Aesthetics Stand-level aesthetics Landscape-level aesthetics References

380 380 380 380 383 385 391 392 395 399 403 403 404 408 409 410 411 415 417

10 Growth and Yield Introduction Growth of an Oak Annual phenology Diameter growth Height growth Survival rates Stand Growth Growth and yield in even-aged stands Growth and yield in uneven-aged stands Growth and Yield Models Modelling methods Stand-level models for oaks Stand table projection models Individual-tree-level models for oaks OAKSIM TWIGS Forest Vegetation Simulator Estimating ingrowth Model evaluation Volume Equations Regional Patterns in Oak Yield and Productivity References

424 424 425 425 426 432 437 439 439 446 447 447 448 451 451 452 453 455 455 459 459 460 461

Appendices Appendix 1. Common and Scientiﬁc Names of Species Appendix 2. Forest Cover Types of Eastern United States Dominated by Oaks or Oaks Mixed with Other Species Appendix 3. Forest Cover Types of Western United States Dominated by Oaks or Oaks Mixed with Other Species

468 475 480

Contents

Appendix 4. Formulae for Converting Site Index for Indiana, Kentucky, Ohio, and West Virginia Appendix 5. Converting Site Indexes for Four Regions Appendix 6. Converting Site Indexes, Yellow-poplar to Oak Appendix 7. Height/dbh Site Index Curves Index

ix

483 484 486 487 489

Preface

The earth is to be seen neither as an ecosystem to be preserved unchanged nor as a quarry to be exploited for selﬁsh and shortrange economic reasons, but as a garden to be cultivated for the development of its own potentialities for the human adventure. (René Dubos, 1976)1

This book is written for forest and wildlife managers, ecologists, silviculturists, environmentalists, students of those ﬁelds, and others interested in sustaining oak forests for their many tangible and intangible values. The focus is on the oaks of the United States. Although the approach is fundamentally silvicultural, it is based on the premise that effective and environmentally sound management and protection of oak forests and associated landscapes should be grounded in ecological understanding. Although the subject is inherently scientiﬁc and technical, we have striven to make it generally accessible by minimizing the use of technical jargon. Where technical terms are necessary for efﬁcient expression of concepts, we have ﬁrst deﬁned them. Much has been written about the ecology and silviculture of oaks. So much so that the related body of literature represents, in one sense, an informational ‘embarrassment of riches’. The embarrassment derives primarily from the paucity of synthesis within and across two broad ﬁelds of study. The ﬁrst is ecology, which is the scientiﬁc study of the processes and relations among organisms and between 1Symbiosis

organisms and their environment including associated energy transformations. The second is silviculture, which is the art and science of producing, tending, and sustaining forests. Although the literature on North American oaks dates to the colonial period, most of it was written within the last 50 years, and a large proportion of that within the last 25 years. However, much of this literature resides in relatively obscure scientiﬁc and technical journals, proceedings of professional and scientiﬁc meetings, government publications, and other sources that are often difﬁcult to locate and retrieve. But even with ready access to this disparate information, its synthesis into an holistic framework of knowledge is a daunting task. This book attempts to ease, if not eliminate, those problems. Although ecology has become a household word, silviculture has not. Nevertheless, silvicultural practices have shaped the character of the landscape wherever oaks and associated forests occur, which includes much of the United States. Those practices often have produced negative public reactions and sometimes even deleterious ecological consequences. Increasing economic demand for oak wood nevertheless makes timber harvesting and its aftermath an ever more conspicuous feature of the landscape. Moreover, the distinction between designed silvicultural practices and purely exploitative logging practices is not always apparent, especially to the public.

between the earth and humankind. Science 193(4252), 459–462 (1976).

xi

xii

Preface

Contemporary philosophies on how oak forests and associated resources should be managed range from narrowly preservationist or narrowly utilitarian to more inclusive and integrative multiple-value philosophies. One of the objectives of this book therefore is to present ecological and silvicultural concepts that can be used to address an array of problems deﬁned by various perceptions of how oak forests should be treated. The current trend in managing forests and forested landscapes is away from a narrow focus on sustaining timber and other commodity outputs and towards a broader philosophy of sustaining desired ecological states. This shift in the forest management paradigm has been wrought by and is consistent with changing social values, scientiﬁc advances in ecology and society’s increasing awareness of environmental problems and expressed concerns on how those problems affect us collectively and individually. Consistent with the new paradigm, this book is designed and intended not so much as a how-to-do-it management manual as it is a source of ideas on how to think about oak forests as responsive ecosystems. Armed with that understanding, we believe managers and conservators of oak forests will be better positioned to adapt to changing social values and simultaneously to build and act on co-evolving ecological and silvicultural information. The book is divided into three sections. The ﬁrst contains three chapters on the ecological characteristics and distribution of oak species and the various kinds of oak forests in the United States, differences among them and how they have been classiﬁed, their natural development, and the relation of oak forests to environment and related environmental concerns. The next two chapters on regeneration ecology provide the critical interface between oak ecology and silviculture. Understanding the regeneration ecology of the oaks is paramount to silviculturists because of widespread difﬁculties in regenerating and thus sustaining oak forests.

The second section comprises three chapters covering site productivity and stand development. An understanding of the productive capacity of oak forests is central to a broad spectrum of issues related to their management and potentialities, not only for timber but also for wildlife and other values. The chapters on stand development, self-thinning and stand density present concepts that are key to the application of silvicultural methods. The third section comprises four chapters on silvicultural methods and the growth and yield of oak forests. Silivcultural methods include traditional even-aged and uneven-aged methods as well as non-traditional methods for multi-resource management and conservation. Regeneration methods are discussed in relation to the apparent regeneration strategies that have evolved in the oaks and how those strategies vary among oak-dominated ecosystems. The approach to regeneration thus is less prescriptive and more ecologically principled than that typically presented in silviculture textbooks and ‘how-to’ guides. Throughout the book, accepted common names of trees follow Little’s (1979) Check List of Native and Naturalized Trees of the United States. Scientiﬁc names of trees and other organisms are listed in Appendix 1. We express our appreciation and indebtedness to all the ecologists, foresters, wildlife biologists, soil scientists, entomologists, pathologists and others, past and present, who have contributed to our collective knowledge of the oaks. We are hopeful that this compilation will make some small contribution to a more ecosystem-centred approach to managing and conserving oaks in the many forests and plant communities in which they occur. Paul S. Johnson Stephen R. Shiﬂey Robert Rogers July 2001

Acknowledgements

The authors gratefully acknowledge the following for their invaluable assistance, without which this book would not have been possible. For their technical assistance and for creating a productive and congenial work environment, we are indebted to North Central Research Station technicians and support staffers Tim Bray, Dianne Brooks, Kenneth Davidson, Laura Herbeck, Kevin Huen, James Lootens, Marilyn Magruder, Allison Ramsey, Hoyt Richards and Neal Sullivan. Special thanks are accorded Lynn Roovers for countless hours spent on graphics and other details, and to William Dijak for his diligence and talent in preparing the maps for Chapter 1. For selflessly contributing their ideas, art work, photographs, reviews, and technical and scientiﬁc knowledge, we are deeply indebted to USDA Forest Service foresters and scientists Robert Cecich, Daniel Dey, Jeffrey Goelz, Gerald Gottfried, Kurt Gottschalk, David Graney, James Guldin, Jay Law, Edward Loewenstein, David Loftis, Robert McQuilkin, Ross Melick, Paul Murphy, Felix Ponder, Ivan Sander, Susan Stout, Richard Teck, Gary Z. Wang, Dale Weigel, Daniel Yaussy and John Zasada. We also thank all our partners in the University of Missouri School of Natural Resources for their valuable advice and support. Special thanks are accorded to Drs John Dwyer, H.E. Garrett, David Larsen, Bernie Lewis,

Nancy Loewenstein, Rose-Marie Muzika, Stephen Pallardy and Mr Dustin Walter for their reviews and ideas. A special thanks to Drs W. Carter Johnson, William Kearby and Walter Koenig for allowing us to reprint their marvellous photos, and Drs Jeffrey Ward (University of Connecticut), Carl Ramm (Michigan State University), William Parker (Ontario Forestry Research Institute), Willard Carmean (Lakehead University, Thunder Bay, Ontario) and an anonymous California reviewer for their constructive reviews and suggestions. Thanks to the Pioneer Forest of Salem Missouri and Leo Drey and Clint Trammel for allowing us to use their forest inventory data to evaluate uneven-aged oak silviculture. Thanks to the USDA Forest Service and the North Central Research Station Assistant Directors Donald Boelter and David Shriner, and Project Leader Frank Thompson, for allowing us the freedom to pursue and complete the project. Perhaps it goes without saying that the book could not have been written without the many researchers, past and present, who have contributed to the rich body of literature in oak ecology and silviculture; but we do not take their dedication and contributions to forest science for granted. Finally, we thank our families for their patience, forgone summer vacations and enduring support.

xiii

Introduction

Conflicting Environmental Philosophies Ecology is the scientiﬁc study of the interrelations among living things and their environment. Ecological knowledge effects an awareness of precarious interdependencies among the myriad organisms, large and minuscule, between organisms and non-living components of ecosystems, and the pervasive human impacts that threaten these relations. Ecology thus obviates our dependency on, and our relation to, natural processes and systems. Perhaps no science more so than ecology has generated more knowledge with implications relating to ethics, morality and human behaviour. In contrast, silviculture is the art and science of tending forests to meet human needs. Because silviculture is usually directly involved in the extraction of biomass, it produces disturbances along with associated ecological side effects. Silviculture is thus based on the planned use of controlled and directed disturbances to achieve deﬁned human objectives. Ideally, it should be based on scientiﬁc principles which ensure that speciﬁed silvicultural goals are consistent with preserving or improving a forest’s ecological qualities, are compatible with its natural dynamic and thereby provide reasonable assurance of the forest’s sustainability. Like its parent discipline, forestry, silviculture evolved out of 17th century Europe in response to purely utilitarian needs, especially for the timber required

for sustaining the large naval armadas required for projecting colonial power in the late 18th century. Paramount among these concerns in Britain and France was a ready supply of pine and oak for ship masts and hulls. However, in the United States, serious concern over a declining forest resource did not occur until the late 19th century. By then the forests of eastern United States had been decimated by exploitative logging. A small but politically inﬂuential group of conservationists feared the same would happen to the western forests. This prompted the setting aside of forest reserves in the early 1890s from what remained of the public domain in the west. In 1897, the Organic Act was passed, which speciﬁed that the purpose of the reserves was ‘to improve and protect the forest within the reservation, or for the purpose of securing favourable conditions of water ﬂows, and to furnish a continuous supply of timber for the use and necessities of citizens of the United States’ (United States Congress, 1897). This landmark legislation speciﬁed that the forest reserves were intended for managed use, not for wilderness preservation. Following the recommendations of the American Forest Congress of 1905, the reserves were transferred from the Department of Interior to the Department of Agriculture. Known as the Transfer Act, it provided that funds from the sale of products or the use of land in the reserves be used for managing and developing the forest reserve system. 1

2

Introduction

This change heralded the implementation of an ambitious programme of scientiﬁc forest management under the direction of Gifford Pinchot, the ﬁrst Chief of the USDA Forest Service. At that time, forestry was virtually an unknown discipline in the United States and forestry curricula in US universities were just emerging. Although politically controversial in its day, the conservation movement was hailed by its founders as not only environmentally wise, but also economically beneﬁcial (Pinchot, 1987). Pinchot and the founders of the early forest conservation movement envisioned a scientiﬁcally based forestry that would not only provide conservation beneﬁts but would also result in the economic stability of rural communities in forested regions. Such beneﬁts would accrue, they argued, from the application of scientiﬁcally derived sustained yield principles, which would ensure for perpetuity the even ﬂow of timber and other commodities originating from the forest (Pinchot, 1987). Because the scientiﬁc underpinnings of sustained yield were largely invested in silviculture, and because silviculture has historically been justiﬁed on economic grounds, silviculture philosophically straddled agronomy (i.e. growing trees as crops) and economics. However, modern silviculture has been broadened to include not only sustaining timber yields, but also sustaining non-commodity values including oldgrowth forests, biodiversity, wildlife habitat and aesthetics. In this wider context, silviculture assumes application to a panoply of values that transcend economic utilitarianism. Despite the differences between the two disciplines, contemporary silviculture as it has been applied to most North American forests, remains naturally allied with and dependent upon ecology for much of its scientiﬁc underpinnings. The schism between silviculturists and some ecologists nevertheless runs deep. One source of this disunion emanates from the ecologists’ traditional focus on studying ecological processes in ecosystems largely unaffected or minimally affected by

humans and drawing conclusions therefrom. In contrast, silviculturists depend on scientiﬁcally based knowledge of disturbance-mediated mechanisms to control and direct forest ecosystem processes for human beneﬁt. Recovery from such disturbances is predicated on the assumption that forests are inherently resilient, i.e. capable of rapidly returning to their previous or other silviculturally directed state. The silviculturist’s anthropocentric view of the forest is anathema to those who adhere to the biocentric view, which elevates nature to a position superior to human self-interest (Devall and Sessions, 1985; Chase, 1995; Ferry, 1995; Fox, 1995). The biocentrist’s agenda is centred on maintaining ‘natural’ ecosystems, including forest, in states free from human interference, and the need for establishing the pre-eminence of those states. From that perspective, human-mediated disturbance is seen as a disrupter of fragile ecosystems and the intended order of things. Moreover, such disruptions can potentially produce species extinctions and other irreversible environmental effects. The biocentric view therefore holds that the best way to preserve nature, wherever some vestige of it remains, is to leave it alone (Devall and Sessions, 1985; Chase, 1995). Humans are viewed as just one of many organisms in the biosphere no more important than any other – and like all component organisms should be subordinate to the healthy functioning of the interactive whole, i.e. the ecosystem. Biocentrism is therefore egalitarian among organisms and premised on an inherent right to life of all species and life forms. By extension, maintaining ecosystems in their ‘natural’ state becomes a social imperative. A biocentrist thus may view silviculture, along with other human interferences in the development of forests, as ecologically threatening, if not ruinous. The biocentric interpretation of the ‘message’ from ecology is thus at irreconcilable odds with the interpretation from silviculture. Biocentrism nevertheless now occupies a position of social and political prominence (Chase, 1995).

Introduction

The connections between ecology and silviculture none the less are apparent and important, especially when silviculture is applied to forests of natural origin. In that setting, silviculture by itself may not introduce new species or populations (i.e. new genetic material) from outside the forest. Human energy expenditures are often limited only to those required in cutting and removing trees. Such relatively non-intensive practices have characterized the silviculture applied to oak forests of the United States. There, oak silviculture has largely followed an ecological model whereby forests are managed by directing their continually changing states, or ecological successions, through manipulation of existing on-site vegetation and propagules. This approach relies on periodic timber harvesting and usually natural regeneration to maintain or periodically recreate desired ecological states. It contrasts with the more intensive agronomic model used in growing pine plantations and other monotypes. The latter approach usually depends on artiﬁcial regeneration, the introduction of new and ‘improved’ genotypes, exotic species, and other intensive and energy-expensive cultural methods like those used in agriculture, horticulture and agroforestry (growing trees intermixed with agricultural or horticultural crops). Nevertheless, the silvicultural methods that have been applied to oaks span the entire range of approaches from ecological to agronomic. In the public’s view, silviculture is an often confusing and controversial subject exacerbated by the claims of some environmentalists that it is an ecologically damaging enterprise that ‘seeks to accept “tree farms” in place of natural forests … The usual approach … is to seek ever more intensive management, which spawns even more problems’ (Devall and Sessions, 1985, p. 146). By comparison, there is seemingly little controversy and confusion over the reason to preserve something in its natural state free from human interference if it is otherwise threatened with extinction – even though the method or means of preservation may be debatable. Likewise, the reason for the cultivation and harvest of a corn ﬁeld is easily understood and

3

accepted because of its purely utilitarian value, and its physical origins borne of human endeavour. Socially, silviculture is a more complicated issue. It is vulnerable in appearance, conceptually and often physically, seen as conforming to neither preservation nor agronomy. It is neither ﬁsh nor fowl, yet is often identiﬁed as disruptive if not exploitative of nature. To the non-silviculturist, application of the ecological model to silviculture may sometimes be difﬁcult to distinguish from purely exploitative and environmentally damaging practices. However, such exploitation is not the intent of, nor does it constitute, silviculture. Silviculture is not synonymous with timber harvesting, yet is dependent upon it. The objective of modern silviculture is to create and maintain forests by design that produce material and nonmaterial beneﬁts to humans without sacriﬁcing their sustainability. Silvicultural intentions nevertheless are not ecologically infallible. A given silvicultural application, despite best intentions, may be inconsistent with ecological realities because of our incomplete knowledge and understanding of ecosystems. Poorly applied silviculture therefore can produce unintended and negative long-term ecological consequences. The possibilities for such outcomes impose serious responsibilities on silviculturists in the practice of their art and science. When silviculture is applied to ‘natural’ ecosystems, the intent, some would say, is to improve on nature by tinkering with it. But the biocentrist would argue that humans cannot improve upon nature – a notion consistent with the theological view that ‘man cannot improve upon God’s handiwork’. And much ecological knowledge and theory is purported to support that perception. Perhaps it is the proximity of the existing ‘near-natural’ state to the intended silviculturally created state that concerns those whose sentiments might be to ‘leave well enough alone’. The biocentrist might argue that silviculture promises only ‘a kinder, gentler rape of the forest’. These views may be further bolstered by an awareness of the shrinkage of natural ecosystems globally and its consequences.

4

Introduction

The fragmentation of today’s landscape into discrete blocks of forests spatially detached from human development may further reinforce the perception of the separation of humans and forest. This outlook is reﬂected in the Latin origin of the word forest, foris, which means outside. This etymology suggests a human view of forests evolving from deep historical and psychological roots, and one in which forests are functionally disconnected from humans. Even as late as the 18th century, the forest was perceived as something ‘beyond’ the boundary of European culture (Bonney, 1996). Today, most of the US population resides in urban areas. There, sources of basic human-sustaining resources are physically distanced from and foreign to everyday experience. A perception of forests as functionally and spatially distant from humans may be further reinforced by the common acknowledgement that, in some cases, the physical separation of humans from nature is necessary to preserve rare or endangered species and habitats. It is generally accepted that such preservation is a democratically mandated function of government. The results are commonly and favourably experienced annually by millions of visitors to national and state parks, wildlife refuges, and designated wilderness areas in national forests and other federal lands. It is generally understood that the role of humans there is restricted to that of protector and spectator, but not interloper. Contrasting with such models of the separation of humans and nature is the historical relation between humans and oaks, which are characterized by connectedness. From the oak’s perspective, those connections have produced both beneﬁcial and harmful effects. The ecological evidence, as later discussed, nevertheless indicates that sustaining and thus preserving many oakdominated ecosystems will require human intervention. Humans and oaks have been closely associated throughout history. Before the arrival of Europeans, native Americans set ﬁres, both accidentally and intentionally, which often burned out of control over enormous areas (Grimm, 1983;

Pyne, 1982, 1997; Guyette et al., 1999). Periodic ﬁres were repeated over centuries in regions indigenous to the oaks, which includes much of North America. Their frequent occurrence created extensive areas of open-grown forests favourable to the survival of the relatively light-demanding but ﬁre-tolerant oaks. It was a disturbance cycle that, in time and space, is unlikely to be repeated. Humans thus have had a prominent effect in shaping the nature and extent of the oak’s habitat, and perhaps even its evolution. But those events have been largely relegated to history. Much of what today remains of the oak forests of the United States is a legacy of an earlier disturbance history that was partially, if not largely, dependent on ﬁre. Unlike the ecologist, the silviculturist has traditionally viewed forests from a utilitarian perspective that emphasized timber production. Accordingly, failure to harvest forests at their inherent sustainable capacity to produce wood (sustained timber yield) is deemed wasteful. A theological counterpart is seemingly expressed by the biblical admonition for man to exert dominion over the earth (Genesis 1:28). An economic analogue is expressed in Adam Smith’s 1776 treatise on the inherent value of the individual pursuit of economic selfinterest (Smith, 1870). Collectively, these beliefs and values, largely borne of the Enlightenment, have dominated the thinking and institutions of Western civilization for over 200 years. Self-interest prevails among private forest owners today, whether ownership goals are economic or non-economic. To a lesser extent, economic objectives dominate the management of many publicly owned lands, including the national forests. In the past, agency mandates, operating budgets and incentives tied to timber sales, produced powerful inducements to emphasize timber production, albeit within calculated sustained yield limits. Only within the last few decades has this philosophy been seriously challenged. Such material utilitarianism reduces forests to collections of trees having only commercial value. Other values are consequently diminished. The

Introduction

American conservationist, Aldo Leopold (1966, p. 251) expressed concern for this philosophy by asserting that ‘… a system of conservation based solely on economic self-interest is hopelessly lopsided. It tends to ignore, and thus eventually to eliminate, many elements in the land community that lack commercial value, but that are (as far as we know) essential to its healthy functioning.’

Silviculture: a Consilient Discipline The practice of silviculture therefore is caught in a web of competing values arising from different philosophies, ranging from biocentrism to economic utilitarianism. Unlike ecology, silviculture is directly connected to social institutions and conventions apart from science. Lying within its parent discipline, forest management, it is subject to the legal and social constraints of environmental law and policy operating within democratic processes (at least in the United States and other democratic countries where silviculture is practised). Within the context of democracy, silviculture is therefore socially integrative, i.e. in its application it must consider values borne of diverse social and political interests. Silviculture nevertheless lies at the core of forest resource management because its application results in direct physical action on the forest. This is also where fundamental scientiﬁc analysis is most needed. However, silviculture does not stand ﬁrmly by itself as a scientiﬁc discipline. This results in part from its strong connections to social and political institutions, and in part from its interdisciplinary qualities as a science. Within the biological domain of science, silviculture is most closely allied to ecology. However, it is also heavily dependent on plant physiology and genetics, plant pathology, entomology, and applied mathematics and statistics. Among the physical sciences, it borrows knowledge from geology, climatology, hydrology and soil science. It is also closely allied to other resource management disciplines including wildlife, ﬁsheries, water and air

5

quality management. Silviculture therefore is inherently scientiﬁcally integrative. Silviculture consequently depends on linking knowledge and theories across many disciplines, both scientiﬁc and nonscientiﬁc, to form what Wilson (1998) terms ‘… a common groundwork of explanation’. If we accept that such linkages comprise consilience, we might consider that silviculture ﬁts Wilson’s context, i.e. it comprises a hybrid domain of knowledge in which consilience is implicit. Because of silviculture’s socioeconomic connections, this consilience extends to other branches of learning including the social sciences and humanities. These connections can be represented by a series of concentric circles representing the social hierarchies within which silviculture exists. With silviculture at its centre, each ring of the social hierarchy bounds all the great areas of knowledge, including the biological, physical and social sciences as well as the humanities (Fig. I.1). This representation emphasizes the consilient nature of silviculture by placing it at the locus of all knowledge comprising its context. It represents an ideal, a unity of learning in which subjects that have been traditionally compartmentalized are breached in Wilson’s words, to ‘… provide a balanced clearer view of the world as it really is … A balanced perspective cannot be acquired by studying disciplines in pieces but through pursuit of the consilience among them … The enterprise is important for yet another reason: it gives ultimate purpose to intellect’ (Wilson, 1998, p. 13). Despite the complexities of silviculture’s complete context, our intent in the following pages is to present a synthesis of the ecological and silvicultural knowledge of oak forests in the United States. It is not to resolve the environmental issues surrounding oak forests, which fall into the social, legal, political and managerial domains represented by the concentric circles surrounding silviculture in Fig. I.1. The silvicultural context is nevertheless broad, and not limited to narrowly deﬁned economic or commodity-production objectives. Consistent with the view of silviculture

6

Introduction

Humanities

Physical sciences cy ra oc l em ta D en licy nm po ro d vi an En w la

t st en re m Fo age an

m t st en re m Fo ge l a ta an en licy m nm po ro d vi n a En w cy la ra oc

em

D

Social sciences

Biological sciences

Fig. I.1. Silviculture’s relation to other disciplines. Concentric circles represent the social hierarchy within which silviculture exists in a democratic society. With silviculture at its centre, each ring of the hierarchy bounds all the major areas of knowledge, including the biological, physical and social sciences, and the humanities.

as a consilient discipline, we view the role of the silviculturist as just one of many possible players in the management of oak forests. Unlike the biocentrist, we infer no moral imperative to create or maintain oak forests in speciﬁed states other than those that are perceived, as best we can discern, as sustainable, beneﬁcial, and pleasing to humankind, and that provide habitat for the many plant and animal species naturally associated with oaks. We believe these goals are consistent with the philosophy of land stewardship and wise use as proposed by earlier generations of conservationists, from which the more recent philosophy of ecosystem management has evolved. Our intention is to present information that can lead to an understanding of, and solutions to, silvicultural problems

related to oak forests. Moreover, we hope that this information fosters an informed and amiable dialogue and trust among foresters, land managers and owners, environmentalists, students and others interested in oak forests. The subject therefore is presented from a silvicultural perspective. The approach comprises a comprehensive view of forests as providing important social, spiritual and economic needs. Such an approach requires anticipating and managing for change, both predictable and unpredictable. This notion is consistent with Botkin’s (1990) call for a ‘new management’, wherein conservation and utilization of forest resources are compatible parts of an integrated ecosystem approach. It contrasts with the ‘old management’ in

Introduction

which conservation was too often subordinate to timber and other commodity production. The central concern of the new forest management, or ecosystem management (Salwasser, 1994), is the sustainability of forested ecosystems and associated human values in a continually changing mosaic of landscape patterns. The resulting

7

management and silviculture therefore must accommodate the complexities of the inevitably and continually changing ecological states that comprise a forested landscape. It also recognizes that such changes occur with or without human interference, and that we have both potentialities and limitations in controlling these changes.

References Bonney, W. (1996) Troping trees. In: Schultz, K.L. and Calhoon, K.S. (eds) The Idea of the Forest. Peter Lang, New York, pp. 119–146. Botkin, D.B. (1990) Discordant Harmonies. Oxford University Press, New York. Chase, A. (1995) In a Dark Wood: The Fight Over Forests and the Rising Tyranny of Ecology. Houghton Miﬂin, Boston. Devall, B. and Sessions, G. (1985) Deep Ecology. Gibbs M. Smith, Layton, Utah. Ferry, L. (1995) The New Ecological Order. University of Chicago Press, Chicago. Fox, W. (1995) Toward a Transpersonal Ecology. State University New York Press, Albany, New York. Grimm, E.C. (1983) Chronology and dynamics of vegetation change in the prairie-woodland region of southern Minnesota, U.S.A. New Phytologist 93, 311–350. Guyette, R., Dey, M. and Dey, D.C. (1999) An Ozark ﬁre history. Missouri Conservationist 60, 4–7. Leopold, A. (1966). A Sand County Almanac. Ballantine, New York. Pinchot, G. ([1947] 1987) Breaking New Ground. Island Press, Washington, DC. Pyne, S.J. (1982) Fire in America. Princeton University Press, Princeton, New Jersey. Pyne, S.J. (1997) America’s Fires: Management on Wildlands and Forests. Forest History Society, Durham, North Carolina. Salwasser, H. (1994) Ecosystem management: can it sustain diversity and productivity? Journal of Forestry 92(8), 6–10. Smith, A. (1870) An Inquiry Into the Nature and Causes of the Wealth of Nations. London. United States Congress (1897) Surveying the public lands. US Statutes at Large 30, Ch. 2, pp. 32–36. Wilson, E.O. (1998) Consilience: The Unity of Knowledge. Knopf, New York.

1 Oak-dominated Ecosystems

Introduction A truly ecological perspective recognizes that humans and their activities are part of nature, and that enhancing all aspects of their lives – including their surroundings – begins with cooperation between individuals, based on mutual trust … Rather than halting or reversing disturbances, we should embrace change. Rather than excluding man from the garden, we should welcome his cultivation of it. (Alston Chase, 1995)

From earliest times, oaks have held a prominent place in human culture. Their uses have included wood for fuel, acorns for hog fodder and ﬂour meal for human consumption, bark for tanning, wood strips for weaving baskets, charcoal for smelting ore, timbers for shipbuilding, mining timbers, railroad ties, pulpwood for paper, and lumber and laminates for furniture, panelling and ﬂooring. Through the mid-19th century, oak was the wood of choice for shipbuilding in Europe and America. For that reason, oak forests and even individual trees were treated as critical national assets. During the Revolutionary War, the poor condition of the British ﬂeet, which lacked replacements and repairs due to shortages of suitable oak timbers, may have contributed to the war’s outcome (Thirgood, 1971). In the 17th century, alarm over the depletion of timber supplies, especially oak, prompted passage and enforcement of laws mandating the protection, culture and establishment of forests in several European countries. In turn, those events inﬂuenced the development of scientiﬁc silviculture, as we know it today. 8

Modern as well as ancient man has beneﬁted from the oak’s relation to wildlife. Wherever oaks occur as a prominent feature of the landscape, wildlife populations rise and fall with the cyclic production of acorns. Numerous species of birds and mammals are dependent on acorns during the food-scarce autumn and winter months. Even human cultures have relied on oaks as a staple food. Acorns were an important part of the diet of Native Americans in California before the 20th century (Kroeber, 1925) (Fig. 1.1). Today, the ecological role of oaks in sustaining wildlife, biodiversity and landscape aesthetics directly affects the quality of human life. The demand for wood products from oaks nevertheless continues to increase and compete with other less tangible values. Some have proposed that forests, including those dominated by oaks, are best allowed to develop naturally, free from human disturbance. What should the balance be among timber, wildlife, water, recreation and other forest values? Is there some middle ground that adequately sustains multiple goals? Informed answers and perspectives require an understanding of the ecology of oaks and the historical role that humans have had in that ecology, especially the comparatively recent role of humans in the ‘protection’ of oak forests from ﬁre. A prerequisite to such understanding is a general knowledge of the oak’s geographical occurrence, taxonomic diversity, adaptations to diverse environments, and the historical changes in its environment.

Oak-dominated Ecosystems

Fig. 1.1. Native American collecting acorns as shown in Hutchings’ California Magazine in 1859. Acorns were a staple food of most California tribes before the end of the 19th century. They were gathered in conical woven baskets, which could hold a bushel or two of the nuts. Although the acorns of many species were eaten, favoured species were California black and California live oaks (Pavlik et al., 1991). After removing the shell (pericarp), acorns were ground into a flour, leached of tannins by soaking in running water, and then used to make a variety of foods including porridge and bread. Acorns were so highly valued that they sometimes provoked inter-tribal ‘acorn wars’. They were also widely utilized as food by Native Americans in the eastern United States. (Courtesy of the Bancroft Library, University of California, Berkeley.)

The Taxonomy of Oaks Taxonomically, the oaks are in the genus Quercus in the family Fagaceae (beech family). The Fagaceae probably originated in the montane tropics from which its members migrated and diverged into the current living genera by the late Cretaceous period (about 60 million years ago) 1

(Axelrod, 1983). By that time, mammals and birds had only recently evolved. Rapid speciation of oaks commenced in the middle Eocene epoch (40–60 million years ago). This was in response to the expansion of drier and colder climates, and subsequently to increased topographic diversity in the late Cenozoic era (< 20 million years ago) and ﬂuctuating climates during the Quaternary period (< 2 million years ago) (Axelrod, 1983). Their fruit, the acorn, distinguishes the oaks from other members of the beech family (e.g. the beeches and chestnuts). With one exception, all plants that produce acorns are oaks. The exception is the genus Lithocarpus, which includes the tanoak of Oregon and California. Although represented by only one North American species, Lithocarpus is represented by 100–200 species in Asia (Little, 1979). Lithocarpus may be an evolutionary link between the chestnut and the oak (McMinn, 1964; cf., Miller and Lamb, 1985, p. 200). Worldwide there are about 400 species of oaks, and they are taxonomically divided into three groups: (i) the red oak group (Quercus section Lobatae1); (ii) the white oak group (Quercus section Quercus2); and (iii) the intermediate group (Quercus section Protobalanus3) (Tucker, 1980; Nixon, 1997). All three groups include tree and shrub species. The red oaks and white oaks include evergreen and deciduous species, whereas the intermediate oaks are all evergreen. The red oaks are found only in the Western Hemisphere where their north–south range extends from Canada to Colombia. In contrast, the white oaks are widely distributed across the Northern Hemisphere. The intermediate group comprises only ﬁve species, all of which occur within southwestern United States and northwestern Mexico. Many of the world’s oaks occur in regions with arid climates, including Mexico, North Africa and Eurasia, where they are often limited in stature to shrubs and small trees. About

Subgenus Erythrobalanus in earlier classifications. Subgenera Lepidobalanus and Leucobalanus in earlier classifications. 3 Subgenus Protobalanus in earlier classifications. 2

9

10

Chapter 1

80% of the world’s oaks occur below 35° north latitude and fewer than 2% (six or seven species) reach 50° (Axelrod, 1983). The most reliable distinction between the white oaks and red oaks is the inner surface of the acorn shell. In the white oaks it is glabrous (hairless) or nearly so, whereas in the red oaks it is conspicuously tomentose (hairy or velvety) (Tucker, 1980). In the intermediate group, this characteristic is not consistent among species. The leaves of the white oaks are usually rounded and without bristle tips whereas the leaf lobes of the red oaks are usually pointed and often bristle-tipped. To many silviculturists, ecologists and wildlife biologists, the most important difference between the white oaks and red oaks is the length of the acorn maturation period. Acorns of species in the white oak group require one season to mature whereas species in the intermediate and most of the red oak group require two seasons. The white oaks and intermediate oaks are characterized by the presence of tyloses (occlusions) in the latewood vessels (waterconducting cells) whereas tyloses are usually absent in the red oaks. These vesselplugging materials confer greater decay resistance to the wood of the white and intermediate oaks than the red oaks. Other morphological features that differentiate the three groups and species within them are presented in various taxonomic treatments (e.g. Tucker, 1980; Jensen, 1997; Manos, 1997; Nixon and Muller, 1997) and ﬁeld identiﬁcation guides (e.g. Miller and Lamb, 1985; Petrides, 1988; Petrides and Petrides, 1992). These sources also include range maps. In addition, the Silvics of North America, Vol. 2 (Burns and Honkala, 1990) provides information on the silvics and geographic ranges of 25 oaks. Of the more than 250 oak species occurring in the Western Hemisphere, the largest number occurs in Mexico and Central America. About ten species occur in Canada. For the United States species, the most complete and authoritative taxonomic treatment of the oaks is in the Flora of North America North of Mexico, Vol. 3 (Flora of North America Editorial

Committee, 1997), which lists 90 species of oaks native to the continental United States. However, we follow the taxonomic nomenclature of Little’s (1979) Checklist of United States Trees because of its widespread use in North American forestry literature. This checklist recognizes 58 native oak species plus nine varieties. Of these, about ten species are shrubs or shrub-like forms. More than 80 hybrids also have been described (Little, 1979; Tucker, 1980).

The Geographic Distribution of US Oaks Species ranges and groupings The oaks are widely distributed across the United States (Fig. 1.2). According to Little (1979), about 40 species and varieties occur east of the 100th meridian and about 30 species and varieties occur to the west. Only two species, chinkapin oak and bur oak, are common to both regions. Bur oak extends to the northwest whereas chinkapin extends to the southwest beyond the 100th meridian. The western oaks fall into three geographically distinct groups. One group is comprised of the west Texas oaks (nine species and varieties), and a second includes the southwestern oaks (16 species) that occur in New Mexico, Arizona, Utah, Colorado and Nevada. A third group is comprised of the Paciﬁc Coast oaks (about 13 tree species plus several shrubby species) occurring largely in California, Oregon and Washington. Within the United States, numbers of oak species vary regionally. Based on a count of the number of oak species that occur within 6000 square mile areas, oak species ‘richness’ reaches a maximum of 20 species in the southeast (Aizen and Patterson, 1990) (Fig. 1.3). There, the ranges of several narrowly distributed North American oak species overlap with the ranges of several widely distributed species. Although the range of an oak species is positively correlated with its acorn size, the reason for this is unknown (Aizen and Patterson, 1990).

Oak-dominated Ecosystems

11

100 M21tb

M331

M241

211

332

241

263 M261

211

M333 M334

211

22tb

341

221a 341

342 261

211

M334

M332 M241

221a

251

M341

262

331

M341 M334

M341 342

M262

221b

M221

332 323

M222

311 M311

313 232

M311 314

321

252

M231 232 231

232 231

411

Fig. 1.2. The distribution of oaks in conterminous United States. The shaded areas represent the aggregated vegetation cover types within which oaks frequently occur as important species at a scale of 1 km2. The 100th meridian demarcates the approximate division between eastern and western oaks. Generated from Advanced Very High Resolution Radiometer satellite images (1990) and an associated system of land cover classification (USDA Forest Service, 1993; Powell et al., 1994). Ecoregion boundaries are from Bailey (1997). See Tables 1.2 and 1.3 for index to numbered ecoregions and oak species found in each. (Map compiled by W.D. Dijak, USDA Forest Service, North Central Research Station, Columbia, Missouri.)

Forest cover types (or simply cover types) are combinations of tree species that tend to spatially reoccur at stand-level scales (e.g. < 100 acres). The resulting categories are thus silviculturally useful in differentiating among different kinds of oak stands. Categorization of United States forests based on deﬁned cover types was begun by the Society of American Foresters in 1929. There are 145 deﬁned cover types in the United States and Canada (Eyre, 1980). These include 31 with ‘oak’ in the cover type name or in the list of species that deﬁne the type (Appendices 2 and 3). Of these, 23 oak types occur east and eight occur west of the 100th meridian. In addition, many of the non-oak cover types include one or more oak species as common associates. The geographic extent of individual cover types ranges from tens of millions of acres (e.g. the white oak–black oak–northern red oak cover type of the eastern US) to rel-

atively restricted areas (e.g. the northern pin oak cover type of the upper Lake States and the Mohr oak cover type of Texas and Oklahoma). Other types such as the live oak type of the South and the bur oak cover type in the Great Plains occur within long narrow belts associated with coastal plains and river corridors, respectively. Many of the western oak cover types, especially those in California, form belts that follow the Coastal and Sierra Nevada mountain ranges and foothills surrounding the Central Valley. Oaks occur in environments ranging from extremely wet and humid (e.g. the overcup oak–water hickory cover type of southern ﬂood plains), to mesic (moist) upland forests receiving 50 or more inches of precipitation per year (e.g. the yellow–poplar–white oak–northern red oak cover type), to Mediterranean climates that receive 10 inches or less precipitation per year (e.g. the blue

12

Chapter 1

Fig. 1.3. The geographic distribution of numbers of oak species in eastern United States and Canada. The isolines were drawn from a grid comprised of 78 ⳯ 78 square mile cells within which the number of oak species were counted based on Little’s (1971, 1977) range maps. The greatest concentration of oak species (15 to 20) occurs in the southeastern United States where the ranges of several narrowly distributed species overlap the ranges of several widely distributed species. (Redrawn from Aizen and Patterson, 1990, used with permission.)

oak–digger pine cover type). Oaks occur in even drier climates where they form shrub vegetation such as the chaparral of southern California and the semi-desert scrub woodland vegetation of the interior southwest. Western cover types such as the canyon live oak cover type include closed-canopy stands in the northern part of their range and savanna-like woodlands in the south. Oak forests therefore range from closed canopy upland and lowland forests with trees greater than 120 ft tall to xeric (droughty) scrublands dominated by dwarf trees and shrubs. Some oaks, such as Georgia oak and McDonald oak are conﬁned to very small geographical ranges and a narrow range of habitat conditions. Others such as white

oak are widely distributed and occur over a broad range of climates and habitat conditions. A species’ ﬂexibility in occupying different habitats is implicit in the deﬁnition of species niche. The term denotes the speciﬁc set of environmental and habitat conditions that permit the full development and completion of the life cycle of an organism (Helms, 1998). The oaks occupy many niches because of the wide range of environmental conditions within which they can collectively occur. However, the niche of an individual species is more limited. Niche differentiation among the oaks and associated species is often evident from the way species segregate along environmental gradients such as the soil moisture gradient (Fig. 1.4). Oaks also differ in their ecologi-

Oak-dominated Ecosystems

cal amplitude, i.e. the range of habitat conditions that a species can tolerate (Allaby, 1994). The ecological amplitude of a species often forms a bell-shaped curve when illustrated diagrammatically (Fig. 1.4). However, some species, such as bur oak, occur in both bottomlands and dry uplands but are nearly absent at intermediate points along the moisture gradient (Curtis, 1959; Johnson, 1990). The species composition of forests is continually changing as a result of forces both internal (autogenic) and external (allogenic) to the forest. Changes are often gradual and frequently result in the replacement of one tree species by another in the process of ecological succession. The vegetation and other organisms within the forest thus effect autogenic change. For example, shade-tolerant species growing beneath the main forest canopy may gradually replace dominant species of lesser shade-tolerance that are unable to regenerate under their own shade. In contrast, allogenic change occurs as a result of

Importance value

Black oak

13

changes in climate, defoliation by exotic insects and pathogens, the movement of soil by wind and water, or from other forces originating outside the forest. Autogenic and allogenic factors sometimes jointly affect the direction and rate of succession. Moreover, disturbances such as windthrow, insect and disease outbreaks, and timber harvesting can accelerate succession or alter its direction. Although the oaks are relatively intolerant of shade, species vary substantially in this attribute. In some habitats, oaks are vulnerable to successional replacement by more shade tolerant species. Compared to many of their competitors, oak seedlings grow more slowly during their ﬁrst few years after initial establishment. When young oaks are overtopped and heavily shaded by other vegetation, few survive for very long. On the other hand, the oaks tend to be relatively drought tolerant, and often survive in habitats that limit the development of species of lesser drought tolerance. Oaks also can produce vigorous

Northern red oak White oak

Sugar maple American beech

Bur oak

Xeric

Mesic Compositional index

Fig. 1.4. Changes in the relative importance of six tree species in the upland forests of southern Wisconsin in relation to the regional soil moisture gradient. Species’ importance is quantitatively expressed by an importance value, which is an index of species’ importance based on its frequency of occurrence, density and basal area relative to other species within a stand. Although there is much overlap among species’ importance value curves, no two species behave exactly the same way with respect to the moisture gradient. The length of the gradient spanned by a species’ range of importance values together with the shape of its importance value curve reflects its niche with respect to the gradient. Importance value curves also define the ecological amplitude of a species, i.e. the range of conditions it can tolerate and the magnitude of its importance in relation to the gradient under the prevailing (i.e. relatively undisturbed) stand conditions. The moisture gradient shown is inferred from the species composition of a series of relatively undisturbed stands (see Curtis, 1959). (Adapted from Curtis, 1959, used with permission.)

14

Chapter 1

sprouts that often outgrow competitors. The balance of these factors thus determine the relative permanence of oaks within a given cover type. In the eastern half of the United States, oaks are often relatively permanent members of cover types on drier sites. In the absence of disturbance, many of the pine and oak–pine cover types occurring on dry habitats are successional to oaks because the oaks are somewhat more shade tolerant than the pines. This successional pattern creates silvicultural problems in maintaining pure pine stands in the south and other regions where oaks and pines co-occur (Burns and Barber, 1989). In bottomlands and mesic uplands, shadetolerant or faster growing species often successionally displace the oaks. Such displacement creates silvicultural problems in perpetuating oaks in these forests. The relative permanence of an oak species within a given cover type (i.e. its resistance to successional replacement by other species) is likely to be highly variable if the cover type spans a broad range of environments. For example, the white oak cover type occurs across a wide range of site conditions from dry to moist. Whereas the type tends to be relatively permanent on dry sites, it is successional to other types on the more mesic sites. Cover type designations, although useful, largely fail to consider these and other ecological factors that determine changes in species composition and how those changes vary spatially (e.g. in relation to climate and site quality), and temporally (e.g. in relation to plant succession and disturbance). Consequently, two or more stands representing a single cover type may represent quite different ecologies with respect to the successional status of oaks, physical environment, understorey vegetation, forest regeneration, fauna and other factors. Forest inventories and satellite imagery have been used to describe the geographic distribution of forest types in the United States (e.g. Fig. 1.5). These maps identify broad cover type groups that are aggregates of the stand cover types described above.

Four groupings widely used to delineate oak forests at the regional scale are: the oak–hickory group, the oak–pine forest group, the oak–gum–cypress group (bottomland forests), and the western hardwood group that includes the western oaks as a subset (Fig. 1.5). However, the names commonly applied to the resulting species aggregations can be misleading. For example, hickory is absent throughout much of the northern part of the range delineated as oak–hickory (Fig. 1.5). Moreover, other forest cover types dominated by oaks are also included within the delineated oak–hickory area. The term ‘oak–hickory’ nevertheless is widely used in reporting forest resource statistics at the regional level even though it is an ecologically imprecise term. Oaks also occur as ecologically and silviculturally important components of many nonoak forests (e.g. pine forests and maple–beech–birch forests). In the eastern United States the oak– hickory, oak–pine and the oak–gum–cypress cover type groups collectively covered 187 million acres or 52% of the timberland4 in 1997. That is an increase from 162 million acres and 45% of eastern timberland in 1953 (USDA Forest Service, 2000). At 124 million acres, the oak–hickory group is the largest cover type in the United States. The western oaks are also signiﬁcant geographically and ecologically. Western hardwood forests (including oaks, tanoak, red alder and aspen) cover 43 million acres or 12% of western forestland. Oaks comprise about 23% of the cubic volume of growing stock trees in the eastern United States and about 1% in the western United States (USDA Forest Service, 2000).

Distribution of oaks by hierarchically classified ecoregions Climate and landform strongly inﬂuence the distribution of oaks. Locally, the distribution of oaks is inﬂuenced by factors such as physiography, soil moisture and geology.

4 Timberland is forest land that is producing, or is capable of producing, more than 20 feet3 acre–1 year–1 of industrial wood crops under natural conditions, and that is not withdrawn from timber use, and that is not associated with urban or rural development. Currently inaccessible and inoperable areas are included.

Oak-dominated Ecosystems

(a)

15

Long. 100

240

210

210 330 340 220

250

320 230 260

310

410

Oak–hickory (b)

240

210

210 330 340 220

250

320 230 260

310

410

Oak–gum–cypress (c)

240

210

210 330 340 250

220

320 230 260

310

Western hardwoods 410

Oak–pine

Fig. 1.5. The major areas of oak–hickory, oak–pine, oak–gum–cypress, and western hardwoods (shaded areas) by state and ecoregion Divisions. In the western US, the map shows the composite western hardwood group that includes oaks, tanoak, red alder, cottonwood and aspen. Numbered ecoregion boundaries on the map are from Bailey (1997) and are summarized in Tables 1.2 and 1.3. Generated from Advanced Very High Resolution Radiometer satellite images at a scale of 1 km2 (1990) and an associated system of land cover classification (USDA Forest Service, 1993; Powell et al., 1994). (Map compiled by W.D. Dijak, USDA Forest Service, North Central Research Station, Columbia, Missouri.)

16

Chapter 1

These and other factors have been used to structure a hierarchical ecological classiﬁcation system (McNab and Avers, 1994; Bailey, 1995, 1997, 1998). This system recognizes the increasing detail necessary to explain the spatial arrangement of forests at increasingly smaller spatial scales (Table 1.1). It thus provides an objective basis for the regional delineation of ecosystems into successively smaller and more homogeneous units. The hierarchical ecological units range in size from continents to a few acres. The larger units are often referred to as ecoregions; the smallest units are often equivalent to forest stands. Domains, Divisions and Provinces form the larger ecoregions (Table 1.1). These are climatic and climatic–physiographic regions that cover millions to tens of thousands of square miles. Provinces are further subdivided into smaller units termed Sections, Subsections, Landtype Associations (LTAs), Ecological

Landtypes (ELTs) and Ecological Landtype Phases (ELTPs). These units range in size from thousands of square miles for Sections to less than 10 acres for some Ecological Landtype Phases. Ecological Landtypes and Ecological Landtype Phases are important silviculturally because they often correspond to individual stands, which are the objects of silviculture. The oaks occur in all three Domains (major climatic regions) of the 48 contiguous states: Humid Temperate, Dry and Humid Tropical (Bailey, 1997). The latter occurs only in the southern tip of Florida. The three Domains are further subdivided into 11 climatic Divisions. Within each Division, mountainous areas with elevational zonation of vegetation are also identiﬁed. Although oaks naturally occur in all 11 of the Divisions, within each Division the distribution of the four major oak forest types is closely related to Division boundaries (Fig. 1.5; Tables 1.2 and 1.3).

Table 1.1. Hierarchy of ecological units used to classify forest ecosystems in the United States.a Ecological unit

Scale (reference size)b

Delineating factorsc

Domain

Millions to tens of thousands of square miles (subcontinent)

Macroclimate, ocean temperature and currents, geomorphology

Division

Millions to tens of thousands of square miles (multi-state)

Geomorphology, climate

Province

Millions to tens of thousands of square miles (multi-state, state)

Geomorphology, climate

Section

1000s of square miles (state, multi-county, National Forest)

Geomorphology, climate, vegetation

Subsection

10s to 100s of square miles (multiple counties, National Forest Ranger District)

Geomorphology, climate, vegetation

Landtype association

10s to 1000s of acres (landscape, watershed)

Landforms, species composition of overstorey, soil associations

Ecological landtype

10s to 100s of acres (multiple stands)

Landform, natural vegetative communities, soils

Ecological landtype phase

1 to 10s of acres (stand)

Soils, landscape position, natural vegetative communities

a

Adapted from McNab and Avers (1994), Bailey (1995) and Cleland et al. (1993); see also Figs 1.2, 1.5 and 1.6. b Indicates a familiar unit of comparable size for reference purposes. This reference unit is not used to delineate the ecological unit. c Some of the factors used to distinguish among ecological units at a given level. Classification complexity typically increases with decreasing unit size.

Oak-dominated Ecosystems

17

Table 1.2. The ecoregion domains, divisions and provinces in the eastern conterminous United States where oaks are found and the principal species occurring in each. Ecoregions from Bailey (1995). Division and province boundaries are shown in Fig. 1.2. Division

Province

– – – – – – – – – – – – – – – – – – – 200 Humid Temperate Domain – – – – – – – – – – – – – – – – – – – 210 Warm Continental M210 Warm Continental Mountains

10 oak species:

220 Hot Continental M220 Hot Continental Mountains

22 oak species:

230 Subtropical M230 Subtropical Mountains

31 oak species:

250 Prairie

211 Mixed deciduous coniferous forests M211a Mixed forest–coniferous forest–tundra, medium M211b Mixed forest–coniferous forest–tundra, high

bear, black, bur, chestnut, chinkapin, n. pin, n. red, scarlet, swamp white, white 221a Broadleaved forests, oceanic 221b Broadleaved forests, continental M221 Deciduous or mixed forest–coniferous forest–meadow M222 Broadleaf forest–meadow

basket, bear, black, blackjack, bur, cherrybark, chestnut, chinkapin, n. pin, n. red, overcup, pin, post, scarlet, shingle, Shumard, s. red, swamp chestnut, swamp white, water, willow, white 231 Broadleaved–coniferous evergreen forests 232 Coniferous–broadleaved semi-evergreen forests M231 Mixed forest–meadow province

Arkansas, bear, black, blackjack, bluejack, bur, Chapman, cherrybark, chestnut, chinkapin, Durand, Georgia, laurel, live, myrtle, Ogelthorpe, n. red, Nuttall, overcup, pin, post, scarlet, shingle, Shumard, s. red, swamp chestnut, swamp white, turkey, water, white, willow 251 Forest-steppes and prairies province 252 Prairies and savannas province

20 oak species:

black, blackjack, bluejack, bur, chinkapin, Durand, live, n. pin, n. red, overcup, laurel, pin, post, s. red, shingle, Shumard, swamp chestnut, swamp white, water, white

– – – – – – – – – – – – – – – – – – – 400 Humid Tropical Domain – – – – – – – – – – – – – – – – – – – 410 Savanna

4 oak species:

411 Open woodlands, shrubs and savanna 412 Semi-evergreen forests 413 Deciduous forests province Chapman, live, laurel, myrtle

– – – – – – – – – – – – – – – – – – – Intrazonal Regions – – – – – – – – – – – – – – – – – – – R Riverine forest

The 11 ecoregion Divisions within the conterminous United States are further subdivided into 44 Provinces (Fig. 1.2). Provinces are delineated based on broad vegetation groups and related regional landforms. Oak forests and woodlands commonly occur in 23 Provinces (Tables 1.2 and 1.3). Province boundaries are useful in delineating oak distributions in some

parts of the United States. For example, Province boundaries correspond with the spatial distribution of the oak forests and woodlands encircling California’s Central Valley. Province boundaries also separate the oak–pine forests of the Piedmont (Province 232) from the wetter oak habitats of the Coastal Plain and the lower Mississippi ﬂood plain (Province 231 and

18

Chapter 1

Table 1.3. The ecoregion domains, divisions and provinces in the western conterminous United States where oaks are found and the principal species occurring in each. Ecoregions from Bailey (1995). Division and province boundaries are shown in Fig. 1.2. Division

Province

– – – – – – – – – – – – – – – – – – 200 Humid Temperate Domain – – – – – – – – – – – – – – – – – – 240 Marine M240 Marine Mountains

3 oak species: 260 Mediterranean

M260 Mediterranean Mountains

13 oak species:

241 Mixed forests M241 Deciduous or mixed forest–coniferous forest–meadow M242a Forest–meadow, medium M242b Forest–meadow, high Oregon white, California black, canyon live 261 Dry steppe 262 Mediterranean hardleaved evergreen forests, open woodlands and shrub 263 Redwood forests M261 Mixed forest–coniferous forest–alpine meadow M262 Mediter. woodland or shrub–mixed or conif. forest–steppe or meadow M263 Shrub or woodland–steppe–meadow blue, California black, California scrub, canyon live, coast live, Dunn, Engelmann, interior live, island live, McDonald, Oregon white, turbinella, valley

– – – – – – – – – – – – – – – – – – 300 Dry Domain – – – – – – – – – – – – – – – – – – 310 Tropical/ Subtropical Steppe 311 Coniferous open woodland and semideserts 312 Steppes 313 Steppes and shrubs 314 Shortgrass steppes M310 Tropical/ Subtropical M311 Steppe or semidesert–mixed forest–alpine meadow or steppe Steppe Mountains 16 oak species: Arizona, canyon live, Chisos, Dunn, Emory, Gambel, Gray, Havard, Lacey, lateleaf, Mohr, sandpaper, silverleaf, Toumey, wavyleaf, turbinella 320 Tropical/ Subtropical Desert

M320 Tropical/ Subtropical Desert Mountains 22 oak species:

330 Temperate Steppe M330 Temperate Steppe Mountains

2 oak species: 340 Temperate Desert M340 Temperate Desert Mountains 3 oak species:

321 Semideserts 322 Oceanic semideserts 323 Deserts on sand M321 Semidesert–shrub–open woodland–steppe or alpine meadow M322 Desert or semidesert–open woodland or shrub–desert or steppe Arizona, chinkapin, Chisos, Dunn, Durand, Emory, Gambel, Graves, gray, Havard, Lacey, lateleaf, live, Mexican blue, Mohr, netleaf, post, sandpaper, silverleaf, Toumey, turbinella, wavyleaf 331 Steppes 332 Dry steppes M331 Forest-steppe–coniferous forest–meadow–tundra M332 Steppe–coniferous forest–tundra M333 Steppe–coniferous forest M334 Steppe–open woodland–coniferous forest–alpine meadow bur, Gambel 341 Semideserts 342 Semideserts and deserts M341 Semidesert–open woodland–coniferous forest–alpine meadow

Gambel, turbinella, wavyleaf

Oak-dominated Ecosystems

Riverine Forest). Province boundaries are also useful in separating the regions where oaks occur from those where they do not. In contrast to the coarser levels of the classiﬁcation hierarchy (Domains through Subsections), which have been delineated nationally, classiﬁcation of the ELT and ELTP levels is incomplete across much of the oak range. Even though classiﬁcation systems down to the ELT or ELTP have been developed for millions of acres, they include only a small fraction of the total area of oak forests. ELTs or ELTPs are usually mapped in the ﬁeld based on differences in soils, physiography and vegetation (including herbs and shrubs). The species composition of the herbaceous layer is often used to distinguish among different ELT or ELTP units because of the ﬁdelity of some herbaceous species (‘indicator’ species) to speciﬁc biophysical conditions. Accordingly, the presence or absence of one or more indicator species can be used to differenti-

19

ate among otherwise similar ELTs or ELTPs. Shrubs are also sometimes used as indicator species. Compared to the herbaceous layer, the composition of the tree layer often recovers slowly from disturbances. Moreover, the tree component may not recover to its pre-disturbance composition. Joint consideration of physical and biological factors and their interactions provide a basis for identifying ecologically homogeneous land units that are silviculturally relevant and useful in delineating management units (Barnes et al., 1982). Ecological classiﬁcation provides a broader ecological context for understanding why oaks occur where they do, and how those occurrences change with time, disturbance and other factors. At the broadest scale the oak forests of the United States can be divided into four eastern and two western groups based on species associations, ecological conditions and successional relations (Fig. 1.6).

Pacific Mediterranean–Marine Region

Coast Range

340 260 Sierra Nevada

Central Valley

320 310

Southwestern Desert–Steppe Region

Edwards Plateau Cross Timbers

Forest–Prairie Transition Region

330

240

Driftless Area 210 Northern Hardwood Region 210 Ohio Valley

250

220 Central Hardwood Region

Southern Pine–Hardwood Region 230

Lower Mississippi Ozark Highlands Valley Boston Mountains

Allegheny Plateau Appalachian Mountains Coastal Plain Piedmont Cumberland Plateau Highland Rim Piedmont Coastal Plain

410

Fig. 1.6. The six regions where oaks commonly occur: Northern Hardwood Region; Central Hardwood Region; Southern Hardwood–Pine Region; Forest–Prairie Transition Region, Southwest Desert–Steppe Region; and Pacific Mediterranean–Marine Region. Numbers correspond to Ecoregion Divisions (Figs 1.2 and 1.5) (Bailey, 1997). Not considered by the above regional groupings are the ranges of Gambel and bur oak, which extend into Division 330, and the ranges of Gambel, turbinella and wavyleaf oaks, which extend into Division 340. The shading shows the distribution of oaks from Fig.1.2.

20

Chapter 1

Boundaries between regions follow Division boundaries in the hierarchical ecological classiﬁcation system. These regional groupings are useful ecologically and silviculturally because they identify areas with broadly similar macroclimates and species associations. Regional differences in the application of silvicultural methods are closely related to corresponding differences in species composition, environmental factors and other ecological conditions. The six forest regions are described below in relation to the Domains, Divisions and Provinces of the hierarchical ecological classiﬁcation system described in the preceding section. However, the regional designations do not explicitly identify the lowland and riparian forests occurring within them. There, along the major rivers and streams within the Southern Pine–Hardwood Region, the oaks attain their greatest size and growth.

Eastern Oak Forests The Northern Hardwood Region Geographic extent The Northern Hardwood Region includes the northern halves of Minnesota, Wisconsin, Michigan and much of the northeastern United States including New Hampshire, Vermont and Maine in their entirety. It includes two ecoregion Provinces: Mixed Deciduous– Coniferous Forests Province (211) and Mixed Forest–Coniferous Forest–Tundra, High Province (M211b) within the Warm Continental Division (Figs 1.2 and 1.6; Table 1.2). The Region extends 1300 miles from west to east and covers 123 million acres, about three-quarters of which is forested. Braun (1972) called this area the Hemlock–White Pine–Northern Hardwood Region. She recognized two major subsections, the Great Lakes–St Lawrence and the Northern Appalachian Highlands. The western and eastern portions of the Northern Hardwood Region share many of the same species, but they differ ecologically and

silviculturally (Godman, 1985). Those differences are due in part to the inﬂuence of the Appalachian Mountains in the eastern part of the Northern Hardwood Region. More than 1.5 million non-industrial private forest owners own approximately half of the forests in the Northern Hardwood Region. Corporate and other private owners hold an additional 25% (Birch, 1996). There are 11 national forests in the region (primarily in the Lake States) that cover 6.5 million acres.

Climate, physiography and soil Precipitation typically ranges from 24 to 45 inches per year although as much as 70 inches occurs in some mountainous areas in the eastern part of the region. Snowfall of 60 to 100 inches per year is common throughout the region. More than 100 inches of snowfall occurs at some of the higher elevations, and snowfall exceeds 400 inches in some locales near the Great Lakes. Mean annual temperature ranges from 35° to 52°F (2 to 11°C) and the growing season lasts from 100 to 160 days (Fig. 1.7) (McNab and Avers, 1994). The region is characterized by low relief with numerous lakes, depressions, morainic hills, drumlins, eskers, outwash plains and other glacial landforms. Variation in the depth and type of glacial deposits and associated heights of water tables are important factors in the identiﬁcation of silviculturally relevant ELTPs (Fig. 1.8). Elevations in the mountainous areas range from 1000 to 4000 feet with individual peaks exceeding 5000 feet. Valleys in the mountainous areas include outwash plains and lakes resulting from glaciation (Bailey, 1995). Soils have formed in diverse organic and mineral materials including peat, muck, marl, clay, silt, sand, gravel and boulders in various combinations. At lower elevations in New England and along the Great Lakes, Spodosols are common. Inceptisols and Alﬁsols dominate at lower elevations elsewhere. In the mountainous zones the soils are primarily Spodosols (Bailey, 1995).

Oak-dominated Ecosystems

Iron mountain, MI

Fort Wayne, IN

Boone, NC

42˚F 30 in.

50˚F 49 in.

52˚F 55 in.

21

Atlanta, GA 62˚F 49 in.

Fargo, ND

41˚F 19 in.

Key West, FL 78˚F 40 in.

Fig. 1.7. Representative climates for selected ecoregion Divisions in the eastern United States. Mean monthly precipitation is shown by the solid lines (right axis) and temperature by dashed lines (left axis). Mean annual values are given above each graph. Division boundaries are shown in Figs 1.2 and 1.5. (Ecoregion and climatic data from Bailey, 1995.)

22

Chapter 1

ELTPs on dry ice-contact and sand hills

ELTPs on outwash plains 1

10

11

12

20

21

ELTPs on mesic ice-contact and sand hills 34

35

24

25

WATER TABLE

WATER TABLE

32

22

ELTPs on herb-poor moraines 37

40

42

43

WATER WATER

Fig. 1.8. Ecological landtype phases (ELTPs) for the upland forests of the Huron–Manistee National Forests in the lower peninsula of Michigan (Province 211: Mixed Deciduous–Coniferous Forests Province). Site productivity generally increases with increasing ELTP value. ELTP 1: Northern pin oak/white oak – Deschampsia type; ELTP 10: Black oak/white oak – Vaccinium type; ELTP 11: Black oak/white oak – Vaccinium type with loamy sand to sandy loam bands in substrata; ELTP 12: Black oak/white oak – Vaccinium type with perched water table at 6–15 ft; ELTP 20: Mixed oak/red maple – Trientalis type; ELTP 21: Mixed oak/red maple – Trientalis type with loamy sand to sandy loam bands in substrata; ELTP 22: Mixed oak/red maple – Trientalis type with perched water table at 6–15 ft; ELTP 24: Mixed oak/red maple – Trientalis type with perched water table at 3.5–6 ft; ELTP 25: Mixed oak/red maple – Trientalis type with coarse loamy substrata; ELTP 32: Northern red oak/red maple – Viburnum type with perched water table at 6–15 ft; ELTP 34: Northern red oak/red maple – Viburnum type with perched water table at 3.5–6 ft; ELTP 35: Northern red oak/red maple – Viburnum type with fine loamy substrata; ELTP 37: Northern red oak/red maple – Desmodium type with sandy loam over fine loamy substrata; ELTP 40: Sugar maple/beech – Maianthemum type; ELTP 42: Sugar maple – Maianthemum type with perched water table at 6–15 ft; ELTP 43: Sugar maple/northern red oak – Maianthemum type with fined texture substrata. (Adapted from Cleland et al., 1993.)

Oak-dominated Ecosystems

Forest history The forests of the Northern Hardwood Region were strongly inﬂuenced by the aboriginal people who lived there. For thousands of years before the arrival of Europeans, Native Americans used ﬁre and land clearing to shape the forest to meet their needs. Accounts of early European settlers indicate that Native Americans burned large portions of the landscape each year (Pyne, 1982). Dry fuels in the late spring before ‘greenup’ and again after leaf fall during ‘Indian summer’ provided favourable conditions for burning. Fires often eliminated forest understorey layers, which in turn encouraged the growth of edible berries and increased forage that attracted edible wildlife. A combination of ﬁre and tree cutting or girdling also was used by Native Americans to create and maintain openings for cultivated crops (Pyne, 1982). Maintaining an open understorey condition by burning also helped defend Indian villages from surprise enemy attacks. Repeated burning helped maintain large areas of oak savannas and barrens. Topography greatly inﬂuenced the spatial distribution of ﬁres, and the frequency and intensity of burning were lower on wet or mesic sites and at higher elevations. This produced a landscape mosaic of diverse species composition. The land clearing and burning practices used in New England were carried westward by settlers who immigrated to the Lake States. Oaks were known as indicators of fair to good conditions for agriculture. Oak forests therefore were often girdled, felled and burned in preparation for agriculture. Over time, enormous areas in the Northern Hardwood Region were cleared for agriculture. Ultimately, however, it was logging that had the greatest impact on the region’s forests. Logging gradually accelerated with the inﬂux of Europeans to New England in the 17th century. Demands for forest products were initially modest in the developing agrarian society. Local timber harvesting supplied wood for homes, barns, fences, heating and cooking, and making potash and tannin. Forests were considered more

23

an impediment to agriculture than a valued resource. However, this changed with industrialization during the 19th century (Williams, 1989). The industrialization of America provided the capacity and economic incentive to exponentially increase lumber production from less than 1 to nearly 45 billion board feet between 1800 and 1900. That lumber, which was principally white pine and other softwoods, came primarily from the Northern Hardwood Region. In 1839, 30% of the nation’s lumber (on a value basis) came from New York. Combined, New York, Pennsylvania, New Jersey and New England produced two-thirds of the nation’s lumber. As supplies of white pine dwindled in the eastern part of the region, timber production moved westward. Although New York reached peak production in 1849, the northeastern states together had by then dropped to half of the national lumber output. The shift in lumber production from the Northeast to the Lake States occurred between 1840 and 1860. Lake States harvest reached a peak of 10 billion board feet annually in 1889. Relatively level terrain, easy access and high demand facilitated the rapid rise in timber harvesting across the Lake States. By 1940, Lake States production dropped below a billion board feet as the timber industry moved south (Williams, 1989). Although white pine was the preferred species, oaks and other hardwoods were utilized locally where they were abundant. Oaks were in less demand by the logging industry. Because of their high density, oak logs could not be ﬂoated down rivers as easily as white pine and required overland transportation to avoid losses (Williams, 1989). Over time, repeated timber harvesting removed the trees of greatest economic value leaving behind stands of inferior quality and composition. Farmers emigrating westward subsequently completed the land clearing. As a preferred fuelwood, the oak’s utilization for that purpose signiﬁcantly affected the region’s forests during the early agricultural period. A colonial family used 20–60 cords of wood annually for

24

Chapter 1

heating and cooking. Although the per capita volume of wood used for fuel decreased over time, the total volume increased because of a growing population. In 1880, more than half of America’s energy needs were still met with fuelwood (Whitney, 1994). Iron furnaces in the region were ﬁred with hardwood charcoal. Less than 1% of the fuelwood burned from 1800 to 1930 was used to produce charcoal for iron smelting, but large areas of forest surrounding iron furnaces were greatly affected (Whitney, 1994). A typical 18th-century smelting operation consumed 100 acres of forest annually to produce charcoal. Because forest regrowth could be repeatedly harvested for this purpose every 25 years, about 2500 acres of forest were required to sustain production of the ironworks. By the late 19th century, charcoal production for large ironworks required annual harvests of 1000– 2000 acres, and some large companies owned 100,000 acres of forest adjacent to their smelters for that purpose (Whitney, 1994). The large clearcuts surrounding the ironworks radically changed the age structure of the forests and also inﬂuenced their species composition. Harvesting hardwoods for fuel favoured the development of hardwood sprouts and increased the relative proportion of hardwood trees in areas that were not converted to agricultural land (Whitney, 1994). During the last half of the 19th century, many cleared acres marginally suited to agriculture were abandoned and subsequently reverted to forest. During the latter half of the 20th century, the combination of abandoned agricultural lands and natural regeneration of cutover lands resulted in large increases in timber volumes throughout the region. In New England, forest volumes increased by 16% between 1970 and 1982 (cubic foot basis) (Seymour, 1995). The increase was predominantly in hardwoods that as a group increased in volume by 24%. The increase in oak volume (16%) was low relative to other hardwoods. In 1992, net annual growth in the Northern Hardwood Region remained at more than twice the annual harvest (Powell et al., 1994).

Oaks as components of the region’s forests The forests of the Northern Hardwood Region today are dominated by more than half a dozen recognized northern hardwood forest types comprising various combinations of sugar maple, red maple, beech, paper birch, yellow birch and eastern hemlock. Although northern red oak typically occurs as a minor component within these types, it sometimes forms pure or nearly pure stands (Fig. 1.9A). Wherever it occurs, it is a valuable and desirable species for timber, acorn production and species diversity. White, black, northern red and chestnut oaks also occur in the southern portions of the region. Oaks are thus a relatively small component of northern hardwood forests. They are most abundant and attain their best development in the southern parts of the region including New York, Massachusetts, northern Pennsylvania, and central and southern Minnesota, Wisconsin and Michigan. There the oak and mixed-hardwood forests grade into the oak–hickory forests of the Central Hardwood Region. In both New England and the Lake States, about 11% of the forestland is classiﬁed as oak–hickory or oak–pine. These forest types include 17 billion cubic feet of growing stock (inclusive of oaks and associated species). Throughout the Northern Hardwood Region, sugar maple, red maple and aspen are the most abundant hardwoods. Conifer forests (red and eastern white pines, spruce and balsam ﬁr) also exceed oak in acreage and volume (Powell et al., 1994). Oaks often reach their greatest density on sites that have been repeatedly disturbed by ﬁre, timber harvesting and other events. After burning or timber harvesting, oaks originating from vigorous seedling sprouts and stump sprouts often dominate stands. However, in the absence of disturbance the oak forests of the Northern Hardwood Region are usually successional to other hardwoods on all but the poorest sites. On the poorer sites, oaks are often relatively permanent members of the forest.

Oak-dominated Ecosystems

25

A

B

Fig. 1.9. (A) A 130-year-old stand of northern red oak in the Northern Hardwood Region of northern Wisconsin (Province 211: Mixed Deciduous–Coniferous Forests Province; Southern Superior Uplands Section). The absence of oak reproduction and a sparse sub-canopy of shade tolerant red and sugar maples are indicators of what is likely to eventually replace the oaks in the absence of disturbance. (B) Xeric northern pin oak–white oak/Deschampsia type (see Fig. 1.8) on deep outwash sand in the northern lower peninsula of Michigan (Province 211: Mixed Deciduous–Coniferous Forests Province; Northern Great Lakes Section). This oak stand is mixed with jack pine; oak site index is ≤50 ft. (USDA Forest Service, North Central Research Station photographs.)

There, oaks frequently invade and successionally replace established pine stands (Seymour, 1995). The loss of the American chestnut to chestnut blight fungus in New England oak forests began in the early 1900s (Fig. 1.10). This increased the relative importance of oaks because oaks often captured the growing space vacated by dying American chestnuts.

Today, the single-tree selection method of silviculture is often applied to northern hardwood forests dominated by shade tolerant species such as sugar maple. This practice focuses on maintaining stands of high quality trees while largely relying on the natural regeneration of shade tolerant species to sustain the silvicultural system. Although this system favours the develop-

26

Chapter 1

pine, form relatively stable forest types of low productivity (Fig. 1.9B).

The Central Hardwood Region Geographic extent

Fig. 1.10 A standing dead American chestnut (minus bark). Chestnut was a common associate and dominant member of eastern oak forests throughout the Appalachians from Maine to Alabama and westward to Missouri. The chestnut blight, which decimated the species throughout its range, permanently altered the ecology of eastern oak forests. The blight was first identified in New York in 1904. Fifty years later it spanned the entire natural range of chestnut. Oaks and associated hardwoods quickly captured the growing space vacated by dead and dying chestnuts. (USDA Forest Service, North Central Research Station photograph.)

ment of high quality oaks in stands where oaks are already present, regenerating oaks beneath the relatively closed canopies of selection forests is usually difﬁcult in this region. On the poorer sites, oaks may develop beneath a pine overstorey and eventually displace the less shade tolerant pine through natural succession or the exposure of oak reproduction in the understorey to full light after timber harvest. On deep sandy soils of the upper Lake States, stands of northern pin, black and white oaks, often mixed with jack

The Central Hardwood Region includes the predominantly deciduous broadleaf forests of Central United States. The region lies entirely within the Humid Temperate Domain. The region includes the two Hot Continental Divisions (Divisions 220 and M220), and intergrades with the eastern part of the Forest–Steppes and Prairies Province (251) of the Prairie Division (250) (Figs 1.2 and 1.6; Table 1.2). The Hot Continental Division is subdivided into two provinces: Broadleaf Forests, Oceanic (221a); and Broadleaf Forests, Continental (221b). The Hot Continental Mountains Division (M220) also is divided into two provinces: Deciduous or Mixed Forest–Coniferous Forest–Meadow (M221), and Broadleaf Forest–Meadow (M222). The Northern Hardwood Region, the Southern Pine–Hardwood Region, the Forest–Prairie Transition and the western edge of the Appalachian Mountains bound the Central Hardwood Region. The Central Hardwood Region extends 1200 miles from southwest to northeast and covers approximately 220 million acres; about half the region is forested. Approximately threequarters of the forest area in the Central Hardwood Region is in non-industrial private ownership. Within that ownership, most holdings are 50 acres or smaller (Birch, 1996). There are seven national forests in the region comprising about 4 million acres distributed across the southern half of the region in Arkansas, Missouri, Illinois, Indiana, Ohio, Kentucky and Tennessee.

Climate, physiography and soil The climate in the Central Hardwood Region is hot continental with warm summers and cold winters. Mean annual temperature ranges from 40 to 65°F (4–18°C), with the warmer temperatures in the south.

Oak-dominated Ecosystems

Annual precipitation ranges from 20 inches in the northwest to 65 inches in the southeast and reaches as much as 80 inches on some Appalachian peaks (Fig. 1.7). Precipitation occurs throughout the year, but tends to be somewhat greater in spring and summer. Droughts may occur during the summer when evapotranspiration is high. Frost-free periods range from 100 days in the northern Appalachians to 220 days in the southern part of the region (Bailey, 1995). Topography is diverse in this region. Elevations in the Appalachian Highlands (province M221) range from 300 to 6000 ft with as much as 3000 ft of local relief. Further west (province 221a), the hills and low mountains of the dissected and uplifted Appalachian Plateau (including the Allegheny and Cumberland Plateau) range from 1000 to 3000 ft in elevation. In the western half of the Central Hardwood region (province 221b), most of the land is rolling but varies from extensive, nearly level areas to areas like the Ozark Highlands where relief reaches 1000 ft. Most of the northern portions of this province were glaciated with the exception of the driftless area of southwestern Wisconsin and adjacent states. Major soils are Alﬁsols, Inceptisols, Mollisols and Ultisols (Bailey, 1995).

Forest history The utilization and exploitation of forests in the Central Hardwood Region has passed through various historical phases (Hicks, 1997). Even before the arrival of Europeans, humans inﬂuenced the nature and extent of the region’s forests (Whitney, 1994). The use of ﬁre to control vegetation by Native Americans signiﬁcantly inﬂuenced the extent and character of presettlement forests (Pyne, 1982; DeVivo, 1991; Olson, 1996). These human-caused alterations of the landscape continued for thousands of years before the arrival of Europeans (Hicks, 1997). After settlement by Europeans, human impacts on the forest expanded and intensiﬁed. Burning, grazing, exploitative timber harvesting and

27

clearing of forests for agriculture occurred on an unprecedented scale. These practices occurred about 200 years earlier in the eastern part of the Central Hardwood Region than in the western part. Historically, different human disturbances were further confounded by intrinsic ecological differences among oak forests within the various ecoregion provinces. Each subregion of the Central Hardwood Forest has its own unique combination of disturbance history, climate, physiography, soils, species associations and successional possibilities. This complicates generalizing the application of silvicultural methods to oak forests across the region. As in the Northern Hardwood Region, the loss of American chestnut to chestnut blight increased the relative proportion and importance of oaks throughout the region. Shortly after 1900, the disease became epidemic and within 40 years it had invaded the entire natural range of the chestnut (Kuhlman, 1978). The loss represents one of the greatest recorded changes in a natural population of plants caused by an introduced organism (Liebhold et al., 1995). The chestnut comprised 25% of the eastern hardwood forest that covered 200 million acres. In the Appalachians, it was the most ecologically and economically important tree species (Kuhlman, 1978). There and in other regions, it grew faster and taller than associated oaks. Before the blight, chestnut was especially important in moist upland forests where it sprouted vigorously and often increased in dominance after logging. In 1900, half the standing timber in Connecticut was chestnut, which was largely comprised of young stands of stump sprout (coppice) origin (Smith, 2000). Although American chestnut provided only about 1% of the nation’s hardwood lumber even at the peak of its importance, its loss (beginning in the early 1900s) had a signiﬁcant impact on local economies in the Appalachians. There its nuts and bark (for tannin) provided scarce cash income, and its wood was valued for a variety of uses (Youngs, 2000). The practice of silviculture in the Central Hardwood Region dates back to the

28

Chapter 1

genesis of North American forestry in the late 19th century (Fernow, 1911; Pinchot, 1987). From then until the 1960s, the major emphasis was on uneven-aged silviculture (Roach, 1968). During the 1960s, the emphasis shifted to even-aged silviculture, especially clearcutting, and this emphasis persisted for about 20 years (Roach and Gingrich, 1968; Johnson, 1993a). Where applied, hardwood silviculture in the region usually follows the ‘ecological model’, which relies on the existing forest vegetation and its natural regeneration capacity. Silvicultural prescriptions are usually focused on controlling stand structure and species composition using cutting methods such as those recommended by Roach and Gingrich (1968). This approach contrasts with the more intensive ‘agronomic model’ of silviculture based on artiﬁcial regeneration, the introduction of improved genotypes, use of herbicides and fertilizer, prescribed burning, and other intensive cultural methods like those commonly used in the silviculture of pine monotypes in the south and elsewhere. Where applied, silviculture in the Central Hardwood Region has usually focused on growing high quality sawtimber. During the course of stand management (but before ﬁnal harvest of even-aged stands), this requires ‘leaving the best’ and ‘cutting the worst’ at each harvest. In evenaged silviculture, each timber harvest concentrates on removing small, sub-canopy trees and poor quality trees in the main canopy, with concomitant attention to species composition. Similarly, in unevenaged silviculture, timber removals are concentrated on poor quality trees, but cutting occurs across a wide range of diameter classes in order to create and maintain the uneven-aged stand structure. In both systems the objective is the improvement of the quality and the economic value of the residual stand. Today, only a small fraction of the forests of the Central Hardwood Region receive systematic silvicultural treatment. This is largely due to the pattern of forest ownership, which is characterized by numerous small tracts owned by private

individuals. Many forest owners are uninterested in silviculture or lack information on its beneﬁts (Bliss et al., 1994, 1997; English et al., 1997). Consequently, the systematic application of silviculture has largely been limited to industrial forests and public lands. The predominant methods of timber harvesting on private lands are probably commercial clearcutting and other forms of high-grading. Not to be confused with silviculturally prescribed methods, these methods consist of harvesting all trees with commercial value without regard to regeneration needs and future stand condition. Such malpractice typically leaves stands of highly variable residual stocking comprised of trees of poor vigour, low quality and undesirable species composition. These practices persist and continue to impact negatively on the quality of the region’s forests. Nevertheless, annual forest growth for the region exceeds annual harvest, and total standing volume of timber has increased steadily since the 1950s (Powell et al., 1994).

Oaks as components of the region’s forests The predominant oaks are black, white, scarlet, chestnut, post, northern red, southern red and bur oak (Fig. 1.11). These species typically occur in various combinations with hickories, sassafras, ﬂowering dogwood, blackgum, black cherry, red maple, and other upland oaks and deciduous tree species. The Ozark Highlands Section of the region, which covers southern Missouri, and parts of northeastern Oklahoma, northern Arkansas and southwestern Illinois, comprises one of the largest contiguous areas dominated by the oak–hickory association in the Central Hardwood Region. Many oak–hickory forests of today may have originated from extensive ﬁre-maintained oak savannas of the presettlement period; these formed closed canopy forests when ﬁres were suppressed (Johnson, 1993a; Olson, 1996). The oak cover types of the Central Hardwood Region include various combinations of oaks, hickories and other tree

Oak-dominated Ecosystems

29

Fig. 1.11 A mature black–northern red–white oak stand on a good site in the Central Hardwood Region of southeastern Ohio (Province 211a: Broadleaved Forests, Oceanic Province). (USDA Forest Service, North Central Research Station photograph.)

species that vary geographically (Appendix 2). Although hickories are common and persistent members of this forest type, they seldom represent more than a small proportion of trees in the main canopy of a mature forest (Braun, 1972). Oak–hickory forests develop on relatively dry sites where oaks persist as dominant members of the forest through successive disturbance events. This persistence is facilitated by the oaks’ drought tolerance and by light intensities in dry ecosystems that are sufﬁcient for the regeneration of the relatively shade-intolerant oaks (Bourdeau, 1954; Carvell and Tryon, 1961; Abrams, 1990). Oaks and hickories are found together on the drier sites throughout the region and comprise a commonly occurring species association. These forests dominate the landscape in the western part of the region. Elsewhere, oaks and hickories as a group commonly occur on dry ridges and southfacing slopes. On the more mesic sites, oaks are often interspersed with other hardwoods. Slope position and aspect strongly inﬂuence the spatial distribution of these forests and are thus useful in deﬁning ELTPs in much of the region (Figs 1.12 and 1.13).

From southern Illinois eastward and in northern Arkansas, the more mesophytic forests of the Central Hardwood Region generally include more complex species mixtures than found in drier oak forests (Fig. 1.14). Although oaks commonly share dominance with non-oaks on these sites, in the absence of recurrent ﬁre and grazing the oaks are often successionally displaced by more moisture-demanding and more shade tolerant non-oaks (Jokela and Sawtelle, 1985; Lorimer, 1985, 1989; Nowacki et al., 1990; Abrams, 1992). Understoreys of these stands are typically lacking in oak reproduction, especially large oak seedling sprouts. Over time, the dominance of oaks decreases while the proportion of non-oaks increases. The latter include various combinations of maples, American beech, black cherry, white ash, American basswood and yellowpoplar. Timber harvesting may accelerate the successional replacement of the oaks (Abrams and Nowacki, 1992). Diverse mixtures of hardwoods are common throughout the Ohio Valley, the Cumberland Plateau and Highland Rim areas of Tennessee and Kentucky, the Appalachian and Allegheny Plateau of

30

2.1, 2.2

1.3 sinkholes

Chapter 1

1.1, 1.2

1.1 Ultic rocky, 1.2 Silty RO Summits 2.1 Ultic, rocky RO/UG Shoulder (some 2.2)

3.1 Exposed, rocky RO/UG Ultic backslope (some 3.2 Alfic) 7.1 to 7.3 Cherty, non-cherty and glade, variable depth to dolomite 2.4 UG Alfic crypt reef bench

3.1, 3.2

Roubidoux (RO)

4.1 Protected RO/UG Ultic backslope (some 4.2 Alfic) 8.1 Cherty protected, variable depth to dolomite 2.3 UG Ultic crypt reef bench 4.1, 4.2

Upper Gasconade (UG)

Cryptozoan Reef

Upper Gasconade (UG) 10 Footslope

10 Footslope Exposed aspects 135–315˚

Protected aspects 315–135˚

Fig. 1.12. Ecological landtype phases (ELTPs) for the Ozark Highlands of Missouri (Province 221b: Broadleaved Forest, Continental Province; Ozark Highlands Section, upland ELTPs in the Current, Eleven Point and Black River Landtype Associations). Aspect, landform and bedrock geology are factors in the classification system. ELTP 1.1, 1.2, 1.3, 2.1, 2.3 and 3.1: pine–oak/Vaccinium dry ultic (chert) woodland; ELTP 2.2 and 3.2: mixed oak–pine/Desmodium, Vaccinium dry-mesic alfic (chert) woodland; ELTP 4.1: mixed oak–hickory/dogwood/Desmodium dry-mesic ultic (chert) forest; ELTP 4.2: mixed oak (white, red) dogwood dry-mesic alfic (chert) forest; ELTP 7.1: post oak (blackjack oak, pine) bluestem xeric chert woodland; ELTP 7.2: red cedar–hardwood/redbud dry dolomite woodland; ELTP 7.3: bluestem, Missouri coneflower dolomite glade; ELTP 8.1: mixed oak–sugar maple/redbud dry-mesic dolomite forest; ELTP 10: mixed oak (white)/dogwood drymesic alfic (chert) footslope forest. (From Nigh et al., 2000, used with permission.)

Oak-dominated Ecosystems

31

Brown County Hills Subsection

Crawford Upland and Escarpment Subsections

Fig. 1.13. Ecological landtype phases (ELTPs) for the forests of the Brown County Hills, Crawford Upland and Escarpment Subsections of southern Indiana (Province 221b: Broadleaved Forests, Continental Province; Interior Low Plateau, Shawnee Hills Section). Oaks and pines typically dominate the exposed (hotter) aspects whereas sugar maple, American beech, yellow-poplar and other shade tolerant hardwoods dominate the protected (cooler) aspects. (Adapted from Van Kley et al., undated.)

32

Chapter 1

Fig. 1.14. A large white oak (47 inches dbh) in Dysart Woods in southeastern Ohio (Province 221a: Broadleaved Forests, Oceanic Province). This 55-acre old-growth oak forest is dominated by white and northern red oaks and is the largest known remnant of the original mixed mesophytic forest of the Central Hardwood Region in southeastern Ohio. (USDA Forest Service, North Central Research Station photograph.)

western West Virginia and western Pennsylvania, the southern Lake States, and other parts of the region. Speciﬁc combinations of canopy dominants often form distinct geographic species groupings. Examples include the beech–maple forests of central Indiana and eastern Ohio, the maple–basswood–northern red oak forests of the driftless area of southwestern Wisconsin, and the black cherry–ash– yellow-poplar forests of the Allegheny Plateau of Pennsylvania. Toward the eastern end of the region, eastern white pine and eastern hemlock may be locally important members of mixed hardwood forests. Mixtures of oak and mesophytic species also occur in northern Arkansas in the Broadleaf Forest–Meadow Province (M222)

(Fig. 1.2; Table 1.2) (Braun, 1972). However, unlike the mixed mesophytic forests further to the east, yellow-poplar is absent. These are the most mesophytic forests in the western end of the region. Some of these combinations are formally recognized as cover types (Eyre, 1980); others form mixtures that are only locally distinguished silviculturally. It is within these mesic, mixed hardwood stands that northern red oak, one of the most commercially valuable tree species of the region, reaches its best development. It is also within these forests that the oaks are also the most difﬁcult to regenerate silviculturally (Carvell and Tryon, 1961; Arend and Scholz, 1969; Trimble, 1973; Loftis, 1988; Johnson, 1993b, 1994a,b). Mesophytic mixed hardwood forests generally occur where oak site index (chapter 4) is ≥ 65 ft at an index age of 50 years. Oak–pine mixtures occur most frequently in the southern and eastern parts of the region and are closely correlated with ﬁre and succession in old ﬁelds, heavily disturbed hardwood stands, and pine plantations. Oak–pine mixtures represent an early- to mid-stage in the succession toward oak–hickory or mixed hardwood forests. In the absence of ﬁre or other disturbances, oak–pine forests may change successionally from predominantly shortleaf pine, pitch pine or Virginia pine to hardwoods as the more shade tolerant hardwoods replace the intolerant pines (Cunningham and Hauser, 1989; Shefﬁeld et al., 1989; Smith et al., 1989; Orwig and Abrams, 1994). The oak–pine mixtures are important for maintaining biodiversity as well as economic timber production (Phillips and Abercrombie, 1987; Cooper, 1989; Kerpez and Stauffer, 1989; Leopold et al., 1989). Consequently, there is increasing interest in methods to create and maintain oak–pine forests (Waldrop, 1989). Speciﬁc combinations of oaks and pine vary with subregion and site quality. Because the pines tend to be associated with the driest (xeric) sites, the associated oaks often include species such as post oak and blackjack oak. On somewhat less xeric

Oak-dominated Ecosystems

sites, pines are commonly associated with black, white, scarlet, southern red or chestnut oaks. In the extreme northwestern part of the region in Minnesota and Wisconsin, jack pine and northern pin oak commonly occur together. Stands of eastern redcedar are closely afﬁliated with the oak–pine mixtures. Eastern redcedar is a common invader of old ﬁelds and glades (Lawson, 1990). It may eventually form dense pure stands if succession is allowed to progress unimpeded by disturbance. However, such stands are short-lived. As the redcedar matures and forms canopy gaps conducive to hardwood or pine regeneration, stands may succeed to oak–pine and oak–hickory mixtures.

The Southern Pine–Hardwood Region Geographic extent The Southern Pine–Hardwood Region includes broadleaved forests, conifer forests and various hardwood–pine mixtures. The region includes the two Subtropical Divisions (230 and M230) of the Humid Temperate Domain (Figs 1.2 and 1.6; Table 1.2). The region covers approximately 270 million acres of which 60% are forested. The extent of the Southern Pine–Hardwood Region is best illustrated by the joint ranges of the oak–pine and oak–gum cypress forest types (Fig. 1.5B and C). The region extends 1300 miles from eastern Texas to Virginia and occurs in a band extending 200–400 miles inland from the coast. At its northern boundary, the Southern Pine–Hardwood Region meets the Central Hardwood Region. Nearly 90% of the forest area in this region is privately owned. Four million non-industrial private forest owners control about 60% of all timberland. About 45% of this ownership is comprised of tracts smaller than 100 acres (Birch, 1996). The 25 national forests in the region comprise 9 million acres of timberland (Powell et al., 1994).

33

Climate, physiography and soil Annual precipitation in the region ranges from about 40 to 60 inches and is well distributed throughout the year. Mean annual temperature ranges from 60 to 70°F (16 to 21°C) and the growing season from 200 to 300 days (Bailey, 1995) (Fig. 1.7). The Southern Pine–Hardwood Region includes four major physiographic regions: the Piedmont (Province 232), the Coastal Plain (Province 231), the Interior Highlands (Province M231) and the lower Mississippi Valley (Riverine Intrazonal Province (R)) (Fig. 1.2). Gentle slopes characterize 50–80% of the area. Elevations range from sea level to 600 ft in the Coastal Plain, 300–1000 ft in the Piedmont, and up to 2600 ft in the Ouachita Mountains of the Interior Highlands. Numerous low-gradient streams, lakes, swamps and marshes characterize the ﬂat Coastal Plain. The wet habitats along the Coastal Plain, the Mississippi Valley and other major rivers support bottomland forest types that are largely absent from the Piedmont and Interior Highlands. The principal soil groups are Ultisols, Spodosols, Vertisols and Entisols, all of which tend to be low in fertility. Exceptions are the Inceptisols, which occur in the alluvial bottoms of the Mississippi River (Bailey, 1995).

Forest history Here, as in other regions, ﬁre greatly impacted the early forests. Fire was an essential tool for maintaining agricultural openings, eliminating brush and hardwood reproduction from pine forests, and increasing forage for grazing. Native Americans regularly burned the forests where they lived. Increased burning associated with European settlement increased the proportion of pine in the region relative to earlier periods (Pyne, 1982; Skeen et al., 1993). Fire was combined with land clearing to open the hardwood forests of the south for agriculture. Even today, forest burning is a prominent practice throughout the region.

34

Chapter 1

In the Piedmont and alluvial river bottoms, vast areas were cleared for agriculture before industrial logging peaked in the region (Sargent, 1884; Hodges, 1995). Logging and the production of naval stores began on a small scale in the 1600s. But by 1880, forests accessible by water and close to population centres were heavily cut over. Charles Mohr noted the rapidity at which the cypress swamps were being logged in some localities and the apparent lack of forest regeneration. He observed that ‘the large number of logs harvested shows clearly with what activity the destruction of these treasures of the forest is being pushed; and the reports, as of heavy thunder, caused by the fall of the mighty trees, resounding at short intervals from near and far, speak of its rapid progress’ (Sargent, 1884). However, Mohr also noted that immense areas of pine forest remained unaffected by logging and that many former hardwood forests that were earlier cleared for agriculture had reverted to pine after their abandonment. The South did not become the centre of the US logging industry until shortly after logging peaked in the Lake States in 1890. By 1900, lumber production in the South exceed that of the Lake States and by 1910 the South produced half of all US lumber. The movement of large lumber companies to the Southern Pine–Hardwood Region coincided with technological advances that increased the speed with which logs could be removed from the woods and transported to the mills (Williams, 1989). Steam powered stationary skidders and loaders were mounted on boats and railcars. As rail lines were extended into the southern forest, logging trains followed and systematically removed virtually all timber within the long reach of a cable skidder mounted on a rail car. The joint enterprise of rail construction and logging greatly accelerated the harvest of southern pines (Williams, 1989). By 1925 southern lumber production began to decline and western production increased. Following the Great Depression, timber production in the South never returned to the levels of 1910–1930, and

the bulk of US timber production moved west. The subsequent establishment of southern paper mills coupled with successful ﬁre prevention and a reduction in openrange grazing accelerated the reforestation of one million cutover acres. This gave rise to the South’s ‘third forest’ which today again produces a greater volume of wood than any other region of the United States.

Oaks as a component of the region’s forests The oak forests in this region can be divided into upland and lowland types. Were it not for the complex spatial intermingling of upland and lowland forests, they could be treated as two ecologically distinct regions. The upland and bottomland oak forests of the region differ substantially in species composition, ecology and the application of silvicultural practices. The Southern Pine–Hardwood Region today includes about 172 million acres of timberland. Of the broadly deﬁned oak forests recognized in national inventories, the oak–hickory group occurs on 55 million acres or one-third of the region’s timberland. Oak–gum–cypress and oak–pine each occur on an additional 16% of the timberland. Thus, oak forests collectively cover more than 60% of the region. Loblolly-shortleaf pine and longleaf-slash pine make up most of the remaining forest acreage. A more detailed cover type classiﬁcation (Eyre, 1980) recognizes 63 cover types that occur within the region (Walker, 1995) – 15 of those include oaks as primary species, and several others include oaks as important associated species (Appendix 2). Southern silviculture has largely focused on pine, especially on industrial forestlands. There, intensive silviculture is commonly practised to maximize timber and wood ﬁbre yields through site preparation, planting genetically improved seedlings, frequent thinning, prescribed burning, and the use of fertilizers, herbicides and pesticides. However, annual softwood removals are nearly equal to annual growth and may soon exceed annual growth (Walker, 1995).

Oak-dominated Ecosystems

The importance of pine in the Piedmont is related to the region’s history – the historical sequence of lumbering, land clearing and farming deforested large areas that were abandoned before 1930 and burned frequently. This disturbance favoured the establishment of pine forests, which greatly increased in acreage relative to other species. Oaks and other hardwoods occur in most natural southern pine stands, and on these sites they increase in importance through succession. Fire suppression, silvicultural thinnings and partial harvests often accelerate this trend (Skeen et al., 1993). The large oak–hickory acreage in the region, the increasing hardwood volumes in pine–hardwood mixtures, and the nearly full utilization of the annual pine growth in the region has recently shifted the utilization of the region’s forests towards the hardwoods. Much of this change has resulted from the utilization of hardwood chips for paper production and composite products. Chips can be made from low quality, small diameter (> 4 inches) hardwoods. This technology created new markets for the abundant low-quality trees that previously had been considered a silvicultural liability. However, this utilization capability has also raised concerns about the potential for overutilization of hardwoods, especially well formed, small hardwood trees that comprise the future hardwood growing stock for solid hardwood products. The region’s oak–hickory forests attain their best development along the border separating the Southern Pine–Hardwood Region and the Central Hardwood Region. Closely related to the oak–hickory forests are mixtures of oak and pine (Fig. 1.15). These mixed forests are increasingly recognized for their importance in maintaining forest biodiversity and their historical importance in the region. The oak–pine type, which occurs on 16% of the timberland of the region, rates high in aesthetic appeal and species richness compared to even-aged pine stands. However, relatively little is known about the longterm management and productivity of

35

Fig. 1.15. A black oak–white oak–shortleaf pine stand in the Ozark Highlands of Missouri (Province 221b: Broadleaved Forests, Continental Province; Ozark Highlands Section). (USDA Forest Service, North Central Research Station photograph.)

oak–pine mixtures for lumber, ﬁbre or other values. In the absence of disturbance, the oaks tend to successionally displace the pines and harvesting the pine often accelerates the process. Southern bottomland hardwoods commonly include 11 species of oaks (cherrybark, Delta post, laurel, Nuttall, overcup, pin, Shumard, swamp chestnut, water, white and willow oaks) (Hodges, 1995). These oaks occur in mixture with other bottomland species along the major rivers of the Coastal Plain as well as the lower reaches of the Mississippi, Arkansas, Missouri, Ohio and Wabash Rivers (Fig. 1.16). In total, southern bottomland hardwoods cover more than 27 million acres (16% of the region’s timberland) and are physiographically and ecologically distinct from surrounding upland oak–hickory, oak–pine and pine forests.

36

Chapter 1

Hodges (1995) reduced the number to three types (cottonwood–willow, baldcypress–tupelo, mixed bottomland hardwoods), and the USDA Forest Service usually reports only two types (oak–gum–cypress, elm–ash–cottonwood) in regional summaries.

The Forest–Prairie Transition Region Geographic extent

Fig. 1.16. A cherrybark oak–sweetgum bottomland stand near the Tombigbee River in Alabama (Province 232: Coniferous-Broadleaved Semievergreen Forests Province). (USDA Forest Service, Southern Research Station photograph.)

Although bottomland forests are relatively ﬂat, elevational differences of only a few feet alter soil formation processes, soil moisture regimes and species composition. Thus, changes in species composition are often associated with relatively minor differences in physiography (Fig. 1.17). Moreover, ﬂoodplain physiography can quickly and frequently change as a result of scouring and deposition of sediments. These factors, coupled with the high tree species diversity of bottomland forests, complicate classifying forest types and developing silvicultural prescriptions appropriate to each. Up to 70 tree species occur in southern bottomland forests (Putnam et al., 1960), and species mixtures often change over short distances within stands. Consequently, species associations are difﬁcult to classify meaningfully into more than a few broad types. Although Eyre (1980) listed 14 bottomland cover types (six named for oaks),

Within the United States, the Forest–Prairie Transition Region extends from southern Texas northward to Minnesota and North Dakota (Fig. 1.6). The region coincides with two ecoregion provinces: Forest–Steppes and Prairies (251) and Prairies and Savannas (252) within the Prairie Division (250) (Fig. 1.2, Table 1.2). The region includes Braun’s (1972) Grassland or Prairie Region, Forest–Prairie Transition and Prairie Peninsula Sections, which fall within her Oak–Hickory Forest Region. As its name implies, the Forest–Prairie Transition Region is transitional between the eastern forests and the prairies and dry woodlands of the Dry Domain to the west. On its eastern border, the region adjoins the Northern Hardwood Region, the Central Hardwood Region and the Southern Pine–Hardwood Region. The Forest–Prairie Transition Region spans 1400 miles in latitude and varies in width from as little as 100 miles along the Canadian border to 600 miles between eastern Nebraska and western Indiana. The region includes approximately 191 million acres, about 7% of which are forested. Between Canada and Oklahoma, forests cover 5% of the landscape with most of the remainder devoted to tilled cropland or pasture. The forest cover increases to 13% in parts of Oklahoma and Texas. Most of the forestland in the Forest–Prairie Transition Region is privately owned. Although there are three national grasslands within the region, only 15,000 acres of national forest land (in central Missouri) are included.

River base level

R ID Sweetgum, hickory G Red oak, swamp chestnut oak E Winged elm, black tupelo gum FL SL A Sweetgum, water oak,willow oak O T U Green ash, nuttall oak G H Overcup oak Water hickory

SW AM P Cypress, water tupelo

River birch LE VE Beech, yellow poplar E Sycamore, sweetgum, spruce pine Oaks FL AT SL Sweetgum, oaks, hickories O Blackgum, winged elm U G H Cypress, swamp tupelo FL AT Overcup oak, willow oak Nuttall oak TE R R AC White oak, red oak E Hickories, sweetgum Yellow poplar Loblolly pine

Minor Bottom

37

BA R

River base level

BA R Willow FR O Cottonwood N T Elm, sycamore

Major Bottom

Pecan, sugarberry F SL LA T O Nutall oak, green ash U G H Sugarberry, elm, red maple R ID Willow G Overcup oak, water hickory E Sweetgum, wateroak Willow oak, green ash FL AT Overcup oak, water hickory

Oak-dominated Ecosystems

Fig. 1.17. The topographic distribution of southern bottomland oaks and associated species in major and minor stream valleys of the Southern Pine–Hardwood Region (Province 232: ConiferousBroadleaved Semi-Evergreen Forests Province). (Reprinted from Hodges and Switzer, 1979, by permission of Society of American Foresters, Bethesda, Maryland. Not for further reproduction.)

Climate, physiography and soil Precipitation within this vast region varies from less than 20 inches per year in the north to 55 inches along the gulf coast of Texas. One-half to two-thirds of the precipitation typically falls during the growing season and snowfall is common north of Texas. From north to south, mean annual temperature ranges from 36° to 70°F (2 to 21°C) with corresponding growing seasons ranging from 111 to 320 days (McNab and Avers, 1994) (Fig. 1.7). Throughout the region precipitation is largely offset by evapotranspiration, creating soil moisture conditions in many localities that are marginal for tree growth.

Most of the region comprises gently rolling plains, although high rounded hills occur and steep bluffs border some river valleys. Elevations range from sea level to 2000 ft. Local relief is less than 165 ft throughout most of the region, but it reaches 500 ft in the Flint Hills of Kansas (McNab and Avers, 1994). Soils are predominantly Mollisols although Vertisols occur on the prairies, and Alﬁsols occur on savannas and within the Mississippi Valley (Bailey, 1995).

Forest history Native Americans who were largely nomadic inhabited the region for at least 10,000 years. Crops were cultivated as

38

Chapter 1

early as 1000 years ago. A few large Native American communities developed in the major river valleys. One of these was Cahokia (near present-day St Louis), which ﬂourished between AD 1000 and 1400 with an estimated population of 25,000. The forests in the region were an important resource for both nomadic people and larger permanent communities. The demise of Cahokia may have been caused by the exhaustion of the surrounding forests that were used for fuel and for the construction and maintenance of a 2mile-long perimeter wall around the city (Lord, 1999). Frequent ﬁres were essential to the maintenance of prairie and savanna vegetation in many parts of the Forest–Prairie Transition Region, and Native Americans burned the grasslands and woodlands where they lived. Grazing and trampling by herds of bison and other ungulates also were important in maintaining prairies and preventing the encroachment of forests and other woody vegetation. By mid-19th century, European settlers began farming the prairies and draining prairie wetlands. The latter produced some of the nation’s most productive agricultural lands. Trees were largely conﬁned to riparian corridors, steep slopes and scattered savannas. With the exclusion of ﬁre and the elimination of free-ranging ungulates, forests frequently encroach upon abandoned ﬁelds and pastures. In 1884, Sargent (1884) stated that, ‘Dakota, with the exception of its riverlands and the small territory between the north and south forks of the Cheyenne River, is practically destitute of timber. The bottoms of the principal streams contain extensive groves of hardwood.’ In Iowa he observed that ‘since the ﬁrst settlement of the state the forest area has increased by the natural spread of trees over ground protected by ﬁre, and by considerable plantations of cottonwood, maples, and other trees of rapid growth made by farmers to supply fuel and shelter’. Further south, in Texas, Charles Mohr noted, ‘The timber growth immediately west of the Brazos is stunted and scanty; large areas of grass land intervene between

the scrubby woods until all at once ligneous growth disappears and the seemingly boundless prairie, in gently undulating swells expands before the view on all sides’ (Sargent, 1884). Since that time, farms have been established on virtually all the lands suitable for row crops or forage production (McNab and Avers, 1994). Depending on the farm economy, the forested acreage has decreased or increased as forests and woodlands were cleared to create more farm land, or as marginal farm land reverted to forest through tree planting or abandonment.

Oaks as components of the region’s forests The best forest development in this region occurs on its eastern border where it abuts the Northern Hardwood Region, the Central Hardwood Region and the Southern Pine–Hardwood Region. Of the 7% of the Forest–Prairie Transition Region that is forested, three-quarters is classiﬁed as oak–hickory or oak–pine. Few of the savannas that formerly occupied the transition zone between forest and prairie exist today. The prairie ﬁres that historically restricted the extent of the region’s forests have been replaced by agricultural practices that now limit most forests to riparian areas or to slopes unsuitable for forage or other crops (Fig. 1.18). Before the mid-19th century, ﬁre was the primary regulator of the distribution of tree species in the region. Narrow bands of forest along streams and ravines, sometimes called gallery forests, provided refuges for trees from the frequent ﬁres that burned across the prairie. Oaks dominated many of these forests. With the advent of farming in this region, the frequency of wildﬁres was greatly reduced. This allowed the gallery forests to expand into untilled areas that were formerly covered by native grasses (Abrams and Gibson, 1991). However, the invading woody species were generally species such as American elm, hackberry and eastern redcedar rather than oaks. The

Oak-dominated Ecosystems

39

A

B

Fig. 1.18. (A) Aerial view of the distribution of forests in the Forest–Prairie Transition Region (Province 251: Forest-Steppes and Prairies Province) of northwestern Missouri. Throughout much of this ecoregion, forests are largely restricted to narrow belts occupying steep slopes along rivers and drainages interspersed with agricultural lands. (B) Forested bluffs dominated by oaks (background) along the Missouri River in central Missouri fronted by cultivated bottomland fields. Before settlement by Europeans, these bottomlands were covered by lowland forests dominated by American elm, silver maple, green ash, eastern cottonwood, bur oak and pin oak. (USDA Forest Service, North Central Research Station photographs.)

reduction in wildﬁres also allowed those species to increase in abundance within existing forests that were formerly dominated by oaks, especially on the more mesic sites. In much of this region, frequent ﬁres are required to prevent the displacement of the oaks by other species (Penfound, 1968; Abrams, 1988; Abrams and Gibson, 1991).

Oak–hickory forests extend from the Central Hardwood Region westward across eastern Oklahoma and into northern Texas (Fig. 1.5A). From east to west the forests become increasingly scrubby and open. An exception is the relatively dense oak forest of the Cross Timbers Region. In Texas, the Cross Timbers comprise two bands of scrubby oak woodland extending 175 miles

40

Chapter 1

southward from the Oklahoma border. These bands are 20–50 miles wide and separated by the Fort Worth Prairie. Forest cover occurs along outcrops of sandy soils of greater porosity than adjacent prairie soils (Braun, 1972). The Cross Timbers were prominent landmarks for westward travellers who otherwise traversed relatively open landscapes (Dyksterhuis, 1948). Although the heavier forest cover in the Cross Timbers area of Texas is somewhat evident from Fig. 1.6, the two distinct strips of woodland are not distinguishable at the resolution shown. Post oak and blackjack oak are the dominant tree species and account for 60% and 20% of the trees, respectively. Except in ﬂoodplains, these oaks seldom exceed 12 inches in diameter and 30–45 ft in height. At one time, the herbaceous vegetation in the Cross Timbers was probably similar to that of the surrounding prairie, but grazing during the last century has greatly altered the species composition of the herbaceous layer (Dyksterhuis, 1948). The Cross Timbers vegetation extends northward through Oklahoma and eventually disappears in southern Kansas. Except for the Cross Timbers Region, the upland woodlands of eastern Oklahoma were formerly post–blackjack oak savannas maintained by frequent ﬁres. Grazing and a reduction in burning have since reduced grass cover and facilitated the establishment of dense tree reproduction in many areas; post and blackjack oaks dominate most stands. Although the average basal area of these forests historically has been relatively low, in the absence of burning it has increased from 49 ft2 acre1 in 1957 (Rice and Penfound, 1959) to 80 ft2 acre1 in 1993 (Rosson, 1994). From Kansas northward there are few forests, but where they do occur, oaks often dominate (Figs 1.2 and 1.6). Many of the oak forest types and conditions occurring in the Central Hardwood Region extend westward through the central portion of the Forest–Prairie Region. The central part of the region is capable of supporting forest vegetation and is successional to forest in

areas protected from cultivation. However, because agriculture is the dominant land use, forests are usually restricted to riparian corridors, wet areas, steep slopes and highly erodible lands, or other sites unsuited to agriculture. Nevertheless, oaks and other hardwoods often develop into commercially valuable stands in those parts of the region lying within Illinois, Iowa, northern Missouri and eastern Kansas. Bur oak is the dominant oak species in the northern reaches of the Forest–Prairie Transition Region. It is the only major oak species with a natural range that extends across western Minnesota and into the Dakotas. Bur oak is well adapted to this region because its deep taproot makes it resistant to drought and able to invade prairie grasslands (Johnson, 1990). Its thick bark makes it highly resistant to ﬁres that eliminate most other woody species. Bur oak also thrives on moist alluvial bottoms that support dense hardwood forests in the northern portion of the Forest–Prairie Transition Region. Here, the bur oak type covers approximately 2% of the land area and is the principal forest type. Cottonwood, quaking aspen and American elm are other abundant hardwoods in the northern part of the Forest–Prairie Transition Region.

Western Oak Forests The Southwestern Desert–Steppe Region Geographic extent The Southwestern Desert–Steppe Region includes the scattered oak forests of Arizona, New Mexico, southern Utah, west Texas and southwest Oklahoma (Figs 1.2, 1.6; Table 1.3). Although the range of Gambel oak extends northward as far as southern Wyoming, the oaks there are a small component of the vegetation. Forests and woodlands cover about 20% of the area, but only 7% of this is considered productive forest. Soil moisture deﬁciencies limit the distribution of oaks and other plant life throughout the region. Oaks

Oak-dominated Ecosystems

occur as scattered trees and in open woodlands. Their distribution within the region is often limited to discontinuous elevational zones that provide the required regime of precipitation and temperature. The region lies entirely within the Dry Domain and comprises parts of three Divisions: Tropical/Subtropical Steppe (310), Tropical/Subtropical Steppe Mountains (M310) and Tropical/Subtropical Desert (320). Included are six ecoregion Provinces: Coniferous Open Woodland and Semideserts (311), Steppes and Shrubs (313), Shortgrass Steppes (314), Steppe or Semidesert–Mixed Forest–Alpine meadow or Steppe (M311), Semideserts (321), and Deserts on Sand (323) (Fig. 1.2; Table 1.3). The Southwestern Desert–Steppe Region extends 1200 miles from northwestern Arizona to the Gulf of Mexico in southern Texas. It varies from 300 to 700 miles in width, and encompasses roughly 250 million acres including the Mojave Desert, the Sonoran Desert, the Painted Desert, the Colorado Plateau, the southern Rocky Mountains, Texas High Plains and the Edwards Plateau. Within the region, oak forests are widely scattered and cover only a small fraction of the landscape (Fig. 1.2). The federal government or Native Americans own two-thirds of the forests and woodlands in Arizona and New Mexico, but in Texas and Oklahoma most are privately owned (Powell et al., 1994).

Climate, physiography and soil A deﬁning characteristic of this region is a rate of surface evaporation that exceeds precipitation. The climate varies from dry to desert. Annual precipitation ranges from less than 10 inches to 30 inches (Bailey, 1995). Even in areas with greater precipitation, high rates of evaporation limit moisture availability. Average annual temperature ranges from 40 to 70°F (4 to 21°C). Although temperatures decrease with increasing elevation, mean monthly temperatures generally exceed 32°F (0°C) (Fig. 1.19). Elevation ranges from sea level along the southern Texas Gulf Coast to 7000 ft in the Colorado Plateau; some

41

mountain peaks are substantially higher. Soils are variable throughout the region and include Mollisols, Aridisols and dry Entisols (Bailey, 1995).

Forest history As in other regions of the United States, Native Americans customarily burned the forests and woodlands where they lived. Lightning was also a common cause of combustion. These ﬁres maintained an open understorey in the extensive ponderosa pine forests of higher elevations. In 1880 alone, about 75,000 acres burned – which accounted for 0.1 to 1% of the woodland within the settled area (Sargent, 1884). Beginning in the mid-19th century, European settlers were drawn to the region by opportunities for mining and livestock production. Lands were not suitable for agriculture, and the great land clearing that decimated the eastern oak forests did not occur here. However, logging, grazing and changes in ﬁre regimes changed the species composition of forests and woodlands. In recent decades, the suppression of ﬁres has increased the amount of tree reproduction, especially conifers, and decreased grasses and forbs growing beneath forest canopies (Long, 1995). Throughout the region, oaks have historically had little commercial value. In 1884, Sargent (1884) described the forests in and around New Mexico: ‘The deciduous trees of this entire southwestern region, often of considerable size, are generally hollow, especially the oaks; they are of little value for any mechanical purpose, although affording abundant and excellent fuel.’ Then, as now, ponderosa pine was the principal timber species.

Oaks as components of the region’s forests Forests and woodlands cover only a small portion of the total area in the region, and the oaks comprise only a small percentage of that. In Arizona and New Mexico, only 15% of the land base is forested. Only 3% of the area of those two states can produce more than 20 ft3 of timber per acre per year,

42

Chapter 1

Astoria, OR

51˚F 76 in.

Pasadena, CA 62˚F 19 in.

Tahoe, CA 42˚F 31 in.

Abilene, TX

65˚F 25 in.

Brawley, CA 72˚F 2 in.

Colorado Springs, CO

48˚F 14 in.

Fig 1.19. Representative climates for selected ecoregion Divisions in the western United States. Mean monthly precipitation is shown by the solid lines (right axis) and temperature by dashed lines (left axis). Mean annual values are given above each graph. Periods of drought are indicated where the precipitation line falls below the temperature line (e.g. as in Division 260). Division boundaries are shown in Figs 1.2 and 1.5. (Ecoregion and climatic data from Bailey, 1995.)

Oak-dominated Ecosystems

and virtually none of that is oak forest. Commercial forests include ponderosa pine (75%), Douglas-ﬁr (13%), spruce–ﬁr (9%) and aspen (3%). The only recognized oak cover type here is western live oak (Appendix 3) (Eyre, 1980). It occurs at elevations from 4000 to 6000 ft in the foothills and lower mountain slopes of Arizona and New Mexico. At higher elevations, the western live oak cover type gives way to ponderosa pine and pinyon–juniper, with oak–conifer mixtures occurring in the transition. At lower elevations the western live oak type yields to an open growth of shrubby evergreen oaks. Mesquite and desert vegetation typically occurs below that. Characteristic species of the western live oak type include Emory, Arizona white, Mexican blue and silverleaf oaks (Eyre, 1980) (Fig. 1.20). Ajo oak, Dunn oak, grey oak and Havard oak also occur in Arizona and New Mexico. At the eastern end of the Southwestern Desert–Steppe Region towards the High Plains and Edwards Plateau of west-central Texas, precipitation increases and oaks become more prominent. The Mohr (shin) oak forest type covers more than 8 million acres in Texas where it develops best under

43

20–25 inches of precipitation annually (Eyre, 1980). However, that amount of precipitation represents the upper end of the range for the region (e.g. see Fig. 1.19, Division 310). Other oaks that occur in west-central Texas include Arizona white, blackjack, bur (marginally), chinkapin, Durand, Emory, Havard, Lacey, live, sandpaper, Texas and Texas live oaks.

The Pacific Mediterranean–Marine Region Geographic extent The Paciﬁc Mediterranean–Marine Region includes the oak forests and woodlands of California, Oregon and Washington (Fig. 1.6). The region lies within the western portion of the Humid Temperate Domain and includes the Mixed Forest–Coniferous Forest–Alpine Meadow Province (M261) and the Mediterranean Woodland or Shrub–Mixed Coniferous Forest–Steppe or Meadow Province (M262) within the Mediterranean Mountains Division (M260) (Fig. 1.2). Oaks also occur within the Coast Ranges of California, which includes the Mediterranean Hardleaved Evergreen Forest,

Fig. 1.20. Emory oak woodland in the Peloncillo Mountains of southwestern New Mexico, Coronado National Forest, New Mexico (Province M321: Semideserts Province). (USDA Forest Service, Rocky Mountain Research Station photograph.)

44

Chapter 1

Open Woodlands and Shrubs Province (262) and the Redwood Province (263). At its northern extent, the Paciﬁc Mediterranean– Marine Region also reaches the Mixed Forest Province (241) of the Marine Division (240) in Oregon and Washington. The Paciﬁc– Mediterranean–Marine Region also includes California’s Central Valley (province 261) and the mountainous zones of Washington and northern Oregon (M261) where oaks are not abundant. The Paciﬁc Mediterranean–Marine Region extends nearly 900 miles from Washington to southern California but less than 200 miles from the Paciﬁc Ocean to the eastern slopes of the Sierra Nevada Mountains. Although the region covers about 75 million acres, the oaks are limited to relatively narrow elevational zones. In California, Oregon and Washington, slightly less than half the timberland is publicly owned (Powell et al., 1994). In contrast with the eastern United States, most of the privately owned timberland in this region is held by corporations rather than by nonindustrial private owners (Birch, 1996). However, the ownership of oak forests and woodlands does not follow this trend; about three-quarters of that acreage is in non-industrial private ownership (Thomas, 1997).

Climate, physiography and soil Climate is strongly inﬂuenced by the Paciﬁc Ocean and by the Coast and Sierra Nevada Ranges, which dominate the physiography of the region. Elevations range from sea level to more than 14,000 ft. In the mountain ranges, increasing elevation is associated with decreasing temperatures and variation in precipitation. For a given elevation, precipitation is generally greater on western slopes than on eastern slopes. Latitude also inﬂuences climate so that a given climatic zone occurs, from north to south, at progressively higher elevations. However, mountainous topography creates climatic irregularities and discontinuities, and the distribution of oaks and associated tree species varies accordingly. Most of the precipitation occurs during the autumn, winter and spring. Annual pre-

cipitation generally ranges from 10 to more than 60 inches in the ecological provinces where oaks occur. Temperature extremes and moisture stress are reduced near the coast where fog supplements precipitation and the ocean reduces ﬂuctuations in temperature. Elsewhere the region’s Mediterranean climate is characterized by 2 to 4 months of drought during the summer (Table 1.3, Fig. 1.19). Low precipitation generally occurs at lower elevations and on the east faces of mountain ranges. Soils include Ultisols, Alﬁsols, Mollisols, Entisols and Inceptisols (Bailey, 1995).

Forest history The historical importance of oaks is recorded in ancient bedrock mortars that were used by Native Americans to grind acorns into ﬂour. Acorns were a staple food of Native Americans in this region, and Biswell (1989) suggests that oaks were so important to their diet that they burned oak woodlands to both encourage oak reproduction and to facilitate acorn gathering. Although human-caused ﬁres have been historically associated with the oaks of the region for thousands of years, there is uncertainty about what proportion of the landscape was regularly affected by humans. During the post-settlement period of 1850–1950, the mean interval between ﬁres in the oak–pine forests of the foothills of central California was 8 years (Stephens, 1997). Commercial logging in the region has largely focused on the conifers. In 1884, Sargent (1884) stated: The forests of California, unlike those of the Atlantic States, contain no great store of hardwoods. The oaks of the Paciﬁc forests, of little value for general mechanical purposes, are unﬁt for cooperage stock. No hickory, gum, elm, or ash of large size is found in these forests, California produces no tree from which a good wine cask or wagon wheel can be made. The cooperage business of the state, rapidly increasing with the development of grape culture, is entirely dependent upon the forests of the Atlantic region for its supply of oak.

Oak-dominated Ecosystems

Sargent further noted that large quantities of chestnut oak (sic tanoak), once common in the northern Coast Range of California, are ‘now becoming scarce and in danger of speedy extermination’ due to utilization by the tanning industry. Sargent’s reference to the oaks of Oregon and Washington is slightly less disparaging. In the Willamette Valley, he noted that Oregon white oak woodlands were becoming re-established after reductions in ﬁre frequency. Along the Yakima River in Washington, he noted that Oregon white oaks were limited to 15 ft in height and 6 inches in diameter. The logging industry on the Paciﬁc Coast was established in the 18th century under Hispanic inﬂuence. Through the middle of the 19th century the relatively small industry served markets in South America, Australia and the Paciﬁc Rim (Williams, 1989). The gold rush of 1849 and the completion of the transcontinental railroad opened additional markets, but the increase in lumber production in this region occurred gradually, beginning about 1900 when the large timber companies and railroads moved west after exhausting the ready supply of timber in the Lake States. Increases in timber production in the region continued into the Great Depression, but output eventually dropped by 75%. By 1950, however, annual timber production in the West exceeded 16 billion board feet annually, which was greater than that produced in other regions of the United States. Today lumber production in this region lags signiﬁcantly behind that of the south. Harvest of hardwood growing stock has remained nearly constant since 1976, but the volume of hardwoods harvested annually is only about 5% of the region’s total. Historically, oak forests were little affected by commercial logging, but locally they were widely utilized for ﬁrewood and fence posts. Ranchers and farmers had the greatest inﬂuence on the oak woodlands of the foothills and lower slopes as a consequence of clearing them for agriculture and grazing. Sargent (1884) noted:

45

The permanence of the mountain forests of California is severely endangered, moreover, by the immense herd of sheep, cattle, and horses driven to the mountains every year, at the commencement of the dry season, to graze. From the foothills to the highest alpine meadows, every blade of herbage and every seedling shrub and tree is devoured.

In California, oak woodlands were reduced from an estimated 10–12 million acres to about 7 million acres today (Thomas, 1997). The oak woodlands are predominantly owned by farmers and ranchers, and between 1945 and 1970 the primary loss of woodland acreage resulted from conversion to rangeland. Invasion of non-native grasses and the suppression of ﬁre have created problems in maintaining oak woodlands and savannas (see Chapter 9 for details of savanna restoration and management). More recently, the greatest losses of oak woodland have resulted from suburban residential development (Bolsinger, 1988). This has given rise to concern for property damage from the wildﬁres historically associated with the oak woodlands.

Oaks as components of the region’s forests Most of the region’s oak forests and woodlands occur in California where they account for approximately one-quarter of the wooded acres. Oaks surround California’s Central Valley in the foothills of the Sierra Nevada, Cascade and Klamath Ranges (Figs 1.5 and 1.6). Although oaks were formerly abundant within parts of California’s Central Valley (province 261), their distribution has been greatly reduced there (Grifﬁn, 1977). Oaks also occur on the western slopes of the Coast Ranges in central and southern California. The range of Oregon white oak extends northward into central Oregon and Washington in the Willamette Valley and the Puget lowlands between the Cascade and Coast Ranges. Included are 18 species of oak trees and shrubs plus additional hybrids (Bolsinger, 1988; Thomas, 1997). Eight oak species that reach tree size are abundant: California black, blue, interior live, coast live, canyon live, valley, Oregon white and Engelmann oaks (Plumb and McDonald, 1981).

46

Chapter 1

Western oak forests are often categorized as either timberland (forests suitable for commercial wood production and capable of producing at least 20 ft3 acre1 year1 of merchantable volume), or woodlands (sites of lower productivity primarily utilized for forage and ﬁrewood). In California, only about 1 in 4 acres of hardwood forest qualiﬁes as timberland. Oak woodlands are sparsely covered with trees compared to oak timberlands. The statewide volume of oaks in woodlands and in timberlands is nevertheless nearly equal because the acreage of woodlands is approximately three times that of timberlands (Table 1.4). Three-quarters of the oak woodlands are grazed and these account for about onethird of California’s total forage (Thomas, 1997). Only about 500,000 board feet of hardwood lumber was produced in California in 1992 (Ward, 1995). The combined effects of temperature and precipitation (which latitude, elevation, slope and aspect affect) regulate the distribution of oaks. In the Paciﬁc Mediterranean–Marine Region, many oak forests and woodlands are restricted to elevational zones in the transition between grassland and chaparral at lower elevations and coniferous forest at higher elevations. Mean temperatures within the region increase with decreasing latitude, and the

oaks occur at higher elevations at lower latitudes. Due to the interaction of climate and mountainous topography, the distribution of oaks in this region is more geographically restricted than in the eastern United States. Several classiﬁcation schemes have been proposed for the complex vegetation relationships that occur in the Paciﬁc Mediterranean–Marine Region (e.g. Grifﬁn, 1977; Paysen et al., 1980, 1982; Barbour, 1988; Allen, 1990) (Fig. 1.21). Eyre (1980) recognized ﬁve oak cover types and two additional types where oaks commonly occur in mixtures with other species (Appendix 3). The Oregon white oak type is found in the northern portion of the Paciﬁc Mediterranean–Marine Region from northern California to Vancouver Island. This type occurs at lower elevations (0–3900 ft) and primarily in inland valleys or lower slopes between the Coast Ranges and the Cascade or Sierra Nevada Ranges (Eyre, 1980). The type makes its best development in the vicinity of the Willamette Valley where closed-canopy Oregon white oak stands developed from former oak savannas when periodic ground ﬁres were excluded (Thilenius, 1968). The species also occurs in mixtures with other hardwoods and conifers including California black oak, canyon live oak, ponderosa pine and Douglas-ﬁr (Appendix 3).

Table 1.4. Standing oak volumes in California timberlands and woodlands. Although oaks make up 63% of the total volume of California’s hardwoods, oak timberlands (commercial forest lands) comprise only 8% of the 50 billion cubic feet total volume (softwoods plus hardwoods on California timberlands).a

Species California black oak Canyon live oak Blue oak Coast live oak Oregon white oak Interior live oak California white (valley) oak Engelmann oak Total oak Total hardwoods (all species) a

Volume in timberlands (million ft3)

Volume in woodlands (million ft3)

Total (million ft3)

Species total as a proportion of all oaks (%)

2,254 1,302 1 126 211 45 34 0 3,973 7,661

277 731 1,112 755 389 508 164 10 3,946 4,855

2,531 2,033 1,113 881 600 553 198 10 7,919 12,516

32 26 14 11 8 7 3 0 100 –

Adapted from Shelly (1997) and Bolsinger (1980, 1988).

Oak-dominated Ecosystems

47

14,500 12,000 Alpine

Meadow

Subalpine forest Lodgepole pine forest

6000

4000

2000

Jeffrey pine forest

Red fir forest

8000

White fir forest

Upland live oak woodland

d an wl k Lo oa d e n liv dla o wo

d xe ral Mi par a ch

Juniper woodland

Elevation (ft)

10,000

Ponderosa pine forest Mixed evergreen forest and black oak woodland

Chamise chaparral

Blue oak woodland

Mesic

Xeric Topographic moisture gradient

Fig. 1.21. Relation of oak forests to elevation, moisture gradients and other forest types found in the Pacific–Mediterranean–Marine Region of California (Ecoregion Provinces 261, 262, 263 and M261). Oak forests and woodlands are usually found above the chaparral zone and below the ponderosa pine zone. (Redrawn from Barbour (1988) and Vankat (1982).) Reprinted with permission of Cambridge University Press.)

California black oak attains a greater volume (Table 1.4) and is distributed across a greater area than the other California oaks (Plumb and McDonald, 1981). The California black oak type occurs from central Oregon to the Mexican border across elevations ranging from 200 to 8000 feet with corresponding annual precipitation of 25–85 inches annually. Best development of the forest type occurs in the northern half of California in the Klamath and Cascade Mountains and the Coast and Sierra Nevada Ranges. There the forest type is found at elevations between 1500 and 3000 ft with corresponding annual precipitation between 30 and 50 inches (Eyre, 1980). After disturbance, this species maintains itself through sprouting to form even-aged stands. On suboptimal sites it is successional to other forest types. Associated species include other oaks, ponderosa pine, Douglas-ﬁr and Paciﬁc madrone (Appendix 3).

Canyon live oak occurs from the Willamette Valley to the Baja Peninsula and east into Arizona at elevations from near sea level in the north to 9000 feet in the south (Eyre, 1980). It comprises about one-quarter of California’s oak volume and is second only to California black oak in this regard (Table 1.4). Canyon live oak forms pure stands on very steep slopes and dry canyon bottoms. Elsewhere it occurs in mixture with Douglas-ﬁr, ponderosa pine and other conifers. The species is shade tolerant when young and often maintains itself in relatively stable communities (Eyre, 1980). The blue oak–digger pine forest type surrounds California’s Central valley at elevations between 500 and 5000 ft, although blue oak occasionally extends to the valley ﬂoor (Fig. 1.22). This forest type occurs between the valley grasslands and the montane forests above, where it can

48

Chapter 1

Fig. 1.22. Blue oak woodland in the Sierra Nevada Range (Province M261: Dry Steppe Province). (USDA Forest Service, North Central Research Station photograph.)

endure a meagre 10 inches of annual precipitation (Eyre, 1980). Forest cover ranges from 30 to 80% with canopy heights between 15 and 50 ft. Associated species include California live oak, interior live oak, valley oak and California black oak (Barbour, 1988). At low elevations blue oak and valley oak mixtures develop savanna communities. Valley oak savannas extend into the Central Valley where they make their best development on alluvial soils (Grifﬁn, 1977). The California coast live oak forest type (sometimes referred to as southern oak woodland) occurs on the west side of the Coast Range in the southern two-thirds of California. It extends inland on north-facing slopes of narrow valleys and other cool sites. This type occurs at elevations of up to 3000 ft in the northern part of its range and to 5000 ft in the southern portion. Although it can form pure, closed canopy stands, it is

considered a woodland type and commonly occurs in savannas comprised of scattered oaks or in mixture with conifers (Appendix 3). California coast live oak is long-lived, moderately shade tolerant, and forms relatively permanent woodlands. When trees reach about 8 inches dbh they are also highly resistant to ﬁre (Eyre, 1980). The ecological importance of California’s oak woodlands and timberlands is receiving increased attention (Pillsbury et al., 1997). Although their value for commercial products is low, their importance to wildlife, water quality, aesthetics, soil protection, recreation and fuelwood is widely acknowledged (Helms and Tappeiner, 1996). A principal silvicultural problem related to the oak woodlands of the Paciﬁc Mediterranean–Marine Region is ensuring that the regeneration of oaks is sufﬁcient for replacing trees periodically lost to natural mortality and timber harvesting.

References Abrams, M.D. (1988) Effects of prescribed ﬁre on woody vegetation in a gallery forest understory in northeastern Kansas. Transactions of the Kansas Academy of Science 91, 63–70. Abrams, M.D. (1990) Adaptations and responses to drought in Quercus species of North America. Tree Physiology 7, 227–238. Abrams, M.D. (1992) Fire and the development of oak forests. BioScience 42, 346–353.

Oak-dominated Ecosystems

49

Abrams, M.D. and Gibson, D.J. (1991) Effects of ﬁre exclusion on tallgrass prairie and gallery forest communities in eastern Kansas. USDA Forest Service General Technical Report SE SE-69, pp. 3–10. Abrams, M.D. and Nowacki, G.J. (1992) Historical variation in ﬁre, oak recruitment, and post-logging accelerated succession in central Pennsylvania. Bulletin of the Torrey Botanical Club 119, 19–28. Aizen, M.A. and Patterson, W.A. (1990) Acorn size and geographical range in the North American oaks (Quercus L.). Journal of Biogeography 17, 327–332. Allaby, M. (ed.) (1994) The Concise Oxford Dictionary of Ecology. Oxford University Press, New York. Allen, B.H. (1990) Classiﬁcation of oak woodlands. Fremontia 18(3), 22–25. Arend, J.L. and Scholz, H.F. (1969) Oak forests of the Lake States and their management. USDA Forest Service Research Paper NC NC-31. Axelrod, D.I. (1983) Biogeography of oaks in the Arcto-Tertiary Province. Annals of the Missouri Botanical Gardens 70, 629–657. Bailey, R.G. (1995) Descriptions of the ecoregions of the United States. USDA Forest Service Miscellaneous Publication 1391. Bailey, R.G. (1997) Ecoregions of North America, 1:15,000,000 scale map (rev.). USDA Forest Service, Washington, DC. Bailey, R.G. (1998) Ecoregions. Springer-Verlag, New York. Barbour, M.G. (1988) Californian upland forests and woodlands. In: Barbour, M.G. and Billings, W.D. (eds) North American Terrestrial Vegetation. Cambridge University Press, New York, pp. 131–164. Barnes, B.V., Pregitzer, K.S., Spies, T.A. and Spooner, V.H. (1982) Ecological forest site classiﬁcation. Journal of Forestry 80, 493–498. Birch, T.W. (1996) Private forest-land owners of the United States, 1994. USDA Forest Service Resource Bulletin NE NE-134. Biswell, H.H. (1989) Prescribed Burning in California Wildlands Vegetation Management. University of California Press, Berkeley. Bliss, J.C., Nepal, S.K., Brooks, R.T., Jr and Larsen, M.D. (1994) Forestry community or granfalloon. Journal of Forestry 92(9), 6–10. Bliss, J.C., Nepal, S.K., Brooks, R.T., Jr and Larsen, M.D. (1997) In the mainstream: environmental attitudes of Mid-South forest owners. Southern Journal of Applied Forestry 21, 37–43. Bolsinger, C.L. (1980) California forests: trends, problems, and opportunities. USDA Forest Service Resource Bulletin PNW PNW-89. Bolsinger, C.L. (1988) The hardwoods of California’s timberlands, woodlands, and savannas. USDA Forest Service Resource Bulletin PSW PSW-148. Bourdeau, P. (1954) Oak seedling ecology determining segregation of species in Piedmont oak–hickory forests. Ecological Monographs 24, 297–320. Braun, E.L. (1972). Deciduous Forests of Eastern North America. Hafner, New York. Burns, R.M. and Barber, J.C. (1989) Silviculture of southern pines. USDA Forest Service General Technical Report WO WO-55, pp. 31–38. Burns, R.M. and Honkala, B.H. (1990) Silvics of North America. USDA Forest Service Agriculture Handbook 654, Vol. 2. Carvell, K.L. and Tryon, E.H. (1961) The effect of environmental factors on the abundance of oak regeneration beneath mature oak stands. Forest Science 7, 98–105. Chase, A. (1995) In a Dark Wood: the Fight Over Forests and the Rising Tyranny of Ecology. Houghton Miﬂin, Boston, Massachusetts. Cleland, D.T., Hart, J.B., Host, G.E., Pregitzer K.S. and Ramm, C.W. (1993) Ecological Classiﬁcation and Inventory System of the Huron-Manistee National Forests. USDA Forest Service, Washington, DC. Cooper, A.W. (1989). Ecology of the pine–hardwood type. USDA Forest Service General Technical Report SE SE-58, pp. 3–8. Cunningham, R.J. and Hauser, C. (1989) The decline of the Missouri Ozark forest between 1880 and 1920. USDA Forest Service General Technical Report SE SE-58, pp. 34–37. Curtis, J.T. (1959) The Vegetation of Wisconsin. University of Wisconsin Press, Madison. DeVivo, M.S. (1991) Indian use of ﬁre and land clearance in the southern Appalachians. USDA Forest Service General Technical Report SE SE-69, pp. 306–310.

50

Chapter 1

Dyksterhuis, E.J. (1948) The vegetation of the western Cross Timbers. Ecological Monographs 18, 325–376. English, B.C., Bell, C.D., Wells, G.R. and Roberts, R.K. (1997) Stewardship incentives in forestry: participation factors in Tennessee. Journal of Forestry 21(1), 5–10. Eyre, F.H. (ed.) (1980) Forest Cover Types of the United States and Canada. Society of American Foresters, Washington, DC. Fernow, B.E. (1911) A Brief History of Forestry. University Press, Toronto. Flora of North America Editorial Committee. (1997) Flora of North America North of Mexico, Vol. 3. Oxford University Press, New York. Godman, R.M. (1985) What are northern hardwoods. In: Hutchinson, J.G. (ed.) Northern Hardwood Notes. USDA Forest Service, North Central Research Station, St Paul, MN, pp. 1.1–1.3. Grifﬁn, J.R. (1977) Oak woodland. In: Barbour, M.G. and Major, J. (eds) Terrestrial vegetation of California. John Wiley & Sons, New York, pp. 383–415. Helms, J.A. (ed.) (1998) The Dictionary of Forestry. Society of American Foresters, Bethesda, Maryland. Helms, J.A. and Tappeiner. (1996) Silviculture in the Sierra. In: Status of the Sierra, Vol. II, Assessments and Scientiﬁc Basis for Management Options. Sierra Nevada Ecosystem Project ﬁnal report to Congress. University of California, Davis, pp. 439–476. Hicks, R.R., Jr (1997) A resource at the crossroads: a history of the central hardwoods. USDA Forest Service General Technical Report NC NC-188, pp. 1–22. Hodges, J.D. (1995) The southern bottomland hardwood region and the brown loam bluffs subregion. In: Barrett, J.W. (ed.) Regional Silviculture of the United States. John Wiley & Sons, New York, pp. 227–270. Hodges, J.D. and Switzer, G.L. (1979) Some aspects of the ecology of southern bottomland hardwoods. In: Gateway to Opportunity: Proceedings of the Society of American Foresters Convention. Society of American Foresters, Washington, DC, pp. 360–365. Jensen, R.J. (1997) Quercus Linnaeus sect. LOBATAE Loudon, red or black oaks. In: Flora of North America North of Mexico (Flora of North America Editorial Committee). Oxford University Press, New York, pp. 447–468. Johnson, P.S. (1990) Quercus macrocarpa Michx. Bur oak. USDA Forest Service Agriculture Handbook 654, Vol. 2, pp. 686–692. Johnson, P.S. (1993a) Perspectives on the ecology and silviculture of oak-dominated forests in the central and eastern states. USDA Forest Service General Technical Report NC NC-153. Johnson, P.S. (1993b) Sources of oak reproduction. USDA Forest Service General Technical Report SE SE-84, pp. 112–131. Johnson, P.S. (1994a) The silviculture of northern red oak. USDA Forest Service General Technical Report NC NC-173, pp. 33–68. Johnson, P.S. (1994b) La sylviculture du chene rouge aux USA. In: Timbal, A., Kremer, A., LeGoff, N. and Nepveu, G. (eds) Le chene rouge d’Amerique. Institut National de la Recherche Agronomique, Paris, pp. 272–283. Jokela, J.J. and Sawtelle, R.A. (1985) Origin of oak stands on the Springﬁeld Plain: a lesson on oak regeneration. Proceedings of the Central Hardwood Forest Conference V. University of Illinois, Urbana-Champaign, pp. 181–188. Kerpez, T.A. and Stauffer, D.F. (1989) Avian communities of pine–hardwood forests in the Southeast: characteristics, management, and modeling. USDA Forest Service General Technical Report SE SE-58, pp. 156–169. Kroeber, A.L. (1925) Handbook of the Indians of California. Bureau of American Ethnology of the Smithsonian Institution Bulletin 78. Kuhlman, E.G. (1978) The devastation of American chestnut by blight. In: MacDonald, W.L., Cech, F.C., Luchok, J. and Smith, H.C. (eds) Proceedings of the American Chestnut Symposium. West Virginia Books, Morgantown, pp. 1–3. Lawson, E.R. (1990) Juniperus virginiana L. Eastern redcedar. USDA Forest Service Agriculture Handbook 654, Vol. 1, pp. 131–140. Leopold, B.D., Weaver, G.H., Cutler, J.D. and Warren, R.C. (1989) Pine–hardwood forests in northcentral Mississippi: an ecological perspective. USDA Forest Service General Technical Report SE SE-58, pp. 211–222.

Oak-dominated Ecosystems

51

Liebhold, A.M., MacDonald, W.L., Bergdahl, D. and Mastro, V.C. (1995a) Invasion by exotic forest pests: a threat to forest ecosystems. Forest Science Monograph 30. Little, E.L., Jr (1971) Atlas of United States Trees, Volume 1, Conifers and Important Hardwoods. USDA Forest Service Miscellaneous Publication 1146. Little, E.L., Jr (1977) Atlas of United States Trees, Volume 4, Minor Eastern Hardwoods. USDA Forest Service Miscellaneous Publication 1342. Little, E.L., Jr. (1979) Checklist of United States trees. USDA Agriculture Handbook 541. Loftis, D.L. (1988) Regenerating red oak in the southern Appalachians: predictive models and practical applications. PhD dissertation, North Carolina State University, Raleigh. Long, J.N. (1995) The middle and southern Rocky Mountain region. In: Barrett, J.W. (ed.) Regional Silviculture of the United States. Wiley & Sons, New York, pp. 3335–3386. Lord, L. (1999) The Americas. US News and World Report (16–23 August), 84–87. Lorimer, C.G. (1985) The role of ﬁre in the perpetuation of oak forests. Proceedings of Challenges in Oak Management and Utilization. University of Wisconsin, Madison, pp. 8–25. Lorimer, C.G. (1989) The oak regeneration problem: new evidence on causes and possible solutions. University of Wisconsin Forestry Research Analyses 8. McMinn, H.E. (1964) An Illustrated Manual of California Shrubs. University of California Press, Berkeley. McNab, W.H. and Avers, P.E. (1994) Ecological subregions of the United States: section descriptions. USDA Forest Service Administrative Publication WO-WSA WO-WSA-5. Manos, P.S. (1997) Quercus Linnaeus sect. PROTOBALANUS (Trelease) A. Camus, Intermediate oaks. In: Flora of North America North of Mexico, Vol. 3. Oxford University Press, New York, pp. 468–471. Miller, H.A. and Lamb, S.H. (1985) Oaks of North America. Naturegraph Publishers, Happy Camp, California. Nigh, T.A., Grabner, J.K., Becker, C. and Kabrick, J.M. (2000) Ecological Classiﬁcation of the Current River Hills Subsection: Draft Manual. Missouri Department of Conservation, Jefferson City. Nixon, K.C. (1997) Quercus Linnaeus, oak. In: Flora of North America North of Mexico, Vol. 3. Oxford University Press, New York, pp. 445–447. Nixon, K.C. and Muller, C.H. (1997) Quercus Linnaeus sect. QUERCUS, White oaks. In: Flora of North America North of Mexico, Vol. 3. Oxford University Press, New York, pp. 471–506. Nowacki, G.J., Abrams, M.D. and Lorimer, C.G. (1990). Composition, structure, and historical development of northern red oak stands along an edaphic gradient in north–central Wisconsin. Forest Science 36, 276–292. Olson, S.D. (1996) The historical occurrence of ﬁre in the Central Hardwoods, with emphasis on southcentral Indiana. Natural Areas Journal 16, 248–256. Orwig, D.A. and Abrams, M.D. (1994) Land-use history (1720–1992), composition, and dynamics of oak–pine forests within the Piedmont and Coastal Plain of northern Virginia. Canadian Journal of Forest Research 24, 1216–1225. Pavlik, B.M., Muick, P.C., Johnson, S. and Popper, M. (1991) Oaks of California. Cachuma Press, Los Olivos, California. Paysen, T.E., Derby, J.A., Balck, H., Jr, Bleich, V.C., and Mincks, J.W. (1980) A vegetation classiﬁcation system applied to southern California. USDA Forest Services General Technical Report PSW PSW-45. Paysen, T.E., Derby, J.A. and Conrad, C.E. (1982). A vegetation classiﬁcation system for use in California. USDA Forest Service General Technical Report PSW PSW-63. Penfound, W.T. (1968) Inﬂuence of a wildﬁre in the Wichita Mountains Wildlife Refuge, Oklahoma. Ecology 49, 1003–1006. Petrides, G.A. (1988) A Field Guide to Eastern Trees. Houghton Mifﬂin, Boston, Massachusetts. Petrides, G.A. and Petrides, O. (1992) A Field Guide to Western Trees. Houghton Mifﬂin, Boston, Massachusetts. Phillips, D.R. and Abercrombie, J.A., Jr (1987) Pine–hardwood mixtures – a new concept in regeneration. Southern Journal of Applied Forestry 11, 192–197. Pillsbury, N.H., Verner, J. and Tietje, W.D. (1997) Proceedings of symposium on oak woodlands: ecol-

52

Chapter 1

ogy, management, and urban interface issues. USDA Forest Service General Technical Report PSW PSW-160. Pinchot, G. (1987) Breaking New Ground. Island Press, Washington, DC. Plumb, T.R. and McDonald, P.M. (1981) Oak management in California. USDA Forest Services General Technical Report PSW PSW-54. Powell, D.S., Faulkner, J.L., Darr, D.R., Shu, Z. and MacCleery, D.W. (1994) Forest resources of the United States, 1992. USDA Forest Service General Technical Report RM RM-234 (rev.). Putnam, J.A., Furnival, G.M. and McKnight, J.S. (1960) Management and inventory of southern hardwoods. USDA Forest Service Agriculture Handbook 181. Pyne, S.J. (1982) Fire in America. Princeton University Press, Princeton, New Jersey. Rice, E.L. and Penfound, W.T. (1959) The upland forests of Oklahoma. Ecology 40, 593–608. Roach, B.A. (1968) Is clear cutting good or bad? Keep Tennessee Green Journal 8, 4–5, 12–14. Roach, B.A. and Gingrich, S.F. (1968) Even-aged silviculture for upland central hardwoods. USDA Forest Service Agriculture Handbook 355. Rosson, J.F., Jr (1994) Quercus stellata growth and stand characteristics in the Quercus stellata–Quercus marilandica forest type in the Cross Timbers region of central Oklahoma. In: Proceedings North American Conference on Barrens and Savannas. US Environmental Protection Agency, Great Lakes National Program Ofﬁce, Chicago, Illinois, pp. 329–333. Sargent, C.S. (1884) Report on the Forests of North America (exclusive of Mexico). Government Printing Ofﬁce, Washington, DC. Seymour, R.S. (1995) The northeastern region. In: Barrett, J.W. (ed.) Regional Silviculture of the United States. Wiley & Sons, New York, pp. 31–80. Shefﬁeld, R.M., Birch, T.W., Leatherberry, E.C. and McWilliams, W.H. (1989) The pine–hardwood resource in the Eastern United States. USDA Forest Service General Technical Report SE SE-58, pp. 9–19. Shelly, J.R. (1997) An examination of the oak woodland as a potential resource for higher-valued wood products. USDA Forest Service General Technical Report PSW PSW-160, pp. 445–455. Skeen, J.N., Doerr, P.D. and Van Lear, D.H. (1993) Oak–hickory–pine forests. In: Martin, W.H., Boyce, S.G. and Echternacht, A.C. (eds) Biodiversity of the Southeastern United States. John Wiley & Sons, New York, pp. 1–34. Smith, D.M. (2000) American chestnut: ill-fated monarch of the eastern hardwood forest. Journal of Forestry 98(2), 12–15. Smith, H.C., Lamson, N.I. and Miller, G.W. (1989) An esthetic alternative to clearcutting? Journal of Forestry 87(3), 14–18. Stephens, S.L. (1997) Fire history of a mixed oak–pine forest in the foothills of the Sierra Nevada, El Dorado County, California. USDA Forest Service General Technical Report PSW PSW-160, pp. 191–198. Thilenius, J.F. (1968) The Quercus garryana forests of the Willamette Valley, Oregon. Ecology 49, 1124–1133. Thirgood, J.V. (1971) The historical signiﬁcance of oak. Proceedings of Oak Symposium (USDA Forest Service Northeastern Forestry Experimental Station), pp. 1–18. Thomas, J.W. (1997) California’s oak woodlands: where we have been, where we are, where we need to go. USDA Forest Service General Technical Report PSW PSW-160, pp. 3–9. Trimble, G.R., Jr (1973) The regeneration of Central Appalachian hardwoods with emphasis on the effects of site quality and harvesting practice. USDA Forest Service Research Paper NE NE282. Tucker, J.M. (1980) Taxonomy of California oaks. USDA Forest Service General Technical Report PSW PSW-44, pp. 19–29. USDA Forest Service (1993) Forest Type Groups of the United States (map). USDA Forest Service, Washington, DC. USDA Forest Service (2000) Resources Planning Act (RPA) Statistical Tables, 1997. USDA Forest Service, Washington, DC. Vankat, J.L. (1982) A gradient perspective on the vegetation of Sequoia National Park, California. Madroño 29, 200–214. Van Kley, J.E., Parker, G.R., Franzmeier, D.P. and Randolph, J.C. (undated). Field Guide: Ecological Classiﬁcation of the Hoosier National Forest and Surrounding Areas of Indiana. USDA Forest Service, Hoosier National Forest, Beford, Indiana.

Oak-dominated Ecosystems

53

Waldrop, T.A. (ed.) (1989) Proceedings Pine–Hardwood Mixtures: a Symposium on Management and Ecology of the Type. USDA Forest Service General Technical Report SE SE-58. Walker, L.C. (1995) The southern pine region. In: Barrett, J.W. (ed.) Regional Silviculture of the United States. John Wiley & Sons, New York, pp. 271–334. Ward, F.R. (1995) California’s forest products industry: 1992. USDA Forest Service Resource Bulletin PNW PNW-206. Whitney, G.G. (1994) From Coastal Wilderness to Fruited Plain. Cambridge University Press, Cambridge, UK. Williams, T. (1989) Incineration of Yellowstone. Audubon 1989(1), 38–89. Youngs, R.L. (2000) A right smart little jolt: loss of the chestnut and a way of life. Journal of Forestry 98(2), 17–21.

2 Regeneration Ecology I: Flowering, Fruiting and Reproduction Characteristics

Introduction Ecologically, the terms ‘regeneration’ and ‘reproduction’ are closely associated. In the narrow biological sense, regeneration refers to the regrowth of lost or destroyed parts of organs. However, regeneration also can be used in a population context to refer to ‘rebirth’, e.g. the rebirth of a forest after its destruction by natural or human causes. The latter meaning is useful in ecology and silviculture because it connotes population process, and has been so used in the literature of those ﬁelds, albeit inconsistently (cf. Harper, 1977; Keeley, 1981; Bartolome et al., 1987; Muick and Bartolome, 1987; Helms, 1998). We herein use regeneration to refer to the ecological processes involving the establishment, growth and population changes of juvenile trees, co-occurring plants, and their propagules rather than to a physiological process (e.g. see Grubb, 1977). In this context, juvenile trees are those not yet capable of ﬂowering and producing seed, which for most of the oaks takes 15–25 years.1 Similarly, the term ‘reproduction’ has a narrow biological deﬁnition referring to sexual or asexual mechanisms by which organisms generate others of their own 1Among

kind.2 Like ‘regeneration’, ‘reproduction’ also can more generally imply the process of reproducing something. We have chosen to use the term reproduction to refer to individual or populations of juvenile trees (or other plants) already reproduced. Unlike regeneration, reproduction in this context does not connote process. Although the two terms have been treated as synonymous in the silvicultural literature, with reproduction relegated to obsolescence by Helms (1998), we believe it is conceptually important to distinguish between objects (young trees) and process (forest renewal). The regeneration of oak forests accordingly can be deﬁned as a multifaceted ecological process. It includes the ﬂowering, fruiting and seed dispersal of mature trees, as well as the germination, seedling establishment, growth and population changes of oak reproduction and associated plants. Forest regeneration thus involves time frames related to stages of stand development. In even-aged silviculture (Chapter 7), the regeneration period for oaks (at the stand scale) may span the last two and the ﬁrst two decades of a rotation,3 during which reproduction becomes established and develops into the new stand. In con-

the oaks of eastern United States, exceptions include sprout-origin bear oaks and chestnut oaks, both of which can produce acorns as early as the end of the third growing season. Sawtooth oak, an Asian species, can produce acorns by age 3 on trees that are only 1 m tall (Nakashizuka et al., 1997). Certain seedling propagation methods combined with selecting progenies with early flowering can be used to obtain early acorn production in some species. 2The American Heritage Dictionary, 3rd edn. Houghton-Mifflin, Boston, Massachusetts. 3A rotation is the period between the establishment and final harvest of an even-aged stand. 54

Regeneration Ecology I: Reproduction

trast, uneven-aged silviculture (Chapter 8) and old-growth forests (Chapter 9) are characterized by a regeneration period that is essentially continuous or frequently periodic. Within any time frame, each step of the regeneration process is beset with uncertainties and unknowns that have contributed to our inability, in some cases, to successfully manage and sustain oak forests. Yet, oaks have thrived throughout North America for millennia seemingly without the help of humans. Much of this apparent enigma can be untangled by considering what we know about the regeneration process, which begins with the oak ﬂower.

Flowering Oaks of the United States fall into one of three species groups. Species in the white oak group require one growing season to complete their reproductive cycle, while most of the species in the red oak group require two growing seasons. The ﬁve species in the intermediate group, which are found only in western North American, also require two growing seasons (Tucker, 1980). Oaks are monoecious, i.e. they produce male and female ﬂowers on the same tree. All trees that have attained ﬂowering age and size usually bear both kinds of ﬂowers in abundance each year. In many species, ﬂowering does not occur until trees are 15–25 years old. The annual regularity of ﬂower production in oak contrasts with its irregular acorn production, which may range from bumper acorn crops in some years to poor or no crops in others. In a sexually mature oak, buds of several types occur along the outer branches comprising current-, 1-year-old and 2-year-old shoots. Some buds are strictly vegetative, i.e. they produce only leaves and other vegetative structures. Other buds are the progenitors of male and female ﬂowers. Some produce only male ﬂowers (catkins) whereas others produce both male ﬂowers and vegetative structures. Female ﬂowers arise from tissues located in leaf axils (Cecich and Larsen, 1997).

55

In white oak, strictly vegetative buds account for about half of all buds and the remaining half represent approximately equal numbers of male and female buds (Cecich and Larsen, 1997). The distribution of buds also varies with position along the branch. About 72% of all buds occur on current-year shoots, 26% on 1-year-old shoots and 2% on 2-year-old shoots. The latter are all vegetative buds occurring in leaf axils. Among the buds that produce ﬂowers, female ﬂower buds occur only on current year shoots whereas male ﬂowers are approximately equally represented on 1year-old and current shoots. However, in the red oak group, female ﬂowers in different states of development may be present on both current year and 1-year-old shoots at the same time. In both the white oak and red oak groups, female ﬂowers occur throughout the crown but are most abundant in the upper crown. In contrast, male ﬂowers occur mainly in the top and middle portions of the crown (Cecich and Larsen, 1997). Of the species native to the United States, ﬂowering is most completely described for white oak (Turkel et al., 1955; Sharp and Chisman, 1961; Stairs, 1964; Mogensen, 1965, 1975; Sharp and Sprague, 1967; Merkle et al., 1980; Feret et al., 1982), although our knowledge of ﬂowering in other oaks is growing (Cecich, 1997). Accordingly, the following account of the ontogeny of oak ﬂowering focuses on white oak. Discussion of other oaks is limited to illustrating differences among species and species groups.

Male flowers The male (staminate) ﬂowers of oaks develop in the axils of scale leaves of the current vegetative buds, or separate male buds bear the staminate ﬂowers (Fig. 2.1A). Species in the white oak group initiate catkin primordia in the buds of shoots produced the year before acorn maturation. Species in the red oak group initiate catkin primordia in the buds of shoots formed two years before acorn maturation. In white oak, catkin initiation occurred in late May

56

Chapter 2

(A)

Bud scales

Catkin

Staminate flowers

Bract

(B)

Inflorescence Perianth Filament

Pollen sacs

(C)

Pollen sac

Filament

Pollen grains

Fig. 2.1. The male (staminate) oak flower. (A) Catkins bearing numerous flowers originate from the axils of bud scales. (B) A single male flower with pollen sacs. (C) Pollen sacs split open to release pollen grains when environmental conditions are favourable. (From Cecich, 1994.)

in Virginia (Merkle et al., 1980), and about a month later in Pennsylvania (Sharp and Chisman, 1961). In Virginia trees, the catkin primordia were macroscopically distinguishable by early August. Individual staminate ﬂowers formed within the primordial catkins in late June or early July and were structurally complete before the onset of dormancy in October. These struc-

tures resumed development in mid-March and the catkins emerged from the bud in early April; further north in Pennsylvania they emerged in late April. In eastern forests, swelling and opening of male ﬂower buds usually is later in the white oak group than in the red oak group. When catkins in the white oak group are emerging or slightly drooping (semi-

Regeneration Ecology I: Reproduction

pendent), pollen dispersal is already complete. Emergence of white oak catkins occurs when daily temperature minima are 50°F (10°C) or higher for 10 days (Sharp and Chisman, 1961). They emerge from the base of the inner scales of buds clustered at the ends of the woody twigs of the previous year’s growth. Numbers of catkins range from three to ten or more per twig tip and are erect shortly after emerging. Within a few days they become semipendent and are usually fully pendent and in full bloom, or anthesis, within 5–12 days (Fig. 2.2). The pollen matures while catkins are elongating. The catkins appear ﬁrst in the topmost branches and emergence progresses downward. Meiosis, the process that reduces chromosome numbers by one-half, begins in the male ﬂowers of white oak when catkins have elongated slightly beyond the bud scales. In the same geographic area, scarlet oak and bear oak (both in the red oak group) begin similar activity about 2 weeks earlier at the time of early bud swell (Stairs, 1964). The diploid (2n) chromosome number, i.e. the number before reduction division in cells takes place, is 24 for all oak species investigated. In white oak, the individual male ﬂowers (Fig. 2.1B) within the catkins reach anthesis 11–16 days after emergence from the bud (Sharp and Chisman, 1961). The basal portions of the catkins mature ﬁrst. The time from catkin emergence to the completion of pollen shedding (Fig. 2.1C) ranges from about 11 to 19 days. The topmost branches shed pollen ﬁrst in forestgrown trees, but the reverse may occur in open-grown trees. Pollen shedding occurs when catkins are 3–6 inches long. Individual trees shed most of their pollen within a 48-hour period if weather is favourable. In New York, pollen shedding in white oak lags that of scarlet oak and bear oak by 2 weeks (Stairs, 1964). About the time white oak catkins are semipendent, new unfolded leaves appear. However, there is a lag in further leaf expansion if catkins contain pollen. At the time of pollen shedding, usually during the last 2 weeks of May in central Pennsylvania, leaf length averages about 2 inches in white oak

57

Fig. 2.2. Northern red oak catkins in full bloom in mid-May (northern Wisconsin). Leaves are not yet fully expanded. (Authors’ photo.)

and 3 inches in chestnut oak (Sharp and Chisman, 1961). The day after pollen dispersal, leaf area increases by about 50%. It is thus possible to identify trees that have shed pollen by the state of leaf development. The period of arrested leaf expansion may favour pollen dispersal by minimizing canopy interference. Dispersal of oak pollen is by wind and usually occurs before there is signiﬁcant insect activity. This contrasts with many associated hardwoods, which ﬂower later and are insect pollinated. Pollen dispersal in white oak occurs only on days when relative humidity is less than 45% for several hours (Sharp and Chisman, 1961). To complete pollen shedding, both white oak and chestnut oak require 2 to 3 such days with air temperatures of 63 to 69°F (17 to

58

Chapter 2

21°C). In contrast, dwarf chinkapin oak (a shrubby member of the white oak group) only needs a few hours of such weather to disperse pollen. A growth chamber study conﬁrmed the adverse effects of high humidity on bear oak (a shrubby member of the red oak group), which produced no immature (ﬁrst-year) acorns when humidity during the ﬂowering period was 61–70% (Wolgast, 1972). When humidity was reduced to 38–50%, a signiﬁcantly larger proportion of the ﬂowers (10.5% of 467) produced immature acorns. A positive association between high relative humidity and acorn production was reported for white oak in Missouri when relative humidity was averaged over the 1week pollination period (Cecich and Sullivan, 1999). This apparent discrepancy with other studies points out the importance of the observed time interval used to express the effect of relative humidity (or other weather factors). If only a few hours of low humidity occurring over a few days during the pollination period are sufﬁcient for effective pollen shedding (Sharp and Chisman, 1961), measurements of relative humidity should then represent this time-dependent sensitivity if the intent is to directly relate relative humidity to pollination and acorn production. Otherwise relative humidity measurements and their apparent effects become potentially confounded with other factors. Such confounding arises from correlated but indirect relations between humidity and other weather variables such as air temperature. A light wind of 5–8 miles per hour is further conducive to pollen shedding. However, in closed-canopy forests, the canopy may act as a windbreak that minimizes the inﬂuence of light winds on the ripening catkins. The canopy nevertheless offers little protection against strong winds during dry rainless periods. In open ﬁelds, oaks shed pollen on the windward side several hours before shedding occurs on the leeward size when weather conditions are otherwise favourable. High winds on humid, cool days do not induce pollen shedding and prolonged cool, wet weather

can cause over-ripened catkins to drop intact with anthers ﬁlled with pollen. Although male ﬂowers can tolerate light frosts, temperatures below 25°F (4°C) are lethal (Sharp and Chisman, 1961). In northern Utah, a 5 May freeze averaging 27°F (3°C) across 37 Gambel oak sites killed semipendent male ﬂowers (Neilson and Wullstein, 1980). Subsequent acorn production was less than in nearby areas unaffected by the freeze. In bear oak in Massachusetts, the number of male ﬂowers per branch at or near the time of anthesis decreased from the top to the bottom of a slope representing a 66 ft (20 m) vertical gradient (Aizen and Kenigsten, 1990). Within a given topographic position, the number of male ﬂowers also decreased with decreasing tree height (i.e. taller trees produced more male ﬂowers). A higher incidence of low temperatures and frost occurred at lower points along the combined vertical topographic and tree-height gradient. Thus frosts, prolonged desiccating winds, and prolonged cool humid weather can be detrimental to male ﬂowering and pollen dispersal. Any of these events may result in reduced or failed acorn production.

Female flowers In white oak in Virginia, the inﬂorescences bearing female (pistillate) ﬂowers are ﬁrst identiﬁable, but only microscopically, in early August of the year before acorn maturation (Merkle et al., 1980). These primordial inﬂorescences usually arise from the axils of the top three or four leaves within the developing bud. There they remain invisible to the unaided eye and continue to develop until late September or early October. Resumption of development occurs the following spring when dramatic changes occur during the last week of March and the ﬁrst week of April. Each inﬂorescence then elongates rapidly, usually producing two to three functional ﬂowers, but sometimes up to ﬁve. At this stage, ﬂowers remain within the swollen buds.

Regeneration Ecology I: Reproduction

Like male ﬂowers, the time of appearance and development of female ﬂowers varies with the weather and thus year. In central Pennsylvania, the leaves of white oak usually emerge from the buds during the second week of April (Sharp and Sprague, 1967). As the buds expand they carry with them the female inﬂorescence. A single elongating stalk then pushes the one to ﬁve pistillate structures upward. The individual female ﬂowers become visible within 5–10 days after emergence of the male catkins (Merkle et al., 1980). At this time, the pendent catkins are about 50% of their ﬁnal length, new vegetative shoots are about 2–3 inches long, and new leaves are 1–2 inches long (Sharp and Sprague, 1967). At this stage of development, the ﬂowers resemble miniature acorns that envelope all the requisite, preformed ﬂoral structures including the cupule (which develops into the ‘cup’ of the mature acorn). The stalks of pistillate white oak ﬂowers continue to elongate until the ﬂowers are mature. Maturation occurs when the three stalk-like styles bearing the pollen-receptive organs, the stigmas, extend beyond the surrounding ﬂoral structures (the perianth) (Fig. 2.3a). Pollen grains released into the air by the male ﬂowers land on the stigmas, germinate, and produce pollen tubes (Fig. 2.3b). A series of events involving callose plugs then occur, which isolate the contents of each pollen tube from other pollen grains (Fig. 2.3c–d). Pollen tubes then stop growing in 2–3 weeks (Fig. 2.3e). There is then a pause in pollen tube growth, which is resumed in 2 weeks in the white oak group and after about 13 months in the red oak group. Pollen tubes then enter the locules (Fig. 2.3f), and the egg of one ovule is fertilized; the other ﬁve eggs and ovules soon die. Approximately 1 month after pollination, or about 1 July in Pennsylvania, meiosis and fertilization have taken place (Turkel et al., 1955; Sharp and Sprague, 1967). Fertilization of Gambel oak (a white oak) in Arizona also occurs then (Brown and Mogensen, 1972). In central Missouri, fertilization of white oak ﬂowers occurs during mid-June (Cecich, 1997). Fertilization of black and northern red oak

59

flowers, which require two growing seasons for ovule maturation, occurs during mid-June of the second year for black oak and 2 weeks later for northern red oak (Fig. 2.3g). Although not all eggs of the six ovules in each ovary may become fertilized, fertilization occurs at nearly the same time on those that do (Mogensen, 1965). Of 30 ovaries studied in three species (Gambel, white and black oak), nearly half the 180 ovules observed aborted because they were not fertilized (Mogensen, 1975). Thus, on average, 2.8 ovules per ovary were fertilized. The remaining aborted ovules were approximately equally divided between those with zygote or embryo failure (28%) and those without an embryo sac (26%). However, all but one fertilized embryo normally aborts very soon after fertilization to produce a one-seeded fruit. Although the reason for the abortion of fertilized ovules is not well understood, the functional ovule may be fertilized before the others. This, in turn, may suppress the growth of the other ovules in a process analogous to apical dominance in stems (Mogensen, 1975). Occasionally, more than one ovule develops to seed maturity within the same ovary to produce a multiple-seeded acorn. Some trees are especially prone to producing these (Coker, 1904; Buchholz, 1941). In four of ﬁve species observed in Illinois, from 1 to 2% of acorns were multiseeded, which suggests that multiseeds may be common in most oaks (Hosner, 1959). By mid-summer of the year of acorn maturation, acorn enlargement and embryo development begin in most species. The generalized sequence of acorn maturation events in central Missouri is illustrated in Fig. 2.3h–l. However, rates of acorn maturation may vary among species, climates, weather conditions and other factors. In white oak and chestnut oak in Pennsylvania, the acorn and cupule begin rapid growth in late July. By the ﬁrst week in August the acorn begins to emerge from the cup. By mid-August, the acorns are about one-third ﬁlled out. By the last week in August they are full size (Fig. 2.4). By mid-September mature acorns are usually

60

Chapter 2

a – the flower

b – pollen grains

c – 1st callose plug

d – 3rd plug

e – pollen tubes cease growing

f – locules are entered

g – fertilization

h – embryo enlarging

i – 1 July

j – 15 July

k – 29 July

l – mature acorn

Regeneration Ecology I: Reproduction

dropping from the cups. In years of heavy acorn production, all acorns on a stalk may mature, but often only one develops and the remainder die (Sharp and Sprague, 1967).

Factors Affecting Acorn Production Most oak species produce good acorn crops one year in three or four (Olson, 1974). Although consistent annual production of male and female ﬂowers is an inherent characteristic of oaks, the large annual variation in acorn production is at least partially controlled by environmental factors (Sharp and Sprague, 1967; Sork and Bramble, 1993). Weather-related factors directly inﬂuence the early ﬂowering process, as discussed in the previous section. Those and other factors also may impact the later stages of acorn development. The total number of female ﬂowers produced and their per cent survival together explain about 90% of the variation in acorn production in black, northern red and white oaks in Missouri (Sork and Bramble, 1993). In contrast, factors related to site productivity such as soil nutrients, topographic position and site index (Chapter 4) appear to have little or no inﬂuence on acorn production (Tryon and

61

Carvell, 1962; Wolgast, 1972). Yet few studies have rigorously examined site productivity as a factor in acorn production in most species. Some trees appear to be under complete genetic control. Some never produce acorns even when they grow in favourable environments and occupy superior crown positions (Wood, 1934; Downs and McQuilkin, 1944; Sharp, 1958; Sharp and Sprague, 1967). Thus environment, genetics and the interactive effects of those factors are all potential determinants of acorn production.

Weather Good white oak acorn crops in Pennsylvania occurred when temperatures in late April were above normal but followed by a sudden drop in temperature in early May (Sharp and Sprague, 1967). A 10-day period in late April with warm nights followed by a 2-week period with cool nights produced good acorn set. Poor acorn production occurred when there was an even progression in mean daily April and May temperatures. In a Missouri study, warm spring temperatures during the year of acorn maturation were positively correlated with the size of acorn

Fig. 2.3. (Opposite) The development of a pollinated pistillate flower to a mature acorn. (a) The flower contains six ovules (two shown) in each of three locules (chambers) of the ovary. Three styles with stigmatic tips contain the transmitting tissue through which the pollen tubes reach the locules. (b) Pollen grains, released into the air by male (staminate) flowers, land on the stigmas, germinate, and produce pollen tubes. (c) A callose plug isolates the contents of each pollen tube from other pollen grains, which soon fall from the stigma. (d) Successively formed plugs isolate the growing tips of pollen tubes from earlier-formed remnants, which maintains turgor pressure in the tip. The third plug appears in the pollen tube after about 1 week. (e) Pollen tubes stop growing in 2–3 weeks. After a 2-week pause, the pollen tubes of species in the white oak subgenus resume growth towards the locules. In the red oaks, the pause lasts about 13 months. Thus, the red oaks require two growing seasons for pollinated flowers to mature into acorns whereas white oaks require only one growing season. (f) Pollen tubes enter the locules at an open space called the paracarpous region. This space allows pollen tubes from any style to randomly enter any of the six ovules. (g) Fertilization of the egg in one ovule, which occurs several days after the pollen tube enters the locule, produces an embryo. The other five eggs and ovules soon die. (h) The embryo enlarges and other parts of the flower differentiate into the cup and shell of the acorn. (i) By early July, the flower is recognizable as a young acorn. The bright green shell begins to protrude from the cup in response to the growing embryo. (j) By mid-July, the embryo is heart-shaped. The two lobes of the ‘heart’ represent the early stages of cotyledon development. (k) By late July, the embryonic axis (root tip, shoot tip and hypocotyl) appear between the two cotyledons. (l) Mature acorns begin to drop from their branches in September. The acorns are filled with starch, lipids and proteins. (Drawings and text contributed by Dr Robert A. Cecich, USDA Forest Service, North Central Research Station, Columbia, Missouri.)

62

Chapter 2

Fig. 2.4. A cluster of nearly mature northern red oak acorns (left) in late August of their second year of development. Acorns in their first year of development are visible in the leaf axils of the current-year’s shoot (right); flower styles are still intact and visible. (Authors’ photo.)

crops in white, black and northern red oaks (Sork and Bramble, 1993). In contrast, summer drought and late spring frost the previous year were negatively associated with acorn production. In another Missouri study, high daily maximum temperatures were negatively associated with white oak acorn production and ﬂower survival and black oak ﬂower survival (Cecich and Sullivan, 1999). Growth chamber studies of immature bear oak acorns revealed that the acorns did not develop under conditions of high humidity (Wolgast, 1972). This effect may have been directly related to a failure in pollen shedding (Sharp and Chisman, 1961). Other investigators have reported little or no evidence that relative humidity, precipitation, drought, site quality and wind velocity affect acorn yield (Downs and McQuilkin, 1944; Sharp and Sprague, 1967; Feret et al., 1982). These disparate ﬁndings may be related to inherent differences among species. But even for a given species, discrepancies in ﬁndings may arise from differences in how the duration, intensity and frequency of potential predictors of acorn production are measured and quantitatively expressed. Foresters and physiologists have long speculated about frost as a cause of failure in

acorn production. However, the frequency of killing frosts in central Pennsylvania indicates that during any given 50-year period, frost accounts for only 4 years of complete acorn crop failures and 4 years of severely reduced crops (Sharp and Sprague, 1967). In white oaks in Missouri, a late spring freeze reduced the number of white oak ﬂowers to about 20% of non-frost years and emergence of surviving ﬂowers was about 2 weeks later than normal (Cecich and Sullivan, 1999). In immature bear oak acorns in New Jersey, spring temperatures of 29 to 32°F (2°C to 0°C) for 3 consecutive days in late May reduced acorn yields (Wolgast and Trout, 1979). Acorn yield increased as the proportion of the crown not damaged by frost increased (Fig. 2.5).

Premature abscission Most female ﬂowers of oaks succumb to premature separation (abscission) from their stalks (peduncles) before they can develop into acorns. Although premature abscission occurs primarily during the pollination–fertilization period, it continues through the acorn maturation period (Turkel et al., 1955; Williamson, 1966) (Fig. 2.6). About 52% of acorns of several species caught in acorn traps from mid-

Regeneration Ecology I: Reproduction

63

30

300

25

250 Pounds 200

20

150

15

100

10

50

5

0 0.0

0.1

0.2

0.3

0.4

0.5

0.6

Pounds of acorns per acre

Number of acorns per acre

Number

0 0.7

Proportion of crown area undamaged Fig. 2.5. Bear oak acorn production (fresh weight) in New Jersey in relation to crown frost damage. (From Wolgast and Trout, 1979, used with permission.)

summer on in Missouri were immature (Christisen and Kearby, 1984). These acorns fell continuously from early July through August. Observed losses from premature abscission in white oak range from 68 to 99% of female ﬂowers observed early in the growing season (Williamson, 1966; Feret et al., 1982). Yet the lower abscission rate represented a ‘good’ acorn crop. Substantial premature abscission thus can occur even in years of high acorn production. Premature abscission is common not only to oaks but to many other hardwoods including willow, poplar, basswood, black locust and elm (Kramer and Kozlowski, 1979; Kozlowski and Pallardy, 1997). Fruitset can be diminished by premature ﬂower abscission, limited pollination or fertilization, inadequate nutrition, embryo abortion, premature abscission of young fruits and other factors (Greulach, 1973). Premature abscission of fruits also may involve competition for photosynthates, which can occur among fruits or between fruits and vegetative growth (Abbott, 1960; Luckwill, 1977).

In the oaks, tree hoppers and other insects have been implicated in premature abscission (Cecich et al., 1991; Cecich, 1994). Hail storms also may damage pistillate oak ﬂowers (Cecich, 1997; Cecich and Sullivan, 1999). The long slender peduncles (stems) of pistillate white oak ﬂowers may be especially susceptible to hail damage. However, the long styles of black oak ﬂowers may be more susceptible to hail damage than the sessile styles of white oak ﬂowers. If the styles of oak ﬂowers are broken off before pollen tubes cease growing, ﬂowers will not be fertilized and will abort. Hail also can dislodge fragile catkins and thus eliminate the pollen source if hail occurs before pollen is shed. Failed or limited pollination in any species can result from inadequate production of pollen caused by unfavourable weather conditions or a limited period of stigma receptivity to pollen. Also, fertilization failure may result from pollen sterility, failure of ovary development, slow pollen tube growth and failure of meiosis. If pollen tube growth is slow, the tube may not reach the embryo sac or sperm viability may be lost before the tube reaches the sac.

64

Chapter 2

fertilized ﬂowers, only 26–29% (5–11% of all ﬂowers) developed into acorns during the 4 study years. However, among individual clones, that percentage ranged from 0 to 42%. Statistical analysis showed that clones explained a signiﬁcant proportion of variation in ﬂowering, fertilization and fruiting. But acorn yield was most strongly correlated with clonal differences in percentage of ﬂowers fertilized and only weakly correlated with ﬂower abundance. Though there was signiﬁcant variation among clones in ﬂower abundance, the greater limitation to acorn production was low fertilization rate.

140 A

Number of live flowers

120 100 80 60 40 20 0 May

July

Sep

May

July

Sep

Month 90 B

Variation in acorn production

80 Number of live flowers

70 60 50 40 30 20 10 0 Apr

May

June

July

Aug

Sep

Date

Fig. 2.6. Survival of female black oak and white oak flowers in central Missouri trees. (A) Black oak flowers. The line connected by circles is the average of five flower-bearing branches in the upper crowns of eight trees in 1991/92; unconnected circles show the range. The line connected by triangles is the average of 11 trees in 1992/93; the unconnected triangles show the range. (B) White oak flowers in nine trees in 1992; the mean and range are shown as in (A). (Data courtesy of Robert A. Cecich, USDA Forest Service, North Central Forest Experiment Station.)

A white oak clonal seed orchard in Tennessee demonstrated the strong genetic control over premature abscission (Farmer, 1981). Among 111 trees representing 31 clones, the percentage of fertilized ﬂowers averaged over all clones ranged from 19 to 38% 7–11 years after orchard establishment. But among individual clones, that percentage ranged from 0 to 75%. Among

The most prominent and consistent attributes of acorn production are large year-toyear and tree-to-tree variation (e.g. see Koenig et al., 1994; Nakashizuka et al., 1997). This variation occurs within species and locales. In white oak and northern red oak in West Virginia, tree-to-tree variation was signiﬁcantly greater than year-to-year variation in trees (Tryon and Carvell, 1962). In bear oak, tree-to-tree variation in acorn production was under strong genetic control (Wolgast, 1972). Forest-grown trees of three acorn-producing ranks (low, medium and high) transplanted to a homogeneous nursery environment conformed to their original rankings. In contrast, soil factors related to the forest environment did not explain a signiﬁcant proportion of the observed variation in acorn production before transplanting. White oak similarly demonstrated strong genetic control over ﬂowering and acorn production in a clonal seed orchard (Farmer, 1981). The large variation in time of acorn fall among individual willow oaks and water oaks during the same year suggests that factor also may be largely under genetic control (Cypert and Webster, 1948). Although there are a few reported studies on the allocation of net annual biomass production to acorns, on average the proportion apparently is relatively small but highly variable among trees. During a

Regeneration Ecology I: Reproduction

5-year study of 29 open-grown sawtooth oaks in Japan, acorns accounted for an estimated 2.8% of net annual biomass production when averaged across years (Nakashizuka et al., 1997). However, among trees this proportion varied from 0 to 20%. The proportion of production allocated to acorns was independent of tree size. The absolute allocation to acorn production increased in proportion to whole-tree production but only after whole-tree production reached 0.76 kg (1.7 lb) per year. Above that threshold, acorn production was closely proportionate to leaf production. Other factors being equal, large trees generally produce more acorns than small trees (Downs and McQuilkin, 1944; Goodrum et al., 1971), and open-grown trees produce more acorns than trees growing in a closed-canopy forest (Sharp, 1958; Sharp and Sprague, 1967). Because both tree size and stand density can be measured and controlled silviculturally, these relations have practical application in managing oak forests for acorn production. There also are large differences among species in acorn-producing potential. Methods for assessing and predicting acorn crops based on these and other relations together with silvicultural guidelines for sustaining acorn production are presented in Chapter 9.

Periodicity Year-to-year variation in acorn production within an individual tree or population of trees can result from the single or joint effects of variation in inherited factors that control the timing and frequency of ﬂowering and fruiting, and environmental factors. Environmental factors that can reduce acorn production below a tree’s inherent potential often occur as essentially random events, which may obscure the expression of any inherently regular pattern, or periodicity, in acorn production. Obscuring factors include weather events unfavourable to ﬂowering and acorn production (Sharp and Sprague, 1967; Sork and Bramble, 1993; Koenig et al., 1996; Cecich, 1997), and insect damage to leaves (Crawley and Akhteruzzaman, 1988), ﬂowers (Cecich et

65

al., 1991) and acorns (Myers, 1978; Christisen and Kearby, 1984). Because of confounding environmental factors, determining the existence of periodicity in natural environments usually requires the use of statistical methods that can reveal inherent cycles, or alternatively, experimental studies in controlled environments designed to eliminate or separate the confounding factors. Periodicity in acorn production thus refers to an inherent potential of a tree to recurrently produce acorn crops at a ﬁxed time interval. Periodicity may characterize an individual tree or a population of trees of the same species. Even though an individual tree may be inherently able to produce acorns at a regular interval, the population as a whole may not if it includes trees with no discernible pattern of acorn production, trees that produce at regular intervals but of different lengths, or trees that produce at the same interval length but which are out of phase with other trees in the population. Periodicity in the occurrence of large or above average acorn crops, or ‘masting’, is also a possibility. The occurrence of periodicity and various patterns of periodicity is usually determined by observing trees known to be consistent, albeit not necessarily proliﬁc, acorn producers (e.g. see Koenig et al., 1994). The occurrence of periodicity and periodicity characteristics in acorn production vary among oak species. In some species in the white oak group, acorn production is noticeably periodic. In the absence of adverse weather, acorn crops tend to occur every other year and good crops about once every 4 years (Downs and McQuilkin, 1944; Goodrum et al., 1971; Beck, 1977; Myers, 1978; Christisen and Kearby, 1984; Koenig et al., 1994). This pattern suggests that white oaks require at least 1 year to recover after a large investment in acorn production. Periodicity in the white oak group also is apparently synchronized, with all fruitful trees within a species, locale and year producing at their inherent individual, but often environmentally constrained, capacity (Sharp, 1958; Myers, 1978).

66

Chapter 2

For four California oak species, selected weather factors (temperature and precipitation) explained 38–78% of the variation in mean annual acorn production (Koenig et al., 1996). In valley and blue oaks, which mature in one growing season, mean annual April temperature (which includes the fertilization period) was positively associated with acorn production. In canyon live and coast live oaks, which mature in two growing seasons, total precipitation from September through the next acorn fall was positively associated with acorn production. The inﬂuence of weather factors thus may obscure any tendency in periodicity as conﬁrmed by the predominance of reported irregularities in acorn production (Sharp, 1958; Grisez, 1975; Godman and Mattson, 1976; Sork et al., 1993). Periodicity tends to be less consistently expressed in the red oaks than in the white oaks (Sharp, 1958; Tryon and Carvell, 1962; Beck, 1977). For example, in northeastern Wisconsin northern red oak bumper crops (deﬁned as 91% or more of maximum potential) occurred at 7-year intervals (Godman and Mattson, 1976). Yet, such crops occurred as few as 2 years apart over the 21-year study period. Good or better crops (60% or more of potential) occurred at 3-year intervals. Other species in the red oak group behave similarly (Sharp, 1958; Christisen and Kearby, 1984). In the red oak group, the combined effects of apparently asynchronous acorn production and unpredictable weather often obscure any tendency towards periodicity. This suggests that the apparent asynchrony of acorn production among individual red oaks may result in lower year-to-year variation in acorn production at the population level than in the white oaks (Tryon and Carvell, 1962). Acorn production among species is generally asynchronous (Sork and Bramble, 1993; Koenig et al., 1994). However, among ﬁve species of California oaks observed for 12 years, acorn production was relatively synchronous within species despite large variation among trees in numbers of acorns produced (Koenig et al., 1994). The same study found no evidence of periodicity in

the occurrence of large crops (‘masting cycles’) at the population level but found that such cycles did occur in individual trees. Individual trees of some species also produced large crops in successive years, but within a population such occurrences occurred no more often than expected by chance. There was no evidence of masting cycles at the population level in any of the ﬁve species studied, which included two species in the white oak group, two in the red oak group (including one evergreen oak), and one evergreen oak in the intermediate group. The investigators noted that regular masting cycles have yet to be demonstrated for any oak species. One method of detecting inherent periodicity in acorn production in individual trees is based on the correlation between the acorn production of a tree in one year with its production in other years. The correlation coefﬁcients for various time intervals or ‘lag times’ (e.g. 1, 2, 3 and 4 years before and after a speciﬁed time) thus can be used to statistically identify time intervals associated with acorn production (positive correlations), time intervals associated with the absence of production (negative correlation), or the absence of correlation (0 correlation). For example, the individual-tree correlation coefﬁcients for a 1-year time lag during a 10-year study period can be derived from the nine pairs representing year 1 vs. year 2, year 2 vs. year 3, . . . year 9 vs. year 10. Similarly, the correlations for 2-, 3- and 4-year or longer time lags can be calculated. The method can be used to derive, for each tree, a series of correlation coefﬁcients each representing a given time lag. If this were done for several trees, the correlation coefﬁcients for each tree then could be evaluated. A general consistency among individuals in the signs of the correlation coefﬁcients (positive or negative) and their statistical signiﬁcance (i.e. the likelihood that they differ from 0) provides a basis for accepting or rejecting the hypothesis of inherent periodicity in acorn production. For a given species, the average correlation coefﬁcient of several trees for a given time lag can be used as the basis for testing the hypothesis. Accordingly, if

Regeneration Ecology I: Reproduction

prior acorn production does not affect current acorn production, the mean correlation coefﬁcient should not differ signiﬁcantly from zero for any given time lag. A positive correlation would indicate a signiﬁcant positive relation between a given time lag and acorn production, and a signiﬁcant negative correlation would indicate a negative relation, i.e. inhibition of acorn production. This method of assessing periodicity also has the advantage of removing differences related to possible asynchrony in acorn production among individual trees among years because correlations are derived for each tree by lag interval rather than year by itself. Such statistical tests provided empirical evidence of an inherent 3-year cycle in white oak acorn production, i.e. years of high production tended to be followed by two successive years of little or no production (Sork and Bramble, 1993). There also is evidence of inherent 2- and 4-year cycles for black oak and northern red oak, respectively (Fig. 2.7). Signiﬁcant negative correlations tended to occur in intervening years. Because all three species were observed in the same locale and yet had different cycles, variation in weather does not provide a likely explanation for variation in apparent periodicity. Evidence for periodicity was strongest for white oak because none of the numerous, measured weather variables were correlated with its apparent 3-year periodicity cycle. Among California oaks that have been similarly studied, a 3-year cycle was statistically evident in blue oak (a member of the white oak group) and canyon live oak (a member of the red oak group with a 1-year acorn maturation period) (Koenig et al., 1991) (Fig. 2.8A). Evidence for periodicity was less conclusive in valley oak (a member of the white oak group), canyon oak (intermediate group) and California black oak (red oak group) (Fig. 2.8B). Plausible explanations for inconclusive results include: (i) the ‘real’ absence of periodicity; (ii) real but weakly expressed periodicity; (iii) the overriding inﬂuence of one or more environmental factors during a given study

67

period; (iv) errors in measuring acorn crop size; (v) a sample size that is too small (i.e. too few trees observed) to substantiate that a ‘real’ association as expressed by a correlation coefﬁcient of a given value is statistically signiﬁcant; and (vi) possible genetic variation in periodicity within a species. Statistical relations (Figs 2.7 and 2.8) suggest that periodicity may be of two types: manifest and latent, i.e. readily expressed and seldom expressed. Accordingly, evidence of manifest periodicity would express itself as a consistent temporal pattern of positive and negative correlations such as that observed in blue oak (Fig. 2.8A). In contrast, latent periodicity would occur in species or populations of trees with an inherent periodicity that is seldom or never expressed because of acute sensitivity to adverse environmental events or frequent exposure to such events. Demonstrating the existence of latent periodicity would require experiments that eliminate or control the factors obscuring periodicity. In the absence of such experiments we can only speculate about the nature of possible genetic variation in periodicity. The occurrence of populations that include trees with both high and low lagtime consistencies (e.g. as expressed by the correlation patterns in Figs 2.7 and 2.8) nevertheless would be consistent with a population that possesses high genetic variation in periodicity. An alternative hypothesis is that, among trees that produce acorns, some individuals are inherently periodic whereas others are inherently non-periodic.

Spatial variation within tree crowns Acorns are usually unevenly distributed throughout the tree crown. However, acorns in open-grown trees were more evenly distributed than acorns in trees in closed stands. In closed stands, most acorns occurred on branches exposed to light (Sharp and Sprague, 1967). Post (1998) found that even in open-grown northern red oaks, production was greatest in the lower section of the south-facing side of crowns. Shading on one or more

Chapter 2

Number of years prior to current year

68

4

3

2

1

–0.8

–0.6

–0.4

–0.2

0.0

0.2

0.4

0.6

0.8

Correlation of current with prior year’s acorn production

Black oak Northern red oak White oak Fig. 2.7. Correlation of current acorn production (mature acorns per tree) with prior years’ production for three species in Missouri. The significant positive correlations (right side of graph) suggest 2-, 3- and 4year acorn production cycles (time lags) for black, white and northern red oaks, respectively, based on 13, 15 and 12 trees (n), respectively, observed for 8 years. Each correlation coefficient shown is the average of the individual-tree coefficients within a species/time-lag interval (see text for discussion). Asterisks (*, **, ***) indicate significance at P < 0.05, < 0.01, and < 0.001, respectively, based on t tests with n 1 degrees of freedom. (Redrawn from Sork et al., 1993, used with permission.)

sides of the crown resulted in a non-discernible pattern of production, although branches exposed to light tended to produce more acorns than those that were not. In a relatively open-grown stand (6 oaks acre1) of coast live oak on a south-facing slope in California, acorn numbers were about two times greater on south-facing sides of crowns than northeast- and northwest-facing sides (Lewis, 1992). Although numbers of acorns in lower and midcrowns were about three times greater than in upper crowns, the opposite trend occurred in trees in another stand that were pruned and irrigated (Lewis, 1989). Such inconsistencies suggest that environmental factors may inﬂuence the withincrown distribution of acorns.

Acorn destroying insects and other organisms may prefer different parts of the crown and thereby inﬂuence the distribution of sound acorns. The infestation of coast live oak acorns by ﬁlbert weevils and ﬁlbert worms was 1.5–3 times greater on northeasterly side of crowns than on south and northwesterly sides (Lewis, 1992). Preferences for cooler crown aspects may reﬂect insect avoidance of overheating. Moreover, the occurrence of acorns that were split open (possibly from bursting from within) occurred on south and northwesterly sides of crowns, making them vulnerable to invasion by ants, wasps, microbial pathogens and other parasites (Lewis, 1992). The numerous confounded biotic and physical factors that apparently can inﬂuence the spa-

Number of years prior to current year

Regeneration Ecology I: Reproduction

6

A

5 4 3 Valley oak Blue oak Coast live oak

2 1 –0.4

Number of years prior to current year

69

6

–0.2

0.0

0.2

0.4

B

5 4 3 2 Canyon live oak California black oak

1 –0.4

–0.3

–0.2

–0.1

0.0

0.1

0.2

0.3

Correlation of current with prior year’s acorn production Fig. 2.8. Correlation of current acorn production with prior years’ production for five species in California. (A) Correlations for three species requiring 1 year for acorns to mature based on 87, 57 and 63 valley, coast live and blue oaks, respectively. For blue and coast live oaks, the patterns of significant correlations suggest a 3-year cycle. For valley oak, the pattern of correlations suggest that it is not strongly cyclic. (B) Correlations for two species requiring 2 years for acorns to mature based on 21 trees of both species. For both species in graph B, the correlations are inconclusive in relation to cyclic patterns because of the absence of significant positive correlations. For all species, correlations are based on the number of acorns counted in a 30-second sampling period on each tree in each of 10 consecutive years. Asterisks (*, **, ***) indicate significance at P < 0.05, < 0.01 and < 0.001 respectively, based on binomial tests. (Redrawn from Koenig et al., 1991.)

tial distribution of acorns within crowns (especially sound acorns) thus complicates generalizing about such distributions other than that they vary greatly among trees.

Acorn Predation and Dispersal Many organisms consume acorns including insects, millipedes, fungi, birds and mammals. Consumption by one or more biotic agents is

often so complete that in any given year few acorns remain to germinate and become seedlings. Although this consumption contributes to oak regeneration failures (Marquis et al., 1976; Galford et al., 1991c), acorns are valuable food for many birds and mammals because of their high caloric content, nutritiousness and availability during seasons when other food is often scarce. Oak forests thus can be managed speciﬁcally for acorn production to beneﬁt wildlife (Chapter 9).

70

Chapter 2

The term dispersal is used to refer to the transport of acorns from their place of origin (i.e. tree crowns or the ground directly beneath them) to some other location irrespective of the distance, mode of transport or subsequent fate of the acorns. Because of their relatively large size and mass, acorns are not dispersed by wind. Although gravity and water may affect localized dispersion, birds and mammals are the most important dispersers of acorns. Dispersal by animals may carry acorns to places that are more favourable or less favourable for germination and seedling establishment than their site of origin. But even under the most favourable circumstances, the oaks pay a price for animal-mediated dispersal because at least some of the dispersed acorns are consumed by the dispersers. Whether an oak beneﬁts from dispersal by a given disperser depends on: • Where the acorns are dispersed • The number dispersed • The pattern of dispersal (scattered or concentrated) • The proportion of acorns consumed by the disperser (including acorns consumed at their place of origin before dispersal has occurred). With respect to the oak, dispersal is ineffective if acorns are largely dispersed to habitats unfavourable for maintaining acorn viability and seedling establishment. If the proportion of total numbers of acorns dispersed is very small, the impact of the dispersing agent (whether positive or negative) on oak regeneration may be negligible. Numbers of dispersed acorns being equal, dispersers that scatter individual acorns or small groups of acorns are likely to beneﬁt oak regeneration more than dispersers that cache large quantities of acorns in one or a few locations. Dispersers that consume virtually all the acorns they disperse clearly represent some loss to potential seedling establishment. Dispersers also may consume signiﬁcant quantities of acorns at their place of origin. If proportionately few acorns are consumed by a disperser, dispersal may be advantageous to the oak if acorns are dispersed in signiﬁcant numbers to habitats favourable for seedling establishment.

Insects: destroyers of acorns Insects may begin destroying acorn crops during ﬂowering. However, the focus of this section is on insects as destroyers of already formed acorns (from immature to mature) through the germination period. During these stages of development, there is a predictable sequence of damaging events and agents involving not only insects but other organisms including bacteria, fungi, algae, protozoa, nematodes, mites and other organisms (Winston, 1956). Insects often initiate damage before acorns mature. For example, the adult female Curculio weevil can chew through the immature acorn’s shell to deposit her eggs inside the acorn. The developing larvae then consume part or all of the acorn’s interior. Invading insects also can carry pathogenic fungi and bacteria into the acorn, which may kill the embryo even though the insect itself may not. Direct insect damage occurs during the several weeks required for the larvae to develop. The developing larvae usually consume most acorns by mid- to late-autumn while they remain on the tree or shortly after they fall. The larvae eventually cut exit holes in the shell of the acorn, making it accessible to other organisms. One of these is the acorn moth, which lays its eggs on or in the exit hole. Their larvae feed on the remainder of the embryo and on the faeces of the previous occupant. After acorns fall, fungi and other decomposers take over the ﬁnal stages of destruction until the acorn is thoroughly decayed and eventually incorporated into the soil (Winston, 1956). Only the acorns that escape this fate are available to wildlife and forest regeneration. Insects typically destroy 50% or more of acorns annually (Gibson, 1982; Kearby et al., 1986). In some years, destruction approaches 100%. Although many kinds of insects destroy acorns (Winston, 1956), most of the damage through the germination period is caused by a relatively small number of insect groups (Gibson, 1972, 1982; Kearby et al., 1986). These include:

Regeneration Ecology I: Reproduction

• Acorn weevils (Curculio spp. and Conotrachelus spp.) • Moths (the ﬁlbertworm moth and the acorn moth) • Acorn gall wasps (cynipids) • Nitidulid sap beetles. Acorn insects are sometimes studied using acorn traps and ground emergence traps from which the insects and acorns can be identiﬁed and counted (Fig. 2.9). The rela-

71

tive importance of the various acorn insects expressed as a percentage of insectinfested acorns is indicated by state and regional studies (Table 2.1). Curculio weevils in any given year typically account for 50–80% of infested acorns, and Conotrachelus weevils for about 5–40%. The moths and gall wasps each typically account for about 0–20% of infestations. However, these percentages can vary

Fig. 2.9. Acorn collection traps (on stakes) and insect emergence traps (on ground) used to study acorn production and acorn insects in a Missouri forest. (Photo courtesy of Dr William H. Kearby, Regional Entomologist (Retired), Wisconsin Department of Natural Resources.) Table 2.1. The relative importance of the four major acorn-insect groups expressed as a percentage of insect-infested acorns. Species, region and years observed White oak

Insect group

Curculio weevils Conotrachelus weevils Filbertworm and acorn mothe Acorn gall wasps a

Ohio 1961a 79.1 15.1 5.7 0.1

Range-wide 1962b 1963c 59.5 38.0 2.5 <0.1

57.5 37.8 4.7 0

Average of seven species in Missouri 1973–1976d 54.4 5.3 21.0 19.3

From acorns collected in three counties; may include immature acorns (from Gibson, 1972). From acorns collected in ten states; may include immature acorns (from Gibson, 1972). c From acorns collected in 13 states and the District of Columbia; may include immature acorns (from Gibson, 1972). d From mature acorns collected in four or five areas (depending on year); species include black, northern red, scarlet, Shumard, blackjack, post and white oaks (from Kearby et al., 1986, Table 9). e Both are moths (Lepidoptera) but the filbertworm is a primary invader and the acorn moth is a secondary invader of acorns. b

72

Chapter 2

greatly among oak species, stages of acorn development, locations and years (Gibson, 1972, 1982; Kearby et al., 1986). The major acorn insects can be divided into primary and secondary invaders. Primary invaders directly enter the acorn whereas secondary invaders require a crack or other opening in the acorn for entry (Murfeldt, 1894; Winston, 1956; Kearby et al., 1986). The most important primary insect invaders are the Curculio acorn weevils, the ﬁlbertworm moth and the cynipid gall wasps. Some species of fungi also are primary invaders. Although there are numerous secondary insect invaders of acorns, the most important are the Conotrachelus acorn weevils and the acorn moth.

Primary invaders The Curculio weevils comprise the most important group of acorn destroyers. Among the 27 American Curculio species, 23 breed and feed on acorns; 22 species do so exclusively on oaks (Gibson, 1969). They occur throughout the oak range in the United States. Two to three weeks before acorns mature, the adult female uses the tiny mouthparts at the end of her long, slender snout to chew a small hole in the acorn shell (Fig. 2.10A). She then makes a chamber near the inner surface of the shell, turns around to deposit an egg, and pushes it through the hole into the chamber (Brown, 1980; Kearby et al., 1986). Although only one egg is usually laid in each chamber, the female may lay several eggs in a single acorn. The eggs hatch in 5–14 days and the larvae then feed on the acorn’s contents. After maturing, the larvae exit the acorn through a circular hole they cut in the shell (Fig. 2.10B). Most larvae exit the acorn in the autumn although occasionally they remain there until spring. After leaving the acorn, the larvae burrow as deep as 12 inches into the soil where they form pupal cells. Most larvae overwinter in the soil for 1 or 2 years, whereas others remain there longer (Gibson, 1969). The transformation of pupae into adults occurs in the spring or

early summer. After the adults emerge from the soil and mate, the females complete the life cycle by depositing their eggs in the new acorns (Myers, 1978). Curculio weevils are vulnerable to parasitism by other insects, fungi, bacteria and nematodes (Gibson, 1969; Kearby et al., 1986). In a Missouri study, as many as 17% of Curculio weevils were parasitized by other insects (Kearby et al., 1986). Predators of adult Curculio weevils include more than 80 bird species (Brooks and Cotton, 1929). The short-tailed shrew is an important predator of weevil larvae in the soil (Brooks, 1910). In a 4-year Missouri study, Curculio weevils infested 31% of more than 10,000 mature acorns of seven species collected from seed traps (Kearby et al., 1986). These weevils occurred in 43% of white oak acorns and 25% of black oak acorns. By individual years, infestations ranged from 10 to 54% for white oak and 19 to 40% for black oak. In coast live oak in California, 27% of acorns collected in a single acorn season were infested with the ﬁlbert weevil, a species of Curculio (Lewis, 1992). The ﬁlbertworm moth is an important primary-invading moth (and is not closely related to the ﬁlbert weevil) (Fig. 2.10C). The larvae feed in acorns and a wide variety of other nuts and fruits (Murtfeldt, 1894; USDA Forest Service, 1985). The adult moth usually oviposits her eggs through the bottom of the acorn cup while the acorn is still forming (Gibson, 1972). Unlike the larvae of the Curculio weevils, which are born inside the acorn, the larvae of the ﬁlbertworm moth must bore into the acorn. A bacterial infection of acorns called drippy nut disease is often associated with this boring activity (Hildebrand and Schroth, 1967). The moth larvae commonly exit through the acorn shell and cup after consuming the interior of the acorn and then may enter a nearby acorn (Kearby et al., 1986). The larvae usually require two or three acorns for the completion of their development (Winston, 1956). Pupation normally takes place in the leaves and other debris on the forest ﬂoor.

Regeneration Ecology I: Reproduction

A

C

73

B

D

E

F

Fig. 2.10. Acorn-destroying insects. (A) An adult Curculio weevil on a black oak acorn. The adult female uses the mouthparts at the end of her long, slender snout to chew a small hole in the acorn shell. She then makes a chamber near the inner surface of the shell, turns around and deposits an egg, and pushes it through the hole into the chamber. (B) A Curculio larva emerging from an exit hole in a black oak acorn. (C) The adult filbertworm moth (about five times actual size). (D) Larval chambers and larvae of a stone gall wasp. The gall and its occupants completely fill the inside of this black oak acorn. (E) A pip gall protruding from beneath the cup of a black oak acorn. The inset photo shows the exit hole of the adult gall wasp, which is visible only after removing the gall and acorn cup. (F) The adult acorn moth (about five times actual size). This insect is a secondary invader of acorns, i.e. it requires a pre-established crack or other opening in the acorn shell as an entryway. The other insects shown are primary invaders. (Photos courtesy of Dr William H. Kearby, Regional Entomologist (Retired), Wisconsin Department of Natural Resources.)

Reported moth damage to acorns often combines that of the ﬁlbertworm moth with the acorn moth. Combined moth attacks in Missouri ranged from 4 to 64% of all insect attacks on mature acorns, depending on year and location, and often comprised 20–40% of attacks (Kearby et al., 1986). Of those, most were attributed to

the ﬁlbertworm moth. For immature acorns, moths accounted for 85% or more of all insect attacks in some locations and years, and most of those were attributed to the ﬁlbertworm moth. Although moths generally infest less than 10% of acorns (Korstian, 1927; Gibson, 1971, 1972), in some instances

74

Chapter 2

infestation rates have exceeded 50% (Sidney, 1948; Gibson, 1962). Moths infested 2.5% of acorns in a range-wide study of white oak (Gibson, 1972). In a similar study of northern red oak, up to 34% of acorns were infested with moth larvae (Gibson, 1982). The ﬁlbertworm moth infested 27% of coast live oak acorns in a California study (Lewis, 1992). In a 4-year Missouri study, moths infested 11% of mature acorns averaged across all species. Moth infestation rates on mature acorns were higher in species in the red oak group (14–22%) than in species in the white oak group (5–8%). Infestation rates also were higher in years of low acorn production than in years of high production (Kearby et al., 1986). The acorn gall wasps are members of the Cynipidae family of insects. Over 700 species are found in the United States and Canada, and about three-quarters of them induce gall formation on oaks (USDA Forest Service, 1985). The adult wasps deposit their eggs in the tissues of all parts of the oaks, from the roots to the ﬂowers. Gall formation apparently results from the reaction of the cambium and other living tissues to larvae-induced stimuli. The gall wasps have alternating sexual and asexual generations. Among the species that affect acorns, adults of the asexual generation oviposit their eggs in oak buds and ﬂowers in the spring, which causes galls to develop in or on the acorn. The sexual generation oviposits in or on immature acorns in June. The life cycle usually requires 2 years to complete (Weld, 1922). Gall wasps that attack acorns fall into three groups based on the location of the gall on the acorn: • Galls that form on or in the acorn cup but not in the acorn itself • ‘Pip’ galls, which form between the acorn and the cup • ‘Stone galls’, which form inside the acorn (Myers, 1978; Kearby et al., 1986). Some gall wasps produce growths resembling small acorns on the side of the acorn cup (Gibson, 1972). These galls usually do not affect acorn viability but may cause premature acorn abscission if they are numerous.

The galls occurring inside the acorn are usually caused by species of Callirhytis wasps. Some species cause stone galls, which are embedded in, but separate from, the interior material of the acorn (i.e. the cotyledon). They may not affect acorn viability unless they form in the area containing the rudimentary root and foliar structures, which lie near the pointed end of the acorn. Some stone galls are comprised of groups of connected larval chambers occupied by several soft white larvae (Fig. 2.10D). ‘Pip’ galls, also caused by Callirhytis wasps, cause the acorn’s cotyledons (the large ﬂeshy kernel inside the acorn shell) to shrivel, which destroys their viability and value to wildlife (Kearby et al., 1986) (Fig. 2.10E). In a 4-year Missouri study of four to ﬁve oak stands, gall wasps accounted for 19% of all insect attacks on mature acorns (all oak species combined) and ranged from 5 to 56% of attacks among years and locations (Kearby et al., 1986). Infestation rates averaged 11% of acorns, but ranged from 3 to 32% among locations and years. On average, 6% of white oak, 10% of black oak and 45% of scarlet oak acorns were infested by gall wasps. Infestation rates of gall wasps on immature acorns averaged 10% across all oak species observed and ranged from 2 to 30% among locations and years. In a range-wide study of northern red oak, gall wasp infestations ranged from 0 to 37% of acorns (Gibson, 1982). The relation between insect damage and acorn crop size is unclear. Although some reports indicate an inverse relation, other reports indicate the opposite, while yet others report no relation (Sidney, 1948; cf. Beal, 1952; Brezner, 1960; Janes and Nichols, 1967; Gibson, 1972; Beck, 1977; McQuilkin and Musbach, 1977). Such inconsistencies may result from interactions between acorn crop size, insect population size and other factors (Myers, 1978; Kearby et al., 1986). Potentially important interactive factors include competition among insects, the searching ability of insects to locate hosts, predation and parasitism among insects, preference of an

Regeneration Ecology I: Reproduction

insect for one host species over another and weather conditions adverse to insects. The Curculio weevils illustrate one type of competitive effect. An adult Curculio usually does not oviposit in an acorn already attacked by another Curculio. Consequently, as the number of weevils and infestations increase, the longer it may take an individual weevil to ﬁnd a suitable host (Kearby et al., 1986).

Secondary invaders In the United States, there are three species of Conotrachelus weevils that breed in acorns (Gibson, 1964, 1982; Kearby et al., 1986). However, a single species accounted for 93% of all Conotrachelus infestations in a range-wide study of white oak acorns (Gibson, 1972). Conotrachelus life histories are similar to the acorn-infesting Curculio weevils. The adults lay their eggs in acorns that have been cracked or opened by biotic agents or other factors. The larvae feed on the interior of the acorn, and after maturing cut an exit hole in the shell. Whereas each Curculio larva usually chews its own exit hole, all Conotrachelus larvae within an acorn usually exit through a common hole (Gibson, 1971, 1972). After the larvae exit the acorn, they drop to the ground and overwinter in the soil where pupation occurs. After pupation in spring and early summer is complete, the adults emerge from June to August, depending on region (Brooks, 1910; Gibson, 1964). The adults live longer than Curculio adults (Gibson, 1964). Parasites of the Conotrachelus weevils include tachinid ﬂies and other insects (Pierce, 1908; Gibson, 1964). Conotrachelus infestation rates are usually higher in the acorns of white oaks than red oaks (Gibson, 1964; Kearby et al., 1986), which is consistent with their role as secondary invaders. The shell-splitting of white oak acorns during autumn germination allows the weevils to easily gain entry into the acorns’ interiors. Acorn infestation rates are therefore likely to be highest after acorns fall to the ground and germinate. Consequently, Conotrachelus populations are likely to be underestimated

75

based on acorns collected in the autumn from acorn traps (Kearby et al., 1986). In years of poor white oak acorn production, Conotrachelus weevils may become more prominent in red oak acorns. In the autumn, adult weevils can enter red oak acorns that have been damaged but not completely consumed by wildlife and other insects, or that have otherwise cracked open. Some weevils also can overwinter in the adult stage (Gibson, 1964) and therefore oviposit in germinating or damaged red oak acorns in the spring. Among acorns collected throughout the range of white oak in two different years, Conotrachelus weevils occurred in 38% of all insect-infested acorns in both years, whereas Curculio weevils occurred in 58 to 60% of such acorns in each year (Gibson, 1972). In a Missouri study, Conotrachelus weevils accounted for 11% of all insectinfested white oak acorns while Curculio weevils accounted for 66% (Kearby et al., 1986). In a range-wide study of autumncollected northern red oak acorns, Conotrachelus were present in less than 3% of the acorns (Gibson, 1982). Over a 4year period in Missouri, Conotrachelus weevils on average occurred in 7% of white oak acorns, but occurred in less than 1% of black oak acorns based on autumn acorn collections (Kearby et al., 1986). However, data from autumn-collected acorns may grossly underestimate the impact of Conotrachelus weevils not only on acorns but on newly germinated seedlings. A central Pennsylvania study revealed that nearly all of the northern red oak acorns observed in the spring were infested with Conotrachelus larvae (Galford et al., 1991c). Moreover, the overwintering adult weevils of some species feed on the emerging roots and shoots of germinating acorns in the spring, which may kill most of the newly developing seedlings that are so attacked (Galford et al., 1988, 1991c; Galford and WeissCottrill, 1991). The acorn moth, known primarily as a secondary invader of acorns, is now also recognized as a primary pest (Galford, 1986) (Fig. 2.10F). It can be a primary pest

76

Chapter 2

of germinating acorns in both spring and autumn. However, in the autumn, it is primarily a secondary pest, invading acorns damaged by other insects or animals. The larvae also can enter previously undamaged acorns by chewing through the base (cup end, Fig. 2.3L) of the acorn, but apparently only in immature acorns (Galford et al., 1991c). Because damage is conﬁned to the cup ends (and thus furthest from the rudimentary root and leaf structures), this type of infestation has little affect on acorn viability. Acorn moth larvae are also more attracted to damaged than undamaged acorns, based on experiments comparing acorns cut in half with uncut acorns (Galford and Weiss-Cottrill, 1991). The larvae, which live in the forest litter, apparently are attracted to the cut acorns because of the volatile substances they emit. Damaged acorns thus may become primary targets for both feeding and reproducing insects. The acorn moth overwinters in the larval stage, and in the Central Hardwood Region may begin to feed in February, making it one of the ﬁrst acorn pests to become active. The acorn moth was ranked fourth in overall importance among insects that feed on germinating acorns and new seedlings in Ohio (after two species of Conotrachelus weevils and a nitidulid sap beetle) (Galford et al., 1988). During acorn germination, the moth larvae feed on the emerging and developing roots (radicles) and then enter the acorn through the splitting shell to complete their development (Galford and Weiss-Cottrill, 1991). In a Pennsylvania study that protected surfacesown acorns from large mammal predation by wire screens, insects destroyed 87% of the acorns and developing seedlings (Galford et al., 1991b). The acorn moth and Conotrachelus weevils were the primary causes of destruction. Thus, even when mammal predation is low and acorns are abundant, high rates of seedling establishment may not occur because of insect predation (Galford and Weiss-Cottrill, 1991). Two species of nitidulid sap beetles are known to feed on germinating acorns (Galford et al., 1991a,b). Adults of both

species consume elongating radicles during germination and then oviposit in the acorns. At least one of the two species also feeds on the tips of developing shoots (Galford, 1987). The combined adult and larval feeding destroys the acorn. They are thus both primary and secondary pests of acorns (Galford et al., 1991a). In Ohio, the insect is one of the most serious pests of germinating red oak acorns in the spring. When the overwintering adults become active during acorn germination, they feed heavily on the radicles. It is the most abundant of the acorn insects in southern Ohio in the spring (Galford et al., 1988). A millipede (Ptyoiulus impressus) also has been observed to feed on the radicles of germinating acorns; these accounted for 17% of the damage to developing acorns of northern red oak in a Pennsylvania study (Galford et al., 1992). Among white, chestnut and northern red oaks, the millipede showed a preference for northern red oak based on experimental feeding tests. In Pennsylvania, millipedes become active a few days before insect activity in late winter or early spring. Under natural conditions, oak seedling establishment is largely limited to acorns that have been buried by rodents (Korstian, 1927; McQuilkin, 1983). Based on ﬁeld experiments, acorns buried 1 inch deep in the soil have a greater chance of escaping insect predation than acorns on the soil surface. In a Pennsylvania study where northern red oak acorns were protected from mammal predation by screens, 40% of those sown on the soil surface and covered only with the current leaf litter were destroyed by insects, whereas only 25% of acorns placed 1 inch below the soil surface were destroyed (Auchmoody et al., 1994). Conotrachelus weevils and nitidulid sap beetles accounted for 97% of the destruction. In a similar Pennsylvania study, insects destroyed 92% of surface-sown acorns beneath mammal-excluding screens, but only 16% of those planted 1 inch deep in the soil (Galford et al., 1991c). The burial of acorns by natural mechanisms thus may account for the origin and subsequent regeneration of many, if not most, oak forests.

Regeneration Ecology I: Reproduction

Most of the acorn-damaging insects, both primary and secondary invaders, live in the forest ﬂoor for at least a part of their life cycle. Consequently, the use of ﬁre to control these insects has been proposed (Wright, 1987). Some of the Conotrachelus weevils live in the forest ﬂoor in both spring and autumn, which is also when prescribed burning is most feasible in oak forests of the eastern United States. Similarly, the ﬁlbertworm moth, acorn moth and nitidulid sap beetles are present in the forest ﬂoor in various stages of their life cycles during those periods. A preliminary study of ﬁre to control these insects in Ohio suggests that burning can reduce populations of several acorn insects, including Conotrachelus and nitidulid sap beetles (Wright, 1987). However, the investigators emphasized that the necessary ecological changes brought about by ﬁre are difﬁcult to determine based on short-term burning experiments. Fire is nevertheless recognized as a natural regulator of insects (Komarek, 1970; Ahlgren, 1974; Martin and Mitchell, 1981).

Rodents: predation and dispersal Among the most important predator/dispersers of acorns are the rodents, including squirrels, mice, voles, chipmunks and gophers. Even before acorns fall, about 10–25% of acorns are taken directly from the trees by birds and squirrels (Downs and McQuilkin, 1944; Dalke, 1953; Burns et al., 1954; Gysel, 1957, Beck, 1977; Myers, 1978). After acorns fall, predation from rodents together with other animals such as racoons, deer, wild turkey and blue jays often consume most of the acorn crop. Signiﬁcant numbers of acorns survive only in years of above average production. Acorns of species in the red oak group have a higher tannin content and therefore are less palatable than species in the white oak group. Tannins also may inhibit the action of digestive enzymes (Short, 1976). Thus, when both kinds of acorns are available, rodents will preferentially choose the white oaks (Smith, 1962; Short, 1976). Although acorns comprise a large propor-

77

tion of the diet of squirrels, other foods are necessary to satisfy their protein and phosphorus requirements (Short, 1976). Grey and fox squirrels carry acorns up to 650 ft or more from their source (Barnett, 1977), dispersing them to numerous scattered, single-nut caches in the litter or soil. This method of caching is called ‘scatterhoarding’ (Cahalane, 1942; Brown and Yeager, 1945; Stapanian and Smith, 1978). Although the dispersers and their competitors consume many cached acorns, some are not and survive to germinate and develop into seedlings. The proportion of acorns consumed thus represents the ‘cost’ of dispersal to the oak. Scatterhoarding decreases the concentration of acorns under the parent tree and thereby reduces the likelihood that other animals such as wild turkey and deer will eat most acorns. Acorn dispersal also reduces direct competition of oak seedlings with the parent tree and produces a more even distribution of seedlings. White-footed mice, deer mice and voles are also common dispersers of acorns. Their caches are often comprised of one or a few acorns placed in the litter, shallow tunnels in the soil, or sometimes in tree stumps and cavities. Small rodents such as white-footed mice and deer mice are generally more destructive of acorn crops than grey or fox squirrels because they often consume all of the acorns they cache (Shaw, 1968b; Marquis et al., 1976; DeLong and Yahner, 1991; Gribko and Hix, 1993). Some rodents are ‘larder hoarders’, sonamed because of their habit of concentrating acorns and other seeds in large quantities in one or a few places. The red squirrel concentrates acorns in a few large caches in tree cavities or under the litter (Layne, 1954). Flying squirrels and chipmunks primarily cache acorns in underground burrows, although both animals sometimes cache acorns under the litter or in shallow holes in the soil (Schwartz and Schwartz, 1959). Except for acorns cached in tree cavities, under rocks or other large objects, and deep burrows in the soil, the cache environment often provides favourable conditions for overwintering, germination and initial seedling establish-

78

Chapter 2

ment. Acorns cached in the forest ﬂoor or beneath it in the surface layer of the mineral soil are protected from desiccation, freezing, overheating from solar insolation and ﬁre, and to a more variable extent against predation by birds, insects and mammals. Whether burial is beneﬁcial may partly depend on climate. In Britain, frequently saturated soils associated with an oceanic climate and related fungal attacks on acorns greatly reduced the viability of buried acorns (Shaw, 1968). In the Piedmont of North Carolina, burial of white oak acorns increased survival from predation, but did not increase germination (Barnett, 1977). In contrast, burial was clearly advantageous to acorns in a blue oak savanna and a coast live oak–valley oak woodland in the seasonally dry foothills of California Coastal Range (Grifﬁn, 1971). A covering of leaf litter, by itself, has usually been shown to beneﬁt acorn survival (e.g. Korstian, 1927; Barrett, 1931). However in Iowa, acorns planted 1 inch deep in the autumn on plots with litter removed produced 2.4–3 times more seedlings (per 100 acorns sown) the following spring than acorns planted to the same depth with leaf litter replaced over the acorn (Krajicek, 1960). The advantage of litter removal occurred in plots where rodents were excluded by screens as well as those that were not screened. The contrasting results of these studies point out the difﬁculty in generalizing about the effects of natural environments on acorn viability and germination. The advantages of acorn burial are nevertheless documented by several studies. In Pennsylvania, rodent predation of northern red oak acorns planted 1 inch deep in the soil surface (and then covered with leaf litter) was 78% whereas losses were 100% for surface-sown acorns (Galford et al., 1991c). Surface-sown acorns were also more vulnerable to insect predation; insects destroyed 8% of buried acorns whereas 92% of surface-sown acorns were destroyed when both were protected from rodents. In another Pennsylvania study, predation of northern red oak acorns sown on the soil

surface was about three times that of acorns buried 1 inch in the soil (91 to 34%) (Auchmoody et al., 1994). Predation was mainly by chipmunks and mice. Predation of blue oak acorns by rodents in an oak woodland in California that were buried 1 inch in the soil was about half that of surface-sown acorns, based on the proportion of acorns that produced seedlings (Fig. 2.11). The primary predators were mice and gophers. The advantages of acorn burial also inﬂuenced 3-year seedling survival rate (Fig. 2.12). Predation on established seedlings protected from cattle and deer was largely attributable to gophers; by then, mice were not a signiﬁcant factor. In contrast, 3-year seedling survival in a more open blue oak savanna was unrelated to acorn burial (Fig. 2.12). Instead, the extremely high seedling mortality rates, which were related largely to grass competition in the savanna, overrode or masked rodent predation as a signiﬁcant factor in seedling survival (Davis et al., 1991). There is thus considerable evidence that acorn burial, under many prevailing conditions, favours acorn survival and oak seedling establishment. Although the advantage of burial is not obtained without cost, it has been proposed that scatterhoarding by rodents, especially squirrels, may be essential in sustaining oak forests (Korstian, 1927). The destruction of unburied acorns by insects can be 100%, even when mammals and birds are excluded as predators (Galford et al., 1991c). Autumn germination of acorns of species in the white oak group affords them a degree of protection against rodent predation. Within a few days after germination begins, 30–50% of the nutrient value of the acorn has been translocated to the developing radicle (Fox, 1982). Squirrels dislike radicles and avoid eating them (Smith and Follmer, 1972), although they may detach the acorn itself from the radicle and then eat the acorn. The radicle and attached plumule (the primordial shoot) thus remain intact so that the acorn may yet develop into a seedling (Fox, 1982). Autumn germination thus provides some acorns an ‘escape’ from predation. However, grey squirrels have developed a

Regeneration Ecology I: Reproduction

79

All acorns (both years) n=4500 P=0.38

1985 acorns n=2280 P=0.27

Surface acorns n=1140 P=0.17

Mice n=660 P=0.11

No mice n=480 P=0.26

1986 acorns n=2220 P=0.50

Buried acorns n=1140 P=0.36

Gophers n=540 P=0.30

No gophers n=600 P=0.42

Surface acorns n=1140 P=0.36

Mice n=660 P=0.31

No mice n=480 P=0.50

Buried acorns n=1080 P=0.62

Gophers n=510 P=0.50

No gophers n=570 P=0.73

Fig. 2.11. The proportion of blue oak acorns that produced seedlings (P) in relation to acorn position (on the soil surface or buried 1 inch beneath the soil surface) and predation by mice and gophers. The chart is based on a direct seeding experiment of 4500 (n) visibly sound acorns sown in two different years in a blue oak woodland on a north-facing slope in the Central Coastal Range of California. Overstorey crown cover was 65% (128 trees per acre); ground cover was grasses and forbs. Various types of exclosures were used to selectively limit access to acorns by mice, gophers, birds, deer and cattle. The chart shows only the factors that significantly reduced P (based on statistical criteria). It is also hierarchical, i.e. the factors considered decrease in importance from top to bottom, based on statistical analysis. The bottom stratum of the chart represents the independent effect of each animal, i.e. mice in the absence of other animals, and gophers in the absence of other animals. For example, the lower right box of the chart shows that of the 1080 acorns experimentally buried in 1986, 73% of those protected from Botta pocket gophers produced seedlings (exclusive of mouse predation). When gophers were not excluded (adjacent box), only 50% of acorns produced seedlings. The higher P values in 1986 were attributed to greater acorn viability and less droughty conditions that year. (From Borchert et al., 1989, used with permission.)

counter tactic known as ‘notching’ (Wood, 1938; Fox, 1982). Before acorn germination occurs, squirrels sometimes chew off the pointed end of the acorn, which contains the embryonic components that develop into the radicle and plumule. This fate may befall 50–60% of cached acorns (Fox, 1982). Notching usually prevents germination and thereby preserves the nutrient value of the remaining acorn. In an Illinois study, only 18% of notched acorns germinated compared to 94% of unnotched acorns (Wood, 1938). These tactics and counter tactics have been hypothesized as coevolved traits between the white oaks and squirrels (Barnett, 1977; Fox, 1982). Squirrels are not known to notch the acorns of the springgerminating red oaks (Fox, 1982).

The occurrence of an acorn crop initiates a chain reaction that ramiﬁes through an oak forest in ways that are not limited to factors involved only in oak regeneration. Rodents, deer and many other animal populations rise and fall with acorn production cycles (Van Dersal, 1940; Brown and Yeager, 1945; Goodrum et al., 1971; Pfannmuller, 1991). High rodent populations are also linked to increased transmission of Lyme disease from ticks to other mammals including humans (Ostfeld et al., 1996; Jones et al., 1998). Forest-dwelling rodents also impact other ecological processes that are not always apparent. For example, white-footed mice and shrews prey on gypsy moth larvae and pupae (Campbell, 1981), and are thought to play a

80

Chapter 2

All seedlings (both sites) n=2842 Ps=0.24

Oak woodland n=1610 Ps=0.37

Gophers n=543 Ps=0.22

Surface acorns n=136 Ps=0.09

Buried acorns n=407 Ps=0.27

Oak savanna n=1232 Ps=0.06

No gophers n=1067 Ps=0.44

Surface acorns n=404 Ps=0.30

Mice n=669 Ps=0.03

No mice n=563 Ps=0.10

Buried acorns n=663 Ps=0.53

Fig. 2.12. The proportion of blue oak seedlings that survived for 3 years (Ps) in an oak woodland and an oak savanna in the Central Coastal Range of California in relation to acorn position (on the soil surface or buried 1 inch beneath the soil surface) and predation by mice and gophers. The overstorey crown cover is 65% (128 trees per acre) in the woodland and 15% in the savanna (42 trees per acre). On both sites, the ground cover was grasses and forbs. Sample size (n) is the initial number of seedlings observed in their first year. The low average survival of seedlings in the savanna was attributed to grass competition. The chart is hierarchical and is based on the methodology described in Fig. 2.11. (From Davis et al., 1991.)

signiﬁcant role in controlling gypsy moth populations (Liebhold et al., 2000). In years of low to moderate gypsy moth abundance, predation by rodents may help to maintain those levels and thus reduce the rate of build-up in the gypsy moth population (Liebhold et al., 2000). In turn, this increases the time between major outbreaks and thus the defoliation of oaks and other susceptible plants (see Chapter 7). When defoliation does occur, it causes mortality or reduced vigour of oaks, and thus reduces acorn production itself (Gottschalk, 1990) – the initiator of the chain reaction. Such chain reactions, however, do not imply that all acorn predators are bound to acorns or without strategies for coping with the great temporal variability in acorn production. Insect predators such as some of the acorn weevils have a 2-year life cycle so that a single acorn crop failure does not destroy the entire insect population. Bird predators can move long distances (band-tailed pigeons and blue jays) and thus avoid acorn

scarcity or ﬁnd alternative foods. Deer can use alternative foods and white-footed mice can eat not only acorns but many different kinds of seeds as well as insects. Populations of acorn predators and dispersers also may remain relatively stable in forests where several species of oaks co-occur (Carmen et al., 1987). Because of the asynchronous acorn production among oak species, the probability of occurrence of a complete acorn crop failure decreases as the number of oak species increases. What was once perceived as a onedimensional negative effect, i.e. reduced oak regeneration potential from rodent predation, is now recognized as a more complex multidimensional forest process involving many organisms and possible effects. Rodents and other seed dispersers are important not only to the oaks, but also in their larger role as dispersers of many kinds of seeds, which in turn collectively contribute to maintaining forest biodiversity and resilience (Healy, 1988).

Regeneration Ecology I: Reproduction

Birds: predation and dispersal Birds rank among the most important dispersers and consumers of acorns. In the United States, acorns are eaten by more than 30 species of birds (Table 2.2). They com-

prise a large proportion of the diet of wood ducks, wild turkeys, band-tailed pigeons, blue jays, scrub jays, Stellar’s jay, redheaded woodpeckers and acorn woodpeckers (Koenig, 1980; Pavlik et al., 1991; Fuchs et al., 1997; Johnson et al., 1997). Although

Table 2.2. Some avian acorn eaters and their acorn dispersal habits.a Dispersal

Species American crow Band-tailed pigeon Clark’s nutcracker Common grackle Ducks Mallard Wood duck Hooded merganser Jays Blue Scrub Steller’s Grouse Greater prairie chicken Ruffed Sharptailed Mourning dove Quail California Northern bobwhite Ring-necked pheasant Titmice Plain Tufted White-breasted nuthatch Wild turkey Woodpeckers Acorn Arizona Golden-fronted Hairy Lewis’s Northern flicker Nuttall’s Pileated Red-bellied Red-headed Yellow-billed magpie a From

No dispersalb

81

To tree crevices and bark

To tree cavities

To ground caches

x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x xc x

x xc x

Van Dersal (1940), Kilham (1958), Bent (1964a,b,c), Griffin (1971), Scott et al. (1977), Koenig (1980, 1990), Gullion (1984), Petrides (1988), Pavlik et al. (1991), Fuchs et al. (1997), and other sources. b These species consume acorns directly from or near seed trees and generally do not disperse or cache acorns. c Acorns are often cached as broken pieces.

82

Chapter 2

other birds also eat acorns, acorns seldom make up a major portion of their diet. Blue jays are especially effective in dispersing acorns from closed-canopy oak stands to open or semi-open habitats such as old ﬁelds and forest edges, which often provide a favourable environment for seedling establishment (Darley-Hill and Johnson, 1981; Harrison and Werner, 1984; Johnson and Webb, 1989; Deen and Hodges, 1991). Acorns may be dispersed up to 2.5 miles from their origin, and individual jays may cache in the ground up to 3000 acorns per year (Johnson and Adkisson, 1986) (Fig. 2.13). Although only a small proportion of cached acorns survive to germinate and grow to maturity, the large numbers dispersed, a scattered dispersal pattern, and transport to favourable habitats facilitates the maintenance and even the expansion of the distributional range of the oak. Acorn dispersal by jays thereby contributes to maintaining biodiversity where forests are highly fragmented – a characteristic of much of the contemporary United States landscape. Dispersal by jays also has been proposed as an explanation for the rapid migration of oaks as the climate warmed during early postglacial times (Johnson and Webb, 1989). Scrub jays, which are widely distributed across the southwest and Florida, similarly disperse and ground-cache acorns. They disperse even greater numbers of acorns per bird than blue jays, but over shorter distances (De Gange et al., 1989). A single bird may cache up to 5000 acorns in a single season but will consume only about one-third of those cached (Pavlik et al., 1991). Other birds that ground-cache and consume acorns include the tufted titmouse (Kilham, 1958; Bent, 1964a), Steller’s jay and yellow-billed magpie (Pavlik et al., 1991). Several species of woodpeckers also consume and cache acorns. Some species store acorns in tree crevices, bark and cavities, which largely represent a loss with respect to potential seedling establishment. The acorn woodpecker of the southwest is a specialist in storing acorns in trees. Acorns are the most important item in their

diet, and are consumed directly on seed trees from time of ripening in the autumn until none remain. The acorn woodpecker also stores large quantities of acorns by drilling numerous acorn-size holes in the bark of trees. The holes are seldom deeper than the thickness of the bark (Bent, 1964b; Koenig, 1980). These ‘granary’ trees consequently suffer little or no apparent damage from the activity, even though a single tree may be riddled with thousands of holes (Fig. 2.14). Pines are favoured as granary trees but other species including oaks are also used. During the autumn acorn-gathering period, the woodpeckers carry acorns from nearby seed trees to the granaries where each acorn is placed into an excavation. As the acorns loosen from dehydration during the winter, the birds relocate acorns to holes of the appropriate size. The acorn woodpecker’s diet consists almost exclusively of acorns during the winter (MacRoberts and MacRoberts, 1976). The woodpecker consumes virtually all acorns stored in granaries. Granary acorns remaining in spring are consumed into the warmer months until supplies are exhausted. The birds form cooperative breeding units of up to 15 usually related birds that maintain and expand their granaries through successive generations (MacRoberts and MacRoberts, 1976; Hannon et al., 1987). The relatively ﬁxed number of granaries within a locale apparently limits bird population density. Estimates of the proportion of the annual acorn crop consumed by this woodpecker range from less than 10% in years of bumper acorn crops to over 50% in years of low acorn production (Koenig, 1980). On average, acorns provide about 25% of the annual caloric requirements of acorn woodpeckers and possibly up to 80% during the winter months. These calories are thus available during the critical winter period when other foods are least available. Acorn woodpecker populations consequently rise and fall with annual ﬂuctuations in acorn production. The maintenance of a cooperative social structure and spring breeding success also depend on winter acorn stores (Hannon et al., 1987).

Regeneration Ecology I: Reproduction

83

Fig. 2.13. A single blue jay can disperse several thousand acorns per year. They can carry up to two white oak acorns or five pin oak acorns in their expandable throat and oesophagus for up to 2 miles from their source. After filling its throat to capacity, a jay may place one last nut in its bill before departing (lower photo) (Johnson and Adkisson, 1986). (Photos by permission of Dr W. Carter Johnson, South Dakota State University, Brookings.)

High acorn woodpecker populations in coastal California in the winter are associated with a high overall abundance of oaks and the joint occurrence of at least ﬁve oak species within the territories of overwintering groups of birds (Bock and Bock, 1974). This assures a high probability of acorn

production from at least one species every year that is adequate for sustaining a high bird population. In the semidesert regions of Arizona, New Mexico and Texas, populations of acorn woodpeckers do not attain densities as high as those observed in the Paciﬁc Coast region. This difference may

84

Chapter 2

peckers consume acorns but are not known to cache them (Bent, 1964b; Scott et al., 1977).

Oak Seedling Establishment Germination and initial establishment

Fig. 2.14. Acorns cached in a granary tree by acorn woodpeckers in California. (Photo courtesy of Dr Walter Koenig, University of California, Berkeley.)

be related to the relatively few oak species that, within the semidesert regions, are likely to occur jointly at the relatively small spatial scales of bird territories. The red-headed woodpecker of eastern forests also caches acorns in trees. Like the acorn woodpecker, red-headed woodpecker densities are correlated with acorn abundance and may be more sensitive than blue jays to annual ﬂuctuations in acorn crops (Smith, 1986). Unlike the acorn woodpecker, the red-headed woodpecker favours caching acorns in tree crevices and deep cavities. They will often seal a cavity entrance with tightly compacted wood particles to conceal the cache from competitors (Kilham, 1958). Other woodpeckers known to cache acorns in trees include the red-bellied woodpecker of eastern forests (Kilham, 1963; Bent, 1964b), and Lewis’s woodpecker of the southwest (Bock and Bock, 1974). Several other species of wood-

The acorns of the two major oak groups, the white oaks and the red oaks, differ in their germination characteristics. The white oaks initiate germination in autumn, whereas the red oaks usually do not germinate until spring. In the white oaks, the radicle (immature root) develops rapidly after autumn germination. Acorns of blue oak and valley oak (both in the white oak group) in California have even been observed to start radicle growth in autumn while still hanging from the tree (Grifﬁn, 1971). However, in all the white oaks, the epicotyl (immature stem) usually does not develop until the following spring. In a small proportion of germinants, shoots may begin to develop in the autumn but overwinter as short succulent stems that require additional chilling to develop further (Farmer, 1977). White oaks thus undergo epicotyl dormancy, which is not broken until the acorn or developing germinant undergoes prolonged exposure to low temperatures such as those normally occurring over winter (Farmer, 1977). Acorns of the red oak group undergo embryo dormancy. Germination does not occur until acorns are chilled for a sufﬁcient time. The requisite chilling period is usually met during overwintering. Under artiﬁcial conditions, embryo dormancy can be broken by storing acorns for 30–60 days at 32–41°F (0–5°C) in moist sand or polyethylene bags (Olson, 1974; Bonner and Vozzo, 1987). In northern red oak, emergence of the radicle in the spring typically begins after about 10 days of sufﬁciently warm temperatures. About 5 days later, the plumule (immature shoot) starts to develop; within 20 days, leaves are unfolding (Fig. 2.15). Acorn germination is usually not delayed beyond the ﬁrst

Regeneration Ecology I: Reproduction

85

Fig. 2.15. Stages in the development of a germinating northern red oak acorn. (A) Emergence of the radicle (immature root) of a ripened acorn after 10 days in a favourable germination environment; (B) radicle development after 15 days; plumule (immature shoot) is beginning to emerge; (C) leaves are unfolding after 20 days. (From Korstian, 1927, used with permission.)

autumn in the white oaks or the ﬁrst spring in the red oaks. However, in a California study, 10% of winter-sown blue oak germinants did not produce emergent shoots from the soil until the second spring after acorns were planted (Tecklin and McCreary, 1991). These unusual events occurred during an exceptionally cold and wet year. Acorns none the less do not accumulate in the forest ﬂoor as do seeds of some other trees and woody plants such as yellow-poplar, white ash, pin cherry, black cherry and briars. To survive over the winter, acorns require protection from desiccation, low temperatures, and predation by insects, rodents and other animals. White oak acorns maintain viability at moisture contents above 40%, whereas red oaks maintain viability down to about 25% (Korstian, 1927). For both species, critical moisture levels can be reached after air-

drying acorns at room temperature for 2–5 days. In one study, the radicles of germinated white oak and chestnut oak acorns were killed by freezing temperatures, whereas ungerminated acorns were not damaged by temperatures a few degrees below freezing (Korstian, 1927). Although temperatures of 20°F (−7°C) for 1 week killed ungerminated white oak acorns, black and northern red oak acorns were not damaged (Aikman, 1934). Germinating white oak acorns with exposed radicles also were killed by temperatures of 20°F (−7°C) for 1 week, but did survive a 12hour exposure to that temperature. Because acorns fall in the autumn during or somewhat before leaf fall, they are usually protected from desiccation and freezing by the current year’s leaf fall (Korstian, 1927; Barrett, 1931). In northern latitudes, deep snow cover during the winter also protects acorns from desiccation and freezing.

86

Chapter 2

The proportion of sound acorns that can germinate is highly variable and even under laboratory conditions ranges from about 50 to 100% depending on species and other factors (Olson, 1974). Conditions that favour acorn germination and seedling establishment in the ﬁeld include a moist, friable soil that can be easily penetrated by the developing radicle. A covering of leaf litter sufﬁcient to prevent surface soil drying and acorn desiccation and freezing also creates favourable conditions (Korstian, 1927). However, a covering of leaves sometimes can be detrimental depending on its thickness and other factors (Nichols, 1954; Krajicek, 1955, 1960; Lockhart et al., 1995). A forest ﬂoor with a thick layer of matted, undecomposed leaves (F-layer) can hinder entry of the radicle into soil. More than 2–3 inches of litter can create a physical barrier to epicotyl emergence, and cause etiolated (pale and fragile) shoots (Barrett, 1931). In France, northern red oak seedling densities exceeding 100,000 per acre were attributed to rapid soil turnover by earthworms and associated acorn burial that created favourable conditions for both survival and germination (Steiner et al., 1993). Even after a bumper acorn crop, numbers of oak seedlings that become established vary greatly depending on predation and overwintering and germination conditions. When conditions are unfavourable, seedling establishment may completely fail. In contrast, hundreds of thousands of new seedlings per acre have been observed in some instances (e.g. Barrett, 1931; R.L. Johnson, 1975). More commonly, numbers range from about 1000 to 10,000 per acre under favourable conditions in oak forests of the eastern United States (e.g. Scholz, 1955; Tryon and Carvell, 1958; McGee, 1967; Johnson, 1974). In some regions, oak seedling establishment even in good seed years may average only a few hundred per acre. For example, after a good black oak acorn crop in the Ozark Highlands of Missouri, 270 seedlings per acre became established (Sander, 1979). In a year when black oak acorn crops were rated fair to good, about 150 seedlings per acre were

established (McQuilkin, 1983). Most of those originated from squirrel-buried acorns; 5–10% originated from acorns cached by mice or voles in or just below the forest ﬂoor. In oak woodlands in California’s Coastal Range, the establishment of large numbers of new oak seedlings occurs only in a wet spring (Grifﬁn, 1971). In that region, burial of acorns under soil or litter in shady places is required to mitigate dry autumn or winter weather and to improve chances of seedling establishment.

Early growth Oak seedlings quickly develop a strong taproot, which usually grows to several inches in length within a few weeks after germination begins in the autumn (white oaks) or spring (red oaks) (Fig. 2.15). By the end of the ﬁrst growing season, taproots are typically about 20 inches long (Holch, 1931; Carpenter and Guard, 1954). In the spring, root growth increases rapidly as soil temperatures rise to an optimum near 75°F (24°C). Because soils are often cooler than this during the early growing season, rate of soil warming is an important factor in oak seedling establishment and growth (Larson, 1970). The early development of a large taproot and delayed shoot growth is characteristic of all oaks (Fig. 2.15). Early root development is facilitated by the acorn itself, which supplies from its cotyledons (the large ﬂeshy interior of the acorn) nutrients to the developing radicle. By the time these food reserves are exhausted, the seedling has typically developed three or four fully expanded leaves and is fully independent of the acorn. The taproots have developed the lateral roots necessary for absorbing water and nutrients. The taproot functions primarily as a food storage organ, whereas the ﬁne lateral roots function mainly to absorb water and nutrients. The taproot is largely non-functional in absorption because it is covered by thick corky (suberized) tissue that is essentially impermeable to water (Carpenter and Guard, 1954).

Regeneration Ecology I: Reproduction

Studies of several oak species have shown that large acorns are positively correlated with the following attributes: high germination percent, rapid shoot emergence, high seedling survival and rapid growth, large root mass and root:shoot ratio and rapid recovery from herbivory (defoliation and browse damage) (Korstian, 1927; McComb, 1934; Tripathi and Khan, 1990; Tecklin and McCreary, 1991; Bonﬁl, 1998). These attributes should theoretically confer survival and growth advantages to the large-seeded individual. There is evidence that acorn predators, including dispersers such as blue jays, prefer small over large acorns. This would seem to confer a further advantage to the large-seeded individual in terms of escape from predation. However, a species may generally beneﬁt as much or more from the advantages of dispersal and thus small seed size. Within a species, some trees also produce larger acorns than others. However, there is great variation in acorn size within a parent tree, and thus much overlap in acorn size among parent trees (Kriebel, 1965; Tecklin and McCreary, 1991). Few studies have monitored acorn size effects for more than one or two growing seasons. However, the advantage of large acorn size based on mean heights of 12-year-old half-sib families and mean family seed weight in European oaks persisted after 12 years (Johnsson, 1952). However, the correlation coefﬁcient between height growth and acorn size decreased from 0.94 after the ﬁrst growing season to 0.47 in the fourteenth year. Apparent acorn size effects thus diminished in importance with time, suggesting that other factors gradually assume more importance. The ﬁrst burst of shoot growth, or ‘ﬂush’, may last only about a week. During this time, the seedling typically grows to a height of about 6 inches. The shoot then enters a ‘resting’ (or lag) phase during which a new terminal bud is formed. It is from this bud that the next ﬂush of growth normally begins. Under favourable growing conditions such as those occurring in an irrigated nursery bed, the resting phase typically lasts from 2 to 4 weeks, but is

87

highly variable among seedlings within the bed. The length of the resting phase under forest conditions is even more variable and usually longer (Johnson, 1979a). Under ﬁeld conditions where seedlings are subjected to water stress, competition from other plants or low light, ﬁrst-year seedlings typically remain in the ﬁrst lag stage during the entire growing season. Shoot elongation then does not resume until the following spring. Under more favourable conditions, oak seedlings may ﬂush several times during a single growing season. There are three discrete stages of shoot development for each ﬂush of growth: (i) a period of rapid shoot elongation (stem linear stage); (ii) rapid leaf expansion (leaf linear stage); and (iii) resting (lag) stage. Because multiple ﬂushes can occur during a single growing season, these stages are successively referred to as 1-, 2- and 3-leaf linear stages, 1-, 2- and 3-lag stages, etc. (Hanson et al., 1986; Dickson, 1991). The term ‘linear’ refers to the essentially linear increase in stem length or leaf area during each observed growth interval (Fig. 2.16A and B). Shoot growth in oak is thus episodic, occurring in a series of bursts of growth (ﬂushes) each followed by a resting period. Root growth, on the other hand, is potentially continuous (Fig. 2.16A and B). Once initiated, it continues until environmental conditions become unfavourable. Thus, under ﬁeld conditions, the growth of oak roots may exhibit periodicity (Teskey, 1978; Reich et al., 1980). The proportion of photosynthates that are transported from the leaves of oak seedlings varies with the stage of shoot development. For example, in northern red oak, a seedling in the middle of its second ﬂush (i.e. 2-leaf linear stage) transports about 90% of the photosynthates produced from ﬁrst-ﬂush leaves to developing leaves of the second ﬂush (Fig. 2.17A). After the leaves of the second ﬂush have fully expanded and the shoot is in the resting (2lag) stage between ﬂushes, about 95% of the currently produced photosynthate is translocated to the lower stem and roots (Fig. 2.17B). Because oak seedlings spend

Chapter 2

6

A

B

5 25 4 20 3 15 2 10 1 5

Leaf area (dm2)

Cumulative root elongation (dm) Cumulative shoot elongation (dm)

88

Study day

0 10

20

30

40

50

Study day 60

70

10

20

30

40

50

60

70

50

100

150 200 250 300 350

Cumulative root elongation (dm)

C 240 200 160 120 80 40

70 60 50 day y 40 Stud

0 10 Lea 20 f ar ea 30 (dm 2 40 )

30

Fig. 2.16. Shoot, leaf and root growth of northern red oak seedlings transplanted to root observation chambers. (A) This seedling flushed twice during the 70-day study period. Of the 61 days of observable shoot and root growth, the seedling was in the resting (lag) stage 43% of the time. Leaf area culminated at 22 dm2 (341 inches2 ) by day 58. After a 29-day delay, roots grew at a continuous, nearly linear rate. (B) This seedling flushed only once during the study period. Of the 61 days of observable shoot and root growth, the seedling was in the resting (lag) stage 82% of the time. Leaf area increased to 28 dm2 (435 inches2) by day 20 and thereafter remained constant. Root elongation was initiated on day 12. Roots elongated slowly until day 30 but grew rapidly thereafter. (C) Relation between cumulative root elongation and leaf area. The model is based on a population of 60 seedlings grown in root observation chambers for 70 days from late March to early June. The response surfaces shown are for container-grown seedlings with intact shoots (upper surface) and for shoots clipped 15 cm (6 inches) above the root collar (lower, partially hidden surface). Initial stem size of seedlings is held constant at 11 mm (0.4 inches) in basal diameter measured 2 cm above the root collar, and at 50 cm (20 inches) length. Response surfaces for bare-root seedlings (not shown) lie below those of container-grown seedlings for a given shoot clipping treatment. Time (study day), leaf area, shoot clipping, seedling type and initial shoot size explained 78% of the variation in root elongation based on linear regression. (Reprinted from Johnson et al., 1984, by permission of Society of American Foresters, Bethesda, Maryland. Not for further reproduction.)

Regeneration Ecology I: Reproduction

(A)

(B)

5%

90% 5%

89

5% 95%

Fig. 2.17. The transport of photosynthates from the leaves of a northern red oak seedling. Percentages are based on the allocation of 14C 48 hours after exposure to 14CO2. (A) A seedling in the middle of its second flush (2-leaf linear stage). In this stage of shoot development, stem elongation is complete but leaf expansion is incomplete; carbon transport is largely from first-flush leaves to developing secondflush leaves. (B) Leaves of the second flush are fully expanded and the shoot is in the resting (2-lag) stage. During this stage, carbon transport is largely from second-flush leaves to the lower stem and roots. This pattern of carbon allocation favours the development of a large root : shoot ratio. (From Dickson, 1991.)

much of their time in the resting stage (Fig. 2.16A and B), this pattern of carbon allocation favours the development of a large root:shoot ratio. With each successive ﬂush during the growing season, an oak seedling increases its leaf area, which in turn increases root growth (Fig. 2.16C). Thus, seedlings that ﬂush frequently quickly develop large leaf and root surface areas. Although lateral roots usually comprise only a small proportion of a seedling’s total root mass, they comprise most of the root’s absorptive surface area. This results from the greater surface area per unit mass of small-diameter roots than large-diameter roots (Fig. 2.18A). Moreover, as root diameters increase, the greater the proportion of non-absorptive corky tissue that covers the roots (Carpenter and Guard, 1954). The absorptive capacity of a taproot is therefore even proportionately lower than its small surface area would indicate. All plants lose water to the atmosphere through microscopic pores in their leaves (stomata) during the physiological process of transpiration (Kozlowski et al., 1997). Water lost to transpiration is normally replaced through root absorption. The rate

of water lost through transpiration can be expressed as grams of water per square centimetre of leaf area per day. In northern red oak seedlings grown in a greenhouse, transpiration rate increased as the ratio of total root surface area to leaf surface area increased (Fig. 2.18B). Thus, when water is non-limiting (as in frequently watered seedlings), its absorption from the soil and conduction through the plant to the atmosphere increases as root surface area increases relative to leaf area (other factors being equal). Under conditions of unlimited water supply, stomata normally remain open during the day. As the transpirational stream passes through open stomata, carbon dioxide (CO2) is simultaneously taken up from the atmosphere. The CO2 is then transformed through photosynthesis and other physiological processes into the various components of the living plant. Large, actively growing roots thus favour efﬁcient water absorption and sustained transpiration and photosynthesis. Although species differ in how they allocate carbon to the various parts of the plant, early investment in building a large root system and high root:shoot ratio appears to be an important physiological

90

Chapter 2 1000 1000 surface RootRoot surface areaarea (cm2(cm ) 2)

800

A

Lateral roots 2 mm diameter

A

Lateral roots 2 mm diameter

800 600 600

Lateral roots 2 mm diameter

400

Lateral roots 2 mm diameter

400 200

Taproots

200

Taproots

0 0

0

5

10

25

30

35

0

5

10 Root 15 mass 20(g) 25

15

20

30

35

Root mass (g)

Leaf transpiration (g of water cm–2 day–1)

0.16

B

0.12

0.08

0.04

0.00 0.0

0.4

0.8

1.2

1.6

2.0

Root area : leaf area ratio Fig. 2.18. Root relations in northern red oak seedlings. (A) Relation between root surface area and root mass of seedlings for three root size classes. Based on 192 seedlings excavated for measurement one growing season after planting in a clearcut. Lateral root surface areas were estimated from electronic area meter measurements; taproot areas from length and diameter measurements (authors’ data). (B) Relation between leaf transpiration and the ratio of root surface area to leaf surface area in well-watered seedlings in a greenhouse experiment (1 cm2 = 0.16 inches2) (redrawn from Parker, 1949, with permission, copyright by American Society of Plant Biologists.).

characteristic common to all oak seedlings. One of the mechanisms by which these high ratios are maintained is through recurrent shoot dieback (see the following section on seedling sprouts). Although root and leaf areas are important morphological characteristics of oak seedlings, by themselves they tell us little about how oak seedlings function physiologically or how oak species differ from each other physiologically with respect to

water use. However, oaks vary in the efﬁciency at which they can take up carbon dioxide from the atmosphere while simultaneously losing water through transpiration. One measure of water-use efﬁciency in plants is provided by the ratio of waterloss resistance to carbon dioxide uptake resistance, which can be determined experimentally (Wuenscher and Kozlowski, 1971). The higher this ratio, the less water will be lost per unit of carbon dioxide

Regeneration Ecology I: Reproduction

ﬁxed. The ratio is also sensitive to leaf temperature, an important environmental variable. For any given species, this ratio generally increases with increasing leaf temperature up to a threshold value and then declines. Comparisons among sugar maple and black, bur, white and northern red oak seedlings demonstrated that the oaks were more efﬁcient in their water use than sugar maple up to a threshold leaf temperature of about 95°F (35°C) (Wuenscher and Kozlowski, 1971). Under most conditions, the water-use efﬁciency of black oak exceeded that of sugar maple by a factor of four or more. Bur oak ranked second in efﬁciency, and was followed by white and northern red oaks, which behaved similarly. The high water-use efﬁciency of black oak at high leaf temperatures (104°F or 40°C) is consistent with its observed ability to grow on hot dry sites. Conversely, the low wateruse efﬁciency of sugar maple is consistent with its occurrence on only relatively moist sites and its absence on dry sites. Oaks as a group are generally regarded as drought tolerant even though there is much variation among species in this characteristic. Because plants are most vulnerable during the seedling stage of development, species evolve adaptive traits that confer a competitive advantage within their regeneration niche (Chapter 3). Oak seedlings have evolved several physiological and morphological characteristics that confer drought tolerance including large seeds that provide food reserves for a protracted period during and after germination, the rapid development of a long taproot, ability to photosynthesize and conduct water through the xylem under high water stress, ﬂexibility in maintaining high root:shoot ratios through recurrent shoot dieback, physiological plasticity that facilitates adjustment to varying water stresses, and large genetic variation in drought tolerance within species (Kriebel, 1988; Matsuda et al., 1989; Abrams, 1990; Kubiske and Abrams, 1992; Bragg et al., 1993; Pallardy and Rhoads, 1993; Rice and Struve, 1997). Among seven commonly occurring species in the eastern United States,

91

McQuilkin (1983) ranked them from most to least drought tolerant based on the literature as follows: blackjack oak > post oak > scarlet oak > chestnut oak > black oak > white oak (Bourdeau, 1954; Mowbray and Oosting, 1968; Racine, 1971; Wuenscher and Kozlowski, 1971; Seidel, 1972; Fralish et al., 1978; Eyre, 1980; Parker et al., 1982). Oak seedlings are also relatively intolerant of shade, although species differ somewhat. For example, among ﬁve common upland oaks in the eastern United States, their shade tolerances have been ranked as follows: white oak > chestnut oak > northern red oak > black oak > scarlet oak (Burns and Honkala, 1990). Rates of photosynthesis in oak seedlings increase with increasing light intensity up to about onethird of full sunlight, but increase little with further increases in light intensity (Kramer and Decker, 1944; Bourdeau, 1954; Loach, 1967; McGee, 1968; Musselman and Gatherum, 1969; Phares, 1971; Shafer, 1971; Wuenscher and Kozlowski, 1971). Light levels under dense forest canopies often fall below 2% of that in the open (Hanson et al., 1986). Under those conditions, oak seedlings cannot live for long (Hanson et al., 1986; Crunkilton et al., 1992; Vivin et al., 1993). Low light levels also proportionately reduce the allocation of carbon to roots and increase the proportion allocated to aboveground parts of the seedling (Gottschalk, 1985; Kolb and Steiner, 1990). In turn, this impedes the development of a large root:shoot ratio even if the seedling survives. Shade tolerance in the oaks may not be as ﬁxed as species’ shade tolerance ratings might infer. For example, oaks have adapted to variation in light conditions by adjusting their time of spring budbreak according to their exposure to light the previous year. Seedlings or saplings growing beneath a forest canopy begin ﬂushing about 1 week earlier in the spring than open-grown oaks, and before the overstorey begins to leaf out (McGee, 1976, 1988, 1997). Understorey oaks therefore begin spring growth when both light and soil moisture conditions are favourable. This pattern of spring budbreak was common to all six of the oak species

92

Chapter 2

observed (white, black, scarlet, post, chestnut and northern red oaks) (McGee, 1997). Other understorey species including hickories, white ash, red maple and sugar maple also began budbreak and growth earlier under shade than in the open. However, some species began growth sooner in the open (serviceberry) or showed no consistent pattern between budbreak and the presence of a forest canopy (sassafras, yellow-poplar and blackgum) (McGee, 1986). In northern red oak seedlings, time of budbreak is also associated with elevation of acorn source. Across the elevational range of 1400–4600 ft above sea level in western North Carolina, acorns from the lowest elevation source ﬂushed 11 days earlier, on average, than the highest elevation source (McGee, 1974, 1997). This difference in ﬂush date was consistent among the four planting sites, which ranged from 1500 to 5400 ft in elevation. These apparent inherited differences in ﬂushing time thus represent genetic adjustments to the timing of budbreak that coincide with ‘safe’ periods of growth initiation. A later onset of growth with increasing elevation provides a frost avoidance mechanism, whereas the earlier onset of budbreak at lower elevations is consistent with spring frost risks and the competitive advantage of the early onset of growth at lower elevations. Collectively, these relations in ﬂushing pattern in oaks have practical silvicultural implications with respect to potential frost damage associated with early ﬂushing related to seed source/planting site relations, and the timing of overstorey removal (McGee, 1997).

Seedling Sprouts Shoot dieback and root : shoot ratio Seedling sprouts are seedlings whose shoots have died back and resprouted one or more times. They are often the predominant form of oak reproduction growing beneath the forest canopy (Fig. 2.19). Sprouts can originate from dormant buds located anywhere along the stem between the root collar and

A

B

Fig. 2.19. Shoot dieback in oak seedling sprouts growing beneath a closed forest canopy in central Missouri. (A) Partial dieback (leafless portion of stem) in a northern red oak seedling sprout. New shoots from dormant buds were initiated 10 inches above ground and near ground line. (B) Complete dieback in a black oak seedling sprout to ground line (leafless stem on right) with two basal sprouts. Both seedling sprouts are approximately 30 inches tall (to tips of tallest dead stems), and both had previously died back several times. (USDA Forest Service, North Central Research Station photograph.)

Regeneration Ecology I: Reproduction

the terminal bud cluster. Dieback and resprouting appear to be important processes in the life of oak reproduction. Those processes are facilitated by a ‘bud bank’ comprised of a reservoir of visible and (to the naked eye) invisible buds that are continually being formed on new branches. In addition, older buds persist at the root collar and on older stems and branches (Wilson, 1993). Shoot dieback is common in all oaks growing under a forest canopy, but is especially prominent in Mediterranean and semidesert climates and elsewhere in droughty uplands. In those environments, oak stands are dominated by species with morphological and physiological adaptations to surviving repeated burning and water stress (Wuenscher and Kozlowski, 1971; Grimm, 1984; Abrams, 1990). Oak reproduction growing beneath a forest canopy is subject to stresses that periodically reduce shoot mass and leaf area through the process of recurrent shoot dieback. Surviving seedling sprouts thus

93

tend to develop larger root: shoot ratios as they age. In turn, high root: shoot ratio and large root mass enable oak reproduction to opportunistically respond to forest disturbances. The growth potential of oak reproduction is not expressed until it is released from the growth-inhibiting shade of the parent stand. This occurs when the forest canopy is partially or completely destroyed by disturbances resulting from ﬁre, windthrow, insects, disease, timber harvesting or other events. Then if the requisite size of the established oak reproduction has been obtained, it can produce two or more long ﬂushes of shoot growth each growing season after disturbance (Johnson, 1979a; Dickson, 1991). Variation in the root : shoot ratio within populations of oak reproduction produces great variation in shoot growth patterns among individual trees of different basal diameters (which in turn are correlated with root mass) (Fig. 2.20). Whether oaks become an important part of the succeed-

Net cumulative shoot growth (ft)

6 5

a

4

b

3

c d

2 e 1 0 1

3

5

7

9

11

13

15

17

Growth week Apr

May

June

July

Aug

Month Fig. 2.20. Patterns of shoot growth in oak sprouts for a range of parent-stem basal diameters (and thus root : shoot ratios). The growth shown occurred the second year after clearcutting in a stand in the Ozark Highlands of Missouri. (a) Black oak sprout originating from a 7-inch diameter stump; (b) black oak sprout originating from an 8-inch diameter stump (the sprout died back from frost during weeks 4 and 5 and subsequently flushed twice); (c) white oak sprout originating from a 1-inch diameter parent stem; (d) black oak sprout originating from a 4-inch stump; (e) white oak sprout originating from a parent stem less than 0.6 inches in diameter. (Adapted from Johnson, 1979a.)

94

Chapter 2

ing stand thus depends on the pre-disturbance development of signiﬁcant numbers of seedling sprouts with large roots and high root : shoot ratios. Oak reproduction is otherwise usually at a competitive disadvantage. This is especially true of the reproduction of the drought tolerant upland species which, even under optimal conditions, grow slowly until roots are large enough to support rapid shoot growth. Recurrent shoot dieback thus appears to be an important aspect of the evolutionary development and adaptive strategy of oaks.

Occurrence of shoot dieback Recurrent shoot dieback in oak reproduction differs from the ‘dieback and decline’ of older trees, natural pruning or twig abscission. The latter is a normal, nonpathological abscission process occurring in current twigs of mature oaks (Millington and Chaney, 1973). The phenomenon known as dieback and decline is common to older (non-juvenile) oaks. It is believed to be initiated by drought stress or insect defoliation, which may set up conditions conducive to disease and further insect attack (Staley, 1965; Nichols, 1968; Ammon et al., 1989; McCracken et al., 1991; Wargo and Haack, 1991; Dwyer et al., 1995; Jenkins and Pallardy, 1995). Dieback and decline of older trees usually result in the complete death of the tree (see Chapter 7). Shoot dieback also can be directly caused by insect damage to shoots of both juvenile and older oaks. But in juvenile oaks, resprouting regardless of the cause of dieback often follows shoot dieback. Whenever oak reproduction grows under a forest canopy, recurrent shoot dieback is a normal part of its life cycle. Dieback is especially prevalent on dry sites where reproduction survives long enough to accumulate over several acorn crops. There, oak seedlings and seedling sprouts must endure low light intensities together 4The

with high water stresses. Under those conditions, oak reproduction is likely to subsist near its compensation point 4 and thus on the edge of death (Hanson et al., 1986; Crunkilton et al., 1992; Vivin et al., 1993). This process reinforces the oak’s strategy of translocating to the root system most of the photosynthate produced during its juvenile growth period (Fig. 2.17). Through periodic reduction of shoot mass, the shoot can be minimized as a sink for the modest amount of photosynthate it produces when growing in shade (Hanson et al., 1986; Crunkilton et al., 1992; Dey and Parker, 1996). Seedling sprouts that survive recurrent shoot dieback thus eventually develop large root systems and large root : shoot ratios. The consequences of recurrent dieback enable oak reproduction to respond opportunistically to natural or silvicultural events that eliminate the shade-imposing overstorey. Shoot dieback in oak reproduction is linked to several factors including growing season water stress and spring frost (Bourdeau, 1954; Johnson 1979a; McGee, 1988). When soil moisture is ample such as in irrigated nursery beds, shoot dieback is uncommon. In contrast, recurrent dieback is characteristic of oak reproduction growing beneath a forest canopy (Liming and Johnston, 1944; Merz and Boyce, 1956; Tryon and Carvell, 1958; Sander, 1971; Abrams, 1990; McClaran and Bartolome, 1990; Crow, 1992). Experiments with seedlings of ﬁve upland oak species grown in pots showed that shoot dieback occurred in some seedlings when they were deprived of moisture (Bourdeau, 1954). After shoot dieback, new shoots developed from surviving dormant buds located below the point of dieback in some individuals of all but one of the species observed; some seedlings died. Water stress thus can directly result in seedling mortality or in shoot dieback followed by the growth of new shoots from surviving buds located below the dieback (Vivin et al., 1993). Succulent spring shoot growth is also susceptible to late spring

compensation point occurs at the light level where carbohydrate breakdown through respiration balances carbohydrate gain through photosynthesis.

Regeneration Ecology I: Reproduction

frosts. Frost-induced diebacks in the spring are usually shortly followed by the growth of new shoots from dormant buds below the dieback. Although shoot dieback in oak reproduction commonly occurs during the growing season as a result of spring frosts and summer water stress, it also occurs during the dormant season. Winter dieback and mortality of northern red oak seedlings occurring between November and spring was observed at the onset of spring growth in an Ohio greenhouse experiment of potted seedlings grown from seed (Wright et al., 1989). After overwintering out of doors, both mortality and dieback were greater in seedlings experimentally subjected to high water-stress and artiﬁcially imposed root injury (roots partially removed by severing) the previous summer than in seedlings that were only moderately stressed or that had uninjured roots. In a Missouri study, shoot dieback of English oak seedlings planted in a clearcut occurred between time of planting in late October and the completion of the ﬁrst ﬂush the following May (Johnson, 1981). Moreover, the average length of dieback was greater for bare root seedlings than for container grown seedlings with intact roots. Similarly, in autumn-planted northern red oak, the frequency of winter dieback of 10 cm or more was greater in bare root than in container grown seedlings. Frequency of dieback also increased with increasing shoot length, decreased with increasing basal diameter of stems for both types of seedlings, and for a given seedling size was more frequent in clearcuts than in shelterwoods (Fig. 2.21). Dormant season dieback thus was inﬂuenced by seedling characteristics (size and morphology) and the presence or absence of a forest canopy. Collectively, these observations conﬁrm that shoot dieback in oaks does occur during the winter. They further show that at least some of the observed variation in the frequency and amount of it is associated with factors that increase or decrease a seedling’s vulnerability to winter desiccation including:

95

• Large numbers of functional small roots at the start of the dormant season (greater in container grown than in bare root or root-injured seedlings), which are associated with a low frequency or small amount of dieback. • Large shoot diameters (positively correlated with root mass and root surface area), which are negatively associated with frequency of dieback. • Long stems, which are positively associated with frequency of dieback. • The presence of a forest canopy (potentially reducing frequency and/or intensity of freeze–thaw cycles), which is associated with reduced frequency of dieback. The observed occurrence of shoot dieback in oak reproduction is consistent with the principle of plant segmentation. According to this principle, the design of the plant’s water transport system (i.e. its hydraulic architecture) favours the preservation of the lower stem over the more vulnerable and less essential shoot tips (Zimmermann, 1983). By extension, shoot dieback in oak reproduction may be related to seasonal cycles of water stress involving: • Loss of ﬁne roots to desiccation (root shedding) or injury in late summer (Head, 1973; Joslin and Henderson, 1987; Wright et al., 1989; Yin et al., 1989), leading to reduced absorption of water during late summer and the following dormant season. • Xylem dysfunction (the formation of xylem embolisms, i.e. air blockages, and tyloses) occurring during both dormant and growing seasons, leading to excessively reduced hydraulic conductivity especially in the terminal sections of stems (Zimmermann, 1983; Tyree, 1989; Cochard et al., 1992; Sperry and Sullivan, 1992; Tyree and Cochard, 1996). • Bud desiccation and mortality and the failure of the stem to initiate cambial growth in the spring between dead buds and the next lower live bud.

Chapter 2

0.6 Bare-root 0.4 Containergrown 0.2

80

ho

60

ls

12 Initia 10 l sho ot dia 8 mete r (mm )

ot

14

itia

0.0

len gt h

(c m )

100

6

In

Probability of shoot dieback >10 cm

96

Fig. 2.21. Estimated probabilities of occurrence of winter shoot dieback (10 cm or more) in northern red oaks autumn-planted under shelterwoods in the Ozark Highlands of Missouri. Estimates are shown in relation to initial seedling size and type of nursery stock. Dieback was observed shortly after the completion of spring flushing of the first field growing season. The response surface shown is for trees planted under shelterwoods thinned to 60% stocking based on Gingrich’s (1967) stocking equation. The model is based on 2950 planted trees and is given by the logistic regression equation: P = 1/(1 + exp(–(–6.1098 + 0.0483•H – 0.1510•D – 0.4251•S – 1.1932•T))), where P is the estimated probability of shoot dieback ≥ 10 cm between the time of planting (late October) and spring flushing (mid-May), H is initial tree shoot length (cm), D is initial basal diameter 2 cm above the root collar (mm), S is presence of shelterwood (shelterwood absent = 0, shelterwood present = 1), and T is type of nursery stock (bare-root = 0; container-growing = 1). For all parameter estimates, P < 0.001 except S (P = 0.01). Based on the Hosmer–Lemeshow goodness-of-fit test, differences between observed and estimated probabilities did not differ significantly (P = 0.0001). Estimates for bare-root stock represent the average of 2+0 seedlings and 1+1 transplants, which did not differ significantly ( = 0.05). Estimated dieback probabilities for trees planted in clearcuts are 0.01 to 0.11 larger than for trees planted under shelterwoods, depending on initial shoot diameter and length. (From authors’ data.)

Survival of buds through summer droughts and the dormant season is essential to initiating cambial growth the next spring (Romberger, 1963). This growth is initiated by the downward translocation of growth regulators originating in buds (Wareing, 1951). If buds in the terminal cluster die from winter desiccation (or any other cause), subsequent cambial initiation and thus a new annual ring will fail to develop between the terminal cluster and the next lower bud (Zasada and Zahner, 1969). If a new growth ring fails to develop, there can be little or no upward conduction

of water from the roots because older annual rings in oaks are largely nonfunctional (Zimmerman, 1983). Shoots may then die back to the next lower living bud (Fig. 2.19). The downward progression of shoot dieback in oak reproduction frequently can be observed in the spring when new ﬂushes of shoot growth from terminal buds fail to develop and one or more of the normally suppressed lateral buds produce new shoots. In upland oak forests, there is much variation in the amount and frequency of seasonal shoot dieback (Crow, 1992). In dry

Regeneration Ecology I: Reproduction

upland oak forests, this is reﬂected in the large variation in differences between root and shoot ages among individuals in the same population (Merz and Boyce, 1956; Powell, 1976). Some of this variation may be associated with differences in root : shoot ratios, absolute root size and absorptive capacity, and thus the ability of roots of oak reproduction to replace water losses in stems and buds before and during the dormant season. Despite its role in reinforcing the apparent root-centred growth and survival strategy of oak reproduction, recurrent shoot dieback exacts a price. Loss of shoots during the dormant season can negatively affect spring root growth. In northern red oak seedlings, the artiﬁcial removal of seedling shoots 6 inches above the root collar in late autumn reduced spring root growth of seedlings transplanted to greenhouse root observation chambers by up to 59% (Fig. 2.16C). Similar results have been observed in other oak species (Lee et al., 1974; Farmer, 1975, 1979). The collective evidence suggests that loss of shoots during the dormant season, whether through artiﬁcial removal or natural dieback, can reduce regulatory root growth promoters that originate in shoots and buds (Vogt and Cox, 1970; Carlson, 1974; Larson, 1975; Farmer, 1979). In turn, reduced root growth in the spring may predispose reproduction to water stress and thus shoot dieback later in the growing season (Johnson, 1979b). Also, during the shoot regrowth period after shoot dieback, there is a greater allocation of growth to shoots than to roots (Cobb et al., 1985; Kruger and Reich, 1989). This consequently lengthens the time required for roots to reach the requisite size for supporting, under ﬁeld conditions, multiple ﬂushing and thus rapid height growth. The physiological effects of shoot losses vary among oak species (Lee et al., 1974) and with other factors. In a young clearcut in Wisconsin, rates of transpiration and net photosynthesis (per unit leaf area) in northern red oak seedlings with pruned shoots (cut off 1 inch above the root collar to simulate dieback) were 30% greater than in unpruned seedlings (Kruger and Reich,

97

1989). However, leaf areas of pruned seedlings were 25% lower than those of unpruned seedlings. The joint effects of increased photosynthetic rate and reduced leaf area thus were largely compensatory. The higher photosynthesis and transpiration rates of the pruned seedlings were attributed to possible (but unmeasured) increases in water availability or other factors associated with increased root : shoot ratio resulting from shoot pruning (Kruger and Reich, 1989). This conclusion is consistent with the behaviour of northern red oak seedlings grown in a greenhouse (Parker, 1949). Transpiration rate (and thus photosynthetic rate) increased as root surface area: leaf area ratio increased (Fig. 2.18B). Shoot pruning of cherrybark oaks growing under forest canopies in Mississippi, combined with mid-storey and understorey competition control, signiﬁcantly increased the height growth of seedlings and seedling sprouts based on 2-year observations (Lockhart et al., 1991). However, pruned seedlings in most cases did not completely regain the height lost to pruning. Survival also was lower in the pruned population of seedlings. The investigators nevertheless concluded, based on physiological and morphological measurements, that pruning-simulated dieback and subsequent resprouting confers a growth advantage to cherrybark oak reproduction. Low to moderate losses of leaves and shoots to defoliating insects and browsing animals can similarly result in growth and survival advantages to oak seedlings (Wright et al., 1989; Welker and Menke, 1990). But defoliation also can cause depletion of starch reserves in oak roots, which is augmented by drought (Parker and Patton, 1975). In blue oak seedlings in a California savanna, rapidly induced water stress combined with severe defoliation caused 100% seedling mortality after 2 years, whereas survival under slower rates of induced water stress combined with complete defoliation was associated with 80% survival (Welker and Menke, 1990). Defoliation of oak seedlings in the autumn also may reduce their growth the following spring (Larson, 1975).

98

Chapter 2

Despite some apparent negative effects, the widespread occurrence of shoot dieback in oak reproduction at the population level may be consistent with a growth and survival strategy that sacriﬁces large numbers of individuals for a high probability of obtaining a few individuals with large roots and attendantly rapid height growth potentials. Whatever the physiological explanation, shoot dieback in oak reproduction occurs within populations that are highly variable physiologically, morphologically and genetically. Moreover, individual seedlings and seedling sprouts occupy relatively small, homogeneous microenvironments within which there is a high probability of occurrence of insufﬁcient soil moisture, nutrients or light. Spatial variability of those resources at small scales thus may explain much of the observed variation in the frequency and amount of oak shoot dieback.

Stump Sprouts and Related Growth Forms Definitions and origins Once a seedling has died back and resprouted, its continued survival can lead to other growth forms. Various names have been applied to these growth forms, depending on how and where they originate on the parent stem or root system. One growth form is the stump sprout. These sprouts originate from dormant buds at or near the base of the stump of a cut tree (Fig. 2.22A and B). However, they also can arise from the bases of trees topkilled by ﬁre (Fig. 2.22C). In the Central Hardwood Region, stump sprouts are often deﬁned as those originating from a cut tree 2 inches dbh and larger (Roach and Gingrich, 1967). However, that deﬁnition has not been universally adopted. Biologically, the distinction between a stump sprout and a seedling sprout is arbitrary because all oaks, from small seedlings to large standing trees, have some potential to produce basal sprouts when the parent stem is cut. Moreover,

when wind, ﬁre or other factors destroy an oak stand, sprouts may develop from the bases of trees that have broken off or from standing trees with dead tops. Sprouts also can arise from dormant buds on the crowns of large, mature root systems having no aboveground stumps (Fig. 2.22D). This growth form, sometimes called a ‘grub’ (Curtis, 1959), is common in savannas, old ﬁelds and other disturbanceperpetuated communities where recurrent ﬁre or decay destroys stumps and where, simultaneously, low overstorey density maintains high light levels sufﬁcient for sprout survival over many decades. Such disturbances favour the accumulation of this form of reproduction, which like seedling sprouts, recurrently die back and resprout. The dieback process extends over a longer period than would be possible under the low light conditions of a closed canopy forest. Where favourable light conditions are maintained, oak grubs may attain ages of several hundreds of years (Curtis, 1959). Like seedling sprouts, stump sprouts and grubs originate from dormant buds at or near the root collar. These buds, connected to the pith of the tree by elements called bud traces, remain just beneath the bark by annually elongating the width of the annual ring (Liming, 1942). They do not develop further unless their vascular connections to the crown are severed by cutting or are otherwise interrupted. As long as the crown of the parent tree is alive, living buds under the bark usually remain in a dormant state imposed by growth-suppressing regulators translocated from the crown (Vogt and Cox, 1970). When buds fail to elongate each year, they are lost as a potential source of sprouts. Sometimes these losses are offset by buds that multiply by branching (Kramer and Kozlowski, 1979). The rate of bud branching and bud mortality changes with the age and size of the tree. The balance of these processes partially determines the spatial distribution and number of sprouts per stump that develop after cutting the parent tree. Also,

Regeneration Ecology I: Reproduction

99

A

B

C

D

Fig. 2.22. Basal sprouting in oaks. (A) A newly initiated black oak stump sprout originating from a basal bud on a recently cut tree. (B) A ten-stem clump of northern red oak stump sprouts 10 years after the parent tree was cut; the dominant stem is 27 ft tall. (C) Basal sprouts of scarlet oak originating after the top of the parent tree was killed by fire. (D) An excavated white oak seedling sprout with a large root (or ‘grub’) many times the mass of the shoots (from a sandy outwash plain in northern Lower Michigan). This grub probably originated after the parent tree was top-killed by fire and persisted through recurrent sprouting in the understorey long after the top and stump disappeared. The root section shown is about 6 ft long; scale is shown by hard hat near the root tip (centre foreground). (USDA Forest Service, North Central Research Station photographs.)

100

Chapter 2

some buds fail to produce shoots after the parent tree is cut because of the physical resistance of the bark to shoot emergence or inhibition by sprouts already emerged (Wilson, 1968). Physical resistance increases with increasing bark thickness and thus tree diameter. As trees become older and larger, their ability to sprout consequently decreases (Johnson, 1977; Weigel and Johnson, 1998). Stool sprouts and root sprouts differ from the other forms of vegetative reproduction because they originate from buds formed from callous tissue around wounds or other tissues (Kramer and Kozlowski, 1979). These are called adventitious buds and unlike dormant buds do not have bud traces extending to the pith of the tree. Stool sprouts develop adventitiously from the cut or wounded surfaces of stumps. They are usually weakly attached to the stump and are therefore often short-lived. Root sprouts, as the name implies, similarly originate from roots. In many oaks, stool sprouts and root sprouts are unimportant ecologically and silviculturally because they tend to be either short-lived or rarely occur. California black oak often produces stool sprouts, but not sprouts from dormant buds, when sprouts originate from high stumps of large trees (McDonald, 1990b). Other oaks reported to produce stool and/or root sprouts include canyon live oak (Thornburgh, 1990), blue oak (McDonald, 1990a), Oregon white oak (Stein, 1990), water oak (Adams, 1983) and swamp chestnut oak (Edwards, 1990a). Some arid-region oaks have evolved special belowground structures that produce sprouts. In Gambel oak, there are three distinct root-like belowground structures: lignotubers, rhizomes and true roots (Tiedemann et al., 1987). Lignotubers are burl-like structures with adventitious buds. These buds are the primary source of new shoots when tree crowns are killed. Several lignotubers may be connected by rhizomes, which have fewer buds. Rhizomes facilitate the development of wide spreading clones (Muller, 1951) that quickly develop from

the seedling state (Christensen, 1955) (Fig. 2.23). Later, the physical separation of individual lignotubers resulting from the death or destruction of connected lignotubers within a clone represents a form of plant multiplication (and thus population growth) that is important in the regeneration of rhizomatous oaks. Lignotubers and rhizomes are anatomically similar to stem wood in that both possess a pith, buds and bud traces. In contrast, roots are devoid of those features. Rhizomes in oaks are associated with arid and semi-arid climates (Muller, 1951), where environment is unfavourable for seedling establishment. In addition to Gambel oak, other rhizomatous oaks in the United States include Havard, sandpaper, live, Vasey, Texas live, Brewer, turbinella, Mohr, Ajo and huckleberry oaks (Muller, 1951). Except for live oak, all are shrubby species conﬁned to the arid southwest.

Frequency of sprouting The proportion of oak stumps that produce sprouts after trees are cut has been related to parent tree diameter and age, and site quality (Johnson, 1977; Weigel and Johnson, 1998). In general, sprouting decreases with increasing tree diameter, age and site quality (Fig. 2.24). But other factors such as season of cutting and shading also can affect stump sprouting in hardwoods. For some species of oaks, stump sprouting is greater for trees cut or killed during the dormant season than during the growing season (Clark and Liming, 1953). However, some of the live oaks of the western United States sprout proliﬁcally regardless of season of cutting (Longhurst, 1956). Few studies have evaluated the effect of shading on the stump sprouting of oaks or other hardwoods. However, all stumps sprouted in a 28-yearold water oak plantation in Louisiana regardless of thinning intensity (Gardiner and Helmig, 1997). In contrast, sugar maple stumps exposed to full light sprouted more frequently than shaded stumps (Church, 1960)

Regeneration Ecology I: Reproduction

101

(A)

(B)

(C)

Fig. 2.23. Regeneration by rhizomes in live oak. (A) Aerial shoot originating from a rhizome; (B) the distribution of rhizomes in a single clone (black dots indicate locations of aerial shoots); (C) a clump of trees in a single clone. (From Muller, 1951, used with permission.)

Sprout growth and mortality Young stump sprouts that arise from polesize and larger parent trees are, in effect, mature root systems connected to juvenile shoots. Even though stump sprouts start out as small as a new seedling (Fig. 2.22A), their large root systems buffer them from many of the adverse competition and site effects associated with the more limited site resources available to smaller reproduction. Stump sprouts thus have the potential to grow rapidly. During

their ﬁrst decade, open-grown stump sprouts in eastern United States may produce four or more ﬂushes of shoot growth per year totaling 3 ft or more even under droughty conditions (Johnson, 1979a; Reich et al., 1980; Cobb et al., 1985) (Fig. 2.20). Even when shaded by an overstorey of 34 ft2 acre1 of basal area, water oak stump sprouts grew at a rate of 1.7 ft year1 for the ﬁrst 5 years. However, this rate slowed to 0.75 ft year1 by age 7 when overstorey density had increased to 52 ft2 acre1 (Gardiner and Helmig, 1997).

102

Chapter 2

1.0

0.8

B 70

0.8

70

110

90

70 0.6

0.6

0.4

90

Parent tree age 50

0.4

70

0.2

0.2 White oak 0.0

90

110

110

110

Chestnut oak 0.0

0

4 8 12 16 20 Parent tree dbh (inches)

1.0

24

0

50

50 70

50 70

0.6

70

0.6

90 90

70 0.4

0.4

90

0.2

90 110

110 0.2

110

110

Scarlet and northern red oaks

Black oak 0.0

0.0 0

4

24

D

0.8 50

4 8 12 16 20 Parent tree dbh (inches)

1.0

C

0.8 Sprouting probability

50 90

Site index (ft) 60 75

50 Sprouting probability

1.0 50

A

8 12 16 20 24 Parent tree dbh (inches)

28

0

4

8 12 16 20 24 Parent tree dbh (inches)

28

Fig. 2.24. Estimated sprouting probabilities (P) for oak stumps in relation to site index and parent tree diameter and age. P is the probability that a stump will have at least one living sprout at the end of the first growing season after cutting. (A) White oak: P = {1/[1 + exp(–(–7.6179 – 1.5760•ln(dbh) + 2.7069•ln(SI) – 0.00667•ln(dbh)•age))]}. (B) Chestnut oak: P = {1/[1 + exp(–(4.5719 – 1.5760•ln(dbh) + 2.7069•SI – 0.00667•ln(dbh)•age))]}. (C) Black oak: P = {1/[1 + exp(–(–1.7718 – 0.0014•dbh•age + 0.0469•SI))]}. (D) Scarlet and northern red oaks: P = {1/[1 + exp(–(–1.1012 – 0.0014•age•dbh + 0.0469•SI))]}, where dbh is dbh of the parent tree (inches), SI is black oak site index in feet at base age 50 (from Carmean et al., 1989), and age is age of the parent tree in years. Based on trees in southern Indiana. (From Weigel and Johnson, 1998.)

Regeneration Ecology I: Reproduction

The large root mass of stump sprouts and their large carbohydrate storage and absorptive capacity, coupled with other physiological factors, facilitate multiple ﬂushing in oaks. In contrast most other growth forms do not produce multiple ﬂushes. Single-ﬂush trees include mature oaks, shaded seedlings and seedling sprouts, and small seedlings and seedling sprouts under water stress (Cook, 1941; Johnston, 1941; Kienholz, 1941; Longman and Coutts, 1974; Borchert, 1976; Buech, 1976). The number of shoot ﬂushes in oak stump sprouts declines with sprout age. In scarlet oak stump sprouts, numbers of ﬂushes decreased from an average of about two per growing season the ﬁrst year to about one by the fourth growing season (Cobb et al., 1985). By the ﬁfth year, numbers of ﬂushes approached that of the single-ﬂush mature tree. Borchert (1976) hypothesized that the progression from multiple to single ﬂushes as trees grow larger may be attributable to a declining root : shoot ratio that results in increasingly longer periods for roots and shoot to restore ‘functional balance’ following periods of shoot elongation and leaf expansion. The number and spatial distribution of sprouts around the stump also inﬂuence sprout growth. Excavation of English oak stump sprouts in Russia showed that the development of numerous sprouts that are well distributed around the stump can maintain the complete parent-tree root system (Kharitonovich, 1937). Such sprouts grew rapidly in comparison to sprouts with stems on only one side of the stump. The latter often withered and died unless the new sprouts developed an independent root system. Even then, sprouts were less vigorous than those connected to the living parent root system. An even distribution of sprouts around the stump increases the likelihood that the vascular connections of sprouts to roots facilitate the translocation of photosynthates to all sides of the

103

parent root system (Kharitonovich, 1937; Wilson, 1968; Kramer and Kozlowski, 1979). Numerous sprouts per stump were correlated with rapid early height growth in ﬁve species of oaks in the Ozark Highlands of Missouri (Johnson, 1977). Northern red oak stump sprouts in Wisconsin behaved similarly (P.S. Johnson, 1975). However, the initial advantage of high clump density may be quickly lost as competition between stems within clumps intensiﬁes with age. In northern red oak, 20 or more stems per clump are commonly initiated. The resulting crowding of stems induces a rapid decrease in clump density. Through this self-thinning process, numbers of live stems per clump typically decline to about four by age 15 (Fig. 2.25). Multiple stems may persist for 50 years or longer (Fig. 2.26). Similar rates of self-thinning were observed in water oak stump sprouts (Gardiner and Helmig, 1997). The negative effect of persistent multiple stems on the growth rate of individual stems has been demonstrated by clump thinning studies (Haney, 1962; Wendel, 1975; Lamson, 1983; Johnson and Rogers, 1984; Lowell et al., 1989). The diameter of the parent tree, and correlatively the size of the root system, affects the growth of oak stump sprouts. For ﬁve species of oaks in the Ozark Highlands, the correlation between stump diameter and the heights of the dominant stem within young sprout clumps in clearcuts was negative for all species (Johnson, 1977). Whether the relation is positive or negative, however, may depend on the range of stump diameters observed. For example, average 4year shoot elongation of the tallest stem in black oak and white oak clumps ranging from small seedling sprouts up to large stump sprouts increased as stump diameter increased up to a threshold diameter of 6 inches (Johnson, 1979b). For larger stumps, sprout height growth

104

Chapter 2

Number of living stems per clump

24

Parent tree dbh (inches) 10 15 20

20

16

12

8

4

0 4

8

12

16

20

24

Clump age (years)

Fig. 2.25. Number of living stems in 382 northern red oak sprout clumps in relation to sprout age and parent tree dbh. Each data point represents 1 to 14 clumps in clearcut stands in southwestern Wisconsin. Lines are mean trends based on a regression model that accounts for 33% of the variation in number of stems. (From P.S. Johnson, 1975, by permission of Society of American Foresters, Bethesda, Maryland. Not for further reproduction.)

Fig. 2.26. Four stems persisted in this 45-year-old clump of northern red oak stump sprouts in southeastern Minnesota. The largest stem is 14 inches dbh. (USDA Forest Service, North Central Research Station photograph.)

decreased with increasing stump diameter (Fig. 2.27A). The stump diameter associated with maximum sprout growth becomes more evident as the observed range of stump diameters increases. Although this general pattern of height growth in oak sprouts is common to ﬁve species in the Ozark Highlands of Missouri, species vary in the diameter (6–12 inches) associated with maximum growth (Fig. 2.27B). The maximum growth rate of reproduction along the continuum of parent stem diameters thus may represent the diameter where the optimum root : shoot ratio occurs most frequently in a genetically, physiologically and morphologically heterogeneous population of trees in a highly variable environment. Other factors also may be signiﬁcant sources of variation in the height growth of oak sprouts. These include variation in site quality (Fig. 2.28), stand density and thus competition from surrounding trees, insect and frost damage, animal browsing and season of cutting. The effects of these factors may vary greatly among oak species and regions.

Regeneration Ecology I: Reproduction

105

16 White oak

A

Black oak

4-year net shoot growth (ft)

14 12 10 8 6 Stump sprouts 4 2 0

2

4 6 8 10 Stump diameter (inches)

12

16 B

5th-year sprout height (ft)

14

White oak Scarlet oak

12

10

8

Black oak

6

Blackjack oak 5

10

15

20

Post oak 25

Stump diameter (inches)

Fig. 2.27. Height growth of oak stump sprouts (the tallest stem per clump) in relation to stump diameter in the Ozark Highlands of Missouri. (A) Predicted and observed 4-year net shoot growth of 43 black oak and 41 white oak sprouts during the first 4 years after clearcutting. Sprouts from stumps ≥ 2.4 inches in diameter represent ‘stump sprouts’ originating from the parent stand (‘overstorey’ trees). These stumps were measured 12 inches above ground. Sprouts from stumps < 2.4 inches in diameter (measured 4 to 6 inches above ground) represent ‘advance reproduction’, i.e. reproduction growing beneath the overstorey at the time of clearcutting. The estimate (curved line) illustrates the continuous non-linear relation between shoot growth and stump diameter across the two arbitrarily defined growth forms (i.e. advance reproduction and stump sprouts). The unexplained variation in shoot growth (R 2 = 0.39) may largely reflect the imperfect correlation between stump diameter and root size – the presumed (but unmeasured) primary ‘causal’ factor. Growth of the two species did not differ ( = 0.05). (From Johnson, 1979a.) (B) Predicted sprout heights of five species in 5-year-old clearcuts. Heights (H5) are based on the regression model: H5 = b0 [exp[–(b1D + b2•1/D)]] where D is stump diameter measured 6 inches above ground, and b1 and b2 are parameters estimated by regression analysis. (From Dey, 1991.)

106

Chapter 2

TSC

60

1.0

Sprout height (ft)

50

0.55 0.1

40 30 20 10 0 4

8

12

16

20

24

Sprout age (years) Fig. 2.28. Estimated and observed height of the tallest stem within young clumps of northern red oak stump sprouts in southwestern Wisconsin in relation to site quality expressed by the topographic site coefficient (TSC). TSC ranges from 0.1 (poorest sites) to 1.0 (best sites) and is based on depth of soil, aspect and slope position (see Chapter 4). (From P.S. Johnson, 1975, by permission of Society of American Foresters, Bethesda, Maryland. Not for further reproduction.)

References Abbott, D.L. (1960) The bourse shoot as a factor in the growth of apple fruits. Annals of Applied Biology 48, 434–438. Abrams, M.D. (1990) Adaptations and responses to drought in Quercus species of North America. Tree Physiology 7, 227–238. Abrams, M.D. and Kubiske, M.E. (1990) Leaf structural characteristics of 31 hardwood and conifer tree species in central Wisconsin: inﬂuence of light regime and shade-tolerance rank. Forest Ecology and Management 31, 245–253. Adams, J.C. (1983) Water oak regeneration in the South’s upland bottomland. USDA Forest Service General Technical Report SE SE-24, pp. 177–179. Ahlgren, I.F. (1974) The effect of ﬁre on soil organisms. In: Kozlowski, T.T. and Ahlgren, E.E. (eds) Fire and Ecosystems. Academic Press, New York, pp. 67–72. Aikman, J.M. (1934) The effect of low temperature on the germination and survival of native oaks. Iowa Academy of Science 41, 89–93. Aizen, M.A. and Kenigsten, A. (1990) Floral sex ratios in scrub oak (Quercus ilicifolia) vary with microtopography and stem height. Canadian Journal of Forest Research 68, 1364–1368. Ammon, V., Nebeker, T.E., Filer, T.H., McCracken, F.I., Solomon, J.D. and Kennedy, H.E., Jr (1989) Oak decline. Mississippi Agriculture & Forestry Experiment Station Technical Bulletin 151. Auchmoody, L.R., Smith, H.C. and Walters, R.S. (1994) Planting northern red oak acorns: Is size and planting depth important? USDA Forest Service Research Paper NE NE-693. Barnett, R.J. (1977) The effect of burial by squirrels on germination and survival of oak and hickory nuts. American Midland Naturalist 98, 319–330. Barrett, L.I. (1931) Inﬂuence of forest litter on the germination and early survival of chestnut oak, Quercus montana Willd. Ecology 12, 476–484. Bartolome, J.W., Muick, M.C. and McClaran, P. (1987) Natural regeneration of California hardwoods. USDA Forest Service General Technical Report PSW PSW-100, pp. 26–31. Beal, J.A. (1952) The more important insects of Duke Forest and the Piedmont Plateau. Duke University School of Forestry Bulletin 44, pp. 18–24.

Regeneration Ecology I: Reproduction

107

Beck, D.E. (1977) Twelve-year acorn yield in southern Appalachian oaks. USDA Forest Service Research Note SE SE-244. Bent, A.C. (1964a) Life Histories of North American Jays, Crows, and Titmice, Part II. Dover, New York. Bent, A.C. (1964b) Life Histories of North American Woodpeckers. Dover, New York. Bent, A.C. (1964c) Life Histories of North American Nuthatches, Wrens, Thrashers, and Their Allies. Dover, New York. Bock, C.E. and Bock, J.H. (1974) Geographical ecology of the acorn woodpecker: diversity versus abundance of resources. American Naturalist 108, 694–698. Bonﬁl, C. (1998) The effects of seed size, cotyledon reserves, and herbivory on seedling survival and growth in Quercus rugosa and Q. laurina (Fagaceae). American Journal of Botany 85, 79–87. Bonner, F.T. and Vozzo, J.A. (1987) Seed biology and technology of Quercus. USDA Forest Service General Technical Report SO SO-66. Borchert, M.I., Davis, F.W., Michaelsen, J. and Oyler, L.D. (1989) Interactions of factors affecting seedling recruitment of blue oak (Quercus douglasii) in California. Ecology 70, 389–404. Borchert, R. (1976) Differences in shoot growth patterns between juvenile and adult trees and their interpretation based on systems analysis of trees. Acta Horticulturae 56, 123–130. Bourdeau, P. (1954) Oak seedling ecology determining segregation of species in Piedmont oak– hickory forests. Ecological Monographs 24, 297–320. Brezner, J. (1960) Biology, ecology, and taxonomy of insects infesting acorns. University of Missouri Agriculture Experiment Station Research Bulletin 726. 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. Forest Ecology and Management 56, 29–41. Brooks, F.E. (1910) Snout beetles that injure nuts. West Virginia Agriculture Experiment Station Bulletin 128, pp. 143–185. Brooks, F.E. and Cotton, R.T. (1929) The chestnut Curculios, Curculio proboscideus Fab. and C. auriger Casey. USDA Technical Bulletin 130. Brown, L.G. and Yeager, L.E. (1945) Fox squirrels and gray squirrels in Illinois. Bulletin of Illinois Natural History Survey 23, 449–536. Brown, L.R. (1980) Insects feeding on California oak trees. USDA Forest Service General Technical Report PSW PSW-44, pp. 184–194. Brown, R.C. and Mogensen, H.L. (1972) Late ovule and early embryo development in Quercus gambelii. American Journal of Botany 59, 311–316. Buchholz, J.T. (1941) Multi-seeded acorns. Transactions of the Illinois Academy of Science 34, 99–101. Buech, R.R. (1976) Tree shoot elongation in northern Wisconsin and relationships with temperature and precipitation. Canadian Journal of Forest Research 6, 487–498. Burns, P.Y., Christisen, D.M. and Nichols, J.M. (1954) Acorn production in the Missouri Ozarks. University of Missouri Agriculture Experiment Station Bulletin 611. Burns, R.M. and Honkala, B.H. (1990) Silvics of North America. USDA Forest Service Agriculture Handbook 654, Vol. 2. Cahalane, V.H. (1942) Caching and recovery of food by the western fox squirrel. Journal of Wildlife Management 6, 338–352. Campbell, R.W. (1981) Population dynamics. In: Doane, C.C. and McManus, M.L. (eds) The Gypsy Moth: Research Toward Integrated Pest Management. USDA Forest Service Technical Bulletin 1584, Washington, DC, pp. 65–214. Carlson, W.C. (1974) Root initiation induced by root pruning in northern red oak. OARDC Research Summary. Ohio Agricultural Research and Development Center 74, pp. 14–16. Carmen, W.J., Koenig, W.D. and Mumme, R.L. (1987) Acorn production by ﬁve species of oaks over a seven year period at the Hastings Reservation, Carmel Valley, California. USDA Forest Service General Technical Report PSW PSW-100, pp. 429–434. Carpenter, I.W. and Guard, A.T. (1954) Anatomy and morphology of the seedling roots of four species of the genus Quercus. Journal of Forestry 52, 269–274. Cecich, R.A. (1994) The reproductive biology of Quercus with an emphasis on Q. rubra. USDA Forest Service General Technical Report NC NC-173, pp. 11–19. Cecich, R.A. (1997) Inﬂuence of weather on pollination and acorn production in two species of Missouri oaks. USDA Forest Service General Technical Report NC NC-188, pp. 252–261.

108

Chapter 2

Cecich, R.A. and Larsen, D.R. (1997) Bud demography and bifurcation in Quercus alba L. and Quercus velutina Lam. Diversity and Adaptation in Oak Species, Proceedings of Working Party 2.08.05, Oct. 12–17, 1997, Genetics of Quercus, of the International Union of Forest Research Organizations (Pennsylvania State University, University Park) 2, pp. 68–75. Cecich, R.A. and Sullivan, N.H. (1999) Inﬂuence of weather at time of pollination on acorn production of Quercus alba and Q. velutina. Canadian Journal of Forest Research 29, 1817–1823. Cecich, R.A., Brown, G.L. and Piotter, B.K. (1991) Pistillate ﬂower abortion in three species of oak (abstract). USDA Forest Service General Technical Report NE NE-148, p. 578. Christensen, E.M. (1955) Ecological notes on the mountain brush in Utah. Proceedings of Utah Academy of Science, Arts and Letters 32, 107–111. Christisen, D.M. and Kearby, W.H. (1984) Mast measurement and production in Missouri (with special reference to acorns). Missouri Department of Conservation Terrestrial Series 13. Church, T.W. Jr (1960) Factors affecting the development and survival of sugar maple sprouts. Proceedings of 1960 Society of American Foresters National Convention, pp. 32–35. Clark, F.B. and Liming, F.G. (1953) Sprouting of blackjack oak in the Missouri Ozarks. USDA Forest Service Central States Forest Experiment Station Technical Paper 137. Cobb, S.W., Miller, A.E. and Zahner, R. (1985) Recurrent shoot ﬂushes in scarlet oak stump sprouts. Forest Science 31, 725–730. Cochard, H., Granier, N.B. and Aussenac G. (1992) Vulnerability to air embolism of three European oak species (Quercus petraea (Matt.) Liebl., Q. pubescens Willd., Q. robur L.). Annales des Sciences Forestieres 49, 225–233. Coker, W.C. (1904) Multi-seeded acorns. Botanical Gazette (Chicago) 37, 61–62. Cook, D.B. (1941) The period of growth in some northeastern trees. Journal of Forestry 39, 956–959. Crawley, M.J. and Akhteruzzaman, M. (1988) Individual variation in the phenology of oak trees and its consequences for herbivorous insects. Functional Ecology 2, 409–415. Crow, T.R. (1992) Population dynamics and growth patterns for a cohort of northern red oak (Quercus rubra) seedlings. Oecologia 91, 192–200. Crunkilton, D.D., Pallardy, S.G. and Garrett, H.E. (1992) Water relations and gas exchange of northern red oak seedlings planted in a central Missouri clearcut and shelterwood. Forest Ecology and Management 53, 117–129. Curtis, J.T. (1959) The Vegetation of Wisconsin. University of Wisconsin Press, Madison. Cypert, E. and Webster, B.S. (1948) Yield and use by wildlife of acorns of water and willow oaks. Journal of Wildlife Management 12, 227–231. Dalke, P.D. (1953) Yields of seeds and mast in second growth hardwood forest, south central Missouri. Journal of Wildlife Management 17, 378–381. Darley-Hill, S. and Johnson, W.C. (1981) Acorn dispersal by the blue jay (Cyanocitta cristata). Oecologia 50, 231–232. Davis, F.W., Borchert, M., Harvey, L.E. and Michaelsen, J.C. (1991) Factors affecting seedling survivorship of blue oak (Quercus douglasii H. & A.) in central California. USDA Forest Service General Technical Report PSW PSW-126, pp. 81–86. Deen, R.T. and Hodges, J.D. (1991) Oak regeneration in abandoned ﬁelds: presumed role of the blue jay. USDA Forest Service General Technical Report SE SE-70, Vol. 1, pp. 84–93. DeGange, A.R., Fitzpatrick, J.W. and Layne, J.N. (1989) Acorn harvesting by Florida scrub jays. Ecology 70, 348–356. Delong, C.A. and Yahner, R.H. (1991) Impact of small mammals on regeneration of northern red oak. USDA Forest Service General Technical Report NE NE-148, p. 581. Dey, D.C. (1991) A comprehensive Ozark regenerator. PhD dissertation, University of Missouri, Columbia. Dey, D.C. and Parker, W.C. (1996) Regeneration of red oak (Quercus rubra L.) using shelterwood systems: ecophysiology, silviculture and management recommendations. Ontario Forest Research Institute Forest Research Information Paper 126. Dickson, R.E. (1991) Episodic growth and carbon physiology in northern red oak. Proceedings of the Oak Resource in the Upper Midwest Conference. University of Minnesota, pp. 117–124. Downs, A.A. and McQuilkin, W.E. (1944) Seed production of southern Appalachian oaks. Journal of Forestry 42, 913–920. Dwyer, J.P., Cutter, B.E. and Wetteroff, J.J. (1995) A dendrochronological study of black and scarlet oak decline in the Missouri Ozarks. Forest Ecology and Management 75, 60–75.

Regeneration Ecology I: Reproduction

109

Edwards, M.B. (1990a) Quercus michauxii Nutt. Swamp chestnut oak. USDA Forest Service Agriculture Handbook 654, Vol. 2, pp. 693–696. Eyre, F.H. (ed.) (1980) Forest Cover Types of the United States and Canada. Society of American Foresters, Washington, DC. Farmer, R.E., Jr (1975) Dormancy and root regeneration of northern red oak. Canadian Journal of Forest Research 5, 176–185. Farmer, R.E., Jr (1977) Epicotyl dormancy in white and chestnut oaks. Forest Science 23, 329–332. Farmer, R.E., Jr (1979) Dormancy and root growth capacity of white and sawtooth oaks. Forest Science 25, 491–494. Farmer, R.E., Jr (1981) Early growth of black cherry, oaks, and yellow-poplar in southern Appalachian plantings. Tree Planters’ Notes 32(3), 12–14. Feret, P.P., Kreh, R.E., Merkle, S.A. and Oderwald, R.G. (1982) Flower abundance, premature acorn abscission, and acorn production in Quercus alba L. Botanical Gazette (Chicago) 143, 216–218. Fox, J.F. (1982) Adaptation of gray squirrel behavior to autumn germination by white oak acorns. Evolution 36, 800–809. Fralish, J.S., Jones, S.M., O’Dell, R.K. and Chambers, J.L. (1978) The effect of soil moisture on site productivity and forest composition in the Shawnee Hills of southern Illinois. In: Balmer, W.E. (ed.) Proceedings of Soil Moisture–Site Productivity Symposium. USDA Forest Service, State and Private Forestry, Southeastern Area, Asheville, North Carolina, pp. 263–283. Fuchs, M.A., Krannitz, P.G., Harestad, A.S. and Bunnell, F.L. (1997) Seeds that ﬂy on feathered wings: acorn dispersal by Steller’s jays. USDA Forest Service General Technical Report PSW PSW-GTR-160, pp. 648–650. Galford, J.R. (1986) Primary infestation of sprouting chestnut, red, and white oak acorns by Valentinia glandulella (Lepidoptera: Blastobasidae). Entomological News 97, 109–112. Galford, J.R. (1987) Effect of Stelidota octomaculata (Coleoptera: Nitidulidae) on germinating acorns under laboratory conditions. Proceedings of the Central Hardwood Forestry Conference VI. University of Tennessee, Knoxville, pp. 419–422. Galford, J.R. and Weiss-Cottrill, D. (1991) Response of insects to damaged and undamaged germinating acorns. USDA Forest Service Research Paper NE NE-656. Galford, J.R., Peacock, J.W. and Wright, S.L. (1988) Insects and other pests affecting oak regeneration. Society of American Foresters Publication 88-03, pp. 219–225. Galford, J.R., Williams, R.N. and Beacom, M. (1991a) Notes on the biology and hosts of Stelidota ferruginea (Coleoptera: Nitidulidae). USDA Forest Service Research Paper NE NE-654. Galford, J.R., Williams, R.N. and Daugherty, A. (1991b) Life history and notes on the biology of Stelidota octomaculata (Coleoptera: Nitidulidae). USDA Forest Service Research Paper NE NE-644. Galford, J.R., Auchmoody, L.R., Smith, H.C. and Walters, R.S. (1991c) Insects affecting establishment of northern red oak seedlings in central Pennsylvania. USDA Forest Service General Technical Report NE NE-148, pp. 271–280. Galford, J.R., Auchmoody, L.R., Walters, R.S. and Smith, H.C. (1992) Millipede damage to germinating acorns of northern red oak. USDA Forest Service Research Paper NE NE-667. Gardiner, E.S. and Helmig, L.M. (1997) Development of water oak stump sprouts under a partial overstory. New Forests 14, 55–62. Gibson, L.P. (1962) Acorn weevils and associated insects of Ohio: their biology and ecology. MS thesis, Ohio State University, Columbus. Gibson, L.P. (1964) Biology and life history of acorn-infesting weevils of the genus Conotrachelus (Coleoptera: Curculionidae). Annals of the Entomological Society of America 57, 521–526. Gibson, L.P. (1969) Monograph of the genus Curculio in the New World (Coleoptera: Curculionidae), Part I. United States and Canada. Miscellaneous Publication of the Entomological Society of America 6, 239–285. Gibson, L.P. (1971) Insects of bur oak acorns. Annals of the Entomological Society of America 64, 232–234. Gibson, L.P. (1972) Insects that damage white oak acorns. USDA Forest Service Research Paper NE NE-220. Gibson, L.P. (1982) Insects that damage northern red oak acorns. USDA Forest Service Research Paper NE NE-492. Gingrich, S.F. (1967) Measuring and evaluating stocking and stand density in upland hardwood forests in the Central States. Forest Science 13, 38–53.

110

Chapter 2

Godman, R.M. and Mattson, G.A. (1976) Seed crops and regeneration problems of 19 species in northeastern Wisconsin. USDA Forest Service Research Paper NC NC-123. Goodrum, P.D., Reid, V.H. and Boyd, C.E. (1971) Acorn yields, characteristics, and management criteria of oaks for wildlife. Journal of Wildlife Management 35, 520–532. Gottschalk, K.W. (1985) Effects of shading on growth and development of northern red oak, black oak, black cherry, and red maple seedlings. I. Height, diameter, and root/shoot ratio. Proceedings of the 5th Central Hardwood Forest Conference. University of Illinois, pp. 189–195. Gottschalk, K.W. (1990) Gypsy moth effects on mast production. Proceedings of the Workshop on Southern Appalachian Mast Management. USDA Forest Service & University of Tennessee, pp. 42–50. Greulach, V.A. (1973) Plant Function and Structure. Macmillan, New York. Gribko, L.S. and Hix, D.M. (1993) The effect of small rodents on northern red oak acorns in north-central West Virginia. USDA Forest Service General Technical Report NC NC-161, pp. 400–408. Grifﬁn, J.R. (1971) Oak regeneration in the upper Carmel Valley, California. Ecology 552, 862–868. Grimm, E.C. (1984) Fire and other factors controlling the Big Woods vegetation of Minnesota in the mid-nineteenth century. Ecological Monographs 54, 291–311. Grisez, T.J. (1975) Flowering and seed production in seven hardwood species. USDA Forest Service Research Paper NE NE-315. Grubb, P.J. (1977) The maintenance of species-richness in plant communities: the importance of the regeneration niche. Biology Review 52, 107–145. Gullion, G. (1984) Grouse of the North Shore. Willow Creek Press, Oshkosh, Wisconsin. Gysel, L.W. (1957) Acorn production on good, medium, and poor oak sites in southern Michigan. Journal of Forestry 55, 570–574. Haney, G.P. (1962) Growth of thinned white oak sprouts. Journal of Forestry 60, 644–645. Hannon, S.J., Mumme, R.L., Koenig, W.D., Spon, S. and Pitelka, F.A. (1987) Poor acorn crop, dominance, and decline in numbers of acorn woodpeckers. Journal of Animal Ecology 56, 197–207. Hanson, P.J., Dickson, R.E., Isebrands, J.G., Crow, T.R. and Dixon, R.K. (1986) A morphological index of Quercus seedling ontogeny for use in studies of physiology and growth. Tree Physiology 2, 273–281. Harper, J.L. (1977) Population Biology of Plants. Academic Press, London. Harrison, J.S. and Werner, P.A. (1984) Colonization by oak seedlings into a heterogeneous successional habitat. Canadian Journal of Botany 62, 559–563. Head, G.C. (1973) Shedding of roots. In: Kozlowski, T. (ed.) Shedding of Plant Parts. Academic Press, New York, pp. 237–293. Healy, W.M. (1988) Effects of seed-eating birds and mammals on Appalachian hardwood regeneration. Society of American Foresters Publication 88-03, pp. 104–111. Helms, J.A. (ed.) (1998) The Dictionary of Forestry. Society of American Foresters, Bethesda, Maryland. Hildebrand, D.C. and Schroth, M.N. (1967) A new species of Erwinia causing the drippy nut disease of live oaks. Phytopathology 57, 250–253. Holch, A.E. (1931) Development of roots of tree seedlings. Ecology 12, 259–298. Hosner, J.F. (1959) Multi-seeded acorns. Journal of Forestry 57, 127. Janes, D.J., and Nichols, J.M. (1967) Acorn production of nine commercial species of oak in Missouri. School of Forestry, University of Missouri (unpublished manuscript). Jenkins, M.A. and Pallardy, S.G. (1995) The inﬂuence of drought on red oak group species growth and mortality in the Missouri Ozarks. Canadian Journal of Forest Research 25, 1119–1127. Johnson, P.S. (1974) Survival and growth of northern red oak seedlings following a prescribed burn. USDA Forest Service Research Note NC NC-177. Johnson, P.S. (1975) Growth and structural development of red oak sprout clumps. Forest Science 21, 413–418. Johnson, P.S. (1977) Predicting oak stump sprouting and sprout development in the Missouri Ozarks. USDA Forest Service Research Paper NC NC-149. Johnson, P.S. (1979a) Shoot elongation of black oak and white oak sprouts. Canadian Journal of Forest Research 9, 489–494. Johnson, P.S. (1979b) Growth potential and ﬁeld performance of planted oaks. Proceedings of the 1979 J.S. Wright Forestry Conference. Purdue University, pp. 113–119.

Regeneration Ecology I: Reproduction

111

Johnson, P.S. (1981) Early results of planting English oak in an Ozark clearcut. USDA Forest Service Research Paper NC NC-204. Johnson, P.S. and Rogers, R. (1984) Predicting 25th-year diameters of thinned stump sprouts of northern red oak. Journal of Forestry 82, 616–619. Johnson, P.S., Novinger, S.L. and Mares, W.G. (1984) Root, shoot, and leaf area growth potentials of northern red oak planting stock. Forest Science 30, 1017–1026. Johnson, R.L. (1975) Natural regeneration and development of Nuttall oak and associated species. USDA Forest Service Research Paper SO SO-104. Johnson, W.C. and Adkisson, C.S. (1986) Airlifting the oaks. Natural History 10, 41–43, 48–49. Johnson, W.C. and Webb, T. III. (1989) The role of blue jays (Cyanocitta cristata L.) in the postglacial dispersal of fagaceous trees in eastern North America. Journal of Biogeography 16, 561–571. Johnson, W.C., Adkisson, C.S., Crow, T.R. and Dixon, M.D. (1997) Nut caching by blue jays (Cyanocitta cristata L.): implications for tree demography. American Midland Naturalist 138, 357–370. Johnsson, H. (1952) Ungdomsutvecklingen hos stjälkek, druvek, och rödek. Svenska Slogsvardsföreningens Tidskrift 2, 168–193. Johnston, J.P. (1941) Height-growth periods of oak and pine reproduction in the Missouri Ozarks. Journal of Forestry 39, 67–68. Jones, C.G., Ostfeld, R.S., Richard, M.P., Schauber, E.M. and Wolff, J.O. (1998) Chain reactions linking acorns to gypsy moth outbreaks and Lyme disease risk. Science 279, 1023–1026. Joslin, J.D. and Henderson, G.S. (1987) Organic matter and nutrients associated with ﬁne root turnover in a white oak stand. Forest Science 33, 330–346. Kearby, W.H., Christisen, D.M. and Myers, S.A. (1986) Insects: their biology and impact on acorn crops in Missouri’s upland forests. Missouri Department of Conservation Terrestrial Series 16. Keeley, J.E. (1981) Reproductive cycles and ﬁre regimes. USDA Forest Service General Technical Report WO WO-26, pp. 231–277. Kharitonovich, F.M. (1937) Reforestation by coppice shoots. Sbornik Polezaschitnye Lesnye Polosy (Moscow) (in Russian), pp. 175–239. Kienholz, R. (1941) Seasonal course of height growth in some hardwoods in Connecticut. Ecology 22, 249–258. Kilham, L. (1958) Sealed-in winter stores of red-headed woodpeckers. Wilson Bulletin 70, 107–113. Kilham, L. (1963) Food storing of red-bellied woodpeckers. Wilson Bulletin 75, 227–234. Koenig, W.D. (1980) Acorn storage by acorn woodpeckers in an oak woodland: an energetics analysis. USDA Forest Service General Technical Report. PSW PSW-44. Koenig, W.D. (1990) Oaks, acorns and the acorn woodpecker. Fremontia 18, 77–79. Koenig, W.D., Carmen, W.J., Stanback, M.T. and Mumme, R.L. (1991) Determinants of acorn productivity among ﬁve species of oaks in central coastal California. USDA Forest Service General Technical Report PSW PSW-126, pp. 136–142. Koenig, W.D., Mumme, R.L., Carmen, W.J. and Stanback, M.T. (1994) Acorn production by oaks in central coastal California: variation within and among years. Ecology 75, 99–109. Koenig, W.D., Knops, J.M.H., Carmen, W.J., Stanback, M.T. and Mumme, R.L. (1996) Acorn production by oaks in central coastal California: inﬂuence of weather at three levels. Canadian Journal of Forest Research 26, 1677–1683. Kolb, T.E. and Steiner, K.C. (1990) Growth and biomass partitioning of northern red oak and yellowpoplar seedlings: effects of shading and grass root competition. Forest Science 36, 34–44. Komarek, E.V. (1970) Insect control–ﬁre for habitat management. Proceedings of the Tall Timbers Fire Ecology Conference 2, pp. 157–171. Korstian, C.F. (1927) Factors controlling germination and early survival in oaks. Yale University School of Forestry Bulletin 19. Kozlowski, T.T. and Pallardy, S.G. (1997) Physiology of Woody Plants, 2nd edn. Academic Press, San Diego. Krajicek, J.E. (1955) Rodents inﬂuence red oak regeneration. USDA Forest Service Central States Forest Experiment Station Note, Note 91. Krajicek, J.E. (1960) Some factors affecting oak and black walnut reproduction. Iowa State Journal of Science 34, 631–634. Kramer, P.J. and Decker, J.P. (1944) Relation between light intensity and rate of photosynthesis of loblolly pine and certain hardwoods. Plant Physiology 19, 350–358.

112

Chapter 2

Kramer, P.J. and Kozlowski, T.T. (1979) Physiology of Woody Plants. Academic Press, New York. Kriebel, H.B. (1965) Parental and provenance effects on growth of red oak seedlings. Proceedings, 4th Central States Forest Tree Improvement Conference, pp. 19–25. Kriebel, H.B., Merritt, C. and Stadt, T. (1988) Genetics of growth rate in Quercus rubra: provenance and family effects by the early third decade in the North Central U.S.A. Sylvae Genetica 37, 193–198. Kruger, E.L. and Reich, P.B. (1989) Comparative growth and physiology of stem-pruned and unpruned northern red oak. USDA Forest Service General Technical Report NC NC-132, 302 pp. Kubiske, M.E. and Abrams, M.D. (1992) Photosynthesis, water relations, and leaf morphology of xeric versus mesic Quercus rubra ecotypes in central Pennsylvania in relation to moisture stress. Canadian Journal of Forest Research 22, 1402–1407. Lamson, N.I. (1983) Precommercial thinning increases diameter growth of Appalachian hardwood stump sprouts. Southern Journal of Applied Forestry 7 (May), 93–97. Larson, M.M. (1970) Root regeneration and early growth of red oak seedlings: inﬂuence of soil temperature. Forest Science 16, 442–446. Larson, M.M. (1975) Pruning northern red oak nursery seedlings: effects on root regeneration and early growth. Canadian Journal of Forest Research 5, 381–386. Layne, J.N. (1954) The biology of the red squirrel, Tamiasciurus hudsonicus loquax (Bangs), in central New York. Ecological Monographs 24, 227–267. Lee, C.I., Moser, B.C. and Hess, C.E. (1974) Root regeneration of transplanted pin and scarlet oak. New Horizons 1974. Horticultural Research Institute, pp. 10–13. Lewis, V.R. (1989) Host–insect interactions in the California oakworm–coast live oak ecosystem. PhD Dissertation, University of California, Berkeley. Lewis, V.R. (1992) Within-tree distribution of acorns infested by Curculio occidentis (Coleoptera: Curculionidae) and Cydia latiferreana (Lepidoptera: Tortricidae) on the coast live oak. Environmental Entomology 21, 975–982. Liebhold, A., Elkinton, J., Williams, D. and Muzika, R.M. (2000) What causes outbreaks of the gypsy moth in North America? Population Ecology 42, 257–266. Liming, F.G. (1942) Blackjack oak in the Missouri Ozarks. Journal of Forestry 40, 249–252. Liming, F.G. and Johnston, J.P. (1944) Reproduction in oak–hickory forest stands of the Missouri Ozarks. Journal of Forestry 42, 175–180. Loach, K. (1967) Shade tolerance in tree seedlings. 1. Leaf photosynthesis and respiration in plants raised under artiﬁcial shade. New Phytologist 66, 607–621. Lockhart, B.R., Hodges, J.D., Toliver, J.R. and Karr, B.L. (1991) Morphology, gas exchange, and carbon-14 allocation patterns in advance cherrybark oak reproduction – preliminary results. USDA Forest Service General Technical Report SE SE-70, Vol. 2, pp. 513–523. Lockhart, B.R., Shelton, M.G. and Guo, Y. (1995). Effects of forest ﬂoor amount and composition and acorn placement on emergence and early establishment of white oak. USDA Forest Service General Technical Report SRS SRS-20, pp. 447–453. Longhurst, W.M. (1956) Stump sprouting of oaks in response to seasonal cutting. Journal of Range Management 9, 194–196. Longman, K.A. and Coutts, M.P. (1974) Physiology of the oak tree. In: Morris, M.G. and Perring, F.H. (eds) The British Oak, its History and Natural History. Botanical Society of the British Isles, Classey Ltd, Faringdon, UK, pp. 194–221. Lowell, K.E., Garrett, H.E. and Mitchell, R.J. (1989) Early thinning can improve your stand of coppice-regenerated oak. USDA Forest Service General Technical Report NC NC-132, pp. 53–58. Luckwill, L.C. (1977) The control of growth and fruitfulness of apple trees. In: Luckwill, L.C. and Cutting, C.V. (eds) Physiology of Tree Crops. Academic Press, New York, pp. 237–254. MacRoberts, M.H. and MacRoberts, B.R. (1976) Social organization and behavior of the acorn woodpecker in central coastal California. Ornithological Monographs 21. Marquis, D.A., Eckert, P.L. and Roach, B.A. (1976) Acorn weevils, rodents, and deer all contribute to oak-regeneration difﬁculties in Pennsylvania. USDA Forest Service Research Paper NE NE-356. Martin, R.E. and Mitchell, R.G. (1981) Possible, potential, probable, and proven ﬁre–insect interactions. Proceedings 1980 Society of American Foresters National Convention, pp. 138–144. Matsuda, K., McBride, J.R. and Kimura, M. (1989) Seedling growth form of oaks. Annals of Botany 64, 439–446. McClaran, M.P. and Bartolome, J.W. (1990) Comparison of actual and predicted blue oak age structures. Journal of Range Management 43, 61–63.

Regeneration Ecology I: Reproduction

113

McComb, A.L. (1934) The relation between acorn weight and the development of one year chestnut oak seedlings. Journal of Forestry 32, 479–484. McCracken, F.I., Ammon, V., Solomon, J.D. and Nebeker, T.E. (1991) Oak decline in the lower Mississippi River Valley. USDA Forest Service General Technical Report SE SE-70, Vol. 1, pp. 299–306. McDonald, P.M. (1990a) Quercus douglasii Hook. & Arn. Blue oak. USDA Forest Service Agriculture Handbook 654, Vol. 2, pp. 631–639. McDonald, P.M. (1990b) Quercus kelloggii Newb. California black oak. USDA Forest Service Agriculture Handbook 654, Vol. 2, pp. 661–671. McGee, C.E. (1967) Regeneration in southern Appalachian oak stands. USDA Forest Service Research Note SE SE-72. McGee, C.E. (1968) Northern red oak seedling growth varies by light intensity and seed source. USDA Forest Service Research Note SE, SE-90. McGee, C.E. (1974) Elevation of seed sources and planting sites affects phenology and development of red oak seedlings. Forest Science 20, 160–164. McGee, C.E. (1976) Differences in budbreak between shade-grown and open-grown oak seedlings. Forest Science 22, 484–486. McGee, C.E. (1986) Budbreak for twenty-three upland hardwoods compared under forest canopies and in recent clearcuts. Forest Science 32, 924–935. McGee, C.E. (1988) Spring weather, canopy removal, and early budbreak threaten oak seedlings. Journal of the Elisha Mitchell Science Society 104, 108–115. McGee, C.E. (1997) The effect of oak budbreak patterns on shade tolerance and regeneration. Diversity and Adaptation in Oak Species, Proceedings of Working Party 2.08.05, Oct. 12–17, 1997, Genetics of Quercus, of the International Union of Forest Research Organizations. Pennsylvania State University, University Park, 2, pp. 279–287. McQuilkin, R.A. (1983) Silvical factors affecting the regeneration of oaks and associated species in Pennsylvania. Proceedings Regenerating Hardwood Stands. Pennsylvania State University, pp. 37–65. McQuilkin, R.A. and Musbach, R.A. (1977) Pin oak acorn production on green tree reservoirs in southeastern Missouri. Journal of Wildlife Management 41, 218–225. Merkle, S.A., Feret, P.P., Croxdale, J.G. and Sharik, T.L. (1980) Development of ﬂoral primordia in white oak. Forest Science 26, 238–250. Merz, R.W. and Boyce, S.G. (1956) Age of oak ‘seedlings’. Journal of Forestry 54, 774–775. Millington, W.F. and Chaney, W.R. (1973) Shedding of shoots and branches. In: Kozlowski, T.T. (ed.) Shedding of Plant Parts. Academic Press, New York, pp. 149–204. Mogensen, H.L. (1965) A contribution to the anatomical development of the acorn in Quercus L. Iowa State Journal of Science 40, 221–255. Mogensen, H.L. (1975) Ovule abortion in Quercus (Fagaceae). American Journal of Botany 62, 160–165. Mowbray, T.B. and Oosting, H.J. (1968) Vegetation gradients in relation to environment and phenology in a southern Blue Ridge Gorge. Ecological Monographs 38, 309–344. Muick, P.C. and Bartolome, J.W. (1987) Factors associated with oak regeneration in California. USDA Forest Service General Technical Report PSW PSW-100, pp. 86–91. Muller, C.H. (1951) The signiﬁcance of vegetative reproduction in Quercus. Madroño 11, 129–137. Murtfeldt, M.E. (1894) Acorn insects, primary and secondary. Insect Life (USDA Division of Entomology) 6, 318–324. Musselman, R.C. and Gatherum, G.E. (1969) Effects of light and moisture on red oak seedlings. Iowa State Journal of Science 43, 273–284. Myers, S.A. (1978) Insect impact on acorn production in Missouri upland forests. PhD dissertation, University of Missouri, Columbia. Nakashizuka, T., Takahashi, Y. and Kawaguchi, H. (1997) Production-dependent reproductive allocation of a tall tree species Quercus serrata. Journal of Plant Research 110, 7–13. Neilson, R.P. and Wullstein, L.H. (1980) Catkin freezing and acorn production in gambel oak in Utah, 1978. American Journal of Botany 67, 426–428. Nichols, J.M. (1954) Direct seeding of oak in Missouri. University of Missouri Agriculture Experiment Station Bulletin 609. Nichols, J.O. (1968) Oak mortality in Pennsylvania. Journal of Forestry 66, 681–694.

114

Chapter 2

Olson, D.F., Jr (1974) Quercus L. Oak. USDA Forest Service Agriculture Handbook 450, 88 pp. Ostfeld, R.S., Jones, C.G. and Wolff, J.O. (1996) Of mice and mast. BioScience 46, 323–330. Pallardy, S.G. and Rhoads, J.L. (1993) Morphological adaptations to drought in seedlings of deciduous angiosperms. Canadian Journal of Forest Research 23, 1766–1774. Parker, J. (1949) Effects of variations in the root–leaf ratio on transpiration rate. Plant Physiology 24, 739–743. Parker, J. and Patton, R.L. (1975) Effects of drought and defoliation on some metabolites in roots of black oak seedlings. Canadian Journal of Forest Research 5, 457–463. Parker, W.C., Pallardy, S.G., Hinckley, T.M. and Teskey, R.O. (1982) Seasonal changes in tissue water relations of three woody species of the Quercus–Carya forest type. Ecology 63, 1259–1267. Pavlik, B.M., Muick, P.C., Johnson, S. and Popper, M. (1991) Oaks of California. Cachuma Press, Los Olivos, California. Petrides, G.A. (1988) A Field Guide to Eastern Trees. Houghton Mifﬂin, Boston, Massachusetts. Pfannmuller, L.A. (1991) Signiﬁcance of oaks and oak forest communities for nongame wildlife. Proceedings of The Oak Resource in the Upper Midwest Conference. University of Minnesota, pp. 56–64. Phares, R.E. (1971) Growth of red oak (Quercus rubra L.) seedlings in relation to light and nutrients. Ecology 52, 669–672. Pierce, W.D. (1908) A list of parasites known to attack American Rhynchophora. Journal of Economic Entomology 1, 380–396. Post, L.S. (1998) Seed management in Tennessee: development of seed zones for Tennessee and distribution and protection of northern red oak (Quercus rubra L.) acorns. MS thesis, University of Tennessee, Knoxville, Tennessee. Powell, D.S. (1976) Sprouting ability of advance reproduction of undisturbed forest stands in West Virginia. MS thesis, University of West Virginia, Morgantown. Racine, C.H. (1971) Reproduction of three species of oak in relation to vegetational and environmental gradients in the southern Blue Ridge. Bulletin of the Torrey Botanical Club 98, 297–310. Reich, R.B., Teskey, R.O., Johnson, P.S. and Hinckley, T.M. (1980) Periodic root and shoot growth in oak. Forest Science 26, 590–598. Rice, C. and Struve, D.K. (1997) Seedling growth form and water use of selected oak species. Diversity and Adaptation in Oak Species, Proceedings of Working Party 2.08.05, Oct. 12–17, 1997, Genetics of Quercus, of the International Union of Forest Research Organizations. Pennsylvania State University, University Park, 2, pp. 269–278. Roach, B.A. and Gingrich, S.F. (1967) Upland hardwoods can be grown efﬁciently. Pulp and Paper (April). Romberger, J.A. (1963) Meristems, growth and development in woody plants. USDA Forest Service Technical Bulletin 1293. Sander, I.L. (1971) Height growth of new oak sprouts depends on size of advance reproduction. Journal of Forestry 69, 809–811. Sander, I.L. (1979) Regenerating oaks. Proceedings of the National Silviculture Workshop (USDA Forest Service), pp. 212–221. Scholz, H.F. (1955) Effect of scariﬁcation on the initial establishment of northern red oak reproduction. USDA Forest Service Lake States Forest Experiment Station Technical Note 425. Schwartz, C.W. and Schwartz, E.R. (1959) The Wild Mammals of Missouri. University of Missouri Press, Columbia. Scott, V.E., Evans, K.E., Patton, D.R. and Stone, C.P. (1977) Cavity-nesting birds of North American Forests. USDA Forest Service Agriculture Handbook 511. Seidel, K.W. (1972) Drought resistance and internal water balance of oak seedlings. Forest Science 18, 34–40. Shafer, J.D. (1971) Shade tolerance of scarlet oak seedlings. MS thesis, University of Missouri, Columbia. Sharp, W.M. (1958) Evaluating mast yields in the oaks. Pennsylvania State University Agriculture Experiment Station Bulletin 635. Sharp, W.M. and Chisman, H.H. (1961) Flowering and fruiting in the white oaks. I. Staminate ﬂowering through pollen dispersal. Ecology 42, 365–372. Sharp, W.M. and Sprague, V.G. (1967) Flowering and fruiting in the white oaks. Pistillate ﬂowering, acorn development, weather, and yields. Ecology 48, 243–251.

Regeneration Ecology I: Reproduction

115

Shaw, M.W. (1968) Factors affecting the natural regeneration of sessile oak (Quercus petraea) in North Wales. II. Acorn losses and germination under ﬁeld conditions. Journal of Ecology 647–660. Short, H.L. (1976) Composition and squirrel use of acorns of black and white oak groups. Journal of Wildlife Management 40, 479–483. Sidney, C. (1948) Acorn weevils of the North Carolina Piedmont: Their biology and method of sampling infestation. MS thesis, Duke University, Durham, North Carolina. Smith, C.C. and Follmer, D. (1972) Food preferences of squirrels. Ecology 53, 82–91. Smith, K.G. (1986) Winter population dynamics of blue jays, red-headed woodpeckers, and northern mockingbirds in the Ozarks. American Midland Naturalist 115, 52–62. Smith, R.L. (1962) Acorn consumption by white-footed mice. West Virginia University Agricultural Experiment Station Bulletin 482T. Sork, V.L. and Bramble, J.E. (1993) Prediction of acorn crops in three species of North American oaks: Quercus alba, Q. rubra and Q. velutina. Annales des Sciences of Forestieres 50 (suppl. 1), 128s–136s. Sork, V.L., Bramble, J. and Sexton, O. (1993) Ecology of mast-fruiting in three species of North American deciduous oaks. Ecology 74, 528–541. Sperry, J.S. and Sullivan, J.E.M. (1992) Xylem embolism in response to freeze–thaw cycles and water stress in ring-porous, diffuse-porous, and conifer species. Plant Physiology 100, 605–613. Stairs, G.R. (1964) Microsporogenesis and embryogenesis in Quercus. Botanical Gazette (Chicago) 125, 115–121. Staley, J.M. (1965) Decline and mortality of red and scarlet oaks. Forest Science 11, 2–17. Stapanian, M.A. and Smith, C.C. (1978) A model for seed scatterhoarding: coevolution of fox squirrels and black walnuts. Ecology 59, 884–896. Stein, W.I. (1990) Quercus garryana Dougl. ex Hook. Oregon white oak. USDA Forest Service Agriculture Handbook 654, Vol. 2, pp. 650–660. Steiner, K.C, Abrams, M.D. and Bowersox, T.W. (1993) Advance reproduction and other stand characteristics in Pennsylvania and French stands of northern red oak. USDA Forest Service General Technical Report NC NC-161, pp. 473–483. Tecklin, J. and McCreary, D.D. (1991). Acorn size as a factor in early seedling growth of blue oaks. USDA Forest Service General Technical Report PSW PSW-126, pp. 48–53. Teskey, R.O. (1978) Inﬂuence of temperature and moisture on root growth of white oak. MS thesis, University of Missouri, Columbia. Thornburgh, D.A. (1990) Quercus chrysolepis Liebm. Canyon live oak. USDA Forest Service Agriculture Handbook 654, Vol. 2, pp. 618–624. Tiedemann, A.R., Clary, W.P. and Barbour, R.J. (1987) Underground systems of Gambel oak (Quercus gambelii) in central Utah. American Journal of Botany 74, 1065–1071. Tripathi, R.S. and Khan, M.L. (1990) Effects of seed weight and microsite characteristics on germination and seedling ﬁtness in two species of Quercus in a subtropical wet hill forest. Oikos 57, 289–296. Tryon, E.H. and Carvell, K.L. (1958) Regeneration under oak stands. West Virginia University Agricultural Experiment Station Bulletin 424T. Tryon, E.H. and Carvell, K.L. (1962) Acorn production and damage. West Virginia University Agricultural Experiment Station Bulletin 466T. Tucker, J.M. (1980) Taxonomy of California oaks. USDA Forest Service General Technical Report PSW PSW-44, pp. 19–29. Turkel, H.S., Rebuck, A.L. and Grove, A.R., Jr (1955) Floral morphology of white oak. Pennsylvania State University Agriculture Experiment Station Bulletin 593. Tyree, M.T. (1989) Cavitation in trees and the hydraulic sufﬁciency of woody stems. Annales des Sciences Forestieres (suppl.) 46, 330s–337s. Tyree, M.T. and Cochard, H. (1996) Summer and winter embolism in oak: impact on water relations. Annales des Sciences Forestieres 53, 173–180. USDA Forest Service (1985) Insects of eastern forests. Miscellaneous Publication 1426. Van Dersal, W.R. (1940) Utilization of oaks by birds and mammals. Journal of Wildlife Management 4, 404–428. Vivin, P., Aussenac, G. and Levy, G. (1993) Differences in drought resistance among 3 deciduous oak species grown in large boxes. Annales des Sciences Forestieres 50, 221–233. Vogt, A.R. and Cox, G.S. (1970) Evidence for the hormonal control of stump sprouting by oak. Forest Science 16, 165–171.

116

Chapter 2

Wareing, P.F. (1951) Growth studies in woody species IV. The initiation of cambial activity in ringporous species. Physiologia Plantarum. 4, 546–562. Wargo, P.M. and Haack, R.A. (1991) Understanding the physiology of dieback and decline diseases and its management implications for oak. Proceedings of The Oak Resource in the Upper Midwest Conference, University of Minnesota, pp. 147–158. Weigel, D.R. and Johnson, P.S. (1998) Planting northern red oak in the Ozark Highlands: A shelterwood prescription. USDA Forest Service Technical Brief TB-NC TB-NC-6. Weld, L.H. (1922) Note on American gallﬂies of the family Cynipidae producing galls on acorns with description of new species. Proceedings of the US National Museum 61, pp. 1–32. Welker, J.M. and Menke, J.W. (1990) The inﬂuence of simulated browsing on tissue water relations, growth and survival of Quercus douglasii (Hook and Arn.) seedlings under slow and rapid rates of soil drought. Functional Ecology 4, 807–817. Wendel, G.W. (1975) Stump sprout growth and quality of several Appalachian hardwood species after clearcutting. USDA Forest Service Research Paper NE NE-329. Williamson, M.J. (1966) Premature abscissions and white oak acorn crops. Forest Science 12, 19–21. Wilson, B.F. (1968) Red maple stump sprout development the ﬁrst year. Harvard Forest Paper 18. Wilson, B.F. (1993) Compensatory shoot growth in young black birch and red maple trees. Canadian Journal of Forest Research 23, 302–306. Winston, P.W. (1956) The acorn microsere, with special reference to arthropods. Ecology 37, 120–132. Wolgast, L.J. (1972) Mast production in scrub oak (Quercus ilicifolia) on the coastal plain in New Jersey. PhD dissertation, Rutgers University, Princeton, New Jersey. Wolgast, L.J. and Trout, J.R. (1979) Late spring frost affects yields of bear oak acorns. Journal of Wildlife Management 43, 239–240. Wood, O.M. (1934) A brief record of seed productivity for chestnut oak in southern New Jersey. Journal of Forestry 32, 1014–1016. Wood, O.M. (1938) Seedling reproduction of oak in southern New Jersey. Ecology 19, 276–293. Wright, S.L. (1987) Managing insects affecting oak regeneration by prescribed burning. USDA Forest Service General Technical Report SE SE-46, pp. 186–192. Wright, S.L., Hall, R.W. and Peacock, J.W. (1989) Effects of simulated insect damage on growth and survival of northern red oak (Quercus rubra L.) seedlings. Environmental Entomology 18, 235–239. Wuenscher, J.E. and Kozlowski, T.T. (1971) Relationship of gas-exchange resistance to tree-seedling ecology. Ecology 52, 1016–1023. Yin, X., Perry, J.A. and Dixon, R.K. (1989) Fine-root dynamics and biomass distribution in a Quercus ecosystem following harvest. Forest Ecology and Management 27, 159–177. Zasada, J.C. and Zahner, R. (1969) Vessel element development in the earlywood of red oak (Quercus rubra). Canadian Journal of Botany 47, 1965–1971. Zimmermann, M.H. (1983) Xylem Structure and the Ascent of Sap. Springer-Verlag, New York.

3 Regeneration Ecology II: Population Dynamics

Introduction This chapter is about the establishment and development of populations of juvenile oaks. Variability is a normal characteristic of tree populations, and it can be described in relation to speciﬁc tree attributes. For example, an oak forest can be described by the size or age distributions of its member trees, and how those distributions vary in time and space. Populations of one tree species also interact with other species, each with unique ecological requirements and competitive advantages and disadvantages that lead to variation in patterns of establishment, growth and survival. Population variability is further increased by forest disturbances. Predicting the responses of tree populations to forest disturbances, whether natural or of human origin, is fundamental to the practice of silviculture. The ﬁtness or suitability of adult trees to the physical characteristics of an ecosystem (i.e. the site) are often cited as explanations for the observed distribution of species across environmentally heterogeneous landscapes. However, relatively little emphasis has similarly been accorded to species’ regeneration requirements. The concept of the regeneration niche attempts to ﬁll that void. In contrast to the more general notion of species’ niche (Chapter 1), regeneration niche refers speciﬁcally to regeneration events and associated ecological conditions. Regeneration niche therefore considers

both the time and place where there is a high probability that a mature tree will be replaced by another tree of the same species (Grubb, 1977). Regeneration niches are ephemeral; their abundance and locations vary over time. The resource requirements of seedlings and seedlings sprouts (e.g. light, nutrients, heat and soil moisture) differ from those of mature trees. Therefore, it has been proposed that niche differences among species coexisting in the same ecosystem may be best expressed during the vulnerable early stages of plant establishment (Grubb, 1977; Latham, 1992; Veblen, 1992). Because all autotrophic plants require essentially the same kinds of resources (but not necessarily the same amounts), niche differentiation among species is likely to be strongly expressed during a life history period when one or more of those resources are limiting. Different species, even among the oaks, adapt to gradients in resource availability and competition in different ways, including how they allocate their growth to stems, roots and leaves (Loach, 1967; Gottschalk, 1985, 1987; Matsuda and McBride, 1986; Matsuda et al., 1989; Kolb and Steiner, 1990; Latham, 1992; Pallardy and Rhoads, 1993; Walters et al., 1993; Callaway and Mahall, 1996; Rice and Struve, 1997). Seedling populations may stratify by species along resource gradients that occur at very small spatial scales. At a spatial scale that corresponds to the area occupied by an adult tree, the distribution 117

118

Chapter 3

of resources may be highly variable (patchy). But at a smaller spatial scale corresponding to the area occupied by a seedling, the same resources may appear very uniform and either favour or inhibit seedling development (Latham, 1992). This variability in the spatial distribution of regeneration niches has been offered as one explanation for the maintenance of diversity in ecosystems (Grubb, 1977; Latham, 1992), and the coexistence of multiple tree species (Veblen, 1992). For the oaks, the critical regeneration events include ﬂowering, fruiting, seed dispersal and germination, seedling establishment, dieback and sprouting, and the growth of oak reproduction. Moreover, numerous biotic and abiotic factors affect the regeneration niche. An oak’s regeneration niche therefore may not pertain to just one ecological factor, but rather to a suite of factors that change over time and result in a corresponding range of regeneration success probabilities. Over time, it is the coupling of regeneration events with ecological conditions that determines probabilities of regeneration success for oaks.

Regeneration Strategy A species’ regeneration niche, i.e. when and where its regeneration requirements are met, is partially expressed by the mechanisms that have evolved to facilitate its regeneration. However, an advantage associated with a particular mechanism in one environment may be a disadvantage in another environment. In a sense, regeneration is one of the problems that a species solves during its evolution. The resulting solution constitutes the evolutionary compromise that deﬁnes a species’ regeneration strategy.1 Strategies pertaining to oaks can be considered in relation to reproductive mechanisms, the accumulation and ﬂuctuation of populations of reproduction, and responses to site and forest disturbance. 1This

Reproductive mechanisms: seeding and sprouting Seed production and sprouting are two reproductive mechanisms, or tactics, employed by oaks in their regeneration strategy. Although all oaks rely to some extent on both seeding and sprouting, they differ in their dependence on one mechanism versus the other. Even within a species, regeneration tactics may vary among habitats and disturbance regimes. The large number of oak species, their wide distribution and the disparate habitats they occupy require that regeneration strategies among the oaks vary substantially. At one extreme, oaks of the arid southwestern United States may regenerate almost exclusively by sprouting. Among those species, some reproduce largely from root-like lignotubers and rhizomes (Tiedemann et al., 1987). One such species is Gambel oak, whose seedlings are often only a minor source of reproduction (Harper et al., 1985). Hinckley oak, a shrubby rhizomatous oak restricted to a natural range of a fraction of an acre in western Texas, is known to regenerate only from rhizomes (Muller, 1951). Similarly, the shrub oaks of the ﬁre-prone southern California chaparral depend heavily on their ability to sprout after burning. But even when regeneration is largely dependent on sprouting, some new seedlings eventually must be produced to replace the trees and rootstocks that are inevitably lost to mortality, if a species is to persist. In the Ozark Highlands of southern Missouri, oak regeneration also largely depends on sprouting (Liming and Johnston, 1944; Johnson, 1979; Dey et al., 1996a). These forests occur on relatively droughty sites and are often dominated by some combination of black, white, scarlet, post, southern red and blackjack oaks. Under mature closed-canopy stands, the total density of oak reproduction often varies from 1000 to 3000 seedlings and seedling sprouts per acre. After a good

and related anthropomorphisms are used for conceptual convenience and should not be interpreted as implying that trees ‘plan’ their evolution or exercise choices in the sense that humans do.

Regeneration Ecology II: Population Dynamics

acorn crop, 150–300 new oak seedlings per acre may become established (Sander, 1979a). But the continued dominance of the oaks largely depends on a relatively small number of oak seedling sprouts (200–400 per acre) with large root systems capable of supporting rapid shoot growth after disturbances that release growing space formerly occupied by trees in the overstorey (Johnson, 1979; Sander et al., 1984). Except on the most mesic habitats within the Ozark Highlands, non-oaks seldom successionally displace oaks. Throughout much of the eastern United States, northern red oak occupies the mesic middle ground between wet and dry extremes in soil moisture. Northern red oak expresses its relatively ﬂexible regeneration strategy by producing thousands of new seedlings per acre following bumper acorn crops (Johnson, 1974). In northern red oak forests in France, the number of oak seedlings and seedling sprouts growing beneath the forest canopy may exceed 100,000 per acre (Steiner et al., 1993). Northern red oak also has ﬂexible seedbed requirements (Crow, 1992), potentially rapid shoot growth (Farmer, 1975), the ability to regenerate from seedlings established after ﬁnal harvest (Johnson et al., 1989), moderate shade tolerance (Sander, 1990), and the capacity to sprout from large stumps of overstorey trees (P.S. Johnson, 1975; Wendel, 1975; Weigel and Johnson, 1998). But unlike the more xerophytic oaks of the Missouri Ozarks, the relatively mesophytic red oak frequently fails to regenerate because of its vulnerability to successional displacement by the more shade tolerant species with which it typically co-occurs (Johnson, 1976; Crow, 1988; Loftis, 1990a; Nowacki et al., 1990; Lorimer, 1993). In bottomland forests of southern United States, seeding appears to be a more important regeneration tactic than in other oak-dominated ecosystems in North America. For example, water oak seedlings established after ﬁnal harvest can become important members of the succeeding stand (Golden and Loewenstein, 1991; Loewenstein, 1992; Loewenstein and Golden, 1995). But the bottomland oaks

119

also can sprout proliﬁcally, and this facilitates their development in stands disturbed by ﬁre and grazing (Aust et al., 1985). Being neither obligate seeders nor sprouters, these oaks have very ﬂexible regeneration strategies. The relative importance of seeding as a species’ regeneration tactic is sometimes revealed by the number of seedlings that become established after a heavy acorn crop. For example, more than 100,000 Nuttall oaks per acre became established after a bumper acorn crop in one bottomland forest (R.L. Johnson, 1975). If only 0.1% of those seedlings (100 per acre) were competitively successful and well distributed, they would occupy most of the available growing space within two decades following a disturbance that removed the overstorey. Nuttall oak’s seeding strategy is complemented by its rapid height growth (Johnson, 1981), which is necessary for it to successfully compete with the fast-growing and persistent competitors occurring on bottomlands. A species’ seeding strategy also may be reinforced by the dispersal of acorns by rodents and jays to habitats favourable to seed protection, germination, and seedling growth (McQuilkin, 1983; Harrison and Werner, 1984; Sork, 1984; Johnson and Webb, 1989). Nevertheless oaks often fail to regenerate in southern bottomland forests. Competing green ash, sweetgum, and other tree reproduction may outgrow and suppress oak reproduction shortly after standinitiating disturbances occur (Johnson and Krinard, 1983; Aust et al., 1985). Moreover, periods of high oak reproduction density in bottomland forests are often followed by periods of low seedling density because of low seedling survival rates and infrequent acorn crops (R.L. Johnson, 1975). Prolonged periods with little or no oak advance reproduction thus frequently occur. Consequently, domination of bottomland forests by oaks is often limited to one generation. Bottomland oaks nevertheless possess characteristics that, under certain conditions, favour their regeneration over associated species. For example, water oak can

120

Chapter 3

persist in ﬂoodplain forests because, unlike its non-oak competitors, its acorns often germinate after ﬂooding occurs. Oak seedlings also avert ﬂood-caused uprooting because of their long taproots and ability to resprout after their tops are damaged (Streng et al., 1989). Because of this persistence, oak reproduction is often relatively well represented in the older and larger size classes of the total reproduction complex. In an East Texas ﬂoodplain forest, water oak seedlings ranged from 5% of all 1-year-old seedlings to 32% of all 5-yearold seedlings (Fig. 3.1). Like oak reproduction in uplands, the older and larger reproduction in bottomlands has the highest probability of capturing growing space after a canopy disturbance. These anomalies in regeneration strategy among the oaks emphasize the difﬁculty of generalizing their regeneration ecology across species and habitats. Different oaks have solved their regeneration problems in different ways. Some species are more ﬂexible than others (Fig. 3.2). In turn, each species’ environment, physiology and genetics shapes its ﬂexibility in regeneration tactic.

100

2,574

1,218

698

Successful regeneration of oaks partly depends on the species that oaks must compete with. The regeneration strategies of competing species may differ greatly from the oaks. As a result, non-oaks may be competitively advantaged or disadvantaged, depending on their adaptations to the particular environment. In addition to basal sprouting and producing seedlings from current seed, non-oaks also may regenerate from root sprouts (e.g. aspens, beech, sassafras and sweetgum) and from seed stored in the forest ﬂoor (e.g. black cherry, white and green ashes, and yellow-poplar). The seeds of some species such as pin cherry and briars, unlike the oaks, remain buried in the forest ﬂoor and soil where they can accumulate for decades (Marks, 1974; Whitney, 1986). Seeds of other proliﬁc seed producers, such as yellow-poplar, black cherry and sassafras, are shorter-lived but nevertheless may be present in high numbers (Wendel, 1977). When environmental conditions are favourable, such as after clearcutting or burning, stored seeds of non-oaks may germinate in enormous numbers and produce dense populations that often outgrow co-occurring oaks. 267

110 Water oak Sweetgum Ironwood Red maple American elm Other

Seedlings (%)

80

60

40

20

0 1

2

3 4 Seedling age (years)

5

Fig. 3.1. Species composition of reproduction under an east Texas floodplain forest in relation to seedling age. Numbers of trees in each age class are shown above bars. Other species include blackgum, deciduous holly, American holly and Sebastian bush. (Adapted from Streng et al., 1989, with permission.)

Regeneration Ecology II: Population Dynamics

121

100

Relative dependence on seeding

Nuttall oak 80 Northern red oak 60 White oak 40

Black oak

20 Post oak Gambel oak 0 0

20

40

60

80

100

Relative dependence on sprouting Fig. 3.2. Conceptualized dependence of six oak species on seeding versus sprouting as a regeneration tactic. Diameters of circles are proportionate to each species’ postulated flexibility in selecting the alternative tactic.

After heavy thinning or complete overstorey removal, competition from non-oak root sprouts and stump sprouts may suppress and kill established oak reproduction (Beck, 1970; Beck and Hooper, 1986). Grasses, sedges, ferns, vines and shrubs also can seriously interfere with the regeneration of oaks (Jarvis, 1964; Beck, 1970; Hanson and Dixon, 1985; Bowersox and McCormick, 1987; Horsley, 1991; Smith and Vankat, 1991; Johnson 1992). In the oak savannas and woodlands of Oregon and California, competition from exotic annual grasses that have replaced native perennial bunchgrasses is a probable cause of oak regeneration failures (Gordon et al., 1989; Welker and Menke, 1990; Barnhart et al., 1991; Danielsen and Halvorson, 1991; Pavlik et al., 1991; Riegel et al., 1992). Whether or not oaks regenerate successfully thus depends in no small part on the competition environment.

Accumulation of oak reproduction In many oak-dominated ecosystems, the oak reproduction beneath the parent

stand includes seedlings and seedling sprouts that originate from several acorn crops. This accumulation of reproduction results from the combined effects of periodic seed production, the relatively large food reserves in acorns that sustain seedlings through the ﬁrst year, the high sprouting capacity of seedlings, drought tolerance and the ability of seedlings to persist under at least moderate shade. A small proportion of the seedlings originating from a single acorn crop (cohort) may survive for several decades. Silviculturists sometimes refer to the resulting accumulation of seedlings and seedling sprouts as advance reproduction because, in the management of even-aged forests, it is present in advance of ﬁnal harvest. Its presence and development largely determines the importance of oak after the occurrence of natural or human events that greatly reduce parent stand density. The capacity of oak reproduction to accumulate over several acorn crops may compensate, in part, for the inability of acorns to remain viable for more than a few months during a single dormant season.

122

Chapter 3

The term accumulation, as used here, is not meant to imply a continuous, unending increase in oak reproduction density with time. Rather, it refers to the episodic addition of new seedlings to one or more established cohorts. The total population of seedlings and seedling sprouts beneath an oak stand responds to the prevailing overstorey condition, which is continually changing and affecting established oak reproduction and other understorey vegetation. However, the changes in oak reproduction occur much faster than in the parent stand. Consequently, relatively short cycles of the birth, growth and death of individual cohorts of oak reproduction are embedded in longer-term changes in overstorey species composition and size structure. An accumulated population of oak reproduction can sometimes capture all or some of the growing space in canopy gaps or larger openings. Its success in doing so usually depends on the presence of seedling sprouts with relatively large roots. True seedlings and small seedling sprouts usually have root systems that are too small to support rapid shoot growth (Sander, 1971). And stump sprouts from overstorey trees, by themselves, may not be numerous enough to capture all of the available growing space (Sander et al., 1984). Nevertheless, it is the number, size and spatial distribution of all three classes of reproduction that express the total oak regeneration potential of a stand (Sander et al., 1984). Because this potential is implicit in the advance reproduction and the parent stand itself, the new stand is essentially a ‘legacy’ (sensu Franklin et al., 1989) of the parent stand. Sustaining oak-dominated forests thus largely depends on perpetuating propagules from one generation to the next. Whether or not the overstorey is eliminated in small gaps or over large areas, the resulting spatial units of reproduction are even-aged.2 Although oaks have not always been regarded as species well adapted to 2However,

capturing small canopy gaps (Ehrenfeld, 1980; Crow, 1988), they have the potential to do so if advance reproduction of sufﬁcient size is present. For example, 10-ft tall advance reproduction of northern red oak captured canopy gaps as small as 1/25 acre in a mixed sugar maple–oak stand in southwestern Wisconsin (Lorimer, 1983). This is the approximate area occupied by the crowns of ﬁve dominant or co-dominant northern red oaks 15 inches dbh in a fully stocked stand. The accumulation of oak advance reproduction is most obvious in xeric uplands where recurrent dieback and resprouting of reproduction produce multiple stems of varying size, some of which are attached to large root systems. Large roots may be several decades older than their living stems. The true age of a seedling sprout thus is recorded in the rings of its taproot, and not its stem, which is set back to age zero with each complete dieback.3 In a southeastern Ohio oak–hickory stand, the roots of oak reproduction were up to 32 years older than the stems (Merz and Boyce, 1956). In West Virginia, root ages of northern red and white oaks were up to 8 and 10 years older, respectively, than their stems (Fig. 3.3). Accordingly, the range of ages of the roots of advance reproduction and the number per acre of older living roots is an indicator of an oak stand’s capacity to accumulate oak reproduction.

Variation in the accumulation process In the oak forests of the eastern United States, the accumulation of oak reproduction generally increases with decreasing site quality and overstorey density. For oak forests in West Virginia, the relation is represented by the decreasing average age of root systems of oak reproduction (age measured at the root collar) with increasing overstorey density and site index (Fig. 3.4A). Accumulation of oaks tends to be

if their age structure were based on age from germination (‘age from birth’), they could be uneven-aged, depending on the length of the reproduction accumulation period. 3Roots of oaks are difficult to age because their annual rings are not easily distinguishable even through a light microscope. Accurate age determination therefore usually requires special techniques such as Xradiography (Renton et al., 1974; Powell, 1976).

Regeneration Ecology II: Population Dynamics

A

20

0.8 0.6

16

0.4

12 8

Mean root age (years)

24

Root age (years)

40

Root diameter (inches)

30

20

Site index

A

35 50 65

80 10

0.2 4

True seedlings (1:1)

0 60

0 0 24

4

8

12

16

70 80 Crown cover (%)

90

B 0.8 0.6

16 12 8 0.4 4 0.2

True seedlings (1:1)

0.8 Root diameter (inches)

20 Root age (years)

123

Site index

B

35 0.6 50 0.4 65

0 0

4

8 12 Stem age (years)

80

16

Fig. 3.3. Relation of root age to stem age and root diameter for (A) white oak, and (B) northern red oak reproduction in a random sample of 56 stems of each species beneath West Virginia oak stands. Based on the linear regression models: (A) Root age = 1.97 + 0.671*SA + 10.0*RD; R 2 = 0.86. (B) Root age = 0.199 + 0.735*SA + 8.98*RD; R 2 = 0.90. In both equations, SA = stem age and RD = root diameter. Root diameters were measured just below the root collar. Estimates are shown for the approximate range of observed values of stem age and root diameter. (Adapted from Powell, 1976, used with permission.)

greatest on the drier sites and where overstorey density is low. Root diameter is positively correlated with root age (when both are measured at the root collar) so the accumulation process is similarly evident in the relation of reproduction basal diameter to site quality and overstorey density (Fig. 3.4B). However, oaks in arid and semi-arid regions behave differently. Because of

0.2

60

70 80 Crown cover (%)

90

Fig. 3.4. Mean age (A) and diameter (B) of roots of oak reproduction in relation to overstorey crown cover and site index in West Virginia forests. Ages and diameters were measured just below the root collar on a random sample of 46 plots. Estimates are based on linear regression models: (A) Root age = 63.8 – 0.391*CC – 0.205*SI – 0.241*SLP; R 2 = 0.74. (B) Root diameter = 1.99 – 0.0172*CC–0.0185*SI + 0.000185*CC*SI – 0.00285*SLP; R 2 = 0.80. For both models, CC = crown cover (%), SI = oak site index (ft at an index age of 50 years), SLP = slope percentage. For this graph slope percentage is held constant at 25, and each data point represents seven to ten trees. Estimates are shown for the approximate range of observed values of crown cover and site index. The estimates represent the average of the sample of scarlet, chestnut, black, white and northern red oak reproduction. (Adapted from Matney, 1974, used with permission.)

124

Chapter 3

extreme heat, solar radiation and moisture deﬁciency, establishment and survival of oak reproduction is often favoured on the less severe sites such as north-facing slopes and under partial shade rather than in canopy gaps (Callaway and D’Antonio, 1991; Callaway et al., 1991; Williams et al., 1991; Callaway, 1992; Keeley, 1992). Wherever the accumulation of oak reproduction occurs, it results from survival rates sufﬁcient for reproduction to build up over several successive acorn crops. Because the size and age of oak advance reproduction are positively correlated, properties of the accumulation process are sometimes implicit in the size distribution of the reproduction. The regeneration dynamics of black oak–white oak/Vaccinium forests in northern Lower Michigan (Host and Pregitzer, 1991) (Fig. 1.8) provide an example. These forests occur on droughty outwash sands where the site index for black oak ranges from about 50 to 60 ft at an index age of 50 years (Cleland et al., 1993). The density of oak reproduction averages about 11,000 seedlings or seedling sprouts per acre (Johnson, 1992). However, the frequency of occurrence of black oak and white oak reproduction in successively larger basal diameter classes decreases at an exponential rate (Fig. 3.5). This relation suggests that only a fraction of the reproduction in each size class survives to become a member of the next larger size class. The recruitment or ingrowth of oak reproduction into successively larger size classes involves three population processes that jointly determine the accumulation rate: (i) the periodic establishment of new seedlings (seedling input); (ii) growth rate; and (iii) survival rate. If, in a given ecosystem, all three processes are in equilibrium so that seedling input balances seedling mortality and seedling growth remains relatively constant, then the associated size and age distribution of survivors should remain constant. However, constancy is an unlikely attribute of an oak forest. Variable weather, ﬂuctuating acorn production, ﬂuctuating populations of acorn consumers, changes

in stand composition and structure that affect acorn production, seedling growth and seedling survival are some of the factors that rule against constancy. The diameter distribution of oak reproduction at one point in time nevertheless reﬂects, in some way, the recent history of seedling input, drain and growth in a given stand. Population curves like those in Fig. 3.5 therefore might be interpreted as representing a transient recruitment rate of oak reproduction into successively larger size classes under recently prevailing conditions. Such rates accordingly would reﬂect the nature of the accumulation process. A transient recruitment rate can be estimated by ﬁtting certain mathematical functions to observed ﬁeld data. The negative exponential function has been widely used to express the constant exponential decrease in numbers of individuals within a cohort per unit time. For the black oak curve in Fig. 3.5, the slope coefﬁcient of the function (3.1) deﬁnes the rate of change in oak seedling and seedling sprout density in relation to increasing basal diameter. However, a more useful expression of the curve’s slope is given by the constant negative exponential rate, k, where k = eb

[3.1]

e is the base of the natural logarithm, and b is the curve’s slope coefﬁcient. Equation 3.1 thus restates b as a negative exponential rate of depletion for increasing basal diameter classes. Accordingly, k corresponds to the survival rate of seedlings or seedling sprouts per unit of increase in diameter (as opposed to change over time). Therefore k represents the probability of a seedling or seedling sprout surviving long enough to attain a basal diameter of 1 inch. For the black oak data, k is approximately 0.000001. However, because 1 inch is beyond the observed data range, it may be more meaningful to derive a survival rate per 0.1 inch of basal diameter increase. If we call this rate k0.1, then k0.1 = eb/10

[3.2]

which yields a rate of 0.27 per 0.1 inch.

Regeneration Ecology II: Population Dynamics

125

Overstorey trees (no./acre)

Seedlings and seedling sprouts per acre

2800 2400 2000 1600 1200

20 16 12 8 4 0 2

800

6 10 14 18 22 Dbh (inches)

400 0 0.0

0.1

0.2

0.3

0.4

0.5

Basal diameter (inches)

White oak

Black oak

Fig. 3.5. The density of black oak and white oak seedlings and seedling sprouts in relation to their basal diameter under a relatively undisturbed stand on a droughty outwash sand in northern Lower Michigan (black oak–white oak/Vaccinium type as defined by Host and Pregitzer, 1991). The black oak curve is given by the negative exponential model: N = 6581.96(e–13.107d ) where N is number of seedlings and seedling sprouts per acre, e is the base of the natural logarithm, and d is the diameter of seedlings or seedling sprouts measured in inches at the ground line. The white oak curve is given by the power function model: N = 3.81d –1.886 The inset graph shows the diameter distribution of the overstorey, which is at 96% stocking based on Gingrich’s (1967) stocking equation. (Authors’ data.)

If current stand conditions were sustained, we would expect about 27% of the population of black oak reproduction to survive to be recruited into the next larger 0.1 inch basal diameter class, 7% (0.272) into the second larger 0.1 inch class, and 0.5% (0.273) in the third larger class. For larger diameter classes, recruitment drops to 0.1% or less. In contrast to black oak’s constant recruitment rate, white oak’s rate continually changes with increasing basal diameter, and this relationship is better described mathematically by the power function (Fig. 3.5). The probability that a white oak seedling or seedling sprout of a given initial size survives to grow into the next larger size class increases as size increases. Assuming that rates are essen-

tially constant over short intervals of basal diameter, the recruitment rate from the 0.1 to the 0.2 inch diameter class is about 18% whereas the rate from the 0.2 to 0.3 inch class is 30% (Table 3.1). The magnitude of increase in recruitment rates with increasing size of reproduction (Table 3.1) suggests that, for a given number of initially established seedlings, white oak is a more aggressive accumulator of seedlings than is black oak in this ecosystem. This is consistent with white oak’s greater shade tolerance, which is also evident in the differences between the overstorey diameter distributions of the two species (Fig. 3.5 inset). The relatively large numbers of white oaks in the small overstorey diameter classes largely

126

Chapter 3

Table 3.1. The estimated probability (P) that a seedling or seedling sprout of white oak advance reproduction of an initial basal diameter survives to a specified future basal diameter class in black oak–white oak/Vaccinium forests in northern Lower Michigan.a Future basal diameter class (inches) Initial basal diameter (inches)

0.2

0.3

0.1 0.2 0.3 0.4

0.18 – – –

0.16 0.30 – –

0.4

0.5

0.15 0.27 0.41 –

0.14 0.24 0.37 0.52

P

aBased on fitting the observed frequency distribution of reproduction by basal diameter classes at one point in time to the power function equation: N = 3.81D1.886, where N is the estimated number of seedlings and seedling sprouts per acre, and D is basal diameter in inches at the ground line for a seedling or for the largest stem for rootstocks with multiple stems. P is derived from Equation 3.2. The values of P shown assume that rates of change within 0.1 inch intervals of basal diameter are constant. (Authors’ data.)

represent suppressed trees. Where a canopy gap occurs, larger seedling sprouts may be recruited into the overstorey. The lower end of the overstorey diameter distribution therefore may include both recently recruited trees, and trees declining in growth and vigour that are succumbing to inter-tree competition. The paucity of black oak in the smaller diameter classes reﬂects its shade intolerance and inability to endure in the sub-canopy, and its low probability of recruitment into the overstorey. Where overstorey density remains uniformly high, the reproduction of both species remains in a largely suppressed state. To remain in that state, the oak reproduction must be continually renewed in endless short cycles of seedling input and mortality. In conceptualizing the reproduction accumulation process, it is convenient to label ecosystems where oak reproduction has a natural propensity to accumulate over long periods as intrinsic accumulators and those with little such propensity as recalcitrant accumulators. Oak reproduction probably accumulates, at least at times, over several acorn crops wherever oaks naturally occur. Accumulation is facilitated in part by the relatively large food reserves of acorns that sustain

seedlings, even when they grow under low light, through the critical ﬁrst year of establishment. During germination, roots emerge before shoots (Fig. 2.15). Rapid taproot growth, deep penetration, recurrent shoot dieback and resprouting favour root development over shoot development and thus reinforce the accumulation process. The accumulation of oak reproduction even in mesic and wet forests therefore may be more pronounced than that of associated non-oaks (e.g. in the bottomland oak forests of Texas) (Fig. 3.1). A more rigorously deﬁned classiﬁcation of accumulator types incorporates the known or assumed capacity of oak advance reproduction to replace the parent stand and its relative constancy in doing so. In the eastern United States, a high capacity for replacement of overstorey oaks by oak reproduction occurs most frequently in the drier ecosystems. There, the accumulation of oak reproduction is often intrinsic and relatively independent of forces external to the ecosystem itself. Oaks in such ecosystems tend to be self-replacing, and therefore are not readily displaced successionally by non-oaks. These ecosystems therefore are relatively resilient, i.e. they tend to return quickly to their predisturbance state after disturbance.

Regeneration Ecology II: Population Dynamics

One of the largest North American ecosystems of this type is the Ozark Highlands of Missouri, where oak reproduction characteristically accumulates in the understorey for decades, even in the absence of large-scale disturbances. The accumulation of oak reproduction is nevertheless strongly correlated with topographic features such as slope position and aspect (Sander et al., 1984; Dey, 1991; Dey et al., 1996b), which collectively inﬂuence light, heat and soil moisture. Oak forests that similarly accumulate oak reproduction occur on the drier sites in the Ohio Valley (Minckler and Woerheide, 1965), the Appalachians (Trimble, 1973; Ross et al., 1986) and the Lake States (Johnson, 1966, 1992; Arend and Scholz, 1969). Ecosystems that are intrinsic accumulators do not necessarily accumulate large numbers of oak reproduction. For example, about 1200 seedlings and seedling sprouts per acre characterize some Ozark oak stands (Sander, 1979b). Only 150–300 new oak seedlings per acre may become established, even after a good acorn crop (Sander, 1979b; McQuilkin, 1983). In contrast, more than 100,000 Nuttall oak seedlings per acre may become established after a single bumper acorn crop in bottomland forests that are recalcitrant accumulators (R.L. Johnson, 1975). Thus, large numbers of oak seedlings by themselves may not identify ecosystems that intrinsically accumulate oak reproduction. These anomalies emphasize the degree to which regeneration processes can vary among different kinds of oak forests. Recalcitrant accumulation of oak reproduction is characteristic of mesic and hydric ecosystems. There, oak reproduction may accumulate only after prolonged or recurrent disturbances spanning successive acorn crops. In the absence of disturbance, these oak forests usually are successionally displaced by non-oaks (Carvell and Tryon, 1961; Trimble, 1973; R.L. Johnson, 1975; Loftis, 1990a; Nowacki et al., 1990; Will-Wolf, 1991; Abrams and Nowacki, 1992). In the absence of disturbance, a dense overstorey with a subcanopy of small trees often develops

127

(Braun, 1972; Loftis, 1990b). Vertical stratiﬁcation of tree crowns together with high total stand density creates extremely low light intensities on the forest ﬂoor. Nevertheless, reported oak reproduction densities in North American forests exceed 50,000 seedlings and seedling sprouts per acre in mesic habitats (Tryon and Carvell, 1958) and 100,000 per acre in bottomlands (R.L. Johnson, 1975). But such ecosystems also may be frequently depauperate of oak reproduction (Sander, 1983). When oak reproduction does occur, few age cohorts are represented because of low survival rates. Most of the seedlings from a single cohort often die before the next good acorn crop occurs because respiration rates of seedlings growing in deep shade often exceed their photosynthetic rates (Hanson et al., 1987; Dey and Parker, 1996). Microsites where oak seedlings become established in large numbers may not be where they ultimately survive (Johnson, 1966; Harrison and Werner, 1984). Whereas the cool, moist microclimate of northeastfacing slopes may favour initial establishment, the more favourable places for long-term survival, root development and thus accumulation occur on the more southerly or neutral aspects (Carvell and Tryon, 1961; Sander et al., 1984; Walters, 1990). In eastern forests, there is a general inverse relation between site quality and regeneration success: the better the site the greater the competition and the more difﬁcult it is to regenerate oaks (Arend and Scholz, 1969; Trimble, 1973; Lorimer, 1989; Loftis, 1990b). The accumulation of oak reproduction in relation to combined site and overstorey factors is indirectly illustrated by the demographics of reproduction density in upland forests of southeastern Ohio (Walters, 1990). There, forests are comprised of black, white, scarlet, northern red and chestnut oaks mixed with maples, American beech, yellow-poplar and other hardwoods. The region is characterized by heavily dissected low hills with relatively short slopes (200–600 ft long) and slope gradients ranging from less than 10 to over 50%. Site quality is highly variable and

Chapter 3

In the same region, the combined density of maple, American beech, yellow-poplar and other non-oak reproduction differs markedly from the oaks. The non-oaks attain highest densities on north-facing slopes and lowest densities on south-facing slopes. Maximum and minimum densities of the non-oaks tend to occur at 1° and 181° azimuth, respectively (Fig. 3.6B). But like the oaks, non-oak densities change with changing slope gradient and overstorey density. For a given aspect, highest densities occur on steep slopes and under low overstorey densities. Collectively, these landscape-level models illustrate important

A 1.5

10% slopes

1.0 30% slopes 360 270 180 90

re

es

)

0.5

h

ut im Az

80 70 Over 60 0 store 50 40 y sto cking (%)

(d

eg

0.0

B

30% slopes

4 3

10% slopes

80 70 Over 60 0 store 50 40 y sto cking (%)

re

0

360 270 180 90

eg

1

es )

2

(d

Seedlings and seedling sprouts (thousands/acre)

Seedlings and seedling sprouts (thousands/acre)

depends on soil and topographic factors. Oak reproduction densities tend to be greatest on hot southwesterly facing slopes and least on cool northeasterly facing slopes (Fig. 3.6). Maximum and minimum densities of oak reproduction greater than 1 ft tall occur at azimuths of 203° and 23°, respectively. However, azimuth effects change with slope gradient and overstorey density. On a given aspect, oak reproduction density is greatest on gentle slopes and decreases with increasing slope gradient. Moreover, for any given combination of slope and azimuth, reproduction density decreases with increasing overstorey density.

Az im ut h

128

Fig. 3.6. Estimated densities of oak (A) and non-oak (B) reproduction beneath stands in southeastern Ohio in relation to overstorey stocking, slope aspect (degrees azimuth) and slope gradient (%). Includes reproduction at least 1 ft tall and up to 1 inch dbh. Oaks include black, white, scarlet, northern red and chestnut oaks. Non-oaks include red maple, yellow-poplar, hickories, sugar maple, American beech, white ash, American basswood, yellow buckeye, sycamore, black gum, black walnut, elms and aspen. The models explain 25% and 51% of the variation in oak and non-oak reproduction density, respectively. Stocking was defined by stocking (relative density) equations. (Adapted from Walters, 1990, used with permission.)

Regeneration Ecology II: Population Dynamics

differences between the demographics of oak and non-oak reproduction. The environmental conditions that favour successful oak regeneration, and thus the accumulation of oak reproduction, can be conceptualized as a regeneration window (Fig. 3.7). The window is relatively narrow in mesic sites where the slow initial growth of oak reproduction must be matched by a light intensity sufﬁcient for seedling growth but low enough to discourage the development of competitors. At the expense of the oaks, which are intermediate in shade tolerance, a lower light intensity favours tolerant species and a higher light intensity favours intolerant species. As site becomes drier, the window expands in response to the drought tolerance of the oaks that occur there and the concomitant exclusion of long-lived competitors that are less tolerant of drought. Reduced stand density also tends to create warmer and drier microclimates than those prevailing in the undisturbed forest

129

(Dey and Parker, 1996). ‘Xerifying’ disturbances such as ﬁre, grazing and insect defoliation may consequently widen the oak’s regeneration window. The regeneration window emphasizes the gradational nature of the oak regeneration process with respect to light and moisture. The propensity of ecosystems to accumulate oak reproduction is similarly gradational. Accordingly, it is useful to consider an intermediate category: ambivalent accumulators of oak reproduction. These can be deﬁned as ecosystems with uncertain or pliant tendencies towards accumulating oak reproduction. Such ecosystems would be expected to lie near the centre of the regeneration window (xero-mesic ecosystems). There, even minor disturbances could produce pronounced shifts in the accumulation of oak reproduction. Identifying these forests is important because they are likely to be especially amenable to silvicultural manipulation (Bakken and Cook, 1998).

Open

Light gradient

The oak regeneration window

Closed canopy Moist

Dry Moisture gradient

Fig. 3.7. The ‘regeneration window’ for oaks in eastern United States forests in relation to light and soil moisture. The window (open area) defines the region most favourable for successful oak regeneration. This region is relatively narrow on moist sites but widens with increasing dryness. On moist sites, intermediate light intensities offer the best opportunities for oak seedling survival and growth. Lower light intensities are insufficient to meet the oak’s minimum light requirements for photosynthesis but are sufficiently high to inhibit the development of many competitors. As site conditions become drier, the window widens because of the oak’s drought tolerance and the exclusion of competitors that are less drought resistant. (Adapted from Hodges and Gardiner, 1993.)

130

Chapter 3

Categorizing ecosystems by their propensity to accumulate oak reproduction can be incorporated into ecological classiﬁcation systems to characterize oak’s replacement potential within each ecological unit. Population characteristics useful for deriving such ratings include the expected number and size distribution of oak reproduction (basal diameters and/or heights), and the relative constancy of those characteristics over time within an ecological classiﬁcation unit. Most classiﬁcation systems generally describe reproduction characteristics and the successional status of oaks within classiﬁcation units (e.g. Smalley, 1978, 1984; Hix, 1988; Kotar et al., 1988; Cleland et al., 1993). To date, none speciﬁcally considers the replacement potential of oaks based on observed reproduction characteristics and associated successional relations, but oak accumulation types can sometimes be inferred from existing descriptions of the ecological units (Table 3.2).

Disturbance effects The presence of established oaks in the overstorey of a stand is often attributed to past stand disturbances. This is especially true in stands that are recalcitrant accumulators of oak reproduction even though there is often little or no evidence of previous disturbances that might explain the origin of the oaks. In the absence of future disturbances of the appropriate frequency and intensity, such oak stands are destined to successional replacement by other species. Past forest disturbances are often recorded in the spatial variability of overstorey density and size structure. In turn, variability in the overstorey usually is associated with variability in oak reproduction density. This is the case in xeric oak forests on outwash sands in northern Lower Michigan. There, more than half of the spatial variation in the density of white oak and black oak reproduction is explained by variation in overstorey density and the basal area of large oaks (Johnson, 1992) (Fig. 3.8). In this example, large trees are deﬁned as those at least 12 inches dbh

(white oak) or 14 inches dbh (black oak) and are of a size generally considered to be better acorn producers than small trees (Downs, 1944). The reproduction density of both white and black oak increases as the basal area of large trees increases. Reproduction of these two species also differs in reaction to overstorey density (Fig. 3.8). Whereas white oak reproduction density decreases with increasing overstorey density, black oak reproduction density gradually increases with increasing overstorey basal area before declining at about 85 ft2 acre−1. Moreover, the two species differ in their apparent sensitivity to overstorey density. The rapid decrease in white oak reproduction density per unit increase in overstorey density contrasts with relatively small changes in black oak reproduction density with change in overstorey density. The two species similarly differ in their sensitivity to the basal area of large diameter trees. White oak reproduction density increases more rapidly per unit basal area of large trees than does black oak. This relation may be associated with white oak’s greater acorn production per unit crown area than black oak’s (Myers, 1978). Consequently, the total maximum density of oak reproduction occurs under conditions that favour white oak reproduction. Despite its regeneration disadvantages, black oak typically maintains a position of dominance in these forests. This persistence may be related to its height and diameter growth, which are greater than that of white oak during the ﬁrst 100 years of stand development (Trimble, 1960; Carmean, 1971). However, other factors, including stand history and site quality also inﬂuence oak reproduction density (Carvell and Tryon, 1961; Arend and Scholz, 1969; Ross et al., 1986; Nowacki et al., 1990; Walters, 1990). Although relations between overstorey density and oak advance reproduction density imply disturbance effects (e.g. Figs 3.7 and 3.8), disturbance effects are directly considered in a predictive model for West Virginia forests (Fig. 3.9). The model expresses disturbance as a subjectively derived index ranging from 0 (none) to 16

Regeneration Ecology II: Population Dynamics

131

Table 3.2. Inferred oak reproduction accumulation types for upland plant associations on the HuronManistee National Forests in Michigan.a See Fig. 1.8 for a schematic illustration of the landforms listed in this table. Plant association Northern pin oak–white oak/ Deschampsia Black oak–white oak/ Vaccinium Mixed oak–red maple/ Trientalis Northern red oak–red maple/ Viburnum Northern red oak–red maple/ Desmodium Sugar maple– beech/ Maianthemum Sugar maple– white ash/ Osmorhiza Sugar maple– white ash/ Caulophyllum

Soil and landform characteristics

Characteristic overstorey/understorey compositionb

Site index (ft)c

Inferred accumulation typed

Excessively well-drained sands of outwash plains

Northern pin, black, and white oaks; jack pine/oaks, black cherry

N. Pin oak: 48

Intrinsic

Excessively well-drained sands of outwash plains

Black, northern pin, and white oaks/red maple, oaks

Black oak: 50–56

Intrinsic

Well to excessively well-drained sands of overwashed moraines and kame terraces Moisture-enriched, well-drained sands of moraines and ice-contact topography

White, black, and northern red oaks; red maple/red maple, witch-hazel Northern red and white oaks; red maple/red maple, black cherry, flowering dogwood, and witch-hazel Northern red and white oaks; red maple/red maple, black cherry, flowering dogwood, and witch-hazel Sugar maple, American beech, northern red oak, red maple/ sugar maple, red maple, American beech Sugar maple, white ash, northern red oak, American basswood, red maple/sugar, American beech Sugar maple, white ash, northern red oak, American basswood, red maple/sugar maple, American beech

N. Red oak: Ambivalent 61–65

Moderately well-drained sandy loams over loamy substrata of ground moraines and glacial lake beds Well-drained morainal medium to fine sands

Moisture- and nutrient-enriched morainal sands Well-drained to moderately well-drained sands over fine loamy material

N. Red oak: Recalcitrant 77

N. Red oak: Recalcitrant 85

N. Red oak: Recalcitrant 76–88

N. Red oak: Recalcitrant 86 N. Red oak: Recalcitrant 76–88

aAdapted

from Cleland and others (1993). composition of relatively undisturbed natural stands. Understorey includes trees and shrubs 1–3 inches dbh. an index age of 50 years. dBased on Cleland and others (1993), Host and Pregitzer (1991), Johnson (1992), and other sources. bCharacteristic cAt

(heavy). It considers intensity of disturbance and time since disturbance events that include burning, grazing and logging (Carvell and Tryon, 1961). Although the proportion of variation in reproduction density is unspeciﬁed, the factors included in the model (light, aspect and disturbance) are statistically signiﬁcant. For a given disturbance intensity, the density of oak reproduction increases with increasing heat associated with slope aspect progressing from northeast (coolest) to southwest (hottest), and increasing light. Historically, ﬁre was the pre-eminent disturbance factor that shaped and maintained environments favouring oaks. The extensive oak savannas the early European settlers encountered in the midwestern United States were products of a centurieslong ﬁre history on a spatial scale unlikely

to be repeated. Although the ecological disturbances that created and sustained those savannas have long disappeared, their effects are still evident in some regions. Within the Forest–Prairie Transition Region of the Midwest, many of today’s closedcanopy oak forests probably originated from the savannas present before settlement by Europeans (Beilmann and Brenner, 1951; Grimm, 1984; Guyette and Cutter, 1991; Abrams, 1992). The historical record indicates that throughout much of North America during the presettlement and early settlement eras, ﬁre created and sustained the conditions necessary for perpetuating oaks where they would not otherwise have occurred (Day, 1953; McClaran and Bartolome, 1989; Johnson, 1993; Lorimer, 1993; Guyette and Dey, 1995).

132

Chapter 3

16

SppBA White oak Black oak

Seedlings and seedling sprouts (thousands/acre)

16 12

13 10

8 7 4 28 16 4

4

0 60

70

80

90

100

Stand basal area (ft2/acre) Fig. 3.8. Estimated densities of black oak and white oak reproduction in relation to total stand basal area and the basal area of large overstorey trees of the same species (SppBA) in xeric oak forests in northern Lower Michigan. The trees comprising SppBA are the presumed primary acorn producers. For white oak, SppBA includes trees ≥12 inches dbh, and for black oak includes trees ≥14 inches dbh. Based on linear regression models that account for 55% and 63% of the observed variation in reproduction density of white oak and black oak, respectively. (Modified from Johnson, 1992, used with permission from Elsevier Science.) 40

50

Moderate disturbance

30

25

20

50

Light (%)

Oak reproduction density (thousands/acre)

Light disturbance

10 25 5 5 0 0 NE

45 N&E

90 SE & NW

135 S

180 SW

Aspect scale and aspect Fig. 3.9. The estimated density of oak seedlings and seedling sprouts beneath mixed oak stands in West Virginia in relation to slope aspect, light and disturbance. Includes white, black, northern red, scarlet and chestnut oak reproduction from 1 ft tall to 0.6 inch dbh. Aspect scale is displayed in degrees departure from northeast (45° azimuth). Light is expressed as the amount measured by a photoelectric cell beneath the overstorey as a percentage of that measured in the open. Disturbance is expressed as a qualitative index based on an arbitrary scale from 0 (none) to 16 (heavy) that considers intensity of and time since grazing, fire and logging; estimates for values of 2 (light disturbance) and 6 (moderate disturbance) are shown. (Adapted from Carvell and Tryon, 1961 by permission of the Society of American Foresters, Bethesda, Maryland. Not for further reproduction.)

Regeneration Ecology II: Population Dynamics

When ﬁres were frequent and intense, open-grown oak savannas were created. Those events produced communities comprised of a few large thick-barked, ﬁreresistant overstorey trees per acre, a unique and diverse grass and forb ﬂora, and abundant oak reproduction largely comprised of seedling sprouts (Curtis, 1959; Haney and Apfelbaum, 1990). Oak savannas, which once dominated the landscape throughout much of the central and eastern states, are today rare plant communities (Haney and Apfelbaum, 1990). Following the cessation of burning and grazing, some savannas developed into closed-canopy forests; many others were converted to pasture and cultivated ﬁelds (Beilmann and Brenner, 1951; Curtis, 1959; Thilenius, 1968; Whitney, 1994). Recurrent ﬁre promotes the accumulation of oak reproduction in various ways. Fire can destroy seed stored in the forest ﬂoor and soil and thereby eliminate or reduce post-ﬁre competition from some shrub, herb and other tree species (Fig. 3.10). However, in some ecosystems, that effect may be offset by the presence of species whose seeds are resistant to ﬁre and whose germination is stimulated by ﬁre and the post-ﬁre environment. The lat-

133

ter include several chaparral species associated with oaks of southwestern United States (Keeley, 1991) and pin cherry in the eastern forests (Wendel, 1990). In other species, seeds may be stored in the humus and soil where they are relatively unaffected by ﬁre. Burning also can reduce overstorey and understorey density (DeSelm et al., 1991), which thereby increases light on the forest ﬂoor and reduces competition for other site resources. Within a species, small-diameter trees are the most vulnerable to ﬁre because their thinner bark offers less insulation against potentially lethal heat (Harmon, 1984; Hengst and Dawson, 1993). Although species vary widely in bark thickness for a given diameter (Fig. 3.11), other physical properties of bark also vary among species. The effectiveness of bark in insulating a tree from heat depends on three physical properties: (i) thermal conductivity (ability to transfer heat); (ii) speciﬁc heat (ability to absorb heat); and (iii) thermal diffusivity (ratio of thermal conductivity to the product of speciﬁc heat and bark density) (Martin, 1962). Consequently, the survival of trees of a given diameter subjected to a given duration and intensity of heat varies among

Fig. 3.10. Fire eliminates many of the oak’s competitors by destroying or reducing seed stored in the forest floor and also by killing established fire-sensitive competitors. (USDA Forest Service, North Central Research Station photograph.)

134

Chapter 3

1.0

Blackgum

Bark thickness (inches)

A 0.8 0.6

Chestnut oak

0.4

Sourwood Virginia pine

0.2

Red maple 0.0 0

4

8

12

16

Bark thickness (inches)

2.0

20 Bur oak

B 1.6

White oak 1.2

Northern red oak

0.8

Sugar maple

0.4 Yellow poplar 0.0 0

10

20

30

40

Dbh (inches) Fig. 3.11. Relation between estimated bark thickness and dbh of trees in eastern United States. (A) Five species in dry pine–oak forests in the southern Appalachians. Shaded circles identify the bark thickness and the correlated dbh needed to assure that 50% of trees are not top-killed by a ‘low-intensity’ fire. The proportion of the variation in bark thickness accounted for by dbh ranged from 69% (Virginia pine) to 96% (blackgum). (Adapted from Harmon, 1984.) (B) Five species in Illinois. The proportion of the variation in bark thickness accounted for by dbh ranged from 56% (northern red oak) to 93% (bur oak). (Adapted from Hengst and Dawson, 1993.)

species. Such differences were observed in pine–oak stands on dry sites in the southern Appalachians where low-intensity ﬁres primarily consumed leaf litter. There, the bark thickness required for 50% survival of a population of trees ranged from less than 0.1 inch to over 0.2 inch at dbh, depending on species (Fig. 3.11). Although chestnut oak is generally considered resistant to ﬁre damage, the amount of bark required for the survival of chestnut oak is greater than that required for some non-oaks such as red maple (generally regarded as ﬁre-sensitive). But because the chestnut oak can grow faster in diameter than red maple (at least on dry sites), the former can more

quickly attain an effective ﬁre-insulating bark thickness (Harmon, 1984). In the Ozark Highlands of southeastern Missouri, oak species differed in their response to ﬁre frequency over a 33- to 35year study period (Huddle and Pallardy, 1995). During that time, experimental plots were burned annually, periodically (every 4 years), or not burned. For plots burned annually or periodically, average survival of post oak (a white oak) was greater than that of red oaks (black, scarlet and southern red oaks combined) after statistically accounting for differences in initial (preburn) diameter between the two species groups. Among all treatments, tree survival

Regeneration Ecology II: Population Dynamics

increased with increasing initial dbh up to 7 or 8 inches regardless of species and treatment. For post oak, ﬁre frequency had no signiﬁcant effect on the relation between survival and initial dbh. In contrast, the relation between survival of red oaks and initial dbh differed signiﬁcantly among the three treatments. Differences in survival generally increased with increasing dbh. Among trees less than 7 inches dbh, lowest survival occurred within periodically burned plots and highest survival occurred within unburned plots. Hickories also responded negatively to burning (Fig. 3.12). Differences among species’ growth characteristics thus may interact with tree size, bark, and other species’ characteristics to determine survival of burned trees. Moreover, ﬁre frequency and pre-burn conditions may further affect these differences. The long-term survival of trees subjected to low-intensity surface ﬁres in the Great Smoky Mountains National Park in eastern Tennessee is largely dependent on tree growth rate and bark thickness. Fire intervals of less than 14 years favour the relatively fast-growing species (pines,

blackgum, sourwood and chestnut oak) over slow-growing species (red maple and hickories) and some fast-growing but relatively thin-barked species such as scarlet oak (Harmon, 1984). Longer intervals favour red maple, hickories and the thinbarked scarlet oak by allowing them to attain a sufﬁcient size and bark thickness to resist ﬁre damage. Moreover, once the more ﬁre-sensitive species reach a safe size, similar future ﬁres are less likely to effectively alter overstorey composition. In mixed white oak–yellow-poplar stands in the North Carolina Piedmont (red oak site index 75–80), responses to highintensity ﬁres illustrate the superior resistance of oaks to top-kill (Maslen, 1989). The burns, described as strip head ﬁres, produced ﬂames 3–10 ft high that consumed most of the litter and caused wounding of the residual stand. Seven years after a single burn, top-kill from burning was observed as well as natural dieback (unrelated to burning) on paired burned and unburned areas. Although all reproduction of all species was top-killed by burning, oak saplings and small poles (trees > 12 ft

1.0

Post oak Red oaks (control)

0.8 Survival probability

135

Red oaks (annual burn)

0.6

Red oaks (periodic burn)

0.4

Hickories (control)

0.2 Hickories (all burns)

0.0 0

2

4 6 Initial dbh (inches)

8

Fig. 3.12. Estimated survival probabilities for oaks and hickories in relation to initial (pre-treatment) dbh and frequency of burning during a 33–35 year period in southeastern Missouri. Periodic burns occurred every 4 years. Species in the red oak group include black, scarlet and southern red oaks; hickories include shagbark, mockernut and black hickories. Survival of post oaks in burned plots did not differ significantly (P < 0.05) from trees in control plots. Probability estimates were derived by logistic regression. (From Huddle and Pallardy, 1996.)

136

Chapter 3

tall but ≤ 6 inches dbh) were more resistant to top-kill by ﬁre than non-oaks of the same size (Fig. 3.13A). In unburned areas, the probability of top-kill among non-oaks was consistently lower than that of oaks across all size classes observed except the largest (Fig. 3.13B). Because virtually all the hardwood species associated with oaks can sprout, top-kill from one or a few single low-intensity burns may not effectively reduce or

eliminate these competitors (Johnson, 1974; Nyland et al., 1983; Walters, 1990). However, through repeated low-intensity burns (Langdon, 1981) and sometimes through single high-intensity burns (Maslen, 1989), differences among species in ﬁre-caused mortality of tops, root systems, dormant buds near the root collar and decay resistance confer a competitive advantage to the oaks (Fig. 3.14). Oaks therefore are ‘ﬁre persistent’ compared to

1.0

Oaks (n = 176) Non-oaks (n = 215)

Probability of no top-kill

A Burned 0.8

0.6

0.4

0.2

0.0

0 1–2 ft

0

0

3–5 ft 6–12 ft >12 ft 2–6 in 7–12 in –<2 in

1.0

Probability of no top-kill

B Unburned

Oaks (n = 193) Non-oaks (n = 210)

0.8

0.6

0.4

0.2

0.0 1–2 ft

3–5 ft 6–12 ft >12 ft 2–6 in 7–12 in –<2 in Tree size class

Fig. 3.13. Probability of observing no top-kill of trees in immature mixed white oak/yellow-poplar stands on mesic sites (red oak site index 75–80) in the North Carolina Piedmont. (A) After a ‘high-intensity’ prescribed burn, and (B) on adjacent unburned areas. Size classes are expressed as heights (ft) for reproduction (trees <2 inches dbh) and diameters (dbh, inches) for overstorey trees. Based on 794 trees observed before and 7 years after burning. On burned areas, top-killed trees include those that later sprouted. On unburned areas, top-kill resulted from natural dieback and mortality. (Adapted from Maslen, 1989.)

Regeneration Ecology II: Population Dynamics

many associated species, which may be equally proliﬁc sprouters. This persistence may be partly related to the concentration of dormant buds near the root collar, which is often located an inch or more below the soil surface where it may be protected from lethally high ﬁre temperatures (Korstian,

1927). Such protection may be facilitated by burial of acorns by rodents (Galford et al., 1991). Barring soil erosion or deposition, the position of the root collar, and thus many dormant buds capable of sprouting, remains ﬁxed at the original acorn burial depth throughout the life of the tree.

1.0 A

137

Oaks (n = 176)

Burned

Survival probability

Non-oaks (n = 215) 0.9

0.8

0.7 1–2 ft

3–5 ft

6–12 ft

>12 ft 2–6 in 7–12 in – < 2 in

1.0

Survival probability

B

Oaks (n = 193) Non-oaks (n = 210)

Unburned

0.9

0.8

0.7 1–2 ft

3–5 ft

6–12 ft

>12 ft 2–6 in 7–12 in – < 2 in Tree size class

Fig. 3.14. Survival probabilities for trees in immature mixed oak stands on mesic sites (red oak site index 75–80) in the North Carolina Piedmont. (A) After a ‘high-intensity’ prescribed burn, and (B) on adjacent unburned areas. Size classes are expressed as heights (ft) for reproduction (trees <2 inches dbh) and diameters (dbh, inches) for overstorey trees. Survivors include trees that resprouted after top-kill by fire. Based on 794 trees observed before and 7 years after burning. (Adapted from Maslen, 1989.)

138

Chapter 3

Increased light on the forest ﬂoor resulting from ﬁre increases rates of photosynthesis and growth of the oak reproduction that survives burning. More light also allocates proportionately more photosynthate to oak roots than shoots (Kolb and Steiner, 1990). Physiological investment in root systems may be further reinforced by ﬁre-induced top-kill, which maintains a high root:shoot ratio and a low shoot mass that might otherwise comprise a greater sink for the relatively modest amount of photosynthate produced by oaks growing under an overstorey. The effectiveness of ﬁre in reducing the height of surviving reproduction is illustrated in a Tennessee study (Fig. 3.15). Burning also increases the number of stems produced by basal sprouting (Thor and Nichols, 1973), which may favour high leaf area:shoot mass ratios that promote root development. In a Wisconsin study, the density of northern pin oak reproduction in canopy gaps was unaffected by a single spring ﬁre, whereas red maple and black cherry reproduction were signiﬁcantly reduced (Reich et al., 1990). Moreover, rates

of photosynthesis after burning were greater for the oak than for associated species. The oak’s higher photosynthetic rates were attributed primarily to increased nitrogen concentrations in leaves, which were maintained throughout the growing season after burning, but not maintained in maple and cherry. Photosynthetic rates of oak reproduction that sprout after artiﬁcial shoot removal (and by extension, top-dieback from burning) also may be greater than in reproduction with intact shoots (Kruger and Reich, 1989). Reduction of leaf litter by burning also facilitates direct contact between acorn and mineral soil, which in some cases may ease the penetration of the radicle into the soil and increase the rate of seedling establishment (Krajicek, 1960; see also Chapter 2). The forest ﬂoor also harbours insects that consume acorns and destroy oak seedlings. Burning can reduce the habitat available for such insects (Wright 1987; Galford et al., 1988) as well as for rodents (Hannah, 1987; Van Lear and Watt, 1993). Overall, ﬁre promotes the accumulation of oak reproduction (Fig. 3.16).

Mean tree height (ft)

10

Control Periodic burns Annual burns

8

6

4

2

0 Southern red oak

Scarlet oak

Post oak

Blackgum

Species Fig. 3.15. Mean heights of tree reproduction on three burn treatments beneath a dry oak forest in the eastern Highland Rim of Tennessee. Annual burns were made every year for 8 years and two periodic burns were made 5 years apart. All burns were made during late winter. Heights were measured on the annual and periodic burns 4 months and 1.5 years, respectively, after the last burn. (Adapted from Thor and Nichols, 1973, used with permission.)

Regeneration Ecology II: Population Dynamics

Conversely, ﬁre can kill small oak reproduction (Johnson, 1974) and depending on timing, can destroy acorns (Auchmoody and Smith, 1993). Korstian (1927) ranked the acorns of ﬁve species of eastern oaks in decreasing order of heat resistance as follows: northern red oak, chestnut oak, black oak, scarlet oak and white oak. This ranking is correlated with the thickness of the acorn shell (pericarp) of each species. However, given the amount of heat released in a typical leaf litter ﬁre, differences among species in heat resistance are relatively minor and may be inconsequen-

139

tial (Korstian, 1927). Low-intensity fall burns in Pennsylvania killed 40–50% of northern red oak acorns (Auchmoody and Smith, 1993). Fire also can seriously wound, degrade and kill overstorey oaks that are the acorn producers (Loomis, 1973, 1974; Rouse, 1986; Maslen, 1989). Variation in the survival and density of oak reproduction after burning also may be related to the presence of other species. The vulnerability of turkey oak to ﬁre in the Florida sandhills is partially dependent on its proximity to longleaf pine (Rebertus et al., 1989). The oak, which is a superior

Recurrent burning

Kills some acorns and small oak seedlings

Reduces . . .

Established understorey competition

Overstorey density

Seedbank of oak competitors

Shoot mass of oak reproduction

Leaf litter

Acorn destroyers and leaf feeders (insects, voles, mice, fungi)

Which increases . . .

Light on the forest floor

Net photosynthesis

Photosynthetic investment in roots of reproduction

Survival

Acorn contact with soil

Seedling establishment

. . . Which favours the accumulation of oak reproduction

Fig. 3.16. Potential effects of fire on the accumulation of oak reproduction.

Acorn viability and germination

140

Chapter 3

competitor to the pine in the absence of ﬁre, is sometimes eliminated by ﬁre when it grows under adult pines. There, high ﬁre temperatures generated by the pine litter may kill turkey oak (Williamson and Black, 1981). Whether an oak survives a ﬁre depends on complex relations among several factors including distance to nearest pine, the size (dbh) of both the oak and nearest pine, and ﬁre intensity. In turn, the distance/size effects result from the combustibility of the pine leaf litter, the depen-

dence of oak sprouting on tree size and other factors. Because of their complexity, simple linear relations cannot describe these effects. Moreover, these factors have a different effect on oak crown-kill from that on post crown-kill sprouting. The probability that an oak survives a ﬁre depends on both its probability of escaping crown-kill and its probability of resprouting given that it is crown-killed. Adding those probabilities produces complex survival responses that differ by ﬁre intensity (Fig. 3.17).

(A)

(B)

Fig. 3.17. The probability of turkey oak survival after burning in the Florida sandhills in relation to turkey oak dbh and distance to nearest longleaf pine. (A) After a ‘mild’ fire. (B) After a ‘hot’ fire. Estimates are based on logistic regression equations. Dbh of pine, also a significant estimator, was held constant at 25 cm. (Redrawn from Rebertus et al., 1989, used with permission.)

Regeneration Ecology II: Population Dynamics

Turkey oak survival rates are lowest for trees of 3–4 inches dbh for both low- and high-intensity ﬁres. Trees of 3–4 inches dbh apparently have neither the sprouting capacity nor the resistance to crown-kill that smaller or larger trees, respectively, have when they are close to pines (Rebertus et al., 1989). However, as distance to nearest pine increases, the dependence of oak survival on dbh decreases. Distance to nearest pine is irrelevant in the ‘hot’ ﬁre (Fig. 3.17B). However, in the ‘mild’ ﬁre, the model indicates that oak survival probabilities are lowest for small diameter trees at a distance of about 60 ft (18 m) from a pine and increase in either direction from that distance (Fig. 3.17A). That effect may be related to delayed mortality of oaks weakened but not initially killed by ﬁre, and the consequently poor vigour of the sprouts from those trees. Many trees apparently weakened but not killed by ﬁre ultimately die. In contrast, oaks close to pines are topkilled by ﬁre, while those somewhat further away and thus only moderately affected may gradually deteriorate and ﬁnally die from reduced carbohydrate reserves; those furthest away are often unaffected. These relations emphasize the joint inﬂuence of the initial (pre-burn) size of the oak and pine, spatial relations between species, and intensity of ﬁre as determinants of ﬁre-related effects of one species upon another. Although spatial variations in the amount and type of fuels are important determinants of oak mortality, season of burning may be an even more important factor based on observations from another study of longleaf pine savannas in the Florida sandhills (Platt et al., 1991). There, ﬁre top-killed 80–90% of oaks in spring burns but only about half that in autumn/winter and summer burns (Platt et al., 1991). In contrast, pine mortality was 40% or less and did not vary signiﬁcantly with season of burn. Spring burns therefore should shift the competitive advantage to pine whereas burning in other seasons should have little if any differential effect. Length of ﬁre-free intervals also

141

may play a role in the ﬁre relations between oak and pine (Platt et al., 1991). Known relations between oaks and other species and the resultant spatial and temporal patterns of tree distribution and ﬁre characteristics emphasize the difﬁculties in accurately assessing the regeneration capacity of oaks in natural ecosystems as they relate to ﬁre. Despite the oak’s dependence on ﬁre, a ﬁre-free interval of sufﬁcient duration is needed for periodic recruitment of oak reproduction into the overstorey. Under natural conditions, recruitment is necessary to replace losses of overstorey trees that succumb to natural mortality from insects, disease, blowdown and ﬁre. Wherever oaks occur, the process of accumulation of oak advance reproduction beneath forest canopies varies across moisture gradients and disturbance regimes. In oak forests of the eastern United States, the accumulation process tends to be intrinsic in xeric ecosystems and recalcitrant in mesic and hydric ecosystems. In any ecosystem, disturbances can potentially modify the process through temporary xeriﬁcation of habitat and reduction of competition. In mesic and hydric ecosystems, disturbance may amplify accumulation, depending on the type, intensity and frequency of the disturbance event. However, single disturbances such as windthrow or heavy timber harvesting may simply release advance reproduction of shade tolerant species and accelerate succession toward dominance by non-oaks (Abrams and Scott, 1989; Abrams and Nowacki, 1992). In contrast, frequent burning, grazing or other disturbances may over time transform even the most mesic ecosystems into accumulators of oak reproduction by widening the oak’s regeneration window. Heavily disturbed mesic and hydric oak forests are often speciesrich mixtures that temporarily coexist until, in the absence of continued disturbance, shade tolerant or other aggressive mesophytes or hydrophytes assert their dominance (Fig. 3.18). The accumulation of oak reproduction beneath the parent stand is an important

142

Chapter 3

Moisture gradient

Oak reproduction accumulation type

Disturbance intensity/ frequency

Low

Xeric

Xero-mesic

Intrinsic

Ambivalent

Med

High

Low

Med

High

Mesic

Recalcitrant

Low

Med

High

Potential number of species in overstorey

Fig. 3.18. A conceptual model of some determinants of variation in species richness and predictability of the overstorey composition in upland oak forests in humid temperate regions. The three accumulation types define the natural propensity for the accumulation of oak advance reproduction, which decreases from intrinsic to recalcitrant. The three disturbance regimes define intensity and/or frequency of disturbance from low to high, which in turn are associated with increasing site xerification and correlated shifts in the accumulation of oak reproduction. The model assumes that the probability of maintaining oak in the overstorey within a given segment of the moisture gradient generally increases with increasing disturbance intensity or frequency. Potential tree species richness increases from xeric to mesic as basic site resources (e.g. moisture and nutrients) increase in availability and become less important in keeping species from becoming established in the overstorey. Potential overstorey species richness and thus a high uncertainty of obtaining a given species’ mix increases from light to dark shading of circles. Within each accumulation type, the potential number of overstorey species is postulated as maximum at an intermediate (‘medium’) disturbance level and minimal at a high disturbance level.

attribute of the regeneration dynamics of oak-dominated ecosystems. Although the extent to which this accumulation occurs varies greatly among oak-dominated ecosystems, it largely determines the permanence of oaks as canopy dominants. Even within the same deﬁned ecological unit, accumulation rates may vary with a stand’s successional status, disturbance regime, species composition, competitive

relations among species present and other factors.

Fluctuation in population density Like the process of reproduction accumulation, ﬂuctuation in reproduction density is determined by rates of seedling establishment and survival. Whereas xeric

Regeneration Ecology II: Population Dynamics

ecosystems in humid temperate regions tend to be characterized by accumulation and relative stability in oak reproduction density, greater ﬂuctuation and instability characterize reproduction density in mesic and hydric ecosystems. The latter tend to periodically produce large numbers of seedlings with low rates of survival, which consequently results in large ﬂuctuations in population density. One cause of ﬂuctuating reproduction density is the irregular production of acorns and new seedlings (Chapter 2) (Cecich, 1991). On average, most species produce a good acorn crop once every 3–5 years, although longer intervals frequently occur (Olson, 1974). Numerous biotic and abiotic factors inﬂuence acorn viability, germination, initial seedling establishment and survival. For example, dry weather, droughty soils and freezing temperatures can reduce acorn viability and germination (Korstian, 1927). Acorn crops also are frequently destroyed by unpredictable but frequent infestations of acorn weevils and other insects (Christisen and Kearby, 1984; Galford et al., 1988). Most of the remaining acorns may be consumed by rodents, deer, birds and other animals (Marquis et al., 1976; Sork et al., 1983; Galford et al., 1988; Borchert et al., 1989; Davis et al., 1991). Consequently, even after a bumper acorn crop, few seedlings may become established (Sullivan, 2001). Among the few viable acorns that escape predation, many fall into microsites unfavourable for germination and seedling establishment. Consequently, large numbers of new oak seedlings occur as unpredictable population waves associated with bumper acorn crops and a patchy spatial distribution. Although the frequency of occurrence of seedling crops and numbers of new seedlings produced are predictable only probabilistically, events that subsequently affect seedling survival are more predictable. For example, overstorey density and other stand characteristics inﬂuence survival of northern red oak seedlings originating from a single cohort. In general, reducing overstorey density increases seedling survival and growth (Beck, 1970; Loftis, 1988b, 1990a; Crow, 1992). A dense

143

layer of lower-storey trees, shrubs or ground cover also can reduce seedling survival and growth (Scholz, 1955; Beck, 1970; Loftis, 1990a; Lorimer et al., 1994). Other factors that can adversely affect oak seedling survival include animal browsing, insect defoliation, droughty soils, inadequate light and frost (Korstian, 1927; Hanson et al., 1987; McGee, 1988; Gottschalk, 1990; Crow, 1992). In North Carolina, the numbers of surviving northern red oak seedlings that originated from a single acorn crop declined at a constant exponential rate during their ﬁrst 9 years beneath the parent stand (Fig. 3.19A). If the same mathematical relation is applied to other reported survival data, results reveal annual survival probabilities for northern red oak ranging from 0.69 to 0.97 (Scholz, 1955; Beck, 1970; Crow, 1992; Loftis, 1988a) (Fig. 3.19B–D). These annual rates correspond to 5-year survival probabilities that range from 0.16 to 0.86, respectively. Although these comparisons represent relatively short time intervals, they emphasize the great variation in survival among seedlings of the same species in various regions and under different stand densities and light and competition environments. These rates also establish a range over which one might expect survival of northern red oak reproduction to commonly occur. Although the short-term survival rates of individual age cohorts are of interest, the accumulated population of reproduction from multiple cohorts over longer periods is often of greater interest. The most direct and accurate way to determine the combined survival rates is to observe the survival of multiple cohorts of oak reproduction for several decades, but observations of that duration are not available. Although short-term observations provide only glimpses of the dynamics of oak reproduction density, they illustrate the range of possibilities. Reported densities of northern red oak advance reproduction in North American forests range from about 400 to 9000 seedlings and seedling sprouts per acre under overstoreys that are predominantly red oak. However, most reported values range from 1000 to 4000

144

Chapter 3

1.0 Sub-canopy removed k = 0.921

A

Overstorey release B k = 0.869 Low release k = 0.902

Survival probability

0.8

0.6 Control k = 0.788

0.4 Sub-canopy intact k = 0.824

0.2

Overstorey & low release k = 0.869

0.0 0

2

4

6

8

10

0

1

2

3

4

5

6

Survival probability

1.0

0.8

Clearcut k =0.983

C

D Shrub k = 0.974

0.6

Grass k = 0.917

Thinned k = 0.884 0.4 Fern k = 0.692

Unthinned k =0.815

0.2

0.0 1

2

3

4

5

Seedling age (years)

6

1

2

3

4

5

Seedling age (years)

Fig. 3.19. Survival curves for northern red oak seedlings originating from single acorn crops (cohorts) expressed as a negative exponential annual survival rate (k) from age 0 (germination) or age 1 year. Survival probability (P) = ky, where y = years since the initial population census in year 0. (A) Under thinned and unthinned mesic forests in North Carolina. (Adapted from Loftis, 1988a.) (B) In four overstorey/understorey release treatments in mesic forests in North Carolina. (Adapted from Beck, 1970.) (C) In clearcut, thinned and unthinned stands treated with a herbicide in dry-mesic forests in northeastern Wisconsin. (Adapted from Crow, 1992.) (D) Competing with three classes of ground cover under fullystocked mesic forests in southwestern Wisconsin. (Adapted from Scholz, 1955.)

(Tryon and Carvell, 1958; Scholz, 1959; Johnson, 1974; Loftis, 1983; Beck and Hooper, 1986; Nowacki et al., 1990; Steiner et al., 1993). However, densities in forests in France frequently exceed 100,000 per acre (Steiner et al., 1993). Longer-term ﬂuctuations in oak reproduction density can be estimated by assuming that acorn crops, and thus seedling crops, can be estimated proba-

bilistically. Based on this idea, a simulation model was developed using reported survival rates and population densities as benchmarks for simulating the ﬂuctuation in the density of northern red oak reproduction. The model is based on only two factors: seedling recruitment and seedling survival rates (Rogers and Johnson, 1998). The model itself makes no assumptions about what causes variation in either fac-

Regeneration Ecology II: Population Dynamics

tor, but it does assume that variation in recruitment can be speciﬁed as a probability. To derive a seedling recruitment rate, the model uses an exponential distribution function to specify the probability (P) that a given number of new seedlings per acre (n) will be added to the existing population of seedlings and seedling sprouts. The function is expressed by: P = 1 e (n/I)

[3.3]

where I is a modiﬁer of the distribution function called the seedling recruitment constant; and e is the base of the natural logarithm. Reasonable values of I range upwards to 2000 or larger, but these values are not equivalent to numbers of seedlings. Rather, I is a model constant that can be adjusted to calibrate the recruitment function of the model for different types of red oak forests, site conditions and other factors. To directly express the number of new seedlings (n) to be added to the existing population, Equation 3.3 can be restated:

n = ln(1−P) • I

145

[3.4]

The term n is introduced into the simulator by randomly generating a probability, P, such that all values of P occur with equal probability, and using it to solve Equation 3.4 for the number of new seedlings to be added to the population in any given year (Fig. 3.20). A probabilistic method for generating n was used because the factors known to affect acorn production and seedling recruitment (including weather and acorn predation) are essentially unpredictable over time. To complete the model, a survival rate, K, must be brought into Equation 3.4 such that: ni+1 = K•ni ln(1−Pi)•I

[3.5]

In this formulation, the model states that the number of seedlings and seedlings sprouts present the next year (ni+1) equals the number surviving to the present (K•ni ) plus any new seedlings added to the population during the next year (ln(1−P)•I).

Seedlings recruited per acre (n)

10,000

8000

6000

4000

I =1000 2000

I =1500 I =500

0 0.0

0.2

0.4

0.6

0.8

1.0

Probability Fig. 3.20. The probability distribution function for determining the number of seedlings per acre (n) recruited, i.e. added to the existing population of northern red oak reproduction, in a given year (based on Equation 3.4). Numbers, n, are shown for three recruitment rates corresponding to different values of I, the seedling recruitment constant in the SIMSEED model. (Adapted from Rogers and Johnson, 1998.)

146

Chapter 3

By applying the model, ﬂuctuations in population density over various durations can be graphically described through computer simulation (Fig. 3.21). These simulations do not model the dynamics of the overstorey stand through time. Rather, they describe the ﬂuctuation in reproduction density that could be expected over time assuming that potential acorn production and other stand factors remain essentially constant. The simulation model projects innumerable hypothetical situations. For example, we might consider a stand with an overstorey that is predominantly northern red

oak and near maximum acorn-producing capacity (Fig. 3.21A). Seedling survival and input rates for this simulation were chosen to simulate the population ﬂuctuation in a relatively favourable environment, e.g. a moderately reduced overstorey and subcanopy density, and sparse ground cover. This simulation shows that reproduction density ﬂuctuates primarily within the range of 6000–12,000 seedlings and seedling sprouts per acre, with occasional surges to 15,000 or more. The simulation of a hypothetical stand with a lower seedling survival and recruitment rate produces a

16,000 Seedlings per acre

A 12,000

8000

4000

0 0

25

50

75

100

125

150

175

200

25

50

75

100

125

150

175

200

16,000

Seedlings per acre

B 12,000

8000

4000

0 0

Years Fig. 3.21. Simulated fluctuation in the density of northern red oak reproduction in two hypothetical steady states. (A) A state characterized by a moderately high seedling survival rate (K = 0.8) and seedling recruitment rate (I = 1500); (B) A state characterized by a lower seedling survival rate (K = 0.7) and seedling recruitment rate (I = 1000). The model is based on Equation 3.5. The model assumes there are no seedlings at the beginning of the simulation period, and that potential acorn production and other stand factors remain constant. (Adapted from Rogers et al., 1993, with permission.)

Regeneration Ecology II: Population Dynamics

population density ﬂuctuating primarily within the range of 2000–5000 seedlings and seedling sprouts per acre (Fig. 3.21B). Those rates would characterize stands with low acorn producing capacity (e.g. resulting from the harvesting or death of acorn producers), a high overstorey density, or dense ground vegetation (e.g. heavy fern cover) that results in a low seedling survival rate. Both simulations illustrate the hypothetical ﬂuctuations in populations of oak reproduction and the problems associated with expecting oak reproduction density to be adequate for regeneration in any randomly selected year. Thus, a single inventory of advance reproduction may not reﬂect the reproduction population density occurring most of the time. The simulations also illustrate that, despite all the factors that hinder oak seedling establishment, years with abundant oak reproduction will occur. These hypothetical simulations also tend to equilibrate about a given density range – which in real stands would be determined by ecosystem characteristics, including those that are considered by ecological classiﬁcation systems (see Chapter 1).

Regeneration Potential Both the accumulation of oak reproduction and its ﬂuctuation in density beneath the canopy of the parent stand are important properties of the regeneration ecology of oaks. The outcome of those processes largely determines the amount and size of the oak reproduction at any given time and its capacity to capture growing space that becomes available when disturbance reduces overstorey density. Regeneration potential is not a static property of a forest but changes with time as stand characteristics and disturbance events occur. The latter can usually be only speciﬁed probabilistically. A stand’s oak regeneration potential is actually realized through the reduction of overstorey density to a level sufﬁcient to permit the recruitment of oak reproduction into the overstorey. In the absence of dis-

147

turbances that reduce overstorey density, self-thinning (Chapter 6) regulates stand density. Although stands undergoing selfthinning continually produce small canopy gaps as trees succumb to natural mortality, stand density is maintained near a constant upper limit of crowding by the expanding crowns and roots of survivors. Consequently, numbers of trees per unit area continually decrease as their average size increases. Although the resulting competition-induced mortality occurs among trees of all crown classes, proportionately more trees die in the inferior, i.e. intermediate and suppressed, crown classes (Lorimer, 1981; Ward and Stephens, 1994). In some ecosystems, shade tolerant species such as sugar maple may be recruited into the sub-canopy of an oak overstorey that is at maximum density (McGill et al., 1999). This recruitment may be facilitated by maple’s regeneration niche that allows it to exploit above- and belowground resources that are largely unavailable to the oaks. But even the shade tolerant sugar maple requires some reduction in canopy density to continue its growth to a dominant or co-dominant crown position (Godman et al., 1990). In contrast, some shade tolerant species of limited stature such as ﬂowering dogwood and sourwood are adapted to the sub-canopy where they can successfully complete their entire life cycle (McLemore, 1990; Overton, 1990). The oak reproduction, however, is largely relegated to the understorey in the absence of disturbances that reduce overstorey density. Disturbances that reduce stand density range from the death of a single tree to larger-scale stand and landscape-level disturbances resulting from timber harvesting, ﬁre, wind and insect epidemics. Every tree species has evolved a regeneration strategy that enables it to capitalize on the growing space released by reductions in overstorey density. Some species have been described as exploitative in their ability to capture growing space, whereas others are conservative (Bormann and Likens, 1979). The upland oaks, and to a lesser extent the bottomland oaks, fall into the latter category

148

Chapter 3

because of the relatively long time required to develop a root system sufﬁcient for rapid height growth following a reduction in overstorey density. Oaks that fail to develop the requisite root size are often doomed to displacement by other species that have a greater capacity to exploit the growing space.

Regeneration mode Overstorey disturbance at some spatial scale is inevitable. Veblen (1992) used the term regeneration mode to refer to the disturbance scale at which a species’ regeneration is most likely to occur. While recognizing that regeneration occurs along a continuum of scales, Veblen deﬁned three modes: gap-phase, catastrophic and continuous. The gap-phase regeneration mode refers to regeneration occurring in small canopy gaps of less than 1 acre that originate from the death of one or a few trees (Watt, 1947). This mode is usually related to natural disturbances originating from within (i.e. endogenous to) a forest but may also include outside forces such as wind. Included is the localized mortality of individual trees from insects, disease, drought and windthrow. At the smallest scale, individual overstorey trees die as they grow older and create gaps in the forest canopy. If the tree creating the gap is large and dies in place, it creates a relatively large, somewhat circular opening in the forest canopy. If the tree is wind-thrown, it may create a long, narrow gap by toppling adjacent trees as it falls. Gap size and shape consequently vary depending on the nature of the disturbance. Gaps may develop slowly when they are caused by a progressive decline in tree vigour before mortality ﬁnally occurs. They also may progressively expand where tree mortality is related to contagious pathogens such as oak wilt. The release of space and resources essential for plant growth therefore may be relatively slow. Small gaps or slow gap development may favour the reproduction of shade tolerant

species, whereas the sudden creation of a large gap may favour less tolerant species or a mixture of tolerant and intolerant species. Depending on the species present, a new canopy gap may be quickly exploited by advance reproduction, new sprouts originating from stumps or roots, or new seedlings. Various sources and species of tree reproduction, in addition to shrubs and herbaceous species, respond to the new space and inﬂuence the competitive outcome. However, the gap immediately begins to close as the crowns of trees bordering the gap expand. Gap closure rates depend on the growth rates of the species in the gap perimeter (Fig. 3.22). The creation of a gap consequently initiates a race to its own extinction. If the gap is large enough, one or more trees originating from new or advance reproduction may be recruited into the overstorey to replace the former occupants. Success in capturing growing space depends on the duration of favourable light and moisture conditions as well as the suite of other factors that inﬂuence the growth and survival of reproduction. In a circular gap opening with diameter equal to the height of surrounding trees, after 10 years of crown closure the amount of direct solar radiation is reduced to 30% compared to that at the time the opening was created. The corresponding reduction is about 70% in gaps with diameters half the height of surrounding trees (Table 3.3). Tree species differ in their ability to capture the space created by canopy gaps. Shade and drought tolerance, growth rate, regeneration strategy, longevity, maximum attainable tree size, gap size and shape, and other factors determine this ability. Some species are well adapted to quickly exploit the space created by a new gap. In northern hardwood forests in New England, pin cherry may quickly, but only temporarily, exploit large canopy openings even when that species is not present as advance reproduction or as adult trees. Its temporary dominance is facilitated by hundreds of thousands of bird-disseminated seeds per acre buried in the forest ﬂoor and soil

Regeneration Ecology II: Population Dynamics

A

149

B

Crown closure rate (ft/decade)

14 A 0

12

20 feet

10

Northern red oak Tree bole

8

Crown encroachment into opening (10-year radial growth)

6 Yellow-poplar 4 Tree crown

2 0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

Diameter growth rate (inches/decade) Fig. 3.22. Crown closure rates for canopy gaps. (A) Relation between estimated 10-year rate of crown closure (R) and average 10-year diameter growth (G) of trees bordering a gap. For northern red oak, R = 1.822G + 5.36; for yellow-poplar, R = 1.096G + 2.90. (B) Crown encroachment of northern red oak border trees into a hypothetical 1/8-acre canopy gap during the 10-year period after gap formation. (Adapted from Trimble and Tryon, 1966, by permission of Society of American Foresters, Bethesda, Maryland. Not for further reproduction.)

Table 3.3. Estimated reduction in direct solar energy received in circular openings 5 and 10 years after their origination.a Reduction (%) in direct solar energy Initial diameter of opening (no. of tree heights)b 0.5 1.0 2.0

After 5 years

After 10 years

21 17 4

71 30 7

aBased

on a computer simulation model (Fischer, 1979, used with permission). The simulation assumes perimeter trees are 20 inches dbh and 75 ft tall at the time the opening is created. Values shown are relative to solar radiation in the open and assume that growth rates of perimeter trees are 0.3 inch per year in dbh and 0.8 ft per year in height. bBased on heights of trees surrounding the canopy opening.

where they may remain viable for up to 50 years (Marks, 1974). When a canopy opening is created, the conditions trigger the germination of the pin cherry seeds. Because of its rapid early growth, it may dominate openings for two to three decades (Marks, 1974). However, the species does not grow to large size, is

extremely intolerant of shade and is relatively short-lived. Consequently, pin cherry’s dominance of forest openings rapidly declines during the second decade after its establishment. This decline then may facilitate replacement by more shade tolerant species with a more conservative regeneration strategy.

150

Chapter 3

In northern hardwood and other mixedspecies forests, shade tolerant hardwoods such as sugar maple and American beech, and sometimes the moderately tolerant northern red oak, may dominate canopy gaps within three decades (Lorimer, 1983). But the success of the oak largely depends on its growth rate, and thus the size of its advance reproduction at the time of the disturbance. The regeneration strategy of oaks therefore is usually dependent on a relatively long pre-disturbance investment in their root systems. This conservative strategy may disadvantage an oak when it competes with species having a more exploitative regeneration strategy. Consequently, the advantages of the oaks’ regeneration strategy are most likely to be expressed in environments that conform to those deﬁned by the oaks’ regeneration window (Fig. 3.7). The silviculturist can simulate the gapphase mode of regeneration by creating small openings in the canopy using group selection and group shelterwood methods of regeneration (Smith, 1986; see also Chapters 7 and 8). Oaks are potentially capable of capturing canopy gaps if the gaps are large enough and advance repro-

duction of large size is present (Fig. 3.23). Northern red oak reproduction about 10 ft tall captured 1/25-acre canopy gaps in a sugar maple–red oak stand in southern Wisconsin (Lorimer, 1983). Ecosystems where oak reproduction intrinsically accumulates are also likely sites for oaks to regenerate by the gap-phase mode. The catastrophic regeneration mode occurs over a brief period and a relatively large area (e.g. > an acre). It usually involves a disturbance originating from forces outside (exogenous to) the stand such as ﬁre, tornadoes, hurricanes and timber harvest. These events abruptly increase light, moisture and nutrients, which are also accompanied by signiﬁcant changes in microclimate. This mode of regeneration often facilitates the temporary coexistence of trees of all shade tolerances (Roach and Gingrich, 1968; Dunn et al., 1983; Smith, 1986). Coexistence is facilitated by shade tolerant species originating from advance reproduction together with less tolerant species originating from new or buried seed and sprouts. In some mesic forests, this may produce stands of high tree diversity (Loftis, 1983; McGee, 1987; Smith and Miller, 1987). Even-aged meth-

Fig. 3.23. A northern red oak sapling growing in a canopy gap in a xero-mesic oak-mixed hardwood stand in northern Wisconsin. This sapling’s potential to eventually dominate the gap depends on its height growth and the rate of crown closure of surrounding trees. Adjacent trees in the main canopy include northern red oak, bigtooth aspen, paper birch, red maple and white ash. (USDA Forest Service, North Central Research Station photograph.)

Regeneration Ecology II: Population Dynamics

ods of silviculture such as clearcutting, shelterwood, seed tree and their variants simulate the catastrophic mode of regeneration through the removal of all or most of the overstorey in one or a few steps (Marquis and Johnson, 1989). Oaks are well adapted to the catastrophic mode of regeneration, especially in xeric ecosystems where oak reproduction intrinsically accumulates. However, in mesic and hydric forests, successful regeneration of oaks may depend on disturbances from ﬁre, grazing or ﬂooding that precede a later catastrophic event that signiﬁcantly reduces overstorey density. The intensity, frequency and recency of minor disturbances combined with later catastrophic disturbances collectively direct ecological succession (GlennLewin and van der Maarel, 1992). The continuous regeneration mode pertains to species that can attain maturity in the absence of a canopy opening. In the oak forests of the Eastern Deciduous Region of the United States, these include shade tolerant species such as ﬂowering dogwood, redbud, American hornbeam, serviceberry and sourwood. These species usually are relegated to the sub-canopy because their maximum attainable heights seldom exceed 35 ft (Burns and Honkala, 1990). However, they do not require canopy gaps for seed production, germination or early growth. Terborgh (1985) theorized that the tops of the relatively ﬂat, spreading crowns of ﬂowering dogwood and redbud occur at a predictable distance below the main canopy. That distance occurs where the beams of solar radiation penetrating different canopy gaps most frequently intersect along the daily solar path during the growing season. Terborgh reasoned that the distinct layers of vegetation in a forest may be adaptations to relatively uniform light ﬁelds that occur at ﬁxed distances below the main tree canopy, which in turn are related to the shape, size and distribution of tree crowns and gaps, together with angles of direct-beam solar radiation. Accordingly, vegetation would occur as distinct vertical layers, as commonly observed, rather than as a vertical continuum of tree crowns (Fig. 3.24).

151

Although Veblen (1992, p. 166) offers that ‘… the continuous mode of regeneration appears to be rare’, some shade tolerant species such as sugar maple and American beech are capable of growing directly into the overstorey if canopy density is not extremely high (Spurr and Barnes, 1980; Canham, 1989; Godman et al., 1990). In oak forests, high overstorey densities reduce light reaching the forest ﬂoor to levels that are usually insufﬁcient for sustaining the growth of oak reproduction into the overstorey. The regeneration of oaks thus depends on forest disturbance, the number, size and spatial distribution of oak advance reproduction present at the time of disturbance, and other factors associated with site quality and competition (Sander, 1971; Sander and Clark, 1971; Trimble, 1973; Sander et al., 1984). When overstorey density is reduced, the oak advance reproduction may capture the vacated growing space. The resulting recruitment of oaks into the overstorey is slow in small canopy gaps that admit little light to the forest ﬂoor and more rapid in larger openings – provided that adequate reproduction is present (Fig. 3.25). Other factors being equal, the growth rate of oak advance reproduction after overstorey removal depends on its pre-disturbance basal diameter (Sander, 1971; Johnson, 1979), which is correlated with total root mass (Canadell and Rodà, 1991) (Fig. 2.27). In oak stands that accumulate oak advance reproduction, stand regeneration potential and succession largely depend on the composition and structure of that reproduction. This type of succession follows the ‘initial ﬂoristics’ model (Egler, 1954). Succession is driven by an initial ﬂoristics when the condition of a plant community at the time of disturbance largely controls its future development. Similar ideas are implicit in the ‘legacy’ concept of Franklin et al. (1989) wherein the future state of an ecosystem is perceived as partly, if not largely, inherited from propagules carried over from the previous state.

152

Chapter 3

Fig. 3.24. Flowering dogwood (seen here in full spring bloom) forms a distinct and permanent subcanopy in this oak–hickory forest. (USDA Forest Service, North Central Research Station photograph.)

A

B

C

Fig. 3.25. Recruitment of oak reproduction into the overstorey. (A) No recruitment; advance reproduction (trees below the thin horizontal line) is suppressed by high overstorey density. (B) Some recruitment occurs in the centre of small canopy gaps. (C) Recruitment accelerates with increasing gap size. Maximum height growth and recruitment occur in centres of gaps where light is maximal and competition from border trees is minimal.

The dependence of the future condition of a stand on its initial state is an important concept in the development and application of predictive regeneration models (e.g. Sander et al., 1984; Loftis, 1990a; Marquis et al., 1992; Johnson and Deen, 1993) (Fig. 3.26). When overstorey density is reduced by ﬁre or timber harvesting, the overstorey itself is often an important component of the initial

state of an oak stand because a signiﬁcant proportion of the reproduction may originate from dormant buds at the bases of top-killed trees or stump sprouts (P.S. Johnson, 1975, 1977; Wendel, 1975; Lamson, 1988; Weigel and Johnson, 1998). The importance of this source of reproduction to the future stand depends on the proportion of trees that produce basal or stump sprouts. In general, older

Regeneration Ecology II: Population Dynamics

153

ASPECT

NW (neutral)

0.37 0.42

NE (cool)

0.28 0.10

0.13

0.27 0.15

0.33 0.12

0.10

0.20 0.08

Lower Middle Upper Slope position

0.32 0.37 0.23

0.37

0.12

0.13

0.42 0.15

0.13

0.09

0.27 0.10

SW (hot)

SE (neutral)

Initial size of advance reproduction (height, basal diameter) 4 ft, 0.5 in

P

2 ft, 0.25 in

P

P = probability of attaining intermediate or larger crown class 21 years after overstorey removal.

Fig. 3.26. Estimated probabilities that oak advance reproduction will attain an intermediate or larger crown class 21 years after complete overstorey removal in the Ozark Highlands of Missouri. Probabilities are shown in relation to aspect for two preharvest size classes (4 ft tall, 0.5 inch in basal diameter, and 2 ft tall, 0.25 inch in basal diameter). The probabilities illustrate the predictive value of the size of oak advance reproduction (i.e. the ‘initial state’) as predictors of a forest’s future state. Probabilities apply to black, white, scarlet, northern red and post oaks, and are based on the predictive regeneration model ACORn. (Adapted from Dey et al., 1996b.)

and larger trees are less likely to produce stump sprouts than smaller and younger trees (Roth and Sleeth, 1939; P.S. Johnson, 1975, 1977; Ross et al., 1986) (Fig. 2.24). But even in older stands, stump sprouts may contribute signiﬁcantly to the future occupancy of growing space (stocking) because, despite their small numbers, growth and survival rates of stump sprouts are usually high (Johnson and Rogers, 1984; Lamson, 1988).

The ecological signiﬁcance of the predisturbance state of an oak forest is not realized until the inhibiting effect of the overstorey is released. The effect accordingly has been referred to as the ‘inhibition’ model of succession (Connell and Slatyer, 1977). Accordingly ‘… replacement occurs only when resources are released by the damage or death of the previous occupant [of the site]’ (Connell and Slatyer, 1977,

154

Chapter 3

p. 1138). An oak forest’s regeneration potential therefore is often encoded in the composition and structure of its advance reproduction and the overstorey. But this potential is not expressed until it is freed from the inhibitory effect of the overstorey. In a mature oak stand, site resources are partitioned among the various vegetative layers. Large trees utilize resources (e.g. soil moisture and nutrients) from the relatively large and heterogeneous spaces they occupy, whereas smaller trees and other low vegetation are limited to relatively small and homogeneous microsites. The early postdisturbance years after overstorey removal are characterized by the direct competition of oak reproduction with a chaotic mixture of

herbaceous, shrub and other tree reproduction, which collectively occupy relatively narrow above- and below-ground strata. Under these conditions, herbaceous and shrub species interfere directly with the growth and survival of tree reproduction. This period is characterized by rapid competitive sorting among and within species. During this period, the composition and structure of the future forest are often neither obvious nor apparently predictable. The oaks may win or lose in the competitive sorting process, depending on antecedent states and events (Fig. 3.27). Nevertheless, a more orderly partitioning of resources gradually reemerges but only becomes visually conspicuous after tree crowns again vertically stratify.

Forest type Non-oak Accumulation of oak reproduction: RECALCITRANT

Mixed oak Oak

Accumulation mediated by disturbance Oaks win Mixed outcome Oaks lose or no oaks present

Not mediated Accumulation of oak reproduction: INTRINSIC

Competitive sorting

Advance tree reproduction and seed bank inhibited by overstorey

Disturbance-mediated release from overstorey inhibition Fig. 3.27. Disturbance-mediated successional pathways in relation to intrinsic and recalcitrant patterns of accumulation of oak reproduction. In the eastern United States, accumulation is usually an intrinsic characteristic of xeric forests. In contrast, accumulation is generally recalcitrant in mesic and hydric forests in the absence of disturbances such as recurrent fire. In the western United States (in Mediterranean and semi-desert climates), accumulation may be favoured in the more moist ecosystems and microhabitats. The outcome of competitive sorting among species during the first two decades after disturbance is largely determined by advance reproduction characteristics and site factors. In eastern United States, oaks typically win the post-disturbance race to capture growing space in xeric ecosystems. In mesic and hydric ecosystems, oak’s ascendance to dominance is more variable, and highly dependent on the type, frequency and intensity of disturbance, which in turn influences the accumulation process. Oak reproduction also can occur and accumulate beneath non-oak types via acorn dispersal by birds and mammals.

Regeneration Ecology II: Population Dynamics

Putting the necessary information on oak regeneration potential into a practical and useful framework for predicting the future composition and structure of oak stands requires quantitative regeneration models.

Modelling theory and objectives In their broadest sense, models represent ‘… in some way the form and/or the function of real-world entities and processes’ (Kimmins, 1987, p. 460). Models therefore may range from unrevealed thoughts to more externally represented ideas expressed by words, pictures, graphs, charts, mathematical equations and computer simulations (Kimmins, 1987). The latter has largely made modelling the valuable tool that it has become for predicting and understanding the behaviour of ecosystems generally and forests speciﬁcally. Many predictive models have been developed including the successional models of ecologists (e.g. Shugart, 1984; Urban and Shugart, 1992; Botkin, 1993), and the growth and yield models of silviculturists (e.g. Belcher et al., 1982; Hilt, 1985; also see Chapter 10). It has been suggested that complete knowledge of the state of an ecosystem (i.e. its multidimensional structure or ‘spacestate’) at any given instant cannot be fully speciﬁed because of the large number of factors, living and non-living, that deﬁne ecosystems (Margalef, 1963; Jørgensen, 1990). Even if such a model could be constructed, its complexity would probably render it useless. This produces the conundrum that an ecosystem model can only approach reality through increasing complexity, which in turn increasingly reduces its utility. Nevertheless, relatively simple ecosystem models can have surprisingly predictive power (Urban and Shugart, 1992), and such models have become widely used to predict forest behaviour. It is none the less important to recognize the limitations of models. The suitability of a model should be evaluated by how well it

155

meets the objectives of its intended application (Buchman and Shiﬂey, 1983; Blake et al., 1990; Bruce, 1990). Forests represent hierarchies of biotic organization ranging from the cellular level upwards to the individual organism level and extending further upwards to species’ populations, associations of species and beyond. Depending on its purpose, a model should be designed to consider speciﬁc levels of detail within hierarchies (Allen and Starr, 1982). Models developed for silviculturists usually focus on predicting population phenomena such as time-dependent changes in the number and size of trees in a stand. Providing a silvicultural explanation of phenomena at the population level typically requires consideration of detail at least one level lower, i.e. for the individual trees in the population (Botkin, 1993). If the ‘explanatory’ level of a model is set too far below the ‘predictive’ level, the detail required by the model may exceed the limits of available knowledge or the ability to meaningfully build such details into a model. On the other hand, a more hierarchically detailed model may have more explanatory power. The ‘bottom-up’ approach (aggregating model components from ﬁne to coarser scales) is consistent with the scientiﬁc method and the building of scientiﬁc theories (Lewis, 1990). It is also the most common approach used in modelling forest dynamics. However, a ‘top-down’ approach to modelling also is possible and begins with a relatively coarse scale, but adds details in subordinate levels (Landsberg, 1986). In addition to hierarchical structure, models also possess other properties that determine their usefulness for meeting deﬁned objectives. Generality, realism and accuracy are three such properties (Levins, 1966; Sharpe, 1990; Botkin, 1993). Generality refers to the range of situations that a model can be applied to. Realism refers to the qualitative similarity between model projections and the real world (e.g. as demonstrated by the similarity in the shapes of projected and actual response curves). Accuracy4 refers to the quantitative

4We use the term accuracy, as Botkin (1993) did, instead of precision as originally used by Levins (1966) and later by Sharpe (1990) to discuss these model properties. In the context of models, accuracy is more consistent with its parallel meaning in statistics and the sciences in general. Also see Sokal and Rohlf (1969).

156

Chapter 3

closeness of model projections to the real world. The goal or objective of the model determines the relative importance of each of these properties. However, there are limitations to each of the various approaches to modelling because each tends to rely on one of three essentially mutually exclusive approaches that are either: (i) statistical, (ii) mathematical, or (iii) science-based. As a consequence, one of the three properties is likely to be sacriﬁced in any one model (Levins, 1966; Sharpe, 1990). In their construction, statistically derived models require relatively large amounts of data. Fitting curves to data using regression analysis often is used to derive relations among variables. The resulting models tend to lack generality because their efﬁcacy in application is limited to the conditions represented by the data on which they are based. Nevertheless, within that limitation, such models are potentially (but not necessarily) accurate and realistic. Most of the models developed for silvicultural applications, including regeneration and growth and yield models, fall into this category. In contrast, mathematical models are derived from abstract relations assumed to describe systems within narrowly deﬁned conditions. They may require no data and are often designed to investigate generalized theories of ecosystem behaviour such as stability and response characteristics (e.g. Gatto and Rinaldi, 1987). Such models may possess accuracy and generality, but often sacriﬁce realism. Realism is lost because of the recognized unrealistic assumptions about the system that must be made in order to conform the system to the model’s known limitations. Mathematical models are usually used to explore theory and are not generally applied to predict outcomes for real ecosystems (Prentice and Helmisaari, 1991). Sharpe (1990) deﬁned a third class of models that he termed science-based models. They are more common to ecology than silviculture and are derived from available information on species’ characteristics such as growth rates, longevity, reaction to stress factors, and other natural history

characteristics. In their construction, science-based models require less data than statistical models but more data than mathematical models. The approach has been widely used to develop forest succession models (e.g. Shugart, 1984; Prentice and Helmisaari, 1991; Botkin, 1993) and to a lesser extent regeneration models (e.g. Waldrop et al., 1986). Science-based models are potentially realistic and general but sacriﬁce accuracy. Accuracy is lost because of the generalized database from which they are derived. Each of these three classes of models is also associated with speciﬁc methods and problems of validation (Sharpe, 1990). The users of models in forestry are largely forest managers who are often concerned with the consequences of silvicultural actions. Accurately predicting the outcome of silvicultural actions is important to managers. However, accurate prediction does not infer that ecological or physiological processes are well understood or even explicitly considered in model development. On the contrary, models with high accuracy often are not based on the mechanisms that directly explain the predicted phenomenon. High accuracy and low explanatory power characterize most of the current forest growth and yield models (Sharpe, 1990; Prentice and Helmisaari, 1991). Such models are largely empirical in that the variables used to make predictions are only indirectly related to the actual mechanisms that govern stand development. For example, a model predicting growth of individual trees may employ predictors such as the initial dbh, the crown class, the ratio of the living crown to its total live height (live crown ratio) and some measure of site quality (e.g. site index). Of course, the underlying causes of tree growth (and death) are rates of photosynthesis, respiration, uptake of water and nutrients, and other physiological mechanisms and the stress factors that limit those processes. Models based on causative physiological mechanisms are sometimes referred to as process models (Blake et al., 1990; Bruce, 1990). There is strong interest in the

Regeneration Ecology II: Population Dynamics

development of ecosystem models that directly consider physiological processes such as photosynthesis and carbon allocation in trees (Dixon et al., 1990). Their potential advantages lie in their greater generality of application and explanatory power over empirical models (Blake et al., 1990; Isebrands et al., 1990; Sharpe, 1990). However, the advantages of generality are likely to be counterbalanced by reduced accuracy and increased complexity of design and efﬁciency of application (Levins, 1966; Sharpe, 1990). It also can be argued that all models are empirical because causation at some scale (e.g. molecular) is not addressed by the model.

Stand-level regeneration models: purpose, problems and limitations From an ecological perspective, regeneration models could be deﬁned as predictors of the outcome of short-term secondary succession associated with planned disturbances. They are designed to function as silvicultural tools for assessing the adequacy of the regeneration potential of stands for meeting pre-deﬁned regeneration goals. Successional models usually do not fulﬁl that function adequately because they usually do not provide the stand-level accuracy needed by silviculturists (Waldrop et al., 1986). Growth and yield models for established stands generally qualify as possessing accuracy but they usually do not explicitly consider regeneration. For example, changes in tree populations over the ﬁrst two decades after ﬁnal harvest of even-aged stands usually are not included in (or are only coarsely simulated by) growth and yield models. This leaves an informational void on stand development for one-ﬁfth to one-fourth of the 80to 100-year rotations typically used in managing oak forests (Chapter 7). Like all models, regeneration models are imperfect in mimicking real ecosystems because the models are simpliﬁcations of reality. Although regeneration models are largely empirical, they may be more precise and realistic than some

157

other kinds of forest dynamics models. However, they typically possess little generality and explanatory power because they are usually statistically derived and based on correlations rather than on causative mechanisms. Several predictive regeneration models nevertheless have been developed for oak and associated mixed hardwood forests (Sander et al., 1984; Waldrop et al., 1986; Loftis, 1990a; Marquis et al., 1992; Johnson and Deen, 1993; Dey et al., 1996a,b). While these models differ in their details and complexity, they are all based partly or wholly on the successional concepts of initial ﬂoristics (Egler, 1954) and inhibition (Connell and Slatyer, 1977). In other words, they assume that the future state of a forest is encoded in its current or initial state. Moreover, overstorey removal or its reduction in density is required to release the future stand from the inhibiting effects of the present overstorey. Because such models are predictors of short-term secondary succession, they are potentially useful as tools for guiding silvicultural decisions. In application, regeneration models are usually designed for speciﬁc geographic regions or ecosystems, and they require quantitative information on the initial state of a stand including current vegetation and site characteristics. For models to be silviculturally practical, the initial state, or predictor variables, must be easily measurable and have predictive power. Those requirements create practical and theoretical problems in the development of regeneration models including: (i) the selection of a relatively small but useful set of easily measured predictors; and (ii) the relatively large variation in tree growth and survival during the post-disturbance regeneration period. These problems are further confounded by the relatively long period of observation (up to two decades) required to obtain data from which to build some types of models. Predictors of the post-disturbance state usually include measurements and/or counts of advance reproduction obtained from sample plots within stands (e.g.

158

Chapter 3

Sander et al., 1984; Marquis et al., 1992; Dey et al., 1996b). Information on advance reproduction heights and/or diameters is a requirement for applying many regeneration models. Although those variables are easily measured attributes of reproduction, they may not be the most important determinants of a stand’s future state. For example, the growth potential of an oak seedling may be more directly and accurately related to its root mass and leaf area (Johnson, 1979). But because those measurements are impractical to obtain in silvicultural practice, height and diameter are the measurements of necessity. Other sources of variation considered in some models include the density and size of competing shrubs and herbs. The ‘initial ﬂoristics’ approach used in prediction also may limit the predictive power of models because it downplays potential contributions of trees originating from seed and stump sprouts.

Fluctuations in weather, insect and mammal populations that affect regeneration processes may be impossible or difﬁcult to predict. Models therefore may lack predictive power because of their inability to consider important factors, both known and unknown. Building regeneration models is also complicated by the inherent variation associated with the disturbed states that characterize regenerating forests. The sudden post-disturbance increase in light triggers the germination of seeds of many tree, shrub and herbaceous species. Some seedlings arise from seed buried in the forest ﬂoor while some are carried in from outside the disturbed area. Advance tree reproduction also is released from suppression and new sprouts may develop from dormant basal buds of cut trees. This produces a chaotic mix of established plants and new propagules that are suddenly competing for space

INPUT Stand and site characteristics: Overstorey Advance reproduction

REGENERATION SIMULATOR

Overstorey harvested

OUTPUT Regenerated stand characteristics: Species Tree size distributions Stocking

Fig. 3.28. Data input and output elements for a representative regeneration simulation model. Input requirements and output vary among available models (see Chapter 7).

Regeneration Ecology II: Population Dynamics

in the new environment. The result is a relatively unstable ecological state that produces changes in stand composition and structure that are difﬁcult to predict. During this period, stand development often is so uncertain that the outcome may best be expressed probabilistically (e.g. Fig. 3.26). Despite these problems, regeneration models attempt to bring order to the chaos of this least predictable stage of stand development. Regeneration models generally fall into one of two categories: (i) decision guides for evaluating the adequacy of a stand’s regeneration potential for meeting deﬁned silvicultural objectives; and (ii) simulation models that predict future stand composi-

159

tion and structure. The former are largely limited to providing a ‘yes’ or ‘no’ answer to the question of the adequacy of regeneration potential. Such models also may be partially or wholly based on experience or ‘expert opinion’. This class of models may range from the relatively simple to the complex, and their application often does not require a computer. In contrast, simulation models predict future stand composition and structure. Because of their complexity, application requires computer software (Fig. 3.28). Both types of regeneration models have been developed for oak and mixed forests in several regions (Chapter 7).

References Abrams, M.D. (1992) Fire and the development of oak forests. BioScience 42, 346–353. Abrams, M.D. and Nowacki, G.J. (1992) Historical variation in ﬁre, oak recruitment, and post-logging accelerated succession in central Pennsylvania. Bulletin of the Torrey Botanical Club 119, 19–28. Abrams, M.D. and Scott, M.L. (1989) Disturbance-mediated accelerated succession in two Michigan forest types. Forest Science 35, 42–49. Allen, T.F.H. and Starr, T.B. (1982) Hierarchy: Perspectives for Ecological Complexity. University of Chicago Press. Arend, J.L. and Scholz, H.F. (1969) Oak forests of the Lake States and their management. USDA Forest Service Research Paper NC NC-31. Auchmoody, L.R. and Smith, H.C. (1993) Survival of northern red oak acorns after fall burning. USDA Forest Service Research Paper NE NE-678. Aust, W.M., Hodges, J.D. and Johnson, R.L. (1985) The origin, growth and development of natural, pure, even-aged stands of bottomland oak. USDA Forest Service General Technical Report SO SO-54, pp. 163–170. Bakken, P.N. and Cook, J.E. (1998) Regeneration potential of six habitat types common to northcentral Wisconsin. Northern Journal of Applied Forestry 15, 116–123. Barnhart, S.J., McBride, J.R. and Warner, P. (1991) Oak seedling establishment in relation to environmental factors at Annadel State Park. USDA Forest Service General Technical Report PSW PSW126, pp. 25–30. Beck, D.E. (1970) Effect of competition on survival and height growth of red oak seedlings. USDA Forest Service Research Paper SE SE-56. Beck, D.E. and Hooper, R.M. (1986) Development of a southern Appalachian hardwood stand after clearcutting. Southern Journal of Applied Forestry 10, 168–172. Beilmann, A.P. and Brenner, L.G. (1951) The recent intrusion of forests in the Ozarks. Annals of the Missouri Botanical Garden 38, 261–282. Belcher, D.M., Holdaway, M.R. and Brand, G.J. (1982) A description of STEMS, the stand and tree evaluation and modeling system. USDA Forest Service General Technical Report NC NC-79. Blake, J.I., Somers, G.L. and Ruark, G.A. (1990) Perspectives on process modeling of forest growth responses to environmental stress. In: Dixon, R.K., Meldahl, R.S., Ruark, G.A. and Warren, W.G. (eds) Process Modeling of Forest Growth Responses to Environmental Stress. Timber Press, Portland, Oregon, pp. 9–17. Borchert, M.I., Davis, F.W., Michaelsen, J. and Oyler, L.D. (1989) Interactions of factors affecting seedling recruitment of blue oak (Quercus douglasii) in California. Ecology 70, 389–404. Bormann, F.H. and Likens, G.E. (1979) Pattern and Process in a Forested Ecosystem. Springer-Verlag, New York.

160

Chapter 3

Botkin, D.B. (1993) Forest Dynamics, an Ecological Model. Oxford University Press, New York. Bowersox, T.W. and McCormick, L.H. (1987) Herbaceous communities reduce the juvenile growth of northern red oak, white ash, yellow poplar, but not white pine. Proceedings of the Central Hardwood Forest Conference VI. University of Tennessee, Knoxville, pp. 39–42. Braun, E.L. (1972) Deciduous Forests of Eastern North America. Hafner, New York. Bruce, D. (1990) Development of empirical forest growth models. In: Dixon, R.K., Meldahl, R.S., Ruark, G.E. and Warren, W.G. (eds) Process Modeling of Forest Growth Responses to Environmental Stress. Timber Press, Portland, Oregon, pp. 191–199. Buchman, R.G. and Shiﬂey, S.R. (1983) Guide to evaluating forest growth projection systems. Journal of Forestry 81, 232–234, 254. Burns, R.M. and Honkala, B.H. (tech. coords.) (1990) Silvics of North America. USDA Forest Service Agriculture Handbook 654, Vol. 2. Callaway, R.M. (1992) Effect of shrubs on recruitment of Quercus douglasii and Quercus lobata in California. Ecology 72, 2118–2128. Callaway, R.M. and D’Antonio, C.M. (1991) Shrub facilitation of coast live oak establishment in central California. Madroño 38, 158–169. Callaway, R.M. and Mahall, B.E. (1996) Variation in leaf structure and function in Quercus douglasii trees differing in root architecture and drought history. International Journal of Plant Sciences 157, 129–135. Callaway, R.M., Nadkarni, N.M. and Mahall, B.E. (1991) Facilitation and interference of Quercus douglasii on understory productivity in central California. Ecology 72, 1484–1499. Canadell, J. and Rodà, F. (1991) Root biomass of Quercus ilex in a montane Mediterranean forest. Canadian Journal of Forest Research 21, 1771–1778. Canham, C.D. (1989) Different responses to gaps among shade-tolerant tree species. Ecology 69, 548–550. Carmean, W.H. (1971) Site index curves for black, white, scarlet and chestnut oaks in the Central States. USDA Forest Service Research Paper NC NC-62. Carvell, K.L. and Tryon, E.H. (1961) The effect of environmental factors on the abundance of oak regeneration beneath mature oak stands. Forest Science 7, 98–105. Cecich, R.A. (1991) Seed production in oak. Proceedings of The Oak Resource in the Upper Midwest Conference. University of Minnesota, St Paul, pp. 125–131. Christisen, D.M. and Kearby, W.H. (1984) Mast measurement and production in Missouri (with special reference to acorns). Missouri Department of Conservation Terrestrial Series 13. Cleland, D.T., Hart, J.B., Host, G.E., Pregitzer, K.S. and Ramm, C.W. (1993) Ecological Classiﬁcation and Inventory System of the Huron–Manistee National Forests. USDA Forest Service, Washington, DC. Connell, J.H. and Slatyer, R.O. (1977) Mechanisms of succession in natural communities and their role in community stability and organization. American Naturalist 111, 1119–1144. Crow, T.R. (1988) Reproductive mode and mechanisms for self-replacement of northern red oak (Quercus rubra) – a review. Forest Science 34, 19–40. Crow, T.R. (1992) Population dynamics and growth patterns for a cohort of northern red oak (Quercus rubra) seedlings. Oecologia 91, 192–200. Curtis, J.T. (1959) The Vegetation of Wisconsin. University of Wisconsin Press, Madison. Danielsen, K.C. and Halvorson, W.L. (1991) Valley oak seedling growth associated with selected grass species. USDA Forest Service General Technical Report PSW PSW-126, pp. 9–13. Davis, F.W., Borchert, M., Harvey, L.E. and Michaelsen, J.C. (1991) Factors affecting seedling survivorship of blue oak (Quercus douglasii H. & A.) in central California. USDA Forest Service General Technical Report PSW PSW-126, pp. 81–86. Day, G.M. (1953) The Indian as an ecological factor in the northeastern forest. Ecology 34, 329–346. DeSelm, H.R., Clebsch, E.E.C. and Bennie, J.C. (1991) Effects of 27 years of prescribed ﬁre on an oak forest and its soils in middle Tennessee. USDA Forest Service General Technical Report SE SE-70, Vol. 1, pp. 409–417. Dey, D.C. (1991) A comprehensive Ozark regenerator. PhD dissertation, University of Missouri, Columbia. Dey, D.C. and Parker, W.C. (1996) Regeneration of red oak (Quercus rubra L.) using shelterwood systems: ecophysiology, silviculture and management recommendations. Ontario Forest Research Institute Forest Research Information Paper 126. Dey, D.C., Johnson, P.S. and Garrett, H.E. (1996a) Modeling the regeneration of oak stands in the Missouri Ozark Highlands. Canadian Journal of Forest Research 26, 573–583.

Regeneration Ecology II: Population Dynamics

161

Dey, D.C., Ter-Mikaelian, M., Johnson, P.S. and Shiﬂey, S.R. (1996b) Users’ guide to ACORn: a comprehensive Ozark regeneration simulator. USDA Forest Service General Technical Report NC NC-180. Dixon, R.K., Meldahl, R.S., Ruark, G.A. and Warren, W.G. (eds) (1990) Process Modeling of Forest Growth Responses to Environmental Stress. Timber Press, Portland, Oregon. Downs, A.A. (1944) Estimating acorn crops for wildlife in the southern Appalachians. Journal of Wildlife Management 8, 339–340. Dunn, C.P., Guntenspergen, G.R. and Dorney, J.R. (1983) Catastrophic wind disturbance in an oldgrowth hemlock–hardwood forest, Wisconsin. Canadian Journal of Botany 61, 211–217. Egler, F.E. (1954) Vegetation science concepts I. Initial ﬂoristic composition. A factor in old-ﬁeld vegetation management. Vegetatio 4, 412–417. Ehrenfeld, J.G. (1980) Understory response to canopy gaps of varying size in a mature oak forest. Bulletin of the Torrey Botanical Club 107, 29–41. Farmer, R.E., Jr (1975) Growth and assimilation rate of juvenile northern red oak: effects of light and temperature. Forest Science 21, 373–381. Fischer, B.C. (1979) Managing light in the selection method. Proceedings 1979 J.S. Wright Forestry Conference. Purdue University, West Lafayette, Indiana, pp. 43–53. Franklin, J.F., Perry, D.A., Schowalter, T.D., Harmon, M.E., McKee, A. and Spies, T.A. (1989) Importance of ecological diversity in maintaining long-term site productivity. In: Perry, D.A., Meurisse, R., Thomas, B. et al. (eds) Maintaining the Long-term Productivity of Paciﬁc Northwest Forest Ecosystems. Timber Press, Portland, Oregon, pp. 82–97. Galford, J.R., Peacock, J.W. and Wright, S.L. (1988) Insects and other pests affecting oak regeneration. Society of American Foresters Publication 88-03, pp. 219–225. Galford, J., Auchmoody, L.R., Smith, H.C. and Walter, R.S. (1991) Insects affecting establishment of northern red oak seedlings in central Pennsylvania. USDA Forest Service General Technical Report NE NE-148, pp. 271–280. Gatto, M. and Rinaldi, S. (1987) Some models of catastrophic behaviour in exploited forests. Vegetatio 69, 213–222. Gingrich, S.F. (1967) Measuring and evaluating stocking and stand density in upland hardwood forests in the Central States. Forest Science 13, 38–53. Glenn-Lewin, D.C. and van der Maarel, E. (1992) Patterns and processes of vegetation dynamics. In: Glenn Lewin, D.C., Peet, R.K. and Veblen, T.T. (eds) Plant Succession, Theory and Prediction. Chapman & Hall, New York, pp. 11–59. Godman, R.M., Yawney, H.W. and Tubbs, C.H. (1990) Acer saccharum Marsh. Sugar maple. USDA Forest Service Agriculture Handbook 654, Vol. 2, pp. 78–91. Golden, M.S. and Loewenstein, E.F. (1991) Regeneration of tree species 7 years after clearcutting in a river bottom in central Alabama. USDA Forest Service General Technical Report SE SE-70, Vol. 1, pp. 76–83. Gordon, D.R., Welker, J.M., Menke, J.W. and Rice, K.J. (1989) Competition for soil water between annual plants and blue oak (Quercus douglasii) seedlings. Oecologia 79, 533–541. Gottschalk, K.W. (1985) Effects of shading on growth and development of northern red oak, black oak, black cherry, and red maple seedlings. I. Height, diameter, and root/shoot ratio. Proceedings Central Hardwood Forest Conference V. University of Illinois, Urbana–Champaign, pp. 189–195. Gottschalk, K.W. (1987) Effects of shading on growth and development of northern red oak, black oak, black cherry, and red maple seedlings. II. Biomass partitioning and prediction. Central Hardwood Forest Conference VI. University of Tennessee, Knoxville, pp. 99–110. Gottschalk, K.W. (1990) Gypsy moth effects on mast production. Proceedings Workshop on Southern Mast Management. University of Tennessee, Knoxville, pp. 42–50. Grimm, E.C. (1984) Fire and other factors controlling the Big Woods vegetation of Minnesota in the mid-nineteenth century. Ecological Monographs 54, 291–311. Grubb, P.J. (1977) The maintenance of species-richness in plant communities: the importance of the regeneration niche. Biological Reviews 52, 107–145. Guyette, R.P. and Cutter, B.E. (1991) Tree-ring analysis of ﬁre history of a post oak savanna in the Missouri Ozarks. Natural Areas Journal 11(2), 93–99. Guyette, R.P. and Dey, D.C. (1995) A history of ﬁre, disturbance, and growth in a red oak stand in the Bancroft District, Ontario. Ontario Forest Research Institute Forest Research Information Paper, Paper 119.

162

Chapter 3

Haney, A. and Apfelbaum, S.I. (1990) Structure and dynamics of Midwest oak savannas. In: Sweeney, J.M. (ed.) Management of Dynamic Ecosystems. The Wildlife Society, North Central Section, West Lafayette, Indiana, pp. 19–30. Hannah, P.R. (1987) Regeneration methods for oaks. Northern Journal of Applied Forestry 4, 97–101. Hanson, P.J. and Dixon, R.K. (1985) Allelopathic inhibition of northern red oak by interrupted fern and goldenrod. Central Hardwood Forest Conference V. University of Illinois, UrbanaChampaign, pp. 269–274. Hanson, P.J., Isebrands, J.G. and Dickson, R.E. (1987) Carbon budgets of Quercus rubra L. seedlings at selected stages of growth: Inﬂuence of light. Proceedings of the Central Hardwood Forest Conference VI. University of Tennessee, Knoxville, pp. 269–276. Harmon, M.E. (1984) Survival of trees after low-intensity surface ﬁres in Great Smoky Mountains National Park. Ecology 65, 796–802. Harper, K.T., Wagstaff, F.J. and Kunzler, L.M. (1985) Biology and management of the Gambel oak vegetative type: a literature review. USDA Forest Service General Technical Report INT INT-179. Harrison, J.S. and Werner, P.A. (1984) Colonization by oak seedlings into a heterogeneous successional habitat. Canadian Journal of Botany 62, 559–563. Hengst, G.E. and Dawson, J.O. (1993) Bark thermal properties of selected central hardwood species. USDA Forest Service General Technical Report NC NC-161, pp. 55–75. Hilt, D.E. (1985) OAKSIM: an individual-tree growth and yield simulator for managed, even-aged, upland oak stands. USDA Forest Service Research Paper NE NE-562. Hix, D.M. (1988) Multifactor classiﬁcation and analysis of upland hardwood forest ecosystems of the Kickapoo River watershed, southwestern Wisconsin. Canadian Journal of Forest Research 18, 1405–1415. Hodges, J.D. and Gardiner, E.S. (1993) Ecology and physiology of oak regeneration. USDA Forest Service General Technical Report SE SE-84, pp. 54–65. Horsley, S.B. (1991) Using Roundup and Oust to control interfering understories in Allegheny hardwood stands. USDA Forest Service General Technical Report NE NE-148, pp. 281–290. Host, G.E. and Pregitzer, K.S. (1991) Ecological species groups for upland forest ecosystems of northwestern Lower Michigan. Forest Ecology and Management 43, 87–102. Huddle, J.A. and Pallardy, S.G. (1996) Effects of long-term annual periodic burning on tree survival and growth in a Missouri Ozark oak–hickory forest. Forest Ecology and Management 82, 1–9. Isebrands, J.G., Rauscher, H.M., Crow, T.R. and Dickmann, D.I. (1990) Whole-tree growth process models based on structural-functional relationships. In: Dixon, R.K., Meldahl, R.S., Ruark, G.A. and Warren, W.G. (eds) Process Modeling of Forest Growth Responses to Environmental Stress. Timber Press, Portland, Oregon, pp. 96–112. Jarvis, P.G. (1964) Interference by Deschampsia ﬂexuosa (L.) Trin. Oikos 15, 56–78. Johnson, P.S. (1966) Vegetational developments after logging in southern Michigan xeric oak forests. PhD dissertation, Michigan State University, East Lansing. Johnson, P.S. (1974) Survival and growth of northern red oak seedlings following a prescribed burn. USDA Forest Service Research Note NC NC-177. Johnson, P.S. (1975) Growth and structural development of red oak sprout clumps. Forest Science 21, 413–418. Johnson, P.S. (1976) Modal development of regeneration in clearcut red oak stands in the Driftless Area. Proceedings of the Central Hardwood Forest Conference I. University of Southern Illinois, Carbondale, pp. 455–475. Johnson, P.S. (1977) Predicting oak stump sprouting and sprout development in the Missouri Ozarks. USDA Forest Service Research Paper NC NC-149. Johnson, P.S. (1979) Shoot elongation of black oak and white oak sprouts. Canadian Journal of Forest Research 9, 489–494. Johnson, P.S. (1992) Oak overstory/reproduction relations in two xeric ecosystems in Michigan. Forest Ecology and Management 48, 233–248. Johnson, P.S. (1993) Perspectives on the ecology and silviculture of oak-dominated forests in the central and eastern states. USDA Forest Service General Technical Report NC NC-153. Johnson, P.S. and Rogers, R. (1984) Predicting 25th-year diameters of thinned stump sprouts of northern red oak. Journal of Forestry 82, 616–619. Johnson, P.S., Jacobs, R.D., Martin, A.J. and Godell, E.D. (1989) Regenerating northern red oak: three successful case histories. Northern Journal of Applied Forestry 6, 174–178.

Regeneration Ecology II: Population Dynamics

163

Johnson, R.L. (1975) Natural regeneration and development of Nuttall oak and associated species. USDA Forest Service Research Paper SO SO-104. Johnson, R.L. (1981) Oak seeding – it can work. Southern Journal of Applied Forestry 5, 28–33. Johnson, R.L. and Deen, R.T. (1993) Prediction of oak regeneration in bottomland forests. USDA Forest Service General Technical Report SE SE-84, pp. 146–155. Johnson, R.L. and Krinard, R.M. (1983) Regeneration in small and large sawtimber sweetgum–red oak stands following selection and seed tree harvest: 23-year results. Southern Journal of Applied Forestry 7, 176–184. Johnson, W.C. and Webb, T. III. (1989) The role of blue jays (Cyanocitta cristata L.) in the postglacial dispersal of fagaceous trees in eastern North America. Journal of Biogeography 16, 561–571. Jørgensen, S.E. (1990) Ecosystem theory, ecological buffer capacity, uncertainty and complexity. Ecological Modelling 52, 125–133. Keeley, J.E. (1991) Seed germination and life history syndromes in the California chaparral. Botanical Review 57, 81–116. Keeley, J.E. (1992) Recruitment of seedlings and vegetative sprouts in unburned chaparral. Ecology 73, 1194–1208. Kimmins, J.P. (1987) Forest Ecology. Macmillan, New York. Kolb, T.E. and Steiner, K.C. (1990) Growth and biomass partitioning response of northern red oak genotypes to shading and grass root competition. Forest Science 36, 293–303. Korstian, C.F. (1927) Factors controlling germination and early survival in oaks. Yale University School Forestry Bulletin 19. Kotar, J., Kovach, J.A. and Locey, C.T. (1988) Field Guide to Forest Habitat Types of Northern Wisconsin. University of Wisconsin and Wisconsin Department of Natural Resources, Madison. Krajicek, J.E. (1960) Some factors affecting oak and black walnut reproduction. Iowa State Journal of Science 34, 631–634. Lamson, N.I. (1988) Role of stump sprouts in regenerating Appalachian hardwood stands. Society of American Foresters Publication 88-03, 31–37. Landsberg, J.J. (1986) Physiological Ecology of Forest Production. Academic Press, New York. Langdon, O.G. (1981) Some effects of prescribed ﬁre on understory vegetation in loblolly pine stands. Symposium Proceedings: Prescribed Fire and Wildlife in Southern Forests. Clemson University, Charleston, South Carolina, pp. 143–153. Latham, R.E. (1992) Co-occurring tree species change rank in seedling performance with resources varied experimentally. Ecology 73, 2129–2144. Levins, R. (1966) The strategy of model building in population biology. American Scientist 54, 421–431. Lewis, R.W. (1990) Theories, speculation, and the structure of knowledge. Speculations in Science and Technology 13, 1–17. Liming, F.G. and Johnston, J.P. (1944) Reproduction in oak–hickory forest stands of the Missouri Ozarks. Journal of Forestry 42, 175–180. Loach, K. (1967) Shade tolerance in tree seedlings. 1. Leaf photosynthesis and respiration in plants raised under artiﬁcial shade. New Phytologist 66, 607–621. Loewenstein, E.F. (1992) Relationships between stem origin and growth characteristics of three bottomland hardwood species on a seven-year-old clearcut area in the Coastal Plain of Alabama. MS Thesis, Auburn University, Auburn, Alabama. Loewenstein, E.F. and Golden, M.S. (1995) Establishment of water oak is not dependent on advance reproduction. USDA Forest Service General Technical Report SRS SRS-1, pp. 443–446. Loftis, D.L. (1983) Regenerating southern Appalachian mixed hardwood stands with the shelterwood method. Southern Journal of Applied Forestry 7, 212–217. Loftis, D.L. (1988a) Regenerating oaks on high-quality sites, an update. Society of American Foresters Publication 88-03, pp. 199–209. Loftis, D.L. (1988b) Regenerating red oak in the southern Appalachians: predictive models and practical applications. PhD dissertation, North Carolina State University, Raleigh. Loftis, D.L. (1990a) Predicting post-harvest performance of advance red oak reproduction in the southern Appalachians. Forest Science 36, 908–916. Loftis, D.L. (1990b) A shelterwood method for regenerating red oak in the southern Appalachians. Forest Science 36, 917–929. Loomis, R.M. (1973) Estimating ﬁre-caused mortality and injury in oak–hickory forests. USDA Forest Service Research Paper NC NC-94.

164

Chapter 3

Loomis, R.M. (1974) Predicting the losses in sawtimber volume and quality from ﬁres in oak–hickory forests. USDA Forest Service Research Paper NC NC-104. Lorimer, C.G. (1981) Survival and growth of understory trees in oak forests of the Hudson Highlands, New York. Canadian Journal of Forest Research 11, 689–695. Lorimer, C.G. (1983) Eighty-year development of northern red oak after partial cutting in a mixedspecies Wisconsin forest. Forest Science 29, 371–383. Lorimer, C.G. (1989) The oak regeneration problem: new evidence on causes and possible solutions. University of Wisconsin Forest Resource Analyses No. 8. Lorimer, C.G. (1993) Causes of the oak regeneration problem. USDA Forest Service General Technical Report SE SE-84, pp. 14–39. Lorimer, C.G., Chapman, J.W. and Lambert W.D. (1994) Tall understorey vegetation as a factor in the poor development of oak seedlings beneath mature stands. Journal of Ecology 82, 227–237. Margalef, R. (1963) On certain unifying principles in ecology. American Naturalist 97, 357–374. Marks, P.L. (1974) The role of pin cherry (Prunus serotina L.) in the maintenance of stability in northern hardwood ecosystems. Ecological Monographs 44, 73–88. Marquis, D.A. and Johnson, R.L. (1989) Silviculture of eastern hardwoods. USDA Forest Service General Technical Report WO WO-55, pp. 9–15. Marquis, D.A., Ernst, R.L. and Stout, S.L. (1992) USDA Forest Service General Technical Report NE NE-96. Marquis, D.A., Eckert, P.L. and Roach, B.A. (1976) Acorn weevils, rodents, and deer all contribute to oak-regeneration difﬁculties in Pennsylvania. USDA Forest Service Research Paper NE NE-356. Martin, R.E. (1962) Thermal properties of bark. Forest Products Journal 13, 419–426. Maslen, P. (1989) Response of immature oaks to prescribed ﬁre in the North Carolina Piedmont. USDA Forest Service General Technical Report SO SO-74, pp. 259–266. Matney, T.G. (1974) Oak regeneration in natural upland oak stands of West Virginia. MS thesis, West Virginia University, Morgantown. Matsuda, K. and McBride, J.R. (1986) Difference in seedling growth morphology as a factor in the distribution of three oaks in Central California. Madroño 33, 207–216. Matsuda, K., McBride, J.R. and Kimura, M. (1989) Seedling growth form of oaks. Annals of Botany 64, 439–446. McClaran, M.P. and Bartolome, J.W. (1989) Fire-related recruitment in stagnant Quercus douglasii populations. Canadian Journal of Forest Research 19, 580–585. McGee, C.E. (1987) Clearcutting in upland hardwoods: panacea or anathema? Proceedings of the Central Hardwood Forest Conference VI. University of Tennessee, Knoxville, pp. 21–29. McGee, C.E. (1988) Spring weather, canopy removal, and early budbreak threaten oak seedlings. Journal of the Elisha Mitchell Science Society 104, 108–115. McGill, D.W., Rogers, R., Martin, A.J. and Johnson, P.S. (1999) Measuring stocking in northern red oak stands in Wisconsin. Northern Journal of Applied Forestry 16, 144–150. McLemore, B.F. (1990) Cornus ﬂorida L. ﬂowering dogwood. USDA Forest Service Agriculture Handbook 654, Vol. 2, pp. 278–283. McQuilkin, R.A. (1983) Silvical factors affecting the regeneration of oaks and associated species in Pennsylvania. Proceedings of Regenerating Hardwood Stands. Pennsylvania State University, University Park, pp. 37–65. Merz, R.W. and Boyce, S.G. (1956) Age of oak ‘seedlings’. Journal of Forestry 54, 774–775. Minckler, L.S. and Woerheide, J.D. (1965) Reproduction of hardwoods 10 years after cutting as affected by site and opening size. Journal of Forestry 63, 103–107. Muller, C.H. (1951) The signiﬁcance of vegetative reproduction in Quercus. Madroño 11, 129–137. Myers, S.A. (1978) Insect impact on acorn production in Missouri upland forests. PhD dissertation, University of Missouri, Columbia. Nowacki, G.J., Abrams, M.D. and Lorimer, C.G. (1990) Composition, structure, and historical development of northern red oak stands along an edaphic gradient in north-central Wisconsin. Forest Science 36, 276–292. Nyland, R.D., Abrahamson, L.P. and Adams, K.B. (1983) Use of prescribed ﬁre for regenerating red and white oak in New York. Proceedings of the 1982 Society of American Foresters National Convention, pp. 163–167. Olson, D.F., Jr (1974) Quercus L. Oak. USDA Forestry Services Agriculture Handbook 450, 88pp.

Regeneration Ecology II: Population Dynamics

165

Overton, R.P. (1990) Oxydendron arboreum (L.) DC Sourwood. USDA Forest Service Agriculture Handbook 654, Vol. 2, pp. 497–500. Pallardy, S.G. and Rhoads, J.L. (1993) Morphological adaptations to drought in seedlings of deciduous angiosperms. Canadian Journal of Forest Research 23, 1766–1774. Pavlik, B.M., Muick, P.C., Johnson, S. and Popper, M. (1991) Oaks of California. Cachuma Press, Los Olivos, California. Platt, W.J., Glitzenstein, J.S. and Streng, D.R. (1991) Evaluating pyrogenicity and its effects on vegetation in longleaf pine savannas. Proceedings 17th Tall Timbers Fire Ecology Conference, pp. 143–161. Powell, D.S. (1976) Sprouting ability of advance reproduction of undisturbed forest stands in West Virginia. MS thesis, West Virginia University, Morgantown. Prentice, I.C. and Helmisaari, H. (1991) Silvics of north European trees: compilation, comparisons and implications for forest succession modelling. Forest Ecology and Management 42, 79–93. Rebertus, A.J., Williamson, G.B. and Moser, E.B. (1989) Longleaf pine pyrogenicity and turkey oak mortality in Florida xeric sandhills. Ecology 70, 60–70. Reich, P.B., Abrams, M.D., Ellsworth, D.S., Kruger, E.L. and Tabone, T.J. (1990) Fire affects ecophysiology and community dynamics of central Wisconsin oak forest regeneration. Ecology 71, 2179–2190. Renton, J.J., Lanasa, M.J. and Tryon, E.H. (1974) Radiography for observing wood features. Journal of Forestry 72, 292–293. Rice, C. and Struve, D.K. (1997) Seedling growth form and water use of selected oak species. Diversity and Adaptation in Oak Species, Proceedings of Working Party 2.08.05, Oct. 12–17, 1997, Genetics of Quercus, of the International Union of Forest Research Organizations. Pennsylvania State University, University Park 2, pp. 269–278. Riegel, G.M., Smith, B.G. and Franklin, J.F. (1992) Foothill oak woodlands of the Interior Valleys of southwestern Oregon. Northwest Science 66(2), 66–76. Roach, B.A. and Gingrich, S.F. (1968) Even-aged silviculture for upland central hardwoods. USDA Forest Service Agriculture Handbook 335. Rogers, R. and Johnson, P.S. (1998) Approaches to modeling natural regeneration in oak-dominated forests. Forest Ecology and Management 106, 45–54. Rogers, R., Johnson, P.S. and Loftis, D.L. (1993) An overview of oak silviculture in the United States: the past, present, and future. Annales des Sciences Forestieres 50, 535–542. Ross, M.S., Sharik, T.L. and Smith, D.W. (1986) Oak regeneration after clear felling in southwest Virginia. Forest Science 32, 157–169. Roth, E.R. and Sleeth, B. (1939) Butt rot in unburned sprout oak stands. USDA Technical Bulletin 684. Rouse, C. (1986) Fire effects in northeastern forests: oak. USDA Forest Service General Technical Report NC NC-105. Sander, I.L. (1971) Height growth of new oak sprouts depends on size of advance reproduction. Journal of Forestry 69, 809–811. Sander, I.L. (1979a) Regenerating oaks. Proceedings National Silviculture Workshop. USDA Forest Service, Washington, DC, pp. 212–221. Sander, I.L. (1979b) Regenerating oaks with the shelterwood system. Proceedings of 1979 J.S. Wright Forestry Conference. Purdue University, West Lafayette, IN, pp. 54–60. Sander, I.L. (1983) Oak regeneration in the Central States. Proceedings of the 11th Annual Hardwood Symposium of the Hardwood Research Council, pp. 18–31. Sander, I.L. (1990) Quercus rubra L. Northern red oak. USDA Forest Service Agriculture Handbook 654, Vol. 2, pp. 727–733. Sander, I.L. and Clark, F.B. (1971) Reproduction of upland hardwood forests in the Central States. USDA Forest Service Agriculture Handbook 405. Sander, I.L., Johnson, P.S. and Rogers, R. (1984) Evaluating oak advance reproduction in the Missouri Ozarks. USDA Forest Service Research Paper NC NC-251. Scholz, H.F. (1955) Growth of northern red oak seedlings under variable conditions of ground cover competition. USDA Forest Service Lake States Forest Experiment Station Technical Note 430. Scholz, H.F. (1959) Further observations on seedbed scariﬁcation show beneﬁts to northern red oak were temporary. USDA Forest Service Lake States Forest Experiment Station Technical Note 555.

166

Chapter 3

Sharpe, P.J.H. (1990) Forest modeling approaches: compromises between generality and precision. In: Dixon, R.K., Meldahl, R.S., Ruark, G.A. and Warren, W.G. (eds) Process Modeling of Forest Growth Responses to Environmental Stress. Timber Press, Portland, Oregon, pp. 180–190. Shugart, H.H. (1984) A Theory of Forest Dynamics. Springer-Verlag, New York. Smalley, G.W. (1978) Classiﬁcation and evaluation of forest sites for the Interior Highlands. Proceedings of the Central Hardwood Forestry Conference II. Purdue University, West Lafayette, IN, 257 pp. Smalley, G.W. (1984) Classiﬁcation and evaluation of forest sites in the Cumberland Mountains. USDA Forest Service General Technical Report SO SO-50. Smith, D.M. (1986) The Practice of Silviculture, 8th edn. John Wiley & Sons, New York. Smith, H.C. and Miller, G.W. (1987) Managing Appalachian hardwood stands using four regeneration practices – 34-year results. Northern Journal of Applied Forestry 4, 180–185. Smith, L.L. and Vankat, J.L. (1991) Communities and tree seedling distribution in Quercus rubra- and Prunus serotina-dominated forests in southwestern Pennsylvania. American Midlands Naturalist 126, 294–307. Sokal, R.R. and Rohlf, F.J. (1969) Biometry: the Principles and Practices of Statistics in Biological Research. W.H. Freeman, San Francisco. Sork, V.L. (1984) Examination of seed dispersal and survival in red oak, Quercus rubra (Fagaceae), using metal-tagged acorns. Ecology 65, 1020–1022. Sork, V.L., Stacey, P. and Averett, J.E. (1983) Utilization of red oak acorns in non-bumper crop year. Oecologia 59, 49–53. Spurr, S.H. and Barnes, B.V. (1980) Forest Ecology, 3rd edn. Ronald Press, New York. Steiner, K.C., Abrams, M.D. and Bowersox, T.W. (1993) Advance reproduction and other stand characteristics in Pennsylvania and French stands of northern red oak. USDA Forest Service General Technical Report NC NC-161, pp. 473–483. Streng, D.R., Glitzenstein, J.S. and Harcombe, P.A. (1989) Woody seedling dynamics in an east Texas ﬂoodplain forest. Ecological Monographs 59, 177–204. Sullivan, N.H. (2001) An algorithm for a landscape model of mast production. PhD dissertation, University of Missouri, Columbia. Terborgh, J. (1985) The vertical component of plant species diversity in temperate tropical forests. American Naturalist 126, 760–776. Thilenius, J.F. (1968) The Quercus garryana forests of the Willamette Valley, Oregon. Ecology 49, 1124–1133. Thor, E. and Nichols, G.M. (1973) Some effects of ﬁres on litter, soil, and hardwood regeneration. Proceedings of the 13th Tall Timbers Fire Ecology Conference, pp. 317–329. Tiedemann, A.R., Clary, W.P. and Barbour, R.J. (1987) Underground systems of Gambel oak (Quercus gambelii) in central Utah. American Journal of Botany 74, 1065–1071. Trimble, G.R., Jr (1960) Relative diameter growth rates of ﬁve upland oaks in West Virginia. Journal of Forestry 58, 111–115. Trimble, G.R., Jr (1973) The regeneration of Central Appalachian hardwoods with emphasis on the effects of site quality and harvesting practice. USDA Forest Service Research Paper NE NE-282. Trimble, G.R., Jr and Tryon, E.H. (1966) Crown encroachment into openings cut in Appalachian hardwood stands. Journal of Forestry 64, 104–108. Tryon, E.H. and Carvell, K.L. (1958) Regeneration under oak stands. West Virginia University Agricultural Experiment Station Bulletin 424T. Urban, D.L. and Shugart, H.H. (1992) Individual-based models of forest succession. In: Glenn-Lewin, D.C., Peet, R.K. and Veblen, C.C. (eds) Plant Succession, Theory and Prediction. Chapman & Hall, New York, pp. 249–292. Van Lear, D.H. and Watt, J.M. (1993) The role of ﬁre in oak regeneration. USDA Forest Service General Technical Report SE SE-84, pp. 6–78. Veblen, T.T. (1992) Regeneration dynamics. In: Glenn-Lewin, D.C., Peet, R.K. and Veblen, T.T. (eds) Plant Succession, Theory and Prediction. Chapman and Hall, New York, pp. 152–187. Waldrop, T.A., Buckner, E.R., Shugart, H.H., Jr and McGee, C.E. (1986) FORCAT: a single tree model of stand development following clearcutting on the Cumberland Plateau. Forest Science 32, 297–317. 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. Oecologia 94, 7–16.

Regeneration Ecology II: Population Dynamics

167

Walters, R.S. (1990) Site quality, ﬁre, and herbicide effects on establishment, growth, and development of regeneration three years after partial cutting of oak stands. PhD dissertation, State University of New York, Syracuse. Ward, J.S. and Stephens, G.R. (1994) Crown class transition rates of maturing northern red oak (Quercus rubra L.). Forest Science 40, 221–237. Watt, A.S. (1947) Pattern and process in the plant community. Journal of Ecology 35, 1–22. Weigel, D.R. and Johnson, P.S. (1998) Stump sprouting probabilities for Indiana oaks. USDA Forest Service Technical Brief TB-NC TB-NC–7. Welker, J.M. and Menke, J.W. (1990) The inﬂuence of simulated browsing on tissue water relations, growth and survival of Quercus douglasii (Hook and Arn.) seedlings under slow and rapid rates of soil drought. Functional Ecology 4, 807–817. Wendel, G.W. (1975) Stump sprout growth and quality of several Appalachian hardwood species after clearcutting. USDA Forest Service Research Paper NE NE-329. Wendel, G.W. (1977) Longevity of black cherry, wild grape, and sassafras seed in the forest ﬂoor. USDA Forest Service Research Paper NE NE-375. Wendel, G.W. (1990) Prunus pensylvanica L. f. Pin cherry. USDA Forest Service Agriculture Handbook 654, Vol. 2, pp. 587–593. Whitney, G.G. (1986) A demographic analysis of Rubus idaeus and Rubus pubescens. Canadian Journal of Botany 64, 2916–2921. Whitney, G.G. (1994) From Coastal Wilderness to Fruited Plain. Cambridge University Press, Cambridge, UK. Williams, K., Davis, S.D., Gartner, B.L. and Karlsson, S. (1991) Factors limiting the establishment of a chaparral oak, Quercus durata Jeps., in grassland. USDA Forest Service General Technical Report PSW PSW-126, pp. 70–73. Williamson, G.B. and Black, E.M. (1981) High temperature of forest ﬁres under pines as a selective advantage over oaks. Nature 293, 643–644. Will-Wolf, S. (1991) Role of ﬁre in maintaining oaks in mesic oak maple forests. Proceedings of the Oak Resource in the Upper Midwest Conference. University of Minnesota, St Paul, pp. 27–33. Wright, S.L. (1987) Managing insects affecting oak regeneration by prescribed burning. USDA Forest Service General Technical Report SE SE-46, pp. 186–192.

4 Site Productivity

Introduction Silviculturists use the term site to refer to an area of forested land that is qualitatively characterized by its climate, soil, vegetation, or quantitatively by its productivity. The latter is usually expressed as potential wood production per unit land area per unit of time (Helms, 1998). Climate, soil, vegetation and productivity characteristics may be used singly or in combination to deﬁne site productivity. More generally, site productivity refers to the ability of a deﬁned area to produce objects and their attributes that are required for living organisms or human society to function (Lee, 1989). Lee (1989) used this deﬁnition to refer to ‘resource site quality’. The researcher’s view of site quality may differ from the land manager’s, and Gholz (1988) pointed out that there is no universally accepted deﬁnition. The two prevalent schools of thought nevertheless consider site productivity as either: (i) a hypothetical, optimal or potential level of productivity, or (ii) an index or relative measure of actual productivity. Although production capability is sometimes called ‘yield’, production and yield are not the same (Smith, 1986). Yield usually refers to material that is usable. In the case of timber, yield refers to the amount of wood that is actually harvested and removed from the site. In contrast, production usually refers to all material that has resulted from tree growth whether or not it is harvested and removed (Zahner and Myers, 1984). 168

Both production and yield on a given site are affected by forest conditions. For a given species or species mix, conditions that commonly reduce stand yields below their maximum potential include low stocking, insect or disease damage, high grading and tree defects. Both the actual and potential production and yield are of interest to silviculturists. Site productivity inﬂuences stand development not only in its effect on growth and yield, but also on regeneration and other ecological processes. Site also inﬂuences stand development such as rates of change in numbers of trees, basal area and average tree diameter. For a given age, undisturbed stands on good sites have fewer trees, but higher basal areas and average stem diameters than do stands on poor sites. This results from the faster growth of trees on good sites. For example, oaks on poor sites (site index 50) in the Piedmont were almost 4 inches smaller in mean dbh at stand age 50 (6.7 inches dbh) than oaks growing on good sites (site index 90) (10.4 inches dbh). The poor-site stands also carried lower basal areas (94 ft2 acre1 vs. 113 ft2 acre1) and twice the number of trees (900 acre−1 vs. 450 acre−1). The amount of clear lumber that develops in an oak largely depends on its rate of height growth and thus on site quality. Rapid height growth increases clear bole length and minimizes the size of the knotty core – and thus increases tree value (Carmean and Boyce, 1973; see also Chapter 10). Site quality also affects other wood properties including hardness, shrinkage, strength and yield of cellulose (Zahner, 1970).

Site Productivity

Measures of Site Productivity Site productivity can be measured and expressed in various ways. Silviculturists commonly estimate the periodic increment of wood or timber production expressed as volume or weight increase over a year, a decade or a rotation. Units of measure include various volumetric measures1 and weight measures in pounds and tons. In this context, site productivity can be deﬁned as ‘the productive capacity of a site, usually expressed as volume production of a given species’ (Society of American Foresters, 1995). The term ‘site quality’ is generally used when productivity is expressed as qualitative classes (e.g. poor, medium, good), or on a relative scale. Ecologists view forest productivity somewhat differently. They are often interested in accumulated biomass (dry weight of organic matter) per unit of land area and rates of increase and decrease of biomass per unit area. They also may separate biomass by plant parts such as bolewood, branches, bark, leaves, buds, roots and reproductive structures (Cannell, 1982; Kozlowski et al., 1991). Changes in biomass or weight of forest trees are used to study growth, nutrient cycling and energy ﬂow in forests. An understanding of ecosystem production dynamics is important to both ecologists and silviculturists and consequently has been the subject of numerous workshops and symposia (e.g. Hennessey et al., 1986; Cole and Gessel, 1988). Components of productivity of interest to ecologists include: ● Gross primary productivity (GPP). GPP is the increase per unit area in dry weight of organic material produced by photosynthesis that remains in the plant plus the weight of dry matter lost by plant respiration. ● Net primary productivity (NPP). NPP is the increase per unit area in the sum of three component measures: (i) the 1Volumetric

169

increase in standing biomass, including leaves, stems, roots and reproductive structures; (ii) litterfall; and (iii) the amount of biomass consumed by animals and microbial decomposers. Annual above-ground net primary production (ANPP) may be the most robust measure of productivity because it includes virtually all commercially harvested material as a subset (although in some cases portions of root systems, or components of below-ground net primary productivity, also are harvested) (Gholz, 1988). ● Net ecosystem production (NEP). NEP is GPP minus loss of dry matter due to heterotrophic respiration of microbes and other non-photosynthetic organisms (Kozlowski et al., 1991). These components of productivity vary greatly among different kinds of oak forests (Table 4.1). The ratio of below-ground to aboveground biomass of individual oaks, and thus the spatial distribution of net production, varies among ecosystems. In Mediterranean climates, the roots of multiple-stemmed oaks of coppice origin may comprise 90% or more of total tree biomass, whereas the roots of single-stemmed (non-coppice) oaks typically make up less than 30% of tree biomass (Whittaker and Woodwell, 1968; Canadell and Rodà, 1991). For singlestemmed Holm oaks growing in Mediterranean Spain, root : shoot ratio was higher on dry sites than on mesic sites based on roots 0.4 inch and larger in diameter (Fig. 4.1). This suggests that site quality, itself, can affect the ratio of below-ground to above-ground biomass in oaks. The collective evidence indicates that the ratio of above-ground to below-ground biomass in oak forests depends on several factors including the reproductive origin of oaks and related disturbance history, stand or tree age, site quality and species composition.

units commonly used in the United States include board feet, cords and cunits. One board foot (bf) is a piece of sawn wood 1 inch (2.54 cm) thick, 12 inches (30.48 cm) wide and 12 inches long. One cord is a stacked pile of wood contained within a space measuring 4 ft (1.2 m) deep, 4 ft high and 8 ft (2.4 m) long, which equals 128 ft3 (11.9 m3). One cunit equals 100 cubic feet (2.8 m3).

170

Chapter 4

Table 4.1. Above-ground biomass and net productivity of trees in selected oak forests.a Biomassb (with foliage) Forest type, location

Net productivityb (with foliage)

Stand age (years)

Trees per hectarec

(t ha1)d

(t ha1)d

serrata/crispata oak, Japang

272.7 [276.9]

8.95 [13.75]

–

561

Post/blackjack oak, Oklahomah

174.4 [179.2]

7.33 [10.59]

80

English/sessile oak with beech, Francei

167.4 [171.0]

7.16 [10.76]

66

Mixed oak/hickory, Tennesseej

133.2 [137.9]

5.4 [10.1]

30–80

–

Northern pin oak, Minnesotak

120.8 [124.2]

5.26 [8.77]

45–50

Northern red oak, Great Smoky Mt, USAl

132.2 [135.0]

4.7 [7.5]

(mature)

Sessile/European turkey oak, Hungarym

199.5 [202.9]

3.00 [6.71

65–68

9.0 [9.4]

0.39 [1.67]

117

Mexican blue/Emory oak, Arizonan

Tree height (m)e

Basal area (m2 ha1)f

15.1

39.8

2600

–

18.3

958

22

32.1

12–25

25.8

1788

c. 15

26.5

2660

14

24.6

17.4

15.1

5.3

4.0

–

190

a

From summaries in Cannell, 1982. Includes boles, branches and bark of trees; does not include fruits, woody litterfall or understorey vegetation. c Minimum diameters (dbh) included in calculations vary among studies; 1 ha = 2.471 acres. d 1 tonne ha1 = 0.446 English ton acre1. e One metre = 3.281 ft. f 1 m2 ha1 = 4.356 ft2 acre1. g Katagiri and Tsutsumi, 1975, 1976, 1978. h Johnson and Risser, 1974. i Kestemont, 1971 (plantation). j Harris et al., 1973; Harris and Henderson, 1981. Species include chestnut, white, northern red and black oaks. k Reiners, 1972. l Whittaker, 1963. m Jakucs, 1981. n Whittaker and Niering, 1975. b

As we might expect, the NPP of oakdominated ecosystems in regions with Mediterranean or desert climates is much lower than those in more humid regions. In dry regions, oaks and other woody plants are often restricted in form to small trees or

shrubs. In a Mexican blue oak/Emory oak desert shrub community in Arizona, estimated above-ground NPP was 0.4 t ha1 year1 (Whittaker and Niering, 1975). This rate is about 3% of that of the aboveground component of a typical oak forest

Site Productivity

171

Dbh (inches)

Biomass of large roots (kg)

100

2.4

3.9

5.5

7.1

8.7

18

22

80

60

40 Xeric Mesic 20

0 6

10

14 Dbh (cm)

Fig. 4.1. Biomass of large diameter roots (≥1 cm (0.4 inch)) of single-stemmed Holm oak in northeastern Spain estimated from dbh measured 1.3 m (4.3 ft) above ground. Mean root biomass of oaks on the two sites differed significantly (P = 0.022) based on analysis of covariance. Log10 (mesic site biomass) = –1.393 + 2.451.log10 (dbh), R 2 = 0.81, n = 20; log10 (xeric site biomass) = –0.448 + 1.734.log10 (dbh), R 2 = 0.71, n = 12. For both equations, root biomass is in kilograms and dbh is in centimetres. (From Canadell and Rodà, 1991, used with permission.)

in eastern United States. However, the roots and associated structures (lignotubers) of oaks growing in such climates typically comprise 65–85% of total plant mass (Rundel, 1980). Much of the production in these ecosystems therefore may lie below ground. The accurate measurement of the biomass of roots in trees is difﬁcult, especially in oaks, which have taproots that can grow to great soil depths. Moreover, it is even more difﬁcult to obtain accurate measurements of root NPP because of the rapid turnover of ﬁne roots. Reported results therefore tend to underestimate NPP of roots (Cannell, 1982). Economists often express forest site productivity as the capacity of an area to produce ﬁnancial return through timber production. Financial return can be measured in several ways including net present value, internal rate of return or soil expectation value. Such economic measures of productivity often include assumptions about the initial state of the forest, duration

of the economic evaluation period, inﬂation rate, discount rate, management practices, management expenses and the timing of periodic expenses and income. Consequently, ﬁnancial return can be estimated with great detail and associated complexity. Nevertheless, relatively simple measures such as per cent value increase or per cent volume increase are often useful for comparing site productivity among stands of similar initial condition. Because economic return is derived from forest volume, estimating site productivity in terms of yield is usually an intermediate step in estimating economic site productivity.

Relation of Site Productivity to Ecological Classification Determination of site productivity occurs within the broader context of regional factors that inﬂuence productivity. For example, climate and geology limit the upper

172

Chapter 4

range of oak site productivity. Likewise, localized topographic features such as slope steepness, slope position and aspect inﬂuence productivity. Experienced silviculturists often take for granted that these factors are related to site productivity and mentally account for them when assessing a site qualitatively. Ecological classiﬁcation systems (ECS) provide a hierarchical framework for placing individual sites within a regional context. ECS provides a means for grouping similar ecosystems based on relations between the characteristics of living organisms and physical features of a site. The ﬁnal product is a set of deﬁned ecological groupings of plants and environmental factors that are repeated across the landscape at various spatial scales ranging from single stands to large regions (Chapter 1). However an ecologist’s perspective of a forest site is apt to be different from a silviculturist’s. Ecologists are likely to consider a site as a relatively uniform geographical unit characterized by certain stable combinations of physical and climatic factors, whereas foresters generally view a site as a land unit characterized by a speciﬁed productive capacity of timber or other forest products. Despite their differences, these views are complementary (Schonau, 1988). An ecological classiﬁcation system provides information on the ecological context of a site. This, in turn, helps the silviculturist not only to evaluate site productivity, but to better understand and predict the response of the forest to silvicultural treatments. Likewise, localized quantitative estimates of site productivity (e.g. site index) can guide the ecologist in deﬁning the spatial extent of ecological classiﬁcation units. Killian (1984) proposed that ecological classiﬁcation precede the determination of site productivity. He reasoned that the goal of site classiﬁcation should be to identify the possibilities and risks to forest management and to predict potential yields, thereby assisting both short- and long-term local and regional forest planning as well as general land-use planning. That objective is more user orientated than that usu-

ally encompassed by ecological site classiﬁcation, which some believe should focus on interrelations among ecosystem components, i.e. climate, physiography–geology, soils and their biota, and vegetation (Barnes, 1984). Regardless of viewpoint, ecological classiﬁcation and site quality are closely linked. Whereas the term site productivity emphasizes the factors that inﬂuence tree growth, ecological classiﬁcation emphasizes the factors that determine the abundance of species and the occurrence of natural groupings of plant species in relation to environment. Moreover, ecological classiﬁcation usually considers both woody and herbaceous vegetation in deﬁning ecological units. Herbaceous vegetation is particularly important because of the relatively large number of herbaceous species present and the ‘ﬁdelity’ of some of those species to the factors that distinguish one ecological classiﬁcation unit from another. The presence and/or abundance of many herbaceous species is often less affected by disturbance factors and associated changes in the successional status of a forest than are trees and other woody species. There are potential advantages of ecological classiﬁcation over traditional site classiﬁcation. In their development, the latter focuses narrowly on forest productivity whereas the former produces natural ecological groupings that can be used to identify ecological units within which productivity, succession, tree regeneration, and responses to silvicultural and natural disturbances are likely to be similar (Chapter 1). However, site productivity estimates that are based solely on ecological classiﬁcation units are likely to be less accurate than those based on direct determination of site productivity at a speciﬁc location within the classiﬁcation unit.

Productivity and Related Selfsustaining Properties of Oak Forests Silvicultural practices in oak forests of the United States have been, by agricultural standards, extensive rather than intensive.

Site Productivity

With few exceptions, oak silviculture has followed an ‘ecological’ model based on managing natural vegetation and plant propagules in place. Unlike the ‘agronomic’ model, new genetic material (e.g. genetically ‘improved’ trees) is seldom introduced; herbicides and fertilizers are used only to a limited extent if at all. Although there are exceptions, the usual objective is to control stand composition, structure, growth and quality largely through timber cutting practices. This approach has been and continues to be largely driven by the economics of oak timber production, which is characterized by low returns on investment (e.g. Dwyer et al., 1993). Even when potential economic returns from more intensive silvicultural practices are deemed acceptable, investment in such practices may be discouraged by long deferrals on returns plus associated risks of ‘crop’ damage or loss from insects, disease, drought, ﬁre and other uncertainties, one or more of which have a high likelihood of occurrence over the relatively long 80- to 100-year production periods usually required for oaks. Moreover, there is growing social demand to manage publically owned forests for a wide range of products and values that transcend narrow timber production objectives. The control of stand composition, structure and density by cutting (timber harvesting) methods alone is not unique to oak forests, but characterizes hardwood silviculture in the United States in general. Although we have considerable knowledge of how to apply a more ‘agronomic’ or culturally intensive model to hardwood silviculture, such methods have seldom been applied outside of research studies. Contemporary oak silviculture therefore lies along the boundary of ‘natural’ and ‘unnatural’ (i.e. human-inﬂuenced) ecological systems. The resulting silviculture is therefore heavily dependent on natural ecological processes that are especially signiﬁcant with respect to forest productivity and its sustainability. Forests, whether inﬂuenced by humans or not, are endowed with certain attributes that ensure a high capacity for self-sustaining productive

173

capacity (e.g. as expressed by ANPP). What effect, then, does the removal of trees from the forest have on forest site productivity?

Effects of timber harvesting on site productivity The site factors most vulnerable to irreversible change are those associated with the soil. Careless removal of trees from a site can cause soil erosion and nutrient losses. Erosion is the physical loss of soil particles caused by wind and water action. Loss of nutrients also accompanies the physical loss of soil. Because the upper soil layers, which are richest in soil nutrients, are eroded ﬁrst and most severely, erosion can have a major impact on soil fertility. However, the loss of nutrients also can occur in the absence of the physical loss of soil particles when nutrient ions become dissolved in surface runoff, are lost in nutrient solutions (leachates) that percolate through and from the soil, or are lost through the removal of vegetation. Erosion is nevertheless a normal geologic process that results in some soil losses even in undisturbed old-growth forests. The estimated prehistoric rate of erosion in forested parts of southern Michigan, an area comprised of extensive oak forests, was estimated at 0.05 t acre1 year1 (Davis, 1970). In the Appalachian Mountains of West Virginia, annual soil erosion losses from hardwood forests range from 0.05 to 0.10 t acre1 year1 for both undisturbed and clearcut forests (Patric, 1976). This compares to ‘acceptable’ rates of soil loss from agricultural lands that range from 1 to 5 t acre1 year1 (Patric, 1977). Moreover, cropland is usually cultivated annually, whereas a managed oak forest is usually logged less frequently than one year in ten. The view that the absence or reduction of tree cover, by itself, causes soil erosion in forests of the eastern United States is largely unfounded (Patric 1976, 1978; Mills et al., 1987). However, to understand this issue, it is important to distinguish between effects related purely to the tem-

174

Chapter 4

porary absence of the forest canopy following timber harvesting, and associated effects related to skid trails and logging road construction. Soil and site effects associated with the removal of trees, per se, are related to the characteristics of forest soils. Four attributes of forest soils are especially important in moderating soil erosion and soil nutrient losses: (i) the forest ﬂoor, (ii) soil structure, (iii) soil inﬁltration rate, and (iv) the dynamic nature of the biomass of forest soils. In eastern deciduous forests, the surface of the forest ﬂoor consists of a layer of undecomposed or partially decomposed leaf litter and other organic debris. The litter layer provides protection against the kinetic energy of rainfall, preventing particle detachment and sealing of pores of the underlying mineral soil. In oak forests, this layer is replenished every year with 1–2 tons per acre of leaf fall, plus another 0.5 ton of woody litter fall (Rodin and Basilevich, 1968; Rochow, 1974). Forest cover thus maintains soil porosity by continually returning leaf litter and woody debris to the forest ﬂoor, which in turn is continually decomposing and being incorporated into the soil by soil organisms. Forest soils consequently develop and maintain a physical structure and macropore space associated with high soil porosity. High porosity, in turn, is associated with high inﬁltration rates and low surface runoff. Inﬁltration rates of 50 or more inches per hour are common in forest soils in the eastern United States, whereas rainfall intensities rarely exceed 2 inches per hour (Patric, 1978). When the water storage capacity of the soil is reached after prolonged precipitation, the water moves laterally through the pores of the soil to streams. The removal of trees from a site, by itself, does not destroy soil porosity provided the site is quickly revegetated. Soil porosity nevertheless can be greatly reduced by logging road construction and

log skidding practices that scrape, gouge or compact the surface layers of the soil. The risk of accelerated soil erosion from a managed forest thus depends largely on logging practices rather than on the temporary absence or reduction of tree cover. The burst of shrub, herb and sprout growth that occurs after overstorey removal tends to mitigate effects that might otherwise result in soil and site degradation. This ﬂush of growth buffers disturbed forests against nutrient losses and facilitates their recovery. For example, 5–6 years after clearcutting a northern red oak stand in southwestern Wisconsin, the estimated biomass of ﬁne roots2 was 70% larger than in an adjacent uncut stand (Yin et al., 1989). Most of this biomass was comprised of shrub and herbaceous species. In this case, not only was root biomass maintained, but temporarily increased after timber harvesting. Such responses ensure protection against both soil and nutrient losses during periods when trees are reestablishing their dominance. Timber harvesting also does not necessarily alter the amount of litter on the ﬂoor of oak forests (Carmean, 1959). Net losses of nutrients from a site nevertheless can occur when nearly all vegetation is removed and nutrients move out of the system at rates higher than they are replaced (Bormann and Likens, 1979). Such losses may occur when entire trees (boles, branches and foliage) are removed to a central location within a stand for chipping, and the nutrient-rich branches and foliage are piled along the roadside. Windrowing, i.e. pushing slash into long narrow piles, is a site preparation technique that also can remove nutrients from large fractions of the forest. If slash piles are burned, the resulting localized areas of intense heat may produce signiﬁcant losses of nitrogen. Pushing slash into piles or windrows before burning often is accompanied by the redistribution of topsoil and minerals, which in turn may decrease site

2 In this study, fine roots were defined as those of 2–10 cm in diameter. The biomass of these roots, which were largely concentrated within the upper 30 cm (12 inches) of soil, fluctuated with the seasons.

Site Productivity

quality over a portion of the stand. In highly porous soils in regions of high rainfall and warm temperatures that are already low in organic matter, timber harvesting may accelerate nutrient losses from leaching, especially nitrogen. The important issue, however, may not be the amount of mineral nutrients removed from the site, per se, but how that amount compares with that available in the soil for forest regrowth after timber harvesting. This amount depends on initial soil fertility, the relative amounts of minerals in the ecosystem that occur in the soil, forest ﬂoor and above-ground biomass, plus the rate at which the soil mineral pool is replenished by decomposition of slash, nutrient deposition from the atmosphere, mineral weathering and nitrogen ﬁxation by soil organisms. Logic would seem to indicate that, because a major portion of a forest’s biomass is in trees, timber harvesting would remove a proportionately large amount of nutrients. On the contrary, nutrient losses from timber harvesting are relatively low. For example, in an oak–hickory forest in Tennessee, the proportion of nutrients in trees was relatively small compared to that present in the entire ecosystem even though trees accounted for 67% of the biomass. Trees comprised 6% of the nitrogen, 2% of the phosphorus, 1% of the potassium and 16% of the calcium (Binkley, 1986). If only tree boles are considered, the percentages are even smaller: 1% or less of nitrogen, phosphorus and potassium, and 11% of calcium. Compared to harvesting only bolewood, whole-tree harvesting removes two to three times more nitrogen, phosphorus and potassium, and about one-third more calcium than does harvesting only bolewood. Whole-tree removal is also potentially more serious in infertile soils, where nutrient deﬁciencies already exist. Nutrient losses from whole-tree harvesting and the time required for their replacement, and the economics of wholetree harvesting are discussed in more detail by others (e.g. Waring and Schlesinger, 1985; Binkley, 1986).

175

Modifying site productivity through fertilization Although the nutrient capital of an oak forest is sufﬁcient to sustain tree growth at some given rate, few forest soils provide nutrient levels that are optimal for tree growth (Smith, 1986; Kozlowski, et al., 1991). Suboptimal levels of nutrients in forest soils can be caused by poor land-use practices that preceded the establishment of an existing forest, or low natural soil fertility. The nutrients that are below optimal levels under oak stands are usually the same elements found in most commercial fertilizers: nitrogen, phosphorus and potassium. Of these, nitrogen is often the most important. Nitrogen is naturally added to soils by rainfall and can be ﬁxed from the atmosphere by some bacteria and other organisms in the soil and in the root nodules of some forest plants. Although oak forests often have an abundant supply of nitrogen, much of it is unavailable for plant growth at any given time because it is tied up in organic matter. Unfortunately, any excess nitrogen does not accumulate, but is changed by bacteria to nitrate (NO3) and then leached from the soil by rainwater, or used by denitrifying bacteria as a source of oxygen. Phosphorus is sometimes deﬁcient in southeastern oak forests growing on highly leached soils, wet soils and very sandy soils, and potassium is occasionally deﬁcient in highly leached sandy soils. When fertilizers are applied to oak forests, trees often respond better to the application of other nutrients when they are applied with nitrogen. For example, calcium fertilization of poor sites in Pennsylvania increased stand volume growth of oaks by 10%, whereas nitrogen and calcium together increased growth by more than 40% (Ward and Bowersox, 1970). Application rates for nitrogen and other nutrients for oak forests should be matched to soil nutrient characteristics of the site. Although the height growth response of oaks to fertilization is uncertain, diameter growth can be increased by more than 30% (Ward and Bowersox, 1970; Graney, 1987). Responses to fertilization may last for 6 years, but largely disappear after 10 years.

176

Chapter 4

Methods of Evaluating Site Quality Site index In North America, site index is the most commonly used method of expressing forest site quality. Site index is deﬁned as the average height of dominant or both dominant and co-dominant trees at a standard, or index, age. Tree height growth of dominant and co-dominant trees in even-aged, fully stocked forest stands is closely related to volume growth. Site index therefore is an indirect measure of site productivity because, by itself, it does not directly express productivity. It is nevertheless useful as an index of productivity because of its correlation with productivity and relative ease of measurement. Most timber yield tables for even-aged oak stands report yields by site index classes (Table 4.2). Similarly, growth and yield models often include site index as one of the variables used to predict yield (see Chapter 10). Height growth (and thus tree height at a speciﬁed age) is a useful indirect measure of site quality because it is relatively independent of stand density. Height growth of Table 4.2. Gross yields per acre by site index classes for normala even-aged oak stands in southwestern Wisconsin.b Stand age (years)

Site index (ft at base age 50) 45

55

65

gross yield (ft3)c 20 40 60 80 100 120 140 160 a

480 1050 1550 2000 2350 2650 2900 3050

850 1750 2550 3300 3900 4350 4750 4950

1150 2400 3600 4600 5500 6200 6700 7000

Normal oak stands are relatively undisturbed stands at average maximum density (Chapter 6). b From Gevorkiantz and Scholz, 1948. c Based on gross volume, excluding bark, of all trees 0.6 inch dbh and larger including tops and limbs suitable for cordwood. Species are predominantly black, white and northern red oaks.

dominant and co-dominant trees is reduced only at the extremes of stand crowding (Carmean, 1975; Lloyd and Jones, 1983; Lanner, 1985; Jones, 1986). Site index can be determined directly from observations of trees growing on the site, or it can be estimated indirectly from physical site characteristics. Direct determination of site index Direct determination of site index requires knowing the heights and ages of dominant and co-dominant trees. This information then is referenced to a set of hypothetical height growth curves, called site indexcurves, which are indexed to a common age (Fig. 4.2). For oak species in the United States, the usual index age is 50 years. A site index of 65 thus indicates that the dominant and co-dominant trees on a site attain an average height of 65 ft at 50 years. A set of site index curves therefore represents patterns of height growth of dominant and co-dominant trees for different sites. Their purpose is to relate observed heights of trees of any age to their expected heights at age 50. Any factors that cause actual tree height growth to differ from the patterns expressed in the curves used to determine site index will introduce error into the estimated site index. Consequently, it is important to ensure that the site index curves utilized are appropriate for the site being evaluated. For example, three sets of site index curves applicable to white oak differ somewhat for site index 70. Graphic comparison of these curves illustrate that the closer the observed tree age is to the index age, the less discrepancy in site index there is among the curves (Fig. 4.3). However, when site index estimates are based on trees younger than 30 years or older than 70 years, differences due to choice of site index curves can be substantial (Carmean, 1979). Accurately determining site index requires: (i) the presence of trees that are reliable indicators of site quality, and (ii) the availability of suitable site index curves

Site Productivity

140

177

100 90

120

80 70 60 80 50 40

60

Site index (ft)

Tree height (ft)

100

30 40 20 0 20

40

60 80 Tree age (years)

100

120

Fig. 4.2. Site index curves (index age 50) for black oak in the unglaciated uplands of southeastern Ohio, eastern Kentucky, southern Indiana and southern Missouri. Dashed-line curves represent values beyond the observed range. (From Carmean, 1971, 1972.)

120

100

Tree height (ft)

80

60

40 Carmean et al. (1989) Graney and Bower (1971)

20

Schnur (1937) 0 20

40

60

80

100

120

Tree age (years) Fig. 4.3. Site index 70 curves for white oak (Carmean, 1971; Graney and Bower, 1971), and a composite curve (applicable to white oak) for upland oaks in the eastern United States (Schnur, 1937). Although all three curves have an identical site index of 70 ft at index age 50, the height–age curves diverge elsewhere. Differences in estimated site index resulting from application of different curves can be substantial, especially when observed tree age differs from the index age by more than 20 years.

178

Chapter 4

that accurately characterize the height growth pattern of the trees observed (Carmean et al., 1989). Suitable trees for determining site index include free-growing, uninjured dominant and co-dominant trees. Such trees most frequently occur in even-aged, fully stocked stands that have not been high-graded by logging, heavily grazed, or otherwise damaged. To determine site index, heights and ages of dominant and co-dominant trees must be ascertained. Tree ages can be obtained from increment cores that reveal annual rings as well as patterns of diameter growth. Trees that are reliable indicators of site index form annual rings during early life that are wide and even. Trees with narrow rings formed in early life indicate suppression of growth and therefore should not be used to determine site index. Trees with forks and other bole defects may indicate earlier top breakage or dieback and therefore may be unreliable indicators of site index. Trees selected for site index determinations also should be from well-stocked even-aged stands comprised of dominant and co-dominant trees whose ages do not differ by more than 10 years. Greater differences can produce highly variable patterns of height growth, which in turn may cause large errors in site index estimation. For example, even-aged upland oak stands in the Missouri Ozarks sometimes include dominant and co-dominant trees appreciably younger or older than most of the trees in those crown classes, i.e. the predominant age class. Because young oaks tend to grow faster in height, they will indicate a higher site index than oaks in the predominant age class if a common age is assumed. Conversely, appreciably older oaks grow more slowly in height and will indicate a lower sites index than trees in the predominant age class if a common age is assumed (McQuilkin, 1975). Site index determined from individual dominant and co-dominant trees may vary even on small plots that appear to be homogeneous in soil and microsite. This variation in site index can produce problems in determining the number of site index trees required to obtain reliable site

index estimates (Carmean, 1975; Lloyd and Haﬂey, 1977; Lloyd, 1981). The requisite number of site trees depends on several factors, including the desired accuracy of the site index estimates, size of the stand and variation in site index. Sampling trees for site index determination thus is more complicated in large tracts with highly variable stand and site conditions. Ideally, large tracts should be stratiﬁed into smaller areas that are relatively homogeneous topographically, edaphically and ecologically. Site index then should be determined within each stratum. Even then, large errors in estimating site index can occur when site index curves are used outside their intended area of application. The height growth of oak coppice stands may differ markedly from seedling-origin stands. This is especially true during the early years of stand development when oak coppice beneﬁts the most from the relatively large parent-tree root systems they are connected to. Large roots and the concomitant capacity of young coppice shoots to produce multiple long ﬂushes (Fig. 2.20) together with clump density effects (Chapter 2) may mask the early expression of site quality effects. Site effects may not be apparent until stands are 20–25 years old. By age 50, the productive capacity of coppice stands in the South Carolina Piedmont was well differentiated. On the best sites (oak site index 90), volumes and above-ground biomass were nearly twice that of stands growing on the poorest sites (oak site index 50) (Zahner and Myers, 1984). The height growth curves for these coppice stands differed substantially from the conventional site index curves developed for upland oaks in the Piedmont region of the southeastern United States (Olson, 1959). The latter overestimated site index by as much as 18 ft when applied to coppice stands (Zahner et al., 1982). Site index curves thus were developed speciﬁcally for coppice-origin stands of that region (Fig. 4.4). Although stand site index is usually expressed as the average site index calculated from qualiﬁed sample trees and reported to the nearest foot, greater preci-

90

90

80

80

70

70

60

60

50

50

40

40

Site index (ft)

Tree height (ft)

Site Productivity

30 20 10

20

30

40

50

Tree age (years) Fig. 4.4. Site index curves (index age 50) for young mixed oak stands of sprout origin in the South Carolina Piedmont. (From Zahner et al., 1982, used with permission.)

sion should not be attributed to the estimate than is statistically justiﬁed. The precision of site index estimates made from site index equations was determined for white, black and scarlet oaks in Missouri. Estimates for trees 20 years younger to 20 years older than index age were precise to within 3–4 ft with 95% conﬁdence when based on ten sample trees. Similar results were observed for oaks (chestnut, white, northern red, black and scarlet) in northwestern West Virginia (Lamson, 1980). Errors in estimating site indices from site index equations arise from three sources: (i) errors in estimating individual sample tree site indices from site index/height regressions; (ii) variation among sample tree heights within even-aged stands; and (iii) measurement errors (McQuilkin and Rogers, 1978). Methods of collecting data for deriving site index curves have evolved greatly since the ﬁrst site index curves were published. The earliest curves were based on only a few plots with total height and age measurements taken from a few selected trees. These height and age data were used to calculate an average ‘guiding curve’ representing the average height growth pattern of a given species. Graphical or propor-

179

tional methods then were used to produce a set of anamorphic curves, within which each curve has the same shape. Anamorphic curves assume that the pattern of height growth is similar for all levels of site quality, and for all climates, soils and topographic conditions within the region of intended application (Carmean, 1970). This assumption is inappropriate for oaks because the shape of the height curve often varies with site quality (Carmean, 1970, 1975). More recently developed site index curves are based on the assumption that the shape of tree height growth curves can differ by site classes. Such curves can be derived through stem analysis, which requires that sample trees be felled and cut into sections to determine the progression of height growth over time. Stem analysis combined with non-linear regression analysis produces polymorphic site index curves. This is now the most widely used method for developing site index curves for species like the oaks that express polymorphic patterns of growth. For the oaks, polymorphic site index curves have largely replaced anamorphic curves. Methods for constructing site index curves have been described by others (Burkhart et al., 1981; Clutter et al., 1983; Borders et al., 1984; Biging, 1985; Avery and Burkhart, 1994). A comprehensive compilation of 127 site index curves for species in the eastern United States, including 21 sets of curves for oaks, is presented by Carmean and others (1989). Each set of curves is based on an expanded form of the Chapman–Richards non-linear function (Ek, 1971; Payandeh, 1974a, 1974b; Monserud and Ek, 1976). One variant of the model estimates height from site index and age whereas the other estimates site index from height and age. Site index comparisons among species Dominant and co-dominant trees of different species growing on the same site are likely to have different height growth patterns. For example, a comparison of site index curves for white, scarlet and red oaks

180

Chapter 4

able for the major upland oaks and associated species in the eastern United States (Fig. 4.6; Appendices 4–6). Site index for a given species is not necessarily uniformly higher or lower than for another species. For example, on good sites the yellow-poplar site index (index age 50) may be as much as 30 ft greater than that of white oak. However, as site quality decreases the difference between yellowpoplar and white oak site indexes also decreases. At site indices below 70 ft, the white oak site index may exceed that of yellow-poplar (Olson and Della-Bianca, 1959; Carmean and Hahn, 1983).

at a common site index of 70 shows that white oak height growth lags behind the other species for the ﬁrst 40 years. But by age 80, white oak height surpasses the other oaks (Fig. 4.5). Consequently, separate site index curves are generally required for each species. This is evident from species comparisons in the Missouri Ozarks. There, the observed site index for white oak on a given site is about 4 ft less than that of black oak; the scarlet oak site index is about 3 ft greater than that of black oak on the same site (McQuilkin, 1974). The co-occurrence of species across a wide range of sites makes possible the development of equations that can be used to convert the observed site index of one species to the estimated site index of another species. Such site index conversions are useful when the species of interest is absent and a common site index basis for all stands or locations within stands is desired. Such conversions are facilitated by graphs or equations designed for this purpose (Nelson and Beaufait, 1956; Trimble and Weitzman, 1956; Doolittle, 1958; Olson and Della-Bianca, 1959; McQuilkin, 1974; Carmean and Hahn, 1983). They are avail-

Indirect estimation of site index from soil and topographic factors Site index curves cannot be applied to sites where suitable trees are absent or where no trees are present. To solve that problem, equations have been developed in some regions to estimate oak site index from soil, topographic and other factors. These equations typically account for 70–85% of observed variation in site index. However, they are often not used in practical applica-

120

Tree height (ft)

100

80

60

40 White oak N. red and black oaks Scarlet oak

20

0 20

40

60

80

100

120

Tree age (years) Fig. 4.5. Comparison of site index curves for scarlet oak, white oak and red/black oaks in the Central Hardwood Region. The curves shown are for site index 70 ft (index age 50). (From Carmean et al., 1989.)

Site Productivity

181

110 100

Site index (ft)

90 80 70 60 50 40

Black oak White oak Scarlet oak Northern red oak Yellow-poplar

30 Fig. 4.6. Chart for converting the site index (index age 50) of one species to another in even-aged upland oak and yellow-poplar stands in the central states. The site index of species absent from a stand can be estimated from species present by using this nomogram. Site indexes for all species are read from the vertical axis. For example, assume that height and age measurements of several dominant and co-dominant yellow-poplars indicate a site index of 94 for that species. The corresponding site index of another species is read by moving vertically downward from 94 on the yellow-poplar curve to the curve of the species of interest. At that interception point, the unknown site index is read horizontally across on the vertical axis. On this chart, site indexes for scarlet, black, northern red, chestnut and white oak corresponding to yellow-poplar site index 94 are approximately 89, 88, 86 and 83, respectively. Conversions also can be derived from equations (see Tables 4.3, 4.4 and 4.5). (From Carmean and Hahn, 1983, by permission of Society of American Foresters, Bethesda, Maryland. Not for further reproduction.)

tion because they require information on one or more soil characteristics such as texture, horizon or soil thickness, and soil chemical properties that are difﬁcult or inconvenient to measure. Also, many of the equations are derived from observed values of site index derived from anamorphic curves that may not accurately describe tree height growth in the region to which the equation applies. Even though such equations are often adequate for practical applications, they only provide an estimate of site index, which itself is only an indirect measure of forest productivity. Thus, these equations are two steps removed from direct determination of productivity (Leary, 1985). Equations for estimating site index nevertheless can provide insight into the factors that inﬂuence the height growth of oaks and therefore oak site productivity. Most of these equations include as predictors one or more topographic factors including slope posi-

tion, aspect, slope gradient and slope shape. Although these factors, by themselves, have no direct effect on tree growth, they are correlated with more directly causative factors. For example, aspect and slope gradient jointly determine the amount of solar radiation received on a slope (Frank and Lee, 1966; Swift, 1976), and thus account for factors more directly related to tree growth such as evapotranspiration, leaf temperature and ﬂuctuation in microclimate. Aspect and slope gradient are correlated with site quality in most oak forests. The better sites usually occur on north to east aspects and the poorer sites on south to west aspects (Carmean, 1965; Hannah, 1968; Hartung and Lloyd, 1969; Graney, 1978; Auchmoody and Smith, 1979). Slope position and slope shape are also related to soil moisture. Lower slopes, for example, usually have higher site indices due to subsurface water ﬂow from upper slopes.

Chapter 4

One predictive site index equation (Auchmoody and Smith, 1979) considers the interactive effects of slope gradient and aspect. In this equation, the effect of aspect on site index is modiﬁed by slope gradient. Site quality accordingly decreases as aspect departs from the cool and moist microenvironment in the northeast quadrant at 81° azimuth. At this most favourable azimuth, slope gradient has no effect. But as aspect changes in either direction towards the least favourable aspect at 261° in the southwest quadrant (180° from 81°), the effect of increasing slope gradient is to reduce site index due to associated increases in solar radiation (Fig. 4.7). Black oak in southern Ohio responded similarly to slope and aspect (Carmean, 1965, 1967). However, in that region, the most favourable and least favourable aspects occurred at 45° and 225°, respectively. The effects of aspect may be asymmetrical, i.e. the most favourable and least favourable aspects may not be separated by 180°. In an Ohio study, the observed opti-

mum aspect occurred at 45° azimuth whereas south, southwest and west aspects (180° to 270°) were almost equally unfavourable (Carmean, 1965, 1967).3 The occurrence of the most favourable aspect in the northeast quadrant was associated with thicker litter layers, thicker A1 horizons high in base saturation, less acidic A and upper B horizons, greater available soil nitrogen and thicker colluvial accumulations than soils on southwest aspects. Thus, soil development processes related to tree growth also appear to be correlated with variation in topography and solar radiation. Relations between aspect and observed tree growth may not be entirely consistent with the actual incident radiation received on a slope. Commonly observed relations between topographic factors and tree growth suggest that other factors can modify solar radiation effects. These factors include soil development processes that, although partially dependent on incident solar radiation, do not necessarily change linearly with solar radiation.

70

0 10

Site index (ft)

25 68

% slope

182

50 66

64

62 0

60

81

120

180

261

240

300

360

Aspect (degrees azimuth) Fig. 4.7. Relation between oak site index (index age 50), aspect and four slope gradients in upland forests in northwestern West Virginia. For comparative purposes, site index has been arbitrarily set at 70 for the most favourable aspect (81° azimuth) for each slope gradient. For each gradient, site index is minimized at 261° azimuth. (Adapted from Auchmoody and Smith, 1979). 3 Stage (1976) describes methods for mathematically identifying and specifying interactive aspect and slope gradient effects on tree growth.

Site Productivity

Like topographic factors, soil factors commonly incorporated into site index prediction equations are often associated with variables thought to be more causally related to tree growth. For example, soil texture and stone content are correlated with available moisture. Likewise, soil pH and per cent base saturation are correlated with the availability of nitrogen and other macronutrients, and soil depth or depth to an impenetrable horizon is correlated with the effective volume of the rooting zone. The effectiveness of a given variable as a predictor of productivity may vary among regions, ecosystems within regions and species. Such variability is reﬂected in the differences among existing sets of site index curves. The development of satisfactory site index estimation equations for hardwood sites on alluvial soils in the South has been shown to be difﬁcult, if not unfeasible (Broadfoot, 1969). The apparent reason is the large number of interacting factors (including ﬂuctuating water tables, aeration, and associated soil rooting space and nutrient availability problems) that inﬂuence site productivity in bottomlands. To overcome these problems, Baker and Broadfoot (1979) developed a ﬁeld guide for classifying sites occupied by, or potentially suitable to, cherrybark, Nuttall, Shumard, water, willow and swamp chestnut oaks. Their method is based on a matrix of soil factors that inﬂuence tree growth in southern bottomlands. Application requires evaluating four major soil factors: (i) soil physical condition; (ii) moisture availability during the growing season; (iii) nutrient availability; and (iv) aeration. Each of these factors is further comprised of speciﬁc soilsite properties whose qualitative or quantitative properties are subdivided into three relative site quality classes: best, medium and poor. A site quality rating (SQR) is then assigned to each soil-site property and site quality class for each of the six bottomland oak species considered by the method. Site suitability and quality for a given species is then evaluated by summing its SQRs across all soil-site properties considered by the method.

183

Estimating site index from tree height and diameter Intensive forest management requires measures of productivity that are sensitive to site differences while retaining validity over time-dependent changes in the tree crop. Site index attempts to satisfy these requirements by extrapolating measured tree height at a given age to tree height at a reference age. Although site index is the most widely used method for assessing site quality in North America, the method has been criticized (Gevorkiantz and Scholz, 1944; Jones, 1969; Gholz, 1988; Avery and Burkhart, 1994). Potential problems in the application of site index include extrapolation errors, measurement errors and possible insensitivity to production expressed as volume (Stout and Shumway, 1982). It is frequently difﬁcult to obtain accurate tree ages from increment cores and to obtain accurate heights of standing trees. Some trees are difﬁcult to age from an increment core because the pith or growing centre of the tree bole is often missed when it is non-circular (Lamson, 1987). In slow-growing trees, growth rings are often obscure. Variability in ages among trees within the same stand also contributes to uncertain age estimates. Although any one of these factors could introduce large errors into estimating site index, age variability is potentially the most serious limitation of site index as an estimator of volume productivity. Moreover, height alone is used in the site index method, even though tree and stand volume depend on both diameter and height. Foresters nevertheless have been reluctant to discard site index for measures of productivity based on diameter growth, which although sensitive to site quality, is also strongly inﬂuenced by stand density. Studies nevertheless have shown that, across a wide range of initial stand densities, yields per unit area tend to converge with time toward a site-speciﬁc maximum value (Drew and Flewelling, 1977; Harper, 1977). This convergence is caused by physiological adjustments in the height and diameter growth of trees to changes in available growing space. When stand den-

184

Chapter 4

sity is low, diameter growth increases and height growth decreases, and conversely when stand density is high, diameter growth decreases and height growth increases (Gevorkiantz and Scholz, 1944). Because the two trends are compensatory, they tend to produce approximately the same volume in the average dominant tree. Theoretically, using both diameter and height to estimate site productivity should compensate for variation in stand density. Based on this concept, tree dbh, height and age can be used to place stands in site index classes based on tree ‘volume index’. A tree’s volume index is the product of its basal area (or squared diameter) and height. The volume index method was originally developed for mixed-oak stands in southwestern Wisconsin (Gevorkiantz and Scholz, 1944). However, the method is potentially applicable to any forest type for which site index and average heights and diameters of dominant trees in fully stocked (‘normal’) stands are known for a

range of tree ages. Such information is often provided in conjunction with published site index, stand and yield tables. From that information, site index classes based on volume index can be graphically related to the average age of dominant trees (Fig. 4.8). The volume index of any observed stand then can be referenced to the graph. To reliably estimate a stand’s volume index, at least 25 dominant trees of approximately the same age should be measured. The observed stand’s volume index is calculated by multiplying the average basal area by the average height of the sample trees. The stand’s site index class then can be determined by referencing its volume index to the reference curves. Volume index may be especially useful for estimating site quality in understocked oak stands. However, volume indices may need to be ‘corrected’ by reducing volume index when stocking is below 50% or above 100% of normal (Gevorkiantz and Scholz, 1944).

65 160

Volume index

120

55

80

50

Site index class (ft)

60

40 45 0 20

40

60

80

100

120

140

Average age of dominant oaks (years) Fig. 4.8. Site index classes (index age 50) derived from volume index and age of dominant oaks in mixed-oak stands in southwestern Wisconsin. Classes are represented by areas between the curves. A stand’s volume index is calculated by multiplying the average basal area of dominant trees (ft2) by their average height (ft). For example, by reference to the chart, 90-year-old trees with a volume index of 80 represent a site index class of 55. In this region, the maximum site index for oak is about 70 ft. (From Gevorkiantz and Scholz, 1944, by permission of Society of American Foresters, Bethesda, Maryland. Not for further reproduction.)

Site Productivity

Unfortunately, the volume index method does not eliminate the problem of age variability because it requires an age measurement. There is, however, a method for estimating site quality that, in ﬁeld application, does not require determining tree age (Stout and Shumway, 1982). The method is based on families of height–diameter curves formulated for site classes by species. Such curves have been derived for white, black and northern red oaks. They were generated by substituting coefﬁcients for S and b (from Appendix 7) into the following equation for tree height in feet (H): H = 4.5 + S(1 – ebD)

[4.1]

where D is dbh in inches, and S and b are site-speciﬁc and species-speciﬁc coefﬁcients, respectively. Equation 4.1 thus can be used to generate families of species-speciﬁc height–diameter curves associated with different site index classes. Site index estimation using this technique requires measuring the heights and diameters of several dominant or co-dominant trees and referencing the paired measurements to height–diameter site index curves (Fig. 4.9). For example, a red oak 90 ft tall with a dbh of 20 inches lies within site class 80 according to Fig. 4.9.

185

Site evaluation alternatives to site index Methods based on soil and physiography Another approach to evaluating oak sites is to place sites into discrete productivity categories or on a continuous site quality scale other than site index. Like soil-site equations for estimating site index, these methods are usually based on soil and topographic factors. A categorical site evaluation method applicable to oak forests in southern Michigan is based on: (i) soil texture; (ii) the presence of moist layers (high water table or ﬁne textured materials) within 4–10 ft of the soil surface; (iii) slope steepness; and (iv) slope position (Gysel and Arend, 1953). High productivity is associated with sites with ﬁne textured subsoils, gentle mid to lower slopes and moisture retentive layers within the upper 4–10 ft of the soil. The associated site evaluation scheme places sites in one of ﬁve productivity classes ranging from very poor to very good, which in turn, are associated with the average volume of dominant and co-dominant oaks (Table 4.3).

110 90 80 70

90

60 80 50 70

Site index (ft)

Tree height (ft)

100

60 50 40 6

10

14

18

22

26

30

Dbh (inches) Fig. 4.9. Site index curves (index age 50) for northern red oak in the central Appalachians based on tree height and dbh. Application requires the measurement of several dominant or co-dominant trees in each stand. (Adapted from Stout and Shumway, 1982, by permission of Society of American Foresters, Bethesda, Maryland. Not for further reproduction.)

186

Chapter 4

Table 4.3. Site classification system for upland oak stands in southern Michigan.a

Texture of subsoil Fine (loams, clay loams and clays)

Medium (sandy loams and loamy sands)

Coarse (sands)

Position of moist layers in substrata

Topography

Highc High

Flatd Rollinge

High

Hillyf

Low Low

Flat Rolling

Low

Hilly

High High

Flat Rolling

Low Low

Flat Rolling

Low

Hilly

High High

Flat Rolling

Position on slope

Site class

Average volume of dominant/ co-dominant oaks [ft3 (bd ft)]b

— Upper Middle Lower Bottom Upper Middle Lower

Very good Good Good Good Very good Medium Good Good

54 (210) 35 (130) 35 (130) 35 (130) 54 (210) 24 (70) 35 (130) 35 (130)

— Upper Middle Lower Bottom Upper Middle Lower — Middle

Medium Medium Medium Good Very good Poor Medium Good Good Good

24 (70) 24 (70) 24 (70) 35 (130) 54 (210) 17 (43) 24 (70) 35 (130) 35 (130) 35 (130)

— Upper Middle Lower Bottom Upper Middle Lower — Middle

Very poor Very poor Poor Medium Good Very poor Poor Good Good Good

12 (25) 12 (25) 17 (43) 24 (70) 35 (130) 12 (25) 12 (25) 17 (43) 17 (43) 17 (43)

a From

Gysel and Arend, 1953. Average volume of 80-year-old black oaks and northern red oaks. c For the fine-textured subsoil class, ‘high’ refers to the presence of a fine-textured (clayey) subsoil; for other subsoil classes, ‘high’ refers to the presence of a water table within 4–10 ft of the soil surface. d Slope steepness of 5% or less. e Slopes are of moderate length with relatively broad ridges and valleys. f Slopes are relatively steep with narrow ridges and valleys. b

The topographic site coefﬁcient (TSC) represents a continuously scaled relative measure of forest site productivity applicable to the Driftless Area of southwestern Wisconsin and adjacent southeastern Minnesota, northeastern Iowa and northwestern Illinois (Johnson, 1975). It can be used to assess site productivity on upland soils ranging in texture from sandy loam to silt loam. TSC integrates into a single value the effects of soil depth, slope position and aspect on

soil moisture and therefore on tree growth. The index value is based on soil plus parent material depth to bedrock (from a minimum of 10 to a maximum of 50 inches) weighted by an index of average growing season soil moisture associated with slope position and aspect (azimuth). TSC is scaled from 0.1 (poorest sites) representing sites that occupy southwest-facing upper slopes on thin soils, to 1.0 (best sites) that occupy low northeastfacing slopes on deep soils (Fig. 4.10).

Slope position

Site Productivity

NE N

Aspect SE NW

187

Soil depth (inches) ≤10

S SW

20

30

≥50

40

Lower

Middle Upper

Level

3.0

2.5 2.0 1.5 1.0 [cos(azimuth-45)]+2

0.1

0.2

Poor

0.3 0.4 0.5 0.6 0.7 0.8 Topographic site coefficient Medium

0.9

1.0

Good

Site quality Fig. 4.10. Topographic site coefficient (TSC) in relation to aspect, slope position and soil depth. To find TSC, locate aspect on the transformed azimuth or compass scale. Then proceed upwards until the appropriate slope position is intersected. Use the LEVEL line for level topography and all slopes less than 15%. Next, proceed horizontally to the right until the appropriate soil depth is intersected. Finally, proceed downward and read the corresponding TSC value. (From Johnson and Rogers, 1982, used with permission.)

TSC thus provides an empirical method for assessing forest site quality where it is impossible to obtain site index because of the absence of trees or the lack of suitable trees for its determination. Potential applications include the assessment of site quality in young clearcuts and other harvested areas. TSC also has been used as a predictor of early stand development after timber harvesting (Johnson, 1976). In addition, the method can be used to assess the suitability of forest or non-forest sites for tree planting and for predicting the growth of natural reproduction. For example, the height growth of northern red oak stump sprouts (Fig. 2.28) and planted hardwoods in clearcuts in southwestern Wisconsin is related to TSC (Johnson, 1975; Johnson and Rogers, 1980, 1982, 1984, 1985). Although the exact relation between TSC and site index is unknown, TSC values span the approximate northern red oak site index

range of 45–70 ft based on the site index curves of Gevorkiantz (1957). In addition to the methods mentioned above, information on oak forest productivity in relation to physiography and soil taxa is contained in many county and regional soil survey manuals published by the USDA Natural Resources Conservation Service (formerly USDA Soil Conservation Service). A site classiﬁcation system for the south-central part of the Central Hardwood Region was developed largely from such information. The classiﬁcation system is applicable to the northern Cumberland Plateau, central ridge and valley, and interior low plateau highland rim sections, which lie within portions of the broadleaved forest, oceanic province (221a, Fig. 1.2) and the broadleaved forest, continental province (221b, Fig. 1.2) (Smalley, 1979, 1982, 1984, 1986). This system has much in

188

Chapter 4

common with the ecological classiﬁcation systems discussed in Chapter 1. The system divides the landscape into ‘landtypes’ based on commonalities in geology, physiography, soils and vegetation. Landtype descriptions are provided and each is rated with respect to forest productivity (site index and mean annual growth of forest stands), plant competition, seedling mortality, equipment limitations, erosion and windthrow hazards, and tree species desirability. Biophysical methods Biophysical methods relate primary environmental variables such as air temperature, precipitation, relative humidity and incident solar radiation to the productive potential of forests. In can be argued that such variables represent the underlying forces that are causally related to variation in forest productivity. Accordingly, productivity estimated from such variables would not be subject to the vagaries of changing stand conditions as are the more common empirical indices like site index. Because site index, by deﬁnition, is tree height at a speciﬁed age, events that alter the pattern of tree height growth may diminish the accuracy and therefore the usefulness of site index. Although foresters assume site index to be constant for a given forest site, certain silvicultural practices such as thinning and fertilizing can affect height growth and therefore the estimate of site index. Biophysical methods of site evaluation attempt to relate forest productivity to factors more directly related to productivity. The biophysical approach is based on the hypothesis that the capability of the land to produce wood depends on the localized characteristics of the continuum representing the movement of water through the soil and plant, and into the atmosphere. Any one of three continuum elements (soil–plant–atmosphere) can potentially limit growth and thus site productivity. The approach assumes that the quantity of soil water consumed in forest transpiration is directly proportional to

wood production (Czarnowski, 1964). For a given species, transpiration rate is dependent on the rate of soil water absorption and the leaf–atmosphere relationships that control water loss. Soil depth, moisture content and texture are factors that determine soil water availability, while variation in soil temperature mediates absorption rates at any given moisture level. However, the driving force for transpiration occurs at the leaf surface and is dependent on vapour pressure differences between leaf and air. For a given ambient vapour pressure, transpiration depends on leaf temperature. Consequently, during periods of intense radiation, even trees growing in moist soils can be severely water stressed if transpiration exceeds absorption. The relation between these two processes thus provides a biophysical basis for explaining differences in site productivity such as those that occur between north- and south-facing slopes. The extent to which these processes are related to (and thus are predictable from) more easily measured soil and topographic factors forms the basis for most other methods of forest site evaluation. For the same species, differences in productivity between two sites with similar soils but different topographies can be substantial. For example, basal area growth of northern red oak in the Appalachian Mountains of West Virginia was more than three times greater on north-facing slopes than on south-facing slopes even though the physical and chemical properties of the soils were similar. The difference in productivity therefore was largely ascribed to differences in radiant and thermal energy regimes (Lee and Sypolt, 1974). Unfortunately, data for relating site productivity to primary environmental variables are difﬁcult to obtain. One difﬁculty is the measurement of key environmental variables such as relative humidity over sufﬁciently long periods (greater than 1 year). Moreover, correlations between above-ground net primary production (ANPP) and single or even multiple envi-

Site Productivity

ronmental factors are extremely variable. Because of these problems, it is difﬁcult to estimate the direct effect of environmental variables on carbon allocation and ﬁxation and to model short-term ANPP for develop-

189

ing a site quality measure that has predictive and interpretative power (Gholz, 1988). Examples of such ‘physiological process’ models of ANPP driven by environmental factors are presented by Gholz (1988).

References Auchmoody, L.R. and Smith, H.C. (1979) Oak soil–site relationships in northwestern West Virginia. USDA Forest Service Research Paper NE NE-434. Avery, T.E. and Burkhart, H.E. (1994) Forest Measurements. McGraw-Hill, New York. Baker, J.B. and Broadfoot, W.M. (1979) Site evaluation for commercially important southern hardwoods. USDA Forest Service General Technical Report SO SO-26. Barnes, B.V. (1984) The ecological approach to ecosystem classiﬁcation. Proceedings of the Symposium Site and Productivity of Fast-growing Plantations (IUFRO), pp. 69–89. Biging, G.S. (1985) Improved estimates of site index curves using a varying-parameter model. Forest Science 31, 248–259. Binkley, D. (1986) Forest Nutrition Management. John Wiley & Sons, New York. Borders, B.E., Bailey, R.L. and Ware, K.S. (1984) Slash pine site index from a polymorphic model by joining (splining) nonpolymorphic segments with an algebra difference method. Forest Science 30, 411–423. Bormann, F.H. and Likens, G.E. (1979) Pattern and Process in a Forested Ecosystem. Springer-Verlag, New York. Broadfoot, W.M. (1969) Problems in relating soil to site index for southern hardwoods. Forest Science 15, 354–364. Burkhart, H.E., Cao, Q.V. and Ware, K.D. (1981) A comparison of growth and yield prediction models for loblolly pine. Virginia Polytechnic Institute and State University, School of Forestry and Wildlife Resources. Canadell, J. and Rodà, F. (1991) Root biomass of Quercus ilex in a Mediterranean forest. Canadian Journal of Forest Research 21, 1771–1778. Cannell, M.G.R. (comp.) (1982) World Forest Biomass and Primary Production Data. Academic Press, London. Carmean, W.H. (1959) Litter weight not reduced following clearcutting of poor-site oak stands. Journal of Forestry 57, 208–209. Carmean, W.H. (1965) Black oak site quality in relation to soil and topography in southeastern Ohio. Soil Science Society of America Proceedings 1965, pp. 308–312. Carmean, W.H. (1967) Soil survey reﬁnements for predicting black oak site quality in southeastern Ohio. Soil Science Society of America Proceedings 1967, pp. 805–810. Carmean, W.H. (1970) Site quality for eastern hardwoods. USDA Forest Service Research Paper NE NE-144, pp. 36–56. Carmean, W.H. (1971) Site index curves for black, white, scarlet and chestnut oaks in the Central States. USDA Forest Service Research Paper NC NC-62. Carmean, W.H. (1972) Site index curves for upland oaks in the Central States. Forest Science 18, 109–120. Carmean, W.H. (1975) Forest site quality evaluation in the United States. Advances in Agronomy 27, 209–269. Carmean, W.H. (1979) Soil-site factors affecting hardwood regeneration and growth. In: Holt, H.A. and Fischer, B.C. (eds) Regenerating Oaks in Upland Hardwood Forests. Purdue University, West Lafayette, Indiana, pp. 61–74. Carmean, W.H. and Boyce, S.G. (1973) Hardwood log quality in relation to site quality. USDA Forest Service Research Paper NC NC-103. Carmean, W.H. and Hahn, J.T. (1983) Site comparisons for upland oaks and yellow-poplar in the Central States. Journal of Forestry 81, 736–739.

190

Chapter 4

Carmean, W.H., Hahn, J.T. and Jacobs, R.D. (1989) Site index curves for forest tree species in the eastern United States. USDA Forest Service General Technical Report NC NC-128. Clutter, J.L., Fortson, J.C., Pienaar, L.V., Brister, G.H. and Bailey, R.L. (1983) Timber Management: A Quantitative Approach. John Wiley & Sons, New York. Cole, D.W. and Gessel, S.P. (1988) Forest Site Evaluation and Long-term Productivity. University of Washington Press, Seattle. Czarnowski, M.S. (1964) Productive Capacity of Locality as a Function of Soil and Climate with Particular Reference to Forest Land. Louisiana State University Press, Baton Rouge. Davis, M.B. (1970) Erosion rates and land use history in southern Michigan. Geological Society of America Abstracts 2, 533. Doolittle, W.T. (1958) Site index of scarlet and black oak in relation to southern Appalachian soil and topography. Forest Science 3, 114–124. Drew, J.T. and Flewelling, J.W. (1977) Some recent Japanese theories of yield–density relationships and their application to Monterey pine plantations. Forest Science 23, 517–534. Dwyer, J.P., Dey, D.C. and Kurtz, W.B. (1993). Proﬁtability of precommercially thinning oak stump sprouts. Northern Journal of Applied Forestry 10, 179–183. Ek, A.R. (1971) A formula for white spruce site index curves. University of Wisconsin Forestry Research Note 161. Frank, E.C. and Lee, R. (1966) Potential solar beam irradiation on slopes: tables for 30° to 50° latitude. USDA Forest Service Rocky Mountain Forest and Range Experiment Station RM RM-18. Gevorkiantz, S.R. (1957) Site index curves for red oak in the Lake States. USDA Forest Service Lake States Forest Experiment Station Technical Note 485. Gevorkiantz, S.R. and Scholz, H.F. (1944) Determining site quality in understocked oak forests. Journal of Forestry 42, 808–811. Gevorkiantz, S.R. and Scholz, H.F. (1948) Timber yields and possible returns from the mixed-oak farmwoods of southwestern Wisconsin. USDA Forest Service Lake States Forest Experiment Station Publication 521. Gholz, H.L. (1988) Problems in the biophysical determination of forest site quality. In: Cole, D.W. and Gessel, S.P. (eds) Forest Site Evaluation and Long Term Productivity. University of Washington Press, Seattle, pp. 12–21. Graney, D.L. (1978) Site quality relationships for the oak–hickory forest type. Proceedings of 1978 Society of American Foresters National Convention, pp. 339–343. Graney, D.L. (1987) Ten-year growth of red and white oak crop trees following thinning and fertilization in the Boston Mountains of Arkansas. USDA Forest Service General Technical Report SO SO-42, pp. 445–449. Graney, D.L. and Bower, D.R. (1971) Site index curves for red and white oaks in the Boston Mountains of Arkansas. USDA Forest Service Research Note SO SO-121. Gysel, L.W. and Arend, J.L. (1953) Oak sites in southern Michigan: their classiﬁcation and evaluation. Michigan State University Technical Bulletin 236. Hannah P.R. (1968) Estimating site index for white and black oaks in Indiana from soil and topographical factors. Journal of Forestry 66, 412–416. Harper, J.L. (1977) Population Biology of Plants. Academic Press, London. Harris, W.F. and Henderson, G.S. (1981) In: Reichle, D.E. (ed.) Dynamic Properties of Forest Ecosystems. Cambridge University Press, London, pp. 658–661. Harris, W.F., Goldstein, R.A. and Henderson, G.S. (1973) Analysis of forest biomass pools, annual primary production and turnover of biomass for a mixed deciduous forest watershed. In: IUFRO Biomass Studies. University of Maine, College of Life Sciences and Agriculture, Orono, pp. 43–64. Hartung, R.E. and Lloyd, J. (1969) Inﬂuence of aspect on forests of the Clarksville soil in Dent County, Missouri. Journal of Forestry 67, 178–182. Helms, J.A. (ed.) (1998) The Dictionary of Forestry. Society of American Foresters, Bethesda, Maryland. Hennessey, T.C., Dougherty, P.M., Kossuth, S.V. and Johnson, J.D. (1986) Stress Physiology and Forest Productivity: Proceedings of the Physiology Working Group of the Society of American Foresters. Kluwer Academic, Boston. Jakucs, P. (1981) In: Reichle, D.E. (ed.) Dynamic Properties of Forest Ecosystems. Cambridge University Press, London, 586pp.

Site Productivity

191

Johnson, F.L. and Risser, P.G. (1974) Biomass, annual net primary production, and dynamics of six mineral elements in a post oak–blackjack oak forest. Ecology 55, 1246–1258. Johnson, P.S. (1975) Growth and structural development of red oak sprout clumps. Forest Science 21, 413–418. Johnson, P.S. (1976) Eight-year performance of interplanted hardwoods in southern Wisconsin oak clearcuts. USDA Forest Service Research Paper NC NC-126. Johnson, P.S. and Rogers, R. (1980) Predicting growth of individual stems within northern red oak sprout clumps. Proceedings of the Central Hardwood Forest Conference III. University of Missouri, Columbia, pp. 420–439. Johnson, P.S. and Rogers, R. (1982) Hardwood interplanting in the upper Mississippi Valley. Proceedings of the Hardwood Regeneration Conference. University of Minnesota, St Paul, pp. 90–109. Johnson, P.S. and Rogers, R. (1984) Predicting 25th-year diameters of thinned stump sprouts of northern red oak. Journal of Forest 82, 616–619. Johnson, P.S. and Rogers, R. (1985) A method for estimating the contribution of planted hardwoods to future stocking. Forest Science 31, 883–891. Jones, E.P., Jr (1986) Slash pine plantation spacing study – age 30. USDA Forest Service General Technical Report SE SE-43, pp. 45–49. Jones, J.R. (1969) Review and comparison of site evaluation methods. USDA Forest Service Research Paper RM RM-51. Katagiri, S. and Tsutsumi, T. (1975) The relationship between site condition and circulation of nutrients in forest ecosystems. III Aboveground biomass and nutrient contents of stands. Journal of the Japanese Forestry Society 57, 412–419. Katagiri, S. and Tsutsumi, T. (1976) The relationship between site condition and circulation of nutrients in forest ecosystems. IV The amount of mineral nutrient returned to forest ﬂoor. Journal of the Japanese Forestry Society 58, 79–85. Katagiri, S. and Tsutsumi, T. (1978) The relationship between site condition and circulation of nutrients in forest ecosystems. V The differences in nutrient circulation between stands located in upper part of slope and lower part of slope. Journal of the Japanese Forestry Society 60, 195–202. Kestemont, P. (1971) Productivité primaire des taillis simples et concept de nécromasse. In: Durigneaud, P. (ed) Productivity of Forest Ecosystems. UNESCO, Paris, pp. 271–279. Killian, W. (1984) Site classiﬁcation and mapping: principles and trends. Proceedings of the IUFRO Symposium Site and Productivity of Fast Growing Plantations 1, pp. 51–68. Kozlowski, T.T., Kramer, P.J. and Pallardy, S.G. (1991) The Physiological Ecology of Woody Plants. Academic Press, San Diego, California. Lamson, N. (1980) Site index prediction tables for oak in northwestern West Virginia. USDA Forest Service General Technical Report NE NE-462. Lamson, N.I. (1987) Estimating northern red oak site-index class from total height and diameter of dominant and codominant trees in central Appalachian hardwood stands. USDA Forest Service Research Paper NE NE-605. Lanner, R.M. (1985) On the insensitivity of height growth to spacing. Forestry Ecology and Management 13, 143–148. Leary, R.A. (1985) Interaction Theory in Forest Ecology and Management. Martinus Nijoff, Dordrecht, Netherlands. Lee, R.G. (1989) The concept and measurement of multi-resource site quality. USDA Forest Service General Technical Report NC NC-135, pp. 32–39. Lee, R. and Sypolt, C.R. (1974) Toward a biophysical evaluation of forest site potential. Forest Science 20, 145–154. Lloyd, F.T. (1981) How many tree heights should you measure for natural Atlantic Coastal Plain loblolly site index? Southern Journal of Applied Forestry 5, 180–183. Lloyd, F.T. and Haﬂey, W.L. (1977) Precision and the probability of misclassiﬁcation in site index estimation. Forest Science 23, 493–499. Lloyd, F.T. and Jones, E.P., Jr (1983) Density effects on height growth and its implications for site index prediction and growth projection. USDA Forest Service General Technical Report SE SE24, pp. 329–333. McQuilkin, R.A. (1974) Site index prediction table for black, scarlet, & white oaks in southeastern Missouri. USDA Forest Service Research Paper NC NC-108.

192

Chapter 4

McQuilkin, R.A. (1975) Errors in site index determination caused by tree age variation in even-aged oak stands. USDA Forest Service Research Note NC NC-185. McQuilkin, R.A. and Rogers, R. (1978) A method for determining the precision of site index estimates made from site index prediction functions. Forest Science 24, 289–296. Mills, W.L. Jr, Fischer, B.C. and Reisinger, T.W. (1987) Upland hardwood silviculture: a review of the literature. Purdue University Agricultural Experiment Station Bulletin 527. Monserud, R.A. and Ek, A.R. (1976) Site index curves and equations for several northern hardwood forest species. University of Wisconsin School of Natural Resources Bulletin R2771. Nelson, T.C. and Beaufait, W.R. (1956) Studies in site evaluation for southern hardwoods. Society of American Foresters Proceedings, pp. 67–70. Olson, D.F., Jr (1959) Site index curves for upland oak in the southeast. USDA Forest Service Southeastern Forest Experiment Station Research Note 125. Olson, D.F., Jr and Della-Bianca, L. (1959) Site index comparisons for several tree species in the Virginia–Carolina Piedmont. USDA Forest Service Station Paper SE SE-104. Patric, J.H. (1976) Soil erosion in the eastern forest. Journal of Forestry 74, 671–677. Patric, J.H. (1977) Soil erosion and its control in eastern woodlands. Northern Logger and Timber Processor 25, 4–5, 22–23. Patric, J.H. (1978) Harvesting effects on soil and water in the eastern hardwood forest. Southern Journal of Applied Forestry 2, 66–73. Payandeh, B. (1974a) Formulated site index curves for major timber species in Ontario. Forest Science 20, 143–144. Payandeh, B. (1974b) Nonlinear site index equations for several Canadian timber species. Forestry Chronicle 47, 194–196. Reiners, W.A. (1972) Structure and energetics of three Minnesota forests. Ecological Monographs 42, 71–94. Rochow, J.J. (1974) Litter fall relations in a Missouri forest. Oikos 25, 80–85. Rodin, L.E. and Basilevich (1968) World distribution of plant biomass. In: Eckardt, F.E. (ed.) Functioning of Terrestrial Ecosystems at the Primary Production Level Proceedings of Copenhagen Symposium. UNESCO, Liège, Belgium, pp. 45–52. Rundel, P.W. (1980) Adaptations of Mediterranean-climate oaks to environmental stress. USDA Forest Service General Technical Report PSW PSW-44, pp. 43–54. Schnur, G.L. (1937) Yield, stand, and volume tables for even-aged upland oak forests. USDA Technical Bulletin 560. Schonau, A.P.G. (1988) Problems in using vegetation or soil classiﬁcation in determining forest site quality. In: Cole, D.W. and Gessel, S.P. (eds) Forest Site Evaluation and Long Term Productivity. University of Washington Press, Seattle, pp. 3–11. Smalley, G.W. (1979) Classiﬁcation and evaluation of forest sites on the southern Cumberland Plateau. USDA Forest Service General Technical Report SO SO-23. Smalley, G.W. (1982) Classiﬁcation and evaluation of forest sites on the Mid-Cumberland Plateau. USDA Forest Service General Technical Report SO SO-38. Smalley, G.W. (1984) Classiﬁcation and evaluation of forest sites in the Cumberland Mountains. USDA Forest Service General Technical Report SO SO-50. Smalley, G.W. (1986) Classiﬁcation and evaluation of forest sites on the Northern Cumberland Plateau. USDA Forest Service General Technical Report SO SO-60. Smith, D.M. (1986) The Practice of Silviculture, 8th edn. John Wiley & Sons, New York. Society of American Foresters (1995) Silviculture Terminology. Society of American Foresters, Bethesda, Maryland. Stage, A.R. (1976) An expression for the effect of aspect, slope, and habitat type on tree growth. Forest Science 22, 457–460. Stout, B.B. and Shumway, D.L. (1982) Site quality estimation using height and diameter. Forest Science 28, 639–645. Swift, L.W., Jr (1976) Algorithm for solar radiation on mountain slopes. Water Resources Research 12, 108–112. Trimble, G.R., Jr and Weitzman, S. (1956) Site index studies of upland oaks in the northern Appalachians. Forest Science 2, 162–173. Ward, W.W. and Bowersox, T.W. (1970) Upland oak response to fertilization with nitrogen, phosphorous, and calcium. Forest Science 16, 113–120.

Site Productivity

193

Waring, R.H. and Schlesinger, W.H. (1985) Forest Ecosystems: Concepts and Management. Academic Press, Orlando, Florida. Whittaker, R.W. (1963) Net production of heath balds and forest heaths in the Great Smoky Mountains. Ecology 46, 176–182. Whittaker, R.H. and Niering, W.A. (1975) Vegetation of the Santa Catalina Mountains, Arizona. V Biomass, production and diversity along the elevation gradient. Ecology 56, 771–790. Whittaker, R.H. and Woodwell, G.M. (1968) Dimension and production relations of trees and shrubs in the Brookhaven Forest, New York. Journal of Ecology 56, 1–25. Yin, X., Perry, J.A. and Dixon, R.K. (1989) Fine-root dynamics and biomass distribution in a Quercus ecosystem following harvest. Forest Ecology and Management 27, 159–177. Zahner, R. (1970) Site quality and wood quality in upland hardwoods: theoretical considerations of wood density. Proceedings of 3rd North American Forestry and Soils Conference, pp. 477–497. Zahner, R. and Myers, R.K. (1984) Productivity of young Piedmont oak stands of sprout origin. Southern Journal of Applied Forestry 8, 102–108. Zahner, R., Myers, R.K. and Churchill, L.A. (1982) Site index curves for young oak stands of sprout origin. Clemson University Forestry Bulletin 35.

5 Development of Natural Stands

Introduction Stands are the basic units of forest management. Although they can vary widely in area, oak stands typically range from 2 to 40 acres. By deﬁnition, a stand covers a relatively homogeneous area with respect to vegetation, soil and site quality. Stand boundaries are usually delineated by a combination of stand characteristics including age structure (even-aged or uneven-aged), predominant tree size (seedling, sapling, pole or sawlog) and species composition. Site homogeneity is usually evaluated through the use of site evaluation (Chapter 4) and ecological classiﬁcation methods (Chapter 1).

Forest Canopy Layers The canopy of a mature oak stand can be divided into two broadly deﬁned layers: overstorey and understorey. The overstorey consists of a main canopy and a subcanopy. The main canopy includes the upper layers of tree crowns, which intercept most of the sunlight (Fig. 5.1). Beneath the main canopy there is often a subcanopy of sapling-size trees (1–5 inches dbh), and beneath that an understorey layer that includes tree reproduction, shrubs and herbaceous plants. Understorey species are usually shade tolerant because only about 1–5% of the sunlight received by the main canopy reaches the forest ﬂoor.

The understorey is nevertheless the domain of advance tree reproduction including oak seedlings and seedling sprouts. For silvicultural purposes, it is convenient to deﬁne an upper size limit for tree reproduction. In the Central Hardwood Region, for example, 1.5 inches dbh is commonly used to deﬁne this limit; larger trees are thus deﬁned as members of the overstorey. However, this limit was probably set more by mensurational convention rather than by biological considerations. Overstorey trees can be further categorized by their relative canopy position, or crown class. Four crown classes are generally recognized: dominant, co-dominant, intermediate and overtopped (or suppressed) (Fig. 5.1). The crowns of dominant trees extend partially above the general level of the main canopy where they receive light both from above and from the sides. These trees are the tallest in a stand, and they usually have the largest crowns. The crowns of co-dominant trees receive full sunlight directly from above but little from the sides. These are among the taller trees in the stand, and their crowns deﬁne the upper level of the main canopy. The crowns of intermediate trees occupy the lower part of the main canopy. They are among the smaller trees in the stand, and receive light only on limited portions of crown tops. The crowns of overtopped trees lie completely below the main canopy in the subcanopy. Little direct sunlight is received by any part of the crowns of overtopped trees, which are often ﬂat-topped and irregular in shape. 194

Development of Natural Stands

195

D

D

C C

C I

I

I

Main canopy

I

Gap

Overstorey O

O

O

O

O

Subcanopy Understorey

Fig. 5.1. Forest canopy layers can be subdivided into overstorey (including the main canopy and subcanopy), and understorey (including tree reproduction, shrubs and herbaceous vegetation). Associated tree crown classes are: D = dominant, C = co-dominant, I = intermediate, O = overtopped (suppressed). The capacity of oak reproduction (especially seedling sprouts) to capture canopy gaps depends on their root size at the time the canopy gap is created, their inherent growth rate (which varies among oak species), gap size, competition from other vegetation and other factors.

The way trees are spatially arranged is called stand structure. Stand structure can be described by various qualitative and quantitative attributes of forests. Crown class is an example of a qualitative attribute that is useful in describing a tree’s relative crown position. By deﬁnition, crown classes infer that trees in forests are vertically stratiﬁed. However, the concept of stand structure also encompasses the areal distribution of trees including their ages and sizes, and the distribution of other stand components such as herbaceous vegetation and coarse woody debris (Helms, 1998). Although age and size structure are related to vertical stratiﬁcation, they can also be quantitatively deﬁned by frequency distributions of tree ages or sizes. The frequency distribution of tree ages is used to distinguish even-aged from uneven-aged stands and to measure other age-speciﬁc properties of stands. Stand size structure is often described by the frequency distribution of tree diameters; this is usually referred to simply as the diameter distribution. Because the diameter distribution of a stand changes as the stand develops, it is a useful diagnostic for assessing stand development. Stand structure in an even

broader context can include how closely packed the trees are. Various measures of stand density and stocking have been developed for quantifying this aspect of stand structure (Chapter 6). Stands also vary in species composition, i.e. the relative proportions of species present. Collectively, the various measures of stand structure and species composition can provide the silviculturist with essential information about the current state of a stand and, by extension, silvicultural options for its management. Central to an understanding of these options is an understanding of forest disturbance events and the related ecological processes and patterns of stand development.

Disturbance Forests continually change as a result of events originating from inside and outside the forest. These events are ubiquitous and in a broad sense can be termed ‘disturbances’. Many forest disturbances are ecologically and silviculturally important because they affect the current structure and composition of a stand as well as its

196

Chapter 5

future development. Disturbances have been variously deﬁned as ‘any relatively discrete event in time that disrupts ecosystem, community, or population structure and changes resources, substrate availability, or the physical environment’ (White and Pickett, 1985); ‘as a discrete, punctuated killing, displacement, or damaging of one or more individuals (or colonies) that directly or indirectly creates an opportunity for new individuals (or colonies) to become established’ (Sousa, 1984); or as ‘a force that kills at least one canopy tree’ (Runkle, 1985). Disturbances increase the resources available to surviving trees in the vicinity of the disturbance. The type, size, severity and frequency of disturbances greatly affect how oak stands develop.

Disturbance type Forest disturbances can be classiﬁed as exogenous or endogenous. Exogenous disturbances originate from forces external to the stand and may damage or kill vigorous as well as unhealthy trees. Examples include disturbances caused by weather, ﬁre and human activity including silviculture. Endogenous disturbances originate within the stand, and tree death and treefall are primary causes. It is not always apparent which type of disturbance has occurred. For example, stresses resulting from the competition between trees (an endogenous factor) may predispose a tree to a fatal insect defoliation or windthrow (exogenous factors). The ecological impact of a disturbance may depend more on its size and severity than its origin (White and Pickett, 1985). Exogenous and endogenous factors represent endpoints of disturbance size that can range from the creation of extensive forest canopy openings with abrupt boundaries to small changes in stand structure associated with the toppling of a single overstorey tree (White and Pickett, 1985). Although exogenous factors typically produce larger disturbances than endogenous factors, endogenous events nevertheless can have substantial localized impacts. For example, when a large, senescent tree falls

it may shear the crowns of surrounding trees, topple or crush other trees in the line of fall, disturb the soil at points of impact, and uplift a mound of soil in its root mass (Bormann and Likens, 1979).

Disturbance size and frequency of occurrence Disturbances of one type or another occur frequently during the life of every oak stand, and each one changes the structure and composition of the stand to some degree. Reice (1994) suggested that to be in recovery from their last disturbance is the normal state for most natural plant communities, and that they rarely attain a state of equilibrium. This is certainly true of oak forests, which owe their very existence to disturbance. Most are therefore in a state of recovery and transition related to their disturbance history. Disturbances vary greatly in their area, severity and frequency, and in associated ecological responses (Sousa, 1984). Even seemingly small disturbances such as animal browsing or leaf litter ﬁres may signiﬁcantly alter a forest if they are widespread and frequent. Disturbances that create large openings in the forest canopy are of particular interest to ecologists and silviculturists because such openings signiﬁcantly alter the course of stand development. Many silvicultural practices fall into this category, and silvicultural treatments can be thought of as disturbances that are designed to control future stand composition and structure. From a silvicultural perspective, forest disturbances can be categorized according to their relative size and impact on stand development: (i) gap-scale disturbances; (ii) incomplete stand-scale disturbances; and (iii) stand-initiating disturbances. Gap-scale disturbances are the smallest and can be endogenous or exogenous in origin. The other two categories result from exogenous forces. The impact of a disturbance on a stand should be evaluated in the context of stand structure and composition. A disturbance of a given type

Development of Natural Stands

and size will usually elicit a different response from a young stand than from a mature stand or among stands that differ in species composition. Gap-scale disturbances occur when a single tree or a small group of trees are lost from the main canopy. The resulting canopy gaps increase light and soil moisture available to trees within the gap and to trees adjacent to the gap. When stands are well stocked and pole-size or smaller, canopy gaps created by the death of small trees are quickly ﬁlled through crown expansion of the surrounding trees. During later stages of stand development, gaps created by the death or harvest of individual trees often are larger and thus may persist for decades. Silvicultural methods designed to create gap-scale disturbances include some types of thinning (Chapter 7), and single-tree and group selection methods (Chapter 8). During a century of even-aged oak stand development, more than 95% of the trees initially present will die. Most of those deaths are caused by inter-tree competition and consequent self-thinning of stands (Chapter 6). As a result, canopy gaps are a common occurrence in oak stands. Although trees of all crown classes die from self-thinning, mortality rates are lower for dominant and co-dominant trees than for trees in subordinate crown classes. For example, in New England stands, 9% and 33% of northern red oaks in dominant and co-dominant crown classes, respectively, died between stand ages 25 and 55. In contrast, 67% and 90% of trees in intermediate and suppressed classes, respectively, died over the same period (Ward and Stephens, 1994). When an individual tree dies, it leaves a canopy gap that is roughly proportionate in area to its basal area. The vacated space is subsequently captured by the expanding crowns of surrounding trees and/or by tree reproduction. Larger gap-scale disturbances occur when small clusters of trees die at the same time, or when individual trees with large crowns die. As gap size increases, the more abrupt and extensive are the associated stand changes. The size of the gap, rate of crown

197

expansion of surrounding trees, and the growth rate of trees within the gap all inﬂuence future stand composition and structure (Hibbs, 1982). Incomplete stand-scale disturbances are larger and more disruptive to the stand than gap-scale disturbances. They are of exogenous origin and caused by events that affect a large proportion of a stand’s area. An incomplete stand-scale disturbance removes enough overstorey trees to initiate the development of a new age class of trees within the stand. The resulting stand age structure is often irregular (i.e. the tree age classes occupy unequal areas), and there may be large variation in the size and number of openings in the canopy. Silvicultural practices that can create incomplete standscale disturbances include seed tree, diameter limit, shelterwood and group selection methods. Fires of sufﬁcient intensity to kill patches of overstorey trees also can result in incomplete stand-scale disturbances. The largest and most severe disturbances are those that initiate a new stand. These disturbances remove most or all of the overstorey. Some residual canopy trees may remain, but the net result is the creation of a new, even-aged stand. High wind, severe ﬁre, and other destructive events that are an acre or larger can cause stand-initiating disturbances. Silvicultural treatments such as clearcutting and shelterwood removal harvests (Chapter 7) are also stand-initiating disturbances. A disturbance of this type resets stand age to or close to zero. In evenaged stands, stand age is thus a measure of the time elapsed since the previous standinitiating disturbance. In managed forests, silvicultural treatments are typically implemented at 10–30 year intervals. Gap-scale disturbances within a stand generally occur with greater frequency than that and are anticipated in silvicultural prescriptions. However, standinitiating disturbances or incomplete stand-scale disturbances are infrequent events that are generally not anticipated in a managed stand except as a prescribed timber harvest. Silvicultural prescriptions nevertheless can be designed to reduce the negative consequences of extensive natural

198

Chapter 5

disturbances such as severe ﬁre, wind or insect damage, which inevitably affect some managed stands. When a disturbance creates an opening of any size in the forest, it increases the availability of resources in and around the opening. Woody vegetation responds in three ways. First, the crowns of trees surrounding the opening grow laterally to capture the space. Although the rate of lateral crown expansion depends on many factors, oaks expand at rates from 6 to 12 inches per year in eastern forests (Fig. 3.22; Trimble and Tryon, 1966; Hibbs, 1982). Crowns of northern red oaks often expand more rapidly than crowns of co-occurring species. In general, crowns of mature trees tend to expand more slowly than young, vigorous trees. And mature oaks in the superior crown classes may not expand at all (Sampson, 1983). Large trees in mature stands nevertheless leave large gaps when they die. As stands mature, there are more opportunities for such gaps to form, and those that persist for decades may facilitate the establishment of oak and other tree species in the understorey. A second response to gap formation is an increase in the height growth of preestablished tree reproduction and subcanopy trees within the gap. This response, in turn, can facilitate the successional replacement of canopy dominants by trees that have persisted in intermediate and suppressed crown classes and that have the capacity to opportunistically expand into canopy gaps. Gap formation accordingly can potentially accelerate the recruitment of shade-tolerant trees and even some fastgrowing, shade-intolerant trees into the overstorey. Whether or not such a replacement process occurs varies with the species composition and the physiological capacity of subordinate canopy layers to capture and hold canopy gaps. A third response to canopy gap formation is the establishment of new trees (and other vegetation). If the openings are large enough and persist long enough, trees that become established in a gap after its formation may eventually grow into the overstorey. In a hemlock–mixed hardwood

forest, seedlings (including oaks) required a canopy opening with a diameter at least half the height of the surrounding trees to facilitate growth into the overstorey (Hibbs, 1982). In smaller openings the crown expansion of surrounding trees usually closed the gap before new trees reached the main canopy. Not only does a canopy gap close laterally with time (Fig. 3.22), the minimum area required to sustain a tree within the gap increases as the tree itself increases in size. A tree therefore must reach the level of the main canopy before its own minimum growing space requirements exceed the gap’s diminishing size. Thus, recruitment of new trees within a gap depends not only on the species, initial size, vigour and growing space requirements of trees within the gap, but also on the size, crown expansion rate and spatial arrangement of trees surrounding the gap. Disturbances that eliminate oaks in the main canopy without also eliminating shade-tolerant trees in the sub-canopy may accelerate succession towards dominance by shade-tolerant species (Abrams and Scott, 1989; Abrams and Downs, 1990; Abrams and Nowacki, 1992). However, moderate overstorey disturbances sometimes favour the oaks. For example, prolonged drought reduced numbers of yellow-poplars, which consequently beneﬁted competing oaks in an Ohio stand (Hilt, 1985). Where oaks and yellow-poplar co-occur, yellow-poplar typically overtops the oak within 10 years of a stand-initiating disturbance (Loftis, 1983, 1990; Beck and Hooper, 1986; Weigel and Johnson, 1999). Consequently, when oaks compete with yellow-poplar, the oaks usually lose. Exceptions occur when a disturbance such as drought or ﬁre selects against yellow-poplar and alters the usual course of succession. In the absence of timber harvesting, the average rate of canopy disturbance across a wide range of temperate forest conditions is about 1% per year and ranges from about 0.5 to 2% (Runkle, 1985). The average time between successive replacement of trees in the main canopy is therefore about 100 years (range 50–200 years). This is consistent with the observed longevity of oaks.

Development of Natural Stands

Although individual oaks may live 400 years or longer, attaining such ages is rare. Even in old-growth oak forests, only a small fraction of trees survive more than 200 years. The distribution of disturbances varies in both time and space. When a disturbance occurs, it temporarily reduces the likelihood of a subsequent disturbance of the same type. For example, high winds that remove the most susceptible trees from a stand usually also reduce the likelihood of additional wind damage for several years thereafter. Likewise, periodic ﬁres that eliminate the most ﬁre-sensitive trees and reduce fuels decrease the likelihood of intense ﬁre damage for several years thereafter. Large or severe disturbances therefore tend to occur at infrequent intervals and are followed by years with below-average rates of disturbance. Suppression of disturbances (e.g. through ﬁre suppression) may increase the likelihood of a severe disturbance in the future. Frequent low-intensity disturbances that remove some trees may increase the growth rate and vigour of surviving trees, and also may reduce the number and impact of subsequent disturbances (Waring and Schlesinger, 1985). However, this is not always the case. Certain combinations of events (e.g. windthrow followed by ﬁre) may create conditions (e.g. increased fuels) capable of producing catastrophic events. Just as the size and species composition of a stand are affected by a disturbance, the reverse is also true. The area, spatial pattern and intensity of a natural disturbance may be modiﬁed by the size and species composition of the vegetation and by the landform where it occurs (Reice, 1994). For example, certain species and certain topographic positions are especially susceptible to wind and ﬁre damage. Temporal and spatial variation in disturbance events also tends to be interrelated (Runkle, 1985). During years when the total disturbed area is relatively large, disturbances are likely to be spatially clustered. This happens either because a few disturbances cover an exceptionally large area or because a large number of smaller disturbances occur within a

199

ﬁxed area, and many disturbed areas are adjacent to one another. As stands change over time, their susceptibility to disturbance also may change (White and Pickett, 1985). Small trees are usually more susceptible to ﬁre damage than large trees. Large trees with welldeveloped crowns are more vulnerable to wind damage than small trees. Changes in species composition that occur during the course of stand development also may alter a stand’s susceptibility to disturbance by agents such as ﬁre, ﬂooding, insects, disease or silvicultural activity.

Development of Even-aged Stands An even-aged stand consists of a cohort of trees comprising a single age class. This usually means that tree ages differ by no more than 20 years. In silvicultural applications, a more speciﬁc deﬁnition is sometimes preferred. For example, Smith (1986) deﬁned an even-aged stand as one where the difference in age between the youngest and oldest trees does not exceed 20% of the rotation (Chapter 7). Although the biological age of an oak (i.e. its age from germination) is most accurately expressed by the number of annual rings just below its root collar (Chapter 3), stand age is conventionally deﬁned by the average age of tree boles. This age usually corresponds to the number of years since the previous standinitiating disturbance. Even-aged stands progress through a relatively predictable series of developmental stages until the next stand-initiating disturbance or incomplete stand-scale disturbance occurs. Deﬁning these stages is useful in understanding the development of oak forests even though the duration of each stage and the accompanying changes in stand structure, density and species composition may differ from stand to stand. Although various terms have been used to deﬁne the stages of stand development (e.g. Bormann and Likens, 1979), we herein follow the terminology of Oliver and Larson (1996) as modiﬁed by Oliver (1997). Four stages have been deﬁned: (i) the stand initi-

200

Chapter 5

ation stage; (ii) the stem exclusion stage; (iii) the understorey reinitiation stage; and (iv) the complex stage (Fig. 5.2). The complex stage of development also has been called the old-growth stage (Oliver, 1981; Oliver and Larson, 1996). The progression of an even-aged oak stand through the various stages of development is accompanied by changes in stand structure and often by changes in species composition. Changes in stand size structure include the accumulation of biomass (Bormann and Likens, 1979), vertical stratiﬁcation of tree crowns and correlated changes in diameter distributions. For some regions, these changes are generalized in published stand tables that list by site classes the expected numbers of oaks and other species

Stand initiation stage

Stem exclusion stage

by tree diameter and age classes in wellstocked stands (e.g. Schnur, 1937; Gevorkiantz and Scholz, 1948). Such tables thus represent models of average expected change in stand structure associated with different stages of stand development. Unlike changes in stand size structure, which follow a similar progression as stands age (Fig. 5.2), changes in species composition often vary from stand to stand, even within similar ecological settings. Species composition in even-aged oak stands largely depends on the size and initial species composition of tree reproduction and propagules present at the time of the previous stand-initiating disturbance. The trees and the other vegetation present at that time develop into the new

Understorey reinitiation stage

Complex stage

Fig. 5.2. Stages of stand development that occur after a major disturbance that destroys all or most of the parent stand. Stand initiation stage: immediately after the disturbance, pre-established reproduction grows rapidly and new trees and other plants appear. In oak stands this stage typically lasts 10–20 years. Stem exclusion stage: no new trees appear and many die from crowding. Trees are well stratified into crown classes by the end of this period. This stage usually begins after the stand reaches 10 years of age and concludes before age 70. Understorey reinitiation stage: tree reproduction becomes re-established under the maturing overstorey. Re-establishment of trees in the understorey is facilitated by the death of individual trees in the main canopy. Canopy gaps are of sufficient size and frequency to significantly increase light on the forest floor. This stage typically begins after age 50 and concludes before age 120. Complex stage: natural mortality of large overstorey trees produces irregular canopy gaps and accelerates the recruitment of reproduction and sub-canopy trees into the overstorey and main canopy, respectively. This stage marks the transition from an even-aged stand to an uneven-aged stand. Oak forests typically require 100 years or longer to reach the complex stage of development. The stated durations for the four stages of stand development are representative of oak forests of the eastern United States and assume that no significant stand-scale disturbances occur. Actual durations of stages of development vary with species composition, site productivity and other factors. (Adapted from Oliver and Larson, 1996; Oliver, 1997.)

Development of Natural Stands

oak stand. Differences among tree species in shade tolerance, longevity and growth potential inﬂuence how the species composition of the stand will change over time. Changes in composition are also inﬂuenced by disturbances such as extreme weather events, disease, animal browsing and human activity. Often these events are only predictable probabilistically. Consequently, young stands of similar composition and structure may follow different developmental trajectories and thus differ greatly in composition and structure at maturity. Structural and compositional changes in forests are attributes of the broader process of ecological succession. Although ecological succession is generally regarded as the natural change in the composition, structure and function of an ecosystem, successional concepts are also applicable and central to silviculture. In this context, silviculture is about directing ecological succession. Ecological succession is of two types: primary and secondary. Primary succession involves long-term ecosystem changes, usually beginning with bare ground devoid of vegetation, and often spanning thousands of years. Related changes in vegetation are associated with changes in the parent material, soil and other factors that are interactively modiﬁed through time. Secondary succession involves shorterterm changes occurring after a disturbance. Secondary successions therefore usually begin with vegetation, plant propagules and other organisms already in place. Thus, there is a direct connection between ecological succession and silvicultural practice, which seeks to design and create disturbances that direct secondary succession in speciﬁc ways. The study of secondary forest succession has traditionally focused on changes occurring over relatively long periods and in the absence of signiﬁcant exogenous disturbance. However, one objective of silviculture is to anticipate and control successional changes in stand structure and composition that result from stand disturbances. In turn, such disturbances result from endogenous events associated with natural stand development (e.g. the death

201

of individual canopy trees), from natural exogenous events (e.g. wildﬁre or weather) or from human intervention (e.g. silviculture). Silvicultural prescriptions are thus applied at various stages of stand development with the intent of directing the course of change in stand structure and species composition in speciﬁc ways. Before prescribing silvicultural treatments, it is therefore important ﬁrst to understand how even-aged stands progress through the stages of structural development in the absence of major exogenous disturbances.

The stand initiation stage The development of an even-aged stand begins with the stand initiation stage. In oak forests, this stage may last up to 20 years. In temperate regions it is characterized by a brushy mass of woody vegetation comprised of thousands of trees and shrubs per acre often mixed with a luxuriant growth of vines and herbaceous plants (Gingrich, 1971) (Fig. 5.3A). It is also a period of rapid change with intense competition among trees and other plants for growing space. Standing biomass is small relative to later stages of development, but the rate of biomass increase is high. During this stage, the quantity of dead biomass is often larger than during other stages of stand development. This is due to the standinitiating disturbance itself, which (except for ﬁre) usually leaves a large residue of tree boles and branches on the forest ﬂoor (Bormann and Likens, 1979; Jenkins and Parker, 1997; Spetich et al., 1999). During the stand initiation stage, gaps in the new vegetative cover may persist for a decade or longer as new trees and other vegetation become established. New tree seedlings and herbaceous vegetation initially require little growing space, and numerous small openings in the developing forest vegetation provide ‘safe sites’ for their establishment (Harper, 1977). These are places where seeds ﬁnd the necessary conditions for germination and growth free from predators, competitors and pathogens. Changes in the number, species and size of

202

Chapter 5

A

B

C

D

Fig. 5.3. The four stages of stand development representative of oak stands in the Ozark Highlands of Missouri: (A) stand initiation stage; (B) stem exclusion stage; (C) understorey reinitiation stage; and (D) complex stage. (USDA Forest Service, North Central Research Station photographs.)

trees during the stand initiation stage are difﬁcult to predict accurately. This is due to the numerous and essentially random

events that inﬂuence the timing and spatial patterns of seed dispersal, germination and seedling survival. Stand development

Development of Natural Stands

during this period is subject to great natural variation, and predictions of stand development during this stage are often speciﬁed qualitatively (Johnson and Deen, 1993) or probabilistically (Chapter 3; also see Johnson and Sander, 1987; Loftis, 1990; Dey, 1991; Dey et al., 1996). Although the composition of the future stand is strongly inﬂuenced by the species initially present at this stage of development, the relative dominance of the various species usually changes as the stand matures. Early theories of forest succession (e.g. Clements, 1916) proposed that each species modiﬁes the site to make it more favourable for the establishment and growth of succeeding species. However, studies of mixed oak–hardwood forests in New England have shown that most of the trees that form the overstorey originated within one or two decades after a major disturbance (Oliver and Stephens, 1977; Oliver, 1978). Oaks that attain membership in the main canopy are usually established before the stand-initiating disturbance occurs (Chapter 2). Species that are shade tolerant, longlived and capable of growing to large size may persist in subordinate canopy positions for decades. They may gradually make their way into the main canopy by opportunistically responding to gap-size disturbances. At stand initiation, the presence of 100 seedlings or seedling sprouts per acre of a late-successional species might seem insigniﬁcant. However, if that population persistently captures canopy gaps, over time it may eventually dominate the main canopy. Different species that become established at approximately the same time therefore may exert dominance at different stages of stand development (Egler, 1954; Oliver and Larson, 1996; also see Chapter 3). The overstorey of a mature even-aged oak stand is largely comprised of species present at the end of the stand initiation stage. During the two stages of stand development that follow the stand initiation stage, few if any new trees are added to the overstorey. Consequently, the composition of an even-aged stand at the end of the

203

stand initiation stage is a good indicator of the composition of the overstorey of the future stand. However, a species’ relative abundance at this stage is often a poor indicator of its future importance. During the ensuing stem exclusion stage, species composition usually shifts toward the species that are best adapted to the site and able to attain and persist in dominant and co-dominant crown positions. Dominance by oaks at the end of the stand initiation stage therefore does not, in itself, ensure their continued dominance.

The stem exclusion stage In temperate regions, crown closure in oak stands has usually occurred by the beginning of the second decade after a stand-initiating disturbance. By that time, trees have stratiﬁed into well-deﬁned crown classes and natural mortality has changed the initially clumped spatial distribution of trees to a more random distribution (Rogers, 1983). This stage of stand development is termed the stem exclusion stage because few, if any, new stems are added to the population of overstorey trees (Oliver, 1980; Oliver and Larson, 1996) (Figs 5.2 and 5.3B). Mortality rates are high, especially among trees in intermediate and suppressed crown classes. The combined growth, competition and mortality of trees during this stage produce spatial adjustments in the main canopy that lead to full or nearly full crown closure and a corresponding full utilization of growing space (Gingrich, 1967, 1971). It is usually not until after the stem exclusion stage begins that growth and yield tables or other predictive models are applicable (Chapter 10). During this and subsequent stages, patterns of stand development and species differentiation are more predictable than during the stand initiation stage. Dense populations of saplings (trees 1–5 inches dbh) and pole-size trees (5–10 inches dbh) dominate the stem exclusion stage. During this stage, the expanding crowns of dominant and co-dominant trees quickly ﬁll the space vacated by dead and dying

204

Chapter 5

oaks to eventually grow taller than subcanopy competitors (e.g. ﬂowering dogwood) that may initially overtop oaks. Persistence thus involves negative impacts on two or more populations, but one population emerges the ‘winner’ because it persists and is less negatively affected than the other. However, the persistence advantage of one species over another often changes across environmental gradients such as the moisture gradient. The diameter distribution of an evenaged stand changes continually as the stand develops. For example, in 10-yearold upland oak stands of the eastern United States, diameter distributions form reverse J-shaped curves comprised of thousands of trees per acre (Fig. 5.4). By the time stands reach a mean diameter of 8 inches dbh and in the absence of disturbance, a bell-shaped diameter distribution comprised of a few hundred trees per acre has formed (Schnur, 1937). On sites of average quality, this occurs at about stand age 85, which is well into the understorey reinitiation stage of development. In the absence of exogenous disturbance,

trees of the inferior crown classes. Even if tree reproduction is present in the understorey, it is usually unable to capture the small, transient canopy gaps that occur during this stage. Little light reaches the forest ﬂoor and understorey vegetation decreases as the herbaceous and low woody vegetation that ﬂourished during the stand initiation stage die from suppression beneath the rising level of the overstorey. During the stem exclusion stage, oaks can sustain a position of dominance in three ways: (i) through inherently faster growth than competitors; (ii) through an initially superior crown position; or (iii) through persistence. Persistence involves survival and continued growth of a species when the population of competitors fares less well under the same conditions. For example, proportionately more oaks may survive drought than faster-growing and potentially long-lived competitors such as red maple or yellow-poplar (Hilt, 1985). Oaks are relatively persistent following disturbance events such as ﬁre in uplands or ﬂood scouring in bottomlands. In some ecosystems, persistence also may allow

800 10* 700 Trees per acre

3000

Trees per acre

600 20 500

2000 1000

400 0 300 30

200

40

100

1

5 9 13 Dbh (inches)

17

50 60 80

*Stand age

0 2

4

6

8 10 12 Dbh (inches)

14

16

Fig. 5.4. Diameter distributions of trees for a time-series of even-aged upland oak stands on average sites in the eastern United States. Inset: The reverse J-shaped diameter distribution resulting from summing the number of trees in each dbh class across all even-aged diameter distributions (for 10-year age classes from 10 to 100). (Adapted from Schnur, 1937.)

Development of Natural Stands

numbers of trees decline with time through the process of self-thinning (Chapter 6). The development of a bell-shaped diameter distribution is not a universal characteristic of even-aged oak stands. Because oak stands are usually mixtures of species, forest canopies in later stages of development tend to stratify vertically by species. This stratiﬁcation results from species’ differences in shade tolerance, longevity, growth rate and maximum attainable size. Slower growing, shade-tolerant trees may persist for decades in subordinate crown classes (Oliver, 1980; Hibbs, 1983). Differential rates of establishment and growth among species in even-aged stands may produce reverse J-shaped and other types of diameter distributions often associated with uneven-aged stands. However, the correlation between tree age and diameter is often low – and near zero in truly even-aged stands. Diameter distributions, by themselves, therefore are not reliable indicators of age distributions, especially in stands comprised of several species (Oliver, 1980; Loewenstein, 1996; also see Chapter 8). The development of an Arkansas sweetgum–red oak stand illustrates the transition from the stand initiation stage to the stem exclusion stage of stand development (Johnson and Krinard, 1983, 1988). This bottomland stand in the Southern Hardwood–Pine Region originated from a timber harvest that left only a few seed trees per acre. Unmerchantable trees were killed and areas inﬂuenced by seed trees were excluded from the analysis. Consequently the results essentially describe the development of a stand after clearcutting. Estimated site index for sweetgum was 100 ft at 50 years. Numbers of trees and stand basal area increased for the ﬁrst 15 years (Fig. 5.5A). By age 18, the declining number of trees and increasing basal area marked the onset of the stem exclusion stage (Fig. 5.5B). The transition from stand initiation stage to stem exclusion stage is also apparent from the simultaneous changes in basal area and numbers of trees (Fig. 5.5C). Between stand ages 3 and 9, numbers of

205

trees rapidly increased as trees ﬁlled the growing space vacated by the harvested parent stand. By age 15, the new stand had attained its maximum number of trees. Thereafter, the trajectory reversed direction, i.e. basal area continued to increase but numbers of trees decreased. This change in direction also marked the beginning of the stem exclusion stage and the onset of competition-induced self-thinning (Chapter 6). Despite the decreasing numbers of trees, the stand continued to increase in basal area after age 15. Although the basal area of undisturbed stands usually increases until stands are in the complex stage of development, the periodic rate of basal area increment slows with increasing stand age. This slowing is indicated by the smaller annual increments of basal area that occurred after age 15 (Fig. 5.5B and C). It is during the stem exclusion stage that oaks must capture growing space from their competitors if they are to dominate the later stages of stand development. When numbers of trees decline, it is usually the smallest trees that die ﬁrst. The vacated growing space and thus the greatest basal area increment usually accrues to large trees of superior crown classes. The time-dependent change in the relation between basal area and numbers of trees also can vary among species in the same stand. For example, sweetgum and river birch followed nearly identical trajectories through stand age 9 (Fig. 5.5D). Because river birch declined in both numbers of trees (after age 9) and in basal area (after age 18), it was nearly excluded from the stand by age 29. Although the trajectory for American hornbeam was similar, it did not decline in numbers of trees until later, which may be related to its greater shade tolerance. Although numbers of red oaks also declined after age 18, their survival rate was greater than that of other species. The combination of rapid growth and a low mortality rate thus enabled the oaks to attain a dominant position in the stand even though they were relatively few in number. Nevertheless, after 29 years of stand development, sweetgum and other non-oaks accounted for more basal area and trees per acre than the oaks.

206

Chapter 5

2500

160 B

A 140

1500

1000 Stand initiation stage

Stem exclusion stage

Basal area (ft2 acre–1)

Trees per acre

2000 120 100

Stand initiation stage

80

Stem exclusion stage

60 40

500 20 0

0 0

5

10

15

20

25

0

30

5

10

160

(18)

100

(15)

80 60 (9)

40

30 Red (29) oaks

(29)

20

River birch

1000

(18)

American hornbeam

(18)

(18) (18)

(29)

(9)

(6)

0 500

(18)

(29)

(3) 0

D

Other major species

(6)

0

30

Sweetgum (29)

40

10

20

25

50

(23)

Basal area (ft2 acre–1)

Basal area (ft2 acre–1)

60

C

(29)

120

20

Stand age (years)

Stand age (years)

140

15

1500

2000

2500

Trees per acre

0

200

400

600

Trees per acre

Fig. 5.5. Change in number and basal area of trees ≥1 inch dbh in an even-aged sweetgum–red oak stand in southeastern Arkansas during its first 29 years of development. (A) During the first 15 years, the increasing number of trees ≥1 inch dbh is characteristic of the stand initiation stage of development. After age 15, the decreasing number of trees ≥1 inch dbh marks the stem exclusion stage of development. (B) Basal area increased through both stages of development. (C) Change over time in basal area and number of trees for all species combined. Plotted lines trace changes from stand age 3 to 29 (stand ages are shown in parentheses). (D) Change over time in basal area and number of trees for five species groups. All species begin near the origin at stand age 3. The plotted age sequence through 29 years is the same as shown in (C). Red oaks include cherrybark, water and willow oaks. (Adapted from Johnson and Krinard, 1988.)

Stocking charts are commonly used to quantify the degree of tree crowding (and thus the utilization of growing space) in relation to basal area and trees per acre

(Chapter 6). Although stocking charts are used primarily to evaluate stand density at a single point in time, they are also useful in illustrating stand development through

Development of Natural Stands

time (e.g. Fig. 5.5C and D). Accordingly, they can be used to show how stands change in composition and structure, and in the proportionate utilization of growing space by different species. For example an even-aged sweetgum–red oak stand in a southern bottomland maintained a reverseJ diameter distribution during its ﬁrst 29 years of development (Fig. 5.6). But over time the shape of the diameter distribution gradually ﬂattened as average tree diameter increased and as numbers of trees in the smaller diameter classes decreased. After the stand entered the stem exclusion stage, the shape of the diameter distribution of the oaks departed from the overall diameter distribution. By stand age 23, the largest oaks in the stand formed a bell-shaped diameter distribution, whereas smaller oaks maintained a reverse J-shaped distribution (Fig. 5.6). Oaks larger than 6 inches dbh formed a ‘hump’ in the diameter distribution, which in previous years formed a reverse J-shape. By age 29, the red oaks thus diverged into two distinct populations: a population of main canopy trees and a population of sub-canopy trees. The structures of these two sub-populations appear destined to diverge further as the largest oaks continue to increase in dominance and the smaller oaks become increasingly suppressed beneath the rising level of the main canopy. Changes in species composition commonly occur after oak stands enter the stem exclusion stage and trees begin to differentiate into well-deﬁned crown classes. Such changes are reﬂected in differences in proportions of species across diameter classes. Although after stand age 15 the total number of trees per acre decreases with increasing stand age and mean stand diameter (Fig. 5.5A and C), the number of trees of any given species may increase or decrease as a proportion of all trees as the stand matures. For example, sweetgum and other non-oaks were the only trees in the largest diameter classes (≥11 inches dbh) at stand age 18, whereas the largest trees in the red oak group (cherrybark, water and willow oaks) were all less than 7 inches dbh at the same age. However, by stand age 29 red

207

oaks represented half the trees in the largest (14 inches) dbh class. Although total numbers of trees decreased by nearly 50% between stand ages 18 and 29, numbers of oaks decreased by only 13%. Because of the rapid diameter growth of the largest oaks, they were among the largest trees by age 29. Thus, it was not until the stand was more than 10 years into the stem exclusion stage that the oaks emerged as a major component of the main canopy. Although sweetgum still dominated the stand after 29 years, the red oaks had increased in dominance at the expense of other species. Oak stands on droughty sites or sites affected by recurrent, low-intensity disturbances (e.g. oak savannas maintained by periodic burning) may not experience a stem exclusion stage. In oak savannas, periodic burning maintains a relatively open canopy, which in turn maintains relatively high light intensities in the understorey. Although oak and other tree reproduction usually accumulate under these conditions, their recruitment into the overstorey is inhibited by burning unless there is a ﬁre-free period of sufﬁcient duration to permit their recruitment into the overstorey (Chapter 9). Under these conditions, recruitment of oaks into the overstorey is limited more by the disturbance regime than by competition or light.

The understorey reinitiation stage During the understorey reinitiation stage, tree reproduction becomes re-established beneath the parent stand (Figs 5.2 and 5.3C). In oak forests, this reproduction usually becomes a major component of the new stand that develops after the next stand-initiating disturbance. Many factors inﬂuence which species become established in the understorey and consequently which species are likely to dominate after a stand-initiating disturbance. In oak forests, light and soil moisture rank among the most important of these factors (Fig. 3.7). This is also the stage of development when oak stands usually attain economic maturity.

208

Chapter 5

All trees

Trees per acre

Trees per acre

Trees per acre

Trees per acre

Trees per acre

Trees per acre

1600

Age 3

120

1200

90

800

60

400

30

0 1600

120

Age 6

90

800

60

400

30

0 1600

0 120 Age 9

1200

90

800

60

400

30

0 1600

Age 15

0 120

1200

90

800

60

400

30

0 1600

Age 18

0 120

1200

90

800

60

400

30

0 1600

Age 3

0

1200

Age 23

0 120

1200

90

800

60

400

30

0

0 120

1600 Trees per acre

Red oaks

Age 29

1200

90

800

60

400

30

Age 6

Age 9

Age 15

Age 18

Age 23

Age 29

0

0 0 2 4 6 8 10 12 14 16 Dbh (inches)

0 2 4 6 8 10 12 14 16 Dbh (inches)

Fig. 5.6. Change in diameter distributions in an even-aged sweetgum–red oak stand in southeastern Arkansas during its first 29 years of development. Until age 15, the stand was in the stand initiation stage of development and thereafter was in the stem exclusion stage. By age 23, the red oaks formed two distinct populations: large trees (6–10 inches dbh) occupying the main canopy, and smaller trees occurring in the sub-canopy. The number of trees is plotted over the midpoints of 2-inch dbh classes. Red oaks include cherrybark, water and willow oaks. (Adapted from Johnson and Krinard, 1988.)

Development of Natural Stands

Consequently, it is a critical period from a silvicultural perspective. Compared to the stem exclusion stage, trees in the main canopy are larger and fewer in number during the understorey reinitiation stage. At this stage of stand development, some of the main canopy trees periodically produce large quantities of seed. Large crowns are important determinants of acorn production (Chapter 9) and thus the establishment of oak seedlings (Fig. 3.8). Moreover, when large trees die they create larger canopy openings than those created during earlier stages of stand development. The crowns of large main canopy trees also expand more slowly than they did during earlier stages of stand development. Consequently, canopy gaps remain open for longer periods, and this often results in light intensities near the forest ﬂoor that are sufﬁcient for seedling establishment and growth. The consequent spatial heterogeneity of the main canopy creates spatial variation in the amount of light reaching the forest ﬂoor. This produces microenvironments favourable for the establishment and growth of oak and other tree reproduction. Established trees in subordinate crown positions also beneﬁt from the increased growing space when large overstorey trees die. In oak forests, the onset of the understorey reinitiation stage typically occurs one or two decades before the end of a silvicultural rotation, i.e. stand age at ﬁnal harvest, which commonly ranges from 80 to 120 years in the eastern United States. Progression to this stage of development may be accelerated by silvicultural actions or by natural disturbances. The establishment of oak reproduction during the understorey reinitiation stage is usually a prerequisite to successful oak regeneration (Chapter 3). The tree reproduction, shrubs and herbaceous vegetation that develop during this period set the stage for regeneration cuttings. If the stand is not regenerated through timber harvesting or other disturbances, some of the reproduction established during this stage of development may eventually grow into the overstorey as the stand approaches the next

209

stage of development – the complex stage. The successional replacement of oaks by more shade tolerant species is one of the most pervasive problems associated with oak silviculture in mesic and hydric forests. The understorey reinitiation stage is therefore a critical time for intervening silviculturally if the objective is to maintain or increase the proportion of oak in the future stand (Chapter 7). However, not all oak forests are successional to non-oaks. In the Ozark Highlands and similar xeric oak forests in the eastern United States, the successional displacement of oaks is limited by the inability of other hardwoods to persist in the superior crown classes beyond the stand initiation stage. Although non-oaks on these sites may aggressively ﬁll canopy gaps immediately after disturbance, their dominance is usually shortlived. In the Ozark Highlands, differences among species in their competitive capacity are reﬂected in their probabilities of attaining an intermediate-or-better crown class after clearcutting. For a given initial (preharvest) basal diameter, these probabilities are higher for oaks than for other hardwoods 15 years after cutting (Fig. 5.7A). By that time, oaks dominate the stands and species are stratiﬁed into well-deﬁned crown classes (Fig. 5.7B). This outcome is the result of the collective inﬂuence of initial ﬂoristics, overstorey inhibition and competitive sorting processes (Chapter 3) that control secondary succession in oak-dominated ecosystems such as the Ozark Highlands. Oak forests in other ecoregions may develop differently. An example is provided by a composite analysis of four mixed oak–hardwood stands in Connecticut during the stem exclusion and understorey reinitiation stages (Ward et al., 1999). When ﬁrst measured, the stands were 25 years old, compositionally similar, and oaks comprised 40% of the basal area. Other major species included red maple, yellow birch, black birch and white ash. Mean site index was 67 ft at age 50. At age 25, the stands were well into the stem exclusion stage of development. During the next 70 years, numbers of trees decreased

210

Chapter 5

0.8 A

P

Oaks

0.6

Hickories and Blackgum

0.4

Sassafras

0.2

Flowering dogwood

0 0

0.4

0.8

1.2

1.6

2

2.4

Preharvest basal diameter of reproduction (inches) B O O

H

O O

BG O

O S

BG

D

BG

S D

S

S D O

O = Oaks H = Hickories BG = Blackgum S = Sassafras D = Flowering dogwood Fig. 5.7. (A) The estimated probability (P) that 15 years after overstorey removal a genet of advance reproduction (i.e. an individual seedling or seedling-sprout) will occupy an intermediate-or-greater crown class in upland oak stands in the Ozark Highlands of Missouri. Size of advance reproduction (basal diameter) at the time of clearcutting or other major disturbance is a key determinant of future stand composition. Estimates of P are based on logistic regression equations for average sites. (B) Vertical stratification of species after 15 years illustrates the probabilities in (A). Oaks are predominant in the upper crown classes while non-oaks are largely relegated to the sub-canopy. The four species groups shown represent the most frequently occurring hardwood species within this ecological region. Oaks are predominantly black, white, scarlet and post oaks. (From authors’ data.)

from 1400 to 600 acre−1 while basal area simultaneously increased from 70 to 120 ft2 acre1 (Fig. 5.8A and B). Drought com-

bined with defoliation by gypsy moth and canker worm reduced stand basal area between ages 55 and 65, which in turn

Development of Natural Stands

1600

130 A

B Basal area (ft2 acre–1)

1400 Trees per acre

211

1200 1000 800 600

120 110 100 90 80 70 60

400 20

40

60

80

100

20

40

Stand age (years)

60

80

100

Stand age (years)

130

50 C

D

(95) Red oaks

(95)

120

(85)

100 (55) 90

Basal area (ft2 acre–1)

Basal area (ft2 acre–1)

40 110

(75) (35)

(65)

80

(55) (95) Birches

30

20

(55)

Maples

(25)

(25)

(95)

(25) 60 500

(25)

White oaks

10

70

(95) (25)

0 700

900

1100

Trees per acre

1300

1500

0

100

200

300

400

Trees per acre

Fig. 5.8. Seventy-year change in number and basal area of trees ≥0.5 inches dbh in unthinned oak–mixed hardwood stands in Connecticut. The values shown are averages of four stands of similar composition and structure. (A) Number of trees. Stands were in the stem exclusion stage of development until age 55. Between ages 65 and 75, the number of trees increased due to recruitment into canopy openings resulting from tree mortality caused by gypsy moth and canker worm defoliation in the previous decade (ages 55–65). (B) Stand basal area. Oaks comprised 37% of stand basal area at stand age 25 and 45% by age 95. The decrease in basal area between ages 55 and 65 was caused by tree mortality. (C) Trajectories over time in the relation between basal area and number of trees for all species combined. Stand age at each point is shown in parentheses. (D) Trajectories for four species groups. Canopy gaps created between ages 55 and 65 accelerated the recruitment of new trees into the overstorey between ages 65 and 75 and marked the beginning of the understorey reinitiation stage of stand development. The number of birches and maples increased after canopy gaps were formed. Basal area of red oaks declined temporarily after age 55, but then increased from age 65 to 95. By age 95, red oaks dominated that stand in terms of basal area, but comprised fewer than 46 trees per acre. White oaks declined in both number of trees and basal area. The red oaks include northern red, black, and scarlet oaks; white oaks include white and chestnut oaks; birches include yellow, black and paper birches; maples include red and sugar maples. (Adapted from Ward et al., 1999; additional data courtesy of Jeffrey S. Ward, Connecticut Agricultural Experiment Station, used with permission.)

212

Chapter 5

accelerated the establishment and ingrowth of reproduction into the overstorey. The total number of overstorey trees increased in the following decade. Defoliation by gypsy moth and elm spanworm occurred episodically after age 65, but those events did not reduce total stand basal area. Both the absolute and relative basal area of the white oak group decreased markedly between stand ages 55 and 65 after insect defoliation and drought. Although those disturbances were not sufﬁcient to initiate a new stand, they did alter species composition by creating conditions that increased numbers of maples and birches. Decreases in basal area and subsequent increases in numbers of trees between ages 55 and 75 are evident from the trajectory of change for all species combined (Fig. 5.8C). Trajectories for individual species-groups reveal a sharp increase in numbers of birches and maples between ages 65 and 75 following the decrease in oak basal area a decade earlier (Fig. 5.8D). The diameter distribution of all trees combined retained a reverse J-shape for the entire 70-year observation period (Fig. 5.9). Like the bottomland oak stand in Fig. 5.6, the shape of the diameter distribution ﬂattened as the stand aged. The number of stems in the smallest diameter class decreased by 600 per acre over 70 years; numbers of trees in the largest diameter classes increased by 10–15 per acre over the same period. At stand age 25, the diameter distribution of the oaks was similar to the distribution of all species combined. But between ages 35 and 65, the oak diameter distributions became increasingly bell-shaped (Fig. 5.9). During that time, oaks decreased from 320 to 55 per acre and remained at approximately 50 per acre until age 95. Although the largest oaks increased in size, few smaller oaks grew into the overstorey. The range of oak diameters also increased over time. By age 95, oaks were the largest trees present and occurred in all diameter classes. Numbers of oaks in the smallest diameter classes increased modestly after age 65 (understorey reinitiation stage), but

birch, maple and beech were 4–12 times more abundant than the oaks in the small diameter classes. Although numbers of birches and maples per acre continually increased over time as a proportion of the total number of trees, the oaks collectively showed the greatest relative increase in basal area. White oak declined in both numbers of trees and basal area following the disturbances occurring between ages 55 and 65. The diameter distribution of oaks present at the time of stand establishment thus formed a continually changing population in terms of numbers of trees and basal area. Although the changes in the size distribution of the oak component of the stand are consistent with expectations for unthinned oak forests based on normal stand tables (Fig. 5.4; Schnur, 1937), they differ fundamentally from the overall diameter distribution of the composite stand (Fig. 5.9). This and the preceding example of stand development (Fig. 5.6) illustrate how species composition and stand structure of oak forests change over time and how those changes can differ among oak forests with different species mixtures, site characteristics and disturbance histories. In both examples, oaks were not the predominant species at the start of the stem exclusion stage of stand development. However, by persistently capturing growing space, the largest oaks were able to maintain rapid diameter growth and increase in basal area relative to other species present. Although reverse J-shaped diameter distributions are most often associated with uneven-aged forests, they also occur in certain even-aged forests – including those beyond the stem exclusion stage of development. These stands often originate following the disturbance of stands comprised of both shade tolerant and shade intolerant species. In subsequent stages of stand development, reverse J-shaped diameter distributions may evolve through the ingrowth of shade tolerant reproduction into the overstorey. Meanwhile, the less tolerant oaks may develop a bell-shaped diameter distribution embedded within the whole-stand diameter distribution.

Development of Natural Stands

Trees per acre

Trees per acre

Trees per acre

Trees per acre

Trees per acre

Trees per acre

Trees per acre

1027 500

All trees Age 25

400

182 80

213

All oaks Age 25

60

300

40

200

20

100 0 500

Age 35

400

0 80

Age 35

60

300

40

200

20

100 0 500

Age 55

400

0 80

Age 55

60

300

40

200

20

100 0 500

Age 65

400

0 80

Age 65

60

300

40

200

20

100 0 500

Age 75

400

0 80

Age 75

60

300

40

200

20

100 0 500

Age 85

400

0 80

Age 85

60

300

40

200

20

100 0 500

Age 95

400

0 80

Age 95

60

300

40

200

20

100

0

0 0

4

8 12 16 20 24 Dbh (inches)

0

4

8 12 16 20 24 Dbh (inches)

Fig. 5.9. Seventy-year change in diameter distributions in unthinned oak–mixed hardwood stands in Connecticut. Values are based on the average of four stands of similar composition and structure. The left-hand column of graphs shows all trees and the right-hand column shows the oaks. Stands were in the stem exclusion stage of development until age 55. Basal area decreased between ages 55 and 65 due to mortality from gypsy moth and canker worm defoliation. During the next decade, the number of trees per acre increased as new trees became established after the reduction in stand density. Trees per acre are plotted over the midpoints of 2-inch dbh classes. (Adapted from Ward et al., 1999; additional data courtesy of Jeffrey S. Ward, Connecticut Agricultural Experiment Station, used with permission.)

214

Chapter 5

Although the diameter distribution changes for the entire stand are somewhat predictable and are useful for describing patterns of stand development, they explain little about the interactive relations between oaks and other species. Predicting the outcome of the competitive struggle between oaks and associated species is especially important when the co-occurring species differ greatly in growth rate and silvical characteristics. Determining the silvicultural prescription that assures the desired representation of oaks in the future stand is a common silvicultural problem in tending young, even-aged stands comprised of oaks mixed with other species. An appropriate silvicultural action requires an understanding of how species interact within shared environments. Large differences in rates of height growth of two desirable, long-lived species growing on the same site can create difﬁculties in maintaining both species. In young stands, species’ differences in height growth may not always be apparent from the current stand structure because those differences may not yet be expressed. And in older stands, the disturbance history that produced the height differences among species may not be evident. Comparing the site index of one species with another is one way to assess the ecological and silvicultural signiﬁcance of species’ height growth differences (Chapter 4). Differences in site index as small as 5 ft may ultimately result in the suppression of the slower-growing species, assuming both species originate about the same time. Such differences, in turn, may limit silvicultural options. Large growth-rate differences between an oak and a co-occurring species may be silviculturally problematic if the oak is the slower-growing species and both species are to be grown to sawlog size, or if the primary objective is to sustain long-term acorn production. However, site index comparisons, by themselves, do not always predict how co-occurring species will interact. Virtually all oak stands eventually become vertically stratiﬁed by species, and even for the same initial species mix, patterns of height stratiﬁcation may differ. The development of various species mixtures

provides examples of how vertical stratiﬁcation can differ among ecosystems. For example, the early pattern of height growth of co-occurring northern red oak, black birch and red maple in young even-aged stands in New England was similar. Not until the end of the third decade did northern red oak emerge as the dominant species (Oliver, 1978) (Fig. 5.10A). A different pattern of stratiﬁcation occurs in the mixed forest type that is transitional between the Allegheny Plateau and the Northern Hardwood Region to the north. There, forests are commonly, but only temporarily, dominated by northern red oak. The oaks and other intolerant and mid-tolerant species retain their bell-shaped distribution as they develop (Stout, 1991). In contrast, sugar maple and other tolerant species form reverse J-shaped diameter distributions as their reproduction accumulates in the understorey, which ultimately leads to the successional displacement of the oaks. In co-occurring cherrybark oak and sweetgum stands in minor river bottoms in Mississippi, it takes about 40 years for the oak to dominate the sweetgum (Clatterbuck and Hodges, 1988). The oak thereafter grows faster when directly competing with sweetgum than when not competing with sweetgum (Fig. 5.10B). This effect may be partially related to differences in the crown shape of the two species. Whereas oak crowns are wider at the top than the bottom (excurrent in shape), sweetgum crowns are the opposite (decurrent in shape). Because the height growth of the two species continues to diverge with age, there is minimal interference between sweetgum crowns and oak crowns. The persistence of the sweetgum also promotes long, clear boles on the oaks. Cherrybark oaks growing in competition with sycamore do not fare as well. The extremely rapid height growth and somewhat excurrent crown of the sycamores quickly suppress the oaks unless 20 ft or more separates the two species (Fig. 5.10C). Similar interspeciﬁc relations occur when scarlet and white oaks grow in competition with yellow-poplar (Fig. 5.10D). Although these examples are generally consistent with

Development of Natural Stands

A

70 Northern red oak

50

120 Black birch

Red oak

40 Red maple

30 20

Close spacing

B

Black birch Red maple

Tree height (ft)

Tree height (ft)

60

Sweetgum (close spacing)

Wide spacing

Cherrybark oak

Oak (close spacing) 100

Cherrybark oak Sweetgum

80

Nearest competitor

Oak (wide spacing) 60

10

215

>18ft

0 0

10 20 30 40 50 60

40 10 20 30 40 50 60 70 Stand age (years)

Stand age (years)

Dominant

60

Codominant

20 0

Yellow-poplar stand 1 100 stand 2

Sycamore Cherrybark oak

Intermediate

Suppressed

0 4 8 12 16 20 24 Stand age (years)

Tree height (ft)

Tree height (ft)

Sycamore Cherrybark 80 oak

40

120

C Crown class

100

Stand 1

80 60 40 20

0 10 20 30 40 Distance (ft)

D Stand 2 Yellowpoplar

Scarlet oak Scarlet oak stand 1

White oak

White oak stand 1

0 1900 1940 1980 1920 1960 Year of origin

Fig. 5.10. Height growth and vertical stratification of oaks in mixed-species stands. (A) Even-aged stands of northern red oak, red maple and black birch in New England. Red oaks emerge as dominants 20–30 years after disturbance. (Adapted from Oliver, 1978.) (B) Even-aged stands of cherrybark oak and sweetgum in minor river bottoms in central Mississippi. Close spacing between sweetgum and oak accelerates the height growth of oaks; wide spacing results in slower height growth and shorter clear-bole length of oaks. (Adapted from Clatterbuck and Hodges, 1988.) (C) Plantation-grown cherrybark oak and American sycamore in a minor river bottom in Arkansas. Close spacing between oaks and sycamores results in suppression or reduced height growth of oaks. (Adapted from Oliver et al., 1990.) (D) Representative trees in two natural stands of mixed oak and yellow-poplar of unspecified age structure in the North Carolina Piedmont. In this region, yellow-poplar outgrows oaks where yellow-poplar site index exceeds 65 ft. (Adapted from O’Hara, 1986, by permission of Society of American Foresters, Bethesda, Maryland. Not for further reproduction.)

site index comparisons between species, site index alone does not explain the magnitude of such species-speciﬁc interactions.

The complex stage In the absence of timber harvesting or other exogenous disturbances that eliminate the overstorey, even-aged stands progress toward the complex stage (Oliver, 1997)

(Figs 5.2 and 5.3D). This stage evolves as the consequence of the natural mortality of large overstorey trees that create canopy gaps that occur at irregular times and locations within a stand. Because these gaps are often large, crown expansion of trees adjacent to a gap is often insufﬁcient to ﬁll the gap. This lag in crown closure allows sub-canopy trees and established reproduction to ﬁll gaps. As new canopy gaps occur, they are ﬁlled by new age classes of trees.

216

Chapter 5

As this gap-ﬁlling process continues in the absence of a stand-initiating disturbance, an uneven-aged stand eventually evolves. The complex stage of stand development includes old-growth forests, which was the original term applied to this stage (Oliver and Larson, 1996). However, there is a basis for distinguishing between the two. All old-growth forests are complex, but not all complex forests are old-growth. Deﬁnitions of old-growth oak forests are usually based on overstorey age, stand disturbance history and structural characteristics such as the presence of old trees, snags and down wood (see Chapter 9; also Meyer, 1986; Parker, 1989; Martin, 1992; Batista and Platt, 1997; Greenberg et al., 1997; Murphy and Nowacki, 1997; Tyrell et al., 1998). The various deﬁnitions of old-growth all assume that human inﬂuences in such forests have been minimal. However, older second-growth forests (managed and unmanaged) may have complex structures without meeting the strict deﬁnition of old growth.

Development of Uneven-aged Stands As an even-aged stand advances towards the complex stage of development, it gradually evolves into a multi-aged population as a result of stand maturation and associated gap formation and ﬁlling. Canopy gaps become more numerous until the stand forms a mosaic of old trees and gaps ﬁlled with younger trees of various ages and species. As trees reﬁll the gaps, standwide diameter distributions gradually change. In the absence of a stand-initiating disturbance, the overall diameter frequency distribution may change from bell-shaped to irregularly shaped or reverse J-shaped. Regardless of the shape of the diameter distribution, the normal evolution of stand structure from even- to uneven-aged eventually produces an uneven-aged collection of dispersed, even-aged groups of trees, each occupying a relatively small area. Over time, the various tree age classes become visually indiscernible.

The resulting uneven-aged stand structure can be generalized from the diameter distributions of a time series of even-aged stands consisting of 10-year age classes with each class comprising an equal area (Fig. 5.4). If numbers of trees in each age class are summed across all diameter classes, numbers of trees in successively larger diameter classes decrease exponentially (Fig. 5.4 inset). This type of diameter distribution therefore could result from either a collection of even-aged stands or from a single uneven-aged stand. However, in the latter, the even-aged spatial units are so small and intermingled that they are largely indistinguishable. During the complex stage of development, differences in shade tolerances and growth rates among the oaks themselves may produce gradual changes in the proportions of oaks among the crown classes and thus corresponding shifts in species composition. For example, through the stem exclusion stage, the relatively shadetolerant and slow growing white oak tends to lag behind co-occurring red oaks in capturing growing space. However, during the understorey reinitiation and complex stages of development, the white oaks that persist in the inferior crown classes gradually ascend to canopy dominance as canopy gaps are formed (Shiﬂey et al., 1995; Spetich, 1995). Although old-growth forests comprised of oaks mixed with other species often form reverse J-shaped diameter distributions, the oaks themselves may not conform to that distribution. Structural and compositional changes during the complex stage largely depend on differences in the rates at which co-occurring species of tree reproduction are recruited into the overstorey and how persistent they are there. Where non-oaks ﬁll most of the canopy gaps created during the complex stage, those species will eventually predominate among the smaller dbh classes. Depending on site quality and other factors, they may or may not ascend to dominance, inhibit oak regeneration or successionally displace the oaks. In the absence of disturbances that reinitiate the establishment and

Development of Natural Stands

217

eter distributions in this ecosystem and their silvicultural maintenance largely depend on the species composition and density of stands (Chapter 8).

development of oak reproduction, a bellshaped diameter distribution of oaks embedded within an overall reverse Jshaped distribution nevertheless is a harbinger of the impending successional displacement of the oaks (Fig. 5.11). In the Ozark Highlands, crown stratiﬁcation among the oaks persists into the complex stage (Shiﬂey et al., 1995). Large numbers of oaks in the smaller diameter classes, even in old-growth stands, reﬂect the oak’s permanence in these forests (Fig. 5.12A and B). In mature, relatively undisturbed second-growth forests, diameter distributions approach a reverse J-shape (Fig. 5.12C and D). For all species combined, the reverse J-shape is even more pronounced because of the high density of permanent subcanopy species such as ﬂowering dogwood and blackgum (Fig. 5.12 inset graphs). Except during the stand initiation stage, the non-oaks are largely relegated to the sub-canopy (Dey et al., 1996). The speciﬁc characteristics of diam-

Disturbance–Recovery Cycles When forest disturbance is limited to gapscale events, stand development follows the sequence illustrated in Fig. 5.2. However, stand-initiating events that eliminate all or most of the overstorey can occur during any stage of stand development. These events return stands to the stand initiation stage of development. In contrast, incomplete stand-scale disturbances may eliminate only a portion of the overstorey and leave signiﬁcant numbers of trees standing. Although incomplete stand-scale disturbances change the stage of stand development, they do not return the stand to the stand initiation stage. Rather, they create a mixed stage of stand

80 3 Trees per acre (oaks)

Trees per acre (all species)

70 60 50 40

2

1

0 30

4

8 12 16 20 24 28 32 36 Dbh class (inches)

20 10 0 4

8

12

16

20

24

28

32

36

Dbh class (inches) Fig. 5.11. The diameter distribution of trees in Spitler Woods, an old-growth forest in central Illinois. The bell-shaped (normal) diameter distribution of the oaks (inset) is obscured by the reverse J-shaped distribution of all tree species. This mesic forest is dominated by sugar maple, oaks and American basswood. The oaks (white, northern red, black, chinkapin, bur and shingle oaks) comprise 33% of stand basal area but only 6% of trees. Other important species include elms, hickories, hackberry, Ohio buckeye, black walnut and white ash. (Author’s data.)

Chapter 5

50 Red oaks

30

Oak stocking = 69%

White oaks

20

B

400

Stocking = 101%

300 200

2

6 10 14 18 22 26 Dbh class (inches)

10

20

400 Stocking = 103%

300 200 100 0 2

6 10 14 18 22 26 Dbh class (inches)

10

0 2

6

10

14

18

22

26

0 2

6

10

14

18

22

26

50 C

40

Trees per acre (total)

50 Trees per acre (oaks)

Oak stocking = 71%

30

100 0

40

Trees per acre (total)

Trees per acre (oaks)

40

Trees per acre (total)

50 A

D

400

Oak stocking = 64%

30

300

Stocking = 103%

200

400

Oak stocking = 71%

30

100

20

40

Trees per acre (total)

218

300

Stocking = 87%

200 100

0 2 6 10 14 18 22 26 Dbh class (inches)

10

20

0 2

6 10 14 18 22 26 Dbh class (inches)

10

0

0 2

6

10

14

18

22

26

Dbh class (inches)

2

6

10

14

18

22

26

Dbh class (inches)

Fig. 5.12. Diameter distributions of oaks and all trees (inset graphs) in four relatively undisturbed oak forests of the Ozark Highlands of Missouri. (A) A 120-acre old-growth stand with trees up to 200–250 years old; human disturbance has been limited to periodic fires occurring more than 40 years ago. (B) A 330-acre old-growth stand with trees up to 140 years old (a few exceed 200 years). There is little evidence of disturbance during the last 40 years. (C) A second-growth 4000-acre forest dominated by 70- to 90-year-old oaks with some shortleaf pine. Stands were disturbed by fire, grazing and logging 45 years prior. (D) A second-growth 3000-acre forest disturbed by fire, grazing and logging 45 years prior. Intensity/recentness of disturbance (increasing from A to D) is reflected in the increasing number of oaks in the smaller diameter classes. All four forests form mosaics of xeric to xero-mesic associations dominated by various mixtures of oaks, hickories and shortleaf pine growing on well- to excessively drained cherty soils. Red oaks include black, scarlet, southern red and northern red oaks; white oaks include white, post and chinkapin oaks. Important subordinate species include flowering dogwood, blackgum, sassafras, elms and red maple. (From Shifley et al., 1995.)

development. Mixed-stage stands resulting from natural events often form mosaics of younger trees developing in large canopy openings interspersed with patches of older trees. They often form irregularly spaced tree populations of variable size and age structure. Mixed-stage stands also can result from various types of timber harvesting.

Stands in the mixed stage of development are distinguishable from stands in other stages by: (i) the spatial scale of disturbance, which is greater than gap scale, and (ii) stand density, which is below average maximum density (see Chapter 6). Relatively young trees also may predominate in stands during the mixed stage. Low density, highly disturbed stands of all descriptions therefore fall into

Development of Natural Stands

the mixed stage of development. Examples of stands in the mixed stage include oak savannas (Chapter 9), stands resulting from indiscriminate timber harvests (e.g. ‘high-grading’), as well as some designed silvicultural practices. The latter include heavily thinned stands, low-density shelterwoods and stands managed by group selection or single tree selection at low stand densities (Chapters 7 and 8). Oak forests in the mixed stage of development are ubiquitous because the events that create them are so common. The complete spectrum of forest developmental stages thus includes the mixed stage plus the four stages previously deﬁned. Collectively they represent points in a potentially endless series of disturbance–recovery cycles initiated by standscale and gap-scale disturbances. These cycles follow speciﬁc sequences determined by the developmental stage of the stand at the time of disturbance and the type and spatial scale of disturbance (Fig. 5.13). Disturbance and recovery cycles provide a conceptual framework for ecological process and silvicultural practice. Silviculture involves the planned application of controlled disturbances to create and maintain (at least temporarily) speciﬁc species compositions, age distributions and size distributions. An ecological understanding of disturbance–recovery cycles is central to achieving silvicultural objectives. However, silvicultural control of stand composition and structure is often complicated by unpredictable natural disturbance events that lie beyond the direct control of the silviculturist. Even when forests are intensively managed, unwelcome disturbances are a part of silvicultural reality. Frequent stand-initiating disturbances can maintain a stand in the stand initiation stage indeﬁnitely. Such disturbance regimes were common and extensive in many regions of North America before European settlement. For example, in the prairie–forest border region of southwestern Wisconsin, plant communities described as ‘oak scrub’ failed to develop beyond the sapling stage because of frequent wildﬁres that recurred for centuries (Marks, 1942; Grimm, 1984; Chumbley

219

et al., 1990). As long as these ﬁres persisted, the stands of oak scrub persisted. The ﬁres arrested succession, and thus maintained stands in the stand initiation stage of development. But by the early 20th century, wildﬁres were controlled and the oak scrubs quickly developed into closed canopy forests (Marks, 1942; Curtis, 1959). An incomplete stand-scale disturbance will transform most stands into the mixed stage of development. Exceptions are young, even-aged stands that are in the stand initiation stage of development. They remain in the stand initiation stage of development following incomplete standscale disturbances because the existing trees and the new trees recruited following the disturbance are within the same age class. If a stand is already in the mixed stage of development, frequent incomplete stand-scale disturbances (such as the ﬁres that historically maintained oak savannas) may keep it in that state indeﬁnitely. Not all stands progress through the stages of development according to the illustrated sequence (Fig. 5.13). For example, oak chaparral and other shrub-like oak communities may not conform to the ﬁve-stage model presented. Disturbances are of many types and intensities, and follow many temporal–spatial patterns. Even distinguishing between gap-scale and stand-scale disturbances is artiﬁcial because real-world disturbance scales are continuous. Disturbances also may affect only some parts of a stand, or two types of disturbances (e.g. windthrow and disease-induced mortality) may affect different parts of a stand simultaneously. Some forests are also more resilient than others, i.e. they have a greater capacity to return to their previous state after disturbance. Consequently, oak forests differ in their disturbance recovery rates and in related changes in stand structure and species composition. The 35-year history of an oak–hickory stand in the Ozark Highlands of Missouri provides an example of disturbance and recovery in that ecoregion. The stand is part of a 160,000-acre forest that has been managed by the single tree selection system since exploitative timber harvesting reduced aver-

220

Chapter 5

Stand initiation stage

Stem exclusion stage

Mixed stage

Understorey reinitiation stage

Complex stage

Key: Type of disturbance

Stand-scale: stand-initiating

Stand-scale: incomplete

Gap-scale

Fig. 5.13. Types of forest disturbances and their relation to stages of stand development and disturbance–recovery cycles. Stands in all developmental stages are subject to gap-scale disturbances originating from the natural mortality of individual trees (dotted arrows). As average tree size (and thus gap size) increases with increasing stand age, stands progress towards the uneven-aged state (complex stage) via the gap-wise replacement of main canopy trees by sub-canopy trees and reproduction. However, stand-scale disturbances can occur at any stage of stand development. When stand-initiating disturbances eliminate all or most of the overstorey, stands abruptly return to the stand initiation stage (solid arrows). Disturbances that eliminate only a fraction of the overstorey, but leave a significant number of trees standing, are termed incomplete stand-scale disturbances. These disturbances produce a mixed stage of stand development comprised of both new and older trees that develop into multi-tiered stands of irregular age structure (dashed arrows). The mixed stage may advance to the complex stage in the absence of further stand-initiating or incomplete stand-scale disturbances, return to the stand initiation stage after a stand-scale disturbance, or remain in the mixed stage as a result of further incomplete stand-scale disturbances.

Development of Natural Stands

age forest-wide stand densities to low levels in the early 1950s (Loewenstein et al., 1995). The recovery of this forest is illustrated by data from four 0.2-acre plots in a single stand that was inventoried every 5 years from 1957 to 1992. During that time, numbers of trees ≥5 inches dbh doubled and stand basal area increased from 33 to 73 ft2 acre1 (Fig. 5.14A and B). This increase occurred even though 17 ft2 of basal area was harvested between 1972 and 1977. During this 35-year period, oaks maintained a nearly constant proportion (48%) of total basal area. Changes in stand structure and species composition are described by changes in basal area and numbers of trees per acre for all species combined and for individual species or groups (Fig. 5.14C and D). Although the number of trees and basal area were reduced by the 1972–1977 timber harvest, the stand quickly resumed its preharvest trajectory of change (Fig. 5.14C). However, within the overall pattern of stand development, there was great variation among species in recovery responses. Whereas numbers of trees and basal areas of all species increased between 1957 and 1972, the 1972–1977 timber harvest reduced the number of large black and scarlet oaks proportionately more than other species. Black and scarlet oak basal areas nevertheless remained relatively constant from 1977 to 1992 (Fig. 5.14D). In contrast, the white oaks and shortleaf pine increased in basal area relative to black and scarlet oaks. The continual increase in the basal area of white oak on the study plots paralleled forestwide changes that occurred during the same period (Loewenstein, 1996; Wang, 1997). Patterns of recovery from disturbance were also evident from changes in diameter distributions. Even though the distributions of all species combined retained a reverse Jshape during the 35-year period, the distributions of the oaks did not (Fig. 5.15). From 1957 to 1972, the oak diameter distributions formed two or more peaks. Oaks in the smallest (6 inch) dbh class nevertheless were more numerous than in any other class throughout the observed period. For the oaks, reverse J-shaped distributions became increasingly prominent with time.

221

Over time, numbers of white oaks increased in the smaller diameter classes and by 1992 white oak dominated the 6–8 inch dbh classes (Fig. 5.15). This change reﬂects white oak’s shade tolerance and its capacity in this ecosystem to sustain rates of ingrowth into successively larger diameter classes in numbers sufﬁcient to maintain a reverse J-shaped diameter distribution. The reduction in stand density to 47 ft2 of basal area per acre between 1972 and 1977 was an important factor in sustaining white oak ingrowth and thus in perpetuating an uneven-aged stand structure. The less shade tolerant black and scarlet oaks decreased in abundance in the smaller diameter classes although they retained their earlier representation in the larger diameter classes. Stand density, diameter distributions and species composition in this uneven-aged stand thus changed continually during the 35year period. The forest’s resilience is reﬂected in its rapid rate of recovery from exploitative timber harvests in the 1950s and a silviculturally designed harvest in the 1970s. Since then, oaks have maintained their dominance, stand basal area has continued to increase and a reverse J-shaped diameter distribution consistent with sustaining an uneven-aged stand structure has developed. Comprehensive analyses of this forest have similarly indicated that continued application of the single tree or group selection method is likely to sustain the oak’s dominance (Loewenstein, 1996; Wang, 1997). However, successful application of the method in this and ecologically similar forests will require reducing stocking to about 50 ft2 of basal area per acre every 15–20 years in order to sustain adequate rates of oak recruitment into the overstorey (Larsen et al., 1999). In effect, this requires using silvicultural control of stand structure and density to suspend stands in the mixed stage of development. From this it should not be inferred that oak forests are generally amenable to uneven-aged silviculture. On the contrary, the intrinsic oak regeneration characteristics of an ecosystem largely determine its capacity to sustain uneven-aged oak populations and thus its suitability to uneven-aged management (see Chapters 3 and 8).

222

Chapter 5

80

180 A

B Basal area (ft2 acre–1)

Trees per acre

160 140 120 100

70

60

50

40

80 1955

1965

1975

1985

30 1955

1995

1965

1975

Year

1995

Year

40

80 C

D

(92) Basal area (ft2 acre–1)

Basal area (ft2 acre–1)

1985

70 (72) (87)

60

(82)

(67)

50

(77)

(62) 40

Shortleaf pine (72)

120

140

Trees per acre

160

180

(57)

(57) (72)

10

White oaks

Hickories

0 100

(92)

Black and scarlet oaks (92)

20

(57)

30

(72)

30

(57) 80

(92)

0

20

40

60

80

Trees per acre

Fig. 5.14. Change in number and basal area of trees ≥5 inches dbh during a 35-year period in an oak stand managed by the single-tree selection method in the Ozark Highlands of Missouri. The stand is dominated by black, scarlet and white oaks; black oak site index ranges from approximately 55 to 65 ft at an index age of 50 years. The stand density was greatly reduced by an exploitative harvest prior to 1957 which placed the stand in the mixed stage of development (Fig. 5.13). This sequence of stand development illustrates recovery from that disturbance and response to a designed silvicultural disturbance between 1972 and 1977. Number of trees (A) and basal area (B) increased during all intervals except 1972–1977 when a timber harvest removed some trees 15 inches dbh and larger. (C) Trajectories over time in the relation between basal area and number of trees for all species combined. The plotted line traces changes from 1957 to 1992 (year shown in parentheses). The stand was in the mixed stage of development during the 35-year period, and low stand basal area facilitated recruitment of trees into the overstorey. (D) Trajectories for the four species groups. The timber harvest reduced the number and basal areas of black/scarlet oaks and shortleaf pine. Although drought-induced mortality after 1977 further reduced the number and basal area of black and scarlet oaks, by 1992 values were similar to those in 1962. Both the number and basal area of white oaks (white and post oaks combined) continually increased over time. (Author’s analysis; data courtesy of Pioneer Forest, Salem, Missouri.)

Development of Natural Stands

Trees per acre

Trees per acre

Trees per acre

Trees per acre

Trees per acre

Trees per acre

Trees per acre

Trees per acre

All trees

223

Oaks 1957

1957

60

60

40

40

20

20

0

White/post oak Black/scarlet oak

0 1962

1962

60

60

40

40

20

20

0

0 1967

1967

60

60

40

40

20

20

0

0 1972

1972

60

60

40

40

20

20

0

0 1977

1977

60

60

40

40

20

20

0

0 1982

1982

60

60

40

40

20

20

0

0 1987

1987

60

60

40

40

20

20

0

0 1992

1992

60

60

40

40

20

20

0

0 6

10 14 18 22 Dbh (inches)

26

6

10

14 18 22 Dbh (inches)

26

Fig. 5.15. Thirty-five-year change in diameter distributions in an uneven-aged oak forest in the Ozark Highlands of Missouri managed by the single-tree selection method. The diameter distributions for all species (left-hand column of graphs) maintained a reverse J-shape throughout the period. However, the number of trees in the smaller diameter classes continually increased during the period. By 1992, white oaks dominated the 6- to 10-inch dbh classes. The number of black and scarlet oaks in the smaller diameter classes increased between 1957 and 1972 but thereafter declined. (Data courtesy of Pioneer Forest, Salem, Missouri.)

224

Chapter 5

References Abrams, M.D. and Downs, J.A. (1990) Successional replacement of old-growth white oak by mixed mesophytic hardwoods in southwestern Pennsylvania. Canadian Journal of Forest Research 20, 1864–1870. Abrams, M.D. and Nowacki, G.J. (1992) Historical variation in ﬁre, oak recruitment, and post-logging accelerated succession in central Pennsylvania. Bulletin of the Torrey Botanical Club 119, 19–28. Abrams, M.D. and Scott, M.L. (1989) Disturbance-mediated accelerated succession in two Michigan forest types. Forest Science 35, 42–49. Batista, W.B. and Platt, W.J. (1997) An old-growth deﬁnition for southern mixed hardwood forests. USDA Forest Service General Technical Report SRS SRS-9. Beck, D.E. and Hooper, R.M. (1986) Development of a southern Appalachian hardwood stand after clearcutting. Southern Journal of Applied Forestry 10, 168–172. Bormann, F.H. and Likens, G.E. (1979) Pattern and Process in a Forested Ecosystem. Springer-Verlag, New York. Chumbley, C.A., Baker, R.G. and Bettis, E.A. III. (1990) Midwestern holocene paleoenvironments revealed by ﬂoodplain deposits in northeastern Iowa. Science 249, 272–274. Clatterbuck, W.K. and Hodges, J.D. (1988) Development of cherrybark oak and sweet gum in mixed, even-aged bottomland stands in central Mississippi, USA. Canadian Journal of Forest Research 18, 12–18. Clements, F.E. (1916) Plant succession: An analysis of the development of vegetation. Carnegie Institute Washington Publication 242. Curtis, J.T. (1959) The Vegetation of Wisconsin. University of Wisconsin Press, Madison. Dey, D.C. (1991) A comprehensive Ozark regenerator. PhD dissertation, University of Missouri, Columbia. Dey, D.C., Ter-Mikaelian, M., Johnson, P.S. and Shiﬂey, S.R. (1996) Users guide to ACORn: a comprehensive Ozark regeneration simulator. USDA Forest Service General Technical Report NC NC180. Egler, F.E. (1954) Vegetation science concepts I. Initial ﬂoristic composition. A factor in old-ﬁeld vegetation management. Vegetatio 4, 412–417. Gevorkiantz, S.R. and Scholz, H.F. (1948) Timber yields and possible returns from the mixed-oak farmwoods of southwestern Wisconsin. USDA Forest Service Lake States Forest Experiment Station Publication 521. Gingrich, S.F. (1967) Measuring and evaluating stocking and stand density in upland hardwood forests in the Central States. Forest Science 13, 38–53. Gingrich, S.F. (1971) Management of young and intermediate stands of upland hardwoods. USDA Forest Service Research Paper NE NE-195. Greenberg, C.H., McLeod, D.E. and Loftis, D.L. (1997) An old-growth deﬁnition for western and mixed mesophytic forests. USDA Forest Service General Technical Report SRS SRS-16. Grimm, E.C. (1984) Fire and other factors controlling the Big Woods vegetation of Minnesota in the mid-nineteenth century. Ecological Monographs 54, 291–311. Harper, J.L. (1977) Population Biology of Plants. Academic Press, London. Helms, J.A. (ed.) (1998) The Dictionary of Forestry. Society of American Foresters, Bethesda, Maryland. Hibbs, D.E. (1982) Gap dynamics in a hemlock-hardwood forest. Canadian Journal of Forest Research 12, 522–527. Hibbs, D.E. (1983) Forty years of forest succession in central New England. Ecology 64, 1394–1401. Hilt, D.E. (1985) Where has all my yellow-poplar gone? Northern Journal of Applied Forestry 2, 67–69. Jenkins, M.A. and Parker, G.R. (1997) Changes in down dead wood volume across a chronosequence of silvicultural openings in southern Indiana forests. USDA Forest Service General Technical Report NC NC-188, pp. 162–169. Johnson, P.S. and Sander, I.L. (1987) Quantifying regeneration potentials of Quercus forests in the Missouri Ozarks. USDA Forest Service General Technical Report NC NC-120, Vol. 1, pp. 377–385. Johnson, R.L. and Deen, R.T. (1993) Prediction of oak regeneration in bottomland forests. USDA Forest Service General Technical Report SE SE-84, pp. 146–155.

Development of Natural Stands

225

Johnson, R.L. and Krinard, R.M. (1983) Regeneration in small and large sawtimber sweetgum–red oak stands following selection and seed tree harvest: 23-year results. Southern Journal of Applied Forestry 7, 176–184. Johnson, R.L. and Krinard, R.M. (1988) Growth and development of two sweetgum–red oak stands from origin through 29 years. Southern Journal of Applied Forestry 12, 73–78. Larsen, D.R., Loewenstein, E.F. and Johnson, P.S. (1999) Sustaining recruitment of oak reproduction in uneven-aged stands in the Ozark Highlands. USDA Forest Service General Technical Report NC NC-203. Loewenstein, E.F. (1996) An analysis of the size- and age-structure of a managed uneven-aged oak– forest. PhD dissertation, University of Missouri, Columbia. Loewenstein, E.F., Garrett, H.E., Johnson, P.S. and Dwyer, J.P. (1995) Changes in a Missouri Ozark oak hickory forest during 40 years of uneven-aged management. USDA Forest Service General Technical Report NE NE-197, pp. 159–164. Loftis, D.L. (1983) Regenerating red oak on productive sites in the southern Appalachians: a research approach. USDA Forest Service General Technical Report SE SE-24, pp. 144–150. Loftis, D.L. (1990) Predicting post-harvest performance of advance red oak reproduction in the southern Appalachians. Forest Science 36, 908–916. Marks, J.B. (1942) Land use and plant succession in Coon Valley, Wisconsin. Ecological Monographs 12, 114–133. Martin, W.H. (1992) Characteristics of old-growth mixed mesophytic forests. Natural Areas Journal 12, 127–135. Meyer, J. (1986) Management of old growth forests in Missouri. Missouri Department of Conservation Habitat Management Series 3. Murphy, P.A. and Nowacki, G.J. (1997) An old-growth deﬁnition for xeric pine and pine–oak woodlands. USDA Forest Service General Technical Report SRS SRS-7. O’Hara, K.L. (1986) Developmental patterns of residual oaks and oak and yellow-poplar regeneration after release in upland hardwood stands. Southern Journal of Applied Forestry 10, 244–248. Oliver, C.D. (1978) The development of northern red oak in mixed stands in Central New England. Yale University School of Forestry and Environmental Science Bulletin 91. Oliver, C.D. (1980) Even-aged development of mixed-species stands. Journal of Forestry 78, 201–203. Oliver, C.D. (1981) Forest development in North America following major disturbances. Forest Ecology and Management 3, 153–168. Oliver, C.D. (1997) Hardwood forest management in the United States: Alternatives for the future. Proceedings of the Annual Hardwood Symposium. National Hardwood Lumber Association, Memphis, Tennessee, 25, pp. 45–58. Oliver, C.D. and Larson, B.C. (1996) Forest Stand Dynamics. John Wiley & Sons, New York. Oliver, C.D. and Stephens, E.P. (1977) Reconstruction of a mixed species forest in central New England. Ecology 58, 562–572. Oliver, C.D., Clatterbuck, W.K. and Burkhardt, E.C. (1990) Spacing and stratiﬁcation patterns of cherrybark oak and American sycamore in mixed, even-aged stands in the southeastern United States. Forest Ecology and Management 31, 67–79. Parker, G.R. (1989) Old-growth forests of the central hardwood region. Natural Areas Journal 9, 5–11. Reice, S.R. (1994) Nonequilibrium determinants of biological community structure. American Scientist 82, 424–435. Rogers, R. (1983) Spatial pattern and growth in a Missouri oak–hickory stand. PhD dissertation, University of Missouri, Columbia. Runkle, J.R. (1985) Disturbance regimes in temperate forests. In: Pickett, S.T.A. and White, P.S. (eds) The Ecology of Natural Disturbance and Patch Dynamics. Academic Press, San Diego, California, pp. 17–33. Schnur, G.L. (1937) Yield, stand, and volume tables for even-aged upland oak forests. USDA Technical Bulletin 560. Shiﬂey, S.R., Roovers, L.M. and Brookshire, B.L. (1995) Structural and compositional differences between old-growth and mature second-growth forests in the Missouri Ozarks. USDA Forest Service General Technical Report NE NE-197, pp. 23–36. Smith, D.M. (1986) The Practice of Silviculture, 8th edn. John Wiley & Sons, New York. Sousa, W.P. (1984) The role of disturbance in natural communities. Annual Review of Ecology and Systematics 15, 353–391.

226

Chapter 5

Spetich, M.A. (1995) Characteristics and spatial pattern of old-growth forests in the Midwest. PhD dissertation, Purdue University, West Lafayette, Indiana. Spetich, M.A., Shiﬂey, S.R. and Parker, G.R. (1999) Regional distribution and dynamics of coarse woody debris in Midwestern old-growth forests. Forest Science 45, 302–313. Stout, S.L. (1991) Stand density, stand structure, and species composition in transition oak stands of northwestern Pennsylvania. USDA Forest Service General Technical Report NE NE-148, pp. 194–206. Trimble, G.R., Jr and Tryon, E.H. (1966) Crown encroachment into openings cut in Appalachian hardwood stands. Journal of Forestry 64, 104–108. Tyrell, L.E., Nowacki, G.J., Crow, T.R., Buckley, D.S., Nauertz, E.A., Niese, J.N., Rollinger, J.L. and Zasada, J.C. (1998) Information about old growth for selected forest type groups in the eastern United States. USDA Forest Service General Technical Report NC NC-197. Wang, Z. (1997) Stability and predictability of diameter distributions in a managed uneven-aged oak forest. PhD dissertation, University of Missouri, Columbia. Ward, J.S. and Stephens, G.R. (1994) Crown class transition rates of maturing northern red oak (Quercus rubra L.). Forest Science 40, 221–237. Ward, J.S., Anagnostakis, S.L. and Ferrandino, F.J. (1999) Stand dynamics in Connecticut hardwood forests, the old series plots (1927–1997). Connecticut Agricultural Experiment Station Bulletin 959. Waring, R.H. and Schlesinger, W.H. (1985) Forest Ecosystems: Concepts and Management. Academic Press, Orlando, Florida. Weigel, D.R. and Johnson, P.S. (1999) Planting red oak under oak/yellow-poplar shelterwoods: A provisional prescription. USDA Forest Service General Technical Report NC NC-210. White, P.S. and Pickett, S.T.A. (1985) Natural disturbance and patch dynamics: An introduction. In: Pickett, S.T.A. and White, P.S. (eds) The Ecology of Natural Disturbance and Patch Dynamics. Academic Press, San Diego, California, pp. 3–13.

6 Self-thinning and Stand Density

Introduction Self-thinning is the natural process whereby numbers of trees per unit area decrease as average tree size increases over time. It is a process intrinsic not only to oak forests but to all forest and plant communities whose composition and structure are inﬂuenced by competition for growing space. Whereas self-thinning is a process, the term stand density refers to various expressions of the absolute or relative amounts of an attribute of tree populations (e.g. numbers of trees or stand basal area) per unit land area. As might be expected, the two concepts are closely connected. Together, they rank among the most important concepts in forest ecology and silviculture.

Self-thinning The principle of self-thinning is most easily described by the temporal changes that occur in the numbers of trees in undisturbed even-aged stands. However, selfthinning also occurs in uneven-aged stands. According to this principle, the ﬁnite growing space of a stand is occupied by progressively fewer trees as average tree size increases with stand age. Trees at a competitive disadvantage die from crowding and suppression as stands approach a limiting number of trees of a given average size that can coexist within an area. As stands reach the stem exclusion stage of development (Chapter 5), tree crowns

expand to ﬁll the available growing space. Crown expansion continues until an upper limit of tree crowding is reached. Thereafter, stands follow a relatively predictable course of density-dependent tree mortality as numbers of trees per unit area decrease with increasing average tree size. It is generally assumed that the combined effects of crown expansion and tree mortality are compensatory so that canopy closure is always maintained except in the presence of ‘irregular’ mortality. The latter may be caused by such factors as air pollution, high winds, ﬂooding, epidemic insect and disease outbreaks, and other factors.

Reineke’s model Reineke’s model for deﬁning average maximum stand density expresses the negative relation between number of trees per unit area and average stand diameter in undisturbed, even-aged stands (Reineke, 1933). Plotting the logarithm of number of trees over the logarithm of mean stand diameter produces a straight line. The relation is given by: log(N) = a + b[log(D)],

[6.1]

where N is number of trees per unit area, D is the diameter (dbh) of the average tree, and a and b are constants for a given species or group of species, where the constant b deﬁnes the slope of the line. The non-linear analogue of Equation 6.1 is given by: N = aDb

[6.2] 227

228

Chapter 6

The relation has often been used to describe the average maximum limits of stand density and, by extension, to provide a relative measure or index of stand density (Reineke, 1933). Similar models of self-thinning based on the relation between numbers of trees and tree height are used in European forestry. Such models can be empirically derived by regression analysis and other statistical methods (Weller, 1987b) using data from temporary or permanent ﬁeld plots from undisturbed stands encompassing a wide range of average stand diameters within a given forest type. Data from permanent plots with repeated measurements are preferred because periodic mortality is actually observed, which reduces assumptions about the self-thinning process (Zeide, 1987). Stands selected to deﬁne a line or limit of average maximum stand density should be at or near the upper limits of stand density with respect to their

average diameter. The resulting line showing number of trees per acre by mean stand dbh is sometimes interpreted as a self-thinning line, or line of 100% relative density (Fig. 6.1). The line provides a useful deﬁnition of the upper limits of stand density because the number of trees per unit area and mean dbh are highly correlated. Reineke (1933) postulated that the coefﬁcient b, which determines the slope of the self-thinning line (Equations 6.1 and 6.2), assumes a value close to 1.605 for all tree species. For even-aged, upland oak forests in the eastern United States dominated by white, black, scarlet and chestnut oaks, the estimated slope coefﬁcient was 1.5 based on temporary plot data from undisturbed stands ranging in site index from 50 to 80 ft (Fig. 6.1). Data from permanent plots in similar oak stands in the Central Hardwood Region produced a self-thinning line with a slope coefﬁcient of 1.57 (Fig. 6.1).

3000

Trees per acre

2000

1000 600 400

200 From Schnur From Gingrich

100 60 2

3

4

6

8

12

16

20

Average dbh (inches) Fig. 6.1. Self-thinning lines (lines of average maximum stand density) for even-aged upland oak forests in the eastern United States based on the Reineke model. The dotted line is derived from Schnur’s (1937) stand table for mixed upland oak stands widely distributed across the eastern United States. The solid line is derived from Gingrich’s (1971) stand table for mixed upland oak stands in the Central Hardwood Region. The arrow from a hypothetical disturbed stand represented by the dot on the graph illustrates a typical trajectory of convergence with the Gingrich-based self-thinning line. Trajectories of stands below the self-thinning line generally move slightly downward from left to right. The downward trend results from competition-induced mortality, which occurs even in stands below average maximum density. After convergence with the self-thinning line and in the absence of further disturbance, the trajectory continues along the self-thinning line. The slope coefficient for the Schnur-based self-thinning line is 1.50 and for the Gingrich-based line is 1.57.

Self-thinning and Stand Density

Over time, stands lying below the selfthinning line will grow and move towards the line. On approaching the self-thinning line, stand development trajectories converge with the self-thinning line. Stand development then proceeds along the line from upper left (younger stands) to lower right (older stands) (Fig. 6.1). However, density-dependent mortality does not occur among trees with equal probability; it is concentrated among the suppressed trees. The overall rate of mortality thus is greatest during the stem exclusion stage of stand development, which is when a large proportion of trees succumb to suppression.

The 3/2 rule Another approach to deﬁning the self-thinning line is based on the relation between average total plant biomass and number of plants per unit area in single-species populations undergoing density-dependent mor-

229

tality (Yoda et al., 1963). The power function model, similar to Equation 6.2, is used to describe the relation. However, in this case the model expresses the relation between average plant dry weight (biomass), w, and number of plants per unit area (N) such that: w = aNb

[6.3]

where a and b are usually estimated by regression from experimental data or ﬁeld observations. Alternatively, the relation can be expressed as total plant weight (W) per unit area by: W = aNb

[6.4]

The 3/2 power relation is displayed with N on the horizontal axis, in contrast to Reineke’s model, where N is displayed on the vertical axis. Self-thinning for the 3/2 power relation thus graphically proceeds from lower right (younger stands) to upper left (older stands) along the self-thinning line (Fig. 6.2).

Mean dry weight of tree bole (kg)

600 200 100 50 Self-thinning line 20 10 5 3 1

0.3 100

200

500

1000

2000

4000

7000

Trees per acre Fig. 6.2. A self-thinning line (line of average maximum stand density) for normally stocked even-aged upland oak forests in the eastern United States based on the relation between average dry weight of tree bole (inside bark) and number of trees per acre. The relation is conceptually similar to Reineke’s model (Fig. 6.1), but differs in format. The arrow from a hypothetical disturbed stand represented by the dot on the graph illustrates a possible trajectory of convergence with the self-thinning line. As stand biomass increases and the number of trees per acre decreases over time, the stand trajectory moves upward and to the left along the self-thinning line. (Adapted from Schnur, 1937.)

230

Chapter 6

Numerous studies have shown that b, the slope coefﬁcient in Equation 6.3, approximates 1.5 (and equivalently, b = 0.5 in Equation 6.4) for many plant species including herbs, shrubs and trees (Yoda et al., 1963; Harper, 1977; Miyanishi et al., 1979; White, 1985; Weller, 1987a). The relation consequently has become known variously as the ‘3/2 power law of self-thinning’, the ‘self-thinning rule’ and the ‘3/2 rule’. However, the relation is herein referred to as a rule rather than a law because of its demonstrated lack of generality (Sprugel, 1984; Weller, 1987a; Zeide, 1987; Norberg, 1988; Lonsdale, 1990) and the absence of a supporting theory (Hutchings, 1983). Discrepancies between observed slope values and 1.5 nevertheless have been interpreted as experimental error because of the coefﬁcient’s presumed generality (Miyanishi et al., 1979; White, 1981). The rule also is purported to be independent of environmental factors (Yoda et al., 1963; White and Harper, 1970), and to be applicable to species mixtures as well as single-species stands (Westoby, 1984; White, 1985). The 3/2 rule can be interpreted geometrically. The rule assumes that plant weight, w, is proportional () to plant volume, v, which in turn is proportional to any linear plant dimension on which volume depends raised to the third power: v w = aN b

[6.5]

If we select crown diameter (Cd) as a linear dimension of interest, then Cd 3 aN b. To conceptualize the relation geometrically, it is convenient to consider Cd 3 proportional to a cylinder representing the ‘exclusive space’ of a tree (Norberg, 1988). Then Cd2, which is proportional to the cylinder’s cross-sectional area, can be used to represent crown area. Further, the cylinder’s height is assumed proportional to crown diameter. This three-dimensional space conceptually envelopes the tree, extending downward from the top of the crown to its corresponding ‘exclusive ground area’ and into its soil space. The volume of exclusive space also can be viewed as a hexagonal column, which conceptually allows for sym-

metrical packing of trees without producing crown overlap or unoccupied area as in circular crown areas (Fig. 6.3). The explanatory value of this simpliﬁed geometric view and its relation to the 3/2 rule is apparent from the geometric relation between the volume of a cylinder and its diameter (i.e. volume is equal to the cylinder’s squared diameter raised to the 3/2 power). Note that this relation only holds when the cylinder’s height is proportional to its diameter. To satisfy the geometric analogy for the 3/2 power rule, the thinning-rule model must provide a measure of Cd2. Such a measure is given by N, the number of trees per unit area (Equation 6.3). Because the reciprocal of N represents the area occupied by the average tree, N is related to crown area and thus crown diameter (Cd 2). For every unit increase in crown area, the exclusive space (volume) of a tree increases by 3/2. So, given a ﬁnite amount of growing space, the number of trees (N) in that space must decrease at a rate of 3/2 per unit increase in crown area. To conform to this geometric model, however, a tree must maintain the same height-todiameter ratio during self-thinning (Fig. 6.3). Accordingly, the various tree structures, including bole and crown, must remain proportionately similar during selfthinning (Yoda et al., 1963). Such constancy of proportions is known as isometry or geometric similarity (McMahon and Bonner, 1983; Norberg, 1988). Under the 3/2 power rule, coefﬁcient b (Equation 6.3) is assumed to be 3/2 for all species. In contrast, coefﬁcient a varies among species and determines the intercept (or elevation) of the thinning lines. This coefﬁcient has been termed the ‘packing constant’ (Norberg, 1988) because it reﬂects the proportion of space occupied by plant biomass and the average plant biomass per unit of ground area. Coefﬁcient a thus increases with increasing density, or packing, of plant parts within a tree’s exclusive space. Coefﬁcient a also has been theorized to be related to the mass of mechanical tissue (bolewood in the case of trees) required to support a unit area of canopy (Givnish, 1986).

Self-thinning and Stand Density

231

D

H

Fig. 6.3. The ‘exclusive space’ of closely packed trees. D is the diameter of the ‘exclusive ground area’ associated with each tree’s crown area and H is tree height. The thinning rule theory implies that the ratio of D : H remains constant throughout stand development. (Redrawn from Norberg, 1988, with permission from the University of Chicago.)

Relation between Reineke’s model and the 3/2 rule The diameter of a tree raised to some power between 2 and 3 equals its volume. For many species, the value of the exponent has been shown to be near 2.5 (Yoda et al., 1963). Reineke’s model and the 3/2 rule therefore are related by the approximate relation between average tree volume, v, and tree diameter, D, where v = D 2.5

[6.6]

From Equation 6.2 it then follows that D 2.5 N 2.5/b

[6.7]

When b = 1.605 in Equation 6.1 (Reineke’s postulated constant), the relation becomes: v = N1.56

[6.8]

where the exponent approximates 3/2 (Zeide, 1985). In postulating a constant of 1.605, Reineke was implying, intentionally or not, that the relation between a

tree’s growing space and its diameter is not constant (i.e. not isometric). To be consistent with the assumption of constancy of tree proportions (isometry) inherent in the 3/2 power rule, the Reineke model must assume a slope constant of 2.

Oak forests and the 3/2 rule Knowledge of the limiting relation between numbers of trees and volume per tree in oak stands is silviculturally useful. The relation can be used as a standard against which other stands can be measured. The 3/2 power rule attempts to describe this limiting relation in general terms for a wide range of plant communities. But to what extent do oak forests conform to the 3/2 power rule? Evaluation of the rule can be divided into two questions. First, does the mathematical form of the model (i.e. the power function) adequately express the relation? Second, is the slope coefﬁcient of 3/2 universally

232

Chapter 6

applicable? If the answer to the ﬁrst question is no, then the second becomes irrelevant. However, rejection of 3/2 as a universal slope coefﬁcient does not, by itself, negate the utility of the model form. Evaluating conformance of oak forests to the 3/2 power rule is complicated by three factors. First, oak yield tables are the only comprehensive sources of information. But yield tables seldom report total tree biomass or volume; most yield tables include only bole volume, which must substitute for bole biomass. Moreover, even if bole mass were measured directly, it is unlikely to be a constant fraction of total tree mass (Sprugel, 1984). Although information on the allocation of biomass to below- and above-ground portions of oaks is sketchy, there is evidence that allocation varies greatly with site quality and between trees of coppice and non-coppice origin (Canadell and Rodà, 1991). Second, most oak stands are comprised of a mixture of species. This confounds the effects of competition within and among species, which have fundamentally different explanations in relation to the thinning rule (Zeide, 1985; Norberg, 1988; Weller, 1989). Variation in wood density among species can also introduce error into coefﬁcient estimates. Third, yield tables available to evaluate self-thinning relations for oak have been smoothed by hand-ﬁt curves or other unspeciﬁed methodologies. As such, they represent models, not data (Weller, 1987a). This obscures variation in the original data and may introduce other possible errors (Lonsdale, 1990). Despite these problems, it may be of interest to evaluate the 3/2 rule in relation to existing yield tables. Based on three oak yield tables for the midwestern and eastern United States, the line representing the relation between bole volume and number of trees produced slope coefﬁcients that ranged from 1.7 to 2.2 (Fig. 6.4). Those estimates were based on substituting bole volume for w in Equation 6.3. The yield table values ﬁt the power function with negligible error (Fig. 6.4), which indicates that their developers deemed this function to adequately describe the self-thinning line.

The tenuous conclusions from the above empirical evidence indicate that self-thinning lines in oak forests: (i) show varying proximity to the theoretical 3/2 slope; and (ii) are nominally, if not statistically, more negative than 3/2. Other investigators have concluded that self-thinning lines based on yield tables are likely to have slopes steeper than 3/2 because only portions of trees (boles) are represented, and because they include only trees above some minimum size (Harper, 1977; Lonsdale, 1990). The latter factor may explain, in part, why the yield tables for the Connecticut oak stands (Fig. 6.4C), limited to stems 2 inches dbh and larger, produce a steeper thinning line than either the upland oak or the Wisconsin yield tables, which included stems 0.6 inch dbh and larger (Fig. 6.4A and B). More recent data based on permanent plot data from unthinned upland oak stands representing three site classes in the oak–hickory region produced self-thinning lines that were quite different from those derived from earlier yield tables (Gingrich, 1971). The newer data suggest that each site class produces a separate concavedownward curve when both axes are transformed to logarithmic scale (Fig. 6.5). This pattern is consistent with those derived from Douglas-ﬁr yield tables based on longterm observations from permanent sample plots (Curtis, 1982; Zeide, 1987). However, it is possible that at least part of the concave pattern could be caused by incomplete utilization of growing space in younger oak stands (e.g. stands with less than 500 ft3 per acre of volume and more than 1200 trees per acre in Fig. 6.5). Yield tables for English oak stands in England also show that the self-thinning line plotted on log–log scale does not form a straight line. The slope coefﬁcient for fully stocked stands up to 100 years old is 1.48, which approximates to the hypothetical value of 3/2 (White and Harper, 1970). However, in stands between 100 and 150 years old, the coefﬁcient shifts to 1. This shift indicates that bole volume per unit of ground area remains constant for English oak with continued tree growth

Self-thinning and Stand Density

100

100 B Average tree volume (ft3)

Average tree volume (ft3)

A

10

b = –3/2 1

10

b = –3/2 1

b = –1.68

b = –1.83 0.1 100

233

1000 Trees per acre

10,000

0.1 100

1000 Trees per acre

10,000

100 Average tree volume (ft3)

C

10

b = –3/2

1

b = –2.16 0.1 100

1000 Trees per acre

10,000

Fig. 6.4. Self-thinning lines for oak stands derived from published yield tables compared to the 3/2 thinning rule line. (A) From yield tables for even-aged upland oak stands in the central and eastern states. (From Schnur, 1937.) (B) From yield tables for mixed hardwood stands dominated by northern red oak in southwestern Wisconsin. (From Gevorkiantz and Scholz, 1948.) (C) From yield tables for mixed oak stands in Connecticut. (From Frothingham, 1912.) Heavy lines are fit by linear regression to table values (shown by dots); b is the slope coefficient. Each regression fit is based on the model: log10(v) = a + log10[b(N )] where v is the volume of the average trees, N is stand density (trees acre−1 and log10 is the logarithm to the base 10. Regression estimates are averaged across the site classes given in each yield table.

and self-thinning (Harper, 1977; Norberg, 1988). Accordingly, the volume of bolewood lost to mortality would be compensated by the gap-ﬁlling of survivors. A slope of 1 also could indicate stagnating height growth in old stands, or root competition and soil physical constraints on root expansion (Norberg, 1988). The emerging evidence from permanent plot data collectively indicates that the thinning line as expressed by the 3/2 thinning rule does not form a straight line

over the life of a stand. If we accept that evidence, the power function equation generally, and thus the 3/2 slope coefﬁcient speciﬁcally, cannot realistically describe self-thinning in oak stands. Moreover, there is evidence that the relation is not independent of site effects (Zeide, 1987). Conformity of tree growth to the slope coefﬁcient of 3/2 requires that trees maintain geometric similarity and constant proportions, or isometry, among their various components as they grow. Alternatively,

234

Chapter 6

Stand volume (ft3 acre–1)

4000 2000 1000 500

200 100 60 100

Site index 55 ft Site index 65 ft Site index 75 ft –3/2 rule

200

300

500

1000 1500

3000

Trees per acre Fig. 6.5. Stand volume in relation to stand density for three site index classes in unthinned (normal) stands in upland hardwoods dominated by mixed oaks in the Central Hardwood Region. The stands represent mixtures of black, white, scarlet and chestnut oaks. The dotted line is the theoretical self-thinning line (0.5 from Equation 6.4) assumed by the thinning rule model; other lines represent table values based on models derived from permanent plots. (Adapted from Gingrich, 1971, Table 1.)

elastic similarity occurs when tree components proportionately change with increasing tree size (allometry) (McMahon and Bonner, 1983). Constant ratios among tree dimensions such as crown diameter and bole diameter in relation to changing overall tree size thus represent geometric similarity for those dimensions. For a species or species group to intrinsically conform to the 3/2 rule, ratios among tree components must remain constant (isometric) as tree size increases. A test of isometry is provided by the ratio of crown diameter to dbh in relation to changing tree size. These relations are deﬁned by tree-area equations for oaks in stands at average maximum density. The equations indicate that the crown diameter : dbh ratio for forest-grown trees declines from about 30 : 1 for trees of small diameter to about 17 : 1 for trees of large diameter (Fig. 6.6d and e). For open-grown trees, ratios of crown diameter to dbh vary even more (Fig. 6.6a–c). In either case, small oaks have proportionately more crown area than large oaks. Such proportionate changes in tree dimensions are not consistent with geomet-

ric similarity and hence the 3/2 rule. In fact, bole diameters must increase at a proportionately faster rate than crown area (and correlatively crown mass) to prevent trees from collapsing under their own weight (McMahon, 1973). True isometry in oaks or any other tree species therefore is unlikely. Moreover, predictable changes in crown diameter : dbh ratios, by themselves, provide an alternative basis for deﬁning a self-thinning line and corresponding measures of relative density, as discussed later in this chapter. Self-thinning consequently appears to be heavily inﬂuenced by tree geometry, which is continually changing to meet requirements for structural resistance to bole breakage as crown mass increases (McMahon, 1973). Moreover, trees grow with great physical plasticity to take advantage of their changing competition environment (Sorrensen-Cothern et al., 1993). For example, one side of a tree crown may expand into the gap created by the death or removal of one of its neighbours, resulting in an expanded but nonsymmetrical crown. Trees that survive

Self-thinning and Stand Density

45

a b

40 Crown diameter : dbh ratio

235

35

c d e

30

25

20

15 0

4

8

12 16 Dbh (inches)

20

24

Fig. 6.6. Crown diameter: dbh ratios in relation to dbh. (a) Open-grown pin oak. (From Krajicek, 1967.) (b) Open-grown upland oaks and hickories. (From Krajicek et al., 1961.) (c) Open-grown bur oak. (From Ek, 1974.) (d) Northern red oak in stands at average maximum density. (From McGill et al., 1991.) (e) Upland oaks and hickories in stands at average maximum density. (From Gingrich, 1967.)

self-thinning acquire new resources (space, light, soil moisture and nutrients) as a consequence of spatial adjustments resulting from the death of neighbouring trees and differential growth rates among competing survivors. Associated with these newly acquired resources are changes in the allocation of growth to the various parts of trees, which thereby inﬂuence their proportions, shape and competitive relations with neighbours (Sorrensen-Cothern et al., 1993). The capacity of oaks to ﬁll irregular canopy spaces, and thus their conformity to allometric growth, may be further reinforced by their upward spreading (decurrent) crowns and weak apical control of lateral branching. Despite apparent limitations of the 3/2 self-thinning rule to describe the underlying geometric relations for trees in evenaged oak stands, the general pattern of rapidly decreasing numbers of trees with increasing mean size is well established. Self-thinning formulae such as the 3/2 rule and Reineke’s model describe standlevel changes in the number of trees with

increasing tree size. These formulae are useful expressions of a relatively predictable process that has practical silvicultural value for deﬁning limits of stand density in relation to average tree size. In turn, those limits can be used as a standard or index for expressing the relative density of any other stand. This leads to the subject of stand density and how it can be measured and expressed in oak forests.

Stand Density and Stocking Terminology Silviculturists are often interested in three related measures of stand density: absolute density, relative density and stocking. These measures can be used to describe a stand relative to some standard of comparison or to some condition that meets a silvicultural objective. Silvicultural decisions are often based on such measures of stand density, and the desired condition of the stand after treatment is usually described

236

Chapter 6

by these measures. Although the terms absolute density, relative density and stocking have not always been used consistently, some general conventions and deﬁnitions have been established (Walker, 1956; Bickford et al., 1957; Gingrich, 1964; Husch et al., 1982; Ernst and Knapp, 1985). Absolute density (or simply density in common usage) is a quantitative, objective measure of one or more physical characteristics of a forest stand expressed per unit area. Measures of absolute density are expressed quantitatively as a tree count, area, volume or mass. Ecologists usually use the term density to refer exclusively to the number of individuals per unit area. In forestry, however, the term can refer to any of several measures of site occupancy, including number of trees, basal area or volume per unit area. Measures of density are usually restricted to trees larger than some minimum size, usually expressed as a minimum dbh. Specifying this minimum size is important because absolute density usually differs with differences in the minimum measured tree size. Measures of relative density provide additional information by comparing an absolute density to a reference value. An example of a measure of relative density is the ratio of the number of trees of a given species per acre to the total number of trees per acre. Expressed as a percentage, this value has long been used by ecologists to deﬁne the relative density of a species within a speciﬁed area. Forest regeneration and growth can be greatly affected by relative stand density. Consequently, silviculturists have developed various methods of expressing relative density. Virtually all silvicultural deﬁnitions of relative density involve ratios. For example, Reineke’s (1933) stand density index provides a reference line describing the maximum number of trees per acre for stands of a given mean dbh (Fig. 6.1). The maximum number of trees decreases rapidly as the mean stand diameter increases. For any given stand, the observed number of trees and mean dbh can be used to compute the ratio (or percentage) of the number of the maximum

(reference) number of trees indicated by the stand density index relation (Reineke, 1933; Schnur, 1937). Other common measures of relative oak stand density are based on tree–area ratios or stocking per cent (Chisman and Schumacher, 1940; Gingrich 1967; Ernst and Knapp, 1985). In their application to oak forest types in North America, most measures of relative density are designed to compare one or more absolute measures of stand density to a standard. The standard is often based on an observed maximum absolute density for undisturbed natural stands at a comparable stage of development, but it may be based on other limits or reference conditions. For example, crown competition factor (CCF) estimates stand density relative to a minimum tree crown area per acre below which trees do not fully utilize available growing space (Krajicek et al., 1961). The method of French ‘normes’ compares the observed number of trees and the mean height of dominant trees to both a maximum density and the minimum number of trees necessary to maintain dbh growth below 2 mm (0.08 inch) per year, which by European standards is considered most desirable for veneer production (Oswald, 1982). Other measures of relative density that generally have not been applied to oak forests, but could be, include: Curtis’s (1982) relative density index (references observed basal area per acre to that of an undisturbed stand with the same quadratic mean diameter); Wilson’s (1946) relative spacing index (references the observed number of trees per acre to the number of trees in an undisturbed stand having the same dominant height); and Drew and Flewelling’s (1977) relative density index (references number of trees per unit area to the volume of the average tree). The latter method is analogous to the graphical format for expressing the 3/2 power rule discussed earlier. Comprehensive reviews of measures of relative density include those by Curtis (1970) and Stout and Larson (1988). Stocking is a subjective term used to describe the adequacy of any observed level of stand density with respect to a silvicultural goal (Bickford et al., 1957;

Self-thinning and Stand Density

Gingrich, 1964). The terms overstocked, understocked and fully stocked are used to describe stocking adequacy relative to a speciﬁed silvicultural goal. Accordingly, a stand may be overstocked (too dense) for one silvicultural objective and fully (i.e. appropriately) stocked for another, or may be overstocked at one age and understocked at another. In contrast to the term stocking, the term stocking per cent is a measure of relative density speciﬁcally associated with the Gingrich-style stocking diagram. This diagram combines measures of absolute and relative density into a single graphical format (Gingrich, 1967). Stocking per cent is a widely used measure of stand density in North American oak silviculture. It is based on the relation between tree size and associated growing space requirements discussed later in this chapter. The word stocking is often used incorrectly to refer to stocking per cent (a measure of relative density). This sometimes creates confusion because, as discussed later, full stocking is synonymous with complete utilization of growing space, which covers a wide range of stocking percentages on the Gingrich stocking diagram. Normal stocking is a term used to describe undisturbed even-aged stands that are at or near maximum density for their age. Normally stocked stands are characterized by a lack of gaps in the forest canopy and a relatively uniform spacing between stems. Basal area and cubic foot volume are at or near their maximum for a given stand age and site quality. Normally stocked stands (sometimes simply called normal stands) usually are identiﬁed subjectively based on these criteria. Observations of the number, basal area and volume of trees per acre in normally stocked stands across a wide range of stand age and site quality classes have been used to develop normal yield tables. These tables specify the expected maximum basal area and maximum cubic foot volume for unmanaged stands of a given age and site class. In addition to their application to yield estimation, the tabulated values can be used as reference conditions to estimate the relative density of other stands.

237

Maximum and minimum growing space There are limits to the amount of growing space a tree of a given bole diameter can occupy. Although this may seem self-evident, the concept is central to quantifying stand density and stocking per cent in oak stands. The actual amount of space that a tree occupies is difﬁcult to measure because it includes crowns and roots that overlap in three dimensions with other trees. Fortunately, for many silvicultural purposes, a tree’s growing space can be adequately estimated as a circular area, or tree area, representing the crown. In this context, tree area is interpreted geometrically as a tree’s area of inﬂuence or potential inﬂuence concentric to the tree bole; it is also highly correlated with dbh. Estimates of the maximum area that a tree of a given dbh can occupy are usually developed from crown and dbh measurements of open-grown trees. In contrast, estimates of the minimum area that a tree requires are usually developed from measurements of tree diameters in normally stocked stands. Trees that are open-grown throughout their lives develop the largest crowns possible for their dbh and species. Consequently, open-grown trees have often been used to estimate the maximum area a tree of given species and dbh can occupy. There is a high correlation between bole diameter and crown area of open-grown trees. This relation has led to the development of equations for estimating the crown areas of opengrown trees from dbh for various oaks and associated species in several regions in the eastern United States. The results have shown that the relation between maximum crown width and bole diameter is often linear or nearly linear (Krajicek et al., 1961; Krajicek, 1967; Ek, 1974). An example is the maximum crown width equation applicable to oaks and hickories in the Central Hardwood Region, which is given by: CWmax = 3.12 + 1.829D

[6.9]

where CWmax is the estimated crown width (ft) of an open-grown upland oak or hickory, and D is tree dbh (inches) (Krajicek et al.,

238

Chapter 6

1961). Assuming tree crowns are circular, squaring both sides of Equation 6.9 and multiplying by π/4 deﬁnes maximum crown area (CAmax) in relation to dbh so that: CAmax = 7.645 + 8.965D + 2.627D2 [6.10] CAmax therefore is the approximate circular crown area (ft2 in vertical projection) of an open-grown upland oak or hickory. Maximum crown width equations also have been derived for other species and regions (Table 6.1). An exponent in the diameter term of some equations indicates non-linearity in the relation. As in the derivation of Equation 6.10, equations in Table 6.1 can be similarly expressed as crown area. Graphical presentation of equations facilitates comparisons among species. For example, open-grown black walnut trees have larger crowns than oaks and hickories for a given diameter, whereas shortleaf pines have smaller crowns. The maximum crown width of sugar maple

may be larger or smaller than that of oaks and hickories, depending on dbh (Fig. 6.7). Assuming that maximum crown width equations adequately express the maximum amount of above-ground growing space that a tree of a given diameter can occupy, we can estimate the fewest trees of a given dbh required to completely occupy an acre, i.e. 43,560/CAmax. Alternatively, the maximum tree area for all the trees on any acre can be calculated by summing their individual maximum crown areas (Equation 6.10 and Table 6.1). When the sum of the maximum crown areas equals the area of an acre (43,560 ft2), the stand is said to have a maximum tree–area ratio (TARmax) of 100%. This represents the condition where tree–area satisﬁes the minimum requirements for full utilization of growing space given that tree crowns are, for their dbh, maximally extended. Maximum tree–area ratio (TARmax) therefore is a relative measure of stand density that deﬁnes the percentage of an area

Table 6.1. Equations for estimating open-grown crown widths from dbh for oaks and some commonly associated species.

Species (location) American elm (Wisconsin) American basswood (Wisconsin) Black cherry (Wisconsin) Black oak (Wisconsin) Black walnut (unspecified) Black walnut (Wisconsin) Bur oak (Wisconsin) Green ash (Wisconsin) Jack pine (Quebec) Loblolly pine (unspecified) Northern red oak (Wisconsin) Oaks and hickories (Iowa) Pin oak (unspecified) Red maple (Wisconsin) Shagbark hickory (Wisconsin) Shortleaf pine (Missouri) Sugar maple (Wisconsin) Sugar maple (Eastern US) Sweetgum (unspecified) White oak (Wisconsin) a Crown

Maximum crown widtha 2.829 + 3.456D 0.8575 0.135 + 3.703D 0.7307 0.621 + 7.059D 0.5441 4.504 + 2.417D 4.873 + 1.993D 4.901 + 2.480D 0.942 + 3.539D 0.7952 4.755D 0.7381 2.036 + 1.736 4.78 + 1.56D 2.850 + 3.782D 0.7968 3.12 + 1.829D 9.06 + 1.525D 4.776D 0.7656 2.369 + 3.548D 0.7986 2.852 + 1.529D 0.868 + 4.150D 0.7514 12.08 + 1.32D 2.65 + 1.975D 3.689 + 1.838D

Source Ek, 1974 Ek, 1974 Ek, 1974 Ek, 1974 Krajicek, 1967 Ek, 1974 Ek, 1974 Ek, 1974 Vezina, 1963 Roberts and Ross, 1965 Ek, 1974 Krajicek et al., 1961 Krajicek, 1967 Ek, 1974 Ek, 1974 Rogers, 1983 Ek, 1974 Smith and Gibbs, 1970 Krajicek, 1967 Ek, 1974

width in feet given tree dbh (D) in inches; corresponds to CWmax in text. Assuming tree crowns are circular in cross-section, maximum crown area in square feet is equal to (CWmax)2•π/4.

Self-thinning and Stand Density

where summations (∑) are over all trees per acre, Di is the dbh of tree i and N is number of trees per acre. Note that ∑Di2 is equal to the stand basal area in square feet per unit area divided by π/576. A CCF of 100 (or equivalently, ∑TARmax = 100%) therefore is usually interpreted as the approximate lowest density at which a stand fully utilizes above-ground growing space. Stands with CCFs below 100 are certain to have canopy gaps. CCF values near 200 have been observed for undisturbed oak–hickory stands (Krajicek et al., 1961). Just as trees have a maximum area they can occupy, they also have a minimum tree area that is necessary for good physiological function and survival. However, minimum tree area is derived quite differently from its maximum tree-area counterpart. Unlike maximum tree area, which can be estimated from open-grown trees, minimum tree area is difﬁcult to observe directly for individual trees. Minimum tree area requirements nevertheless can be estimated from data obtained from undisturbed, normally stocked, even-aged stands. Estimation is based on deriving minimum tree–area ratio (TARmin) equations

(e.g. an acre) that would be utilized by trees when all tree crowns are fully extended. When TARmax is less than 100%, the trees present would not utilize the available growing space even when their crowns are fully extended. Consequently, reducing TARmax below 100% by thinning will, at least temporarily, result in unutilized growing space. Calculating TARmax can be simpliﬁed by dividing equations for open-grown crown areas (e.g. Equation 6.10, or Table 6.1) by 435.6, the area comprising 1% of an acre. For Equation 6.10, TARmax is given by: TARmax = 0.0175 + 0.0205D + 0.00603D2 [6.11] where TARmax is the maximum percentage of an acre that a tree of a given dbh (D) can occupy. The sum of TARmax for all the trees on an acre is sometimes referred to as crown competition factor (CCF) (Krajicek et al., 1961) and is calculated by summing TARmax for individual trees as follows: CCF = ∑(0.0175 + 0.0205Di + 0.00603Di2) [6.12] = 0.0175N + 0.0205∑Di + 0.00603∑Di2 [6.13] 60

Oak and hickory Shortleaf pine Sugar maple Black walnut

Maximum crown width (ft)

50

40

30

20

10

0 0

4

8

239

12

16

20

24

28

Dbh (inches) Fig. 6.7. Estimated open-grown crown widths of oaks and hickories and three commonly associated species in relation to bole diameter (dbh). See also Table 6.1. (Adapted from Ek, 1974 (sugar maple), Krajicek, 1967 (black walnut), Krajicek et al., 1961 (oak–hickory), Rogers, 1983b (shortleaf pine).)

240

Chapter 6

that express tree growing space requirements for normally stocked stands (Chisman and Schumacher, 1940). Like TARmax, TARmin expresses tree area in per cent of an acre. Just as maximum tree area is a linear function of diameter and diameter squared (Equations 6.9 and 6.10), a tree’s minimum tree area can be similarly expressed by: TARmin = c0 + c1D + c2D2

[6.14]

where TARmin is the estimated minimum per cent of an acre required by a tree of a given dbh (D) in a normally stocked forest. Unlike the maximum tree–area coefﬁcients, the coefﬁcients for the minimum tree–area equation are not derived from measurements of crown diameters. Instead, they are estimated by regression by assuming the sum of the tree areas for all trees on an acre of undisturbed, normally stocked forest is equal to 43,650 ft2, or 100% of an acre (Chisman and Schumacher, 1940; Gingrich, 1967). Minimum tree area then can be expressed directly as a percentage of an acre. This expression has been termed stocking per cent (S%) (Gingrich, 1967) and is given by: S% = ∑ (b0 + b1Di + b2D22)

[6.15]

= b0N + b1∑Di + b2∑Di2

[6.16]

where summations (∑) are over all trees per acre, S% is the percentage of an acre ﬁlled by the minimum tree areas of all trees on that acre, Di is the dbh of tree i, N is the number of trees per acre, and b0, b1 and b2 are coefﬁcients (usually estimated by regression). The stocking percentage represented by a single tree can be derived by solving Equation 6.16 for N = 1. When stocking percentage equations are expressed on a per acre basis (e.g. Equation 6.15 or 6.16), equations for minimum tree area in square feet can be derived by multiplying each term in the equation by 435.6 (the number of square feet in 1% of an acre). Although stocking percentage is usually the relative density measure of choice, rescaling to square feet facilitates comparing the estimated tree areas representing

maximum and minimum growing space (Fig. 6.8). Other factors being equal, the closer a tree’s crown is to its maximum size for its dbh, the faster the tree’s diameter and gross volume growth. The minimum tree–area ratio is reportedly independent of stand age and site quality (Chisman and Schumacher, 1940; Gingrich, 1967), and can be applied to mixed as well as to pure stands. The methodology for deriving minimum tree–area equations is described in more detail by others (Chisman and Schumacher, 1940; Gingrich, 1967; Roach, 1977; Ernst and Knapp, 1985; Stout and Nyland, 1986). Oak and associated forest types are seldom comprised of a single species. In applying stocking equations, it is therefore important to recognize differences in tree–area ratios among species. In developing stocking equations, coefﬁcients for individual species can be derived by incorporating species-speciﬁc terms into stocking equations so that the Equation 6.15 expands to the more general form: no. spp.

no. trees

no. trees

S% = ∑ (b0j Nj b1j ∑ Dij b2j ∑ Dij2) j=1 i=1 i=1 [6.17] where the outer summation (∑) is over all species; the inner summations are over all i trees of species j; b0j, b1j and b2j are coefﬁcients speciﬁc to species j; Nj is the number of trees of species j; and Dij is the diameter of tree i of species j (Roach, 1977). In forests such as the oak–hickory type of the Central Hardwood Region, the minimum tree–area ratios of the predominant species do not differ signiﬁcantly (Krajicek et al., 1961; Gingrich, 1967). A single set of coefﬁcients therefore can be used to represent the major species of the forest type. In other forest types, tree–area ratios differ appreciably among species (Roach, 1977; Stout and Nyland, 1986; Zhang et al., 1995). When such differences occur, separate coefﬁcients for individual species or species groups can improve the accuracy of relative density equations. This is the case in the Allegheny hardwood forests of Pennsylvania, which are comprised of mixed stands of black cherry, yellow-

Self-thinning and Stand Density

241

3000

Crown area (ft2)

2500

2000 Maximum crown area 1500

1000 Minimum crown area

500

0 2

6

10

14 18 22 Dbh (inches)

26

30

Fig. 6.8. Estimated maximum and minimum tree areas in relation to bole diameter (dbh) for upland oaks and hickories in the Central Hardwood Region. The area between the two lines represents the approximate biological range of crown areas for individual trees. (From Gingrich, 1967.)

poplar, red maple, white ash, sugar maple, black birch, yellow birch, American beech, oaks and other species. Analysis of species-speciﬁc tree–area ratios identiﬁed three species groups with signiﬁcantly different tree–area ratios (Stout and Nyland, 1986; Zhang et al., 1995). Stocking equations based on tree–area ratios also have been derived for northern red oak and various forest types of the eastern United States that often include oaks (Table 6.2). The silvicultural value of tree–area ratios and stocking percentage for deﬁning relative density is reinforced by their demonstrated independence of stand age and site quality (Chisman and Schumacher, 1940; Gingrich, 1967). They also have been shown to be little inﬂuenced by variation in stand structure (Gingrich, 1967). Minimum tree–area ratio equations and equivalent stocking percentage equations can be used to calculate an approximate average upper limit of stand density corresponding to that for normally stocked stands. This upper limit, which can be graphically expressed as a line on a stand density diagram, has been termed average maximum density or average max-

imum competition (Ernst and Knapp, 1985). This line is generally interpreted as the level of density that stands tend to return to in the absence of disturbance (Gingrich, 1967; Ernst and Knapp, 1985). Nevertheless, stand density is likely to ﬂuctuate around this line due to variation in weather, minor outbreaks of insects and disease, and other factors associated with ‘regular’ tree mortality. A potential deﬁciency of all measures of maximum relative density is the necessarily subjective selection of stands or plots used to derive tree–area ratios.

Stand density diagrams Stand density diagrams are graphical representations of the equations and variables that deﬁne relative stand density. In practical application, measures of relative density are often more convenient when displayed as diagrams than as equations. The most widely used type of stand density diagram in North American oak and hardwood silviculture is the one developed by Gingrich (1967).

242

Chapter 6

Table 6.2. Equations for estimating stocking per cent (minimum tree–area ratio as a per cent of an acre) by species group. Species (location)

Stocking per cent per acre a

Source

Upland oaks and hickories (Ohio, Kentucky, Missouri, Iowa) Northern red oak (Wisconsin) Sugar maple American beech (Allegheny Plateau) Black cherry Yellow-poplar (Allegheny Plateau) Red maple American basswood White ash (Allegheny Plateau) Black walnut (Unspecified) Shortleaf pine (Missouri)

0.00507N + 0.01698ΣD + 0.00317ΣD2

Gingrich, 1967

0.02476N + 0.004182ΣD + 0.00267ΣD2

McGill et al., 1991, 1999

0.003082N + 0.006272ΣD + 0.00469ΣD2

Stout and Nyland, 1986

0.02794N + 0.01545ΣD + 0.000871ΣD2

Stout and Nyland, 1986

0.01798N + 0.02143ΣD + 0.001711ΣD2

Stout and Nyland, 1986

0.01646N + 0.01347ΣD + 0.002757ΣD2

Schlesingerb

0.008798N + 0.009435ΣD + 0.00253ΣD2

Rogers, 1983

is number of trees per acre, ΣD is sum of diameters (inches/acre), and ΣD2 is sum of squared diameters (square inches/acre). Note that ΣD2 is equal to basal area (ft2/acre) divided by 0.005454. Minimum tree area in ft2 (TARmin) can be obtained by multiplying the calculated stocking per cent by 435.6, the number of square feet in 1% of an acre. bUnpublished equation by Richard C. Schlesinger, USDA Forest Service, Columbia, Missouri. aN

The Gingrich diagram Gingrich’s (1967) stand density diagram is based on tree–area ratio equations. The equations, in turn, are based on data from white, black, scarlet and chestnut oak stands in the Central Hardwood Region. Gingrich’s diagram incorporates measures of absolute density, relative density and stocking percentage into one graph. The measures of absolute density used are number of trees per acre (horizontal axis) and basal area per acre (vertical axis). Quadratic mean stand diameter (i.e. the diameter of the tree of average basal area) is also shown in relation to basal area and trees per acre. For any given stand, observed values of basal area and trees per acre can be used to determine stocking per cent directly from the diagram (Fig. 6.9). Gingrich’s diagram graphically deﬁnes the line representing average maximum stand density in relation to basal area, trees

per acre and mean stand diameter. The line is based on the minimum tree–area equation for upland oaks and hickories (Table 6.2). On the diagram, the line is labelled ‘100% stocking’ or ‘A-level stocking’ (Fig. 6.9). Similarly, the maximum tree–area ratio equation (Equation 6.11) deﬁnes the fewest number of trees of a given diameter sufﬁcient to completely occupy all the growing space on an acre. On the diagram, this reference line is labelled ‘B-level stocking’. It could also be called the ‘line of imminent competitioninduced mortality’ (Drew and Flewelling, 1977). Over a wide range of stand conditions, B-level occurs between 57% and 59% of A-level stocking. Another line, labelled C-level, deﬁnes the relative density at which a stand on an average site requires 10 years to attain B-level stocking. The steep slope of the C-level line reﬂects the relatively rapid rate of increase in

Self-thinning and Stand Density

243

7

110

6

Av e

A

100

et

ree

Ov

90

dia

me

ter

4

ers

toc

80

(in

.)

ke

d

Fu lly s

toc

ke

d

7 in 0 g pe 80 rc en t

B

60

90

70

10 0

11 0

3

50 Un d

C

ers

toc

ke

d

40

30

6 St 0 oc k

40

50

Basal area per acre (ft2)

rag

5

20 1

3

5

7

9

11

13

15

Hundred trees per acre 15

140 130

A

13 Avera ge 12 tre 11 ed iam 10 ete 9 r (i Ov n.) ers 8 toc ked

11 0 Fu

90 80

80

lly sto cke d

50

kin rst

ock ed

St oc

de

60

Un

50

C

gp

70

er

B 70

nt

90

10

100

60

7

0

110

ce

Basal area per acre (ft2)

120

40 50

100 150 200 250 300 350 400 Trees per acre

Fig. 6.9. Relation of basal area, number of trees and average tree diameter to stocking per cent in upland oak stands in the Central Hardwood Region. Values are for trees ≥ 2 inches dbh. Lines of average tree diameter correspond to the dbh of the tree of average basal area or quadratic mean diameter. Upper panel is for stands with quadratic mean tree diameters from 3 to 7 inches dbh; lower panel is for stands with quadratic mean diameters from 7 to 15 inches dbh. The area between curves A and B represents the range of stocking where trees fully utilize growing space. (From Gingrich, 1967.)

244

Chapter 6

stocking (and basal area) in stands with small mean diameters. Qualitative levels of stocking (overstocked, fully stocked and understocked) are also identiﬁed on the stocking diagram. Stands are overstocked when they fall above A-level, fully stocked when they are between A- and B-levels, and understocked when they are below B-level. Stands within the fully stocked range therefore are considered to completely utilize growing space. Although stocking per cent can be calculated directly from the associated tree–area ratio equation (Table 6.2), the diagram facilitates rapid estimation when basal area and number of trees per acre are known. Stocking per cent is also related to canopy cover immediately after thinning and subsequent crown closure (Fig. 6.10). Unless residual stand densities are reduced to levels below 50% stocking, the canopy gaps created by thinning quickly close through crown expansion and correlated dbh growth. The stocking diagram illustrates the inadequacy of basal area alone as an expression of relative stand density. If basal area and stocking per cent were equivalent measures of density, the stocking per cent lines would parallel the horizontal basal area lines – thereby signifying isometry between tree crown area and basal area. Instead, the slope of the stocking per cent lines shows that trees can endure more crowding as they increase in diameter. This relation also implies that trees of large diameter require less space to support a unit of basal area than trees of small diameter. The density diagram is designed for rapid estimation of stocking per cent from two easily obtained measures of stand density: basal area and trees per acre. Note, however, that the stocking per cent equation (Equation 6.16 and Table 6.2) includes three terms: number of trees per acre (N), sum of diameters per acre (∑D) and sum of diameters squared per acre (∑D2) the latter value also is equivalent to 0.005454•basal area per acre. The density diagram incorporates only number of trees and basal area with no direct measure of sum of diameters. Nevertheless, it is important to consider both ∑D and ∑D2 when estimating

stocking per cent because together they account for variation in stand structure. This can be seen from the relation between the variance in tree diameters (2) and the difference or ‘discrepancy’ between arithmetic mean stand diameter (AMD) and quadratic mean stand diameter (QMD): 2 = QMD2 AMD2

[6.18]

where AMD is the arithmetic mean stand diameter (i.e. ∑D/N) and QMD is the tree of average basal area (i.e. [(∑D2/N)0.5]). The greater the variance in tree diameters, the greater the difference between arithmetic mean stand diameter and quadratic mean stand diameter. Moreover, if either AMD or QMD is given, the other can be algebraically derived from knowledge of 2. A valuable property of the stocking per cent equation (Equation 6.16 and Table 6.2) therefore is the indirect incorporation of information about the variance of tree diameters via inclusion of ∑D and ∑D2. The loss of accuracy in calculating stocking per cent that results from not considering ∑D would not be signiﬁcant if all the trees were of equal or nearly equal diameter. But this is seldom if ever the case. When Gingrich (1967) derived the stocking per cent lines shown on his diagram, he incorporated estimates of the variance of tree diameters. He based these estimates on the observed negative linear relation between 2 and QMD. The stocking per cent lines shown on the density diagram therefore reﬂect average or ‘representative’ stand diameter distributions that change as the arithmetic mean stand diameter changes. Stocking per cent derived directly from the stocking per cent equation may be as much as 4% greater than indicated by the density diagram when the range of tree diameters is small. Where there is a wide range of diameters (e.g. in uneven-aged stands), stocking per cent calculated from the equation may be as much as 6% lower than that indicated by the stand density diagram. Differences of this magnitude are usually of little practical importance. However, not all density diagrams based on tree–area ratios incorporate estimates of the variance of diameters. Developers of dia-

Self-thinning and Stand Density

245

A

B

C

Fig. 6.10. Representative canopy cover of oak stands in the Ozark Highlands for three levels of stocking: (A) 100%; (B) 80%; and (C) 60%. Photos for 60% and 80% stocking were taken immediately after thinning. (USDA Forest Service, North Central Research Station photograph.)

grams may wittingly or unwittingly assume that variance in diameters, and thus stand structure, is inconsequential in estimating tree–area ratios. When information on individual tree diameters is available, the tree–area equation or stocking per cent equation will produce estimates that are

more accurate than estimates read from the stocking diagram. Gingrich’s stand density diagram was developed from observations in forests comprised largely of white, black, scarlet and chestnut oaks. Other oaks may have different growing space requirements and

246

Chapter 6

thus different A-level and B-level stocking lines. For example, northern red oak attains 100% stocking at substantially higher levels of absolute density than those represented by Gingrich’s equation. Alternative stocking equations and stand density diagrams therefore have been developed for northern red oak (Sampson, 1983; Sampson et al., 1983; McGill et al., 1991, 1999; Stout, 1991) (Fig. 6.11). Based on the equation of McGill and others (Table 6.2), a stocking per cent of 100 for northern red oak equates to stocking percentages ranging from 120 to 140 on the Gingrich diagram (Fig. 6.9). Assuming the associated tree–area equations are accurate, white, black, scarlet and chestnut oaks and hickories in the Central Hardwood Region require 20–40% more growing space per tree than that required for northern red oak in Wisconsin. Two other stand density diagrams based on tree–area ratios have relevance to oak forests in the eastern United States. One is the stand density diagram for Allegheny hardwoods developed by Roach (1977). Those forests often include northern red oak as an associated species. The related stand density diagram is unique in that it accommodates adjustments in A-level stocking associated with changes in species composition. Because of the relatively narrow crowns of black cherry, white ash and yellow-poplar, the reference line for Alevel stocking increases as the percentage of those species increases. A stand density diagram based on tree–area ratios also has been developed for southern bottomland hardwood forests that include oaks as component species (Goelz, 1995). The diagram is based on the stand tables of Putnam and others (1960) for mixed bottomland stands in the lower Mississippi Valley, lower Piedmont and southern Coastal Plain. In those forests, cherrybark, laurel, Nuttall, overcup, pin, Shumard, water, willow and swamp chestnut oaks are commonly associated with other bottomland species. Despite their widespread use, some of the properties and assumptions incorporated into stand density diagrams based on tree–area ratios are not always well under-

stood or appreciated. Some characteristics worth remembering are presented below. ● Crown closure is assumed to occur when the sum of the corresponding crown areas of open-grown trees equals 43,560 ft2, i.e. ∑TARmax = 43,560 ft and CCF = 100. This is assumed to be the point of crown closure where inter-tree competition begins. However, it is difﬁcult to perceive of a situation where this assumption would be literally true. Irregularities in tree spacing virtually ensure that tree crowns will form overlaps and gaps. The assumption nevertheless provides a quantiﬁable and demonstrably useful reference line that estimates the minimum number of trees necessary to occupy the available growing space. ● On a good site, trees of a given diameter may be younger than trees of the same diameter on a poor site. Trees of the same size nevertheless require the same amount of growing space, regardless of their ages (Krajicek et al., 1961; Gingrich, 1967). ● Stand density can exceed 100% stocking. This gives rise to the term ‘overstocked’. However, in the absence of disturbance, stands are expected to move toward the line of A-level stocking regardless of where they initially fall on the density diagram. ● According to current silvicultural theory, oak stands maintained in the ‘fully stocked’ zone will produce approximately the same total gross volume or biomass growth regardless of where they are maintained within that zone. However, other factors being equal, individual trees in oak stands maintained near B-level stocking will grow faster and usually produce greater total net merchantable volume than trees in stands maintained near A-level. Exceptions include oak stands grown exclusively for wood ﬁbre or stands containing certain other hardwoods. The net merchantable growth per acre of stands containing a large proportion of black cherry (Nowak, 1996) or yellow-

Self-thinning and Stand Density

7

160

6

Av e

A 140

5 O

Basal area per acre (ft2)

247

120 Fu

100

lly

st

oc

ke

ra

ve

ge

rs

to

tre

ck

e

di

am

4

et

ed

er

(in

.) 3

d

80

00

0 11

1 90

60

B Un

40

80 ent rc pe g in

70 de

rs

to

ck

60 ock St

50 ed

40 30

20 100

400

700

1000

1300

1600

Trees per acre

15

180

13

A

toc

ked

90

ly s

10

Ful

140

0

11 0

160

Ave rag e tr 12 ee 11 dia me 10 ter (in. 9 Ove ) rsto 8 cke d 7

120

7 pe 0 rc en 80 t

Basal area per acre (ft2)

200

50

B rsto

cke

30

d

40

Un de

60

St

80

6 oc 0 kin g

100

40 0

100

200 300 400 Trees per acre

500

600

Fig. 6.11. Relation of basal area, number of trees and average tree diameter to stocking per cent in northern red oak stands in Wisconsin. Values are for trees ≥ 2 inches dbh. Lines of average tree diameter correspond to the dbh of the tree of average basal area or quadratic mean diameter. Upper panel is for stands with quadratic mean tree diameters from 3 to 7 inches dbh; lower panel is for stands with quadratic mean diameters from 7 to 15 inches dbh. The area between curves A and B represents the range of stocking where trees fully utilize growing space. (From McGill et al., 1991, used with permission.)

248

Chapter 6

poplar (Trimble, 1968; Beck, 1986) may decrease with decreasing residual stand density. For those species, post-thinning rates of crown expansion and crown densiﬁcation may be insufﬁcient to compensate for loss of growth related to the removal of trees (Nowak, 1996). The general rule of equal growth across the ‘fully stocked’ zone therefore may not hold for some species mixes that commonly include oaks. ● There is evidence, based on long-term forest growth studies, that gross volume growth increases with increasing stand density up to a stand’s maximum biological limits of density (Zeide, 2001). Accordingly, there would be no such thing as an excessive number of trees – as implied by the labelling of the ‘overstocked’ zone of the Gingrich and similar stocking charts. In a general biological context, the concept of ‘overstocking’ therefore may be misleading. Nevertheless, the use of thinning to increase the rate of residual stand growth to more quickly obtain merchantable timber products other than wood ﬁbre generally holds. • Tree–area ratio equations assume that interactions among species do not affect tree area. However, it would seem plausible that co-occurring species that appreciably differ in how they utilize site resources would more efﬁciently ‘share’ a unit space than co-occurring species with very similar resource requirements (McGill et al., 1999). If this were true, we might expect TARmin for two species with very different resource requirements to be smaller (and therefore their joint absolute maximum densities to be larger) when the two species are interacting than when they are growing apart. For example, northern red oak stands with a subcanopy of sugar maple have higher average maximum absolute densities (in basal area and trees per acre) than pure red oak stands (McGill et al., 1999). This suggests that minimum tree–area ratios for a given species may change as species composition changes.

● Stands at identical stocking per cent may develop differently, depending on their structure and history. For example, the growth of a previously undisturbed stand reduced to 70% stocking by windthrow of mostly large trees is likely to be different from a similar stand that was thinned to the same stocking by removing mostly small trees. The general direction of the response should be similar for both stands: increased stocking and basal area, and decreased number of trees. But rates of stand growth and increases in stocking would be expected to differ substantially because of differences in stand structure and composition. Stand density, therefore, does not deﬁne a unique stand condition even within the same forest type and site class. Use of stocking per cent or other measures of stand density to make accurate growth predictions thus requires specifying additional stand characteristics.

Other diagrams Stand density diagrams other than those based on tree–area ratio have been developed, and some of these have been applied to oak forests. One example is represented by Oswald’s (1982) diagram for English (pedunculate) oak and sessile oak in France. This diagram is a variant of Reineke’s (1933) stand density index. However, mean tree diameter as shown on the horizontal axis in the Reineke model is replaced by the mean height of the 100 tallest trees per hectare (Fig. 6.12). Another example is the stand density diagram for upland hardwoods developed by Kershaw and Fischer (1991), which is based on a format developed for Douglas ﬁr (Drew and Flewelling, 1977). It incorporates the same principles as the Reineke and Gingrich diagrams, but shows trees per acre in relation to mean board foot volume per tree. It differs from other oak stand density diagrams by directly incorporating information on merchantable products.

Self-thinning and Stand Density

249

Trees per hectare

2000 1600 1400 1200 1000 800

M

ax im

um

bi

ol

600 500 400

og

ic

al

de

ns

ity

300 250 200 170 140 120 100 85 70

Density index 140 120 100 85 70 16 18 20 22 24 26 28 30 32 34 Mean height of dominant trees (m)

Fig. 6.12. A stand density diagram for sessile and English oaks (non-coppice stands) in France. Stand density is expressed as the relation between trees per hectare (upper storey trees ≥7 cm in diameter at 1.4 m) and the mean height of the 100 largest-diameter trees per hectare. The uppermost line represents maximum biological stand density. Densities below the lowermost line result in annual ring widths exceeding 2 mm, which European foresters regard as undesirable for oak veneer logs. In this diagram, stand densities are indexed to the number of trees per hectare in stands with dominant trees averaging 35 m tall. Density indices between 70 and 140 identify the operational range for residual stand densities; the usual after-thinning range for oaks is between 70 and 100. French foresters refer to these indices as normes. (Redrawn from Oswald, 1982, used with permission.)

Density diagrams and stand growth In addition to their conventional use in managing stand density, density diagrams can be used to show how different stand attributes simultaneously change through time. This application can be illustrated by plotting actual or projected changes in stand basal area and trees per acre on the density diagram. Some growth and yield simulation models provide this as a graphic display option. Stand growth can be displayed in a variety of formats and units, depending on silvicultural objectives (Fig. 6.13). In the absence of disturbance and at densities below 100% stocking (i.e. below the A-level stocking line), we would expect stands to increase in basal area and decrease in number of trees per acre. Similarly, stands above 100% stocking would be expected to gradually decrease in number of trees without substantially increasing in basal area. In either case, the

expectation is for stands to move through time toward the 100% relative density line. But because the reference line of maximum density is only an approximation of an average state that ﬂuctuates, the growth of any given stand can be expected to vary about the line after reaching it. A relation between relative density and stand growth is implicit in oak stocking diagrams. Gross volume growth is assumed to be maximum and nearly constant when stands are fully stocked, i.e. maintained between A- and B-level reference lines. Except as noted above, merchantable volume growth is often maximized by maintaining stand density near B-level. Stocking charts also can be used in conjunction with growth and yield models to compare stand growth responses to various combinations of initial basal area and trees per acre (Leary and Stanﬁeld, 1986; Goelz, 1991). Although the relative density lines shown on stand density diagrams are independent of site

Chapter 6

7 st

oc

ke d

100

80

100 120 140

275 Trees per acre

Fu lly

110

60

B

110

120

300

t

130

Av A er 13 ag e 12 tre e 11 di O am A ver 10 et st er oc 9 (in ke .) d 8

1 per 00 cen

Basal area per acre (ft2)

140

Stand density index

15

250

225

Sto 90 ckin g

250

90 100

200

300

400

6

Trees per acre

7

8

9

10

11 12

Average dbh (inches)

22 20

16

120 110 100

me

18

olu

dv

an

St ex

ind

Average tree volume (ft3)

C

90 80 14

12 200

70

220 240 260 Trees per acre

280

300

Fig. 6.13. Seventeen-year growth of an upland oak stand in Missouri (shown by arrows) displayed on three types of stand density diagrams. To better illustrate stand growth, only an enlarged part of each diagram is shown. (A) Gingrich’s (1967) diagram for oak–hickory forests of the Central Hardwood Region. (B) A Reineke diagram of stand density index for upland oaks of the eastern United States. An index of 100 represents average maximum density (the dotted self-thinning in Fig. 6.1). (Adapted from Schnur, 1937.) (C) A diagram based on the relation between average tree volume and number of trees per acre for upland oak stands of the eastern United States. A stand volume index of 100 represents average maximum density. This diagram follows the format in Fig. 6.4. (Adapted from Schnur, 1937.)

quality and stand age, rates of stand growth are not. Consequently, when growth is plotted on a density diagram, the results pertain to speciﬁc site and stand conditions.

Analysis of estimated oak growth rates based on Dale’s (1972) growth and yield model applied to a wide range of relative density classes showed that the model pre-

Self-thinning and Stand Density

dicted maximum net cubic-foot volume yield at basal areas as much as 20 ft2 acre1 below B-level on the Gingrich stand density diagram (Leak, 1981). Nevertheless, stands maintained below Blevel are likely to incur defects in bole quality associated with the resulting increase in epicormic branching (Dale, 1972; Sonderman, 1985). Stand density diagrams and equations therefore should be considered guides rather than rigid rules applicable to all situations. An understanding of the principles of self-thinning and stand density are prerequisite to knowledgeable application of silvi-

251

cultural systems. Whereas the preceding chapters have focused largely on ‘natural’ ecological events, including forest disturbance, the following chapters focus on silviculture. The silviculture of North American oaks is, and has been, largely centred on the manipulation of the structure, composition and regeneration of natural stands. From an ecological perspective, a silvicultural system therefore represents a planned disturbance or series of disturbances designed to achieve speciﬁc goals. In subsequent chapters the principles discussed in this chapter are applied to silvicultural methods for oak forests.

References Beck, D.E. (1986) Thinning Appalachian pole and small sawtimber stands. Society of American Foresters Publication 86-02, pp. 85–98. Bickford, C.A., Baker, F.S. and Wilson, F.G. (1957) Stocking, normality, and measurement of stand density. Journal of Forestry 55, 99–104. Canadell, J. and Rodà, F. (1991) Root biomass of (Quercus ilex) in a montane Mediterranean forest. Canadian Journal of Forest Research 21, 1771–1778. Chisman, H.H. and Schumacher, F.X. (1940) On the tree–area ratio and certain of its applications. Journal of Forestry 38, 311–317. Curtis, R.O. (1970) Stand density measures: an interpretation. Forest Science 16, 403–414. Curtis, R.O. (1982) A simple index of stand density for Douglas-ﬁr. Forest Science 28, 92–94. Dale, M.E. (1972) Growth and yield predictions for upland oak stands 10 years after initial thinning. USDA Forest Service Research Paper NE NE-241. Drew, J.T. and Flewelling, J.W. (1977) Some recent Japanese theories of yield–density relationships and their application to Monterey pine plantations. Forest Science 23, 517–534. Ek, A.R. (1974) Dimensional relationships of forest and open grown trees in Wisconsin. University of Wisconsin Forestry Research Note July. Ernst, R.L. and Knapp, W.H. (1985) Forest stand density and stocking: concepts, terms, and the use of stocking guides. USDA Forest Service General Technical Report WO WO-44. Frothingham, E.H. (1912) Second-growth hardwoods in Connecticut. USDA Forest Service Bulletin 96. Gevorkiantz, S.R. and Scholz, H.F. (1948) Timber yields and possible returns from the mixed-oak farmwoods of southwestern Wisconsin. USDA Forest Service Lake States Forest Experiment Station Publication 521. Gingrich, S.F. (1964) Criteria for measuring stocking in forest stands. Proceedings of 1964 Society American Foresters and National Convention, pp. 198–201. Gingrich, S.F. (1967) Measuring and evaluating stocking and stand density in upland hardwood forests in the Central States. Forest Science 13, 38–53. Gingrich, S.F. (1971) Management of young and intermediate stands of upland hardwoods. USDA Forest Service Research Paper NE NE-195. Givnish, T.J. (1986) Biomechanical constraints on self-thinning in plant populations. Journal of Theoretical Biology 119, 139–146. Goelz, J.C.G. (1991) Generation of a new type of stocking guide that reﬂects stand growth. USDA Forest Service General Technical Report SE SE-70, Vol. 1, pp. 240–247. Goelz, J.C.G. (1995) A stocking guide for southern bottomland hardwoods. Southern Journal of Applied Forestry 19, 103–104.

252

Chapter 6

Harper, J.L. (1977) Population Biology of Plants. Academic Press, London. Husch, B., Miller, C.I. and Beers, T.W. (1982) Forest Mensuration. Wiley, New York. Hutchings, M. (1983) Ecology’s law in search of a theory. New Scientist 16, 765–767. Kershaw, J.A., Jr and Fischer, B.C. (1991) A stand density management diagram for sawtimber-sized mixed upland central hardwoods. USDA Forest Service General Technical Report NE NE-148, pp. 414–428. Krajicek, J.E. (1967) Maximum use of minimum acres. Proceedings 9th Southern Forest Tree Improvement Conference, pp. 35–37. Krajicek, J.E., Brinkman, K.A. and Gingrich, S.F. (1961) Crown competition – a measure of density. Forest Science 7, 35–42. Leak, W.B. (1981) Do stocking guides in the Eastern United States relate to stand growth? Journal of Forestry 79, 661–664. Leary, R.A. and Stanﬁeld, D. (1986) Stocking guides made dynamic. Northern Journal of Applied Forestry 3, 139–142. Lonsdale, W.M. (1990) The self-thinning rule: dead or alive? Ecology 71, 1373–1388. McGill, D., Martin, J., Rogers, R. and Johnson, P.S. (1991) New stocking charts for northern red oak. University of Wisconsin Forestry Research Notes 277. McGill, D.W., Rogers, R., Martin, A.J. and Johnson, P.S. (1999) Measuring stocking in northern red oak stands in Wisconsin. Northern Journal of Applied Forestry 16, 144–150. McMahon, T. (1973) Size and shape in biology. Science 179, 1201–1204. McMahon, T.A. and Bonner, J.T. (1983) On Size and Life. Scientiﬁc American Books, New York. Miyanishi, K., Hoy, A.R. and Cavers, P.B. (1979) A generalized law of self-thinning in plant populations. Journal of Theoretical Biology 78, 439–442. Norberg, R.A. (1988) Theory of growth geometry of plants and self-thinning of plant populations: geometric similarity, elastic similarity, and different growth modes of plant parts. American Naturalist 131, 220–256. Nowak, C.A. (1996) Wood volume increment in thinned, 50- to 55-year-old, mixed-species Allegheny hardwoods. Canadian Journal of Forest Research 26, 819–835. Oswald, H. (1982) Silviculture of oak and beech high forests in France. Proceedings of Broadleaves in Britain, Future Management & Research Symposium, pp. 31–39. Putnam, J.A., Furnival, G.M. and McKnight, J.S. (1960) Management and inventory of southern hardwoods. USDA Forest Service Agriculture Handbook 181. Reineke, L.H. (1933) Perfecting a stand-density index for even-aged forests. Journal of Agricultural Research 46, 627–638. Roach, B.A. (1977) A stocking guide for Allegheny hardwoods and its use in controlling intermediate cuttings. USDA Forest Service Research Paper NE NE-373. Roberts, E.G. and Ross, R.D. (1965) Crown area of free-growing loblolly pine and its apparent independence of age and site. Journal of Forestry 63, 462–463. Rogers, R. (1983) Guides for thinning shortleaf pine. USDA Forest Service General Technical Report SE SE-24, pp. 217–225. Sampson, T.L. (1983) A stocking guide for northern red oak in New England. MS thesis, University of New Hampshire, Durham. Sampson, T.L., Barrett, J.P. and Leak, W.B. (1983) A stocking chart for northern red oak in New England. University of New Hampshire Agricultural Experiment Station Research Report 100. Schnur, G.L. (1937). Yield, stand, and volume tables for even-aged upland oak forests. USDA Technical Bulletin 560. Smith, H.C. and Gibbs, C.B. (1970) A guide to sugarbush stocking. USDA Forest Service Research Paper NE NE-171. Sonderman, D.L. (1985) Stand density – a factor affecting stem quality of young hardwoods. USDA Forest Service Research Paper NE NE-561. Sorrensen-Cothern, K.A., Ford, E.D. and Sprugel, D.G. (1993) A model of competition incorporating plasticity through modular foliage and crown development. Ecological Monographs 63, 277–304. Sprugel, D.G. (1984) Density, biomass, productivity, and nutrient-cycling changes during stand development in wave-regenerated balsam ﬁr forests. Ecological Monographs 54, 165–186. Stout, S.L. (1991) Stand density, stand structure, and species composition in transition oak stands of northwestern Pennsylvania. USDA Forest Service General Technical Report NE NE-148, pp. 194–206.

Self-thinning and Stand Density

253

Stout, S.L. and Larson, B.C. (1988) Relative stand density: why do we need to know? USDA Forest Service General Technical Report INT INT-243, pp. 73–79. Stout, S.L. and Nyland, R.D. (1986) Role of species composition in relative density measurement in Allegheny hardwoods. Canadian Journal of Forest Research 16, 574–579. Trimble, G.R. Jr (1968) Multiple stems and single stems of red oak give same site index. Journal of Forestry 66, 198. Vezina, P.E. (1963) More about the crown competition factor. Forestry Chronicle 39, 313–317. Walker, N. (1956) Growing stock volumes in unmanaged and managed forests. Journal of Forestry 54, 378–383. Weller, D.E. (1987a) A reevaluation of the 3/2 power rule of plant self-thinning. Ecological Monographs 57, 23–43. Weller, D.E. (1987b) Self-thinning exponent correlated with allometric measures of plant geometry. Ecology 68, 813–821. Weller, D.E. (1989) The interspeciﬁc size–density relationship among crowded plant stands and its implications for the 3/2 power rule of self-thinning. American Naturalist 133, 20–41. Westoby, M. (1984) The self-thinning rule. Advances in Ecological Research 14, 167–225. White, J. (1981) The allometric interpretation of the self-thinning rule. Journal of Theoretical Biology 89, 475–500. White, J. (1985) The thinning rule and its application to mixtures of plant populations. In: Studies on Plant Demography. Academic Press, New York, pp. 291–309. White, J. and Harper, J.L. (1970) Correlated changes in plant size and number in plant populations. Journal of Ecology 58, 467–485. Wilson, F.G. (1946) Numerical stocking in terms of height. Journal of Forestry 44, 758–761. Yoda, K., Kira, T., Ogawa, H. and Hozumi, K. (1963) Self-thinning in overcrowded pure stands under cultivated and natural conditions. Journal of Biology 14, 107–129. Zeide, B. (1985) Production of unmanaged bottomland hardwoods in Arkansas. Proceedings of the Central Hardwood Forest Conference V. University of Illinois, Urbana-Champaign, pp. 118–124. Zeide, B. (1987) Analysis of the 3/2 power law of self-thinning. Forest Science 33, 517–537. Zeide, B. (2001) Thinning and growth: a full turnaround. Journal of Forestry 99(1), 20–24. Zhang, L., Oswald, B.P., Green, T.H. and Stout, S.L. (1995) Relative density measurement and species composition in the mixed upland hardwood forests of North Alabama. USDA Forest Service General Technical Report SRS SRS-1, pp. 467–472.

7 Even-aged Silvicultural Methods

Introduction Two silvicultural systems of managing forests are generally recognized: even-aged and uneven-aged (Smith, 1986). This chapter focuses on the methods used in the even-aged silviculture of oak forests. The complete implementation of either system can lead to a regulated forest theoretically capable of sustaining to perpetuity an even ﬂow of timber products and other values. The traditional objective of even-aged management is to regulate a forest by managing the stands within it as a mosaic of different age classes. The trees in each stand are allowed to grow to a speciﬁc age called the rotation age. On reaching rotation age, a stand is renewed or regenerated by a ﬁnal harvest that requires the application of one of several even-aged regeneration methods. Even-aged forest management is said to be based on area control because it relies on regulating forest yield by creating and maintaining stands of various age classes, with each class occupying an approximately equal area of the forest. A regulated even-aged forest thus consists of a balanced distribution of stand age classes that is maintained through time. This arrangement also sustains an even and continuous ﬂow of wood products and a constant proportion of stands in each age class. Diameter frequency distributions of mature even-aged oak stands are often bell-shaped, i.e. normally distributed (Fig. 5.4). There also is growing interest in maintaining forests in 254

other speciﬁed conditions for social and ecological reasons. Such considerations may require a perspective different from that considered by silviculturists in the past. Although there are only two types of management systems, three types of tree age distributions have been recognized: (i) even-aged; (ii) two-aged; and (iii) unevenaged (Smith, 1986). Even-aged stands are deﬁned as those where the difference in age between the oldest and youngest trees does not exceed 20% of the rotation. Uneven-aged stands contain at least three age classes intermingled on the same area. Two-aged stands are comprised of two age classes of trees. In silvicultural practice, a new stand resulting from the ﬁnal harvest of an evenaged stand is usually considered even-aged regardless of the actual distribution of tree ages in the new stand. This convention requires a ﬂexible deﬁnition of the evenaged state. One deﬁnition employs the term cohort to refer to all the trees, arising anew or from advance (pre-established) reproduction, that originate from a silvicultural or natural event that produces a canopy gap or large opening in the forest (Oliver and Larson, 1996). Members of a cohort are considered even-aged regardless of their actual biological ages. This deﬁnition differs from that used in plant population biology, where cohort usually denotes membership in a group of plants originating from a single seed crop (Harper, 1977). The broader silvicultural deﬁnition facilitates reference to single- and multiple-

Even-aged Silvicultural Methods

cohort stands and thus many commonly encountered tree populations resulting from natural and silvicultural regeneration events (Fig. 5.13). Spatial scale also can be considered in determining whether a stand is even-aged or uneven-aged. For example, several cohorts of trees may occur within spatial scales smaller than that normally deﬁned as a stand. The deﬁnition of the even-aged state therefore may vary depending on the spatial scale that is relevant to management objectives. Deﬁning the uneven-aged state similarly depends on considerations of spatial scale (Chapter 8).

Natural Regeneration Methods Stands under even-aged management are regenerated at or near the end of the rotation by one of three silvicultural methods: (i) clearcutting; (ii) shelterwood; or (iii) seed tree methods. In their usual application, the parent stand is completely removed in one or more steps that eventually leave the site free of overstorey shade. Overstorey removal allows new reproduction to become established and the advance reproduction, if present, to develop in full light. All three methods can be used for regenerating a wide range of species from shade tolerant to shade intolerant. However, their use is often necessary for regenerating intolerant and mid-tolerant species such as the oaks. There also are variations within each method that provide additional ﬂexibility for attaining regeneration and other objectives. Although evenaged regeneration methods are applied at the end of the rotation (and thus the end of the life of the parent stand), their application is intended to initiate the new stand.

The clearcutting method Clearcutting probably is the easiest to apply and most economically efﬁcient of the regeneration methods. In its simplest application to oak forests, the method requires only the removal of the overstorey

255

to release the oak reproduction beneath it. The success of the method depends on preexisting (‘advance’) reproduction to replace the parent stand. New reproduction of nonoaks from seed stored in the forest ﬂoor, root sprouts and newly arriving wind-disseminated seeds also can become a part of the new stand. The method’s simplicity derives from the one-step removal of the overstorey. Its economic efﬁciency obtains from the minimization of logging costs per unit of tree volume harvested. However, for environmental and social reasons, the method also is the most controversial of the regeneration methods. Consequently, before the method is applied, several factors should be weighed, including: (i) the ecosystem-speciﬁc suitability of clearcutting for meeting oak regeneration requirements; (ii) predictability of the regeneration outcome; and (iii) economic, environmental and social considerations.

Suitability to oak regeneration requirements Clearcutting has been used successfully to regenerate oaks in many of the drier oak forests of the Central Hardwood Region and elsewhere where oak advance reproduction intrinsically accumulates (Roach and Gingrich, 1968; Sander, 1977; Johnson, 1993). In mature oaks stands (Fig. 7.1A), such accumulation may be sufﬁcient for replacing the parent stand after clearcutting. Regeneration success depends on a reproduction establishment period of a decade or more before the ﬁnal harvest is made (Sander, 1971, 1977). Where regeneration guidelines are available, stand regeneration potential can be objectively evaluated from an inventory of the overstorey (from which stump sprouts originate) and advance reproduction (Fig. 7.1B). Oak advance reproduction of the requisite size and spatial distribution must be present at the time of ﬁnal harvest if oaks are to become a major part of the next stand (Sander et al., 1984). If stand regeneration potential is deemed adequate, all trees ≥2 inches dbh should be cut if the management objective is exclusively timber production. However, clearcutting a stand without carefully considering its

256

Chapter 7

A

B

Fig. 7.1. (A) A mature even-aged mixed-oak stand in central Missouri (Central Hardwood Region). The stand is fully stocked and dominated by northern red, white and black oaks. Site index for red oak ranges from 65 to 70 ft at an index age of 50 years. Successfully regenerating such stands to oak by clearcutting depends on the oak regeneration potential. (B) An inventory of the oak advance reproduction coupled with the application of a regional regeneration guide (e.g. Dey et al., 1996b) provides an objective basis for predicting future stand composition and the utilization of growing space by oaks and other species. (USDA Forest Service, North Central Research Station photographs.)

regeneration potential may result in a new stand with few oaks. Where the oak regeneration potential is low, clearcutting will abruptly shift species composition from oak to non-oak. A common outcome is a mixture of less desirable species and poorly distributed stocking of oak stump sprouts (Fig. 7.2A). Cutting only trees with commercial value usually leaves stands of poor quality and undesirable species (Fig. 7.2B). The

residual trees are likely to capture much of the growing space at the expense of more desirable but smaller reproduction. Where management requires retaining overstorey trees to meet non-timber objectives, retained trees should be selected by silvicultural design rather than by the logger (Smith et al., 1989). When properly applied, the method is called clearcutting with reserves and creates, at least temporarily, a two-aged stand (Helms, 1998). A few

Even-aged Silvicultural Methods

257

A

B

Fig. 7.2. (A) Twenty years after clearcutting, this stand in southwestern Wisconsin (Central Hardwood Region) is largely comprised of northern red oak stump sprouts; site index for northern red oak is 65 ft. Stocking of oak stump sprouts is poorly distributed and the remaining stocking is poor quality black cherry and American elm. (B) A commercial clearcut 5 years after harvesting an oak stand in southeastern Ohio (Central Hardwood Region). Although the residual trees are of poor quality, their crowns and the growing space they occupy will expand and hinder the development of the established reproduction. Some residual trees can be retained in clearcuts for non-timber objectives, but it should be by design rather than ‘loggers choice’. (USDA Forest Service, North Central Research Station photographs.)

trees per acre can be retained for aesthetic purposes, den trees and snags for wildlife, and acorn production for wildlife. However, oaks retained as seed trees for regeneration purposes are usually ineffective after clearcutting.

Immediately after clearcutting, the site may temporarily appear to be inadequate in tree reproduction and vegetative cover (Fig. 7.3A). However, within a year, a dense new growth of herbaceous and woody species quickly occupy the site. Some of

258

Chapter 7

A

B

C

Fig. 7.3. (A) A clearcut oak stand in the Central Hardwood Region during the first year after harvest. New growth develops rapidly from advance reproduction of trees and shrubs, stump sprouts, seed stored in the forest floor and seed disseminated from surrounding vegetation. (B) Three years after clearcutting, this stand in the Ozark Highlands of Missouri (Central Hardwood Region) was well stocked with oak seedlings and seedling sprouts. (C) Twenty years after clearcutting, this stand in the Ozark Highlands was at 100% stocking and dominated by white, black and scarlet oaks. (USDA Forest Service, North Central Research Station photographs.)

Even-aged Silvicultural Methods

this emergent vegetation originates from plants and root systems present before harvest. Other plants originate from the seed bank in the forest ﬂoor and from windblown and animal-dispersed seed from adjacent sites. The sudden ﬂush of growth holds soil and nutrients in place even in steep terrain. Soil erosion and site deterioration after clearcutting are usually the result of improper logging road and skid trail design, which are largely avoidable (Chapter 4). Usually within 2 or 3 years, trees begin to dominate the site (Fig. 7.3B); by stand age 20, trees fully occupy the growing space (Fig. 7.3C). In the Ozark Highlands, 400–600 stems of oak reproduction per acre from 3 to 5 ft tall may provide adequate future stocking, depending on site factors (Sander et al., 1984). However, oak stump sprouts can compensate for deﬁciencies in stocking from advance reproduction. Stump sprouts may be especially important in previously unthinned stands where numerous smalldiameter white oaks often comprise a subordinate canopy layer. Previously thinned stands are therefore likely to produce few stump sprouts if thinning has been properly applied by concentrating removals in the overtopped and intermediate crown classes and rotation ages are 80 years or longer. Clearcutting is not an effective regeneration method in all xeric oak forests. For example, black oak and white oak stands on droughty outwash sands in northern Lower Michigan often fail to regenerate after clearcutting even when there is abundant oak advance reproduction (Johnson, 1992a). In that region, regeneration failures occur even where oak advance reproduction is ﬁve to ten times greater per unit area than under similar oak stands in the Ozark Highlands. The site index, composition, structure, stocking and yield of oak stands in both regions are similar. The failure of clearcutting to regenerate the Michigan forests appears to be related to: (i) the small size of the oak reproduction (mostly less than 1 foot tall); and (ii) its relatedly slow growth and high mortality after overstorey removal. The latter is associated with the post-clearcutting development of a dense

259

mat of sedges (Carex spp.), which in turn kills most of the established oak reproduction (Johnson, 1992a; Congdon, 1993). These failures suggest the need for regeneration prescriptions that increase light to levels sufﬁcient for the development of oak advance reproduction without stimulating the development of sedges. Regeneration failures related to clearcutting these northern oak forests also may be related to the removal of the frost-protecting overstorey canopy. In contrast, populations of oak advance reproduction in the Missouri forests, although of lower density than those in Michigan, include hundreds of large stems per acre that, after clearcutting, can rapidly capture growing space. Moreover, woody plants are the primary competitors of oak reproduction in the Ozark forests and late spring frosts occur there less frequently than in the northern forest. The contrasting regeneration ecology of these superﬁcially similar oak forests emphasizes the sensitivity of oak regeneration to factors that may not measurably affect growth and yield. Such differences in ecosystem reactions to clearcutting, and disturbances in general, point out the importance of distinguishing among ecologically deﬁned classes of oak forests (Chapter 1). In mesic oak forests, clearcutting usually fails to restore oaks to their preharvest level of importance (McGee and Hooper, 1970; Johnson, 1976; Beck and Hooper, 1986; Loftis, 1988; Stanturf et al., 1997; Jenkins and Parker, 1998). There, clearcutting often accelerates succession towards shade tolerant hardwoods such as sugar maple, red maple and American beech, or fast growing intolerant species such as yellow-poplar, white ash and black cherry (Abrams and Nowacki, 1992; Jenkins and Parker, 1998). Even when oak advance reproduction is abundant, it typically is suppressed after overstorey removal by the growth of non-oak stump sprouts and other competition (McGee and Hooper, 1970; Beck and Hooper, 1986). For example, 20 years after clearcutting an oak/yellow-poplar stand on a very productive site (yellowpoplar site index 100+ ft) in the southern Appalachians, the stand was dominated by

260

Chapter 7

A survey of 29 clearcut oak stands in the Ohio Valley 5–26 years after harvest showed that the amount of oak reproduction varied with time since clearcutting and site quality. The oak volume of the parent stands, a mixture of upland oaks (chestnut, white, black and scarlet oaks), comprised at least 60% of the total volume. On poor sites (oak site index 50–59 ft), dominant and codominant oaks became increasingly more

yellow-poplar and other non-oaks (Beck and Hooper, 1986). This conversion occurred even though there were more than 5000 stems per acre of oak advance reproduction (Fig. 7.4). Nevertheless, oaks sometimes succeed after clearcutting mesic forests if a large proportion of the overstorey oaks sprout (P.S. Johnson, 1975; Wendel, 1975), or where competition is not severe or is controlled (Johnson et al., 1989; Jacobs and Wray, 1992). 10,000

Trees per acre

A

Sweet birch

1,000

Other Yellow-poplar Oaks

100

Red maple

10 0

5

10

15

20

500 B

Trees per acre

400 300 200 Other Yellow-poplar

100

Red maple

0

Oaks Sweet birch

5

10

15

20

Stand age (years) Fig. 7.4. Twenty-year change in stand composition after clearcutting a cove hardwood stand in the southern Appalachian Mountains of North Carolina (Central Hardwood Region). (A) All reproduction ≥4.5 ft tall. (B) Free-to-grow (not overtopped) reproduction. ‘Other’ species include black locust, eastern hemlock, white ash, black cherry, hickories, American basswood, blackgum, sassafras, sourwood and flowering dogwood. The oak group includes northern red, black, chestnut and white oaks. The parent stand was 53% oak and 33% yellow-poplar by volume. (From Beck and Hooper, 1986.)

Even-aged Silvicultural Methods

abundant with stand age and accounted for 64% of all stems in stands 15 years old and older (Hilt, 1985a). On medium sites (oak site index 60–69 ft), oaks attained moderate importance after clearcutting and changed little with increasing stand age. On good sites (oak site index 70–80 ft), dominant and co-dominant oaks decreased with time and accounted for only 11% of stems in stands 15 years old and older, by which time yellow-poplar, black cherry and ash dominated the sites (Fig. 7.5). In this region, the compositional outcome is largely determined by the interaction of time and site quality: oaks ultimately emerge as dominants within two decades on poor sites (Hilt, 1985b). Yellow-poplar and other site-demanding hardwoods ultimately emerge as dominants on the more productive sites. On sites of intermediate quality, oaks often regenerate to intermediate levels of importance. The compositional and structural changes occurring after clearcutting are expressions of secondary succession. Each

silvicultural ‘outcome’ represents a point along one of several possible trajectories or successional pathways. Poor (xeric) sites typically possess few pathways (i.e. future compositional and structural possibilities) because many species, especially those that are sensitive to soil moisture stress, are excluded from those ecosystems by their failure to initially colonize and/or by high mortality rates. In mesic ecosystems, successional pathways are more numerous because virtually all species from those adapted to dry sites (xerophytes) to those that are more moisture-demanding (mesophytes) are physiologically capable of surviving and growing there. It is thus the competition environment that largely excludes oaks from the mesic ecosystems. Conversely, the moisture-demanding nonoaks usually are excluded from xeric ecosystems (Wuenscher and Kozlowski, 1971; Abrams, 1992). However, the common presence of oaks in mesic ecosystems suggests that certain kinds of disturbances can favour the oaks. Based on the outcome

70

Oaks as a percentage of all trees

261

Site index 50–59

60 50 40 30

60–69

20 70–80

10 0 5–9

10–14 Stand age-class (years)

15+

Fig. 7.5. Change in the relative proportion of oaks in relation to site index in mixed oak stands in southeastern Ohio (Central Hardwood Region). The oaks include white, black, scarlet and chestnut oaks. Other species present include red and sugar maples, black cherry, white ash, hickories, bigtooth aspen, flowering dogwood, sassafras and eastern hophornbeam. The maples and yellow-poplar dominated stands on the two better site classes by stand age 15. (Adapted from Hilt, 1985a.)

262

Chapter 7

of numerous reported trials, it is nevertheless apparent that clearcutting, by itself, often fails to maintain oaks at preharvest levels of stocking in mesic ecosystems. Despite its importance, the presumed requirement for the presence of oak advance reproduction may not be an irrevocable rule. This may be especially true in some mesic and hydric ecosystems. Silvicultural alternatives to relying on advance reproduction have been described for northern red oak in the Driftless Area of southwestern Wisconsin and adjacent states (Johnson et al., 1989; Bundy et al., 1991; Jacobs and Wray, 1992). There, the site index of red oak in the more mesic ecosystems ranges from about 65 to 70 ft (Gevorkiantz and Scholz, 1948; Gevorkiantz, 1957). On those sites, thousands of red oak seedlings per acre may become established in a single year following a good acorn crop (Scholz, 1955; Johnson, 1974). Numbers of seedlings also can be increased by mechanical soil scariﬁcation timed to coincide with an acorn crop. Scariﬁcation effectively reduces competition (Bundy et al., 1991). It also facilitates direct contact between germinating acorns and a mineral seedbed, which can increase initial seedling establishment if done before leaf fall. Scariﬁcation in combination with direct-seeded northern red oak acorns produced similar increases in seedling establishment under a shelterwood in Pennsylvania (Zaczek et al., 1997). After overstorey removal, the growth of red oak reproduction on productive sites is potentially rapid and less dependent on the development of a large root system and a long establishment period than is the case for the more xerophytic oaks. Thus, large numbers of red oak seedlings and a favourable postharvest environment can combine to effect successful regeneration. However, unlike many other mesic ecosystems, the density of competing trees in Driftless Area stands is relatively low and often comprised of species such as black cherry, paper birch, elms and hickories (Johnson, 1976; Martin et al., 1992). After clearcutting, the rapid decline of non-oaks from the main canopy is similar to that

which occurs in young red oak stands in New England (Oliver, 1978; Hibbs and Bentley, 1984). Soil scariﬁcation has been used in the Driftless Area to effectively reduce competition and simultaneously prepare a patchy seedbed of mineral soil (Bundy et al., 1991; Jacobs and Wray, 1992). One technique uses a tractor-mounted blade set 6 inches above ground. Careful application of the treatment during a good acorn crop results in mechanical breakage of low vegetation, including tree reproduction and shrubs, while simultaneously effecting only moderate soil disturbance. In one trial, the method increased the height and leaf area of red oak reproduction two growing seasons after treatment (1 year after clearcutting) (Bundy et al., 1991). A similar treatment consisting of mechanical uprooting of low vegetation during logging, or in a separate operation before or after logging, also has been proposed (Jacobs and Wray, 1992). Experience has shown that oak advance reproduction survives the treatment because of its deep taproot and capacity for sprouting after top injury. The recommended time to apply the treatment is in the autumn during a good seed year before acorns drop. Timing the treatment with acorn fall can increase numbers of seedlings as well as reduce competition. Applying herbicides to the understorey before clearcutting and concurrently with a good acorn crop also can facilitate oak regeneration. Even though the herbicide kills some of the oak advance reproduction, one trial resulted in about 1000 red oak stems per acre 0.5 inch dbh and larger 11 years after clearcutting; about 27% were co-dominant or larger (Johnson et al., 1989). Strategies for regenerating northern red oak in the Driftless Area are discussed in more detail by Jacobs and Wray (1992). In New England and the Lake States, northern red oak sometimes regenerates under pine stands if there is a nearby red oak seed source (Cline and Lockard, 1925; Crow and Isebrands, 1986). When the pine is harvested, the oak advance reproduction then may capture the site (Sampson et al., 1983). Such pine-to-oak successions may

Even-aged Silvicultural Methods

be facilitated by the dispersal of acorns by animals. Small mammals disperse red oak acorns up to about 60 ft (Sork, 1984) while blue jays disperse them up to 2.5 miles (Fig. 2.13). Although most of the dispersed acorns are consumed by animals, even the small surviving proportion of blue jay-dispersed acorns can produce signiﬁcant numbers of oak seedlings because of the large numbers of acorns that are dispersed and the favourable germination and growth environments they are carried to (Johnson and Webb, 1989). In bottomland oak forests, success in regenerating oaks by clearcutting has been mixed, ranging from successful to unsuccessful (e.g. Bowling and Kellison, 1983; Gresham, 1985a,b). There, prediction of stand replacement after clearcutting may be confounded by ﬂooding, which may affect oak and non-oak advance reproduction differently. Physiological tolerance of bottomland oaks to inundation varies among species and tolerance is dependent on length of the ﬂood period, depth of inundation, and other factors (McKevlin, 1992). Several bottomland non-oaks such as water hickory, water tupelo, swamp tupelo, green ash and red maple are more physiologically tolerant of inundation than other oaks (Hosner and Minckler, 1960, 1963; Hosner and Boyce, 1962; Broadfoot and Williston, 1973; McKnight et al., 1981; McKevlin, 1992). However, ﬂoods of short to moderate duration limited to the dormant season may favour oak reproduction. For example, water oak seedlings in an East Texas ﬂoodplain survived winter and early spring ﬂooding because, unlike their shallow-rooted competitors, some of the oak seedlings germinated after ﬂooding, survived physical damage, resisted uprooting during ﬂooding, and resprouted when damaged (Streng et al., 1989). Consequently, oaks may comprise a relatively large proportion of the older, and thus larger, pool of advance reproduction in bottomlands because of their deep roots and relatively high survival and sprouting rates (Streng et al., 1989) (Fig. 3.1). Like the upland oaks, it is the larger advance reproduction that is

263

most likely to capture the growing space after clearcutting. The outcome largely depends on the species composition and size distribution of the advance reproduction. Different reactions to ﬂooding among species and the mix of species present at the time of ﬂooding thus may largely determine the regeneration outcome in bottomland oak forests. In some cases, bottomland oaks have successfully regenerated from seedlings established after clearcutting (Golden and Loewenstein, 1991; Golden, 1993; Nix and Lafaye, 1993). In two bottomland sites in South Carolina, cherrybark, shumard and water oak seedlings established after clearcutting resulted in over 500 oaks per acre that attained canopy dominance 5 years after clearcutting (Nix and Lafaye, 1993). Success was attributed to a moderately good acorn crop followed by early winter logging in wet weather, which in turn scariﬁed the soil and buried many acorns. Despite such successes, it would be silviculturally more prudent to rely on oak advance reproduction of the requisite size and number. One guideline for regenerating bottomland hardwoods deems clearcutting a viable option only if 200–500 seedlings per acre of desirable species are present as advance reproduction (McKevlin, 1992). The composition and structure of bottomland clearcuts at the end of the ﬁrst decade after cutting may not reﬂect the potential of oaks to ultimately dominate stands. For example, water and swamp chestnut oaks collectively increased from 32 to 44% of total stand basal area from stand age 14–22 years in a bottomland forest in Mississippi. This gain was made at the expense of sweetgum, ironwood, pine, blackgum and other species, whose collective basal areas remained relatively unchanged, but whose survival rates were lower than the oaks during that period (Fig. 7.6). As a result, the oaks emerged as the major dominant species by stand age 22. The emergence of bottomland oaks from inferior to dominant crown classes is related to spacing between oaks and competitors. When cherrybark oak com-

264

Chapter 7 60 Basal area (ft2 acre–1)

A 50

Oaks

40 Sweetgum

30 20

Ironwood Pine

10

Other Blackgum

0 14

19

22

1.00 B

Survival probability

0.95 Oaks (k = 0.977)

0.90 0.85

Other hardwoods (k = 0.959)

0.80 0.75 0.70 14

16

18

20

22

Stand age (years) Fig. 7.6. Stand development after clearcutting a bottomland oak/mixed hardwood stand in east-central Mississippi (Southern Pine–Hardwood Region). (A) Change in basal area with stand age. The oaks are predominantly water oak but include swamp chestnut oak; pine includes loblolly and spruce pines; other includes magnolia, elms, hickory and red maple. (B) Survival of oaks compared to the average of all other hardwoods. The curves are based on the negative exponential rates (k) calculated from observed 22nd-year survival (authors’ analysis). Actual survival between stand ages 14 and 22 was not observed. (Adapted from Bowling and Kellison, 1983, by permission of Society of American Foresters, Bethesda, Maryland. Not for further reproduction.)

petes with sweetgum in Mississippi lowlands, height growth and thus the outcome of competition between the two species depends on the average spacing between dominant/co-dominant trees (Clatterbuck and Hodges, 1988). When spacing between is less than 18 ft, sweetgum initially grows faster than oak. However, by stand age 32, the oaks are signiﬁcantly taller than the sweetgum under these competitive conditions. In contrast, where the spacing between dominant and co-dominant competitors

is greater than about 18 ft, the height growth of oak is similar to that of competitors (Fig. 5.10B). The faster growth of oak under the more competitive (‘restricted’) condition also resulted in greater average clear bole length of oak at stand age 40 (40 ft) than in the unrestricted mode (23 ft). The ultimate height growth advantage of oak in the restricted mode occurred even though the oaks were 3–6 ft shorter than sweetgum during the brushy stand initiation stage of stand development.

Even-aged Silvicultural Methods

Because site quality and competition vary greatly among different kinds of oak forests, ecological classiﬁcation also is a potential silvicultural tool for assessing which stands are unlikely to regenerate. Deﬁned ecological classes of forests can be used to distinguish among oak stands that appear to be similar (e.g. similar cover types) but that behave differently in their regeneration characteristics. But even within a deﬁned ecological class, it is desirable to have a more speciﬁc basis for assessing an individual stand’s regeneration potential before clearcutting is applied because of the temporal and spatial variation in reproduction density and size. So how can the required reproduction characteristics be identiﬁed given that a stand’s regeneration potential is not realized until after the overstorey is harvested? Although experienced silviculturists may be able to assess the regeneration outcome from visual examination of the advance reproduction and overstorey, more objective methods are often desirable.

Regeneration models One method for objectively assessing regeneration potential involves the use of predictive regeneration models. Among the models applicable to clearcutting oak forests are those developed for the Ozark Highlands (Sander et al., 1976, 1984; Johnson and Sander, 1987; Dey, 1991; Dey et al., 1996b). The region is transitional to the Great Plains and includes the southern half of Missouri and extends into northern Arkansas and northeastern Oklahoma (McNab and Avers, 1994). The upland forests there are typically dominated by various combinations of black, white, scarlet, northern red, southern red, post, and blackjack oaks, and other hardwoods that are sometimes mixed with shortleaf pine and eastern redcedar (Braun, 1972). Site index for black oak and scarlet oak ranges from about 40 to 80 ft at an index age of 50 years (McQuilkin, 1974). The predominant associated hardwoods include hickories, sassafras, blackgum and ﬂowering dogwood. Although shortleaf pine and eastern

265

redcedar are frequently absent from individual stands, the associated hardwoods are usually present. In these relatively dry forests, oak reproduction typically accumulates beneath the canopy of the parent stand for several decades (Liming and Johnston, 1944). There, the shoots of oak reproduction repeatedly die back and the roots of survivors slowly attain large size. Typical of ecosystems where oak reproduction accumulates, the oak advance reproduction remains largely in a suppressed state until disturbance substantially reduces overstorey density. The reproduction present at the time of disturbance is an important component of the initial state of the new stand because of its potential for capturing growing space after disturbance. The oaks of the Ozark Highlands are ‘persistent’ (sensu Veblen, 1992) because their presence is sustained over successive generations. This persistence contrasts with the more mesic oak forests to the north and east and with the bottomland oak forests to the south where dominance by oaks often may last only one generation because of successional replacement by long-lived, faster growing or more shade tolerant hardwoods (R.L. Johnson, 1975; Loftis, 1990b; Nowacki et al., 1990). In contrast, the dominance of the hardwoods associated with oaks in the Ozark Highlands usually lasts for only two decades after overstorey removal. The high mortality rates and limited development of the non-oaks in this ecosystem relegates them to saplings and reproduction in older stands (Braun, 1972; Dey, 1991). Even though these forests are often called oak–hickory forests, the hickories usually comprise a minor part of the overstorey. During the ﬁrst decade after overstorey removal, the composition of the reproduction largely depends on the mix of seed and plants present when the disturbance occurs. Although this initial mix itself is relatively unpredictable, the non-oaks are quickly relegated to the sub-canopy as competition intensiﬁes and crowns close. Within two decades, the oaks have reemerged as the dominant species. This pattern of stand development conforms to the

266

Chapter 7

‘competitive sorting’ model described by Margalef (1963, 1968) and Peet (1992). However, in the Ozark Highlands the outcome is inﬂuenced by topography. The ascendence of oaks to dominance by competitive sorting is rapid and predictable on hot southwest-facing slopes and on neutral southeast and northwest slopes. It is less certain on cool northeast slopes where the pre-disturbance accumulation of oak reproduction is less pronounced and shade tolerant competitors are sometimes capable of replacing the oaks or slowing their re-emergence to dominance (Sander et al., 1984). The outcome is probabilistically speciﬁable for each species and depends on their size at the time of clearcutting (Fig. 7.7). Even though the non-oaks of the Ozark Highlands are usually unimportant as dominant components of older stands, they nevertheless persist from one generation to the next. Their abundance and early rapid growth after clearcutting also interfere with the re-emergence of the oaks. The oaks and associated hardwoods nevertheless re-establish a ‘compositional equilibrium’ (sensu Veblen, 1992) within two decades of disturbance. The consistency of this pattern of stand redevelopment simpliﬁes the prediction of changes in species composition and therefore modelling the regeneration process. In the Ozark Highlands, predicting changes in species composition after clearcutting thus is less a question of losing the oaks through successional displacement than it is of how the proportions of the various oak species will change and how quickly and completely the oaks will capture growing space in the regenerated stand. ADVREGEN is a probabilistic individual-tree regeneration model applicable to the Ozark Highlands. It was developed to assess the adequacy of the oak, hickory and blackgum regeneration potential in stands considered for harvesting by clearcutting (Sander et al., 1984; Johnson and Sander, 1987). However, the model is applicable to any method of regeneration requiring complete overstorey removal on areas of about one-third of an acre or larger. The model provides a simple ‘yes’ or ‘no’ answer to

the question of whether a given stand will be adequately stocked 20 years after ﬁnal harvest. The criterion for deﬁning minimum adequacy of stocking assumes that by stand age 20 there must be at least 221 dominant and co-dominant oaks per acre averaging 4.5 inches dbh (the average expected dbh of dominant and co-dominant trees in 20-year-old stands). That level of stocking equates to C-level stocking on Gingrich’s (1967) stocking chart (Fig. 6.9). The trees in stands at C-level stocking will fully utilize growing space within 10 years (i.e. by about stand age 30) assuming that trees are well distributed (Gingrich, 1967; see also Chapter 6). Application of the model is relatively simple and is facilitated by the ADVREGEN computer program or by referring to related tables (Sander et al., 1984). To use the model, the height and basal diameter of the largest stem of advance reproduction must be measured on 1/735-acre (4.3-ft radius) ﬁeld plots and the slope position and aspect of each plot must be speciﬁed. A 1/735-acre plot equals the minimum growing space required for a tree 4.5 inches dbh as deﬁned by Gingrich (1967). Co-dominant and dominant trees averaging 4.5 inches dbh occur in stands averaging about 3 inches dbh (based on trees 1.6 inches dbh and larger). Because 3 inches is the smallest mean stand diameter shown on Gingrich’s (1967) stocking chart, it represents the earliest point in stand development (approximately stand age 20 on average sites) that the utilization of growing space by trees (stocking) can be practically determined. ADVREGEN also facilitates estimating the probable contribution of stump sprouts to future stocking. To do that, the diameters of overstorey oaks must be sampled and stand site index must be determined. The regeneration model is based on estimates of the probability that a seedling or seedling sprout of a given initial size (basal diameter and height) will survive and grow to dominant or co-dominant crown class after clearcutting. For a given species, large initial diameters and heights are associated with high probabilities. The

Even-aged Silvicultural Methods

267

A

Fig. 7.7. (A) Large oak advance reproduction such as this 7-ft tall seedling sprout in a Central Hardwood stand has a high dominance probability, i.e. a high chance of becoming a dominant or co-dominant tree after complete overstorey removal. (USDA Forest Service, North Central Research Station photograph.) (B) Dominance probabilities for advance reproduction (seedling or seedling sprout) of different species in the Ozark Highlands of Missouri (Central Hardwood Region). In these examples, dominance probability is the probability that a tree will attain an intermediate-orlarger crown class 21 years after clearcutting. In each case, probabilities are for 6-ft-tall advance reproduction with 1-inch basal diameters growing on neutral aspects (southeast- or northwestfacing) mid-slopes. The five species groups shown are the predominant hardwoods within this ecoregion. Probabilities were generated by the regeneration model ACORn (Dey et al., 1996b). 0.6

B

Dominance probability

0.5

0.4

0.3

0.2

0.1

0.0 Hickory

Sassafras

Blackgum

resulting estimates, called dominance probabilities, are generated by a predictive equation (Fig. 7.8). Dominance probabilities are extended to stand age 20 by assuming an annual negative exponential ‘retainment’ rate of 0.99 (Johnson and Sander, 1987). A ‘stocking value’ for each

Dogwood

Oaks

plot is then calculated from the reciprocal of the 20-year dominance probability and the binomial probability distribution (Sander et al., 1984; Johnson and Sander, 1987). If a stand’s mean stocking value equals or exceeds 30%, the stand is deemed adequately stocked.

268

Chapter 7

0.8 a

0.6 0.5 0.4 0.3 1.75 1.50

0.2 b

0.1

1.25

Ba sa ld iam et er (in ch es )

Dominance probability

0.7

1.00

0.0 5

0.75

4

3

2

Heigh

t (ft)

0.50 1

0 0.25

Fig. 7.8. Estimated dominance probabilities for oak advance reproduction in relation to initial (preharvest) size of reproduction in the Ozark Highlands of Missouri (Central Hardwood Region). Dominance probability is here defined as the probability that seedling or seedling sprout of a given height and basal diameter immediately before clearcutting will be dominant or co-dominant 5 years later. (a) Probabilities for mid-slopes on neutral aspects (southeast- and northwest-facing slopes); (b) Probabilities for lower slopes on cool aspects (northeast-facing slopes). Probabilities for other aspect/slope combinations lie between the two response surfaces shown, which represent averages for black, white, scarlet, post and blackjack oaks. (Derived from a logistic regression model from Sander et al., 1984.)

The ADVREGEN model is simple in concept and application because the deﬁnition of a dominance probability integrates growth and survival into a single value. The speciﬁed size of the inventory plot, which considers the future growing space requirements of trees, also simpliﬁes prediction and ﬁeld application. Although the model does not explicitly consider competition effects, such effects are implicit in the effects of topographic factors (advance reproduction) or site index (stump sprouts) on dominance probabilities. The limitations of ADVREGEN include the lack of speciﬁcation of the composition and structure of the future stand, the use of a single stocking criterion, restriction to stands where there will be complete overstorey removal (i.e. where all trees 2 inches dbh and larger will be cut), regeneration stock-

ing estimates that are limited to oaks, hickories and blackgum, and a lack of generality of application to conditions not considered by the data from which the model was built. Model development and application are discussed in more detail elsewhere (Sander et al., 1984; Johnson and Sander, 1987). A similar regeneration model applicable to northern red oak reproduction in the southern Appalachians predicts 20th-year dominance probabilities from preharvest basal (ground line) diameter and site index. Probabilities increase with increasing basal diameter and decreasing site index across the site index range of 70–90 ft (Table 7.1). The smaller dominance probabilities associated with the higher site indices are attributable to the increasing competition from yellow-poplar and other vegetation

Even-aged Silvicultural Methods

269

Table 7.1. Twentieth-year dominance probabilitiesa for northern red oak advance reproduction in the southern Appalachians. Basal diameterb of advance reproduction (inches) 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1–1.5 1.6–2.0

Oak site indexc (probability) 70 0.01 0.02 0.04 0.06 0.09 0.13 0.17 0.21 0.25 0.29 0.38 0.46

80 0 0.01 0.02 0.03 0.04 0.06 0.09 0.12 0.15 0.18 0.29 0.41

90 0 0 0.01 0.01 0.02 0.03 0.04 0.06 0.08 0.11 0.19 0.34

aThe

probability that a stem of red oak advance reproduction will be dominant or co-dominant 20 years after overstorey removal. (From Loftis, 1990a.) bMeasured at ground line. cHeight in feet at an index age of 50 years based on the curves of Olson (1959).

associated with increasing site quality. In application, the expected future numbers of dominant and co-dominant red oaks per unit area can be calculated by multiplying the observed number of seedlings in each of several basal diameter classes by the dominance probability for each class and summing those products across all diameter classes. Because northern red oak is usually a minor to moderately important component of the mixed mesophytic forests of the region, the model itself makes no assumptions about the adequacy of future red oak stocking. A model called ACORn (A Comprehensive Ozark Regenerator), like ADVREGEN, is applicable to the Ozark Highlands. It can be used to predict the composition and structure of stands 21 years after complete overstorey removal (Dey et al., 1996a,b). This probabilistic individual-tree, distance-independent model simulates the development of oak stands including all the oaks native to the Ozarks and the four major associated hardwoods (hickories, blackgum, ﬂowering dogwood

and sassafras). ACORn makes no assumptions about the stocking adequacy of the future stand. Instead, the model predicts the future distribution of tree diameters by species (Fig. 7.9). The output from ACORn can be used as input for existing growth and yield models. This facilitates projecting stand growth and change through the next rotation. Like ADVREGEN, the application of ACORn requires an inventory of the advance reproduction and overstorey together with information on site quality as expressed by slope position, aspect and site index. Because of its complexity, using the model requires a computer. The model can be applied to ‘real’ stands or used to compare various hypothetical situations. The ACORn model assumes that the height growth of reproduction, including advance reproduction and stump sprouts, is related to the apparent size of the root system rather than reproduction origin, per se. Accordingly, all forms of reproduction including new seedlings, seedling sprouts and stump sprouts represent a continuum of growth potential determined by root

Chapter 7

24

A

Data input: overstorey

20

Trees per acre

White oak 16 Black oak 12 Other species 8 4 0

Advance reproduction (genets/acre)

2.5

4.5

6.5

8.5 10.5 12.5 14.5 Dbh class (inches)

16.5

18.5

B

Data input: advance reproduction

500 400 300 200 100 0

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9

1

1.1 1.2

Basal diameter class (inches)

24

C

400

Model output: regenerated stand at age 21

2* Trees per acre

270

300

8

*Stocking % in dbh class

200 15 100

22 15 4

3

1

6.5

7.5

8.5

0 0.5

1.5

2.5

3.5 4.5 5.5 Dbh class (inches)

Even-aged Silvicultural Methods

size, which in turn is correlated with basal tree diameter. Rate of height growth at ﬁrst increases rapidly with increasing initial basal diameter, reaches a maximum, and then decreases (Fig. 2.27). For oak advance reproduction, initial height also is considered by the model. However, for various physiological and ecological reasons, initial size relations explain only a small to moderate proportion of the variance in the postharvest growth of reproduction. To produce realistic simulations of the size distributions of trees, ACORn jointly employs the regression estimates of future tree heights and the observed variation around those estimates (Dey, 1991). The prediction errors are used to estimate the probability that a seedling, seedling sprout or a stump sprout will attain a speciﬁed 21st-year diameter based on their initial size and site-related factors. These probabilities then are reduced by multiplying them by estimated survival probabilities. The resulting individual-tree probabilities facilitate predicting future diameter distributions of trees that survive and grow to stand age 21 based on a preharvest inventory of the advance reproduction and the overstorey. ACORn is limited to predicting stand composition and structure at stand age 21, and like ADVREGEN is limited to predicting the outcome after complete overstorey

271

removal. Also, contributions to future stocking from reproduction originating from seed cannot be accounted for. The latter eliminates prediction of two regionally important species, eastern redcedar and shortleaf pine. FORCAT is a regeneration model designed to simulate the development of clearcut stands on the Cumberland Plateau (Waldrop et al., 1986). The Cumberland Plateau is centred in eastern Tennessee and extends northwards into eastern Kentucky and southwards into northern Alabama (Smalley, 1979, 1982, 1984, 1986). This is a ﬂoristically diverse region that includes forests dominated by oaks, pines and numerous other hardwoods. Site index for upland oaks ranges from about 50 to 80 ft at an index age of 50 years. Variation in the region’s undulating to steep topography, together with soil depth and soil texture, largely controls site quality and therefore species composition, succession and regeneration dynamics. The drier sites, which are characterized by southerly-facing slopes, dry ridges and thin soils, are usually dominated by oaks and hickories sometimes mixed with shortleaf pine, Virginia pine and eastern redcedar. Common associated hardwoods include blackgum, red maple, sourwood, ﬂowering dogwood and sassafras (Waldrop et al., 1986). On the drier sites, oak reproduction

Fig. 7.9. (Opposite) Graphical representation of input and output for the regeneration model ACORn (A Comprehensive Ozark Regenerator) (Dey et al., 1996b). The model predicts, from a preharvest stand inventory (tree list), the 21st-year composition, structure and stocking of oak stands that are proposed for complete overstorey removal in the Ozark Highlands of Missouri (Central Hardwood Region). (A) Diameter distribution of overstorey trees. ACORn requires an inventory of the diameters of overstorey trees (listed by species) and an estimate of stand site index. This facilitates predicting contributions to the new stand from stump sprouts originating from cut overstorey trees. The graph represents a mature black oak/white oak stand at 84% stocking (based on Gingrich, 1967) growing on a southeast-facing slope (black oak site index 63). The stand is dominated by black oak. (B) Distribution of basal diameters of advance (preharvest) reproduction. The model requires an inventory of the advance reproduction (a list of basal diameters and heights by species) to predict its contribution to the future stand. Information on the slope position and aspect of each reproduction sample plot or the stand as a whole also is required. In this example, white oak reproduction predominates. (C) Model output: the regenerated stand. Based on input from A and B, the model predicts the new stand will be dominated by white oak, which contributes 72% stocking of the total projected 94% at stand age 21. Other species include hickories, blackgum, sassafras and flowering dogwood. The computer model produces graphical output similar to that shown in (C) as well as tables. Output can be obtained by crown classes, source of reproduction (stump sprout or advance reproduction) and user-specified species or species-groups. (Adapted from Dey et al., 1996b.)

272

Chapter 7

often naturally accumulates. Stand structure, composition and succession therefore are similar to the Ozark Highlands and other xeric oak forests of the Central Hardwood Region. On the richer sites on north- and eastfacing slopes, valley bottoms and coves, stands are often dominated by diverse mixtures of tolerant and intolerant mesophytes such as yellow-poplar, sugar maple, American beech, black cherry and white ash mixed with minor to moderate proportions of northern red, white and other oaks (Braun, 1972). Lush subcanopies of ﬂowering dogwood, sourwood, bigleaf magnolia, umbrella magnolia, American hornbeam and other shade tolerant species are common (Carpenter, 1976). Because oak advance reproduction usually does not accumulate under those conditions, regenerating oaks is problematic on these sites. FORCAT predicts future stand composition and structure from measurements of the initial state, i.e. dbh, height and species composition of the parent stand overstorey. The model can consider up to 33 tree species and can simulate the composition and structure of stands of any age up to 100 years. The model user also can begin the simulation at any stand age beginning with the preharvest state or from subsequent years. Unlike ACORn, FORCAT does not directly consider the presence of advance reproduction. Therefore, in application, FORCAT does not require measurements or counts of reproduction. Regeneration events are simulated by various model subroutines that probabilistically account for sprouting from harvested overstorey trees, seed production and germination. Regional variation in temperature and soil moisture effects are considered by constraining the maximum growth potential of each species based on growing-season degree days and soil moisture days above speciﬁed critical thresholds. Site-speciﬁc effects are considered by providing separate values of a diameter growth constant for each species in each of 20 different ‘land types’ in the MidCumberland Plateau as deﬁned by Smalley

(1982). Tree survival is predicted from tree age and expected diameter growth rate. The model also can simulate the effects of prescribed burning. In tests of FORCAT’s precision, predictions became increasingly more accurate with increasing stand age (up to 100 years). Its questionable ability to accurately predict the composition of young stands (e.g. at age 20) may be silviculturally problematic. FORCAT’s generalized structure makes it adaptable to a wide range of species, site and ecological conditions like those characteristic of the Cumberland Plateau. This ﬂexibility is derived from the model’s lineage to FORET, a gap model designed to predict long-term forest succession in Tennessee (Shugart and West, 1977). Another regeneration model covers the Allegheny Hardwood Region from the Allegheny Plateau of northwestern Pennsylvania (ecoregion province 211, Fig. 1.2) northwards into New York where it is transitional to the Northern Hardwood Region (ecoregion province 221a, Fig. 1.2) (Marquis et al., 1992). To the south, the applicable area extends into West Virginia and Maryland (ecoregion provinces 221a and M221, Fig. 1.2). Complete silvicultural guidelines, including the regeneration model are available for three commonly occurring hardwood forest types within the region: cherry–maple, beech–birch– maple and oak–hickory (Marquis et al., 1992). The latter occur primarily on the drier sites. However, one or more species of oak (including northern red, white, black, scarlet and chestnut oaks) are potential components of all three types. Important non-oaks include black cherry, red maple, sugar maple, American beech, yellow birch, sweet birch, hickories, white ash, yellow-poplar, pin cherry, striped maple eastern hophornbeam and other species (Marquis et al., 1975). Serious regeneration problems often occur throughout the region because of heavy deer browsing and competition from ferns, grasses and shade tolerant sub-canopy tree species (Marquis, 1974, 1981; Horsley, 1982; Horsley and Marquis, 1983).

Even-aged Silvicultural Methods

Comprehensive guidelines are available for prescribing silvicultural treatments for stands within the region (Marquis et al., 1992). Application of the guidelines is facilitated by a computer program, SILVAH, which incorporates criteria for evaluating stand regeneration potential. SILVAH also produces printed output containing data summaries and stand diagnostics. The regeneration model included in SILVAH is applicable to stands under even-aged management. Using the model requires preharvest counts of reproduction by species groups within sample plots of 6 ft radius. The counts are weighted by height and vigour classes and adjusted for interfering factors including intensity of deer browsing, soil characteristics, and competition from non-commercial tree species and other plants. The resulting rating then is used to determine whether a reproduction plot is stocked (Marquis et al., 1992). A stand’s regeneration potential is deemed adequate if at least 70% of plots are rated as stocked. Regeneration from sapling-size trees (0.5–6 inches dbh) also can be considered. Three commercially valuable species groups are recognized: black cherry, oaks and other ‘desirable’ species. Other tree species are categorized as ‘undesirable’ and are treated by the model as an ‘interfering’ factor in the regeneration of the commercially valuable species. There also is a provision for evaluating regeneration after ﬁnal harvest. Regeneration adequacy can be projected by using the SILVAH program or by hand cal-

273

culation using the tables and decision rules in Marquis et al. (1992). The regeneration model was developed from the relation between the observed density of advance reproduction before ﬁnal harvest and stand stocking after harvest (Grisez and Peace, 1973; Marquis and Bjorkbom, 1982). The model emphasizes black cherry because of its regional importance and considers other species, e.g. oaks and other ‘desirable’ species, as rather broad species groups. Stocking criteria for oaks were largely adapted from those developed in other regions and modiﬁed by general observations on oak regeneration within the Allegheny region (Marquis et al., 1992). Because the model is in part statistically derived, it possesses some of the characteristics of other statistically based models as discussed above. However, the model differs from the models previously discussed because of its partly statistical, partly expert opinion design. In southern bottomland oak stands, regeneration potential can be assessed by assigning a score of 1–3 to trees (both advance reproduction and overstorey) within 1/100-acre plots (Johnson and Deen, 1993). Small advance reproduction is assigned a value of 1 whereas larger reproduction is assigned a value of 3; conversely, small overstorey trees are assigned a value of 3 and larger trees a value of 1 or 0, which reﬂects their stump sprouting potential (Table 7.2). Because the scoring system is relatively simple, an inventory of advance reproduction and the overstorey

Table 7.2. Scoring system for assessing the adequacy of the regeneration potential of 1/100-acre plots in southern bottomland hardwood stands.a Advance reproduction height (ft) <1.0 1.1–2.9 >3.0 Score aTo

1

2

3

Overstorey tree dbh (inches)b <5.5 5.6–10.5 10.6–15.5 3

2

1

determine if a plot is adequately stocked, each tree (reproduction and overstorey) within the plot is assigned the score given in the table. A plot is deemed adequately stocked with ‘desirable’ reproduction if the tree scores sum to 12 or more. Scores for overstorey trees reflect their stump sprouting potential. The system is based on studies in stands dominated by mixed oaks (willow, water, cherrybark, overcup, white, laurel, swamp chestnut, Shumard, southern red, Nuttall and northern red) and green ash in east central and southern Mississippi. (From Johnson and Deen, 1993.) bIncludes trees ≥2 inches dbh.

274

Chapter 7

can be rapidly and efﬁciently obtained. Hand calculation of plot and stand stocking is relatively simple and an example tally sheet is presented in Johnson and Deen (1993). Based on assessments of reproduction immediately before and 1–3 years after complete overstorey removal in 118 plots in nine stands, 78% of plots were identiﬁed as correctly categorized by the preharvest scoring system. Because of the ephemeral nature of advance reproduction in bottomlands, the preharvest inventory should be made within 1 year of overstorey removal. Model development and characteristics are similar to that of the regeneration model in SILVAH.

The shelterwood method The objective of the shelterwood method is to create conditions favourable for the establishment and development of tree reproduction of the desired species beneath the parent stand. Stand density is reduced to leave only those trees needed to produce adequate shade and protection for a new age class to develop (Helms, 1998). Control of overstorey density thus functions to control light and the development of understorey competition, moderate temperatures, reduce wind velocity and associated drying effects, and provide soil protection. The method can include a sequence of cuttings including: (i) a preparatory cut to facilitate crown expansion of residual trees and increase seed production; (ii) an establishment cut to prepare the seedbed for seedling establishment; and (iii) a removal cut to release established reproduction from the overstorey. A feature common to all stages of the shelterwood method is the retainment of the best trees for the ﬁnal removal cut. There are several variants of the shelterwood method based on how cuttings are carried out spatially. In the uniform shelterwood method, trees are cut uniformly throughout the stand. In the group shelterwood method, trees are cut in groups or patches; in the strip shelterwood method, trees are cut in narrow strips (Helms,

1998). In the shelterwood with reserves method, trees are retained after reproduction is established to obtain goals other than regeneration. Sometimes the method also involves controlling understorey vegetation. All of these methods are potentially applicable to oak forests.

Suitability to oak regeneration requirements The shelterwood method frequently has been advocated as a method suited to regenerating oaks (Korstian, 1927; Scholz, 1952; Sander, 1979; Smith, 1986; Hannah, 1987; Beck, 1991). The method is potentially applicable in the troublesome mesic and hydric ecosystems because it focuses on controlling stand density near the end of the rotation – a critical period for reinforcing the accumulation of oak reproduction. Controlling light and competition are key features of the shelterwood’s application to oaks. Light under unthinned hardwood stands may be only 1% of full sunlight (Dey and Parker, 1996, 1997). For northern red and many other oaks, the minimum amount of light required to produce carbohydrates sufﬁcient to sustain existing tissues (i.e. the physiological compensation point) is about 2–5% of full sunlight (Hanson et al., 1987). However, maintaining positive rates of shoot growth requires light above 20% of full sunlight (Gottschalk, 1994). In red oak stands in Ontario, shelterwoods that reduce stand density to 70% and 50% crown cover increased light at seedling level to 25% and 50%, respectively (Dey and Parker, 1996). Thus, shelterwoods of appropriate density should in theory provide the light environment needed for adequate oak seedling growth. In northern climates such as the Lake States, the method also can be used to protect oak seedlings from frost damage (Teclaw and Isebrands, 1993a). However, effectively applying the shelterwood method to oak regeneration often involves more than controlling overstorey density. Reducing overstorey density often stimulates undesirable understorey vegetation

Even-aged Silvicultural Methods

and thus may defeat the objective (Schuler and Miller, 1995). The timing of shelterwood cuts with acorn crops and a shelterwood period that assures adequate seedling development are also important but potentially complicating factors in applying the method. For these reasons, shelterwood prescriptions for oak forests need to be ecosystem-speciﬁc. The following section presents some examples of the method’s application to oaks in several regions.

Applications in oak forests Among the variants of the method, the uniform shelterwood method has been most commonly applied to oak forests. The method involves cutting uniformly throughout the stand primarily to remove overtopped and intermediate crown classes, and secondarily to remove other main canopy trees until the desired stand density is obtained (Fig. 7.10A). The desired shelterwood density and degree of understorey control may vary by forest type and site quality. The shelterwood can be removed in a single harvest or a series of harvests over intervals of time sufﬁcient to obtain the desired reproduction development (Fig. 7.10B). A prescription for applying the method was developed for productive sites (site index 70–90 ft) in mixed hardwood forests in the southern Appalachians (Loftis, 1983a, 1990a,b). These forests typically have dense sub-canopies of tolerant hardwoods comprising 15–30% of total stand basal area. The prescription calls for a twocut shelterwood in which stand density is reduced to 60–70% of average maximum stand basal density, depending on site index. This is accomplished by removing trees from the sub-canopy by applying a herbicide, which also prevents their sprouting. The method thus eliminates the sub-canopy and thereby increases light without creating large gaps in the main canopy. On site indices of 70, 80 and 90 ft, recommended residual basal areas are 60%, 65% and 70% of initial basal area, respectively. Reducing stand density below these limits increases the establishment of

275

yellow-poplar seedlings to densities that are incompatible with regenerating red oak. Recommendations are to make the initial shelterwood cut at least 10 years before the end of the rotation. By applying this prescription, the survival rate (Fig. 3.19A) and growth of red oak seedlings can be substantially increased. The method also is compatible with minimizing risks from oak decline because it causes minimal site disturbance and overstorey canopy reduction. Although northern red oak seedlings under red oak stands in Appalachian forests commonly number 1000 to 2000 per acre (Tryon and Carvell, 1958; Loftis, 1983b), seedlings often average only 0.2 inch or less in basal diameter (Loftis, 1988). If the above shelterwood prescription were applied, relations among seedling survival, growth and dominance probabilities indicate that 1000 well-distributed seedlings per acre would result in 10 to 40 dominant and codominant red oaks per acre 20 years after ﬁnal overstorey removal across the site index range of 90–70 ft, respectively. Even though these estimates are hypothetical and remain to be veriﬁed by long-term ﬁeld tests, they nevertheless illustrate the potentials and problems in maintaining red oak in regenerated stands in this and other regions where yellow-poplar is present. Success in maintaining the species largely depends on site quality, stand density and competition control. The documented successes and failures in applying the shelterwood method in mesic ecosystems in the eastern United States emphasize the importance of controlling competition, including interfering herb, shrub and tree layers (Johnson et al., 1989; Horsley, 1991; Crow, 1992; Jacobs and Wray, 1992; Schuler and Miller, 1995). Herbicides may provide the most efﬁcient method of competition control, but opposition to their use has prompted evaluation of other methods. And soil scariﬁcation, especially on steep terrain, may not always be feasible or environmentally acceptable. The shelterwood method is sometimes effective in regenerating northern red oak in the Driftless Area of southwestern Wisconsin (Johnson et al., 1989). Based on

276

Chapter 7

A

B

Fig. 7.10. (A) The uniform shelterwood method applied to a 90-year-old stand in the Ozark Highlands of Missouri (Central Hardwood Region); site index for black oak is 75 ft. The stand was thinned to 60% stocking largely through the removal of trees in suppressed and intermediate crown classes. The shelterwood is composed of northern red, white and black oaks. (B) Five years after reducing overstorey density to 60% stocking, a dense understorey of oak and other reproduction was present beneath this shelterwood in southeastern Ohio (Central Hardwood Region); black oak site index is 65 ft. (USDA Forest Service, North Central Research Station photographs.)

two case histories (red oak site index 65–70 ft), reduction in overstorey density to 70–80 ft2 per acre 10 years before shelterwood removal resulted in 200–400 dominant and co-dominant red oaks per acre during the second decade after shelterwood removal. However, early results from a study on a similar site in southeastern Minnesota suggested that the method may be more effective when reduction in overstorey density

coincides with a good acorn crop and competition is reduced by soil scariﬁcation (Bundy et al., 1991). Application of the shelterwood method in mesic oak forests in southern Michigan (Hill and Dickmann, 1988) and southern Wisconsin (Lorimer, 1989) also produced encouraging results. The minimum duration of the shelterwood period, i.e. the length of time between the shelterwood cut and shelterwood

Even-aged Silvicultural Methods

removal, is uncertain. Results from the Driftless Area suggest that 1 or 2 years may be sufﬁcient if competition is controlled. Results from one case history in southwestern Wisconsin even indicate that northern red oak can be successfully regenerated with little or no oak advance reproduction provided that the ﬁnal harvest occurs during the dormant season following a good acorn crop (Johnson et al., 1989). Although those results challenge the axiom that advance reproduction is required, the safest approach nevertheless would be to obtain adequate oak reproduction before the seed source is removed in the ﬁnal harvest. The potential for regenerating oaks using prescribed burning in combination with the shelterwood method has been demonstrated in principle (Nyland et al., 1983; Brose and Van Lear, 1998a,b; Van Lear and Waldrop, 1988; Keyser et al., 1996). Its primary value may be on highly productive sites where oak regeneration is most problematic. Burning is unlikely to beneﬁt stands on dry sites where oak reproduction naturally accumulates. Wherever it is used, success will depend on matching the frequency, timing and intensity of burns with other stand characteristics. Based on a review of the literature, Van Lear and Waldrop (1988) concluded that frequent low-intensity back ﬁres (i.e. ﬁres burning against the wind) with low ﬂame heights would be the most useful for building up oak reproduction under a shelterwood. Such burns do not damage the overstorey or site. In contrast, high-intensity head ﬁres (i.e. ﬁres burning with the wind) are likely to cause excessive overstorey mortality because the ﬂames produced may reach into tree crowns. Such ﬁres also may cause site damage. Single low-intensity burns are not likely to effectively increase oak regeneration potential (Johnson, 1974; Nyland et al., 1983; Merritt and Pope, 1991; WillWolf, 1991). A single burn may actually stimulate the germination of yellow-poplar seed (Shearin et al., 1972), which often remains viable in the forest ﬂoor in large numbers for 4 years or longer (Clark and Boyce, 1964). In contrast, repeated prescribed burns under pine stands have been

277

shown to increase the build-up of oak reproduction at the expense of the pine and ﬁre-sensitive hardwoods such as yellowpoplar (Clatterbuck, 1998). When consistent with other management objectives, shelterwoods can be maintained until a two-aged stand, or ‘irregular’ shelterwood develops (Smith, 1986; Beck, 1991). Although seldom used, the irregular shelterwood method is potentially compatible with managing oak forests for continuous acorn production (Chapter 9) and other objectives such as maintaining a continuous overstorey canopy while retaining the essential characteristics of even-aged stands. Similarly, the group shelterwood method has been proposed and described with respect to its possible application to oaks, but has seldom been used (Stroempl and Secker, 1993).

The seed tree method The seed tree method is an even-aged method that usually leaves ten or fewer seed-producing trees on every acre (Smith, 1986). In its conventional application, the intent is to provide seed after most of the overstorey is removed. The method is generally not recommended for regenerating oaks because the seed trees provide too little reproduction too late. In the few reported cases where the method has been applied to oak forests, the seed trees contributed little to regeneration (DeBell et al., 1968; Johnson and Krinard, 1983). Nevertheless, the method may be useful if sustaining acorn production on harvested areas is important. It can provide for substantial acorn production for wildlife if good seed producers are retained. However, identifying the seed producers requires long-term records on seed production, which few forest managers are likely to have (Chapter 9). The method nevertheless creates more structural diversity and visual appeal than a clearcut. When applied for this purpose, the seed tree method does not differ from clearcutting with reserves. Sharp (1958) observed that fewer than 30% of white oaks produced acorns and

278

Chapter 7

many of those were poor producers. If this is typical of most oak species, leaving ten good acorn producers per acre would retain 40% or more of the acorn producing capacity of the original stand. This assumes that there are about 75 oaks in the overstorey at the end of the rotation. Moreover, under the open-grown conditions created by the seed tree method, the crowns of individual seed trees can potentially expand to their maximum area and branch density to maximize acorn production per tree. However, the crowns of some seed trees may degenerate from crown dieback because of their sudden exposure to full light (Smith, 1986). However, the snags that ultimately develop provide valuable habitat for cavity nesting birds, and the standing dead wood provides habitat for organisms essential to maintaining biodiversity (Franklin et al., 1989; Hansen et al., 1991). Moreover, seed trees could be quickly converted to dead snags by girdling once it was determined they had little or no value for acorn production or other purposes.

Artificial Regeneration Methods Oaks can sometimes be successfully regenerated by seeding and planting. These methods can be applied to old ﬁelds and other open areas devoid of trees, or combined with natural regeneration methods such as clearcutting and shelterwood methods. In any given situation, the feasibility of planting depends on the species to be planted, site and cost factors. Regardless of method, success in planting depends in part on obtaining oak seedlings that are capable of performing adequately in a given environment.

Oak nursery stock Producing high quality seedlings is essential to planting success. Assuring seedling quality begins with collecting acorns from stands or trees known to be adapted to the planting site. Ideally, seed should be obtained from trees of genetically superior quality, or if that

is not possible, from sources known to produce acceptable seedlings suited to the intended planting sites. Gains in ﬁeld performance from genetically superior seed sources are potentially great (LaFarge and Lewis, 1987; Buchschacher et al., 1991; Rink and Coggeshall, 1995; Kormanik et al., 1997; Schlarbaum et al., 1997a,b). Although such oak seedlings are not yet generally available, they may soon be (Schlarbaum et al., 1997a). After acorns are collected, their viability and germination potential can be maintained by placing them in water immediately after collection and discarding those that ﬂoat. The method assumes that the ‘ﬂoaters’ largely represent weeviled or desiccated acorns and that the ‘sinkers’ largely represent sound acorns. However, in some years, many sound acorns may lose enough water through natural drying before they are collected to cause them to ﬂoat. The percentage of sound acorns can be determined by cutting open a sample of ‘sinkers’ and ‘ﬂoaters’ to identify the proportion of weeviled or otherwise damaged acorns in each population. The population of ﬂoaters then can be discarded or saved depending on their internal appearance. In either case, hydration of acorns for 16–24 hours ensures full imbibition of water without harming the acorn. Acorns can be sown immediately after collecting or placed in cold storage if sowing is to be delayed. Species in the white oak group preferably should be sown shortly after collection in the autumn; they cannot be successfully stored for more than 4–6 months (over winter) (Bonner and Vozzo, 1987). However, some species in the red oak group have been successfully stored for up to 3 years. For either species group, the best storage environment is one that: • maintains acorn moisture content above 30% for species in the red oak group, and at 45–60% for species in the white oak group; • maintains temperatures near but above freezing (34–36°F or 1–2°C); and • allows some gas exchange with the atmosphere (Bonner, 1973; Bonner and Vozzo, 1987).

Even-aged Silvicultural Methods

Storing acorns in 4–10 mil polyethylene bags allows the necessary exchange of carbon dioxide and oxygen, yet provides an effective barrier to moisture loss. However, thinner polyethylene (1.75 mil) or cloth bags may be better for white oak acorns because of their greater air requirements (Rink and Williams, 1984). A common practice is to store acorns in cans or drums with a polyethylene liner bag. To maintain adequate aeration, the container tops should not be completely closed. Additional aeration of acorns during storage often occurs when weevil larvae emerge from acorns and eat through the polyethylene bags. However, larvae do not attack intact acorns during storage. Weevil infestation therefore cannot increase during storage. Although there are methods for killing larvae (Olson, 1974), the risks from doing so outweigh the advantages. One method, which requires immersing acorns in hot water (49°C) for 40 min, may kill the acorns. Another, and even less desirable method, is to fumigate the acorns with methyl bromide, carbon disulphide or thiamine bisulphate. These treatments also can harm acorns if not done properly. The general recommendation is to avoid all such treatments (Bonner and Vozzo, 1987). Controlling seedling density in the nursery is also important to ensure seedling quality. Seedbed density can be controlled by sowing mature acorns of known germination capacity and high moisture content. High seedbed densities (e.g. > six seedlings per square foot) will result in small seedlings with limited growth capacity. Obtaining a desirable density therefore requires adjusting sowing rates based on the percentage of acorns expected to germinate. After sowing, the seedbed should be mulched to prevent acorn desiccation and to protect acorns from freezing. In addition, it is usually necessary to protect seedbeds from bird and mammal predation until germination is complete and leaves are about onehalf fully expanded. If acorns are inadvertently oversown, seedbed density can be reduced by removing (‘roguing’) seedlings (preferably the smaller ones) until the desired bed density is obtained. Roguing should be done early in the ﬁrst growing season.

279

Seedlings can be lifted any time after they are fully dormant in the autumn. Cold storage at 33–38°F (0.5–3°C) and proper packaging of seedlings from time of lifting to time of planting is essential for maintaining physiological quality of seedlings (Webb and von Althen, 1980). Seedlings can be planted in the autumn, winter or spring provided they are dormant. Seedling size and root morphology at the time of planting are important determinants of the physiological quality of seedlings and thus planted tree performance (Johnson, 1988; Kormanik, 1989; Schultz and Thompson, 1989, 1990, 1991; Kaczmarek and Pope, 1993; Teclaw and Isebrands, 1993a,b; Thompson and Schultz, 1995; Dey and Parker, 1997b; Kormanik et al., 1998). Based on studies in the Eastern Hardwood Region, oak seedlings with relatively large basal diameters (e.g. ≥3/8 inch 1 inch above the root collar) perform better than smaller seedlings. Planting seedlings that have been grown for 1 year in the nursery then lifted and transplanted back to the nursery (called 1+1 transplants) have been shown to outperform 1- or 2-year-old seedlings that are not transplanted in the nursery (Johnson, 1984). Similar but more cost-effective advantages can be obtained by undercutting seedlings in the nursery the ﬁrst or second year (Schultz and Thompson, 1990; Buchschacher et al., 1991; Weigel and Johnson, 1998a,b, 1999). Seedlings can be undercut by drawing a tractor-mounted blade or other root-cutting device at a prescribed soil depth – usually 6–8 inches below the soil surface (Fig. 7.11A). This practice severs the tap root and stimulates the development of lateral roots in the nursery (Fig. 7.11B) and thus the number of new roots that develop after ﬁeld planting (Johnson, 1988). The combined effect of undercutting and low seedbed density can greatly increase the number of large lateral roots that are essential for post-planting root regeneration and thus successful ﬁeld performance (Table 7.3). However, to obtain the full advantage of undercutting, the blade used to lift seedlings should be set below the undercutting depth. Undercutting should not be confused with root pruning,

280

Chapter 7

A

B

Fig. 7.11. (A) Undercutting northern red oak seedlings in the nursery with a tractor-drawn blade in early summer of their second year. (B) Undercutting stimulates lateral root development (0 = not undercut, 1 = once undercut, 2 = twice undercut). These seedlings are also root pruned to a common length (10 inches) to facilitate field planting. Because undercutting retards overall seedling growth, a given nursery bed will yield fewer undercut seedlings of a given diameter than not-undercut seedlings (Schultz and Thompson, 1990). A favourable time to undercut is when seedlings are in the resting (lag) phase between first and second flushes of shoot growth. The date of occurrence of this stage may occur any time from late spring to mid-summer, depending on nursery latitude, weather, nursery management practices and other factors. (USDA Forest Service, North Central Research Station photographs.)

which is done after lifting to reduce roots to a common length (Fig. 7.11B) to accommodate proper placement in a planting hole of the same depth.

Removing seedling tops 6–8 inches above the root collar just before planting further improves planted tree performance when trees are planted under shelterwoods.

Even-aged Silvicultural Methods

281

Table 7.3. Response of northern red oak seedlings to variation in nursery bed density and undercutting treatments.a

Bed density/ (undercutting treatment)b 3/ft2 (undercut) (not undercut) 6/ft2 (undercut) (not undercut) 12/ft2 (undercut) (not undercut) Statistical significancee

Mean seedling Mean seedling heightc diameterc (inches) (inches) 15.3 20.9 14.9 19.1 15.2 18.1 D, U, DU

0.32 0.39 0.29 0.33 0.27 0.30 D, U, DU

Mean number of lateral roots > 0.04 inch (1 mm) diameter Normal Wound 1st order rootsd Total 13.4 10.8 10.6 8.8 8.7 6.6 D, U

6.1 — 5.8 — 6.2 —

19.5 10.8 16.4 8.8 14.9 6.6

aFrom

Schultz and Thompson (1990); based on data from the Illinois State Forest Nursery. beds were reduced to the prescribed seedling densities in the spring before the second growing season. Seedlings were undercut during the second growing season at a depth of 6 inches when taproots at that depth were 0.25 to 0.5 inches diameter. cMeasured after two growing seasons in the nursery; diameters were measured 0.5 inch above the root collar. dWound roots are roots > 0.04 inch (1 mm) in diameter that develop at or near the undercutting wound. eVariables statistically significant at = 0.01: D = bed density; U = undercutting treatments; DU = interaction between D and U. bNursery

Although shoot clipping may not beneﬁt seedlings planted in the open (including new clearcuts), there is no evidence it reduces seedling performance (Johnson, 1984, 1988, 1989). For a given type of oak seedling, outplanting performance varies with site quality, competition from established tree reproduction and other vegetation (Johnson and Rogers, 1985).

Oak plantation establishment Oaks can be established as pure or mixed plantations on old ﬁelds or other open areas largely devoid of forest vegetation. On these sites, mechanical planters and site preparation equipment can often be used. Subsequent mechanical weeding and row thinning methods such as those commonly applied to pine plantations also can be used. In the United States, probably more is known about oak plantation establishment in southern bottomlands than elsewhere. This has partly resulted from the availability of bottomland agricultural ﬁelds that were abandoned because of frequent ﬂood-

ing. These areas were originally covered by oaks and other lowland hardwoods and rank among the most productive forests within the temperate region. Oak site index on southern bottomlands ranges from 80 to 100 ft or more at an index age of 50 years. The productivity of bottomland forests therefore may sometimes justify relatively high investments in site preparation, planting and weeding. Species that have been successfully planted include Nuttall, water, cherrybark, Shumard, willow, white and swamp chestnut oaks. Recommendations are to plant seedlings of 3/8-inch caliper or larger with roots pruned to a length of about 8 inches. Bottomland oaks grow best on moist, welldrained soils of medium texture and high fertility. Because each bottomland oak species has speciﬁc site requirements, successful oak planting requires careful matching of species to site (Fig. 1.17). Site preparation is often necessary to obtain acceptable growth and to facilitate the use of mechanical planters or seeders. Sites that have been under cultivation for a long time may require breaking plough

282

Chapter 7

pans or other compacted soil layers with specialized equipment. Discing, mowing and herbicides can be used to speed the growth of planted oaks during the ﬁrst few years. However, even without weed control, trees growing on site index 80–90 ft can be expected to reach 15–20 ft in height in 10–15 years under normal conditions (Kennedy, 1993). Cleaning may be required after that to release planted trees from overtopping competitors. Direct seeding also has been successfully used in southern bottomlands, especially in the lower Mississippi River Valley. More than ten oak species have been successfully direct-seeded (Johnson and Krinard, 1987). As in planting, species growth requirements should be matched to site. An array of experiences in planting and direct seeding of bottomland oaks are reported in more detail elsewhere (e.g. Johnson, 1981; Krinard and Francis, 1983; Kennedy and Krinard, 1985; Johnson and Krinard, 1987; Krinard and Johnson, 1988; Wittwer, 1991; Bullard et al., 1992; Schweitzer et al., 1997). Oak plantations are sometimes established using ‘tree shelters’. Tree shelters are transparent or translucent plastic tubes that protect individual trees from animal damage and also create a greenhouse effect around each tree (Potter, 1991). They are commercially available in a range of materials, durability, sizes (diameter and height) and colours (Windell, 1991). Sizes typically range from 3 to 6 inches in diameter and 4 to 5 ft in height. Five-foot shelters are recommended where deer browsing is severe or snow is deep (Kays, 1996; Schuler and Miller, 1996). Commonly available colours include white, green, tan and brown. Some are circular in cross section whereas others are square or rectangular. Trees in shelters usually grow faster than non-sheltered trees. Accelerated growth may be due to increased air temperature, carbon dioxide, and reduced wind inside the shelter (Potter, 1988, 1991; Mayhead and Jones, 1991; Windell, 1991; Minter et al., 1992). Although ambient air temperatures inside shelters may exceed 38°C (101°F), leaf temperatures on actively

transpiring trees may be 7–11 degrees lower (Potter, 1988). However, inside and outside temperatures of shelters with perforated walls may not differ (Minter et al., 1992). Several studies have demonstrated the effectiveness of shelters in increasing the survival and height growth of oak seedlings. The period of accelerated growth is most dramatic while seedlings are still in the tree shelter, and slows after they emerge. Emergence may take 2 years or longer depending on tree shelter height and other factors. In southern Britain, sheltered sessile oak seedlings grew ﬁve times faster in height than unsheltered seedlings over 4- and 5-year periods (Potter, 1988, 1991). In the United States, reported height growth of northern red oak in tree shelters ranges from about 30 to 230% greater than that of non-sheltered seedlings (cf. Lantagne, 1991; Teclaw and Isebrands, 1991; Zastrow and Marty, 1991; Minter et al., 1992; Bardon and Countryman, 1993; Smith, 1993; Walters, 1993; Gillespie et al., 1996; Ponder, 1997). Variation in growth responses among studies may be related to differences in competition, site quality, climate and weather, shelter design and colour, overhead shade, seedling quality, length of period observed herbivory and other factors. However, tree shelters have not always accelerated the growth of planted oaks (Lantagne, 1996; Teclaw and Zasada, 1996; Lantagne and Miller, 1997). Tree shelters also may increase oak seedling survival (Potter, 1988, 1991; Lantagne, 1991, 1996; Zastrow and Marty, 1991; Bardon and Countryman, 1993; Smith, 1993). Because of the relatively high cost of the tree shelters and their installation and maintenance, it may be difﬁcult to justify the use of shelters unless animal or other types of damage are severe, future tree value is high, or other factors merit the investment. Shelters nevertheless can reduce the cost of applying herbicides because the shelter itself protects the seedling from herbicide damage. Shelters also protect against damage from mechanical equipment such as mowers and string trimmers, and facilitate

Even-aged Silvicultural Methods

later ﬁnding planted trees for inspection and cultural treatments. Tree shelters can be used in a variety of situations ranging from conventional plantation establishment in old ﬁelds and other treeless areas, to enrichment plantings (see the following section) in clearcuts and under shelterwoods. When tree shelters are used under shelterwoods, shelters that block little light are recommended (Potter, 1991; Schuler and Miller, 1996). Other factors being equal, shelterwood density itself affects the growth of seedlings in tree shelters. In a Wisconsin study, the height growth of northern red oak seedlings in tree shelters decreased with increasing shelterwood density (Teclaw and Isebrands, 1991). But even under high-density shelterwoods (100% crown cover), seedlings in tree shelters grew faster than unsheltered seedlings. Where late spring frosts are frequent and severe (such as in the Northern Hardwood Region), tree shelters used in combination with shelterwoods may be necessary to obtain any growth advantage from the tree shelter. Tree shelters did not beneﬁt northern red oak seedlings planted in clearcuts in northern Wisconsin because of dieback caused by late spring frosts, but did beneﬁt seedlings under shelterwoods that provided 50% or 75% crown cover (Teclaw and Zasada, 1996). However, tree shelters had a detrimental effect on the height growth of northern red oak planted under shelterwoods in a northern Michigan study (Lantagne and Miller, 1997). Tree shelters also have potential application in sheltering natural reproduction (Potter, 1991; Kittredge et al., 1992), direct seeding with acorns (Smith, 1993; Walters 1993; Bailey et al., 1996; Schuler and Miller, 1996), rehabilitating partially failed plantings (Potter, 1991; Gillespie et al., 1996), and revegetating strip-mined lands (Farley et al., 1996). Seedlings grown in tree shelters develop a nearly columnar stem with practically no taper (Potter, 1988). Although rapid height growth helps keep the sheltered tree above competitors and out of reach of browsing animals, stems are usually so weak that the

283

tree shelter is needed for physical support. While this acceleration in stem growth occurs at the expense of root growth, the delay in root growth is more than compensated for by the third year after planting (Ponder, 1996). It nevertheless may take 5 years or longer for the tree to develop sufﬁcient diameter to stand by itself (Potter, 1988, 1991). Tree shelters therefore must be sufﬁciently durable to last until trees can emerge from the shelter. Many of the commercially available tree shelters are made of polypropylene, which is not durable in sunlight unless an ultraviolet inhibitor is added (Potter, 1988; Windell and Haywood, 1996). After trees emerge from the top of the tree shelter, height growth slows and more growth is allocated to stem diameter (Schuler and Miller, 1996; Strobl and Wagner, 1996; Windell and Haywood, 1996). The top rim of a shelter also must be smooth to prevent abrasion and damage to the thin stems. Although promising in principle, much remains to be learned about tree shelters and their application to establishing oaks. Guidelines for using tree shelters with northern red oak, including direct seeding and shelterwood plantings, are presented by Schuler and Miller (1996).

Enrichment planting Enrichment planting is deﬁned as planting to improve the proportion of desirable species or to increase biodiversity by establishing young trees among existing forest growth (Helms, 1998). The method potentially can be combined with any natural regeneration method. In oak forests, experience in enrichment planting largely has been limited to clearcutting and shelterwood methods.

Planting in clearcuts Oaks can be successfully planted in clearcuts despite the numerous reported failures of such plantings (e.g. Hilt, 1977; Loftis, 1979; McGee and Loftis, 1986). Opportunities for planting oaks in clearcuts vary among ecosystems. In

284

Chapter 7

upland oak stands in the eastern United States, probably the best opportunities for planting the major oak species occur within the oak site index range of 60–75 ft. There, site quality is sufﬁcient to meet the growth requirements of most oak species, but competition is not too severe. Outside this site index range, it may be difﬁcult to justify planting oaks for ecological and economic reasons. Enrichment plantings are most effectively used in conjunction with existing natural regeneration guides. Guidelines for evaluating the natural regeneration potential of oak stands are available for several regions (see ‘Regeneration models’, this chapter). Planting costs accordingly can be minimized by planting only the number of trees needed to attain a given future stocking goal (Johnson and Rogers, 1985; Johnson et al., 1986). However, this strategy requires information on the expected performance of the planted oaks. The outcome will depend on site quality, overstorey and understorey composition and structure of the present stand, planting stock size, animal browsing and other factors. Other factors being equal, planting oaks in clearcuts has been less effective than planting oaks under shelterwoods. One reason for planting failures in clearcuts is the delayed onset of root growth in bare-root seedlings (Johnson et al., 1984). Although container-grown seedlings do somewhat better in clearcuts (Johnson et al., 1986), the use of tree shelters (as discussed above) in combination with competition control may provide better results provided that the additional costs can be justiﬁed. Because of the problems in planting oaks in clearcuts, methods of oak enrichment planting are discussed in more detail in the following section on planting under shelterwoods. Mixtures of oak and pine can be created by combining clearcutting with enrichment planting. Where oak regeneration dynamics favours the persistence of oaks, one method for establishing such mixtures is to plant pine. The method has been shown to produce oak–pine mixtures with aesthetic and wildlife beneﬁts as well as cost-efﬁ-

cient timber production beneﬁts. The method has been employed in the Southeast as a low-cost alternative to the high site preparation costs required in establishing pure pine stands. It is especially applicable to cut-over pine and mixed oak–pine stands that have not regenerated to pine, but where the natural oak regeneration potential is high. The site index for white oak on these sites typically ranges from 60 to 70 ft. One method of planting pine in oak clearcuts, developed in the Georgia Piedmont, is called the ‘fell and burn technique’ (McMinn, 1986). The method requires commercial clearcutting followed by the felling of all residual trees greater than 5 ft tall during mid-April to early June. At that time, most of the residual hardwoods are near or at full leaf expansion. The timing of this treatment is critical because: (i) the felled vegetation is needed to carry a summer burn; and (ii) cutting at the peak of leafout reduces the carbohydrate reserves of the hardwood sprouts, which in turn slows their regrowth. By cutting trees after they have leafed out, they also dry more thoroughly through ‘transpirational drying’. The felled trees are then allowed to dry until early July or later, when turkey and quail nesting is completed. The stand is then burned, which produces a high-intensity ﬁre over a moist fuelbed. The result is additional top-kill and carbohydrate reserve depletion of the oaks and other hardwoods. Although burning does not reduce the number of hardwood sprouts, it does reduce their height growth (McMinn, 1986). Burning removes about 80% of the forest ﬂoor, but leaves about two-thirds of the root mat intact to prevent erosion. In late winter or early spring after burning, about 450 genetically improved pine seedlings per acre are hand planted. In the Piedmont, both shortleaf pine and loblolly pine are planted, but in the Coastal Plain, loblolly pine is preferred. Total costs in applying the method are about half those of conventional pine regeneration methods. The method results in high survival and rapid growth of the planted pines because

Even-aged Silvicultural Methods

competition from hardwoods is effectively reduced, but not eliminated. The method thus produces a mixture of pine, oaks and other hardwoods. Wildlife beneﬁts include browse for deer and cover for many other species. As stands mature, the oaks also provide mast. Other beneﬁts include improved insect and disease resistance compared to pure pine stands. Although the prescription calls for removing all residual stems before burning, some residual trees probably could be retained to reinforce structural diversity without detracting from other objectives.

Planting under shelterwoods Oak planting can be combined with the shelterwood method to supplement the natural regeneration potential of a stand. Like natural oak reproduction, planted oaks also can beneﬁt from a shelterwood. The period under a shelterwood allows seedlings time to become established before they compete with the surge of competition that develops after complete overstorey removal. For bare-root nursery stock, this recovery period is critical because the physiological disruptions to seedlings from lifting, handling and planting delay the initiation of root and shoot growth the ﬁrst year after planting (Johnson et al., 1984; Johnson, 1988; Struve and Joly, 1992). Shelterwoods of appropriate density thus provide planted oaks with sufﬁcient light while allowing time for them to re-establish and expand their root systems before ﬁnal shelterwood removal (Dey and Parker, 1996, 1997a). In turn, this increases the proportion of planted trees that can successfully compete with other established trees after shelterwood removal. Although the use of tree shelters is also feasible under shelterwoods (as discussed above), in many if not most situations the additional expense of tree shelters is not necessary if appropriate silvicultural methods are used (as discussed below). Exceptions may include areas where deer browsing or other animal damage is severe. A practical problem in integrating planting with the shelterwood method is quanti-

285

tatively expressing and accurately predicting the expected outcome. Unlike plantings designed to create a monotype (i.e. a single-species stand), planting oaks under shelterwoods can take advantage of a stand’s natural regeneration potential. The number of planted trees (and thus planting costs) required to obtain a given future stocking thereby can be minimized. However, planting trees among naturally established trees and vegetation complicates predicting the overall regeneration outcome. Uncertainties are largely the consequence of unknowns related to the competitive struggle between planted trees and competitors in a relatively heterogeneous physical and biotic environment. The two major determinants of planting success, planted tree survival and growth, have traditionally been treated as separate and independent responses. In establishing monotypes by planting, survival by itself may provide a sufﬁcient measure of planting success – especially when intensive preplanting site preparation eliminates competitors. In a shelterwood setting, it is more useful to deﬁne a planted tree’s competitive capacity, i.e. its expected capacity to survive and grow at a rate sufﬁcient to attain and maintain dominance among its competitors over a speciﬁed period. Accordingly, it makes little difference whether a planted tree dies or grows so slowly that it becomes overtopped by competitors. In either case, the seedling has failed silviculturally. A suitable quantitative expression of competitive capacity accordingly would consider survival and growth simultaneously. Ideally, this quantiﬁcation also would account for the effects of nursery stock quality and planting environment. The latter includes site quality, anticipated changes in competition and planned modiﬁcations of the environment before and after planting including weeding, control of shelterwood density and number of years planted trees remain under the shelterwood. Dominance probabilities (see ‘Regeneration models’, this chapter) provide a useful and convenient quantitative expression of competitive capacity. In a

286

Chapter 7

shelterwood context, dominance probability is the probability that an individual planted tree attains dominance or co-dominance a given number of years after planting or after shelterwood removal. These probabilities are likely to change through time as a result of differences in the survival and growth of planted trees and their competitors. In principle, dominance probabilities can increase, decrease, remain constant or reverse their direction through time. They therefore provide a potentially ﬂexible quantitative expression of a planted tree’s competitive capacity. Dominance probabilities can be calculated as observed averages for a speciﬁc type of seedling in a given planting environment and number of years after planting. Alternatively, they can be estimated by various statistical methods (e.g. Weigel and Johnson, 1998a,b, 2000). The silvicultural value of dominance probabilities (P) lies not so much in the probabilities themselves, but in their reciprocals (i.e. 1/P). The reciprocals deﬁne how many seedlings are needed to obtain one competitively successful tree at a future point in time. These reciprocals are sometimes called planting factors (Johnson, 1984; Johnson et al., 1986; Weigel and Johnson, 1998a,b). The shelterwood method provides ﬂexibility in controlling stand density (and therefore overhead light), timing the removal of the shelterwood and controlling understorey competition. Four important questions in planting oaks under shelterwoods are: • What level of overhead shading (or shelterwood density) is appropriate? • Is control of understorey vegetation necessary? • How long should the shelterwood be retained? • After ﬁnal overstorey removal, is further silvicultural intervention necessary to ensure satisfactory survival and growth of planted trees? Reduction in overstorey density is usually necessary for satisfactory growth of trees planted under shelterwoods. Residual

stand densities near 50–60% stocking or 70–80% crown cover often provide the requisite conditions (Gottschalk and Marquis, 1982; Pubanz and Lorimer, 1992; Dey and Parker, 1996, 1997a). Thinning below those levels often stimulates the growth of understorey competitors at the expense of planted trees. Controlling low shade from shrubs, small trees, stump sprouts and herbaceous vegetation may be as important as controlling high shade (Horsley, 1991; Pubanz and Lorimer, 1992; Teclaw and Isebrands, 1993a,b). Control of understorey vegetation provides the dual advantages of increasing light during the shelterwood period as well as reducing competition after shelterwood removal. The length of time planted trees should remain under a shelterwood therefore partly depends on how quickly the understorey competition develops and whether additional competition control is economically justiﬁable. The effect of shelterwood density, understorey density and various seedling factors is demonstrated by the performance of northern red oaks planted in the Boston Mountains of northern Arkansas (Central Hardwood Region). Based on dominance probabilities 11 years after planting and 8 years after complete shelterwood removal, northern red oaks planted under low to moderate shelterwood densities (40% and 60% stocking) outperformed trees planted under higher shelterwood densities (80% stocking) (Fig. 7.12). For a given initial seedling size and type, dominance probabilities increased with time and with increasing intensity of weed control. Probabilities decreased with increasing site quality (as expressed by site index). The latter effect is related to the greater abundance and rapid growth of competitors on the better sites. There, the dominant competitors are blackgum, red maple, black cherry, ﬂowering dogwood and other shade tolerant species. For a given planting environment, seedling characteristics that inﬂuenced planting success included initial seedling size (expressed as basal stem diameter measured 1 inch above the root collar), and

Even-aged Silvicultural Methods

whether or not the tops of planted seedlings were cut off (clipped) 8 inches above the root collar before planting. Dominance probabilities increased with increasing initial basal stem diameter (i.e. caliper) (Fig. 7.12), and other factors being equal, probabilities were greater for clipped than for unclipped seedlings The effect of the timing of overstorey removal is demonstrated by a 19-year study on planting northern red oak in the Ozark Highlands of Missouri. In this region, opportunities for planting red oak under shelterwoods often occur on north- and east-facing slopes, where site index for black oak typically ranges from 60 to 75 ft. Although site quality there is favourable for northern red oak, the natural oak regeneration potential is often low (Fig. 3.26). The number of years planted trees remain under a shelterwood signiﬁcantly affects dominance probabilities. In the Missouri study, highest probabilities occurred when shelterwoods were retained for 10 years. Shelterwoods retained for 3 or 6 years were second best, while those retained 0 years (i.e. trees planted in clearcuts) were least successful (Fig. 7.13). Dominance probabilities also increased with increasing initial shoot caliper. For a given initial caliper, 2-year-old transplants outperformed 2-yearold seedlings and clipped trees performed better than unclipped trees (Fig. 7.14A). Where oaks occur with yellow-poplar, planting oaks under shelterwoods represents a special problem. Yellow-poplar outgrows most co-occurring oaks, is long-lived and regenerates aggressively from seed and sprouts after complete or moderately heavy overstorey removal (Beck, 1991; Beck and Della-Bianca, 1981). The capacity of oaks to successfully regenerate naturally in competition with yellow-poplar is therefore severely limited. As a consequence, oaks are declining in abundance where the two species co-occur. Attempts to regenerate mixed oak and yellow-poplar stands by planting have produced a long history of planting failures, especially where site quality is high (e.g. red oak site index ≥75 ft) (Olson and Hooper, 1972; Russell, 1973; McGee and Loftis, 1986). In mixed oak and yellow-poplar stands

287

in West Virginia (oak site index 60–70 ft) that were planted to northern red oak after clearcutting, 30–50% of planted seedlings were considered competitively successful after 5 years (Wendel, 1980). Northern red oaks planted under mixed oak/yellowpoplar shelterwoods in southern Indiana (black oak site index 75 ft) responded similarly (Weigel and Johnson, 2000). In this experiment, the shelterwood was thinned to 60% stocking and retained for 3 years. Five years after shelterwood removal (8 years after planting), dominance probabilities of planted trees ranged from <0.10 to >0.55, depending on initial seedling caliper, shoot clipping and undercutting treatments (Fig. 7.15). However, these probabilities declined rapidly during the next 5 years as a result of suppression from overtopping yellow-poplar. Ten years after shelterwood removal, dominance probabilities for all classes and initial sizes of planted seedlings declined to <0.10. Although dominance probabilities indicated that preplanting and postharvest herbicide treatments temporarily created a favourable environment for 35% or more of some types and sizes of planted oaks, yellow-poplar dominated the planting sites 10 years after ﬁnal harvest (Fig. 7.14B). On these and similar sites, successful oak planting requires controlling competitors, especially yellow-poplar, no later than the end of the early release interval combined with planting nursery stock with the requisite growth potential (Fig. 7.15). Prescribed burning has been effectively used in conjunction with the shelterwood method to increase the natural regeneration potential of oaks at the expense of the more ﬁre-sensitive yellow-poplar, maples and other competitors (Nyland et al., 1983; Keyser et al., 1996; Brose and Van Lear, 1998a,b; Clatterbuck, 1998). Weigel and Johnson (2000) accordingly proposed a prescription for planting northern red oak on yellow-poplar sites that calls for relatively large undercut and topclipped planting stock coupled with competition control designed to eliminate subsequent suppression (Fig. 7.15). The objective is thus to preserve the gains obtained during the early release interval.

288

Chapter 7

Site index 18 m (59 ft)

0.8 W2

0.6

W1

0.4

W0

0.2 0.0 11 10 Ye ar 9 sa 8 fte rp 7 lan 6 tin g

4

ia Init

20 ) 16 mm 12 er ( p i l 8 ca ing edl e s l

11th-year dominance probability

Dominance probability

0.8 D4060 D80

W2 W2 W1

0.6

W1 W0

0.4

W0 0.2

0.0 4

8

12

16

20

Initial seedling caliper (mm)

Site index 24 m (79 ft)

0.8 0.6

W2

0.4

W1 W0

0.2 0.0 10 9

Ye a

rs

16

8

aft

er

12

7

pla

nti

6

ng

4

I

8

g

al s niti

lin eed

20

r ipe cal

)

(m m

11th-year dominance probability

Dominance probability

0.8 D4060 D80

W2

0.6

W2 W1

0.4

W1 W0

0.2

W0

0.0 4

8

12

16

20

Initial seedling caliper (mm)

Fig. 7.12. Estimated dominance probabilities for 2-year-old northern red oak seedlings planted under shelterwoods in the Boston Mountains of northern Arkansas (Central Hardwood Region). The estimates are based on logistic regression analysis and are shown for two values of red oak site index in relation to shelterwood density (expressed as stocking per cent, D), years after planting (three-dimensional graphs), initial seedling caliper (stem diameter 2 cm above the root collar), and weed control (W) treatments; W0 = no weed control; W1 = one herbicide application (the winter before spring planting) + cutting stems of all tree reproduction ≥30 cm (1 ft) tall to 3.8 cm (1.5 inches) dbh 30 cm above the ground the winter before overstorey removal; W2 = two herbicide applications before final shelterwood removal. The shelterwoods were removed 3 years after planting. Site index was estimated from Graney and Bower (1971). The three-dimensional graphs show probabilities for shelterwoods at 40% and 60% stocking (D4060), which did not differ significantly ( = 0.001); the 80% shelterwood density (D80) differed significantly from the average of D40 and 60. The two-dimensional graphs show 11th-year probabilities. All estimates shown are for seedlings with shoots clipped 20 cm (8 inches) above the root collar at time of planting; estimates for unclipped seedlings are significantly smaller other factors being equal. All predictors are significant at = 0.001. The model is based on 4128 planted trees distributed across five study sites. (From authors’ data and analysis.)

Even-aged Silvicultural Methods

1 + 1 transplants

Years under shelterwood 10 10 3, 6 3, 6

0.8 Dominance probability

289

Clipped Unclipped 0.6

0 0 0.4

0.2

0.0 4

6

8

10

12

14

16

1 + 0 seedlings 0.8 10

Dominance probability

Clipped Unclipped

10 3, 6 3, 6

0.6

0.4

0 0

0.2

0.0 4

6

8

10

12

14

16

Initial shoot caliper (mm) Fig. 7.13. Estimated dominance probabilities for planted northern red oak seedlings and transplants 16 years after shelterwood removal and 19 years after planting in the Ozark Highlands of Missouri (Central Hardwood Region); black oak site index is 60 to 70 ft. A dominance probability is here defined as the likelihood that a planted tree will attain dominance or co-dominance 16 years after shelterwood removal. Probabilities are statistically adjusted to a common number of years after shelterwood removal (16) based on a ‘time effect’ variable. Actual observed time-since-overstorey-removal ranged from 9 years (10-year treatment) to 19 years (0-year or clearcut treatment). Probabilities are based on logistic regression analysis and are shown in relation to the number of years planted trees remained under the shelterwood (0, 3, 6 and 10), initial shoot caliper and shoot clipping treatments. Probabilities for 3- and 6-year shelterwood periods did not significantly differ ( = 0.01). Probabilities for 10-year and 3- and 6-year shelterwood treatments differed significantly from the clearcut treatment. Clipped seedlings performed significantly better than unclipped seedlings. (Author’s analysis of USDA Forest Service data.)

290

Chapter 7

A

B

Fig. 7.14. (A) A planted northern red oak in the Ozark Highlands of Missouri (Central Hardwood Region) that is co-dominant 13 years after shelterwood removal. The tree remained under the shelterwood (at 60% stocking) for 3 years. This tree was a 1-year-old seedling with a clipped top and a caliper of 11 mm (approx. 3/8 inch) at time of planting. Based on estimated dominance probabilities (Fig. 7.13), about 42% of such seedlings would be expected to attain dominance or co-dominance 16 years after shelterwood removal. (B) A planted northern red oak (held in hand) in an intermediate crown class 10 years after shelterwood removal in a mixed oak–yellow-poplar stand in southern Indiana (Central Hardwood Region). Ten years after shelterwood removal, 90% of planted trees had died or were overtopped by yellow-poplars such as the two large trees in the foreground and the one in the background (Weigel and Johnson, 2000). (USDA Forest Service, North Central Research Station photographs.)

Several regionally speciﬁc prescriptions for planting oaks under shelterwoods have been developed based on dominance probabilities (e.g. Johnson et al., 1986; Weigel and Johnson, 1998a,b, 2000; Johnson, 1992b). Collectively, these prescriptions emphasize the need to consider site quality and initial seedling characteristics as discussed above. Examples include prescriptions for planting northern red oak and white oak under shelterwoods in the Ozark Highlands (Weigel and Johnson, 1998a,b), which call for: • Planting on sites with a site index of 60 ft or greater for black oak.

• Reducing overstorey density to 60% stocking based on Gingrich’s (1967) stocking chart by thinning from below (i.e. concentrating removals on sub-canopy trees down to 2 inches dbh) (Fig. 7.16A). • Treating the planting site with an effective herbicide before planting by targeting woody plants between 1/2 inch and 2 inches dbh. • Planting top-clipped seedlings that average at least 1/4 inch in caliper measured 1 inch above the root collar (Fig. 7.16B). • Removing the shelterwood after three growing seasons. For either species, best results are

Even-aged Silvicultural Methods

291

Years after planting 7

8

9

10

11

12

13

9

10

0.7 Initial seedling caliper (inches)

Dominance probability

0.6

U0C0 1.0

0.5

0.75

0.4

1.0 0.5

0.3

0.75

0.2

0.5 0.25

0.1

0.25 Early release interval

0.0

U1C1

4

Suppression interval

5

6

7

8

Years after shelterwood removal Fig. 7.15. Estimated dominance probabilities for 2-year-old northern red oak seedlings planted under mixed oak and yellow-poplar shelterwoods in southern Indiana (Central Hardwood Region); the site index for black oak is 75 ft. Shelterwoods were retained for 3 years. Probabilities are shown in relation to years after shelterwood removal and initial seedling caliper. Dominance probabilities are shown for the nominally ‘best’ and ‘worst’ undercutting (U) and shoot clipping (C) treatments. (U0C0: not undercut, not top-clipped; U1C1: undercut during the first growing season in the nursery, shoots clipped 8 inches above the root collar before planting). A herbicide was applied to the planting sites once before planting and once after planting using a directed spray that targeted individual woody stems 0.5 to 2 inches dbh. (From Weigel and Johnson, 2000.)

obtained when shelterwoods are planted in the spring with seedlings that are at least 1/4 inch in caliper and top-clipped. Based on the planting factors (the reciprocals of dominance probabilities) derived for this region, it would be necessary to plant 220 red oak seedlings averaging 1/4 inch in caliper to obtain 100 competitively successful trees 10 years after shelterwood removal; 180 undercut white oak seedlings of the same size would meet the same goal 8 years after shelterwood removal (Table 7.4). Although numbers of planted trees required to meet a given future stocking goal decrease with increasing initial seedling caliper (Table 7.4), the larger seedling and planting costs

may not justify the expense in the Ozark Region (Weigel and Johnson, 1998a,b). In contrast, planting red oaks in yellowpoplar stands will require seedlings ≥1/2 inch in caliper to be competitively successful even during the early release interval (i.e. the ﬁrst 5 years after shelterwood removal) (Fig. 7.15). Herbicide applications should target individual woody stems between 0.5 and 2 inches dbh. This approach minimizes the amount of herbicide applied while effectively reducing the primary understorey competitors of planted trees. When herbicides are applied to individual stems as a directed spray, the desirable natural reproduction already present can be preserved.

292

Chapter 7

A

B

Fig. 7.16. (A) Planting oaks under a shelterwood thinned to 60% stocking in the Ozark Highlands of Missouri (Central Hardwood Region). (B) A 2-year-old top-clipped northern red oak seedling prepared for planting. The top has been clipped 8 inches above the root collar. This seedling was undercut in the nursery the first year. The taproot and lateral roots were pruned to a common length of 8 inches below the root collar to facilitate planting. After planting, most new roots develop from large-diameter (≥2 mm) firstorder laterals and at or near the point of root pruning. (USDA Forest Service, North Central Research Station photographs.)

Alternatively, it may be feasible in some situations to avoid the use of herbicides by adjusting planting factors and initial seedling caliper upwards to compensate for expected losses to increased competition (Johnson, 1992b). Alternative methods of weed control such as prescribed burning also may provide effective weed control

(Brose and Van Lear, 1998a,b; Weigel and Johnson, 2000). Other factors being equal, attaining a given stocking goal requires more intensive weed control on good sites than on poor sites (Fig. 7.12). In the Ozark Highlands, shelterwoods can be retained for at least 10 years without adversely affecting survival or response to

Even-aged Silvicultural Methods

293

Table 7.4. Number of planted oaks in the Ozark Highlands of Missouri needed to obtain one dominant or co-dominant tree 8 or 10 years after shelterwood removal (planting factor) in relation to initial seedling caliper and nursery undercutting treatment. Planting factor

Initial seedling caliper (inch)a 1/4 3/8 1/2 5/8 3/4 7/8 1

White oakb (8 years after shelterwood removal)

Northern red oakc (10 years after shelterwood removal)

Not undercut

Undercutd

Not undercut or undercute

3.2 2.4 2.0 1.8 1.6 1.5 1.4

1.8 1.5 1.4 1.3 1.2 1.2 1.2

2.2 2.0 1.9 1.8 1.8 1.8 1.8

aMeasured

1 inch above the root collar. on estimated dominance probabilities 8 years after shelterwood removal for 3-year-old seedlings with tops clipped 8 inches above the root collar after lifting and roots pruned to a common length of 10 inches before planting. (From Weigel and Johnson, 1998a.) cBased on estimated dominance probabilities 10 years after shelterwood removal for 2-year-old seedlings with tops clipped 8 inches above the root collar after lifting and roots pruned to a common length of 8 inches before planting (Fig. 7.16B). (From Weigel and Johnson, 1998b.) dUndercut the second year in the nursery. eUndercut the first year in the nursery. bBased

overstorey release (Fig. 7.13). However, this relatively long shelterwood period may not be advisable in regions where understoreys quickly redevelop after treatment. Shelterwood densities of 40–60% stocking provide the requisite light conditions when coupled with weed control (Fig. 7.12). Although the combined costs of weed control and planting under shelterwoods are relatively high (Weigel and Johnson, 1998a,b), investment analysis of Iowa and Missouri forests indicates that the method may be a viable economic alternative to allowing stands to succeed naturally to species of lower economic value (Countryman and Miller, 1989). The method also capitalizes on a stand’s natural regeneration potential and thereby minimizes planting costs while maintaining diverse mixtures of tree species. Future advances in seedling quality through genetic selection of seedlings with rapid growth and roots with numerous large lat-

eral roots may make shelterwood plantings and other oak planting strategies an even more dependable regeneration practice (Kormanik et al., 1997, 1998).

Intermediate Cuttings Definitions and theory Intermediate cuttings include any treatment or method of tending designed to improve the growth, quality, vigour or composition of a stand. They include thinning, cleaning, release cutting, and various types of improvement cutting including salvage and sanitation cuttings as discussed below. The objective of thinning is to concentrate growth on fewer trees per acre than would occur in unthinned stands (Chapter 6), and in the process to improve stand composition and tree quality. Several types

294

Chapter 7

of thinning methods have been described including: low thinning, crown thinning, selection thinning, mechanical thinning and free thinning, which are deﬁned as follows (Smith, 1986): • Low thinning, often called ‘thinning from below’, removes trees in the lower crown and diameter classes. • Crown thinning removes trees from the middle and upper ranges of crown and diameter classes. The objective is to favour the development of the best trees in these same classes. • Selection thinning removes dominant trees to stimulate the growth of trees in the lower crown classes. The method can be used to eliminate poorly formed dominants and release less vigorous trees of better form. • Mechanical thinning is any method based primarily on the spacing between trees (e.g. row thinning in plantations), and is usually applied to young stands with relatively undifferentiated crown classes. • Free thinning is designed to free crop trees without regard to their position in the crown canopy. The method is applicable to stands with very irregular structure and composition. All of these thinning methods are potentially applicable to even-aged oak stands. However, low thinning is the most frequently used method and is the focus of the discussion on thinning presented below. Although low thinning in oak forests concentrates on removing trees in the lower crown and diameter classes, some trees in the higher classes also may be removed to obtain the desired stocking. Thus, the practical difference between low and high thinning is usually one of degree of removal of trees in low versus high crown classes. Methods other than low thinning may be dictated by existing stand structure and composition and other factors. Regardless of the thinning method used, at the end of the rotation, high-value trees should occupy most of the growing space. This objective does not rule out incorporating other management objectives such as

retaining some cull or other trees of low economic value that are of value for wildlife and other non-timber objectives. When thinning maintains stand densities that fully utilize growing space, there is theoretically no loss in total standing tree biomass or volume yield by the end of the rotation. In fact, yields of merchantable timber are likely to be increased by thinning because thinnings recover merchantable products that would otherwise be lost to natural mortality from self-thinning (Chapters 6 and 10). Not everybody agrees on how and when thinnings should be made in oak stands. However, silviculturists generally agree that stands originating primarily from seedlings and seedling sprouts should be treated differently from those originating primarily from stump sprouts. The latter require early intermediate cuttings if tree quality matters. The former may sometimes beneﬁt from no thinning for 40 years or longer (Carvell, 1971). Conversely, oak stands are usually not thinned later than 60 years in the eastern United States. After that, growth responses are usually no better than in unthinned stands. However, it is sometimes desirable to salvage trees that have succumbed to insects, disease or windthrow late in the rotation. Under normal conditions, silviculture applied late in the rotation should be aimed at preparing for regeneration as discussed in the shelterwood method section of this chapter. When thinnings are concentrated in the lower crown classes throughout the life of an oak stand, the ﬁnal crop trees will be of large diameter and thus low stump sprouting capacity (Fig. 2.24). Although many of the small-diameter oaks removed in thinning will sprout, the moderate-to-high overstorey densities usually maintained in well-managed oak stands are not conducive to stump sprout survival to the end of the rotation (Gardiner and Helmig, 1997). Intensive management for sawlogs therefore can be expected to progressively reduce the importance of oak stump sprouts and thereby force a reliance on regeneration from oak seedlings and seedling sprouts. Lengthening the rotation

Even-aged Silvicultural Methods

will have a similar effect, because for a given diameter, older oaks have lower sprouting probabilities than younger oaks (Fig. 2.24). In addition to thinning, other types of intermediate cuttings are also applicable to oak forests. These include cleanings, which are applied to stands not past the sapling stage (including stand initiation and early stem exclusion stages; Fig. 5.3). The objective of a cleaning is to free the favoured trees from the less desirable trees of the same age class that overtop them or are likely to (Smith, 1986; Helms, 1998). Liberation cuttings are similar to cleanings, but involve the removal of older trees left from a previous stand or which for other reasons were present before the new stand was established. Both types of cuttings, by deﬁnition, are applicable to stands not past the sapling stage. Improvement cuttings are deﬁned as those made in stands past the sapling stage to improve their composition and quality by removing from the main canopy trees of undesirable species, form and condition (Smith, 1986). They are often conducted preliminary to, or simultaneously with, regular thinnings. Improvement cuttings are applicable to many oak stands that have not previously been managed, which often include undesirable species or trees of poor quality. Improvement cuttings include salvage cuttings made to remove trees that are in danger of being killed or that have already been damaged or recently killed by injurious agents other than competition. Examples of their application to oak forests include cutting dead or dying trees affected by gypsy moth defoliation and oak decline. When such cuttings are made to pre-empt damage to high risk trees, they are called presalvage cuttings. Sanitation cuttings remove trees that have been attacked by, or appear to be in danger of attack by, certain insects and diseases to prevent their spread to other trees (Smith, 1986). Salvage, presalvage, sanitation and other types of cuttings can be used to control several insect and disease problems in oak forests as discussed below.

295

Application Thinning and cleaning Intermediate cutting focuses on maintaining an adequate density of dominant and co-dominant trees, and improving species composition and stand quality. Because oak stands are usually comprised of a wide range of tree diameters and crown classes, intermediate cuttings that remove only the largest trees automatically promote the smaller ones to crop trees. However, such trees have grown so slowly and often are of such low vigour that they usually respond poorly to thinning. This lengthens the rotation and also may produce dysgenic effects, i.e. effects that are detrimental to the genetic qualities of the future stand. The objective of thinning is to minimize rotation length by growing trees as rapidly as is consistent with maintaining tree quality and full utilization of growing space. This is usually accomplished by restricting cutting to the lower half of the diameter range (Roach and Gingrich, 1968) and maintaining moderate stand densities (e.g. near B-level as deﬁned by the stocking charts described in Chapter 6). However, trees substantially larger than the dominant and co-dominant trees in the main stand, which may represent an older age class of trees persisting from an earlier stand condition, are also candidates for removal. This is especially true if such trees are numerous enough to seriously hinder the development of the predominant age class. Trees to be retained thus should be selected on the basis of species, size, stem form, stand structure, apparent vigour as evidenced by crown condition, and spacing. Thinning can be applied to stands before they produce merchantable products. This is likely to be the case in stands not yet pole-size (4–9 inches dbh). Any treatment applied to stands not past the sapling stage that is designed to free the favoured trees from less desirable trees of the same age class that overtop them or are likely to do so is called a cleaning (Helms, 1998). In oak forests of the eastern United States, the sapling period ranges from about 5 to 15 years, depending on site

296

Chapter 7

quality and species composition. In general, this period spans the late stand initiation stage to the early stem exclusion stage of stand development (Figs 5.2 and 5.3). Because the cost of thinning during this period is usually not recovered in the sale of timber products, the term ‘precommercial’ thinning is often applied. The objective of a cleaning is not to remove all undesirable stems, but to select potential crop trees and to release only those that need it. Opportunities for precommercial thinning occur in the early stem exclusion stage of stand development when trees are sapling-size and tree crown positions are well established (Fig. 7.17A). Young high-density stands of desirable species composition facilitate selecting crop trees at an optimum spacing from among the best individuals (Fig. 7.17B). A spacing of 15–20 ft between crop trees usually provides acceptable stocking by the time trees are large enough for a commercial thinning. Smaller trees are unlikely to grow fast enough to keep up with the rapidly developing main canopy. Because cleaning is expensive and costs are carried for long periods at compound interest, the application of the method is usually limited to the better sites. There nevertheless is much disagreement and uncertainty over the efﬁcacy of precommercial thinning. For example, results from thinning West Virginia oak stands during the sapling stage (at 15 years) indicated there was little advantage to precommerical thinning (Carvell, 1971). The thinnings adjusted spacing, favoured the desired species and removed poorly formed trees. Ten years later when plots were compared to plots receiving no thinning, the cleaned plots were only slightly better in composition. Even when signiﬁcant improvements in growth, quality and species composition result, carrying the costs of cleaning at compound interest over most of the rotation often renders the practice economically unfeasible. Exceptions may occur where there are markets for small products. Cleanings applied to stands in the stand initiation stage of development (Fig. 5.3A) may be even more problematic. During that

period, crown classes are changing rapidly and unpredictably. This was observed in a West Virginia study (red oak site index 70 ft) where removing competitors within a 5-ft radius of 9-year-old crop trees did not significantly increase height or diameter growth or length of clear stem, nor did it prevent crown-class regression during the 5-year study period (Lamson and Smith, 1978). Similar results were observed in 7-year-old stands (oak site index 60 ft) (Trimble, 1974). When cleanings are applied to stands in the sapling stage of development, recommendations are to do so conservatively by creating canopy gaps no more than 5 ft wide on three or four sides of selected crop trees (Sampson et al., 1983). Despite these problems and uncertainties, cleanings can potentially increase yields. In the Central Hardwood Region, cubic-foot yields of oak stands can be increased by 50% or more when thinning is begun at age 10 (a cleaning) than when begun at age 60 (Gingrich, 1971). Although economic considerations may preclude cleanings, oak stands often can be profitably thinned as early as stand age 25 where there is a market for fuelwood or pulpwood (Carvell, 1971). Even if returns are slightly less than costs, thinning may be justiﬁed by the early acceleration of the growth of crop trees. Early thinning therefore should not be focused on early proﬁts, but on increasing the growing space for the ﬁnal crop trees. Not only will the growth rates of residual trees be increased, but early thinning also provides the opportunity to select trees of the best species and highest quality without the largely unrecoverable expense incurred in cleanings. The ﬁrst thinning should remove undesirable species and poorly formed stump sprouts and other trees. Additional trees may need to be cut to improve spacing among dominant and co-dominant trees. Even when the ﬁrst thinning is compelled by an economic return, the primary goal should be the long-term improvement of stand composition, structure and growth. A maxim of thinning is never to compromise the quality and growth potential of the stand.

Even-aged Silvicultural Methods

297

A

B

Fig. 7.17. (A) An unthinned sapling-size oak stand in southeastern Ohio (Central Hardwood Region). Stocking is at or near 100% and comprised of many oaks of good form. Such stands are good candidates for precommercial thinning. (B) A thinned sapling-size mixed upland oak stand in southeastern Ohio thinned to a prescribed density. Only the largest and best-formed crop trees were retained on this research plot. This is an area-wide thinning with very uniform spacing between trees. In practice, precommercial thinnings are usually not applied area-wide as illustrated here. Rather, thinning is restricted to a prescribed distance around selected crop trees that leaves intervening unthinned areas. (USDA Forest Service, North Central Research Station photographs.)

298

Chapter 7

The ﬁrst thinning is thus the ﬁrst step in accelerating the growth of crop trees. Periodic thinnings are thereafter needed to sustain high growth rates. However, thinning should not reduce stand density below full utilization of growing space (e.g. as deﬁned by regional stocking guides). Cutting to lower densities creates unused growing space and thus reduced yields, may encourage an undesirable understorey and may reduce tree quality by encouraging epicormic branching on tree boles (Fig. 7.18). Even though maximum yields may occur when stocking is below B-level (Dale, 1972; Leak, 1981), the lower limit of thinning is usually set at B-level to ensure bole quality (Dale and Sonderman, 1984). Thinning below B-level risks increasing epicormic branching on the lower bole with consequent loss of tree value. Even light thinnings reduce competition for growing space and thereby reduce the natural pruning of trees. However, the number of branches a tree retains is determined partly by genetics and partly by its growth rate coupled with competition and side shading. Accordingly, selecting crop trees should favour those with clean boles and rapid growth. Stand density also can potentially inﬂuence other bole characteristics such as forking and stem taper. Nevertheless, there was no evidence that thinning over a wide range of stocking levels from <50 to >75% (based on Fig. 6.9) signiﬁcantly affected stem taper of dominant and co-dominant oaks during a 12year study period in the Central Hardwood Region (Hilt and Dale, 1979). A different strategy for applying intermediate cuttings is recommended for northern red oak stands in central New England. In this region, where red oak site index ranges from about 55 to 75 ft, red oak commonly occurs in association with red maple and birch (Oliver, 1978, 1980; Hibbs and Bentley, 1983). Although the maple and birch typically outnumber and outgrow the oaks during the ﬁrst two decades of stand life, red oak ultimately emerges to form the dominant canopy because of its greater survival and sustained growth rates (Oliver, 1978; Hibbs, 1981). Only about

Fig. 7.18. The lower bole of this black oak is covered with ‘epicormic branches’. They originated from dormant buds under the bark that ‘sprouted’ when the tree was suddenly exposed to high light intensity after heavy thinning. Epicormic branches that persist and grow reduce bole quality and tree value. Maintaining adequate lateral shading of boles by controlling stand density, and thus light, is essential to minimizing epicormic branching. Because the propensity for epicormic branching in oaks is inherited, crop trees should be selected from among those with few such branches at time of thinning. Oak stump sprouts also arise from dormant buds. But stump sprouts originate at or near the tree base after the tree is cut or top-killed and then develop into tree boles (Fig. 2.22). (USDA Forest Service, North Central Research Station photograph.)

twice the number of red oak crop trees required at the end of the rotation are needed at stand age 25 for the stand to be considered adequately stocked with red oak (Hibbs, 1981). During the ﬁrst 45 years, rapid height growth and the development of long, branchless boles are encouraged by maintaining high stand densities. Although early thinning increases diameter growth,

Even-aged Silvicultural Methods

merchantable height is reduced. The resulting value lost to reduced merchantable height is not compensated by gains in diameter growth (Hibbs and Bentley, 1984). First thinnings are therefore delayed to stand age 45. Stands are then thinned to Blevel stocking (Sampson et al., 1983) every 10 years until shelterwood cuttings are made to obtain oak reproduction (Hibbs and Bentley, 1983). The recommended rotation, assuming current utilization technology, is about 95 years when diameters of individual trees reach ﬁnancial maturity, which occurs at 21–25 inches dbh depending on site quality (Hibbs and Bentley, 1984). Thinnings and improvement cuts thus are not continued indeﬁnitely. As trees increase in size and cuttings remove only a few trees per acre, relatively few trees in the residual stand will beneﬁt from thinning. Late in the rotation, cutting only a few trees per acre may harvest periodic growth but will increase the growing space of only a few residual trees. This largely defeats the intent of thinning. A practical rule is to terminate thinnings for the purpose of stand density control by the time a stand reaches 75% of rotation age (Roach and Gingrich, 1968). For upland oaks in the eastern United States, this corresponds to about 60–70 years on average sites, and less on better sites. Guidelines for thinning even-aged oak stands are presented in Roach and Gingrich (1968). Although these guidelines are speciﬁcally designed for the Central Hardwood Region, their research-based development makes them broadly applicable in principle to many other kinds of oak forests. Included in the guidelines is a decision key for both thinning and regenerating Central Hardwood oak stands. The effects of thinning on the growth and yield of oak stands are further discussed in Chapter 10.

Tending oak coppice (stump sprouts) In some species such as northern red and scarlet oaks, multiple stems often persist within clumps of stump sprouts for decades (Figs 2.26 and 7.19A). Northern

299

red oak clumps often retain four or more live stems for 20 years and longer (Fig. 2.25). As coppice stems grow larger, within-clump competition increases and reduces the diameter growth of individual stems (Johnson and Rogers, 1984). One or more stems within each clump usually occupy and maintain a dominant or codominant crown position. In central Appalachian forests, an estimated 80–90% of northern red oak clumps contain at least one stem of potentially high future value (Lamson, 1976). Thinning within clumps increases the growth rate of the residual stems. The amount of increase depends on species, the initial diameters of the retained stems, clump age, site quality and clump density. The latter effect can be quantitatively expressed by the ratio of basal area of a given stem to total clump basal area (relative crop stem basal area or RBA) after thinning (Johnson and Rogers, 1980, 1984). Thus it is not the number of stems left in a clump, per se, that determines the diameter growth of a crop stem, but other factors being equal, how much within-clump basal area competition the crop stem experiences. Consequently, thinning clumps to one stem (RBA = 1.0) results in maximum diameter growth of any given crop stem. Diameter growth of northern red oak sprouts can be substantially increased by early clump thinning (Johnson and Rogers, 1980). In unthinned clumps on good sites, a stem that is 1 inch dbh at age 5 will grow to about 6.5 inches dbh at age 25; the same stem in a clump thinned to one stem at age 5 will grow to about 10.8 inches dbh – or 1.7 times larger. Estimated 25-year diameters indicate that early clump thinning to one stem is likely to produce economically mature stems 20 or more years earlier than trees of seedling or seedling sprout origin. For a given clump age, the largest stem within a clump usually produces the greatest growth over a given time interval. However, the largest stem may not always be the most desirable stem when stem quality and the preferred low basal attachment to the stump are considered. Good diameter growth from smaller stems can be

300

Chapter 7

A

B

C

D

Fig. 7.19. (A) A 75-year-old northern red oak coppice stand in the Driftless Area of southwestern Wisconsin (Central Hardwood Region); site index for northern red oak is 70 ft. The stand has never been thinned and is largely comprised of multiple-stemmed clumps. (B) A 20-year-old clump of northern red oak sprouts. All six living stems in this clump are seriously degraded by ‘sweep’ (stem curvature); the three stems on the right are fused at the base. (C) Cross-section of the base of a 20-year-old northern red oak stump sprout. The two larger stems are fused together but still remain separated by bark; they envelop a small dead stem. (D) In this 20-year-old northern red oak clump, the stump on the left enveloped three sprouts that fused together and developed a common cambium. (USDA Forest Service, North Central Research Station photographs.)

Even-aged Silvicultural Methods

obtained when clumps are thinned early to one stem. For example, in 5-year-old clumps, stems with diameters one-third that of the largest stem will on average be only 6% smaller 20 years later. So if thinned early, stems substantially smaller than the largest stem can be selected as crop stems without sacriﬁcing much growth. Conversely, stems that are considerably larger than the average dominant stem for a given age may respond little to clump thinning because their RBAs before thinning are already close to 1.0. The longer thinning is delayed, the greater the reduction in expected 25-year diameters. For example, for average size stems on average sites, delaying thinning from age 5 to age 10 will reduce 25-year diameter by about 12%; delaying thinning to age 15 or 20 will reduce 25-year diameter by 23% and 30%, respectively (Table 7.5). To sustain maximum growth of thinned sprout clumps, competitors outside the clump should be removed about every 10 years. A simple guideline for

determining an appropriate thinning radius around a crop tree is provided by the ‘D + 1’ rule, i.e. the crop stem dbh in inches plus 1 where each inch of dbh equals 1 ft radius. This distance approximates the maximum tree area for upland oaks in the Central Hardwood Region (Gingrich, 1967) and therefore ensures a crop stem approximately the maximum growing space it can use at the time of thinning. A more accurate spacing for a crop tree of any given dbh and desired residual stand density can be obtained by applying a method called ‘rule thinning’ (Rogers and Johnson, 1985). Improvements in stem quality as well as increases in growth are generally needed to justify investments in precommercial clump thinning. Early clump thinning can eliminate or minimize many of the defects that later develop in oak coppice. The most common defects are: • Decay. • Seams or radial shakes. • Poor stem form caused by sweep.

Table 7.5. Effects of delaying thinning of northern red oak sprout clumps based on estimated 25-year diameters.a Stem dbh at age 5 (inches) Site index 50b 0.5 1.0 1.5 2.0 Site index 60b 0.5 1.0 1.5 2.0 Site index 70b 0.5 1.0 1.5 2.0 aAdapted

Estimated 25-year dbh (inches) when clumps are thinned to one stem at age: 5 10 15 20 8.3c 8.5 8.8 9.1

Average unthinned clumps

7.1 7.6 8.1 8.8

6.1 6.8 7.7 8.4

5.2 6.3 7.2 8.2

4.4 5.8 6.9 8.0

9.4 9.7 10.0 10.3

7.9 8.5 9.1 9.9

6.6 7.5 8.5 9.5

5.4 6.8 8.0 9.2

4.4 6.1 7.6 8.9

10.6 10.8 11.2 11.6

8.7 9.4 10.2 10.9

7.1 8.2 9.4 10.6

5.7 7.2 8.7 10.3

4.7 6.5 8.3 10.0

from Johnson and Rogers (1980, 1984). feet at an index age of 50 years; based on the site index curves of Gevorkiantz (1957). cThe values given assume that stand density is maintained at a level that provides the maximum amount of growing space a tree of a given diameter can utilize. bIn

301

302

Chapter 7

• Branching and forking in the lower bole. • Weak attachment of sprout to the root system. DECAY In oak coppice, the two primary entry points of decay-causing organisms are through heartwood connections with the parent stump or with dead companion stems (Fig. 7.19C). Heartwood connections between sprouts and stump can be minimized by selecting crop stems that originate at or below ground line (Roth and Hepting, 1943, 1969). Decay entry through companion stems can be largely eliminated by thinning clumps to one stem before the largest stems exceed about 2 inches dbh (about 10 years old or less in most eastern oak forests). In sprouts larger than this, the point of sprout origin is often obscured and the faster growing stems begin to fuse with and envelop slower growing stems. Fusion of oak stems in unthinned clumps eventually produce heartwood unions (Fig. 7.19D). Removing one stem from such a union consequently creates an entryway for heartwood decay organisms into remaining stems (Roth and Sleeth, 1939). Stems forming V-shaped unions are apt to have heartwood unions, especially after fused stems reach 3 inches dbh; stems with U-shaped unions are less likely to have heartwood unions. Stems forming V-shaped unions therefore should not be selected as crop stems when clumps are thinned (Stroempl, 1983). Even when decay develops, it tends to be restricted, or compartmentalized, by a ‘barrier zone’ formed by the cambium after wounding (Shigo, 1979). Because of compartmentalization, the potential for spread of decayed wood in oak coppice is limited. Dead companion stems and basal stubs nevertheless predispose crop stems to other defects such as seams. A precaution applicable to oaks in the red oak group is to avoid clump thinning during spring and early summer in areas of known oak wilt activity. The stumps created by thinning can become entryways for oak wilt disease into interconnected crop stems (Kuntz and Drake, 1957; also see the section on oak wilt, later in this chapter).

SEAMS Seams, also known as radial shakes, are separations of wood tissue along the radial plane of living trees (Butin and Shigo, 1981). In stump sprouts, radial shakes are commonly associated with the occurrence of dead stems. Shakes originate from barrier zones that create planes of structural weakness along which radial shakes can be initiated. Stresses such as wind and sudden temperature drops may cause further radial separations and result in obvious external vertical seams or ‘frost cracks’. These usually occur in the lower 6 ft of the bole (Fig. 7.20). Radial shakes are serious defects because the separations often continue for the life of the tree – and they are only rarely

Fig. 7.20. A seam or radial shake on the lower bole of a northern red oak. Such defects are common in oaks and seriously reduce tree quality because the cracks extend from the centre to the outside of the bole and are seldom overgrown, i.e. they usually ‘grow’ with the tree. Radial shake is common in oak stump sprouts but can be reduced by early clump thinning to one or two stems. (USDA Forest Service, North Central Research Station photograph.)

Even-aged Silvicultural Methods

A

303

Fig. 7.21. (Opposite) (A) An unthinned 5-yearold northern red oak sprout clump. The more than ten living stems completely envelop the base of the 95-year-old parent tree. (B) Thinning clumps early to one stem (shown here at age 5) maximizes diameter growth by eliminating within-clump competition. This practice also eliminates basal fusion between companion stems and reduces the incidence of decay, seams and sweep. Selected crop stems should be attached to the parent tree at or below the ground line. At this age, most stem quality problems associated with oak stump sprouts have not yet developed. (USDA Forest Service, North Central Research Station photographs.)

ated by wounds, care should be taken not to injure crop stems during clump thinning. Dead branch stubs are also initiation points for radial shake, so early pruning of crop stems may reduce the incidence of seams.

B

SWEEP The crowns of stems in high density clumps have a tendency to lean away from the clump centre causing stem curvature or ‘sweep’ in the lower bole (Fig. 7.19B). Even moderate sweep is considered a serious defect in young hardwoods (Sonderman, 1979). Consequently, stems with sweep should be removed when thinning clumps. The earlier clumps are thinned, the less likely it is that sweep will become a serious problem. BRANCHING AND FORKING IN THE LOWER BOLE

compartmentalized. Consequently, radial shake is a serious problem if high value sawlogs are the desired product. The incidence of radial shake in oak stump sprouts can be reduced by early clump thinning to one or two stems (Fig. 7.21). Where seams already occur, affected stems should be removed during clump thinning. Because radial shake also can be initi-

Branching and forking in the lower bole (i.e. below 17 ft) contributes to serious stem degrade in young hardwoods (Sonderman, 1979). Because low stand density exacerbates both problems, maintaining adequate stand density throughout the life of the stand is critical for growing high quality sawtimber. However, stump sprouts by themselves seldom provide adequate stand stocking. Adequate stocking after regeneration cutting usually requires an abundance of advance reproduction of seedlings and seedling sprouts before ﬁnal harvest. When total stand regeneration potential is adequate, stump sprouts often develop into stems of satisfactory quality because overall stand density promotes natural pruning and minimizes epicormic branching on

304

Chapter 7

boles (Fig. 7.18). Although clump density in unthinned clumps is often high enough to produce at least one stem of potentially high quality (Lamson, 1976), trees around the clump must be relied upon to control epicormic branching and thus crop stem quality after clump thinning. Clump thinning slows natural pruning to some extent (Lamson, 1983). Clear-bole length of thinned sapling-size sprouts of northern red oak 5 years after thinning averaged about 2 ft less than unthinned clumps. However, a combination of clump thinning and manual pruning can offset losses in clear bole length associated with thinning alone. Low stand density or excessive crop tree release also may result in epicormic branching of boles. However, species differ in their propensity to produce epicormic branches. Northern red oak and white oak produce large numbers of epicormic branches whereas red maple produces relatively few; American basswood is intermediate between the oaks and red maple (Trimble and Seegrist, 1973). Clump thinning to one or two stems did not increase epicormic branching of sapling-size northern red oak stump sprouts in West Virginia (Lamson, 1983). Epicormic branching also tends to be greater among intermediate and suppressed than among co-dominant and dominant trees (Trimble and Seegrist, 1973). Clumps that show little propensity for epicormic branching before thinning are the preferred candidates for clump thinning (Ward, 1966). Epicormic branching of boles may appear to be more damaging than it actually is for the following reasons: • Most sprouting occurs on the upper stem, but the greatest volume and value of the tree are usually in the lower log. • Intermediate and overtopped trees sprout the most, but the most valuable trees are dominant and co-dominant. • Many small epicormic branches are short-lived and therefore do not cause signiﬁcant degrade (Trimble and Seegrist, 1973).

WEAK BASAL ATTACHMENT Despite care in selecting crop stems that originate low and thus appear to have sound basal attachments, some of those stems may eventually break off at the base. In the overall population of coppice stands, the most common cause of stem mortality (not to be confused with clump mortality) is stem breakage at the point where a stem is attached to the stump (Lamson, 1983). In part, this is a natural consequence of clump self-thinning. However, the unique characteristics of coppice-origin stems produce a self-thinning dynamic unlike that occurring between individual trees. For example, stems within sprout clumps often break because of constricted vascular connections caused by included stump bark, which can girdle the growing sprout (Wilson, 1968). Although the attachments may appear to be strong when viewed externally, the interior vascular connections may be small and weak. Although there is no practical way to non-destructively identify stems with this defect, loss of selected crop stems due to such breakage appears to be a minor problem in most oak coppice stands (Johnson and Rogers, 1980; Lamson, 1983). Before thinning oak sprout clumps, factors related to the timing of thinning, residual clump density, stand selection, crop stem selection and overall stand density should be considered according to the following general guidelines. TIMING OF THINNING AND RESIDUAL CLUMP DENSITY Thin to one or two stems per clump as early as practical and preferably before age 10 (Fig. 7.21). STAND SELECTION Select stands that are on average or better sites that are well stocked; neither poor sites nor poorly stocked stands will yield the high value trees needed to justify the expenditure of precommercial thinning. Thin clumps before crop stems reach 3 inches dbh (about age 15 in many upland eastern forests) and preferably before they reach 2 inches dbh (about age 10). Beyond 3 inches dbh, options in crop stem selection are few and most of the potential growth gain from clump thinning is lost.

Even-aged Silvicultural Methods

STEM SELECTION In clumps with potential crop stems 3 inches dbh or less, select as the crop stem the largest single stem that is well attached to the stump at or below ground line and that is not connected to another stem. For stems between 2 and 3 inches dbh, discriminate against those with sweep, crook and seams. For smaller stems, consideration of stem quality factors is probably neither necessary nor meaningful. If two crop stems are retained, the stems should be far enough apart so that during their expected life span they do not fuse together at a common base (Roth, 1956; Stroempl, 1983) (Fig. 7.19C and D). For stems larger than 3 inches dbh, the lower 17 ft should: (i) be free of sweep, crook, seams, forks and visible decay; (ii) be free of V-shaped connections with other stems; and (iii) show little evidence of epicormic branching. These factors being equal, select the stem with the fewest branches in the lower bole.

CROP

THINNING AROUND CROP STEMS Growth and quality can be optimized by thinning around crop stems to provide adequate growing space but at the same time leaving enough shade to minimize epicormic branching and forking, and to promote the development of straight lower boles. Appropriate thinning radii can be derived from known maximum growing space requirements for each species (Chapter 6). Subsequent thinnings (e.g. at 10-year intervals) around each crop stem will be required to maintain near-maximum growth throughout the rotation. Such croptree thinning also can be integrated with area-wide thinning to obtain a relatively uniform overall stand density (Rogers and Johnson, 1985). Despite the potential tree growth and quality advantages provided by clump thinning, economic returns may not justify the practice. For example, an economic analysis of a mixed oak stand in the Ozark Highlands (oak site index 63 ft) indicated that the projected response of sprout clumps to thinning to one stem at age 5 was not economically feasible when costs

305

were carried to the end of a 60-year rotation (Dwyer et al., 1993).

Special Problems: Reducing Insect and Disease Impacts Intermediate cuttings and regeneration cuttings in oak stands are sometimes necessitated by insect and disease problems. Among those posing the most serious and widespread threats to oak forests are the gypsy moth, oak decline and oak wilt.

Gypsy moth Throughout much of the oak range in the eastern United States, the gypsy moth is a devastating forest pest that defoliates and damages millions of acres of forest each year. Annual damage and control of the insect costs tens of millions of dollars (USDA Forest Service, 2001). The insect is native to Europe and Asia and was accidentally introduced near Boston, Massachusetts, in the late 1860s. It has since expanded its range southwesterly and westerly so that it now extends over much of the eastern United States (Fig. 7.22). Defoliation is caused by the larval (caterpillar) stage of the gypsy moth (Fig. 7.23A). Defoliation has many consequences including tree mortality, loss of tree growth, undesirable changes in the species composition of forests, decreased aesthetic quality of forests, reduced water quality and the nuisance of numerous caterpillars around homes in affected areas (Campbell and Sloan, 1977; Twery, 1991; Liebhold et al., 1995a; USDA Forest Service, 2001). Each new generation of the gypsy moth begins during late summer when the females lay clusters of eggs, usually on tree trunks, where they remain through the winter (Leonard, 1981). The eggs hatch in the spring to form the ﬁrst larval stage, or instar. The newly hatched larvae are about 3 mm long, light weight, and have long hairs that greatly increase their surface area. The larvae leave the vicinity of the eggs and move upward into tree crowns to feed. As they move, they

306

Chapter 7

1998

2010

Fig. 7.22. The distribution of the gypsy moth in the United States in 1998 (light shading up to heavy solid line). The dark shading represents the predicted transition zone (i.e. boundary area of discontinuous occurrence). The predicted 2010 distribution assumes that eradication and suppression practices are applied. The insect is also present in portions of eastern Canada. The current rate of spread is about 13 miles year−1. This spread is attributed to natural movement of wind-borne larvae in their first stage (instar) of development and the accidental dispersion of insects by humans. (USDA Forest Service map.)

spin silken threads to which they remain attached. The young larvae hang from these threads, which break when caught by the wind. The suspended larvae are then blown aloft and disperse to other trees. The larvae feed on the foliage of trees and other vegetation during the night and rest in bark crevices and other protected places during the day. Exceptions occur during population peaks, when feeding occurs all day. The larvae go through ﬁve to six moults, which occur at about weekly intervals. As the larvae grow, their consumption of foliage increases exponentially. During the last instar they eat more than during all the other stages combined. By the end of all larval stages, a single larva has consumed about 11 ft2 (1 m2) of foliage, has increased in length to about 50–90 mm, and in weight by more than 1000-fold (Leonard, 1981). The end of the 8-week larval stage marks the beginning of the pupal stage. The pupae live in a sparse silken cocoon on tree trunks for about 2 weeks and then emerge in midsummer, usually in July, as adult moths (Fig. 7.23B). Mating and egg laying occur during the approximately 1-week life span

of the adult. The male moth locates the female with a pair of feathery antennae, which detects the sex pheromone emitted by the female. Although the female is unable to ﬂy, wind dispersal during the ﬁrst instar together with inadvertent dispersal of various life stages via human transport are effective mechanisms of range expansion (Leonard, 1981). Gypsy moth populations are cyclical. Extended periods of low insect density are followed by rapid population explosions, which may last for several years. A population crash then follows (Montgomery and Wallner, 1988). When insect densities attain high levels, susceptible trees are often completely defoliated. Several successive years of defoliation coupled with other associated biological and physical stress factors often result in tree mortality. Although less than 20% of defoliated trees usually die from an infestation, localized mortality can be extremely high. Despite suppression and eradication programmes by federal and state agencies, which have been effective in slowing the spread of the gypsy moth, the insect continues to expand its range. Not only do immediate economic

Even-aged Silvicultural Methods

A

B

Fig. 7.23. (A) Gypsy moth larvae (last instar) feeding on an oak leaf. (B) Adult female gypsy moths. The adults live for about a week during which time they mate and lay their eggs on tree bark and other objects. Compared to the white or cream coloured, flightless females, the male moths are mottled brown with black wing markings, have a slimmer abdomen and can fly. (USDA Forest Service photos.)

losses result from gypsy moth defoliations, but often there are associated long-term changes in forest composition away from the oaks (a preferred host), and towards less valuable species. Such changes in species composition reduce acorn production, exacerbate the oak regeneration problem, negatively affect animal populations

307

that depend on acorns, and thus reduce forest biodiversity. Although the larvae are capable of feeding on the foliage of hundreds of North American plant species (Liebhold et al., 1995b), the most common hosts within the insect’s current range are oaks and aspens (Gansner and Herrick, 1985). Tree species that are readily eaten by gypsy moth larvae during all larval stages are categorized as susceptible, whereas species eaten when preferred foliage is not available or only by some larval stages are categorized as resistant. Susceptible (preferred) hosts include the oaks and many associated hardwoods (Table 7.6). Immune species are deﬁned as those that are rarely or never eaten. Shrubs and herbaceous plants are also eaten by the gypsy moth and they also vary in susceptibility. Because oaks rank among the preferred hosts of the gypsy moth, stands with a high proportion of oaks are susceptible to heavy defoliation (Gottschalk, 1991). However, not all oaks are equally susceptible. For example, chestnut, black and scarlet oaks are more susceptible than northern red oak (Gansner and Herrick, 1985). Guidelines for evaluating the susceptibility of forests to defoliation are based on the percentage basal area of oaks and other preferred (susceptible) hosts (Gottschalk, 1993). In all species, heavy defoliation may cause trees to refoliate during the same year, which in turn depletes carbohydrate reserves and reduces tree vigour. Weakened trees are then vulnerable to attack by secondary organisms such as Armillaria root disease and the twolined chestnut borer (Dunbar and Stephens, 1975; Wargo, 1977; Gansner and Herrick, 1985; Gottschalk et al., 1989; Burrill et al., 1999). Trees at highest risk are those in suppressed (overtopped) and intermediate crown classes and within a given crown class, trees with poor crowns (Herrick and Gansner, 1987). The most signiﬁcant economic effect of the gypsy moth on oaks is not mortality, but reduced growth, yield and wood quality. Gypsy moth outbreaks occur at 7–10 year intervals (Gottschalk et al., 1989). It may take 3 years for stand growth to recover from a single defoliation (Twery,

308

Chapter 7

Table 7.6. Gypsy moth susceptibility ratings for some common overstorey and understorey species of the eastern United States.a Susceptible: Overstorey: Understorey: Resistant: Overstorey:

Understorey:

Immune: Overstorey:

Understorey:

Species readily eaten by gypsy moth larvae during all larval stages American basswood, bigtooth aspen, grey birch, mountain-ash, oak, paper birch, pear, quaking aspen, river birch, sweetgum, tamarack, willow American hazel, eastern hophornbeam, hawthorn, speckled alder, witch-hazel Species eaten when preferred foliage is not available and/or only by some larval stages American beech, American chestnut, American elm, black cherry, black walnut, boxelder, butternut, eastern cottonwood, eastern hemlock, hickory, Norway maple, pine, red maple, sassafras, spruce, sugar maple, sweet birch, sweet cherry American hornbeam, beaked hazel, blackhaw, chokecherry, flowering dogwood, pawpaw, serviceberry, sourwood Species rarely eaten American holly, American sycamore, ash, baldcypress, balsam fir, blackgum, black locust, cucumbertree, eastern redcedar, Fraser fir, hackberry, honeylocust, horsechestnut, Kentucky coffeetree, northern catalpa, Ohio buckeye, pin cherry, red mulberry, silver maple, slippery elm, yellow buckeye, yellow-poplar alternate-leaf dogwood, American elder, common juniper, common persimmon, grey dogwood, mountain maple, nannyberry, possumhaw, redbud, red-osier dogwood, roundleaf dogwood, spicebush, striped maple

aThe

list includes commonly occurring species in each category and is not exhaustive for any category. Adapted from Gottschalk (1993) and Liebhold and others (1995).

1987). For two defoliations per decade, loss of volume increment has been estimated at about 10%, exclusive of mortality. For two defoliations per 5 years, estimated growth losses are about 19% (Twery, 1987). However, such losses may be offset by subsequent gains in growth of the survivors resulting from the thinning effect. To minimize loss in value of sawtimber-size oaks killed by defoliation, trees should be salvaged within the ﬁrst year after death (Garges et al., 1984).

Control It has long been proposed that silviculture can be used to reduce the susceptibility (defoliation potential) and vulnerability (tree mortality) of forests to gypsy moth defoliation (Fiske, 1913; Clement and Munro, 1917; Behre, 1939; Bess et al., 1947; Gottschalk, 1993). Although this premise has been questioned for lack of a 1Current

rigorously developed scientiﬁc basis (Muzika and Liebhold, 2000), there nevertheless is evidence that certain silvicultural practices are effective in reducing stand susceptibility and vulnerability (Gottschalk, 1989; Liebhold et al., 1998; Grushecky et al., 1998; Burrill et al., 1999; Davidson et al., 1999). Silvicultural guidelines for coping with the problem accordingly have been developed (Gottschalk, 1993). These guidelines are based on the presumed or apparent beneﬁts of altering stand composition to favour less susceptible species, improving the vigour of the stand by thinning, and removing trees via pre-outbreak thinnings (sanitation cuttings) that are likely to die from defoliation. These guidelines are summarized below. However, the concerned reader is urged to review the guide itself, together with other current information on the gypsy moth.1 Three categories of defoliation immi-

information on the gypsy moth is available from the USDA Forest Service gypsy moth web site at http://www.fs.fed.us/ne/morgantown/4557/gmoth

Even-aged Silvicultural Methods

nence are considered in the application of silvicultural guidelines: • Defoliation not imminent within 1–3 years. • Defoliation imminent within 1–3 years or now occurring. • Defoliation has occurred. The ﬁrst category usually applies to stands located outside generally affected areas, whereas the last two categories usually apply to stands within affected areas (Gottschalk, 1993). Forest pest management specialists can be consulted on the location and size of the area currently infested, expected trends in insect populations, and the time when the infestation is likely to occur in a given stand or locale. After gypsy moth populations have been established in an area, inventories of egg masses per acre are used to determine potential population levels. Egg masses are evident in infested areas for 7–8 months, which allows time for forest managers to plan and apply silvicultural treatments before outbreaks occur. After the imminence of defoliation is ascertained, a stand inventory and analysis is required to determine stand susceptibility to defoliation. In this context, susceptibility is deﬁned as the probability of defoliation by gypsy moth given that the insect is present (Gottschalk, 1993). Stand susceptibility can be rated based on the factors that affect defoliation potential. These include the percentage of stand basal in oaks, the percentage in highly preferred oaks (e.g. black and chestnut oaks), and average stand dbh and crown condition (Fig. 7.24). Stands with a large proportion of highly preferred hosts for gypsy moth, especially when the crowns of those species are large and in good condition, are attractive to feeding insects. Potential defoliation in such stands is therefore high (Fig. 7.24). Other factors also inﬂuence stand susceptibility. Susceptible sites are associated with places where tree growth is slow such as shallow or sandy soils, dry ridge tops and wherever drought stress is frequent. Some stages of larval development

309

ﬁnd refuge in speciﬁc tree features such as bark ﬂaps, large dead branches, bole wounds and holes, dead sprout stubs, deep bark ﬁssures and other protected sites (Bess et al., 1947; Campbell et al., 1975a,b; Gottschalk, 1993). Stand susceptibility therefore can be reduced by removing these structural features. However, recommendations are to concentrate removals of such refuges that are high in trees before reducing those low in trees or on the ground where they may provide predator habitat (Gottschalk, 1993). Susceptible sites also have fewer small mammal predators such as mice and shrews and are less desirable habitats for those animals than the less susceptible sites (Yahner and Smith, 1991). Environmental stresses that can increase susceptibility include damage associated with frost, other insects and logging. Bands of susceptible forests often develop around residential areas, recreational areas, roads, rights-of-way and other areas at the interface of forests and human activity. The associated disturbances can create gypsy moth refuges and stresses in trees that increase their susceptibility to defoliation (Campbell et al., 1976; McManus and Houston, 1979; Gottschalk, 1993). Crown condition (Table 7.7) is an important determinant of stand susceptibility and is a better indicator of a tree’s capacity to recover from defoliation than is crown class (or tree size) (Gottschalk and MacFarlane, 1993). Crown condition is therefore a determinant of stand vulnerability, which is deﬁned as the probability of occurrence of stand damage given that defoliation has occurred (Gottschalk, 1993). The criterion for ‘damage’ is usually based on tree mortality. Other factors being equal, defoliated trees in subordinate crown classes are more likely to die than those in dominant or co-dominant classes. However, trees in co-dominant or dominant crown classes with poor crowns are more likely to die from defoliation than trees in the intermediate crown class with a good crown (Campbell and Valentine, 1972; Campbell and Sloan, 1977; Crow and Hicks, 1990).

310

Chapter 7

Defoliation potential

% BA in good crown condition

≥70% % BA in chestnut and black oaks

% BA in good crown condition

<70%

≥60%

40%

<60%

31%

≥30%

29%

<30%

22%

≥40%

32%

10–39%

23%

≥7 inches

24%

<7 inches

15%

20–49%

18%

<20%

9%

≥50% % BA in good crown condition

≥40% % BA in chestnut and black oaks

% BA in oak

<40%

Mean stand dbh (inches)

<50%

Defoliation level Heavy Moderate Light

≥30% 20–29% <20%

Fig. 7.24. Estimated gypsy moth defoliation potential based on a central Pennsylvania model. Defoliation predictors are percentage of stand basal area (BA) in oaks and the percentage in black and chestnut oaks, percentage of BA in good crown condition and average stand dbh. (Adapted from Herrick and Gansner, 1986.)

There are several regional models for estimating individual tree mortality associated with gypsy moth defoliation. These include the Ridge and Valley area of Pennsylvania (Herrick and Gansner, 1987), New England (Campbell and Valentine, 1972; Valentine and Campbell, 1975) and the Pocono Mountains of Pennsylvania (Herrick, 1982; Gansner and Herrick, 1984). The models differ in the predictors used to estimate mortality. For example,

the New England model uses species, crown condition, crown class, defoliation intensity and defoliation frequency to estimate the probability of mortality. The Poconos model is based on crown condition, species and slope aspect as determinants of vulnerability (Fig. 7.25). Stand-level mortality models have been developed for New England (Valentine and Campbell, 1975), the Pocono Mountains, New Jersey (Kegg, 1974) and the central

Even-aged Silvicultural Methods

311

Table 7.7. Crown condition classes for ranking risk of tree mortality to gypsy moth defoliation.a Crown condition class/ (mortality risk)

Crown characteristics

Poor (high)

≥50% of crown branches are dead; foliage density, size and colouration are subnormal; or epicormic bole sprouting is heavy

Fair (moderate)

25–49% of crown branches are dead; foliage density, size and colouration are subnormal; or some epicormic bole sprouting is evident

Good (low)

<25% of crown branches are dead; healthy foliage; little or no epicormic bole sprouting

aAdapted

from Gottschalk (1993) and Gottschalk and MacFarlane (1993).

Crown condition

Good

Fair

Poor

Species

Aspect

Species

Other than white oak group

White oak group

N, NE, E, SE, or S

SW, NW, W, level

Other than oaks

Oaks

2%

9%

11%

43%

62%

86%

Fig. 7.25. Estimated probabilities of defoliation-related mortality based on model for the Pocono Mountains of Pennsylvania. (Adapted from Herrick, 1982, and Gansner and Herrick, 1984.)

Appalachian Ridge and Valley area (Crow and Hicks, 1990). Both individual tree and stand-level models require stand inventories that include the information necessary for mode implementation. Stand susceptibility and vulnerability ratings together with other stand information facilitate the application of silvicultural decision-making guidelines and

prescriptions (Gottschalk, 1993). Some prescriptions for coping with the gypsy moth are preventive whereas others are designed to minimize timber losses after gypsy moth defoliation has occurred. Preventive silvicultural measures include presalvage cuttings, which are designed to anticipate and minimize the impact of defoliation. Presalvage cuttings focus on

312

Chapter 7

reducing stand vulnerability by harvesting highly vulnerable trees before they are defoliated and die. Gottschalk (1993) deﬁned three types of presalvage cuttings applicable to stands susceptible to gypsy moth defoliation: presalvage thinning, presalvage harvest and presalvage shelterwoods. Presalvage thinning can be applied to well-stocked stands that are more than 15 years from maturity. Attaining the desired stand condition may require reducing relative stand density (stocking) by 50% or more. This exceeds the usual recommendation not to remove more than 35% of stocking in a single cut (Roach and Gingrich, 1968). However, in stands with more than 50% of basal area in preferred host species, normal thinning prescriptions are unlikely to sufﬁciently reduce stand vulnerability. Presalvage thinning is implemented 1–3 years before defoliation to allow stands to recover from possible (and usually temporary) stresses associated with thinning itself (Gottschalk, 1993). The prescription prioritizes removals (from highest to lowest) according to species and crown condition as follows: • • • •

Oaks with poor crowns. Non-oaks with poor crowns. Oaks with fair crowns. Non-oaks with fair crowns.

These priorities can be integrated with normal thinning procedures for maintaining adequate residual stocking, stand structure and species composition to the extent possible but within constraints set by the gypsy moth control objectives. A presalvage harvest can be used in stands near maturity or that are poorly stocked (e.g. below C-level stocking; Fig. 6.9), but do have adequate regeneration potential. The objective is to harvest the stand in a single cut before defoliation occurs in order to take advantage of the presence of adequate advance reproduction and to preserve stump sprouting potential. A presalvage harvest can be applied to stands where defoliation is occurring or is expected to occur. Where regeneration potential is inadequate, a presalvage shel-

terwood cutting can be used to encourage the development of reproduction. This strategy may be appropriate when stand susceptibility and vulnerability are both low to moderate. Establishment of desirable reproduction may require 5 years or longer. The shelterwood should retain desirable species of low to moderate vulnerability, and trees in good crown condition that are potentially good seed producers. Depending on stand regeneration dynamics, this may require some compromise in the composition of the regenerated stand and accepting fewer oaks. Where the latter is unacceptable, underplanting shelterwoods with oaks (as discussed earlier in this chapter) may be a viable option. Two types of sanitation cuttings have been proposed for stands with high defoliation potentials: sanitation thinning and sanitation conversion (Gottschalk, 1993). By deﬁnition, sanitation cuttings are designed to prevent the spread and establishment of damaging organisms (Smith, 1986). Sanitation thinning, in contrast to presalvage thinning (designed to reduce stand vulnerability), is accomplished by cutting trees to reduce stand susceptibility. The method is applicable to stands where defoliation is occurring or where the imminence of defoliation is several years away. Sanitation thinnings eliminate trees that are current or prospective sources of infestation (Gottschalk, 1993). This process includes removing preferred host species and structural features that provide refuges for the insect, and improving habitat for predators and parasites of the gypsy moth. Sanitation thinning is applicable to stands with less than 50% of basal area in preferred host species. In such stands, the goal is to reduce susceptibility by reducing the basal area of preferred host species to 20% or less (Gottschalk, 1993). Priorities for removing trees are (highest to lowest): • Preferred host species trees with numerous structural features that provide refuges for larvae. • Trees in poor crown condition. • Trees in fair crown condition.

Even-aged Silvicultural Methods

Sanitation conversion cuttings are intended to reduce susceptibility and/or vulnerability by converting stands from preferred host species to more resistant or immune species before infestations occur. They are applicable to highly susceptible or vulnerable stands, which are likely to incur either frequent, heavy defoliation or high mortality. On poor sites with a large oak component, such stands may be amenable to natural conversion to pines or pines mixed with resistant hardwoods. On the better sites, conversion of stands to mixed hardwoods containing a high percentage of resistant species may be practical. In either case, it may be necessary to create a shelterwood to obtain adequate reproduction of desirable species. The method is potentially adaptable to natural and artiﬁcial regeneration or a combination of the two. Although converting an oak stand to other species may reduce stand value, conversion may be the more economically desirable strategy in the long run because it reduces susceptibility and lowers stand protection costs. The gypsy moth guidelines deﬁne ﬁve types of salvage cuttings: salvage thinning, salvage cutting, salvage harvest, salvage shelterwood and salvage conversion. They are applicable to stands where defoliation has already occurred and signiﬁcant mortality has occurred or is imminent. The objective of salvage thinning is to capture the economic value of dead trees and simultaneously to thin remaining live trees to reduce stand susceptibility and vulnerability. Dead trees should be salvaged within a year of their death (Garges et al., 1984). Suitable stands include those with greater than C-level stocking (Fig. 6.9) and which are at least 10 years from maturity, and contain live trees comprising more than 60% stocking. Such stands therefore would be adequately stocked for management to maturity. Thinning the live trees provides an opportunity to improve stand vigour, growth and quality while supporting the economic feasibility of salvaging the dead trees. Recommendations are to remove no more than 35% of live-tree stocking in a single cut. Priorities for tree removal (from highest to lowest) are:

313

• Dead trees. • Oaks with poor crowns that are likely to die. • Other species with poor crowns that are likely to die. • Trees with fair crowns. These priorities can be integrated with the usual thinning objectives of maintaining adequate residual stand stocking, removing less desirable trees before higher quality trees and maintaining a desirable stand structure. The occurrence of dead trees provides opportunities for maintaining or increasing numbers of den trees for mammals and cavity trees for birds. Trees with structural features associated with larvae refuges also can be removed in thinning. Salvage cutting objectives are similar to salvage thinning objectives, i.e. both are aimed at the economic salvage of dead and dying trees. However, salvage cutting is speciﬁcally applicable to stands that are between 30% and 60% stocking of live trees. In this case, the live component of the stand is not thinned because stand density is already at or below full utilization of growing space. In stands below 30% stocking, recommendations are to defer cutting until they can be thinned or re-examined in the future. Cutting priorities are simply to remove dead and dying trees. Options include leaving a few dead trees per acre to provide animal dens and cavities. Regardless of whether stands are given a salvage thinning or are deferred from cutting due to understocking, the residual stands are likely to be less susceptible and less vulnerable to the gypsy moth than before the last defoliation. A salvage harvest is similar to a presalvage harvest, i.e. both are applicable to stands with low stocking (e.g.
314

Chapter 7

shelterwood cuttings. However, the former is applicable to stands in advanced stages of defoliation and mortality. Similarly, the application of salvage conversion cuttings can be used to salvage dead and dying trees and to convert those stands to species not preferred by the gypsy moth. The silvicultural guidelines for forests threatened by the gypsy moth (Gottschalk, 1993) also include recommendations for deferring cutting, stand re-examination and criteria for prioritizing the use of pesticides. However, silviculture is only one aspect of a comprehensive programme of gypsy moth management that also includes eradication, suppression, slowing the spread and biological control of the insect. Eradication involves detecting and destroying new, isolated populations outside the insects’s generally established area, i.e. beyond the ‘front’. Eradication methods vary, but include spraying with chemicals or biological pesticides, mating disruption or mass-trapping of insects (USDA Forest Service, 2001). Direct suppression is used to minimize gypsy moth effects within the established population area. Materials used in suppression include the chemical pesticide diﬂubenzuron (Dimilin®) and biological pesticides including a bacterium (Bacillus thuringiensis) and ‘Gypchek’ (a formulation of a naturally occurring gypsy moth virus) (Podgwaite et al., 1991). Biological control involves the use of natural enemies of the gypsy moth (predators, parasites and diseases) to reduce populations of the insect. Vertebrate predators include mice, voles, shrews, toads and birds including the blue jay and blackbilled cuckoo; invertebrate predators include certain spiders, beetles and parasitoids including wasps and ﬂies (Campbell and Sloan, 1977; Campbell, 1981; Grushecky et al., 1998; Liebhold et al., 1998). Two disease enemies that are already established within the gypsy moth range in North America are the fungal pathogen Entomophaga maimaiga and the nucleopolyhedrosis virus (NPV) (Lewis, 1981; Onken, 1995), from which Gypchek is formulated. E. maimaiga and NPV are

the principal natural enemies that kill large numbers of gypsy moth larvae (Reardon and Hajek, 1993). Another proposed method of biological control is based on reducing the effects of defoliation by introducing or encouraging certain native fungi that aggressively compete with Armillaria for dead tree and stump food reservoirs, but unlike Armillaria do not attack living trees (Burrill et al., 1999). Currently, the USDA Forest Service and its cooperators are working to slow the gypsy moth spread along its 1200-mile front. To do this, pheromone traps are systematically located within the insect’s expanding range to detect isolated populations that develop ahead of the front. These insect colonies are then suppressed or eradicated using the most efﬁcacious methods available (USDA Forest Service, 2001).

Oak decline A malady of oaks called oak decline is also widespread and a potentially serious threat to millions of acres of oak forests. Its effects range from partial crown dieback to tree death. This phenomenon is believed to be initiated by drought stress and/or defoliation, which is subsequently exacerbated by further insect and disease attack (Wargo and Haack, 1991). Maintaining species diversity is believed to be important in minimizing the impact of this stress-mediated disease (Houston, 1987). The most susceptible stands are those of advanced age occurring on drier sites (red oak site index <70ft), and largely comprised of trees in the red oak group (Starkey et al., 1989). Because tree age and site quality are the most important factors that predispose oaks to decline, the ratio of site index to tree or stand age is a useful index of their susceptibility to decline; low ratios are associated with a high susceptibility (Oak and Starkey, 1991). Although much remains to be learned about oak decline, possible preventive or remedial silvicultural treatments include: (i) thinning early to reduce future competition-induced stress and to favour less sus-

Even-aged Silvicultural Methods

ceptible species; (ii) reducing rotation age; (iii) favouring less susceptible species when regenerating stands; and (iv) controlling defoliating insects (Starkey et al., 1989). There is some evidence that thinning stands where decline is already present increases its severity. Possible explanations include an increased food base for Armillaria root disease associated with the root systems of harvested and logging-injured trees, reduced soil moisture related to soil compaction from logging and increased temperatures on the forest ﬂoor related to reduced overstorey density (Starkey et al., 1989). Thinning nevertheless reduced the incidence of oak decline in a Missouri study (Dwyer et al., 1995). Although thinning should theoretically increase the vigour of residual trees, they may be temporarily stressed by their sudden exposure to increased sunlight, wind, raised water tables, winter injury and increased soil temperatures. In turn, these stresses may reduce their resistance to pathogens such as Armillaria root disease before the beneﬁts of thinning occur (Wargo and Harrington, 1991). A silvicultural strategy was developed for mixed oak stands in the Ozark Highlands that are at risk for oak decline (Dwyer and Kurtz, 1994). Stands dominated by scarlet and black oaks of advanced age, which are common in this region, are especially susceptible to oak decline (Wetteroff, 1993; Jenkins and Pallardy, 1995). These stands occur on sites ranging from oak site index 60 to 70 ft (McQuilkin, 1974). The strategy calls for removing oaks with 30% or greater crown dieback as a ﬁrst priority measure, and secondly, removing smaller overtopped scarlet and black oaks. Thinning is deferred to stand age 40 at which time stand density is reduced to 65% stocking based on Gingrich’s (1967) stocking chart (Fig. 6.9). This thinning is expected to remove about 450 board feet of sawtimber and 6 cords of pulpwood per acre. A second thinning is made at stand age 50 by reducing stocking to 64%, which removes about 950 board feet of sawtimber and 2 cords of pulpwood per acre. At stand age 60, a shelterwood

315

harvest that reduces stocking to 44% is recommended. The ﬁnal harvest is at age 63, which is expected to remove about 3300 board feet of sawtimber and 4 cords of pulpwood per acre. The overall strategy is based on comparing various thinning alternatives and is designed to maximize the net present worth of a stand after taking into consideration the risk of declinerelated mortality, returns from thinning, the carrying costs of thinning and ﬁnal crop tree value. Both clearcutting and shelterwood methods are potentially useful for regenerating oak stands of advanced age that are decline-susceptible or that are already affected by decline. Clearcutting may be the more appropriate choice in stands largely comprised of red oaks and where the current stand regeneration potential is sufﬁcient to meet management objectives, or supplemental oak planting is feasible. In stands already affected by oak decline, the stump sprouting potential of the overstorey has already been reduced. Expected contributions to future stocking from oak stump sprouts therefore need to be proportionately reduced to account for the number of decline-affected trees (see Fig. 2.24 and also details of regeneration models applicable to clearcutting, this chapter). Trees on the edges of clearcuts may be predisposed to oak decline because of stresses related to sudden crown exposure and possible damage to roots from harvesting machinery (Mason et al., 1989; Wargo and Harrington, 1991). Where advance reproduction is inadequate, application of the shelterwood method may be appropriate. This assumes that the shelterwood itself is of adequate density, composition, and of sufﬁciently low susceptibility to decline to meet regeneration objectives (see ‘The shelterwood method’, this chapter). Many of the silvicultural strategies for coping with gypsy moth as discussed in the preceding section of this chapter are applicable in principle to stands at risk for, or already in, oak decline. These include the prescriptions for thinning, regeneration, salvage and sanitation cuttings.

316

Chapter 7

Oak wilt Oak wilt is a potentially serious disease of oaks in the eastern United States. The disease is caused by a fungus that disrupts the oak’s vascular system and causes leaf wilting and tree death. Although species in the red oak subgenus are the most vulnerable, the white oaks are not immune. The disease occurs throughout the Central Hardwood Region, the Forest–Prairie Region, parts of the Northern Hardwood Region and to a limited extent in the Southern Pine–Hardwood Region (Gibbs and French, 1980). The disease is spread through root grafts that connect neighbouring trees, and overland by certain sap-feeding insects such as nitidulid beetles and tree-wounding insects such as oak bark beetles. The oak bark beetles are believed to be the most important vectors of the oak wilt fungus throughout the known range of the disease (Rexrode, 1977). The nitidulid beetles also may be important vectors in some areas, especially where spore-producing mats frequently form on diseased trees (Dorsey et al., 1953; Gibbs and French, 1980). Red oaks usually die within a year of infection, and often within a few days to 2 weeks after wilting becomes visible (Bruhn et al., 1991). The severity of the disease varies geographically and is greatest in the northwestern part of its range. Its rapid expansion in parts of Minnesota and Wisconsin during the last 60 years may be related to relatively new and highly susceptible host populations of northern pin oak (Gibbs and French, 1980). Northern pin oak commonly forms pure or nearly pure stands of coppice origin on sandy outwash plains and associated glacial landforms that are characterized by frequent droughts and low soil fertility. Many of these forests developed after logging and ﬁre destroyed the original, more diverse forest communities of the region. Oak wilt is characterized by expanding clusters of dead trees, or epicentres caused by the transmission of the disease through naturally occurring root grafts between trees. These grafts are most common within the same oak species, but they can also

occur between different oak species (Gibbs and French, 1980). In northern pin oak stands, the probability of occurrence of root grafts between any two trees is related to their size, the distance between them and soil characteristics (Bruhn et al., 1991; Bruhn and Heyd, 1992). The spread of oak wilt through root grafts is greater in shallower than in deeper soils, and greater in sandy than in ﬁner-textured soils (Garin, 1942; Gillespie and True, 1959; Bruhn et al., 1991). In Minnesota, more than 95% of diseased oaks are infected via root grafts (Juzwik, 1983). The maximum observed annual rate of spread via root grafts in northern pin oak stands on sandy soils in Michigan was about 40 ft (Bruhn and Heyd, 1992). However, root transmission may be less important in the southern part of the oak wilt range than further north (Gibbs and French, 1980). In infected trees, fungal structures called mats can form on the boles and large branches of living oaks and also on cut logs with attached bark. In the northwestern portion of the oak wilt range, these mats are common and typically form in the spring, usually in May and early June, on trees that died from the disease the previous summer. Insects are attracted to the mats, possibly by their fruity odour, and carry the fungal spores to fresh wounds on healthy oaks a mile or more away (Gibbs and French, 1980; Juzwik, 1983). New epicentres are thus created by insects that carry the fungal spores from infected to uninfected trees. Infection is initiated by spores carried to fresh wounds on tree boles and branches. Humans often unwittingly spread the fungus by transporting diseased wood with attached bark to unaffected areas. Spring is the most active time of the year for overland spread of oak wilt because it is then that most mats are produced, insect vectors are most abundant, and trees are most susceptible and easily wounded (Bruhn and Heyd, 1992). Methods for controlling oak wilt fall into two categories: (i) controlling the expansion of established epicentres via root grafts; and (ii) reducing overland spread by insects. Established methods for

Even-aged Silvicultural Methods

controlling epicentre expansion have been developed for northern pin oak in Michigan (Bruhn and Heyd, 1992). Epicentres and individual diseased trees can be located from large-scale (e.g. 1:1250) infrared transparency photographs taken in late July or early August. On ﬁlm, healthy trees appear bright red; sick trees appear pink; wilted trees appear tan; and leaﬂess trees are black (Bruhn and Heyd, 1992). Once epicentres are located, their expansion can be controlled by constructing ‘root graft barriers’. The objective of a barrier is to contain the underground spread of the pathogen. Such barriers are typically used in residential areas and parks where their relatively high cost of construction can be justiﬁed. They can be created by either of two methods: (i) soil fumigants, and (ii) deep root-severing trenches dug with a vibratory plough of the type used to lay telephone cable. In urban areas with numerous underground utilities, a soil fumigant such as Vapam can be used to chemically kill trees with interconnected roots (Bruhn and Heyd, 1992). However, in addition to the environmental risks imposed by the use of a toxic chemical, another disadvantage of the fumigant treatment is that it can cause cankers on the lower boles of healthy trees inside the barrier. Although the cankers may not kill trees, they impair their health. Where feasible, the vibratory plough is therefore likely to be the treatment of choice. A plough blade capable of cutting a 5-ft deep trench is required for effective barrier construction. A strategy for barrier placement was developed from research and experience in northern pin oak stands in Michigan (Bruhn et al., 1991). The method is based on a model that estimates the probability of oak wilt transmission between a diseased and an apparently healthy tree. Application of the model requires knowing the distance between such pairs of trees and their diameters (dbh). The intertree distances for infection are determined from the sum of the paired-tree diameters. These distances have been calculated for 95% and 99% probabilities that an apparently healthy tree will not contract oak wilt from an infected

317

tree within a year of the infected tree’s death (Fig. 7.26). Distances for multiplestemmed sprout clumps are based on the sum of the diameters of all living stems within the clump. These distances vary with soil type, and are greater in sands than in ﬁner-textured soils (Fig. 7.26). Using the intertree distances, effective root graft barriers can be located and constructed around oak wilt epicentres. Recommendations are to create two concentric barriers, inner and outer, based on the intertree distances for 95% and 99% probabilities of disease containment, respectively (Fig. 7.26). The inner barrier is located just outside the distance to the apparently healthy tree based on the 95% probability. To assure separating root systems of interconnected trees, recommendations are to position the inner barrier close to the border trees within the inner barrier, rather than midway between those trees and the closest trees outside the inner barrier. The outer barrier corresponds to distances associated with the 99% probability. The barrier boundaries then are marked with ﬂagging to guide the plough operator or fumigant applicator. The barriers are constructed within a month or two of establishing inner and outer boundaries. Because oak roots usually do not grow beneath long-established and well-travelled roads, barriers along such roadways may be unnecessary (Bruhn and Heyd, 1992). Barriers should be created before trees in the epicentre are felled. After root graft barriers are constructed, all oaks inside the inner (95%) barrier are felled as soon as possible; trees between the inner and outer barriers are not cut (Fig. 7.27). An appropriate herbicide should be applied to the stumps of all cut oaks to prevent sprouting and perpetuating living, diseased root systems within the treated area. Infected trees are marked for sanitary treatment to prevent overland spread of the disease. These treatments need to be completed before fungal mats form the following spring. Sanitary treatments include completely covering the ground on treated sites with plastic to prevent insects from reaching fungal mats. To be effective, the covering should be sealed around its edges.

318

Chapter 7

G99%

Distance between trees (ft)

120

100

P99% G95%

80 P95% 60

40

20

0 0

10

20

30

40

50

Sum of dbhs (inches) Fig. 7.26. Relation of the distance between pairs of oaks and the sum of their dbhs for 95% and 99% probabilities that the apparently uninfected member of the pair will not contract oak wilt from the infected member within a year of the latter tree’s death. Relations are based on a model developed in the Upper Peninsula of Michigan (Northern Hardwood Region) in stands that were predominantly northern pin oak. The study sites occupied Grayling and Pemene soil series, which differ in texture and soil (solum) thickness. The Grayling soil is formed in sands occurring on glacial outwash and lake plains and has a solum thickness of 15–30 inches. The Pemene soil is formed in loamy sand glacial till on ground or end moraines and has a solum thickness of 24–48 inches. The two soils and associated probabilities are labelled G95% and G99% for the Grayling soil and P95% and P99% for the Pemene soil. (Adapted from Bruhn and Heyd, 1992, and Bruhn et al., 1991, by permission of Society of American Foresters, Bethesda, Maryland. Not for further reproduction.)

In this enclosed environment, competing fungi will kill the oak wilt fungus (Bruhn and Heyd, 1992). The covering then can be removed at the end of the following summer. Alternatively, the wood from infected trees can be removed to non-vulnerable sites several miles from the infection site. This option should be supervised by oak wilt experts familiar with the local distribution of oaks. Infected wood also can be destroyed before mats form. This requires debarking or destruction of infected wood. Wood can be sold to a saw mill or other utilization plant provided that the buyer understands that the wood is diseased and should be utilized before the next spring. However, the oak wilt fungus cannot live on lumber. The mill preferably should be one that is several miles from the nearest vulnerable oak stand. Oaks within the inner

barrier not showing disease symptoms also should be felled as soon as possible after barrier establishment. If they are felled before the next spring, fungal mats will not form on them and they can be marked and exempted from sanitary treatment. Additional details, explanations and precautions on implementing the root graft barrier method of oak wilt control are presented in Bruhn and Heyd (1992). Root graft barriers may be impractical or unnecessary in many silvicultural applications. Precautionary measures nevertheless may be advisable in known wilt-prone areas including those lying outside the extremely susceptible northern pin oak forests of the Lakes States. Such measures include avoiding pruning, stand and sprout-clump thinning, logging and other possible injury-producing activities to oaks from

Even-aged Silvicultural Methods

319

99% barrier 95% barrier No trees are cut

Cut all oaks and apply sanitation treatments

Fig. 7.27. Relation of inner (95%) and outer (99%) root graft barriers to cutting and sanitation treatments for controlling oak wilt around epicentres of infection. Only trees within the 95% barrier (see Fig. 7.26) are cut. Appropriate sanitation treatments are then applied (see text). (Adapted from Bruhn and Heyd, 1992.)

May to July. Injuries create entryways for spore-carrying insects into the wood of boles and branches, which then can become sites for the establishment of new epicentres. In areas of known oak wilt activity, clump thinning during this period should be avoided because of possible disease transmission through interconnections of the stumps of cut stems with crop stems. Thinning sprout clumps in northern pin oak coppice stands in Wisconsin in late May and early June resulted in a 23% infection rate of residual stems; 19% became infected when thinning was done in mid-June (Kuntz and Drake, 1957). There also is evidence that the overland spread of oak wilt can be reduced by injecting herbicides into trees already infected with oak wilt. One study showed that the herbicide cacodylic acid (hyroxydimethylarsine oxide) when pressure injected into the xylem rendered infected oaks unsuitable sites for oak bark beetle breeding and the formation of fungus mats (Rexrode, 1977). The treatment reduced the number of infected trees with oak bark beetles by 75%, and the number of trees producing fungal mats by 61%. However, the method is not effective in containing disease transmission via root grafting. Although oak wilt has caused heavy

damage to oak stands in some areas, it has not become a serious pathogen in many other areas where it occurs possibly because the insect vectors are not efﬁcient transmitters of the disease and the occurrence of mycelial mats is infrequent (Gibb and French, 1980; USDA Forest Service, 1985).

Economic, Environmental and Social Considerations The clearcutting method Despite its simplicity and economic and technical efﬁcacy, clearcutting has generally not received public acceptance (Bliss, 2000). The method also has received criticism for allegedly scientiﬁc reasons, especially with respect to issues related to forest fragmentation, endangered habitats, soil erosion and long-term decline in forest productivity (see ‘Effects of timber harvesting on site productivity’ in Chapter 4). In addition to being perceived as unattractive in appearance, clearcutting is widely believed to accelerate soil erosion and runoff, and to increase stream sedimentation. Probably no other silvicultural practice has received more public

320

Chapter 7

condemnation than clearcutting. This social reaction to clearcutting, in turn, has led to the widespread abandonment of clearcutting in favour of often misapplied uneven-aged silvicultural methods. Ironically, clearcutting’s rise to prominence during the late 1960s and 1970s began with dissatisfaction with unevenaged silviculture. Disappointment with single-tree selection, combined with early regeneration successes with clearcutting, ushered in the clearcutting era in the Central Hardwood Region in the 1960s. The practice was reinforced by the seminal publication Evenaged Silviculture for Upland Central Hardwoods by Roach and Gingrich (1968). This publication represented a synthesis of years of even-aged silvicultural research and experience. The shift from unevenaged to even-aged silviculture thus was not borne of inexperience with alternative silvicultural systems, nor of the absence of a scientiﬁc basis for the technical efﬁcacy of even-aged silviculture. Clearcutting was enthusiastically accepted by silviculturists for several reasons. In the Central Hardwood Region, it met the ecological requirements of the commercially valuable, shade intolerant species, including the oaks, white ash, black cherry and yellowpoplar. Clearcutting was also economically efﬁcient. Logging, road building and administrative costs were minimized because large timber volumes could be harvested from relatively small areas. Other factors contributing to the acceptance of clearcutting among forest managers included its endorsement by wildlife biologists, the development of even-aged stocking guides and prescriptions for its application (Gingrich, 1967; Roach and Gingrich, 1968; Roach, 1977), the simplicity of creating balanced distributions of stand age classes, and other perceived advantages (McGee, 1987). Clearcutting thus became the most widely recommended and applied regeneration method in the region, albeit with mixed success in oak forests. In oak forests, it was most successful in the drier ecosystems such as the Ozark Highlands of

Missouri. There, the method’s success coincided with the natural occurrence of large pre-established oak seedling-sprouts (Sander, 1971; Sander and Clark, 1971; Sander et al., 1984; Johnson, 1993a). The method was less successful in regenerating oaks in the more productive ecosystems of the Ohio Valley, Appalachians and elsewhere (Gammon et al., 1960; Johnson, 1976; Loftis, 1983b; Beck and Hooper, 1986). Even there, other commercially valuable species usually replaced the oaks. Clearcutting none the less often accelerated the succession of oak-dominated stands to stands with less oak or no oak (Abrams and Downs, 1990; Abrams and Nowacki, 1992). The rise of environmental activism in the 1970s focused public attention on real or perceived negative consequences of clearcutting, including diminished aesthetic value, biodiversity, old growth, certain wildlife values, and soil and water values. To counter these concerns, clearcutting was modiﬁed in various ways: size of clearcuts was reduced; snags, cull trees and uncut islands of trees were retained to enhance wildlife and aesthetic values; the removal of non-commercial residual trees was deferred; cuts were shaped to ﬁt more aesthetically into the landscape; and uncut strips were left where clearcuts bordered roads, lakes, streams and other sensitive areas (USDA Forest Service, 1973, 1980; Evans and Conner, 1979; Smith et al., 1989). Despite these modiﬁcations, clearcutting continued to fall into public disfavour. The public demanded systems that focused less on producing commodities and more on preserving aesthetics, biodiversity and other intangible values (Perry and Maghembe, 1989; Salwasser, 1990; Gale and Cordray, 1991; Hansen et al., 1991; Kessler, 1991). Attitudes towards clearcutting nevertheless may be more complex than commonly realized. For example, in a survey of nonindustrial private forest (NIPF) owners in the mid-South, nearly half of the respondents considered clearcutting to be an acceptable practice on NIPF lands, whereas only 14% considered the practice accept-

Even-aged Silvicultural Methods

able on government-owned lands (Bliss et al., 1997). With respect to its application on private lands, a signiﬁcantly larger proportion of college-educated respondents (52%) considered it an acceptable practice than did less educated respondents (42%). Moreover, the responses of NIPF owners were similar to those of the general public (Bliss et al., 1994). Despite the use of clearcutting as a possible solution to many regeneration problems, the widespread adoption of clearcutting itself produced unanticipated social and political consequences. One outcome was the ‘Monongahela Decision’ of 1970, which imposed uneven-aged silviculture as the primary management tool on the Monongahela National Forest in West Virginia. The decision reinforced organized intervention by the critics of clearcutting, which resulted in a major reduction in its application on all the National Forests. But it also had a wider effect. As Hicks (1997) noted: ‘… it imposed the will of the public over that of foresters, who had always been regarded as the experts in forest resource management’. The solution to an innocent silvicultural problem thus spiralled into the more profound problem of determining who plays a dominant role in planning the use of public forests. The clearcutting controversy has demonstrated that solutions to silvicultural problems are likely to transcend the scientiﬁc and technical approaches familiar to foresters. Foresters (especially those managing public lands) no longer have exclusive control of their domain. They now must engage in public dialogue on diverse and often competing forest uses. Effective dialogue requires recognizing and considering widely differing perspectives of forest values, some of which may be anathema to traditional forestry theory and practice. The emergence of this valuecentred debate has precipitated a profound change in silvicultural practice: the replacement of the economically focused sustainable timber yield paradigm by the more ecologically centred sustainable forest paradigm. However, it may not be the abandonment of clearcutting that charac-

321

terizes the new paradigm as much as it represents a search for a wider range of methods for meeting unprecedentedly diverse social, economic and environmental demands. The contemporary social complexity of silvicultural issues extends in part from an increasingly comprehensive human view of forests as a part of the ‘commons’, which includes not only trees but the air, water, soil, settings for human spiritual renewal and other values that society demands from forests (Walter and Johnson, 2001). Accordingly, it might seem appropriate, especially in a democratic society, to form forestry policy around popular perceptions of what should comprise the forest commons. Although this strategy might ensure the implementation of socially acceptable practices, it may not produce socially desirable outcomes. A forestry based solely on a publically agreed upon visualization may be no more sustainable than the one the public earlier disagreed with. In developing and recommending silvicultural practices, forest managers therefore could beneﬁt from a better understanding of how social views of the forested landscape evolve. A ﬁrst step towards this goal would be to recognize that the personal idealization of nature, the human landscape experience, and the dislike of disruption and change are all parts of human culture. Only when silviculture is given its full deﬁnition of comprising more than scientiﬁc and technical solutions can a socially desirable and socially acceptable framework for its application be formulated. Perhaps the greatest challenge to silviculturists and forest managers today is to deﬁne the line that divides these two goals. If nothing else, the clearcutting controversy has brought this social, scientiﬁc and technical dilemma into focus.

The Shelterwood and seed tree methods The uniform shelterwood and seed tree methods suffer from much the same social drawbacks as clearcutting, i.e. regenerated

322

Chapter 7

stands look much like clearcuts after the ﬁnal shelterwood cut or, in the case of clearcuts with reserves or seed trees, few trees per acre remain. Perhaps as Murphy and others (1993) suggested, the shelterwood method suffers from a ‘guilt of association’ with other even-aged methods. The method is none the less potentially ﬂexible because shelterwoods can be retained, at least hypothetically, until a two-aged stand develops (Smith, 1986; Beck, 1991). In theory, a shelterwood can

be removed in steps that are so gradual that eventually the regenerated stand developing beneath it grows up and becomes indistinguishable from the shelterwood before the latter is completely removed. The resulting ‘irregular shelterwood’ thus may have more aesthetic appeal than the regular shelterwood or clearcutting methods. The method is also potentially applicable to stands dedicated to acorn production and other biodiversity objectives as discussed above.

References Abrams, M.D. (1992) Fire and the development of oak forests. BioScience 42, 346–353. Abrams, M.D. and Downs, J.A. (1990) Successional replacement of old-growth white oak by mixed mesophytic hardwoods in southwestern Pennsylvania. Canadian Journal of Forest Research 20, 1864–1870. Abrams, M.D. and Nowacki, G.J. (1992) Historical variation in ﬁre, oak recruitment, and post-logging accelerated succession in central Pennsylvania. Bulletin of the Torrey Botanical Club 119, 19–28. Bailey, J.K., Zaczek, J.J. and Steiner, K.C. (1996) A comparison of four tree shelter systems using planted seedlings and direct seeded acorns of northern red oak at three sites in Pennsylvania. USDA Forest Service General Technical Report NE NE-221, 64 pp. Bardon, R.E. and Countryman, D.W. (1993) Survival and growth for the ﬁrst-growing season of northern red oak (Quercus rubra L.) seedlings underplanted in mixed upland hardwood stands in south central Iowa. USDA Forest Service General Technical Report NC NC-161, pp. 195–209. Beck, D.E. (1991) The shelterwood method, a research perspective. USDA Forest Service Proceedings 1990 Genetics/Silviculture Workshop. Washington, DC, pp. 252–258. Beck, D.E. and Della-Bianca, L. (1981) Yellow-poplar: characteristics and management. USDA Forest Service Agriculture Handbook 583. Beck, D.E. and Hooper, R.M. (1986) Development of a southern Appalachian hardwood stand after clearcutting. Southern Journal of Applied Forestry 10, 168–172. Behre, C.E. (1939) The opportunity for forestry practice in the control of gypsy moth in Massachusetts. Journal of Forestry 37, 546–551. Bess, H.A., Spurr, S.H. and Littleﬁeld, E.W. (1947) Forest site conditions and the gypsy moth. Harvard Forest Bulletin 22. Bliss, J.C. (2000) Public perceptions of clearcutting. Journal of Forestry 98(12), 4–9. Bliss, J.C, Nepal, S.K., Brooks, R.T., Jr and Larsen, M.D. (1994) Forestry community or granfalloon. Journal of Forestry 92(9), 6–10. Bliss, J.C., Nepal, S.K., Brooks, R.T., Jr and Larsen, M.D. (1997) In the mainstream: environmental attitudes of Mid-South forest owners. Southern Journal of Applied Forestry 21, 37–43. Bonner, F.T. (1973) Storing red oak acorns. Tree Planters’ Notes 24(3), 12–13. Bonner, F.T. and Vozzo, J.A. (1987) Seed biology and technology of Quercus. USDA Forest Service General Technical Report SO SO-66. Bowling, D.R. and Kellison, R.C. (1983) Bottomland hardwood stand development following clearcutting. Southern Journal of Applied Forestry 7, 110–116. Braun, E.L. (1972) Deciduous Forests of Eastern North America. Hafner, New York. Broadfoot, W.M. and Williston, J.H. (1973) Flooding effects on southern forests. Journal of Forestry 71(9), 584–587. Brose, P.H. and Van Lear, D.H. (1998a) Effects of seasonal prescribed ﬁres on hardwood advance regeneration in shelterwood stands. USDA Forest Service General Technical Report SRS SRS20, pp. 310–314.

Even-aged Silvicultural Methods

323

Brose, P.H. and Van Lear D.H. (1998b) Responses of hardwood advance regeneration to seasonal prescribed ﬁres in oak-dominated shelterwood stands. Canadian Journal of Forest Research 28, 331–339. Bruhn, J.N. and Heyd, R.L. (1992) Biology and control of oak wilt in Michigan red oak stands. Northern Journal of Applied Forestry 9, 47–51. Bruhn, J.N., Pickens, J.B. and Stanﬁeld, D.B. (1991) Probit analysis of oak wilt transmission through root grafts in red oak stands. Forest Science 37, 28–44. Buchschacher, G.L., Tomlinson, P.T., Johnson, P.S. and Isebrands, J.G. (1991) Effects of seed source and cultural practices on emergence and seedling quality of northern red oak nursery stock. USDA Forest Service General Technical Report SE, SE-70, Vol. 1, pp. 126–130. Bullard, S., Hodges, J.D., Johnson, R.L. and Straka, T.J. (1992) Economics of direct seeding and planting for establishing oak stands on old-ﬁeld sites in the South. Southern Journal of Applied Forestry 16, 34–40. Bundy, P.P., Alm, A.A. and Baughman, M.J. (1991) Red oak regeneration following ‘scariﬁcation’ and harvesting: a case study. Northern Journal of Applied Forestry 8, 173–174. Burrill, E.A., Worrall, P.M. and Stehman, S.V. (1999) Effects of defoliation and cutting in eastern oak forests on Armillaria spp. and a competitor, Megacollybia platyphylla. Canadian Journal of Forest Research 29, 347–355. Butin, H. and Shigo, A.L. (1981) Radial shakes and ‘frost cracks’ in living oak trees. USDA Forest Service Research Paper NE NE-478. Campbell, R.W. (1981) Population dynamics. In: Doane, C.C. and McManus, M.L. (eds) The Gypsy Moth: Research Toward Integrated Pest Management. USDA Forest Service Technical Bulletin 1584, Washington, DC, pp. 65–214. Campbell, R.W. and Sloan, R.J. (1977) Natural regulation of innocuous gypsy moth populations. Environmental Entomology 6, 315–322. Campbell, R.W. and Valentine, H.T. (1972) Tree condition and mortality following defoliation by the gypsy moth. USDA Forest Service Research Paper NE NE-236. Campbell, R.W., Hubbard, D.L. and Sloan, R.J. (1975a) Patterns of gypsy moth occurrence within a sparse and numerically stable population. Environmental Entomology 4, 535–542. Campbell R.W., Hubbard, D.L. and Sloan, R.J. (1975b) Location of gypsy moth pupae and subsequent pupal survival in sparse, stable populations. Environmental Entomology 4, 597–600. Campbell, R.W., Miller, M.G., Duda, E.J., Biazak, C.E. and Sloan, R.J. (1976) Man’s activities and subsequent gypsy moth egg-mass density along with forest edge. Environmental Entomology 5, 273–276. Carpenter, S.B. (1976) Stand structure of a forest in the Cumberland Plateau of eastern Kentucky ﬁfty years after logging and burning. Castanea 41, 325–336. Carvell, K.L. (1971) Silvicultural aspects of intermediate cuttings. USDA Forest Service NE Forest Experiment Station Oak Symposium Proceedings, pp. 60–64. Clark, F.B. and Boyce, S.G. (1964) Yellow-poplar seed remains viable in the forest litter. Journal of Forestry 62, 564–567. Clatterbuck, W.K. (1998) Use of prescribed ﬁre to promote oak regeneration. USDA Forest Service General Technical Report SRS SRS-20, pp. 315–318. Clatterbuck, W.K. and Hodges, J.D. (1988) Development of cherrybark oak and sweet gum in mixed, even-aged bottomland stands in central Mississippi, USA. Canadian Journal of Forest Research 18, 12–18. Clement, G.E. and Munro, W. (1917) Control of the gypsy moth by forest management. USDA Bureau of Entomology Bulletin 484. Cline, A.C. and Lockard, C.R. (1925) Mixed white pine and hardwood. Harvard Forest Bulletin 8. Congdon, L.D. (1993) Natural regeneration of oak in northern lower Michigan. MS thesis, University of Missouri, Columbia. Countryman, D.W. and Miller, H.R. (1989) Investment analysis of upland oak stands with sugar maple understories: management for oak vs. conversion to sugar maple in Iowa and Missouri. Northern Journal of Applied Forestry 6, 165–169. Crow, G.R. and Hicks, R.R., Jr (1990) Predicting mortality in mixed oak stands following spring insect defoliation. Forest Science 36, 831–841. Crow, T.R. (1992) Population dynamics and growth patterns for a cohort of northern red oak (Quercus rubra) seedlings. Oecologia 91, 192–200.

324

Chapter 7

Crow, T.R. and Isebrands, J.G. (1986) Oak regeneration–an update. Proceedings Integrated Pest Management Symposium for Northern Forests. Cooperative Extension Service, University of Wisconsin, Madison, pp. 73–86. Dale, M.E. (1972) Growth and yield predictions for upland oak stands 10 years after initial thinning. USDA Forest Service Research Paper NE NE-241. Dale, M.E. and Sonderman, D.L. (1984) Effect of thinning on growth and potential quality of young white oak crop trees. USDA Forest Service Research Paper NE NE-539. Davidson, C.B., Gottschalk, K.W. and Johnson, J.E. (1999) Tree mortality following defoliation by the European gypsy moth (Lymantria dispar L.) in the United States: a review. Forest Science 45, 74–84. DeBell, D.S., Langdon, O.G. and Stubbs, J. (1968) Reproducing mixed hardwoods by a seed-tree cutting in the Carolina Coastal Plain. Southern Lumberman 217, 121–123. Dey, D.C. (1991) A comprehensive Ozark regenerator. PhD dissertation, University of Missouri, Columbia. Dey, D.C. and Parker, W.C. (1996) Regeneration of red oak (Quercus rubra L.) using shelterwood systems: ecophysiology, silviculture and management recommendations. Ontario Forest Research Institute Forest Research Information Paper 126. Dey, D.C. and Parker, W.C. (1997a) Overstory density affects ﬁeld performance of underplanted red oak (Quercus rubra L.) in Ontario. Northern Journal of Applied Forestry 14, 120–125. Dey, D.C. and Parker, W.C. (1997b) Morphological indicators of stock quality and ﬁeld performance of red oak (Quercus rubra L.) seedlings underplanted in a central Ontario shelterwood. New Forests 14, 145–156. Dey, D.C., Johnson, P.S. and Garrett, H.E. (1996a) Modeling the regeneration of oak stands in the Missouri Ozark Highlands. Canadian Journal of Forest Research 26, 573–583. Dey, D.C., Ter-Mikaelian, M., Johnson, P.S. and Shiﬂey, S.R. (1996b) Users guide to ACORn: a comprehensive Ozark regeneration simulator. USDA Forest Service General Technical Report NC NC-180. Dorsey, C.K., Jewell, F.F., Leach, J.G. and True, R.P. (1953) Experimental transmission of oak wilt by four species of Nitidulidae. Plant Disease Reporter 37, 419–420. Dunbar, D.M. and Stephens, G.R. (1975) Association of twolined chestnut borer and shoestring fungus with mortality of defoliated oak in Connecticut. Forest Science 21, 169–174. Dwyer, J.P. and Kurtz, W.B. (1994) Pre-outbreak management recommendation for 60-year-old declining oak stands in the Ozarks. Northern Journal of Applied Forestry 11, 98–101. Dwyer, J.P., Cutter, B.E. and Wetteroff, J.J. (1995) A dendrochronological study of black and scarlet oak decline in the Missouri Ozarks. Forest Ecology and Management 75, 60–75. Evans, K.E. and Conner, R.N. (1979) Snag management. USDA Forest Service General Technical Report NC NC-51, pp. 214–225. Farley, M.E., Perry, P.S. and Woyar, P.R. (1996) Valley coal tree shelter ﬁeld trial. USDA Forest Service General Technical Report NE NE-221, pp. 60–63. Fiske, W.F. (1913) The gypsy moth as a forest insect with suggestions as to its control. USDA Bureau of Entomology Circular 164. Franklin, J. (1989) Toward a new forestry. American Forests 95, 37–44. Franklin, J.F., Perry, D.A., Schowalter, T.D., Harmon, M.E., McKee, A. and Spier, T.A. (1989) Importance of ecological diversity in maintaining long-term site productivity. In: Perry, D.A., Meurisse, R., Thomas, B., Miller, R., Boyle, J. and Means, J. (eds) Maintaining the Long-term Productivity of Paciﬁc Northwest Forest Ecosystems. Timber Press, Portland, Oregon, pp. 82–97. Gale, R.P. and Cordray, S.M. (1991) What should forests sustain? Eight answers. Journal of Forestry 89(5), 31–36. Gammon, A.D., Rudolph, V.J. and Arend, J.L. (1960) Regeneration following clearcutting of oak during a seed year. Journal of Forestry 58, 711–715. Gansner, D.A. and Herrick, O.W. (1984) Guides for estimating forest stand losses to gypsy moth. Northern Journal of Applied Forestry 1, 21–23. Gansner, D.A. and Herrick, O.W. (1985) Host preferences of gypsy moth on a new frontier of infestation. USDA Forest Service Research Note NE NE-330. Gansner, D.A., Herrick, O.W. and Ticehurst, M. (1985) A method for predicting gypsy moth defoliation from egg mass counts. Northern Journal of Applied Forestry 2, 78–79.

Even-aged Silvicultural Methods

325

Gardiner, E.S. and Helmig, L.M. (1997) Development of water oak stump sprouts under a partial overstory. New Forests 14, 55–62. Garges, L.D., Laboskey, P., Blankenhorn, P.R. and Rishel, L.E. (1984) Lumber recovery from gypsy moth-killed red and white oak trees. Forest Products Journal 34, 45–50. Garin, G.I. (1942) Distribution of roots of certain tree species in two Connecticut soils. Connecticut Agricultural Experiment Station Bulletin 454. Gevorkiantz, S.R. (1957) Site index curves for red oak in the Lake States. USDA Forest Service Lake States Forest Experiment Station Technical Note 485. Gevorkiantz, S.R. and Scholz, H.F. (1948) Timber yields and possible returns from the mixed-oak farmwoods of southwestern Wisconsin. USDA Forest Service Lake States Forest Experiment Station Publication 521. Gibbs, J.N. and French, D.W. (1980) The transmission of oak wilt. USDA Forest Service Research Paper NC NC-185. Gillespie, R., Rathfon, R. and Myers, R.K. (1996) Rehabilitating a young northern red oak planting with tree shelters. Northern Journal of Applied Forestry 13, 24–29. Gillespie, W.H. and True, R.P. (1959) Three factors which inﬂuence the local spread of oak wilt in ﬁve northeastern counties of West Virginia. Plant Disease Reporter 43, 588–593. Gingrich, S.F. (1967) Measuring and evaluating stocking and stand density in upland hardwood forests in the Central States. Forest Science 13, 38–53. Gingrich, S.F. (1971) Management of young and intermediate stands of upland hardwoods. USDA Forest Service Research Paper NE NE-195. Golden, M.S. (1993) Establishment of regeneration in small patch clearcut openings in an Alabama River ﬂoodplain forest. USDA Forest Service General Technical Report SO SO-93, pp. 161–167. Golden, M.S. and Loewenstein, E.F. (1991) Regeneration of tree species 7 years after clearcutting in a river bottom in central Alabama. USDA Forest Service General Technical Report SE SE-70, Vol. 1, pp. 76–83. Gottschalk, K.W. (1989) Effects of previous stand management on mortality following gypsy moth defoliation: preliminary results. USDA Forest Service General Technical Report SO SO-74, pp. 573–578. Gottschalk, K.W. (1991) Gypsy moth: impacts and silvicultural options. Proceedings The Oak Resource in the Upper Midwest Conference. University of Minnesota, St Paul, pp. 159–169. Gottschalk, K.W. (1993) Silvicultural guidelines for forest stands threatened by the gypsy moth. USDA Forest Service General Technical Report NE NE-171. Gottschalk, K.W. (1994) Shade, leaf growth and crown development of Quercus rubra, Quercus velutina, Prunus serotina and Acer rubrum seedlings. Tree Physiology 14, 735–749. Gottschalk, K.W. and MacFarlane, W.R. (1993) Photographic guide to crown condition of oaks: use for gypsy moth silvicultural treatments. USDA Forest Service General Technical Report NE NE-168. Gottschalk, K.W. and Marquis, D.A. (1982) Survival and growth of planted red oak and white ash as affected by residual overstory density, stock size, and deer browsing. Proceedings 4th Central Hardwood Forest Conference. University of Kentucky, Lexington, pp. 125–140. Gottschalk, K.W., Gansner, D.A. and Twery, M.J. (1989) Impacts of gypsy moth on oak timber resources or will there be oak in 2001. Proceedings 17th Annual Hardwood Symposium of the Hardwood Research Council. Graney, D.L. and Bower, D.R. (1971) Site index curves for red and white oaks in the Boston Mountains of Arkansas – USDA Forest Service Research Note SO SO-121 Gresham, C.A. (1985a) Clearcutting not enough for early establishment of desirable species in Santee River swamp. Southern Journal of Applied Forestry 9, 52–54. Gresham, C.A. (1985b) Pine and hardwood regeneration alternatives for harvested bottomland hardwood stands. USDA Forest Service General Technical Report SO SO-54, pp. 87–92. Grisez, T.J. and Peace, M.R. (1973) Requirements for advance reproduction in Allegheny hardwoods – an interim guide. USDA Forest Service Research Note NE NE-180. Grushecky, S.T., Liebhold, A.M., Greer, R. and Smith, R.L. (1998) Does forest thinning affect predation on gypsy moth (Lepidoptera: Lymantriidae) larvae and pupae? Environmental Entomology 27, 268–276. Hannah, P.R. (1987) Regeneration methods for oaks. Northern Journal of Applied Forestry 4, 97–101. Hansen, A.J., Spies, T.A., Swanson, F.J. and Ohmann, J.L. (1991) Conserving biodiversity in managed forests. BioScience 41, 382–392.

326

Chapter 7

Hanson, P.J., Isebrands, J.G. and Dickson, R.E. (1987a) Carbon budgets of Quercus rubra L. seedlings at selected stages of growth: inﬂuence of light. Proceedings 6th Central Hardwood Forest Conference. University of Tennessee, Knoxville, pp. 269–276. Harper, J.L. (1977) Population Biology of Plants. Academic Press, London. Helms, J.A. (ed.) (1998) The Dictionary of Forestry. Society of American Foresters, Bethesda, Maryland. Herrick, O.W. (1982) Hazard rating forest trees threatened with gypsy moth invasion. In: Coping With the Gypsy Moth. The Pennsylvania State University, University Park, pp. 38–42. Herrick, O.W. and Gansner, D.A. (1986) Rating forest stands for gypsy moth defoliation. USDA Forest Service Research Paper NE NE-583. Herrick, O.W. and Gansner, D.A. (1987) Mortality risks for forest trees threatened with gypsy moth infestation. USDA Forest Service Research Note NE NE-338. Hibbs, D.E. (1981) Development of the forest canopy – a key to silviculture. Northern Logger and Timber Processor (May), 62, 78. Hibbs, D.E. and Bentley, W.R. (1983) A management guide for oak in New England. University of Connecticut Cooperative Extension Bulletin 83–12. Hibbs, D.E. and Bentley, W.R. (1984) A growth model for red oak in New England. Canadian Journal of Forest Research 14, 250–254. Hicks, R.R., Jr (1997) A resource at the crossroads: a history of the central hardwoods. USDA Forest Service General Technical Report NC NC-188, pp. 1–22. Hill, J.P. and Dickmann, D.I. (1988) A comparison of three methods for naturally reproducing oak in southern Michigan. Northern Journal of Applied Forestry 5, 113–117. Hilt, D.E. (1977) Introduction of black walnut and northern red oak seedlings in an upland hardwood forest in southeastern Ohio. USDA Forest Service Research Note NE NE-241. Hilt, D.E. (1985a) Species composition of young Central Hardwood stands that develop after clearcutting. Proceedings 5th Central Hardwood Forest Conference. University of Illinois, ChampaignUrbana, pp. 11–14. Hilt, D.E. (1985b) Where has all my yellow-poplar gone? Northern Journal of Applied Forestry 2, 67–69. Hilt, D.E. and Dale, M.E. (1979) Stem form changes in upland oaks after thinning. USDA Forest Service Research Paper NE NE-433. Horsley, S.B. (1982) Development of reproduction in Allegheny hardwood stands after herbicideclearcuts and herbicide-shelterwood cuts. USDA Forest Service Research Note NE NE-308. Horsley, S.B. (1991) Using Roundup and Oust to control interfering understories in Allegheny hardwood stands. USDA Forest Service General Technical Report NE NE-148, pp. 281–290. Horsley, S.B. and Marquis, D.A. (1983) Interference by weeds and deer with Allegheny hardwood reproduction. Canadian Journal of Forest Research 13, 61–69. Hosner, J.F. and Boyce, S.G. (1962) Tolerance to water saturated soil of various bottomland hardwoods. Forest Science 8, 180–186. Hosner, J.F. and Minckler, L.S. (1960) Hardwood reproduction in the river bottoms of southern Illinois. Forest Science 6, 67–77. Hosner, J.F. and Minckler, L.S. (1963) Bottomland hardwood forests of southern Illinois – regeneration and succession. Ecology 44, 29–41. Houston, D.R. (1987) Recognizing and managing diebacks/declines. Proceedings Integrated Pest Management Symposium for Northern Forests. University of Wisconsin, Madison, pp. 153–166. Isebrands, J.G. and Crow, T.R. (1985) Techniques for rooting juvenile softwood cuttings of northern red oak. Proceedings 5th Central Hardwood Forest Conference. University of Illinois, Champaign-Urbana, pp. 228–233. Jacobs, R.D. and Wray, R.D. (1992) Managing oak in the Driftless Area. Minnesota Extension Service Bulletin NR-BU NR-BU-5900-S. Jenkins, M.A. and Pallardy, S.G. (1995) The inﬂuence of drought on red oak group species growth and mortality in the Missouri Ozarks. Canadian Journal of Forest Research 25, 1119–1127. Jenkins, M.A. and Parker, G.R. (1998) Composition and diversity of woody vegetation in silvicultural openings of southern Indiana forests. Forest Ecology and Management 109, 57–74. Johnson, P.S. (1974) Survival and growth of northern red oak seedlings following a prescribed burn. USDA Forest Service Research Note NC NC-177. Johnson, P.S. (1975) Growth and structural development of red oak sprout clumps. Forest Science 21, 413–418.

Even-aged Silvicultural Methods

327

Johnson, P.S. (1976) Modal development of regeneration in clearcut red oak stands in the Driftless Area. Proceedings 1st Central Hardwood Forest Conference. University of Southern Illinois, Carbondale, pp. 455–475. Johnson, P.S. (1984) Responses of planted northern red oak to three overstory treatments. Canadian Journal of Forest Research 14, 536–542. Johnson, P.S. (1988) Undercutting alters root morphology and improves ﬁeld performance of northern red oak. Proceedings 10th North American Forestry and Biology Workshop, pp. 316–323. Johnson, P.S. (1989) Growing hardwood nursery stock for planting on forest sites with special reference to northern red oak. Proceedings 1989 Northeastern Area Nurserymen’s Conference, pp. 46–62. Johnson, P.S. (1992a) Oak overstory/reproduction relations in two xeric ecosystems in Michigan. Forest Ecology and Management 48, 233–248. Johnson, P.S. (1992b) Underplanting northern red oak in Missouri without herbicides. USDA Forest Service General Technical Report NC NC-152. Johnson, P.S. (1993) Perspectives on the ecology and silviculture of oak-dominated forests in the central and eastern states. USDA Forest Service General Technical Report NC NC-153. Johnson, P.S. (1994) How to manage oak forests for acorn production. USDA Forest Service Technical Brief TB-NC TB-NC-1. Johnson, P.S. and Rogers, R. (1980) Predicting growth of individual stems within northern red oak sprout clumps. Proceedings 3rd Central Hardwood Forest Conference. University of Missouri, Columbia, pp. 420–439. Johnson, P.S. and Rogers, R. (1984) Predicting 25th-year diameters of thinned stump sprouts of northern red oak. Journal of Forestry 82, 616–619. Johnson, P.S. and Rogers, R. (1985) A method for estimating the contribution of planted hardwoods to future stocking. Forest Science 31, 883–891. Johnson, P.S. and Sander, I.L. (1987) Quantifying regeneration potentials of Quercus forests in the Missouri Ozarks. USDA Forest Service General Technical Report NC NC-120, Vol. 1, pp. 377–385. Johnson, P.S., Novinger, S.L. and Mares, W.G. (1984) Root, shoot, and leaf area growth potentials of northern red oak planting stock. Forest Science 30, 1017–1026. Johnson, P.S., Dale, C.D., Davidson, K.R. and Law, J.R. (1986) Planting northern red oak in the Missouri Ozarks: a prescription. Northern Journal of Applied Forestry 3, 66–68. Johnson, P.S., Jacobs, R.D., Martin, A.J. and Godell, E.D. (1989) Regenerating northern red oak: three successful case histories. Northern Journal of Applied Forestry 6, 174–178. Johnson, R.L. (1975) Natural regeneration and development of Nuttall oak and associated species. USDA Forest Service Research Paper S0 SO-104 Johnson, R.L. (1981) Oak seeding – it can work. Southern Journal of Applied Forestry 5, 28–33. Johnson, R.L. and Deen, R.T. (1993) Prediction of oak regeneration in bottomland forests. USDA Forest Service General Technical Report SE SE-84, pp. 146–155. Johnson, R.L. and Krinard, R.M. (1983) Development of seven hardwood species in small forest openings, 22-year results. Southern Journal of Applied Forestry 7, 153–156. Johnson, R.L. and Krinard, R.M. (1987) Direct-seeding of southern oaks – a progress report. Proceedings 15th Hardwood Research Council Annual Hardwood Symposium, pp. 10–16. Johnson, W.C. and Webb, T., III. (1989) The role of blue jays (Cyanocitta cristata L.) in the postglacial dispersal of fagaceous trees in eastern North America. Journal of Biogeography 16, 561–571. Juzwik, J. (1983) Factors affecting overland transmission of Ceratocystis fagacearum in Minnesota. PhD dissertation, University of Minnesota, St Paul, Minnesota. Kaczmarek, D.J. and Pope, P.E. (1993) Covariate analysis of northern red oak seedling growth. USDA Forest Service General Technical Report SO SO-93, pp. 351–356. Kays, J.S. (1996) Deer protection for small forest plantations: comparing costs of tree shelters, electric fencing and repellents. USDA Forest Service General Technical Report NE NE-221, pp. 5–12. Kegg, J.D. (1974) Quantitative classiﬁcation of forest stands as related to their vulnerability to loss following initial outbreak of the gypsy moth, Porthetria dispar (L.), with equations for forecasting basal area losses. MS thesis, Rutgers University, New Brunswick, NJ. Kennedy, H.E., Jr (1993) Artiﬁcial regeneration of bottomland oaks. USDA Forest Service General Technical Report SE SE-84, pp. 241–249. Kennedy, H.E., Jr and Krinard, R.M. (1985) Shumard oaks successfully planted on high pH soils. USDA Forest Service Research Note SO SO-321.

328

Chapter 7

Kessler, W.B. (1991) New perspectives for managing forested lands: keeping pace with society. Proceedings The Oak Resource in the Upper Midwest Conference. University of Minnesota, St Paul, pp. 34–38. Keyser, P.D., Brose, P.H., Van Lear, D. and Burtner, K.M. (1996) Enhancing oak regeneration with ﬁre in shelterwood stands: preliminary trials. Transactions 61st North American Wildlife and Natural Resource Conference, pp. 215–219. Kittredge, D.B., Kelty, M.J. and Ashton, M.S. (1992) The use of tree shelters with northern red oak natural regeneration in southern New England. Northern Journal of Applied Forestry 9, 141–145. Kormanik, P.P. (1989) Importance of ﬁrst-order lateral roots in the early development of forest tree seedlings. Proceedings Interrelationships between Microorganisms and Plants in Soil. Institute of Microbiology, Czechoslovak Academy of Science, pp. 157–169. Kormanik, P.P., Sung, S.S., Kormanik T.L., Zarnoch, S.J. and Schlarbaum, S.E. (1997) Heritability of ﬁrst-order lateral roots in ﬁve Quercus species: effect on 1-0 seedling quality evaluation. In: Steiner, K.C. (ed.) Proceedings Diversity and Adaptation in Oak Species. Pennsylvania State University, University Park, pp. 194–200. Kormanik, P.P., Sung, S.S., Kass, D.J. and Schlarbaum, S.E. (1998) Effect of seedling size and ﬁrstorder lateral roots on early development of northern red oak on mesic sites. USDA Forest Service General Technical SRS SRS-20, pp. 247–252. Korstian, C.F. (1927) Factors controlling germination and early survival in oaks. Yale University School of Forestry Bulletin 19. Krinard, R.M. and Francis, J.K. (1983) Twenty-year results of planted cherrybark oak on old ﬁelds in brown loam bluffs. Tree Planters’ Notes 34(4), 20–22. Krinard, R.M. and Johnson, R.L. (1988) Stand parameters of a 27-year-old water oak plantation on old ﬁeld loessial soils. USDA Forest Service Research Note SO SO-348. Kuntz, J.E. and Drake, C.R. (1957) Tree wounds and long-distance spread of oak wilt. Phytopathology 47, 22. LaFarge, T. and Lewis, R.A. (1987) Phenotypic selection effective in a northern red oak seedling seed orchard. Proceedings 19th Southern Forest Tree Improvement Conference, pp. 200–207. Lamson, N.I. (1976) Appalachian hardwood stump sprouts are potential sawlog crop trees. USDA Forest Service Research Note NE NE-229. Lamson, N.I. (1983) Precommercial thinning increases diameter growth of Appalachian hardwood stump sprouts. Southern Journal of Applied Forestry 7 (May), 93–97. Lamson, N.I. and Smith, H.C. (1978) Response to crop-tree release: sugar maple, red oak, black cherry, and yellow-poplar saplings in a 9-year-old stand. USDA Forest Service Research Paper NE NE-394. Lantagne, D.O. (1991) Tree shelters increase heights of planted northern red oaks. USDA Forest Service General Technical Report NE NE-148, pp. 291–298. Lantagne, D.O. (1996) Effects of tree shelters on planted red oak in Michigan. USDA Forest Service General Technical Report NE NE-221, pp. 24–28. Lantagne, D.O. and Miller, R. (1997) Tree shelters fail to enhance height growth of northern red oak in the Upper Peninsula. USDA Forest Service General Technical Report NC NC-188, 351 pp. Leak, W.B. (1981) Do stocking guides in the Eastern United States relate to stand growth? Journal of Forestry 79, 661–664. Leonard, D.E. (1981) Bioecology of the gypsy moth. In: Doane, C.C. and McManus, M.L. (eds) The Gypsy Moth: Research Toward Integrated Pest Management, USDA Forest Service Technical Bulletin 1584, Washington, DC, pp. 1–29. Lewis, F.G. (1981) Gypsy moth nucleopolyhedrosis virus: Introduction. In: Doane, C.C. and McManus, M.L. (eds) The Gypsy Moth: Research Toward Integrated Pest Management, USDA Forest Service Technical Bulletin 1584, Washington, DC, pp. 503–514. Liebhold, A.M, MacDonald, W.L., Bergdahl, D. and Mastro, V.C. (1995a) Invasion by exotic forest pests: a threat to forest ecosystems. Forest Science Monograph 30. Liebhold, A.M., Gottschalk, K.W., Muzika, R., Montgomery, M E., Young, R., Day, K. and Kelley, B. (1995b) Suitability of North American tree species to the gypsy moth: a summary of ﬁeld and laboratory tests. USDA Forest Service General Technical Report NE NE-211. Liebhold, A.M., Muzika, R.M. and Gottschalk, K.W. (1998) Does thinning affect gypsy moth dynamics? Forest Science 44, 239–245.

Even-aged Silvicultural Methods

329

Liming, F.G. and Johnston, J.P. (1944) Reproduction in oak–hickory forest stands of the Missouri Ozarks. Journal of Forestry 42, 175–180. Loftis, D.L. (1979) Northern red oak performs poorly in North Carolina planting. USDA Forest Service Research Note SE SE-277. Loftis, D.L. (1983a) Regenerating southern Appalachian mixed hardwood stands with the shelterwood method. Southern Journal of Applied Forestry 7, 212–217. Loftis, D.L. (1983b) Regenerating red oak on productive sites in the southern Appalachians: a research approach. USDA Forest Service General Technical Report SE SE-24, pp. 144–150. Loftis, D.L. (1988) Regenerating oaks on high-quality sites, an update. Society of American Foresters Publication 88-03, pp. 199–209. Loftis, D.L. (1990a) Predicting post-harvest performance of advance red oak reproduction in the southern Appalachians. Forest Science 36, 908–916. Loftis, D.L. (1990b) A shelterwood method for regenerating red oak in the southern Appalachians. Forest Science 36, 917–929. Lorimer, C.G. (1989a) The oak regeneration problem: new evidence on causes and possible solutions. University of Wisconsin Forestry Research Analyses 8. Margalef, R. (1963) On certain unifying principles in ecology. American Naturalist 97, 357–374. Margalef, R. (1968) Perspectives in Ecological Theory. University of Chicago Press, Chicago. Marquis, D.A. (1974) The impacts of deer browsing on Allegheny hardwood regeneration. USDA Forest Service Research Paper NE NE-308. Marquis, D.A. (1981) Survival, growth, and quality of residual trees following clearcutting in Allegheny hardwood forests. USDA Forest Service Research Paper NE NE-477. Marquis, D.A. and Bjorkbom, J.C. (1982) Guidelines for evaluating regeneration before and after clearcutting Allegheny hardwoods. USDA Forest Service Research Note NE NE-307. Marquis, D.A., Grisez, T.J., Bjorkbom, J.C. and Roach, B.A. (1975) Interim guide to regeneration of Allegheny hardwoods. USDA Forest Service General Technical Report NE NE-19. Marquis, D.A., Ernst, R.L. and Stout, S.L. (1992) Prescribing silvicultural treatments in hardwood stands of the Alleghenies (revised). USDA Forest Service General Technical Report NE NE-96. Martin, J., McGill, D. and Phibbs, A. (1992) Red oak regeneration in southwest Wisconsin. University of Wisconsin Forestry Research Note 279. Mason, G.N., Gottschalk, K.W. and Hadﬁeld, J.S. (1989) Effects of timber management practices on insects and diseases. In: Burns, R.M. (ed.) The Scientiﬁc Basis for Silvicultural and Management Decisions in the National Forest System. USDA Forest Service General Technical Report WO WO-55, pp 165–169. Mayhead, G.J. and Jones, D. (1991) Carbon dioxide concentrations within tree shelters. Quarterly Journal of Forestry 75, 228–232. McGee, C.E. (1987) Clearcutting in upland hardwoods: panacea or anathema? Proceedings 6th Central Hardwood Forest Conference. University of Tennessee, Knoxville, pp. 21–29. McGee, C.E. and Hooper, R.M. (1970) Regeneration after clearcutting in the southern Appalachians. USDA Forest Service Research Paper SE SE-70. McGee, C.E. and Loftis, D.L. (1986) Planted oaks perform poorly in North Carolina and Tennessee. Northern Journal of Applied Forestry 3, 114–116. McKevlin, M.R. (1992) Guide to regeneration of bottomland hardwoods. USDA Forest Service General Technical Report SE SE-76. McKnight, J.S., Hook, D.D., Langdon, O.G. and Johnson, R.L. (1981) Flood tolerance and related characteristics of trees of the bottomland forests of the southern United States. In: Clark, J.R. and Benforado, J. (eds) Wetlands of Bottomland Hardwood Forests: Proceedings of a Workshop on Bottomland Hardwood Forest Wetlands of the Southeastern United States. Elsevier Science, New York, pp. 29–69. McManus, M.L. and Houston, D.R. (1979) The homeowner and the gypsy moth: guidelines for control. USDA Home and Garden Bulletin Bull. 227. McMinn, J.W. (1986) Transpirational drying of red oaks, sweetgum, and yellow-poplar in the Upper Piedmont of Georgia. Forest Products Journal 36, 25–27. McNab, W.H. and Avers, P.E. (1994) Ecological subregions of the Unites States: section descriptions. USDA Forest Service Administrative Publication WO-WSA WO-WSA-5. McQuilkin, R.A. (1974) Site index prediction table for black, scarlet, and white oaks in southeastern Missouri. USDA Forest Service Research Paper NC NC-108.

330

Chapter 7

Merritt, C. and Pope, P.E. (1991) The effect of environmental factors, including wildﬁre and prescribed burning, on the regeneration of oaks in Indiana. Purdue University Department of Forestry and Natural Resources Agricultural Experiment Station Bulletin 612. Minter, W.F., Myers, R.K. and Fischer, B.C. (1992) Effects of tree shelters on northern red oak seedlings planted in harvested forest openings. Northern Journal of Applied Forestry 9, 58–63. Montgomery, M.E. and Wallner, W.E. (1988) The gypsy moth: a westward migrant. In: Berryman, A.A. (ed.) Dynamics of Forest Insect Populations, Patterns, Causes, Implications. Plenum Press, New York, pp. 353–375. Murphy, P.A., Shelton, M.G. and Graney, D.L. (1993) Group selection – problems and possibilities for the more shade-intolerant species. USDA Forest Service General Technical Report NC NC-161, pp. 229–247. Muzika, R.M. and Liebhold, A.M. (2000) A critique of silvicultural approaches to managing defoliating insects in North America. Agricultural and Forest Entomology 2, 97–105. Nix, L.E. and Lafaye, A. (1993) Successful regeneration of quality oaks in a red river bottomland stand of South Carolina. USDA Forest Service General Technical Report SO SO-93, pp. 81–85. Nowacki, G.J., Abrams, M.D. and Lorimer, C.G. (1990) Composition, structure, and historical development of northern red oak stands along an edaphic gradient in north-central Wisconsin. Forest Science 36, 276–292. Nyland, R.D., Abrahamson, L.P. and Adams, K.B. (1983) Use of prescribed ﬁre for regenerating red and white oak in New York. Proceedings 1982 Society American Foresters National Convention, pp. 163–167. Oak, S.W. and Starkey, D.A. (1991) Measuring physiological maturity from site quality and age: observations from oak decline areas in southern upland forests. Proceedings Southwide Forest Disease Workshop. Duke University, Durham, NC. Oliver, C.D. (1978) The development of northern red oak in mixed stands in Central New England. Yale University School of Forestry and Environmental Science Bulletin 91. Oliver, C.D. (1980) Even-aged development of mixed-species stands. Journal of Forestry 78, 201–203. Oliver, C.D. and Larson, B.C. (1996) Forest Stand Dynamics. John Wiley & Sons, New York. Olson, D.F., Jr (1959) Site index curves for upland oak in the southeast. USDA Forest Service Southeastern Forest Experiment Station Research Note 125. Olson, D.F., Jr (1974) Quercus L. Oak. USDA Forest Service Agriculture Handbook 450, 88 pp. Olson, D.F., Jr and Hooper, R.M. (1972) Northern red oak plantings survive well in southern Appalachians. Tree Planters’ Notes 23(1), 16–18. Onken, A. (1995). Gypsy moth nucleopolyhedrosis virus (NPV) literature review. Gypsy Moth News (39), 7–10. Peet, R.K. (1992) Community structure and ecosystem function. In: Glenn-Lewin, D.C. and Peet, R.K. (eds) Plant Succession. Chapman & Hall, New York, pp. 103–151. Perry, D.A. and Maghembe, J. (1989) Ecosystem concepts and current trends in forest management: time for reappraisal. Forest Ecology and Management 26, 123–140. Podgwaite, J.D., Reardon, R.C., Kolodny-Hirsch, D.M. and Walton, G.S. (1991) Efﬁcacy of ground application of the gypsy moth (Lepidoptera: Lymantriidae) nucleopolyhedrosis virus product, gypchek. Journal of Economic Entomology 84, 440–444. Ponder, F., Jr. (1996) Tree shelter effects on stem and root biomass of planted hardwoods. USDA Forest Service General Technical Report NE NE-221, pp. 19–23. Ponder, F., Jr (1997) Survival and growth of hardwood seedlings following preplanting-root treatments and treeshelters. USDA Forest Service General Technical Report NC NC-188, pp. 332–340. Potter, M.J. (1988) Treeshelters improve survival and increase early growth rates. Journal of Forestry 86(8), 39–41. Potter, M.J. (1991) Treeshelters. Forestry Commission Handbook 7. London. Pubanz, D.M. and Lorimer, C.G. (1992). Oak regeneration experiments in southwestern Wisconsin: two-year results. University of Wisconsin-Madison College of Agriculture and Life Science Publication R3552. Reardon, R.C. and Hajek, A.E. (1993) Entomophaga maimaiga in North America: a review. USDA Forest Service NA-TP TP-15-93. Rexrode, C.O. (1977) Cacodylic acid reduces the spread of oak wilt. Plant Disease Reporter 61, 972–975.

Even-aged Silvicultural Methods

331

Rink, G., and Coggeshall, M.V. (1995) Potential height gain from selection in a ﬁve-year-old white oak progeny test. Southern Journal of Applied Forestry 19, 10–13. Rink, G. and Williams, R.D. (1984) Storage technique affects white oak acorn viability. Tree Planters’ Notes 35(1), 3–5. Roach, B.A. (1977) A stocking guide for Allegheny hardwoods and its use in controlling intermediate cuttings. USDA Forest Service Research Paper NE NE-373. Roach, B.A. and Gingrich, S.F. (1968) Even-aged silviculture for upland central hardwoods. USDA Forest Service Agriculture Handbook 355. Rogers, R. and Johnson, P.S. (1985) Rule thinning: a ﬁeld method for meeting stocking goals in oak stands. Proceedings 5th Central Hardwood Forest Conference. University of Illinois, Champaign-Urbana, pp. 106–110. Roth, E.R. (1956) Decay following thinning of sprout oak clumps. Journal of Forestry 54, 26–30. Roth, E.R. and Hepting, G.H. (1943) Origin and development of oak stump sprouts as affecting their likelihood to decay. Journal of Forestry 41, 27–36. Roth, E.R. and Hepting, G.H. (1969) Prediction of butt rot in newly regenerated sprout oak stands. Journal of Forestry 67, 756–760. Roth, E.R. and Sleeth, B. (1939) Butt rot in unburned sprout oak stands. USDA Technical Bulletin 684. Russell, T.E. (1973) Survival and growth of bar-slit planted northern red oak studied in Tennessee. Tree Planters’ Notes 24(3), 6–9. Salwasser, H. (1990) Gaining perspective: forestry for the future. Proceedings 1990 Society of American Foresters National Convention, pp. 55–63. Sampson, T.L., Barrett, J.P. and Leak, W.B. (1983) A stocking chart for northern red oak in New England. University of New Hampshire Agriculture Experiment Station Research Report 100. Sander, I.L. (1971) Height growth of new oak sprouts depends on size of advance reproduction. Journal of Forestry 69, 809–811. Sander, I.L. (1977) Manager’s handbook for oaks in the North Central States. USDA Forest Service General Technical Report NC NC-37. Sander, I.L. (1979) Regenerating oaks. Proceedings National Silviculture Workshop. USDA Forest Service, pp. 212–221. Sander, I.L. and Clark, F.B. (1971) Reproduction of upland hardwood forests in the Central States. USDA Forest Service Agriculture Handbook 405. Sander, I.L., Johnson, P.S. and Watt, R.F. (1976) A guide for evaluating the adequacy of oak advance reproduction. USDA Forest Service General Technical Report NC NC-23. Sander, I.L., Johnson, P.S. and Rogers, R. (1984) Evaluating oak advance reproduction in the Missouri Ozarks. USDA Forest Service Research Paper NC NC-251. Schlarbaum, S.E., Kormanik, P.P., Tibbs, T. and Barber, L.R. (1997a) Oak seedlings: quality improved available now – genetically improved available soon. Proceedings 25th Annual Hardwood Symposium. National Hardwood Lumber Association, Memphis, TN, pp. 123–129. Schlarbaum, S.E., Barber, L.R., Cecich, R.A., Cox, R.A., Grant, J.F., Kormanik, P.P., LaFarge, T., Lambdin, P.L., Lay, S.A., Post, L.S., Profﬁtt, C.K., Remaley, M.A., Saxton, A.M., Stringer, J.W. and Tibbs, T. (1997b) Research and development activities in a northern red oak (Quercus rubra L.) seedling seed orchard. Proceedings Diversity and Adaptation in Oak Species. Pennsylvania State University, University Park, pp. 185–192. Scholz, H.F. (1952) Age variability of northern red oak in the Upper Mississippi woodlands. Journal of Forestry 50, 518–521. Scholz, H.F. (1955) Effect of scariﬁcation on the initial establishment of northern red oak reproduction. USDA Forest Service Lake States Forest Experiment Station Technical Note 425. Schuler, T.M. and Miller, G.W. (1995) Shelterwood treatments fail to establish oak reproduction on mesic forest sites in West Virginia – 10-year results. USDA Forest Service General Technical Report NE NE-197, pp. 375–387. Schuler, T.M. and Miller, G.W. (1996) Guidelines for using tree shelters to regenerate northern red oak. USDA Forest Service General Technical Report NE NE-221, pp. 37–45. Schultz, R.C. and Thompson, J.R. (1989) Nursery cultural practices that improve hardwood seedling root morphology. Proceedings 1989 Northeastern Area Nurserymen’s Conference, pp. 17–39. Schultz, R.C. and Thompson, J.R. (1990) Nursery practices that improve hardwood seedling root morphology. Tree Planters’ Notes 41(3), 21–32.

332

Chapter 7

Schultz, R.C. and Thompson, J.R. (1991) The quality of oak seedlings needed for successful artiﬁcial regeneration in the central states. Proceedings The Oak Resource in the Upper Midwest Conference. University of Minnesota, St Paul, pp. 180–186. Schweitzer, C.J., Stanturf, J.A., Shepard, J.P., Wilkins, T.M., Portwood, C.J. and Dorris, L.C., Jr (1997) Large-scale comparison of reforestation techniques commonly used in the Lower Mississippi Alluvial Valley: ﬁrst year results. USDA Forest Service General Technical Report NC NC-188, pp. 313–320. Sharp, W.M. (1958) Evaluating mast yields in the oaks. Pennsylvania State University Agriculture Experiment Station Bulletin 635. Shearin, A.T., Bruner, M.H. and Goebel, N.B. (1972) Prescribed burning stimulates natural regeneration of yellow-poplar. Journal of Forestry 70, 482–484. Shigo, A.L. (1979) Tree decay an expanded concept. USDA Forest Service Agriculture Information Bulletin 419. Shugart, H.H. and West, D.C. (1977) Development of an Appalachian deciduous forest succession model and its application to assessment of the impact of the chestnut blight. Journal of Environmental Management 5, 161–179. Smalley, G.W. (1979) Classiﬁcation and evaluation of forest sites on the southern Cumberland Plateau. USDA Forest Service General Technical Report SO SO-23. Smalley, G.W. (1982) Classiﬁcation and evaluation of forest sites on the Mid-Cumberland Plateau. USDA Forest Service General Technical Report SO SO-38. Smalley, G.W. (1984) Classiﬁcation and evaluation of forest sites in the Cumberland Mountains. USDA Forest Service General Technical Report SO SO-50. Smalley, G.W. (1986) Classiﬁcation and evaluation of forest sites on the Northern Cumberland Plateau. USDA Forest Service General Technical Report SO SO-60. Smith, D.M. (1986) The Practice of Silviculture, 8th edn. John Wiley & Sons, New York. Smith, H.C. (1993) Development of red oak seedlings using plastic shelters on hardwood sites in West Virginia. USDA Forest Service Research Paper NE NE-672. Smith, H.C., Lamson, N.I. and Miller, G.W. (1989) An esthetic alternative to clearcutting? Journal of Forestry 87(3), 14–18. Sonderman, D.L. (1979) Guide to the measurement of tree characteristics important to the quality classiﬁcation system for young hardwood trees. USDA Forest Service General Technical Report NE NE-54. Sork, V.L. (1984) Examination of seed dispersal and survival in red oak, Quercus rubra (Fagaceae), using metal-tagged acorns. Ecology 65, 1020–1022. Stanturf, J.A., Auchmoody, L.R. and Walters, R.S. (1997) Regeneration responses of oak-dominated stands to thinning and clearcutting in northwestern Pennsylvania. USDA Forest Service General Technical Report NC NC-188, pp. 321–331. Starkey, D.A., Oak, S.W., Ryan, G.W., Taintner, F.H., Redmond, C. and Brown, H.D. (1989) Evaluation of oak decline areas in the South. USDA Forest Service Southern Region Protection Report R8PR R8-PR 17. Streng, D.R., Glitzenstein, J.S. and Harcombe, P.A. (1989) Woody seedling dynamics in an east Texas ﬂoodplain forest. Ecological Monographs 59, 177–204. Strobl, S. and Wagner, R.G. (1996) Early results with translucent tree shelters in southern Ontario. USDA Forest Service General Technical Report NE NE-221, pp. 13–23. Stroempl, G. (1983) Thinning clumps of northern hardwood stump sprouts to produce high quality timber. Ontario Forest Research Institute Forest Research Information Paper 104. Stroempl, G. and Secker, P.W. (1993) Guide to the group shelterwood cutting method for regenerating northern red oak. Ontario Forest Research Institute Forest Research Information Paper 120. Struve, D.K. and Joly, R.J. (1992) Transplanted red oak seedlings mediate transplant shock by reducing leaf surface area and altering carbon allocation. Canadian Journal of Forest Research 22, 1441–1448. Teclaw, R.M. and Isebrands, J.G. (1991) Artiﬁcial regeneration of northern red oak in the Lake States. Proceedings The Oak Resource in the Upper Midwest Conference. University of Minnesota, St Paul, pp. 187–197. Teclaw, R.M. and Isebrands, J.G. (1993a) An artiﬁcial regeneration system for establishing northern red oak on dry-mesic sites in the Lake States, USA. Annales des Sciences Forestieres 50, 543–552.

Even-aged Silvicultural Methods

333

Teclaw, R.M. and Isebrands, J.G. (1993b) Artiﬁcial regeneration of northern red oak in the Lake States with a light shelterwood: a departure from tradition. USDA Forest Service General Technical Report NC NC-161, pp. 185–194. Teclaw, R. and Zasada, J. (1996) Effects of two types of tree shelters on artiﬁcial regeneration of northern red oak in northern Wisconsin. USDA Forest Service General Technical Report NE NE–221, p. 68. Thompson, J.R. and Schultz, R.C. (1995) Root system morphology of Quercus rubra L. planting stock and 3-year ﬁeld performance in Iowa. New Forests 9, 225–236. Trimble, G.R., Jr (1974) Response to crop-tree release by 7-year-old stems of red maple stump sprouts and northern red oak advance reproduction. USDA Forest Service Research Paper NE NE-303. Trimble, G.R., Jr and Seegrist, D.W. (1973) Epicormic branching on hardwood trees bordering forest openings. USDA Forest Service Research Paper NE NE-261. Tryon, E.H. and Carvell, K.L. (1958) Regeneration under oak stands. West Virginia University Agricultural Experiment Station Bulletin 424T. Twery, M.J. (1987) Changes in vertical distribution of xylem production in hardwoods defoliated by gypsy moth. PhD dissertation, Yale University, New Haven, Connecticut. Twery, M.J. (1991) Effects of defoliation by gypsy moth. USDA Forest Service General Technical Report NE NE-146, pp. 27–39. USDA Forest Service. (1973) National forest landscape management: Vol. 1 – Concepts and principles. USDA Agriculture Handbook 434. USDA Forest Service. (1980) National forest landscape management: Vol. 2 – Timber. USDA Forest Service Agriculture Handbook 559. USDA Forest Service. (2001) Gypsy Moth in North America. (Internet web site, http://www.fs.fed.us/ne/morgantown/4557/gmoth, Morgantown, West Virginia.) Valentine, H.T. and Campbell, R.W. (1975) A simulation model of gypsy moth forest interaction. Forest Science 21, 233–238. Van Lear, D.H. and Waldrop, T.A. (1988) Effects of ﬁre on natural regeneration in the Appalachian Mountains. Society of American Foresters Publication 88-03, pp. 56–70. Veblen, T.T. (1992) Regeneration dynamics. In: Glenn-Lewin, D.C., Peet, R.K. and Veblen, T.T. (eds) Plant Succession, Theory and Prediction. Chapman & Hall, London, pp. 152–187. Waldrop, T.A., Buckner, E.R., Shugart, H.H., Jr and McGee, C.E. (1986) FORCAT: a single tree model of stand development following clearcutting on the Cumberland Plateau. Forest Science 32, 297–317. Walter, W.D. and Johnson, P.S. (2001) Sustainable silviculture for Missouri’s oak forests. USDA Forest Service General Technical Report NC (in press). Walters, R.S. (1993) Protecting red oak seedlings with tree shelters in northwestern Pennsylvania. USDA Forest Service Research Paper NE NE-679. Ward, W.W. (1966) Epicormic branching of black and white oaks. Forest Science 12, 290–296. Wargo, P.M. (1977) Armillaria mellea and Agrilus bilineatus and mortality of defoliated oak trees. Forest Science 23, 485–492. Wargo, P.M. and Haack, R.A. (1991) Understanding the physiology of dieback and decline diseases and its management implications for oak. Proceedings The Oak Resource in the Upper Midwest Conference. University of Minnesota, St Paul, pp. 147–158. Wargo, P.M. and Harrington, T.C. (1991) Host stress and susceptibility. In: Shaw, C.G., III and Kile, G.A. (eds) Armillaria Root Disease. USDA Forest Service Agriculture Handbook 691, Washington, DC, pp. 88–101. Webb, D.P. and von Althen, F.W. (1980) Storage of hardwood planting stock: effects of various storage regimes and packaging methods on root growth and physiological quality. New Zealand Journal of Forest Science 10, 83–96. Weigel, D.R. and Johnson, P.S. (1998a) Planting white oak in the Ozark Highlands: A shelterwood prescription. USDA Forest Service Technical Brief TB-NC TB-NC-5. Weigel, D.R. and Johnson, P.S. (1998b) Planting northern red oak in the Ozark Highlands: A shelterwood prescription. USDA Forest Service Technical Brief TB-NC TB-NC-6. Weigel, D.R. and Johnson, P.S. (2000) Planting red oak under oak/yellow-poplar shelterwoods: A provisional prescription. USDA Forest Service General Technical Report NC NC-210. Wendel, G.W. (1975) Stump sprout growth and quality of several Appalachian hardwood species after clearcutting. USDA Forest Service Research Paper NE NE-329.

334

Chapter 7

Wendel, G.W. (1980) Growth and survival of planted northern red oak seedlings in West Virginia. Southern Journal of Applied Forestry 4, 49–54. Wetteroff, J.J. (1993) Management and prediction of red oak decline in the Missouri Ozarks. MSc thesis, University of Missouri, Columbia. Will-Wolf, S. (1991) Role of ﬁre in maintaining oaks in mesic oak maple forests. Proceedings The Oak Resource in the Upper Midwest Conference. University of Minnesota, St Paul, pp. 27–33. Wilson, B.F. (1968) Red maple stump sprout development the ﬁrst year. Harvard Forest Paper 18. Windell, K. (1991) Tree shelters for seedling protection. USDA Forest Service Technical and Development Program Publication 9124-2834-MTDC. Windell, K. and Haywood, J.D. (1996). Intermediate results of a treeshelter durability study. USDA Forest Service General Technical Report NE NE-221, pp. 46–56. Wittwer, R.F. (1991) Direct seeding of bottomland oaks in Oklahoma. Southern Journal of Applied Forestry 15, 17–22. Wuenscher, J.E. and Kozlowski, T.T. (1971) Relationship of gas-exchange resistance to tree-seedling ecology. Ecology 52, 1016–1023. Yahner, R.H. and Smith, H.R. (1991) Small mammal abundance and habitat relationships on deciduous forested sites with different susceptibility to gypsy moth defoliation. Environmental Management 15, 113–120. Zaczek, J.J., Harding, J. and Welﬂey, J. (1997) Impact of soil scariﬁcation on the composition of regeneration and species diversity in an oak shelterwood. USDA Forest Service General Technical Report NC NC-188, pp. 341–348. Zastrow, D.E. and Marty, T.L. (1991) Tree shelter experiences. Proceedings The Oak Resource in the Upper Midwest Conference. University of Minnesota, St Paul, pp. 198–205.

8 Uneven-aged Silvicultural Methods

Introduction The traditional objective of uneven-aged silviculture is to create and maintain unevenaged stands for the sustained, even ﬂow of forest products. A regulated uneven-aged forest is one where all the individual stands or management units within it are unevenaged and, under continued management, collectively produce an even, sustainable yield of timber. Ideally, each individual stand is a self-contained sustained yield unit. By deﬁnition, an uneven-aged stand contains at least three age classes of trees that are closely intermingled on the same area (Smith, 1986). In reality, uneven-aged stands are usually comprised of even-aged groups of trees, but the groups are so small that visually it is difﬁcult to tell where one

group begins and another ends. An identifying characteristic of such stands is the vertical crown stratiﬁcation of trees within a relatively small area (Fig. 8.1). Maintaining the uneven-aged state requires periodic recruitment of reproduction into the overstorey. This must occur at least three times over a time span corresponding to age of the oldest trees retained in the stand. Recruitment of reproduction must be sufﬁcient to replace trees that die or that are periodically harvested and to maintain desired levels of stand density, yield and tree quality. Timber yields are obtained by periodically harvesting individual trees or small groups of trees within stands; the overstorey is never completely removed. Unlike even-aged silviculture, there is no rotation.

Fig. 8.1. Uneven-aged stands are characterized by an intermingling of trees of at least three age classes. 335

336

Chapter 8

Instead, forests are regulated through the periodic control of stand structure and density by timber harvesting. The number of years between harvests is termed the cutting cycle (Fig. 8.2). The periodic harvests reduce stand density to desired levels and manipulate stand size structure. They can also be utilized to modify species composition. In contrast to even-aged silviculture where stand structure changes constantly over a rotation, the goal of uneven-aged silviculture is to create a desired stand size structure and maintain it over time. Periodic harvests typically remove 25% or less of standing volume and usually remove trees across a wide range of diameter and

crown classes. One goal of uneven-aged silviculture is to continually improve the quality of the residual stand by removing the poor quality trees and retaining the high quality trees at each harvest. To attain nontimber objectives, den trees, snags and culls can be retained to satisfy wildlife, biodiversity and aesthetic objectives. In contrast to even-aged forests which are regulated through area control, uneven-aged forests are regulated through control of stand structure. The traditional goal is to create a stand structure that is sustainable over time and that produces a relatively constant yield of timber. The method therefore is sometimes said to be based on volume control.

Uneven-aged cutting cycle

A 1

2

3

4

5

6

Stand volume

100

0 0

20

40

B

60

80

100

120

Even-aged rotation 1

2

3

4

Stand volume

100

0 0

100

200 300 400 Years Fig. 8.2. Change in relative stand volume for forests under uneven-aged and even-aged silviculture: (A) volume change over six 20-year cutting cycles in an uneven-aged stand; and (B) volume change over four 100-year rotations in an even-aged stand. In this example, two intermediate cuttings are made within each rotation.

Uneven-aged Silvicultural Methods

However, it is only through the maintenance of a speciﬁed stand structure that effective control of volume is possible. The requisite stand structure is generally based on a diameter frequency distribution characterized by decreasing numbers of trees with increasing tree diameter. Bell-shaped, skewed bell-shaped or other irregular diameter distributions are not, as far as we know, sustainable naturally or silviculturally (Leak, 1996). In a fully regulated uneven-aged forest, the trees in each stand typically maintain a smooth, reverse Jshaped diameter frequency distribution. Two silvicultural methods are used in uneven-aged management: (i) single-tree selection, and (ii) group selection.

The Single-tree Selection Method Principles of application The single-tree selection method is a complete silvicultural system and a regeneration method. The periodic harvest of individual trees is carried out so that the

337

desired stand age and size structure, species composition and stocking are maintained. In principle, the maximum size of the canopy gaps resulting from harvesting single trees is limited to the crown areas of the largest trees cut. For oaks 16–20 inches dbh, the resulting canopy gaps range from about 1/100–1/30 acre, depending on tree dbh and stand density. In practice, however, larger canopy gaps are often created when small groups of dead and dying trees are salvaged. Sustaining stand size and age structure with the single-tree selection method depends on the recruitment (ingrowth) of reproduction within canopy gaps into the overstorey. Each pulse of ingrowth forms the heartbeat of the three-stage cycle of reduction in stand density, recruitment and stand growth (Fig. 8.3). The single-tree selection method differs from even-aged methods in the spatial scale and frequency of disturbance, and associated regeneration dynamics. An uneven-aged stand could be deﬁned as one where, after full regulation is attained, the forest-wide structure reoccurs at a ‘small’

Pulse of reproduction recruitment into overstorey

Regeneration process

Periodic reduction in stand density through harvesting

Increase in stand density through growth

Fig. 8.3. The single-tree selection method is sustained through the three-stage cycle of periodic reduction in stand density, followed by a pulse of reproduction recruitment into the overstorey, and then a renewed increase in stand density that continues to the end of the cutting cycle. Regeneration is a ubiquitous and continuous process in the well-managed uneven-aged forest.

Chapter 8

scale (e.g. <1 acre) at high probability (e.g. with ≥90% frequency) in space and time. This probabilistic view of the uneven-aged state recognizes the existence and acceptance of some departure from the ‘ideal’ or desired stand structure. This is an important qualiﬁcation because it recognizes that the silviculturally idealized uneven-aged state need not occur at all times nor everywhere within a stand. In applying the single-tree selection method, ﬁve stand characteristics should be considered at each periodic harvest in each stand or management unit: (i) the distribution of tree ages; (ii) the distribution of tree diameters; (iii) stand stocking; (iv) reproduction; and (v) species composition. With the exception of age distribution, all of the above stand characteristics also are considered at various times during the life of a managed even-aged stand. In principle, the need to assess all of these factors at every harvest entry makes the application of the single-tree selection method more complex than any even-aged method. However, the application of the single-tree selection method is usually simpliﬁed by assuming the desired stand age distribution follows the creation and maintenance of the desired diameter distribution. Consequently, stand structure in single-tree selection silviculture is usually described quantitatively by tree diameter distributions rather than age distributions. Much attention therefore is given to deﬁning the desired diameter distribution of an unevenaged forest. Uneven-aged stands can assume innumerable diameter distributions – some of which are also characteristic of even-aged stands. In general, uneven-aged diameter distributions are placed into two categories: balanced and unbalanced (or irregular) diameter distributions. Balanced diameter distributions form relatively smooth ‘depletion curves’ representing a continual decline in numbers of trees as diameter increases, whereas unbalanced distributions can assume any shape (Fig. 8.4). Uneven-aged stands with unbalanced diameter distributions are common to many oak forests, especially those repeatedly dis-

Trees per acre

338

Unbalanced

Balanced

Dbh Fig. 8.4. Examples of balanced and unbalanced uneven-aged diameter distributions.

turbed by undesigned timber harvesting or disturbances from ﬁre and wind. In natural uneven-aged stands, balanced diameter distributions commonly occur in old growth and other relatively undisturbed forests, especially those dominated by shade tolerant species. However, even-aged stands may form, at some point in their development, diameter distributions similar to uneven-aged stands. Thus, the shape of a diameter distribution, by itself, does not conﬁrm the existence of the uneven-aged state (Lorimer and Krug, 1983). Despite the ambiguities of using diameter distributions to infer stand age structure, the classical application of the single-tree selection method is based on creating and maintaining balanced diameter distributions. Balanced diameter distributions provide the basis for systematically sustaining even yields of timber – and the uneven-aged state itself. There is no established theory on how to sustain unbalanced structures. But even if they were sustainable, unbalanced structures are unlikely to produce even yields of timber over time. Stands with unbalanced structures may drift towards the even-aged state – especially stands comprised of relatively shade intolerant species. From a silvicultural perspective, a stand with an unbalanced structure is like a ship adrift, i.e. its future structural state is indeterminate.

Uneven-aged Silvicultural Methods

Balanced diameter distributions form the basis for the traditional approach to uneven-aged silviculture. They are generally referred to as reverse J-shaped distributions and are characterized by decreasing numbers of trees in successively larger diameter classes. However, the speciﬁc shape of such distributions may vary and different names have been applied to certain departures from the classical reverse-J shape. For example, reverse-J distributions with a somewhat ﬂattened mid-section are sometimes called rotated sigmoid distributions (e.g. Goff and West, 1975; Lorimer and Frelich, 1984; Leak, 1996). Regardless of their speciﬁc quantitative properties, such distributions result from natural forest development that involves three simultaneous processes: (i) tree mortality; (ii) tree growth; and (iii) ingrowth, i.e. the growth of reproduction into the overstorey. With respect to mortality, a reverse Jshaped distribution represents a depletion process caused when a proportion of trees in each diameter class fails to survive and grow into the next larger class. In managed stands, development over time is altered by a fourth process: the periodic reduction in stand density through timber harvesting. Harvesting reduces the rate of natural mortality from self-thinning, increases the growth rate of individual trees in the residual stand and accelerates the ingrowth of reproduction into the overstorey. Although various types of diameter frequency distributions representing depletion curves have been used to characterize natural and managed uneven-aged forests (Goff and West, 1975; Lorimer and Frelich, 1984; Hansen and Nyland, 1987; Leak, 1996), the distribution most widely used in silviculture to describe a balanced diameter distribution is mathematically deﬁned by the negative exponential function.

The negative exponential diameter distribution Figure 5.4 illustrates how the collective diameter distribution of a series of evenaged stands forms a negative exponential diameter distribution. In this context, the

339

negative exponential distribution could be equated to a forest-wide redistribution of trees in each age class to each acre of the forest. Each cohort or age class accordingly would occupy the same amount of space within each uneven-aged stand as it did across an entire regulated even-aged forest. Thus, the diameter distribution characteristics of the regulated even-aged forest are preserved, but at a much smaller spatial scale. Because the actual spatial scale is unspeciﬁed in uneven-aged silviculture, control of stand structure is substituted for area control. The negative exponential distribution provides an objective basis for organizing an uneven-aged forest that otherwise might be a chaos of tiny and spatially unrecognizable even-aged units. Although a great variety of uneven-aged, unbalanced stand structures are possible, a balanced distribution of diameters is required for the forest as a whole if there is to be a reasonable assurance of sustaining the uneven-aged state. We know of no theory or system for perpetuating a balanced collection of unbalanced stands where diameter distributions depart from the reverse J-shape or similar distributions representing depletion curves. Where sustained yield is not required, it may be sufﬁcient under some circumstances to maintain stands in loosely deﬁned unevenaged states and to obtain irregular yields whenever such yields are consistent with broader management objectives. For example, this strategy may be applicable to forests where some yield is desired but where an even ﬂow of timber products is impractical or secondary in importance to maintaining non-timber values. But perpetuating even a single stand in an unbalanced and silviculturally sustainable size structure is problematic without a systematic procedure for doing so. For these reasons, we present the single-tree selection method based on the application of the negative exponential diameter distribution to the individual stand, the basic silvicultural unit. The theory of the selection method assumes that within each stand a negative exponential diameter distribution in some

340

Chapter 8

form is silviculturally stable and sustainable in perpetuity. This does not necessarily mean that the distribution would occur naturally and remain stable in the absence of silviculture, but it does imply that the distribution is sustainable under the imposed silvicultural system. Moreover, the common occurrence of the negative exponential diameter distribution in old-growth hardwood forests and its relation to the process of self-thinning (Chapter 6) provide an underlying ecological basis for its application in single-tree selection silviculture. The negative exponential frequency distribution forms a straight line when the logarithm of the number of trees per unit area is plotted against tree diameter. Numbers of trees decrease at a constant exponential rate as tree diameter increases. In its non-linear form, the relation is given by: Ni = aebdi

[8.1]

where Ni is the number (frequency) of trees per acre within dbh class di, e is the base of the natural logarithm, and a and b are parameters that deﬁne stand density and slope shape, respectively. The parameter b is always negative and expresses the constant exponential decrease in N per unit change in diameter. Parameter b also can be used to derive the associated exponential depletion rate k from the relation: k = eb

[8.2]

Thus, for a given value of b, k is the proportion of trees of a given diameter class that occur in the next larger diameter class. In general the proportion, p, of trees of given diameter, di, that occur at a diameter that is n units larger than di can be computed as: p = kn

[8.3]

If we assume that k is a rate of change that represents a steady state in a stand at average maximum density, p is the proportion of trees surviving from di to di +n. 1There

Self-thinning is deﬁned as densitydependent mortality in stands at average maximum density (Chapter 6). The quantitative expression of self-thinning has largely evolved from observations in evenaged stands that are continually increasing in mean diameter. Although an unevenaged stand at average maximum density is not necessarily changing in mean diameter, it is undergoing self-thinning. In that context, self-thinning as a rate could be deﬁned by k, the constant probability that a tree of a given diameter class survives to the next larger class. An analogous deﬁnition of self-thinning rate for stands that are changing in average diameter would be the variable probability that a randomly selected tree in a stand of a speciﬁed mean diameter survives to a mean stand diameter one diameter unit larger (from Equation 6.2).1 Unlike the constant rate k, this probability is variable in relation to mean stand diameter. Both deﬁnitions of self-thinning apply only to stands that are at average maximum density. Silviculturists usually re-express the rate b as the ratio of the number of trees in one dbh class to the number in the next larger class. This ratio, or quotient (q), is given by (eb)−1 or k−1 for diameter classes of unit width. The resulting values have traditionally been referred to simply as q values (or q factors) by silviculturists.2 As discussed later, the value of q in ﬁeld applications is affected by the width of the diameter classes used to describe the diameter distribution, and quotients for any diameter class of width w are given by qw. For an observed diameter distribution, q can be derived by dividing the number of trees in one diameter class by those in the next larger class and repeating the process for the range of observed diameters; q then can be estimated as the average of all quotients. However, q can be more efﬁciently and accurately derived using readily available computer regression programs. These

is also a mathematical connection between Equations 6.2 and 8.1 (Zeide, 1984) and thus the two definitions. expression q is sometimes referred to as de Liocourt’s q, after its 19th century originator, the French forester de Liocourt (de Liocourt, 1898; Meyer et al., 1952). 2The

Uneven-aged Silvicultural Methods

programs also usually generate goodnessof-ﬁt statistics that are helpful in assessing how well the negative exponential function ﬁts an observed or hypothetical diameter distribution. They also can be used to ﬁt other mathematical functions that may better characterize stand structure (Leak, 1996). High values of q produce rapidly descending curves representing relatively large numbers of small trees and small numbers of large trees. Small values of q produce slowly descending curves representing relatively few small trees and relatively large numbers of large trees. In silvicultural applications in North American hardwood forests, commonly recommended values of q range from 1.1 to 1.4 based on 1 inch dbh classes (1.2 to 2.0 for 2 inch classes) (Trimble, 1970; Leak, 1978, 1987; Smith, 1980; Smith and Lamson, 1982; Leak et al., 1987; Law and Lorimer, 1989). Tree diameters in old-growth hardwood stands (including some dominated by oaks) often follow the negative exponential distribution (Meyer and Stevenson, 1943; Lorimer, 1980; Richards et al., 1995; Shiﬂey et al., 1995). Old-growth stands therefore may provide silviculturally valuable information on the diameter distribution possibilities for uneven-aged forests at or near average maximum density. Distributions may vary from extremely smooth (balanced) to extremely irregular (unbalanced), and from steep to gently sloping. Even within the same type of oldgrowth forest, the slope shape of the diameter distribution (and thus q) may vary appreciably depending on stand history, disturbance events, species composition, rates of succession and other factors (Meyer et al., 1952; Lorimer, 1980). Early studies of undisturbed, old-growth stands have shown that the parameter estimates a and b (Equation 8.1) derived from the diameter distribution of those stands are positively correlated (Meyer and Stevenson, 1943; Meyer et al., 1952). This correlation is a consequence of the natural development of such stands and the upper limits of stand density as discussed in

341

Chapter 6. Because parameters a and b deﬁne stocking and slope shape, respectively, and because old-growth stands are usually at or near average maximum density, any increase in the number of trees in one section of the diameter distribution must be accompanied by a decrease in the number of trees in another section. Such changes in the shape of the diameter distribution may result from slow, successional changes in species composition and associated changes in survival and diameter growth rates. They also may be related to sudden gap-scale disturbances that affect both stand composition and structure. Whatever the cause, the shape of the diameter distribution must self-adjust over time for stands to maintain themselves near average maximum density. For any ﬁxed stand density, the relation can be illustrated by comparing a series of negative exponential distributions and varying q (Fig. 8.5). As q (slope steepness) increases or decreases, a also must increase or decrease to maintain a constant stand density (Fig. 8.5, inset). The structure of undisturbed old-growth stands can provide valuable information on the natural dynamics and thus the shape characteristics (b or q) of uneven-aged diameter distributions for a given forest type. However, old-growth forests usually have stand densities at or near the maximum attainable for the site. In the case of oak forests, undisturbed old-growth stands also are relatively rare. Old-growth stands therefore are likely to be of limited value as silvicultural models for uneven-aged stand structure because, as discussed later in this chapter, successful application of the single-tree selection method in oak forests requires maintaining stands at relatively low densities. At stand densities below average maximum levels, it is hypothetically possible to silviculturally create, if not sustain, many different combinations of slope shape (q) and stand density (a). Uneven-aged management requires measurement and monitoring of the shape of a stand’s diameter distribution to evaluate how it conforms to a desired uneven-aged distribution. One method of doing this is

342

Chapter 8

160

q 1.4

300 240

Trees per acre

120

a

180 120

1.3

60

80

0 1.10 1.15 1.20 1.25 1.30 1.35 1.40 q 40

1.2

1.1 0 2

6

10 Dbh (inches)

14

18

Fig. 8.5. The negative exponential diameter distributions for four values of q (for 1 inch dbh classes) and a fixed stand density (80 ft2 acre−1 of basal area). Inset: The relation between a (from Equation 8.1) and q for the same distributions.

by visual inspection of graphs and by curve ﬁtting. Stands that do not initially show some predisposition toward a reverse Jshaped diameter distribution are often poor candidates for the single-tree selection system. Curve ﬁtting based on stand inventory therefore can be used to prioritize stands for placement under uneven-aged silviculture or to evaluate how they have responded to earlier silvicultural attempts to bring them into conformity with target diameter distributions, or ‘guiding curves’. As the term implies, the purpose of the guiding curve is to guide timber harvesting. An appropriate curve would be one that is compatible with the natural dynamics of the forest as well as with management objectives and constraints. In principle, periodic harvesting removes only the number of trees in each diameter class that are in excess of the guiding curve (Fig. 8.6). Selecting a guiding curve requires not only choosing a slope shape (q), but also a residual stand density (associated with a in Equation 8.1), and the diameter range across which the residual stand density is calculated. In choosing these values, both ecological and economic realities should be considered. A prerequisite to selecting a guiding curve is an understanding of how to specify a curve

that meets management goals for stand structure and density. Curve speciﬁcation thus differs from ﬁtting curves to existing tree data or selecting a guiding curve.

Specifying the distribution Specifying a guiding curve refers to deﬁning a diameter frequency distribution that meets predeﬁned goals for stand structure and stand density over the range of tree diameters to be considered. The current diameter distribution of a given stand may or may not be similar to the guiding curve speciﬁed for that stand by the silviculturist. A traditional approach to specifying the guiding curve has been ﬁrst to select the largest dbh class in the residual stand, i.e. the largest tree that will remain after a harvest. This diameter depends on species, site quality, management objectives and other factors. To derive a trial guiding curve, the number of trees in the largest dbh class can be arbitrarily set to one per acre. The number of trees in each of the smaller diameter classes then can be calculated by multiplying the next larger class by the selected q. From this, the basal area in each dbh class and total basal area or stocking can be calculated. The outcome

Uneven-aged Silvicultural Methods

343

70 60

Trees per acre

50 40 30 20 10 0 2

4

6

8

10 12 14 16 18 20 Dbh (inches) Fig. 8.6. The diameter distribution of an uneven-aged oak stand at 86% stocking (bars) and the guiding curve representing a q of 1.3 (for 1 inch dbh classes) and 60% stocking over the dbh range of 2–18 inches. The number of trees in each dbh class above and to the right of the guiding curve would be periodically harvested to bring the stand back to the guiding curve. Stocking is based on Gingrich (1967).

provides a trial stand structure with a shape speciﬁed by the value of q. However, the resulting stand density may be larger or smaller than desired. The trial curve can be adjusted upwards or downwards based on the ratio between its current basal area (or stocking) and the desired basal area (or stocking) until the appropriate curve is obtained (Marquis, 1978). Moser (1976) presented a more direct and accurate method for deriving a guiding curve. His method produces an exact curve for any pre-speciﬁed q and stand density. The method also accommodates expressing stand density as basal area, stocking per cent based on the tree–area ratio (Chisman and Schumacher, 1940; Gingrich, 1967), or crown competition factor (Krajicek et al., 1961). The two latter measures of stand density are given by: Density = c1∑Ni c2∑diNi

c3∑di2Ni

i i i [8.4] where i represents the ith diameter class, the summation is over all diameter classes, di and Ni are as deﬁned in Equation 8.1, and the coefﬁcients c1, c2 and c3 are derived by species or species groups using appropriate methods (Chisman and Schumacher, 1940;

Krajicek et al., 1961; Ernst and Knapp, 1985; Stout and Larson, 1988; also see Chapter 6; Stout and Nyland, 1986). Coefﬁcients applicable to several oak forest types are presented in Table 8.1; the single coefﬁcient for basal area, C3, is independent of species. An alternative expression of Equation 8.4 can be derived by substituting the righthand side of Equation 8.1 for Ni in Equation 8.4 so that: Density = c1∑aebdi c2∑diaebdi i

i

c3∑di2aebdi i

[8.5]

Then, solving for a (the density parameter):   a = Density/ ∑ c1 + c2di + c3di2 ebdi    i

(

)

[8.6]

Because b = −ln(q), Equation 8.6 can be restated as:   a = Density/ ∑ c1 + c2di + c3di2 edi ( − ln(q ))/w    i

(

)

[8.7] where w is diameter class width (Moser, 1976).

344

Chapter 8

Table 8.1. Coefficients for Equation 8.4 related to measures of stand density for several oak forest types.a

c1

Measure of stand density

Region/forest type

Basal area Stocking per cent

All Central Hardwood/ oak–hickoryb Wisconsin/ northern red oakc Wisconsin/ northern red oakd Allegheny Plateau of NW PA/ mixed oak–northern hardwoodse Central and southern bottomlands/pin oakf Central Hardwood/ Oak–hickoryg

Stocking per cent Stocking per cent Stocking per cent

Crown competition factor Crown competition factor

Coefficient c2

c3

0 0.00507

0 0.01698

0.005454 0.00317

0.01360

0.00930

0.00320

0.02476

0.04182

0.00267

0.0068718

0.016787

0.0019797

0.1480

0.04982

0.004193

0.01750

0.02050

0.00600

aSee

Chapter 6 for a discussion of measures of stand density. Gingrich (1967). Based on data from stands that were predominantly oak (white, scarlet, northern red and chestnut oaks); includes hickories and associated species in Central Hardwood forests. cFrom McGill et al. (1991). Based on data from stands that were predominantly northern red oak. Includes only contributions to stocking from ‘main canopy’ trees (intermediate, co-dominant and dominant crown classes). dFrom McGill et al. (1991). Based on data from stands that were predominantly northern red oak. Includes contributions to stocking from all trees including trees other than northern red oak. eFrom Stout (1991). Based on data from northwestern Pennsylvania, a transition zone between the Allegheny Hardwood Region and the Northern Hardwood Region. Associated stands are defined as those with at least 25% of basal area in oak and at least 25% in northern hardwood species, and at least 65% of basal area is comprised of those species plus the common associates of each. fFrom Krajicek (1967). (See Chapter 6 for a definition of crown competition factor.) gFrom Krajicek et al. (1961). bFrom

Equation 8.7 therefore can be used to calculate the stand density parameter a for speciﬁed values of stand density and of q for given diameter ranges. Moreover, it can be applied to various measures of stand density for several oak and other forest types (Tables 6.2 and 8.1). A more convenient method of specifying a guiding curve that does not require calculating the parameter a is given by the following equation:   N max = RSD/ ∑ Ai ⋅ q( Dmax − Di )/w ) [8.8]  i  where Nmax is the number of trees to be retained in the largest diameter class, RSD is the selected residual stand density (expressed as basal area per unit area, stocking per cent or crown competition

factor), i represents the ith dbh class, the summation is over all diameter classes, Ai is the basal area or tree area (expressed in the same units as RSD) of the tree of dbh Di, q is the selected stand structure represented by the guiding curve, Dmax is the maximum tree dbh represented by the guiding curve, Di is the midpoint dbh of the ith diameter class and w is dbh class width. After calculating Nmax (the number of trees in the largest diameter class), the number of trees in the other diameter classes can be calculated by starting with the largest diameter class and sequentially multiplying q times the number of trees in the diameter class to obtain the number in the next smaller diameter classes. Equation 8.8 is convenient for specifying a guiding curve using a computer spreadsheet program.

Uneven-aged Silvicultural Methods

space changes among diameter classes as q and stand density change (Fig. 8.7B). For high values of q, a relatively large proportion of growing space is occupied by small trees. In contrast, low values of q produce less skewed diameter distributions, and large trees comprise proportionately more growing space for a given stand density. For

Using Equations 8.1 and 8.8, families of curves can be generated to illustrate graphically how varying stand density and q jointly affect diameter distributions (Fig. 8.7A). The resulting curves facilitate comparison of alternative stand structures. The method also can be extended to describe graphically how the allocation of growing

Trees per acre per dbh class

180

150

345

Stocking % 100

A

80

q = 1.2 120

q = 1.4 60

90 40 60

30

100 80 60 40

0 2

4

6

8

10

12

14

16

18

20

Dbh (inches)

Stocking per dbh class (%)

12

Stocking %

B

10

q = 1.2 q = 1.4

8

100

6

80 60

4 2

40 100 80 60 40

0 2

4

6

8

10

12

14

16

18

20

Dbh (inches) Fig. 8.7. Distribution of numbers of trees and stocking by dbh classes for two q values and four stocking levels. (A) Families of curves for negative exponential diameter distributions for 1 inch dbh classes. Maximum tree dbh is set at 20 inches for q = 1.2, and 16 inches for q = 1.4. (B) Related distribution of stocking. Stocking per cent (relative stand density) is based on Gingrich’s (1967) minimum tree area equation for oak–hickory forests in the Central Hardwood Region.

346

Chapter 8

high values of q, reducing stand density reduces the variability among diameter classes relative to the amount of growing space occupied by any one diameter class. The stocking distribution curves thus provide a more detailed view of how growing space is allocated than do frequency distribution curves. Stocking distribution curves (Fig 8.7B) may be especially useful for assessing the allocation of growing space among trees in the larger diameter classes, which are often not apparent from frequency distribution curves (Fig 8.7A). Although sawtimber-size trees are relatively few in number, minor differences in their numbers per acre may represent large differences in periodic yield and economic value. The curves also can be used to examine graphically the effect of varying maximum tree diameter and/or varying the range of diameters across which stocking is deﬁned. As noted earlier in this section, the value of q (i.e. the ratio of the number of trees in one diameter class to the next larger diameter class) depends upon the width of the diameter classes used to describe the diameter distribution. For a given value of q, a given stand density and a given diameter class width, doubling the the diameter class width requires squaring the value of q to maintain the same slope in the diameter distribution curve (Fig. 8.8). In general, increasing the diameter class width by any multiple, m, requires recomputing q as qm in order to obtain comparable guiding curves. In applications, diameter class width used to inventory a stand and to compute q values for the current stand conditions and for guiding curves need to be considered. Equations 8.7 and 8.8 explicitly account for diameter class width.

Correlation between tree age and diameter It is commonly assumed that stands with balanced negative exponential diameter distributions represent correspondingly balanced age distributions. However, the age distribution of a stand can only be conclusively determined by directly observing tree ages (Lorimer and Krug, 1983). Even

though it is usually impractical to obtain the information necessary to describe accurately a stand’s age distribution, it is instructive to examine the possible relations between tree age and diameter in uneven-aged stands. One way to observe the relation between age and diameter distributions is to compare graphically their frequency distributions. The relative similarity or dissimilarity in the shape of the graphs reﬂects the degree of agreement between tree age and diameter. However, such comparisons may be misleading unless trees down to the very smallest diameters are included. If they are not, the ages of trees below the minimum observed diameter are not accounted for. The result is an incomplete and biased age frequency distribution (Loewenstein et al., 2000). Because of low to moderate correlation between tree age and diameter, the ages of the excluded trees may range from very young to relatively old. Discrepancies between the graphically apparent and actual age distributions of all trees may be especially great in uneven-aged stands because of the preponderance of small-diameter trees. Alternatively, the relation between tree age and diameter can be assessed by calculating the correlation coefﬁcient between observed ages and diameters. Stands can be categorized based on the magnitude of the coefﬁcient: (i) high correlation coefﬁcients (close to 1) associated with balanced all-aged stands (category 1 stands); (ii) mid-range correlation coefﬁcients associated with irregular uneven-aged stands (category 2 stands); and (iii) low correlation coefﬁcients (close to 0) associated with even-aged stands (category 3 stands). The correlation between tree age and diameter in an even-aged stand should be very small because trees have the same age regardless of dbh. Nevertheless, reverse J-shaped diameter distributions are not uncommon in even-aged stands. They are the rule in very young even-aged stands (Fig. 5.4), and also may occur after timber harvests in even-aged stands that leave numerous small-diameter trees and remove most large-diameter trees. Diameter distributions appearing to be uneven-aged therefore may be even-aged.

Uneven-aged Silvicultural Methods

347

100 A

90

q = 1.44 (2 inches dbh classes)

Trees per acre

80 70 60 50 40 30 20 10

q = 1.2 (1 inch dbh classes) 2

6

10

14

18

Dbh (inches)

100 B

Trees per acre

q = 1.44 (2 inch dbh classes)

10

q = 1.2 (1 inch dbh classes)

1 2

6

10

14

18

Dbh (inches)

Fig. 8.8. (A) Equivalent dbh frequency distributions for two diameter-class widths: 1 inch and 2 inches (w in Equation 8.8). The relation is illustrated for a q of 1.2 (for 1 inch dbh classes) and the corresponding value of q = 1.44 (or 1.22) for 2 inch classes. (B) When the slopes of the two distributions are linearized by logarithmic transformation of the vertical axis (as shown), the curves are parallel, i.e. they have a common slope (parameter b in Equation 8.1). However, this is only true if the dbh ranges of the two distributions share the same diameter class mid-points for the smallest and largest dbh’s considered. The greater number of trees per acre in each 2 inch dbh class elevates the curve above that for 1 inch classes (i.e. the trees per acre in two 1 inch classes must be combined to equate to a 2 inch dbh class). There are also 7% more trees per acre represented by the curve for 2 inch dbh classes. This difference results from the wider dbh interval for the smallest and largest dbh classes for the 2 inch curve (as shown).

348

Chapter 8

Trees per acre

Category 1 stands represent the theoretically ideal state under uneven-aged management where the observed frequency distributions of both diameter and age would form a relatively smooth, exponentially declining curve. Sustaining such distributions requires the periodic recruitment of trees (reproduction) into the smallest overstorey size class. This recruitment and that of existing overstorey trees into successively larger diameter classes must occur at approximately the same rate across the entire range of diameters in order to maintain the negative exponential curve shape. To realize this objective, the selected guiding curve must be consistent with a forest’s growth and regeneration dynamics. Thus, the guiding curve is usually based on what is known about the dynamics of the forest to be managed. An ideal guiding curve would be one that remains relatively stable from one cutting cycle to the next (Fig. 8.9). The selected guiding curve accordingly should maintain its shape and also sustain the desired residual stand density, species composition, volume growth, tree quality and other stand attributes deemed important. After selecting a guiding curve, the

diameter distribution can be created and maintained by periodically reducing stand density through timber harvesting, typically at intervals of 10–20 years. Maintaining the distribution requires cutting trees across a wide range of diameter classes, not just the largest. Failure to maintain a stable diameter distribution over time can result from inadequate ingrowth of reproduction of desired species into the overstorey, failing to harvest trees in smaller diameter classes, changes in species composition, irregular mortality and other factors. Changes in diameter growth rates may be associated with changes in species composition that are often the result of succession and difﬁcult to control silviculturally. Setting a high minimum diameter for harvested trees also will result in poor control over both the shape of the dbh distribution and stand density. The portion of the diameter distribution below the minimum harvest dbh tends to develop towards its natural ecological equilibrium, which does not necessarily preserve the desired overall shape of the distribution. Any of these problems can produce stands that fall into category 2.

En

Beg inni ng

do

fc

of c u

u tt

in g

cy

c le

q tting

cyc le

1.7

1.7

Dbh Fig. 8.9. The theory and application of guiding curves assume that the shape of the selected diameter distribution, and thus q, remains relatively constant during the cutting cycle.

Uneven-aged Silvicultural Methods

Category 2 stands arise from factors that fall into two general groups: • Variation (dispersion) of diameters within a single age class (cohort). • Spatial and temporal variation related to site factors and stand characteristics. Variation in diameters within a cohort is related to competition within and among cohorts and will occur even in the absence of other sources of environmental variation. This dispersion results from genetic variation in growth rate and other factors such as susceptibility to insects and disease, and imperfect spacing of trees. Imperfect spacing, which is partially related to clumped spatial distributions of reproduction, eventually results in unequal growth rates among trees (Rogers, 1983). Small trees within a single young age cohort have relatively limited diameter dispersion. But this dispersion increases with time as the mean diameter of the cohort increases. The magnitude of dispersion is evident from stand tables for unmanaged even-aged oak stands at average maximum density in the Central Hardwood Region (e.g. Fig. 5.4). Within 80-year-old cohorts on average sites, diameters span a 12–14 inch range (Schnur, 1937). Although the dynamics of an even-aged stand are not the same as those of an uneven-aged stand, evenaged stand tables illustrate the principle of diameter dispersion within age cohorts that occurs even with uniform stand conditions and limited disturbance. This dispersion tends to be compounded in uneven-aged stands due to competition within and among age cohorts and by mixtures of species with different growth rates. Consequently, in a naturally developing uneven-aged stand, a high correlation among age and diameter would be unlikely if for no other reason than dispersion of tree diameters within age cohorts. Periodic harvesting tends to reduce within-cohort diameter dispersion if cutting were concentrated on trees of low vigour and slow growth across the entire diameter range. This practice compresses the left tail of each cohort’s diameter distribution and thus reduces the diameter range

349

of each age cohort. This tends to increase the correlation between tree diameter and age within the residual stand. When trees are harvested across only a portion of the diameter range, we might expect the correlation coefﬁcient between age and dbh to fall somewhere between low and high extremes, depending on how much of the diameter range is unregulated. For example, in a forest where only trees 11 inches dbh and larger are cut, stand structure below 11 inches dbh is uncontrolled and free to develop naturally. In one such uneven-aged oak forest in the Ozark Highlands, tree diameter explained 43% of the variation (r = 0.65) in the ages of oaks 2 inches dbh and larger after two cutting cycles (Fig. 8.10A). The associated diameter distribution nevertheless is relatively well balanced despite the irregular age distribution (Fig. 8.10B). Relatively balanced diameter distributions therefore are possible within very irregular age distributions. Irregularity in the timing and intensity of harvesting also may reduce the correlation between tree age and diameter by increasing heterogeneity in stand stocking. Because diameter growth is very sensitive to stand density, anything that increases spatial heterogeneity in stand density is likely to decrease the correlation between tree age and dbh. Other factors that increase spatial heterogeneity in stand density include irregular mortality, natural disturbances, cutting cycles that ﬂuctuate with market conditions and inconsistent application of cutting guidelines and prescriptions. Low correlations between age and dbh occur in even-aged stands that are in silvercultural transition to uneven-aged structure, but have progressed through only one or two cutting cycles. Normal variation in the oak regeneration process (Chapter 3) also may inﬂuence the correlation between age and dbh in uneven-aged stands. Most applications of uneven-aged silviculture are likely to begin with stands in various irregular uneven-aged states with a range of diameter distributions ranging from poorly balanced to well balanced. Theoretically, an ecologically appropriate

350

Chapter 8

100 A

Tree age (years)

80

60

40

20

0 4

8

12

16

20

Dbh (inches)

140 B Number of trees

120

Number of trees

100 80 60

160 140 120 100 80 60 40 20 0

0

20

40

60

80 100 120 140

Tree age class (years) 40 20 0 2

4

6

8

10

12

14

16

18

20

2

Dbh (inches) Fig. 8.10. Age and diameter distributions of 578 randomly selected oaks in a 640 acre uneven-aged oak forest in the Ozark Highlands of Missouri. (A) The bivariate distribution of ages and diameters of trees 100 years old or less (98% of the sample population). Dbh accounts for 43% of the variation in tree age; the correlation coefficient is 0.65. (B) The observed frequency distribution of trees in 1 inch dbh classes (bars) and the negative exponential function fitted to the observed frequencies (curved line); q = 1.32 for 1 inch dbh classes. (Inset) The age frequency distribution of the same population. The age distribution is incomplete and biased because observations are truncated at the 2 inch dbh class (see text for related discussion). This forest has been managed for 40 years by the single-tree selection method and has been through two cutting cycles. The predominant species are white, scarlet, black, northern red and chinkapin oaks. (Adapted from Loewenstein, 1996, used with permission.)

Uneven-aged Silvicultural Methods

and consistently applied guiding curve should gradually increase the correlation between tree age and dbh. However, it is unlikely that the correlation would ever approach unity for the reasons cited, and because monitoring tree ages and cutting to maintain an actual age distribution is impractical.

Sustainability and stability Silviculturally, the term sustainability usually connotes a forest or stand attribute that can be maintained to perpetuity. This includes the traditional concept of sustained yield, which implies that timber yield is sustainable indeﬁnitely through a combination of controlled periodic harvesting and other silvicultural practices. The sustainability concept also can be applied to other forest attributes such as wildlife habitat, water yield, biodiversity and aesthetic values. Regardless of context, sustainability implies permanence and continuity. It is nevertheless difﬁcult to verify scientiﬁcally that a given forest attribute such as a particular species composition and size structure is indeed sustainable. This follows for two reasons. First, the concept of sustainability infers a centuries-long time period that lies beyond managerial or scientiﬁc experience. Second, ecosystems are continually changing and thus are not intrinsically stable. Silviculture nevertheless implies some degree of control over natural processes and therefore predictability regarding when and where certain forest conditions will occur. Silviculture is used to direct and sometimes ‘suspend’ succession within a narrow range of ecological states that satisfy human objectives. Speciﬁcally, unevenaged silviculture attempts to suspend stands and forests at a relatively constant species composition, age structure and size structure. This makes uneven-aged silviculture more complex than even-aged silviculture. The term silvicultural stability is used here to refer to an ecological state or narrow range of states that can be maintained through silviculture.

351

An ideal uneven-aged silvicultural system would create an uneven-aged state that is both economically viable and ecologically sustainable. And once the desired state has been created, it must be maintained by periodically harvesting trees in excess of requirements for an ecologically appropriate guiding curve. The sustainability of the resulting tree diameter distribution will depend on both ecological and economic factors. To be ecologically sustainable, both the stand density and the shape of diameter distribution must be consistent with site factors and the biological characteristics of the tree species present, including growth and survival rates, shade tolerance and regeneration dynamics. An appropriate stand size structure also must be consistent with sustaining the selected residual density, which in turn must provide periodic timber harvests that satisfy economic and operational requirements. Silvicultural application of the negative exponential distribution assumes that only a fraction of the trees in any one diameter class grow into the next larger class and that this fraction is constant across all diameter classes. Sustainability of stand structure requires a sustained ﬂow of reproduction into the overstorey. Timber yields are realized through the periodic harvest of the trees in each dbh class that are excess to the silviculturally prescribed guiding curve (Fig. 8.6), and these harvests are necessary to maintain the desired diameter distribution. Trees therefore must be harvested across a wide range of tree diameters, not just the largest diameters. Concentrating each harvest on only the largest diameter classes is unlikely to sustain the necessary uneven-aged stand structure or stocking. This practice, sometimes called ‘selective cutting’, results in poor control of stand density, stand structure and tree quality, and is unlikely to be sustainable. Despite the similarity between the terms ‘selective cutting’ and ‘selection cutting’ (or ‘single-tree selection’), the two practices have little in common silviculturally. Too often this difference is not recognized or understood.

352

Chapter 8

Maintenance of the desired stand structure and composition may not be assured even when trees are harvested across the entire range of diameters in accordance with a guiding curve. Doing so assumes that the resulting tree survival and growth will, through successive harvests, conform to the artiﬁcially imposed q. But conformity requires that q remain constant across the entire range of diameters as stand density rebounds after each harvest. However, there is no a priori reason to assume that this will happen. Survival rates among trees of different diameter classes may change as a function of stand density. Thus, if thinning to a speciﬁed q and a residual stand density results in higher rates of survival among small-diameter trees than large-diameter trees, the resulting q value over the low end of the diameter range eventually would assume a lower value than that of trees in the upper end of the diameter range. Maintaining the desired q also requires a rate of ingrowth of trees into the overstorey that may not conform to that assumed by the selected q. Such non-conformities may be an indication that uneven-aged silviculture and the single-tree selection method are: (i) inappropriate, or (ii) appropriate but the selected guiding curve is inconsistent with stand dynamics. From outward appearances, the singletree selection method may seem to represent the most ‘natural’ of the silvicultural systems. The naturalness of the method nevertheless may be deceptive on two counts. First, the success of the method depends on a silviculturally controlled rate of recruitment of natural reproduction into the overstorey that must continually balance periodic removals. Sustained regeneration of suitable species is therefore essential to the method’s success. Second, the method requires relatively intensive control of stand structure and density. Sustaining the system consequently depends on perpetuating the three-stage cycle of periodic reduction of stand density, recruitment of reproduction into the overstorey and stand growth (Fig. 8.3). For all but the most shade tolerant species, greater silvicultural control is required to apply and sustain effectively the

single-tree selection method than any other silvicultural method. This is the major problem in applying the method to the relatively shade intolerant oaks with their erratic seed production cycles, seedling establishment and other regeneration uncertainties. Despite these problems, there is evidence that the method is suited to some North American oak forests.

Applicability to oak forests Based on experience in the oak-dominated forests of the Central Hardwood Region, Gingrich (1967, p. 47) was pessimistic about successfully applying the single-tree selection method to oak forests: Unlike northern hardwood stands, upland hardwood stands (except for short periods after heavy cutting) rarely have high frequencies in the small diameters because such stands lack the tolerant species needed to maintain the inverse, J-shaped distribution. Indeed, there is no evidence that the common species comprising upland hardwood stands can maintain the inverse, J-shaped distribution naturally, and formal research attempts to maintain it by cutting have failed.

Roach (1968, p. 12) expressed similar pessimism: Selection cutting is a good forestry theory, although it is still only a theory and for timber production has not yet been either proved or disproved. I dislike to use it in the mixed oaks and upland hardwoods for three reasons: The expense required for its proper application is exorbitant … We do not get satisfactory reproduction of preferred species under it. And when reproduction is slow or uncertain, selection cutting is very difﬁcult to regulate for sustained yield, and the chances for error are numerous and serious.

If we were to heed the advice of these experienced forest scientists, we would not undertake the uneven-aged management of oak forests, or perhaps do so only apprehensively. However, the oaks cover a broad geographic area and include many different forest types, each with different ecological attributes.

Uneven-aged Silvicultural Methods

Possibilities and limitations Roach and Gingrich raise two ecologically and silviculturally important issues related to the uneven-aged silviculture of oaks: (i) sustaining the regeneration process; and (ii) creating and sustaining a reverse J-shaped (negative exponential) diameter distribution. Both problems are central to sustaining uneven-aged stands. Maintaining a negative exponential diameter distribution requires the periodic establishment and development of oak reproduction. Although it may be possible to create the requisite diameter distribution in the absence of adequate regeneration, it cannot be sustained unless there is periodic recruitment of reproduction into the overstorey. Sustained, periodic recruitment of oak reproduction into the overstorey determines the success or failure of the single-tree selection method in oak forests. The applicability of the method to a given stand or forest accordingly depends on the intrinsic regeneration characteristics of a forest, especially its capacity to accumulate, or build up, oak reproduction over successive acorn crops (Chapter 3). Such accumulation is favoured on the drier sites where oaks tend to persist.

353

In contrast, oaks are less likely to regenerate successfully on the more mesic sites because the requisite accumulation of oak reproduction is less likely to occur there. On those sites, maintaining oaks is difﬁcult regardless of regeneration method. Additional problems are created by the selection forest’s multilayered canopy and the perpetual shade it casts on developing oak reproduction. Even in stands that are intrinsic accumulators of oak reproduction, the recruitment dynamics of oak reproduction into the overstorey are poorly understood. Generally the movement of oak reproduction into the overstorey will be greatest shortly after harvest and then decline as stand density and crown closure increase. Decline in recruitment also may occur as the population of large reproduction ready for recruitment into the overstorey is temporarily exhausted by its own ingrowth into the overstorey. Signiﬁcant recruitment of oaks into the overstorey therefore may occur only during a fraction of a cutting cycle in any given stand (Fig. 8.11). This fraction may not necessarily represent consecutive years, depending on weather and other environmental or biotic factors

Fig. 8.11. Oak reproduction growing beneath an uneven-aged forest canopy on a site that favours the accumulation of oak seedlings and seedling sprouts over successive acorn crops. Growth of reproduction into the overstorey depends on the creation of canopy gaps and the presence of seedling sprouts with large root systems and the capacity for rapid shoot growth. Even when canopy conditions are favourable, significant recruitment may occur only during a fraction of a cutting cycle, depending on the availability of ‘recruitment ready’ reproduction.

354

Chapter 8

It is consequently unlikely that a reverse Jshaped dbh distribution will consistently extend down into the reproduction size classes. Oak reproduction density and size in a selection forest varies with overstorey density (Fig. 8.12), which in a given stand is correlated with time since last harvest. Moreover, oak reproduction is characterized by relatively unpredictable shoot growth, dieback, resprouting and mortality – which may be only partially controlled by stand density (Liming and Johnston, 1944; Johnson, 1979; Crow, 1992; Lorimer et al., 1994; Dey and Parker, 1996; Dey et al., 1996). Consequently, the distribution of reproduction is often best expressed probabilistically. Diameter distributions of unmanaged even-aged oak stands typically form a bellshaped or normal distribution by the time they reach a mean diameter of about 8 inches (Chapter 5). However, in oak-dominated ecosystems that are successional to shade tolerant non-oaks, the tolerant species typically develop a sub-canopy with a reverse J-shaped diameter distribu-

tion. The oak component, which continually shrinks in importance as shade tolerant species capture the site, maintains its bell-shaped diameter distribution until it virtually disappears as a component of the stand. This pattern of successional displacement frequently occurs in oak forests that are successional to shade tolerant species such as red and sugar maples and American beech or to aggressive intolerant species such as yellow-poplar (Lorimer, 1980, 1981, 1983, 1984; McGee, 1984; Nowacki et al., 1990; Cho and Boerner, 1991). The diameter distribution of the overall stand may be reverse J-shaped, but the oaks themselves retain a normal diameter distribution that may be obscured at a scale that depicts the overall distribution (Fig. 5.11). The oak’s inability to maintain a negative exponential diameter distribution because of its shade intolerance and related regeneration failures ultimately seals its successional fate in forests that are recalcitrant accumulators of oak reproduction (Fig. 3.18).

0.7 0.6 0.5 Oak reproduction (no. acre–1) 0.4

P

≥100 0.3 ≥200

0.2 0.1

≥400 0.0 0

20

40

60

Overstorey basal area

80 (ft2

100

120

acre–1)

Fig. 8.12. The estimated probability of occurrence (P) of oak reproduction densities of at least 100, 200 or 400 genets per acre that are at least 4.5 ft tall and up to 1.6 inches dbh in relation to overstorey density in an uneven-aged forest in the Ozark Highlands of Missouri. The oak reproduction is predominantly white, black and scarlet oaks. Based on a logistic regression and reproduction densities observed at a scale of 1/50 acre (from Larsen et al., 1997, used with permission).

Uneven-aged Silvicultural Methods

Moreover, reduction of stand density by itself is unlikely to resolve the problem of inadequate oak regeneration in stands undergoing successional displacement. Reduction in stand density may actually hasten the rate of displacement (Abrams and Nowacki, 1992; Jenkins and Parker, 1998). Exceptions may occur where wildﬁres, prescribed burning or other disturbances recurrently reduce stand density and simultaneously facilitate the build-up of oak reproduction. Prescriptions that reproduce these effects may be feasible under even-aged silviculture, where they can be effectively applied once or twice towards the end of the rotation (see Chapter 7). But to effectively sustain the requisite build-up of oak reproduction and its recruitment into the overstorey of an uneven-aged stand, these techniques may need to be applied during each cutting cycle. The economic practicality of such intensive practices and attendant risks are questionable. Practical limitations therefore would appear to outweigh possibilities for effectively applying the single-tree selection method to forests that are recalcitrant accumulators of oak reproduction and have strong successional patterns that favour non-oaks. Much of the experience in applying single-tree selection nevertheless has been obtained in such oak forests. The collective experience has produced a history of documented failures in applying the single-tree selection method to oaks and other relatively intolerant hardwoods (Gingrich, 1967; Roach and Gingrich, 1967; Trimble, 1970, 1973; Schlesinger, 1976; Smith, 1980; Della-Bianca and Beck, 1985; Smith and Miller, 1987). In turn, those results have often been generalized as a silvicultural principle that single-tree selection is prone to failure in all oak forests (e.g. Sander, 1977; Hibbs and Bentley, 1983; Marquis and Johnson, 1989; Marquis et al., 1992). Applying the single-tree selection method to oak forests that intrinsically accumulate oak reproduction represents other possibilities. Because these forests are generally the more xeric and less productive ecosystems, we might question the

355

economics of applying this relatively costly silvicultural method to them. From a strictly ecological perspective, it nevertheless would appear to be more feasible to apply the method there than to stands that are recalcitrant accumulators for two related reasons: (i) regeneration dynamics, and (ii) diameter distributions. Whereas the regeneration dynamics of xeric oak forests have been earlier discussed (Chapter 3), the signiﬁcance of diameter distributions requires explanation. Relatively undisturbed xeric oak stands of the eastern United States are often comprised of numerous small-diameter white oaks. This condition is exempliﬁed by, but is not unique to, the white oak–black oak forests growing on deep sands in northern Lower Michigan. These forests, which range in site index from about 50 to 60, often form unbalanced uneven-aged stands (Johnson, 1992). Where there has been little or no previous management, the diameter distribution of white oak approaches a negative exponential distribution. In contrast, black oak (often mixed with northern pin oak) typically forms a bell-shaped diameter distribution (Fig. 8.13A). These inherently different diameter distributions reﬂect species’ differences in shade tolerance, survival rates and recruitment rates into the overstorey. The composite size structure of a typical stand approaches, but does not attain, a balanced negative exponential distribution. Diameter distributions are often deﬁcient in the 2–3 inch diameter classes (Fig. 8.13B). Similar diameter distributions have been reported for relatively undisturbed oak stands in the Ozark Highlands of Missouri (Shiﬂey et al., 1995). How then might the single-tree selection method be applied to these oak forests? To answer, assume that Fig. 8.13B characterizes the current size structure of one such forest. Note that there is a deﬁciency of trees in the two smallest dbh classes. A plausible explanation for this deﬁciency might be that the smaller oaks are dying from suppression at a relatively high rate and that there is insufﬁcient recruitment of reproduction into the overstorey at the prevailing overstorey density (81%

356

Chapter 8

32 28 Trees per acre

A

Black oak White oak

24 20 16 12 8 4 0

80

2

4

8

10

12

14

16

18

Stocking % (q =1.3)

70 Trees per acre

6

20

B

80

60

60

50

50

40 30 20 10 0 2

4

6

8

10

12

14

16

18

20

Dbh class (inches) Fig. 8.13. Representative size structure of black oak–white oak stands in northern Lower Michigan. (A) The distribution of diameters by species for a stand at 81% stocking (33% white oak, 48% black oak, with mean diameters of 5.1 and 8.0 inches, respectively). (B) The composite diameter distribution (bars) and guiding curves for three levels of stocking (based on Gingrich, 1967) over the dbh range of 2–20 inches. (Adapted from Johnson, 1992.)

stocking). If that were the case, we should be able to increase the survival and recruitment rate through appropriate periodic reductions in stand density. Numbers of small oaks accordingly would eventually move upwards towards a balanced negative exponential distribution such as that represented by the 50% or 60% stocking curves illustrated in Fig. 8.13B. This is the case in similar oak forests of the Ozark Highlands of Missouri. There, sustain-

ing the requisite recruitment of oak reproduction requires reducing total stand density to about 50% stocking at the beginning of each 15 to 20 year cutting cycle (Larsen et al., 1997, 1999; Loewenstein et al., 2000). Although this is 5–10% below the minimum stocking usually recommended for oak forests in the Central Hardwood Region (Roach and Gingrich, 1968; Sander, 1977), stand densities rebound to about 75–80% stocking by the end of a 20 year cutting cycle on average sites.

Uneven-aged Silvicultural Methods

Such densities, combined with a guiding curve based on a q of 1.2–1.3 (for 1 inch dbh classes), appear to be consistent with the natural regeneration and stand dynamics of these forests (Larsen et al., 1999). A 155,000 acre forest in the Ozark Highlands dominated by white, black and scarlet oaks has been successfully managed for more than 40 years using the single-tree selection system (Fig. 8.14). In this forest, only trees ≥10 inches dbh are harvested. Analyses of continuous forest inventory records have shown that, after two cutting cycles, diameter distributions above the minimum cutting threshold have maintained a balanced negative exponential diameter distribution that reoccurs at a small (<1 acre) scale (Loewenstein, 1996; Loewenstein et al., 2000). However, the shape of the distribution below the threshold ﬂuctuates signiﬁcantly (Wang, 1997). This ﬂuctuation may be caused by the

357

interdependency of stand densities above and below the threshold diameter, which in turn is perpetuated by periodic cutting (Wang, 1997). Accordingly, a partial cut of trees ≥10 inches dbh instantaneously steepens the diameter distribution curve, which increases q. Further curve steepening may result from reduced mortality of small trees and increased recruitment of reproduction into the overstorey associated with the reduction in the density of trees above the 10 inch threshold diameter. The steepening trend eventually reverses itself as stand density above the threshold gradually rebounds. This rebounding, in turn, slows the recruitment rate of reproduction into the overstorey and increases the mortality rate of small trees as suppression by larger trees intensiﬁes. This trend gradually ﬂattens the diameter distribution, which reduces q. The result is an oscillation between steepening and ﬂattening of the diameter distribution

Fig. 8.14. An uneven-aged oak stand in the Ozark Highlands of Missouri. By maintaining moderately low stand densities and using the single-tree selection method, an uneven-aged forest comprising a reverse J-shaped diameter frequency distribution representing all age classes of oaks can be sustained. Unlike the oak forests of many other ecoregions, these distributions are silviculturally sustainable because of the oak regeneration dynamics characteristic of the region. (USDA Forest Service, North Central Research Station photograph.)

358

Chapter 8

about the 10 inch harvest threshold diameter, which is perpetuated by periodic partial cutting above the threshold. The threshold diameter effectively acts as fulcrum about which the shape of the diameter distribution below the threshold diameter oscillates (Fig. 8.15). The portion of the stand ≥10 inches in diameter, while largely controlled silviculturally, is also responding to natural stand dynamics. Trees smaller than 10 inches dbh, while largely controlled by natural stand dynamics, are also inﬂuenced by the periodic reduction on stocking of larger trees. White oak plays a key role in this process because of its predominance in the small diameter classes (including reproduction), shade tolerance and persistence as a canopy dominant. The application of the single-tree selection method in the Ozark Highlands thus produces a self-limiting instability of the diameter distribution characterized by an

oscillation in q below the minimum cutting diameter. This ﬂuctuation in slope shape tends to return q to a relatively stable value once stand densities return to about 70% stocking (Wang, 1997). Self-limiting instability of the diameter distribution therefore appears to be a characteristic of the forest’s response to the silvicultural system. Regardless of the explanation, there is substantial evidence that the single-tree selection method can be used to sustain uneven-aged oak stands in the Ozark Highlands.

General considerations Apart from the preceding example, the long-term sustainability of the single-tree selection method in oak forests that are intrinsic accumulators of oak reproduction remains to be veriﬁed in practice. Nevertheless, factors consistent with the

70 50 Merchantable trees

20

q =

1.

.28

uc

35

10

T

=1

Fl

q

Trees per acre

30

8

tu

at

ion

in

q

q Gu

6 Unmerchantable trees 4

6

1.

3

id q ing = 1. curv 3 e

4 2

=

8

10

12

Dbh (inches) Fig. 8.15. Fluctuation in q for trees smaller than 10 inches dbh during a single cutting cycle in an uneven-aged oak forest in the Ozark Highlands. Diameter distributions are shown in logarithmic scale to emphasize the magnitude of fluctuation. Trees below the 10 inch threshold dbh (T) are unmerchantable and are not harvested. In this ecosystem the threshold diameter appears to act as a fulcrum about which the shape of the diameter distribution to the left of T oscillates. The resultant ‘self-limiting instability’ of q below T is perpetuated by periodic partial cutting to a guiding curve applied to trees with dbh T (see text). (Adapted from Wang, 1997, used with permission).

Uneven-aged Silvicultural Methods

method’s application to those forests include a natural regeneration process that sustains an adequate oak regeneration potential and a naturally occurring diameter distribution that often approaches a negative exponential distribution. Although it is often assumed that oak stands are typically even-aged, they can occur as uneven-aged populations at the relatively small spatial scales (e.g. ≤1 acre) required in selection silviculture (Loewenstein, 1996; Loewenstein et al., 2000). Both regeneration and stand structure therefore are potentially adaptable to creating and sustaining negative exponential diameter distributions in these ecosystems. Such distributions are most likely to occur when stands are managed at moderately low residual densities and low q values. Logical candidates for uneven-aged silviculture accordingly include oak stands with a pre-existing wide range of diameters and ages that occur on sites that are intrinsic accumulators of oak reproduction. In selecting a guiding curve for application to oak forests that are potentially compatible with the single-tree selection method, two factors are of paramount importance: (i) the steepness of the curve slope as expressed by q, which determines the proportion of small to large diameter trees; and (ii) the height (intercept) of the guiding curve, which determines stand density. An appropriate guiding curve therefore would be one that deﬁnes a moderately low overall stand density (e.g. 60% stocking) and q values within the range of 1.1–1.3 for 1 inch dbh classes. This combination would appear to be consistent with the light requirements for oak reproduction and the natural dynamics of these ecosystems. However, similar oak forests in other regions may respond differently. Guiding curves for values of q that exceed 1.3 (based on 1 inch dbh classes) are unlikely to be compatible with the natural dynamics of oak stands that intrinsically accumulate reproduction because natural regeneration in those stands will be unlikely to produce enough oak reproduction or recruit it to the overstorey.

359

The success of the method will partly depend on a satisfactory growth response of trees in inferior crown classes to periodic reductions of stand density. At least two studies have shown that the response of overtopped white oaks to release is highly variable and in part depends on tree vigour as evidenced by crown and stem form, diameter and previous growth rate (Schlesinger, 1978; McGee and Bivens, 1984). McGee and Bivens (1984) concluded that ‘… overtopped white oak trees are not good candidates for crop trees. Even if high-quality overtopped oaks are present, their performance following release will at best be variable and their potential to produce high-quality products questionable.’ These conclusions are consistent with the oak’s weak apical dominance (Zimmermann and Brown, 1971; Oliver and Larson, 1996) and its associated inability to regain apical dominance after prolonged suppression. This physiological characteristic, by itself, would seem to work against successful application of the single-tree selection method. Oaks in inferior crown classes also have a propensity to produce epicormic branches following release. Both characteristics are potentially problematic in sustaining tree quality, and thus in providing a sound economic basis for the method (Trimble and Seegrist, 1973; McGee and Bivens, 1984). Nevertheless, it could be argued that many of these negative effects might be avoided or minimized by managing oak stands at relatively low densities and concentrating periodic removals on trees of low vigour within each diameter class. The long-term outcome of such practices nevertheless remains undetermined. Application of the single-tree selection method to a large forest property that covers a wide range of sites from xeric to mesic also may not be uniformly sustainable because of associated variation in oak regeneration potential. This may require mixing even-aged with unevenaged silvicultural methods to accommodate site-related variation in oak regeneration potential.

360

Chapter 8

Field implementation The guidelines outlined below assume the objective is to sustain oaks as an important stand component using the single-tree selection method. Where the method is deemed inappropriate and oaks are the desired species, we recommend other silvicultural methods such as shelterwood (Chapter 7) or group selection methods discussed later in this chapter. In some cases, the silviculturist may choose to apply the single-tree selection method and to accept implicitly the eventual successional displacement of an existing oak component by other species. However, the following guidelines assume that the objective is to perpetuate oak-dominated stands in ecosystems showing some propensity to sustain oaks in a negative exponential or similar diameter distribution. PRELIMINARY INVENTORY A preliminary stand inventory is requisite to: (i) deciding whether the single-tree selection method is

ecologically appropriate; (ii) describing the current diameter distribution of the stand; and (iii) selecting an appropriate guiding diameter distribution curve. The overstorey of each stand should be inventoried to provide a quantitative description of its size structure, density and species composition. Inventories should record diameters by species and include trees down to the 2 inch dbh class. Including small-diameter trees is important in assessing the sustainability of oaks with the single-tree selection method. Diameter frequency distributions should then be prepared for the stand as a whole and for individual species or species groups. For stands or species groups that approximate a negative exponential distribution, a preliminary estimate of q can be obtained by determining the proportion of basal area in the large (e.g. sawtimber) size classes (Fig. 8.16). Figure 8.16 or similarly derived curves together with point sample estimates of basal area (Husch et al., 1982) can facilitate a quick approximation of q.

1.5

1.4

q

1.3

1.2

1.1 15

25 35 45 55 65 Per cent of basal area in sawtimber

75

Fig. 8.16. Relation between q (for 1 inch dbh classes) and percentage of stand basal area in sawtimber (trees ≥11 in dbh) for diameter distributions that conform exactly to the negative exponential distribution. The closer the observed distribution is to the negative exponential, the more accurate is the estimate of q derived from the curve. The curve shown is based on trees 2 inches dbh and larger. The relation facilitates quick preliminary estimates of q from point samples of basal area in stands with negative exponential or similar diameter distributions. Values of q for other dbh class-widths in inches (w) are given by qw.

Uneven-aged Silvicultural Methods

Advance reproduction also should be inventoried. If a regional regeneration guide appropriate to the stand is available (see Chapter 7), it may be necessary to follow its recommended inventory procedure to obtain an adequate evaluation of the quantity of advance reproduction. Even though such guides are usually designed for even-aged applications, directly or indirectly they may provide useful information on the sustainability of oaks. Advance reproduction inventories should include an assessment of the number of trees in the 1 inch dbh class or other large reproduction in comparison to the number of trees speciﬁed for the 2 inch dbh class by the guiding curve. One indication of adequate regeneration potential would occur when the product of q (from the guiding curve) and the density of trees in the 2 inch dbh class approximates the density of reproduction in the 1 inch dbh class. INTERPRETING THE OVERSTOREY INVENTORY For a given species, a diameter frequency distribution that forms a bell-shaped curve indicates that there has not been sustained recruitment of that species into the overstorey. The relative abundance of oaks in the smaller diameter classes thus provides one indicator of the potential sustainability of oaks under the single-tree selection method. This is particularly true if white oak, one of the more shade tolerant oaks in eastern forests, is a major stand component. If its diameter distribution is more-or-less bell-shaped and lacking in small trees, it is unlikely that the oaks will be sustainable as an uneven-aged population. The presence of large numbers of shade tolerant nonoaks, especially maples, that form reverse Jshaped diameter distributions spanning a range of diameters from small to large is good evidence that the successional displacement of oaks is already underway. Wherever oaks share canopy dominance with shade tolerant competitors such as sugar maple or fast-growing intolerant competitors such as yellow-poplar, the singletree selection method is unlikely to sustain the oaks. However, poor representation of overstorey maples in the smaller diameter

361

classes, even when maple reproduction is present, may indicate a limited potential for maple to displace the species that currently dominate the site. Although site index is not always a good indicator of oak regeneration potential, in general it is difﬁcult to regenerate oaks in uneven-aged stands where oak site index exceeds 65 (Weitzman and Trimble, 1957; Trimble, 1973). Local or regional ecological classiﬁcation systems sometimes provide additional information on the relative persistence of oaks within deﬁned ecosystem types (see Chapter 1). Generalized interpretations of diameter distributions as discussed above often may be complicated by variation in stand density, disturbance history, site quality, species composition and other factors (Lorimer and Krug, 1983). Variation in any of these factors may produce conditions that are misleading with respect to the sustainability of oaks. Even in ecosystems that are intrinsic accumulators of oak reproduction, time since last disturbance may affect diameter distributions. We nevertheless propose that wherever numerous smalldiameter oaks are growing together with large-diameter oaks at moderate to high relative stand densities, it is likely that oaks can sustain a negative exponential diameter distribution. A stand’s current age and size structure are often valuable indicators of the relative difﬁculty or ease of silviculturally creating and sustaining a balanced uneven-aged structure. INTERPRETING

THE

ADVANCE

REPRODUCTION

An inventory of the advance reproduction of a stand can provide insight into its replacement potential, and how that potential varies spatially within the stand. Local or regional regeneration guides such as those described in Chapter 7 may be helpful. Local ecological classiﬁcation and site evaluation guides also may be helpful in assessing a stand’s regeneration and successional potential (e.g. Gysel and Arend, 1953; Smalley, 1978, 1979, 1982, 1984, 1986, 1988; Allen, 1987, 1990; Hix, 1988; Kotar et al., 1988; Jones, 1991; Cleland et al., 1993; Bakken and Cook, 1998; Van Kley et al., undated). INVENTORY

362

Chapter 8

The presence of several hundred large (e.g. ≥2 ft tall) oak seedling sprouts per acre, which commonly occurs in the drier oak–hickory forests in much of the eastern United States, might be a good indicator that a stand is an intrinsic accumulator of oak reproduction and thus potentially sustainable under single-tree selection. An abundance of small oak seedlings, by itself, does not indicate that a stand is necessarily an intrinsic accumulator of oak reproduction (Chapter 3). In some oak stands, abundant maple reproduction may be a good indicator of the maple’s potential to successionally displace oaks. However, this is not always true, especially on xeric or xero-mesic sites. In mesic and hydric sites the presence of maple or other shade tolerant reproduction and aggressive intolerant species is usually a clear indication of the successional potential of other species to displace the oaks. Distinguishing which process prevails may be evident from observing nearby stands on similar sites. Where a local or regional ecological classiﬁcation system is available, the prevailing process may be apparent from observing stands within the same ecological class. Moreover, the ecological classiﬁcation system itself may specify the successional status and thus relative persistence of oaks in a deﬁned ecological unit. Old-growth stands also can provide useful benchmarks for assessing the successional status of oaks on similar sites. The presence of non-oak reproduction with site- or life-historyrestricted growth potentials such as red and sugar maples, blackgum, sassafras, sourwood and ﬂowering dogwood may not, by itself, restrict oak’s regeneration potential. Their potential to do so is often speciﬁed by or can be inferred from local or regional ecological site classiﬁcation systems. Personal silvicultural experience in the same or similar ecosystems sometimes may provide the best basis for decisions. Often there is no absolute assurance of the desired outcome, only reasonable and informed judgements. If the current structure and composition of the overstorey and

reproduction is judged on the best available criteria to be suitable for single-tree selection, the next step is to deﬁne an appropriate guiding diameter distribution curve. SELECTING A GUIDING DIAMETER DISTRIBUTION

The guiding curve applied to regulate a stand’s structure should be consistent with its ecological conditions and stand dynamics as earlier discussed. It also should include diameters down to about 2 inches dbh. Managing stands at relatively low q values (1.1–1.3 for 1 inch dbh classes) and periodically reducing residual stand densities to about 50 ft2 of basal area per acre is generally consistent with oak regeneration dynamics. For a given stand density, low q values allocate proportionately more growing space to the larger and more valuable crop trees than do higher q values (Fig. 8.7B). However, in mesophytic and hydric stands it may be difﬁcult to maintain low q values. There, reproduction of shade tolerant or aggressive intolerant non-oaks is likely to capture much of the growing space made available after each reduction in stand density. This, in turn, may steepen the diameter distribution curve and increase q. Low residual stand density also is likely to exacerbate this effect. Given those conditions, the best course may be to abandon pretensions of sustaining oaks with the single-tree selection method. Other techniques such as the shelterwood (Loftis, 1990) or group selection (Murphy et al., 1993; Miller et al., 1995) methods may be better options. Selecting a guiding curve also requires selecting the diameter of the largest tree to be retained after a harvest cut. The value selected should be based on site quality, species’ longevity, local timber markets and ﬁnancial maturity of trees. Other factors being equal, trees should be grown to larger diameters on good sites than on poor sites. However, choices are sometimes limited by species characteristics. For example, scarlet oak is vulnerable to dieback and other causes of early mortality, regardless of site quality (Starkey and Oak, 1989; Oak et al., 1996). The minimum diameter for the guiding curve also must be speciﬁed because CURVE

Uneven-aged Silvicultural Methods

the diameter range and q must be set to obtain a guiding curve that conforms to both the desired q and residual stand density (e.g. by application of Equation 8.8). A guiding curve also must specify the diameter class width (w in Equation 8.8) to which the curve is applicable. Appropriately applied, Equation 8.8 can transform implementation of the single-tree selection method from an applied art form to a more objective science. It also can be coupled with mathematical growth models, linear programming, goal programming, and other management tools to optimize cutting cycle length, cutting schedules for converting irregular stands to the target structure, species mixes and stocking for value growth (Adams and Ek, 1974; Adams, 1976; Hann and Bare, 1979; Buongiorno et al., 1995). However, in their application to uneven-aged oak forests, these management techniques should be constrained by ecological realities. There are few reasons to believe there is great ﬂexibility in using the single-tree selection method to manage oaks. Where the objective is to sustain oaks, the method is constrained by the need to:

363

Graphical representation of the guiding curve (e.g. Figs 8.5 and 8.6) is unlikely to provide a practical ﬁeld guide for marking trees to be harvested. It is usually more efﬁcient to mark trees based on the desired residual basal area of a few broad size classes (e.g. saplings, poles and two sawtimber classes). In the ﬁeld, a running tally of basal area then can be efﬁciently obtained by size classes using a prism or other point sampling device. Marking to the correct residual density per acre requires knowing for each size class the basal area that conforms to the guiding curve as well as the actual basal area present. After the guiding curve has been speciﬁed, the corresponding residual stand basal area can be calculated for each diameter class of interest. Of the total stand basal area represented by the guiding curve, the proportion to be retained in each size class will depend on q and the range of diameters considered. For any guiding curve based on the negative exponential distribution and a given q, the per cent of total stand basal area to be retained in a given dbh class will be the same, regardless of the selected residual density (Table 8.2). Combined with stand inventories, these percentages can be used to specify the actual residual and harvest basal areas per acre for any dbh class. To ensure that the goals of the guiding curve are met, frequent checks on residual stand basal area should be made during the course of timber marking. Trees should be marked for removal over the full range of merchantable size classes and concentrated in size classes with surplus trees. Although the general silvicultural rule is to remove MARKING GUIDELINES

• Limit the method’s application to ecosystems that are intrinsic accumulators of oak reproduction. • Maintain relatively low residual stand densities by reducing total stand density to about 50 ft2 of basal area per acre as frequently as practical (e.g. 15–20 year cutting cycles). • Use a guiding curve with a q value of 1.2–1.3 based on 1 inch dbh classes (Larsen et al., 1999).

Table 8.2. The percentage distribution of basal area by tree-size classes for selected q values.a

q value Tree-size class (dbh, inches)

1.1

1.2

1.3

1.4

1.5

1.7

37 43 17 3

54 37 8 1

Per cent of stand basal area Saplings (≤5) Poles (6–10) Small sawtimber (11–16) Large sawtimber (17+)

5 21 42 32

11 30 39 20

19 38 32 11

28 42 24 6

aPercentages are applicable to the dbh range of 2–20 inches and all stand densities; q values are for 1 inch dbh classes.

364

Chapter 8

trees across the entire diameter range that are in excess of the guiding curve (Fig. 8.6), this may not always be necessary. The need to do so may vary among ecosystems and the stand dynamic unique to each. A good rule none the less is to remove trees at least across the range of diameters from 10 inches and larger, and within this range to leave the number of trees prescribed by an appropriate guiding curve. Removal of some undesirable pole-size trees will improve the survival of higher value residual trees and also concentrate growth on those stems (Smith and Miller, 1987). Within a dbh class, removals should be concentrated on trees of low vigour as evidenced by poor crown and stem form. Spacing and species composition should also be considered. The general rule is ‘cut the worst and leave the best’. In previously unmanaged stands with irregular diameter distributions, it may be necessary to work gradually towards the desired diameter distribution over several cutting cycles. To maintain adequate overall stocking, it is often necessary to compensate for an inadequate number of trees in one dbh class by retaining stocking in excess of the guiding curve in another class. To minimize epicormic branching, it is generally recommended that no more than one-third of total stand basal area be removed in one cut (Sonderman, 1985; Law and Lorimer, 1989). The cutting cycle is the interval between harvests. The length of a cycle will depend on stand growth rate and value, and local wood utilization and logging norms. The time between harvests should allow for accumulation of sufﬁcient volume for an economically operable harvest. In hardwood forests of the eastern United States, economically practical cutting cycles range from about 10 to 20 years. On average upland oak sites in the Central Hardwood Region, stand stocking per cent based on Gingrich’s (1967) stocking equation increases at approximately 1.3% per year on average sites. On those sites, stands periodically reduced to 50% stocking can be expected to increase to about 70% in 15 years. Due to the oak’s relative intolerance

LENGTH OF CUTTING CYCLE

of shade, it is desirable to use short cutting cycles and minimize the length of time stands are at high basal areas and high stocking percentages (e.g. ≥80% stocking). A shorter cutting cycle also provides better opportunities to salvage mortality and losses in bole quality due to logging damage (Smith and Miller, 1987). Consequently, cutting cycles should be as short as is economically feasible. METHOD VARIANTS Variants of the single-tree selection method largely centre around various schemes for harvesting to speciﬁed diameter limits (e.g. Smith, 1980; Smith and Miller, 1987; Miller and Smith, 1993). Although such harvesting is not true singletree selection, the results sometimes approximate cutting to a speciﬁed stand structure. As commonly applied, diameter-limit cutting usually refers to cutting all trees of sawlog size (often 11 inches dbh and larger) at each harvest. Such a method is not likely to be sustainable. However, diameter-limit harvests can be set to higher limits so that a proportion of the sawlog component is always retained. But whether diameter limits are set high or low, there is relatively little control over the density, structure, composition and quality of the residual stand. A variant of the diameter-limit approach, called ‘ﬂexible diameter limit cutting’ has been tested in mixed mesophytic forests containing oaks in the central Appalachians (Miller and Smith, 1993). The method produced results similar to those expected using single-tree selection. Large trees with the potential to increase in value are retained whereas other merchantable trees are cut to improve stand quality. Residual stand density guidelines are used to adjust diameter limits before trees are marked to prevent overcutting and thus sustain yields. Accordingly, merchantable trees are harvested that are of low quality, high risk undesirable species or are not earning an acceptable rate of return. Because rates of return decline for large sawtimber trees, harvesting this largest component of the desirable growing stock is ﬂexibly focused

Uneven-aged Silvicultural Methods

on trees 18 inches dbh and larger. The minimum cutting diameter for those trees varies by species, site quality, the desired rate of return and potential for future improvement in log grade. Each species has its own diameter limit based on potential growth and rate of return. Harvest of other merchantable trees provides for continual improvement in overall stand quality. In effect, all size classes within the range of merchantable diameters are managed on the basis of potential rate of return. Using this method, residual stand structure is left somewhat to chance. After 20 years, the method nevertheless produced a stand structure similar to that which would be expected using the single-tree selection method. MONITORING Periodic inventories of stands or entire forests are necessary for monitoring the effectiveness of the single-tree selection method in meeting objectives for stand size structure, density and species composition. If objectives are not being met, adjustments can be made to the guiding curve or other silvicultural strategies. Determining the feasibility of implementing and continuing the application of the single-tree selection method depends on continually gathering and interpreting information on stand or forest conditions. Such information is essential to making informed silvicultural decisions. It is useful to develop decision-making guidelines applicable to deﬁned ecoregions, and stand and site conditions. An example is provided by guidelines for selecting and managing upland oak stands for single-tree selection in the Ozark Highlands of Missouri (Table 8.3). The guidelines, based on pre-existing stand structure, density and composition, include recommendations for applying the method where appropriate, and refer the user to information on alternative methods where its application is not appropriate. Such guidelines can be detailed or generalized, tailored to meet various management objectives, and modiﬁed as experience and knowledge in their application accumulate.

365

Group Selection Method The group selection method is a modiﬁcation of the single-tree selection method where openings larger than the size of the largest individual trees are made in the forest canopy. Typical opening sizes range from 0.2 to 0.5 acre. The group selection method is a regeneration method usually aimed at obtaining reproduction of intolerant and mid-tolerant species. The method reduces the negative visual impacts of harvesting trees over large areas associated with most even-aged methods. It nevertheless has much in common with the clearcutting and shelterwood methods, but at a smaller spatial scale. Like those methods, consistent success in regenerating oaks depends on obtaining adequate advance reproduction before group openings are created. In effect, the method is a melding of even-aged and uneven-aged systems (Fig. 8.17). In practice, some group openings are unavoidably created in the normal application of the single-tree selection method as a result of irregular mortality, associated salvage cutting and the occurrence of other unplanned irregularities in the forest canopy. Compared to the single-tree selection method the distinguishing feature of the group selection method is the planned creation of larger canopy openings. Despite little documented experience with the method, it has generated considerable interest among hardwood silviculturists. Whereas some have rejected the method (Nelson et al., 1973; Roach, 1974) others have been more optimistic about its application (Minckler and Woerheide, 1965; Clark and Watt, 1971; Rudolph and Lemmien, 1976; Leak and Filip, 1977; Heald and Haight, 1979; Marquis and Johnson, 1989; Minckler, 1989; Jacobs and Wray, 1992; Murphy et al., 1993; Miller et al., 1995). Rejection has centred largely on perceived difﬁculties in yield regulation (Roach, 1974) rather than on regeneration or other silvicultural limitations. Documented applications to oak forests are limited, and restricted to inconclusive or undemonstrated long-term results (e.g. Rudolph and Lemmien, 1976; Heald and Haight, 1979; Hill and Dickmann, 1988; Guldin and Parks, 1989; Golden, 1995).

366

Chapter 8

Table 8.3. Silvicultural decision table and guidelines for applying uneven-aged silviculture to upland oak stands in the Ozark Highlands based on existing stand structure, density and composition.a Stand structure Stand density and composition Adequate

Inadequate

aThe

Adequate

Inadequate

Stand condition I: Stand structure, density and composition are all adequate for applying the single-tree selection method. Oaks form a relatively balanced reverse J-shaped dbh distribution for trees 2 inches dbh and larger. Basal area of oaks ≥2 inches dbh is ≥45 ft2 acre−1.b Subcanopy oaks (2–4 inches dbh) number at least 125 per acre

Stand condition II: Stand density and composition are adequate; stand structure is marginally inadequate. Oaks form a relatively balanced reverse J-shaped dbh distribution for trees ≥6 inches dbh, but not in the smaller dbh classes. Subcanopy oaks (2–4 inches dbh) number >70 but <125 per acre. Basal area of oaks ≥2 inches dbh is ≥45 ft2 acre−1b

Recommended action: Mark the oak component of the merchantable stand to a guiding curve of 1.7c based on a total oak density of 45 ft2 acre−1b (oaks ≥2 inches dbh). If current total stand basal area is ≥85 ft2 acre−1, reduce total stand density to only 60 ft2 acre−1 in the first harvest cut. Across all merchantable diameters, cut the worst trees and leave the best (after considering biodiversity goals)

Recommended action: Same as Stand Condition I, except mark the oak component of the merchantable stand to a q of 1.5.c Gradually increase q to 1.7c in succeeding harvest cuts as stocking of the oak sub-canopy increases

Stand condition III: Overall stand structure (oaks + non-oaks) is adequate, but composition and density are inadequate. Stands form a reverse J-shaped dbh distribution including oaks ≥6 inches dbh. However, the sub-canopy is dominated by non-oak hardwoods. An oak subcanopy (2–4 inches dbh) is nearly or completely absent and/ or total oak basal area (oaks ≥2 inches dbh) is <45 ft2 acre−1b

Stand condition IV: Stand structure and density are inadequate. Total oak basal area (oaks ≥2 inches dbh) is <45 ft2 acre−1b. Stands are often characterized by bell-shaped dbh distributions of black and/or scarlet oaks averaging ≥8 inches dbh. Stocking of sub-canopy oaks is inadequate (i.e. oaks in 2–4 inches dbh classes number <70 per acre)

Recommended action: (1) Apply the group selection method to create conditions favourable to regenerating oaks and developing the desired uneven-aged stand structure and composition. Evaluate the various group selection options (see Table 8.6). (2) Alternatively, apply even-aged silviculture (clearcut or shelterwood methods)

Recommended action: Same as stand condition III

table assumes uneven-aged silviculture is the preferred option and that a wide range of tree sizes from sapling to sawtimber classes are present. bOn average, commercially valuable non-oaks will comprise an additional 5 ft2 of basal area. If non-oak basal area is <3 ft2 per acre or absent, increase the minimum required oak stocking to 50 ft2 of basal area per acre. cBased on 2 inch dbh classes.

Uneven-aged Silvicultural Methods

367

Fig. 8.17. Group openings can provide adequate light for the development of oak and other shade intolerant species that often fail to regenerate under single-tree selection. Because of shading and root competition from the adjacent stand, reproduction grows most rapidly near the centre of the opening.

The method is potentially applicable to oak forests because it can provide the light necessary for oak regeneration (Fig. 8.18). Opening size controls light within the opening, and much attention has been given to deﬁning a minimum opening size suitable for regenerating oaks. Minima for regenerating oaks in the Central Hardwood Region have been deﬁned based on a computer simulation model of the solar radiation received on the forest ﬂoor (Fischer, 1979, 1981). The model expresses solar radiation as a percentage of that received in the open, which varies with surrounding tree height, slope aspect and slope gradient. The estimated opening sizes assume that the minimum light requirement for oaks is one-third of full light. This value is based on the observed minimum light intensity for maximum rates of photosynthesis of northern red oak seedlings (Phares, 1971). Because light in small openings is directly related to the height of the surrounding trees, it is sometimes convenient to express minimum opening size in terms of those heights. Estimates of minimum opening sizes needed to support oak reproduction range from one to two tree heights, depending on

aspect and slope (Table 8.4). Minimum opening size for Central Hardwood forests accordingly would range from about 0.1 to 0.6 acre for tree heights from 60 to 90 ft, respectively, across a wide range of aspects and slopes (Table 8.5). In bottomland oak forests in the South, where trees can attain heights of 125 ft or more, minimum opening size may exceed 1.6 acres. The amount of light initially received in an opening decreases with time because of crown expansion of the residual stand around the opening. Small openings therefore close more rapidly than large openings (Table 3.3). Consequently, crown closure rates probably set the lower limits of practical opening size for oaks to about two tree heights based on the one-third full-light criterion. Group openings of two tree heights also are consistent with earlier subjective assessments of minimum opening size (Trimble and Tryon, 1966; Smith, 1986; Miller et al., 1995), although minima may vary with aspect and slope steepness (Fischer, 1979). Although these calculated values provide an objective basis for deﬁning minimum opening size, light near the forest ﬂoor (i.e. at ‘reproduction height’) may be

368

Chapter 8

Fig. 8.18. The interior of a one-eighth of an acre group opening (approximately one tree height in diameter) in an oak stand in southeastern Ohio 5 years after cutting. Despite the small size of the opening, there is abundant tree reproduction including oaks. The shading effect of the adjacent stand is evident from the short trees near the opening edge (right foreground). (USDA Forest Service, North Central Research Station photograph.)

controlled as much or more by low vegetation as by canopy density (Lorimer et al., 1994). Moreover, herbaceous and shrub species together with competing non-oak tree reproduction may respond more rapidly to canopy openings than the existing oak reproduction. Like the clearcutting method, adequate oak advance reproduction is essential to the successful regeneration of group openings. Oak reproduction therefore must be present in adequate numbers, size and spatial distribution (Sander et al., 1984; Dey, 1991). Only then are oaks likely to be consistently competitive with other established vegetation after the opening is created. Although there is evidence that oak reproduction established after openings are made can contribute to future stand stocking in some situations (Johnson et al., 1989; Jacobs and Wray, 1992; Golden, 1995; Loewenstein and Golden, 1995), it would be prudent to obtain the requisite advance reproduction before cutting. Whereas the lower limits of opening size can be objectively deﬁned, it is less clear what the upper limits should be. Ultimately, the mosaic of group openings determines

the minimum spatial scale at which the uneven-aged state will occur. Because uneven-aged stands include at least three age classes, a complete uneven-aged unit must be at least as large as the average area that will include at least three age classes of trees in the future. For example, contiguous one-third of an acre openings will create approximately 1 acre uneven-aged units, whereas half an acre openings will create 1.5 acre units. As opening size increases, the canopy mosaic becomes increasingly ‘coarse grained’. At some upper size, the method will fail to provide the visual appearance of a continuous, all-aged forest canopy. Openings of half an acre or less generally provide visual effects similar to that of the single-tree selection method (Marquis and Johnson, 1989). Visual impacts nevertheless are only one of many effects determined by opening size. Silviculturists have traditionally focused on the effects of opening size on reproduction and the development of trees bordering the opening. Small openings create a large ratio of perimeter to opening area. For example, ten one-third of an acre circular

Uneven-aged Silvicultural Methods

369

Table 8.4. Effects of selected aspects, slope gradients and opening sizes on percentage of direct solar radiation received on the forest floor in square openings.a Underlined values approximate or exceed minimum light requirements for maximum photosynthesis of northern red oak. Aspect (azimuth)

Slope gradient

Opening size in tree heights 0.5

1.0

0 0 0 180 180

30 15 0 15 30

2.0

4.0

Percentage direct solar radiationb 15 34 19 47 24 56 30 66 37 66

Degrees 1 4 5 5 12

57 74 86 89 87

aFrom

Fischer (1981). on determining whether a given point on the forest floor within a forest opening is in the shade or sun. When a point is in the sun, the instantaneous direct beam solar radiation at that time is calculated and assumed for a given time interval. Direct solar radiation is then summed for the total time period considered and expressed as a percentage of that received on an open horizontal surface of equivalent area. Calculations are based on solar radiation received on clear, sunny days during the growing season (12 April–30 August) at latitude 40° 14’ 00”N.

bBased

Table 8.5. Area of circular openings in acres in relation to perimeter tree height. Opening size in tree heights (number)

Mean height of dominant and co-dominant perimeter trees (feet) 60

70

80

90

0.07 0.26 0.58 1.04

0.09 0.35 0.80 1.41

0.12 0.46 1.04 1.85

0.05 0.58 1.31 2.34

100

110

120

130

0.18 0.72 1.62 2.89

0.22 0.87 1.96 3.49

0.26 1.04 2.34 4.15

0.31 1.22 2.74 4.88

Acres 1 2 3 4

openings (3.3 acres) create 3.2 times more perimeter (edge) than a single 3.3 acre circular clearcut. Oaks bordering openings are prone to developing epicormic branches when exposed to the high light intensities occurring along edges, which in turn reduces their bole quality and value (Trimble and Seegrist, 1973). At the same time, shade and competition from border trees suppresses reproduction near edges. Other possible edge effects include variation in species composition within openings that is related to seed dispersal distances, browsing of reproduction (Marquis, 1974; Marquis et al., 1976), damage to reproduction by felling of surrounding trees, damage to the surrounding forest by logging within groups and associated pathogenic effects such as oak decline (Oak et al., 1996; Starkey et al.,

1989), and effects on breeding birds (Thompson et al., 1995, 1996; Annand and Thompson, 1997; Thompson and Dessecker, 1997). Collectively, such effects are likely to be expressed at different scales and across different ‘inﬂuence zones’ (Fig. 8.19). There is therefore not a single edge effect, but many such effects. Factors other than regeneration and yield regulation thus may determine opening size or even the feasibility or appropriateness of the group selection method. Assuming the method meets management objectives, it can be applied in various ways. One method is to follow procedures for the single-tree selection method with the added proviso of creating openings in each stand at each entry (Law and Lorimer, 1989). No attempt is made to

370

Chapter 8

1 2 3 4 5 6 Fig. 8.19. Potential influence zones associated with edge effects related to: 1. Epicormic branching of trees in the residual stand. 2. Growth of tree reproduction. 3. Browsing of reproduction by deer. 4. Postharvest dispersal of acorns by rodents. 5. Damage to tree reproduction by felling surrounding trees; postharvest dispersal of acorns by blue jays. 6. Impact of timber harvesting in opening on oak decline in surrounding forest; nest predation and brood parasitism of birds. (Adapted from Bradshaw, 1992, with permission from Elsevier Science.)

keep track of individual groups. It is assumed that groups eventually become spatially indistinguishable or ‘lost’ in the overall forest matrix. This expectation is consistent with the objective of maintaining a balanced stand structure at a relatively small spatial scale. The strategy requires marking the stand to obtain a balanced diameter distribution in all but the group openings. Roach (1974) asserted that maintaining this balance under group selection will be difﬁcult at best. He argued that it will become increasingly difﬁcult with successive cutting cycles to ﬁt additional group openings into a stand. Unless the number and size of group openings are relatively small, a visibly apparent group-opening matrix may persist and eventually characterize the stand. According to Roach, this would eventually create practical problems in marking stands. As trees originating in groups become more prominent with each successive cutting cycle, the marking guides initially used between groups to maintain a reverse-J diameter distribution would over time increasingly shift to

accommodate marking to some other criterion within the even-aged groups. Roach’s concern thus implies that groups remain identiﬁable, at least through several cutting cycles. The resulting confusion, he argued, would not be conducive to the practical considerations of marking stands to maintain a balanced stand structure. Roach further pointed out that there would not necessarily be an equal area of group-opening age classes by the time a stand would otherwise have attained full regulation. This, in turn, would further exacerbate attaining a balanced stand structure and thus result in poor yield regulation. The extent to which these problems materialize may depend on the number and size of group openings created at each harvest. Where it is deemed important to maintain a large representation of intolerant and mid-tolerant species using unevenaged silviculture, the number and/or size of openings may need to be maximal. If only a minor representation of these species is acceptable, then the number of group openings created at each harvest can be minimal and thus less problematic in

Uneven-aged Silvicultural Methods

the stand-wide maintenance of a relatively homogeneous uneven-aged state. Both objectives assume the group selection method is effective in sustaining the desired species composition. Results of applying the method for 38 years in a northern hardwood forest in New Hampshire resulted in a balanced reverse-J diameter distribution after three cutting cycles (Leak and Filip, 1977). Intolerant and mid-tolerant species comprised 25–33% of trees from 4 to 12 inches dbh. The method has been used in the central Appalachians and Ohio valley where it effectively maintained a higher proportion of intolerant and mid-tolerant species than single-tree selection (Minckler and Woerheide, 1965; Schlesinger, 1976; Smith, 1980; Miller et al., 1995; Weigel and Parker, 1995). However, oaks were poorly represented in group openings except in the drier ecosystems. Alternatively, the group selection method can be applied using area regulation within individual stands or management units. As in even-aged silviculture, a rotation length must be established. In this method, group openings are created at each harvest entry according to a ﬁxed schedule set by the cutting cycle, stand area and the conversion period. Accordingly, the total area in group openings and their number must be calculated for each stand at each harvest entry. Area in openings (AO) can be calculated by: AO = CC(SA/R)

[8.9]

where CC = cutting cycle, SA = stand area, and R = rotation length (or conversion period) (Murphy et al., 1993). Number of openings per stand per harvest entry (N) is thus: N = AO/OS

[8.10]

where OS = opening size. The method thus mimics the clearcutting method but at a much smaller spatial scale. A fraction of each stand is regenerated by completely removing the overstorey in groups over the number of harvest entries equal to the conversion period divided by the cutting cycle. At the end of the conversion period,

371

the stand will be a fully regulated unevenaged stand comprised of N even-aged groups. Whether or not the groups are visually distinguishable at that time will depend on group opening size and age. Larger openings can be expected to maintain their identity longer than smaller openings, and for a given opening size, older groups will be more difﬁcult to identify than younger groups. This method may be especially appropriate in previously high-graded stands comprised of overstoreys with little or no acceptable growing stock. Each stand is systematically renewed by removing the overstorey over a given number of harvests equal to the rotation age divided by the cutting cycle. The method also may overcome some of the problems of regulation control as discussed above. How to treat areas between the most recently created openings at each harvest entry nevertheless remains problematic. There are at least three options: 1. Treat each group as an even-aged ministand and thin from below as recommended in even-aged silviculture. 2. Treat each group as part of an unevenaged stand and thin to a speciﬁed stand structure as previously described. 3. Apply ‘free thinning’ to each group whereby cutting favours desired trees of good form regardless of crown position. No attempt is made to create and maintain a pre-deﬁned stand structure. Option 1 implies that the identity of individual groups can be maintained. The method therefore may be most applicable to relatively large group openings (e.g. >0.5 acre). However, maintaining group identity may not be essential in the oldest age classes (e.g. >60 years). At the end of the conversion period, ﬁtting new groups into old groups might be accomplished by considering a combination of factors including species composition, presence or absence of adequate advance reproduction and the biological or economic maturity of the overstorey. The approach may have application in small privately owned forests (e.g. <100 acres) where the owner-manager is inti-

372

Chapter 8

mately acquainted with each silvicultural unit. The approach also may be compatible with planting all or some of the group openings to oaks or other desirable species following the guidelines for shelterwood planting as discussed in Chapter 7. Option 2 implies that each stand will form, with the possible exception of the more recently created openings, a relatively homogeneous uneven-aged unit. This option may be appropriate where the intent is to shift to single-tree selection at the end of the conversion period, and the primary reason for creating groups is to renew a decadent or otherwise undesirable stand. This strategy assumes that maintaining a desirable species mix is feasible using single-tree selection. In applying the method, creating small group openings (e.g. <0.5 acre) may facilitate the transition from group selection to single-tree selection. Option 3 implies that an uncertain future stand structure is acceptable. Uncertainty results from abandoning the goals of creating and maintaining either a balanced stand-wide collection of evenaged groups (option 1), or a balanced, sustainable and relatively homogeneous uneven-aged stand structure represented by option 2. The result is likely to produce an unbalanced, uneven-aged structure whose future state is at best questionable and at worst unpredictable. The method nevertheless may have application where sustaining pre-deﬁned stand structure, species composition and associated values (including regulated yield over time) are not be important. Examples may include riparian zones, forest management primarily for water quality (e.g. municipal watersheds), areas around campgrounds, and other recreational and scenic areas that lie outside yield-regulation zones. This option is likewise applicable to the single-tree selection method. In either case, the principle of ‘cut the worst and leave the best’ should be followed. Given that the goal of a sustainable, pre-deﬁned stand structure is set aside, potential problems in applying the group selection method may largely disappear (Roach, 1974). The trade-off is the

loss of assurance of sustaining a deﬁned structural state or the oaks themselves. The options described above also could incorporate retention of one or more uncut (reserve) trees in openings where openings are relatively large (e.g. >0.5 acre). There are thus many possible variants of the group selection method (Table 8.6). Additional variants of the method also can be designed to meet speciﬁc objectives. Although the success of these methods in oak forests remains to be veriﬁed, their potential value nevertheless is suggested by theoretical considerations, trials in other forest types and preliminary results in oak forests (Minckler and Woerheide, 1965; Clark and Watt, 1971; Rudolph and Lemmien, 1976; Schlesinger, 1976; Leak and Filip, 1977; Fischer and Merritt, 1978; Heald and Haight, 1979; Smith, 1980; Hill and Dickmann, 1988; Toliver and Jackson, 1988; Guldin and Parks, 1989; Law and Lorimer, 1989; Jacobs and Wray, 1992; Murphy et al., 1993; Golden, 1995; Miller et al., 1995; Weigel and Parker, 1995).

Economic, Environmental and Social Considerations Selection silviculture is the least economically efﬁcient of all the silvicultural systems. This largely results from the relatively small amount of timber removed per acre per harvest. Timber marking, administrative and road construction costs per unit of volume removed per harvest are also high. Although group selection may reduce unit costs somewhat, costs nevertheless remain relatively high in comparison to even-aged methods (Shaffer et al., 1993). Moreover, logging damage to the residual stand sometimes may reduce the value of future crop trees (Dwyer et al., 2000). Oak mortality related to oak decline also may be accelerated by selection silviculture, especially where species in the red oak group predominate (Starkey and Oak, 1989). In the group selection method, reduced value of oak logs is associated with high ratios of edge to opening area and resulting epicormic branching and

Uneven-aged Silvicultural Methods

373

Table 8.6. Some variants of the group selection method and their application. Type Structure control

Area control

Option

Description of method

1

Group selection with stand structure control

2

Groups with reserves

All options listed

Group selection with area control

1

Groups maintained as identifiable even-aged units and thinned from below

2

Groups not maintained as identifiable units; balanced uneven-aged structure maintained stand-wide

3

Groups not maintained as identifiable units; free thinning applied standwide

4

Groups with reserves

Objectives and method of implementing Primary objectives: To provide light for regeneration of mid-tolerant and intolerant species while maintaining a relatively continuous and homogeneous canopy cover; to sustain adequate representation of smallest dbh classes. Method: Small group openings (≤0.5 acre) comprising 10% or less of stand area are created at each harvest cut. Requires stand-wide control of stand structure that ultimately produces a balanced reverse J-shaped dbh distribution. To regenerate oaks, requires advance reproduction of oaks in group openings Primary objectives and methods: Same as above except some trees within groups are retained for goals other than regeneration Primary objectives: To provide light for regeneration of mid-tolerant and intolerant species. The selected variant of the method (see below) depends on other objectives. Method: Area control with or without stand structure control. Area in openings (AO) is calculated by: AO = CC(SA/R), where CC = cutting cycle, SA = stand area and R = rotation length (‘conversion period’). Number of openings per stand per harvest entry (N) is N = AO/OS where OS is opening size Primary objectives: To maintain group identity to preserve area regulation within stands. May also be compatible with preharvest underplanting in group openings. Method: Create large groups (>0.5 acre). Thin each group from below as in even-aged silviculture and maintain group identity through most or all of the rotation. May be compatible with planting oaks as a supplement to natural reproduction as in the shelterwood-underplanting method (Chapter 7) Primary objectives: To maintain a relatively homogeneous canopy cover by creating small groups and/or to facilitate the transition to single-tree selection after the stand has been regulated by area control. Applicable to stands with little or no acceptable stocking. Method: Create small groups (≤0.5 acre). Maintain a stand-wide reverse J-shaped dbh distribution Primary objectives: To maintain a relatively continuous canopy cover that does not necessarily form a reverse J-shaped dbh distribution. Applicable to small private ownerships, recreation areas, municipal watersheds and other areas where sustained yield of timber products is not required. Method: Create small group openings (≤0.5 acre). Apply stand-wide free thinning to favour trees of good form regardless of crown position. No attempt is made to create and maintain a pre-defined stand structure Primary objectives: Any of those listed for the last three options with the added provision that some trees in group openings are retained for goals other than regeneration

374

Chapter 8

associated bole degrade (Trimble and Smith, 1970; Trimble and Seegrist, 1973; Smith, 1980). Although earlier experiences in applying single-tree selection to oak forests were largely negative from a technical perspective (Gingrich, 1967; Roach, 1968; Sander, 1978), its apparently successful application to some types of oak forests indicates that it should no longer be universally dismissed as an option for the oaks (Guldin and Parks, 1989; Loewenstein et al., 1995; Loewenstein, 1996; Wang, 1997; Larsen et al., 1999). Where the method is ecologically appropriate, there appear to be few limitations to applying uneven-aged silviculture in one or more of its variants. Some of the economic disadvantages may even be offset by the relatively frequent economic returns provided, which may be an

important management incentive for some forest owners. However, the caveats that apply may, by themselves, be of sufﬁcient importance to preclude the method’s application to a speciﬁc stand or forest. Risks in applying uneven-aged silviculture to oak forests nevertheless are partially mitigated by the maintenance of a more publically acceptable forest condition than that obtained under even-aged silviculture (Hull and Buhyoff, 1986; Gobster, 1994; Herrick and Rudis, 1994). Where unevenaged silviculture is ecologically appropriate, it offers an opportunity to simulate late-successional forest dynamics, the maintenance of certain types of wildlife habitats and aesthetic values that are generally under-represented and decreasing in occurrence across the landscape as a whole (Guldin, 1996).

References Abrams, M.D. and Nowacki, G.J. (1992) Historical variation in ﬁre, oak recruitment, and post-logging accelerated succession in central Pennsylvania. Bulletin of the Torrey Botanical Club 119, 19–28. Adams, D.M. (1976) A note on the interdependence of stand structure and best stocking in a selection forest. Forest Science 22, 180–184. Adams, D.M. and Ek, A.R. (1974) Optimizing the management of uneven-aged forest stands. Canadian Journal of Forest Research 4, 274–287. Allen, B.H. (1987) Ecological type classiﬁcation for California: the Forest Service approach. USDA Forest Service General Technical Report PSW SW-98. Allen, B.H. (1990) Classiﬁcation of oak woodlands. Fremontia 18(3), 22–25. Annand, E.M. and Thompson, F.R. III. (1997) Forest bird response to regeneration practices in central hardwood forests. Journal of Wildlife Management 61, 159–171. Bakken, P.N. and Cook, J.E. (1998) Regeneration potential of six habitat types common to north-central Wisconsin. Northern Journal of Applied Forestry 15, 116–123. Bradshaw, F.J. (1992) Quantifying edge effect and patch size for multiple-use silviculture – a discussion paper. Forest Ecology and Management 48, 249–264. Buongiorno, J., Peyron, J., Houllier, F. and Bruciamacchie, M. (1995) Growth and management of mixed-species, uneven-aged forests in the French Jura: implications for economic returns and tree diversity. Forest Science 41, 397–429. Chisman, H.H. and Schumacher, F.X. (1940) On the tree–area ratio and certain of its applications. Journal of Forestry 38, 311–317. Cho, D.-S. and Boerner, R.E.J. (1991) Canopy disturbance patterns and regeneration of Quercus species in two Ohio old-growth forests. Vegetatio 93, 9–18. Clark, F.B. and Watt, R.F. (1971) Silvicultural methods for regenerating oaks. Proceedings of the Oak Symposium. USDA Forest Service Northeastern Forest Experiment Station, pp. 37–43. Cleland, D.T., Hart, J.B., Host, G.E., Pregitzer K.S. and Ramm, C.W. (1993) Ecological Classiﬁcation and Inventory System of the Huron–Manistee National Forests. USDA Forest Service, Washington, DC. Crow, T.R. (1992) Population dynamics and growth patterns for a cohort of northern red oak (Quercus rubra) seedlings. Oecologia 91, 192–200.

Uneven-aged Silvicultural Methods

375

de Liocourt, F. (1898) De l’amenagement des Sapinieres. Bulletin de la Societé forestière de FrancheComte et du Territorie de Belfort 4, 396–409, 645–647. Della-Bianca, L. and Beck, D.E. (1985) Selection management in southern Appalachian hardwoods. Southern Journal of Applied Forestry 9, 191–196. Dey, D.C. (1991) A comprehensive Ozark regenerator. PhD dissertation, University of Missouri, Columbia. Dey, D.C. and Parker, W.C. (1996) Regeneration of red oak (Quercus rubra L.) using shelterwood systems: ecophysiology, silviculture and management recommendations. Ontario Forest Research Institute Forest Research Information Paper 126. Dey, D.C., Johnson, P.S. and Garrett, H.E. (1996) Modeling the regeneration of oak stands in the Missouri Ozark Highlands. Canadian Journal of Forest Research 26, 573–583. Dwyer, J.P., Dey, D.C., Walter, W.D. and Jensen, R.G. (2000) Logging impact in uneven-aged stands of the Missouri Ozark Forest Ecosystem Project. Proceedings of 1999 Society American Foresters National Convention, pp. 210–222. Ernst, R.L. and Knapp, W.H. (1985) Forest stand density and stocking: concepts, terms, and the use of stocking guides. USDA Forest Service General Technical Report WO WO-44. Fischer, B.C. (1979) Managing light in the selection method. Proceedings 1979 J.S. Wright Forestry Conference. Purdue University of West Lafayette, Indiana, pp. 43–53. Fischer, B.C. (1981) Designing forest openings for the group selection method. USDA Forest Service General Technical Report SO SO-34, pp. 274–277. Fischer, B.C. and Merritt, C. (1978) SHADOS: A computer model to simulate light energy distribution in small forest openings. Proceedings Central Hardwood Forestry Conference II. Purdue University, West Lafayette, Indiana, pp. 302–319. Gingrich, S.F. (1967) Measuring and evaluating stocking and stand density in upland hardwood forests in the Central States. Forest Science 13, 38–53. Gobster, P.H. (1994) The aesthetic experience of sustainable forest ecosystems. USDA Forest Service General Technical Report RM RM-247, pp. 246–255. Goff, F.G. and West, D. (1975) Canopy–understory interaction effects on forest population structure. Forest Science 21, 98–108. Golden, M.S. (1995) Establishment of oak regeneration in group selection openings in a river ﬂoodplain forest. USDA Forest Service General Technical Report SRS SRS-1, pp. 413–418. Guldin, J.M. (1996) The role of uneven-aged silviculture in the context of ecosystem management. Western Journal of Applied Forestry 11, 4–12. Guldin, J.M. and Parks, T. (1989) Development of cherrybark oak in an uneven-aged stand in West Tennessee. USDA Forest Service General Technical Report SO SO-74, pp. 327–331. Gysel, L.W. and Arend, J.L. (1953) Oak sites in southern Michigan: their classiﬁcation and evaluation. Michigan State University Technical Bulletin 236. Hann, D.W. and Bare, B.B. (1979) Uneven-aged forest management: state of the art (or science?). USDA Forest Service General Technical Report INT INT-50. Hansen, G.D. and Nyland, R.D. (1987) Effects of diameter distribution on the growth of simulated uneven-aged sugar maple stands. Canadian Journal of Forest Research 17, 1–8. Heald, R.C. and Haight, R. (1979) A new approach to uneven-aged silviculture and management of mixed conifer–oak forests. California Agriculture (May), 20–22. Herrick, T.A. and Rudis, V.A. (1994) Scenery in the Quachita National Forest. USDA Forest Service General Technical Report SO SO-112, pp. 212–222. Hibbs, D.E. and Bentley, W.R. (1983) A management guide for oak in New England. University of Connecticut Cooperative Extension Bulletin 83–12. Hill, J.P. and Dickmann, D.I. (1988) A comparison of three methods for naturally reproducing oak in southern Michigan. Northern Journal of Applied Forestry 5, 113–117. Hix, D.M. (1988) Multifactor classiﬁcation and analysis of upland hardwood forest ecosystems of the Kickapoo River watershed, southwestern Wisconsin. Canadian Journal of Forest Research 18, 1405–1415. Hull, R.B., IV and Buhyoff, G.J. (1986) The scenic beauty temporal distribution method: an attempt to make scenic beauty assessments compatible with forest planning efforts. Forest Science 32, 271–286. Husch, B., Miller, C.I. and Beers, T.W. (1982) Forest Mensuration. John Wiley & Sons, New York.

376

Chapter 8

Jacobs, R.D. and Wray, R.D. (1992) Managing oak in the Driftless Area. Minnesota Extension Service Bulletin NR-BU NR-BU-5900-S. Jenkins, M.A. and Parker, G.R. (1998) Composition and diversity of woody vegetation in silvicultural openings of southern Indiana forests. Forest Ecology and Management 109, 57–74. Johnson, P.S. (1992) Oak overstory/reproduction relations in two xeric ecosystems in Michigan. Forest Ecology and Management 48, 233–248. Johnson, P.S., Jacobs, R.D., Martin, A.J. and Goder, E.D. (1989) Regenerating northern red oak: three successful case histories. Northern Journal of Applied Forestry 6, 174–178. Johnson, R.L. (1979) Adequate oak regeneration – a problem without a solution? Proceedings 7th Annual Hardwood Symposium of the Hardwood Research Council, pp. 59–65. Jones, S.M. (1991) Landscape ecosystem classiﬁcation for South Carolina. USDA Forest Service General Technical Report SE SE-68, pp. 59–68. Kotar, J., Kovach, J.A. and Locey, C.T. (1988) Field Guide to Forest Habitat Types of Northern Wisconsin. University of Wisconsin, Madison. Krajicek, J.E. (1967) Maximum use of minimum acres. Proceedings 9th Southern Forest Tree Improvement Conference, pp. 35–37. Krajicek, J.E., Brinkman, K.A. and Gingrich, S.F. (1961) Crown competition – a measure of density. Forest Science 7, 35–42. Larsen, D.R., Metzger, M.A. and Johnson, P.S. (1997) Oak regeneration and overstory density in the Missouri Ozarks. Canadian Journal of Forest Research 27, 869–875. Larsen, D.R., Loewenstein, E.F. and Johnson, P.S. (1999) Sustaining recruitment of oak reproduction in uneven-aged stands in the Ozark Highlands. USDA Forest Service General Technical Report NC NC-203. Law, J.R. and Lorimer, C.G. (1989) Managing uneven-aged stands. USDA Forest Service Central Hardwood Note 6.08. Leak, W.B. (1978) Stand structure. Uneven-aged Silviculture and Management in the United States. USDA Forest Service, Washington, DC, pp. 104–114. Leak, W.B. (1987) Characteristics of ﬁve climax stands in New Hampshire. USDA Forest Service Research Note NE NE-336. Leak, W.B. (1996) Long-term structural change in uneven-aged northern hardwoods. Forest Science 42, 160–165. Leak, W.B. and Filip, S.M. (1977) Thirty-eight years of group selection in New England northern hardwoods. Journal of Forestry 75(10), 641–643. Leak, W.B., Solomon, D.S. and DeBald, P.S. (1987) Silvicultural guide for northern hardwood types in the Northeast (revised). USDA Forest Service Research Paper NE NE-603. Liming, F.G. and Johnston, J.P. (1944) Reproduction in oak–hickory forest stands of the Missouri Ozarks. Journal of Forestry 42, 175–180. Loewenstein, E.F. (1996) An analysis of the size- and age-structure of a managed uneven-aged oak forest. PhD dissertation, University of Missouri, Columbia. Loewenstein, E.F. and Golden, M.S. (1995) Establishment of water oak is not dependent on advance reproduction. USDA Forest Service General Technical Report SRS SRS-1, pp. 443–446. Loewenstein, E.F., Garrett, H.E., Johnson, P.S. and Dwyer, J.P. (1995) Changes in a Missouri Ozark oak–hickory forest during 40 years of uneven-aged management. USDA Forest Service General Technical Report NE NE-197, pp. 159–164. Loewenstein, E.F., Johnson, P.S. and Garrett, H.E. (2000) Age and diameter structure of a managed uneven-aged oak forest. Canadian Journal of Forest Research 30, 1060–1070. Loftis, D.L. (1990) A shelterwood method for regenerating red oak in the southern Appalachians. Forest Science 36, 917–929. Lorimer, C.G. (1980) Age structure and disturbance history of a southern Appalachian virgin forest. Ecology 61, 1169–1184. Lorimer, C.G. (1981) Survival and growth of understory trees in oak forests of the Hudson Highlands, New York. Canadian Journal of Forest Research 11, 689–695. Lorimer, C.G. (1983) Eighty-year development of northern red oak after partial cutting in a mixedspecies Wisconsin forest. Forest Science 29, 371–383. Lorimer, C.G. (1984) Development of the red maple understory in northeastern oak forests. Forest Science 30, 13–22.

Uneven-aged Silvicultural Methods

377

Lorimer, C.G. and Frelich, L.E. (1984) A simulation of equilibrium diameter distributions of sugar maple (Acer saccharum). Bulletin of the Torrey Botanical Club 111, 193–199. Lorimer, C.G. and Krug, A.G. (1983) Diameter distributions in even-aged stands of shade-tolerant and midtolerant tree species. American Midland Naturalist 109, 331–345. Lorimer, C.G., Chapman, J.W. and Lambert, W.D. (1994) Tall understorey vegetation as a factor in the poor development of oak seedlings beneath mature stands. Journal of Ecology 82, 227–237. Marquis, D.A. (1974) The impact of deer browsing on Allegheny hardwood regeneration. USDA Forest Service Research Paper NE NE-308. Marquis, D.A. (1978) Application of uneven-aged silviculture and management on public and private lands. Uneven-aged Silviculture and Management in the United States. USDA Forest Service, Washington, DC, pp. 25–61. Marquis, D.A. and Johnson, R.L. (1989) Silviculture of eastern hardwoods. USDA Forest Service General Technical Report WO WO–55, pp. 9–15. Marquis, D.A., Eckert, P.L. and Roach, B.A. (1976) Acorn weevils, rodents, and deer all contribute to oak-regeneration difﬁculties in Pennsylvania. USDA Forest Service Research Paper NE NE-356. Marquis, D.A., Ernst, R.L. and Stout, S.L. (1992) Prescribing silvicultural treatments in hardwood stands of the Alleghenies (revised). USDA Forest Service General Technical Report NE NE-96. McGee, C.E. (1984) Heavy mortality and succession in a virgin mixed mesophytic forest. USDA Forest Service Research Paper SO SO-209. McGee, C.E. and Bivens, D.L. (1984) A billion overtopped white oak – assets or liabilities? Southern Journal of Applied Forestry 8, 216–220. McGill, D., Martin, J., Rogers, R. and Johnson, P.S. (1991) New stocking charts for northern red oak. University of Wisconsin Forestry Research Notes 277. Meyer, H.A. and Stevenson, D.D. (1943) The structure and growth of virgin beech–birch–maple–hemlock forests in northern Pennsylvania. Journal of Agricultural Research 67, 465–484. Meyer, H.A., Recknagel, A.B. and Stevenson, D.D. (1952) Forest Management. Ronald Press, New York. Miller, G.W. and Smith, H.C. (1993) A practical alternative to single-tree selection. Northern Journal of Applied Forestry 10, 32–38. Miller, G.W., Schuler, T.M. and Smith, H.C. (1995) Method for applying group selection in central Appalachian hardwoods. USDA Forest Service Research Paper NE NE-696. Minckler, L.S. (1989) Intensive group selection silviculture in central hardwoods. USDA Forest Service General Technical Report NC NC-132, pp. 35–39. Minckler, L.S. and Woerheide, J.D. (1965) Reproduction of hardwoods 10 years after cutting as affected by site and opening size. Journal of Forestry 63, 103–107. Moser, J.W. Jr (1976) Speciﬁcation of density for the inverse J-shaped diameter distribution. Forest Science 22, 177–180. Murphy, P.A., Shelton, M.G. and Graney, D.L. (1993) Group selection – problems and possibilities for the more shade-intolerant species. USDA Forest Service General Technical Report NC NC-161, pp. 229–247. Nelson, R.E., Young, R.A. and Gilmore, A.R. (1973) Twenty-two years of management of upland hardwoods in southern Illinois. University of Illinois Agriculture Experiment Station Forestry Research Report 73-3. Nowacki, G.J., Abrams, M.D. and Lorimer, C.G. (1990) Composition, structure, and historical development of northern red oak stands along an edaphic gradient in north-central Wisconsin. Forest Science 36, 276–292. Oak, S., Tainter, F., Williams, J. and Starkey, D. (1996) Oak decline risk rating for the southeastern United States. Annales des Sciences Forestieres 53, 721–730. Oliver, C.D. and Larson, B.C. (1996) Forest Stand Dynamics. John Wiley & Sons, New York. Phares, R.E. (1971) Fertilization tests with potted red oak seedlings. USDA Forest Service Research Note NC NC-114. Richards, R.H., Shiﬂey, S.R., Rebertus, A.J. and Chaplin S.J. (1995) Characteristics and dynamics of an upland Missouri old-growth forest. USDA Forest Service General Technical Report NE NE197, pp. 11–22. Roach, B.A. (1968) Is clear cutting good or bad? Keep Tennessee Green Journal 8, 4–5, 12–14. Roach, B.A. (1974) Selection cutting and group selection. State University of New York Applied Forestry Research Institute AFRI Miscellaneous Report 5.

378

Chapter 8

Roach, B.A. and Gingrich, S.F. (1967) Upland hardwoods can be grown efﬁciently. Pulp and Paper (April). Roach, B.A. and Gingrich, S.F. (1968) Even-aged silviculture for upland central hardwoods. USDA Forest Service Agriculture Handbook 355. Rogers, R. (1983) Spatial pattern and growth in a Missouri oak–hickory stand. PhD dissertation, University of Missouri, Columbia. Rudolph, V.J. and Lemmien, W.A. (1976) Silvicultural cuttings in an oak–hickory stand in Michigan: 21-year results. Proceedings of Central Hardwood Forestry Conference I. University of Southern Illinois, Carbondale, pp. 431–453. Sander, I.L. (1977) Manager’s handbook for oaks in the North Central States. USDA Forest Service General Technical Report NC NC-37. Sander, I.L. (1978) Silvicultural systems for the oak–hickory forest type. North America’s Forests: Gateway to Opportunity Society of America Forestry & Canadian Institute of Forestry, pp. 344–348. Sander, I.L., Johnson, P.S. and Rogers, R. (1984) Evaluating oak advance reproduction in the Missouri Ozarks. USDA Forest Service Research Paper NC NC-251. Schlesinger, R.C. (1976) 16 years of selection silviculture in upland hardwood stands. USDA Forest Service Research Paper NC NC-125. Schlesinger, R.C. (1978) Increased growth of released white oak poles continues through two decades. Journal of Forestry 76(11), 726–727. Schnur, G.L. (1937) Yield, stand, and volume tables for even-aged upland oak forests. USDA Technical Bulletin 560. Shaffer, R.M., Brummel, K.R., Reisinger, T.W. and Stokes, B.J. (1993) Impact of group selection silviculture on timber harvesting productivity and cost in Appalachian hardwood timber stands. Northern Journal of Applied Forestry 10, 170–174. Shiﬂey, S.R., Roovers, L.M. and Brookshire, B.L. (1995) Structural and compositional differences between old-growth and mature second-growth forests in the Missouri Ozarks. USDA Forest Service General Technical Report NE NE-197, pp. 23–36. Smalley, G.W. (1978) Classiﬁcation and evaluation of forest sites for the Interior Highlands. Proceedings of Central Hardwood Forest Conference II. Purdue University, West Lafayette, Indiana, 257 pp. Smalley, G.W. (1979) Classiﬁcation and evaluation of forest sites on the southern Cumberland Plateau. USDA Forest Service General Technical Report SO SO-23. Smalley, G.W. (1982) Classiﬁcation and evaluation of forest sites on the Mid-Cumberland Plateau. USDA Forest Service General Technical Report SO SO-38. Smalley, G.W. (1984) Classiﬁcation and evaluation of forest sites in the Cumberland Mountains. USDA Forest Service General Technical Report SO SO-50. Smalley, G.W. (1986) Classiﬁcation and evaluation of forest sites on the Northern Cumberland Plateau. USDA Forest Service General Technical Report SO SO-60. Smalley, G.W. (1988). Soil-site relations of upland oaks in northern Alabama. USDA Forest Service Research Note SO SO-64. Smith, D.M. (1986) The Practice of Silviculture, 8th edn. John Wiley & Sons, New York. Smith, H.C. (1980) An evaluation of four uneven-age cutting practices in central Appalachian hardwoods. Southern Journal of Applied Forestry 4, 193–200. Smith, H.C. and Lamson, N.I. (1982) Number of residual trees: A guide for selection cutting. USDA Forest Service General Technical Report NE NE-80. Smith, H.C. and Miller, G.W. (1987) Managing Appalachian hardwood stands using four regeneration practices – 34-year results. Northern Journal of Applied Forestry 4, 180–185. Sonderman, D.L. (1985) Stand density – a factor affecting stem quality of young hardwoods. USDA Forest Service Research Paper NE NE-561. Starkey, D.A. and Oak, S.W. (1989) Silvicultural implications of factors associated with oak decline in southern upland hardwoods. USDA Forest Service General Technical Report SO SO-74, pp. 579–585. Starkey, D.A., Oak, S.W., Ryan, G.W., Taintner, F.H., Redmond, C. and Brown, H.D. (1989) Evaluation of oak decline areas in the South. USDA Forest Service Southern Region Protection Report R8-PR R8-PR-17.

Uneven-aged Silvicultural Methods

379

Stout, S.L. (1991) Stand density, stand structure, and species composition in transition oak stands of northwestern Pennsylvania. USDA Forest Service Research Paper NE NE-148, pp. 194–206. Stout, S.L. and Larson, B.C. (1988) Relative stand density: why do we need to know? USDA Forest Service Research Paper INT INT-243, pp. 73–79. Stout, S.L. and Nyland, R.D. (1986) Role of species composition in relative density measurement in Allegheny hardwoods. Canadian Journal of Forest Research 16, 574–579. Thompson, F.R., III and Dessecker, D.R. (1997) Management of early-successional communities in Central Hardwood Forests. USDA Forest Service Research Paper NC NC-195. Thompson, F.R., III, Probst, J.R. and Raphael, M.G. (1995) Impacts of silviculture: overview and management recommendations. In: Martin, T.E. and Finch, D.M. (eds) Ecology and Management of Neotropical Migratory Birds. Oxford University Press, New York, pp. 201–219. Thompson, F.R., III, Robinson, S.K., Whitehead, D.R. and Brawn, J.D. (1996) Management of central hardwood landscapes for the conservation of migratory birds. USDA Forest Service General Technical Report NC NC-187, pp. 117–143. Toliver, J.R. and Jackson, B.D. (1988) Recommended silvicultural practices in southern wetland forests. USDA Forest Service General Technical Report SE SE-50, pp. 72–77. Trimble, G.R., Jr (1970) Twenty years of intensive uneven-aged management: effect on growth, yields, and species composition in two hardwood stands in West Virginia. USDA Forest Service Research Paper NE NE-154. Trimble, G.R., Jr (1973) The regeneration of Central Appalachian hardwoods with emphasis on the effects of site quality and harvesting practice. USDA Forest Service Research Paper NE NE-282. Trimble, G.R. Jr and Seegrist, D.W. (1973) Epicormic branching on hardwood trees bordering forest openings. USDA Forest Service Research Paper NE NE-261. Trimble, G.R., Jr and Smith, H.C. (1970) Sprouting of dormant buds on border trees. USDA Forest Service Research Paper NE NE-179. Trimble, G.R., Jr and Tryon, E.H. (1966) Crown encroachment into openings cut in Appalachian hardwood stands. Journal of Forestry 64, 104–108. Van Kley, J.E., Parker, G.R., Franzmeier, D.P. and Randolph, J.C. (undated) Field Guide: Ecological Classiﬁcation of the Hoosier National Forest and Surrounding Areas of Indiana. USDA Forest Service, Hoosier National Forest, Bedford, Indiana. Wang, Z. (1997) Stability and predictability of diameter distributions in a managed uneven-aged oak forest. PhD dissertation, University of Missouri, Columbia, MO. Weigel, D.R. and Parker, G.R. (1995) Tree regeneration following group selection harvesting in southern Indiana. USDA Forest Service General Technical Report NE NE-197, pp. 316–325. Weitzman, S. and Trimble, G.R., Jr (1957) Some natural factors that govern the management of oak. USDA Forest Service Northeastern Forest Experiment Station Paper 88. Zeide, B. (1984) Exponential diameter distribution: interpretation of coefﬁcients. Forest Science 30, 907–912. Zimmermann, M.H. and Brown, C.L. (1971) Trees, Structure and Function. Springer-Verlag, New York.

9 Silvicultural Methods For Multi-resource Management

Introduction Contemporary management and protection of oak forests includes objectives that until recently have not been generally considered in the practice of silviculture. These include restoring and maintaining oak savannas, managing oak forests for acorn production, maintaining old-growth oak forests and creating ‘aesthetic’ oak forests. Silviculture can be applied to oak forests to meet all of these objectives. Moreover, these management objectives are not mutually exclusive. For example, managing for oak savannas may increase plant biodiversity and improve acorn production for wildlife while simultaneously creating a vegetation structure with high aesthetic appeal. This chapter presents silvicultural practices designed to meet several non-commodity objectives.

Oak Savannas Extent and characteristics Oak savannas once covered 27–32 million acres in the prairie–forest transition zone that extended from Minnesota to Texas and eastwards into Ohio (Nuzzo, 1986) (Fig. 9.1). In southeastern Minnesota and northeastern Iowa, extensive oak savannas and related ﬁre-dependent communities existed for thousands of years before the arrival of Europeans (Grimm, 1984; Chumbley et al., 1990). Today, midwestern oak savannas have been reduced to about 0.02% of their pre-

settlement (pre-1840) area (Nuzzo, 1986). Oak savannas also may have been more widespread further east in pre-colonial times (Day, 1953). Savannas also covered millions of acres in California, Oregon and Washington before European settlement. Where savannas occurred, they were usually associated with humans and their use of ﬁre (Beilmann and Brenner, 1951; Day, 1953; Cooper, 1961; Wilhelm, 1973; Komarek, 1974; Little, 1974; Dorney, 1981; Pyne, 1982; Grimm, 1984; Biswell, 1989). Oak savannas in the eastern United States are characterized by low to moderate overstorey densities dominated by oaks and a ground layer rich in grasses and forbs (Haney and Apfelbaum, 1990) (Fig. 9.2). The terms oak barrens, prairie grove, oak opening, oak woodland and brush prairie have all been used to describe savannas or savannalike plant communities. Nuzzo (1986) deﬁned Midwestern oak savannas as communities ‘...dominated by oaks having between 10 and 80% canopy, with or without a shrub layer, with an herbaceous, predominantly grassy ground layer composed of species associated with both prairie and forest communities...’. Oak savannas thus possess characteristics of both prairies and forests and can occur in habitats ranging from xeric to hydric (Curtis, 1959). Their remarkable biodiversity is concentrated in the herbaceous layer, which often includes 250 or more plant species. In contrast, tree species diversity in the overstorey is usually low. The oaks that dominate the overstorey may include any of the species native to a 380

Silvicultural Methods for Multi-resource Management

381

Fig. 9.1. The presettlement distribution of the major areas of oak savannas in the United States before 1840. (From Nuzzo, 1986, used with permission.)

region, but those oak species with greater ﬁre resistance are typically more abundant. In the Midwest, ﬁve classes of oak savannas have been tentatively deﬁned based on their ﬂoristic composition, soils and disturbance regime (Haney and Apfelbaum, 1990). They range from tallgrass savannas on mesic sites with fertile soils to dry, sandy ridges with stunted trees.

Western oak savannas are often included as a component of a broader category of vegetation termed oak woodlands. These woodlands are subdivided into savanna forms (less than 30% crown cover) and woodland forms (more than 30% crown cover) (Grifﬁn, 1977). Total woodland area in California exceeds 9 million acres. Oaks dominate most of this acreage,

Fig. 9.2. A dense and diverse herbaceous layer has been maintained by frequent controlled burning of this post oak–black oak–white oak savanna in the Ozark Highlands of Missouri (Turkey Pen Hollow State Park). (USDA Forest Service, North Central Research Station photograph.)

382

Chapter 9

but the proportion in oak savannas has not been accurately estimated. Hardwoods (primarily oaks) grow on an additional 1.6 million acres classiﬁed as grassland in California. Of that grassland acreage, 7% carry more than ﬁve trees per acre and 38% carry one to ﬁve trees per acre (Bolsinger, 1988). Savannas dominated by Oregon white oak once were common within and surrounding the Willamette Valley. Much of the former savanna is now in agricultural use or has developed into oak woodland and closed-canopy forest (Thilenius, 1968). Oregon white oak savannas also occur in northern California. Valley oak savannas were once common on the alluvial soils of California’s Central Valley. Due to agricultural and urban development, the remaining valley oak savannas are now more commonly found in the foothills surrounding the valley at elevations up to 5000 ft and in central California’s Coast Ranges (Grifﬁn, 1976). On drier sites in this region, blue oak and valley oak co-occur in some of these savannas. Blue oak woodlands cover extensive areas in the foothills

surrounding California’s Central Valley and in the Coast Ranges (blue oak–digger pine type; Eyre, 1980). Some of these represent the more open savanna form of woodland. Non-native grasses that limit oak reproduction often dominate the open understoreys of California’s oak savannas and woodlands (Fig. 9.3). They include a wide range of oak species associations that occur between low elevation valley grasslands and montane coniferous forests (Grifﬁn, 1977; Rossi, 1980). Although species associations of oaks vary over the geographic range of oak woodlands, the commonly occurring oaks include valley, blue, California black, California live, canyon live, interior live, Oregon white and Engelmann oaks (Grifﬁn, 1977). With their open canopies and grassy ground cover, many western oak woodlands are similar in appearance to midwestern savannas (Fig. 9.3). The herbaceous vegetation, originally dominated by bunchgrasses, is now often dominated by introduced grasses and lack the diversity of plant species present before the era of livestock grazing and ﬁre suppression.

Fig. 9.3. A California black oak woodland in Yosemite National Park. This open park-like woodland visually resembles the savannas of the eastern United States. The ecology of most western oak woodlands has been altered by cattle grazing, the introduction of non-native herbaceous species and suppression of wildfires. These practices have reduced herbaceous diversity and increased the difficulty of regenerating oaks. (USDA Forest Service, North Central Research Station photograph.)

Silvicultural Methods for Multi-resource Management

Disturbance processes In both the east and west, the disappearance and decline of oak savannas was hastened by a combination of factors including ﬁre suppression, the building of roads and railroads (which often formed effective ﬁre barriers), livestock and deer browsing, cultivation, land clearing for range and pasture improvement, herbaceous competition including the introduction of alien grasses,

383

and urban expansion (White, 1966; Thilenius, 1968; Grifﬁn, 1976; Rossi, 1980; Nuzzo, 1986; Barnhardt et al., 1987; Reed and Sugihara, 1987; Adams et al., 1991; Davis et al., 1991). In the absence of ﬁre, oak savannas in many ecosystems decline in species diversity and develop closed canopies within several decades (Beilmann and Brenner, 1951; Curtis, 1959; Bray, 1960; Thilenius, 1968; Grifﬁn, 1976; Grimm, 1983; Nuzzo, 1986) (Fig. 9.4). Many of today’s

A

B

Fig. 9.4. A black oak savanna in southern Wisconsin. (A) A diverse herbaceous layer has been maintained by periodic prescribed burning, which keeps the oak reproduction (foreground) in a shrub-like state. (B) After several years without burning, oaks and other hardwoods quickly develop into a closedcanopy forest. (Photographs courtesy of Dr Craig G. Lorimer, University of Wisconsin-Madison.)

384

Chapter 9

closed-canopy forests originated from former savannas, and scattered broad-crowned trees persist in closed-canopy forests as relics from an earlier era (Thilenius, 1968; McClain et al., 1993). The transition from savanna to closed-canopy forest is more rapid on the more mesic and productive sites than on xeric sites. The extensive presettlement oak savannas of the Midwest and the Willamette Valley were largely the products of recurrent wildﬁres occurring over thousands of years before those regions were settled by Europeans (Pyne, 1982; Grimm, 1983; Chumbley et al., 1990; Ladd, 1991; Abrams, 1992). During periods when savannas ﬂourished, indigenous people frequently set ﬁres in forests and prairies to increase forage, improve habitat for game species, drive game, harm enemies and open visual corridors (Pyne, 1982; Whitney, 1994). Indigenous people and early European settlers were unencumbered by present-day legal constraints to burning (White, 1991), and they acted without detailed prescriptions or burning plans. Consequently ﬁres varied greatly in frequency, intensity and extent. Historical accounts indicate that vast areas were periodically burned, often in the autumn after the frost killed the herbaceous vegetation (Pyne, 1982; Ladd, 1991). The few studies of historical ﬁre frequency in midwestern savannas are based on ﬁre-scars on old trees. They show that before settlement by Europeans, ﬁres occurred at average intervals ranging from 4 to 17 years (Henderson and Long, 1984; Guyette and Cutter, 1991; Jenkins, 1997). Periods of increasing human population were associated with shorter intervals between ﬁres (Guyette, 1995). After European settlement and before the era of ﬁre suppression, the average time between ﬁres decreased and the frequency of severe ﬁres declined. Within a given savanna, the area affected by ﬁre varied from year to year. In some midwestern savannas, intense ﬁres that burned 20–80% of the savanna occurred at an average interval of 11 years during the 18th century (Guyette and

Cutter, 1991). Additional low-intensity ground ﬁres that did not leave ﬁre scars also probably occurred. Fires varied in intensity and impact due to variation in weather, quantity and condition of fuels, topography and the location of natural ﬁrebreaks (Bowles and McBride, 1998). Coves and lower slopes dampened the spread of ﬁre and often harboured pockets of dense tree cover surrounded by more ﬂammable savanna vegetation. The result was a mosaic of diverse plant communities that included prairies, savannas, dense patches of oak saplings (‘oak scrub’) and closedcanopy forests. Grimm (1984) suggested that the resulting landscape mosaic represented a ‘ﬁre-probability pattern’ because the various intermingled vegetation types were the products of the frequency and severity of burning. He concluded that the resulting landscape mosaic was relatively stable because each vegetation type was in equilibrium with its speciﬁc ﬁre disturbance regime. Recurrent burning thus represented, at the landscape level, a stabilizing effect because it sustained the prevailing state (cf. Anderson and Brown, 1986). Variation in frequency, intensity and season of burning affects the composition, structure and biodiversity of oak savannas. For example, when a Minnesota oak savanna was burned at intervals ranging from 1 to 12 years, the largest number of plant species occurred in areas burned every 2 years (Tester, 1989). Two years without burning alternating with 2 years of burning was proposed as an optimum burning schedule. The years without ﬁre allow fuels to accumulate so that the subsequent burn is hotter than an annual burn and more effective in limiting the invasion or survival of non-prairie species. This relationship may vary among different types of oak savannas. Burning in different seasons also may produce different results (Van Lear and Waldrop, 1988). Most contemporary prescribed burning is conducted in the spring, even though autumn ﬁres were common historically (Ladd, 1991). In oak savannas in the eastern United States, the absence of ﬁre results in the

Silvicultural Methods for Multi-resource Management

invasion of trees and shrubs and a decline in the total number of plant and bird species that inhabit these communities. In northern Illinois, the succession of tallgrass savanna to closed-canopy forest reduced the number of plant species from more than 300 to fewer than 25 and reduced bird species from 28 to 4 (Haney and Apfelbaum, 1990). Bowles and McBride (1998) reported a decrease in herbaceous species richness (i.e. the total number of herbaceous species) with increasing canopy cover in a mesic oak savanna in northern Illinois. In the Ozark Highlands of Missouri, 64 species of birds occur in a community described as ‘grass and shrub savannah’ in contrast to 39 species in a mature oak–hickory forest with little undergrowth (Evans and Kirkman, 1981). In western savannas characterized by a Mediterranean climate (e.g. those of central California), hot and dry summers can limit canopy closure and also the regeneration of oaks, especially in the presence of competition by non-native grasses and foraging by livestock and deer (White, 1966; Danielsen and Halvorson, 1991). In those ecosystems, the role of ﬁre in maintaining or restoring the savanna vegetation is less certain. Grazing and browsing by elk, bison and deer were signiﬁcant historical factors in sustaining savannas (Jenkins, 1997). Bison require about 30 lb of forage per day, and localized grazing disturbances were probably severe (Evans and Probasco, 1977). In addition to reducing herbaceous cover by grazing, these animals affected vegetation by trampling and wallowing, and by distributing seeds in their dung.

Managing oak savannas The conservation of biodiversity is now widely regarded as a legitimate natural resource management objective (Crow, 1990; Hansen et al., 1991). Oak savannas are characterized by high within-stand diversity, and because of their rarity in the eastern and Midwestern United States, their presence increases landscape diver-

385

sity. In theory, the restoration and maintenance of oak savannas should result from restoring the disturbance regimes that historically created and perpetuated them. However, the restoration process may require several decades. Moreover, strategies for restoring savannas differ from those used to maintain established savannas (White, 1986; McClain et al., 1993). Once restored, a savanna is never ‘ﬁnished’ (Packard, 1991). Burning or other appropriate treatments must be continually used to maintain it lest it revert to a closedcanopy forest or other less desirable state. Silvicultural prescriptions for oak savannas need to be tailored to each type of community and to management objectives. However, savanna management is hindered by a lack of tested silvicultural prescriptions. Although long-term studies of savanna ecology and management are few, there is a growing body of knowledge based on practical experience. For midwestern savannas, much relevant information has been compiled in the proceedings of various conferences on savanna ecology and management (e.g. Burger et al., 1991; Fralish, 1994; Stearns and Holland, 1995; US Environmental Protection Agency, 1995). Compilations for western oak woodlands and savannas are also relevant (e.g. Plumb, 1980; Conrad and Oechel, 1982; Plumb and Pillsbury, 1987; Standiford, 1991; Ffolliott et al., 1992; Leach and Ross, 1995; Pillsbury et al., 1997).

Restoration Opinions differ about the management tools most appropriate for restoring oak savannas. Those opinions usually hinge on which of the following management objectives is selected: • re-establishing (simulating) the disturbance processes that historically created savannas and thereby gradually restoring the savanna state, or • quickly re-creating a semblance of an oak savanna by removing trees to a predetermined canopy density, burning and/or artiﬁcially re-establishing an herbaceous savanna ﬂora by seeding or planting.

386

Chapter 9

The ﬁrst approach is a long-term strategy based on restoring the ecological processes that lead to savanna formation. A restored oak savanna is accordingly expected to gradually emerge over several decades of management. The second approach relies more heavily on intensive management practices including some that historically were not associated with savannas. Rapid restoration of savanna vegetation structure (e.g. within a decade rather than over several decades) may be important in areas where: • Visual appearance and aesthetics are primary concerns. • There is an immediate need to provide a savanna demonstration and teaching ‘laboratory’. • Quickly providing habitat for endangered species such as the savannadependent Karner blue butterﬂy is critical. Restoration can be accomplished most rapidly on sites where the vestiges of a former savanna structure and species composition remain. Indicators of former savannas include: • The presence of old, dominant oaks. These may be surrounded by a dense growth of younger trees that developed after burning ceased (Thilenius, 1968; McClain et al., 1993). • The presence of remnant populations of characteristic savanna herbaceous species (see Mead, 1846; Brendel, 1887; Daniels, 1904; Curtis, 1959; Packard, 1991). • Historical records for the site (e.g. Public Land Survey ﬁeld notes) that provide evidence of a past savanna. Selecting areas with low overstorey and mid-storey tree densities and a remnant herbaceous layer comprised of at least some savanna species reduces the time, expense and difﬁculty of savanna restoration. Potentially useful cultural practices include burning, controlled grazing, tree felling, tree deadening, seeding and planting. Some or all of the last ﬁve of these

practices may not be required, depending on the existing vegetation, the desired speed of restoration and the resources available for restoration. Fire is required in the eastern United States and in some western savannas such as those comprised of Oregon white oak in the Willamette Valley. However, the restorative effect of ﬁre is uncertain in blue oak and valley oak savannas where the climate, an understorey of non-native grasses and abundant acorn consumers and seedling browsers limit oak reproduction (Grifﬁn, 1976). There are three important considerations in developing a savanna burning regime. First, ﬁres should be sufﬁciently frequent and intense to kill some of the woody understorey vegetation and midstorey trees. Although burning eliminates ﬁre-sensitive woody plants, it also promotes the accumulation of oak reproduction (Chapter 3). Second, ﬁres should be timed to favour the development of savanna forbs and grasses. Third, extended ﬁre-free intervals lasting a few years to a few decades are needed periodically to recruit oak saplings into the overstorey to replace trees lost to ﬁre, senescence and other causes. A conservative approach suitable to savanna restoration in the Midwest might begin with relatively light burns that reduce accumulated fuels and gradually eliminate ﬁre-sensitive woody species (Jenkins, 1997). Reintroducing frequent burning in remnants of former savannas should, over time, restore the characteristic vegetation structure and composition. Although relying exclusively on prescribed burning may not be the quickest way to restore a savanna, it is a historically authentic approach. Where a closedcanopy forest has developed, restoration may require frequent burning over several decades to effectively encourage the invasion and build-up of the herbaceous savanna ﬂora (White, 1983). In some cases, intense ﬁres may be necessary to reduce the number of large trees (White, 1983). When fuel loads are high, caution is required during the ﬁrst prescribed burns because of the risk of excessive damage to

Silvicultural Methods for Multi-resource Management

overstorey trees. However, such risks need to be balanced against the failure to sufﬁciently reduce overstorey density and thus to re-establish and maintain the desired herbaceous species. Periodic reductions in stand density to appropriate levels also assure a cycle of recruitment of oak reproduction into the overstorey. To quickly restore savannas, tree felling and deadening are sometimes used to reduce overstorey density. Because those practices usually precede prescribed burning, the resulting dead wood increases the fuel load. Concentrated fuels create hot spots that can destroy soil organic matter, retard the re-establishment of herbaceous and woody vegetation, and cause soil erosion (Biswell, 1989). Leaving standing dead trees may reduce this problem and also increase the amount of light reaching the forest ﬂoor. However, timber harvesting in some cases can provide revenue and also reduce the hazards associated with standing dead trees. When burning follows timber harvesting or tree deadening, additional ﬁre-related mortality of residual trees should be anticipated. Immediately after a ﬁre, damage to trees can be estimated from the height of scorch marks on tree boles (e.g. Fig. 10.12). If ﬁre damage to residual trees is extensive, a salvage harvest may be warranted. Controlled grazing also provides a potentially useful tool for establishing and maintaining midwestern savannas. However, there are few guidelines speciﬁc to savanna restoration. Historically, ungulates were functional components of oak savannas that facilitated seed dissemination and nutrient cycling, and thereby the maintenance of the savanna ﬂora (Evans and Probasco, 1977). Grazing nevertheless will not eliminate the need for regular burning in midwestern savannas. Possible negative consequences that should be monitored if grazing is used include soil compaction and the introduction of undesirable seed in manure. In the west, livestock grazing has long been associated with oak savannas, but usually not to the beneﬁt of savanna preservation. Cattle consume acorns and

387

browse oak seedlings, which often are scarce and slow to establish. Deer, gophers and other rodents may further exacerbate the effects of cattle grazing. However, even the use of exclosures to eliminate cattle, deer and gophers failed to satisfactorily establish oak reproduction in valley oak savannas in California’s Santa Lucia Mountains (Grifﬁn, 1976). Where seed sources for herbaceous savanna vegetation (e.g. prairie species) are lacking, they can be seeded or planted (Packard, 1991; McClain et al., 1993). Such measures may be the only way to re-establish herbaceous species that have disappeared from the site. However, careful assessment of herbaceous vegetation and prudent selection of potential sites before savanna restoration is initiated may eliminate the need for costly planting. Savannas are often depicted as visually attractive, park-like settings comprised of a few scattered, wide-crowned trees rising above a dense growth of herbaceous vegetation (Figs 9.2–9.4). Historical descriptions and old photographs conﬁrm that at least some savannas conform to this perception. However, the visual qualities of savannas vary over time. During the course of restoration and maintenance, the presence of standing charred and dead trees, down wood and a brushy understorey may periodically become prominent features that reduce the aesthetic qualities of a savanna. Success in oak savanna restoration should be gauged more by progress towards attaining speciﬁc compositional and structural characteristics than by narrowly perceived aesthetic qualities.

Maintenance In savannas in the eastern United States, burning at short intervals (e.g. biennially) over long periods favours the accumulation of oak seedlings and seedling sprouts (Johnson, 1993). Low intensity ﬁres (e.g. ﬁres with low ﬂames moving slowly against the wind or downhill) are usually lethal only to small stems. Recurring ﬁres often eliminate small oak seedlings and ﬁre-sensitive woody species such as

388

Chapter 9

maples, birches, elms, cherries and ashes. However, oaks often escape lethal temperatures by resprouting from the root collar, which may lie an inch or more below the soil surface (Curtis, 1959; Brown, 1960; Niering et al., 1970; Thor and Nichols, 1973). Some of this reproduction eventually develops large roots and the associated capacity for rapid height growth during ﬁre-free intervals (Johnson, 1979). The cycle that sustains the oaks in these savannas therefore is not continual burning at short intervals, but rather a series of ﬁres at short intervals interrupted by a prolonged period without ﬁre (Streng and Harcomb, 1982; Haney and Apfelbaum, 1990; Bowles and McBride, 1998). Fire-free intervals of 10–20 years may be necessary for some of the oak reproduction to periodically reach the savanna overstorey and thereby replace trees lost to mortality (Haney and Apfelbaum, 1990). The resulting intermittent recruitment process will create an overstorey comprised of several age classes that may be separated by two or more decades. Even if the mean ﬁre-free interval is only a few years, recruitment of oak reproduction into the overstorey can be facilitated by a sequence of ﬁres that is: (i) sufﬁciently variable over time so that long ﬁre-free intervals (e.g. 5–20 years) occasionally occur; or (ii) sufﬁciently variable spatially so that some portions of the savanna fail to burn for several years (Rebertus et al., 1993; Rebertus and Burns, 1997). The resulting burning regime accordingly would produce a continually shifting mosaic of tree cover representing a patchwork of recently recruited trees intermingled with trees of older age classes. Although the upper age limit for many oaks is about 250 years, individual trees may survive longer (Burns and Honkala, 1990). The eventual mortality of overstorey oaks in savannas requires their periodic replacement by new trees recruited from the pool of oak advance reproduction. In oak savannas east of the Great Plains that are subjected to frequent burning over many years, this pool of reproduction is likely to be more than adequate. Prescribed burning guidelines are avail-

able for managing oak savannas in Minnesota (Irving, 1971; White, 1986). These and documented experiences with prescribed burning in other forest types and for other management objectives provide a foundation for developing provisional burning prescriptions for savanna management in various regions (e.g. see Thor and Nichols, 1973; Green, 1980; Van Lear and Waldrop, 1988; Brose and Van Lear, 1997). To some extent, ﬁre frequency and intensity are self-regulating in savannas. Intense ﬁres in an established savanna will decrease fuel loads and thereby reduce the intensity of subsequent ﬁres until fuels build up again. Conversely, a long period without ﬁre will increase the likelihood of an intense burn when the next ﬁre occurs. Variation in weather, season of burning and topography introduce additional natural variation into ﬁre intensity and frequency. A more variable oak recruitment problem characterizes western oak savannas. For example, the decline of blue oak savannas throughout California’s foothills appears to be related to oak recruitment rates insufﬁcient for replacing overstorey trees lost to mortality (White, 1966; Bartolome et al., 1987; McClaran and Bartolome, 1989; Phillips et al., 1997; Swiecki et al., 1997a,b). This problem may be related to poor establishment, growth and survival of oak reproduction under the prevailing conditions of cattle grazing, grass competition, limited moisture and sometimes high overstorey density (Adams et al., 1991; Barnhardt et al., 1991; Danielson and Halvorson, 1991; Davis et al., 1991). There is little evidence that other woody plants are invading these communities in the absence of ﬁre. If these conditions persist, grasslands eventually will replace the oak savannas and woodlands. The role of ﬁre in oak regeneration in western savannas is not fully understood. Fire may be less important than cattle grazing and other problems as a cause of poor oak recruitment (e.g. Swiecki et al., 1997a). In fact, burning may beneﬁt the introduced grasses occurring beneath savannas to the

Silvicultural Methods for Multi-resource Management

detriment of oak reproduction. The role of ﬁre in the maintenance of western oak savannas is more variable and more complex than in oak savannas in the East. None the less, it is well established that ﬁre has played a signiﬁcant, if not always major, role historically in the formation of present-day western oak savannas and woodlands (Thilenius, 1968; Grifﬁn, 1976; Rossi, 1980; Plumb and McDonald, 1981; McClaran and Bartolome, 1989). Despite uncertainties about the role of ﬁre in maintaining western oak savannas, prescribed burning is a common denominator in the management of many oak savannas in both the east and the west. However, the effective use of ﬁre is likely to differ among regions and ecosystems. Periodic monitoring and assessment of woody vegetation structure, tree regeneration and recruitment, and the composition of herbaceous vegetation is essential to a long-term programme of savanna management. Regeneration guidelines (Chapters 2 and 7) adapted to meet savanna management objectives and a crown cover chart (described below) can assist managers in interpreting monitoring results. The value of monitoring will increase over time if adequate records are maintained and there is timely analysis and application of information. Designed experiments in savanna restoration are rare due to the long and highly variable sequence of disturbance factors required in savanna management. Consequently, a better understanding of savanna management seems most likely to emerge, at least in the short term, from monitoring ﬁeld practices and adopting new practices as new information warrants.

Estimating crown cover The exceptional biodiversity of midwestern oak savannas is closely associated with their transitional physiognomy between forest and prairie. Savannas may resemble closed-canopy forests at the upper end of their overstorey density range, and may resemble open prairies at the lower end.

389

Maintaining this transitional state between forest and prairie is a primary goal when restoring and managing oak savannas. Savannas that have a crown cover outside the requisite 10–80% range (or are on a trajectory that will soon take them there) require intervention to maintain the savanna state. If an appropriate tree crown cover is not maintained, the desired herbaceous ﬂora will be difﬁcult to restore or maintain. A crown cover chart designed speciﬁcally for oak savannas in the Central Hardwood Region can be used to assess the current condition of the savanna overstorey and to monitor changes over time (Fig. 9.5). Using the chart requires an estimate of the number of trees and stand basal area. Although it is similar to the Gingrich stocking chart (Fig. 6.9), the 100% crown cover isoline (Fig. 9.5) is analogous to the ‘B’ line on the Gingrich chart (Fig. 6.9). Both lines deﬁne the stand conditions at which crown closure is imminent. The crown cover chart was developed from equations that estimate tree crown width from dbh for open-grown oaks and hickories in the Central Hardwood Region (Krajicek et al., 1961; Krajicek, 1967; also see Table 6.1). Studies have shown that for a given species, tree diameter explains 90% or more of the variation in crown width in open-grown trees (e.g. Krajicek, 1967) (Figs 9.6 and 9.7). On the crown cover chart (Fig. 9.5), the number, mean diameter and total basal area of trees is used to estimate crown area. Crown cover percentages are thus read from the chart isoline that intersects the observed numbers of trees and basal area per acre. The chart accordingly provides an estimate of the maximum crown cover that trees can attain for a given basal area and mean diameter. An oak savanna, especially one that has been recently reduced in density (naturally or by thinning), may have an actual crown cover that is less than that indicated by the crown cover chart. This discrepancy occurs when tree crowns have not had sufﬁcient time to fully expand after thinning. However, there is evidence that

390

Chapter 9

80 14 12

Basal area per acre (ft2)

70

10

9

8

60

Ave r

age

7

tr e e

d ia m

6

e te

5

50

r (i n

ch e

s)

4

40 30

10 0 0

10 100

30

20

200

50

40

70

60

20

80

90

3 100

ent

er c

p ver n co

w

Cro

300

400

500

600

700

800

Trees per acre

90 30 26 24

Basal area per acre (ft2)

80

Average tree diameter (in ches) 22 20 18 16

70

14 0 10

90

60 80

50

70 60

40 50

30

40

20

30

Cr

o

c wn

ov

er

r pe

ce

nt

20

10 10

0 0

10

20

30

40

50

60

70

80

Trees per acre Fig. 9.5. Relation between tree crown cover and basal area, number of trees per acre and average diameter (dbh) in open-grown oak–hickory stands in the Central Hardwood Region. The relation assumes that trees are oaks (black, white and northern red) or hickories at maximum crown expansion for a given diameter. Stands at or above 100% crown cover represent closed-canopy forests. Average tree diameter is the diameter of the tree of average basal area. (From Law et al., 1994.)

crown areas of dominant oaks in closedcanopy stands do not differ greatly from those growing in open stands (McGill et al., 1991). Therefore, large-crowned trees

persisting from an earlier open-grown condition should be adequately represented by the crown cover chart even in savannas with closed canopies.

Silvicultural Methods for Multi-resource Management

391

Fig. 9.6. Open-grown trees such as this bur oak develop predictable maximum crown areas that can be estimated from bole diameter. (USDA Forest Service, North Central Research Station photograph.)

7 Elm Oak–hickory

Crown area (acre percentage)

6

Bur oak Sugar maple

5 4

Black cherry

3 2 1 0 0

5

10

15 20 Dbh (inches)

25

30

Fig. 9.7. Estimated crown area (as a percentage of an acre) of open-grown trees in relation to bole diameter (dbh) for five species. (Adapted from Ek, 1974 (American elm, bur oak and black cherry); Krajicek et al., 1961 (oak–hickory); Smith and Gibbs, 1970 (sugar maple).)

Managing Stands for Acorn Production Oak forests are life support systems for the many animals that live there. Acorns, a staple product of oaks forests, are eaten by many species of birds and mammals including deer, bears, squirrels, mice, rab-

bits, foxes, raccoons, grackles, turkeys, grouse, quail, blue jays, woodpeckers and waterfowl. The health and population density of wildlife species often rise and fall with the cyclic production of acorns. The importance of acorns to wildlife is related to several factors including their widespread occurrence, palatability, nutritious-

392

Chapter 9

ness and availability during the food-scarce autumn and winter months. Where maintaining high quality wildlife habitat is a high priority, sustaining acorn production may be of paramount importance. Even though many of the factors that control acorn production are unknown, there are practical methods for increasing and sustaining production.

Assessing and predicting acorn crops Estimating the quantity of acorns produced by trees or stands and predicting the size of future of acorn crops are important tools in managing wildlife populations and scheduling regeneration practices to take advantage of good acorn crops. Variation in acorn production has been related to differences among oak species, differences among oaks of a single species, and differences among stands due to species composition, stand size structure and site characteristics.

Differences among species Differences in acorn production among species are sometimes reﬂected in reported maximum numbers of acorns produced by single trees in one year (Table 9.1). However,

such observations may not provide a reliable basis for comparing species’ potentials and differences if: (i) the observed numbers of trees and time intervals are too small to consistently include a value approaching the species’ maximum potential; (ii) the observed trees are of variable but unspeciﬁed size; or (iii) species’ population characteristics such as the frequency of occurrence of good and poor acorn crops and the proportion of trees that are inherently poor acorn producers are not considered. A more reliable basis for comparing acorn production among species is the average annual number of acorns produced per tree for a population of trees observed over several years (e.g. Christensen, 1955). However, even that method may not provide a satisfactory assessment of species’ differences because individual trees usually produce acorns in proportion to the cross-sectional areas of their crowns (Tryon and Carvell, 1962). Thus, comparative acorn production may be more meaningfully described by numbers of acorns produced per unit of crown area or per unit of basal area based on many trees observed over several years. A study in West Virginia showed that the average annual production of northern red oak acorns over 5 years was nearly twice that of white oak based on the

Table 9.1. Reported maximum numbers of mature acorns per single tree per year for 12 species.a

Species

Acorns per tree

Years observed (trees sampled)

Blackjack oak Black oak Bluejack oak Chestnut oak N. red oak Post oak Scarlet oak S. red oak Swamp chestnut oak Water oak White oak Willow oak

3,100 9,640 5,500 2,000 4,020 2,630 46,000 5,000 1,000 93,000 23,788 52,200

3 (NA) 4 (55) 3 (NA) 3 (55) 4 (13) 4 (4) 7 (NA) 3 (NA) 3 (NA) 3 (15) 4 (4) 3 (15)

aMaximum

numbers from published sources.

State

Reference

Louisiana Missouri Louisiana New Jersey Missouri Missouri N. Carolina Louisiana Louisiana Arkansas Virginia Arkansas

Moody et al., 1954 Christisen and Kearby, 1984 Moody et al., 1954 Wood, 1934 Christisen and Kearby, 1984 Christisen and Kearby, 1984 Downs, 1944 Moody et al., 1954 Moody et al., 1954 Cypert and Webster, 1948 Feret et al., 1982 Cypert and Webster, 1948

Silvicultural Methods for Multi-resource Management

number of mature acorns per unit of crown area (Tryon and Carvell, 1962). In southern Michigan, northern red oaks produced about three times more acorns than white oaks per unit of crown area over a 4-year study period (Gysel, 1957). Where the objective is to relate acorn production to food potentially available for wildlife, the most relevant expression of production may be the total weight of sound acorns per unit of crown area or basal area because acorns of different species differ greatly in size. For example, a bur oak acorn of average size weighs about six times more than a pin oak acorn of average size and an Oregon white oak acorn weighs about 12 times more than a huckleberry oak acorn (Olson, 1974). A study in North Carolina showed that, among ﬁve species, average annual acorn production of northern red oak was greater than the other species observed based on the fresh weight of mature acorns per square foot of basal area per year (Beck, 1977) (Table 9.2). Comparing the acorn producing capacities of different species based on average yields over only a few years also can be misleading. To resolve that problem, recognized differences among species in acorn producing potential can be used to deﬁne indices of relative acorn production for each species. Sharp’s (1958) index bases acorn production on the average number of acorns per branch tip, where a branch tip is the terminal 24 inches of shoot growth on any branch in the upper one-third of the tree crown. For species in the red oak group, a 24-inch branch tip excludes the current year’s growth and is rated at 100% of potential yield if 32 or more acorns are present. For species in the white oak group, a 24-inch branch tip includes the current year’s growth and is rated at 100% of potential yield if 24 or more acorns are present (Table 9.3). This deﬁnition of 100% acorn yield thus considers differences among the two major species groups in the maximum numbers of acorns likely to occur on a typical branch. A similar acorn production index considers inherent differences among species

393

Table 9.2. Average annual weight yield of mature acorns per square foot of basal area in North Carolina.a Acorns (lb ft2) of basal area Species

Mean

Northern red oak White oak Scarlet oak Chestnut oak Black oak

12.2 8.3 3.0 2.4 1.6

12-Year range 0–62.2 0.1–39.8 0–15.4 0–17.6 0–5.4

aBased

on 12 years of observations from six 2/3 acre plots in 63- to 82-year-old even-aged stands. (From Beck, 1977.) Reported values are gross production and include insect-infested acorns.

in acorn-cluster size and percentage of the crown containing acorns (Christisen and Kearby, 1984). A cluster consists of a closely associated group of acorns (e.g. the cluster of second-year acorns in Fig. 2.4) rather than the entire branch tip. In application, a northern red oak with clusters of ﬁve to seven acorns and a black oak with clusters of twice that number both receive the maximum index rating of nine. Both cluster sizes express the average maximum potential for their respective species. The rating system also recognizes that white oaks produce acorns on limbs and branches midway down the trunk or lower, whereas black oak and northern red oaks produce acorns primarily in the upper onethird of their crowns. This index can be used to estimate numbers of acorns per unit crown area for each species (Fig. 9.8). Depending on species, the production index accounts for 75–87% of the variation in the observed numbers of acorns per unit of crown area (Myers, 1978). Other visual acorn crop rating systems also have been proposed (e.g. Graves, 1980; McDonald, 1992; Koenig et al., 1994). Koenig and others (1994) estimated acorn production based on 15-second counts of acorns on each tree by each of two observers. They found close agreement between those estimates and numbers of acorns collected in acorn traps located beneath the same trees. They argued for recording and using acorn counts directly

394

Chapter 9

Table 9.3. Sharp’s (1958) acorn production index.

Production index

Relative abundance rating of acorns

0 1 2 3 4 5

None Trace Poor Fair Good Bumper crop

Average number of acorns observed per branch tipa

Per cent of maximum potential number of acorns per branch tipa

White oaks

Red oaks

0 <2.4 2.4–6.1 6.2–13.2 13.3–18.1 18.2+

0 <3.2 3.2–8.2 8.3–16.2 16.3–24.2 24.3+

0 <10 10–25 26–50 51–75 76–100

aA

branch tip is any terminal 24 inches of shoot growth in the upper one-third of the tree crown. For species in the red oak group, a 24 inch branch tip excludes the current year’s growth and is rated at 100% of potential yield if 32 or more acorns are present. For species in the white oak group, a 24 inch branch tip includes the current year’s growth because all acorns maturing in the current year are on the current year’s shoot growth. White oak branch tips are rated at 100% of potential yield if 24 or more acorns are present.

rather than categorizing them by index classes. Such counts, they pointed out, provide greater ﬂexibility in statistical analyses. Moreover, they advocated counting acorns on the same trees each year to make

Poor

Fair

among-year comparisons more accurate and to decrease the sample size needed to accurately estimate overall productivity. They found that estimating acorn crops based on their visual survey protocol was

Good

Heavy

14 White oak

Acorns per ft2 of crown area

12 10 Black oak 8 6 Scarlet oak N. red oak

4 2 0 1

2

3 4 5 6 7 Acorn production index

8

9

Fig. 9.8. Acorn production per square foot of crown area in relation to an acorn production index for Missouri stands. The index ratings are as follows: (1) few to none; (2) poor – sparsely scattered acorns; (3) poor to fair; (4) fair – evenly distributed single acorns and small clusters; (5) fair to good; (6) good – evenly distributed acorns with numerous small- and medium-sized clusters; (7) good to heavy; (8) heavy – numerous medium- and large-sized clusters throughout the crown; (9) bumper – very high acorn density over a large percentage of the crown. The production index accounted for 75–87% of the variation in stand yield of acorns, depending on species. (From Myers, 1978.)

Silvicultural Methods for Multi-resource Management

16–25 times faster than estimating based on acorn traps. Such time savings potentially facilitate more intensive monitoring of acorn production for a given area or species. Visual surveys such as those described above do not, by themselves, directly yield data on the absolute numbers of acorns produced by trees. However, those numbers can be estimated by simultaneously monitoring acorns in traps randomly set out under a subset of the observed trees (Fig. 2.9). Then the relation between the two measures, i.e. the visual counts and numbers of acorns per unit area from the traps, can be established by regression methods (Koenig et al., 1994). However, this method is only as accurate as the trap data. Sources of error in acorn trap counts originate from the relatively small proportion of ground area that can be practically sampled and animal pilfering of acorns from traps. The latter source of error can be reduced by counting acorn caps in addition to acorns (Auchmoody et al., 1993; Koenig et al., 1994). Another approach to estimating numbers of acorns is based on the bifurcation rate of tree branching. Bifurcation rate (br) is the number of times a tree branches along the route from the ground to any given terminal bud. The number of branching bifurcations remained remarkably constant in a population of 23 oaks (12 white and 11 black oaks) in central Missouri that ranged from 8 to 23 inches dbh (Cecich and Larsen, 1997). The mean bifurcation rate of 8.9 (standard deviation 1.1) did not differ signiﬁcantly between species. Bifurcation rate can be used to estimate the total number of potential acorn-producing branches per tree. If the average number of acorns per branch is known (e.g. as required in determining Sharp’s acorn production index), the total number of acorns per tree can be estimated by multiplication. For the observed population of Missouri trees, the average number of branch tips can be cal-

395

culated as 2br (= 28.9 = 478). Thus, if the average number of acorns per branch is known (including lower branches not usually counted in applying Sharp’s method), this average can be multiplied by 2br to estimate whole-tree acorn production.1 To be useful, such estimates would require intensive measurements of branching in trees and accounting for possible sources of local or regional variation in the bifurcation rate of tree branching before they could be conﬁdently applied to large populations of trees (e.g. an entire stand, forest or ecoregion). The method none the less deserves further testing in designing more accurate methods of estimating acorn crops. Even the most reﬁned applications of the methods described above may not eliminate major errors in estimating acorn production. Errors are most likely to occur when estimates are applied to stands where the size of trees (stand structure) and other stand characteristics differ markedly from the stands in which the acorn production model was developed.

Effects of tree size and stand characteristics The acorn-producing potential of individual trees in forests is partially related to stand density and the associated exposure of tree crowns to light. Other factors being equal, oaks growing in the open produce more acorns than those growing in closedcanopy forests (Sharp, 1958; Sharp and Sprague, 1967). Open-grown trees have: (i) maximum crown area for a given stem diameter (Krajicek et al., 1961); (ii) large numbers of branches per unit of crown area and thus large numbers of buds from which acorns can arise (Sharp, 1958); and (iii) full crown exposure to light and thus maximum photosynthesis per unit leaf area (Verme, 1953). For black oaks, acorn production was 11 times greater in portions of

1For species in the red oak group, the correct multiplier is 2br1 because acorns are borne on 2-year-old branches. Also, the multiplier 2br can be adjusted to account only for branches in the upper one-third of the crown as required in using Sharp’s (1958) acorn inventory system.

396

Chapter 9

the crown exposed to full light than in shaded portions (Verme, 1953). Similarly, exposed crown areas of white oak and northern red oak produced ﬁve times more acorns than shaded crown areas (Post, 1998). Northern red oaks shaded on one or more sides by other trees tended to produce more ﬂowers and immature acorns on branches receiving full sunlight. Relative stand density (stocking) is an indirect measure of crown competition. Stocking can vary greatly even among stands with closed canopies (Chapter 6). As stocking increases, the mean crown area per tree decreases, which in turn reduces the mean ratio of tree crown area to tree diameter. Stocking charts and equations that relate stand basal area and mean stand diameter to relative stand density (stocking percentage) are expressions of these relations (e.g. Figs 6.9 and 6.11). Healy (1997) related stocking percentage to differences in acorn production of individual northern red oaks in a mixed northern red oak stand in Massachusetts (initial age 42 years). Two levels of initial stocking were observed over 6 years: 40 and >100% based on Gingrich’s (1967) stocking chart (Fig. 6.9). At 40% stocking, trees in the thinned stand did not fully utilize growing space, and thus represented a somewhat open-grown stand condition. Based on acorn sampling beneath the crowns of 60 dominant and co-dominant red oaks in each treatment, trees in the thinned stand produced signiﬁcantly more acorns than trees in the unthinned stand based on 6-year average production. Although the proportion of trees bearing acorns and the proportion of sound acorns were similar in thinned and unthinned stands in most years, there was great year-to-year variation in total production. In years of poor production, the proportion of sample trees bearing acorns was greater in the thinned than the unthinned stand. A companion study of eight stands ranging from 62 to 82 years old compared stand-wide acorn production in thinned and unthinned stands (Healy, 1997). When averaged over the 3 year study period,

acorn production in thinned stands (averaging 71% stocking) and unthinned stands (averaging 100% stocking) did not differ signiﬁcantly. The lack of signiﬁcance between the 3-year means was attributed to large year-to-year variation in acorn production. However, when acorn yields for each of the three years were considered individually, annual yields were consistently greater (nominally) in thinned than in unthinned stands. As in the individual tree study, the relative effects of thinning were greatest during years of poor acorn production and least during years of high production. Apparent gains from thinning ranged from 59 to 93% based on the dry weight of acorns per acre and from 42 to 94% based on numbers of sound acorns per acre (Healy, 1997). Thinning also can reduce acorn yields per acre even though individual tree yields may increase (e.g. Harlow and Eikum, 1963). This effect may be attributable to two factors: • A reduction in total oak crown area. However, if residual stand density is within the range representing full utilization of growing space (e.g. between A- and B-levels of stocking in Fig. 6.9), oak crowns can expand after thinning to capture the growing space vacated by thinned trees. At those stocking levels, reduction in total oak crown area should be temporary (assuming the retained trees are healthy). • The chance removal during thinning of one or more of the relatively few but inherently good acorn producers per acre. Although the remaining producers may each increase in yield, the increase may not compensate for the removal of good producers. If increasing or maximizing acorn production is the primary goal of thinning, good producers should be identiﬁed before thinning and then retained (see the section below on ‘Guidelines for sustaining acorn production’). Reported effects of thinning, where thinning is applied without regard to individual-tree acorn production characteristics, therefore should be

Silvicultural Methods for Multi-resource Management

carefully evaluated. Even more generally, the reporting of no signiﬁcant differences in acorn production related to stand stocking (or any other stand characteristic) may be the result of inadequate sampling of acorns. Vertical stratiﬁcation of trees produces recognizable canopy crown classes (Smith, 1986). Some investigators have noted that trees in the dominant and co-dominant crown classes produce more acorns per unit of crown area than trees in the intermediate and suppressed crown classes. The small, shaded crowns of trees in lower crown classes develop few branches per unit crown area, and this reduces their acorn-bearing potential (Kittredge and Chittenden, 1929; Moody, 1953, unpublished; Gysel, 1956). Thus, expressing acorn production as yield per unit of crown area without considering variation in stand density and structure (i.e. the distribution of crown classes) obscures potential differences related to variation in crown class and crown exposure. In most stands, crown class is correlated with bole diameter, which in turn is correlated with crown area and acorn produc-

tion. Thus, bole diameter can account for signiﬁcant variation in acorn production that is related to both crown class and crown area. In southern Appalachian forests, this relation has been used to develop models for estimating average annual acorn production of dominant and co-dominant trees from tree diameter (Fig. 9.9). In the south, similar models were developed to estimate average annual fresh-weight yields of mature acorns from tree diameters (Goodrum et al., 1971). Dbh explained 48–81% of the variation in acorn yield, depending on species. In the same study, predictive models based on crown radius explained 76–94% of the variation in total acorn yield (Fig. 9.10). Although crown radius produces more accurate estimates of acorn yield than bole diameter, the former is more difﬁcult to measure. For that reason, bole diameter is usually the predictor of choice. Few studies have evaluated the effects of site quality, i.e. measures of forest productive capacity (Chapter 4), on acorn production. A 4-year southern Michigan study indicated that site quality had no effect on white oak or black oak acorn production when production was expressed as num-

4000 Scarlet oak

Acorns per tree

3000

Black oak

2000

White oak 1000 Chestnut oak N. red oak 0 10

14

18

22

397

26

30

34

Dbh (inches) Fig. 9.9. Average annual production of mature acorns in the southern Appalachians in relation to tree diameter based on a 7-year study of dominant and co-dominant trees. (From Downs, 1944, used with permission.)

398

Chapter 9

50 White oak

Acorn yield (pounds/tree)

40 Water oak 30

20 Blackjack oak

Post oak Southern red oak

10

0 5

10

15 20 Crown radius (ft)

25

30

Fig. 9.10. Estimated average annual yield (fresh weight) of sound mature acorns in relation to crown radius for five species in the upper coastal plain of Texas and Louisiana. Estimates are based on 63–507 trees observed for 6–18 consecutive years, depending on species. Crown radius explained 48–81% of the variation in yield, depending on species. Models were based on the observed weight of acorns per size class (diameter or crown radius class) per year averaged over all trees in the size class. (From Goodrum et al., 1971, used with permission.)

bers of fully developed acorns or acorn weight per unit of crown area (Gysel, 1957). Similarly, there was no evidence that site quality as measured by site index and topographic factors in West Virginia forests inﬂuenced the production of northern red and white oak acorns (Tryon and Carvell, 1962). Other factors correlated with site quality nevertheless may affect acorn survival and seedling establishment. For example, rich, moist habitats provide better conditions than drier forests for organisms such as fungi that can initiate decay in acorns after they have fallen (Winston, 1956). Wildlife that consume acorns also may prefer certain habitats and associated site quality factors over others (Gysel, 1957). All the predictive models discussed above express acorn yield on an average annual basis and ignore the enormous yearto-year variation in acorn production. Although such models provide useful estimates of average acorn production over relatively long periods such as a decade, they

cannot accurately estimate acorn production for any given year. A basis for predicting yields in advance of a few weeks before acorn drop has yet to be developed (Cecich and Sullivan, 1999). Lack of progress in developing such models is due to inadequate knowledge of the ﬂowering and fruiting process and the numerous factors that inﬂuence ﬂowering and acorn production. Many of these factors, including weather and insect events, essentially occur at random and can only be accounted for probabilistically. One proposed model for predicting annual variation in populations of pistillate ﬂowers and acorns in white oak is based on the probabilities of occurrence of events known to reduce acorn production (Larsen and Cecich, 1997). These include the relative frequency of low relative humidity (below 60%) during the 1-week pollination period, hail storms that destroy ﬂowers, insect populations that reduce numbers of ﬂowers (tree hoppers) and acorns (acorn weevils), summer droughts

Silvicultural Methods for Multi-resource Management

that cause premature ﬂower and acorn abscission, and genetic factors related to ﬂower fertilization. A similar model incorporates these and other factors into a hierarchical framework that estimates acorn production at stand and landscape scales (Sullivan, 2001). The model considers the interactions between landform and the weather events listed above, and also tree sizes and ages, the relative proportion of red oaks and white oaks, and differences in their ﬂowering and fruiting phenologies. Simulations through time consider the lagged effects of weather conditions on the 2-year development cycle of red oak acorns and on the annual development cycle of white oak acorns. Aggregate probability estimates of acorn production for stands can be linked to a landscape model that simulates future patterns of forest age structure and species composition in response to disturbance by wind, ﬁre and timber harvest (Mladenoff and He, 1999; Shiﬂey et al., 2000; Sullivan, 2001). The result is a probabilistic model that can be applied to ten acres or tens of thousands of acres. Although probabilistic models are not able to predict a given future years’ acorn crop with a high degree of certainty, they are useful in estimating the probability of a good (or poor) acorn crop in a given future year, the expected frequency of good and poor crops over a decade or other deﬁned period, and the probability of occurrence, by species, of 2 or 3 consecutive years of poor acorn crops. The latter estimates are useful to wildlife managers concerned with minimizing overall ﬂuctuations in acorn production. Such predictive models are also useful for assessing the relative importance of the various factors known to limit acorn production and thus the regeneration of oak forests.

Guidelines for sustaining acorn production Acorn production in established stands can be sustained, and perhaps even increased, by applying the following guidelines:

399

• Before the ﬁrst thinning (e.g. at stand age 25–30 years), identify and reserve the good acorn producers in each stand. To do this, record the acorn production of candidate trees for 5 years or more. If that is impractical, assess the acornproducing capacity of individual trees by observing production during a single year in which a good to excellent acorn crop occurs for one or more of the major species present. However, in the red oak group, many good producers may be overlooked because not all trees of those species may produce well in the same year. Criteria for identifying good acorn producers are given by Sharp (1958) (Table 9.4). In the oak–hickory region, the best time to rank trees is from 10 to 25 August, which is before acorn consumers have eaten or cached many acorns. Acorns are best observed with binoculars on bright days when they are silhouetted against the sky (Fig. 9.11). • During thinning, retain a mixture of oak species to minimize the impact of the large year-to-year ﬂuctuation in acorn production in any one species. • Thin around the identiﬁed acorn producers to expose their crowns to full light on all sides. This facilitates crown expansion and increases branch density. Table 9.4. An acorn production ranking system for individual oaks.a Average number of acorns per branchb Ranking

White oak group

Red oak group

Excellent Good Fair Poor

18+ 12–17 6–11 <5

24+ 16–23 8–15 <8

aNote that in any one year, excellent producers may not reach their potential because of unfavourable environmental factors. (Adapted from Sharp, 1958.) bBased on the terminal 24 inches of healthy branches in the upper one-third of crowns. For species in the red oak group, a 24 inch branch tip excludes the current year’s growth; for species in the white oak group, a 24 inch branch tip includes the current year’s growth.

400

Chapter 9

Fig. 9.11. If the 24 acorns on this northern red oak branch were average for the tree as a whole during a bumper acorn-crop year, the tree would be ranked as an excellent acorn producer (Table 9.4). In applying the ranking system, acorns are counted (with the aid of binoculars) on the terminal 24 inch lengths of healthy branches in the upper one-third of the crown. (USDA Forest Service, North Central Research Station photograph.)

Branch density increases acorn production per unit of crown area because of the increase in numbers (density) of acorn-bearing branches (Verme, 1953). Among potential acorn producers, dominant and co-dominant trees will be the most efﬁcient producers. Area-wide thinning is not required because only 20 or fewer good seed producers are likely to occur per acre even in pure oak stands. However these seed producers typically will be dominant and co-dominant trees, and they may account for proportionately more basal area and stocking than their numbers alone indicate. • Adjust the rotation age (or in unevenage management, adjust the maximum tree diameter) to increase the number of trees in the size classes that typically produce the most acorns. Maximum production will vary by species. For example, production in northern red oak peaks when tree dbh reaches 20

inches and then it declines in larger trees (Fig. 9.9). In contrast, white oak production is maximized at about 26 inches. Many other species, however, do not exhibit well-deﬁned diameterrelated peaks in production, at least within the diameter ranges that have been reported (Downs, 1944; Goodrum et al., 1971). Large, senescent trees are usually poor acorn producers (Huntley, 1983). The guidelines presented above can be applied to many kinds of oak forests, both upland and lowland. The latter include green tree reservoirs, which are bottomland oak stands managed speciﬁcally for acorn production and waterfowl habitat. They thus require special management methods not common to upland oak forests. Green tree reservoirs are artiﬁcially ﬂooded in winter (usually November to February) to attract migrating waterfowl (Fig. 9.12). The waterfowl feed on fallen acorns, which

Silvicultural Methods for Multi-resource Management

401

A

B

Fig. 9.12. Two pin oak stands in a managed green tree reservoir in southern Illinois. (A) A flooded stand during the dormant season; continuous controlled flooding (to a depth of 1–3 ft) usually occurs from November to February. (B) To minimize oak mortality, stands must remain unflooded during most of the growing season. Several years with no flooding may be required to establish oak reproduction for regenerating older stands that are declining in vigour and acorn production. (USDA Forest Service, North Central Research Station photographs.)

sink to the forest ﬂoor. Pin oak is usually the principal oak, but willow, overcup, water and cherrybark oaks are also common. Many green tree reservoirs are adjacent to the Mississippi River and its major tributaries. Flooding is controlled by artiﬁcial levees and ﬂooding depths are usually 1–3 ft. These areas are relatively ﬂat with poor surface drainage, soils are typically clayey with poor internal drainage even when not ﬂooded, and stands are domi-

nated by oaks. Standing ﬂoodwater during the dormant season is not detrimental to the established trees. During years of good production, acorns are a principal autumn and winter food for mallards, wood ducks and other ducks (McQuilkin and Musbach, 1977). The combination of shallow water and abundant acorns makes green tree reservoirs a preferred winter habitat for ducks, and also provides opportunities for autumn duck hunting.

Chapter 9

Acorns per acre (1000s)

Like other oak forests, acorn production in green tree reservoirs varies from year to year. Over 12 years, average annual acorn production in a Missouri green tree reservoir was 63,000 sound acorns (approximately 150 lb) acre1 year1. However, yields ranged from 8 to 440 lb acre1 year1 (Fig. 9.13). Acorn crops failed completely every 3–4 years and there was only one year with a bumper crop (McQuilkin and Musbach, 1977). Stands that were not ﬂooded produced about one-third more acorns than ﬂooded stands. However, because insect predation of acorns was greater in unﬂooded stands, the number of sound acorns in each case was nearly equal (Minckler and McDermott, 1960; Minckler and Janes, 1965; McQuilkin and Musbach, 1977). The percentage of acorns that were sound ranged from 60 to 90% in ﬂooded stands and from 30 to 60% in unﬂooded stands (McQuilkin and Musbach, 1977). Although the reason for this difference was not apparent, the results demonstrated that annual ﬂooding, at least in this case, did not reduce but increased the production of sound acorns.

In ﬂooded stands, the number of acorns that failed to mature or that were insectdamaged was approximately the same each year. Good acorn yields thus occurred only when bumper crops overwhelmed losses to insects. By the time pin oak stands in green tree reservoirs are 25–35 years old, they are capable of producing large acorn crops. In stands more than 25 years old, production increases as the number of large trees (>11 inches dbh) increases. Acorn production also may increase with increasing stand basal area (up to 90 ft2 acre1). However, green tree reservoirs with basal areas as low as 40 ft2 acre1 can produce good acorn crops if stands are comprised of large-diameter trees (McQuilkin and Musbach, 1977). It is therefore possible to sustain good acorn production and simultaneously meet other management objectives if appropriate stand densities and large-crowned acorn-producing trees are retained during thinning. Sustaining high acorn production in green tree reservoirs requires the periodic replacement of oaks lost to mortality and

180

439

160

390

140

342

120

293

100

244

80

195

60

146

40

98

20

49

0 1956

1958

1960

1962 1964 Year

1966

1968

Acorns per acre (pounds)

402

0 1970

Fig. 9.13. Fluctuation in annual production of sound, mature acorns (predominantly pin oak) over 12 years in a green tree reservoir in southern Missouri. Values are averages for stands with basal areas ranging from 40 to 90 ft2 acre1. (Adapted from McQuilkin and Musbach, 1977, used with permission.)

Silvicultural Methods for Multi-resource Management

timber harvesting. Even when acorns are abundant, establishing new oak seedlings may be problematic when stands are annually ﬂooded. Due to the limited size of green tree reservoirs and their high development costs, lengthy periods without ﬂooding are likely to be unacceptable to the forest or wildlife manager. Reduction of canopy cover, site preparation and cessation of ﬂooding for several years nevertheless may be required to establish oak reproduction and recruit new oak seedlings into the overstorey (Minckler and McDermott, 1960; also see Chapter 3). Although tested regeneration prescriptions are lacking, shelterwood and group selection methods would seem to have potential application in regenerating and managing oaks for acorn production in green tree reservoirs.

Old-growth Oak Forests The term ‘old growth’ is used to denote forests or stands of trees that have remained largely undisturbed by humans over long periods (e.g. more than a century). Although old-growth oak forests comprise only a tiny fraction of today’s oak resource, they are potentially important features of the landscape because of their role in maintaining biodiversity. In addition to the presence of relatively old trees, old-growth forests possess other characteristics that are not well represented in other stages of forest development. Moreover, certain species and life forms are closely bound to these special attributes of old growth. They also provide ecological ‘laboratories’ for studying natural processes of stand development and for establishing benchmarks against which the effects of silvicultural practices can be gauged. Oldgrowth forests also rank high aesthetically and are places where some people ﬁnd psychological and spiritual solace and renewal (Schroeder, 1996). Nevertheless there are many questions concerning objectives and strategies for managing oldgrowth oak forests.

403

Extent and characteristics Estimates of the extent of old-growth forests in the eastern United States range from 500,000 to more than 1.5 million acres (Davis, 1996; Leverett, 1996). However, oaks dominate only a fraction of that. In 1937, an estimated 350,000 acres of ‘virgin’ upland oak forest remained within the Central Hardwood Region (Schnur, 1937). This amounted to 0.3% of the total oak forest acreage. Today, old-growth hardwood forests of all types in the Central Hardwood Region collectively cover about 100,000 acres (Parker, 1989). In Indiana, Illinois, Missouri and Iowa, old-growth forests cover about 22,000 acres, or approximately 0.08% of the forests of those states (Shiﬂey, 1994). Most of those forests include oaks as a major component. In Michigan, Wisconsin and Minnesota, 138,000 acres of oak–hickory forest are at least 120 years old. However, only about 900 of those acres have not been logged to some extent. In some regions, remnants of old-growth oak forests may be more widespread than commonly recognized because certain oldgrowth forest conditions are inconspicuous and have been overlooked. For example, about 19% of the oak forests on steep, droughty south-facing slopes in northwestern Arkansas are dominated by post oaks ranging from 140 to 320 years old (Stahle and Chaney, 1994). These trees, which occupy sites categorized as non-commercial forestland, are typically less than 65 ft tall and between 10 and 24 inches dbh. The harsh environment of those sites naturally maintains small trees and low stand densities. Such old-growth remnants may comprise as much as 0.8% of the forest area of northwestern Arkansas. In addition to old trees, certain compositional, structural and dynamic features characterize old growth forests (Fig. 9.14). These include the continual creation of senescent and standing dead trees, of canopy gaps and attendant forest regeneration, and of down wood in various states of decomposition. These conditions occur

404

Chapter 9

Fig. 9.14. A large northern red oak dominates neighbouring sugar maples and other hardwoods in this mixed mesophytic old-growth forest in southern Indiana. Although smaller trees are more numerous than larger trees in old-growth forests, it is the grandeur of the largest trees that usually catches our eye. Old-growth stands often form and naturally maintain a negative exponential diameter frequency distribution. (USDA Forest Service, North Central Research Station photograph.)

when major disturbances are excluded from a forest over a long period. Accordingly, old growth could be deﬁned as forests that are ‘relatively old and relatively undisturbed by humans’ (Hunter, 1989). However, other criteria are needed for practical deﬁnitions that can be applied to speciﬁc forest types and geographic regions. These criteria include the attainment of tree ages sufﬁcient to represent a relatively stable species composition, mean net stand growth near zero, dominant trees of an age that represents their speciesspeciﬁc average life expectancy, and the absence of timber harvesting at intensities

or frequencies that can signiﬁcantly alter species composition (Hunter, 1989). Investigators generally agree on the major characteristics associated with oldgrowth hardwood forests (Table 9.5). Nevertheless, the great variability of old growth, even within a forest type or region, renders a precise deﬁnition difﬁcult if not impossible. This problem persists despite many written descriptions of individual old-growth stands (Nowacki and Trianowsky, 1993; Tyrrell et al., 1998). The commonly held perception of oldgrowth forests as pristine plant communities unaffected by humans is largely a myth (Day, 1953; Pyne, 1982; Lorimer, 1985; Whitney, 1994; Hicks, 1997; also see Chapter 1). This is especially true of oakdominated old-growth forests, many of which have been affected by human disturbances. Although old-growth forests are noted for their low level of human disturbance, virtually every existing old-growth oak forest in the United States has been periodically burned by Native Americans or by European settlers (Parker, 1989; Ladd, 1991; Whitney, 1994; Guyette, 1995; White and White, 1996). Many also have been disturbed to various degrees by domestic cattle, and others have been selectively logged to a limited extent during the last two centuries. Individually or collectively, such disturbances do not negate the value of these old growth remnants. In fact, the long history of such disturbances raises questions about the degree to which low intensity disturbances such as ﬁres should be considered a normal part of old-growth forest development.

Silvicultural options A silviculture for old-growth forests may at ﬁrst appear to be an oxymoron. If the management objective is simply to maintain and protect existing old growth by minimizing human-caused disturbances, then any silvicultural treatment would be incompatible with that objective. However, other objectives may include: (i) maintaining a species composition that is best

Table 9.5. Characteristics used to define old-growth hardwood stands in the eastern United States and associated values and qualities observed in old-growth oak stands. Observed values and qualities of defining characteristic for

Defining characteristics for old-growth hardwood stands

Source: Parker, 1989

Source: Meyer, 1986

High species richness and diversity Species richness ≥20 canopy trees. Shannon– Weiner index >3 Evenness >0.5 Simpson Index <0.3

Tree species richness 20–40 Herbaceous species richness 17–53 Breeding bird species richness 18–33

Diverse dominant tree species composition

24–39 woody species with dbh ≥0.8 inch

Uneven-aged with canopy species in several size classes

Multilayered canopy. Wide range of tree heights and ages

Uneven-aged with canopy species present in many size and age classes

Several large canopy trees

Decadence evident in tops and boles of large trees

Largest oaks 30–55 inches dbh

Mesic and wet-mesic old-growth oak stands Source: Tyrell et al., 1998

Xeric old-growth oak stands Source: Tyrell et al., 1998

Largest oaks 50–77 inches dbh

Largest oaks 40–48 inches dbh

Oldest trees at least 112 years and as old as 336 years

Oldest oaks 200–400+ years

Oldest trees ≥150 years

Density 770–2250 trees per acre ≥1 inch dbh. Density 90–190 trees per acre ≥4 inches dbh

Density 50–245 trees per acre ≥4 inches dbh Density 2–70 trees per acre ≥20 inches dbh

Density 115–180 trees per acre ≥4 inches dbh

Commercially important species abundant. Evidence of very limited selective logging along perimeter of some tracts

Large, high-quality commercially important trees present (indicating no recent harvest) Oldest trees ≥200 years

Mean age of overstorey 135–210 years. Max. age 190–375 years.

Overstorey density approximately 100 trees per acre ≥4 inches dbh

Overstorey density approximately 65–1707 trees per acre ≥4 inches dbh

Oldest trees ≥100 years

Continued

Silvicultural Methods for Multi-resource Management

Source: Martin, 1992

Upland old-growth oak and mixed hardwood stands in IN, IL, MO, IA Source: Shifley, 1994; Shifley et al., 1995, 1997; Spetich, 1995; Roovers and Shifley, 1997; Spetich et al., 1999

405

406

Table 9.5. Continued

Overstorey basal area

At least 25% stocking

Basal area 91–141 ft2 acre1

Basal area 2–205 ft2 acre1

Basal area 40–120 ft2 acre1

≥110 ft2 acre1

110–150 ft2 acre1

in live trees

for trees ≥1 inch dbh

for trees ≥4 inches dbh

for trees ≥4 inches dbh

Volume 16,000–25,000

≥14 inches dbh

Basal area 83–139 ft2 acre1

56–84 snags/acre ≥1 inch dbh Down wood volume 350–500 ft3 acre1 for stems ≥4 inches diameter

for trees ≥4 inches dbh

board ft per acre Logs and snags present in various sizes and stages of decay

7–17 snags/acre ≥4 inches dbh Dead wood on ground 7–11 tons acre−1 Annual tree mortality 0.6–0.9%

Large snags and large down logs widely distributed Large logs present in streams and drainages

37–128 snags/acre ≥4 inches dbh Snag basal area 7 to 20 ft2 acre1 Snags roughly 10% of live trees by dbh class Volume of snags 140–510 ft3 acre1 Down wood volume 350– 1580 ft3 acre1 for stems ≥4 inches diameter

45–68 snags/acre ≥1 inch dbh 0–36 snags/acre ≥4 inches dbh Snag basal area 3–14 ft2 acre1 for trees ≥4 inches dbh 50–295 down logs/acre ≥4 inches diameter Down wood volume 300– 980 ft3 acre1 for stems ≥4 inches diameter

Tree-fall gaps formed by windthrow; gaps usually <1.2 acres

Gaps are 7–8% of forest; randomly distributed; range from 0.12 to 0.9 acres in size

Various degrees of understorey density and of herbaceous ground cover

Numerous tree-fall gaps

0.7–9% of area in canopy gaps

Plants and animals that prefer old growth Undisturbed soils and soil macropores Little or no evidence of human disturbance

3–8 cavities/acre with smallest opening = 4 inches. Organic matter 5–12% in first 2 inches of soil. Litter dry weight 2–12 tons acre1. Undisturbed by harvest. No appreciable fire or grazing in last 50 years

Chapter 9

Observed values and qualities of defining characteristic for

Defining characteristics for old-growth hardwood stands Overstorey basal area

Silvicultural Methods for Multi-resource Management

407

Fig. 9.15. Old-growth oak forests are characterized by large volumes of down wood – often twice that of mature second-growth forests on comparable sites. Wind is often a factor in tree fall, but trees also fall as a result of senescence, disease and insect damage. The fall of large trees such as the one shown often create canopy gaps that are quickly captured by tree reproduction. (USDA Forest Service, North Central Research Station photograph.)

achieved through periodic burning; (ii) the eradication of exotic species; (iii) proactively managing advanced second-growth forests for certain old-growth characteristics; or (4) expanding the effective area of existing old growth with silviculturally designed buffer zones. Strategies for applying these options are discussed below.

Existing old-growth forests For the few old-growth oak forests that remain, direct silvicultural manipulation will usually be inappropriate. Minimal human disturbance is, after all, a deﬁning characteristic of old-growth stands. Nevertheless, two exceptions to this general rule have been proposed (Parker, 1989; Guldin, 1991). The ﬁrst concerns the control of exotic species. In some cases, it may be feasible and appropriate to eliminate or control exotic plant species that have invaded an old-growth forest. However, a light-handed approach is usually preferable. Intensive cultural operations to remove exotic plants may do as much to degrade the old-growth character of a stand as the exotic species itself. The second exception concerns the role

of ﬁre in the development and maintenance of oak forests. The historical evidence points to ﬁre as a major disturbance factor in oak forests not only through the ﬁrst half of the 20th century, but for centuries earlier (Day, 1953; Little, 1974; Dorney, 1981; Pyne, 1982; Grimm, 1984; Lorimer, 1985; Chumbley et al., 1990; Ladd, 1991; Whitney, 1994; Guyette, 1995). Many old-growth oak forests originated when ﬁres occurred at a greater frequency than was the norm during much of the 20th century (Parker, 1989; Ladd, 1991; Guyette 1995; White and White, 1996). In many regions, these ﬁres were predominantly of human origin and recurred at intervals as short as 2–4 years. Should ﬁre therefore be considered a natural process associated with old-growth forests? And if that is so, at what frequency and intensity? Unless ﬁres are of sufﬁcient intensity to damage or kill a large proportion of overstorey trees, periodic ﬁres will neither move a forest towards nor away from the oldgrowth state. Fire may, however, alter the species composition of an old-growth forest. In the eastern United States, old-growth oak forests on mesic sites are usually successional to species of greater shade toler-

408

Chapter 9

ance such as sugar maple and beech (Parker et al., 1985; Schlesinger, 1989; Abrams, 1992; Shotola et al., 1992; Runkle, 1996; also see Chapters 3 and 5). On these sites, aggressive intolerant species such as yellow-poplar also can displace oaks by invading the naturally formed canopy gaps typical of an advanced old-growth state (Abrams and Downs, 1990). The reduction in ﬁre frequency during the last half of the 20th century coincided with a shift from dominance by oaks to other species in many old-growth forests, especially on mesic sites. Prescribed burning is a silvicultural tool that may increase the regeneration potential of oaks in mesic forests (Chapter 3) and in old-growth oak forests in particular (Parker, 1989; Guldin, 1991). Nevertheless, prescribed burning has seldom been used in old-growth forests for that purpose, presumably because of the risks involved. If an old-growth overstorey is severely damaged or destroyed, its recovery may take a century or more. Prudence dictates a cautious approach. There is little information to guide the detailed development of burning prescriptions to increase oak regeneration in oldgrowth forests. Given that all old-growth hardwood forests are exceedingly rare (whether dominated by oaks or by other species), there may be little basis to assume that an old-growth forest with high oak regeneration potential is of greater ecological or aesthetic value than one that is successional to other tree species. This may be especially true in landscapes comprised of a mosaic of stand ages and structural states that include mature second-growth oak stands.

Forests in transition to old-growth If old-growth oak forests are to become a more prominent feature of the landscape, a substantial acreage of existing [older] second-growth forests must be allowed to develop to the necessary age, size structure and species composition. To help meet this objective, more than 10% of second-growth forests in many national forests have been

set aside to develop old-growth characteristics. Although it is possible to design silvicultural treatments to accelerate the development of some old-growth characteristics, such treatments are rarely desirable. Selective girdling or felling (without removal of cut trees) theoretically could be used to accelerate the development of large overstorey trees and the accumulation of snags and down wood. However, tree deadening and the use of other silvicultural methods is at odds with the objective of minimizing human disturbance. Natural patterns of development eventually produce an old-growth stand without incurring the costs of silvicultural treatments. The issues surrounding prescribed burning in oak forests in transition to oldgrowth are essentially the same as those discussed earlier. Like many remnant oldgrowth forests, many second-growth oak forests are successional to shade tolerant species. Where this is the case, prescribed burning may be useful in maintaining oaks. The large existing acreage of secondgrowth oak forests in transition to oldgrowth provides opportunities for testing alternative ﬁre regimes for sustaining oaks in maturing forests.

Managing second-growth forests for oldgrowth characteristics Multiple management objectives such as producing timber while retaining oldgrowth features can be attained silviculturally with treatments designed to maintain or increase some of the characteristics outlined in Table 9.5. For example, unevenaged silviculture using single-tree or group selection methods can be used to mimic the formation of natural canopy gaps and obtain the subsequent within-gap recruitment of trees into the overstorey. In stands managed according to a negative exponential diameter distribution, the q value (Chapter 8) can be set fairly low (e.g. at 1.1–1.2 for 2 inch dbh classes) and the maximum tree size can be increased so that more large diameter trees are retained (Guldin, 1991; also see Chapter 8). Even though an old-growth oak forest may have

Silvicultural Methods for Multi-resource Management

ten or fewer trees per acre that are larger than 24 inches dbh, that is twice the number that typically occurs in a mature second-growth oak forest (McComb and Muller, 1983; Shiﬂey et al., 1995). Retention and/or creation of large snags will increase the amount of standing dead wood, and the number of den and nesting sites associated with dead standing trees (McComb and Muller, 1983; Allen and Corn, 1990).

Old-growth forests at the landscape scale At the landscape scale, the remaining oldgrowth oak stands provide valuable ecological models of the old-growth state. Although there are similarities between old-growth and mature second growth oak stands in terms of basal area, diameter distribution and number of snags, old-growth oak stands typically have greater volumes of down wood and larger mean and maximum tree diameters than second growth stands (Fig. 9.15). The species composition of old-growth stands also tends to differ from second-growth stands. For example, the proportion of long-lived white oaks and shade tolerant tree species is often greater in old-growth oak stands than in mature second-growth oak stands on similar sites (Muller, 1982; McComb and Muller, 1983; Shiﬂey et al., 1995). Where old-growth stands already exist, it may be feasible to expand their effective area by adding adjacent forest acreage. The additional acreage can be dedicated partially or exclusively to the development of old-growth characteristics. Expanding the size of old-growth oak forests is desirable for two reasons. First, most existing oldgrowth oak forests are small – usually less than a few hundred acres. Remnant oldgrowth oak forests largely occur as scattered small patches in landscapes dominated by second growth or non-forest lands. Increasing the total area of some tracts managed for old growth characteristics can increase landscape diversity. Second, larger old-growth tracts are better able to absorb large-scale natural distur-

409

bances (e.g. tornadoes) and still retain their character. Surrounding an old-growth core area with a buffer zone of second growth can protect the core area while extending its effective size. The second-growth buffer reduces edge effects at the margins of the old-growth and reduces the potential for invasion by plants adapted to edge environments. Such invasions can alter the composition of understorey vegetation inwards by 65 ft or more from the perimeter of the oldgrowth stand (Brothers and Spingarn, 1992). Modiﬁcation of understorey microclimate may extend up to 150 ft from edges to interiors of old-growth hardwood forests (Jacquart et al., 1992). In small stands, this edge effect can greatly reduce the effective area of the old growth core. Abrupt edges around old-growth nevertheless produce trade-offs. Edges usually produce greater densities and diversity of woody stems, and receive more solar radiation than forest interiors. The latter may be especially favourable to oak regeneration, in contrast to old-growth interiors where low light levels near the forest ﬂoor discourage the accumulation of oak reproduction, which in turn hastens the successional displacement of oaks (Brothers, 1993). In addition to protecting the old-growth core, an appropriately managed buffer zone can provide other beneﬁts including: • Suitable habitat for certain plant and animal species that utilize old-growth forests. • Temporary refuge for plants and animals after natural disturbances within the old-growth core. • A source of replacement for old-growth core areas lost to catastrophic disturbances. Many of these beneﬁts can be attained by conventional practices such as managing stands within the old-growth buffer zone on long, staggered, even-aged rotations (Hunter, 1990). Over time as stands in the buffer zone are harvested, the location of the oldest stands will shift although the total acreage by age classes remains constant.

410

Chapter 9

Edge contrast between the old-growth core and the buffer zone should be minimized. But within the second-growth buffer zone, the number, size, shape and harvest sequence of stands can be manipulated to alter the amount and type of forest edge (ecotones), to regulate age differences among adjacent stands and to create natural ﬁrebreaks (Harris, 1984; Hunter, 1990). The spatial arrangement of early- and latesuccessional stands in the buffer zone may signiﬁcantly affect the zone’s wildlife habitat qualities, and choices of buffer management alternatives often depend on wildlife management objectives. General principles and guidelines relevant to various wildlife objectives are presented by Hunter (1990) and Thompson et al. (1995). A buffer zone managed using unevenaged silviculture will often produce more similarity between the buffer and the oldgrowth core than a buffer managed using even-aged silviculture. The amount and type of edge in an uneven-aged buffer zone can be manipulated by controlling the size, number and spatial arrangement of harvested trees and timing of harvests (Guldin, 1991). Uneven-aged silviculture also eliminates large openings that can funnel damaging winds into the core oldgrowth area and maintains a continuous cover of high forest that minimizes edge effects around the perimeter of the oldgrowth core. Corridors of mature forest connecting old-growth tracts may increase ecosystem stability by providing plants and animals with pathways for migration (Harris, 1984). However, an intermingling of old growth with buffer zones and connective corridors may be a difﬁcult condition to obtain where land ownership is highly fragmented and land use is diverse. The concept of managing buffer zones adjacent to old-growth forests can be applied to most landscapes containing both old-growth and second-growth forests to: (i) protect and buffer existing old growth; and (ii) extend some old-growth characteristics into surrounding second-growth forests. The net result can extend the effective size of old-growth forests (Mladenoff et

al., 1994). But because landscape-scale management practices for old-growth forests are largely untested in practice, their application should be monitored for conformance with desired outcomes. Over time, ﬁeld practices can be modiﬁed as experience and ecosystem-speciﬁc conditions dictate.

Aesthetics The aesthetic consequences of forest management are important factors in contemporary silvicultural decision making. In some situations, minimizing negative visual impacts may be the primary consideration in selecting a silvicultural method. It has even been suggested that when people are present in a forest, their experiences are also a part of the ecosystem (Schroeder, 1996). Although the aesthetic qualities of a forest are most commonly perceived visually, the sounds, smells, temperatures and textures of the forest are also part of the experience. No two people entering a forest are likely to have the same aesthetic experience because of differences in their emotional states, prior experiences, personal sense of beauty and a host of other factors. To some, a forest may be perceived as a place of darkness or a physical or economic barrier, while others may perceive the same forest as a place of beauty and source of emotional or spiritual renewal. Forests often represent stability, and old forests can provide a physical and emotional link to earlier times and places. Because human perceptions of a forest vary with individual experiences and values, aesthetic responses to silvicultural treatments may: • Vary widely among individuals. • Be based on strongly held personal values. • Evoke strong emotions. • Be emotionally linked to issues or past incidents that are not readily apparent or intuitive (Schroeder, 1996). Even though aesthetic responses to silvicultural treatments vary among individu-

Silvicultural Methods for Multi-resource Management

als, there are certain silvicultural practices that are generally associated with increasing the aesthetic appeal of a managed forest. There are also silvicultural practices that are widely viewed as decreasing the aesthetic appeal of a forest. Some of these practices apply to individual stands; others involve the spatial arrangement and timing of silvicultural treatments across a landscape. Even though the aesthetic appeal of a forest can be explored through any of the senses, we consider only visual perceptions of forest beauty here. It is the only aesthetic interaction between humans and forests that to date has been investigated quantitatively, and it is the primary means through which most people interact with natural environments (Gobster, 1994).

Stand-level aesthetics Scenic beauty ratings have been used as a measure of the visual attractiveness of a forest. Asking observers to rate the relative attractiveness of various real or photographed forest settings can be used to derive scenic beauty ratings. The silvicultural practices used to create the observed forest conditions, the measured physical characteristics of the forest and other elements in each view are analysed to identify

411

the features that are associated with high or low scenic beauty ratings. Only a few such studies have been conducted speciﬁcally on oak and oak–pine forests, but some general trends are present in most studies (Table 9.6). Some silvicultural practices may actually increase aesthetic appeal. For example, intermediate cuttings are associated with increased scenic beauty except for the period immediately following timber harvesting (Rutherford and Shafer, 1969; Brush, 1979). Large amounts of dead and down wood on the forest ﬂoor of upland hardwood stands are associated with diminished scenic beauty (Vodak et al., 1985; Ribe, 1991). Likewise, scenic beauty decreases with increasing intensity of harvest from no harvest to intermediate thinnings to clearcuts (Vodak et al., 1985). Savannas with large scattered trees, long lines of sight and abundant ground ﬂora are consistently rated among the most aesthetically appealing settings (Balling and Falk, 1982). However, in any given setting the presence of roads, water, scenic vistas, geologic formations or people may overshadow the inﬂuence of vegetation structure and composition on perceptions of scenic beauty. The scenic quality of a stand changes over the course of stand development.

Table 9.6. Forest characteristics associated with high and low scenic beauty ratings.a Forest scenic beauty rating High • Large trees • Open forest understoreys allowing views deep into the forest • Herbaceous vegetation on the forest floor • Appearance of easy travel through the forest • Large openings surrounded by trees • Relatively old forests • Vegetation diversity • Vistas, overlooks, water bodies, unique geological formations • Flowering trees or other flowering vegetation a

Low • • • • • • • • • •

Visible logging slash or other evidence of logging Exposed soil Large volumes of down wood Height of dead wood (logging slash or natural) greater than 3 ft above ground Dead trees Any unhealthy appearance Dense, impenetrable understorey or mid-storey vegetation A predominance of small trees Closed views Unbounded forest openings

From Brush, 1979; Litton and McDonald, 1980; Schroeder and Daniel, 1980; Vodak et al., 1985; Rudis et al., 1988; Ribe, 1991; Gobster, 1994.

412

Chapter 9

Managers can plan for these changes in the same way as they plan for changes in timber volume over time. Simulation studies have been used to explore changes in scenic beauty over a rotation for loblolly pine managed under alternative silvicultural regimes (Hull and Buyhoff, 1986). Although there are no similar reported studies on oak forests, stand age and silvicultural treatments in general strongly inﬂuence aesthetic preferences. Unlike timber volume, however, scenic beauty is not a tangible commodity that can be transported for consumption elsewhere. Diminished scenic beauty in one stand will not necessarily be compensated by increased scenic beauty in another. For some people, the aesthetic value of a forest is intimately tied to its location (e.g. a forest that they regularly visit). In those situations, any change in the forest may be perceived negatively regardless of offsetting improvements in scenic beauty elsewhere. All timber harvesting tends to at least temporarily reduce the aesthetic appeal of a forest by creating slash, skid trails and log landings, and by causing some damage to the residual vegetation. However, some silvicultural practices produce fewer negative visual impacts than others (Table 9.7).

Although few studies have addressed the aesthetics of oak or oak–pine silviculture, ponderosa pine stands with a Gamble oak understorey were associated with high scenic beauty (Brown and Daniel, 1986). Similarly, when oaks were retained as a component of mature shortleaf pine stands, the stands were perceived as more attractive than those with no hardwoods (Gramann and Rudis, 1994). In general, retaining oaks and other hardwoods in predominantly conifer forests will add colour and variation in vegetation structure to the stand. These effects are especially apparent in autumn, winter and spring (McDonald and Huber, 1995). Silvicultural systems can be evaluated with respect to their impacts on scenic beauty (Table 9.8). For a given system, scenic beauty ratings are usually lowest immediately after timber harvesting. In some cases, managing visual impacts during a single timber harvest may be of paramount importance. In other cases it may be more important to consider changes in scenic beauty over an entire rotation. Some silvicultural practices result in frequent but relatively small reductions in scenic beauty during and shortly after harvest (e.g. singletree selection method). Other practices may

Table 9.7. Practices useful in reducing negative visual impacts of timber harvesting.a • Locate stand boundaries so they follow natural landscape features • Soften the contrast between harvest and non-harvest areas using partial harvests to provide a gradual transition from cut to uncut stands • Reduce the apparent size of large openings by retaining groups of trees or shrubs • Retain peninsulas of vegetation that extend from adjacent stands into harvested areas leaving evenly spaced residual trees • Retain vegetation with showy flowers • Retain conifers to increase visual diversity in hardwood forests, especially during the dormant season • Retain oaks and other hardwoods to increase visual diversity in predominantly conifer forests a

From Smith and Kuhr, 1989.

• Minimize the duration of harvest operations to reduce the duration of their visual impact • Harvest during dormant season to minimize the visibility of roads and skid trails and to minimize colour contrasts in vegetation among cut and uncut areas • Remove understorey vegetation to increase visual penetration into the stand • Select residual den trees or snags so they fall in clumps with other vegetation in the foreground or along edges of openings • Design roads, skid trails and landings to minimize the amount of area disturbed and to minimize visibility to other forest users • Lop, chip or scatter slash to reduce its height and visibility • Utilize tops and limbs for products to minimize volume of residue

Table 9.8. Silvicultural systems and their impacts on scenic beauty after timber harvesting.a

Positive impacts on scenic beauty

Negative impacts on scenic beauty

Clearcutting method

May open vistas. Can create visual variety in forest structure and species composition

Creates slash and disturbed soils that may be visible for 2 years or more. Woody and herbaceous reproduction forms a brushy mass that may be visually impenetrable for years and low in scenic beauty

Two-stage shelterwood method

The shelterwood stage may provide open, park-like appearance with aesthetic appeal. Retained overstorey softens the visual impact of the initial harvest. Advance growth of the understorey during the shelterwood stage helps mask the impact of the removal cut. Deferment cutting, which retains the overstorey indefinitely, may enhance aesthetic qualities. Can create visual variety in forest structure and species composition (Smith et al., 1989)

Logging activity, slash and soil disturbance occurs twice (i.e. shelterwood creation and shelterwood removal). Unsightly damage to residual vegetation (overstorey or understorey) may result at both harvests

Three-stage shelterwood method

Same as two-stage shelterwood but third stage leaves some residual overstorey present for a longer period (up to one-third of the rotation) and further reduces the visual impact of harvest. Can create visual variety in forest structure and species composition

Logging activity, slash and soil disturbance occur at three harvests. Unsightly damage to residual vegetation (overstorey or understorey) may result at each of the three harvests

Two-age management

Maintains continuous overstorey cover. Can create visual variety in forest structure and species composition. Residual forest cover can mitigate visual impacts of harvest

Logging activity, slash and soil disturbance occur twice per rotation. Unsightly damage to residual vegetation (overstorey or understorey) may result at each harvest

Seed tree method

Seed trees retained after initial harvest may soften the visual impact of harvesting. May open vistas. Can create visual variety in forest structure and species composition. Deferment cutting which retains the seed trees indefinitely can enhance the aesthetic qualities (Smith et al., 1989)

Logging activity, slash and soil disturbance occur at two harvests (i.e. seed tree retention and removal). Woody and herbaceous reproduction may form a brushy mass that is visually impenetrable for years and of low scenic beauty. The wide, regular spacing often desired for seed trees may be less attractive than other (e.g. clumped) arrangements of residual trees

Continued

Silvicultural Methods for Multi-resource Management

Silvicultural system

413

Table 9.8. Continued 414

Positive impacts on scenic beauty

Negative impacts on scenic beauty

Group selection

Maintains nearly continuous forest overstorey vegetation throughout cutting cycles. Residual vegetation mitigates visual impact. Group opening can provide visual variety in forest structure and species composition. Associated thinning between groups increases visual penetration into the stand and may even increase the scenic beauty relative to stands with no harvest (Rutherford and Shafer, 1969)

Low intensity logging activity occurs frequently (e.g. every decade). May require more roads or skid trails than other harvest methods or allow less opportunity for skid trails to recover between harvests. Group openings have the same characteristics as very small (e.g. up to 1 acre) clearcuts. May obscure scenic vistas

Single-tree selection

Maintains continuous forest overstorey vegetation throughout cutting cycles. Residual vegetation mitigates visual impact of harvest. Associated thinnings may increase visual penetration into the stand. Stands managed using selection harvest may be perceived as more attractive than stands with no harvest (Rutherford and Shafer, 1969)

Low intensity logging activity occurs frequently (e.g. every decade). May require more roads or skid trails than other harvest methods or allow less opportunity for skid trails to recover between harvests. Periodic logging results in soil disturbance and creation of slash. May obscure scenic vistas

Intermediate thinning

Creates views into forest interior. Accelerates development of large diameter trees. Low intensity disturbance relative to regeneration cuts. Thinned stands may be perceived as more attractive than stands with no harvest (Rutherford and Shafer, 1969)

Logging activity, slash and soil disturbance occur periodically. Unsightly damage to residual vegetation (overstorey or understorey) may result at the periodic harvests

No harvest

Develops large tree character. No logging activity, slash or soil disturbance. No periodic fluctuations in scenic beauty due to harvest

May accumulate relatively large volumes of snags and down wood or obscure scenic vistas

a For

each silvicultural system, the described impacts apply to years immediately following harvests except where otherwise noted. Silvicultural systems are listed in approximate order of decreasing impact on scenic beauty. In general, the aesthetic qualities of stands increase with time since harvest, regardless of the system used. (Adapted from Smith and Kuhr, 1989.)

Chapter 9

Silvicultural system

Silvicultural Methods for Multi-resource Management

produce infrequent but large, longer lasting reductions in scenic beauty (e.g. clearcutting). Table 9.8 presents silvicultural systems in approximate order of decreasing negative impact on scenic beauty after timber harvesting. However, human perceptions are complex. For example, a clearcut that opens a spectacular vista may evoke a more favourable aesthetic response than a single-tree selection harvest near a popular recreation area. A stocking chart provides a potentially useful tool for deﬁning regions of stand density and average tree size associated with different levels of aesthetic appeal. Although formal studies linking scenic beauty ratings with stand density and tree size are lacking, some general trends in scenic preferences can be organized on stocking charts. Lowest levels of aesthetic appeal are generally associated with dense stands of small trees. From there, aesthetic appeal increases with increasing average tree size (Fig. 9.16).

Landscape-level aesthetics Landscapes comprising thousands of acres with large vistas (‘viewsheds’), travel corridors and recreation areas present additional aesthetic concerns. For example, a travel corridor ﬂanked by a continuous mature and unharvested forest may have less aesthetic appeal than a corridor containing occasional forest openings that provide a more varied roadside view (Schroeder and Daniel, 1980). Visual variety is known to be aesthetically appealing. Consequently, maximizing the scenic beauty of individual stands does not necessarily maximize the scenic beauty of the landscape, especially if the result is an unvaried landscape. Managing landscapes and vistas for aesthetic appeal typically extends beyond the purview of the silviculturist and into that of the landscape architect. Forest managers nevertheless should be aware of the potential impacts of silvicultural treatments on landscape vistas and methods for increasing the aesthetic appeal of silviculturally

415

treated areas when viewed from a distance. The following guidelines can be applied where the aesthetic appeal of landscape views and travel corridors are important (Smith and Kuhr, 1989; Palmer et al., 1995): • Create landscapes with visual variety; they usually have great appeal. That variety can come from vegetation but may be substantially enhanced by natural landforms, geologic features, water bodies and open spaces. • Create mosaics of vegetation that contain different size classes and densities of trees and other plants. • Concentrate efforts on viewsheds visible from travel corridors, recreation sites or other high-use areas. • Concentrate efforts on the foreground of a viewshed where visibility and potential impact are greater than in the viewshed background. • Design openings that are varied in shape and that have boundaries that follow natural topographic features. • Avoid openings that cross ridgelines; they visually emphasize differences between openings and adjacent areas. • Limit the total area of the landscape view that is harvested. • Retain conifers in areas that are predominantly hardwood and hardwoods in areas that are predominantly conifer to increase visual variety during the dormant season. • Vary the size of forest openings. • Select opening sizes that are appropriate for the landscape view. One study found that scenic beauty ratings were higher when the total harvested area in the middle ground of the scene was divided into cutting units of 10–14 acres in size rather than ones that were smaller (4–5 acres) or larger (20–30 acres) (Palmer et al., 1995). Management of forested landscapes for aesthetic values thus requires an awareness of the different kinds of human reactions to not only the stages of stand development, condition and spatial arrangement, but to their larger landscape setting.

416

Chapter 9 7

110

Av e

rag

6

et

A

100

ree

5

dia

me

ter

(in

ch

es

)

rate

4

80

Mo de

70 60

3 110

Very Low

Low

10

0

90

B 80

50

60

ge

St

40

oc kin g

50 30

ta rc en

40

70

C

pe

Basal area per acre (ft2)

90

20 1

3

5

7

9

11

13

15

Hundred trees per acre

15

140 130

A

h

7 110

Hig

110

10

0

h

ate Mo

80

B

en

ta

80

50

50

kin

60

St oc

C

g

pe

rc

70

70 60

90

der

90

ge

100

Hig

Ver y

Basal area per acre (ft2)

120

Av era ge 13 tre 2 1 ed 1 iam 1 ete 10 r (i nch 9 es) 8

40 50

100 150 200 250 300 350 400 Trees per acre

Fig. 9.16. Inferred relations between scenic preference (very low to very high) and stand stocking based on Gingrich’s stocking chart for upland oak stands in the Central Hardwood Region and other sources. Open stands with large diameter trees generally have higher scenic beauty than stands crowded with small tress. However, many other factors affect the perceived scenic beauty of a given site.

Silvicultural Methods for Multi-resource Management

417

References Abrams, M.D. (1992) Fire and the development of oak forests. BioScience 42, 346–353. Abrams, M.D. and Downs, J.A. (1990) Successional replacement of old-growth white oak by mixed mesophytic hardwoods in southwestern Pennsylvania. Canadian Journal of Forest Research 20, 1864–1870. Adams, T.E., Jr, Sands, P.B., Weitkamp, W.H. and McDougald, N.K. (1991) Blue and valley oak seedling establishment on California’s hardwood rangelands. USDA Forest Service General Technical Report PSW PSW-126, pp. 41–47. Allen, A.W. and Corn, J.G. (1990) Relationships between live tree diameter and cavity abundance in Missouri oak–hickory forest. Northern Journal of Applied Forestry 7, 179–183. Ambuel, B. and Temple, S.A. (1983) Area-dependent changes in the bird communities and vegetation of southern Wisconsin forests. Ecology 64, 1057–1068. Anderson, R.C. and Brown, L.E. (1986) Stability and instability in plant communities following ﬁre. American Journal of Botany 73, 364–368. Auchmoody, L.R., Smith, H.C. and Walters, R.S. (1993) Acorn production in northern red oak stands in northwestern Pennsylvania. USDA Forest Service General Technical Report NE NE-680. Balling, J.D. and Falk, J.H. (1982) Development of visual preference for natural environments. Environment and Behavior 14, 5–28. Barnhardt, S.J., McBride, J.R., Cicero, C., da Silva, P. and Warner, P. (1987) Vegetation dynamics of the northern oak woodland. USDA Forest Service General Technical Report PSW PSW-100, pp. 53–58. Barnhardt, S.J., McBride, J.R. and Warner, P. (1991) Oak seedling establishment in relation to environmental factors at Annadel State Park. USDA Forest Service General Technical Report PSW PSW-126, pp. 25–30. Bartolome, J.W., Muick, P.C. and McClaran, M.T. (1987) Natural regeneration of Californian hardwoods. USDA Forest Service General Technical Report PSW PSW-100, pp. 26–31. Beck, D.E. (1977) Twelve-year acorn yield in southern Appalachian oaks. USDA Forest Service Research Note SE SE-244. Beilmann, A.P. and Brenner, L.G. (1951) The recent intrusion of forests in the Ozarks. Annals of the Missouri Botanical Garden 38, 263–281. Biswell, H.H. (1989) Prescribed Burning in California Wildlands Vegetation Management. University of California Press, Berkeley. Bolsinger, C.L. (1988) The hardwoods of California’s timberlands, woodlands, and savannas. USDA Forest Service Resource Bulletin PSW PSW-148. Bowles, M.L. and McBride, J.L. (1998) Vegetation composition, structure, and chronological change in a decadent Midwestern North American savanna remnant. Natural Areas Journal 18, 14–27. Bray, J.R. (1960) The composition of savanna vegetation in Wisconsin. Ecology 41, 721–732. Brendel, F. (1887) Flora Peoriana: The Vegetation in the Climate of Middle Illinois. J.W. Frank and Sons, Peoria, Illinois. Brose, P.H. and Van Lear, D.H. (1997) Responses of hardwood advance regeneration to seasonal prescribed ﬁres in oak-dominated shelterwood stands. Canadian Journal of Forest Research 28, 331–339. Brothers, T.S. (1993) Fragmentation and edge effects in central Indiana old-growth forests. Natural Areas Journal 13, 268–275. Brothers, T.S. and Spingarn, A. (1992) Forest fragmentation and alien plant invasion of central Indiana old-growth forests. Conservation Biology 6, 91–99. Brown, J.H., Jr (1960) The role of ﬁre in altering the species composition of forests in Rhode Island. Ecology 41, 310–316. Brown, T.C. and Daniel, T.C. (1986) Predicting scenic beauty of timber stands. Forest Science 32, 471–487. Brush, R.O. (1979) The attractiveness of woodlands: perceptions of forest landowners in Massachusetts. Forest Science 25, 495–506. Burger, G.V., Ebinger, J.E. and Wilhelm, G.S. (eds) (1991) Proceedings Oak Woods Management Workshop. East Illinois University, Charleston. Burns, R.M. and Honkala, B.H. (tech. coords.) (1990) Silvics of North America Vol. 2, Hardwoods. USDA Forest Service Agriculture Handbook 654.

418

Chapter 9

Cecich, R.W. and Larsen, D.R. (1997) Bud demography and bifurcation in Quercus alba L. and Quercus velutina Lam. Proceedings 2nd Working Party 2.08.05, Genetics of Quercus, Diversity and Adaption in Oak Species. IUFRO, pp. 68–75. Cecich, R.A. and Sullivan, N.H. (1999) Inﬂuence of weather at time of pollination on acorn production of Quercus alba and Q. velutina. Canadian Journal of Forest Research 29, 1817–1823. Christensen, E.M. (1955) Ecological notes on the mountain brush in Utah. Proceedings Utah Academy of Science, Arts, and Letters 32, 107–111. Christisen, D.M. and Kearby, W.H. (1984) Mast Measurement and Production in Missouri (with special reference to acorns). Missouri Department of Conservation Terrestrial Series 13. Chumbley, C.A., Baker, R.G. and Bettis, E.A. III. (1990) Midwestern holocene paleoenvironments revealed by ﬂoodplain deposits in northeastern Iowa. Science 249, 272–274. Conrad, C.E. and Oechel, W.C. (tech. coords.) (1982) Proceedings of Symposium on Dynamics and Management of Mediterranean-Type Ecosystems. USDA Forest Service General Technical Report PSW PSW-58. Cooper, C.F. (1961) The ecology of ﬁre. Scientiﬁc American 204, 150–160. Crow, T.R. (1990) Conservation biology and landscape ecology: new perspectives for resource managers. Proceedings 1990 Society of American Forestry National Convention, pp. 47–54. Curtis, J.T. (1959) The Vegetation of Wisconsin. University of Wisconsin Press, Madison. Cypert, E. and Webster, B.S. (1948) Yield and use by wildlife of acorns of water and willow oaks. Journal of Wildlife Management 12, 227–231. Daniels, F.P. (1904) Ecology of the ﬂora of Sturgis, Michigan and vicinity. Report of Michigan Academy of Science, Arts and Letters 4, 145–159. Danielsen, K.C. and Halvorson, W.L. (1991) Valley oak seedling growth associated with selected grass species. USDA Forest Service General Technical Report PSW PSW-126, pp. 9–13. Davis, F.W., Borchert, M., Harvey, L.E. and Michaelsen, J.C. (1991) Factors affecting seedling survivorship of blue oak (Quercus douglasii H. & A.) in Central California. USDA Forest Service General Technical Report PSW PSW-126, pp. 81–86. Davis, M.B. (1996) Eastern Old-Growth Forests. Island Press, Washington, DC. Day, G.M. (1953) The Indian as an ecological factor in the northeastern forest. Ecology 34, 329–346. Dorney, J.R. (1981) The impact of Native Americans on presettlement vegetation in southeastern Wisconsin. Transactions of Wisconsin Academy of Science, Arts and Letters 69, 26–36. Downs, A.A. (1944) Estimating acorn crops for wildlife in the southern Appalachians. Journal of Wildlife Management 8, 339–340. Ek, A.R. (1974) Dimensional relationships of forest and open grown trees in Wisconsin. University of Wisconsin Forestry Research Note July. Evans, K.E. and Kirkman, R.A. (1981) Guide to bird habitats of the Ozark Plateau. USDA Forest Service General Technical Report NC NC-68. Evans, K.E. and Probasco, G.E. (1977) Wildlife of the prairies and plains. USDA Forest Service General Technical Report NC NC-29. Eyre, F.H. (ed.) (1980) Forest Cover Types of the United States and Canada. Society of American Foresters, Washington, DC. Feret, P.P., Kreh, R.E., Merkle, S.A. and Oderwald, R.G. (1982) Flower abundance, premature acorn abscission, and acorn production in Quercus alba L. Botanical Gazette (Chicago) 143, 216–218. Ffolliott, P.F., Gottfried, G.J., Bennett, D.A., Hernandez, V.M., Ortega-Rubio, A. and Hamre, R.H. (tech. coords.) (1992) Ecology and management of oak and associated woodlands: perspectives in the southwestern United States and northern Mexico. USDA Forest Service General Technical Report RM RM-218. Fralish, J. (comp.) (1994) Living in the edge. Proceedings of the North American Conference on Savanna and Barrens. US Environmental Protection Agency. (Electronically available via internet: http://www.epa.gov/glnpo/oak/oak94/index.html) Gingrich, S.F. (1967) Measuring and evaluating stocking and stand density in upland hardwood forests in the Central States. Forest Science 13, 38–53. Gobster, P.H. (1994) The aesthetic experience of sustainable forest ecosystems. USDA Forest Service General Technical Report RM RM-247, pp. 246–255. Goodrum, P.D., Reid, V.H. and Boyd, C.E. (1971) Acorn yields, characteristics, and management criteria of oaks for wildlife. Journal of Wildlife Management 35, 520–532. Gramann, J.H. and Rudis, V.A. (1994) Effects of hardwood retention, season of year, and landform on

Silvicultural Methods for Multi-resource Management

419

the perceived scenic beauty of forest plots in the Ouachita Mountains. USDA Forest Service General Technical Report SO SO-112, pp. 223–228. Graves, W.C. (1980) Annual oak mast yields from visual estimates. USDA Forest Service General Technical Report PSW PSW-44, pp. 270–274. Green, L.R. (1980) Prescribed burning in California oak management. USDA Forest Service General Technical Report PSW PSW-44, pp. 136–142. Grifﬁn, J.R. (1976) Regeneration in Quercus lobata savannas, Santa Lucia Mountains, California. American Midland Naturalist 95, 422–435. Grifﬁn, J.R. (1977) Oak woodland. In: Barbour, M.G. and Major, J. (eds) Terrestrial Vegetation of California. John Wiley & Sons, New York, pp. 383–415. Grimm, E.C. (1983) Chronology and dynamics of vegetation change in the prairie–woodland region of southern Minnesota, USA. New Phytologist 93, 311–350. Grimm, E.C. (1984) Fire and other factors controlling the Big Woods vegetation of Minnesota in the mid-nineteenth century. Ecological Monographs 54, 291–311. Guldin, J.M. (1991) Silvicultural practices applied to old-growth stand management. Proceedings of Restoration of Old Growth Forests in the Interior Highlands of Arkansas and Oklahoma Conference. Winrock International, Morrilton, AR, pp. 171–190. Guyette, R.P. (1995) A History of Wildland Fire in the Current River Watershed. University of Missouri School of Natural Resources, Columbia. Guyette, R.P. and Cutter, B.E. (1991) Tree-ring analysis of ﬁre history of a post oak savanna in the Missouri Ozarks. Natural Areas Journal 11, 93–99. Gysel, L.W. (1956) Measurement of acorn crops. Forest Science 2, 305–313. Gysel, L.W. (1957) Acorn production on good, medium, and poor oak sites in southern Michigan. Journal of Forestry 55, 570–574. Haney, A. and Apfelbaum, S.I. (1990) Structure and dynamics of midwest oak savannas. In: Sweeney, J.M. (ed.) Management of Dynamic Ecosystems. The Wildlife Society, North Central Section, West Lafayette, Indiana, pp. 19–30. Hansen, A.J., Spies, T.A., Swanson, F.J. and Ohmann, J.L. (1991) Conserving biodiversity in managed forests. BioScience 41, 382–392. Harlow, R.F. and Eikum, R.L. (1963) The effect of stand density on the acorn production of turkey oaks. Proceedings of the Annual Conference of the Southeastern Association of Game and Fish Commissioners 16, pp. 126–133. Harris, L.D. (1984) The Fragmented Forest. University of Chicago Press, Chicago, Illinois. Healy, W.M. (1997) Thinning New England oak stands to enhance acorn production. Northern Journal of Applied Forestry 14, 152–156. Henderson, N.R. and Long, J.N. (1984) A comparison of stand structure and ﬁre history in two black oak woodlands in northwestern Indiana. Botanical Gazette 145, 222–228. Hicks, R.R., Jr (1997) A resource at the crossroads: a history of the central hardwoods. USDA Forest Service General Technical Report NC NC-188, pp. 1–22. Hull, R.B. and Buhyoff, G.J. (1986) The scenic beauty temporal distribution method: an attempt to make scenic beauty assessments compatible with forest planning methods. Forest Science 32, 271–286. Hunter, M.L., Jr (1989) What constitutes an old-growth stand? Journal of Forestry 87, 33–35. Hunter, M.L., Jr (1990) Wildlife, Forests and Forestry. Regents/Prentice Hall, Englewood Cliffs, New Jersey. Huntley, J.C. (1983) Squirrel den tree management: reducing incompatibility with timber production in upland hardwoods. USDA Forest Service General Technical Report SE SE-24, pp. 488–495. Irving, F.D. (1971) Field instruction in prescribed burning techniques at the University of Minnesota. Proceedings of the 10th Tall Timbers Fire Ecology Conference, pp. 323–331. Jacquart, E.M., Armentano, T.V. and Spingarn, A.L. (1992) Spatial and temporal tree responses to water stress in an old-growth deciduous forest. American Midland Naturalist 127, 158–171. Jenkins, S.E. (1997) Spatial demography of an Ozark savanna. PhD dissertation, University of Missouri, Columbia. Johnson, P.S. (1979) Shoot elongation of black oak and white oak sprouts. Canadian Journal of Forest Research 9, 489–494. Johnson, P.S. (1993) Perspectives on the ecology and silviculture of oak-dominated forests in the central and eastern states. USDA Forest Service General Technical Report NC NC-153.

420

Chapter 9

Kittredge, J. and Chittenden, A.K. (1929) Oak forests of northern Michigan. USDA Forest Service Lake States Forest Experiment Station Special Bulletin 190. Koenig, W.D., Knops, J.M.H., Carmen, W.J., Stanback, M.T. and Mumme, R.L. (1994) Estimating acorn crops using visual surveys. Canadian Journal of Forest Research 24, 2105–2112. Komarek, E.V. (1974) Effects of ﬁre on temperate forests and related ecosystems: southeastern United States. In: Kozlowski, T.T. and Ahlgren, C.E. (eds) Fire and Ecosystems. Academic Press, New York, pp. 251–277. Krajicek, J.E. (1967) Maximum use of minimum acres. Proceedings of the 9th Southern Tree Improvement Conference, pp. 35–37. Krajicek, J.E., Brinkman, K.A. and Gingrich, S.F. (1961) Crown competition – measure of density. Forest Science 7, 35–42. Ladd, D. (1991) Reexamination of the role of ﬁre in Missouri oak woodlands. Proceedings Oak Woods Management Workshop. East Illinois University, Charleston, pp. 67–80. Larsen, D.R. and Cecich, R.A. (1997) Model of white oak ﬂower survival and maturation. USDA Forest Service General Technical Report NC NC-188, pp. 262–268. Law, J.R., Johnson, P.S. and Houf, G. (1994) A crown cover chart for oak savannas. USDA Forest Service Technical Brief TB-NC TB-NC-2. Leach, M.K. and Ross, L. (eds) (1995) Proceedings 1995 Midwest Oak Savanna and Woodland Ecosystems Conference. US Environmental Protection Agency. (Electronically available via internet: http://www.epa.gov/glnpo/oak/oak95/) Leverett, R. (1996) Deﬁnitions and history. In: Davis, M.B. (ed.) Eastern Old-Growth Forests. Island Press, Washington, DC. pp. 3–17. Little, S. (1974) Effects of ﬁre on temperate forests: northeastern United States. In: Kozlowski, T.T. and Ahlgren, C.E. (eds) Fire and Ecosystems. Academic Press, New York, pp. 225–250. Litton, R.B., Jr and McDonald, P.M. (1980) Silviculture and visual resources. Proceedings 1979 Society American Foresters National Convention, pp. 97–102. Lorimer, C.G. (1985) The role of ﬁre in the perpetuation of oak forests. Proceedings Challenges in Oak Management and Utilization. University of Wisconsin, Madison, pp. 8–25. Martin, W.H. (1992) Characteristics of old-growth mixed mesophytic forests. Natural Areas Journal 12, 127–135. McClain, W.E., Jenkins, M.A., Jenkins, S.E. and Ebinger, J.E. (1993) Changes in the woody vegetation of a bur oak savanna remnant in central Illinois. Natural Areas Journal 13, 108–114. McClaran, M.P. and Bartolome, J.W. (1989) Fire-related recruitment in stagnant Quercus douglasii populations. Canadian Journal of Forest Research 19, 580–585. McComb, W.C. and Muller, R.N. (1983) Snag densities in old-growth and second-growth Appalachian forests. Journal of Wildlife Management 47, 376–382. McDonald, P.M. (1992) Estimating seed crops of conifer and hardwood species. Canadian Journal of Forest Research 22, 832–838. McDonald, P.M. and Huber, D.W. (1995) California’s hardwood resource: managing for wildlife, water, pleasing scenery, and wood products. USDA Forest Service General Technical Report PSW PSW-154. McGill, D., Martin, J., Rogers, R. and Johnson, P.S. (1991) New stocking charts for northern red oak. University of Wisconsin Forestry Research Notes 277. McQuilkin, R.A. and Musbach, R.A. (1977) Pin oak acorn production on Green Tree Reservoirs in southeastern Missouri. Journal of Wildlife Management 41, 218–225. Mead, S.B. (1846) Catalogue of plants growing spontaneously in the state of Illinois, the principal part near Augusta, Hancock County. Prairie Farmer 6, 35–36, 60, 93, 199–122. Meyer, J. (1986) Management of old growth forests in Missouri. Missouri Department of Conservation Habitat Management Series 3. Minckler, L.S. and Janes, D. (1965) Pin oak acorn production on normal and ﬂooded areas. University of Missouri Agricultural Experiment Station Research Bulletin 898. Minckler, L.S. and McDermott, R.E. (1960) Pin oak acorn production and regeneration as affected by stand density, structure and ﬂooding. University of Missouri Agricultural Experiment Station Research Bulletin 750. Mladenoff, D.J. and He, H. (1999) Design and behavior of LANDIS: an object-oriented model of forest landscape disturbance and succession. In: Mladenoff, D.J. and Baker, W.L. (eds) Spatial Modeling of Forest Landscape Change: Approaches and Applications. Cambridge University Press, Cambridge, UK, pp. 125–162.

Silvicultural Methods for Multi-resource Management

421

Mladenoff, D.J., White, M.A., Crow, T.R. and Pastor, J. (1994) Applying principles of landscape management to integrate old-growth forest enhancement and commodity use. Conservation Biology 8, 752–764. Moody, R.D., Collins, J.O. and Reid, V.H. (1954) Oak production study under way. Louisiana Conservationist 6(9), 6–8. Mossman, M.J. and Lange, K.I. (1982) Breeding Birds of the Baraboo Hills, Wisconsin: Their History, Distribution, and Ecology. Wisconsin Department of Natural Resources and Wisconsin Society for Ornithology, Madison. Muller, R.N. (1982) Vegetation patterns in the mixed mesophytic forest of eastern Kentucky. Ecology 63, 1901–1917. Myers, S.A. (1978) Insect impact on acorn production in Missouri upland forests. PhD dissertation, University of Missouri, Columbia. Niering, W.A., Goodwin, R.H. and Taylor, S. (1970) Prescribed burning in southern New England: introduction to long-range studies. Proceedings of the 10th Tall Timbers Fire Ecology Conference, pp. 267–286. Nowacki, G.J. and Trianosky, P.A. (1993) Literature on old-growth forests of the eastern North America. Natural Areas Journal 13, 87–105. Nuzzo, V.A. (1986) Extent and status of Midwest oak savanna: presettlement and 1985. Natural Areas Journal 6, 6–36. Olson, D.F. Jr (1974) Quercus L. Oak. USDA Forest Service Agriculture Handbook 450, 88 pp. Packard, S. (1991) Rediscovering the tallgrass savanna of Illinois. Proceedings of the Oak Woods Management Workshop. East Illinois University, Charleston, pp. 55–66. Palmer, J.F., Shannon, S., Harrilchak, M.A., Gobster, P.H. and Kokx, T. (1995) Esthetics of clearcutting: alternatives in the White Mountain National Forest. Journal of Forestry 93(5), 37–42. Parker, G.R. (1989) Old-growth forests of the central hardwood region. Natural Areas Journal 9, 5–11. Parker, G.R., Leopold, D.J. and Eichenberger, J.K. (1985) Tree dynamics in an old-growth, deciduous forest. Forest Ecology and Management 11, 31–57. Peet, R.K. and Loucks, O.L. (1977) A gradient analysis of southern Wisconsin forests. Ecology 58, 485–499. Phillips, R.L., McDougald, N.K., Standiford, R.B., McCreary, D.D. and Frost, W.E. (1997) Blue oak regeneration in southern Sierra Nevada foothills. USDA Forest Service General Technical Report PSW PSW-160, pp. 177–181. Pillsbury, N.H., Verner, J. and Tietje, W.D. (tech. coords.) (1997) Proceedings Symposium on Oak Woodlands: Ecology, Management, and Urban Interface Issues. USDA Forest Service General Technical Report PSW PSW-160. Plumb, T.R. (tech. coord.) (1980) Proceedings Symposium on the Ecology, Management, and Utilization of California Oaks. USDA Forest Service General Technical Report PSW PSW-44. Plumb, T.R. and McDonald, P.M. (1981) Oak management in California. USDA Forest Service General Technical Report PSW PSW-54. Plumb, T.R. and Pillsbury, N.H. (tech. coords.) (1987) Proceedings Symposium on Multiple-Use Management of California’s Hardwood Resources. USDA Forest Service General Technical Report PSW PSW-100. Post, L.S. (1998) Seed management in Tennessee: development of seed zones for Tennessee and distribution and protection of northern red oak (Quercus rubra L.) acorns. MS thesis, University of Tennessee, Knoxville. Pyne, S.J. (1982) Fire in America. Princeton University Press, Princeton, New Jersey. Rebertus, A.J. and Burns, B.R. (1997) The importance of gap processes in the development and maintenance of oak savannas and dry forests. Journal of Ecology 85, 635–645. Rebertus, A.J., Williamson, G.B., Platt, W.J. and Glitzenstein, J.S. (1993) Impacts of temporal variation in ﬁre regime on savanna oaks and pines. Proceedings 18th Tall Timbers Fire Ecology Conference, pp. 215–225. Reed, L.J. and Sugihara, N G. (1987) Northern oak woodlands – ecosystem in jeopardy or is it already too late? USDA Forest Service General Technical Report PSW PSW-100, pp. 59–63. Ribe, R. (1991) The scenic impact of key forest attributes and long-term management alternatives for hardwood forests. USDA Forest Service General Technical Report NE NE-148, pp. 34–54. Roovers, L.M. and Shiﬂey, S.R. (1997) Composition and dynamics of Spitler Woods, an old-growth remnant forest in Illinois (USA). Natural Areas Journal 17, 219–232.

422

Chapter 9

Rossi, R.S. (1980) History of cultural inﬂuences on the distribution and reproduction of oaks in California. USDA Forest Service General Technical Report PSW PSW-44, pp. 7–18. Rudis, V.A., Gramann, J.H., Ruddell, E.J. and Westphal, J.M. (1988) Forest inventory and management-based visual preference models of southern pine stands. Forest Science 34, 846–863. Runkle, J.R. (1996) Central mesophytic forests. In: Davis, M.B. (ed.) Eastern Old-Growth Forests. Island Press, Washington, DC, pp. 161–177. Rutherford, W., Jr and Shafer, E.L., Jr (1969) Selection cuts increased natural beauty in two Adirondack forest stands. Journal of Forestry 67, 415–419. Schlesinger, R.C. (1989) Dynamics of the sugar male component of a white oak yellow-poplar community. USDA Forest Service General Technical Report NC NC-132, pp. 262–266. Schnur, G.L. (1937) Yield, stand, and volume tables for even-aged upland oak forests. USDA Technical Bulletin 560. Schroeder, H. (1996) Ecology of the heart: restoring and sustaining the human experience of ecosystems. In: Ewart, A.W. (ed.) Natural Resources Management: the Human Dimension. Westview Press, Boulder, Colorado, pp. 13–27. Schroeder, H. and Daniel, T.C. (1980) Predicting the scenic quality of forest road corridors. Environment and Behavior 12, 349–366. Sharp, W.M. (1958) Evaluating mast yields in the oaks. Pennsylvania State University Agricultural Experiment Station Bulletin 635. Sharp, W.M. and Sprague, V.G. (1967) Flowering and fruiting in the white oaks. Pistillate ﬂowering, acorn development, weather, and yields. Ecology 48, 243–251. Shiﬂey, S.R. (1994) Old-growth forests in the Central States: observations on a unique resource. Proceedings 1993 Society of American Foresters National Convention, pp. 527–528. Shiﬂey, S.R., Roovers, L.M. and Brookshire, B.L. (1995) Structural and compositional differences between old-growth and mature second-growth forests in the Missouri Ozarks. USDA Forest Service General Technical Report NE NE-197, pp. 23–36. Shiﬂey, S.R., Brookshire, B.L., Larsen, D.R. and Herbeck, L.A. (1997) Snags and down wood in Missouri old-growth and mature second-growth forests. Northern Journal of Applied Forestry 14, 165–172. Shiﬂey, S.R., Thompson, F.R., III, Larsen, D.R. and Dijak, W.D. (2000) Modeling forest landscape change in the Missouri Ozarks under alternative management practices. Computers and Electronics in Agriculture 27, 7–24. Shotola, S.J., Weaver, G.T., Robertson, P.A. and Ashby, W.C. (1992) Sugar maple invasion of an oldgrowth oak–hickory forest in southwestern Illinois. American Midlands Naturalist 127, 125–138. Smith, D.M. (1986) The Practice of Silviculture, 8th edn. John Wiley & Sons, New York. Smith, H.C. and Gibbs, C.B. (1970) A guide to sugarbush stocking. USDA Forest Service Research Paper NE NE-171. Smith, H.C., Lamson, N.I. and Miller, G.W. (1989) An esthetic alternative to clearcutting? Journal of Forestry 87(3), 14–18. Smith, J.E. and Kuhr, J.F. (1989) Visual management of Central Hardwoods. USDA Forest Service Central Hardwood Notes, Note 10.03. Spetich, M.A. (1995) Characteristics and spatial pattern of old-growth forests in the Midwest. PhD dissertation, Purdue University, West Lafayette, Indiana. Spetich, M.A., Shiﬂey, S.R. and Parker, G.R. (1999) Regional distribution and dynamics of coarse woody debris in midwestern old-growth forests. Forest Science 45, 302–313. Stahle, D.W. and Chaney, P.L. (1994) A predictive model for the location of ancient forests. Natural Areas Journal 14, 151–158. Standiford, R.B. (tech. coord.) (1991) Proceedings Symposium on Oak Woodlands and Hardwood Rangeland Management. USDA Forest Service General Technical Report PSW PSW-126. Stearns, F. and Holland, K. (eds) (1995) Proceedings 1993 Midwest Oak Savanna Conference. US Environmental Protection Agency. (Electronically available via internet: http://www.epa.gov/ glnpo/oak/oak93/) Streng, D.R. and Harcombe, P.A. (1982) Why don’t East Texas savannas grow up to forest? American Midland Naturalist 108, 278–294. Sullivan, N.H. (2001) An algorithm for a landscape model of mast production. PhD dissertation, University of Missouri, Columbia.

Silvicultural Methods for Multi-resource Management

423

Swiecki, T.J., Bernhardt, E.A. and Drake, C. (1997a) Stand-level status of blue oak sapling recruitment and regeneration. USDA Forest Service General Technical Report PSW PSW-160, pp. 147–156. Swiecki, T.J., Bernhardt, E.A. and Drake, C. (1997b) Factors affecting blue oak sapling recruitment. USDA Forest Service General Technical Report PSW PSW-160, pp. 157–167. Tester, J.R. (1989) Effects of ﬁre frequency on oak savanna in east-central Minnesota. Bulletin of the Torrey Botanical Club 116, 134–144. Thilenius, J.F. (1968) The Quercus garryana forests of the Willamette Valley, Oregon. Ecology 49, 1124–1133. Thompson, F.R., III, Probst, J.R. and Raphael, M.G. (1995) Impacts of silviculture: overview of management recommendations. In: Martin, T.E. and Finch. D.M. (eds) Ecology and Management of Neotropical Migratory Birds. Oxford University Press, New York, pp. 201–219. Thor, E. and Nichols, G.M. (1973) Some effects of ﬁres on litter, soil, and hardwood regeneration. Proceedings 13th Tall Timbers Fire Ecology Conference, pp. 317–329. Tryon, E.H. and Carvell, K.L. (1962) Acorn production and damage. West Virginia University Agricultural Experiment Station Bulletin 466T. Tyrrell, L.E., Nowacki, G.J., Crow, T.R., Buckley, D.S., Nauertz, E.A., Niese, J.N., Rollinger, J.L. and Zasada, J.C. (1998) Information about old growth for selected forest type groups in the eastern United States. USDA Forest Service General Technical Report NC NC-197. US Environmental Protection Agency (1995) Proceedings Midwest Oak Savanna and Woodland Ecosystems Conference. US Environmental Protection Agency. (Electronically available via internet: http://www.epa.gov/glnpo/oak/oak95/) Van Lear, D.H. and Waldrop, T.A. (1988) Effects of ﬁre on natural regeneration in the Appalachian Mountains. Society of American Foresters Publication 88-03, pp. 56–70. Verme, L.J. (1953) Production and utilization of acorns in Clinton County, Michigan. MS thesis, Michigan State University, East Lansing. Vodak, M.C., Roberts, P.L., Wellman, J.D. and Buhyoff, G.J. (1985) Scenic impacts of eastern hardwood management. Forest Science 31, 289–301. White, A.S. (1983) The effects of thirteen years of annual prescribed burning on a Quercus ellipsoidalis community in Minnesota. Ecology 64, 1081–1085. White, A.S. (1986) Prescribed burning for oak savanna restoration in central Minnesota. USDA Forest Service Research Paper NC NC-266. White, D.H. (1991) Legal implications associated with use and control of ﬁre as a management practice. Proceedings of the 17th Tall Timbers Ecology Conference, pp. 375–384. White, K.L. (1966) Structure and composition of foothill woodland in central Coastal California. Ecology 47, 229–237. White, P.S. and White R.D. (1996) Old-growth oak and oak–hickory forests. In: Eastern Old-Growth Forests. Island Press, Washington, DC, pp. 178–198. Whitney, G.G. (1994) From Coastal Wilderness to Fruited Plain. Cambridge University Press, Cambridge, UK. Wilhelm, G. (1973) Fire ecology in Shenandoah National Park. Proceedings 12th Tall Timbers Fire Ecology Conference, pp. 445–488. Will-Wolf, S. (1991) Role of ﬁre in maintaining oaks in mesic oak maple forests. Proceedings The Oak Resource in the Upper Midwest Conference. University of Minnesota, St Paul, pp. 27–33. Winston, P.W. (1956) The acorn microsere, with special reference to arthropods. Ecology 37, 120–132. Wood, O.M. (1934) A brief record of seed productivity for chestnut oak in southern New Jersey. Journal of Forestry 32, 1014–1016.

10 Growth and Yield

Introduction Although most trees increase in height and diameter each year, some become smaller in response to adverse conditions and others die. These responses all represent growth in the context of forest growth and yield. Yield is the net result of accumulated changes over time. Although yield is usually expressed in units of merchantable volume, yield also can refer to other tree and stand attributes. When forest yield is expressed as a mathematical function of time (or age), growth can be computed as the ﬁrst derivative of the yield function with respect to time. Conversely, yield over a speciﬁed time interval can be determined through numerical or analytical integration of a mathematical growth function (often called a growth curve). Growth and yield relations can be modelled at different scales, e.g. the tree, the stand, the forest or the region. Collectively, the combined changes at any given scale produce the composite changes observed at larger scales. Although it might seem logical simply to combine predictions of growth at ﬁne scales (e.g. the growth of a tree) to compile estimates of growth at coarser scales (e.g. the growth of a forest or a forested landscape), the enormous detail required to do so often renders the approach impractical. Instead, predictive growth models that operate at different scales have evolved more or less independently. The appropriate choice of model

scale depends on: (i) the question or problem to be addressed; (ii) the availability of models applicable to different scales; (iii) the availability of data necessary to apply the models; and (iv) the time and other resources required for model implementation. Although models have proven useful in predicting forest growth and yield, they also have limitations. A model may not be able to predict forest growth as accurately as an experienced forester who can foretell the effects of subtle differences in site quality or forest condition that are not explicitly considered by the model. Models nevertheless have the advantage of speed and consistency. They can quickly be applied to large forest inventories and adjusted or modiﬁed to reﬂect local patterns of growth as new information becomes available. Moreover, improvements in computing hardware and software continue to simplify the application of growth and yield models. A qualitative understanding of factors that affect the growth and yield of trees and stands is requisite to understanding quantitative models for predicting tree and stand development. Predictive models are relevant to oak silviculture because they predict changes in stand structure and composition, and they forecast product outputs. Models are presented below in their order of historical development and increasing complexity beginning with stand-level models and proceeding through individual-tree-based models. Physiological 424

Growth and Yield

models of tree development and models of large-scale forest landscapes change are not addressed.

Growth of an Oak Annual phenology The annual cycle of an oak’s growth begins with root elongation in the spring. Root elongation can continue through the entire growing season and well into autumn (Longman and Coutts, 1974). The rate of root growth of seedlings slows during periods of rapid shoot elongation, but that of mature trees continues throughout the period of shoot elongation (Reich et al., 1980) (Fig. 10.1). For trees with a large root system, the rate of root growth changes at various depths in the soil in response to temperature and moisture conditions (Teskey, 1978). As the soil warms in the spring, root growth may be greatest in the upper soil layers. Maximum root growth in deeper soil layers typically occurs when moisture in the surface layers becomes limiting later in the growing season. Cambial growth begins simultaneously throughout the bole and branches shortly after the onset of root growth and approximately 2 weeks before bud enlargement (Zasada and Zahner, 1969; Longman and Coutts, 1974). Radial growth is initially very slow. It is not until well after bud break, when oak leaves are ‘squirrel-ear size’, that diameter growth is sufﬁcient to

425

be recorded with dendrometer bands (at least a 0.01 inch increase in circumference) (Buchanan et al., 1962). The development of earlywood vessels (i.e. water-conducting cells) continues for approximately 10 weeks before changing to latewood vessels (Zasada and Zahner, 1969). Measurable diameter growth of red oaks in New York began in mid-April and continued at a nearly constant rate until the end of July (Karnig and Stout, 1969). Diameter growth of black oak in Missouri also commenced in April and was 80–90% complete within the ﬁrst 3 months of the growing season; most of that growth occurred in the ﬁrst 2 months (Buchanan et al., 1962). About 80% of annual diameter growth for northern red oak in Michigan was complete when 1000 degree-days above 40°F (4°C) had accumulated for the season (Fuller et al., 1988). Oaks in upper canopy layers continue diameter growth later into the growing season than those in subordinate crown positions. Water-stressed trees in the understorey may even decrease in diameter during the growing season, but they return to original size when moisture is more abundant (Buchanan et al., 1962). Shoot growth in oaks begins from 1 to 5 weeks after the start of cambial growth (Longman and Coutts, 1974; Reich et al., 1980) (Fig. 10.1). Unlike root elongation and cambial growth, which continue for months, shoot growth of non-juvenile trees is complete within 2–4 weeks (Buech, 1976; Reich et al., 1980). The amount of shoot growth can vary from 5 ft or more

Fig. 10.1. The phenology of root, stem and shoot growth in oaks. Thicker bars indicate periods of relatively greater growth. Weather, site conditions, species, tree age or size, and root/shoot ratio can all affect patterns of development. (Adapted from Longman and Coutts (1974) and other sources.)

426

Chapter 10

annually for vigorous young sprouts to virtually none for subordinate branches in mature trees (Reich et al., 1980). When conditions are favourable, oak reproduction may produce two or more ﬂushes of shoot growth after a rest period of 1 to several weeks between ﬂushes (Johnson, 1975, 1979; Reich et al., 1980; see also Chapter 2). However, by age ﬁve the number of ﬂushes is limited to one or two, with multiple ﬂushes occurring only occasionally on the most vigorous seedlings or sprouts (Johnston, 1941; Cobb et al., 1985).

Diameter growth Sources of variation More measurements of the diameter growth of oaks have been made than of any other tree dimension. Despite this wealth of data, only two facts about oak diameter growth are known with certainty: (i) it is inﬂuenced by many factors; and (ii) it is only partially predictable. Diameter growth varies among oak species and is further inﬂuenced by tree

size, competition, stand density, crown position, site quality and weather. Whereas factors such as site quality remain relatively ﬁxed, others such as stand density are continually changing, albeit in a more or less predictable way. Factors such as weather are inherently unpredictable and may contribute to much of the observed variation in diameter growth. Despite the many factors that affect diameter growth, general trends emerge from statistics averaged over large numbers of trees. Northern red oak is among the fastest growing of the upland oaks and its diameter growth rate averages about twice that of chinkapin or post oak (Table 10.1). Although on average the red oaks as a group grow faster in diameter than the white oaks, both groups are characterized by large variation in growth. The diameter growth of oaks is inﬂuenced more by crown position than by differences among species. For example, a change in crown class from co-dominant to intermediate typically reduces diameter growth by 20–50% (Fig. 10.2). The inﬂuence of crown class on diameter growth is

Table 10.1. Observed 10-year diameter growth rates of oaks averaged over a wide range of initial tree diameters.a

Species Northern red oak Black oak Shingle oak Scarlet oak White oak Chestnut oak Swamp white oak Bur oak Chinkapin oak Blackjack oak Post oak California black oak Overcup oak Nuttall oak Willow oak

Average dbh growth (inches/decade)

Years to grow 1 inch in dbh

1.63–2.90 1.50–2.26 1.82 1.55–1.92 1.12–1.78 1.05–1.60 1.4 1.3 0.82–1.09 0.93 0.79 0.67 2.4b 2.3b 2.3b

3–6 4–7 5 5–7 6–9 6–10 7 8 9–12 11 13 15 4b 4b 4b

aPresented in approximate order of decreasing growth rate. Compiled from Bull (1945), Dunlap (1921), Gemmill (1980), Kershaw and Fischer (1985), McDonald (1980), McIntyre (1933), Shifley and Smith (1982), Smith and Shifley (1984), and Trimble (1960, 1969). b Based on well-formed dominant trees and not directly comparable to data for other species.

Growth and Yield

427

Fig. 10.2. Ten-year average dbh growth by crown classes of some common oaks. Red oaks are from Indiana, black and white oaks from Missouri and chestnut oaks from West Virginia. (From Trimble, 1969; Shifley and Smith, 1982; Smith and Shifley, 1984.)

so great that, for trees in intermediate and overtopped crown classes, species-related differences in dbh growth may be indiscernible (Trimble, 1960). Diameter growth of dominant upland oaks can continue at rates of 2–3 inches in dbh per decade to age 70 and for trees up to 30 inches dbh (Patton, 1922; McKnight and Johnson, 1980). In bottomlands, dominant oaks can grow more than 4 inches dbh per decade. Wherever oaks occur, rapid diameter growth largely depends on maintaining a superior crown position. Based on data averaged across a wide range of tree and stand conditions, the mean diameter growth rate of oaks larger than 8 inches dbh remains relatively constant or may decline slightly with increasing dbh (Fig. 10.3). However, such averages ignore effects of crown class. Overtopped trees tend to have small diameters and small diameter increments. Consequently, growth rates for trees less than 8 inches dbh that are based on averages composited across heterogeneous stand conditions tend to be low because the smaller trees largely occur in inferior crown classes. When trees

within a single species and crown class (e.g. all intermediate white oaks) are considered separately, the correlation between initial diameter and diameter growth is seldom statistically signiﬁcant (Trimble, 1960). Site quality also inﬂuences the rate of diameter growth. An increase in site index from 50 to 80 ft (base age 50) can increase diameter growth by 1 inch per decade for dominant and co-dominant upland northern red, black, chestnut and scarlet oaks (Trimble, 1960). But site effects are not always expressed in trees in intermediate and overtopped crown classes. Diameter growth of oaks in bottomlands generally exceeds that in uplands. In southern bottomlands, dominant trees in the red oak group may grow more than 4 inches dbh per decade, and trees in the white oak group may grow nearly 3 inches dbh per decade (McKnight and Johnson, 1980). Fertilizers, although not widely applied to oak stands, also can increase diameter growth. Fifty-year-old northern red and white oaks grew 30–40% faster following application of nitrogen and phosphorus (Ward and Bowersox, 1970; Graney, 1987).

428

Chapter 10

10-year dbh growth (inches)

2.0 Black oak 1.5

White oak

1.0 Post oak 0.5

0.0 2

6

10

14

18

22

26

30

Initial dbh (inches) Fig. 10.3. Ten-year diameter growth of oaks in relation to initial dbh. Based on data from more than 9000 trees in Missouri. (From Shifley and Smith, 1982.)

White oak and black oak responded similarly but were less responsive to fertilization than northern red oak (McQuilkin, 1982). Maximum responses to fertilization usually occur the second year after application, but growth gains may persist for 6 years or longer (Karnig, 1972; Lamson, 1978; McQuilkin, 1982; Graney, 1987). Thinning combined with fertilization can further accelerate diameter growth (Graney, 1987).

Responses to thinning Reducing stand density by thinning usually increases the diameter growth of the remaining trees. This response results from the expansion of tree crowns and roots into space previously occupied by the harvested trees. This expansion, in turn, increases the leaf area and root surface area of individual residual trees and thus their photosynthetic capacity and diameter growth. Thinning also directly affects the relative crown positions of trees, and differences in crown position greatly affect diameter growth (Fig. 10.2). Diameter growth of pole-size and larger oaks increases as residual stand density decreases, even when stand density is

reduced to 30% stocking or less (Dale, 1968; Hilt, 1979; Mitchell et al., 1988) (Fig. 10.4). This response occurs across a wide range of tree ages and site classes (Dale, 1968; Hilt, 1979; Graney, 1987). Diameter growth responses to thinning depend on species, site quality, tree size, tree age, tree crown condition, intensity of thinning and time since thinning. Although most trees respond to thinning with increased diameter growth, young, wellformed trees in dominant and co-dominant crown classes usually respond the most. Trees in overtopped and intermediate crown classes often have small, poorly formed crowns and stems, and do not respond to thinning as much or as consistently as younger trees with good form and superior crown position (McGee, 1981; McGee and Bivens, 1984; Clatterbuck, 1993). Crop-tree thinning concentrates the effect of thinning on selected trees. After thinning, the diameter growth of dominant and co-dominant pole-sized oaks may increase by 50% and the effect may last for several years after treatment (Minckler, 1967; Dale and Sonderman, 1984; Ward, 1995). However, sustained increases in diameter growth of 15–30% over several

Growth and Yield

429

4 <32% Stocking

10-year dbh growth (inches)

40% Stocking 55% Stocking 70% Stocking

3

>80% Stocking

2

1

0 Missouri

Iowa

Fig. 10.4. Ten-year dbh growth of the 40 largest oaks per acre in relation to stand density (expressed as stocking percentage based on Gingrich’s (1967) stocking equation). (Adapted from Hilt, 1979.)

decades are more typically observed, especially at the residual stand densities commonly retained (Hilt, 1979; Mitchell et al., 1988). For dominant and co-dominant pole-sized white oaks, responses to thinning are often greatest the second year after thinning. However, diameter growth responses of trees with small crowns that initially developed in high-density stands may be delayed for 3–5 years. Although diameter growth declines slightly after reaching its peak response to thinning, the effects of thinning can persist for 20 years or longer (Minckler, 1957, 1967; Dale, 1968; Schlesinger, 1978; Mitchell et al., 1988). This long-term response occurs when the residual trees capture and exploit the growing space vacated by harvested trees. The effects of thinning in stands younger than 10 years (often termed weeding) are less certain. Unlike oak stands thinned in the pole-stage, weeding around sapling-size oaks does not consistently improve or maintain their crown class after treatment (Trimble, 1974). Sprouts from cut stems can quickly re-occupy the growing

space that is only temporarily freed by weeding or early thinning. Consequently, weeding does not consistently increase diameter growth of dominant and co-dominant saplings (Trimble, 1974; Lamson and Smith, 1978), but the practice often does increase diameter growth of intermediate and suppressed saplings (Allen and Marquis, 1970). Saplings in intermediate and suppressed crown classes are not usually purposely favoured in weeding operations because they rarely respond with sufﬁcient height growth to move them into the dominant and co-dominant crown classes. Weeding in conjunction with herbicides to control resprouting of cut stems can extend the thinning effect and thus may be more effective than mechanical thinning in accelerating the diameter growth of young stands (Wendel and Lamson, 1987). Thinning clumps of oak stump sprouts can signiﬁcantly accelerate diameter growth of the remaining sprouts. When red oak sprout clumps ranging from 4 to 22 years in age were thinned to a single stem, the residual stem grew about 30% faster in

430

Chapter 10

diameter than a stem of the same size in an unthinned clump (4.1 vs. 3.1 inches per decade). When thinned to two stems, residual red oak sprouts grew 14% faster in diameter than sprouts in unthinned clumps (Johnson and Rogers, 1980). Twenty-ﬁve-year-old white oak stump sprouts thinned to a single stem grew nearly 60% faster than the largest sprout in comparable unthinned sprout clumps (1.6 vs. 1.0 inches per decade) (Haney, 1962). Removal of competing understorey vegetation alone has not consistently increased the diameter growth of oaks. Northern red oaks in thinned stands in New York grew at the same rate whether or not competing understorey vegetation was eliminated with herbicide (Karnig and Stout, 1969). In thinned stands in Missouri and Iowa, black oaks and white oaks increased in diameter growth by 10–20% when understorey vegetation was removed. However, little or no response to understorey vegetation removal was observed in oak stands in Ohio and Kentucky that were similarly treated. These seemingly disparate results may be explained by the drier sites in Missouri and Iowa, where understorey removal may have reduced moisture stress in overstorey trees (Dale, 1975). Although thinning virtually ensures an increase in diameter growth of at least some residual trees, the beneﬁts from thinning depend on several factors including the duration of the response, expected changes in tree quality, and the resultant net economic gain (or loss). In general, heavier thinnings produce the greatest increases in the diameter growth of residual trees. However, thinning oak stands below 55% stocking (based on Gingrich’s (1967) stocking equation) is rarely recommended because the increased growth of residual trees may not compensate for the loss of growing stock and underutilized growing space. Moreover, epicormic branching of oaks tends to increase with thinning intensity, which may reduce the value of residual trees (Dale and Sonderman, 1984; Sonderman, 1984; Sonderman and Rast, 1988).

Value growth A tree’s economic value increases substantially when it crosses a size threshold that deﬁnes the minimum for a product class or log grade. Conversely, it can decrease dramatically when degraded by epicormic branches, logging damage, insects, disease or other factors. Consequently, changes in tree value sometimes provide a more meaningful measure of growth than changes in tree size. But coupling market values with the many variables that inﬂuence tree growth makes it difﬁcult to generalize about value growth. Mean annual value increases of 4.6% (net of inﬂation) have been reported for northern red and black oaks in New England (Gansner et al., 1990) and Pennsylvania (Herrick and Gansner, 1985). For well-formed northern red oak and black oaks between 9 and 13 inches dbh, average rates exceeded 9%. In contrast, average rates for white oak ranged from 0.3 to 1.3%, depending on region, and did not exceed 1.6%. In general, rates of value increase were greatest for trees of good form growing in stands of high site quality and where basal area was less than 60 ft2 acre−1. Among tree size classes, small sawlog-size oaks (8–16 inches dbh) yielded the highest value growth rates. The annual rate of value growth for a tree usually declined after it attained large size and high quality. In Wisconsin, annual value growth rates for northern red oak averaged between 1.2 and 1.9% (net of inﬂation) if increases in tree grade accruing with increased diameter are ignored (Buongiorno and Hseu, 1993). These value growth rates were the highest among all the species included in the study, but due to differences in methodology, those results are not directly comparable to the value growth rates cited above for New England and Pennsylvania. At two sites in Indiana, the average annual value increase for surviving white, northern red and black oaks retained after applying two-stage, irregular shelterwood cuttings ranged from 22 to 28% (including inﬂation) for more than a decade after the

Growth and Yield

21

7

18

6

15

5

12

4

9

3

6

2

3

1

0

0

alone) accounted for at least half the increase in tree value. Other factors being equal, value growth of northern red oak exceeded chestnut oak, which in turn exceeded white oak (Trimble and Mendel, 1969). Site quality also inﬂuences the value growth of individual trees. Knot-free wood is produced at earlier ages and in greater quantity on good sites than on poor sites. The formation of clear wood on a tree bole occurs as concentric sheaths surrounding the knotty core (Fig. 10.5). Dominant and co-dominant upland oak saplings and poles typically produce a clear bole length that is approximately one-third their total height. Dominant and co-dominant trees grow faster in height and produce greater clear bole length at a given age on good sites than on poor sites. For dominant and co-dominant black oaks, a branch-free 17-ft

Tree age (years)

Tree height (ft)

last harvest (Fischer et al., 1980). Residual trees averaged 11 inches dbh, and the value increase for northern red and black oaks was 2% higher than for white oak. However, windthrow at the site destroyed nearly 60% of the shelterwood trees greater than 15 inches dbh. This reduced the combined increase for all residual trees (surviving and windthrown) to between 17% and 21% annually (including inﬂation). Value growth rates (including inﬂation) for white, northern red and chestnut oaks in West Virginia exceeded 40% in some cases, but the rate decreased with increasing tree diameter and with decreasing tree vigour and site quality. Increases tended to level off at values between 4% and 12% annually (including inﬂation) when trees reached 24 inches dbh on site index 80. On average, increases in tree quality over time (as opposed to increases in tree volume

431

Clear bole length at age 7

Knotty wood Clear wood Fig. 10.5. The formation of clear (knot-free) wood as the base of the crown moves up the bole is illustrated for the first 7 years of tree growth. Oak crowns typically extend over the upper two-thirds of the bole in young trees. Clear wood thus forms only on the lower third of the bole after dead branches are shed and branch scars are healed. (Adapted from Carmean and Boyce, 1973.)

432

Chapter 10

log is expected at age 27 for site index 70, but takes nearly twice as long to develop (age 52) for site index 50. Trees growing on good sites also produce more clear wood at a given height on the bole (Carmean and Boyce, 1973). On good sites, the rapid development of a clear log allows more rings of knot-free wood to be added at a given age. For a given age, dominant and co-dominant trees on good sites produce, on average, logs of greater value than dominant and co-dominant trees on poor sites. On the better sites, the value increment associated with the additional bole length and knot-free logs boosts the value growth of individual trees and ultimately the value growth per acre.

Height growth Much information about the height growth of oaks is contained in site index curves (Chapter 4). These curves describe the effect of site quality on maximum height growth, and also conﬁrm that rates of height growth differ among the oaks (e.g. scarlet oak > black oak > white oak) (Doolittle, 1958; McQuilkin, 1974). Unfortunately, site index curves are applicable only to dominant and co-dominant trees that have not been previously suppressed. In any stand, most trees will experience periods of suppression and therefore will be shorter than indicated by site index curves of tree height over age. Height growth of dominant and co-dominant trees is relatively independent of stand density (Chapter 4). The utility of site index as a measure of site quality is premised on this assumption. But independence of height growth and stand density does not hold for very young oaks or trees growing in the lower canopy. After stands younger than 20 years old are thinned, suppressed oaks can double their periodic rate of height growth, whereas those in superior crown classes in the same stands do not respond (Haney, 1962; Ward, 1995). In young stands, thinning may even reduce the height growth of dominant and codominant trees. In 7- to 9-year-old oak

sapling stands, thinning to less than 50% stocking can decrease the rate of height growth by 60% relative to trees in unthinned stands (Allen and Marquis, 1970). The greatest negative impact from early thinning occurs among dominant and co-dominant saplings. However, the reduced height growth of young oaks after thinning is partially compensated by increases in diameter growth (Gevorkiantz and Scholz, 1948; Stout and Shumway, 1982; see also Chapter 7). Oaks more than 30 years old may show little or no height growth response after thinning. In Tennessee for example, the average height growth of overtopped white oaks that were released from the main canopy was only 0.5 ft per year (McGee, 1981; McGee and Bivens, 1984, 1993; Clatterbuck, 1993). This rate of growth was less than that of comparable overtopped white oaks in stands that were not thinned. However, increased diameter growth from thinning may increase net merchantable volume despite slow height growth. Older, overtopped white oaks nevertheless usually do not effectively respond to thinning. Moreover, they often develop and retain epicormic branches after thinning (McGee and Bivens, 1984; Clatterbuck, 1993). The difﬁculty of measuring tree heights is often an incentive to estimate tree height indirectly from other variables. For a given species, a tree’s height is related to its age, diameter, site quality, and sometimes stand density. Because tree diameter and site quality are relatively easy to measure, these factors are commonly used to estimate tree height (Fig. 10.6). Although other factors such as weather, mechanical damage from other trees, browsing, or damage from other biotic agents also can signiﬁcantly inﬂuence height growth, relatively little is known about how those factors are related to height growth. A comprehensive and widely used equation for estimating heights of oaks was developed for the Central Hardwood Region from 2300 felled trees in Illinois, Indiana, Iowa, Kentucky, Missouri and Ohio (Hilt and Dale, 1982). This equation, which is a component of the OAKSIM

Growth and Yield

433

100

Total tree height (ft)

80

60

40

20 6

10

14

18

22

26

Dbh (inches) Fig. 10.6. Heights and diameters of 500 white oaks in southern Missouri. Although the general shape of the height–diameter curve is apparent, each diameter class includes a wide range of heights. Approximately 65% of the variation in oak height can be explained by tree diameter. (From authors’ analysis.)

growth and yield model, predicts total tree height from dbh, age and site index by species (Hilt, 1985a,b, 1987). The equation, which is applicable to white oak and black oak, is constrained to ensure that, other factors being equal: ● Height approaches 4.5 ft as dbh approaches zero. ● Height growth decreases as dbh increases. ● Height growth decreases with increasing age. ● Height increases with increasing site index. ● Height, diameter, site index relationships are jointly consistent with Schnur’s (1937) stand tables. Because of these constraints, the resulting equation is mathematically complex. The equation form is given by: H = 1.37 = ( kH s − 1.37) ⋅    kH − H  D   s s   [10.1] 1 − expln    kH s − 1.37  Ds   

where H = total tree height (m) k = 1.07 for black oaks or 1.12 for white oaks D = tree dbh (cm) Hs = 1.80408 S0.932097[1 – 0.280445 exp(0.0240308A)]2.55529S 0.744034 Ds = 5.49927 S [1 – exp(0.0192593A)]1.25342 for black oak or Ds = 6.40146 S0.631893[1 – exp(0.0227614A)]1.21892 for white oak S = site index (m) A = stand age (years). This equation produces a height–diameter curve similar to Fig. 10.6. However, Equation 10.1 (graphed in Fig. 10.7) expands the height–diameter relation to account for age and site index effects. Although the equation was developed using data from unthinned stands, its application to thinned stands showed that height predictions were within 10% of observed values (Hilt and Dale, 1982). Another height–diameter equation is

434

Chapter 10

Tree height (ft)

120

80

40

50

20 10 Dbh ( inches )

0

(y

Ag e

100

0 30

ea

rs

)

200 150

0

Fig. 10.7. Estimated white oak heights by dbh and stand age for site index 70. Based on Equation 10.1. (Adapted from Hilt and Dale, 1982.)

applicable to ﬁve oak species in mixed hardwood stands in the Appalachians of Georgia, North Carolina, Tennessee and Virginia (Harrison et al., 1986). It accounts for changes in the height–diameter relation as the average height of dominant and codominant trees in the stand increases (Fig. 10.8). In applying the model, the average height of dominant and co-dominant trees can be determined from ﬁeld measurements or, for even-aged stands, can be estimated from site index curves when site index and stand age are known. The equation is of the form: H = 4.5 + Hd[1 + b1eb2 Hd][1 – eb3 D/Hd] [10.2] where H = total tree height (ft) Hd = average height of dominant and codominant trees, which can be measured in the ﬁeld or estimated as Hd = S e[22.0217(1/A – 1/50)] where bi = species-speciﬁc coefﬁcients from Table 10.2 D = tree dbh (inches) S = site index (ft at age 50) A = stand age (years).

Equations that predict tree height in the absence of information on stand age are required when stand age is unknown. This is a common problem in heavily disturbed stands and in stands under uneven-aged management. The following model can be used to predict merchantable heights of oaks in the northeastern United States when age is unknown (Yaussy and Dale, 1991): Hm = b1 Sb2 [1 – eb3 (D – T)]

[10.3]

where Hm = (a) merchantable tree height (ft) of the sawlog portion of the tree to a minimum 9-inch top diameter outside bark (dob) or where terminated by limbs, crook or breakage (T is set to 9); or (b) merchantable height (ft) of the bole to a minimum 4-inch top dob (T is set to 4) D = tree dbh (inches) T = top diameter (inches) outside bark for the height measurement, with (a) T = 9 for sawlog height or (b) T = 4 for bole length to a 4-inch top dob S = site index (ft at base age 50) bi = species-speciﬁc coefﬁcients (Table 10.3).

Growth and Yield

435

80

Tree height (ft)

60 N. red oak Black oak Scarlet oak White oak Chestnut oak

40

20

0 0

8

16 Dbh (inches)

24

Fig. 10.8 Relation between height and diameter for five species of oaks in the southern Appalachians when the mean height of dominant and co-dominant trees is 70 ft. (Adapted from Harrison et al., 1986, by permission of the Society of American Foresters, Bethesda, Maryland. Not for further reproduction.)

This equation was developed from data on more than 4000 oaks. It accounts for 25% of the observed variation in merchantable heights of sawlog-size trees and 50% of the variation in tree bole length. The unexplained variation is due to differences in tree form that cause trees of the same diameter to vary in merchantable height. Predicting merchantable heights with an equation is not recommended as a substitute for ﬁeld measurement of heights when inventorying high-value products in individual stands. However, in large-scale applications such as forest-wide invento-

ries, the equation-generated heights are often adequate (Fig. 10.9). Some height estimation equations also include stand density as an estimator. One such equation, which is applicable to three oak species-groups in the Lake States, estimates total tree height from dbh, site index and stand basal area (Ek et al., 1984). The equation also includes a taper term to facilitate estimation of heights to speciﬁed top diameters. It assumes that, all other factors being equal, trees increase in height as site index, dbh and stand basal area increase. According to the equation, estimated tree

Table 10.2. Species’ coefficients for Equation 10.2.a Coefficients Species Black oak Chestnut oak Northern red oak Scarlet oak White oak aFrom

b1 0.0719 0.2046 0.1967 0.0021 0.0026

b2

b3

0.00029

16.562 13.621 10.530 16.322 14.601

0.00957 0.00107

0.04569 0.04002

Harrison et al., 1986 by permission of the Society of American Foresters, Bethesda, Maryland. Not for further reproduction.

436

Chapter 10

Table 10.3. Species’ coefficients for Equation 10.3.a Coefficients Sawlog length (9 inch top) Species group Black oak Chestnut oak Northern red oak Scarlet oak White oak a

Pulpwood length (4 inch top)

b1

b2

b3

b1

b2

b3

16.636 10.405 25.095 22.206 11.050

0.201 0.308 0.086 0.132 0.283

0.328 0.379 0.403 0.370 0.387

13.901 10.573 34.608 27.724 19.406

0.291 0.338 0.062 0.141 0.206

0.283 0.301 0.294 0.239 0.247

From Yaussy and Dale, 1991.

heights decrease slightly when stand density is reduced by thinning. Height–diameter relations for several oaks of the western United States are described by the following equation: H = 4.5 + exp(b0 b1 Db2)

[10.4]

where H = total tree height (ft) D = tree dbh (inches) bi = regression coefﬁcients that differ by species and region.

This equation was developed for Oregon white oak in the west central Willamette Valley (Wang and Hann, 1988) and for Oregon white oak, California black oak and canyon live oak in southwestern Oregon (Larsen and Hann, 1987). For the latter three species, tree height to the base of the live crown also can be estimated from total height and other stand characteristics (Ritchie and Hann, 1987). A more complicated equation based on dbh, site index and relative crown size can be used

50

Merchantable tree height (ft)

Pulpwood

40 Sawlogs

30

Site index 70 Site index 50 20

10 4

8

12

16

20

24

28

32

Dbh (inches) Fig. 10.9. Estimated merchantable tree heights for sawlog-size (9-inch top dob) and pulpwood-size (4-inch top dob) white oaks in the northeastern United States (see Equation 10.3). (From Yaussy and Dale, 1991.)

Growth and Yield

to estimate the total height of blue oak in California (Standiford, 1997). The total height of upland oaks and hickories in Illinois can be estimated from observed merchantable height by the following equation from Myers and Belcher (1981): H = (0.85 Hm + 30)

of survival rates (or conversely mortality rates) requires the use of a common time interval, usually one year or one decade. Relations between annual and periodic survival and mortality rates with respect to time interval are as follows: Annual survival rate = (periodic survival rate)1/n Periodic survival rate = (annual survival rate)n Annual mortality rate = 1 annual survival rate Periodic mortality rate = 1 periodic survival rate

[10.5]

where H is total height (ft) and Hm is merchantable height (ft) to a 5-inch top diameter outside bark. When merchantable height is known to a 4-inch top diameter inside bark, total heights of oaks in the Lake States can be estimated using a similar methodology described by Gevorkiantz and Olsen (1955, table 11).

Survival rates Although competition results in regular and predictable reductions in numbers of trees per acre as stands develop (Chapters 5 and 6), individual tree mortality is less predictable. Survival rates of individual trees vary by species, tree size, stand density, site quality and crown class. Comparison

437

[10.6] [10.7] [10.8] [10.9]

where n is the number of years in the period and survival and mortality rates are expressed decimally. Mortality rates are computed in terms of the corresponding survival rate and then converted to a mortality rate. When averaged across a wide range of stand conditions, annual survival rates are lowest for small trees. For example, the annual survival of 2-inch dbh oaks in Missouri averages 94%, whereas the average rate for 6-inch trees exceeds 99% (Fig. 10.10). Trees in the white oak group

1.00 Indiana and Illinois 0.99

Annual survival rate

Missouri 0.98

0.97

0.96

0.95

0.94 2

6

10

14

18

22

26

30

Initial dbh (inches) Fig. 10.10. Annual survival rates for oaks (all species) in Indiana, Illinois and Missouri. Higher average site quality in Indiana and Illinois results in faster stand development, greater competition and lower survival rates for small diameter trees than in Missouri. For both data sets, the white oak group comprised 53% of trees, and the red oak group comprised 47%. (From Shifley and Smith, 1982, and Smith and Shifley, 1984.)

438

Chapter 10

generally have higher survival rates than trees in the red oak group (Fig. 10.11). This characteristic is consistent with the greater shade tolerance and longevity of the white oaks. Survival rates for oaks also vary by region. For example, rates for oaks smaller than 6 inches dbh are lower in Indiana and Illinois than in Missouri; the pattern is reversed for larger trees (Fig. 10.10). Such relations may be related to site productivity, which on average is higher in Indiana and Illinois than in Missouri. These patterns of survival reﬂect the faster tree growth, earlier stand closure and earlier competitioninduced self-thinning on better sites. Although oaks die from many causes, the death of a tree often follows a long downward spiral in health and vigour (Franklin et al., 1987; Ammon et al., 1989; Starkey et al., 1989; Wargo and Haack, 1991). Lee (1971) suggests separating mortality into two classes: regular and irregular. Regular mortality results from competition and self-thinning as a consequence of stand development through time. It is the most prevalent type of tree mortal-

ity and is more predictable than other causes of mortality. Irregular mortality is the abrupt, largely unpredictable mortality that results from events such as ﬁre, windstorms, lightning, ice, snow, drought, insect and disease epidemics, and other factors. Although the two classes of mortality are not always separable (e.g. intense competition may predispose a tree to insect attack), it is useful to recognize that the largest and most predictable proportion of tree mortality results from competition. Tree survival following ﬁre varies by species, tree size and ﬁre intensity. Post, white, and chestnut oak are more ﬁre-resistant than black, northern red and scarlet oaks (Loomis, 1973; Regelbrugge and Smith, 1994). Fire intensity is often reﬂected in the scorch height on the bole. When the scorch height in feet is equal to the tree diameter in inches, the probability of mortality ranges from 50 to 90%, depending on species and tree diameter (Fig. 10.12). Oaks are usually better able to survive a dormant-season ﬁre than a growing-season ﬁre of the same intensity.

1.00 White oak group

Annual survival rate

0.99

Red oak group

0.98

0.97

0.96

0.95

0.94 2

6

10

14

18

22

26

30

Initial dbh (inches) Fig. 10.11. Mean annual survival rate in relation to initial diameter for white oak and red oak species groups in Missouri. Rates are based on remeasurement intervals ranging from 7 to 11 years. The white oak group is predominantly white oak (52%) and post oak (40%); the red oak group is predominantly black (57%), northern red (13%) scarlet (13%) and blackjack (12%) oaks. Based on more than 14,000 trees in stands of variable age, density and structure. (From Shifley and Smith, 1982.)

Growth and Yield

439

20 Red oaks (MO)

Scorch height (ft)

15 White/post oaks (MO)

10

Chestnut oaks (VA) 5 Red oaks (VA) 0 1

5

9 Dbh (inches)

13

Fig. 10.12. Bole scorch heights associated with the death of 50% of trees of a given dbh following growing-season fires in Missouri and Virginia. Species in the white oak group (white and post, or chestnut oaks) survived higher scorch heights than species in the red oak group (black, scarlet and northern red oaks) for a given diameter. However, the relation varied greatly within species groups in both states. (Adapted from Loomis, 1973 (Missouri), and Regelbrugge and Smith, 1994 (Virginia)).

Stand Growth Growth and yield in even-aged stands Normal stands (100% stocking) Early yield tables for oaks were based on data from so-called ‘normally’ stocked even-aged stands. These yield tables were derived from inventories of relatively undisturbed stands and were thus at or near maximum density for their age and site quality. Normal yield tables present expected mean volumes per acre by stand age and site classes. The change in volume by age class thus can be used to infer volume growth over time. Normal yield tables therefore represent simple models of stand 1The

change over time. Stand tables, which list numbers of trees by diameter classes for different stand ages and site classes, also can be compiled from the same data. Normal yield and normal stand tables apply to stands at or near the 100% stocking line (sometimes termed average maximum relative density) on Gingrich-style stocking charts1 (e.g. Figs 6.9 and 6.11). Reported yields for normal stands thus represent the special case where stands are at 100% stocking. Yields will be different for stands with stocking below 100%. Stand volume growth and yield depend largely on ﬁve factors: stand age, tree heights, site quality, stocking and merchantability standards (Gevorkiantz and Scholz, 1948). The basal area of even-aged,

terms ‘normal stocking’ and ‘full stocking’ are sometimes used interchangeably. This is potentially confusing because ‘full stocking’ as defined by the Gingrich (1967) stocking chart refers to the range between 58% and 100% stocking (Fig. 6.9). Any stand density within this range represents full utilization of growing space by trees (assuming trees are well distributed). As used herein, ‘normal’ stand density or ‘normal’ stocking is synonymous with 100% stocking as shown on Gingrich’s or similar stocking charts.

440

Chapter 10

normally stocked stands increases over time as mean stand diameter increases. These increases are accompanied by decreases in numbers of trees per acre. A widely cited and comprehensive source of information on the growth and yield of oak stands is Schnur’s (1937) compilation of normal yield and stand tables for upland oaks of the eastern United States. It includes stand and yield tables by age and site index for even-aged, normally stocked stands, and is based on data gathered across much of the oak range of the eastern United States. These tables were compiled from data collected in even-aged stands that were at least 30% oak (predominantly white, black, scarlet, chestnut and northern red oaks). Selected stands spanned a wide range of ages, and many originated after clearcutting. The stands were fully stocked with uniformly distributed trees. Because Schnur’s yield tables are referenced to stand age and site index, they can be presented graphically (Fig. 10.13). However, their derivation limits their application to stands near 100% stocking (i.e. normal stocking). Similar stand and yield tables were developed for normal even-aged oak stands in southwestern Wisconsin (Gevorkiantz and Scholz, 1948). On average, 85% of the merchantable volume in those stands was comprised of northern red and white oak. Numerous mesic species were also present, including sugar maple, American basswood, elm, black walnut, black cherry and ash. The transitional nature of these forests to more northern forest types is reﬂected in the paucity of hickories compared to Schnur’s (1937) study region, where hickories occurred on 70% of sample plots. Gingrich’s (1971a) yield tables for the Central Hardwood Region provide a third comprehensive source of growth and yield data for even-aged oak stands. These tables include thinning effects by age and site index. They also include yield tables for unthinned stands at 100% stocking. Yield estimates from the three sources cited above can be compared graphically (Fig. 10.14). Although estimated basal areas

are similar for all sources, volume estimates differ substantially. Volume differences result from differences in merchantability standards, analytical methodology and stand selection criteria. Because yield tables are simple models of average forest growth and yield, they cannot be expected to predict accurately changes in individual stands. They nevertheless provide useful information on general patterns of oak stand development and provide reasonable average yield estimates for oak stands at 100% stocking. When yield tables (or any other source of forest yield data) are used to make silvicultural and management decisions, their applicability to a speciﬁc stand or area should be veriﬁed. It is especially important to consider merchantability standards and the minimum dbh used in yield table construction, because these can greatly affect volume estimates. Although stand volume increases with age, the number of trees decreases dramatically. Tree mortality is thus relentless in unthinned oak stands. Typically, only about 5% of the trees in a normally stocked 10-year-old oak stand will live to age 100 (Schnur, 1937). The rate of decrease in numbers of trees can be graphically illustrated for an age series of unthinned even-aged oak stands (Fig. 10.15). This decrease is the result of competition-induced tree mortality, or selfthinning. Self-thinning occurs at an earlier age on the better sites (where trees grow rapidly) than on the poorer sites (Chapter 6). Whereas annual survival rates for oaks are about 93% in 10-year-old stands, survival rates increase to 99% by age 80 (Fig. 10.16). Even though these survival rates exceed 90%, they represent exponentially decreasing numbers of trees over time. For example, an annual survival rate of 93% per year reduces the tree population by almost half in one decade (0.9310 = 0.48); this rate would be typical of 10- to 20-year old stands. Even when the annual survival rate is 98%, about half the original number of trees in a stand will die in 35 years (0.9835 = 0.49).

140

A

80 60 40

B

80

100 60 20 30,000

20,000

70 60

10,000

50 40

0 6000 Cubic ft per acre

441

80

C

70 4000

Site index (ft)

Board ft per acre

Basal area (ft2 per acre)

Growth and Yield

60 50 40

2000

0 80

80

Cords per acre

D 60

70 60

40

50 40

20 0 10

20

30

40

50

60

70

80

90

100

Stand age (years) Fig. 10.13. Basal area and volume yields of normal (100% stocked) upland oak stands in the eastern United States. (A) Stand basal area for trees ≥0.6 inch dbh; (B) cubic foot volumes excluding bark for trees ≥0.6 inch dbh; (C) board foot (Scribner) volume to an 8 inch top diameter inside bark; and (D) cordwood volume to a 4 inch top diameter outside bark. (From Schnur, 1937.)

Thinned stands Thinning results in recovery of the volume in trees that otherwise would be lost to mortality due to crowding and self-thinning. Moreover, growth in thinned stands

is concentrated on fewer, faster-growing trees that have been speciﬁcally retained for their potential to produce high-quality products. The objective of thinning is therefore to reduce stand density below 100% stocking in order to allocate more

Chapter 10

Basal area ft2 per acre

442

160 120 80

Gingrich (1971) Schnur (1937) Gevorkaintz and Scholz (1948)

40 0

Cubic ft per acre

5000 4000 3000 2000 1000 0

Board ft per acre

20,000 15,000 10,000 5000

Cords per acre

0 60 40 20 0 20

40

60

80

100

120

140

Stand age (years) Fig. 10.14. Reported yields for normal (100% stocked) upland oak stands from three sources (Schnur (1937) for site index 70, Gevorkiantz and Scholz (1948) for a medium site, and Gingrich (1971a) for site index 65). The Schnur and Gingrich volumes are based on data from mixed oak stands in central and eastern United States. The Gevorkiantz and Scholz volumes are based on data from the driftless area of southwestern Wisconsin, which is predominantly northern red oak. Schnur’s volumes are board feet Scribner to an 8 inch top diameter inside bark (dib), cubic feet excluding bark to a 0.6-inch top dib, and cords to a 4-inch top diameter outside bark (dob). Volumes for Gevorkiantz and Scholz are board feet International quarter-inch to an 8-inch top dib, cubic feet excluding bark to a 0.6-inch top dib, and cords to a 4 inch top dob. Volumes for Gingrich are board feet International quarter-inch to an 8.5-inch top dob, cubic feet of entire stem including bark, and cords to a 4-inch top dob.

Growth and Yield

Site index (ft) 40 60 80

6000

Trees per acre

443

4000

2000

0 20

40

60

80

100

Stand age (years) Fig. 10.15. Trees per acre by stand age and oak site index in normally stocked even-aged upland oak stands in the eastern United States. Includes all trees 0.6 inches dbh and larger. (From Schnur, 1937.) 1.00

Annual survival rate

0.98

0.96

0.94 Site index (ft) 60 (Schnur)

0.92

65 (Gingrich)

0.90 15

25

35

45

55

65

75

85

95

Stand age (years) Fig. 10.16. Annual survival rates of trees by stand age classes for normally stocked even-aged upland oak stands in the eastern United States. (Adapted from Schnur, 1937, and Gingrich, 1971a.)

growing space to fewer but more desirable trees, and thereby accelerate their growth. Normal yield tables do not apply to thinned stands because normal yield tables represent growth relations occurring at or near maximum stand density (i.e. 100% stocking).

Numerical algorithms can be used to predict the rate at which any stand below 100% stocking (thinned or otherwise) approaches normality through time. Yield estimates for thinned stands accordingly can be related to normal yield tables (Davis, 1966). However, a more direct

444

Chapter 10

approach is to use yield tables or yield models that are directly applicable to thinned stands such as Gingrich’s (1971a) yield tables for oak stands in the Central Hardwood Region. These tables are also compatible with the silvicultural recommendations of Roach and Gingrich (1962, 1968) and include projections of stand development for several thinning regimes (Fig. 10.17). The objective of thinning is to accelerate the diameter and merchantable volume growth of the residual stand. The greater the intensity of thinning (down to the minimum stocking required for full utilization of growing space), the greater the diameter growth of the remaining trees. Thinning also results in the utilization of trees that would otherwise be lost through mortality. Thinning can be applied by area-wide thinning (i.e. thinning the stand as a whole), or by crop-tree thinning (i.e. thinning around

selected trees) (Rogers and Johnson, 1985; Mitchell et al., 1988; Lamson et al., 1990). Both methods accelerate the growth of the retained trees. Compared to unthinned oak stands, thinning can increase yields by 25–80% over a rotation. Greatest yield increases occur when thinning begins early in the life of the stand and when stocking is reduced to about 60% at each entry (Fig. 10.18) (Gingrich, 1971a, b). The gross cubic foot volume and basal area growth of white oak stands in the midwestern United States remains relatively constant at residual stand densities ranging from 50 to 120% stocking (Dale, 1968). This is consistent with the Langsaeter principle (Smith, 1962), which posits that, other factors being equal, the gross volume growth of all trees per unit area (including trees that die) remains nearly constant over a wide range of stocking levels (e.g. between A and B levels of stocking as

Cu

Stand yield (board feet acre–1)

mu

lat

ive

16,000

12,000

al idu s Re

8000

d

inne

Unth

4000

Cut

0 30

40

50

60

70

80

Stand age (years) Fig. 10.17. Board-foot yields of thinned and unthinned upland oak stands in the Central Hardwood Region, oak site index 65. Total yield for the unthinned stand is shown by the line labelled ‘Unthinned’. The thinned stand was thinned to 60% stocking every 10 years beginning at age 30. Total volume yield for the thinned stand (labelled ‘Cumulative’) is the sum of two components: the residual growing stock after periodic thinnings (labelled ‘Residual’), and the volume of material harvested during the thinnings (labelled ‘Cut’). Yields for thinned stands are nearly identical to unthinned stands to age 50, but by age 80 the cumulative yield for thinned stands is nearly double that of unthinned stands. (From Gingrich, 1971a.)

Cumulative 80-year yield (board feet acre–1)

Growth and Yield

445

25,000

20,000

Age when thinned (years) 30 40 50 60 Unthinned

15,000

10,000

5000

0 55

65 Oak site index (ft)

75

Fig. 10.18. Effects of site quality and age at first thinning on cumulative board foot yields (cut volume plus residual stand volume) of Central Hardwood oak stands (Gingrich, 1971a). In this example, stands were thinned every 10 years to age 80. Cumulative yields can be increased by starting periodic thinnings early in the rotation.

deﬁned in Figs 6.9 and 6.11). However, recent re-examinations of this principle indicate that, at least in some cases, gross volume growth continually increases with increasing stand density (for further discussion see Nyland, 1996; Smith et al., 1997; Zeide, 2001). In either case, thinning directs the volume growth per acre to a smaller number of faster growing trees, while allowing the harvest and utilization of trees that would otherwise die from inter-tree competition. Consequently, controlling stand density in oak stands through thinning can produce substantial increases in net merchantable volume production (board feet or cubic feet) and in value even though total gross volume growth may be held constant or even decrease relative to unthinned stands. Oak stands in Kentucky and Iowa attained a maximum net growth in basal area of 2.9 ft2 acre1 year1 when they were thinned to a residual stocking of 50–60% based on Gingrich’s stocking chart (Dale, 1968). In comparison, the basal area of unthinned stands at 100% stocking grew at an annual rate of only 1.9 ft2, or about 65% of maximum. Net cubic foot volume

growth (total growth less mortality) reached a maximum at 70–80 ft3 acre1 year1 when residual stocking was maintained at 60%. At 100% stocking, net volume growth was reduced to 50–60 ft3 acre1 year1 (Dale, 1968). Thinning oak stands to 40% stocking may further accelerate the growth of the residual trees and the net volume growth of the stand (Leak, 1981; Hilt and Dale, 1989), but sawtimber quality may suffer from increased epicormic branching. Below 40% stocking, net volume growth per acre decreases because growing space is not fully utilized (Hilt and Dale, 1989). Information on the response of western oaks to thinning is limited. The annual basal area growth of 40–85-year-old California live oak stands increased from 0.5 to 2.1 ft2 acre1 when thinned from 180 ft2 acre1 to 100 ft2 acre1. Basal area growth following thinning to 50 ft2 acre1 was 1.7 ft2 acre1 annually (Pillsbury and Joseph, 1991). Over the 5-year duration of that study, annual cubic and board foot volume growth were greater on the unthinned plots. In a stand of California black oak, tanoak and madrone with an ini-

446

Chapter 10

tial basal area of 200 ft2 acre1, net volume growth was highly variable after thinning to residual basal areas ranging from 85 to 140 ft2 acre1 (Plumb and McDonald, 1981). Maximum annual volume growth was 93 ft3 acre1, but a relationship between density and volume growth was not evident in that small study. Fertilization in combination with thinning can further accelerate basal area and volume growth of oak stands. In the Boston Mountains of northwestern Arkansas, nitrogen fertilization increased basal area and volume growth (ft3 in the bole from ground to tip) by 16–20% relative to unfertilized stands (Graney and Murphy, 1993). This response occurred at residual densities from 40 to 100 ft2 acre1 and at site indices ranging from 50 to 80 ft. Effects of fertilization persisted for up to 10 years after treatment. Historically, most thinning research and the resulting silvicultural recommendations were based on the implicit objective of maximizing net volume or value production. Maximization of one measure of productivity (e.g. cubic foot volume) may not maximize another (e.g. ﬁnancial return). Also, economic yields may not show the smooth response curve characteristic of volume or basal area production. This is especially true for high-value hardwoods. Diameter-based product classes (such as pulpwood vs. sawtimber) are characterized by ﬁnancial yield curves with sharp increases in value whenever a substantial proportion of trees in a stand grow into high-value product classes. Accounting for the time-value of money (i.e. interest rates and inﬂation) further complicates assessment of value growth. Analyses of ﬁnancial returns from thinning (e.g. Utz and Sims, 1981) are dependent on assumptions about costs and revenues. Most recommendations for thinning oak stands are based on yield tables and are intended to produce an economically valuable mix of products including cordwood, sawlogs and, where possible, veneer logs (Roach and Gingrich, 1962, 1968; Gingrich, 1971a; Sander, 1977). Although those recommendations are often appropriate, they may not suit all management objectives. Consequently, a thorough prognosis for a

speciﬁc stand may require a detailed stand inventory and a growth and yield analysis using one of the models described later in this chapter. Yield tables are best suited to characterizing the expected development of an ‘average’ even-aged stand. Growth and yield equations overcome some of the limitations of yield tables by their ability to account for factors such as variation in species mixtures or variation in tree size distributions.

Growth and yield in uneven-aged stands Yield estimation for uneven-aged stands under single-tree or group selection silviculture presents problems separate from those for even-aged stands. Uneven-aged stands are periodically thinned by methods that create and sustain three or more age classes. Stands are typically harvested at 10–20 year intervals called cutting cycles; there is no speciﬁed rotation age. Stand structure goals are usually deﬁned by the use of q values or other criteria that deﬁne a target diameter distribution. If not already in place, the desired stand structure must be created using combinations of practices such as single-tree and group selection methods (Chapter 8). Yield information for uneven-aged oak stands is largely anecdotal. In an 18-year study of a mixed-oak woodlot in central Indiana, annual growth averaged 306 board feet (Doyle log rule) acre1 year1 of which 176 board feet were cut (Bramble and Fix, 1980). Corresponding net cubic foot volume growth was 76 ft3 acre1 year1 of which 41 ft3 were cut. In a 16-year study of oak–hickory stands on poor sites in southern Illinois managed by single-tree selection, periodic annual net board foot volume growth averaged 133 board feet acre1 (Schlesinger, 1976). Of that growth, about half (64 board feet) was harvested. Good sites in the same study produced 218 board feet acre1 year1 of which 149 board feet were harvested. Annual net growth in two West Virginia oak stands managed by the selection method for 20 years averaged 308 board

Growth and Yield

feet acre1 of which 190 board feet were harvested (International quarter-inch log rule). The growth on individual tracts ranged from 228 to 400 board ft acre1 year1. Net cubic foot volume growth (trees between 5 and 11 inches dbh) averaged 55 ft3 acre1 year1 of which 15 ft3 were harvested. Basal area growth averaged 2.2 ft2 year1 (Trimble, 1970). Appalachian hardwood stands in West Virginia that were managed for 30 years using single-tree selection grew at annual rates between 330 and 505 board ft (International quarter-inch log rule). Annual basal area growth ranged from 2.1 to 2.7 ft2 acre1. Northern red oak was a major component of these stands and red oak site index ranged from 60 to 80 ft. These reported periodic growth rates for managed uneven-aged stands are as high or higher than those for unmanaged even-aged stands when the even-aged yields are expressed as the mean annual increment over a full rotation. The mean annual increment of managed even-aged stands may be higher or lower than that of managed uneven-aged stands, depending on site quality, management intensity and rotation length. However, such comparisons are confounded by several factors. First, reports of growth and yield for uneven-aged stands cover relatively short intervals, usually 20 years or less for stands that were mature when uneven-aged management began. Second, the actual harvest from uneven-aged stands is typically half to three-quarters of periodic growth. Over the long term, cumulative periodic harvest, rather than cumulative growth, is a better measure of yield for stands under uneven-aged management. Accordingly, long periods of observation would be needed to adequately compare expected yields of even- and uneven-aged stands. Third, uneven-aged stands are usually comprised of a large proportion of oaks in overtopped and intermediate crown classes. Many of those trees do not respond to periodic reductions in stand density. Finally, issues related to changes in species composition, tree quality and the sustainability of uneven-aged oak forests remain unresolved in many ecoregions (Chapter 8).

447

Growth and Yield Models Modelling methods Any method of forecasting forest growth is a growth and yield model. This includes yield tables and the educated guesses of experienced foresters. As computing technology has become more sophisticated and accessible, yield tables and educated guesses have been largely replaced by analytical models that can quickly provide estimates of periodic growth or cumulative yield from a given set of initial conditions. General modelling theory and objectives are discussed in conjunction with regeneration models in Chapter 3. Most growth and yield models are classiﬁed as statistical models as deﬁned in that chapter. Statistical growth and yield models are usually developed by ﬁtting regression curves to numerous observations of tree or stand growth and yield. The resulting models provide the following advantages over other methods of yield estimation: (i) rapid implementation via computers; (ii) incorporation of more independent (predictor) variables; (iii) simultaneous consideration of the effects of various species mixtures and thinning regimes; and (iv) ﬂexibility in summarizing and reporting growth and yield estimates. Most statistical growth and yield models fall into one of four broad classes: stand-level models, individualtree-level models, stand table projection models, and diameter distribution models. Stand-level models directly predict growth and yield (per acre) for an entire stand. Some stand-level models do this by estimating and combining the components of stand growth: survivor growth, ingrowth, mortality and harvest (Beers, 1962). These models have modest data and computational requirements compared to other types of models. However, ﬂexibility in handling species mixtures and options for summarizing results are limited in this class of models. Individual-tree models predict the growth and survival of a representative sample of individual trees in a stand and accumulate those changes to arrive at an

448

Chapter 10

estimate of stand change. The implementation of this class of models requires more data from the user than other classes. It also offers greater ﬂexibility for predicting the growth of mixed-species stands and for summarizing results. Stand table projection models use stand table projection formulae (e.g. Husch et al., 1982) to forecast forest changes, which are displayed as numbers of trees per unit area by species and/or dbh classes. Stand table projection has long been used to project short-term growth based on local ﬁeld observations of diameter growth. The technique has become more versatile with the development of models that predict growth and survival by species groups and diameter classes. Diameter distribution models deﬁne a stand’s diameter frequency distribution (i.e. numbers of trees by dbh classes) using a probability density function such as the Weibull (Bailey and Dell, 1973). These models predict changes in the shape of a diameter distribution over time. Those changes are then translated into changes in numbers of trees by diameter classes. Characteristics such as basal area or volume can be estimated for the entire stand or by diameter classes using the predicted shape of the diameter frequency distribution. This class of models has shown greatest promise for application to species that grow in relatively pure stands. Their application to oak forests has been very limited. With any model, the appropriate use and interpretation of model output requires knowledge of the class of models to which it belongs. Within each class of models there are similarities in model assumptions, data required as model input, the way data are processed and the details of model output. Every model has a limited range of applicability; model predictions reﬂect the range of conditions included in the data used to calibrate the model.

Stand-level models for oaks One of the earliest statistical growth and yield models for oaks was the stand-level model, GROAK (Dale, 1972, 1973). It is based on a system of ﬁve equations that pre-

dicts annual basal area increment, mean stand diameter, gross cubic foot volume, merchantable cubic foot volume and gross board foot volume. Applying the model requires user-provided information on stand age, site index, current stand basal area and number of trees per acre. The model is relatively easy to understand and its component equations are presented below (Dale, 1972):

B = 3.68521 B A0.75 – ln(B) B A0.8 0.011383 S B A−1.05 [10.10] Cubic Vol = exp(3.09094 0.00930176 S 1.03909 ln(B) – [10.11] 20.11035 A1) Dq = 1.1341 0.0019876 A S

[10.12]

CF ratio = −0.052676 0.7876045 exp [−(1.2987 – 0.08117 Dq)10] [10.13] BF ratio = −0.088414 3.63827 exp [−(2 – 0.125 Dq)4] [10.14] where

B = net annual basal area growth (ft2 acre1) for trees ≥2.6 inches dbh Dq = diameter of the tree of average basal area for trees ≥2.6 inches dbh Cubic Vol = total volume (ft3 acre1) for trees ≥2.6 inches dbh including bark, stump and tip, but not branchwood A = stand age (years) B = basal area of live trees ≥2.6 inches dbh (ft2 acre−1) S = site index (ft at 50 years) CF ratio = ratio of merchantable cubic foot volume (ft3 acre1) to total cubic foot volume. Merchantable volume is for trees ≥ 4.6 inches dbh to a top dob of 4.5 inches; excludes stump, bark and branches. CF ratio = 0 when Dq < 2.3, and CF ratio = 0.735 when Dq > 16. BF ratio = ratio of merchantable board foot volume (ft acre1) to total cubic foot volume. Board foot volumes are in International quarter-inch rule for trees ≥8.6 inches dbh to an 8.5 inches top dob BF ratio = 0 when Dq < 4.8, and BF ratio = 3.55 when Dq > 16. Equation 10.10, which predicts basal area growth, is of key importance in this model (Fig. 10.19). Other equations deﬁne other stand characteristics in terms of basal area.

449

4

3

2 20

1

basal

area

80 100

(ft 2 ac

re –1)

120

ag

e

95

60

d

Initial

(y e

70 40

an

020

ar s)

45

St

Annual basal area growth (ft2 acre–1)

Growth and Yield

Fig. 10.19. Annual basal area growth (ft2 acre1) of upland oak stands in the Central Hardwood Region by age and initial basal area for oak site index 70 (based on Equation 10.10). (From Dale, 1972.)

By adding the predicted basal area growth to the initial basal area and incrementing stand age by 1 year, it is possible to carry the system of equations forward through time, predicting change in basal area and other stand characteristics year by year. Initially released in a computer program with the acronym GROAK, the model has application to estimating yields of evenaged upland oak stands in the eastern United States. Thinning is simulated by reducing stand basal area. An extensive set of tables derived from the model is available in print (Dale, 1972), and can be used to estimate growth and yield under a wide range of thinning regimes. However, the model also can be implemented using spreadsheet software. The model has been incorporated into other simulation programs such as Yield (Hepp, 1982) and WINYIELD (Hepp, 1995), which combined growth and yield models for several forest types into a single menu-driven computer program. Such computer programs provide a convenient user-interface that simpliﬁes entering data for initial stand conditions, offers several output options and allows economic analysis of simulated thinning regimes. Equations 10.10–10.14 are based on growth measurements on permanent plots in Ohio, Kentucky, Missouri and Iowa.

However, Dale (1972) cautions that the predicted maximum cubic foot volume growth (trees ≥2.5 inches dbh) occurs at stand densities between 30% and 60% stocking, which are below densities at which maximum growth reportedly can occur (Dale, 1968; Gingrich, 1971a). A test of the model with data from thinned plots in the Boston Mountains of northwestern Arkansas produced acceptable results, although accuracy was improved by using equation coefﬁcients derived from local growth data (Graney and Murphy, 1991). Equations for the model modiﬁed for stands in the Boston Mountains (Graney and Murphy, 1994) are listed below: B2 = {0.02590 – [0.02590 – B10.71754] (A2/A1)0.71754(2.4884)}1/(0.71754) [10.15] T = B0.94 exp(3.1427 0.015209 S – 37.92/A) [10.16] M = 0.92588 T {1 – exp[0.028105 (24.422) D]}1/0.028105 [10.17] Dq = 2.0672 + 0.0024 A S 0.02734 B [10.18] V = 0.91634 M {1 – exp[0.12375 (0.026524) S D]}1/0.12375 [10.19] FD = 3.4528 V 0.00044738 V2

[10.20]

FS = 5.1686 V 0.00026988

V2

[10.21]

FI = 6.0194 V 0.00032819

V2

[10.22]

Chapter 10

where:

This system of equations predicts basal area and volume growth of even-aged, upland forests in the Boston Mountains of Arkansas that are dominated by black, northern red and white oaks. Based on Equation 10.15, greatest annual basal area growth occurs in young stands with initial basal areas from 50 to 80 ft2 acre1 (Fig. 10.20). However, for stands on average sites that are older than 55 years, board foot volume growth is greatest when basal area is 80 ft2 acre1 or greater. These equations can be implemented using spreadsheet software. A system of stand-level yield equations developed for oak–hickory forests in Minnesota can be used to predict yields by stand age, site index and stand density (Walters and Ek, 1993). It was developed from more than 600 permanent sample plots from a statewide forest inventory. These equations are part of a larger set of yield models for the 14 major forest types in Minnesota, and they can be conveniently used to make combined regional estimates for several forest types. Because sample plots were not screened for uniformity of stocking or stand history, the equations estimate the

2.0

1.5

1.0 20

0.5

ag

asa40 la

Initial

basal

e

70 I20 nitial b

95

rea (f

area

–1 t 260 a

cre)

(ft 2 ac

re –1)

nd

0.0

(y

ea

rs

)

45

sta

Annual basal area growth (ft2 acre–1)

A1 = current stand age A2 = future stand age A = any speciﬁed age B1 = basal area (ft2 acre−1) at age A1, trees ≥2.6 inches dbh B2 = projected basal area (ft2) acre−1 at age A2, trees ≥2.6 inches dbh B = any speciﬁed basal area (ft2) acre−1), trees ≥2.6 inches dbh S = site index (ft at age 50) Dq = quadratic mean dbh (inches) T = total cubic foot volume (ft3 acre1 inside bark from 0.2-ft stump height to top of tree) M = merchantable cubic foot volume (ft3 acre1 inside bark from 0.5-ft stump to 5-inch top dob for trees ≥6 inches dbh) V = sawtimber cubic foot volume (ft3 acre1 inside bark from 1-ft stump to 10inch top dob for trees ≥12 inches dbh) FD = board foot volume (acre−1, Doyle rule) FS = board foot volume (acre−1, Scribner rule) FI = board foot volume (acre−1, International quarter-inch log rule).

80

120

In itia l

450

Fig. 10.20. Annual basal area growth of upland oak stands in the Boston Mountains of Arkansas in relation to initial stand age and initial basal area. The graph illustrates values for site index 70 (based on Equation 10.15). (Adapted from Graney and Murphy, 1994.)

Growth and Yield

average, statewide forest conditions and estimated yields are not equivalent to values expected for stands at 100% stocking. The Walters and Ek models nevertheless can be used to produce yield tables by site and age classes that are similar in format to the normal yield tables of Schnur (1937) and Gevorkiantz and Scholz (1948), and they also can be used to predict yields at variable stand densities. For oak stands at 100% stocking that are ≤ 30 years old and growing on site index 70, estimates of basal area and total cubic foot volume yields (trees ≥4.95 inches dbh) from the Walter’s and Ek equations are higher than those reported by Schnur’s (1937) normal yield tables. In older stands, estimated basal area yields are within 80–90% of the values reported by Schnur although estimated total cubic foot volumes are only within 65% of Schnur’s reported yields. Up to stand age 50, board foot volumes are at least 33% greater than those predicted by Schnur, but at older ages board foot yields are 40% lower than Schnur’s tables. A few stand-level models have been developed for mixed-species forests where oaks are a major component, but do not necessarily comprise most of the stocking. One of these is a stand level model for upland oak–shortleaf pine mixtures in southeast Missouri (Larsen, 1980). Other such models include a red oak–sweetgum model applicable to stands in minor stream bottoms in central Mississippi (Sullivan et al, 1983), and a model for Appalachian mixed hardwoods in North Carolina and Georgia (Bowling et al., 1989). The latter model, in addition to estimating basal area and number of trees per acre, uses the Weibull distribution to generate information on tree diameter distributions. These models are well suited for growth and yield estimation within the region and the range of forest conditions for which they were calibrated.

Stand table projection models Stand table projection is a mensurational technique for projecting growth by manip-

451

ulating a stand table containing numbers of trees by species and dbh classes. Stand tables are commonly constructed from forest inventories. Diameter growth rates are derived from increment cores or from repeated measurements of permanent forest inventory plots. The estimated diameter growth rates are applied to the stand table to forecast how trees will increase in diameter over time (Husch et al., 1982). For short projection periods (e.g. 10 or 20 years), stand table projections based on site-speciﬁc dbh growth data are often very accurate if management practices and forest conditions do not greatly change. Stand table projections can be implemented using spreadsheet software or by programs designed speciﬁcally for that purpose (e.g. Hepp, 1992, 1995).

Individual-tree-level models for oaks Individual-tree growth and survival models can provide information not available from stand-level models. However, they do so at the cost of greater complexity in application and greater data input requirements. This class of models requires a userprovided list of trees comprising a representative sample of trees from each stand for which future yield estimates are desired. The required minimum information for each sample tree includes its species, dbh and the number of trees per acre it represents. The tree list is usually compiled from a stand inventory. Any common sampling procedure can be used to obtain the representative sample of trees. Values are expanded to a per acre basis. Individual-tree models operate directly on the tree list, using estimated periodic tree diameter growth, height growth and survival to update the tree list through time. At any point during the simulation period, the current state of the tree list can be viewed. Stand summaries by species, size classes and product classes also can be generated. These models thus can account for species’ differences in growth and survival. Because the number of species occurring in oak forests may be large,

452

Chapter 10

species are usually grouped according to their silvical characteristics. To simplify model development, the same algebraic equation form is used to predict the growth of each species group. However, each group has a unique set of coefﬁcients that result in different predicted growth rates. Survival models are developed similarly. Because of the many details required for implementing individual-tree models, their use requires computer software. Software may be provided by the model developer or may be available elsewhere. Despite the differences among software programs in the details of individual-tree models, they all follow a similar computational scheme (Fig. 10.21). Of the widely available and used individual-tree models, OAKSIM (Hilt, 1985a), TWIGS (Miner et al., 1988) and FVS (Teck et al., 1996) are applicable to oak forests of the eastern United States and are discussed below. There are other individual-tree models that can be applied to oak forests, but their applications are more limited geographically or to speciﬁc objectives. Included are models for: ● Northern red oak and other species in Vermont (Hughes and Sendak, 1985). ● The economic analysis of different harvesting regimes for red oak in New England (Hibbs and Bentley, 1984). ● Mixed oak and other Appalachian hardwoods immediately after thinning (Harrison et al., 1986). ● Canyon live oak and Oregon white oak in southwestern Oregon (Hann and Larson, 1991). ● Blue oak in California (Standiford, 1997). ● Oak stands of sprout origin in Missouri (Lowell and Mitchell, 1987; Lowell et al., 1987).

● Even-aged stands on upland sites. ● Trees ≥2.6 inches dbh. ● Oaks comprising at least 75% of the stand basal area. ● Stand ages 30–120 years. ● A maximum projection period of 50 years. ● Black oak site index 50–85 ft (age 50). ● Stocking from 20 to 120% based on Gingrich’s (1967) chart.

Set options Specify input/output

Read tree list

Summarize and display tree and stand conditions

Simulate harvest as required

Compute cost/revenue for harvest Analyse financial parameters

Grow the trees Increment diameters Compute tree mortality

OAKSIM OAKSIM is an individual-tree-based model designed to estimate the growth and yield of thinned and unthinned mixed-oak stands in southern Ohio and southeastern Kentucky (Hilt, 1983, 1985a,b). The model is applicable to the following conditions (Hilt, 1985b):

Final summary and exit

Fig. 10.21. Steps for a representative individualtree-based growth and yield simulation model.

Growth and Yield

The following equations deﬁne the model:

0.10

The software for model application incorporates an algorithm that adds random variation to the predicted basal area growth, and also includes constraints that ensure the combined growth and survival of all individual trees is approximately equal to the per acre growth predicted by Dale’s (1972) equations (Equations 10.10 – 10.14) for a comparable stand. OAKSIM computes yields in cubic and board feet (International quarter-inch log rule) to user-speciﬁed top diameters. Additional details on model implementation, volume computations and thinning algorithms are given in Hilt (1985a). The general shape of the growth and survival models (Fig. 10.22) is consistent with the empirical observations presented earlier. OAKSIM can be implemented with various software programs including those of Hilt (1986b), the Northeast Decision Model (Simpson et al., 1996), and YIELD-MS or WINYIELD (Hepp, 1992, 1995).

in g

26

ck

2

1.0 0.8

B

0.6

26 20 14

0.09 0.07

Annua

l dbh

0.05

growth

ch

es

0.2

)

0.4

8 0.03 0.01 2

(in

Probability of survival

20 14 Initia 8 l dbh (inch es)

St o

where

B = average 5-year basal area growth per tree (ft2)

D = periodic 5-year tree diameter growth (inches, computed from B and D) D = initial tree dbh (inches) S = site index (ft at 50 years) Dq = quadratic mean stand diameter (trees ≥2.6 inches dbh) P = stocking per cent (based on Gingrich, 1967) PS = 5-year probability of survival H = tree height (ft) bi, ci = parameter estimates for species groups (see Hilt, 1985a).

20 40 60 80 100 120

%

0.05

bh

[10.26]

0.15

ld

H = 4.5 c0 [1 – exp(c1)]

0.20

(inche

itia

[10.24]

PS = 1 – [1 exp(b0 + 2b1 D b2 D)]1 [b3 b4 S] [10.25]

0.25

s)

In

D = (D2 B/0.005454)0.5 – D

A Annual dbh growth (inches)

B = D2 [6.96762087䡠 106 S1.5731724exp(−0.11839854 Dq – 0.01198244 P)] [10.23]

453

Fig. 10.22. Response surfaces for growth and survival of white oak based on the OAKSIM model. (A) Annual diameter growth for trees growing on site index 70 in relation to initial tree diameter (dbh) and stocking percentage (based on Gingrich’s (1967) stocking equation). The model assumes that initial dbh is equal to quadratic mean stand diameter; (B) annual survival probabilities for white oak in relation to initial dbh and annual diameter growth rate. (From Hilt, 1985a.)

TWIGS TWIGS is an individual-tree-based model designed to estimate growth and yield for established stands in the eastern United States (Belcher et al., 1982; Brand et al., 1987; Shiﬂey, 1987; Miner et al., 1988). The model incorporates equations that predict annual diameter growth and survival for many species groups, including oaks. The approach accommodates a wide range of species mixtures from stands where oaks

Chapter 10

A 0.16 0.12 0.08 0.04

100 80 60

26 18

Db h

40

10

(inc

2 20

hes

Annual survival probability

tions can be used to graphically illustrate representative growth and survival relations produced by the model (Fig. 10.23). Although each regional variant of TWIGS utilizes a different set of mathematical models to estimate annual tree diameter growth and survival probabilities, within each variant the same algebraic model form is used to estimate diameter growth for all tree species. Growth differences among species are accounted for by species-speciﬁc regression coefﬁcients. The

Ba s tre al a es re (ft 2 a o ac f la re – rge 1 ) r

Annual dbh growth (inches)

predominate to those where oaks are a minor component. The TWIGS software facilitates applying the models to individual stands, predicting growth under a wide range of silvicultural practices, and computing measures of economic performance associated with each practice. There are four regional variants of TWIGS. Although the complexity of the component equations and differences among the model’s regional variants prohibits detailed presentation here, the equa-

)

1.00 0.99

B

0.98

100 80

0.97 0.96

Ba s tre al ar es ea (ft 2 of ac larg re –1 er )

454

60 26

Dbh

40

18

(inch

10

es)

2 20

Fig. 10.23. Estimated annual diameter growth and survival of white oak based on the Central States TWIGS model. Although shapes of response surfaces vary for other species, all are based on the same mathematical model form. (A) Estimated annual dbh growth of a white oak in relation to its dbh and the total basal area per acre of trees that are as large or larger in dbh. (B) Estimated annual probability of survival of a white oak in relation to its dbh and the total basal area per acre of trees that are as large or larger in dbh. (Adapted from Shifley, 1987, and Miner et al., 1988.)

Growth and Yield

same is true for tree survival equations and for volume equations used to compute yields within each TWIGS variant. The TWIGS software provides a common interface for implementing any of its variants. TWIGS reports volumes in cubic feet or board feet (International quarter-inch log rule) of merchantable material, and tons or cubic feet of residue by species group. Stocking percentage, diameter distributions, stumpage values (based on user supplied stumpage rates), and a variety of economic performance measures also can be computed. During the course of a simulation, the current state of the tree list can be displayed or saved to compute other stand characteristics. Management options include a variety of thinning rules that can be applied interactively to meet speciﬁc stand density goals and to manipulate species composition. The growth and survival models used in TWIGS also have been incorporated into other programs such as GROW (Brand, 1981), YIELD-MS (Hepp, 1992), WINYIELD (Hepp, 1995), the Northeast Decision Model (Simpson et al., 1996), and the Forest Vegetation Simulator (FVS) (Bush, 1993, 1995; Bush and Brand, 1995; Teck et al., 1996). The FVS model (described in the next section) provides a ﬂexible and powerful tool for predicting and comparing growth, yield and harvesting options for multiple stands managed under different scenarios. Additional information on the accuracy and precision of growth estimates produced by TWIGS is presented elsewhere (Brand and Holdaway, 1983; Crow, 1986; Holdaway and Brand, 1986; Kowalski and Gertner, 1989; Schuler et al., 1993). Schuler and others (1993) compared predictions from TWIGS, OAKSIM and other models for the northeastern United States.

Forest Vegetation Simulator The Forest Vegetation Simulator (FVS) is individual-tree-based modelling software that includes growth, survival and regeneration models applicable to the major forest regions of the United States (Teck et al., 1996). FVS grew out of the Prognosis

455

model and software (Stage, 1973; Wycoff et al., 1982) that were originally developed for coniferous forests in the western United States. Subsequent variants of Prognosis have been developed for other geographical regions and include individual-tree growth and survival models for several western oak species (Table 10.4). Although the speciﬁc details of the oak growth and survival equations differ among the regional variants of FVS, they are all similar in form to those developed for the Prognosis model (Stage, 1973). The models for western oak species function much like OAKSIM (Hilt 1985a,b) and TWIGS (Miner et al., 1988). Consequently, the FVS software was expanded to incorporate the growth and survival models from the Lake States, Central States, Northeast and Southeast variants of TWIGS. FVS includes more than 20 regional variants, each with individual-tree-based growth and survival models for the major species occurring within each region. Nine of the FVS variants provide growth and yield projections for oaks (Table 10.4). The FVS software provides a standard interface with access to most of the individual-tree-based oak growth and survival models applicable in the United States. FVS includes options for simulating harvest treatment, summarizing product output, working with multiple stands and displaying the predicted forest structure in two and three dimensions (Teck et al., 1997).

Estimating ingrowth Ingrowth is the number or volume of trees that periodically grow into the smallest measured tree size class. It is not usually accounted for in oak growth and yield models. For well stocked, even-aged stands that are pole-size or larger, ingrowth can often be disregarded. In those stands, ingrowth will contribute little to stocking or merchantable volume. But for young stands, understocked stands or stands under uneven-aged management, ingrowth can be important. Ingrowth is deﬁned in relation to a

Region of applicability

Lake States

MN, WI, MI

Central States

Oak species included Red oak group

White oak group

Sources of information

28

N. red oak Black oak N. pin oak

White oak Swamp white oak Bur oak Chinkapin oak

Belcher et al., 1982; Blinn et al., 1988 Brand and Holdaway, 1983 Brand et al., 1987; Buchman, 1983 Buchman and Lentz, 1984 Buchman et al., 1983; Bush and Brand, 1995 Holdaway, 1984; Miner et al., 1988; Teck, 1995 Teck et al., 1996; USDA Forest Service, 1979 Online at FVS website

IN, IL, MO, IA

30

Black, N. red Scarlet, Pin Blackjack oak

Chestnut oak Chinkapin oak Post, White oak

Brand et al., 1987; Bush, 1995 Miner et al., 1988; Shifley, 1987; Teck, 1995

Northeast

CT, DE, KY, ME, MD, NA, NH, NJ, NY, OH, PA, RI, VT, WV

28

Black oak Cherrybark oak N. red oak S. red oak Scarlet oak Pin oak

Chestnut oak Swamp chestnut Swamp white oak White oak Bur oak Post Oak

Bush, 1993 Hilt and Teck, 1989 Hilt et al., 1987 Simpson et al., 1995 Teck, 1990; Teck, 1995 Teck et al., 1996; Teck and Hilt, 1991

Southeast Georgia

GA

53a

Black, N. red S. red, Scarlet Shingle, Laurel Overcup oak

Chestnut oak Dwarf post oak Post, White, Water Willow oak

Bolton and Meldahl, 1990a,b Meldahl et al., 1988 Teck et al., 1996

Central Rockies

Central Rocky Mountains

24

All indigenous oaks

All indigenous oaks

Stage, 1973; Teck, 1995 Wycoff et al., 1982; Online at FVS website

Klamath Mountains

Klamath Mountains

11

California black oak

Pacific Northwest Coast

Coastal Pacific Northwest

37

Oregon white oak

Stage, 1973; Teck, 1995 Teck et al., 1996; Wycoff et al., 1982 Online at FVS website

Westside Cascades

Western Cascade Mtns

37

Oregon white oak

Stage, 1973; Teck, 1995; Teck et al., 1996 Wycoff et al., 1982; Online at FVS website

Westside Sierra Nevada

Western Sierra Nevada Mtns

11

aThis

California black oak

Stage, 1973; Teck, 1995; Teck et al., 1996 Wycoff et al., 1982; Online at FVS website

Stage, 1973; Teck, 1995 Teck et al., 1996 Online at FVS website

is the number of site-by-species-group combinations and is not directly comparable to species-groups for other variants.

Chapter 10

Regional variant

No. species groups included

456

Table 10.4. Regional variants of the Forest Vegetation Simulator (FVS), the major oak species included in each variant and additional sources of information. See also Teck et al. (1997) and the FVS website.

Growth and Yield

speciﬁed threshold diameter. Trees below the threshold are not included in stand or forest inventories. Common ingrowth thresholds range from 1 to 9 inches dbh. Across that range, ingrowth can vary greatly. Typically, the average annual number of ingrowth trees declines rapidly as the threshold diameter increases (Shiﬂey, 1990) (Fig. 10.24). However, there is great variation around these averages. Coefﬁcients of variation for the number of ingrowth trees per acre typically range from 100 to 300%. There are few models for estimating ingrowth in oak stands. In thinned oak stands in the Boston Mountains of Arkansas, the average annual number of ingrowth trees per acre for a 2.6 inch dbh threshold was 1.36% of the total number of live trees. These stands covered a wide range of initial ages (Graney and Murphy, 1991). In contrast, the annual rate of ingrowth in thinned stands ranging from 22 to 90 years old in Ohio, Kentucky, Missouri and Iowa was only 0.63% of live trees ≥2.6 inches dbh (Dale, 1973). The following ingrowth model based on the latter study accounted for nearly half of the

observed 1973):

variation

in

ingrowth

(Dale,

I2.6 = 0.09264 0.00000113 A2 – 0.015674 ln(B) – 0.07618 Dq A0.8 0.001019 N A0.8 – 0.00000083 S N [10.27] where I2.6 = ratio of annual number of ingrowth trees per acre for a 2.6 inches dbh threshold to total number of trees per acre ≥2.6 inches dbh A = stand age in years B = basal area of trees ≥2.6 inches dbh (ft2 acre1) Dq = quadratic mean stand diameter (inches, for trees ≥2.6 inches dbh) S = site index (ft at 50 years) N = number of trees ≥2.6 inches dbh (number acre−1). For applications requiring a threshold dbh other than 2.6 inches, the following model can be used to estimate ingrowth for any threshold diameter between 1 and 13 inches dbh (Shiﬂey et al., 1993). The model is based on data from three oak forest types in Indiana, Illinois and Missouri

40

Ingrowth (trees per acre per year)

Oak–hickory Oak–pine Oak–gum–cypress

30

20

10

0 1

3

457

5

7

9

Ingrowth threshold dbh (inches) Fig. 10.24. Average number of trees per acre per year that grow into a given threshold dbh class (ingrowth) for three oak forest types. Based on data from Indiana, Illinois and Missouri. (From Shifley, 1990.)

Chapter 10

400 300 200 100

80 etition factor

comp

db ld

Crow n

h

9

140

(in ch e

5

0 200

s)

1

20

Th re sh o

Ingrowth (trees per acre per decade)

458

13

Fig. 10.25. Number of trees per acre per decade that grow into a given or threshold dbh class (ingrowth) in oak–hickory forests. Crown competition factor is a measure of stand density based on the maximum growing space a tree of a given diameter can utilize (see Chapter 6). (From Shifley et al., 1993, by permission of Society of American Foresters, Bethesda, Maryland. Not for further reproduction.)

(Fig. 10.25); the model equations are as follows: For oak–hickory forests: IT = 0.1032 MaxT – 0.000004211 T (MaxT)2 – 0.0000022 T 2 (MaxT)2 [10.28] For oak–pine forests: IT = 0.09132 MaxT – 0.0000003867 T (MaxT)2 − 0.000001296T 2 (MaxT)2 [10.29] For oak–gum–cypress forests: IT = 0.04949 MaxT 0.00001457 T (MaxT)2 – 0.0000007506 T 2 (MaxT)2 [10.30] where IT

= number of ingrowth trees per acre per decade at threshold T T = ingrowth threshold size between 1 and 13 inches dbh MaxT = an expression for the maximum possible number of ingrowth trees at threshold T, computed as MaxT = [(287 18.5 T - 0.0703 T2) – CCF] and 0.00180303 (3.121.829 T)2 CCF = crown competition factor expressed as a percentage of an

acre (see Krajicek et al., 1961 and Chapter 6). Equations 10.28–10.30 should be selected to match forest type. The ingrowth threshold diameter is speciﬁed as T in the right-hand side of the equation. CCF can be calculated using species-speciﬁc formulae as described by Krajicek and others (1961) or Shiﬂey (1990). However, for oak–hickory forests, CCF can be approximated by: CCF = 0.1755 N 0.02058 D 0.006032 D2 [10.31] where CCF = crown competition factor expressed as a percentage of an acre N = the number of trees per acre larger than the threshold dbh D = the sum of tree diameters per acre larger than the threshold dbh D2 = sum of tree diameters squared per acre for trees larger than the threshold dbh. (This value can be computed as the sum of basal areas of trees larger than the threshold dbh divided by 0.005454.)

Growth and Yield

Model evaluation Using a computer model or a yield table to obtain a growth or yield estimate does not ensure that the estimate will be accurate. Weather ﬂuctuations, genetic variation, variation in tree spacing and site quality, and the many above- and below-ground structural characteristics of trees and stands that are not usually measured, can all affect the accuracy of predictions. The central question in evaluating a growth and yield model is whether or not it provides a better estimate than alternative estimation methods. A model’s validity is therefore relative and its utility should be judged in relation to an educated guess or another model that would be used in its place (Forrester, 1968). Although model suitability depends on many factors, they can be grouped into three categories: (i) application environment; (ii) model performance; and (iii) model design (Buchman and Shiﬂey, 1983). Shortcomings in any one category may justify eliminating a model from further consideration. The application environment includes quality of user support, computer requirements and data handling ability. Model performance includes the analytical evaluation of accuracy and precision associated with the predictions made by the model or its various components. The best evaluations of model performance are those that compare model predictions with observed growth within the locality where the model is to be applied. Quantitative evaluations of model performance are not always easy, but there are some simple procedures that can be used to ensure that performance is acceptable before relying on a model. These include: (i) comparing model results to other published sources of growth and yield information, to other models, or to educated guesses by local experts; (ii) applying the model to a wide range of conditions and noting any unreasonable predictions that result from initial conditions within the model’s range of applicability; and (iii) checking for realistic model responses to

459

simulated silvicultural practices such as thinning. Additional guidance in selecting models can be obtained by following established guidelines for the statistical evaluation of models (e.g. Reynolds et al., 1981). Design considerations include the ability of the model to simulate silvicultural practices, the ease with which it can be adjusted to meet local conditions, and its ability to accommodate a variety of other objectives. Whereas the design characteristics of yield tables are ﬁxed, computer models are more ﬂexible. Some allow customized output tables that help users compare management alternatives. However, such ﬂexibility often carries with it the cost of added complexity. Among the existing growth and yield models for oaks, it is difﬁcult to generalize about the superiority of one versus another. Direct comparisons of models are few, and rigorous quantitative evaluations of models using independent data are even fewer. Before applying any model it would therefore be prudent to: (i) compare conditions for the stands of interest to the conditions reportedly used to develop the model; (ii) study the information in the user’s guide about how the model was tested at the time of its development and what known limitations apply to the model; and (iii) conduct local evaluations of model performance. When a model prediction fails to reproduce local observations, some remedial measures are possible. Simple ratio adjustments to predicted growth or survival rates can be used to adjust individual-tree-based models for individual sites (Stage, 1982; Smith, 1983). The Forest Vegetation Simulator includes software options that facilitate adjusting model estimates to match growth trends observed locally (Teck et al., 1996).

Volume Equations The literature is replete with volume tables and equations applicable to the oaks. The selection of a volume estimation table or equation is often determined by local or

460

Chapter 10

regional custom, or by the similarity between stands of interest and those used to develop a given equation or table. Most volume estimation tables and equations for oaks have not been quantitatively compared with one another. Volume estimation for oak trees and stands is complicated by volume differences among species and inconsistent merchantability standards. Available volume tables and equations often differ in minimum top diameters and units (e.g. board feet, cubic feet, green weight, dry weight). Some tables pertain to individual species, while others (e.g. Gevorkiantz and Olsen, 1955) apply to groups of species. The simplest and most direct volume estimates are obtained by observing merchantable tree height and diameter for individual oak trees and looking up the corresponding volume in a species-speciﬁc or a composite volume table. For large inventories it is often more convenient to use equations that analytically reproduce the entries in the volume tables, which eliminate the need to look up tabular values for each tree (e.g. Beers, 1964). Composite volume tables or equations that are averaged across many species are convenient to use during large inventories. Although information on merchantable tree heights is required in the application of most volume tables and equations, measuring heights of individual trees is usually time consuming. When merchantable tree height is not known, local volume tables based on diameter and site index can be used. However, there is likely to be some loss of accuracy with such tables compared to volume estimates based on measured tree heights. Examples of local volume equations include those of Hahn and Hansen (1991), Raile and others (1982), and Smith and Weist (1982). The most versatile volume estimation equations are those based on taper equations that allow the user to set and/or alter the merchantable top diameter. Taper-

based volume estimation equations for oak include those of Hilt (1980), Martin (1981) and Williams and Wiant (1994). Application of these equations requires an estimate of total tree height and a computer.

Regional Patterns in Oak Yield and Productivity Oak growth and yield can be observed at scales ranging from individual trees to stands to regions. Similar to the regional patterns in the distribution of oak species and oak forests (Fig. 1.2), there are regional patterns in the distribution of the volume of oak timber. In the United States, the total standing volume of oak in 1997 was about 364 trillion board feet or 113 trillion cubic feet (USDA Forest Service, 2000). This resource is largely concentrated in the eastern United States (Fig. 10.26). Although the south central and eastern states contain the highest volumes of oaks, the relatively droughty oak forests of California contain over 5 billion cubic feet. Statewide volumes are inﬂuenced by a state’s size and the acreage of forest within it. Summaries of the average volume of oaks per acre of forest provide a different perspective. For example, Illinois is only 11% forested and the total oak resource is small relative to other states. Yet the volume of oak per acre of forest in Illinois is among the highest of any state, indicating that its forests contain relatively large numbers of oaks with high volumes. The forests of Connecticut, Maryland, Rhode Island, Tennessee, Virginia and West Virginia also have high volumes of oaks per acre of forest. In those states, many oak forests occur on highly productive sites with high standing volumes. Maps from other sources provide additional details on the geographic distribution of volumes of the major oak species in the eastern United States (e.g. Beltz et al., 1992).

Growth and Yield

461

0–1000 1001–2000 2001–3000 3001–4000 4001–6000 > 6000

0–100 101–200 201–300 301–400 401–500 > 500 Fig. 10.26. The distribution of standing oak volumes by state. (A) Total oak cubic volume per state (million ft3); (B) average cubic volume per acre of timberland (ft3). (Adapted from Powell et al., 1994.)

References Allen, R.H. and Marquis, D.A. (1970) Effect of thinning on height and diameter growth of oak and yellow-poplar saplings. USDA Forest Service Research Paper NE NE-173. Ammon, V., Nebeker, T.E., Filer, T.H., McCracken, F.I., Solomon, J.D. and Kennedy, H.E. (1989) Oak decline. Mississippi Agriculture and Forestry Experiment Station Technical Bulletin 151. Bailey, R.L. and Dell, T.R. (1973) Quantifying diameter distributions with the Weibull function. Forest Science 19, 97–104. Beers, T.W. (1962) Components of forest growth. Journal of Forestry 60, 245–248. Beers, T.W. (1964) Composite hardwood volume tables. Purdue University Agricultural Experimental Station Research Bulletin 787. Belcher, D.M., Holdaway, M.R. and Brand, G.J. (1982) A description of STEMS – the stand and tree evaluation and modeling system. USDA Forest Service General Technical Report NC NC-79. Beltz, R.C., Cost, N.D., Kingsley, N.P. and Peters, J.R. (1992) Timber volume distribution maps for the eastern United States. USDA Forest Service General Technical Report WO WO-60.

462

Chapter 10

Blinn, C.R., Rose, D.W. and Belli, M.L. (1988) Guide to analyzing investment options using TWIGS. USDA Forest Service General Technical Report NC NC-123. Bolton, R.K. and Meldahl, R.S. (1990a) Design and development of a multipurpose forest projection system for southern forests. Alabama Agricultural Experiment Station Bulletin 603. Bolton, R.K. and Meldahl, R.S. (1990b) User’s guide to a multipurpose forest projection system for southern forests. Alabama Agricultural Experiment Station Bulletin 604. Bowling, E.H., Burkhart, H.E., Burk, T.E. and Beck, D.E. (1989) A stand-level multispecies growth model for Appalachian hardwoods. Canadian Journal of Forest Research 19, 405–412. Bramble, W.C. and Fix, W.L. (1980) Productivity of a managed woodland in the central states. Journal of Forestry 78, 553–556. Brand, G.J. (1981) GROW – a computer subroutine that projects the growth of trees in Lake States’ forests. USDA Forest Service Research Paper NC NC-207. Brand, G.J. and Holdaway, M.R. (1983) Users need performance information to evaluate models. Journal of Forestry 81, 235–237. Brand, G.J., Holdaway, M.R. and Shiﬂey, S.R. (1987) A description of the TWIGS and STEMS individual-tree-based growth simulation models and their applications. USDA Forest Service General Technical Report NC NC-120, Vol. 2, pp. 950–960. Buchanan, W.D., Liming, F.D. and Harrison, W. (1962) Diameter growth patterns of black oak trees. Journal of Forestry 60, 342–343. Buchman, R.G. (1983) Survival predictions for major Lake States tree species. USDA Forest Service Research Paper NC NC-233. Buchman, R.G. and Lentz, E.L. (1984) More Lake States tree survival predictions. USDA Forest Service Research Note NC NC-312. Buchman, R.G. and Shiﬂey, S.R. (1983) Guide to evaluating forest growth projection systems. Journal of Forestry 81, 232–235. Buchman, R.G., Pederson, S.P. and Walters, N.R. (1983) A tree survival model with application to species of the Great Lakes region. Canadian Journal of Forest Research 13, 601–608. Buech, R. (1976) Tree shoot elongation in northern Wisconsin and relationships with temperature and precipitation. Canadian Journal of Forest Research 6, 487–498. Bull, H. (1945) Diameter growth of southern bottomland hardwoods. Journal of Forestry 43, 326–327. Buongiorno, J. and Hseu, J.S. (1993) Volume and value growth of hardwood trees in Wisconsin. Northern Journal of Applied Forestry 10, 63–69. Bush, R.R. (1993) The northeastern TWIGS variant of the forest vegetation simulator. USDA Forest Service WO Service Center, Fort Collins, CO. Bush, R.R. (1995) The central states TWIGS variant of the forest vegetation simulator. USDA Forest Service WO Service Center, Fort Collins, CO. Bush, R.R. and Brand, G.J. (1995) The Lake States TWIGS variant of the Forest Vegetation Simulator. USDA Forest Service WO Service Center, Fort Collins, CO. Carmean, W.H. and Boyce, S.G. (1973) Hardwood log quality in relation to site quality. USDA Forest Service Research Paper NC NC-103. Clatterbuck, W.K. (1993) Are overtopped white oak good candidates for management? USDA Forest Service General Technical Report SO SO-93, pp. 497–500. Cobb, S.W., Miller, A.E. and Zahner, R. (1985) Recurrent shoot ﬂushes in scarlet oak stump sprouts. Forest Science 31, 725–730. Crow, T.R. (1986) Comparing the performance of two northern hardwood growth projection systems. Northern Journal of Applied Forestry 3, 28–32. Dale, M.E. (1968) Growth response from thinning young even-aged white oak stands. USDA Forest Service Research Paper NE NE-112. Dale, M.E. (1972) Growth and yield predictions for upland oak stands 10 years after initial thinning. USDA Forest Service Research Paper NE NE-241. Dale, M.E. (1973) Growth and yield functions for thinned upland oak stands. PhD dissertation, Iowa State University, Ames. Dale, M.E. (1975) Effect of removing understorey on growth of upland oak. USDA Forest Service Research Paper NE NE-321. Dale, M.E. and Sonderman, D.L. (1984) Effect of thinning on growth and potential quality of young white oak crop trees. USDA Forest Service Research Paper NE NE-539. Davis, K.P. (1966) Forest Management, 2nd edn. McGraw-Hill, New York.

Growth and Yield

463

Doolittle, W.T. (1958) Site index comparisons for several forest species in the southern Appalachians. Soil Science Society of America Proceedings 22, pp. 455–458. Dunlap, F. (1921) Growth of oak in the Ozarks. Missouri Agricultural Experiment Station Research Bulletin 41. Ek, A.R., Birdsall, E.T. and Spears, R.J. (1984). A simple model for estimating total and merchantable tree heights. USDA Forest Service Research Note NC NC-309. Fischer, B.C., Standiford, R.B. and Callahan, J.C. (1980) Value growth rates from overwood trees in the irregular shelterwood method in southern Indiana upland hardwoods. Proceedings Central Hardwood Forests Conference III. University of Missouri, Columbia, pp. 401–407. Forrester, J.W. (1968) Principles of Systems: Text and Workbook, 2nd preliminary edn. Wright-Allen Press, Cambridge, Chs 1–10. Franklin, J.F., Shugart, H.H. and Harmon, M.E. (1987) Tree death as an ecological process. BioScience 37, 550–556. Fuller, L.G., Reed, D.D. and Holmes. M.J. (1988) Modeling northern hardwood diameter growth using weekly climatic factors in northern Michigan. USDA Forest Service General Technical Report NC NC-120, Vol. 1, pp. 467–474. Gansner, D.A., Arner, S.L. and Birch, T.W. (1990) Timber value growth rates in New England. USDA Forest Service Research Paper NE NE-632. Gemmill, B. (1980) Radial growth of California black oak in the San Bernardino Mountains. USDA USDA Forest Service General Technical Report PSW PSW-44, pp. 128–135. Gevorkiantz, S.R. and Olsen, L.P. (1955) Composite volume tables for timber and their application in the Lake States. USDA Technical Bulletin 1104. Gevorkiantz, S.R. and Scholz, H.F. (1948) Timber yields and possible returns from the mixed-oak farmwoods of southwestern Wisconsin. Wisconsin Department of Conservation Publication 521. Gingrich, S.F. (1967) Measuring and evaluating stocking and stand density in upland hardwood forests in the Central States. Forest Science 13, 38–53. Gingrich, S.F. (1971a) Management of young and intermediate stands of upland hardwoods. USDA Forest Service Research Paper NE NE-195. Gingrich, S.F. (1971b) Stocking, growth, and yield of oak stands. Proceedings Oak Symposium. USDA Forest Service Northeastern Forest Experiment Station, pp. 65–73. Graney, D.L. (1987) Ten-year growth of red and white oak crop trees following thinning and fertilization in the Boston Mountains of Arkansas. USDA Forest Service General Technical Report SO SO-42, pp. 445–449. Graney, D.L. and Murphy, P.A. (1991) Growth and yield comparisons for upland oak stands in the Boston Mountains of Arkansas. USDA Forest Service General Technical Report SE SE-70, pp. 229–239. Graney, D.L. and Murphy, P.A. (1993) Growth response of upland oak stands in the Boston Mountains to nitrogen fertilization. USDA Forest Service General Technical Report SO SO-70, pp. 485–495. Graney, D.L. and Murphy, P.A. (1994) Growth and yield of thinned upland oak stands in the Boston Mountains of Arkansas. Southern Journal of Applied Forestry 18, 10–14. Hahn, J.T. and Hansen, M.H. (1991) Cubic and board foot volume models for the Central States. Northern Journal of Applied Forestry 8, 47–57. Haney, G.P. (1962) Growth of thinned white oak sprouts. Journal of Forestry 60, 644–645. Hann, D.W. and Larsen, D.R. (1991) Diameter growth equations for fourteen tree species in southwest Oregon. Oregon State University Forest Research Laboratory Research Bulletin 69. Harrison, W.C., Burk, T.E. and Beck, D.E. (1986) Individual tree basal area increment and total height equations for Appalachian mixed hardwoods after thinning. Southern Journal of Applied Forestry 10, 99–104. Hepp, T.E. (1982) YIELD: timber yield forecasting and planning tool. Southern Journal of Applied Forestry 6, 135–140. Hepp, T.E. (1992) A user manual for YIELD-MS timber planning tool for mixed stands, Version 3. In INFORM 3 Users Manual. Ofﬁce of Natural Resources and Economic Development, Tennessee Valley Authority, Norris, Tennessee. Hepp, T.E. (1995) WINYIELD 1.0 Software and Documentation. Forest Resources Systems Institute (FORS), Florence, Alabama. Herrick, O.W. and Gansner, D.A. (1985). Forest-tree value growth rates. Northern Journal of Applied Forestry 2, 11–13.

464

Chapter 10

Hibbs, D.E. and Bentley, W.R. (1984) A growth model for red oak in New England. Canadian Journal of Forest Research 14, 250–254. Hilt, D.E. (1979) Diameter growth of upland oaks after thinning. USDA Forest Service Research Paper NE NE-437. Hilt, D.E. (1980) Taper-based system for estimating stem volumes of upland oaks. USDA Forest Service Research Paper NE NE-458. Hilt, D.E. (1983) Individual-tree diameter growth model for managed, even-aged, upland oak stands. USDA Forest Service Research Paper NE NE-533. Hilt, D.E. (1985a) OAKSIM: an individual-tree growth and yield simulator for managed, even-aged, upland oak stands. USDA Forest Service Research Paper NE NE-562. Hilt, D.E. (1985b) User’s guide to OAKSIM: an individual-tree growth and yield simulator for managed, even-aged, upland oak stands. USDA Forest Service General Technical Report NE NE-104. Hilt, D.E. (1987) Using OAKSIM: an individual-tree growth and yield simulator for managed, evenaged, upland oak stands. Proceedings 15th Annual Hardwood Symposium of the Hardwood Research Council, Memphis, Tennessee, pp. 67–86. Hilt, D.E. and Dale, M.E. (1982) Height prediction equations for even-aged upland oak stands. USDA Forest Service Research Paper NE NE-493. Hilt, D.E. and Dale, M.E. (1989) Thinning even-aged upland oak stands. USDA Forest Service, North Central Forest Experiment Station Central Hardwood Notes. Note 6.06. Hilt, D.E. and Teck, R.M. (1989) NE-TWIGS: an individual-tree growth and yield projections system for the northeastern United States. The Compiler 7(2), 10–16. Hilt, D.E., Teck, R.M. and Fuller, L.G. (1987) Generalized individual-tree models for the Northeast. The Compiler 5(4), 17–27. Holdaway, M.R. (1984) Modeling the effect of competition on tree diameter growth as applied in STEMS. USDA Forest Service General Technical Report NC NC-94. Holdaway, M.R. and Brand, G.J. (1986) An evaluation of Lake States STEMS85. USDA Forest Service Research Paper NC NC-269. Hughes, G.A. and Sendak, P.E. (1985) Key algorithms used in GRO2: A computer simulation model for predicting tree and stand growth. USDA Forest Service Research Paper NE NE-570. Husch, B., Miller, C.I. and Beers, T.W. (1982) Forest Mensuration. John Wiley & Sons, New York. Johnson, P.S. (1975) Growth and structural development of red oak sprout clumps. Forest Science 21, 413–418. Johnson, P.S. (1979) Shoot elongation of black oak and white oak sprouts. Canadian Journal of Forest Research 9, 489–494. Johnson, P.S. and Rogers, R. (1980) Predicting growth of individual stems within northern red oak sprout clumps. Proceedings of the Central Hardwood Forest Conference III. University of Missouri, Columbia, pp. 420–439. Johnston, J.P. (1941) Height-growth periods of oak and pine reproduction in the Missouri Ozarks. Journal of Forestry 39, 67–68. Karnig, J.J. (1972) Nitrogen fertilizer increases diameter growth of northern red oak. Journal of Forestry 70, 629. Karnig, J.J. and Stout, B.B. (1969) Diameter growth of northern red oak following understory control. Harvard Black Rock Forest, Black Rock Forest Paper No. 30. Kershaw, J.A. and Fischer, B.C. (1985) Diameter growth of sawtimber size trees in Indiana. Purdue University Department of Forestry and Natural Resources Bulletin 181, pp. 21–24. Kowalski, D.G. and Gertner, G.Z. (1989) A validation of TWIGS for Illinois forests. Northern Journal of Applied Forestry 6, 154–156. Krajicek, J.E., Brinkman, K.A. and Gingrich, S.F. (1961) Crown competition – a measure of density. Forest Science 7, 35–42. Lamson, N.I. (1978) Fertilization increases growth of sawlog-size yellow-poplar and red oak in West Virginia. USDA Forest Service Research Paper NE NE-403. Lamson, N.I., Smith, H.C., Perkey, A.W. and Brock, S.M. (1990) Crown release increases growth of crop trees. USDA Forest Service Research Paper NE NE-635. Lamson, N.I. and Smith, H.C. (1978) Response to crop tree release: sugar maple, red oak, black cherry, and yellow poplar saplings in a 9-year-old stand. USDA Forest Service Research Paper NE NE-394.

Growth and Yield

465

Larsen, D.R. (1980) A growth and yield model for managed upland oak–shortleaf pine stands in Missouri. MS thesis, University of Missouri, Columbia. Larsen, D.R. and Hann, D.W. (1987) Height–diameter equations for seventeen tree species in southwest Oregon. Oregon State University Forestry Research Laboratory Research Paper 49. Leak, W.B. (1981) Do stocking guides in the eastern United States relate to stand growth? Journal of Forestry 79, 661–664. Lee, Y.J. (1971) Predicting mortality for even-aged stands of lodgepole pine. Forestry Chronicle 47, 29–32. Longman, K.A. and Coutts, M.P. (1974) Physiology of the oak tree. In: Morris, M.G. and Perring, F.H. (eds) The British Oak. E.W. Classey, Faringdon, UK, pp. 192–221. Loomis, R.M. (1973) Estimating ﬁre-caused mortality and injury in oak–hickory forests. USDA Forest Service Research Paper NC NC-94. Lowell, K.E. and Mitchell, R.J. (1987) Stand growth projection: simultaneous estimation of growth and mortality using a single probabilistic function. Canadian Journal of Forest Research 17, 1466–1470. Lowell, K.E., Mitchell, R.J., Johnson, P.S., Garrett, H.E. and Cox, G.S. (1987) Predicting growth and ‘success’ of coppice-regenerated oak stems. Forest Science 33, 740–749. Martin, A.J. (1981) Taper and volume equations for selected Appalachian hardwood species. USDA Forest Service Research Paper NE NE-490. McDonald, P.M. (1980) Growth of thinned and unthinned hardwood stands in northern Sierra Nevada – preliminary ﬁndings. USDA Forest Service General Technical Report PSW PSW-44, pp. 119–127. McGee, C.E. (1981) Response of overtopped white oak to release. USDA Forest Service Research Note SO SO-273. McGee, C.E. and Bivens, D.L. (1984) A billion overtopped white oak – assets or liabilities? Southern Journal of Applied Forestry 8, 216–220. McIntyre, A.C. (1933) Comparative radial growth of various oaks. Journal of Forestry 31, 341–342. McKnight, J.S. and Johnson, R.L. (1980) Hardwood management in southern bottomlands. Forest Farmer March, 31–35. McQuilkin, R.A. (1982) Fertilization of oaks in the Missouri Ozarks. Proceedings Central Hardwood Forests Conference IV. University of Kentucky, Lexington, pp. 79–92. McQuilkin, R.A. (1974) Site index prediction table for black, scarlet and white oaks in southeastern Missouri. USDA Forest Service Research Paper NC NC-108. Meldahl, R.S., Bolton, R.K. and Eriksson, M. (1988) Development of a mixed species projection system for southern forests. USDA Forest Service General Technical Report NC NC 120, pp. 102–109. Minckler, L.S. (1967) Release and pruning can improve growth and quality of white oak. Journal of Forestry 65, 654–655. Minckler, L.S. (1957) Response of pole-sized white oak trees to release. Journal of Forestry 55, 814–815. Miner, C.L., Walters, N.R. and Belli, M.L. (1988) A guide to the TWIGS program for the North Central United States. USDA Forest Service Research General Technical Report NC NC-125. Mitchell, R.J., Dwyer, J.P., Musbach, R.A., Garrett, H.E., Cox, G.S. and Kurtz, W.B. (1988) Crop tree release of a scarlet–black oak stand. Northern Journal of Applied Forestry 5, 96–99. Myers, C. and Belcher, D.M. (1981) Estimating total-tree height for upland oaks and hickories in southern Illinois. USDA Forest Service Research Note NC NC-272. Nyland, R.D. (1996) Silviculture: Concepts and Applications. McGraw-Hill, New York. Patton, R.T. (1922) Red oak and white ash, a study of growth and yield. Harvard Forestry Bulletin No. 4. Pillsbury, N.H. and Joseph, J.P. (1991) Coast live oak thinning study in the central coast of California – ﬁfth-year results. USDA Forest Service General Technical Report PSW PSW-126, pp. 320–332. Plumb, T.R. and McDonald, P.M. (1981) Oak management in California. USDA Forest Service General Technical Report PSW PSW-54. Powell, D.S., Faulkner, J.L., Darr, D.R., Zhu, Z. and MacCleery, D.W. (1994) Forest resources of the United States. USDA Forest Service General Technical Report RM RM-234. Raile, G.K., Smith, W.B. and Weist, C.A. (1982) A net volume equation for Michigan’s upper and lower peninsulas. USDA Forest Service General Technical Report NC NC-80.

466

Chapter 10

Regelbrugge, J.C. and Smith, D.W. (1994) Postﬁre tree mortality in relation to wildﬁre severity in mixed oak forests in the Blue Ridge of Virginia. Northern Journal of Applied Forestry 11, 90–97. Reich, P.B., Teskey, R.O., Johnson, P.S. and Hinckley, T.M. (1980) Periodic root and shoot growth in oak. Forest Science 4, 590–598. Reynolds, M.R., Jr, Burkhart, H.E. and Daniels, R.F. (1981) Procedures for statistical validation of stochastic simulation models. Forest Science 27, 349–364. Ritchie, M.W. and Hann, D.W. (1987) Equations for predicting height to crown base for fourteen tree species in southwest Oregon. Oregon State University Forestry Research Laboratory Research Paper 50. Roach, B.A. and Gingrich, S.F. (1962) Timber management guide for upland central hardwoods. USDA Forest Service Central States Forest Experiment Station. Roach, B.A. and Gingrich, S.F. (1968) Even-aged silviculture for upland central hardwoods. USDA Forest Service Agriculture Handbook 355. Rogers, R. and Johnson, P.S. (1985) Rule thinning: a ﬁeld method for meeting stocking goals in oak stands. Proceedings Central Hardwood Forest Conference V. University of Illinois, UrbanaChampaign, pp. 106–110. Sander, I.L. (1977) Manager’s handbook for oaks in the north central states. USDA Forest Service General Technical Report NC NC-37. Schlesinger, R.C. (1976) Sixteen years of selection silviculture in upland hardwood stands. USDA Forest Service Research Paper NC NC-125. Schlesinger, R.C. (1978) Increased growth of released white oak poles continues through two decades. Journal of Forestry 76, 726–727. Schnur, G.L. (1937) Yield, stand, and volume tables of even-aged upland oak forests. USDA Technical Bulletin 560. Schuler, T.M., Marquis, D.A., Ernst, R.L. and Simpson, B.T. (1993) Test of four stand growth simulators for the northeastern United States. USDA Forest Service Research Paper NE NE-676. Shiﬂey, S.R. (1987) A generalized system of models forecasting central states tree growth. USDA Forest Service Research Paper NC NC-279. Shiﬂey, S.R. (1990) Analysis and modeling of patterns of forest ingrowth in the north central United States. PhD dissertation, University of Minnesota, St Paul, Minnesota. Shiﬂey, S.R. and Smith, W.B. (1982) Diameter growth, survival, and volume estimates for Missouri trees. USDA Forest Service Research Note NC NC-292. Shiﬂey, S.R., Ek, A.R. and Burk, T.E. (1993) A generalized methodology for estimating forest ingrowth at multiple threshold diameters. Forest Science 39, 776–798. Simpson, B.T., Kollasch, R.P., Twery, M.J. and Schuler, T.M. (1996) NED/SIPS user’s manual: northeast decision model stand inventory processor and simulator version 1.11. USDA Forest Service General Technical Report NE NE-205. Smith, D.M. (1962) The Practice of Silviculture, 7th edn. John Wiley & Sons, New York. Smith, D.M., Larson, B.C., Kelty, M.J. and Ashton, P.M.S. (1997) The Practice of Silviculture: Applied Forest Ecology. John Wiley & Sons, New York. Smith, W.B. (1983) Adjusting the STEMS regional forest growth model to improve local predictions. USDA Forest Service Research Note NC NC-297. Smith, W.B. and Shiﬂey, S.R. (1984) Diameter growth, survival and volume estimates for trees in Indiana and Illinois. USDA Forest Service Research Paper NC NC-257. Smith, W.B. and Weist, C.A. (1982) A net volume equation for Indiana. USDA Forest Service Research Bulletin NC NC-63. Sonderman, D.L. (1984) Quality response of even-aged 80-year-old white oak trees after thinning. USDA Forest Service Research Paper NE NE-543. Sonderman, D.L. and Rast, E.D. (1988) Effect of thinning on mixed-oak stem quality. USDA Forest Service Research Paper NE NE-618. Stage, A.R. (1973) Prognosis model for stand development. USDA Forest Service General Technical Report INT INT-137. Stage, A.R. (1982) Use of self-calibration procedures to adjust general regional yield models to local conditions. Proceedings of Forest Resource Inventory, Growth Models, Management Planning, and Remote Sensing, IUFRO World Congress 5.4.01, Niigata, Japan, pp. 365–375. Standiford, R.B. (1997) Growth of blue oak on California’s hardwood rangelands. USDA Forest Service General Technical Report PSW PSW-160, pp. 169–176.

Growth and Yield

467

Starkey, D.A., Oak, S.W., Ryan, G.W., Taintner, F.H., Redmond, C. and Brown, H.D. (1989) Evaluation of oak decline areas in the South. USDA Forest Service Southern Region Protection Report R8PR R8-PR 17. Stout, B.B. and Shumway, D.L. (1982) Site quality estimation using height and diameter. Forestry Science 28, 639–645. Sullivan, A.D., Matney, T.G. and Hodges, J.D. (1983) Variable density yield tables for red oak–sweetgum stands. USDA Forest Service General Technical Report SE SE-24, pp. 298–301. Teck, R.M. (1990) NE-TWIGS 3.0: an individual-tree growth and yield projection system for the northeastern United States. The Compiler 8(1), 25–27. Teck, R.M. (1995) Forest vegetation simulator submittal system users guide. USDA Forest Service WO Service Center, Fort Collins, CO. Teck, R.M. and Hilt, D.E. (1991) Individual-tree diameter growth model for the northeastern United States. USDA Forest Service Research Paper NE NE-649. Teck, R.M., Moeur, M. and Eav, B. (1996) Forecasting ecosystems with the Forest Vegetation Simulator. Journal of Forestry 94(12), 7–10. Teck, R.M., Moeur, M. and Adams, J. (comps.) (1997) Proceedings Forest Vegetation Simulator Conference. USDA Forest Service General Technical Report INT INT-373. Teskey, R. (1978) Inﬂuence of temperature and moisture on root growth of white oak. MS thesis, University of Missouri, Columbia. Trimble, G.R., Jr (1960) Relative diameter growth rates of ﬁve upland oaks in West Virginia. Journal of Forestry 58, 111–115. Trimble, G.R., Jr (1969) Diameter growth of individual hardwood trees. USDA Forest Service Research Note NE NE-145. Trimble, G.R., Jr (1970) Twenty years of intensive uneven-aged management: effect on growth, yield, and species composition in two hardwood stands in West Virginia. USDA Forest Service Research Paper NE NE-154. Trimble, G.R., Jr (1974) Response to crop-tree release by 7-year-old stems of red maple stump sprouts and northern red oak advance reproduction. USDA Forest Service Research Paper NE NE-303. Trimble, G.R., Jr and Mendel, J.J. (1969) The rate of value increase for northern red oak, white oak, and chestnut oak. USDA Forest Service Research Paper NE NE-129. USDA Forest Service (1979) A generalized forest growth projection system applied to the Lake States region. USDA Forest Service General Technical Report NC NC-49. Utz, K.A. and Sims, D.H. (1981) Investment analysis of upland oak stands. USDA Forest Service Southeastern Area State and Private Forestry Publication SA-FR 12. Walters, D.K. and Ek, A.R. (1993) Whole stand yield and density equations for fourteen forest types in Minnesota. Northern Journal of Applied Forestry 10, 75–85. Wang, C. and Hann, D.W. (1988) Height-diameter equations for sixteen tree species in the central western Willamette Valley of Oregon. Oregon State University Forestry Research Laboratory Research Paper 51. Ward, J.S. (1995) Intensity of precommercial crop tree release increases diameter and crown growth in upland hardwoods. USDA Forest Service General Technical Report NE NE-197, pp. 388–398. Ward, W.W. and Bowersox, T.W. (1970) Upland oak response to fertilization with nitrogen, phosphorous, and calcium. Forest Science 16, 113–120. Wargo, P.M. and Haack, R.A. (1991) Understanding the physiology of dieback and decline diseases and its management implications for oak. Proceedings The Oak Resource in the Upper Midwest Conference. University of Minnesota, St Paul, pp. 147–158. Wendel, G.W. and Lamson, N.I. (1987) Effects of herbicide release on the growth of 8- to 12-year-old hardwood crop trees. USDA Forest Service Research Paper NE NE-598. Williams, T.B. and Wiant, H.V., Jr (1994) Evaluation of nine taper systems for four Appalachian hardwoods. Northern Journal of Applied Forestry 11, 24–25. Wycoff, W.R., Crookston, N.L. and Stage, A.R. (1982) User’s guide to the stand Prognosis Model. USDA Forest Service General Technical Report INT INT-133. Yaussy, D.A. and Dale, M.E. (1991) Merchantable sawlog and bole-length equations for the northeastern United States. USDA Forest Service Research Paper NE NE-650. Zasada, J.C. and Zahner, R. (1969) Vessel element development in the earlywood of red oak (Quercus rubra). Canadian Journal of Botany 47, 1965–1971. Zeide, B. (2001) Thinning and growth: a full turnaround. Journal of Forestry 99, 20–25.

Appendix 1 Common and Scientific Names of Species

Common name

Scientific name

Plants Ajo oak alternate-leaf dogwood American basswood American beech American chestnut American elder American elm American hazel American holly American hornbeam American sycamore apple Arizona white oak Arkansas oak ash baldcypress balsam fir beaked hazel bear oak bigleaf magnolia bigleaf maple bigtooth aspen bishop pine bitternut hickory black ash black birch black cherry blackgum, black tupelo blackhaw black hickory black locust black oak black walnut blackjack oak

Quercus turbinella var. ajoensis (C. H. Muller) Little Cornus alternifolia L. Tilia americana L. Fagus grandifolia Ehrh. Castanea dentata (Marsh.) Borkh. Sambucus canadensis L. Ulmus americana L. Corylus americana Walt. Ilex opaca Ait. Carpinus caroliniana Walt. Platanus occidentalis L. Malus spp. Quercus arizonica Sarg. Quercus arkansana Sarg. Fraxinus spp. Taxodium distichum (L.) Rich. Abies balsamea (L.) Mill. Corylus cornuta Marsh. Quercus ilicifolia Wangenh. Magnolia macrophylla Michx. Acer macrophyllum Pursh Populus grandidentata Michx. Pinus muricata D. Don Carya cordiformis (Wangenh.) K. Koch Fraxinus nigra Marsh. see sweet birch Prunus serotina Ehrh. Nyssa sylvatica Marsh. Viburnum prunifolium L. Carya texana Buckl. Robinia pseudoacacia L. Quercus velutina Lam. Juglans nigra L. Quercus marilandica Muenchh.

468

Common and Scientific Names

Common name

Scientific name

Plants (continued) blue oak blueberry bluejack oak boxelder Brewer oak briar bristlecone fir bur oak butternut California black oak California buckeye California coast live oak California live oak California red fir California scrub oak California white fir California white oak canyon live oak Chapman oak cherrybark oak chestnut oak chinkapin oak Chisos oak chokecherry coast live oak common juniper common persimmon cork oak Coulter pine cucumber tree deciduous holly Delta post oak diamondleaf oak Digger pine Douglas-fir downy oak Dunn oak Durand oak dwarf chinkapin oak dwarf post oak eastern cottonwood eastern hemlock eastern hophornbeam eastern redcedar eastern white pine elder Emory oak Engelmann oak

Quercus douglasii Hook. & Arn. Vaccinium spp. Quercus incana Bartr. Acer negundo L. see Oregon white oak Rubus spp. Abies bracteata D. Don ex Poiteau Quercus macrocarpa Michx. Juglans cinerea L. Quercus kelloggii Newb. Aesculus californica (Spach) Nutt. see California live oak Quercus agrifolia Née Abies magnifica A. Murr. Quercus dumosa Nutt. Abies concolor var. Iowiana (Gord.) Lemm. Quercus lobata Née Quercus chrysolepis Liebm. Quercus chapmanii Sarg. Quercus falcata var. pagodaefolia Ell. Quercus prinus L. Quercus muehlenbergii Engelm. Quercus graciliformis C. H. Muller Prunus virginiana L. see California live oak Juniperus communis L. Diospyros virginiana L. Quercus suber L. Pinus coulteri D. Don Magnolia acuminata L. see possumhaw Quercus stellata var. paludosa Sarg. see laurel oak Pinus sabiniana Dougl. Pseudotsuga menziesii (Mirb.) Franco Quercus pubescens Willd. Quercus dunnii Kellogg Quercus durandii Buckl. Quercus prinoides Willd. see sand post oak Populus deltoides Bartr. ex Marsh. Tsuga canadensis (L.) Carr. Ostrya virginiana (Mill.) K. Koch Juniperus virginiana L. Pinus strobus L. Sambucus spp. Quercus emoryi Torr. Quercus engelmannii Greene

469

470

Appendix 1

Common name

Scientific name

Plants (continued) English oak European turkey oak flowering dogwood Fraser fir Gambel oak Georgia oak giant chinkapin goldenrod grape green ash greenbriar grey birch grey dogwood grey oak hackberry Havard oak hawthorn hazel hickory Hinckley oak Holm oak honeylocust horsechestnut huckleberry oak incense-cedar interior live oak interrupted fern ironwood island live oak jack pine Jeffrey pine Kentucky coffeetree Kermes oak knobcone pine Lacey oak lateleaf oak laurel oak live oak loblolly pine longleaf pine McDonald oak Mexican blue oak mockernut hickory Mohr oak Monterey pine mountain-ash mountain maple mulberry myrtle oak

Quercus robur L. Quercus cerris L. Cornus florida L. Abies fraseri (Pursh) Poir. Quercus gambelii Nutt. Quercus georgiana M. A. Curtis Castanopsis chrysophylla (Dougl.) A. DC. Solidago spp. Vitis spp. Fraxinus pennsylvanica Marsh. Smilax spp. Betula populifolia Marsh. Cornus racemosa Lam. Quercus grisea Liebm. Celtis occidentalis L. Quercus havardii Rydb. Crataegus spp. Corylus spp. Carya spp. Quercus hinckleyi C. H. Muller Quercus ilex L. Gleditsia triacanthos L. Aesculus hippocastanum L. Quercus vaccinifolia Kellogg Libocedrus decurrens Torr. Quercus wislizenii A. DC. Osmunda claytoniana L. see American hornbeam Quercus tomentella Engelm. Pinus banksiana Lamb. Pinus jeffreyi Grev. & Balf. Gymnocladus dioicus (L.) K. Koch Quercus coccifera L. Pinus attenuata Lemm. Quercus glaucoides Mart. & Gal. Quercus tardifolia C. H. Muller Quercus laurifolia Michx. Quercus virginiana Mill. Pinus taeda L. Pinus palustris Mill. Quercus macdonaldii Greene Quercus oblongifolia Torr. Carya tomentosa (Poir.) Nutt. Quercus mohriana Buckl.ex Rydb. Pinus radiata D. Don Sorbus spp. Acer spicatum Lam. Morus spp. Quercus myrtifolia Willd.

Common and Scientific Names

Common name

Scientific name

Plants (continued) nannyberry netleaf oak northern catalpa northern pin oak northern red oak Norway maple Nuttall oak Oglethorpe oak Ohio buckeye Oregon ash Oregon white oak overcup oak Pacific dogwood Pacific madrone paper birch pawpaw pear pecan pin cherry pine pin oak pitch pine ponderosa pine possumhaw post oak quaking aspen redbud red maple red mulberry red-osier dogwood red pine rhododendron river birch roundleaf dogwood sand live oak sand post oak sandpaper oak sassafras sarsaparilla sawtooth oak scarlet oak Sebastian bush sedges serviceberry sessile oak shagbark hickory sheep-laurel shin oak shingle oak

Viburnum lentago L. Quercus rugosa Née Catalpa speciosa Warder ex Englem. Quercus ellipsoidalis E. J. Hill Quercus rubra L. Acer platanoides L. Quercus nuttallii Palmer Quercus oglethorpensis Duncan Aesculus glabra Willd. Fraxinus latifolia Benth. Quercus garryana Dougl. ex Hook. Quercus lyrata Walt. Cornus nuttallii Audubon Arbutus menziesii Pursh Betula papyrifera Marsh. Asimina triloba (L.) Dunal Pyrus spp. Carya illinoensis (Wangenh.) K. Koch Prunus pensylvanica L. f. Pinus spp. Quercus palustris Muenchh. Pinus rigida Mill. Pinus ponderosa Dougl. ex Laws. Ilex decidua Walt. Quercus stellata Wangenh. Populus tremuloides Michx. Cercis canadensis L. Acer rubrum L. Morus rubra L. Cornus stolonifera Michx. Pinus resinosa Ait. Rhododendron spp. Betula nigra L. Cornus rugosa Lam. Quercus virginiana var. geminata (Small) Sarg. Quercus stellata var. margaretta (Ashe) Sarg. Quercus pungens Liebm. Sassafras albidum (Nutt.) Nees Aralia spp. Quercus accutissima Carruthers Quercus coccinea Muenchh. Sebastiana fruiticosa Carex spp. Amelanchier spp. Quercus petraea (Mattuschka) Carya ovata (Mill.) K. Koch Kalmia angustifolia L. see Mohr oak Quercus imbricaria Michx.

471

472

Appendix 1

Common name

Scientific name

Plants (continued) shortleaf pine Shumard oak silverleaf oak silver maple slash pine slippery elm sourwood southern California walnut southern magnolia southern red oak speckled alder spicebush spruce spruce pine striped maple sugar maple sugar pine sugarberry swamp chestnut oak swamp hickory swamp tupelo swamp white oak sweet birch sweet cherry sweetfern sweetgum sycamore Table Mountain pine tamarack tanoak Texas live oak Texas oak Toumey oak turbinella oak turkey oak umbrella magnolia valley oak Vasey oak viburnum Virginia pine water hickory water oak water tupelo waterlocust wavyleaf oak white ash white oak wild grape willow

Pinus echinata Mill. Quercus shumardii Buckl. Quercus hypoleucoides A. Camus Acer saccharinum L. Pinus elliottii Engelm. Ulmus rubra Muhl. Oxydendrum arboreum (L.) DC. Juglans californica Wats. Magnolia grandiflora L. Quercus falcata Michx. Alnus rugosa (Du Roi) Spreng. Lindera benzoin (L.) Blume Picea spp. Pinus glabra Walt. Acer pensylvanicum L. Acer saccharum Marsh. Pinus lambertiana Dougl. Celtis laevigata Willd. Quercus michauxii Nutt. see bitternut hickory Nyssa sylvatica var. biflora (Walt.) Sarg. Quercus bicolor Willd. Betula lenta L. Prunus avium (L.) L. Comptonia peregrina (L.) Coult. Liquidambar styraciflua L. see American sycamore Pinus pungens Lamb. Larix laricina (Du Roi) K. Koch Lithocarpus densiflorus (Hook. & Arn.) Rehd. Quercus virginiana var. fusiformis (Small) Sarg. Quercus shumardii var. texana (Buckl.) Ashe Quercus toumeyi Sarg. Quercus turbinella Greene Quercus laevis Walt. Magnolia tripetala L. see California white oak Quercus pungens var. vaseyana (Buckl.) C. H. Muller Viburnum spp. Pinus virginiana Mill. Carya aquatica (Michx. f.)Nutt. Quercus nigra L. Nyssa aquatica L. Gleditsia aquatica Marsh. Quercus undulata Torr. Fraxinus americana L. Quercus alba L. Vitis spp. Salix spp.

Common and Scientific Names

Common name

Scientific name

Plants (continued) willow oak winged elm witch-hazel yellow birch yellow buckeye yellow-poplar

Quercus phellos L. Ulmus alata Michx. Hamamelis virginiana Betula alleghaniensis Britton Aesculus octandra Marsh. Liriodendron tulipifera L.

Mammals bear bison Botta pocket gopher chipmunk (eastern) deer deer mouse elk flying squirrels foxes fox squirrel grey squirrel mice pocket gophers rabbits raccoon red squirrel shrews squirrels voles white-footed mouse

Family Ursidae Bison spp. Thomomys bottae Tamias striatus Odocoileus spp. Peromyscus maniculatus Cervus spp. Glaucomys spp. Vulpes vulpes and Urocyon cinereoargenteus Sciurus niger Sciurus carolinensis Families Muridae and Cricetidae Family Geomyidae Sylvilagus spp., Lepus spp. Procyron lotor Tamiasciurus hudsonicus Sorex spp. Sciurus spp. and Tamiasciurus spp. Microtus spp. Peromyscus leucopus

Birds Acorn woodpecker American crow Arizona (Strickland’s) woodpecker Band-tailed pigeon black-billed cuckoo blue jay California quail Clark’s nutcracker common grackle golden-fronted woodpecker greater prairie chicken hairy woodpecker hooded merganser Lewis’s woodpecker mallard mourning dove northern (common) bobwhite

Melanerpes formicivorus Corvus brachyrhynchos Picoides stricklandii Columba fasciata Coccyzus erythropthalmus Cyanocitta cristata Callipepla californica Nucifraga columbiana Quiscalus quisula Melanerpes aurifrons Tympanuchus cupido Picoides villosus Lophodytes cucullatus Melanerpes lewis Anas platyrhynchos Zenaida macroura Colinus virginianus

473

474

Appendix 1

Common name

Scientific name

Birds (continued) northern (common) flicker Nuttall’s woodpecker pileated woodpecker plain titmouse quail red-bellied woodpecker red-headed woodpecker ring-necked pheasant ruffed grouse scrub jay sharp-tailed grouse Stellar’s jay tufted titmouse turkey white-breasted nuthatch wild turkey wood duck woodpeckers Yellow-billed magpie

Colaptes auratus Picoides nuttallii Dryocopus pileatus Parus inornatus Family Phasianidae Melanerpes carolinus Melanerpes erythrocephalus Phasianus colchicus Bonasa umbellus Aphelocoma coerulescens Pedioecetes phasianellus Cyanocitta stelleri Parus bicolor Family Meleagrididae Sitta carolinensis Meleagris gallopavo Aix sponsa Family Picidae Pica nuttalli

Insects acorn weevils acorn gall wasps acorn moth canker worm elm spanworm filbertworm moth filbert weevil gypsy moth Karner blue butterfly nitidulid beetles nitidulid sap beetles (on acorns) oak bark beetles tachinid flies twolined chestnut borer

Stelidota octomaculata and S. ferruginea Pseudopityophthorus spp. Family Tachinidae Agrilus bilineatus

Fungi (tree pathogens) Armillaria root disease chestnut blight oak wilt

Armillaria spp. Cryphonectria parasitica Ceratocystis fagacearum

Curculio spp. and Conotrachelus spp. Cynips spp. and Callirhytis spp. Valentinia glandulella Alsophila pometaria, Paleacrita vernata Ennomos subsignarius Cydia latiferreana (formerly Melissopus latiferreanus) Curculio occidentis Lymantria dispar Melissa samuelis Family Nitidulidae

Appendix 2 Forest Cover Types of Eastern United States Dominated by Oaks or Oaks Mixed with Other Speciesa

475

476

White pine–northern red oak–red maple (20)

E. white pine, n. red oak, red maple/ white ash, e. hemlock, birches, black cherry, American basswood, sugar maple, American beech

S. New England, cent. NY, PA, Lake States, s. ON [211, 221a, M221]

Moderately dry to mesic sites; early to midsuccessional; more persistent on dry sites

White pine–chestnut oak (51)

E. white pine, chestnut oak/scarlet, white, post and black oaks; hickories, blackgum; sourwood; red maple; pitch, Table Mountain, Virginia, and shortleaf pines; yellow-poplar; black cherry

Appalachian region from WV to GA, sw. VA, e. TN, w. NC [221a, M221, 232]

Dry to mesic sites; precipitation: 35–80 inches; mostly at 1200–3600 ft elevation; physiographic climax on dry sites

Other northern types Northern pin oak (14) N. pin, white, black, bur, and n. red oaks; jack pine/red pine, e. white pine, quaking and bigtooth aspens, red maple, black cherry

cent. and n. Lower MI, cent. WI, n. WI, e. cent. MN [211, 221b]

Extremely dry sites with acid sandy soils; n. pin oak may be successional to associated oaks when present

Post oak–blackjack oak (40)

Post and blackjack oaks/hickories; black, scarlet, bluejack, s. red, shingle, white, and turkey oaks; shortleaf and Virginia pines; e. redcedar; and others

E. KS southward to TX and eastward to Atlantic Coastal Plain [221a, 221b, M221, M222, 231, 232, 251, 252]

Primarily dry sites where it forms an edaphic climax; subclimax elsewhere

Bur oak (42)

Bur oak/Uplands: n. pin, black, chinkapin and white oaks; shagbark hickory. Bottomlands: hickories, black walnut, e. cottonwood, white ash, American elm, green ash, and others

Sporadic in the prairie–forest transition zone from s. Canada southward to n. MO; westward to SD and WY along rivers [251, 331, 332]

Dry uplands and moist benches on river bottoms; successional to associated species in bottomlands; persistence in uplands depends on fire or other disturbance

Bear oak (43)

Bear oak/scarlet, chestnut, white, black, blackjack, post, and n. red oaks; e. white, Virginia, and shortleaf pines; quaking and bigtooth aspens; sassafras; red maple; and others

Coastal Plain from New England southward to NJ; sporadic in w. VA and e. WV [211, 221a, M221]

Tall shrub-dwarf tree type restricted to poor, dry sites; successional to associated species, depending on fire frequency

Type group

Northern

Pine and hemlock types

Upland oak type

Ecological relations

Appendix 2

Type species/ common associatesb

Region

Central

Geographic distribution (ecoregion province number)c

Type name (number)

Chestnut oak/bear, n. red., s. red, black, post, scarlet, and white oaks; hickories; yellow-poplar; blackgum; sweetgum; black cherry; red and sugar maples; and others

AL and GA northward to NJ, NY, and southern New England [211, M211b, 221a, 231, 232]

Mostly on dry Appalachian uplands from 1500 to 4600 ft where it may form a physiographic climax

White oak–black oak–northern red oak (52)

White oak, black oak, and n. red oak/n. pin, scarlet, s. red, chinkapin, post, and blackjack oaks; hickories; yellow-poplar; blackgum; sugar and red maples; white and green ashes; and others

Great Plains eastward to the Atlantic Coast; Lake States, s. ON., and s. New England southward to Coastal Plain [211, M221, M222, 231, M231, 232, M211b]

Dry to mesic sites; ranges from subclimax to climax depending on site factors

White oak (53)

White oak/n. red, scarlet, black, chestnut, and bur oaks; hickories; blackgum; yellow-poplar; white ash; maples; and others

Across the range of white oaks from MN southward to TX, eastward to the Atlantic Coastal Plain, and north to New England [211, M221, M222, 231, M231, 232, M211b, 252]

Dry to mesic sites; ranges from subclimax to climax depending on site factors

Black oak (110)

Black oak/white, post, blackjack, scarlet, n. red, and chestnut oaks; hickories; blackgum; shortleaf and loblolly pines; yellow-poplar; and others

Central States, Ozark Highlands, n. IN, Lower MI [221a, 221b, 251]

Dry to mesic sites; successional to associated oaks on dry sites and other species on mesic sites

MN, WI, and MI; sporadic in New England and Appalachians from PA southward [211, 221a, 221b, M221]

Moderately dry to mesic sites; usually successional to associated species

s. NY, PA, and sw. New England southward along the Appalachians; also Cumberland and Allegheny Mountains [211, 221a, M221]

From 500 to 4500 ft; successional to mixed oak or shade-tolerant non-oaks depending on site quality

Northern red oak (55) Northern red oak/yellow-poplar, black cherry, sugar maple, white ash, white oak, American beech, and others Other central types

Yellow-poplar– white oak–northern red oak (59)

Yellow-poplar, white oak, n. red oak/ maples, white ash, black cherry, American beech, e. hemlock, black walnut, and others

477

Continued

Forest cover types of eastern United States dominated by oaks or oaks mixed with other species

Chestnut oak (44)

Oak–pine types

Bottomland types

Type species/ common associatesb

Geographic distribution [ecoregion province number]c

Pin oak–sweetgum (65)

Pin oak, sweetgum/red maple, American elm, blackgum, swamp white oak, willow oak, overcup oak, swamp chestnut oak, green ash, hickories, and others

Ohio River Valley from WV through OH, s. IN, s. IL, KY, w. TN; southward from se. MO to central AR [221a, 221b, 232]

Longleaf pine– scrub oak (71)

Shortleaf pine with some combination of white, s. red, black, post, blackjack, and scarlet oaks/blackgum, red maple, hickories, and others

Subclimax type maintained Piedmont, Cumberland Plateau, s. Appalachians below by fire 2000 ft, ne., and n. cent. MS, LA, AR, s. MO, ne. TX, e. cent. OK [221b, M222, M231, 232, 252] .

Shortleaf pine– oak (76)

Shortleaf pine with some combination of white, s. red, black, post, blackjack, and scarlet oaks/ blackgum, red maple, hickories, and others

Piedmont, Cumberland Plateau, s. Appalachians below 2000 ft, ne. and n. cent. MS, LA, AR, s. MO, ne. TX, e. cent. OK [221b, M222, M231, 232, 252]

Dry to moderately dry sites; potentially successional to associated oaks depending on frequency and intensity of disturbance

Virginia pine–oak (78)

Virginia pine and some combination of s. red, scarlet, black, chestnut, white, post, and blackjack oaks/ shortleaf and pitch pines, dogwood, yellow-poplar, blackgum, red maple, hickories, and others

s. PA southward to GA and AL along the foothills of s. Appalachian Mts and Piedmont [221a, M221, 232]

Primarily on old fields of low to moderate site quality; successional to oaks or shade tolerant hardwoods depending onsite quality

Willow oak–water oak–diamondleaf (laurel) oak (88)

Willow, water, and diamondleaf oaks, Nuttall and cherrybark oaks, red maple, green ash, sweetgum, swamp hickory, spruce and loblolly pines, and others

Coastal Plain from se. VA to w. FL westward to e. TX [231]

Most common on flat, poorly drained alluvial floodplains; may be a topographic/ edaphic climax on some sites

Live oak (89)

Live oak/water oak, southern magnolia, sugarberry, American elm, and green ash

s. LA and sw. MS [231]

Narrow belts <1 mile wide in lowest bottoms; may be a topographic/edaphic climax on some sites

Ecological relations Bottomlands; successional to associated species

Appendix 2

Southern

Type group

478

Region

Type name (number)

Highest first-bottom ridges; may be climax on older alluvium

Swamp chestnut and cherrybark oaks/ green ash, hickories, white oak, Delta post oak, Shumard oak, and blackgum

Sweetgum–willow oak (92)

Sweetgum and willow oak/sugarberry, Alluvial floodplains of major green ash, America elm, and Nuttall rivers in AR, LA, MS, AL, e. oak MO, e. TX [231, 232]

Overcup oak– water hickory (96)

Overcup oak and water hickory/ green ash, sugarberry, America elm, waterlocust, red maple, and Nuttall oak

Floodplains of the Gulf and s. Atlantic states; also TN and s. IL [211b, 231, 232]

Floodplains where water stands into the growing season; clay or silty clay soils; poor site quality; permanence depends on poor drainage

Mohrs (‘shin’) oak (67)

Mohrs (shin) oak in mixture with other west Texas oaks

w. cent. TX to sw. OK; sporadic in Mexico [313, 314, 321]

Primarily a shrub type on calcareous soils; the type is best represented where precipitation is 20–25 inches/year

Southern scrub oak (72)

Mixtures of turkey, bluejack, blackjack, sand post, sand live, live, and myrtle oaks

se. Coastal Plain [231]

Dry, infertile, sandy soils formerly occupied by longleaf pine or longleaf pine–scrub oak types; may be self-perpetuating with periodic fire

Self-perpetuating on firstbottom ridges of alluvial floodplains

a Adapted from Eyre (1980) Forest Cover Types of the United States and Canada. Society of American Foresters, Washington, DC. Many other cover types not listed also include oaks as associated species. b The species for which this type is named comprises 50% or more of stand basal area or crown cover; importance of associated species is often highly variable. c Numbers in brackets are ecoregion provinces (Fig. 1.2).

Forest cover types of eastern United States dominated by oaks or oaks mixed with other species

Other southern types

Sporadic over small areas in southern floodplains of major rivers [231, 232]

Swamp chestnut oak–cherrybark oak (91)

479

Appendix 3 Forest Cover Types of Western United States Dominated by Oaks or Oaks Mixed with Other Speciesa

480

North Pacific

Low elevations, interior

Type species/ common associatesb

Geographic distribution (ecoregion province)c

Oregon white oak (233)

Oregon white oak/Douglas-fir, bigleaf maple, Oregon ash, ponderosa pine, incense-cedar, California black oak, Pacific madrone, canyon live oak and tanoak

se. Vancouver Island in BC southward through w. WA, w. OR, n. CA to Santa Cruz Mts [241]

Primarily lower slopes between the Coast and Cascade or Sierra Nevada ranges; forms open savannas and pure closed-canopy stands as well as mixtures; successional to conifers

Douglas-fir–tanoak– Pacific madrone (234)

Douglas-fir, tanoak, Pacific madrone/ canyon live oak, sugar pine, ponderosa pine, giant chinkapin, California black oak, Oregon white oak and others

Coast Ranges of n. and central CA and sw. OR at 500–4000 ft [M261, 262, M262, 263]

Best developed in mild, moist climate; complex successional relations determined by disturbance, species composition and site factors

Bur oak (236)

Similar to type 42 except associates in w. edge of range include ponderosa pine

Along streams in NE, SD, ND and adjacent Canada; extends westward into Black Hills of SD [331, 332]

Primarily stream bottoms; successional to associated hardwoods

California white fir, ponderosa pine, sugar pine, incense-cedar, California black oak, Douglas-fir/California red fir, Pacific madrone, tanoak and Pacific dogwood

W. side of Sierra Nevada range from 3000 to 6000 ft; locally on e. side of Sierra Nevada Range [M261]

Successional status depends on stand composition, associated species, site and disturbance regime

California black oak (246)

California black oak/ponderosa pine, Douglas-fir, incense cedar, knobcone pine, Pacific madrone, tanoak, Oregon white oak, canyon live oak and others

Sporadically in small areas from cent. OR southward through Cascade, Sierra Nevada, and Coast Ranges to near Mexican border [M261, 262, M262]

Persistent subclimax in its optimal range of 1500–3000 ft and 30–50 inches precipitation; elsewhere successional to associated species where it needs fire or other disturbance to hold its own

Canyon live oak (249)

Canyon live oak/Douglas-fir; ponderosa, Jeffrey, sugar, and Coulter pines; bristlecone fir

Northern Cascade, Sierra Nevada, and Coast Ranges southward to extreme s. CA [M241, M261, 262, M262, 263, M263]

From sea level to 9000 ft on steep canyon slopes and drier canyon bottoms; forms relatively stable communities of mixed to pure stands

South Pacific Sierra Nevada mixed conifer (243)

Ecological relations

Continued

481

Type name (number)

Forest cover types of western United States dominated by oaks or oaks mixed with other species

Region

Type name (number)

Type species/ common associatesb

Geographic distribution [ecoregion provincec

Blue oak– digger pine (250)

Blue oak and digger pine/interior live, canyon live, California black, and California scrub, coast live, and valley oaks; California buckeye

Forms a nearly continuous belt around the Central Valley of California between valley grasslands and montane forests at 3000–5000 ft [M261, M262]

Climax type of the oak woodland or foothill woodland of California; commonly forms pure stands of blue oak especially at low elevations; stands variable in density

California coast live oak (255)

California coast live oak/Bishop, Monterey and knobcone pines; Engelmann, interior live, valley, canyon live, and blue oaks; tanoak; southern California walnut, Pacific madrone, bigleaf maple and others

Primarily on the w. side of the Coastal Range in CA; occasionally on E. slopes; up to 3000 ft in the north to 5000 ft in the south [M261, 262, M262, 263]

Tends to form climax stands that form dense, closed-canopy stands resembling a forest; becomes more savanna-like to the south

Western live oak (241)

Emory, Arizona white, Mexican blue, and silverleaf oaks (evergreen); may include some deciduous oaks

Found at 4000–6000+ ft elevation in s. New Mexico and s.e. and central Arizona [313, M313, 321, 322]

Climax type but can be replaced by chaparral on drier sites. Typically open woodland with dry-tropic shrubs, succulents, cacti and grasses. Makes best development on mesic semi-arid soils

Ecological relations

a Adapted from Eyre (1980) Forest Cover Types of the United States and Canada. Society of American Foresters, Washington, DC. Many other cover types not listed also include oaks as associated species. b The species for which this type is named comprises 50% or more of stand basal area or crown cover; importance of associated species is often highly variable. c Numbers in brackets are ecoregion provinces (Fig. 1.2).

Appendix 3

Desert Southwest

482

Region

Appendix 4 Conversion of Site Indexes for Indiana, Kentucky, Ohio and West Virginia

Formulae for converting site index (in feet at base age 50) of one species to another in unglaciated regions of Indiana, Kentucky, Ohio and West Virginia. Equations can be solved forwards or backwards because each was graphically derived to represent the average of the two equations representing each species pair (e.g. the equations for estimating northern red oak from yellow-poplar, and estimating yellow-poplar from northern red oak).a Convert the site index (SI) of the species below

To black oak (BO)

To white oak (WO)

To scarlet oak (SO)

To chestnut oak (CO)

Yellow-poplar (YP)

SIBO = 7.772 + 0.865SIYP

SIWO = 5.826 + 0.831SIYP

–

–

SIWO = 1.233 + 0.924SINRO

SISO = 3.455 + 0.976SINRO

SICO = 0.928 + 0.968SINRO

SISO = 0.858 + 1.037SICO

Northern red oak SIBO = 0.181 + (NRO) 1.018SINRO Chestnut oak (CO)

SIBO = 2.118 + 1.064SICO

SIWO = 2.752 + 1.036SICO

Scarlet oak (SO)

SIBO = 0.473 + 0.988SISO

SIWO = 1.032 + 0.941SISO

White oak (WO)

SIBO = 2.706 + 1.106SIWO

To northern red oak (NRO) SINRO = 1.310 + 0.954SIYP

aFrom: Carmean, W.H. and Hahn, J.T. (1983) Site comparisons for upland oaks and yellow-poplar in the central States. Journal of Forestry 81, 736–739.

483

Appendix 5 Converting Site Indexes for Four Regions

Formulae for converting site indexes (in feet at base age 50) for oaks and associated species from one species to another in four regions. Equations can be solved forwards or backwards.

To black oak (BO) in southeastern Missouric

To northern red oak (NRO) in northern Wisconsin and northern Michigand

SISO = 1.171 + 1.076SIWO

SIBO = 4 + SIWO

–

SIRO = 0.97SISO

–

SIBO = SISO – 3

–

Black oak (BO) SIRO = 1.03SIBO

–

–

–

Chestnut oak (CO)

SIRO = 1.02SICO

–

–

–

Yellow-poplar (YP)

–

SISO = 27.674 + 0.586SIYP

–

–

Shortleaf/ – pitch pine (SP)

SISO = 6.251 + 1.001SISP

–

–

American – basswood (AB)

–

–

SINRO = –26.909 + 1.406SIAB

American elm (AE)

–

–

–

SINRO = 17.252 + 0.713SIAE

Aspens (A)

–

–

–

SINRO = 1.756 + 0.872SIA

Black ash (BA) –

–

–

SINRO = 23.568 + 0.633SIBA

Paper birch (PB)

–

–

–

SINRO = –16.401 + 1.255SIPB

White ash (WA) –

–

–

SINRO = 22.442 + 0.646SIWA

Yellow birch (YB)

–

–

SINRO = 6.247 + 1.008SIYB

Convert the site index (SI) of the species below

To northern red oak (NRO) in the northern Appalachiansa

White oak (WO)

SINRO = 1.05SIWO

Scarlet oak (SO)

484

–

To scarlet/black/ chestnut/northern red oak group (SO) in the southern Appalachiansb

Converting Site Indexes for Four Regions

a From:

485

Trimble, G.R., Jr and Weitzman, S. (1956) Site index studies of upland oaks in the northern Appalachians. Forest Science 2, 162–173. b From: Doolittle, W.T. (1958) Site index comparisons for several forest species in the southern Appalachians. Soil Science Society of America Proceedings 22, pp. 455–458. c From: McQuilkin, R.A. (1976) The necessity of independent testing of soil-site equations. Soil Science Society of America Journal 40, 783–785. d From: Carmean, W.H., Clark, F.B., Williams, R.D. and Hannah, P.R. (1976) Hardwoods planted in old fields favored by prior tree cover. USDA Forest Service Research Paper NC NC-134.

Appendix 6 Converting Site Indexes, Yellow-poplar to Oak

Formulae for converting yellow-poplar site index to oak site indexes in the Virginia– Carolina Piedmont.a To convert yellow-poplar (YP) site index (SI) to that of the species below Black oak (BO)

Use this equation (where SI is in feet at base age 50) SIBO = 39.7 + 0.45SIYP

White and southern red oak group (WO)

SIWO = 36.7 + 0.45SIYP

Scarlet and northern red oak group (SO)

SISO = 44.5 + 0.45SIYP

a From: Olson, D.F., Jr and Della-Bianca, L. (1959) Site index comparisons for several tree species in the Virginia–Carolina Piedmont. USDA Forest Service Southeastern Forest Experiment Station Paper 104.

486

Appendix 7 Height/dbh Site Index Curves

Parameter estimates for site index asymptotes (S) and species coefﬁcients (b) for deriving height/dbh site index curves from Equation 4.1.a Site index class

S

b1

White oakb

40 50 60 70

66.68 75.53 82.57 92.88

0.123 0.123 0.123 0.123

Northern red oakb

50 60 70 80 90

74.28 81.00 89.09 95.50 102.20

0.12 0.12 0.12 0.12 0.12

Black oakb

50 60 70 80 90

84.72 90.10 97.53 100.9 108.3

0.086 0.086 0.086 0.086 0.086

Northern red oakc

60 70 80

79.45 93.331 111.848

0.137 0.109 0.098

Species

aBased

on Equation 4.1: H = 4.5 + S(1 – e–bD) where H = total tree height (ft), D = dbh (inches), S = coefficient for the site-specific asymptote, b = coefficient for the species-specific rate coefficient, e = base of the natural logarithm, 4.5 = correction for D measured at breast height. bFrom: Stout, B.B. and Shumway, D.L. (1982). Site quality estimation using height and diameter. Forest Science 28, 639–645. cFrom: Lamson, N.I. (1987) Estimating northern red oak site-index class from total height and diameter of dominant and co-dominant trees in central Appalachian hardwood stands. USDA Forest Service Research Paper NE NE-605.

487

Index

accumulation see oak reproduction ACORn see models, regeneration acorn crop see acorn production acorn insects 70–77 acorn gall wasps 71, 73–76 acorn weevils 71–73, 75–77 Conotrachelus weevils 75–77 Curculio weevils 71–73, 75 drippy nut disease 72 ﬁre as a regulator of 77 in the forest ﬂoor 77 moths 71–74 acorn moth 71, 73, 75, 77 ﬁlbertworm moth 71–74 nitidulid sap beetles 71, 76 primary invaders 72–75 secondary invaders 75–77 acorn predation and dispersal 69–85, 263 acorn moth 73, 75–76 as primary and secondary invader 76 animal-mediated dispersal 70 birds 81–84, 263 blue jay 82–83, 87, 263 granary trees 82, 84 list of dispersers and consumers 81 scrub jay 82 in semi-desert regions 83 Steller’s jay 82 tufted titmouse 82 woodpeckers 82–84 yellow-billed magpie 82 see also acorn insects burial by rodents 76 destructive organisms 70 dispersal (deﬁned) 70 fungi 70, 78

millipedes on chestnut oak acorns 76 and germinating acorns 76 on northern red oak acorns 76 on white oak acorns 76 rodents 77–81 acorn caching 77–78 acorn ‘notching’ by squirrels 78–79 acorn protection by burial 76, 78 acorn species preferences 77 and asynchronous acorn production 80 and autumn germination of acorns 78 blue oak 78 chipmunks 77 coast live oak woodland 78 cost of dispersal to oaks 77 effects on gypsy moth 79–80 gophers 78–79 larder hoarding 77 and Lyme disease 79 mice 77–79 predation on northern red oak 78 scatterhoarding 77 species’ preferences 77 squirrels 77 valley oak woodland 78 white-footed mice 77 seedling establishment 70 acorn production 61, 62, 64–69, 71, 257, 262–263, 278, 391–403 acorn clusters (photo) 62 acorn collection traps (photo) 71 assessing and predicting crops 392–395 frost damage 63 genetic control 61, 64–65, 67 guidelines for sustaining 399–403 in green tree reservoirs 400–403 managing stands for 391–403 489

490

Index

acorn production continued models 398–399 periodicity 65–69 premature abscission 62–64 in relation to dbh 397 site quality effects 61, 397–398 spatial variation within crowns 67–69 crown aspect 68 insect effects 68 shading 67 stand density effects 67 stand density effects 395–396 thinning effects 396–397 tree size effects 395–398 variation in 64–69 vertical crown stratiﬁcation effects 397 weather 61–62 advance reproduction inventory 256 in natural regeneration 255, 258–259, 262–263 see also oak reproduction, accumulation of; models, regeneration ADVREGEN see models, regeneration aesthetics corridors 415 forest opening size 415 group selection openings 371 human perceptions 410–411 landscape level 415 shape of openings 415 vegetation mosaics 415 viewsheds 415 visual variety 415 scenic beauty ratings 411 scenic quality of stands 411, 413–416 clearcuts 413 group selection 414 intermediate thinning 414 no harvesting 414 seed tree method 413 shelterwoods 413 single-tree selection 414 stand density 416 two-age management 413 species composition effects 412 stand-level aesthetics 411–415 timber harvesting effects 412 age distribution of trees correlation with diameter distribution 346–351 agroforestry 3 agronomic model 3, 173 Allegheny Plateau geographic location (map) 19 oak regeneration 272–273 regeneration models 271–272

allogenic change 13 alluvial soils 183 anthropocentrism 2–3 Appalachian hardwoods oak regeneration 268–269 stump sprouts 299 uneven-aged management 364, 371 area control 254 artiﬁcial regeneration see nursery practices; planting oaks autogenic change 13

biocentrism 2–3 biomass above-ground 170 below-ground 169, 171 birds see acorn predation and dispersal black oak–white oak–shortleaf pine type (photo) 35 black–northern red–white oak type (photo) 29 blue jays see acorn predation and dispersal blue oak woodland (photo) 48 bottomland oak forests regeneration 263–264 regeneration models 273–274 site quality 171 topographic distribution 36–37; (illustrated) 37 see also planting oaks buds female 55 male 55 red oak group 55 vegetative 55 on white oak 55 burning see ﬁre

callus tissue 100 cambial growth 425 canopy gaps crown expansion in northern red oak 198 formation and closure 148–150 light 198 in oak stands 197 see also crown closure rates; regeneration mode catkins see male ﬂowers Central Hardwood Region 19–20, 26–33 climate 26–27 forest history 27–28 geographic extent 26; (map) 19 oak distribution 28–33 physiography 26–27 thinning 29

Index

cherrybark oak–sweetgum type (photo) 36 cleaning 295–296 clearcutting method bottomland forests 263 Cumberland Plateau 271–272 dominance probability 267–268, 285–289 economic and social considerations 320–322 enrichment planting 284–285 natural regeneration 255–274 oak decline 314–315 oak regeneration requirements 255–265 Ozark Highlands 256–271, 286–293 regeneration guides 255, 272–273, 284 stand development 258 with reserves 256–257 see also advance reproduction; models, regeneration; natural regeneration methods; planting oaks cohort 121, 254–255; (deﬁned) 121 competition cherrybark oak and sweetgum 214–215 cherrybark oak and sycamore 214–215 grasses 121 non-oaks 120, 121 in western oak savannas 121 competitive sorting 209 complex stage and crown expansion 215 stand structure 199–202, 215–216, 220 vs. old-growth 216 coppice 169 see also stump sprouts; thinning, stump sprouts cover types 11, 15, 24–26, 28–36, 38–43, 45–48 distribution 14–15 eastern United States 475–479 western United States 480–482 crop trees 295–296, 298–299 Cross Timbers Region 39–40 geographic location (map) 19 crown class and diameter growth 425–426, 428 crown closure rates 149 northern red oak 149 yellow-poplar 149 crown competition factor 239, 246 crown dimensions, open-grown crown width equations 238–239 crown shape 214 crown thinning 294 Cumberland Plateau geographic location (map) 19 oak regeneration 271–272 regeneration models 271–272

491

site evaluation 187 dbh distribution see diameter distribution diameter distribution bell-shaped 205, 213 changes in 213, 214 correlation with age distribution 346–351 in even-aged stands 254 in Ozark Highlands 218 reverse J-shaped 205, 212, 213, 214, 221 see also single-tree selection method diameter frequency distribution see diameter distribution diameter growth by crown class 425–426, 428 by diameter class 427–428 models see models, growth and yield response to thinning 428–430 by site quality 427 by species 426, 429–430 stump sprouts 429–430 diameter-limit cutting 364–365 direct seeding 280–281 distribution of oaks see geographic distribution of oaks disturbance (deﬁned) 196 and topographic position 199 endogenous 196 exogenous 196 gap-scale 196, 197 incomplete stand-scale 196, 197 recovery from 195 size 196–199 spatial clustering of 199 stand-initiating 196 windthrow 195 yellow-poplar 198 disturbance–recovery cycles in black, scarlet and white oak stands 221 forest resilience 221 in Forest–Prairie Transition Region 219 gap-scale events 219 in oak chaparral 219 in oak shrub communities 219 in Ozark Highlands 219 Driftless Area see Wisconsin Driftless Area incomplete stand-scale events 217 stages of stand development 220 stand-initiating events 217 suspended in ‘oak scrub’ stage 219 division see ecoregions domain see ecoregions dominance probability 267–268, 285–289 see also models, regeneration; planting oaks drought 198

492

Index

drought tolerance of oaks 13, 91 species ranking 91

earlywood 425 ecological amplitude 12–13 ecological classiﬁcation 14–22, 30–31, 47, 171–172 accumulation types 131 herbaceous vegetation 172 oak regeneration 265 relation to site quality 171–172 reproduction density 132 see also ecroregions; ecological landtypes ecological landtype 16, 22, 30–31, 37, 47 ecology (deﬁned) xi, 1 ecoregions 14–48 ecosystem management 6–7 Emory oak woodland (photo) 43 epicormic branching 298, 303–304, 359, 369 even-aged stands see stand development, even-aged stands

Fagaceae 9 fertilization of soil 427–428, 446 ﬁre accumulation of oak reproduction 131, 138–139 acorn heat resistance 139 acorn–soil contact 138 as a regulator of insect populations 77 bark thickness 133 chestnut oak 133 damage 387 effect on oak reproduction 355, 357 effects on oaks 4 enrichment planting 284–285, 290, 293 ﬁre temperatures in pine litter 140 ﬁre-free interval 141, 384, 386 in Florida sandhills 139, 140, 141 fuels 386 grasses and forbs 133 habitat for rodents 138 height of reproduction 138 hickories 135 history 23–24, 27, 33–34, 38, 60–61, 65–68, 316, 383 in Tennessee 138 increased light 138 inﬂuence on oak distribution 4, 8, 24, 28–29, 32, 35, 38–40, 46, 48 longleaf pine 139, 141 non-oaks 136 in North Carolina 137 oak sprouting 99, 137, 140–141 oaks as ﬁre persistent 136

in old-growth forests 407–408 in Ozark Highlands 134 pine–oak stands 133 post oak 134–135, 439 prescribed burning 277, 284–285, 290, 293 probability pattern 384 protected root collars 137 red maple 133, 135 red oaks 134–135, 439 reduces overstorey density 133 reduction of leaf litter 138 root:shoot ratio 138 safe tree size 135 in savannas 380, 384, 388–389 scarlet oak 135 spatial variation of fuels 141 sprouting 136 top-kill 135–136 tree survival 134–141, 438–440 tree wounding 135, 139 turkey oak 139–140 type of fuels 141 white oak 439 widening oak regeneration window 141 xeriﬁcation 141 ﬂowers, female 58–61 aborted ovules 59 acorn cup 59–60 acorn enlargement 59–60 black oak 59, 64 bud swelling 58 cupule development 59–60 developmental process (illustration) 60 fertilization 59 Gambel oak 59 in Missouri 59 multiple-seeded acorns 59 northern red oak 59 ovule maturation 59 pollen tubes 59 position in crown 55 as preformed structures 59 primordial inﬂorescences 58 stigmas 59–60 time of appearance 59 white oak 59, 64 ﬂowers, male 55–58 development 56 frost effects 58 Gambel oak 58 maturation 57 meiosis 57 in Pennsylvania 56 pollen dispersal 57–58 position in crown 55 premature abscission 62–64

Index

relative humidity effects 57–58 topography effects 58 white oak group 55 wind effects 58 ﬂowers, pistillate see female ﬂowers ﬂowers, staminate see male ﬂowers FORCAT see models, regeneration forest canopy layers crown classes 194–195 and light interception 195 main canopy 195 overstorey (deﬁned) 195 subcanopy 195 understorey (deﬁned) 195 forest cover type see cover type forest management history 1–2 Forest–Prairie Transition Region 19, 36–40 climate 36 forest history 37–38 geographic extent 19, 36; (map) 19 oak distribution 38–40 physiography 36, (photo) 39 forest productivity see site productivity forest resilience 80, 221 forest type see cover type free thinning 294 FVS see models, growth and yield

geographic distribution of oaks 10–20, 24–26, 28–36, 38–43, 45–48; (map) 19 germination and seedling establishment 84–86 acorn desiccation 85–86 blue oak 84–85 California Coastal Range 86 chestnut oak 85 chilling requirements 84 and earthworms 86 embryo dormancy 84 epicotyl emergence 84–86 and forest ﬂoor characteristics 86 low temperatures 85–86 northern red oak (illustrated) 85 overwintering 84, 86 Ozark Highlands 86 predation by insects 85 predation by rodents 85 radicle development 84 red oaks 85 squirrel-buried acorns 86 snow cover 85 stages of seedling development 85 temperature requirements for 84–85 valley oak 84 white oak 84–85 winter acorn survival 85

493

green tree reservoirs see acorn production GROAK see models, growth and yield group selection method 365–374; (deﬁned) 365 crown closure 367 cutting cycle 371 economic and social considerations 372–374 enrichment planting 371 epicormic branching 369 examples Appalachians 371 New Hampshire 371 Ohio Valley 371 ﬁeld implementation area regulation 371 group openings 369–371 silvicultural guidelines 366–399 treatment between openings 371–372 inﬂuence zones of openings (illustrated) 370 light in openings 367–369 oak regeneration 365 oak reproduction 366–368 opening edge effects 367–370 opening size 365–370 stand structure 366 variants of the method 366, 371–373 visual impacts 365 GROW see models, growth and yield growth cambial 425 diameter see diameter growth root see root growth shoot see shoot growth stand see stand growth stump sprouts see stump sprouts volume see volume growth value see value growth growth curve 424 gypsy moth 79–80, 305–314 and Armillaria 314 and birds 314 control 79–80, 308–314 crown classes for risk ranking 311 defoliation 80, 308–310 economic effects 308 Entomophaga maimaiga 314 eradication 314 geographic distribution 306, 309 impact on trees 305–308 life cycle 305–307 risk rating stands 309–311 trees 308–312

494

Index

gypsy moth continued and rodents 314 silvicultural mitigation 308–314 presalvage cuttings 311–314 salvage cuttings 313–314 sanitation cuttings 312–313 species susceptibility 307–308; (deﬁned) 307 stand susceptibility 309–311; (deﬁned) 309 stand vulnerability 309–311; (deﬁned) 310 suppression 314 thinning 311–313 tree mortality 308, 311, 312 tree species susceptibility 307–308 height growth 432–437 black oak 433–436 blue oak 437 California black oak 436 canyon live oak 436 chestnut oak 434–436 models 269–271, 433–437 northern red oak 434–436 Oregon white oak 436 scarlet oak 434–436 by site index 432 by stand density 432 white oak 433–436 with thinning 432 see also site index curves herbicides 262, 280, 287, 291–293, 318 hierarchical ecological units see ecological land units; ecoregions human culture and oaks 4, 8–9 importance value 13 improvement cutting 295 ingrowth models 455–458 initial ﬂoristics model 151, 209 see also succession insects see acorn insects; acorn predation and dispersal; ﬁre; gypsy moth intermediate cuttings 294–305; (deﬁned) 294 crop trees 295–296, 298–299 scheduling 294–295 types 294–295 see also thinning Lake States oak regeneration 262 shelterwood 274 latewood 425 liberation cutting 295 lignotubers 100, 171

Lithocarpus 9 Lobatae 9 low thinning 294

management paradigm xii philosophy xii mast see acorn production models acorn production 398–399 growth and yield 424, 447–459 diameter distribution 448 evaluation 459 height 433–437 individual-tree 447, 451–455 ingrowth 455–458 stand level 424, 447–451 stand table projection 448, 451 modelling theory 155–157 bottom-up approach 155 as simpliﬁcations of reality 157 top-down approach 155 properties of models 155–157 accuracy 155, 156 complexity 157 efﬁciency of application 157 generality 155 realism 155, 156 regeneration 265–274 accuracy 157 ACORn 153, 269–272 advance reproduction as predictors 158 ADVREGEN 266–269 Allegheny Plateau 271–272 Appalachians 268–269 as predictors of succession 157 bottomland 273–274 buried seed 158 Cumberland Plateau 271–272 data input requirements 159 dominance probabilities 267–268 decision guide models 158 ﬂuctuation in seedling density 145–146 FORCAT 271–272 height growth 269–271 initial ﬂoristics approach 158 limitations 157 measurements required for application 157–159 Northeast Decision Model 260 predictive uncertainty 158 SILVAH 271–274 SIMSEED 145–146 simulation models 158

Index

stand level 157–158 unpredictable factors 158 types of models 155–157 empirical 156 mathematical 156 models in forestry 156 physiological 156 process models 156 statistical 156 science-based 156 mortality see survival

net primary productivity 169–171 by species 170 New England oak regeneration 262 thinning 298–299 uneven-aged management 371 niche see species niche nomenclature for plants and animals cited 468–474 for tree species xii normal stocking see stand growth normal yield 439 Northeast Decision Model see models, growth and yield; models, regeneration Northern Hardwood Region 19–26 climate 20–21 forest history 23–24 geographic extent 19–20; (map) 19 oak distribution 24–26 physiography 20–21 northern pin oak–white oak forest type (photo) 25 northern red oak forest type (photo) 25 nursery practices acorn viability 278 root pruning 279, 281, 293 seedbed density 279 seedling quality 278–279 shoot clipping 279, 287–293 sowing acorns 279 storing acorns 278 transplanting seedlings 279 undercutting 279, 281–282, 293 nutrients, in boles, branches and foliage 174

oak decline 314–319; (deﬁned) 314–315 Armillaria root disease 315 silvicultural treatments 315 oak planting see planting oaks oak range 10–20 oak reproduction, accumulation of 121–147, 151–154 accumulation (deﬁned) 122

495

accumulation types ambivalent accumulators 129, 154 and ecological classiﬁcation 131 intrinsic accumulators 126, 154 recalcitrant accumulators 125, 127, 154 canopy gaps 122–123 dieback 122, 125 distribution of size classes, 124 disturbance 130–142 effects on accumulation of reproduction 133 intensity and frequency 142 mechanisms 130–133 ﬁre effects 131, 139; (illustrated) 139 site quality 130 stand history 130 ecological classiﬁcation 130 in Forest–Prairie Transition Region 131 livestock grazing 133, 141 in oak savannas 131 presettlement era 131 ecosystem resilience 126 future stand composition 152–153 Ozark Highlands 152–153 ﬂuctuation in reproduction density 142–147 importance in regeneration modeling 152 initial ﬂoristics model 151 initial state 152 overstorey density effects 128 overstorey inhibition 153 Ozark Highlands 127 predictive regeneration models 152–153 see also models, regeneration regeneration potential 122, 147–158 regeneration window 129 reproduction density 125 root age and size 122 root growth 125 stand density 122 succession 154 topography effects 128 transient recruitment rate 124 variation in accumulation 122–130 xeriﬁcation 141 see also single-tree selection method oak savannas 78–80, 380–391 biodiversity 384–385 in California 381, 382 characteristics 380–382 crown area by species 391 crown cover 389–391 crown cover chart 389–390 crown cover/dbh relations 391 density range 389

496

Index

oak savannas continued disturbance processes 383–385 deer browsing 383 ﬁre 380, 383–385, 387 introduced grasses 383, 385, 388–389 livestock 383 urban expansion 383 estimating crown cover 389–391 ﬁre 380, 383–385, 387 history 384 intervals 384 probability pattern 384 ﬂoristic composition 78 and oak seedling mortality 78 geographic extent 380–382; (map) 381 grass competition 78 history 380 in California 381, 382 in Illinois 385 introduced plants in 383, 385 maintenance 387–389 blue oak savannas 388 burning and weather 388 burning cycles 388 introduced grasses 389 monitoring 389 oak seedling establishment 388 oak sprouting 388 survival of oak reproduction 388 managing 385–391 in Minnesota 384 in Oregon 382 in Ozark Highlands 381, 385 presettlement history bison 385 Willamette Valley 384 research on 389 restoration 385–387 aesthetics 387 artiﬁcial reestablishment 385, 387 burning 385, 387 canopy density 385 deer browsing 387 ﬁre damage 387 ﬁre-free intervals 386 ﬁre fuels 386, 387 gophers 387 livestock grazing 387 oak recruitment 386 overstorey density 387 strategies 386 timber harvesting 387 tree deadening 387 tree felling 387 valley oak savannas 387 species richness 385 urban expansion 383

western oak savannas and woodlands animal predation on acorns 78–80 blue oak 78–80, 382 blue oak woodland (photo) 48 bunchgrass 382 California black oak woodland 382 California coast live oak 48 classiﬁcation 47, 381–382 climate 385 ﬁre 385 geographic extent (map) 381 introduced grasses 382, 385, 389 livestock 385 oak regeneration 385 oak species 382 rodent predation on acorns 79–80 survival of oak reproduction 388 types 47, 381 valley oak 382 Willamette Valley 384 woodland forms 381, 382 oak seedling survival 142–147 see also models, regeneration; oak reproduction, accumulation of oak wilt 316–320; (deﬁned) 316 control 317–319 epicentres 316 fungal mats 316 root grafts 316 spread 316–318 treatment of 314–320 oak–gum–cypress forest type 14–15 geographic extent (map) 15 oak–hickory forest type 14–15 geographic extent (map) 15 oak–pine forest type 14–15, 25, 35 geographic extent (map) 15 OAKSIM see models, growth and yield Ohio Valley animal disturbances 404 biodiversity 403 buffer zones 410 compared to selection method 341, 362 European settlers 404 exotic species in 407 extent 403–404 ﬁre in 407, 408 den sites 409 forests in transition to old growth 408–409 nesting sites 409 q or q-value 408 snags 409 standing dead trees 409 inﬂuence of Native Americans (Indians) 404 landscape scale 409–410

Index

oak regeneration 260 old-growth forests 32, 341, 362, 403–410; (deﬁned) 403, 404 old-growth characteristics 405–406 silvicultural options 404–410 planting oaks 287 shelterwoods 276 thinning 297 uneven-aged management 371 white oak (photo) 32 overstorey inhibition 209 Ozark Highlands age and diameter distributions 350 geographic location (map) 19 oak regeneration 259, 265–267 oak reproduction 355 planting oaks 286–281, 289–293 regeneration models 256–271 shelterwood 276

497

shoot clipping 279, 287, 290–293 tree shelters 282–285 see also nursery practices pollen dispersal 57–58 in bear oak 58 in chestnut oak 57 in chinkapin oak 58 in white oak 57 prairie 36–40 presalvage cutting 295 prescribed burning see ﬁre preservation 3 productivity and sustainability 172–173 Protobalanus 9 province see ecoregions q or q-factor see single-tree selection; forests in transition to old growth 408–409

Quercus 9 Paciﬁc Mediterranean–Marine Region 19–20, 43–48 climate 44 forest history 44–45 geographic extent 43–44; (map) 19 oak distribution 45–48 physiography 44 phenology 424 pistillate ﬂowers see female ﬂowers planting oaks controlling understorey 286 controlling overstorey density 286–287 direct seeding 280–281 dominance probabilities 285–287 planting in clearcuts 284–285 in group selection openings 371 prescribed ﬁre 284–285, 290, 293 under shelterwoods 285–294 plantation establishment bottomland sites 280–281 cleaning 280 herbicides 280, 287, 291–293 site preparation 280 thinning 280 weeding 280 planting factor 286, 293 regeneration potential 285 seedling size 280, 287–293 shelterwoods Arkansas 287–288 Indiana 287 Ozark Highlands 287, 289–293 prescriptions 290–291 West Virginia 287 with yellow-poplar 287

recruitment of reproduction into overstorey (illustration) 152 regeneration (deﬁned) 54 differentiated from reproduction 54 regeneration methods 255–278, 337, 352–355, 357, 359, 361, 365–373 and succession 259–263 acorn production 262 clearcutting method 255–274 group selection method 365–373 herbicides 262 seed tree method 277–278 shelterwood method 274–277 single-tree selection 337, 352–355, 357, 359, 361 site quality 261 soil moisture 261–262 soil scariﬁcation 262 see also acorn production; models, regeneration regeneration mode 148–155; (deﬁned) 148 catastrophic mode 150–151 continuous mode 151 gap-phase mode148–150 regeneration niche (deﬁned) 118 regeneration potential 147–148, 255–256, 265 regeneration strategy (deﬁned) 118 seeding in arid regions 123 in Nuttall oak 119 root vs. stem age 123 seedling density 119 seedling persistence 120 site quality effects 123 in southern bottomlands 119

498

Index

regeneration strategy continued seeding continued stand density effects 122–123 in water oak 119–120 seeding vs. sprouting dependency 121 species’ ﬂexibility 119–121 sprouting in Ozark Highlands 118 root system size 119 regeneration tactic (deﬁned) 118 see also regeneration strategy regeneration window 129 regulated forest 254 reproduction (deﬁned) 54 see also oak reproduction, accumulation of reproductive mechanisms 118–119 seeding vs. sprouting (illustration) 121 rhizomatous oaks 100, 101; (illustration) 101 rodents see acorn predation and dispersal; ﬁre; gypsy moth root growth in oak seedlings 86–89 periodicity 87, 425 see also germination and seedling establishment root sprouts see stump sprouts root:shoot ratio 89–90, 92–94, 97, 100, 104, 138, 169 and site productivity 169 rotation age 254 runoff 173

safe sites 201 salvage cutting 295 savanna see oak savanna sanitation cutting 295 scientiﬁc names of species cited 468–474 seed buried in forest ﬂoor 120 seed tree method acorn production 278 application to oak forests 278 economic and social considerations 320–322 natural regeneration 277–278 objective 278 tree cavities278 seedling growth 84–92 acorn food reserves 86 acorn size effects 87 in associated non-oaks 92 black oak 92 carbon allocation 89, 91 cotyledon effects 86 delayed shoot growth 86 drought tolerance 91

elevation of seed source 92 ﬂushing 87, 92 frost avoidance mechanism 92 genetic variation 91 lag stage of shoot growth 87 lateral roots 86 leaf area growth 88 leaf linear stage of shoot growth 87 leaf temperature 91 morphological characteristics 91 northern red oak 88, 92 periodicity in oak root growth 87 photosynthesis 87, 89 physiological plasticity 91 resting stage of shoot growth 87 root growth 88 root permeability to water 86 root surface area 89–90 root:shoot ratio 89–90 scarlet oak 92 seed source 92 shade-tolerance rankings 91 shoot growth 87–88 soil temperature 86 stomata 89 suberized roots 86 taproot development 86 time of spring budbreak 91–92 transpiration 89–90 variation in acorn size 87 water stress 87 water-use efﬁciency 90 white oak 92 seedling sprouts (deﬁned) 92 see also oak reproduction, accumulation of; shoot dieback in reproduction selection thinning 294 self thinning (deﬁned) 227 oak forests and −3/2 rule 231–235 allocation of biomass 232 allometry 234 apical control 235 applicability 232–235 based on yield tables for English oak 232 elastic similarity 234 geometric similarity 233 crown diameter:dbh ratios 235 isometry 233–234; (deﬁned) 233 non-symmetrical crowns 234 in old oak stands 233 in the Oak–Hickory Region 232 ratio constancies of tree components 234 self-thinnning lines 233 tree geometry 234 Reineke’s model 227–229

Index

Reineke’s model vs. −3/2 rule 231 the −3/2 rule 229–231 shade tolerance of oaks see also oak reproduction, accumulation of; shelterwood method; shoot dieback in oak reproduction shade tolerance ranking 91 shelterwood method acorn crops 275 controlling stand density 274–275 economic and social considerations 322 enrichment planting 285–293 establishment cut 274 group shelterwood 274 herbicide 275 in yellow-poplar stands 275, 277 natural regeneration 274–277 oak decline 314–315; (deﬁned) 314–315 oak regeneration 274–277 objective 274 preparatory cut 274 prescribed burning 277, 290, 293 regional examples Appalachians 275 Lake States 274 Michigan 276 Minnesota 275–276 Ohio 276 Ontario 274 Ozark Highlands 276 Wisconsin Driftless Area 275–276 removal cut 274 strip shelterwood 274 uniform shelterwood 274–275 with prescribed burning 277 with tree shelters 283 see also planting oaks shoot dieback in reproduction 92–98 and shoot growth 94 as an adaptive strategy 94 in black oak 93 bud desiccation 95 bud survival 96, 97 cambial initiation 96 carbon allocation 97 in cherrybark oak 97 in clearcuts 95 cost of 97 defoliation 97 depletion of starch reserves 97 drought 97 in English oak 95 frequency of dieback 95 induced by stresses 93, 94 and insects 94 loss of ﬁne roots 95 and multiple ﬂushing 97

499

in northern red oak 92, 96, 97 photosynthesis 97 principle of plant segmentation 95 progression of 96 and root growth 97 root growth regulators 97 root:shoot ratio 93, 97 seedling mortality 94 shade effects 93 in shelterwoods 95, 96 simulated by shoot pruning 97 spring frost 94, 95 sprouting from dormant buds 92 in sprouts 354 terminal bud cluster 93 transpiration 97 vs. dieback and decline disease 94 winter dieback 95–96 xylem dysfunction 95 shoot growth, 425 see also height growth shrews 79, 314 SILVAH see models, regeneration silvicultural systems 254 silviculture (deﬁned) xi, 1 agronomic model 3, 173 consilient discipline 5–7 ecological model 3, 173 history 1–2 integrative science 5–6 origin 1 relation to other disciplines 6 social sciences 5–6 see also clearcutting method; group selection method; natural regeneration methods; seed tree method; single-tree method; unevenaged methods; uneven-aged silviculture single-tree selection method 337–366 accumulation of oak reproduction 353, 355, 359, 362–363 canopy gaps 337 compared to old-growth 341, 362 complexity 338 cutting cycle 364 depletion curve 338 diameter distributions balanced 338 effect of timber harvesting 349 guiding curve 342–348, 351–352, 359, 362–364, 366 negative exponential 339–348, 353–359, 360–361 relation to age distribution 346–351 reverse J-shape 339–348, 353–359, 360–361

500

Index

single-tree selection method continued diameter distributions continued rotated sigmoid 339 stability 348–349, 351–352, 358–359, 361–362 sustainability 338–340, 351–352, 355, 357 unbalanced 338 disturbance frequency 337 economic and social considerations 355, 372–374 examples of application Appalachian Highlands 355, 357–358, 364–365 Michigan 355–356 Ozark Highlands 354–355 ﬁeld implementation 360–365 cutting cycle 364 guiding curve 362–363 inventory 360–362 monitoring 364 silvicultural guidelines 366 timber marking guidelines 362–364 variants 364–365 guiding curve 342–348, 351–352, 360 monitoring 341 principles of application 337 q or q-factor (deﬁned) 340 mathematical relations 340–343, 345, 347, 348 relation to stand structure 342–348, 351, 356–363 specifying 342–348, 360–363, 366 selective vs. selection cutting 351 spatial scale 337–339 species succession 351–352, 354–355, 362 stand characteristics to consider 338, 366 stocking 343–344, 356, 363–364, 366 three-stage cycle 337 variants of the method 364–365 site evaluation categorical 185–186 Central Hardwood Region, 187 Cumberland Plateau 187 Driftless Area 186–187 Highland Rim, 187 Interior Plateau 187 moist layers 186 slope position 186 soil texture 186 southern Michigan 186 topographic site coefﬁcient 186–187 topography 186 young clearcuts 187 see also site index

site index accuracy of determining 176, 178 comparison among species 179–180 conversion between species 180 curves anamorphic 179, 181 for comparing species growth 214 comparison among species 179–180 polymorphic 179 direct determination 176–179 from soil and topographic factors 180–183 height growth 432 height/dbh curves black oak 487 northern red oak 487 white oak 487 indirect estimation from soil and topographic factors 180–183 from tree height and diameter 183 precision of estimates 179 in relation to aspect and slope per cent 181–183 species conversion equations 483–486 site productivity 168–189 and ecological classiﬁcation 171–172 biomass 174 in bottomlands 183 clear wood formation 431 economic returns 173 effects of timber harvesting 173 forest ﬂoor 174 inﬁltration rate, 174 nutrient losses 174–175 production vs. yield 168 removal of tree cover 173 root:shoot ratio 169 site (deﬁned) 168 site quality (deﬁned) 168 stand development 168 units of measure 169 gross primary 169 net primary 169 volumetric 169 value increase 171 volume increase 171 wood quality 168 value growth 431 see also site index slope effect on site quality 181–182, 187 soil and site quality base saturation 183 calcium 175 erosion 173 fertility 175 fertilization 175

Index

nitrogen 175 pH 183 Southern Pine–Hardwood Region 19–20, 33–36 climate 33 forest history 33–34 geographic extent 33; (map) 19 oak distribution 34–36 physiography 33 Southwestern Desert–Steppe Region 19–20, 40–43 climate 41 forest history 41 geographic extent 40–41; (map) 19 oak distribution 41–43 physiography 41 species niche 12–14 species richness 10–13, 142 accumulation type 142 disturbance 142 moisture gradient 142 sprouting see also regeneration strategy; stump sprouts staminate ﬂowers see male ﬂowers stand density and stocking and regeneration 236 crown width equations 237–238 effect on bole quality 251 effect on epicormic branching 251 open-grown crowns 237–239 relation to crown competition factors 237 terminology absolute density 235–236 crown competition factor 236 Curtis’s relative density index 236 Drew and Flewelling density index 236 fully stocked 237 Gingrich-style diagram 237 maximum tree area 237 minimum tree area 237 normal stocking 237 normal yield tables 237 overstocked 237 relative density 235, 236 stocking 235–237 stocking per cent 237 tree area (deﬁned) 237 tree-area ratio 238 understocked 237 Wilson’s relative spacing index 236 stand density diagrams for English and sessile oak 248, 249 for tracking stand growth 250 Gingrich diagram (illustrated) 243, 247 arithmetic vs. quadratic mean dbh 244

501

differences among regions 246 effects of species’ interactions 248 fully stocked zone 246 line of imminent competitioninduced mortality 242 for northern red oak 243, 247 stocking and biological growth limits 248 stocking levels 242 for southern bottomland forests 246 variance in tree dbh 244–245 Kershaw and Fischer diagram 248 stocking levels 242–248 see also stand density and stocking stand development 194–251, 227,–251 even-aged stands 199–216; (deﬁned) 254 biological vs. bole age of oaks 199 cohort (deﬁned) 199 complex stage 199–202, 215–216, 220 diameter distributions 212–214 incomplete stand-initiating disturbance 199 mixed stands 209, 212 spatial scale 255 stages of development 199–216 stand age (deﬁned) 199 stand initiation stage 199–203, 220 stand-initiating disturbance 199 stem exclusion stage 199–210, 220 understorey reinitiation stage 199–202, 207–215, 220 uneven-aged stands 216–223 diameter distributions 217, 218, 221, 223 stand growth 439–447 even-aged stands 439–446 normal stands 439–441 survival 440, 443 thinned stands 441–447 uneven-aged stands 446–447 unthinned stands 439–441 see also models, growth and yield stand initiation stage 199–203, 220 stand regeneration potential 147–148 change in 147 overstorey density 147 reproduction root size 148 shade tolerance 147 subcanopy species 147 stand structure 195 stand table projection 448, 451 stem exclusion stage 199–210, 220; (illustrated) 200, 202, 220 change in stand structure basal area 206 numbers of trees 206

502

Index

stem exclusion stage continued change in dbh distribution 208 correlation of age and diameter 205 diameter distributions bell-shaped 205 reverse J-shaped 205 ﬁre effects 207 oak dominance 204 in oak savannas 207 in Ozark Highlands 210 stand structure 199–200, 205 tracking on stocking chart 206 stewardship 6 stocking see group selection method; singletree selection method; stand density and stocking; stand density diagrams; stand growth see stem exclusion stage; thinning stool sprouts (deﬁned) 100 see also stump sprouts stump sprouts 98–106, 299–305; (deﬁned) 98 basal sprouting (illustrated) 99 defects 302–305 dormant buds 98 branching 98 mortality 98 growth 299–302, 429, 430 frequency of sprouting 100, 102 by species 102 estimated probability 102 models 102 parent tree diameter and age 100 season of cutting 100 shading 100 site quality 100 growth and survival 101–105 grub 98–99 height growth 105–106 physical resistance of bark 100 selecting crop stems 305 sprouting after ﬁre (photo) 99 thinning within clumps 299–305 weak basal attachment 100 subsurface water ﬂow 181 succession 13–14, 24–26, 29, 32, 151, 154, 351–352, 354–355, 362 accumulation of oak reproduction 154 competitive sorting 154 disturbance-mediated 154 initial ﬂoristics model 151, 209 seed bank 154 uneven-aged management 351–352, 354–355, 362 survival rates of overstorey oaks 437–439 chestnut oak 438–439 by dbh 437–439 following ﬁre 438–440

post oak 438–439 red oaks 438–439 by site quality 437–438 white oaks 438–439 see also self-thinning sustainability and productivity capacity 173

taxonomy 9–10 thinning for acorn production 396–397, 399 for aesthetics 414 commercial 296–297 crop trees 295–296, 298–299, 428 diameter growth 428–429 epicormic branching 298, 303–304 with fertilization 446 mechanical 294 precommercial 296 in response to gypsy moth 311–313 self-thinning in sprout clumps 103–104 stand growth 440–447 stocking 298–299, 301 stump sprouts (coppice) 299–305, 429 uneven-aged management 352, 371 yield 294, 296 see also intermediate cuttings; selfthinning tree shelters see planting oaks tree–area ratio (deﬁned) 238 differences among species 240 equations by species 240–242 minimum and maximum 241 relation to stand age and site quality 241 see also self-thinning TWIGS see models, growth and yield two-aged stand 254

understorey reinitiation stage light and soil moisture 207 in the Ozark Highlands 209 stand structure 199–202, 207–215, 220 tree reproduction 207 uneven-aged forest (deﬁned) 335 maintaining 335 regulated 335–336 see also stand development, uneven-aged stands uneven-aged silviculture cutting cycle 364 diameter distributions 337–350, 356–362 objective 335–336 stand structure 335–337 volume control 335

Index

see also group selection method; singletree selection method uneven-aged stand (deﬁned) 254 growth and yield 446–447 spatial scale 255 utilitarianism 4 value growth 430–432 black oak 430–431 northern red oak 430–431 white oak 430–431 vertical crown stratiﬁcation 151, 195, 214–215 adaptations to light ﬁelds 151 cherrybark oak and sweetgum 214, 215 cherrybark oak and sycamore 214, 215 ﬂowering dogwood 151 redbud 151 Terborgh theory 151 volume California oaks 46 equations 459–460 growth 439–447 index 184 regional patterns 460–461

West Virginia planting oaks 287 thinning 296

503

western hardwoods 14–15 wildlife use of acorns 8 see also acorn predation and dispersal wind 197, 199 WINYIELD see models, growth and yield Wisconsin Driftless Area geographic location (map) 19 oak regeneration 262 shelterwood 275–276 site evaluation 186–187 stump sprouts 300 topographic site coefﬁcient 186–187

yield effect of thinning 294, 441–447 even-aged stands 439–447 regional patterns 440–443 uneven-aged stands 446–447 see also yield tables YIELD see models, growth and yield YIELD–MS see models, growth and yield yield tables 460–461