Biodiversity in Drylands: Toward a Unified Framework (Long-Term Ecological Research Network Series)

Biodiversity in Drylands: Toward a Unified Framework Moshe Shachak, et al., Editors OXFORD UNIVERSITY PRESS BIODIVER...

Author: Moshe Shachak | James R. Gosz | Stewart T. A. Pickett | Avi Perevolotsky

49 downloads 499 Views 4MB 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

Biodiversity in Drylands: Toward a Unified Framework

Moshe Shachak, et al., Editors

OXFORD UNIVERSITY PRESS

BIODIVERSITY IN DRYLANDS

LONG-TERM ECOLOGICAL RESEARCH NETWORK SERIES LTER Publications Committee Climate Variability and Ecosystem Response at Long-Term Ecological Research Sites Edited by David Greenland, Douglas G. Goodin, and Raymond C. Smith Grassland Dynamics: Long-Term Ecological Research in Tallgrass Prairie Edited by Alan K. Knapp, John M. Briggs, David C. Hartnett, and Scott L. Collins Standard Soil Methods for Long-Term Ecological Research Edited by G. Philip Robertson, David C. Coleman, Caroline S. Bledsoe, and Phillip Sollins Structure and Function of an Alpine Ecosystem: Niwot Ridge, Colorado Edited by William D. Bowman and Timothy R. Seastedt Biodiversity in Drylands: Toward a Uniﬁed Framework Edited by Moshe Shachak, James R. Gosz, Avi Perevolotsky, and Steward T.A. Pickett

BIODIVERSITY IN DRYLANDS: TOWARD A UNIFIED FRAMEWORK Edited by

Moshe Shachak James R. Gosz Steward T.A. Pickett Avi Perevolotsky

1 2005

1 Oxford New York Auckland Bangkok Buenos Aires Cape Town Chennai Dar es Salaam Delhi Hong Kong Istanbul Karachi Kolkata Kuala Lumpur Madrid Melbourne Mexico City Mumbai Nairobi Sa˜o Paulo Shanghai Taipei Tokyo Toronto

Copyright # 2005 by Oxford University Press, Inc. Published by Oxford University Press, Inc., 198 Madison Avenue, New York, New York 10016 www.oup.com Oxford is a registered trademark of Oxford University Press All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior permission of Oxford University Press. Library of Congress Cataloging-in-Publication Data Biodiversity in drylands: toward a uniﬁed framework/ edited by Moshe Shachak . . . [et al.]. p. cm—(Long-Term Ecological Research Network series) ISBN 0-19-513985-2 1. Arid regions ecology. 2. Biological diversity. I. Shachak, Moshe. II. Series. QH541.5.A74B56 2004 577.54—dc22 2003015271

9 8 7 6 5 4 3 2 1 Printed in the United States of America on acid-free paper

In Memory of Gary Allan Polis (28 August 1946–27 March 2000) Robert D. Holt Wendy B. Anderson

G

ary Polis loved deserts. Whenever he dealt with the topics of his major contributions to community ecology—the prevalence of intraguild predation and omnivory; the ubiquity of reticulate food web structures, with many weak and donor-controlled links; and the importance of allochthonous subsidies, detrital pathways, and temporal variability in food web dynamics—he would invariably lace the discussion with concrete examples from desert ecosystems. Among his many books is The Ecology of Desert Communities (1991), which in many ways can be viewed as a natural predecessor to the current volume. Gary’s boyhood fascination with deserts and scorpions broadened into a detailed understanding of food web interactions, providing an entre´e to the ecological community, where he became a leader in food web ecology. It is in this arena that he made the greatest contributions to ecology, and this is one clear instance where an understanding of desert ecology has led to conceptual advances in ecology as a whole. Gary was a superior naturalist. He could develop a sense about a place because he spent huge amounts of time in the deserts he studied. He was accomplished at making acute observations about patterns, and relating them to broader ecological concepts. For example, while collecting data to describe patterns of scorpion diversity and distribution on desert islands in the Gulf of California, he noticed that patterns of spider and lizard abundances and distributions seemed to covary with scorpion distribution patterns. He began to understand the web of interactions that could exist among these three higher-level consumers. Furthermore, he became aware that all of these patterns were intricately tied to the inﬂuences of the

vi In Memory of Gary Allan Polis

surrounding ocean on the food webs of the desert islands, via the subsidy of their ecosystems by materials drawn from the marine environments. Gary’s emphasis on the ubiquity and importance of such subsidies was a major contribution made in his last few years. Again, Gary tapped his passion for desert ecology to inform at a deeper level for the discipline as a whole. The consummate educator, Gary taught everybody: undergraduate and graduate students, colleagues, his own children, dozens of volunteers who assisted him in the ﬁeld, virtually anyone who would listen. Three generations of academic oﬀspring have beneﬁted from Gary’s wonderful insights and worldview directly or indirectly, and his ideas and perspectives continue to resonate through ecology today. Finally, Gary had an amazing talent for bringing together people who might not otherwise interact to create novel syntheses among previously disparate disciplines. His appreciation of people from diverse backgrounds (scientiﬁc and cultural) drew people to him and hence to each other. He ran his own research laboratory in an intellectually inclusive style—inviting students and post-docs with interests in plants, invertebrates, vertebrates, or soils, and ranging from taxonomy to physiology to landscape ecology–aiming toward a synthetic understanding of the desert systems he so loved. In particular, Gary believed that the interplay of temporal variability and spatial heterogeneity was fundamental to understanding desert communities. Valuing biological diversity, scientiﬁc diversity, and the diversity of human perspectives alike. Gary recognized that any healthy assembly of species—or people—must include numerous functional groups to thrive. In this volume, we acknowledge that we are poorer for his loss, but richer for his having been among us.

Reference Polis, G.A. (ed.). 1991. The Ecology of Desert Communities. University of Arizona Press, Tucson, Arizona.

Preface

Why Study Biodiversity in Drylands? Many international conventions have identiﬁed the primary environmental problems of the world as being urbanization, global change, desertiﬁcation, and biodiversity (Vitousek 1994). These important environmental problems are interrelated and dramatically expressed in the world’s drylands. While all are important and actively studied topics, the role and signiﬁcance of biodiversity in drylands is the least well understood scientiﬁcally topic. Biodiversity was coined as a term to promote public and political dialog about the state and future of the world’s biological richness. It is regarded as both a social–political construct and a scientiﬁc concept (Gaston 1996). While its history ensures social signiﬁcance, it leaves the concept devoid of the theory and body of knowledge that typiﬁes scientiﬁc disciplines. The need for the development of biodiversity as a scientiﬁc discipline is clear. First, although the concept is general, and appropriately applied to ecological realms ranging from the genetic to the ecosystem and landscape (Noss and Cooperider 1994), it has most often been tacitly restricted to the topic of species. Management, especially, requires the use of biodiversity to motivate and organize studies of ecosystems and landscapes. Second, while there is a rigorous general deﬁnition that identiﬁes the core aspects of biodiversity as number and diﬀerence in ‘‘biological entities,’’ how this deﬁnition can and should be speciﬁed to the range of ‘‘ecological entities’’ is unclear. Third, the function of biodiversity must be determined in a wide variety of environments. Biodiversity is an important topic that requires a well-developed theory, and a clear strategy for application to management.

viii Preface

The aim of this book is to elucidate the scientiﬁc basis for biodiversity studies and management. We emphasize biodiversity as a powerful, integrative concept, but one that still requires careful articulation and application. Even though this book utilizes many case studies from drylands, it emphasizes the generality of the biodiversity concept. Drylands are experiencing an accelerating rate of change, mainly due to shifts in land use and climate change initiated by humans. These changes aﬀect the distribution and abundance of species, habitats, and ecosystems, thereby creating new landscape mosaics. Biodiversity includes the diversity of organisms in complex assemblages of interacting communities and ecosystems, as changes in global systems accelerate, changes in dryland biodiversity are also accelerating. Understanding biodiversity changes in these systems requires the development of general guiding principles for the study and management of biodiversity of drylands. In order to address the problem we organized a workshop, ‘‘Biodiversity in Drylands: Toward a Uniﬁed Framework for Research and Management’’ (26 June–2 July 1999), at the Blaustein Institute for Desert Research of Ben Gurion University, Israel. The aim of the workshop was to confront three main problems that emerge as a consequence of the broad scope of the biodiversity concept. The ﬁrst problem addressed how to incorporate processes (e.g., foraging, energy and nutrient ﬂows, patch dynamics) into a concept that is based on entities (e.g., individual organisms, species, habitat types, patch types). The second involved how to integrate ecological subdisciplines (e.g., ecosystem, population, landscape ecology) that are involved in biodiversity studies. The third included how to use a theoretical framework that incorporates ecological processes and entities, and integrates across subdisciplines as a guideline for conservation, restoration, and management of biodiversity. We deﬁned a number of objectives for the workshop participants: 1. Develop a conceptual framework that can integrate studies of biodiversity in a network of research and management sites in drylands. 2. Evaluate the state of knowledge of dryland biodiversity and pose questions to guide future research and management. 3. Generate new concepts and ideas needed for a theory and management of biodiversity in drylands. 4. Publish the ﬁndings in a book on biodiversity studies in drylands aimed at students, scientists, managers, and educators. This book represents the ideas, methodologies, and applications stimulated by workshop presentations, discussions, and group eﬀorts. It addresses the question how diversity of entities in ecological systems, through their webs of interactions, aﬀect the performance of the ecosystem. We believe that by addressing this question we have helped advance the aim of seeking a uniﬁed framework for biodiversity studies and management in many types of habitats. A central theme of the book is the relationship between the diversity of organisms and landscapes and the structure and function of the ecosystem.

Preface

ix

We have integrated processes and entities in biodiversity studies based on the relationship between primary production, food web interactions, community processes, resource distribution, habitat structures and species diversity, in relation to scale. We suggest general guidelines for integration across ecological subdisciplines in biodiversity studies. The integration is based on the sets of relationships between organisms as ecosystem engineers and species diversity, microbial and ecosystem processes, species and ecosystem processes, and landscape processes and species diversity. The applied aspects of this book suggest utility of the biodiversity concept for conservation, restoration, and management. The topics encompass how rangeland management and water harvesting support sustainability of biodiversity.

Acknowledgments We are grateful to a number of organizations and individuals whose eﬀorts enabled us to hold our workshop and publish this book. An important organization that assists in studying biodiversity in drylands is the International Arid Lands Consortium (IALC). The IALC was very generous in its support for the workshop in Israel and for the agendum leading to the publication of this book. The IALC supports projects that are intended to lead to a better understanding of the management of fragile dryland ecosystems for sustainable human use. Their policy dictates that understanding biodiversity is critical for managers attempting to achieve the goals of maintenance of ecosystem function, optimizing yield of valuable plant and animal species, or maintaining niches for threatened or endangered species. The IALC experience in encouraging the linkage of biodiversity research with management needs illustrates the problems and potentials of addressing human needs more directly through research in biodiversity. The IALC is a partnership of organizations established in 1990 and dedicated to research, education, and training relative to development, management, restoration, and reclamation of arid and semiarid lands with a primary focus on the Middle East and the Southwestern United States. Fifty research and development projects have been funded from 1993 through 2000, of which 24 relate to a variety of biodiversity issues and concerns. A listing of those projects funded by the IALC since 1993 can be found in Hegwood (2000) and on the website http://ag.arizona.edu/OALS/IALC/Home.html. These projects dealt with the role of landscape and species diversities in the function of water-limited systems. In addition, the IALC supports demonstration projects that represent applications of available knowledge and technologies derived from research and development eﬀorts for the management of sustainable ecological systems. The IALC provides unique opportunities to foster international collaboration for basic and applied research. Sponsors are particularly interested in the

x Preface

transfer of technology to citizens of countries most in need of this technology and in promoting cooperation and peaceful interaction among neighboring nations in the politically troubled Middle East. Please visit the IALC website for more information. We want to thank Ben Gurion University of the Negev for generous support and for providing campus facilities for the workshop; also the Blaustein Institute for Desert Research, and especially the Blaustein Center for Scientiﬁc Cooperation, for ﬁnancial support and providing use of the Sede Boqer facilities for the workshop. We are indebted to the Jewis National Fund for their ﬁnancial support. We are also grateful to Patty Sprott at the Long Term Ecological Research Network Oﬃce in Albuquerque, New Mexico for support with managing and editing the volume. We appreciate the fact that Oxford University Press agreed to publish our book and thank them for their patience. Many individuals assisted in helping this eﬀort become a reality. We are grateful to them for that. They include Menachem Sachs for his encouragement and help in organizing support for the workshop and book; Andy Wilby and Yoram Ayal for helping in the organization and logistics of the workshop, which was signiﬁcant for the success of the workshop; Bertrand Boeken, Yarden Oren, and David Ward for their help in planning and carrying out the ﬁeld trips during the workshop; Yael Kaplan for her secretarial assistance; Bob Waide who agreed to be the link between Ben Gurion University and ILTER; and Sol Brand for helping in editing aspects of the book. In conclusion, we hope that this book will be beneﬁcial to students, scientists, managers, and educators who are concerned with biodiversity issues. We believe that greater understanding of biodiversity of drylands, which occupy 40% of the world’s terrestrial systems and are extremely sensitive to desertiﬁcation, will be the outcome of this book.

References Gaston, K.J. (ed.) 1996. Biodiversity. A biology of numbers and diﬀerence. Blackwell Science, Oxford, UK. Hegwood, D.A. (compiler). 2000. International Arid Lands Consortium: A compendium of funded projects. International Arid Lands Consortium, Tucson, Arizona. Noss, R.F., and A.Y. Cooperider. 1994. Saving Nature’s Legacy: Protecting and Restoring Biodiversity. Island Press, Washington, DC. Vitousek, P.M. 1994. Beyond global warming: ecology and global change. Ecology 75: 1861–1876.

Contents

Contributors 1

xv

Introduction A Framework for Biodiversity Studies

3

Moshe Shachak, James R. Gosz, Avi Perevolotsky, and Steward T.A. Pickett

I

Living Components of Biodiversity: Organisms

2

How Can High Animal Diversity Be Supported in Low-Productivity Deserts? The Role of Macrodetritivory and Habitat Physiognomy

15

Yoram Ayal, Gary A. Polis, Yael Lubin, and Deborah E. Goldberg

3

Biodiversity Along Core–Periphery Clines

30

Salit Kark, Sergei Volis, and Ariel Novoplansky

4

Species Diversity, Environmental Heterogeneity, and Species Interactions 57 William A. Mitchell, Burt P. Kotler, Joel S. Brown, Leon Blaustein, and Sasha R.X. Dall

xii Contents

5

SHALOM A Landscape Simulation Model for Understanding Animal Biodiversity 70 Yaron Ziv, Michael L. Rosenzweig, and Robert D. Holt

6

Spatial Scale and Species Diversity Building Species–Area Curves from Species Incidence

89

William Edward Kunin and Jack J. Lennon

7

Microbial Contributions to Biodiversity in Deserts

109

Peter M. Groﬀman, Eli Zaady, and Moshe Shachak

8

Uniﬁed Framework I Interspeciﬁc Interactions and Species Diversity in Drylands

122

Gary A. Polis, Yoram Ayal, Alona Bachi, Sasha R.X. Dall, Deborah E. Goldberg, Robert D. Holt, Salit Kark, Burt P. Kotler, and William A. Mitchell

II

Ecological Complexes of Biodiversity, Ecosystems, and Landscapes

9

Species Diversity and Ecosystem Processes in Water-Limited Systems 153 Moshe Shachak, Steward T.A. Pickett, and James R. Gosz

10

Linking Species Diversity and Landscape Diversity

167

Bertrand Boeken, Yarden Oren, Shlomo Brandwine, and Sol Brand

11

The Impact of Animals on Species Diversity in Arid-Land Plant Communities 189 Andrew Wilby, Bertrand Boeken, and Moshe Shachak

12

Resource Partitioning and Biodiversity in Fractal Environments with Applications to Dryland Communities 206 Mark E. Ritchie and Han Olﬀ

13

Uniﬁed Framework II Ecosystem Processes: A Link Between Species and Landscape Diversity 220 Moshe Shachak, Robert Waide, and Peter M. Groﬀman

Contents

III

Biodiversity, Conservation, and Management

14

The Effects of Grazing on Plant Biodiversity in Arid Ecosystems David Ward

15

Sustainability in Arid Grasslands New Technology Applications for Management

250

Arian Pregenzer, Robert R. Parmenter, Howard Passell, John R. Vande Castle, Thomas K. Budge, and Gregory Michael Bonito

16

Reconciliation Ecology and the Future of Species Diversity Michael L. Rosenzweig

17

Management for Biodiversity Human and Landscape Effects on Dry Environments

286

Avi Perevolotsky, Moshe Shachak, and Steward T.A. Pickett

18

Uniﬁed Framework III Human Interactions with Biodiversity

305

Anna A. Sher, Bruce M. Kahn, and Christopher R. Dickman

19

Toward a Uniﬁed Framework in Biodiversity Studies Moshe Shachak, James R. Gosz, Avi Perevolotsky, and Steward T.A. Pickett

Index

337

320

266

xiii

233

This page intentionally left blank

Contributors

Yoram Ayal Blaustein Institute for Desert Research Ben Gurion University of the Negev Sede Boqer, Israel 84990

Sol Brand Blaustein Institute for Desert Research Ben Gurion University of the Negev Sede Boqer, Israel 84990

Alona Bachi Department of Ecology and Evolutionary Biology University of Arizona Tucson, AZ 85721-0088

Shlomo Brandwine Blaustein Institute for Desert Research Ben Gurion University of the Negev Sede Boqer, Israel 84990

Leon Blaustein Institute of Evolution University of Haifa, Israel 31905

Joel S. Brown Department of Biological Sciences University of Illinois Chicago, IL 60607

Bertrand Boeken Blaustein Institute for Desert Research Ben Gurion University of the Negev Sede Boqer, Israel 84990

Thomas K. Budge Earth Data Analysis Center University of New Mexico Albuquerque, NM 87131

Gregory Michael Bonito LTER Network Oﬃce University of New Mexico Albuquerque, NM 87106

Sasha R.X. Dall Blaustein Institute for Desert Research Ben Gurion University of the Negev Sede Boqer, Israel 84990 xv

xvi

Contributors

Christopher R. Dickman School of Biological Sciences and Institute of Wildlife Research University of Sydney NSW, Australia 2006 Deborah E. Goldberg Department of Ecology and Evolutionary Biology University of Michigan Ann Arbor, MI 48109-1048 James R. Gosz University of New Mexico Albuquerque NM 87131-1091 Peter M. Groﬀman Institute of Ecosystem Studies Milbrook, NY 12545 Robert D. Holt Department of Zoology University of Florida Gainesville, FL 32611 Bruce M. Kahn Department of Rural Sociology University of Wisconsin Madison, WI 53706 Salit Kark Institute of Life Sciences The Hebrew University of Jerusalem Jerusalem, Israel 91904 Burt P. Kotler Blaustein Institute for Desert Research Ben Gurion University of the Negev Sede Boqer, Israel 84990 William Edward Kunin School of Biology University of Leeds Leeds, United Kingdom LS2 9JT Jack J. Lennon Macaulay Institute Aberdeen, Scotland AB15 8QH

Yael Lubin Blaustein Institute for Desert Research Ben Gurion University of the Negev Sede Boqer, Israel 84990 William A. Mitchell Department of Life Sciences Indiana State University Terre Haute, IN 47809 Ariel Novoplansky Blaustein Institute for Desert Research Ben Gurion University of the Negev Sede Boqer, Israel 84990 Han Olﬀ Department of Environmental Science Wageningen Agricultural University Bornsesteeg 69 6708 PD Wageningen The Netherlands Yarden Oren Sustainable Ecosystems Department Commonwealth Scientiﬁc & Industrial Research Organization (CSIRO) Canberra ACT Australia 2601 Robert R. Parmenter Department of Biology University of New Mexico Albuquerque, NM 87131 Howard Passell Cooperative Monitoring Center Sandia National Laboratory Albuquerque, NM 87185 Avi Perevolotsky Department of Natural Resources Agricultural Research Organization Bet Dagan, Israel 50250 Steward T.A. Pickett Institute of Ecosystem Studies Millbrook, NY 12545 Arian Pregenzer Cooperative Monitoring Center Sandia National Laboratory Albuquerque, NM 87185

Contributors Mark E. Ritchie Department of Biology Syracuse University Syracuse, NY 13244

Michael L. Rosenzweig Department of Ecology and Evolutionary Biology University of Arizona Tucson, AZ 85721-0088

Moshe Shachak Blaustein Institute for Desert Research Ben-Gurion University of the Negev Sede Boqer, Israel 84990

Anna A. Sher Weed Science Program University of California Davis, CA 95616

John R. Vande Castle LTER Network Oﬃce University of New Mexico Albuquerque, NM 87106

xvii

Sergei Volis Institutes for Applied Research Ben Gurion University Beer Sheva, Israel 84105 Robert Waide LTER Network Oﬃce University of New Mexico Albuquerque, NM 87106 David Ward Department of Conservation Ecology University of Stellenbosch Matieland, South Africa 7602 Andrew Wilby WERC Center for Population Biology Silwood Park, Ascot, Berks United Kingdom SL5 7PY Eli Zaady Blaustein Institute for Desert Research Ben Gurion University of the Negev Sede Boqer, Israel 84990 Yaron Ziv Department of Life Sciences Ben Gurion University Beer Sheva, Israel 84105

This page intentionally left blank

BIODIVERSITY IN DRYLANDS

This page intentionally left blank

1 Introduction A Framework for Biodiversity Studies Moshe Shachak James R. Gosz Avi Perevolotsky Steward T.A. Pickett

The Need for a Uniﬁed Framework of Biodiversity Biodiversity is regarded as a scientiﬁc concept, a measurable entity, as well as a social–political construct (Gaston 1996, Wilson 1993). The aim of this volume is to develop the scientiﬁc basis for biodiversity studies, and for the integration of the concept into management practice. We emphasize biodiversity as a powerful, integrative concept—one that requires careful articulation and further conceptualization before application. Diversity is a concept that refers to the range of variation or diﬀerences among a set of entities; biological diversity then refers to variety within the living world. An example of biological diversity is ‘‘species diversity,’’ which is commonly used to describe the number, variety, and variability of the assemblage of living organisms in a deﬁned area or space. However, biodiversity as a concept has evolved. Current deﬁnitions expand the biological diversity concept to emphasize the multiple dimensions and ecological realms in which biodiversity can be observed. These deﬁnitions stress that biodiversity encompasses at least four kinds of diversities: genetic diversity, species or taxonomic diversity, ecosystem diversity, and landscape diversity (McAllister 1991; Solbrig 1993, Stuart and Adams 1991; Groombridge 1992; Heywood 1994, Wilson 1993). Two main problems emerge as a consequence of the broad scope that the biodiversity concept has taken at present. Cast as questions, the problems are: (1) How do we incorporate processes (e.g., foraging, energy and nutrient ﬂows, patch dynamics) into a concept that is based on seemingly static entities (i.e., individual organisms, species, habitat types, patch types)? (2) How do we integrate across ecological subdisciplines (e.g., 3

4

Introduction

ecosystem, population, landscape ecology) and across scales that are involved in biodiversity studies? The two problems are not mutually exclusive. Indeed, they are inseparable and complementary. For example, to determine how species diversity and ecosystem processes interact requires incorporation of entities and processes, as well as integration of community and ecosystem ecology. The focus on both entities and processes reﬂects the long-recognized dichotomy of structure and function in biology and ecology. Clearly, both structure and function must be integrated in order to successfully solve ecological questions. Dealing with biodiversity brings this needed integration into focus. The history of science shows that integration across disciplines is a critical component of scientiﬁc progress (Cohen 1985, Pickett 1999). Integration forces us to ask new questions, ﬁll gaps in understanding, facilitate information ﬂow among disciplines, and bridge dichotomies that arise due to divergence in disciplinary paradigms (Pickett et al. 1994). In addition, integration creates new disciplines, such as applied ecology, and impels us to address issues of scale. Indeed the introduction of the concept of biodiversity has led to new questions that had not been raised within the research agendas of population, community, ecosystem, and landscape ecology until recently (Gaston 1996). The study of biodiversity as an interfacing process between populations and ecosystems is an excellent example of the role of the concept of biodiversity in generating new questions and facilitating information ﬂows across subdisciplines (Levin 1997, Mooney et al. 1996). These interfacing studies address questions regarding the role of organisms (e.g., populations, species, functional groups) in system processes (e.g., nutrient retention, decomposition, production), and how system properties such as stability, resistance, invasibility, and predictability are aﬀected by organismal diversity. The attempts to answer the above questions include experiments in the ﬁeld and in artiﬁcial ecosystems (Naeem et al. 1995, Mooney et al. 1996, Lawton 1994, Naeem et al. 1994, Tilman et. al., 1996), ﬁeld data collection, modeling, and the introduction of new concepts and theories. New empirical data provide evidence that biotic diversity at levels ranging from genetic diversity among populations to landscape diversity is critical for the maintenance of natural and agricultural ecosystems (Schulze and Mooney 1993, Allen-Wardell et al. 1998). In addition to empirically based new ideas, the interfacing studies generated the novel concept of ‘‘functional biodiversity’’ (Grassle et al. 1991, Solbrig 1994, Martinez 1995), ‘‘organisms as ecosystem engineers’’ (Jones et al. 1994, 1997), and ‘‘ecosystem predictability’’ (McGrady-Steed et al. 1997). These new concepts identify important arenas for integration. Functional biodiversity addresses the variety of the relationships between speciﬁc groups of living entities and certain ecological processes. The concept demonstrates the lack of a theoretical framework to guide and prioritize the broad scope of observations, experiments, and management so characteristic of the ﬁeld of ecology (Lawton and Brown 1993). Essentially, the concept reverses the traditional relationship between biotic and abiotic factors. Whereas, classically, ecologists

Introduction

5

have used environmental parameters as predictors or independent variables, the concept of functional biodiversity suggests that the independent variables are biotic. The lexicon for such relationships (Lawton, 1994), the experimental methods, and the response dimensions to be measured in these variables, are not speciﬁed by existing ecological theory (McGrady-Steed et al. 1997). However, the concept of functional biodiversity advanced biodiversity studies by generating hypotheses that address the relationship among physiology, species diversity, and ecosystem function (Solbrig 1994, Martinez 1995, Schulze and Mooney 1993). An important conceptual aspect of functional biodiversity is the conceptual reﬁnement and reinvigoration of the diversity/stability issue, which had been abandoned as unproductive. Formerly, the issue was dealt with descriptively, using indirect measures based on vague concepts. Currently, the issue is being addressed in a functional and experimental manner. The inclusion of functional relationships, process studies, and studies of dynamics of the entities involved are key aspects of the power of the functional approach. The concept of ecosystem engineering addresses the variety of the relationships among organism activities, landscape diversity, ecosystem processes, and species diversity. The concept of organisms as ecosystem engineers also relates to biodiversity questions at the interface between population/community, ecosystem, and landscape ecology. This concept also introduces new feedback pathways between organisms and their environment. Currently we lack a conceptual framework or methodology for understanding engineering– biodiversity relationships. This is because these relationships lie outside of the traditional domains of population and ecosystem models. However, we see initial eﬀort for developing new models that combine population processes and engineering (Gurney and Lawton 1996). The concept of ‘‘ecosystem predictability’’ (McGrady-Steed et al. 1997) identiﬁes variations in ecosystem processes that are subjected to control by the richness of species in those systems. There are other new and important concepts that have arisen from the study of the role of biodiversity in system processes. These include the ‘‘redundant,’’ ‘‘rivet,’’ and ‘‘idiosyncratic’’ hypotheses (Walker 1992, Lawton and Brown 1993, Vitousek and Hooper 1993, Lawton 1997, Ehrlich and Walker 1998). Biodiversity includes more than the interface of species and ecosystem concepts (Martinez 1995). We recognize at least two more interfaces: species and landscape, and ecosystem and landscape. Since the time of Watt (1947), ecologists have recognized the importance of spatial pattern and its relationship to processes. Therefore, it is important to know how spatial patterns are related to species richness. The landscape mosaic redistributes resources. This is the link between ecosystem processes and landscape. This link is vital for determination of species diversity. We assume that the essential role of biodiversity studies is to highlight new, interfacing questions, concepts, and theories. We believe that as was shown for species–ecosystem interfacing, adding new interfaces will contribute to our understanding of the relationship between biodiversity and ecological func-

6

Introduction

tion. Dryland biodiversity studies can be a good start in this direction. Dryland systems are especially amenable to experimentation; they have manageable faunal and ﬂoristic diversities and tractable physical structures. The openness of many dryland ecosystems allows a degree of visualization and understanding diﬃcult to obtain in more complex environments. Although current expansions of the biodiversity concept are new, they have already demonstrated a remarkable ability to integrate formerly disparate areas, generate new data and related concepts, and serve as a foundation for a new theory. Still, the concept requires further elaboration and analysis. This book brings together chapters that focus on the wide range of interfaces between biodiversity and other ecological realms. To make the collection most useful, we present a framework to tie the variety of perspectives, concepts, and connections together. We introduce this framework below.

Toward a Framework for Biodiversity In this section we elaborate on the biodiversity concept and relate the book chapters to this comprehensive concept. We suggest that biodiversity refers to an assemblage of ecological entities, on various scales, appropriate to types and numbers of entities and the diﬀerences and interactions among them. In ﬁg. 1.1, we clarify the concept of ecological entities, their relation to scale and interactions, and generate four focal questions as a foundation for integrative biodiversity. Types and Number of Ecological Entities (Fig. 1.1(II)) Ecologists recognize many types of tangible and abstract ecological entities. These include: genes, species, functional groups, trophic levels, compartments in ecosystem models, resources that organisms use, habitats, and patches. For our purposes, the diversity of entities can be grouped into three types: organism-, resource-, and landscape-related entities. For example, a small unit of soil (1 1 1 m) encompasses organisms such as bacteria and fungi that can be classiﬁed by entities such as genes, species, and functional groups. These entities are mixed with diverse nitrogen resources such as parent material, rainfall, nitrogen ﬁxers and dry deposition. Another group of entities in this mixture is landscape entities, such as diﬀerent soil horizons or soil patches varying in air, water, and solid particle content. A principal question pertaining to all ecological entities is: What are the processes that control the number of a speciﬁc entity and all types of entities in a biosphere unit? More speciﬁcally, we ask: What are the processes that control the number of species, resources, and patches in ecological systems at diﬀerent spatial and temporal scales? (Fig. 1.1(II).) The numbers of ecological entities are usually scale dependent. An increase in spatial scale will increase the number of species and the number of habitats in which they live.

Introduction

7

Figure 1.1 A conceptual model capturing the essence and the focal questions of biodiversity studies. (I) Units of the biosphere are a mixture of ecological entities ( , *: ecological entities such as species and habitats). (II) A ﬁrst set of questions related to the dynamics of the numbers of the entities (?1: ﬁrst question). (III) The entities diﬀer in their properties (*, *, , ; for example, species diﬀer in body mass, while habitats diﬀer in patch types). ?2: The second focal question refers to the relationship between the number and the diﬀerences. (IV) The entities are organized by interactions. ?3: The third focal question focuses on the interactions of depicts the diﬀerent entity types and their eﬀects on number and diﬀerences ( control of the dynamics of one entity type, i.e., species, on another, i.e., patches). (V) The organized entities aﬀect ecosystem processes (primary production, decomposition, etc.). ?4: The fourth focal question deals with the number, diﬀerences, and interactions on system behavior.

8

Introduction

Several chapters in this book address the number of entities and scale relationships. Rosenzweig (chapter 15) explores what is known about the species–area relationship. Kark et al. (chapter 3) present a study of organismal diversity patterns across a distributional range of species, thus revealing areas with especially high genetic and morphological diversity. Ritchie and Olﬀ (chapter 11) suggest a scale-dependent model of species coexistence in relation to body size. Boeken et al. (chapter 9) show how the scale, structure, and deﬁnition of patches are relevant for the distribution of the organisms. Shachak et al. (chapter 8) suggest that hypotheses are dependent on the assumptions of water accessibility and utilization by a diverse plant assemblage and scale of analysis. Mitchell et al. (chapter 4) show scale-dependent biodiversity by comparing small-scale interactions with large-scale patterns. Kunin and Lennon (chapter 6) refer to the relationship between scale and incidence. Differences Among Entities (Fig. 1.1(III)) Entities diﬀer in their properties. Species diﬀer in many properties such as body size, behavior, and abundance. Similarly, patches are distinguished by a suite of properties. Some examples are size, resources, and spatial distribution. In the biodiversity context, the question is: What is the relationship between the number of entities and the degree of diﬀerences among them? This last question is the second focal question for a uniﬁed biodiversity framework. The second focal question integrates types of entities, their number, and the diﬀerences among them. Speciﬁcally, we ask how diﬀerences in a trait or a set of traits of an assemblage of species aﬀect the number of species in a given area. Or: How do diﬀerences in patch properties in a landscape mosaic aﬀect the number of species? Ziv et al. (chapter 5) propose a model that simulates how the number and distribution of diﬀerent-sized habitats aﬀect the number of species, of diﬀerent body mass. Shachak et al. (chapter 8) show how diﬀerences among species using water for biomass production along a gradient of soil moisture can aﬀect the productivity–diversity relationship. Boeken et al. (chapter 9) present a conceptual framework for connecting species diversity and landscapes. They discuss how changes in species assemblages, which diﬀer in abundance and frequency of occurrence in patches, coincide with changes in landscape structure. Ritchie and Olﬀ (chapter 11) discuss how diﬀerences in habitat, food, and resources may contribute to higher species diversity. Wilby et al. (chapter 10) show the role of animals in controlling the diﬀerences in local species assemblage, in terms of the number of individuals and the frequency of species occurrence. Interactions Among Entities (Fig. 1.1(IV)) Each assemblage of ecological entities is characterized by a web of interactions. From a biodiversity perspective, the third focal question is: What is the

Introduction

9

outcome, in terms of numbers and diﬀerences, of the interactions between diﬀerent types of entities? For example, we ask how changes in the number and properties of habitats in a landscape aﬀect the number and properties of species, and vice versa. Several chapters address the question of number, diﬀerence, and interaction. Mitchell et al. (chapter 4) discuss how variation in substrate, slope, solar input, and productivity contribute to species interactions at a local scale and patterns of species diversity. Ayal et al. (chapter 2) present a relationship between landscape heterogeneity in plant cover, food web structure, predation, and species diversity. Boeken et al. (chapter 9) introduce a new method and Ziv et al. (chapter 5) propose a simulation model, for studying species, landscape diversity relationships. In addition to the reciprocal eﬀect of the interactions among various entities, there is another outcome of interentity relationships; their eﬀect on ecosystem processes (ﬁg. 1.1(V)). This generates the fourth focal question: How does the organization of an assemblage of ecological entities control ecosystem processes and how do the ecosystem processes feedback to the number and diﬀerences of ecological entities? The existing studies in relation to entity organization, ecosystem processes and their feedbacks on organization focus on: (1) the relationship between species or functional group diversity and productivity, and (2) organisms as ecosystem engineers. In studies of the eﬀect of diversity on productivity, the assumption is that rates of ecosystem processes are determined by complementarity in resources used by diﬀerent species or functional groups. Under this assumption, the organization of the entities is controlled by the diversity of resources and their distribution in time and space. The study of organisms as ecosystem engineers refers to the modulation of the landscape by organisms that modify resource distribution. This activity reorganizes species and landscape diversity. Shachak et al. (chapter 8) present a conceptual model linking species properties and ecosystem processes for water-limited systems. They integrate species, water, and energy ﬂows to demonstrate the relationships between species diversity and productivity. Wilby et al. (chapter 10) show how animals modify environment structure and reorganize the entities in the system, thereby aﬀecting species number and diﬀerences. Boeken et al. (chapter 9) deal with the relationship between species and landscape diversity and how this relationship aﬀects productivity.

A Framework for Biodiversity Management Biodiversity management refers to human inﬂuences on the organization and interactions of an assemblage of ecological entities. The aim is to have a desired number and variety of ecological entities in a managed biosphere unit. Any biodiversity management is a manipulation of landscape diversity, succession, ecosystem processes, and species diversity (Pickett et al. 1997, 1999). This is the consequence, even when the objective of management is

10 Introduction

to manipulate only one of the components. This suggests that development of a biodiversity framework that integrates among ecological entities interactions with ecosystem processes should enable managers to follow the chain of interactions among landscape diversity, ecosystem processes, and species diversity. Therefore, biodiversity management is a part of ecosystem management. Ecosystem management is the manipulation of ecosystems to satisfy speciﬁed societal values. This deﬁnition is useful because biodiversity is managed to meet human values such as high species and landscape diversity and the goods and services that they provide. Several chapters address management issues. Perevolotzky et al. (chapter 16) present a conceptual model on human–biodiversity relationships in waterlimited systems. They demonstrate how humans have actively managed landscape diversity to direct ecosystem processes for enhancing the distribution and abundance of organisms that provide ecosystem services. Pregenzer et al. (chapter 14) outline four diﬀerent kinds of large-scale data required by land managers for the development of sustainable land-use strategies that can be met with current or future technologies. Ward (chapter 13) contributes to the understanding of the eﬀect of grazing by domestic animals on dryland biodiversity by comparing North American and African studies. Rosenzweig (chapter 15) discusses how to manage ecosystems to stop mass extinction of species. He suggests methods for sharing anthropogenic habitats with species. He also proposes techniques for providing conditions needed by both native species and human society.

Book Organization Most of the chapters in this book refer to the relationship among three components: organism diversity, ecosystem processes, and landscape diversity. However, in many of the chapters there are diﬀerent emphases on these three constituents. Therefore, we grouped chapters within two main parts according to the gist of the particular chapter; whether they focus on the organism, ecosystem, or landscape perspective. We also included a third part dedicated to biodiversity, conservation, and management. In part I, ‘‘Living Components of Biodiversity: Organisms,’’ we determined the order of the chapters according to topics: chapters 2 and 3 are devoted to biodiversity studies on a relatively small scale; chapters 4 and 5 deal with concepts and models; chapter 6 deals with species diversity on a large scale. In part II, ‘‘Ecological Complexes of Biodiversity, Ecosystems, and Landscapes,’’ chapter 8 focuses on the role of ecosystem science in biodiversity, while chapters 9 to 11 emphasize the landscape aspect. In part III, ‘‘Biodiversity, Conservation, and Management,’’ chapters 13 and 14 deal with the relationship between biodiversity and grazing systems; chapter 15 concerns species conservation.

Introduction

11

At the end of each part, ideas for a uniﬁed framework are proposed. Chapter 7 provides a uniﬁed framework for biodiversity from the community and population ecology perspective. Chapter 12 suggests a uniﬁed framework for biodiversity from the ecosystem perspective, while chapter 17 indicates human interactions with biodiversity. The opening chapter suggests a conceptual framework for biodiversity studies and demonstrates how the ensuing chapters contribute to the idea. Chapter 18, the last chapter, suggests a uniﬁed framework for biodiversity studies.

References Allen-Wardell, G. et al. (21 authors). 1998. The potential consequences of pollinator declines on the conservation of biodiversity and stability of food crop yields. Conservation Biology 12: 8–17. Cohen, J.B. 1985. Revolution in Science. Belknap Press, Cambridge, MA. Ehrlich, P., and B. Walker. 1998. Rivets and redundancy. BioScience 48: 387. Gaston, K.J (ed.). 1996. Biodiversity – A Biology of Numbers and Diﬀerences. Blackwell Science, Oxford. Grassle, J.F., P. Lasserre, A.D. McIntyre, and G.C. Ray. 1991. Marine biodiversity and ecosystem function. Biology International Special Issue 23: 1–19. Groombridge, B. 1992. Global Biodiversity: Status of the Earth’s Living Resources. Chapman and Hall, London. Gurney, W.S.C., and J.H. Lawton. 1996. The population dynamics of ecosystem engineers. Oikos 76: 273–283. Heywood, V.H. 1994. The measurement of biodiversity and the politics of implementation. In: P.L. Forey, C.J. Humphries, and R.I. Vane-Wright (eds.), Systematics and Conservation Evaluation, pp. 15–22. Oxford University Press, Oxford. Jones, C.G., J.H. Lawton, and M. Shachak. 1994. Organisms as ecosystem engineers. Oikos 69: 373–386. Jones, C.G., J.H. Lawton, and M. Shachak. 1997. Positive and negative eﬀects of organisms as physical ecosystem engineers. Ecology 78: 1946–1957. Lawton, J.H. 1994. What do species do in ecosystems? Oikos 71: 367–374. Lawton, J.H. 1997. The role of species in ecosystems: aspects of ecological complexity and biological diversity. In: A. Takuya, S.A. Levin, and M. Higashi (eds.), Biodiversity: An Ecological Perspective, pp. 215–228. Springer-Verlag, Berlin. Lawton, J.H., and V.K. Brown. 1993. Redundancy in ecosystems. In: E.D. Schulze and H.A. Mooney (eds.), Biodiversity and Ecosystem Function, pp. 255–270. Springer-Verlag, Berlin. Levin, S.A. 1997. Biodiversity: interfacing populations and ecosystems. In: A. Takuya, S.A. Levin, and M. Higashi (eds.), Biodiversity: An Ecological Perspective, pp. 277–287. Springer-Verlag, Berlin. Martinez, N.D. 1995. Unifying ecological subdisciplines with ecosystem food webs. In: C.G. Jones and J.H. Lawton (eds.), Linking Species and Ecosystems, pp. 166–175. Chapman and Hall, New York. McAllister, D.E. 1991. What is biodiversity? Canadian Biodiversity 1: 4–6. McGrady-Steed, J, P.M. Harris, and P.J. Morin. 1997. Biodiversity regulates ecosystem predictability. Nature 390: 162–165.

12 Introduction

Mooney, H.A., J.H. Cushman, E. Medin, O.E. Sala, and E.D. Schulze. 1996. What we have learned about the ecosystem functioning of biodiversity. In: H.A. Mooney, J.H. Cushman, E. Medlin, O.E. Sala, and E.D. Schulze (eds.), Functional Roles of Biodiversity: A Global Perpective, pp. 475–484. Scope 55, John Wiley and Sons, New York. Naeem, S., L.J. Thompson, S.P. Lawler, J.H. Lawton, and R.M. Woodﬁn. 1994. Declining biodiversity can alter the performance of ecosystems. Nature 368: 734–736. Naeem, S., L.J. Thompson, S.P. Lawler, J.H. Lawton, and R.M. Woodﬁn. 1995. Empirical evidence that declining species diversity may alter the performance of terrestrial ecosystems. Philosophical Transactions of the Royal Society of London B 347: 249–262. Pickett, S.T.A. 1999. The culture of synthesis: habits of mind in novel ecological integration. Oikos 87: 479–487. Pickett, S.T.A., J. Kolassa, and C.G. Jones. 1994. Ecological Understanding: The Nature of Theory and the Nature of Nature. Academic Press, Cambridge. Pickett, S.T.A., R.S. Ostfeld, M. Shachak, and G.E. Likens, (eds.) 1997. The Ecological Basis of Conservation: Heterogeneity, Ecosystems, and Biodiversity. Chapman and Hall, New York. Pickett, S.T.A., M. Shachak, B. Boeken, and J.J. Armesto. 1999. The management of ecological systems. In: T.W. Hoekstra and M. Shachak (eds.), Arid Lands Management: Toward Ecological Sustainability, pp. 8–17. University of Illinois Press, Urbana. Schulze, E.D., and H.A. Mooney (eds.) 1993. Ecosystem Function of Biodiversity. Springer-Verlag, Berlin. Solbrig, O.T. 1993. Plant traits and adaptive strategies: their role in ecosystem function. In: E.D. Schulze and H.A. Mooney (eds.), Biodiversity and Ecosystem Function, pp. 97–116. Springer-Verlag, Berlin. Stuart, S.N., and R.J. Adams. 1991. Biodiversity in Sun Saharan Africa and Its Islands. World Conservation Union, Gland, Switzerland. Tilman, D., D. Wedin, and J. Knops. 1996. Productivity and sustainability inﬂuenced by biodiversity in grassland ecosystems. Nature 379: 718–720. Vitousek, P.M., and D.U. Hooper. 1993. Biological diversity and terrestrial ecosystem biogeochemistry. In: E.D. Schulze and H.A. Mooney (eds.), Biodiversity and Ecosystem Function, pp. 3–14. Springer-Verlag, Berlin. Walker, B.H. 1992. Biodiversity and ecological redundancy. Biological Conservation 6: 18–23. Watt, A. S. 1947. Pattern and process in the plant community. Journal of Ecology 35: 1–22. Wilson, E.O. 1993. The Diversity of Life. Harvard University Press, Cambridge, MA.

Part I

Living Components of Biodiversity: Organisms

This page intentionally left blank

2 How Can High Animal Diversity Be Supported in Low-Productivity Deserts? The Role of Macrodetritivory and Habitat Physiognomy Yoram Ayal Gary A. Polis* Yael Lubin Deborah E. Goldberg

O

n large spatial scales, species diversity is typically correlated positively with productivity or energy supply (Wright et al. 1993, Huston 1994, Waide et al. 1999). In line with this general pattern, deserts are assumed to have relatively few species for two main reasons. First, relatively few plants and animals have acquired the physiological capabilities to withstand the stresses exerted by the high temperatures and shortage of water found in deserts (reviewed by Noy-Meir 1974, Evenari 1985, Shmida et al. 1986). A second, more ecological mechanism is resource limitation. In deserts, the low and highly variable precipitation levels, high temperatures and high evapotranspiration ratios limit both plant abundance and productivity to very low levels (Noy-Meir 1973, 1985, Polis 1991d). This lack of material at the primary producer level should exacerbate the harsh abiotic conditions and reduce the abundance of animals at higher trophic levels by limiting the types of resources and their availability. Animal abundance should be even further reduced because primary productivity is not only low, but also tends to be sporadic in time and space (MacMahon 1981,

Gary Polis died tragically during the preparation of this chapter, before he could ﬁnish contributing his insights and examples based on his broad experience in many desert ecosystems. We dedicate our eﬀorts to Gary, with the deepest sadness and regret for his loss. 15

16 Living Components of Biodiversity: Organisms

Crawford 1981, Ludwig 1986). Herbivores should have diﬃculties tracking these variations (e.g., Ayal 1994) and eﬃciently using the available food resources. Hence, herbivore populations in deserts have low densities relative to other biomes (Wisdom 1991) and most of the primary productivity remains unused (Crawford 1981, Noy-Meir 1985). This low abundance of herbivores should propagate through the food web and result as well in lower abundance of higher trophic levels. The number of individuals and the number of species are not always positively correlated; in particular, some examples of low diversity at high productivity with high densities are well documented (e.g., salt marshes, reviewed by Waide et al. 1999). However, several distinct mechanisms have led to the expectation that when productivity and the number of individuals are low, the number of species is also likely to be low. First, within trophic levels, the ‘‘statistical mechanics’’ model of Wright et al. (1993) may operate. In this model, the amount of energy present determines the probability distribution of population sizes for the members of the species pool in a region. A species with a larger population size has a higher probability of occurrence in a given patch and therefore a higher cumulative probability that it will be present in any patch in a region. As energy supply increases, the probability distribution of population sizes shifts upward and more species are likely to occur in the region. Suﬃciently low food available at the base of the food chain means that fewer trophic levels (and fewer functional groups of species) can be supported because of the ineﬃciency of energy transfer along the food chain (Elton 1927, Lindemann 1942, Fretwell 1977, Oksanen et al. 1981). Fewer trophic levels are likely to lead to fewer species because of fewer possible roles in the community. A third mechanism by which low abundance leads to low diversity explicitly assumes that competition among consumers is an important process. Reduction of the number of species within any given trophic level should also lead to reduction in diversity at higher trophic level because of fewer opportunities for specialization and resource partitioning based on prey type. In addition, low prey density should support only generalist predators that can use most prey types they may encounter. Hence, desert predators are expected to overlap widely in their diets, which conventional competition theory tells us should limit the number of predators that can coexist (Polis 1991d). Despite the logic of these arguments for low species diversity in deserts, empirical studies demonstrate that diversity of at least above-ground desert animal communities is actually quite high (Polis 1991c and review in Polis 1991a). Representatives of almost all terrestrial animal taxa are found in deserts (Polis 1991c). Physiological and behavioral adaptations for desert conditions have appeared many times in evolutionary history, suggesting that these traits are acquired easily by animals. Thus, harsh environmental conditions have not directly limited the number of species in deserts. But, the above energy-based arguments still hold and pose the question how the high

High Animal Biodiversity in Low-Productivity Deserts

17

species diversity observed in deserts can be maintained, based on such a low productivity. In this chapter, we develop several hypotheses about the solutions to this apparent paradox. First, although primary productivity is low, deserts tend to have higher ecological eﬃciencies of transfer to higher trophic levels, and thus a given amount of food at the base can support higher abundances at higher trophic levels. Because several of the arguments described above for why low primary productivity should lead to low diversity are based on low densities of available prey for higher trophic levels, this means that diversity should also be higher than expected, based solely on energy available. The higher net ecological eﬃciencies in deserts are due to two mechanisms: (1) the dominance of macrodetritivores rather than herbivores as the central link between primary and secondary productivity, leading to a higher proportion of the energy in detritus being transferred to animals in above-ground food webs in deserts, and (2) the dominance of the above-ground food web by small arthropods and poikilotherms, with lower metabolic requirements and hence higher eﬃciencies (Turner 1970, McNeill and Lawton 1970). Second, the energy-based arguments for low diversity in deserts assume that resource limitation is ubiquitous and therefore that competition for resources is the dominant ecological interaction in deserts. In contrast, we argue that in many cases macrodetritivores are not limited by their resources and instead, that predator-mediated coexistence is a particularly important diversity-maintaining process in deserts. In particular, we argue that deserts have high spatial heterogeneity of plant cover and thus have high spatial heterogeneity of predation intensity due to use of cover as refugia. This in turn should lead to great potential for predator-mediated habitat partitioning (Holt 1977), which then allows higher diversity than expected, based solely on energy considerations. Because data on food web structure in deserts around the world is rather limited, our ideas are based primarily on our own studies in the Negev Desert (Israel), the North American deserts, and the Namib desert. We hope that our arguments will give new impetus to the study of desert communities to examine the generality of the hypotheses we develop to explain high diversity in deserts despite their low productivity.

Higher Ecological Efﬁciency in Deserts: The Desert Food Web In this section, we focus largely on why animal abundance in deserts is likely to be higher than expected, based solely on the amount of primary productivity available. As described above, the mechanisms that have led to the expectation of low diversity in deserts are direct consequences of this expected low abundance and, therefore, our arguments also directly lead to an explanation of higher than expected diversity.

18 Living Components of Biodiversity: Organisms

Macrodetritivores as the Main Primary Consumers in Desert Food Webs In most terrestrial communities, 80–90% of primary productivity becomes plant litter (detritus) to be decomposed by microbes either in the soil, or on the soil surface. Most of this litter is mineralized or recycled within the soil food web and only a minor portion ﬁnds its way back to the above-ground food web (reviewed by Hairston and Hairston 1993), the one usually studied by community ecologists (Polis and Strong 1996). For deserts, we argue that even more of the primary productivity goes into litter but that less of it is mineralized or stays in the soil food web. Instead, we hypothesize that macrodetritivores transfer the energy and minerals from detritus directly to higher trophic levels in the above-ground food web, leading to greater overall eﬃcient transfer from primary producers to higher trophic levels. As the growing season for plants is short in deserts, desert herbivores need to complete their development within this short period and survive the long period between consecutive growing seasons either by subsisting on a poor resource or by being inactive. Most herbivores in deserts are ephemeral insects (e.g., grasshoppers, moths, beetles) with only one reproductive cycle per year, therefore they cannot track the high between-year variability in productivity levels (Ayal 1994). In years of high productivity, most of the unconsumed plant production dries up and turns to plant litter. In addition, desert soils also should contain high amounts of plant detritus because of the high allocation to roots typical of desert plants. Plants in deserts are limited by soil resources, especially water, and by nitrogen in times of high precipitation (Noy-Meir 1973, 1985). Under such conditions, plants allocate a high proportion of their energy to develop their root system and less to their shoots (high root:shoot ratio, Tilman 1988). Indeed, Cody (1986) found that the root:shoot ratio in desert shrubs is around 2, while mesic habitats typically have root:shoot ratios less than 1 (Tilman 1988). Thus, compared with more mesic areas, deserts should have a higher proportion of primary production going into detritus than to herbivores. In addition to being more abundant, the detritus in deserts should be less likely to be broken down by free-living microbes. The low moisture in deserts during most of the year limits microbial decomposition of both above- and below-ground plant litter (Vishnevetsky and Steinberger 1997). This seasonal limitation of decomposition lent support to Huston’s (1994) argument on the importance of the length of the growing season to productivity (or, in this case, to decomposition). The low decomposition rate results in a rich and reliable resource (above- and below-ground litter) for those animals that use it, namely, macrodetritivorous arthropods such as isopods, termites, and tenebrionid beetles (larvae and adults) (Johnson and Whitford 1975, Crawford 1981, 1991, MacKay 1991). One can say, therefore, that in deserts most of the microbial decomposition takes place within the macrodetritivores’ gut rather than in the soil (Crawford and Taylor 1984). This explains

High Animal Biodiversity in Low-Productivity Deserts

19

the high biomass of macrodetritivores in deserts (Crawford 1991, Mackay 1991). But more important to the current discussion of deserts, the diverse group of macrodetritivores forms an integral part of the above-ground food web (e.g., Polis 1991b, Ayal and Merkl 1994). This contrasts with more mesic communities in which the soil food web has relatively few links to the aboveground web (e.g., earthworms and larvae eaten by moles and birds). Accordingly, macrodetritivores in deserts take the place of herbivores as the major group of primary consumers linking primary producers to higher trophic levels. This replacement has several important implications for energy ﬂow and species diversity in deserts. One critical distinction between herbivores and macrodetritivores is that macrodetritivores have no negative eﬀect on plant productivity. This contrasts the common consumer-producer dynamics in which high consumption by the consumer has a negative eﬀect on the producer dynamics, and the interaction drives both the producer and the consumer populations to low density levels. Thus in herbivore–plant interactions, high eﬃciency of the herbivore results in a reduction in plant biomass and consequently a reduction in plant productivity (e.g., Noy-Meir 1975). However, macrodetritivores consume the plant tissue after its death and thus do not aﬀect plant dynamics or productivity directly. In fact in deserts, macrodetritivores are the main decomposers and thus contribute to nutrient recycling from plant litter back to the soil. In addition, many detritivores are burrowers (e.g., termites, isopods, and tenebrionid larvae) and thus contribute to soil turnover and increase water inﬁltration into the soil. Consequently, macrodetritivores may commonly have positive eﬀects on plant productivity in deserts (e.g., Whitford 1986). The slow rate of microbial decomposition in deserts and the shift of plant litter decomposition to macrodetritivores are unique features of desert communities and are key factors in understanding their structure and function. This shift changes profoundly the proportion of primary productivity channeled into the above-ground food web relative to the proportion channeled in other terrestrial ecosystems. In other terrestrial ecosystems, herbivores use less than 20% of the primary productivity, and more than 80% is lost to microbial decomposition and does not ﬁnd its way back to the above-ground food web (Hairston and Hairston 1993). In deserts, microbial decomposition accounts for less than 10% of plant primary productivity (Whitford 1986). Even if some of the plant litter is lost by physical degradation or transported by wind or water, the majority of the primary productivity is probably still consumed by macrodetritivores and channeled into the above-ground food web (reviewed by Whitford 1986). Thus, despite the low level of primary productivity in deserts, the energy base of the above-ground food web is likely to be much higher than is generally assumed based on patterns found in other terrestrial communities. This could lead to higher diversity either by permitting a greater degree of prey-specialization by consumers within a trophic level or by allowing the addition of more trophic levels.

20 Living Components of Biodiversity: Organisms

Desert Communities Dominated by Small and Poikilothermic Consumers As early as 1927, Elton recognized the importance of the energy base of a community to the length of food chain found in it (his pyramid of numbers, and later, the pyramid of biomass; Bodenheimer 1938). This energy-centered approach to community ecology was elaborated by Lindemann (1942) and currently dominates the discussion on community structure and function (Fretwell 1977, Oksanen et al. 1981, Oksanen 1992, Hairston and Hairston 1993, Hairston 1997, Polis and Strong 1996). However, Elton (1927) also recognized that organism size plays an important role in food chains, although this issue has been relatively neglected in the discussion of food chain length (but see Hutchinson 1959, Yodzis 1984, Cousins 1987, Hairston and Hairston 1993). Carnivores are generally larger than their prey (the gape-limitation hypothesis, Zaret 1980). Thus, a food chain based on small primary consumers (i.e., insects, zooplankton) can include more links than one that starts with even a small mammalian herbivore. The desert food chain that is based on macrodetritivorous insects includes predatory insects, arachnids, and reptiles as primary predators (e.g., Morton and James 1988) and larger reptiles and predatory birds as secondary predators, with mammals (e.g., jackals or coyotes) becoming important only in relatively more productive arid grasslands (Brown 1986). The desert food web is also a poikilotherm-based web. The main primary and secondary consumers are poikilotherms. Food chains based on poikilotherms are energetically more eﬃcient than food chains based on the homeothermic mammals or birds. This is because poikilotherms have lower basic metabolic rates and lower energy requirements for maintenance than homeotherms (Table 41-4 in Ricklefs 1973, Humphreys 1979, McNeill and Lawton 1970). Thus, a given amount of primary productivity will enable more trophic links in a poikilotherm- than in a homoeotherm-based food chain (Yodzis 1984). Reagan et al. (1996) evoked a similar argument when discussing an island tropical forest food web. They suggested that the absence of large herbivores and the dominance of poikilotherms in the food web in the community they studied resulted in longer food chains than in similar continental communities where mammal herbivores and predators are common. Being small and poikilothermic allows members of the desert food web to fulﬁll their energetic demands by utilizing a smaller area of habitat than their larger homeothermic counterparts in more productive habitats. Thus, higher densities of macrodetritivores can live in 100 m2 of desert than can the number of voles in a single hectare of productive meadow or ungulates in an African grassland. Similarly, at the next higher trophic level, more scorpions or lizards can live in the same area in deserts than the weasels or lions of similar trophic position in more mesic productive communities. Hence, all else being equal, food webs made up of largely small or poikilothermic organisms should be able to support higher population densities and thus more trophic levels and more species than larger, homeothermic organisms.

High Animal Biodiversity in Low-Productivity Deserts

21

To summarize, we argue that deserts have greater ecological eﬃciency than many mesic environments. We also argue that, in contrast to the current expectation that deserts are one-link communities (Fretwell 1977, Oksanen et al. 1981, Oksanen and Oksanen 2000), desert communities have at least four links when reduced to their basic trophic levels (ﬁg. 2.1; for a more detailed approach to desert food webs see Polis 1991b). Thus, both the within- and the between-trophic level mechanisms we described earlier to link increasing numbers of individuals with increasing numbers of species may operate, and at least partially account for the observed high diversity of desert animal communities.

Greater Importance of Predation in Deserts: The Role of Plant Cover in Mediating Predator–Prey Interactions and Promoting Diversity In this section, we develop the argument that predation is an important factor in promoting high diversity of animals in deserts. Many attributes of desert organisms and communities suggest that predation is a major force. For example, in desert communities there is a high proportion of predators and a high frequency of cryptic coloration, and activity is frequently limited to times of day when predators are ineﬃcient or inactive (Louw and Seely 1980). In addition, both predatory arthropods (reviewed by Polis and Yamashita 1991) and predatory nonmammalian vertebrates (reviewed by Vitt 1991, Wiens 1991) are common in deserts (Polis 1991b). We ﬁrst argue that the high degree of horizontal redistribution of water characteristic of deserts results in large spatial heterogeneity in productivity, and thus in plant cover. Plant cover is an important refuge for small animals, which are consumed by large predators. This results in high spatial heterogeneity in predation intensity and hence the potential for predator-mediated habitat segregation, yielding enhanced coexistence and diversity.

Figure 2.1 The basic above-ground trophic structure of desert communities. Main groups of macrodetritivores are termites, isopods, and tenebrionids; of primary predators are arachnids and reptiles; of secondary predators are birds. Note that macrodetritivores have no negative eﬀect on plant productivity.

22 Living Components of Biodiversity: Organisms

High Spatial Heterogeneity of Productivity and Plant Cover in Deserts Water is the main resource that limits plant productivity in arid habitats (Noy-Meir 1973, 1985). Because the amount of direct precipitation in arid environments is low, horizontal redistribution of water has a huge impact on local productivity at small and intermediate scales. Redistribution of water depends on the velocity of the rain event, substrate features at the contributing and recipient sites, plant cover, and relief (Yair and Danin 1980). Poor habitats are those with poor absorption capabilities and low water retention capacity (e.g., rocky surfaces or steep slopes). Rich habitats have high absorption and water retention capabilities and are surrounded by large contributing surfaces. Rich patches may be small depressions created by animals digging for food (Alkon 1999) or burrows (reviewed by Whitford and Kay 1999), or washes that collect water from small areas and ones that carry water from large catchments. This redistribution of water results in local gradients of productivity. In turn, the local variation in productivity results in especially strong variation in plant cover and microhabitat physiognomy in arid environments because small changes in the amount of available water determine whether a given site is barren or supports sporadic growth of annuals, a small perennial, or a large shrub. The spatial heterogeneity in productivity and cover within an area depends on the topography of the area. Regions with low relief (e.g., sand ﬁelds, large plains) will have relatively less spatial heterogeneity in productivity and plant cover than regions with high relief (e.g., high dunes, rocky ridges, and wadis). Therefore, areas with low relief should have overall productivity that is more easily predictable by rainfall than areas with high relief.

High Spatial Heterogeneity in Cover Leads to High Variation in Predation Efﬁciency by Large Secondary Predators Secondary predators of the desert food web consist primarily of large predatory birds, with reptiles and large omnivorous mammals becoming important in some areas. Birds can move easily to forage over large areas and exploit areas with episodic rich food, and migrate to other regions in seasons of low food availability (Wiens 1991). Most of these predators are visually oriented and hover, perch, or walk while foraging. Increasing the amount of plant cover in the habitat can therefore greatly reduce the eﬃciency of these secondary predators. Consequently, the intensity of predation by secondary predators in deserts is negatively correlated with the amount of plant cover in the habitat (e.g., Ayal and Merkl 1994, Seely 1985, Kotler 1997). It is interesting to note that in the Namib desert where plants are scarce and cover is minimal, a diverse fauna of tenebrionid beetles subsist on wind-driven detritus, and these macrodetritivores are active at the hottest period of the day. Although competition for resources was evoked to explain this unexpected

High Animal Biodiversity in Low-Productivity Deserts

23

pattern (Hamilton 1971), an alternative and more plausible explanation is predation avoidance. The tenebrionids become active at times when their largely nocturnal avian and mammalian predators cannot be active.

High Heterogeneity in Secondary Predation Efﬁciency Leads to High Heterogeneity in Primary Predator Abundance and Predation Efﬁciency Primary predators in deserts are mainly arthropods (insects, spiders, scorpions, solifugids) and smaller vertebrates (lizards, snakes, shrews, and some insectivorous rodents). These predators hunt within the vegetation cover and utilize cover for their own protection against second-tier predators (e.g., Skutelsky 1996). The abundance and diversity of primary predators should be correlated positively with plant cover (ﬁg. 2.2). This is true for spiders, for which plant cover, density, and structural diversity can strongly inﬂuence species diversity and abundance (Brandt and Lubin 1998, Robinson 1981). Primary predators are more abundant in areas of high plant cover both because the small (largely arthropod) detritivores are more abundant, and because they too will beneﬁt from the structural diversity of plant cover for concealment and defense from their own predators. Finally, many secondary predators and parasitoids are small and also forage within the vegetation layer. Thus, for example, araneophagic and oophagic spiders (Whitehouse

Figure 2.2 The suggested eﬀect of habitat physiognomy on trophic interactions in habitats of relatively low and high productivity within deserts. Arrows point to the aﬀected trophic level and their width indicates the relative strength of the eﬀect. In low-productivity habitats, plant cover is low and secondary predators are highly eﬃcient and limit the density and activity of primary predators. As a result, macrodetritivores are released from predation and become food limited. In highproductivity habitats, plant cover is high and provides shelter to primary predators from secondary predators. As a result, primary predators are both abundant and highly active and limit the densities of macrodetritivores. Hence, in productive habitats much of the plant litter remains unconsumed.

24 Living Components of Biodiversity: Organisms

and Lubin 1998), predatory hemipterans, and parasitoid wasps constitute a second tier of small predators whose abundance should correlate positively with plant cover. This diversity of small predators leads to eﬃcient utilization of the macrodetritivore resource in those habitats rich in vegetation cover, but low predation intensity in nearby habitats with lower vegetation cover. Effects of Heterogeneity in Predation on Regulation and Abundance of Trophic Groups The above-described relationships between predation intensity and plant cover in deserts have a substantial eﬀect on the abundance and regulation of diﬀerent trophic groups in these habitats. In poor habitats (e.g., plains in the Negev Desert), plant cover is low and bird predation on primary predators is intense. Therefore, primary predators are found in low abundance and their activity is limited to areas around refuges such as shrubs (e.g., Skutelsky 1996) or burrows (Shachak and Brand 1983). As a result, macrodetritivores are released from predation, and are relatively abundant and should be limited only by their food resources. In contrast, plant cover is high in rich desert habitats (e.g., wadis) and predation intensity by birds is lower. In rich habitats, primary predators are abundant, leading to high predation on macrodetritivores. This high predation, combined with the high production of detritus, suggests that food is unlikely to be limiting to macrodetritores in rich desert habitats, so competition among the macrodetritivores is unlikely to be important. Effect of Heterogeneity in Predation Intensity on Species Diversity The regulation of abundance and diversity in deserts by means of predation, rather than competition, can inﬂuence species diversity through mechanisms operating either within or between habitats. High predation intensity within rich desert habitats means that food limitation for macrodetritivores and competition for resources among them is unlikely to be important within those habitats. Instead, we suggest that this release from competition directly increases the abundance and potential diversity of macrodetritivores. Hence, the food web based on macrodetritivores within rich desert habitats is also potentially more diverse. The spatial heterogeneity in predation intensity among desert habitats promotes habitat segregation and thus may also increase diversity. For example, Ayal and Merkl (1994) suggested that habitat and size-dependent predation on tenebrionids was the most probable explanation for the observed habitat segregation of tenebrionids in the Negev Desert. They found that in compact soil habitats, body size of tenebrionid adults and amount of plant cover are positively correlated: small species are common in plains with low plant cover, medium-size species are common in slope habitats with intermediate plant cover, and large species are common in wadis with high plant cover.

High Animal Biodiversity in Low-Productivity Deserts

25

This habitat segregation is consistent with the preference of large predatory birds for large tenebrionids, which therefore only survive when plant cover is available as a refuge. In contrast, small tenebrionid species are eaten mainly by scorpions, which are most abundant in the plant-covered wadis but are less eﬀective predators in the plain. Cage experiments with birds and artiﬁcial cover support this hypothesis (Groner and Ayal 2001). And birds were also reported as an important factor in determining prey communities in other deserts (Seely 1991, Wiens et al. 1991, Polis 1991a, Floyd 1996). In addition to spatial heterogeneity of predation, high diversity of detritivores is also likely promoted directly by other aspects of the high spatial heterogeneity of the environment described above, accompanied by habitat specialization of the detritivores (see chapter 7 this volume, Seely 1985). Even if plant detritus is not amenable to food partitioning (Crawford 1981, 1991), the habitat in which it is consumed may promote special adaptations. For example, tenebrionid beetles show diverse morphological adaptations that can be attributed to habitat-substrate specialization (e.g., Medvedev 1965, Louw and Seely 1982) and habitat segregation according to substrate type has been demonstrated in tenebrionids (Ayal and Merkl 1994, Krasnov and Ayal 1995, Seely 1985, 1991). However, despite the high spatial substrate diversity found in deserts, the number of coexisting species is an order of magnitude higher than the available types of substrate, so that direct eﬀects of heterogeneity are unlikely to be the sole explanation of high diversity in macrodetritivores. The predation hypothesis (Ayal and Makl 1994, Groner and Ayal 2001) mentioned above is a possible solution to this apparent paradox of high diversity of macrodetritivores in deserts despite the fact that detritus is a highly accessible resource of low nutritional value and relatively uniform composition.

Conclusions We challenge the idea that low productivity in deserts causes low diversity when the whole community is examined (rather than a speciﬁc guild, as is typically done). We also challenge the idea that desert communities are simple in structure and that biotic interactions are not important for understanding species distribution and abundance in them. We suggest that the desert aboveground food web is based on macrodetritivores as the primary consumers, and because both primary consumers and primary predators are ectotherms, the web is highly eﬃcient energetically. Because of low microbial decomposition rates, macrodetritivores use a high proportion of the primary productivity, become abundant, and form a base to another two trophic levels. The heterogeneity in plant cover adds another dimension to the biotic interactions along the desert food chain. Plant cover mediates the strength of interactions between the secondary and primary predators. In this way plant cover also aﬀects the strength of interactions between the primary predators and the macrodetritivores. Yet the low rate of microbial decomposition probably has

26 Living Components of Biodiversity: Organisms

a negative eﬀect on the diversity of soil organisms in deserts. This low belowground diversity may, in part, counteract the increase in the above-ground diversity that results from the shift of energy from the below- to the aboveground community typical of deserts. However, in contrast to the dominant paradigm about above-ground, desert community structure (Shmida et al. 1986), we suggest that at least the above-ground compartment of desert communities is shaped by biotic interactions, and desert communities as a whole have some unique features compared with other terrestrial communities.

References Alkon, P.U. 1999. Microhabitat to landscape impacts: crested porcupine digs in the Negev Desert highlands. Journal of Arid Environments 41: 183–202. Ayal, Y. 1994. Time-lags in insect response to plant productivity—signiﬁcance for plant insect interactions in deserts. Ecological Entomology 19: 207–214. Ayal, Y., and O. Merkl. 1994. Spatial and temporal distribution of tenebrionid species (Coleoptera) in the Negev Highlands, Israel. Journal of Arid Environments 27: 347–361. Bodenheimer, F.S. 1938. Problems of Animal Ecology. Oxford University Press, Oxford. Brandt, J. and Y. Lubin. 1998. An experimental manipulation of vegetation structure: consequences for desert spiders. Israel Journal of Zoology 44: 201–216. Brown, J.H. 1986. The roles of vertebrates in desert ecosystems. Pp. 51–71. Whitford, W.G., ed., Pattern and Processes in Desert Ecosystems. University of New Mexico Press, Albuquerque. Cody, M. 1986. Structural niches in plant communities. Pp. 381–405 in J. Diamond and T. Case, eds., Community Ecology. Harper and Row, New York. Cousins, S. 1987. The decline of the trophic level concept. Trends in Ecology and Evolution 2: 312–316. Crawford, C. 1991. The community ecology of macroarthropod detritivores. Pp. 89–112 in G. Polis, ed., The Ecology of Desert Communities. The University of Arizona Press, Tucson. Crawford, C., and E. Taylor. 1984. Decomposition in arid environments: role of the detritivore gut. South African Journal of Science 80: 170–176. Crawford, C.S. 1981. Biology of Desert Invertebrates. Springer Verlag, New York. Elton, C. 1927. Animal Ecology. Sidgwick and Jackson, London. Evenari, M. 1985. Adaptations of plants and animals to the desert environment. Pp. 79–92 in M. Evenari, I. Noy-Meir, and D. Godall, eds., Hot Deserts and arid Shrublands. Elsevier, Amsterdam. Floyd, T. 1996. Top-down impacts on creosotebush herbivores in a spatially and temporally complex environment. Ecology 77: 1544–1555. Fretwell, S. 1977. The regulation of plant communities by the food chain exploiting them. Perspectives in Biology and Medicine 20: 169–185. Groner, E., and Y. Ayal. 2001. The interaction between bird predation and plant cover in determining habitat occupancy of darkling beetles. Oikos 93(1): 22–31. Hairston Jr., N., and N. Hairston Sr. 1993. Cause-eﬀect relationships in energy ﬂow, trophic structure, and interspeciﬁc interactions. American Naturalist 142: 379–411.

High Animal Biodiversity in Low-Productivity Deserts

27

Hairston, N.G. 1997. Does food web complexity eliminate trophic-level dynamics? American Naturalist 149: 1001–1007. Hamilton, W.J. III. 1971. Competition and thermoregulatory behavior of the Namib Desert tenebrionid beetle genus Cardiosis. Ecology 52: 810–822. Holt, R. 1977. Predation, apparent competition, and the structure of prey communities. Theoretical Population Biology 12: 197–229. Humphreys, W. 1979. Production and respiration in animal populations. Journal of Animal Ecology 48: 427–453. Huston, M.A. 1994. Biological Diversity: the Coexistence of Species on Changing Landscapes. Cambridge University Press, Cambridge. Hutchinson, G. 1959. Homage to Santa Rosalia or why are there so many kind of animals. American Naturalist 93: 145–159. Johnson, K., and W. Whitford. 1975. Foraging ecology and relative importance of subterranean termites in Chihuahuan desert ecosystems. Environmental Entomology 4: 66–70. Kotler, B.P. 1997. Patch use by gerbils in a risky environment: Manipulating food and safety to test four models. Oikos 78: 274–282. Krasnov, B., and Y. Ayal. 1995. Seasonal changes in darkling beetle communities (Coleoptera, Tenebrionidae) in the Ramon erosion cirque, Negev Highlands, Israel. Journal of Arid Environments 31: 335–347. Lindemann, R. 1942. The trophic-dynamic aspect of ecology. Ecology 23: 399–418. Louw, G., and M.K. Seely 1982. Ecology of Desert Organisms. Longman, London. Ludwig, J.A. 1986. Primary production variability in desert ecosystems. Pp. 5–17 in W.G. Whitford, ed., Pattern and Processes in Desert Ecosystems. University of New Mexico Press, Albuquerque. MacKay, W. 1991. The role of ants and termites in desert communities. Pp. 113–150 in G. Polis, ed., The Ecology of Desert Communities. University of Arizona, Tucson. MacMahon, J.A. 1981. Introduction. Pp. 263–269 in D. Goodall and R. Perry, eds., Arid-Land Ecosystems: Structure, Functioning and Management, Vol. 2. Cambridge University Press, Cambridge. McNeill, S., and J. Lawton. 1970. Annual production and respiration in animal populations. Nature 225: 472–474. Medvedev, G.S. 1965. Adaptations of leg structure in desert darkling beetles. Entomological Review 44: 473–475. Morton, S., and C. James. 1988. The diversity and abundance of lizards in arid Australia: a new hypothesis. American Naturalist 132: 237–256. Noy-Meir, I. 1973. Desert ecosystems: environment and producers. Annual Review of Ecology and Systematics 4: 25–52. Noy-Meir, I. 1974. Desert ecosystems: higher trophic levels. Annual Review of Ecology and Systematics 5: 195–214. Noy-Meir, I. 1975. Stability of grazing systems: an application of predator-prey graphs. Journal of Ecology 63: 459–483. Noy-Meir, I. 1985. Desert ecosystem structure and function. Pp. 93–103 in M. Evenari, I. Noy-Meir, and D. Goodall, eds., Hot Deserts and Arid Shrublands, A. Elsevier Scientiﬁc Publications, Amsterdam. Oksanen, L. 1992. Evolution of exploitation ecosystems.1. Predation, foraging ecology and population-dynamics in herbivores. Evolutionary Ecology 6(1): 15–33. Oksanen, L., and T. Oksanen. 2000. The logic and realism of the hypothesis of exploitation ecosystems (EEH). American Naturalist 155: 703–723.

28 Living Components of Biodiversity: Organisms

Oksanen, L., S.D. Fretwell, J. Arruda, and P. Niemela. 1981. Exploitation ecosystems in gradients of primary productivity. American Naturalist 118: 240–261. Polis, G.A., ed. 1991a. The Ecology of Desert Communities. University of Arizona Press, Tucson. Polis, G.A. 1991b. Food webs in desert communities: complexity via diversity and omnivory. Pp. 383–437 in G.A. Polis, ed., The Ecology of Desert Communities. University of Arizona Press, Tucson. Polis, G.A. 1991c. Complex trophic interactions in deserts: an empirical critique of food-web theory. American Naturalist 138: 123–155. Polis, G.A. 1991d. Desert communities: An overview of patterns and processes. Pp. 1–26 in G.A. Polis, ed., The Ecology of Desert Communities. University of Arizona Press, Tucson. Polis, G.A., and D.R. Strong. 1996. Food web complexity and community dynamics. American Naturalist 147: 813–846. Polis, G.A., and T. Yamashita. 1991. The ecology and importance of predaceous arthropods in desert communities. Pp. 180–222 in G.A. Polis, ed., The Ecology of Desert Communities. Arizona University Press, Tucson. Reagan, D.P., Camilo, R., and R.B. Waide. 1996. The community food web: major properties and patterns of organization. Pp. 463–510 in D.P. Reagan and R.B. Waide, eds., The Food-Web of a Tropical Rain Forest. University of Chicago Press, Chicago. Ricklefs, R. 1973. Ecology. Chiron, Newton, MA. Robinson, J.V. 1981. The eﬀect of architectural variation in habitat on a spider community: An experimental ﬁeld study. Ecology 62: 73–80. Seely, M.K. 1985. Predation and environment as selective forces in the Namib Desert. Pp. 161–165 in E. Vrba, ed., Species and Speciation. Transvaal Museum Monograph No. 4, Transvaal Museum, Pretoria. Seely, M.K. 1991. Sand dune communities. Pp. 348–382 in G.A. Polis, ed., The Ecology of Desert Communities. University of Arizona Press, Tuscon. Shachak, M. and S. Brand. 1983. The relationship between sit and wait foraging strategy and dispersal in the desert scorpion, Scorpio maurus palmtus. Oecologia 60: 371–377. Shmida, A., M. Evenari, and I. Noy-Meir. 1986. Hot desert ecosystems: an integrated view. Pp. 379–387 in M. Evenari, A. Shmida, and I. Noy-Meir, eds., Hot Deserts and Arid Shrublands. Elsevier Science Publications, Amsterdam. Skutelsky, O. 1996. Predation risk and state-dependent foraging in scorpions: Eﬀects of moonlight on foraging in the scorpion Buthus occitanus. Animal Behaviour 52: 49–57. Tilman, D. 1988. Dynamics and Structure of Plant Communities. Princeton University Press, Princeton, NJ. Turner, F.B. 1970. The ecological eﬃciency of consumer populations. Ecology 51: 741–742. Vishnevetsky, S., and Y. Steinberger. 1997. Bacterial and fungal dynamics and their contribution to microbial biomass in desert soil. Journal of Arid Environments 37: 83–90. Vitt, L. 1991. Desert reptile communities. Pp. 249–277 in G. Polis, ed., The Ecology of Desert Communities. Arizona University Press, Tucson. Waide, R.B., M.R. Willig, C.F. Steiner, G. Mittelbach, L. Gough, S.I. Dodson, G.P. Juday, and R. Parmenter. 1999. The relationship between productivity and species richness. Annual Review of Ecology and Systematics 30: 257–300.

High Animal Biodiversity in Low-Productivity Deserts

29

Whitehouse, M.E.A., and Y. Lubin. 1998. Relative seasonal abundance of ﬁve spider species in the Negev desert: Intraguild interactions and their implications. Israel Journal of Zoology 44: 187–200. Whitford, W.G. 1986. Decomposition and nutrient cycling in deserts. Pp. 93–117 in Whitford, W.G., ed., Pattern and Processes in Desert Ecosystems. University of New Mexico Press, Albuquerque. Whitford, W.G., and F.R. Kay. 1999. Bioperturbation by mammals in deserts: a review. Journal of Arid Environments 41: 203–230. Wiens, J. 1991. The ecology of desert birds. Pp. 278–310 in G.A. Polis, ed., The Ecology of Desert Communities. Arizona University Press, Tucson. Wiens, J.A., R.G. Cates, J.T. Rotenberry, N. Cobb, B. Vanhorne, and R.A. Redak. 1991. Arthropod dynamics on sagebrush (Artemisia tridentata)—Eﬀects of plant chemistry and avian predation. Ecological Monographs 61: 299–321. Wisdom, C.S. 1991. Patterns of heterogeneity in desert herbivorous insect communities. Pp. 151–179 in G.A. Polis, ed., The Ecology of Desert Communities. The University of Arizona Press, Tucson. Wright, D.H., D.J. Currie, and B.A. Maurer. 1993. Energy supply and patterns of species richness on local and regional scales. Pp. 66–74 in R. Ricklefs and D. Schluter, eds., Species Diversity in Ecological Communities. University of Chicago Press, Chicago. Yair, A., and A. Danin. 1980. Spatial variation in vegetation as related to the soil moisture regime over an arid limestone hillside, northern Negev, Israel. Oecologia 47: 83–88. Yodzis, P. 1984. Energy ﬂow and the vertical structure of real ecosystems. Oecologia 65: 86–88. Zaret, T. 1980. Predation and Fresh Water Communities. Yale University Press, New Haven, CT.

3 Biodiversity Along Core–Periphery Clines Salit Kark Sergei Volis Ariel Novoplansky

T

he study of biodiversity has received wide attention in recent decades. Biodiversity has been deﬁned in various ways (Gaston and Spicer, 1998, Purvis and Hector 2000, and chapters in this volume). Discussion regarding its deﬁnitions is dynamic, with shifts between the more traditional emphasis on community structure to emphasis on the higher ecosystem level or the lower population levels (e.g., chapters in this volume, Poiani et al. 2000). One of the deﬁnitions, proposed in the United Nations Convention on Biological Diversity held in Rio de Janeiro (1992) is ‘‘the diversity within species, between species and of ecosystems.’’ The withinspecies component of diversity is further deﬁned as ‘‘the frequency and diversity of diﬀerent genes and/or genomes . . .’’ (IUCN 1993) as estimated by the genetic and morphological diversity within species. While research and conservation eﬀorts in the past century have focused mainly on the community level, they have recently been extended to include the within-species (Hanski 1989) and the ecosystem levels. The component comprising within-species genetic and morphological diversity is increasingly emphasized as an important element of biodiversity (UN Convention 1992). Recent studies suggest that patterns of genetic diversity signiﬁcantly inﬂuence the viability and persistence of local populations (Frankham 1996, Lacy 1997, Riddle 1996, Vrijenhoek et al. 1985). Revealing geographical patterns of genetic diversity is highly relevant to conservation biology and especially to explicit decision-making procedures allowing systematic rather than opportunistic selection of populations and areas for in situ protection (Pressey et al. 1993). Therefore, studying spatial patterns in within-species diversity may be vital in deﬁning and prioritizing conservation eﬀorts (Brooks et al. 1992). 30

Biodiversity Along Core–Periphery Clines 31

Local populations of a species often diﬀer in the ecological conditions experienced by their members (Brown 1984, Gaston 1990, Lawton et al. 1994). These factors potentially aﬀect population characteristics, structure, and within-population genetic and morphological diversity (Brussard 1984, Lawton 1995, Parsons 1991). The spatial location of a population within a species range may be related to its patterns of diversity (Lesica and Allendorf 1995). Thus, detecting within-species diversity patterns across distributional ranges is important for our understanding of ecological and evolutionary (e.g., speciation) processes (Smith et al. 1997), and for the determination of conservation priorities (Kark 1999). This is especially important in the face of recent climatic and environmental changes occurring at global, regional, and local spatial scales (Safriel et al. 1994). Much of the scientiﬁc focus at the community level is aimed at detecting areas especially rich in biological diversity, that is, ‘‘diversity hotspots’’ (Myers 1990, Myers et al. 2000). This approach focuses attention on revealing areas rich in species diversity, endemism, and rare and endangered species (Mittermeier et al. 1998). Similarly, at the within-species level, revealing areas with especially high genetic and morphological diversity, and rare or unique genetic structures may be important in setting research and conservation priorities (Kark 1999, Kark et al. 1999).

Core and Periphery Patterns of diversity within species may be studied along ‘‘clines’’ (Brussard 1984, Carson 1959, Da Cunha and Dobzhansky 1954, Lennon et al. 1997, Lesica and Allendorf 1995, Mayr 1963). Within the distribution range of a single species, the terms ‘‘periphery’’ and ‘‘core’’ are often used to refer to the physical location of a population within the range (Brown et al. 1996). Accordingly, ‘‘peripheral’’ populations are those located at the very edge of the distribution, while ‘‘core’’ populations are those found further away from the range boundaries (Brown et al. 1996, Channell and Lomolino 2000a). This geographical distinction may also have ecological relationships (Brussard 1984). ‘‘Marginal’’ areas are the ecologically least favorable (Brown 1984, Brussard 1984, Gaston 1990, Hengeveld and Haeck 1981, Wiens 1989), least predictable, and least suitable parts of the range (Hengeveld and Haeck 1981), in locations where extinction probabilities are relatively high (Lennon et al. 1997). In contrast, in ‘‘central’’ areas, extinction probabilities are lower and conditions are, over time, more favorable and predictable for the species (Brown 1984, Brussard 1984, Gaston 1990, Lesica and Allendorf 1995). There are speciﬁc cases in which the geographical and ecological areas are not congruent (Brussard 1984, Gaston 1990, Lesica and Allendorf 1995), but in many cases they do coincide (Brussard 1984, Hoﬀmann and Blows 1994, Wiens 1989). Core and peripheral populations are expected to experience diﬀerent biotic and abiotic selection pressures (Andrewartha and Birch 1954,

32 Living Components of Biodiversity: Organisms

Brown et al. 1995, Lawton 1995, Lesica and Allendorf 1995). In many cases, the optimal combination of ecological, environmental, and biotic factors for a species is found within the geographical core of its range (Lesica and Allendorf 1995, Wiens 1989). Population densities generally decrease and ﬂuctuate more along such clines toward the periphery (Brown 1984, Brown et al. 1995, Brussard 1984, Caughley et al. 1988, Collins and Glenn 1991, Gaston 1990, Hengeveld and Haeck 1981, Hoﬀmann and Blows 1994, Lomolino and Channell 1995, Vrijenhoek et al. 1985, Wiens 1989), the range tends to become less continuous (Brown et al. 1996), and populations become more isolated, transient (Lomolino and Channell 1995), and patchily distributed (Boorman and Levitt 1973, Carter and Prince 1988), although exceptions to these trends exist (Blackburn et al. 1999, Lawton 1995, Svensson 1992). Fluctuations in population size and growth rate at the periphery may result in very small population size (Brussard 1984), and when dispersal from neighboring populations is limited, may lead to the extinction of local populations (Harrison 1994, Lennon et al. 1997, Thomas and Hanski 1997).

Diversity in Core Versus Periphery: Classical Hypotheses Three main hypotheses concerning trends in genetic diversity across core– periphery clines are found in the literature, each having diﬀerent spatial implications (reviewed in Safriel et al. 1994).

Increasing Diversity from Periphery to Core— the ‘‘Carson Hypothesis’’ The hypothesis, developed by Carson (1959), argues that genetic diversity will increase from the range periphery toward the core. Carson suggested that core populations are more continuous and dense, undergo balancing selection, and are therefore expected to show higher levels of withinpopulation genetic diversity than peripheral populations that are relatively small, fragmented, and isolated (Carson 1959, Mayr 1963). This theory implicitly refers to a neutral model of gene diversity, large core populations maintain higher genetic diversity because they harbor more mutations and because genetic drift is less eﬀective in them compared with small isolated peripheral populations (Lesica and Allendorf 1995). Yet this prediction may also be explained based on selective considerations. Accordingly, diversity of adaptive traits at the periphery is predicted to be lower if only a few genotypes can cope with its extreme conditions (Hoﬀmann and Parsons 1991). This hypothesis is supported by classical papers (e.g., Da Cunha and Dobzhansky 1954) and by more recent studies (Hoﬀmann and Parsons 1991, Parsons 1991, Vrijenhoek et al. 1985, reviewed in Lesica and Allendorf 1995, see table 3.1).

Table 3.1 A partial summary of studies testing trends in within-population diversity in core vs. peripheral populations. Among the papers published between 1978 and 1998 that deal with diversity in peripheral vs. core populations, some show evidence for higher diversity in core populations, supporting the Carson hypothesis, while others show evidence for the opposite trend, supporting the Fisher hypothesis. Several papers show no consistent trends or signiﬁcant differences between core and peripheral populations. Only papers where authors refer to core (or central) and peripheral (or marginal) populations, rather than only one of these, were included. Incomplete data were not ﬁlled in

33

Study Area

Species Studied (Common Name)

Europe

Quercus petrea

Central and western Mediterranean basin; French Atlantic coast region

Quercus ilex (holm oak)

Central and northern Japan—Hokkaido and Honshu Islands

Pinus pumila (stone pine)

Western North America

Number of Populations Studied

Type of Diversity Measured

81

Genetic (allozyme)

Hypothesis Supported

Comments

Source Reference

Inconsistent

Higher heterozygosity in core than in peripheral populations, but higher number of alleles per locus in the periphery than in the core

Zanetto and Kremer (1994)

Genetic (allozyme)

Carson

Six main disjunct regions of the range

Michaud et al. (1995)

18

Genetic (allozyme)

Carson Fisher for some loci

15 enzyme systems, generally lower in peripheral populations (in some alleles higher in peripheral populations)

Tani et al. (1996)

Bromus tectorum (cheatgrass)

6

Genetic (quantitative)

Inconsistent mixed

Introduced species

Rice and Mack (1991)

California

Avena barbata

97

Genetic (allozyme)

Humpshaped

Genetic diversity in polymorphic populations was positively related to microhabitat heterogeneity (spatiotemporal) which had a bell shape. 35 loci

Allard et al. (1978)

Israel

Avena barbata

31

Genetic (allozyme)

Inconsistent

35 loci

Allard et al. (1978) continued

Table 3.1 continued

Study Area

Species Studied (Common Name)

Number of Populations Studied

Type of Diversity Measured

Canada—USA

Carex lasiocarpa

39

Canada—USA

Carex pellita

20

Utah

34

Hypothesis Supported

Comments

Source Reference

Genetic (allozyme)

Fisher

12 loci

McClintock and Waterway (1993)

Genetic (allozyme)

Inconsistent

12 loci

McClintock and Waterway (1993)

Hordeum jubatum

Genetic (allozyme)

Carson

18 loci

Shumaker and Babble (1980)

Turkmenistan and Israel

Hordeum spontaneum (wild barley)

Morphological

Fisher

Higher in peripheral populations in most traits (14 of 18)

Volis et al. (1998)

Israel

Hordeum spontaneum (wild barley)

Genetic (allozyme)

Fisher

Utah

Pseudotsua menziesii

Genetic (allozyme)

Carson

20 loci

Schnabel et al. (1993)

New Zealand

Leptosepermum scoparium (Myrtaceae)

Morphological

Inconsistent mixed

Inconsistent for diﬀerent Traits studied

Wilson et al. (1991)

Australian coast

Drosophila serrata

Genetic (quantitative)

Carson

Blows and Hoﬀmann (1993)

Kenya, Morroco, Italy, Reunion, Australia

Ceratitis capitata (medﬂy)

Genetic (allozyme and DNA)

Carson

Baruﬃ et al. (1995)

17

Nevo and Beiles (1988)

Study Area

Species Studied (Common Name)

Number of Populations Studied

Hypothesis Supported

Comments

Source Reference

23 loci; lower in periphery in allozyme analysis, nonregular yet ‘‘slightly impaired’’ in periphery in morphological traits

Descimon and Napolitano (1993)

35

Morphological and genetic (allozyme)

Carson for genetic and inconsistent for phenotypic

Theba pisana (Helicidae)

Genetic (allozyme)

Fisher

Nevo and Beiles (1988)

Israel

Hyla arborea (Hylidae)

Genetic (allozyme)

Fisher

Nevo and Beiles (1988)

Israel

Bufo viridis (Buﬁnidae)

Genetic (allozyme)

Fisher

Nevo and Beiles (1988)

Israel

Gryllotalpa gryllotalpa (Gryllotalpidae, mole cricket)

Genetic (allozyme)

Humpshaped

Israel

Acomys cahirinus (Muridae)

Genetic (allozyme)

Fisher

Nevo and Beiles (1988)

Israel

Agama stellio (Agamidae)

Genetic (allozyme)

Fisher

Nevo and Beiles (1988)

Arizona vs. Sonora, Mexico

Poeciliopsis occidentalis

Genetic (allozyme)

Carson

Southeastern France

Parnassius mnemosyne (Papilionidae)

Israel

24

Type of Diversity Measured

21 (16 Sonoran)

Diversity ﬁrst increases but declines toward the range extreme periphery

Heterozygosity

Nevo and Beiles (1989)

Vrijenhoek et al. (1985)

36 Living Components of Biodiversity: Organisms

Decreasing Diversity from Periphery to Core— the ‘‘Fisher Hypothesis’’ This hypothesis, resulting from Fisher’s (1930) work, predicts that genetic diversity should decrease from the periphery toward the core of the species’ range. Accordingly, peripheral populations will sustain higher levels of genetic diversity due to ﬂuctuating selection in spatially heterogeneous and unpredictable environments, while core populations will experience stabilizing selection, which reduces genetic diversity (Burger 1988, Fisher 1930, Hoﬀmann and Parsons 1991, Parsons 1989). This theory implicitly refers to adaptive considerations and mainly to the type and strength of the pressures of natural selection. This theory, too, is supported by empirical evidence from a wide array of species and study systems (Hoﬀmann and Parsons 1991, Nevo and Beiles 1988, Parsons 1991, table 3.1). Homogenous Diversity from Periphery to Core— the ‘‘Mayr Hypothesis’’ Mayr suggested that in some cases gene ﬂow from the core may compensate for the eﬀects of local selection and genetic drift at the periphery. In such cases genetic diversity may actually be homogenous throughout the species’ range (Mayr 1963, 1970, table 3.1).

Early Studies Early work done in the 1950s that compared patterns of genetic diversity across the distribution range yielded contradictory results. One of the earliest studies was performed by Da Cunha and Dobzhansky (1954), who compared chromosomal polymorphism in core and peripheral populations of Drosophila. Their hypothesis was that the amount of adaptive polymorphism carried in a population is correlated with the variety of the ecological niches its members exploit. They found that core Drosophila willistoni populations were highly polymorphic relative to those at the periphery, where the species was less common and less ubiquitous than its competitors. They interpreted this result in a ‘‘Carsonian’’ fashion, because their central populations were both richer and more diverse (Da Cunha and Dobzhansky 1954). However, White (1951) found no diminution of chromosomal variability toward the distribution periphery, and other studies found an increase in genetic diversity at the periphery of the range (summarized by Hoﬀmann and Parsons 1991). Although the topic continues to draw attention, no clear pattern has emerged (e.g., Brussard 1984, Parsons 1991). Work focusing on a wide range of animal and plant species has tested these hypotheses using various phenotypic and molecular genetic estimates. Overall, each one of the above-mentioned hypotheses has gained considerable support by empirical evidence (table 3.1). In addition, there are cases in which no obvious spatial trends appear

Biodiversity Along Core–Periphery Clines 37

(Brussard 1984). Reviewing the case of protein electrophoretic diversity, Parsons (1989, p.43) notes that: ‘‘variability levels in central vs. marginal populations have revealed a rather confused situation. For an endangered ﬁsh, Poeciliopsis occidentalis, in Arizona, geographically peripheral populations show less electrophoretic variation than do central populations. In contrast, some Drosophila populations show higher electrophoretic variability at the margins . . . . Hence, comparisons of electrophoretic variability under differing ecological circumstances must be approached with extreme caution.’’

Discrepancies Between the Hypotheses Although some of the discrepancies found among the studies may reﬂect real diﬀerences between the study systems and diﬀerent species, it appears that methodological and conceptual factors may have also contributed to the confusion: 1. Spatial Deﬁnitions of Core and Periphery: Studies of core versus periphery often compare two main distribution areas, one representing periphery and the other, core. Populations often are sampled from two extremes rather than along a continuum. The deﬁnition and logic behind the selection of these two areas very often differ between studies (Antonovics et al. 1994). In some cases, peripheral populations are sampled from the edge of the species continuous distribution range, where population density declines rapidly (Lennon et al. 1997). Yet this region may not represent the very edge of the range. Additional small population patches may occur beyond this region. Alternatively, peripheral populations may be sampled at the very extreme periphery of the range, representing small and isolated fragmented populations (Antonovics et al. 1994). In these cases, populations from intermediate areas of the range, located between the extreme periphery and the core, are not included in the study. These differences in sampling may easily lead to contradictory conclusions, because different sections of the distribution range are compared. While the sampling of the periphery is often inconsistent across studies, the deﬁnition of the term ‘‘core’’ is often even less clear. Thus, for example, some studies refer to the point that is geographically farthest away from all range peripheries, while others divide the range into two equal-sized areas, the one representing ‘‘core’’ and the other ‘‘periphery’’ (e.g., Lomolino and Channell 1995, Channell and Lomolino 2000b). The geometric shape of the range and the patterns of patchiness within it will largely determine which areas will be considered core versus peripheral. Other studies focus on the population density, where high-density populations are considered as central populations, yet these may not always be located at the geographical core. 2. Distance Between the Compared Populations: In some cases the two distribution areas representing core and periphery are compared in populations that are geographically very distant from each other (Brussard 1984),

38 Living Components of Biodiversity: Organisms

sometimes from different continents (see, e.g., comparison of marginal vs. central populations of birds in Møller 1995). Populations from very distant areas may experience different evolutionary and recent histories, causing distinct patterns of genetic diversity within populations. 3. Anthropomorphic Inﬂuences: Differences may arise due to sampling of areas with differing levels of human-related disturbance. Because sampling of the extreme edge of the range may be very difﬁcult, populations studied in these areas are often sampled in nonnatural-resource-rich (e.g., agricultural) and human-impacted areas, especially when the periphery occurs in the desert. Populations in these areas may actually have different patterns of genetic diversity relative to the more natural surrounding environment, where population density is lower and sampling becomes more difﬁcult (see discussion in Kark 1999, Kark et al. 1999). 4. Diversity Estimates Used: Genetic diversity within a population may be affected by gene ﬂow, population dynamics, and random processes, such as genetic drift (Slatkin 1994, Wade and Goodnight 1998, Wright 1932) and natural selection. The interaction between the neutral factors and the type and levels of selection pressure will largely determine the levels of genetic diversity in populations. While Carson’s theory implicitly refers to a neutral model of gene diversity, the Fisher hypothesis of hypervariable marginal populations due to ﬂuctuating environments refers to selected markers and to more complicated genetic models that involve natural selection. Therefore, different trends in genetic diversity may be obtained by studying traits controlled by ‘‘neutral’’ versus ‘‘naturally selected’’ genes (Futuyma, 1997) or when comparing different loci. This may lead to contradictory results in different studies, because different genes and alleles may be subjected to varying pressures of natural selection (Randi et al. in review). Different studies have used diverse genetic and molecular methods to reveal genetic diversity (e.g., chromosome inversions vs. microsatellites) that may be affected to a different extent by random versus selective processes (table 3.1). However, even within a single class of genetic markers, such as allozymes, contradictory results are often revealed between studies testing trends across species ranges (e.g., Brussard 1984, Hoffmann and Parsons 1991). 5. Different Scales Are Studied: This could contribute at both the sampling and analysis steps to differences between studies focusing on unequal scales, or when species of different size and dispersal ability are compared, as discussed below.

A New Integrating Hypothesis Kark has recently proposed a new hypothesis, predicting a hump-shaped unimodal pattern of diversity across the range, with peak genetic diversity levels in intermediate populations, located between the range periphery and the core (Kark 1999). It suggests that part of the discrepancy between empiri-

Biodiversity Along Core–Periphery Clines 39

cal ﬁndings appearing in the literature may be due to partial sampling of core–periphery gradients, representing the increase or decrease phase only. Maximum diversity is predicted to occur in the edge of the species continuous distribution, often congruent with areas of ecological transition (i.e., ecotones). As suggested by Kark (1999), in order to test the proposed hypotheses one would need to: (a) identify an area where steep environmental changes occur across short geographical distances and (b) select species that are distributed along the gradient that include more continuous populations, populations at the edge of their more continuous range, and small and isolated populations at the extreme periphery of their range. In this chapter, we aim to review new studies testing the above hypotheses across a steep climatic gradient in Israel, along which many species reach the edge of their distribution range (YomTov and Tchernov 1988). In the studies presented below, we deﬁne the periphery and core of the range based on population densities. Thus, the area beyond which the population density declines to zero will be considered the range periphery. The area where the species reaches the edge of its continuous distribution will be called the ‘‘turnover zone’’ (see below). The area where population densities are high and distribution is relatively continuous will be considered the core of the distribution range. We argue that the periphery can be more clearly and absolutely deﬁned, and thus delimited on a spatial basis, compared with the core. Therefore, we recommend, when possible, to refer to the distance from the range periphery, rather than to an arbitrary core– periphery dichotomy.

Turnover Zones and Ecotones Areas of ecological transition, that is, ecotones and environmental gradients, have recently received scientiﬁc attention due to their potential importance in processes generating biodiversity (Schneider et al. 1999, Smith et al. 1997) and as potential biodiversity hotspots (Kark et al. 1999). Many species reach the limits of their continuous distributions in these areas of steep ecological transition between ecosystems (Danin 1998, Endler 1982). From here toward the most extreme periphery of the range, populations become very small and isolated, and eventually fade out, marking the edge of the species range (Kark 1999). Thus ‘‘turnover zones’’ of various species often are predicted to occur at areas of ecological transition, leading to congruence of the ecotone and the turnover zone (Kark 1999). In these areas, the environment ﬂuctuates temporally and spatially between more favorable and more extreme (Kark et al. 1999). In this chapter, we will use the term ‘‘ecotone’’ for ecosystem-related transition zones (i.e., areas of transition between ecosystems) and the term ‘‘turnover zone’’ for density and distribution-related changes within a single species (i.e., transition between core and peripheral distribution).

40 Living Components of Biodiversity: Organisms

Core–Peripheral Clines in Israel The Ecogeographical Gradient in Israel Israel comprises a narrow land bridge between Europe, Asia, and Africa, with steep climatic and ecological clines across relatively short distances (Bitan and Rubin 1991, Kadmon and Danin 1997, Yom-Tov and Tchernov 1988). The sharp ecological gradient from Mediterranean to desert ecosystems is congruent with distributional margins of many Mediterranean, Irano-Turanian, and Saharo-Arabian organisms (Danin 1998, Yom-Tov and Tchernov 1988). While mean annual rainfall in the Mediterranean Galilee and Golan Heights in the north may exceed 900 mm per year, only 200–300 km to the south the mean annual rainfall in the southern Negev desert is less than 30 mm and is highly variable among years (Bitan and Rubin 1991, ﬁg. 3.1). An especially steep climatic gradient occurs in the northern Negev ecotone, where mean annual rainfall decreases from over 450 mm to less than 150 mm within a narrow belt of less than 60 km (Danin 1998, Kark et al. 1999). It is here that many Mediterranean, steppe, and desert species reach the edge of their continuous distributions (Bitan and Rubin 1991, Kadmon and Danin 1997, Safriel et al. 1994, Yom-Tov and Tchernov 1988). Thus, the ecological cline in this region oﬀers a unique opportunity to compare geographically proximate populations with very diﬀerent population densities that are potentially connected by dispersal and gene ﬂow. This gradient includes populations along a distribution gradient of increasing distance from the very edge of the distribution range. Furthermore, it allows testing of the presented hypotheses, comparing trends in genetic and phenotypic diversity along core–periphery clines. Earlier Studies Considerable work has been done over the years studying patterns of allozyme, DNA, and chromosmal diversity across the Mediterranean–desert gradient in Israel (Nevo and Beiles 1988, 1989). This work included many diﬀerent groups, from insects to mammals and reptiles. The general pattern suggested by Nevo and Beiles is that diversity increases from the mesic toward the more arid environments. We suggest that an addition to this important work should be to relate the ﬁndings to the species’ distribution range patterns and processes. A certain environment that determines population sizes, gene ﬂow, patterns of local selection, and the resulting phenotypic and genetic diversity (such as the Mediterranean environment), may be very diﬀerent in the degree of stress it presents to a mesic versus a desert species. For example, patterns may diﬀer greatly between desert species, for which the Mediterranean region is the periphery of the range, and for Mediterranean species, for which the desert provides the range periphery. The distribution range reﬂects the response of the species to the diverse environmental and biotic conditions. In many cases, the range periphery represents the area

Biodiversity Along Core–Periphery Clines 41

Figure 3.1 Mean annual rainfall in Israel. The map was generated based on data in the GIS laboratory of the Hebrew University of Jerusalem. Note the sharp changes in rainfall across short geographical distances, and especially the Mediterraneandesert ecotone region where rainfall rapidly declines to the south and to the east.

beyond which the species cannot maintain viable populations at a certain point in time. Therefore, reference to the species’ distribution range, in addition to the environmental variables, may be highly important to our understanding of the ecological and evolutionary processes that determine diversity patterns and to conservation of the species. Here, we discuss some recent case studies from Israel that follow diversity patterns with reference to the species’ distribution range. Recent Work Following is a brief review of some of our recent studies from Israel, testing the above hypotheses on diversity across periphery-to-core (nonperiphery) clines within the distribution range. We look at a study of allozyme diversity

42 Living Components of Biodiversity: Organisms

of a phasianid game bird, the chukar partridge (Alectoris chukar), a quantitative genetics study of an annual legume (Trifolium purpureum) and a perennial clonal grass (Dactylis glomerata), and a study of phenotypic and allozyme diversity of an annual grass, the wild progenitor of cultivated barley, Hordeum spontaneum. All these species have high densities in the Mediterranean region of Israel and their populations become smaller and more isolated toward the arid Negev Desert, which comprises their global distributional periphery.

The Chukar Partridge (Alectoris chukar) Chukar Distribution The chukar partridge (Alectoris chukar) generally inhabits the mesic and semiarid areas, and has relatively large and continuous populations in Mediterranean and steppe regions of Israel (Shirihai 1996). Deserts represent the margins of its range, where it occurs in isolated and sparse populations (Shirihai 1996, Liu Naifa pers. comm.). The chukar is a species indigenous to the region. The extreme desert regions of the southern Negev and Sinai comprise the global southwestern border of its distribution. The chukarcontinuous Mediterranean areas in the north and center of Israel are referred to in this work as part of the ‘‘core.’’ The Mediterranean desert ecotone of the northern Negev is the edge of the chukar-continuous distribution. This area comprises the ‘‘turnover zone’’ of the species’ range in Israel, where rapid thinning of chukar populations occurs across short geographic distances (Shirihai 1996). Marginal to the Negev Highlands in the south of Israel and toward the Sinai Desert, chukar density decreases, distribution becomes discontinuous, and local populations become patchy and isolated (Degen et al. 1984, Pinshow et al. 1982, Shirihai 1996). This area comprises the extreme periphery of the chukar range. An additional isolated population, most probably a relict from the late Pleistocene, is found in the mountains of the southern Sinai desert (see discussion in Kark et al. 1999). Chukars do not possess many physiological adaptations to heat stress (Carmi-Winkler et al. 1987, Frumkin 1983, Kam 1986), especially as compared with the partly sympatric sand partridge (Ammoperdix heyi), which is well adapted to the desert (Carmi-Winkler et al. 1987, Degen et al. 1984, Pinshow et al. 1982). A main limiting factor in the desert is the chukars’ ability to forage long enough to obtain their energy requirements without risking their heat balance. In arid hot environments, extremely high temperatures limit their foraging activity to approximately one hour a day, which cannot suﬃce for their energy and water demands (Carmi-Winkler et al. 1987). Therefore, in extremely arid regions the species usually occurs in small resource-rich patches (Carmi-Winkler et al. 1987, Pinshow et al. 1983, Shirihai 1996). These habitat patches must be rich enough to meet the birds’ energetic needs in the short available foraging time, limited to

Biodiversity Along Core–Periphery Clines 43

the early morning hours (Carmi-Winkler et al. 1987, Degen et al. 1983, Pinshow et al. 1983). The patches must also provide suﬃcient water during the dry months when chukars need to drink water on a regular daily basis (Carmi-Winkler et al. 1987, Degen et al. 1983, Pinshow et al. 1983). As far as is known, chukars do not exhibit long-distance spatial or altitudinal migrations (Alkon 1974, Paz 1987), and available information from marked chukars in Israel suggests that movement of individual birds is usually limited to an area ranging several square kilometers in both northern and southern populations (Alkon 1974, P. Alkon unpublished data).

Trends in Diversity Across the Range Trends in within-population allozyme diversity were studied in the chukar partridge in Israel and were compared with three years of study (1990, 1993, 1995). Five chukar populations were sampled along the gradient in each of two years, 1990 and 1993. Three populations were sampled in both years in order to enable a comparison among years and to test robustness and variability of the trends along short-term time scales. Allozyme diversity in 32 allozyme loci was determined for birds collected in each population (see Kark 1999 for details). Trends revealed were very similar for the two years of study. Genetic diversity, as estimated by the percentage of polymorphic loci, mean number of alleles, and observed and expected heterozygosity increased from the core to the ecotone (Kark et al. 1999). Single and multiloci Hardy–Weinberg and linkage disequilibria increased signiﬁcantly from populations in the Mediterranean region to those at the ecotone, despite the close geographical proximity between populations in these two regions. As predicted, peak diversity was found in the intermediate ecotone area, located between the extreme periphery of the range and the core (Kark et al. 1999). Local populations maintained distinct genetic characteristics even though population genetic data indicated the likelihood of substantial gene ﬂow among populations, supporting the ‘‘divergence with gene ﬂow’’ model of speciation (Rice and Hostert 1993, Smith et al. 1997, see Kark et al. 1999 for discussion). Study populations showed instances of isolation by distance eﬀects in the face of both short distances among populations and the absence of signiﬁcant geographical barriers (Kark et al. 1999). The highly diverse genetic structure of diﬀerent chukar populations across the short geographical gradient given relatively high levels of gene ﬂow is especially interesting. It could be maintained by a combination of stochastic population dynamics factors and natural selection acting on morphophysiological traits and on their linked allozyme loci (Kark et al. 1999). Increased intrapopulation genetic diversity in the ecotone region resulted from the addition of new alleles, not present in less variable populations, and increasing frequency of rare alleles at polymorphic loci. These results suggest a clear trend across the part of the range studied, yet it

44 Living Components of Biodiversity: Organisms

should be emphasized that this sampling did not include the extreme periphery of the range. Following these ﬁndings, 13 chukar populations were sampled in 1995. These included the part of the range previously studied, but also added three populations from the extreme periphery of the range in the central and southern Negev, as described in detail in Kark (1999). Diversity across the more continuous distribution range shows a unimodal trend of genetic diversity from core to periphery. Peak diversity was found in all four populations sampled in the ecotone region of Israel; it decreased toward both the Mediterranean core, as in previous years, and toward the arid range periphery, showing a hump-shaped pattern across the range. Similar patterns were seen at the phenotypic level (Kark et al. 2002). These ﬁndings support the hypothesis presented by Kark (1999), suggesting that diversity will show a hump-shaped unimodal pattern across the distribution range, with peak levels in intermediate (subperipheral) populations located in the turnover zone region.

The Purple Clover (Trifolium purpureum) and the Common Dactyl (Dactylis glomerata) Distribution The purple clover, Trifolium purpureum, is distributed throughout the Mediterranean basin. The common dactyl, Dactylis glomerata, is Mediterranean, Irano-Turanian, and Euro-Siberian. Both species are found throughout the Mediterranean regions of northern and central Israel, and reach the edge of their global range in the ecotone region in the Northern Negev. Unlike the chukar partridge, these plant species do not extend toward the more extreme arid Negev desert. The two species exhibit a patchy distribution in all regions of Israel, including the more mesic northern regions with high local variation. In some cases local density is very high and they are the dominant species (Danino 2000). Due to the limitations of the study design, populations with very low local densities were not included in the sampling. Trends in Diversity Across the Range A greenhouse experiment was conducted with Trifolium purpureum (an annual legume) and Dactylis glomerata (a perennial grass). Details are given in Danino (2000). Parent plants were collected in the ﬁeld at three core populations (Mediterranean, ca. 600 mm mean annual rainfall), and four peripheral (from the ecotone region with ca. 300 mm mean annual rainfall) populations in Israel. Plants (whole clones in Dactylis and individual plants in Trifolium) were collected from an area of circa 1 ha, using a random spatial design. Only individuals above a certain size that were suitable for the

Biodiversity Along Core–Periphery Clines 45

experimental design were sampled in all locations. Therefore, the sampled plants did not represent the entire size range in the population. This limitation was especially important for Dactylis glomerata in which many clones were excluded due to their small size. Sampling aimed to include plants of homogeneous size. Plants (cuttings for Dactylis and seedlings for Trifolium) were grown in pots in a greenhouse at the Sede Boqer Campus, located in the central Negev desert of Israel. The oﬀspring of each parental plant (i.e., ‘‘family’’) were grown under high or low water availability (Danino 2000) for one season (November 1994–April 1995). In all, 73 Dactylis and 90 Trifolium plant families were studied, with at least three replicates from each family in both low and high water treatments. Oﬀspring were measured for the following traits: total shoot biomass; branch or tiller biomass; number of branches or tillers; mean individual branch or tiller biomass; inﬂorescence biomass; number of inﬂorescence; mean individual biomass; mean inﬂorescence length; proportion of aborted inﬂorescence; reproductive eﬀort (inﬂorescence biomass/total shoot biomass). A factorial layout with the main factors being water treatment and population location across the range was used for region (core vs. periphery), population-within-region (three in core, four in periphery), or population alone (seven populations). A nested ANOVA was used on the pooled data to test the eﬀects of the treatment on each of the dependent variables (trait measurements) depicting various growth and allocation characteristics (Danino 2000). A signiﬁcant diﬀerence was found between regions (core and periphery) for inﬂorescence biomass and for number and biomass of tillers in Dactylis and for both total and shoot biomass in Trifolium, with a general tendency toward fewer and bigger inﬂorescence or shoots in the core and more but smaller inﬂorescence or shoots in the ecotone periphery (Danino 2000). The eﬀect of the individual population was highly signiﬁcant for the majority of traits in both species and under both water treatments when calculated across all the populations. In most cases, populations also diﬀered signiﬁcantly within region. Region-by-water interaction was highly signiﬁcant for percent aborted inﬂorescence and for the reproductive eﬀort and marginally signiﬁcant for total biomass and for the number of branches in Trifolium, yet none of these was signiﬁcant in Dactylis. When calculated from the pooled data, population-by-water interactions were signiﬁcant for the vast majority of traits studied in both species. While signiﬁcant diﬀerences were found between the estimated genetic variances of the individual populations, no generalization could be made for such diﬀerences between the core and the periphery. It may be that in species with highly patchy distributions the diﬀerences between core and periphery populations are less than those between diﬀerent populations within the same geographical region. We conclude that the studied populations can respond to natural selective pressures imposed by a reduction in water supply comparable to the conditions created by the low water treatment to potted plants. Furthermore, the large diﬀerences found among the sampled populations imply that each and

46 Living Components of Biodiversity: Organisms

every one of the populations studied may represent a unique collection of genetic backgrounds. This would be interesting to study further. A possible interpretation of these results is that both the Mediterranean and the semiarid populations of Trifulium purpureum and Dactylis glomerata in Israel represent the ‘‘peripheries’’ of their distributions. If this is so, the more favorable regions of these and other species with similar ranges would be found in the temperate regions of Europe or, in the case of Irano-Turanian species, the highlands of central Asia. As noted earlier, populations with very low local densities could not be included in the study due to the design, which required certain sample sizes. These may actually be representing the very extreme environmental patches and may show diﬀerent patterns of phenotypic diversity.

Wild Barley (Hordeum spontaneum) Distribution Wild barley, Hordeum spontaneum, the ancestor of the cultivated barley, is widely distributed across the eastern Mediterranean basin, and western and central Asia to western China, Pakistan, and India. It is one of the main annual components of open park-forests of Quercus ithaburensis in central and northern Israel. It is also abundant in the Mediterranean grasslands of northern Israel, Jordan, southern Syria, and Lebanon (Harlan and Zohary, 1966). The species is seldom found in regions where mean annual rainfall is less than 150 mm, where mean winter temperatures go below 58C, or at altitudes over 1500 m (Harlan and Zohary 1966, Harlan 1968). The species’ range periphery is found in loess or sand deserts. In these environments, low and unpredictable rainfall are the major limiting factors for this species. Plants in these peripheral populations occupy only accumulating runoﬀ wadis or ravines. Population size declines toward the geographical range periphery, but local population density within the wadis is often as high as at the species core (Gutterman 1992, Volis, pers. obs.). Therefore, the scale at which population density is studied and is estimated may largely determine the spatial diversity trends detected in this species. Distinction between wateraccumulating and non-water-accumulating habitats, therefore, is useful, and is the most reliable criterion for deﬁning the species’ range periphery at the local scale (Volis et al. 2001). Barley densities increase and distribution becomes less patchy in the more mesic Mediterranean areas, toward the core of its range. In this region populations are large and are rarely isolated. In the Negev and Judean Deserts, wild barley density decreases, distribution becomes discontinuous and coincides with wadi distribution alone. Local populations become highly patchy and isolated from each other. These desert regions of Israel, similar to the chukar partridge, constitute the southwestern border of species global distribution range.

Biodiversity Along Core–Periphery Clines 47

Trends in Diversity Across the Range Genetic diversity in wild barley was studied at two levels: quantitative (phenotypic) traits and allozymes.

Phenotypic Diversity Ten populations of wild barley were sampled in 1993 in Israel. Of these, ﬁve populations, representing the core, were sampled in central Israel in the mesic Shefela and Judea Hills (mean annual rainfall ranged between 400 and 600 mm). Five peripheral populations were sampled from the Negev and Judean Deserts at the species range periphery (mean annual rainfall between 70 and 160 mm). The ecotone between the Mediterranean and desert ecosystems, which is the turnover zone of this species distribution, was not sampled in this study. Seeds were sown in 1994 and three-week-old plantlets were transplanted into an experimental ﬁeld of the Institutes for Applied Research in Beer Sheva. Located in the northern Negev ecotone region between the two sampled regions, Beer Sheva receives 250 mm of annual rainfall. The morphological traits included: culm length, ﬂag and penultimate leaf length, spike length, awn length, number of nodes, internode length and total tiller height, number of spikelets per spike, and the average spikelet and seed weight. In addition, two phenological traits were measured: (i) the number of days to awn appearance, indicating the onset of the reproductive phase and (ii) the number of days to anthesis. Analysis of phenotypic variation included a nested ANOVA and MANOVA approach with regions as a ﬁxed eﬀect and populations nested within regions as a random eﬀect (for details see Volis et al. 2002). The degree of variation showed opposite patterns for diﬀerent traits in plants of Mediterranean and desert origin. Our explanation of the ﬁndings employs a combined eﬀect of directional and diversifying selection, possibly as a result of temporal heterogeneity. Israeli peripheral populations inhabiting unpredictable desert environments with respect to water availability apparently represent a ‘‘spreading risk’’ strategy with alternative phenotypes present in a population (Ellner 1985, 1987, Kaplan and Cooper 1984). The traits associated with this strategy are those that enhance temporal variation in the germination of the seeds and in the maturation of the plants, including the onset of germination, seed dormancy, and start of reproduction. Days to awn appearance and anthesis (indicating transition to reproductive stage) and ﬂag and penultimate leaf length (determining grain ﬁlling and thus duration of reproductive stage until seed maturation) were more variable in peripheral, than in core populations. Large diﬀerences were found between seed dormancy of Israeli peripheral and core populations (Volis et al. 2004). The core populations were more variable than peripheral populations in most of the other traits. This may be explained by the fact that variability in traits that are not directly related to ﬁtness has a lower ‘‘cost’’ and therefore, variability may be maintained under more favorable conditions.

48 Living Components of Biodiversity: Organisms

Genetic Diversity The same 10 populations analyzed for phenotypic diversity and two additional populations were used to study genetic variability, as estimated using starch-gel electrophoresis of nine water-soluble leaf proteins (Volis et al. 2002). At the ﬁrst stage, populations were pooled into two groups, representing core and periphery. Following this approach, no signiﬁcant diﬀerences were found between the core and the periphery in any of the three estimates of genetic diversity (mean number of alleles per locus, proportion of polymorphic loci, and expected heterozygosity). Following the hump-shaped pattern hypothesis presented by Kark (1999), trends were retested using a continuous approach to the range (rather than the previous more traditional dichotomous approach, in which populations were pooled into core and periphery and the means for each of the regions were compared). A quadratic regression between allozyme diversity and mean annual rainfall was performed. A noticeable trend, although nonsigniﬁcant, appeared (ﬁg. 3.2). Diversity within populations showed a hump-shaped pattern across the rainfall gradient. As in the case of the chukar partridge, a polynomial quadratic regression explained the ﬁndings much better than a linear model. Diversity in all three diversity estimates increased toward the Mediterranean desert ecotone (which was not sampled), where the edge of the species’ continuous distribution edge occurs. These results are in accordance with the hypothesis predicting a hump-shaped curve of within-population diversity across core– periphery clines (Kark 1999).

Figure 3.2 Allozyme diversity in 12 wild barley populations in Israel, estimated by percentage of polymorphic loci (P) and by mean expected heterozygosity based on Hardy–Weinberg equilibrium (He), as a function of mean annual rainfall. Rainfall is strongly correlated with the distance of the population from the species’ range periphery.

Biodiversity Along Core–Periphery Clines 49

Synthesis of Case Studies The three studies presented above tested patterns of within-population diversity across the distribution range of four species with rather similar ranges in Israel, ranging from a phasianid bird to annual and perennial plants. The studies examined diﬀerent measures of genetic and morphological diversity and had diﬀerent experimental and sampling designs. Unlike previous work on aridity gradients in Israel, our main goal in this review is not to compare the ﬁndings of the studies but rather to emphasize examples of the approach taken in these case studies and their implications to the ﬁndings. It is interesting that the two studies that focused on the chukar partridge and the wild barley showed rather similar patterns. When only part of the species’ range in Israel was studied, or when populations were pooled for analysis, rather than compared across a continuum, very diﬀerent patterns appeared, compared with an analysis where patterns were tested across a more continuous distribution gradient extending from the very edge (periphery) toward the interior of the species distribution (core). In addition, as emphasized by the chukar work, sampling of the intermediate ecotone region, located between the core and the periphery of these species’ ranges revealed a hump-shaped pattern that would have otherwise been overlooked. In addition, when the extreme periphery of the chukar range was not sampled, a partial pattern appeared. The barley work at its ﬁrst stage did not detect any diﬀerences between the core and the peripheral populations in their levels of diversity. Yet when analyzed in more detail across a rainfall gradient, representing a distributional cline for this species, a pattern seems to emerge, pointing toward a hump-shaped unimodal pattern of diversity across the range. This work suggests that the scale on which the study focuses may be crucial in determining the patterns detected. Thus even studies of a single species, which deﬁne and sample populations at diﬀerent spatial scales, may lead to diﬀerent conclusions. A study on patchily distributed plants from peripheral habitats which includes ten isolated subpopulations, for example, could lead to a diﬀerent ﬁnding compared with work in which each subpopulation represents a study population. Because diﬀerent organisms diﬀer in the graininess in which they respond to the environment, diﬀerences between studies may also arise when species of diﬀerent body sizes and life histories are studied at one spatial scale. Additionally, populations are often deﬁned based on area size, and therefore the size of the area sampled for each species may largely determine the patterns found. For example, an area of 10 km2 for a medium-sized bird such as the chukar, and a plant species can represent very diﬀerent patterns of patchiness at the local scale, especially closer to the range periphery, where distribution often becomes less continuous. For example, while chukar distribution becomes more patchy toward the range periphery, distribution of the purple clover is very patchy in both the Mediterranean and more arid periphery. Thus it is important to carefully determine the spatial sampling design, based on the study goals. It is also useful to give these details in the methodology

50 Living Components of Biodiversity: Organisms

chapter of the scientiﬁc reports of the work, to enable comparisons between studies. As mentioned earlier, this chapter focuses on patterns in diversity across distribution clines, rather than on environmental clines alone. We believe that the distribution range and its spatial patterns summarize the way that individuals of diﬀerent populations respond to the environments in which they are found. Climatic variables often used to correlate with diversity patterns, such as mean annual temperature, may not reﬂect the environment that an individual really perceives in its microhabitat. The distribution range and population density are the bottom-line summary of the way the environment (both biotic and abiotic) inﬂuences species, and therefore may be a more useful tool, especially in studies that aim to derive recommendations for conservation.

Conservation Implications As discussed above, spatial scale may largely determine the patterns of diversity found in populations across a species’ distribution range. Therefore, the goals of conservation programs should be clearly set, and these goals should be kept in mind when setting conservation priorities. We suggest that mapping patterns of biological diversity across species distribution ranges, and especially across ecological gradients, is desired for maintaining high levels of genetic diversity. Conserving the areas in which the processes that are generating this diversity are occurring may be an important ﬁrst step. Ecological gradients and areas of ecological transition may be good candidates. We suggest that more attention must be paid by conservation biologists toward areas of transition between climatic regions and among ecosystems (such as Mediterranean and desert in the case of Israel). In these areas, many species attain the edge of their continuous distributions. Currently, very little conservation attention is directed toward these areas. For example, the Global 200 program (Olson and Dinerstein 1998) does not emphasize areas of transition or ecological gradients. If these areas are shown to harbor high levels of genetic and morphological diversity in additional species, investing more eﬀort in conserving them may prove to be a cost-eﬀective conservation strategy.

Emphasis on Diversity in Drylands This chapter has focused on the study of diversity across species distribution (core–periphery) clines. In many cases these gradients are congruent with climatic (e.g., increasing aridity) gradients. While various species change their distribution pattern across these climatic gradients, species diﬀer in their distribution patterns. For example, many species reach the edge of their distribution range in the ecotones of Israel. For some species the

Biodiversity Along Core–Periphery Clines 51

Mediterranean desert ecotone of the northern Negev comprises a southern limit to their Palearctic distribution, while for other species this same area is a northern edge to their Saharan distribution or the western margins of their Irano-Turanian distribution (Danin 1998). Peripheral, core and ‘‘turnover zone’’ populations may thus be located in diﬀerent ecological regions. The case studies presented in this chapter were all species with a southern distribution margin in the Negev desert. It would be interesting to continue this work and compare trends in diversity across species with opposing distributions, for which the desert is the core of the range and the Mediterranean region is the periphery. We predict that diversity patterns will correspond to the location across the range (i.e., periphery–core cline) rather than the climatic changes alone (e.g., the degree of aridity). General patterns in diversity, if they exist, are expected to be found across distribution gradients rather than climatic gradients alone. For many Saharan and Arabian species the arid region may actually prove to be a more favorable region of the range where density, diversity, and persistence are higher.

Acknowledgments Our thanks go to Philip Alkon, Avigdor Cahaner, Ayelet Danino, Samuel Mendlinger, Imanuel Noy-Meir, Ettore Randi, and Uriel Safriel for their invaluable collaboration and discussion in various parts of this work, and the Israel Nature and Parks Authority scientists and rangers for their assistance in ﬁeld work. Support for this research was granted to S.K. from the Pontremoli and the Rieger Research Funds through Keren Kayemet LeIsrael (JNF), The Ecology Fund founded by the JNF, The Blaustein International Center for Desert Studies of the Blaustein Institute for Desert Research, Ben-Gurion University of the Negev, and by grants to A.N. from the Israel Science Foundation founded by The Israel Academy of Sciences and Humanities and the International Arid Land Consortium. This is publication number 356 of the Mitrani Department of Desert Ecology.

References Alkon, P.U. 1974. Social behavior of a population of chukar partridge (Alectoris chukar) in Israel. Ph.D. thesis, Cornell University, Ithaca, New York. Allard, R.W., R.D. Miller, and A.L. Kahler. 1978. The relationship between degree of environmental heterogeneity and genetic popymorphism. Pp. 49–73 in A.H.J. Freysen and J.W. Woldendorp, eds., Structure and Functioning of Plant Populations. North-Holland, New York. Andrewartha, H.G., and L.C. Birch. 1954. The Distribution and Abundance of Animals. University of Chicago Press, Chicago. Antonovics, J., P. Thrall, A. Jarosz, and D. Stratton. 1994. Ecological genetics of metapopulations: The Silene-Ustilago plant-pathogen system. Pp. 146–170 in L. Real, ed., Ecological Genetics. Princeton University Press, Princeton. Baruﬃ, L., G. Damiani, C.R. Guglielmino, C. Bandi, A.R. Malacrida, and G. Gasperi. 1995. Polymorphism within and between populations of Ceratitis capitata: Comparison between RAPD and multilocus enzyme electrophoresis data. Heredity 74: 425–437.

52 Living Components of Biodiversity: Organisms

Bitan, A., and S. Rubin. 1991. Climatic Atlas of Israel for Physical and Environmental Planning and Design. The Israeli Ministry of Transport, Tel Aviv. Blackburn, T.M., K.J. Gaston, R.M. Quinn, and R.D. Gregory. 1999 range. Do local abundances of British birds change with proximity to range edge? Journal of Biogeography 26: 493–505. Blows, M.W., and A.A. Hoﬀmann. 1993. The genetics of central and marginal populations of Drosophila serrata. I. Genetic variation for stress resistance and species borders. Evolution 47: 1255–1270. Boorman, S.A., and P.R. Levitt. 1973. Group selection on the boundary of a stable population. Theoretical Population Biology 4: 85–128. Brooks, D.R., R.L. Mayden, and D.A. McLennan. 1992. Phylogeny and biodiversity: Conserving our evolutionary legacy. Trends in Ecology and Evolution 7: 55–59. Brown, J.H. 1984. On the relationship between abundance and distribution of species. The American Naturalist 124: 255–279. Brown, J.H., D.W. Mehlman, and G.C. Stevens. 1995. Spatial variation in abundance. Ecology 76: 2028–2043. Brown, J.H., G.C. Stevens, and D.M. Kaufman. 1996. The geographic range: size, shape, boundaries, and internal structure. Annual Review of Ecology and Systematics 27: 597–623. Brussard, P.F. 1984. Geographic patterns and environmental gradients: the centralmarginal model in Drosophila revisited. Annual Review of Ecology and Systematics 15: 25–64. Burger, R. 1988. The maintenance of genetic variation: A functional analytic approach to quantitative genetic models. Pp. 63–72 in G. de Jong, ed., Population Genetics and Evolution. Springer-Verlag, Berlin. Carmi-Winkler, N., A.A. Degen, and B. Pinshow. 1987. Seasonal time-energy budgets of free-living chukars in the Negev Desert. The Condor 89: 594–601. Carson, H.L. 1959. Genetic conditions that promote or retard the formation of species. Cold Spring Harbor Symposium on Quantitative Biology 24: 87–103. Carter, R.N., and S.D. Prince. 1988. Distribution limits from a demographic viewpoint. Pp. 165–184 in A.J. Davy, M.J. Hutchings, and A.R. Watkinson, eds., Plant Population Ecology. Blackwell Scientiﬁc Publications, Oxford. Caughley, G., D. Grice, R. Barker, and B. Brown. 1988. The edge of the range. Journal of Animal Ecology 57: 771–785. Channell, R., and M. Lomolino. 2000a. Dynamic biogeography and conservation of endangered species. Nature 403: 84–86. Channell, R., and M. Lomolino. 2000b. Trajectories to extinction: spatial dynamics of the contraction of geographic ranges. Journal of Biogeography 27: 169–179. Collins, S. L., and S.M. Glenn. 1991. Importance of spatial and temporal dynamics in species regional abundance and distribution. Ecology 72: 654–664. Da Cunha, A.B., and T. Dobzhansky. 1954. A further study of chromosomal polymorphism in Drosophila willistoni in its relation to the environment. Evolution 8: 119–134. Danin, A. 1998. Man and the natural environment. Pp. 25–37 in T. E. Levy, ed., The Archaeology of Society in the Holy Land. Leicester University Press, Leicester. Danino, A. 2000. Biological responses to global change—diﬀerences between central and peripheral populations, M.Sc. thesis, Ben Gurion University of the Negev (in Hebrew with English summary).

Biodiversity Along Core–Periphery Clines 53 Degen, A.A., B. Pinshow, and P.U. Alkon. 1983. Summer water turnover weights in free-living chukars (Alectoris chukar) and sand partridges (Ammoperdix heyi) in the Negev desert. The Condor 85: 333–337. Degen, A.A., B. Pinshow, and P.J. Shaw. 1984. Must desert chukars (Alectoris chukar sinaica) drink water? Water inﬂux and body mass changes in response to dietary water content. The Auk 101: 47–52. Descimon, H., and M. Napolitano. 1993. Enzyme polymorphism, wing pattern variability, and geographical isolation in an endangered butterﬂy species Biological Conservation 66: 117–123. Ellner, S. 1985. ESS germination strategies in randomly varying environments. Theoretical Population Biology 28: 50–116. Ellner, S. 1987. Alternative plant life history strategies and coexistence in randomly varying environments. Vegetatio 69: 199–208. Endler, J.A. 1982. Problems in distinguishing historical from ecological factors in biogeography. American Zoologist 22: 441–452. Fisher, R.A. 1930. The Genetical Theory of Natural Selection. Clarendon Press, Oxford. Frankham, R. 1996. Relationship of genetic variation to population size in wildlife. Conservation Biology 10: 1500–1508. Frumkin, R. 1983. Thermoregulation and acid base status in heat stressed desert phasianids: A comparison between the chukar partridge (Alectoris chukar) and the sand partridge (Ammoperdix heyi). Ben-Gurion University of the Negev (in Hebrew, abstract in English). Futuyma, D. 1997. Evolutionary Biology. Sinauer Associates, Sunderland, MA. Gaston, K. J. 1990. Patterns in the geographical ranges of species. Biological Reviews of the Cambridge Philosophical Society 65: 105–129. Gaston, K. J., and J. Spicer. 1998. Biodiversity: an Introduction. Blackwell Science, Oxford. Gutterman, Y. 1992. Ecophysiology of Negev upland annual grasses. Pp. 145–162 in G.P. Chapman, ed., Desertiﬁed Grasslands: Their Biology and Management. Linnean Society Symposium. Series 13, Academic Press, London. Hanski, I. 1989. Metapopulation dynamics: does it help to have more of the same? Trends in Ecology and Evolution 4: 113–114. Harlan, J.R. 1968. On the Origin of Barley. U.S.D.A. Agricultural Handbook 338: 9–13. Harlan, R.J., and D. Zohary. 1966. Distribution of wild wheats and barley. Science 153: 1074–1080. Harrison, S. 1994. Metapopulations and conservation. Pp. 111–128 in P. J. Edwards, R.M. May, and N.R. Webb, eds., Large Scale Ecology and Conservation Biology. Blackwell Scientiﬁc Publications, Oxford. Hengeveld, R., and J. Haeck. 1981. The distribution of abundance. II. Models and implications. Proc. Kon. Ned. Akad. Wetensch., C 84: 257–284. Hoﬀmann, A.A., and M.W. Blows. 1994. Species borders: ecological and evolutionary perspectives. Trends in Ecology and Evolution 9: 223–227. Hoﬀmann, A.A., and P.A. Parsons. 1991. Evolutionary Genetics and Environmental Stress. Oxford University Press, Oxford. IUCN. 1993. The Convention on Biological Diversity: An Explanatory Guide. The IUCN Environmental Law Centre (ELC), Bonn. Kadmon, R., and A. Danin. 1997. Floristic variation in Israel: a GIS analysis. Flora 192: 341–345.

54 Living Components of Biodiversity: Organisms

Kam, M. 1986. Energy budget and seasonal water of free chukar (Alectoris chukar) and sand partridges (Ammoperdix heyi) in the Negev Highlands. Ben Gurion University (in Hebrew, abstract in English). Kaplan, R.H. and W.S. Cooper. 1984. The evolution of developmental plasticity in reproductive characteristics: an application of the adaptive coin-ﬂip principle. The American Naturalist 123: 393–410. Kark, S. 1999. Within-population diversity in the distribution range: Partridges as a research model. Department of Evolution, Systematics and Ecology. The Hebrew University of Jerusalem, and Ecology Research Center. http://shum.huji.ac.il/ salit Kark, S., P.U. Alkon, U.N. Safriel, and E. Randi. 1999. Conservation priorities for the chukar partridge in Israel based on genetic diversity across an ecological gradient. Conservation Biology 13: 542–552. Kark, S., T. Mukerji, U.N. Safriel, R. Nissani, and A. Darvasi. 2002. Peak morphological diversity in an ecotone unveiled in the chukar partridge by a novel Estimator in a Dependent Sample (EDS). Journal of Animal Ecology 71: 1015–1029. Lacy, R.C. 1997. Importance of genetic variation to the viability of mammalian populations. Journal of Mammalogy 78: 320–335. Lawton, J.H. 1995. Population abundances, geographic ranges and conservation: 1994 Witherby lecture. Bird Study 43: 3–19. Lawton, J.H., S. Nee, A.J. Letcher, and P.H. Harvey. 1994. Animal distributions: patterns and processes. Pp. 41–58 in P.J. Edwards, R.M. May, and N.R. Webb, eds., Large Scale Ecology and Conservation Biology. Blackwell Scientiﬁc Publications, Oxford. Lennon, J.J., J.R.G. Turner, and D. Connell. 1997. A metapopulation model of species boundaries. Oikos 78: 486–502. Lesica, P., and F.W. Allendorf. 1995. When are peripheral populations valuable for conservation? Conservation Biology 9: 753–760. Lomolino, M.V., and R. Channell. 1995. Splendid isolation: patterns of geographic range collapse in endangered mammals. Journal of Mammalogy 76: 335–347. Mayr, E. 1963. Animal Species and Evolution. Belknap Press/Harvard University Press, Cambridge, MA. Mayr, E. 1970. Populations, Species, and Evolution. Belknap Press, Cambridge, MA. McClintock, K.A. and M.J. Waterway. 1993. Patterns of allozyme variation and clonal diversity in Carex lasiocarpa and C. pellita (Cyperaceae). American Journal of Botany, 80: 1251–1263. Michaud, H., L. Toumi, R. Lumaret, T.X. Li, F. Romane, and F. Di Giusto. 1995. Eﬀect of geographical discontinuity on genetic variation in Quercus ilex (holm oak). Evidence from enzyme polymorphism. Heredity 74: 590–606. Mittermeier, R.A., N. Myers, J.B. Thomsen, and A.B. da Fonseca. 1998. Biodiversity hotspots and major tropical wilderness areas: approaches to setting conservation priorities. Conservation Biology, 12: 516–532. Møller, A.P. 1995. Patterns of ﬂuctuating asymmetry in sexual ornaments of bird from marginal and central populations. The American Naturalist 145: 316–327. Myers, N. 1990. The biodiversity challenge: expanded hotspots analysis. Environmentalist 10: 243–256. Myers, N., R. Mittermeier, C. Mittermeier, G. da Fonseca, and J. Kent. 2000. Biodiversity hotspots for conservation priorities. Nature (London) 403: 853–858.

Biodiversity Along Core–Periphery Clines 55 Nevo, E., and A. Beiles. 1988. Genetic parallelism of protein polymorphism in nature: ecological test of the neutral theory of molecular evolution. Biological Journal of the Linnean Society 35: 229–245. Nevo, E., and A. Beiles. 1989. Genetic diversity in the desert: patterns and testable hypotheses. Journal of Arid Environments, 17: 241–244. Olson, D.M., and E. Dinerstein. 1998. The Global 200: A representation approach to conserving the Earth’s most biologically valuable ecoregions. Conservation Biology 12: 502–515. Parsons, P.A. 1989. Environmental stresses and conservation of natural populations. Annual Review of Ecology and Systematics 20: 29–49. Parsons, P.A. 1991. Evolutionary rates: stress and species boundaries. Annual Review of Ecology and Systematics 22: 1–18. Paz, U. 1987. The Birds of Israel. Steimatzky, Tel-Aviv. Pinshow, B., P.U. Alkon, and A.A. Degen. 1982. Ecophysiological studies on water, energy and thermoregulatory relationships of Negev desert chukars and sand partridges. Israel Journal of Zoology 31: 63. Pinshow, B., A.A. Degen, and P.U. Alkon. 1983. Water intake, existence energy and responses to dehydration in the sand partridge (Ammoperdix heyi) and the chukar (Alectoris chukar), two phasianids of the Negev desert. Physiological Zoology 56: 281–289. Poiani, K., B. Richter, M. Anderson, and H. Richter. 2000. Biodiversity conservation at multiple scales: Functional sites, landscapes and networks. BioScience 50: 133–146. Pressey, R.L., C.J. Humphries, C.R. Margules, R.I. Vane-Wright, and P.H. Williams. 1993. Beyond opportunism: key principles for systematic reserve selection. Trends in Ecology and Evolution 8: 124–128. Purvis, A., and A. Hector. 2000. Getting the measure of biodiversity. Nature 405: 412–419. Randi, E., S. Kark, and C. Tabarroni. In review. Population structure of chukars (Alectoris chukar) distributed across an ecotone: A nested clade analysis of mitochondrial DNA haplotypes. Journal of Molecular Ecology. Rice, K.J., and R.N. Mack. 1991. Ecological genetics of Bromus tectorum. 1. A hierarchical analysis of phenotypic variation. Oecologia 88: 77–83. Rice, W.R., and E.E. Hostert. 1993. Laboratory experiments on speciation: what have we learned in 40 years? Evolution 47: 1637–1653. Riddle, B.R. 1996. The molecular phylogeographic bridge between deep and shallow history in continental biotas. Trends in Ecology and Evolution 11: 207–211. Safriel, U.N., S. Volis, and S. Kark. 1994. Core and peripheral populations, and global climate change. Israel Journal of Plant Sciences 42: 331–345. Schnabel, A., J.L. Hamrick, and P.V. Wells. 1993. Inﬂuence of Quaternary history on the population genetic structure of Douglas-ﬁr (Pseudotsuga menziestii) in the Great Basin. Canadian Journal of Forest Research 23: 1900–1906. Schneider, C., T. Smith, B. Larison, and C. Moritz. 1999. A test of alternative models of diversiﬁcation in tropical rainforests: Ecological gradients vs. rainforest refujia. Proceedings to the National Academy of Sciences 96: 13869–13873. Shirihai, H. 1996. The Birds of Israel. Academic Press, London. Shumaker, K.M., and G.R. Babble. 1980. Patterns of allozymic similarity in ecologically central and marginal populations of Hordeum jubatum in Utah. Evolution 34: 110–116.

56 Living Components of Biodiversity: Organisms

Slatkin, M. 1994. Gene ﬂow and population structure. Pp. 146–170 in L. Real, ed., Ecological Genetics. Princeton University Press, Princeton. Smith, T.B., R.K. Wayne, D.J. Girman, and M.W. Bruford. 1997. A role for ecotones in generating rainforest biodiversity. Science 276: 1855–1857. Svensson, B.W. 1992. Changes in occupancy, niche breadth and abundance of three Gyrinus species as their respective range limits are approached. Oikos 63: 147–156. Tani, N., N. Tomaru, M. Araki, and K. Ohba. 1996. Genetic diversity and diﬀerentiation in populations of Japanese stone pine (Pinus pumila) in Japan. Canadian Journal of Forest Research, 26: 1454–1462. Thomas, C.D., and I. Hanski. 1997. Butterﬂy metapopulations. Pp. 359–386 in I. Hanski and M.E. Gilpin, eds., Metapopulation Biology: Ecology, Genetics, and Evolution. Academic Press, San Diego. United Nations Convention 1992. Earth Summit—convention on biological diversity, Rio de Janeiro. Volis, S., S. Mendlinger, L. Olsvig-Whittaker, U.N. Safriel, and N. Orlovsky. 1998. Phenotypic variation and stress resistance in core and peripheral populations of Hordeum spontaneum. Biodiversity and Conservation 7: 799–813. Volis, S., S. Mendlinger, and N. Orlovsky. 2001. Variability in phenotypic traits in core and peripheral populations of wild barley, Hordeum spontaneum Koch. Hereditas 133: 235–247. Volis, S., S. Mendlinger, Y. Turuspekov, and U. Esnazarov. 2002. Phenotypic and allozyme variation in Mediterranean and desert populations of wild barley, Hordeum spontaneum. Evolution 56: 1403–1415. Volis, S., S. Mendlinger, and D. Ward. 2004. Demography and role of the seed bank in Mediterranean and desert populations of wild barley, Hordeum spontaneum Koch. Basic and Applied Ecology 5: 53–64. Vrijenhoek, R.C., M.E. Douglas, and G.K. Meﬀe. 1985. Conservation genetics of endangered ﬁsh populations in Arizona. Science 229: 400–402. Wade, M., and C. Goodnight. 1998. Perspective: The theories of Fisher and Wright in the context of metapopulations: When nature does many small experiments. Evolution 52: 1537–1553. White, M.J.D. 1951. Structural heterozygosity in natural populations of the grasshopper Trimerotropis sparsa. Evolution 5: 376–394. Wiens, J.A. 1989. The Ecology of Bird Communities. Cambridge University Press, Cambridge. Wilson, J.B., Y. Ronghua, A.F. Mark, and A.D.Q. Agnew. 1991. A test of the low marginal variance (LMV) theory, in Leptospermum scoparium (Myrtaceae). Evolution 45: 780–784. Wright, S. 1932. The roles of mutation, inbreeding, crossbreeding, and selection in evolution. Proceedings of the Sixth International Congress of Genetics 1: 232–248. Yom-Tov, Y., and E. Tchernov. 1988. The Zoogeography of Israel. Dr W. Junk, Dordrecht. Zanetto, A., and Z. Kremer. 1994. Genetic variation in natural populations of oaks in Europe: Quercus petraea and Quercus robur. In Conservation of the Wild Relatives of European Cultivated Plants III Workshop. Gibilmanna–Palermo.

4 Species Diversity, Environmental Heterogeneity, and Species Interactions William A. Mitchell Burt P. Kotler Joel S. Brown Leon Blaustein Sasha R.X. Dall

D

espite their apparent simplicity, arid environments can be quite heterogeneous. From small-scale variation in substrate and slope to large-scale geographic variation in solar input and productivity, drylands and deserts provide organisms with a tremendous range of ecological challenges (Schmidt-Nielsen 1964, Huggett 1995). Any single species is unable to meet all of these challenges equally well. A species will do better in some environments than others because evolution in heterogeneous environments is constrained by ﬁtness tradeoﬀs. Such tradeoﬀs prevent the evolution of a versatile species, competitively superior to all other species across the entire spectrum of heterogeneity (Rosenzweig 1987). Although ﬁtness tradeoﬀs may hinder species’ evolution in heterogeneous environments, they are a blessing for biodiversity. The source of biodiversity that we address in this chapter is the interplay of heterogeneity, tradeoﬀs, and density dependence. While we focus on species interactions at the local scale, our presentation includes a model that predicts changes in local diversity as a function of climate. The model’s predictions are based on changes in the nature of competition wrought by changes in productivity levels and climatic regimes. Cast in terms of evolutionary stable strategies (ESSs), the predictions refer to evolutionary as well as ecological patterns.

Mechanisms of Coexistence A mechanism of coexistence consists of an axis of environmental heterogeneity together with an axis that indicates a tradeoﬀ in the abilities of species 57

58 Living Components of Biodiversity: Organisms

to exploit diﬀerent parts of the axis. In the absence of some kind of heterogeneity, there is only one environmental type, and whatever species is best adapted to it will competitively exclude others. In the absence of a tradeoﬀ, one species could evolve competitive superiority over the full range of heterogeneity, again resulting in a monomorphic community. Consider some examples of mechanisms of species’ coexistence from dryland communities (Kotler and Brown 1988, Brown et al. 1994). For many taxa, spatial heterogeneity in predation risk is a consequence of the pattern of bushy and open areas common in drylands. In certain rodent communities, some species are able to exploit the relatively riskier open microhabitats by virtue of antipredator morphologies (Kotler 1984). One of these morphologies is bipedalism, which confers agility and fast speed on its owner at the cost of maneuverability in shrubs. Another tradeoﬀ is based on rodent body size. Larger rodents possess larger auditory bullae, especially in the family Heteromyidae (Kotler 1984). The size of the bullae reﬂects sensitivity to low ﬁeld vibrations (Webster and Webster 1971, Lay 1974) and the ability to detect approaching predators (Webster and Strother 1972). Large rodents acquire these facilities, however, at the cost of an overall higher metabolic rate, whereas smaller rodents are able to forage proﬁtably on lower seed densities. An example of these tradeoﬀs in action is a six-species rodent community at Tonopah Junction, Nevada, in the Great Basin Desert, which segregates species by body size along a bush-open microhabitat axis of heterogeneity (Kotler 1984). Other examples of mechanisms of species coexistence involve birds. Bushopen microhabitat selection among overwintering sparrows in semiarid grasslands apparently is based on diﬀerences in escape abilities, with some species being more vulnerable than others away from cover (Pulliam and Mills 1977). This can result in coexistence if the most vulnerable species is also the best resource competitor near protective cover. Each species then has a microhabitat in which it can proﬁt more than its competitor. A rodent community in the creosote shrublands of the Sonoran Desert exempliﬁes a diﬀerent sort of tradeoﬀ, this time involving travel eﬃciency versus foraging eﬃciency (Brown 1989a). The salient feature of the environment is the tremendous spatial heterogeneity in seed densities (adjacent patches can vary by as much as 70-fold in seed availability; Price and Reichmann 1987). The tradeoﬀ is mediated via body size. The round-tailed ground squirrel (Spermophilus tereticaudus) uses its larger size and speed to move quickly from patch to patch, visiting many patches in each foraging bout. Its high speed lowers its travel costs in both time and energy. This allows it to behave like a ‘‘cream skimmer’’ (Brown et al. 1994) ﬁnding and exploiting only the richest parts of the richest patches. The smaller Merriam’s kangaroo rat moves less frequently, but uses its lower metabolic rate and foraging costs to exploits patches more thoroughly and to lower seed densities like a ‘‘crumb picker’’ (Brown 1989b, Brown et al. 1994), thereby allowing it to coexist with the larger ground squirrel. A similar example involves the exploitation of a pulsed resource of nectar by solitary bees in the Sonoran Desert (Schaﬀer et al. 1979). Here, body size

Species Diversity, Environmental Heterogeneity, and Species Interactions 59

and wing disc loading appear to be involved in tradeoﬀs between harvest rate and foraging eﬃciency. Larger bees have higher wing disc loading, are faster ﬂiers, and presumably can harvest nectar more quickly, but they have high foraging costs due to the high disc loading. In contrast, because smaller bees have lower wing disc loading, they have less expensive (albeit slower) ﬂight, and presumably lower foraging costs. Bees show strong temporal partitioning, with larger bees foraging earlier in the day when nectar is most abundant, and smaller bees continuing to forage even as nectar is depleted. These examples illustrate the variety of circumstances under which heterogeneity and tradeoﬀs may combine to promote the ecological coexistence of competing species. In the model section of our chapter we employ the framework of heterogeneity and tradeoﬀs to study evolutionarily stable strategies in competitive communities. Using the model, we can predict how biodiversity changes as a function of productivity and climate (heat), parameters that show wide variation among dryland communities. We start with a brief overview of the two patterns of species diversity that motivated the modeling.

The Effects of Productivity and Climate (Heat) on Species Diversity One of the more puzzling patterns in ecology is the hump-shaped relationship between species richness and productivity (Rosenzweig and Abramsky 1993, Waide et al. 1999, Dodson 2000, Gross 2000, Mittelbach et al. 2001). When productivity is low, increased productivity is accompanied by increased diversity. But when productivity is high, increased in productivity corresponds to a decline in diversity. The ﬁrst part of the pattern seems easier to explain: more productivity yields more individuals across species, and the more individuals there are in each species, the less likely a species is to go locally extinct. This has been referred to as the ‘‘more individuals’’ hypothesis (Srivastava and Lawton 1998). The declining part of the productivity–diversity curve is more problematic. Why should higher productivity yield fewer species? Rosenzweig (1995) reviewed nine hypotheses listed in the literature and found problems with most of them. Some lacked a mechanistic basis, others would not hold in evolutionary time, and still others appeared logically circular. There remains a need for a mechanistically based hypothesis that predicts the declining phase of the productivity–diversity curve in both ecological and evolutionary time. In the present chapter, we try to address this need (see next section). The presence of a hump-shaped productivity–diversity curve means that some taxa will reach their maximum diversity in relatively unproductive drylands. For example, North American rodent diversity peaks in the mixed desert grassland of the Sonoran desert, and declines as productivity increases eastward (Rosenzweig 1995). In other cases, the dryland system can include both the peak diversity and part of the declining phase. For example,

60 Living Components of Biodiversity: Organisms

Rosenzweig (1995) ﬁnds the hump-shaped pattern in rodents of the Gobi desert, and Abramsky and Rosenzweig (1984) observe the pattern in Middle Eastern rodent assemblages. A diﬀerent pattern of species diversity shows that diversity increases with environmental heat, such as solar input, temperature, and potential evapotranspiration (Wright et al. 1993). The correlation between diversity and environmental heat may appear at ﬁrst glance to be an artifact of the wellknown correlation between diversity and latitude, but when latitude and heat compete for variance in diversity in regression analyses, heat usually wins (Currie 1991). Why should diversity increase with heat? Just as in the case of the increasing phase of the diversity–productivity curve, one explanation hinges upon population densities being higher where environments are warmer, due to lower metabolic costs of existence. If warmer climates support more total individuals, then species extinction rates will be lower, resulting in higher species diversity (Wright et al. 1993). This version of the ‘‘more individuals’’ hypothesis proposed to explain the increase phase of diversity with resource productivity (Srivastava and Lawton 1998). But there is a problem with this hypothesis. For example, in North America, the higher latitudes are both cooler and less speciose. Yet average population densities are actually higher in the north (Currie and Fritz 1993), rather than lower, as would be expected if the ‘‘more individuals’’ hypothesis accounted for the lower species diversity there. As in the case of the productivity–diversity pattern, the correlation of diversity with environmental heat calls for further work in developing and testing mechanistic models of species diversity. The need for predictive, mechanistic models is made imperative because purely correlational studies suﬀer from the inevitable correlations among the putative explanatory variables, including heat and productivity. Our model includes parameters for both productivity and maintenance cost, which in endotherms should be related to environmental heat. Consequently, we are able to use the model to predict how productivity and environmental heat can drive species diversity via mechanisms of coexistence. Furthermore, because it is an ESS model, the predictions remain valid for evolutionary time.

The Model Our model incorporates environmental heterogeneity and an evolutionary tradeoﬀ. We model environmental heterogeneity as a continuous distribution of habitat types, that is, there exists an inﬁnite variety of habitats. A continuous distribution is probably more realistic than a discrete and ﬁnite distribution. As ecologists, we may distinguish microhabitats by discrete categories such as ‘‘bush’’ and ‘‘open’’ for the sake of ﬁeld experiments and observations, but bush size, percent bush cover, or distance from the bush may be more important, and these are continuous variables.

Species Diversity, Environmental Heterogeneity, and Species Interactions 61

We also assume a continuous distribution of heritable phenotypes in the evolutionarily feasible set. These phenotypes are subject to an evolutionary tradeoﬀ such that each particular phenotype is superior in one habitat over all other phenotypes, and diﬀerent phenotypes are better suited to diﬀerent habitats (ﬁg. 4.1). By assuming a continuous distribution for habitat types and strategies, we avoid setting an a priori limit to the number of species that can coexist in either ecological or evolutionary time. Therefore, any limit to, or change in, diversity predicted by our model will ﬂow from the integrated eﬀects of environmental heterogeneity and ﬁtness tradeoﬀs, together with climate and resource productivity (Mitchell 2000, Mitchell and Porter 2001). In overview, our model has the following characteristics. Individuals bias foraging to those habitats that are more proﬁtable, i.e. where levels of food resource and habitat foraging costs deﬁne habitat proﬁtability. Levels of food resource, and hence individual foraging eﬀort, decrease with the density and foraging eﬀort of intra- and interspeciﬁc competitors (Mitchell et al. 1990). Foraging cost depends on the combination of habitat and phenotype. For this model, we represent the cost function for a given strategy, u, as a quadratic with its minimum value at a habitat in which the strategy pays its lowest foraging cost. Diﬀerent strategies possess diﬀerent cost functions with their minima located in diﬀerent habitats. Thus, for each of the inﬁnite number of habitat types, there exists a strategy that is superior in that habitat to all other strategies (ﬁg. 4.1). Finally, we assume that populations are limited by, and hence ﬁtness dependent on, a combination of food and nonenergetic factors.

Figure 4.1 A representative set of foraging cost curves, Cost ðu; zÞ; used in our model. Diﬀerent foraging cost curves correspond to diﬀerent evolutionary strategies. A cost curve shows the rate of energy expended by an individual of a particular strategy while it forages as a function of habitat of type z. Each strategy pays its lowest foraging cost in one habitat type, and increasingly higher costs as the habitats become increasingly diﬀerent from the lowest cost habitat. A ﬁtness tradeoﬀ results because diﬀerent strategies pay their lowest cost in diﬀerent habitats.

62 Living Components of Biodiversity: Organisms

Details of the Model We assume that ﬁtness is a function of both energetic and nonenergetic terms. The energetic term includes the energy available for reproduction, which is the diﬀerence between the net return from foraging and metabolic maintenance cost. The nonenergetic term represents the other eﬀects of crowding or density dependence aside from energy intake (e.g., limited nest sites or burrow refuges). A general way to express these assumptions is with the function Fitness ¼ ½b ðDaily Netprofit MCÞ ndayseN

ð1Þ

where b = Conversion rate of energy into oﬀspring (oﬀspring/ KJ) Daily Netproﬁt = Daily foraging proﬁt (KJ/(day individual) MC = Daily maintenance cost (KJ/(day individual) ndays = Number of days per generation (days/generation) b = Density-dependent term representing the eﬀects of crowding, independent of resource competition (i.e., limited nest sites) N = Total population density in the community Here, we deﬁne maintenance cost (MC) to be the lowest rate of energy expenditure by a resting individual in its environment. Cooler environments impose higher MC due to thermoregulation. The activity cost attributable to foraging (shown in ﬁg. 4.1), on the other hand, is accounted for in the Daily Netproﬁt, so it does not increase maintenance cost. We must expand the term for Daily Netproﬁt in equation (1) to account for environmental heterogeneity, evolutionary tradeoﬀs, and behavior. We calculate Daily Netproﬁt as the integral over all habitat types of an individual’s net proﬁt in each habitat. The net proﬁt in a habitat is the diﬀerence between energy intake and the energy expended due to foraging, which in turn depend on the combination of the strategy and the habitat type. Letting u represent strategy and z represent habitat, the term for net proﬁt in a habitat becomes Netprofitðu; zÞ ¼ Energy Intakeðu; zÞ Energy Expenditureðu; zÞ

ð2Þ

We assume that individuals deplete resources in a habitat as they forage. Therefore, energy intake from a patch is simply the diﬀerence between the initial and quitting resource densities, multiplied by the per item value of the resource, v, Energy Intakeðu; zÞ ¼ ½Initial Resource DensityðzÞ Quitting Resource Densityðu; zÞv

ð3Þ

Energy expenditure in a habitat is equal to the rate of energy expenditure attributable to foraging multiplied by foraging time in the habitat. As stated previously, and represented in ﬁg. 4.1, we indicate the rate of energy expenditure due to foraging in a habitat by the function Cost (u, z). Foraging time

Species Diversity, Environmental Heterogeneity, and Species Interactions 63

in a habitat is the time it takes the forager to deplete resources from the initial resource density to the quitting resource density. For simplicity, we assume that the rate of harvest, and hence depletion, is proportional to resource density. Letting a be the coeﬃcient relating resource density to the harvest rate, the foraging time becomes 1 Initial Resource DensityðzÞ Foraging timeðu; zÞ ¼ ln ð4Þ a Quitting Resource Densityðu; zÞ The quitting resource density represents the behavioral choice, or patchleaving rule, of the forager. The forager will achieve its greatest ﬁtness by quitting a habitat patch when the rate of return from the patch equals the cost of foraging the patch (Charnov 1976, Brown 1988). In our model there is no missed opportunity cost or predation risk, so the best time for a forager to leave is when the rate of energy harvest equals Cost (u,v). The rate of energy harvest at any resource density is the density multiplied by av, so, the quitting resource density is deﬁned by the equation Quitting Resource Densityðu; zÞ ¼ Costðu; zÞ=av

ð5Þ

By substituting equations (2–5) into equation (1), we can determine the behavior and ﬁtness of any phenotype, u, once we know the initial resource levels for all the habitats. We solve for initial resource levels as a dynamic balance between resource exploitation and resource renewal. To account for exploitation of resources, we envision foragers that move randomly among habitats, so that the frequency of visits to a habitat is proportional to the density of searching foragers. Hence, the opportunity for a habitat to renew between forager visits decreases with forager density. To account for renewal, we model a rate of resource input to all habitats. This rate represents productivity in our model.

Determining the ESS and Community Invasibility Fitnesses in our model are frequency- and density-dependent. We can illustrate this fact graphically by mapping ﬁtness on an adaptive landscape. An adaptive landscape illustrates ﬁtness as a function of strategy, for a range of possible strategies. An evolutionarily stable strategy (ESS) is a strategy that cannot be invaded by a rare alternative strategy; that is, all feasible alternative strategies will possess lower ﬁtness. Hence a resident strategy is an ESS if it resides on a ‘‘peak’’ in the adaptive landscape. Alternatively, if the resident resides on a slope in the adaptive landscape, that strategy can be invaded by an alternative strategy from the ‘‘upslope’’ direction. To generate an adaptive landscape, we ﬁrst introduce one or more species as residents. Then we use equations (1–5) to solve numerically for resident equilibrium population densities, that is, the densities at which resident ﬁtnesses equal one. By virtue of their foraging, the residents generate a resource distribution across the range of habitat types.

64 Living Components of Biodiversity: Organisms

Any nonresident species that tries to invade will encounter the resource levels generated by the foraging of residents. We can use equations (1)–(5) to calculate ﬁtnesses of nonresidents under the assumption that they are too rare to signiﬁcantly impact resource distributions. If the nonresident’s ﬁtness is less than one, it cannot successfully invade. But if its ﬁtness is greater than one, the propagule increases in density until density-dependence reduces its ﬁtness to one. In the process of growing to its equilibrium, the successful invader will reshape the habitat resource distribution. As a result, the adaptive landscape changes to reﬂect the new ﬁtnesses of the remaining set of nonresidents. In some cases a resident will prevent all nonresidents from successfully invading. This resident (or set of residents) will occupy a global peak (or peaks) in the adaptive landscape. If, however, a resident occupies a hillside in the adaptive landscape, it is subject to invasion or directional selection. Other conﬁgurations of the adaptive landscape are possible, as we will show in the next section. All numerical analysis was performed in QuickBasic, version 4.5 (computer program available from the authors). Equilibrium densities were solved using a bisection routine, and the integration over habitats to calculate net foraging proﬁt was performed using Romberg integration with an adaptive step size. Evolution of the Adaptive Landscape If a resident is not at an ESS, the system can be invaded, resulting in a modiﬁed adaptive landscape. Figure 4.2 illustrates this process in a system characterized by a low maintenance cost (high environmental heat) and low productivity. The endpoint of the invasions is a species that resides in a valley of the adaptive landscape. This valley is a stable minimum in that the species is not able to evolve out of it—any slight alteration in strategy by the species results in directional selection for the species to return to the stable minimum (Mitchell 2000). The stable minimum can be invaded by a second species or a reproductive isolate of the original resident. The two species are then subject to divergent selection, and evolve to occupy two distinct peaks in the adaptive landscape (ﬁg. 4.2D). The Effects of Environmental Heat (Maintenance Cost) and Productivity on Diversity Holding productivity constant and increasing maintenance cost (MC) has a dramatic eﬀect on the adaptive landscape and the invasibility of the system (ﬁg. 4.3). At the low MC, a single species evolves to a stable minimum as described above. This adaptive landscape is susceptible to invasion, or competitive ecological speciation (Schluter 2001), followed by divergent selection to yield a two-species ESS. At a higher level of MC however, the ‘‘valley’’ of the landscape ﬂattens out, producing at ﬁrst a small local peak. This local ESS is not subject to directional selection, but it can be invaded by a species whose strategy diﬀers signiﬁcantly. At still higher MC, the ‘‘shoulders’’ of the

Species Diversity, Environmental Heterogeneity, and Species Interactions 65

Figure 4.2 A sequence of adaptive landscapes in a system characterized by a low maintenance cost (MC ¼ 20 KJ/day) and low productivity (Prod ¼ 20). (A) A landscape generated by a single resident species that is not an evolutionarily stable strategy (ESS). The landscape rises sharply to the right of the strategy used by the species, indicating the strategies that could successfully invade the system. (B) The new adaptive landscape that results when the resident species evolves by directional selection or is simply replaced by the invasion of a new species positioned to the right on the adaptive landscape. (C) The endpoint of a single-species evolution is that the species occupies a stable minimum in the adaptive landscape. This adaptive landscape can be invaded by a diﬀerent species, or disruptive selection may promote sympatric speciation. (D) A two-species ESS that results if the stable minimum is invaded.

landscape fall below a ﬁtness value of one, and a single resident will inhabit a global peak, which is an ESS. By changing nothing more than MC, we have reduced the number of species that can coexist, even in evolutionary time. Increasing productivity produces results that are qualitatively similar to the results of increasing MC (ﬁg. 4.3). The increase in productivity reduces the number of species in the ESS from two to one. Mechanistically, what determines whether two species can coexist? Coexistence is possible if the combination of MC and productivity allow diﬀerent species to use the heterogeneity with suﬃcient diﬀerence. Maintenance cost and productivity determine the level of resources required by species to possess a ﬁtness of one. When MC and productivity are low, species are able achieve a ﬁtness of one even when competition holds the encountered resources in habitats to low levels. At low resource levels, habitat overlap among competitors is also low, because the rate of energy harvest

66 Living Components of Biodiversity: Organisms

Figure 4.3 Diversity and community invasibility is reduced by increased maintenance cost and productivity. The eﬀect of increased maintenance cost is shown in the three vertically stacked ﬁgures, where the adaptive landscape changes from stable minimum, to a local peak, to a global peak. The stable minimum is easily invadable by any strategy arbitrarily close to the resident. The global peak is not invadable, and constitutes a single species evolutionarily stable strategy (ESS). The eﬀect of increased productivity is shown by the horizontally arrayed ﬁgures. As with maintenance cost, an increase in productivity changes the landscape from an invadable stable minimum to a single species ESS. The broad arrow indicates that the stable minimum, once invaded, can evolve to a two-species ESS.

from secondary habitats is too close to the energy cost of using those habitats. But when MC is high, species need higher ambient resource levels to achieve a ﬁtness of one. Similarly, if productivity is high, higher resource levels are necessary to counteract the increased density-dependence due to the nonenergetic component of ﬁtness (e.g., limited burrow sites). In either case, the increased resource levels result in secondary habitats increasing in relative proﬁtability, and this encourages greater overlap in habitat use among com-

Species Diversity, Environmental Heterogeneity, and Species Interactions 67

petitors. As overlap increases, competition begins to favor the strategy with the lowest average cost of foraging across all habitat types. For the tradeoﬀ function we used (represented in ﬁg. 4.1), this is the strategy u ¼ 5, which resides in the middle of the strategy space.

Discussion Theories that combine environmental heterogeneity and tradeoﬀs ﬁnd their usefulness not only in helping us to understand coexistence of local competitors, but also in predicting how coexistence can depend on climate and productivity. For environmental heterogeneity to be relevant to diversity, it must induce ﬁtness diﬀerences. Furthermore, it must be accompanied by features of the organisms that induce tradeoﬀs in the ability of individuals to exploit diﬀerent parts of the heterogeneity (Kotler and Brown 1988). Finally, it helps if individuals can select from among the heterogeneity, avoiding areas and times where they are less eﬃcient than competitors. We have shown that the conditions for heterogeneity and tradeoﬀs to promote diversity depend on features of the environment not normally associated with models of local coexistence. Furthermore, our model predicts patterns of diversity that have been observed to hold on a geographic scale. The prediction of increased diversity with environmental heat does not depend on a species-abundance curve. Instead, the prediction is derived from how maintenance cost changes frequency-dependent ﬁtness in a competitive community. The model also avoids a problem that Rosenzweig and Abramsky (1993) had with some models that invoke environmental heterogeneity to explain the decrease in diversity with increased productivity. These authors note that explanations based on the assumption that productivity reduces heterogeneity (e.g., Tilman 1987) work ﬁne in ecological time, but these same explanations ignore the fact that habitat and resource subdivision are evolved responses of organisms, and hence these models may not work on an evolutionary time scale. It is not obvious, therefore, why a reduced range of heterogeneity should result in reduced diversity, rather than evolution of narrower habitat and resource use. Mitchell (2000) showed that even when habitats are continuous, the range of heterogeneity can still set a limit to species diversity. And the model we present here demonstrates that, even when evolution is allowed to operate, increased productivity can still reduce diversity. Dryland species diversity undoubtedly depends on interactions among variables acting over a range of scales. We have focused on species coexistence at the local scale. Regional models, in contrast, simplify the mechanics of local interactions to study the roles of dispersal and local extinction. These models can predict diversity based on a tradeoﬀ in dispersal and competitive ability (e.g., Tilman 1994), but in many such cases competitive abilities will still depend on local, within-patch heterogeneity. In these cases, the coexistence conditions will still be inﬂuenced by climate and productivity.

68 Living Components of Biodiversity: Organisms

In summary, the model that we present here illustrates how the geographically scaled variables can alter the chance that local mechanisms permit coexistence. The results of our model suggest that large-scale correlations between diversity and heat or productivity may be the result of local coexistence mechanisms, which should be tested empirically.

References Abramsky, Z., and M.L. Rosenzweig. 1984. Tilman’s predicted productivity-diversity relationship shown by desert rodents. Nature 309: 150–151. Brown, J.S. 1988. Patch use as an indicator of habitat preference, predation risk, and competition. Behavioral Ecology and Sociobiology 22: 37–48. Brown, J.S. 1989a. Desert rodent community structure: a test of four mechanisms of coexistence. Ecological Monographs 20: 1–20. Brown, J.S. 1989b. Coexistence on a seasonal resource. American Naturalist 133: 168–182. Brown, J.S., B.P. Kotler, and W.A. Mitchell. 1994. Foraging theory, patch use and the structure of a Negev Desert granivore community. Ecology 75: 2286–2300 Charnov, E.L. 1976. Optimal foraging: the marginal value theorem. Theoretical Population Biology 9: 129–136. Currie, D.J. 1991. Energy and large-scale patterns of animal- and plant-species richness. American Naturalist 137: 27–49. Currie, D.J., and J.T. Fritz. 1993. Global patterns of animal abundance and species energy use. Oikos 67: 56–68. Dodson, S.I., S.E. Arnott, and K.L. Cottingham. 2000. The relationship in lake communities between primary productivity and species richness. Ecology 81: 2662–2679. Gross, K.L., M.R. Willig, L. Gough, R. Inouye, and S.B. Cox. 2000. Patterns of species density and productivity at diﬀerent spatial scales in herbaceous plant communities. Oikos 89: 417–427. Huggett, R.J. 1995. Geoecology: an Evolutionary Approach. Routledge, New York. Kotler, B.P. 1984. Predation risk and the structure of desert rodent communities. Ecology 65: 91–96. Kotler, B.P., and J.S. Brown. 1988. Environmental heterogeneity and the coexistence of desert rodents. Annual Review of Ecology and Systematics 19: 281–307. Lay, D.M. 1974. Diﬀerential predation of gerbils (Meriones) by the little owl, Athene brahma. Journal of Mammalogy 55: 608–614. Mitchell, W.A. 2000. Limits to species richness in a continuum of habitat heterogeneity: an ESS approach. Evolutionary Ecology Research 2: 293–316. Mitchell, W.A., and W.P. Porter. 2001. Foraging games and species diversity. Annales Zoological Fennici 38: 89–98. Mitchell, W.A., Z. Abramsky, B.P. Kotler, B. Pinshow, and J.S. Brown. 1990. The eﬀect of competition on foraging activity in desert rodents: Theory and experiments. Ecology 71: 844–854. Mittelbach, G.G, C.F. Steiner, S.M Scheiner, K.L Gross, H.L. Reynolds, R.B. Waide, M.R. Willig, S.I. Dodson, and L. Gough. 2001. What is the observed relationship between species richness and productivity? Ecology 82: 2381–2396.

Species Diversity, Environmental Heterogeneity, and Species Interactions 69 Price, M.V., and O.J. Reichmann. 1987. Spatial and temporal heterogeneity in Sonoran Desert soil seed pools, and implications for heteromyid rodent foraging. Ecology 68: 1797–1811. Pulliam, H.R., and G.S. Mills. 1977. The use of space by wintering sparrows. Ecology: 1393–1399. Rosenzweig, M.L. 1987. Habitat selection as a source of biological diversity. Evolutionary Ecology 1: 315–330. Rosenzweig, M.L., and Z. Abramsky. 1993. How are diversity and productivity related? In Species Diversity in Ecological Communities: Historical and Geographical Perspectives, R. Ricklefs and D. Schluter, eds., pp. 52–65. University of Chicago Press, Chicago. Rosenzweig, M. L. 1995. Species Diversity in Space and Time. Cambridge University Press, Cambridge. Schaﬀer, W.M., D.B. Jensen, D.E. Hobbs, J. Gurvetich, J.R. Todd, and M.V. Schaﬀer. 1979. Competition, foraging energetics, and the cost of sociality in three species of bees. Ecology 60: 976–987. Schluter, D. 2001. Ecology and the origin of species. Trends in Ecology and Evolution 16: 372–380. Schmidt-Nielsen, K. 1964. Desert Animals: Physiological Problems of Heat and Water. Oxford University Press, New York. Srivastava, D.S., and J.H. Lawton 1998. Why more productive sites have more species: an experimental test of theory using tree-hole communities. American Naturalist 152: 510–529. Tilman, D. 1987. Secondary succession and the pattern of plant dominance along experimental nitrogen gradients. Ecological Monographs 57: 189–214. Tilman, D. 1994. Competition and biodiversity in spatially structured habitats. Ecology 75: 2–16. Waide, R.B., M.R. Willig, C.F. Steiner, G.G. Mittelbach, L. Gough, S.I. Dodson, G.P. Juday, and R. Parmenter. 1999. The relationship between primary productivity and species richness. Annual Review of Ecology and Systematics 30: 257–300. Webster, D.B., and D.B Strother. 1972. Middle-ear morphology and auditory sensitivity of heteromyid rodents. American Zoologist 12: 727. Webster, D.B., and M. Webster. 1971. Adaptive value of hearing and vision in kangaroo rat predator avoidance. Brain, Behavior and Evolution 4: 310–322. Wright, D.H., D.J. Currie, and B. A. Maurer. 1993. Energy supply and patterns of species richness on local and regional scales. In Species Diversity in Ecological Communities, R.E. Ricklefs and D. Schluter, eds., pp. 66–74. University of Chicago Press, Chicago.

5 SHALOM A Landscape Simulation Model for Understanding Animal Biodiversity Yaron Ziv Michael L. Rosenzweig Robert D. Holt

T

he ecological complexity of landscape components of biodiversity may be understood by examining relatively simple landscapes such as those of arid and semiarid lands. It is believed that such lands provide easy recognition of their components and a relatively simple interaction between their diﬀerent diversities (Safriel et al. 1989). In general, ecological complexity emerges from the existence of environmental heterogeneity and scaling eﬀects. The eﬀects of scaling include the diﬀerential changes in observed patterns produced by processes that operate and interact at diﬀerent tempospatial scales. For example, interspeciﬁc competition may have a strong inﬂuence on species coexistence and, therefore, diversity, at a local scale, may be insigniﬁcant for determining species diversity compared with a regional scale, where colonization–extinction dynamics may be the major determinant for species diversity. Environmental heterogeneity mainly results from three components: habitat diversity (the number of diﬀerent habitats), habitat size (the size of each habitat’s patch), and habitat patchiness (the distribution of the diﬀerent habitats’ patches in the landscape). Each component may aﬀect species diversity by providing speciﬁc processes for coexistence, colonization, extinction, and population-size dependent eﬀects. Additionally, as emphasized by Kotliar and Wiens (1990), diﬀerent scales (Wiens 1989) should introduce diﬀerent levels of heterogeneity that may inﬂuence the way organisms respond to their environment. Morris (1987) suggested that an organism that does not respond to a particular heterogeneity presented at one scale may respond to the heterogeneity presented at another scale. This concept has led ecologists to accept the idea that ecological processes and patterns are not 70

SHALOM: A Landscape Simulation Model 71

ﬁxed, but rather depend on the scale under study (e.g., Addicott et al. 1987, Kotliar and Wiens 1990, Dunning et al. 1992, Wiens et al. 1993). In this chapter we describe a spatially explicit, multispecies, process-based landscape simulation model, SHALOM (Species-Habitat ArrangementLandscape-Oriented-Model) that has been designed to explore ecological complexity of large scales. After describing the model, we will present several simulation results to demonstrate the strengths of using such models for understanding biodiversity processes and patterns. We believe that this model can serve an important tool for exploring biodiversity in arid and semiarid lands.

Model Design The model is coded in C++ (Stroustrup 1995) using object-oriented programming (Booch 1991, Martin 1995) for designing the diﬀerent components of ecological structure (e.g., species, habitats) as classes of objects. A class is a general template of a particular component of a model, treated as an autonomic unit obtaining its own characteristics and functions (i.e., encapsulation). Object-oriented programming allows us to model natural systems realistically because diﬀerent components of a model can be designed and coded as classes of objects. The model is based on ecological realism. First, it explicitly deﬁnes the processes aﬀecting species, populations, and communities (hence, processbased model); in most cases it goes beyond the simple description of a process to characterize it by its mechanics. Second, it avoids arbitrary functions and arbitrary value assignments by relying on empirical ecological ﬁndings. Finally, many of the processes’ coeﬃcients depend on body size via allometric equations where parameters for these equations come from the empirical literature (see Peters 1983, Schmidt-Nielsen 1984, Calder 1996). This, in turn, ensures that values for many processes of the model are realistic. Figure 5.1 describes the relationship between the diﬀerent classes and the position of the processes between the classes according to the way they are modeled. Note the hierarchical structure of the model: the landscape-scale processes (described below) are invoked by the class landscape directly, while the local-scale processes are invoked at the patch-population level (described below). A detailed description of the model is found in Ziv (1998).

Model’s Classes and Their Characteristics The model deﬁnes seven biological (population, species, community) and physical (cell, patch, habitat, landscape) components that produce an ecological structure as the model’s classes. It uses the current terminology of landscape ecology (e.g., Forman and Godron 1986, Turner 1989) for the terms used here.

72 Living Components of Biodiversity: Organisms

Figure 5.1 The class-relationship diagram of the model. Notice that, consistent with the multiscale design of the model, the landscape-scale processes are positioned between the landscape and the patch classes, while the local-scale processes are positioned at the population-community level.

A landscape (the coarse grain of the model) is the entire area under study composed by a row-by-column matrix of cells. The size of the landscape is determined by its number of cells and the area of each cell in the matrix. Two processes are directly controlled by the landscape: ‘‘catastrophic stochasticity’’ and ‘‘dispersal’’ (detailed description of these processes is given in the ‘‘Model’s Processes’’ section below). A cell is a square in the landscape matrix that serves technically to produce patches. Each cell has an ‘‘area’’ and contains a single habitat type. It is the ﬁne grain of the model. A habitat has relatively homogeneous physical and biological attributes. The physical characteristics are temperature and precipitation, because at large scales the combination of temperature and precipitation distinguishes particular ecosystems and biomes (Holdridge 1947, Lieth and Whittaker 1975). Temperature and precipitation are characterized by their long-term annual mean and standard deviation. These statistics may be linked in a probabilistic manner (the higher the standard deviation, the less likely that the mean is met in a given year). The biological characteristics of a habitat are the list of ‘‘resources’’ it oﬀers as well as the ‘‘resource-proportion distribution’’ of each of these resources. Resources are assumed to be discrete. ‘‘Resourceproportion distribution’’ represents the proportion of each resource

SHALOM: A Landscape Simulation Model 73

in the habitat (e.g., for two resources that occur equally in a habitat, each has a resource-proportion of 0.5). A patch is the area composed of all adjacent cells sharing a habitat type where the local-scale processes take place. Individuals of a species in one patch (population) interact among themselves independently of individuals in adjacent patches. Dispersal may connect patches. Variation in cells and habitat result in patch-speciﬁc characteristics, such as ‘‘energy supply’’ and ‘‘resource-proportion productivity.’’ ‘‘Energy supply’’ is given by multiplying productivity (energy per unit of time per unit of area) by the patch’s area, while productivity is calculated as a linear function of the product temperature times precipitation (Rosenzweig 1968, Lieth and Whittaker 1975). ‘‘Resourceproportion energy supply’’ is the amount of energy per unit time oﬀered by each resource represented in the patch. The resources and their distribution are determined by the patch’s habitat. A species is the sum of all populations in the landscape, that is, a species is a metapopulation. Each species has ‘‘body size,’’ ‘‘niche position’’ (deﬁned by habitat and resource utilization axes described below), and ‘‘dispersal coeﬃcient.’’ Body size plays an important role in the model. ‘‘Birth rate,’’ ‘‘death rate,’’ and ‘‘metabolic rate’’ can be body-size dependent (Y ¼ aM b , where Y is a rate, M is body size, and a and b are coeﬃcients; Calder 1996). Habitat utilization and resource utilization usually play important roles in a species’ niche position. These utilizations reﬂect the physical and biological characteristics of a habitat. Thus the model can compare what is oﬀered by a patch with what is required by a species in it. (This comparison takes place in the class ‘‘population’’ and is called ‘‘species-habitat match.’’) Habitat utilization is deﬁned by the ‘‘temperature’’ and ‘‘precipitation’’ requirements which, for simplicity, determine the species’ niche. The temperature and precipitation requirements of a species are set by each characteristic’s ‘‘mean’’ and ‘‘standard deviation.’’ We assume that the mean represents the value at which a species reproduces best, while the standard deviation represents the species’ tolerance to values that are diﬀerent from the mean. We also assume a tradeoﬀ between maximum performance and tolerance: the higher the standard deviation, the worse the species does at each point in its niche. This tradeoﬀ allows for tolerance–intolerance community organization (see Colwell and Fuentes 1975, Rosenzweig 1991). Temperature and precipitation can be represented by a binormal distribution according to the ‘‘central limit theorem’’ (see Durrett 1991). Hence, a species’ niche is characterized by a binormal space, shaped by the temperature and precipitation’s mean and standard deviation. The lists of ‘‘resources’’ and ‘‘resource-proportion use’’ set the resource utilization of a species. As in class ‘‘habitat,’’ resources are distributed discretely.

74 Living Components of Biodiversity: Organisms

Each species also has a ‘‘dispersal coeﬃcient,’’ which determines the intensity of dispersal when and if it is invoked. The dispersal coeﬃcient is a species-speciﬁc dimensionless value that allows the model to speed up or slow down the movement of populations relative to other populations or relative to the same populations in other simulations. A population is the group of individuals belonging to a given species in a particular patch. Many of the population’s characteristics are determined by the ‘‘species’’ it belongs to. Some of these characteristics do not change during a simulation (‘‘body size,’’ ‘‘birth rate,’’ ‘‘death rate,’’ ‘‘metabolic rate,’’ ‘‘habitat utilization,’’ and ‘‘dispersal coeﬃcient’’). Other characteristics do change according to the requirements and pressures a particular population faces in each ‘‘patch.’’ The information from the patch sets such changes. The population’s ‘‘intrinsic rate of increase’’ (i.e., the maximal growth rate with no intra- and interspeciﬁc competitors) is calculated by subtracting the species’ death rate from its habitat-speciﬁc birth rate. The latter is obtained by multiplying the species–habitat match value (see ‘‘Model’s Processes’’ below) by the species birth rate. ‘‘Initial population size’’ is the number of individuals at the beginning of a run. The model allows initial population sizes to diﬀer. Thus, one can explore how initial conditions may aﬀect the community and landscape (e.g., priority eﬀect; Quinn and Robinson 1987, Lawler and Morin 1993). The ‘‘carrying capacity’’ of a population is its population size at equilibrium in the absence of stochasticity. The list of ‘‘resources’’ used by a population results from the resources used by its species and the resources available in the patch. The population’s resource-proportion use is then rescaled accordingly (considering only the resources that are actually used), maintaining the ratios of all resources used in the patch. For example, if only two resources can be used by the population and they have fundamental proportions (according to the species’ ‘‘resource-proportion use’’) of 0.1 and 0.3 (i.e., 1:3 ratio), then they will be rescaled to have proportions of 0.25 and 0.75 in the population’s diet. A community is the set of nonzero populations in a patch.

Model’s Processes Ecological processes are simulated on two scales—local and landscape—similar to the general separation made by Whittaker and Levin (1977). Localscale processes occur within each patch, while the landscape-scale processes are those that occur across or between patches. This multiscale hierarchy allows most processes to work inside patches and to have a direct impact on population growth. Meanwhile, processes occurring between patches can aﬀect population growth indirectly and at diﬀerent temporal scales.

SHALOM: A Landscape Simulation Model 75

Local-Scale Processes Community-level saturation eﬀect, fðsÞ The community-level saturation eﬀect builds on the ratio between the energy oﬀered by a patch (i.e., energy supply) and the overall energy consumed by all populations in a patch. The energy consumed by all populations in the patch is the sum of each population’s species-speciﬁc energy consumption, which is calculated by multiplying the metabolic rate of the species to which the population belongs by the number of individuals in that population. Because a patch’s energy supply and a species’ metabolic rate share units (energy/time), the division of these two gives a dimensionless variable (e.g., Vogel 1994) that ranges from zero (i.e., no individuals at all) to any positive value. A patch may oﬀer more than one resource. A population may consume all of the patch’s resources or only a subset of them, depending on the population’s list of resources. Each resource’s energy in a patch is determined by its proportion (resource-proportion energy supply) out of the energy supply in that patch. An algorithm sets the relative use of each resource by those species that share it. The community-level saturation eﬀect equation treats each resource one at a time and then sums all resources. The following equation describes the community-level saturation eﬀect on population j, fðsÞj , given its species i, for K resources: fðsÞj ¼

K X S X RPUkl Nl EMi k¼1 l¼1

RPPk

ð1Þ

where l is a population selected from all S existing populations in a patch, RPUkl is the resource-proportion use of resource k by population l, Nl is the size of population l, EMi is the body-size dependent metabolic rate of species i, which population l belongs to, and RPPk is the resource-proportion energy supply of resource k in a patch. The community-level saturation eﬀect is analogous to the carryingcapacity feedback function of the logistic equation (May 1981). However, the model does not assume an arbitrary value for carrying capacity. Instead, the value for carrying capacity comes from calculating the equilibrium of a population when saturation exists. It represents the density-dependent pressure a population experiences from all of a patch’s populations, including its own. Hence, it includes both intra- and interspeciﬁc density dependence. Species–habitat Match, fðmÞ The species–habitat match quantiﬁes how well individuals of a particular population are suited to a particular patch, given the population’s species and the patch’s habitat. The function builds on the overlap between the temperature–precipitation binormal curve of the species and the temperature–precipitation biuniform curve of the habitat. Speciﬁcally, the population’s niche space is given by the following binormal distribution equation:

76 Living Components of Biodiversity: Organisms

(

D1 ¼

"

1 exp 0:5 1 p2

!#) x XiT y XiP y XiP 2 þ 2p SDiT SDiP SDiP pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 2SDiP SDiT ð1 p2 Þ ð2Þ

x XiT SDiT

2

where x and y are values of temperature and precipitation at the patch, XiT is the species’ temperature requirement’s mean, SDiT is the species’ temperature requirement’s standard deviation, XiP is the species’ precipitation requirement’s mean, SDiP is the species’ precipitation requirement’s standard deviation, and p is a covariance between the species temperature and precipitation. The patch’s habitat space is given by the following biuniform distribution equation: D2 ¼ 4SDhT 4SDhP D} 1

ð3Þ

where D} 1 is the highest distribution value of the population’s species niche space, SDhT is the habitat temperature characteristic’s standard deviation, and SDhP is the habitat precipitation characteristic’s standard deviation. The ﬁnal species–habitat match value for a given population in a particular patch (fðmÞj Þ is given by dividing the population’s niche space nested within the patch’s habitat space by the patch’s entire habitat space. The species–habitat match value represents the fraction of the population’s species ability expressed in the particular patch given its habitat. A value of 1 represents a perfect match, while a value of 0 represents no match at all. The above form of calculating species–habitat match provides two major outcomes that we should expect to see in nature. First, the more tolerant a species, the more likely it will match a habitat far away from the species population’s temperature and precipitation mean values. Second, the lower the standard deviation of the habitat’s precipitation and temperature characteristics, the higher the species–habitat match. This should be true because a habitat’s standard deviations are negatively correlated with the probability of getting a particular value at a given time. Higher standard deviations represent a lower probability of any species ﬁnding a given value in a habitat. Ecologically, this should represent a measure of predictability: the lower the standard deviations of the habitat, the better it is for the populations occurring in that habitat. The population dynamics equation (eq. (5)) uses a single value for the species–habitat match. Other functions can be used to get the desired value. When possible, the species–habitat match should be generated with empirically derived functions that use the natural history of the species and more accurate measurements of how well the species thrives in the available habitats. Demographic stochasticity This refers to any change in population size caused by a chance event (resulting from sampling errors), independent of a

SHALOM: A Landscape Simulation Model 77

biological process. It tends to have critical eﬀects when populations sizes are low. We used a simple descriptive equation to model stochastic deviations from the deterministic, body-size-dependent birth and death rates. The deviations are negatively correlated with population size: the larger the population, the lower the deviations are likely to be. Although the equation does not relate to any speciﬁc process (e.g., sex ratio or encounter rate), its behavior does follow the typical expectations of such stochasticity. The equation aﬀects demographic parameters randomly and it is density-dependent (e.g., Diamond 1984, Shaﬀer and Samson 1985, Pimm et al. 1988, Lande 1993). The following equation deﬁnes the population’s stochasticity in birth or death rates, Zj , from a species’ deterministic birth or death rates, zi : ! "ð0:5zi Þ pﬃﬃﬃﬃﬃﬃ Z j ¼ zi Nj

ð4Þ

where e is a random number sampled from a Gaussian probability distribution (with a mean of zero and a symmetrical truncation of two standard deviations, of one unit each), 0.5zi is a scaling term to make each distribution range between zero and twice the highest birth or death rate, g is a demographic stochasticity coeﬃcient that allows for changing the ‘‘intensity’’ of the eﬀect, and Nj is population size. We used a logistic-like continuous-time population growth for the localscale population dynamics. Birth rate and death rate are handled independently. This separation is realistic (Begon et al. 1986) because birth rate and death rate may be limited by diﬀerent processes, such as a need for proteinrich resources for lactating females that are not required by the rest of the population. Overall, the equation by which a given population grows in a patch given the above processes is dNj ¼ Nj bi ffðmÞj ð1 fðsÞj Þþ g Nj di f1 þ fðsÞj g dt

ð5Þ

where fðmÞj and fðsÞj are the species–habitat match eﬀect and the saturation eﬀect, respectively, and (1 fðsÞj Þþ indicates that the latter term cannot take a value lower than zero (see Wiegert 1979). The community-level saturation eﬀect ðfðsÞj ) enters the equation twice. First, we subtract the community-level saturation eﬀect from one as in the carrying-capacity feedback function of the logistic equation (i.e., 1 N=K). This new term models the eﬀect of the community saturation on birth. It is assumed (as in the logistic equation) that birth decreases linearly with an increase in community density. Oversaturation (i.e., 1 fðsÞj < 0 results in no birth. Second, we add one to the community-level saturation eﬀect to model the eﬀect of the community saturation on death. Here also, it is assumed that death rates decrease linearly with an increase in community density.

78 Living Components of Biodiversity: Organisms

The local-scale population dynamics equation with its analytical solution and outcomes for body-size dependent habitat speciﬁcity are found in Ziv (2000). Landscape-Scale Processes Dispersal, fðdÞ This is the movement of individuals from one patch to another (e.g., Levin 1974, Andow et al. 1990, Johnson et al. 1992, Gustafson and Gardner 1996). In the model, individuals of a particular population in a given patch migrate to adjacent patches if they can gain a higher potential ﬁtness there. The dispersal function builds on the optimization principles used for intraspeciﬁc density-dependent habitat selection suggested by Fretwell and Lucas (1969) and Fretwell (1972) (ideal free distribution). In the model, the dispersal process assumes that a population’s individuals can instantly assess the adjacent population’s per-capita growth rate. At each time step, the model calculates the per-capita growth rate of each population. Then it compares that rate with all adjacent populations’ percapita growth rate. Individuals move from patches with relative low percapita growth rate (i.e., low ﬁtness potential) to patches with high per-capita growth rate (i.e., higher ﬁtness potential). This results in equalizing the percapita growth rates of populations of the same species across patches (Fretwell 1972). Dispersal occurs on a continuous-time scale. Hence, dispersal from a given patch to patches that are unadjacent can happen fast in appropriate conditions (e.g., some patches of low potential ﬁtness and a patch of a very high potential ﬁtness). However, there is an implicit distance eﬀect because individuals need to cross the adjacent patches ﬁrst, and because each population in the diﬀerent patches experiences population change due to other processes. This eﬀect can be controlled by changing the species dispersal coeﬃcient such that the rate at which individuals of populations of a given species move agrees with the user’s needs. Catastrophic stochasticity Also known as disturbance-induced extinction (Levin and Paine 1974, Pickett and White 1985, Turner et al. 1989), this is a density-independent loss of individuals due to some event (e.g., extreme cold weather or a drought) that has a random probability of occurrence. Some environments may have a higher probability of being aﬀected by catastrophes than others. Catastrophes may cause the disappearance of entire populations of a given community or only partial disappearance. The same catastrophe may eliminate some species from a patch but only reduce others. A catastrophic event may be very local, such as within a single habitat (e.g., a falling tree in a forest), or may cover an extensive area and include many diﬀerent types of habitats (see Turner et al. 1989). The catastrophic stochasticity of the model SHALOM relies on randomnumber-generating procedures (Press et al. 1995). These allow one to change the probability, intensity, and range of the density-independent loss of individuals and populations. The user sets the following options: the probability

SHALOM: A Landscape Simulation Model 79

function (either uniform or Gaussian) of the catastrophic stochasticity distribution, the threshold (a fraction between 0 and 1) below which catastrophic stochasticity is not invoked, the lower and the upper limits (a fraction between 0 and 1) for population loss once a catastrophic stochasticity is invoked, the probability function (either uniform or Gaussian) of the population loss, and the spatial distribution (either a random or a ﬁxed distribution on a cell, or patch, or the entire landscape) of the catastrophic stochasticity. The two landscape-scale processes aﬀect population growth on two diﬀerent time scales. As mentioned above, dispersal is assumed to occur on a continuous-time scale similar to the continuous-time scale of the local population dynamics. In fact, dispersal at any time step of the model depends on the local-scale per-capita growth rate of each population. Deﬁning the local growth of population j in equation (5) as FðlÞj , the overall population growth, including dispersal, becomes AP X dNj ¼ FðlÞj þ ðfðdÞjl NjðÞ =lðþÞ Þ dt l¼1

ð6Þ

where AP is the number of adjacent patches and NjðÞ=lðþÞ indicates that the per-capita migration is multiplied by the patch’s population size or by the adjacent patch’s population size, depending on the sign of the per-capita movement. A positive per-capita movement means that the particular patch’s per-capita growth rate is higher than the one adjacent. Hence, individuals from the adjacent patch disperse into it. In contrast, a negative per-capita movement means that individuals should disperse into the adjacent patch. Catastrophic stochasticity is simulated on a discrete time scale. Once a year (or on an interval that amounts to a year), the model invokes catastrophic stochasticity. Model Mechanics Before each run of the model, the user assigns the following: the species and their attributes, the habitats and their attributes, and the habitat arrangement in the landscape. Given this information, the model creates the patches as they would appear to organisms in the real world. Having modeled patches and species in the landscape, populations are then created. The species– habitat match of a population is then calculated. The option of invoking demographic stochasticity is set for each population. All populations of a particular patch create the patch’s community. The community monitors the overall saturation eﬀect in a patch as well as the diﬀerent species’ composition and diversity. Once the landscape is completely deﬁned, the model asks for information about the large-scale processes. Dispersal may or may not be invoked by the user. Similarly, catastrophic stochasticity may or may not be invoked. If catastrophic stochasticity is invoked, the model asks for information about its diﬀerent options.

80 Living Components of Biodiversity: Organisms

Following the speciﬁcation of the initial population size for each population and the run time (in years), the model runs a population-growth simulation of the diﬀerent populations in the diﬀerent patches. The Runge–Kutta method (Press et al. 1995) is used to integrate the small steps (dt ¼ 0:001 yr) on a continuous time axis. The model returns the value of population size for each population in the diﬀerent patches every 100 time steps (i.e., 0.1 yr). The information is saved to an output ﬁle for further analysis. At the end of the run, the model calculates the ratio of each population’s size to its carrying capacity and returns values of the number of species and two species-diversity indices: Simpson’s diversity index (Simpson 1949) and Fisher’s alpha (Fisher et al. 1943).

Using the Model: An Example of the Effects of Ecological Processes on Community Structure in a Heterogeneous Landscape How do diﬀerent processes—interspeciﬁc competition, demographic stochasticity, and dispersal—known to aﬀect communities at a local scale, aﬀect species composition and species-diversity patterns in a spatially heterogeneous landscape scale? Many studies have explored various processes that aﬀect communities in heterogeneous landscapes. However, these studies treat each process discretely (e.g., Andow et al. 1990, Dunning et al. 1992, Holt 1992). How the interaction of multiple processes aﬀects community structure is rarely explored, except in the context of metapopulation dynamics. In the following sections, we will describe a simulation design that allows the modeling of several species of diﬀerent body sizes in a very simple heterogeneous landscape without losing track of the species diversity in each patch or in the entire landscape. As will be shown later, this simple simulation will provide enough information to make some sophisticated predictions. Simulation Design We simulated a landscape with 2 2 cells, each 100 m2, having its own unique habitat (total of four habitats). We chose this simple landscape design because the existence of the diﬀerent processes in the current simulation added a tremendous amount of complexity to the model. Thus, the simple landscape design provides focus on the processes’ outcomes. We assigned realistic productivity values for the diﬀerent habitats without a speciﬁc process in mind in order to keep the model as general as possible. Note that also in this simulation, patch and habitat are synonymous. We also assigned diﬀerent species–habitat matches to the diﬀerent habitats, such that habitat 1 was the best habitat and habitat 4 was the worst (species–habitat match ¼ 0.997, 0.987, 0.971, and 0.949 for habitats 1, 2, 3, and 4, respectively). To allow for competitive coexistence between the modeled species,

SHALOM: A Landscape Simulation Model 81

each habitat oﬀered 28 diﬀerent resources. To avoid a speciﬁc resourceproductivity distribution, we assigned an equal productivity for each resource out of the total productivity of the habitat. We simulated a total of 26 species. Species diﬀered in only one characteristic—body size. Body size ranged between 5 g and 1585 g, corresponding to log values of body size ranging between 0.7 and 3.2. We assigned a unique preferred resource to each species and gave it a resource-proportion use of 0.5. Each species could consume two other resources, one on each side of the preferred resource; each of these had a resource-proportion use of 0.25 (e.g., species 1 is able to consume resources 1, 2, and 3 with a resource-proportion use of 0.25:0.5:0.25, species 2 is able to consume resources 2, 3, and 4 with a resource-proportion use of 0.25:0.5:0.25, and so on). Preliminary simulations have shown that this resource allocation was suﬃcient to produce a competitive relationship with resource partitioning without assuming any complex resource-use function. Throughout the simulations, we used the allometric power coeﬃcients known for eutherian mammals for birth rate (0.33), death rate (0:56), and metabolic rate (0.75) (Calder 1996). When catastrophic stochasticity was invoked, we gave the system a 10% chance of suﬀering catastrophic stochasticity in a year (an average of one catastrophe every 10 years). In catastrophic years, stochasticity can aﬀect up to 50% of the landscape with up to 50% loss of population size in those patches aﬀected. These values were chosen after experimenting with many simulation designs. They are high enough to aﬀect population and species distribution (Turner 1987), yet, low enough that no populations are driven to extinction. Other than these ﬁrst-level assignments of values for cells, habitats, and species, no other assignments were made for second-level procedures such as habitat-speciﬁc population abundance, etc. Therefore, any bodysize-dependent patterns that emerge will result only from the basic rules described here. To understand the eﬀects of the diﬀerent ecological processes, we initially explored the patterns emerging from communities not aﬀected by any of the above processes (i.e., in which competition is strictly intraspeciﬁc). We then introduced interspeciﬁc competition, and added, thereafter, demographic stochasticity to interspeciﬁc competition to explore how it changes the predicted patterns. Finally, we allowed dispersal to connect all patches. Results Carrying capacities All habitats were suitable for all species. That is, without any population-reducing processes—interspeciﬁc competition and demographic and catastrophic stochasticities—all populations in all habitats could maintain a persistent population size. Figure 5.2 shows the carrying capacities of the diﬀerent populations in the diﬀerent habitats as well as the species abundance in the entire landscape. Because all populations can persist in

82 Living Components of Biodiversity: Organisms

Figure 5.2 Carrying capacities of species as a function of their body size in the diﬀerent patches (habitats) and in the entire landscape.

all habitats, and because no process other than intraspeciﬁc competition aﬀects population growth, the same population size pattern in all the habitats and the same species diversity pattern across the entire landscape (i.e., the sum of all population sizes of each species) emerge. The only diﬀerence between habitats is that carrying capacities of populations of the same species are lower in habitats with fewer species–habitat matches. The eﬀect of interspeciﬁc competition Here we assumed that resource partitioning occurs such that the most preferred resource is diﬀerent for each species. Because of the overlap in resource use, each resource is consumed by three species. This shared consumption can lead to competitive exclusion. When resources are equally shared by species of diﬀerent body sizes, the larger species outcompetes the smaller species that use the same resources. This outcome results from the lower death rates of larger species. Regardless of the speciﬁc mechanism, this larger-species competitive advantage is consistent with competitive outcomes observed in many real systems (see Kotler and Brown 1988), conﬁrming the eﬀectiveness of the model SHALOM. Due to the modeling of resource partitioning as a deterministic process that does not change between habitats, the same species composition exists in all four habitats of the model as well as for the entire landscape (ﬁg. 5.3). With interspeciﬁc competition, some populations are outcompeted, leaving a discontinuous distribution of body sizes. The absence of a particular species depends on an intratrophic level cascading eﬀect: the largest species depresses the second largest species population size due to the largest species’ competitive advantage. Although the second largest species has the competitive

SHALOM: A Landscape Simulation Model 83

Figure 5.3 Sizes of all populations in the diﬀerent patches (habitats) and in the entire landscape with interspeciﬁc competition. Interspeciﬁc competition deterministically aﬀects all populations the same way. Identical pattern emerge for each patch and for the entire landscape.

advantage over the third largest species, the present but minimal eﬀect of the third largest species on the second largest species is enough to depress the former further to local extinction. Because, in the present model, species share resources only with the species closest in body size, the third largest species, which does not share resources with the largest one, is saved from the potentially dominating eﬀect by the extinction of the second largest species. The process repeats with the fourth, ﬁfth, and sixth largest species, and so on. Because all interactions between all species are taking place simultaneously, the overall eﬀect on the diﬀerent species sometimes results in an absence of a species particular body size in between two coexisting species, each having close body sizes. The two species coexist because the larger species can consume its most preferred resource better, and has a competitive advantage, while the smaller species beneﬁts from the other resource that is no longer used by the now extinct, smaller species. In the end, 12 species coexist in the landscape. Adding demographic stochasticity to interspeciﬁc competition Demographic stochasticity, or the sampling eﬀects regarding sex ratio, litter size, etc., that may promote local extinctions of small populations, exists no matter what other processes aﬀect population growth (Pimm et al. 1988, Lande 1993). With demographic stochasticity, diﬀerent patterns appear in the diﬀerent habitats (ﬁg. 5.4). Populations of larger species are more likely to become extinct because they exist in fewer numbers. However, the particular population that ends up extinct is determined randomly. Once a particular population becomes extinct, its closest competitor in body size

84 Living Components of Biodiversity: Organisms

Figure 5.4 Typical sizes of all populations in the diﬀerent patches (habitats) and in the entire landscape with interspeciﬁc competition and demographic stochasticity. The community structure in each patch is determined by those large populations that escaped extinction. Once a random extinction of a large-body population takes place, the community is well structured according to competitive interactions. Demographic stochasticity increases species diversity at the landscape scale.

beneﬁts from a competitive release and enjoys a higher population size, hence, showing a negative autocorrelation in population size. The rest of the community is now competitively determined by the particular large body-sized species that escaped extinction. Because demographic stochasticity reduces species diversity in each habitat, population size of the survivors, on average, is higher than with interspeciﬁc competition alone; that is, the same resources are now divided among fewer species comprising more individuals. At the landscape scale, more species exist because of the randomness of some extinction in the diﬀerent habitats. Hence, demographic stochasticity increases species diversity at the landscape scale (see also Chesson and Case 1986). Overall, on average, 17:59 1:72 species exist in the landscape. The eﬀect of dispersal with stochastic eﬀects and interspeciﬁc competition Dispersal (e.g., Levin 1974, Johnson et al. 1992) has consistently been shown to have major eﬀects on single-species distributions as well as on multispecies community structures. With dispersal (ﬁg. 5.5), colonists can restore local populations of their species. When the species is competitively subordinate, a permanent recovery is unlikely. However, the recovery of a competitively dominant population has a signiﬁcant eﬀect on community composition. If dispersal is frequent enough, dominant species can establish in all patches and, on average, overcome the stochastic eﬀects that tend to produce locally diﬀerent patterns.

SHALOM: A Landscape Simulation Model 85

Figure 5.5 Typical sizes of all populations in the diﬀerent patches (habitats) and in the entire landscape with interspeciﬁc competition, demographic stochasticity, and dispersal. Dispersal allows dominant populations that became extinct from a particular patch due to a chance event (stochasticity) to recolonize that patch and increase in numbers. As a result, the dominant species in the landscape reestablish their populations in all patches and, on average, overcome the stochastic eﬀects that might locally produce diﬀerent patterns.

Knowing the outcomes (or ﬁngerprints) of the diﬀerent processes (i.e., competition and demographic stochasticity), we can now detect the ﬁngerprints of the diﬀerent processes. As before, demographic stochasticities are responsible for larger discontinuities of body sizes and for the disappearance of the largest species from the landscape (the local extinction of the largest species from all the patches deprives them of colonists that could otherwise restore extinct populations). Dispersal allows dominant species to recolonize habitats in which they have previously become extinct, resulting in homogeneity among habitats in a landscape. At the landscape scale, the existence of dispersal together with demographic stochasticity and interspeciﬁc competition produces the lowest species diversity (5.08 0.598). The main reason for this low species diversity is the ongoing disappearance of small populations that usually belong to species of large body size.

Discussion This model presents a new approach to the study of complex ecological systems. This new approach contributes to our understanding of large-scale ecological processes and patterns by providing us with nontrivial predictions on the combination of spatial heterogeneity and multiple-process interac-

86 Living Components of Biodiversity: Organisms

tions. The example given in this chapter demonstrates this contribution by providing speciﬁc outcomes that could have not been predicted otherwise. Stochasticity depresses mean population sizes and allows diﬀerent habitats to support diﬀerent communities. These diﬀerent communities are determined by which large species becomes locally extinct at random. The local extinction of a large species shifts the maximum body size of the competitively organized community. With both demographic and catastrophic stochasticities, species diversity is higher than with interspeciﬁc competition alone. The eﬀect of demographic stochasticity on species composition diﬀers from that of catastrophic stochasticity. With demographic stochasticity, discontinuities of body sizes are larger, and no species of very similar body size coexist. With catastrophic stochasticity, all of the largest species disappear. Combined, each of the two stochasticities aﬀects species composition in the diﬀerent habitats and in the landscape. Hence, such communities have large discontinuities of body size and none of the largest species. Dispersing individuals move between habitats and reestablish the local populations of their species. Thus, dispersal neutralizes the randomness of the assemblages produced by stochasticity. As a result, each habitat tends toward the same set of species. However, even with dispersal, stochasticity eliminates the largest species and produces large discontinuities in the body size distribution. Loss of randomness in the assemblages means that, at the landscape scale, dispersal reduces species diversity. The predictions presented here about species composition and species diversity demonstrate the usefulness of the current model. The ability to characterize speciﬁc ﬁngerprints of diﬀerent processes and then analyze the joint eﬀect of multiple processes by tracking these ﬁngerprints should help us to understand natural systems better. We recommend that ecologists adopt such an approach for understanding ecological complexity. Ecologists need also to set up studies that will allow them to test whether the predicted patterns produced by particular processes in the model are indeed observed in the ﬁeld.

Note

Send all correspondence to Yaron Ziv.

References Addicott, J.F., Aho, J.M., Antolin, M.F., Padilla, D.K., Richardson, J.S., and Soluk, D.A. 1987. Ecological neighborhoods: scaling environmental patterns. Oikos 49: 340–346. Andow, D.A., Kareiva, P.M., Levin, S.A., and Okubo, A. 1990. Spread of invading organisms. Landscape Ecology 4: 177–188. Begon, M., Harper, J.L., and Townsend, C.R. 1986. Ecology: Individuals, Populations and Communities. Sinauer Associates, Sunderland. Booch, G. 1991. Object-Oriented Design with Applications. Benjamin/Cummings, Redwood City.

SHALOM: A Landscape Simulation Model 87 Calder, W.A. III. 1996. Size, Function, and Life History. Dover, New York. Chesson, P.L., and Case, T.J. 1986. Overview: nonequilibrium community theories: chance, variability, history, and coexistence. In Community Ecology, J. Diamond, and T.J. Case (eds.), pp. 229–239. Harper and Row, New York. Colwell, R.K. and Fuentes, E.R. 1975. Experimental studies of the niche. Annual Review of Ecology and Systematics 6: 281–310. Diamond, J.M. 1984. ‘‘Normal’’ extinctions of isolated populations. In Extinctions, M.H. Nitecki (ed.), pp. 191–246. University of Chicago Press, Chicago. Durrett, R. 1991. Probability: Theory and Examples. Wadsworth and Brooks/Cole Advanced Books and Software, Paciﬁc Grove. Dunning, J.B., Danielson, B.J., and Pulliam, H.R. 1992. Ecological processes that aﬀect populations in complex landscapes. Oikos 65: 169–174. Fisher, R.A., Corbet, A.S., and Williams, C.B. 1943. The relation between the number of species and the number of individuals in a random sample from an animal population. Journal of Animal Ecology 12: 42–58. Forman, R.T.T., and Godron, M. 1986. Landscape Ecology. John Wiley and Sons, New York. Fretwell, S.D. 1972. Populations in a Seasonal Environment. Princeton University Press, Princeton. Fretwell, S.D., and Lucas, H.L.J. 1969. On territorial behavior and other factors inﬂuencing habitat distribution in birds. Acta Biotheoretica 19: 16–36. Gustafson, E.J., and Gardner, R.H. 1996. The eﬀect of landscape heterogeneity on the probability of patch colonization. Ecology 77: 94–107. Holdridge, L.R. 1947. Determination of world plant formations from simple climatic data. Science 105: 367–368. Holt, R.D. 1992. A neglected facet of island biogeography: the role of internal spatial dynamics in area eﬀects. Theoretical Population Biology 41: 354–371. Johnson, A.R., Wiens, J.A., Milne, B.T., and Crist, T.O. 1992. Animal movements and population dynamics in heterogeneous landscapes. Landscape Ecology 7: 63–75. Kotler, B.P., and Brown, J.S. 1988. Environmental heterogeneity and the coexistence of desert rodents. Annual Review of Ecology and Systematics 19: 281–307. Kotliar, N.B., and Wiens, J.A. 1990. Multiple scales of patchiness and patch structure: a hierarchial framework for the study of heterogeneity. Oikos 59: 253–260. Lande, R.L. 1993. Risks of population extinction from demographic and environmental stochasticity and random catastrophes. American Naturalist 142: 911–927. Lawler, S.P., and Morin, P.J. 1993. Temporal overlap, competition, and priority eﬀects in larval anurans. Ecology 74: 174–182. Levin, S.A., 1974. Dispersion and population interactions. American Naturalist 108: 207–228. Levin, S.A., and Paine, R.T. 1974. Disturbance, patch formation, and community structure. Proceedings of the National Academy of Sciences (USA) 71: 2744–2747. Lieth, H., and Whittaker, R.H. 1975. Primary Productivity of the Biosphere. SpringerVerlag, New York. Martin, R.C. 1995. Designing Object-Oriented C++ Applications Using the Booch Method. Prentice Hall, Englewood Cliﬀs. Morris, D.W. 1987. Ecological scale and habitat use. Ecology 68: 362–369. May, R.M. 1981. Theoretical Ecology. Blackwell Scientiﬁc Publications, Oxford. Peters, R.H. 1983. The Ecological Implications of Body Size. Cambridge University Press, Cambridge.

88 Living Components of Biodiversity: Organisms

Pickett, S.T.A., and White, P.S. 1985. The Ecology of Natural Disturbance and Patch Dynamics. Academic Press, New York. Pimm, S.L., Jones, H.L., and Diamond, J. 1988. On the risk of extinction. American Naturalist 132: 757–785. Press, W.H., Teukolsky, S.A., Vetterling, W.T., and Flannery, B.P. 1995. Numerical Recipes in C (second edition). Cambridge University Press, Cambridge. Quinn, J.F., and Robinson, G.R. 1987. The eﬀects of experimental subdivision of ﬂowering plant diversity in a California annual grassland. Journal of Ecology 75: 837–856. Rosenzweig, M.L. 1968. Net primary productivity of terrestrial environments: predictions from climatological data. American Naturalist 102: 67–84. Rosenzweig, M.L. 1991. Habitat selection and population interactions: the search for mechanism. American Naturalist 137: S5–S28. Safriel, U.N., Ayal, Y., Kotler, B.P., Lubin, Y., Olsvig-Whittaker, L., and Pinshow, B. 1989. What’s special about desert ecology? Journal of Arid Environments 17: 125–130. Schmidt-Nielsen, K. 1984. Scaling: Why Is Animal Size So Important? Cambridge University Press, Cambridge. Shaﬀer, M.L. and Samson, F.B. 1985. Population size and extinction: a note on determining critical population sizes. American Naturalist 125: 144–152. Simpson, E.H. 1949. Measurement of diversity. Nature 163: 688. Stroustrup, B. 1995. The C++ Programming Language (second edition). Addison Wesley, Reading, MA. Turner, M.G. 1987. Landscape Heterogeneity and Disturbance. Springer-Verlag, New York. Turner, M.G. 1989. Landscape ecology: the eﬀect of pattern on process. Annual Review of Ecology Systematics 20: 171–197. Turner, M.G., Gardner, R.H., Dale, V.H. and O’Neill, R.V. 1989. Predicting the spread of disturbance across heterogeneous landscapes. Oikos 55: 121–129. Vogel, S. 1994. Life in Moving Fluids: The Physical Biology of Flow (second edition). Princeton University Press, Princeton. Whittaker, R.H. and Levin, S.A. 1977. The role of mosaic phenomena in natural communities. Theoretical Population Biology 12: 117–139. Wiegert, R.G. 1979. Population models: experimental tools for analysis of ecosystems. In Analysis of Ecological Systems, D.J. Horn, G.R. Stairs, and R.D. Mitchell (eds.). Ohio State University Press, Columbus. Wiens, J.A. 1989. Spatial scales in ecology. Functional Ecology 3: 385–397. Wiens, J.A., Stenseth, N.C., Van Horne, B., and Ims, R.A. 1993. Ecological mechanisms and landscape ecology. Oikos 66: 369–380. Ziv, Y. 1998. The eﬀect of habitat heterogeneity on species diversity patterns: a community-level approach using an object-oriented landscape simulation model (SHALOM). Ecological Modelling 111: 135–170. Ziv, Y. 2000. On the scaling of habitat speciﬁcity with body size. Ecology 81: 2932– 2938.

6 Spatial Scale and Species Diversity Building Species–Area Curves from Species Incidence William Edward Kunin Jack J. Lennon

Biodiversity and the Species–Area Relationship This chapter is largely focused on the species–area relationship (SAR), although it may not seem so for much of the time. Bear with us; we will get there in the end. Our aim is to provide insights into how the relationship works, and how it is built. This leads us to take a rather reductionist approach, and to break down the SAR into its component parts. We will spend a substantial section of this chapter examining these pieces and their properties. We will then explore the logic by which the parts are reassembled, and will explore how biological and biogeographical properties of a system may aﬀect the SAR. Before attempting this feat, however, we should begin with a brief discussion of the SAR itself, to explain why it is worth making such a fuss over. The SAR is, after all, only a simple graph: a plot of the number of species found in a sample as a function of the area sampled. Ecologists being an argumentative lot, we cannot even all agree on what this plot should look like; Gleason (1922, see also Williams 1964) argued that the absolute number of species should be plotted as a function of the logarithm of area, whereas Arrhenius (1921, see also Preston 1960) suggested that both species and area should be plotted logarithmically. Connor and McCoy (1979) found cases that ﬁt both models, and two others besides (log species by untransformed area, and neither variable transformed). However it’s plotted, the SAR is not even a particularly attractive or elegant graph—at its best (!) it is simply a straight diagonal line within a tight scatter of datapoints on a rectangular plot. Hardly something to set the pulse racing. 89

90 Living Components of Biodiversity: Organisms

Yet the SAR is exciting stuﬀ; that simple line encapsulates a great deal of information about the diversity of biological systems across a wide range of scales. As Rosenzweig (1995) points out, the SAR brings together local, landscape, and regional measures of ‘‘inventory’’ diversity (a, g, and e diversity, in Whittaker’s 1972 terminology)—extending even to global diversity in some cases; and then unites them with local and regional measures of ‘‘transitional’’ diversity (b and d diversity, respectively) into a single, uniﬁed framework. Hegel would be spinning in his proverbial resting place: here the ‘‘synthesis’’ (the SAR) was developed decades before any of the ‘‘theses’’ or ‘‘antitheses’’ (the various scale-speciﬁc measures of diversity) were devised. We tend to think of the issue of scale-dependence of ecological patterns as a subject of comparatively recent interest (e.g., Tilman and Kareiva 1997), yet here is one of the oldest of ecological tools, devoted speciﬁcally to this most modern of concerns. The problems with the SAR are related to its strengths: it encapsulates so much information in such a simple form that it’s diﬃcult to know how to unpick all of the richness contained within it. How are we to understand the height of the curve, its slope, its curvature (when it is curved) or its straightness (when it’s straight)? How can we explain the tendency for the curve to follow Gleason’s logarithmic model in some cases, and to follow Arrhenius’ power law relationship in others? If we are to understand this tool (we would argue), we need to take it apart into its component pieces. If we can understand these parts and how they ﬁt together, we should gain some insight into how this most fundamental measuring stick for biodiversity works.

Species Distributions and Incidence Curves If the SAR measures species diversity, then its component parts are individual species distributions. As the SAR is devoted to cross-scale patterns, so too these species distributions must be analyzed in a multiscale perspective. To be more precise, the species–area curve can be thought of as the sum of myriad species-speciﬁc ‘‘incidence area’’ curves, that is, the gradual increase in the probability of encountering each species as the area surveyed grows. Each of these curves must rise as we move from ﬁne to coarser scales of analysis, because the probability of sampling any given species will rise as the size of the sampled ‘‘quadrat’’ increases. At the extreme of a vanishingly small (point) quadrat, the area occupied by the species will be its total cover, the fraction of the survey area covered by the species’ tissue. At the opposite extreme, a quadrat the size of the planet will be certain to contain all of the species on earth, and so the probability of each species’ incidence will be 1.0. Between these two extremes, the incidence function for any given species will rise monotonically. At each scale, the probability of including a particular species in a random quadrat will be determined by the species’ ubiquity at that scale, that is, the fraction of all quadrats that size that include

Spatial Scale and Species Diversity 91

the species in question. This fraction is simply the area occupied by the species when analyzed at this scale divided by the total area of the survey. The notion that the area ‘‘occupied’’ by a species diﬀers at diﬀerent scales of analysis may seem a bit peculiar, but it is key to what follows. An essential step toward understanding the scaling properties of diversity is to understand that the quantity of ‘‘stuﬀ’’ measured almost always depends on the scale at which the stuﬀ in question is measured. This can be best understood by means on a well-used analogy: measuring a coastline (see Richardson 1961). It may seem a straightforward matter to measure the length of, for instance, the Israeli coast, but the problem proves quite diﬃcult on closer inspection. The obvious ﬁrst attempt would involve lining up a tape measure on a map of the country, from the end of the Gaza strip to the Lebanese border, some 200 or so kilometers. But the coast is curved, and the tape would not follow it precisely. To get a better estimate, we might push the tape in at the center, making two straight segments, with a sum a bit greater than our ﬁrst estimate. But neither of these segments would quite capture the coast’s shape, so we might wish to divide each of them in half, and then each of the halves yet again, and so on. At each step the apparent length of the coastline would increase. We can measure the ‘‘scale’’ of the analysis by the length of the individual tape segments used. As the scale of our measurement fell to, for instance, 1 km, we would begin to expose the larger bumps in the coastline, with a detour around the Carmel (for example) adding a bit of distance to the total. At a still ﬁner scale (say, 1 m), we would begin to ﬁnd all sorts of zigs and zags in our formerly straight coastline, with detours around rocks and jetties greatly lengthening the overall coast length. Moving to even ﬁner (mm or mm) scales, we might eventually be forced to work our way around each pebble and sand grain on the edge of each beach, vastly expanding the measured distance. The measured length continues to grow as the scale of analysis shrinks. The logic of measuring species distributions across scales is very similar, but here the area grows progressively smaller (rather than larger, as above) as the scale of analysis grows ﬁner. As alluded to above, at extremely ﬁne scale, we would count only areas physically covered by the species in question as ‘‘occupied.’’ At somewhat coarser scales of analysis, we might count any square meter containing at least one individual as ‘‘occupied,’’ but in doing so we would include quite a bit of area that the species was not physically covering. At coarser scales still, all of the square kilometers containing the species might be toted up, again, bringing in a great deal of land that was deemed ‘‘unoccupied’’ at the ﬁner scales of analysis. These diﬀerent scales of analysis are not simply progressively coarser and less accurate ways of measuring distribution; they actually reveal diﬀerent attributes of a species’ spatial distribution. Thus an extremely coarse-scale map (using, e.g., a 100 km grid) would reveal only the broadest outlines of a species’ geographical range, but progressively ﬁner scale maps would begin to reveal geographically isolated subpopulations, habitat preferences, discrete populations, local patterns of aggregation, and so on (Kunin 1998). By plotting the area

92 Living Components of Biodiversity: Organisms

occupied across a range of scales (a ‘‘scale–area’’ curve, Kunin 1998), a variety of aspects of abundance can be displayed simultaneously. Note that this scale–area curve is almost identical to the incidence function we require for building the SAR; we need only divide the area occupied by the total area surveyed. To understand the SAR, then, we need to learn more about these scale–area curves. There are reasons to believe that, for many species, these curves should be approximately straight when plotted on logarithmic axes (Kunin 1998). In the language of geometry, this is to say that these distributions are approximately fractal, at least over a reasonably wide range of scales. For the purposes of our argument, it would be suﬃcient to state this fact and to build oﬀ it, but it is ﬁtting (given the mechanistic tone of this chapter) to propose a mechanism for these approximately fractal species distributions before proceeding further. Fractal Habitats and Species Distributions The ﬁrst step toward understanding the distribution of a species is to map the area of potentially suitable habitat available to it. In principle, at least, one could map out all areas that fall within the fundamental niche (sensu Hutchinson 1957) of a given species. This ‘‘niche map’’ would include all sites at which a pioneering party of immigrants of the species in question would have a positive population growth rate ( > 1). A wide variety of variables might contribute to determining the suitability of a site to a given species, but even without knowing precisely what they are, we can guess something about their distribution in space. We might expect this map to be approximately fractal. Geometric analyses of various aspects of the physical environment, for instance, coastlines (Richardson 1961), rock substrates (Kulatilake et al. 1997), lava ﬂows (Gaonach et al. 1992), topography (Klinkenburg and Goodchild 1992), drainage (Deschaux and Souriau 1996), soil moisture (Pelletier et al. 1997), and rainfall patterns (Onof and Wheater 1996) typically display some form of fractal scaling, at least over certain ranges of scales. Disturbance events such as earthquakes (Hastings and Sugihara 1993), landslides (Pelletier et al. 1997), erosion (Chase 1992), ﬁres (McAlpine and Wotton 1993), and some human activities (Frohn et al. 1996) show evidence of fractal properties as well. As the suitability of a site for a given species depends at least in part on these and other physical factors, we might expect the distribution of habitat acceptable to a given species (i.e., all areas for which its > 1), to be approximately fractal as well. This prospect seems particularly appropriate in arid lands, where approximately fractal drainage and erosion patterns are often starkly evident (without the concealing mantle of vegetation), leading to similarly patchy and complex patterns of habitat availability for many species. The physical niche map of a species will generally be reshaped by biotic factors as well (chapter 11, this volume). Interactions with other species (dominant competitors, natural enemies, mutualists) could alter the map considerably, but so long as these other species also have roughly fractal distributions, the map of habitat

Spatial Scale and Species Diversity 93

remaining available to our focal species is simply the intersection of two fractals, and thus likely to be approximately fractal as well (but see below). Of course, even if we could map all habitable land for a particular species, this would only be the ﬁrst step—the playing board upon which the species’ career is acted out (to mix metaphors badly). The fraction of that map that is actually occupied by a species will be determined in large part by a process of colonization and local extinction within patches, akin to a metapopulation model (Levins 1970, Hanski and Gilpin 1997), but acting on multiple scales. Areas that are particularly isolated (with consequent low rates of colonization), or that are small or poor in quality (both leading to high extinction rates) may be probabilistically unoccupied, even if suitable. The same is true of the interacting species that inﬂuence our focal species’ available niche map—they may be unable to colonize some sites that would be suitable for them, thus perhaps allowing our focal species to persist in some sites from which it would otherwise be excluded. On the other hand, high dispersal of propagules into sites that are marginally unsuitable for our focal species may result in the maintenance of ‘‘sink’’ populations outside of the species original niche map (Pulliam 1988). These processes, taken together, will reshape the original map, removing potentially large areas of unrealized niche space, while perhaps adding a bit of populated area in ‘‘non-niche’’ sites (ﬁg. 6.1). These deletions and additions may aﬀect the species distribution diﬀerently at diﬀerent scales of analysis. If the original niche map can be thought of as approximately fractal, it can be represented as a line on a scale–area graph. This is done by dividing the region into cells at diﬀerent resolutions, and counting up the area occupied, a process which is conceptually akin to the computation of the so-called boxcounting fractal dimension (Db , see, e.g., Hastings and Sugihara 1993, Virkkala 1991, Kunin 1998), one of the simplest forms of fractal analysis. Some species may have highly aggregated clumps of habitat available to them (represented by a relatively ﬂat scale–area curve for their niche maps and high values of Db ), while others occupy a sparsely scattered niche-scape (resulting in a steeply sloped scale–area relationship and a low Db value). The processes of colonization and local extinction will act most strongly at diﬀerent characteristic ranges of scales. Large blocks of habitat may be removed from the map at very coarse scales due to biogeographic scale barriers to dispersal, oceans or mountain ranges cutting oﬀ entire continents of potential niche space. At ﬁner scales of analysis, rather smaller isolation distances are involved, and patches below some critical size or quality threshold may begin dropping oﬀ the list. The tradeoﬀ between isolation and size often found in metapopulation studies (e.g., Thomas and Jones 1993) means that progressively smaller chunks of habitat are eliminated as one goes to ﬁner scales of analysis. Where ‘‘sink’’ populations extend the map into inhospitable terrain, however, they can only do so in sites very close to existing sources of propagules, and so the area they add to the map should generally act on quite ﬁne spatial scales. Taken together, these processes are likely to result in a somewhat ﬂatter (higher Db ) curve for actual species distributions across

94 Living Components of Biodiversity: Organisms

Figure 6.1 Maps (a, b) and scale–area curves (c) for a hypothetical species’ fundamental niche space (a, solid line in c) and realized populations (b, oﬀ line in c). Note that colonization and extinction processes remove larger areas at coarse scales than at ﬁne, while adding area (‘‘sink’’ populations, denoted by gray shading) at relatively ﬁne scales only.

scales than for the corresponding niche map, although the degree of ﬂattening should depend on dispersal ability (ﬁg. 6.1). As noted above, there is an intimate connection between an individual species’ distribution, as represented by a scale–area curve, and the species– area curve commonly used in the ecological literature. If we examine a ﬁnite region, and the area within that region occupied by a particular species rises

Spatial Scale and Species Diversity 95

in some fashion as we move from ﬁne to coarse scales of analysis (as it must), then the fraction of all cells occupied by the species must also rise in a related manner. These fractions can be interpreted as probabilities; a species occupying half of the 10 km cells within a survey region will have a 50% chance of being present in any one arbitrarily chosen cell at that scale. The species–area curve is the sum of such probabilities across species, and so should reﬂect the shape of the scale area curves that make it up. If scale–area curves are approximately linear on logarithmic axes (that is, if distributions are fractal), then it seems reasonable to expect the species area curve to be linear when plotted on logarithmic axes as well (Harte et al. 1999). Thus we may have a mechanistic basis for the power law formulation of the SAR, originally proposed by Arrhenius (1921) and widely accepted in recent decades (e.g. Rosenzweig 1995). Three Opposing Twists Unfortunately, things are not quite that simple. At least three factors can aﬀect the shape of species–area curves, even when all of the species involved are fractally distributed. The ﬁrst concerns variation in the fractal properties of diﬀerent species. So long as all of the species concerned have distributions with the same fractal dimension, the species–area curve they produce will indeed be linear when plotted on logarithmic axes. However, where species diﬀer in their fractal scaling, as many species do, the species–area curve ceases to be linear, growing progressively steeper at increasingly coarse scales (ﬁg. 6.2; see appendix to this chapter for a more mathematical approach). This is counterintuitive for those of us used to working in arithmetic (rather than logarithmic) space; we generally suppose that the sum of a group of straight lines ought to be a straight line itself. Not so in logarithmic space, however: the log of a sum is not equal to the sum of the logs. If two scale–area curves diﬀer in slope, the multiplicative diﬀerence between them changes across scales, and so the degree to which the higher of the two dominates their sum also changes. This can be seen most clearly where scale–area curves of species cross (ﬁg. 6.2). Whichever species has the highest incidence at a given scale has a disproportionate inﬂuence on this sum. Where scale–area curves approach one another and eventually cross, this lead role vanishes and is ultimately replaced by a diﬀerent ‘‘leader,’’ one that, of necessity, will have a steeper scale–area curve than the species it supplanted. The resulting shift toward steeper slopes at coarse scales is certainly not what the extensive literature on species–area curves would lead us to believe occurs in nature. The long-standing dispute between Arrhenius’s power function SAR (1921) and the semilog version championed by Gleason (1922) suggests that the SAR should either be linear or decelerating when plotted on two logarithmic axes; we generally don’t expect to ﬁnd the sort of accelerating function suggested by the argument above! A second eﬀect may counteract the ﬁrst one. One problem with box counting as a way to measure fractals is that they can become ‘‘saturated’’; at

96 Living Components of Biodiversity: Organisms

Figure 6.2 Eﬀects of variation between species in scale–area curves on the properties of the resulting species–area relationship. The scale–area curves illustrated in (a) diﬀer in slope. The summed probability of the two species occurring in a sample (b) does not rise linearly with scale, but curves slightly upward. This can be seen more clearly by plotting the slope of the curve (c), here expressed as the ratio of probabilities between successive scales. Had parallel scale– area curves (i.e., equal fractal dimensions) been used, the summed probability of occurrence would have remained linear.

coarse scales of analysis, common species come to occupy all of the available cells (ﬁg. 6.3). Once they reach this saturated state, such species necessarily cease to increase the area they occupy at still coarser scales, as they already occupy the entire area under consideration. This should result in an abrupt

Spatial Scale and Species Diversity 97

Figure 6.3 Example of map saturation. Two ﬁne-scale grids of similar fractal dimension but diﬀerent abundance are shown on the left. The commoner case (lower row) quickly comes to occupy nearly the entire arena at coarser scales of analysis (center two panels), causing the scale–area curve (on the right) to decelerate.

shift downward in the slope of their scale–area curves. In practice, however, the eﬀect is often much more gradual, with diﬀerent parts of a species’ distribution saturating progressively across a range of scales. This curvature in scale–abundance curves should cause a similar curvature in the SAR as well— resulting in a decelerating species–area curve when plotted on logarithmic axes, consistent with Gleason’s viewpoint. Depending on its relative strength, this eﬀect could partially or wholly counteract the eﬀect described above, perhaps even reversing it. Before leaving the subject, there is a third process that may inﬂuence the shape of species–area curves away from perfect log-linearity: the single-cell anomaly. Real populations (unlike idealized fractals) have ﬁnite distributions. At a suﬃciently coarse scale, the entire sampled distribution of the species will be enclosed in a single cell. At all scales coarser than that, the species will behave as if it has a fractal dimension of zero, causing its scale–area curve to rise very steeply (ﬁg. 6.4). As this phenomenon only takes hold at scales coarser than a species’ regional distribution, we had not anticipated that it would be very important in most SAR analyses. However, it should be noted that rare species make up a large fraction of most community samples, and this process is likely to act most strongly on such species (see below). Thus we are left with a surprisingly complex story to explain a straight line. Even if we begin with approximately fractal species distributions—which are themselves supposed to be linear on such axes—we still must balance oﬀ at least three competing forces, all of which tend to bend these simple lines. To make matters worse, two of the processes (adding dissimilar fractals and the single-cell anomaly) tend to bend the SAR upward, instead of the downward

98 Living Components of Biodiversity: Organisms

Figure 6.4 The single-cell anomaly. Two ﬁne-scale distributions with similar fractal dimensions, but diﬀering in abundance are shown on the left. The rarer site (lower row) has only a single occupied cell at the intermediate scale of analysis. This results in an abruptly steeper scale–area curve at coarser scales (shown on the right).

bend we might expect to ﬁnd from the empirical literature. Where the SAR follows the simple log-linearity of Arrhenius’ power function, it could either be because none of these three forces acts with any signiﬁcant force (leaving us with the simple addition of similar fractals posited by Harte et al. 1999), or else this apparent simplicity must hide a complex core, with the combined eﬀect of upward and downward curvature somehow balancing one another out. Alternatively, if the SAR follows Gleason’s decelerating SAR, it either means that the underlying distributions are not fractal or, if they are, that the impact of map saturation signiﬁcantly outweighs the combined eﬀects of the other two forces.

Testing the Ideas: A Fractal Transect of the Northern Negev To test these ideas, we need a handy SAR to dissect into its component species distributions. As luck would have it, we have just such ﬁeld data to hand. One of us (WEK) had earlier carried out a transect survey of the crucifers of the northern Negev (spanning the range of sites visited at the workshop from which this book originates, see preface for details), and the data are well suited to exploring some of the ideas discussed above. Before proceeding further, a bit of explanation about the transect itself is in order. The data were collected in an attempt to explore the scaling properties of species’ distributions, and in particular the declines in abundance as one approaches the margins of a species’ range. It is widely appreciated that species tend to grow rarer as they approach their range margins (e.g., Brown

Spatial Scale and Species Diversity 99

1984), indeed in the absence of some abrupt barrier such as a coastline, it would be surprising if they did not (see Lennon et al. 1997). But the nature of that decline is not well known; does the local density of a species’ populations decline, or is there simply a decline in the fraction of the landscape that it can occupy? A similar question can be posed in the language of scale–area curves: Does the slope of the curve shift, or just its elevation? To rephrase this yet again into something a biologist might recognize: Does a species remain equally patchy as it becomes rarer, or does its distribution become clumpier or (alternatively) more evenly scattered as it approaches its range margins? To address issues of this sort, one needs distributional data collected across a wide geographical range, but with suﬃcient local intensity to pick up ﬁnescale aggregation patterns. Given a ﬁnite amount of time and labor, these two factors (the spatial extent of the transect and local intensity of sampling) are generally inversely correlated; a ﬁxed total number of quadrats dispersed along a growing transect results in samples scattered ever more widely apart. Again, fractal geometry may come to our assistance, providing an eﬃcient means of collecting samples of information across a wide range of spatial scales. If species distributions are approximately fractal, as we have suggested above, then a fractal transect can be an eﬃcient sampling tool; its overlap with a species’ distribution is simply the intersection of two fractals, a quantity which is well understood in theory: Dint ¼ Dspecies þ Dsample 2

ð1Þ

where Dspecies is the fractal dimension of the species’ distribution, Dsample is the fractal dimension of the transect used to study it, and Dint is the fractal dimension of the intersection between the two—the population captured within the sample (Mandelbrot 1982, J. Halley pers. comm.). Thus, for the simplest case of a linear transect (Dsample ¼ 1) across a species’ range, the resulting data should reveal a fractal dimension of Dint ¼ Dspecies þ 1 2 ¼ Dspecies 1

ð2Þ

Here one need only add 1 to the fractal dimension revealed by the transect to estimate the true fractal dimension of the species’ population. Note that this constrains the range of fractals that can be studied; fractal dimensions cannot be negative, so species exhibiting fractal dimensions less than 1 will not be amenable to study in this case. The same principle should apply to any form of fractal sampling; the application of equation (1) should allow the true fractal dimension to be calculated from that revealed in the sample. The Survey To study the crucifers of the Northern Negev, a variant of the well-known Cantor set (see, e.g., Hastings and Sugihara 1993) was used as a transect (ﬁg. 6.5). The basic unit was a 0:5 0:5 m quadrat. Four of these quadrats were arranged in a line at 1 m spacings to form a sample 5 m in total length. Four sets of such samples were then arranged at 10 m spacings to form a local

100

Living Components of Biodiversity: Organisms

Figure 6.5 Schematic representation of the sampling design used in the Northern Negev transect. Details are given in the text.

survey 50 m in length. Similar surveys were performed at 100 m intervals, covering a total distance of 500 m. In a similar manner, four such sections spaced 1 km apart were surveyed within a 5 km region. A total of ﬁve such regions were examined at 10 km spacings, creating a total transect 65 km in length, stretching from the edge of the Zin canyon (just south of Sede Boqer) northwards to Devira, some 16 km north of Beer Sheva. For most of its length, the transect used a gas pipeline maintenance road as an access route, which resulted in a slight west-to-east shift between survey areas; transect segments, however were oriented directly north–south. Note that the coarsest scale (with ﬁve rather than four sample areas) is a slight variation from the Cantor set pattern; at all other scales each tenfold increase in transect length is accompanied by only a fourfold increase in the area surveyed, giving a fractal dimension of logð4Þ= logð10Þ ¼ 0:602

ð3Þ

Consequently, a sample of this sort can only be used on populations with fairly high fractal dimensions; applying equation (1) suggests that only distributions with D > 1:4 should provide meaningful positive dimension estimates. The transect was designed to study population shifts across range margins, and so it was placed across a rather steep environmental gradient to improve the chances of encountering such boundaries. Over the length of the transect,

Spatial Scale and Species Diversity

101

there is a dramatic shift in water availability, with mean annual rainfall shifting from approximately 100 mm/yr in the south to about 300 mm/yr in the north. This results in a pronounced shift in vegetation from desert to nearly Mediterranean, with a corresponding turnover in species composition. Relatively uniform substrates were sampled, however, with the gaps in the transect so positioned as to skip over the bands of rock outcrops north of Sede Boqer, so that the survey included almost exclusively loess soils. The clusters of samples at diﬀerent scales can be used to devise species– area curve for the transect, as well as to construct incidence curves for each of the species sampled. This is done by considering individual quadrats (0.25 m2 in area), sets of four quadrats (1 m2 total area), sets of four such groups (4 m2), and so on. Note that this does not quite ﬁt Rosenzweig’s (1995) criterion, that SARs ideally should be composed of contiguous samples, but at least the groupings are composed of closest neighbors within the sample. While the resulting compression of space may exaggerate the rate of species turnover, the bias should act uniformly across scales, allowing a rather unnaturally steep but otherwise realistic SAR to be estimated. It should certainly serve adequately for demonstration purposes.

Transect Results Despite the rather unconventional nature of the sampling regime used here, the results appear reassuringly familiar: taken as a whole, the dataset displays a classic power law (Arrhenius) SAR (ﬁg. 6.6), although there appears to be a bit of sigmoid curvature around the line of best ﬁt. A total of 19 species were encountered in the survey, although more than half (10 species) were encountered in fewer than 10 of the 1280 quadrats sampled.

Figure 6.6 The species–area relationship found in the Northern Negev transect samples. The relationship is well represented by a line in log–log space, implying a power law relationship. Note, however, that there is a noticeable curvature to the relationship, with generally steeper slopes at the ﬁnest and coarsest scales, and somewhat shallower slopes at intermediate scales.

102

Living Components of Biodiversity: Organisms

Examining individual species–area distributions, it quickly becomes clear that all three of the problems discussed above act within the sampled biota. First of all, it is apparent that the range of fractal properties exhibited by these species varies considerably, with estimated slopes of scale–area curves ranging from 0.135 to 0.422. This is easily suﬃcient to trigger the ﬁrst of our biases: the accelerating curve produced when dissimilar fractals are added together. Our second bias is also apparent; many of the graphs show noticeable deceleration, presumably due to saturation. This tendency is most notable among the commonest species (which approach saturation most quickly), but the trend is visible even in species that are not particularly common. Finally, several of the rarer species show a distinct upturn at coarse scales, which is often clearly attributable to the single-cell anomaly. However, the very rarest of these species were encountered only in a single quadrat each, providing a strictly linear, but unusually steep, scale–area curve. This is simply the single cell anomaly taken to its extreme, creating an artifactual estimate of Db ¼ 0. For some of these species, the problem may be exacerbated by the sampling technique; these rare scattered species are precisely the sort for which meaningful fractal estimates may be diﬃcult with a lowdimension sample. Until now, we have treated the transect as a single sample, 65 km long, crossing a wide diversity of conditions. However, for our purposes it may be better to consider it as ﬁve shorter (5 km long) transects, each in a diﬀerent environment. We can then plot the SAR for each (ﬁg. 6.7). Now, rather than a single SAR, we have ﬁve diﬀerent curves, which conveniently exhibit a range of behaviors, stretching from pure power law dynamics (Sites I and V) through intermediate levels of curvature (Site II) to Gleasonian logarithmic curves (Sites III and IV). This puts us in a position to try to dissect what aspects of the sites and their individual species distributions contribute to these diﬀering patterns.

Figure 6.7 Species–area relationships for the ﬁve component segments of the northern Negev transect (a) plotted on log–log axes (after Arrhenius 1921); (b) plotted on semilog axes (after Gleason 1922).

Spatial Scale and Species Diversity

103

We can now examine the various sources of curvature, to see which is most likely to be responsible for the diﬀerences observed. All ﬁve sites contain species with markedly diﬀerent fractal properties, suggesting that the dissimilar fractals eﬀect should apply to all of them. However, the variance in slopes is fairly similar in all ﬁve sites, and is in any case uncorrelated with SAR properties, with the two power law sites, I and V, having the highest and lowest variances of the group, respectively. Consequently, we cannot rely on this factor to explain the observed diﬀerences in curve shapes between sites. There is also relatively little diﬀerence between the sites in the degree of ‘‘fractalness’’ of their component species distributions. In the absence of a widely agreed measure for such a property, we will use a rough but convenient measure: the departure of each species’ scale–area curve from the bestﬁt line in log–log space, as measured by the R2 value (table 6.1). While there are some diﬀerence between species and sites in this respect, the overall difference between sites is surprisingly small. Certainly we cannot distinguish the power law sites (I and V, with a mean R2 value of 0.972) from the logarithmic sites (III and IV, with a mean R2 value of 0.971) on this basis. There is a much more striking diﬀerence between sites, however, in the nature of this departure from pure fractal linearity. In the power law sites, there is a mixture of upward and downward (and sigmoid) curvature of scale–area curves, whereas the more Gleasonian sites display almost exclusively downward curvature. This, in turn, reﬂects a diﬀerence in the species–abundance mixtures in the various sites; the straight power law sites had a much greater share of rare species sampled, while the curved (logarithmic function) sites were dominated by a few relatively common species.

General Conclusions It has become apparent, to us at least, that the species–area relationship is a rather more complex beast than we have usually given it credit for. In its most widely accepted (power law) formulation, it is built upon a tacitly fractal perspective on species distributions, a perspective that has gained popularity long after the original model was proposed. Yet, in a way, the departures from pure fractal scaling are as interesting as the fractal pattern itself, and these may hold the key to understanding the diversity of SARs documented in the literature. We have identiﬁed at least three such distorting processes, and have demonstrated through a case study how they may combine to determine the overall shape of the curve. What appears to be simple linearity is more akin to a ceaseﬁre line: a balance of opposing forces that are attempting to bend the curve in diﬀerent characteristic ways. In the example we examined above, the observed diﬀerences in SAR shapes seem to be caused by a shifting balance between what we termed ‘‘map saturation’’ and ‘‘single cell anomalies.’’ These, in turn, may reﬂect shifts in the mix of species, with the former process most aﬀecting common and widely scattered populations, while the latter is characteristic of rare and tightly clustered distributions (ﬁg. 6.8).

104

Living Components of Biodiversity: Organisms

Table 6.1 Summary of scale–area curve properties for species in the Northern Negev transect data. Values are given as R 2 values of linear relationships between log species and log area occupied R2 of correlations between log scale and log area occupied Site I Power law Torularia

0.986 (

Reboudia

0.980 0.973 (

Carrichtera

Site II Intermediate

Site III Exponential

Site IV Exponential

0.962 ( 0.886 ( 0.921 (

0.988 ( 0.870 (

0.980 ( 0.914 (

1*

0.997

0.991 (

0.944 (

0.999 (

0.992

0.990 )

Matthiola

0.946 ( 0.973 (

Enarthrocarpus Biscutella

0.971 ) 0.937 (

0.947 (

Leptoleum

1* 0.992

Erucaria Nasturtiopsis Diplotaxis harra

Site V Power law

0.985 0.966 ( 0.999 (

0.933 ( 0.980 (

D. erucoides

0.963

Hirschfeldia

0.954

Others:

1* 0.939 )

0.997 ( 1* 1*

0.980 0.985

Mean

0.965

0.958

0.964

0.978

0.979

Mean ignoring singletons

0.962

0.953

0.964

0.974

0.972

Summed departure from linearity Decelerating (( ) Mixed () Accelerating ( ) )

P ð1 R2 )

0.272 (71%)

0.284 (86%)

0.143 (98%)

0.115 (88%)

0.094 (49%)

0.020 (5%)

0.035 (11%)

0.003 (2%)

0.016 (12%)

0.098 (51%)

0.090 (24%)

0.010 (3%)

0

0

0

): accelerating curve; (: decelerating curve; : curve with different directions of curvature over different ranges of scales. *Site with only one occupied quadrat of the species indicated. Such species necessarily had perfectly linear scale–area plots, with R2 therefore equal to 1.

There are, of course, limits to how much we can prove on the basis of one transect sample, whether viewed as a whole or decomposed into its ﬁve component sites. However, if the patterns uncovered above generalize to other SARs, there seem to be clear indications that the shape of such curves can reveal interesting and rather subtle bits of information about the communities of organisms they represent. Areas dominated by a few ubiquitous species, showing saturating distributional dynamics, should display tendencies toward Gleasonian SARs, whereas sites with large numbers of rare and restricted species might be expected to produce curves more akin to Arrhenius’ original power law model. Clearly, this is only a ﬁrst step, but it is encouraging to ﬁnd

Spatial Scale and Species Diversity

105

Figure 6.8 Eﬀect of distributional properties on two processes that aﬀect SAR shape. Rare and/or highly aggregated species are particularly prone to single-cell anomalies, as their populations are readily enclosed within a single occupied cell. This creates a steep scale–area curve at coarse scales, biasing estimates of Db downward (as indicated by the arrow). Very common or widely scattered populations, on the other hand, tend to result in map saturation, causing a ﬂattening of scale–area curves and an upward bias in Db estimates (again, indicated by an arrow).

that SARs can be successfully dissected, and that potentially useful information can be found from among their entrails. The SAR may be among the oldest tools in the study of biodiversity, but it looks set to begin a second career.

Acknowledgments The authors would like to thank the Israeli hosts of the 1999 workshop (see preface for details) for their hospitality and encouragement, both at the workshop and subsequently. WEK would also like to thank the British Royal Society for supporting the ﬁeldwork described in the latter half of the chapter (twice!). JJL is supported by a Co-operative Joint Venture Agreement with the U.S.D.A. Forest Service.

Appendix: On the Sum of Scale–Area Curves The following sections provide the mathematical foundations for a few points made verbally in the text, and underline the foundations of a point not now included in the text but which was included in the original oral presentation of this chapter. The SAR for Fractal Species Distributions Let the number of occupied quadrats n at a particular scale s be given by n ¼ n0 ð1=sÞD

106

Living Components of Biodiversity: Organisms

where n0 is a constant (which may be interpreted as the probability that a unit quadrat, of s = 1, is occupied), and D is the box-counting fractal dimension. The probability that a randomly selected quadrat at a particular scale is occupied is the ratio of the number of squares occupied at this scale divided by the total number of squares at this scale. This latter is simply (1/s)2, so the probability is p ¼ nð1=sÞ2 ¼ n0 ð1=sÞD2 ¼ n0 s2D

If species are distributed independently, then the expected number of species in a quadrat of scale s is the sum of the probabilities of the individual species: X RðsÞ ¼ pi Which on expansion of p gives RðsÞ ¼

X

n0i s2Di

This is the fractal SAR. It is not linear in log–log axes unless all Di s are equal, since diﬀerent species make diﬀerent contributions at diﬀerent scales. To demonstrate this, we explicitly solve for the slope of this SAR.

The Slope of the SAR Instantaneous Scale Change Partial diﬀerentiation with respect to s gives the slope of the curve at each s: X R=s ¼ ð2 Di Þn0i s1Di It can be seen from this that the contribution of a particular species to the overall rate of accumulation of species with area depends on two main factors: (i) the abundance of the species at a reference scale (n0) and (ii) its sensitivity to scale, 1 D. A species with a relatively large n0 but with a small D 1 may make a relatively large contribution at coarse scales but a relatively small contribution at ﬁne scales.

Discrete Scale Change If we consider a discrete change in scale, then the change in richness with scale is given by X 2Di R ¼ ai s ; ai ¼ n0i ð1 1=kÞ2Di Where k is the change in linear scale (e.g., k ¼ 10 means a 100-fold change in quadrat area). This equation is very similar to the continuous scale change version.

Spatial Scale and Species Diversity

107

The SAR and Correlated Species Incidence One ﬁnal issue we considered is the degree of spatial correlation between species incidence. In an earlier draft of this chapter (the one presented orally at the workshop), WEK asserted that the scaling patterns of species correlation might aﬀect the SAR in interesting ways, especially in arid lands, where complex patterns of positive and negative correlations at diﬀerent scales might be expected. We tested these ideas by simulating species distributions, with diﬀerent degrees of spatial correlation (from completely independent distributions to strong positive and negative spatial correlations). Spectral synthesis simulation of intercorrelated species distribution patterns (with ﬁxed D ¼ 1:5) gave a ﬁxed SAR, independent of the level of correlation! It appears not to matter whether or not species tend to occur together—if we accumulated species from a randomly chosen point in the landscape, then the same average SAR curve occurs even if species are perfectly correlated, that is, distributed identically.

References Arrhenius, O. 1921. Species and area. Journal of Ecology 9: 95–99. Brown, J.H. 1984. On the relationship between the abundance and distribution of species. The American Naturalist 124: 255–279. Chase, C.G. 1992. Fluvial landsculpting and the fractal dimension of topography. Geomorphology 5: 39–57. Connor, E.F., and E.D. McCoy. 1979. The statistics and biology of the species-area relationship. The American Naturalist 113: 791–833 Deschaux, V., and M. Souriau. 1996. Topography of large-scale watersheds—fractal texture and global drift—application to the Mississippi basin. Earth and Planetary Science Letters 143: 257–267. Frohn, R.C., K.C. McGwire, V.H. Dale, and J.E. Estes. 1996. Using satellite remotesensing analysis to evaluate a socioeconomic and ecological model of deforestation in Rondonia, Brazil. International Journal of Remote Sensing 17: 3233–3255. Gaonach, H., S. Lovejoy, and J. Stix. 1992. Scale invariance of basaltic lava ﬂows and their fractal dimensions. Geophysical Research Letters 19: 785–788. Gleason, H.A. 1922. On the relation between species and area. Ecology 3: 158–162 Hanski, I.A., and M.E. Gilpin. 1997. Metapopulation Biology. Academic Press, San Diego. Harte, J., Kinzig, A., and Green, J. 1999. Self-similarity and the distribution and abundance of species. Science 284: 334–336. Hastings, H.M., and G Sugihara. 1993. Fractals: a User’s Guide for the Natural Sciences. Oxford University Press, Oxford. Hutchinson, G.E. 1957. Concluding remarks. Cold Spring Harbor Symposium on Quantitative Biology 22: 415–427. Klinkenburg, B., and M.F. Goodchild. 1992. The fractal properties of topography—a comparison of methods. Earth Surface Processes and Landforms 17: 217–234. Kulatilake, P.H.S.W., R Fiedler, and B.B. Panda. 1997. Box fractal dimension as a measure of statistical homogeneity of jointed rock masses. Engineering Geology 48: 217–229.

108

Living Components of Biodiversity: Organisms

Kunin, W.E. 1998. Extrapolating species abundance across spatial scales. Science 281: 1513–1515. Lennon, J.J., J.R.G. Turner, and D. Connell. 1997 A metapopulation model of species boundaries. Oikos 78: 486–502. Levins, R. 1970. Extinction. In M. Gerstenharber (ed.), Some Mathematical Questions in Biology. American Mathematical Society 2: 77–107. Mandelbrot, B.B. 1982. The Fractal Geometry of Nature. W.H. Freeman, New York. McAlpine, R.S., and B.M. Wotton. 1993. The use of fractal dimension to improve wildland ﬁre perimeter predictions. Canadian Journal of Forest Research (Journal Canadien de la Recherche Forestiere) 23:1073–1077. Onof, C., and H.S. Wheater. 1996. Analysis of the spatial coverage of British rainfall ﬁelds. Journal of Hydrology 176: 97–113. Pelletier, J.D., B.D. Malamud, T. Blodgett, and D.L. Turcotte. 1997. Scale-invariance of soil moisture variability and its implications for the frequency-size distribution of landslides. Engineering Geology 48: 255–268. Preston, F.W. 1960. Time and space and the variation of species. Ecology 41: 611–627. Pulliam, H.R. 1988. Sources, sinks, and population regulation. The American Naturalist 132: 652–661. Richardson, L.F. 1961. The problem of contiguity: an appendix of statistics of deadly quarrels. General Systems Yearbook 6: 139–187. Rosenzweig, M.L. 1995. Species Diversity in Space and Time. Cambridge University Press, Cambridge. Thomas, C.D., and T.M. Jones. 1993. Partial recovery of a skipper butterﬂy (Hesperia comma) from population refuges: lessons for conservation in a fragmented landscape. Journal of Animal Ecology 62: 472–481. Tilman, D., and P. Karieva (eds.) 1997. Spatial Ecology: The Role of Space in Population Dynamics and Interspeciﬁc Interactions. Princeton University Press, Princeton. Virkkala, R. 1991. Spatial and temporal variation in bird communities and populations in north-boreal coniferous forests: a multi-scale approach. Oikos 62: 59–66. Whittakker, R.H. 1972. Evolution and measurement of species diversity. Taxon 21: 213–251. Williams, C.B. 1964. Patterns and the Balance of Nature. Academic Press, New York.

7 Microbial Contributions to Biodiversity in Deserts Peter M. Groffman Eli Zaady Moshe Shachak

T

he number of species living in the soil may well represent the largest reservoir of biodiversity on earth (Giller 1996, Wardle and Giller 1996, Service 1997). Five thousand microbial species have been described and identiﬁed (Amann and Kuhl 1998), but the actual number of species may be greater than 1 million (American Society for Microbiology 1994), larger even than the number of insect species (Service 1997). Over the last 10 to 15 years, interest in soil biodiversity has soared, driven by advances in molecular techniques that allow for identiﬁcation and analysis of soil microbes, many of which are diﬃcult to extract and culture (Kennedy and Gewin 1997). However, the factors that control soil microbial biodiversity and the links between soil biodiversity and ecosystem function are still unclear (Beare et al. 1995, Schimel 1995, Freckman et al. 1997, Brussard et al. 1997, Wall and Moore 1999). Soil may represent an excellent venue for exploring links between biodiversity and ecosystem function. The vast numbers of species in soil and methodological problems have long necessitated a functional approach in soil studies. As a result, soil functions important to organic matter degradation, nutrient cycling, water quality, and air chemistry are well studied (Groﬀman and Bohlen 1999). As our knowledge of soil biodiversity increases, this information may provide a strong basis for evaluating links between biodiversity and these functions. Evaluating functional diversity of soil communities requires considering how microbes interact with plants and soil fauna to produce patterns of ecosystem processes (Wall and Moore 1999). These interactions vary within and between ecosystems (i.e., across landscapes). Throughout this book, we 109

110

Living Components of Biodiversity: Organisms

suggest that the science of biodiversity must consider links to ecosystem processes and interactions with landscape diversity (Shachak et al. this volume). The need for these links is particularly clear when considering soil biodiversity. There have been relatively few studies of microbial processes in desert soils, and very little analysis of desert soil biodiversity (Parker et al. 1984, Schlesinger et al. 1987, Peterjohn 1991, Fließbach et al. 1994, Zaady et al. 1996a,b, Steinberger et al. 1999). Given this lack of data, we cannot provide a comprehensive review of desert soil biodiversity, with comparison to more mesic ecosytems. Rather, we can review some of the unique features of desert microbial ecology and pose questions about desert soil biodiversity that should be addressed in future research. Our analysis, which is be framed in the context of ‘‘ecosystem processes and landscape biodiversity,’’ which is the theme of this book (Shachak et al. this volume), addresses the importance of (1) wetting events, (2) ecosystem diversity, and (3) landscape diversity as controllers of microbial biomass, activity, and biodiversity in desert soils. We address these features and questions primarily with data that we have collected in the Negev desert in Israel, but focus on results that are relevant to other deserts around the world.

The Importance of Wetting Events as Controllers of Microbial Biomass, Activity and Biodiversity in Desert Soils Many studies have found that much of the microbial activity that occurs in desert soils is concentrated during brief periods of high activity following wetting events (Parker et al. 1984, Schlesinger et al. 1987, Peterjohn 1991, Fließbach et al. 1994, Zaady et al. 1996a,b). During these periods, soils are warm and wet, that is, nearly optimal for microbial activity. Given the short generation times, dispersal possibilities, and survival capacities of microbes, we hypothesize that these brief periods of optimal conditions should allow for the development of diverse microbial communities. In one of the few studies of desert soils using recently developed molecular techniques (phospholipids fatty acid analysis, or PLFA), Steinberger et al. (1999) evaluated microbial biomass and diversity in Judean desert (Israel) soils along a precipitation gradient, during wet and dry periods. They found that biomass and diversity were highest in the wettest sites, during the wettest periods. Our own work in Negev desert (Israel) soils has found that the intensity of soil respiration during wetting events is high enough to deplete soil O2 concentration, providing the anaerobic conditions required for denitriﬁcation (Zaady et al. 1996b). We hypothesize that anaerobic respiration processes like denitriﬁcation are actually more important in arid ecosystems than in more temperate areas (Peterjohn and Schlesinger 1990, Groﬀman et al. 1993, Frank and Groﬀman 1998) because they are so prevalent during the pulses of activity following wetting. The presence of anaerobic respirers like denitriﬁers

Microbial Contributions to Biodiversity in Deserts

111

may represent an important functional diversity, maximizing the extent of nutrient cycling activity that can occur during the brief periods of water availability following rain events. On the other hand, the presence of denitriﬁers fosters gaseous N loss from these systems, which may be a critical constraint to primary production in these ecosystems (Peterjohn and Schlesinger 1990).

Ecosystem Diversity and Microbial Biomass, Activity, and Biodiversity in Deserts Ecosystem diversity is related to the number of (1) compartments that partake in ecosystem processes, (2) trajectories in ecosystem processes, and (3) functional groups involved in ecosystem processes. In its simplest form, nutrient cycling involves the movement of nutrients between plants, animals and inorganic forms in the soil (ﬁg. 7.1). However, as shown in ﬁgure 7.1, ecosystem biodiversity can considerably complicate nutrient cycles. The presence of diﬀerent plants, which produce diﬀerent amounts and types of organic matter (Pastor et al. 1984, Scott and Binkley 1997, Finzi et al. 1998), and the importance of fauna (Beare et al. 1992) can vary greatly in diﬀerent ecosystems. The importance of ecosystem diversity to soil biodiversity and its role in sustaining ecosystem functions are not clear. Our work in the Negev suggests that ecosystem diversity may be important to nutrient cycling functions in aridlands. As shown in ﬁgure 7.1, Negev desert ecosystems have a large number of ecosystem components and complex nutrient cycling patterns. This ecosystem diversity may lead to high soil microbial diversity. For example, snails have long been recognized as an important component of Negev desert ecosystems (ﬁg. 7.2, Shachak et al. 1987, Jones and Shachak 1990, 1994). Due to their high density and high feeding rates they produce large amounts of feces. These feces may greatly facilitate the cycling of nutrients from plants back to inorganic soil pools by functioning as a pool of nutrients that is readily mobilized when increases in soil moisture permit (ﬁg. 7.3, Zaady et al. 1996a). Alternatively, snail feces may detract from soil nutrient cycling functions in that this mobilizable pool of nutrients is subject to gaseous and leaching loss. Thus snails have both positive and negative eﬀects on ecosystem nutrient cycling and retention. Ecosystems with snails clearly have higher N availability during wetting events (ﬁg. 7.3), but they also likely have higher N losses during these periods as well. An important question for future research is to determine if ecosystems with snails have higher soil microbial biodiversity and if this biodiversity inﬂuences nutrient cycling and loss functions during wetting events. More fundamentally, it will be important to examine relationships between ecosystem diversity and nutrient cycling processes in other ecosystems to determine if high ecosystem diversity facilitates nutrient cycling processes in a wide range of systems (e.g., it would be interesting to evaluate deserts with no snails).

112

Living Components of Biodiversity: Organisms

Figure 7.1 Conceptual diagram of a nutrient cycle in a Negev desert ecosystem highlighting the importance of ecosystem diversity (i.e., multiple ecosystem compartments, trajectories, and functional groups) as a regulator of the transformation of organic matter and the importance of microbes (MO) as decomposers.

It is important to note that there are several types of snails in the Negev. There are rock-eating snails that feed on endolithic lichens, soil-eating snails that feed on cyanobacteria in soil crusts, and plant-eating snails that feed on detritus (Yom-Tov 1971, Shachak et al. 1987, Jones and Shachak 1990). Snail consumption of these diverse food sources adds a high diversity of organic materials to the microbial community. It remains to be seen if this high diversity of food sources increases soil biodiversity. In contrast to snail feces, plant litter appears to act as an important sink for nutrients released during wetting events (ﬁg. 7.4, Zaady et al. 1996b). In simulated wetting events with litter, microbial N pools in soil decreased over time, suggesting that N was being sequestered in litter rather than remaining

Microbial Contributions to Biodiversity in Deserts

113

Figure 7.2 Negev desert landscape in Sayeret Shaked Park near beer Sheva in the northern Negev Desert, Israel (31 170N, 34 370E), showing high density of snails.

Figure 7.3 Inorganic N during a simulated wetting event (a 12-day laboratory incubation at 25 C) in Negev desert soils with and without snail feces. Note that a large pulse of inorganic N was released from snail feces but that this pulse disappeared over the course of the incubation. Feces were collected from sites in the rocky hills of the northern Negev Desert, Israel near Sede Boqer (30 520N 34 470E) and were derived from the snail Trochoidea seetzenii feeding on litter the shrub Zygophyllum dumosum. Data from Zaady et al. (1996a).

114

Living Components of Biodiversity: Organisms

Figure 7.4 Chloroform-labile (microbial) N during a simulated wetting event (a 30day laboratory incubation at 25 C) in Negev desert soils with and without plant litter. Note that soil microbial N decreases over time in the litter-amended incubations, suggesting that N released from soil by wetting was sequestered in the litter. Litter was collected from 20 macrophytic patch mounds in Sayeret Shaked Park near Beer Sheva in the northern Negev Desert, Israel, (31 170N, 34 370E), mixed, and then sieved to ﬁve diﬀerent size classes; 250, 500, 1000, 2000 mm. Material less than 250 mm was considered to be soil. Values are mean standard error of all litter size classes combined versus the soil-alone treatment. Data from Zaady et al. (1996b).

in the actively cycling (and more easily lost) soil microbial pool. Litter may counteract some of the stimulatory eﬀects of snail feces, which are often mixed in with litter. The snail and litter examples suggest that arid conditions do not inherently inhibit the development of high ecosystem diversity and the emergence of complex nutrient cycling activity. Indeed, complex nutrient cycles may greatly facilitate nutrient availability and plant growth under arid conditions. However, these complex nutrient cycles may also stimulate nutrient loss. In all ecosystems, there is an inherent tension between processes that facilitate nutrient cycling and availability and those that act to prevent nutrient loss (Bormann and Likens 1979, Vitousek et al. 1998). Determining if and how ecosystem diversity aﬀects this balance is not clear, and represents an important challenge for ecology over the next decade or so. Determining how ecosystem diversity inﬂuences soil biodiversity may be critical to resolving this functional question.

Microbial Contributions to Biodiversity in Deserts

115

Landscape Diversity and Microbial Biomass, Activity, and Biodiversity in Deserts Many desert areas are characterized by high landscape diversity, for example marked patchiness in the distribution of vegetation (Barth and Klemmedson 1978, West 1989, Schlesinger et al. 1990, Allen 1991, Agular and Sala 1999). This landscape-scale diversity is thought to have important eﬀects on desert ecosystem functions related to nutrient cycling and productivity (Shachak et al. 1998). Our data from the Negev conﬁrm that landscape diversity is important to ecosystem function, but suggest that this diversity may have both positive and negative eﬀects on nutrient cycling and retention. In the Negev there are two main types of patches (ﬁg. 7.5, Friedmann and Galun 1974, West 1990, Shachak et al. 1993, Shachak and Boeken 1994, Zaady and Shachak 1994): (1) macrophytic patches consisting of herbs and shrubs on slightly raised mounds and (2) microphytic patches consisting of cyanobacteria, algae, mosses, and lichens that are characterized by the presence of a relatively impermeable soil crust. This landscape diversity likely inﬂuences microbial biodiversity through its eﬀects on key factors that control microbial community composition and function, such as soil moisture and organic matter. The macrophytic patches, which are common to many deserts, are considered to be ‘‘islands of fertility’’ with high levels of water and nutrients (West 1989, Smith et al.1994) relative to the microphytic patches. The extreme diﬀerences in the nature and extent of water and nutrient availability in the diﬀerent patch types likely increases overall microbial biodiversity in aridlands.

Figure 7.5 Conceptual diagram of the Negev desert landscape showing macrophytic patches dominated by shrubs and microphytic patches dominated by cyanobacteria, algae, mosses, lichens, and soil crust. From Zaady et al. (1998).

116

Living Components of Biodiversity: Organisms

Landscape diversity eﬀects on microbial processes have important implications for ecosystem productivity and nutrient retention in aridlands (Shachak et al. 1998, Agular and Sala 1999). For example in the Negev, there is a strong assumption that both major patch types are important components of the landscape nutrient cycle. Plant growth in the macrophytic patches is thought to be strongly limited by N (Evenari 1985), while N-ﬁxing cyanobacteria are thought to contribute signiﬁcant amounts of N to the microphytic patches (West 1990, Evans and Ehleringer 1993). Runoﬀ following rainfall events may transfer N from the crust-covered microphytic patches to the macrophytic patches (Zaady and Shachak 1994). A major area of research is to determine the optimal juxtaposition and abundance of diﬀerent patch types for speciﬁc ecosystem management objectives (Shachak et al. 1998). Some of our research in the Negev raises questions about the role of diﬀerent patch types in landscape nutrient cycling and loss. First, the role of the biological crusts as sources of N to the landscape may be overstated. While there is a general assumption that the crust areas ﬁx N at relatively high rates, and that this N contributes to productivity in the macrophytic mounds, we have found equal or higher rates of N ﬁxation in the mounds than in the crusts (ﬁg. 7.6, Zaady et al. 1998). Second, while the macrophytic patches are clearly ‘‘hotspots’’ of higher plant productivity in the landscape, they support

Figure 7.6 N ﬁxation during a simulated wetting event (a 72-hour laboratory incubation at 25 C) in soils from macro-and microphytic patches in the Negev desert. Soils were collected from nine sites located along a rainfall gradient in the Negev Deset, Israel, ranging from from 50 to 325 mm per annum. At each site, samples were collected from soils underlying shrubs, from cyanobacteria-dominated microphytic crusts (less than 125 mm rainfall) and moss-dominated microphytic crusts (more than 125 mm rainfall). N ﬁxation was measured with an acetylene reduction method. Values are mean standard error.

Microbial Contributions to Biodiversity in Deserts

117

Figure 7.7 Denitriﬁcation during a simulated wetting event (a 72-hour laboratory incubation at 25 C) in soils from macro-and microphytic patches in the Negev desert. Soils were collected from nine sites located along a rainfall gradient in the Negev Deset, Israel, ranging from from 50 to 325 mm per annum. At each site, samples were collected from soils underlying shrubs, from cyanobacteria-dominated microphytic crusts (less than 125 mm rainfall) and moss-dominated microphytic crusts (more than 125 mm rainfall). Denitriﬁcation was measured with an acetylene inhibition method. Values are mean standard error.

high rates of denitriﬁcation, and are therefore also ‘‘hotspots’’ of N loss (ﬁg. 7.7). There is a clear need for more speciﬁc evaluation of the functional role of diﬀerent patch types to determine just how landscape diversity is important to water and nutrient cycling functions in deserts. Future research should also determine if soil microbial biodiversity plays a key role in the functional diﬀerences between diﬀerent patch types.

Conclusions Soil microbial ecology has been dominated by functional studies of processes, especially those related to nutrient cycling. While we know a lot about how ecosystem conditions inﬂuence microbial activity, we know little about the role that microbial diversity and community structure play in this activity. Studies of microbial response to variation in ecosystem conditions should be a productive platform for posing questions about links between biodiversity and ecosystem function. Deserts, where conditions vary greatly in time and space, may be especially valuable venues for such studies.

118

Living Components of Biodiversity: Organisms

Microbial activity in deserts is concentrated in brief periods of optimal conditions during wetting events. A full range of processes (e.g., aerobic and anaerobic) occurs during these events. This functional diversity is important to patterns of nutrient cycling and nutrient loss in desert ecosystems. High ecosystem diversity could facilitate microbial diversity and does facilitate microbial activity in desert ecosystems. A high diversity of ecosystem compartments, trajectories in ecosystem processes, and functional groups facilitates high rates of microbial activity during wetting events, allowing for large amounts of nutrient cycling to occur during these events, at least in the Negev. The links between ecosystem diversity, soil microbial biodiversity, and these nutrient cycling functions should be explored in future research in a wide range of ecosystem types. While aridity may not constrain microbial diversity and activity, it may constrain biotic control over nutrient losses from desert ecosystems. The potential for gaseous and hydrologic loss during wetting events is high. In all ecosystems, there is an inherent tension between processes that facilitate nutrient cycling and availability and those that act to prevent nutrient loss. A favorable balance of these processes may be less common in deserts than in more mesic ecosystems. The role of soil microbial biodiversity in this balance may be important, for example if there are more functional types of microbes in the soil, there may be a better balance between cycling, availability, and loss. The balance between processes that facilitate both nutrient cycling and loss needs to be considered in assessments of patch dynamics and management in desert landscapes. Decisions about managing the abundance and juxtaposition of patch types should consider eﬀects on nutrient losses as well as the nutrient input and plant diversity considerations that usually drive these assessments. Evaluating the eﬀects of soil microbial biodiversity on the nutrient cycling and loss functions in diﬀerent patch types should be a priority for future research.

References Agular M.R., and O.E. Sala. 1999. Patch structure, dynamics and implications for the functioning of arid ecosystems. TREE 14: 273–277. Allen, E.B. 1991. Temporal and spatial organization of desert plant communities. Pp. 295–332 in J. Skujins (ed.). Semiarid Lands and Desert: Soil Resource and Reclamation. Marcel Decker, New York. American Society for Microbiology. 1994. Microbial diversity research priorities. American Society for Microbiology, Washington, D.C. Amman, R., and M. Kuhl. 1998. In situ methods for assessment of microorganisms and their activities. Current Opinion in Microbiology 1: 352–358.

Microbial Contributions to Biodiversity in Deserts

119

Barth, R.C., and J.O. Klemmedson. 1978. Shrub-induced spatial patterns of dry matter, nitrogen, and organic carbon. Soil Science Society of America Journal 42: 804–809. Beare, M.H., R.W. Parmelee, P.F. Hendrix, W. Cheng, D.C. Coleman, and D.A. Crossley, Jr. 1992. Microbial and faunal interactions and eﬀects on litter nitrogen and decomposition in agroecosystems. Ecological Monographs 62: 569–591. Beare, M.H., D.C. Coleman, D.A. Crossley, Jr., P.F. Hendrix, and E.P. Odum. 1995. A hierarchical approach to evaluating the signiﬁcance of soil biodiversity to biogeochemical cycling. Pp. 5–22 in H.P. Collins, G.P. Robertson and M.J. Klug (eds.). The Signiﬁcance and Regulation of Soil Biodiversity, Kluwer Academic Publishers, The Netherlands. Bormann, F.H., and G.E. Likens. 1979. Pattern and Process in a Forested Ecosystem. Springer-Verlag, New York. Brussaard, L., V.M. Behan-Pelletier, D.E. Bignell, V.K. Brown, W. Didden, P. Folgarait, C. Fragoso, D.W. Freckman, V.V.S.R. Gupta, T. Hattori, D.L. Hawksworth, C. Klopatek, P. Lavelle, D.W. Malloch, J. Rusek, B. Soderstrom, J.M. Tiedje, and R.A. Virginia. 1997. Biodiversity and ecosystem functioning in soil. Ambio 26: 563–570. Evans, R.D., and J.R. Ehleringer. 1993. A break in the nitrogen cycle in aridlands? Evidence from d15 N of soils. Oecologia 94: 314–317. Evenari, M. 1985. The desert environment. Pp. 1–22 in M. Evenari, I. Noy-Meir and E.W. Goodall (eds.). Hot Deserts and Arid Shrublands, Vol. 12A: Ecosystems of the World. Elsevier, Amsterdam. Finzi, A.C., N. van Breemen, and C.C. Canham. 1998. Canopy tree-soil interactions within temperate forests: Species eﬀects on soil carbon and nitrogen. Ecological Applications 8: 440–446. Fließbach A., S. Sarig, and Y. Steinberger. 1994. Eﬀects of water pulses and climatic conditions on microbial biomass kinetics and microbial activity in a Yermosol of the central Negev. Arid Soil Research and Rehabilitation 8: 353–362. Frank, D.A., and P.M. Groﬀman. 1998. Denitriﬁcation in a semi-arid grazing ecosystem. Oecologia 117: 564–569. Freckman, D.W., T.H. Blackburn, L. Brussaard, P. Hutchings, M.A. Palmer, and P.V.R. Snelgrove. 1997. Linking biodiversity and ecosystem functioning of soils and sediments. Ambio 26: 556–562. Friedman, E.I., and M. Galun. 1974. Desert algae, lichens and fungi. Pp. 165–212 in G.W. Brown (ed.). Desert Biology, Vol. 2. Academic Press, London. Giller, P.S. 1996. The diversity of soil communities, the ‘‘poor man’s tropical rainforest.’’ Biodiversity and Conservation 5: 135–168. Groﬀman, P.M., and P.J. Bohlen. 1999. Soil and sediment biodiversity: cross-system comparisons and large scale eﬀects. BioScience 49: 139–148. Groﬀman, P.M., C.W. Rice, and J.M. Tiedje. 1993. Denitriﬁcation in a tallgrass prairie landscape. Ecology 74: 855–862. Jones, C.G., and M. Shachak. 1990. Fertilization of the desert soil by rock-eating snails. Nature 346: 839–841. Jones, C.G., and M. Shachak. 1994. Desert snails’ daily grind. Natural History 103: 56–61. Kennedy, A.C., and V.L. Gewin. 1997. Soil microbial diversity: Present and future considerations. Soil Science 162: 607–617.

120

Living Components of Biodiversity: Organisms

Parker, L.W., P.F. Santos, J. Phillips, and W.G. Whitford. 1984. Carbon and nitrogen dynamics during the decomposition of litter and roots of a Chihuahuan desert annual, Lepidium lasiocarpum. Ecological Monographs 54: 339–360. Pastor, J., J.D. Aber, C.A. McClaugherty, and J.M. Melillo. 1984. Aboveground production and N and P cycling along a nitrogen mineralization gradient on Blackhawk Island, Wisconsin. Ecology 65: 256–268. Peterjohn, W.T. 1991. Denitriﬁcation: enzyme content and activity in desert soils. Soil Biology and Biochemistry 23: 845–855. Peterjohn, W.T., and W.H. Schlesinger. 1990. Nitrogen loss from deserts in the western United States. Biogeochemistry 10: 67–79. Schlesinger, W.H., P.J. Fonteyn, and G.M. Marion. 1987. Soil moisture content and plant transpiration in the Chihuahuan desert of New Mexico. Journal of Arid Environments 12: 119–126. Schlesinger, W.H., J.F. Reynolds, G.L. Cunningham, L.F. Huenneke, W.M. Jarrell, R.A. Virginia, and W.G. Whitford. 1990. Biological feedbacks in global desertiﬁcation. Science 247: 1043–1048. Schimel, J. 1995. Ecosystem consequences of microbial diversity and community structure. Pp. 240–254 in F.S. Chapin and A. Ko¨rner (eds.). Arctic and Alpine Biodiversity. Springer-Verlag, Berlin, Heidelberg. Scott, N.A., and D. Binkley. 1997. Foliage litter quality and annual net N mineralization: comparison across North American forested sites. Oecologia 11: 151–159. Service, R.F. 1997. Microbiologists explore life’s rich, hidden kingdoms. Science 275: 1740–1742. Shachak, M., and B. Boeken. 1994. Desert plant communities in human-made patches—implications for management. Ecological Applications 4(4): 702–716. Shachak, M., B. Boeken, and E. Zaady. 1993. Ecological Aspects of the Savannization Project. Annual Progress Report 1993, Jewish National Fund (in Hebrew). Shachak, M., C.G. Jones, and Y. Granot. 1987. Herbivory in rock and the weathering of a desert. Science 236: 1098–1099. Shachak, M., M. Sachs, and I. Moshe. 1998. Ecosystem management of desertiﬁed shrublands in Israel. Ecosystems 1: 475–483. Shachak, M., R. Waide, and P.M. Groﬀman, this volume. Ecosystem processes: a link between species and landscape diversity. Smith, J.L., J.J. Halvorson, and H. Bolton Jr. 1994. Spatial relationships of soil microbial biomass and C and N mineralization in a semi-arid shrub-steppe ecosystem. Soil Biology and Biochemistry 26: 1151–1159. Steinberger, Y., L. Zelles, Q.Y. Bai, M. Von Lu¨tzow, and J.C. Munch. 1999. Phospholipid fatty acid proﬁles as indicators for the microbial community structure in soils along a climatic transect in the Judean Desert. Biology and Fertility of Soils 28: 292–300. Vitousek, P.M., L.O. Hedin, P.A. Matson, J.H. Fownes, and J. Neﬀ. 1998. Withinsystem element cycles, input-output budgets and nutrient limitation. Pp. 432–451 in: M.L. Pace and P.M. Groﬀman (eds.) Successes, Limitations and Frontiers in Ecosystem Science. Springer-Verlag, New York. Wall, D.H., and J.C. Moore. 1999. Interactions underground. BioScience 49: 109–188. Wardle, D.A., and K.E. Giller. 1996. The quest for a contemporary ecological dimension to soil biology. Soil Biology and Biochemistry 28: 1549–1554. West, N.E. 1989. Spatial pattern-functional interactions in shrub-dominated plant communities. Pp. 283–305 in: C.M. McKell (ed.). The Biology and Utilization of Shrubs. Academic Press, London.

Microbial Contributions to Biodiversity in Deserts

121

West, N.E. 1990. Structure and function of microphytic soil crusts in wildland ecosystems of arid to semi-arid regions. Advances in Ecological Research 20: 179–22. Yom-Tov, Y. 1971. The biology of the desert snails. Trochoidea seetzenii and Sphincterochia boissieri. Israel Journal of Zoology 20: 213–248. Zaady, E. and M. Shachak. 1994. Microphytic soil crust and ecosystem leakage in the Negev desert. American Journal of Botany 81: 109. Zaady, E., P.M. Groﬀman, and M. Shachak. 1996a. Release and consumption of nitrogen from snail feces in Negev desert soils. Biology and Fertility of Soils 23: 399–405. Zaady, E., P.M. Groﬀman, and M. Shachak. 1996b. Litter as a regulator of nitrogen and carbon dynamics in macrophytic patches in Negev desert soils. Soil Biology and Biochemistry 28: 39–46. Zaady, E., P.M. Groﬀman, and M. Shachak. 1998. Nitrogen ﬁxation in macro- and microphytic patches in the Negev desert. Soil Biology and Biochemistry 30: 449–454.

8 Uniﬁed Framework I Interspeciﬁc Interactions and Species Diversity in Drylands Gary A. Polis Yoram Ayal Alona Bachi Sasha R.X. Dall Deborah E. Goldberg Robert D. Holt Salit Kark Burt P. Kotler William A. Mitchell

T

he goal of this chapter is to delineate how abiotic conditions, regional processes, and species interactions inﬂuence species diversity at local scales in drylands. There is a very rich literature that bears on this topic, but here we focus on mechanisms that promote or constrain local diversity and ask how these factors apply to deserts. We ask, ‘‘What is diﬀerent about deserts, relative to other habitats, in their patterns of diversity, temporal variability in productivity, and spatial heterogeneity?’’ We assess how such diﬀerences might modify extant theory, and sketch relevant examples. Compared with other biomes, productivity, population densities, and community biomass are much lower in deserts, and temporal heterogeneity is typically higher. Do these diﬀerences imply distinct ecological processes and patterns in deserts? Or, do processes operate in deserts in similar ways as in tropical forests or grasslands? For example, it is often assumed that abiotic factors are more important in deserts. If so, how do abiotic factors modify biotic interactions? How do we integrate physical and biotic interactions? More generally, we ask what should be the main goals and approaches of a research program to understand the role of species 122

Unified Framework I 123

interactions in determining community structure in drylands, as modiﬁed by abiotic factors and regional processes.

What Is Different About Drylands? Relative Diversity of Deserts Deserts are traditionally perceived as relatively simple ecosystems harboring low species diversity. Yet increasing evidence suggests that desert communities can be highly diverse and complex. To our knowledge the only systematic analysis of the relative diversity in desert versus nondesert communities was compiled by Polis (1991a). These data suggest that patterns diﬀer widely among taxonomic groups. In some cases, deserts support high diversity, comparable to or even higher than nonarid areas (see Polis 1991b). For example, while avian (Wiens 1991) and anuran (Woodward and Mitchell 1991) diversities are low compared with other biomes, desert annual plants show extremely high species diversity (Inouye 1991). Ants, succulent plants, lizards, scorpions, and tenebrionid beetles also have relatively high diversity in deserts (Polis 1991a–c, Wiens 1991). But, while very high diversity may occur, local diversity varies greatly in space and time (e.g., ants and annual plants: Danin 1977, Inouye 1991, MacKay 1991). We suggest that deserts lie somewhere in the middle of the spectrum of diversity among biomes of the world, rather than at one extreme as often assumed (Polis 1991a–c). One theme in desert research has been to emphasize clines of diversity along physical and ‘‘aridity’’ gradients. Diﬀerent patterns may result from the fact that studies compare diﬀerent taxa, scales, and ecological regions. Studies focusing on trends in species diversity across physical gradients diﬀer in the type of gradient compared. In some cases, precipitation gradients (e.g., birds; MacKay 1991) are studied. Some workers even compare latitudinal clines or a combination of these factors (Ricklefs and Schluter 1993). Thus, conclusions derived as to general patterns of arid-land species diversity may diﬀer depending on the taxon, geographical area, scale (Rosenzweig 1995), and the particular gradient examined. However, in most cases, species richness was not constant across the aridity gradient. In some systems, species richness declines with aridity (e.g., birds), whereas in others, diversity increases (e.g., scorpions). Important Characteristics of Deserts that Affect Diversity 1. Low productivity. By deﬁnition drylands differ from other biomes in having low annual precipitation (100–250 mm/year) or net negative evapotranspiration (Polis 1991a). Moreover, desert soils are often nutrient-poor (Anderson and Polis 1999). Low water and nutrient availability translates into very low annual net primary productivity (ANPP averages 5–200 g/ m2/yr) (Lieth 1978).

124

Living Components of Biodiversity: Organisms

2. Temporal variability. In addition to being unproductive on average, deserts are also the most variable biome in temperature, precipitation, and productivity (Polis 1991a). Annual variation in precipitation is inversely related to mean precipitation, so interannual variation (unpredictability) in rainfall is higher in arid regions than in biomes with higher annual precipitation (Polis and Yamashita 1991, Polis et al. in press). Annual rainfall may vary by 1–2 orders of magnitude in deserts (e.g., southern California: 34mm–301 mm, Polis and Farley 1980; the Namib: 2.2–134 mm, Seely and Louw 1980; Baja California 0–10 mm in dry years to 150–280 mm in El Nin˜o years, Polis et al. 1997a). Productivity mirrors precipitation during these periods (Seely and Louw 1980, Noy-Meir 1981). For example, the standing biomass in the Namib increased 600–900% (plants and animals respectively) after heavy rains following a 13-year dry period (Seely and Louw 1980). Above-ground ANPP varied by about 20 times between years in the Chihuahuan desert (Ludwig 1986) and 5–43 times on desert islands in the Gulf of California (Polis et al. 1998). Great variation in rainfall drives variation in productivity (Ludwig 1986, 1987) and community structure (see below). Extreme precipitation events (e.g., due to El Nin˜o events) may exert more inﬂuence on community composition and ecosystem patterns and processes than do ‘‘mean’’ conditions (Noy-Meir 1973, 1974, Ludwig 1986, Hobbs and Mooney 1995). Longer term changes in precipitation (at scales from 10 to 106 years) also occur. For example, American rain-shadow deserts were produced from more mesic areas during orogenous events that built the Andes and Sierra Nevada Mountains. Large parts of the southwestern deserts of North America were under great lakes as little as 10,000–12,000 years ago (Benson and Thompson 1987). Temporal heterogeneity at geological scales could exert strong historical inﬂuences. Desert biota probably experience relatively rapid change in patterns of distribution and abundance during wet–dry cycles. We posit that low mean productivity in deserts may be less important to diversity and species interactions than is high variance in productivity. That is, we speculate that chronic low productivity is more easily accommodated by evolutionary adaptation than is unpredictability in production (Polis 1991a, Wiens 1991). Temporal diﬀerences in productivity occur at many diﬀerent scales and signiﬁcantly inﬂuence communities (Wiens et al. 1986, Giller and Gee 1987, Kotler and Brown 1988, Polis 1991a, Polis et al. 1996). Attempting to elucidate the consequences of temporal variation is one of the most challenging, least resolved, but most important tasks of community ecology. 3. Spatial heterogeneity. Deserts have also been postulated to be more spatially heterogeneous than more mesic environments (Crawford 1986). However, unlike temporal heterogeneity, we are not aware of quantitative comparisons of spatial heterogeneity of deserts with other environments except for information on foliage height diversity (Cody 1974). Comprehensive assessments of spatial diversity that include vegetation,

Unified Framework I 125

edaphic, and geological characteristics would be a difﬁcult, but important, endeavor. There seem to be two main lines of reasoning underlying this generalization, both incorporating feedback processes that reinforce heterogeneity over time. First, the paucity of plant cover implies that soil is not well developed. Heterogeneity in the physical nature of the substrate (e.g., rock outcrops, gravel beds, coarse sand, or ﬁne dust) is thus exposed at the surface (rather than buried under soil horizons that may be relatively more uniform in texture, as in more mesic biomes). Surface substrata may diﬀer drastically in water inﬁltration properties and nutrient storage, and therefore in their potential for plant growth and animal activities. This heterogeneity is reinforced in a positive feedback because of the patterns of run-oﬀ of water after rains (Noy-Meir 1981, Yair and Shachak 1987). Run-oﬀ sources (e.g., smooth rock, bare soil with a crust) contribute incident water to patches with higher inﬁltration rate, further increasing soil development in run-on sites (Shachak et al. 1998). Second, the sparse plant cover in drylands causes strong contrasts between local areas where plants occur, and areas where they do not. The ‘‘islands of fertility’’ around individual plants are due to multiple processes that lead to long-term improvements in soil conditions under plants. By forming wind barriers, plants collect ﬁner soil particles and organic debris and add their own litter, thereby increasing nutrient availability. The presence of relatively impermeable plant crusts (lichens, blue-green algae) on bare soil between shrubs provides further horizontal redistribution of water and other materials; crusts lead to run-oﬀ, and shrubs collect run-on water, debris, and seeds. These processes increase water inﬁltration, which further increases microbial activity and nutrient availability. Besides soil modiﬁcation, the presence of adult plants modiﬁes animal activities, for example, by concentrating burrows or animal urine and feces under plants. These activities again reinforce the initial heterogeneity that facilitated the plant’s presence. Thus, ‘‘hot spots’’ of relatively high primary productivity and biological activity juxtaposed to very unproductive interplant areas characterize deserts. Enriched soils and associated plants, in turn, often cascade to exert strong facilitative eﬀects on other organisms. Two general positive eﬀects arise from plants—a trophic eﬀect from their productivity and a refuge eﬀect from their presence in the physical environment. First, plant production travels up the food web to aﬀect secondary productivity and the distribution and abundance of heterotrophs (Noy-Meir 1985, Polis 1999). For example, some desert shrubs support a high diversity of herbivorous insects (Wisdom 1991); in turn, invertebrate predators are attracted to these areas: 26 spider species occur on creosote (Larrea divaricata) and more than 25 species on saltbush (Atriplex canescens) (Polis and Yamashita 1991). Second, many small animals use plants as refuges from both harsh abiotic conditions and from larger predators. For example, Ayal and Merkl (1994) found that large tenebrionid beetle species were most abundant in habitats with high shrub cover, where they were protected from predatory birds (Groner and Ayal 2001). Ayal et al.

126

Living Components of Biodiversity: Organisms

(chapter 2) suggest that this pattern may hold for most detritivores and small herbivores in drylands, and that the structural role of plant cover may outweigh the trophic eﬀect of plants in determining dryland food webs. Kotler (1984) has documented the role of shrubs as patches of refuge from predators. Other biotic processes likewise increase spatial heterogeneity. For example, many desert plants secrete salt and increase soil salinity around their canopies (Liphschitz and Waisel 1982). Animal activity such as burrowing (ants, termites, isopods) change soil properties by increasing water absorption, and adding nutrients through secretion (Crawford 1986). Abandoned burrows serve as refuges for small animals (Heth 1991). Termites increase local decomposition rate by burying dry plant material in their burrows, whereas ants concentrate seeds of their food plants in their nests (Whitford 1991). The crested porcupine is very eﬀective in increasing spatial heterogeneity in the Negev Desert in Israel as a result of digging activity while feeding on underground storage organs of plants. Its diggings trap both water and organic material, creating better germination sites for many species (Alkon and Olsvig-Whittaker 1989, Boeken et al. 1995). We are uncertain if these biotic facilitations are more or less important in deserts, compared with more mesic systems. However, it is important to note that some types of spatial heterogeneity are less prominent in deserts. In particular, foliage height diversity in deserts is low. Consequently, whatever eﬀects such architecture exerts on community characteristics (e.g., increasing bird diversity by providing more spatial niches; Cody 1974) must be less important in deserts. While we have postulated that spatial heterogeneity in soil and geological properties is likely greater in drylands than in more mesic closed-cover plant communities, edaphic conditions and sometimes plant composition in these other communities can also be highly heterogeneous. Nevertheless, we argue that the contrast between sites with no plants at all and some plants—as is found in arid systems—is more dramatic and consequential for the lives of other organisms. As discussed in much of the rest of this chapter, this hypothesized higher spatial heterogeneity in more arid environments could have many important consequences for species interactions, population dynamics, and community structure (chapter 12). With the advent of remote sensing, geostatistics, and other techniques of spatial analysis, comparable data on spatial heterogeneity is beginning to accumulate. For example, geostatistics have been used to describe the magnitude and scale of spatial heterogeneity in old ﬁelds (Robertson et al. 1988) and in agricultural ﬁelds (Robertson et al. 1993). As more such data are published, it will be possible to test with rigor the hypothesis that spatial heterogeneity is greater in arid environments. Consequences of Temporal and Spatial Heterogeneity One key aspect of heterogeneity is the temporally unpredictable ‘‘feast or famine’’ nature of primary productivity and food availability in deserts

Unified Framework I 127

(Polis 1991a, Polis et al. 1998). Under ‘‘bad’’ conditions, precipitation is low or nonexistent and plants grow little, if at all. In ‘‘good’’ periods of adequate to heavy rains, a relatively luxuriant plant growth may occur. Dramatic changes in productivity stimulated Noy-Meir (1973, 1974) to propose his ‘‘pulse-reserve hypothesis’’ as a paradigm for arid areas. Noy-Meir argues that plants and animals grow and establish reserves (e.g., seeds, tubers, tissue, eggs) during good times; these reserves then maintain the population or individual during interim dry periods. Several basic features of desert organisms may have evolved in response to unpredictable productivity; for example, life history strategies (Polis and Farley 1980, Louw and Seely 1982, Polis 1991a) and opportunistic diets (Noy-Meir 1974, Brown 1986, Polis 1991c). Diﬀerences in life history and trophic opportunism exert great inﬂuence on interactions such as competition and predation. For example, opportunistic animals exhibit quick functional responses to prey eruptions/ﬂuctuations but probably cannot tightly regulate particular prey species. Crawford (1986) notes that spatial and temporal patchiness of nutritional reserves in deserts, combined with stochastic arrival of moisture, limit the accuracy of predictions about foraging and the impact of consumers. Spatial patchiness, temporal variation, and aridity clines create a mix of good and bad periods and habitats. Such variation has several consequences at the community level. Perhaps most important, desert systems should be rather dynamic, nonequilibrium communities. ‘‘Hide and seek’’ dynamics in heterogeneous environments allow for local extinctions followed by recolonization (Taylor 1988). Extinctions may be caused by physical disturbances, competitive exclusion, or mortality from predators or pathogens. For example, areas of relatively high productivity (e.g., run-oﬀ areas) can serve as refugia in times of severe drought (Noy-Meir 1981). Heterogeneity allows persistence and coexistence among species that are engaged in interactions that could otherwise lead to exclusion (Caswell 1978, Taylor 1988). Resource limitation and competition may only occur during periods of ‘‘ecological crunches’’ (Wiens 1977, 1991). Periods between crunches are marked by little or no competition and relaxed selection. Inferior competitors or prey can escape elimination by being distributed in periods or places that are enemy free. Heterogeneity spreads the risk of extinction and increases population persistence by decreasing overall susceptibility to various mortality factors. These processes, permitted by heterogeneity, promote diversity. Heterogeneity in production between (micro) habitats may also strengthen demographic and trophic interconnections among habitats. High interpatch variability in productivity likely partitions a species’ population into individuals that live in ‘‘source’’ and ‘‘sink’’ habitats (Pulliam 1988). Source habitats produce a net surplus of individuals. Sink habitats are suboptimal areas where populations are not self-sustaining, but persist due to immigration from sources. Similarly, we expect that marked spatial heterogeneity in productivity creates a landscape where consumer–resource and food web interactions in relatively unproductive habitats are strongly aﬀected by subsidies

128

Living Components of Biodiversity: Organisms

from productive patches (Polis et al. 1997b). In general, nutrients, detritus, prey, and predators should move primarily from productive to less productive habitats. We speculate that strong demographic and trophic spatial interconnections might occur commonly in the patchy environment of deserts. Finally, heterogeneity may slow the speed of evolution and the likelihood that species tightly coevolve. The strength of selection is not constant and gene ﬂow may disrupt locally adaptive changes in gene frequency. A tendency toward trophic generalization suggests that interactions may be relatively less tightly coevolved, leading to the organization of exploiter–victim systems or guilds of potential competitors that are more loosely structured, relative to more homogeneous environments.

Processes Affecting Species Diversity The Relationship Between Regional and Local Diversity Viewed over suﬃciently long time scales, all local ecological communities arise via colonization from larger regional and biogeographical species pools, ﬁltered by matches between species autecological requirements and local environments, and interspeciﬁc interactions such as predation and mutualism (Zobel 1992, Holt 1993, Belyea and Lancaster 1999). Patterns of species diversity, and in particular the relationship of species richness to local environmental factors such as primary productivity or disturbance rates, ultimately reﬂect the interplay of multiple factors operating at many spatial and temporal scales (Huston 1999). At a local scale, interspeciﬁc interactions such as competition, predation, and mutualism (which directly or indirectly require contacts between individuals) can either enhance or depress species richness. At broad spatial scales, speciation, species migration, and regional extinctions inﬂuence the size of the species pool available for colonization into a local focal community (Latham and Ricklefs 1993). For example, the great diversity of tenebrionid beetles in the Namib Desert may be attributed to extensive in situ speciation that occurred in this, the oldest of all extant deserts (Seely 1991). Alternatively, younger or isolated deserts may be depauperate in species diversity because regional pools of desert dwellers may be low or there has been little time for speciation. Species richness at local sites can be viewed as arising from a series of ﬁlters relating species pools at large spatial scales to local communities (Zobel 1992). The ﬁrst stage in explaining diversity at a particular site is simply the size of the regional species pool. All else equal, sites with larger regional pools are likely to have larger local pools—this ‘‘ﬁlter’’ is thus a result of biogeographic and evolutionary (speciation and extinction) processes. These processes are important determinants of local diversity (Ricklefs and Schluter 1993), but beyond the scope of this chapter. Instead, we focus on the ﬁlters between a given regional species pool and the local community, including abiotic limitations, facilitation, competition, and consumer–resource interactions.

Unified Framework I 129

Cornell and Lawton (1992) argued that if local interactions such as competition are important ﬁlters limiting membership in local communities, then one should observe a saturating rather than linear relationship on plots relating local to regional richness. Therefore, the observation that linear relationships seem to be the most common result (e.g., Caley and Schluter 1997) suggests that interspeciﬁc interactions may not be prime determinants of patterns in local species richness. This conclusion is controversial (Huston 1999), and it is not our goal here to address the various interpretations of these relationships. Rather, we emphasize that most theory and empirical research on species interactions in community ecology has focused on the role of species interactions in explaining patterns within a region with the same species pool, that is, the scatter around the lines, rather than with the shape of the lines themselves. Thus, overall linear relationships between local and regional richness across diﬀerent regions does not necessarily imply that interactions have no eﬀect on relative richness among sites within a region. Instead, interactions could still strongly inﬂuence the extent to which local richness deviates (positively or negatively) from the average expected for a given size of regional pool. It is useful to start with a null hypothesis, which is that local species interactions do not constrain (or facilitate) community membership; instead, abiotic conditions act as the primary ﬁlter. We then go on to explore the roles of positive interspeciﬁc interactions, competitive interactions, and consumer– resource interactions in determining community membership. Abiotic and Demographic Filters If one views the regional species pool as analogous to a ‘‘continent,’’ and the focal community as an ‘‘island,’’ the classical dynamic model of island biogeography (MacArthur and Wilson 1967) schematically portrays the interplay of regional and local processes in determining species richness and composition. This classical model views local communities as a balance between colonization of new species from the pool, and extinction of resident species. If species are not interacting, by deﬁnition the rate of colonization and extinction for any given species should be independent of local community composition. Factors that enhance colonization or lower extinction rates lead to higher local species richness, whereas factors that diminish colonization or aggravate extinction should push down species richness, relative to the size of the species pool. The ﬁrst ﬁlter between the regional pool and the local community is the match between a species’ basic niche requirements and local site characteristics. In deserts, compared with more moderate biomes, is there a stronger abiotic ﬁlter because of harsh conditions, low resources, or greater temporal heterogeneity, resulting in lower diversity? If so, this could result in either higher extinction rates or lower colonization rates (depending on the life stage at which the lack of match of requirements and site characteristics occurs) and xeric sites having fewer and mesic sites more species than the average

130

Living Components of Biodiversity: Organisms

expected, given a certain size regional pool. While this idea of a stronger abiotic ﬁlter in extreme environments is old and widespread (e.g., Darwin 1859 for extreme deserts, see Parsons 1996 for more recent discussion), it has also been highly controversial because stronger abiotic ﬁlters in more stressful environments are often taken to mean weaker impacts of competition and/or facilitation on community structure. However, this is not necessarily the case. The controversy may result from a confusion between the factors that determine whether a species is present/absent in a habitat (‘‘distribution’’) versus the factors that determine individual ﬁtness and population dynamics of species (‘‘abundance’’) that can persist in the habitat. For example, it is possible that more species are eliminated from xeric than from mesic sites by physiological intolerance of prevailing conditions (i.e., that abiotic ﬁlters are stronger) but those species that are able to persist compete just as intensely among themselves, as do species that cohabit in mesic conditions. Most existing comparisons of eﬀects of species interactions along favorability or stress gradients test the ‘‘abundance’’ eﬀect rather than the ‘‘distribution’’ eﬀect (see below for review) because they compare competitive intensities among species that naturally co-occur within a habitat, rather than eﬀects of competition on species that do not naturally persist in the habitat. To rigorously test the hypothesis that abiotic ﬁlters are stronger in more xeric environments, one would have to transplant to a habitat species that normally do not occur there and monitor their demography in the absence of potential competitors/facilitators/predators. In addition to complete physiological intolerance, abiotic ﬁlters potentially involve several other distinct demographic mechanisms that could control colonization and extinction rates and thus local richness for a given regional pool (table 8.1). First, because deserts have low productivity, many species are expected to be low in abundance, and/or have low maximal growth rates. All else being equal, these demographic factors tend to lower colonization rates into local communities, and enhance local extinction rates. Consider ﬁrst the eﬀect of low population size. Because of demographic stochasticity, populations with small absolute sizes experience high extinction risk even in favorable environments (Belovsky 1987). If a species has low average abundance in occupied sites, this both reduces the ﬂux of dispersers available for colonizing empty sites, and weakens the ‘‘rescue eﬀect’’ (Brown and Kodric-Brown 1979) in occupied sites. Now consider the eﬀect of low average growth rates. Episodic disturbances can drive species to troughs of low abundance, during which they risk extinction due to demographic stochasticity. Low growth rates lengthen the time period of increased extinction risk. Moreover, chronic mortality factors are more likely to drive species extinct if there is weak compensation with increased growth rates at low densities (e.g., Holt 1985). As noted above, compared with many other biomes, deserts have greater coeﬃcients of variation in production, as well as low overall levels of production. To be present in a local community there obviously has to be a match between a species’ autecological requirements (i.e., its fundamental niche) and

Unified Framework I 131 Table 8.1 Summary of potential effects of low productivity and high temporal and spatial heterogeneity on colonization (C) and extinction (E) rates Characteristic of Drylands Low productivity

High temporal variability

High spatial variability

Processes and Phenomena Aﬀected Low abundance ! increased importance of demographic stochasticity Low abundance ! fewer dispersing propagules Low population growth rate ! low rate of recovery from disturbance

Eﬀect on Demography "E "E #C

Populations at risk under rare ‘‘crunch’’ conditions

"E

Selection for high dispersal rates Selection for demographic buﬀering mechanisms (e.g., seed bank)

" C** # E**

Habitat specialists have fewer suitable patches ! more at risk in metapopulation scenario Greater range of environmental conditions ! more sink populations More potential for habitat selection

"E "E # E**

Note: Increased extinction (" E) and/or decreased colonization (# C) rates would lead to lower species richness in drylands relative to more mesic conditions. **Effects in the opposite direction, resulting in higher diversity in drylands.

the local environment. In highly variable environments such as deserts, the match between these may be ephemeral, leading to transient and spatially variable matches between niche requirements and local conditions. This could contribute to a high local extinction rate. Even for species that are generally adapted to dryland environments, because of temporal heterogeneity (e.g., occasional exceptionally long bouts between rainfall pulses) there may be enhanced extinction risks. For instance, for an annual plant species without a seed bank, its long-term growth rate is the geometric mean ﬁtness across long time-series of environmental ﬂuctuations. A single year of very low ﬁtness can doom the plant to local extinction, even in an overall favorable environment. Finally, as also noted above, for some organisms desert environments may be very patchy. Metapopulation models (see Holt 1997) suggest that habitat specialists may have trouble persisting on scarce habitats, unless they have very low extinction rates, or high colonization rates. If desert environments overall are experienced as highly patchy, this eﬀect could lower local species richness. Given these demographic considerations alone, one expects a lowering of local species richness, because of increased local extinction and decreased local colonization rates, so that deserts would have lowered slopes in regional–local species richness plots. However, a number of factors may mitigate these demographic eﬀects (table 8.1). For instance, the colonization rate of

132

Living Components of Biodiversity: Organisms

empty sites does not vary simply with the abundance of a species in those sites it occupies. Colonization rate is likely to increase with an increasing percapita propensity to disperse, and with an increasing probability of successfully traversing unsuitable landscapes separating habitable sites. As noted earlier, temporal variability that is often weakly correlated in space is a hallmark of desert ecosystems. Evolutionary theories of dispersal suggest that such patterns of spatiotemporal variability tend to favor increased per-capita rates of dispersal (Holt and Barﬁeld 2001, Ronce et al. 2000). If species in desert systems are characteristically strong dispersers, this should tend to enrich local communities by increasing colonization rates. (Many deserts are famous for nomadic behaviors in consumers.) Because plant biomass is low, passive dispersal propagules carried by the wind may also travel much farther before coming to rest due to physical obstruction. We are unaware of studies that speciﬁcally quantify dispersal rates of desert species, compared with other biomes. Furthermore, many adaptations to desert environments obviously mitigate local extinction risks. For plants, such adaptations include seed banks for annuals, and resource storage structures for perennials. Habitat selection by consumers can permit microhabitats of high or stable ﬁtness to be sought out, reducing the risk of local population extinction. Generalist consumers can beneﬁt from their potential use of a wide range of independently varying resources, which buﬀer extinction risk. The ideas sketched above suggest that the relationship between local and regional species richness in desert biomes is inﬂuenced by many factors, some of which depress diversity, and others which may enhance diversity. The net eﬀect of these factors is unclear. Needed are long-term population studies in deserts that will permit a comparison of local extinction rates with other biomes. While we have described these demographic processes under the rubric of abiotic ﬁlters, because they could operate in the absence of interspeciﬁc interactions, they all could be moderated or intensiﬁed by interspeciﬁc interactions, as described in the next three sections.

Positive Interactions Traditional community theory has emphasized how local interspeciﬁc interactions make species persistence and coexistence less likely, thus limiting local species richness. In recent years there has been increasing concern with the potential for positive interspeciﬁc interactions to inﬂuence communities (Bertness and Callaway 1994). Such positive interactions could lead to increasing colonization or decreasing extinction rates with increasing number of species. Can ‘‘diversity beget diversity,’’ and if so, is this process more important in drylands than other biomes? No general theory exists on this theme, so we here outline some initial tentative steps toward a synthesis of facilitation in desert systems.

Unified Framework I 133

Examples of positive interactions in drylands abound (see reviews in Callaway 1995). Here we attempt a rough classiﬁcation of the many distinct mechanisms that have been described to provide a basis for generalizing about their importance in drylands relative to more mesic environments. As with any classiﬁcation, the categories are not entirely discrete and our assignments are often ambiguous. Nevertheless, it provides a heuristic value in allowing hypotheses to be developed about classes of mechanisms. The most general distinction is between trophic interactions (those mediated entirely through eﬀects on consumable resources) and nontrophic interactions (not mediated through eﬀects on consumable resources). Nontrophic positive interactions are very similar in deﬁnition to ecosystem engineering as deﬁned by Jones et al. (1994, 1997), although they emphasize that engineering can have negative inﬂuences as well on some species. We prefer the terminology of nontrophic interactions to emphasize the contrast, and allow further parallel classiﬁcations within trophic and nontrophic types of interactions. Within each of these categories, we can further divide mechanisms of positive interactions into direct (no intermediary involved) and indirect (either abundance- or traitmediated) interactions. Here we focus on two of the possible types of positive interactions: those in which organisms facilitate other organisms in ways other than providing food (nontrophic facilitations, such as habitat amelioration and habitat creation); and those in which organisms indirectly facilitate other organisms by reducing the probability of their becoming food through predation by a third group of organisms (trophic indirect facilitation). Below, we give examples of each of these and review ideas and evidence about how they are expected to vary in xeric relative to mesic environments. The remaining category, trophic direct facilitations, is simply the consumer side of consumer–resource interactions (i.e., the plus side of a þ= interaction). For nontrophic direct facilitations, the best-documented class of examples in drylands is probably habitat amelioration, deﬁned broadly as when organisms improve the physical habitat in terms of resource availability and/or amelioration of stress in some way that beneﬁts other organisms. In plants, this is often more narrowly referred to as ‘‘nursing’’ or the ‘‘nurse plant phenomenon’’ because the facilitator is often an adult, while the facilitatee is often a juvenile (see Callaway 1995, Callaway and Walker 1997 for comprehensive reviews). For example, larger individuals of sessile species can reduce temperatures due to their shading eﬀect—reducing direct negative eﬀects of high temperature and increasing water availability and decreasing water demand (reviewed in Holmgren et al. 1997). Another example is the ‘‘island of fertility,’’ where isolated plants trap organic matter and gradually increase local nutrient supply for other, usually smaller, plants (e.g., GarciaMoya and McKell 1970). Both of these processes in plants can beneﬁt animals as well. Another class of examples of nontrophic direct facilitation is habitat creation, or the creation of new structures by organisms (allogenic ecosystem engineering, sensu Jones et al. 1994). These are also well-documented in drylands, for example, increased bird diversity with increased plant structural diversity (Cody 1993) and the favorable germination sites in pits

134

Living Components of Biodiversity: Organisms

created by porcupine diggings (Gutterman 1982). Nontrophic indirect facilitations would be more complex examples in which these eﬀects propagate through links of other species. For example, a shrub that casts shade might facilitate burrowing activities by rodents, in turn modifying growth conditions for seedlings of other plant species. Facilitation through habitat amelioration has been argued to be more common in less productive environments (Bertness and Callaway 1994). In demographic terms, that means the presence of organisms should decrease extinction rates or increase colonization rates more strongly in less productive environments. We suggest that facilitation through microhabitat creation is also more common in less productive habitats and that the reason is similar for both mechanisms: in an extremely unfavorable habitat or a site with no suitable habitat at all, even a small improvement in absolute terms makes a large percentage improvement (Bertness and Callaway 1994). A speciﬁc model developing this basic argument was presented for nurse plants in drylands by Holmgren et al. (1997). However, Jones et al. (1997) have argued that habitat creation may be more important in environments with extensive plant cover, abundant large animals, and dominant organisms with massive and persistent structures—characteristics of more, rather than less, productive ecosystems. Although previous workers have not always clearly separated eﬀects of habitat creation and habitat amelioration, narrative reviews of facilitation support the idea that nontrophic direct facilitative eﬀects are more common in unproductive environments in plants (Callaway and Walker 1997, Holmgren et al. 1997). The simplest example of direct facilitation in trophic interactions is for a consumer to have a ‘‘donor-controlled’’ relation with its resource population (DeAngelis 1992). This generates the potential for asymmetrical facilitation, given that the consumer clearly requires the resource population to persist, but without reciprocal dependency. This diversity-generating mechanism is usually thought of as operating through chains of specialist consumers, i.e., a greater diversity of resources can support a greater diversity of consumers while keeping degree of resource overlap constant. However, it may be more likely to operate through generalist consumers because of bet-hedging or portfolio eﬀects; if each of various resource populations responds in a species-speciﬁc fashion to environmental variability, a generalist consumer may be able to survive in situations where a specialist on any given resource risks extinction. Ritchie (1999) recently observed this eﬀect of trophic generalization on extinction risk for a dryland herbivore (prairie dogs). The higher risk of extinction of trophic specialists should be exaggerated, given chains of trophic specialization (e.g., a specialist herbivore on a single plant species, supporting a specialist parasitoid, supporting in turn a specialist hyperparasitoid; see Holt et al. 1999). The bet-hedging advantage of trophic generalization should be especially important when resource populations are low in average abundance and strongly variable through time. Thus, we expect specialist consumers to be less common in deserts than in other biomes, particularly at higher trophic levels. Diversity at low trophic ranks can

Unified Framework I 135

facilitate diversity at higher trophic ranks via the buﬀering eﬀects of trophic generalization. This advantage of trophic generalization in turn implies that desert food webs are likely to be highly reticulate. There may also be occasional positive top-down eﬀects upon species richness at low trophic levels. For instance, in situations where competitive exclusion might occur at low trophic ranks (e.g., strong resource competition among rodent species), the scarcity of cover in drylands (chapter 2) could magnify the eﬀect of predators, which in turn can facilitate coexistence (if the more eﬃcient competitor is also more vulnerable to predation, e.g., Kotler 1984). Trophic indirect facilitations include both predator- and competitormediated interactions and both trait- and abundance-mediated indirect interactions. In drylands, one documented mechanism is a trait-mediated indirect eﬀect on predator–prey interactions, whereby one group of organisms provides refuges and therefore reduces predation rates on the target organisms. (This can also be regarded as a form of interference, i.e., nontrophic competition from the point of view of the predator). For example, shrubs in drylands often provide refuges from natural enemies for other plants (e.g., Fuentes et al. 1986), for invertebrate prey species (Ayal and Merkl 1994, Groner and Ayal 2001), and for small mammals (Kotler et al. 1993). Two general hypotheses have been proposed about the importance of indirect facilitation involving refuges. Bertness and Callaway (1994) propose that this type of facilitation is most important at high productivity, consistent with Connell’s (1975) much earlier argument that natural enemies are more important at high productivity than at low or intermediate productivity. In contrast, Ayal et al. (chapter 2) have proposed that refuges created by vegetation for herbivores or intermediate predators are much more important at low productivity. A recent quantitative review (meta-analysis) of the results of ﬁeld removal experiments in plants (Goldberg et al. 1999) compared the frequency and magnitude of positive interactions along productivity gradients. Because it was impossible to extract information on mechanisms of facilitation from most of the studies, the review tested the total summed eﬀects of habitat amelioration, habitat creation, and provision of refuges, even though these diﬀerent mechanisms are predicted to change in diﬀerent ways with productivity. Using the same database as Goldberg et al. (1999), we found a higher frequency of facilitative eﬀects on plants in unproductive environments, consistent with the Bertness and Callaway model for habitat amelioration, but only for growth as the measure of plant response (ﬁg. 8.1, top). In addition, despite this more frequent occurrence of positive interactions, the actual magnitude of interactions (degree to which plants were facilitated or inhibited) did not diﬀer signiﬁcantly for growth responses to neighbors (ﬁg. 8.1, bottom). This suggests that the competitive interactions that do occur at low productivity tend to be stronger than those at higher productivity, that is, that both extremely strong facilitation and extremely strong competition are more likely in unproductive environments. While this is consistent with the higher

136

Living Components of Biodiversity: Organisms

Figure 8.1 A meta-analysis of the eﬀects of neighbors on growth and survival of plants in low versus high productivity environments, using the database from the summary of published experiments reported in Goldberg et al. (1999). Only herbaceous plants with herbaceous neighbors were included. Productivity was estimated by standing crop with values <350 g/m2 being considered as ‘‘low productivity’’ and >350 g/m2 as ‘‘high productivity.’’ Interaction intensity was quantiﬁed as ln(performance with neighbors/performance without neighbors); values >0 indicate facilitation and <0 indicate competition. (A) Percentage of all interactions that were positive. (B) Mean interaction intensity 1 SE. N ¼ 15, 13 for survival at low and high productivity and n ¼ 115, 123 for growth at low and high productivity, respectively.

temporal variability expected for low-productivity arid environments, the data in ﬁg. 8.1 cannot be separated for unproductive environments due to aridity versus those due to low nutrient supply within mesic environments. Even more interesting is the observation that facilitation of survival in plants actually shows the reverse pattern: facilitation was signiﬁcantly more common in more productive sites (ﬁg. 8.1, top). In this case, average interaction

Unified Framework I 137

intensity was consistent with the frequency data, with signiﬁcantly more beneﬁcial interactions at higher productivity (ﬁg. 8.1, bottom). Thus, these results for survival are consistent with the Bertness–Callaway prediction about patterns in refuge provision. It is tempting to speculate that the diﬀerences between survival and growth responses to neighbors along productivity gradients is consistent with diﬀering demographic consequences of habitat stress (aﬀecting growth) and predation (aﬀecting survival), but more information on the actual mechanisms of facilitation in unproductive (especially xeric) habitats is needed to evaluate these ideas.

Negative Interactions Negative interactions between individuals in the same species and between diﬀerent species in the same trophic level are pervasive in arid ecosystems, as in other communities (Connell 1983, Schoener 1983, Fowler 1986, Gurevitch et al. 1992). By negative interactions, we include all situations where increased population density, individual growth rate, or per-capita population growth rate causes other individuals to decrease activity, individual, or population growth rates. Negative–negative interactions (i.e., competition) occur when negative eﬀects are reciprocal among individuals from diﬀerent species (e.g., Schoener 1983, Pimm et al. 1985). Negative interactions between species can constrain biodiversity (Rosenzweig 1978, Connell 1980, Schoener 1983, Holt 1984, Sih et al. 1985). Negative interactions can be divided into trophic and nontrophic interactions. Trophic interactions are both direct (i.e., cannibalism and intraguild predation) and indirect (e.g., exploitation competition for limiting resources). Intraguild predation is widespread in deserts (Polis et al. 1989). Cannibalism is another negative trophic interaction that is widespread across a diversity of taxa, including many desert species (Polis 1980, 1981, 1991c, Polis and Yamashita 1991). Exploitation competition is common among many desert taxa, including plants and granivorous rodents, birds, and ants (e.g., Fowler 1986, Brown and Lieberman 1973, Brown et al. 1979). Negative nontrophic interactions include a great variety of interference competition mechanisms such as territoriality, dominance hierarchies, chemical fouling, and allelopathy. Interference and allelopathic interactions have been argued to be wide spread among plant communities (Crawley 1986), though the extent to which these interactions aﬀect community dynamics and biodiversity is unknown. Likewise, territoriality occurs in many desert taxa, such as spiders (Riechert 1991), nectivorous hummingbirds (Brown and Kodric-Brown 1979), and granivorous kangaroo rats. The exact importance of exploitation competition among desert plants is uncertain. A number of conceptual models predict that abiotic factors are more important than biological interactions such as competition at less favorable or ‘‘stressful’’ ends of gradients (Miller 1967, Kruckeberg 1969, Grime 1973, 1977, Rosenzweig and Abramsky 1986, Keddy 1990, and see Connell

138

Living Components of Biodiversity: Organisms

1975, Menge and Sutherland 1987 for variants also incorporating predation). In contrast, other models suggest that competition is important regardless of environmental favorability, although the limiting resource may well change among environments (MacArthur 1972, Newman 1973, Tilman 1988). The prevailing thought has been that competitive interactions are rare or weak in arid ecosystems, especially for plants. The reason for this view was that population densities in deserts are often low, and organisms rarely come into contact to interfere or compete for resources. Moreover, the extreme conditions found in deserts often lead to notable morphological and physiological adaptations that may allow the use of a unique niche. Finally, as aridity increases and productivity falls, conditions for species coexistence often narrow. The typical outcome is competitive exclusion rather than coexistence, so the remaining species may exhibit a lack of competition. Species may need to possess similar adaptations to cope with arid conditions, either due to a shared phylogenetic history or because the range of potential adaptations is more limited in more extreme conditions. This could force species into recurrent competition. As an example, all annuals must be attuned to episodic rainfall pulses, and so may experience transient spikes of competition for water. However, in theory, niche partitioning could allow competitive coexistence in all but the most extreme cases (Chesson 2000). There is substantial evidence for strong competition due to exploitation of limiting resources in deserts, including habitat selection behavior among coexisting gerbil species in the Negev Desert (Abramsky et al. 1985, 1990, 1991) and their density-dependent activity times (Mitchell et al. 1990), competition for seeds between ants and rodents in the Sonoran Desert (Brown and Davidson 1977, Brown et al. 1979), and competition among perennial plants along a water gradient (Robberecht et al. 1983, Gurevitch 1986, Eissenstat and Caldwell 1988). Competition tends to constrain local diversity, unless there are mechanisms of coexistence (Tilman and Pacala 1993, Chesson 2000). Early models predicted that the amounts of limiting resources and physical factors set the maximum number of species that can coexist (see references in Armstrong and McGehee 1980). Current models predict the potential coexistence of more than one species on a single resource (Tilman 1990, Grover 1997). However, a general rule is that robust coexistence requires mutual invasibility: two species can coexist when each can invade a community that consists of its competitors at equilibrium, that is, each can increase when rare. This requires two ingredients. First, there must be an axis of environmental heterogeneity. Second, there must be a tradeoﬀ such that each species has a part of the axis where it is superior to its competitors (Brown 1989, Tilman and Pacala 1993), that is, niche diﬀerentiation. Environmental heterogeneity can occur in space, time, and resource attributes. Common niche axes in deserts include habitat diﬀerences based on substrate type, sand versus rock (Shmida and Wilson 1985), or vegetation density (wadi bottom vs. hillside), microhabitat (bush vs. open areas; Brown

Unified Framework I 139

and Lieberman 1973, Rosenzweig 1973, Kotler 1984), temporal resource variation (Kotler et al. 1993, Ziv et al. 1993), temporal diﬀerences in foraging costs due to seasonal activity of predators (Brown 1989), and spatial variation in resource abundance (Brown 1989). Water has a role in most of these. It is not clear if there are distinctive tradeoﬀs common in deserts that are rare in other biomes. However, the openness of the desert environment does lead to large spatial heterogeneity in predation risk (Ayal et al., chapter 2), so tradeoﬀs in competitive abilities in safe habitats and microhabitats versus ability to avoid predators in risky ones clearly contribute to species coexistence (Kotler et al. 1988). Also, extreme variation in precipitation and thus resource availability enhances the importance of tradeoﬀs in maintenance eﬃciency (via torpor or life history) to survive between pulses of productivity, versus foraging eﬃciency to continue to forage proﬁtably even at low resource density. Extreme spatial variability in resource availability promotes tradeoﬀs in travel and harvest rate versus foraging eﬃciency (Brown 1989). Tradeoﬀs of recruitment abilities under diﬀerent environmental conditions in regards to storage eﬀects may be very important in deserts (Warner and Chesson 1985). Consider some examples that illustrate these general features of deserts. An example of bush/open microhabitat selection includes the interaction among overwintering sparrows in semiarid grasslands (Pulliam and Mills 1977). Bird species diﬀer in escape abilities, with some bird species being more vulnerable than others away from cover. This can result in coexistence, provided that the most vulnerable species is also the best resource competitor near protective cover. The result is that each species has a microhabitat in which it is superior in competition. Coexistence on a spatially variable resource occurs between a pair of granivorous rodents in the Sonoran Desert (Brown 1989). Merriam’s kangaroo rat (Dipodomys merriami) and the roundtailed ground squirrel (Spermophilus tereticaudus) coexist in a creosote ﬂat because each can best exploit a diﬀerent distribution of resource patches. The high travel speed and travel eﬃciency of the ground squirrel allows it move frequently among patches and to better discover spatially dispersed, rich seed patches; the smaller body size of the kangaroo rat confers a lower overall metabolic cost of foraging and allows it to forage proﬁtably in poorer resource patches than can the ground squirrel. Coexistence on a temporally variable resource may describe interactions between annual and perennial plants (Shmida and Ellner 1984, Brown 1989). Coexistence is possible between a species superior at maintenance (a species with a low ﬁxed cost of existence that can survive from one peak in resource production to the next) and a foraging eﬃciency specialist (a species with a low variable cost of foraging that can forage proﬁtable even on relatively depleted resources). Annuals have extremely low costs of existence in their dormant seed stage, while perennials may be able to photosynthesize even when surface moisture is unavailable. Another example may be guilds of pollinators, such as bees, that compete for the daily pulses of nectar (Schaﬀer et al. 1979).

140

Living Components of Biodiversity: Organisms

Consumer–Resource Interactions (Predators, Pathogens, Parasitoids, Herbivores) In strong consumer–resource interactions, one expects a tendency toward instability when the resource itself is a biotic, renewing population. Such instability can lead to population cycles, chaotic dynamics, or localized extinctions. Strong predator–prey interactions certainly exist in desert biomes. However, in contrast to other biomes (e.g., boreal forest), we are unaware of any convincing case of pronounced, regular predator–prey cycles. It is useful to consider some reasons why this pattern (presuming for the sake of argument that it exists) may arise. Predator–prey cycles tend to be promoted by the conjunction of several factors: (1) high prey productivity (the ‘‘paradox of enrichment’’); (2) weak direct density-dependence; (3) trophic specialization; (4) temporal and spatial homogeneity. It seems likely that in many desert systems, one or more of these needed ingredients could be lacking. Low productivity systems are ones where by deﬁnition resources will be limiting for many species. It thus may be the case that many prey populations in deserts have insuﬃcient basic productivity to sustain predator–prey cycles. As noted above, many consumers have conspicuous interference competition, including cannibalism. Such mechanisms of direct density-dependence dampen consumer responses to changes in prey numbers. There are relatively few detailed food web studies in deserts, but those that have been attempted (e.g., Polis 1991c) reveal that many species are relatively generalized in their diet, and the webs as a whole are highly reticulate. In some circumstances, these factors reduce the likelihood of predator–prey cycles. Finally, spatial heterogeneity at small spatial scales can create refuges, which are broadly stabilizing. Temporal heterogeneity often implies that populations are most limited by ‘‘crunch’’ years of low resource abundance. This weakens the capacity of natural enemies to overexploit their victims during typical years. McNaughton et al. (1991) have observed that proportionately less primary production goes through herbivory in drylands than in other ecosystems. There are two basic explanations for this, both involving trophic dynamics: (1) a bottom-up explanation emphasizes the role of low overall plant productivity and temporal variability in productivity. Simple resource consumer models (Oksanen et al. 1981) suggest that herbivores persist with diﬃculty at extremely low productivity and, even if they do persist, it is at low abundances where they consume proportionately less standing crop. This eﬀect is exaggerated by temporal variability in productivity. Strong temporal variation in production leads to consumer populations regulated primarily by food availability in crunch years (Wiens 1977). This implies that in more typical years, there will be fewer herbivores than expected, given ambient production, and therefore total consumption will be proportionately less in highly variable environments in most years (Andrewartha and Birch 1954, 1984). (2) An alternative, top-down explanation is that the low plant cover in drylands

Unified Framework I 141

results in greater predation (or at least predation risk) on herbivores and consequently less herbivory (chapters 2 and 4, herein). We argued above that deserts tend to be dominated by generalist rather than specialist consumers. Numerical and aggregative responses by generalist consumers can put at risk the local persistence of prey populations, particularly of species with low intrinsic growth rates (Holt and Kotler 1987, Holt and Lawton 1994). This observation, taken alone, might suggest that predators will often reduce species richness of their prey in deserts. However, theory suggests that such eﬀects due to apparent competition often decline with decreasing productivity (Holt et al. 1994, Holt 1997). Moreover, given strong temporal variability, the abundance of generalist predators will be set during periods when overall prey abundance is low. Incorporating temporal variability into predator–prey models often reduces average predator abundance (Abrams et al. 1998). For generalist predators with saturating functional responses, in non-‘‘crunch’’ years a superabundance of prey will be present; alternative prey can then experience indirect mutualism, mediated through the saturated functional response of their shared predator (Holt 1997).

A Comment on Desert Food Webs and Diversity In the above paragraphs, we have reviewed theory and evidence on pairwise species interactions, emphasizing patterns and processes that are more conspicuous or important in deserts. But all these interactions are of course embedded in species-rich, complex food webs. As noted above, desert food webs often appear to be highly reticulate. In principle, increasing the number of links in food webs can increase the potential for indirect interactions. In practice, the complexity of such linkages can diﬀuse interactions, making it diﬃcult to discern in empirical studies the net eﬀect of one focal species upon another. The net sign of an interaction is likely to vary as a function of many factors, including the local species composition, habitat heterogeneity, and position along aridity gradients. Rather than review general theory on this topic, or speculate about how this body of theory may or may not bear on desert ecosystems, we simply note that describing and analyzing the dynamics of entire desert food webs is a challenge for future research.

Acknowledgments This chapter originated with lengthy discussion sessions at Sede Boqer at the meeting in 1999 (see preface for details), during which all the listed authors provided useful insights and text. Gary Polis, Deborah Goldberg, and Bob Holt took the lead in writing up these discussions, with considerable input from Burt Kotler. Our intent was to produce this chapter, then a much more detailed and reference-rich review paper. Unfortunately, this process was incomplete at the time of Gary’s death, and after some unavoidable delays, Deborah and Bob worked with the rough drafts and notes made available to them from Gary’s ﬁles so as to produce

142

Living Components of Biodiversity: Organisms

the current document. In our editing, we have tried to remain faithful to the collaborative and collegial spirit of the initial discussions, as well as to the overall vision that Gary had regarding the nature of desert communities. We have no doubt that the chapter would have evolved considerably beyond its current form had Gary been here to participate in this process.

References Abrams, P.A., R.D. Holt, and J.D. Roth. 1998. Apparent competition or apparent mutualism? Shared predation when populations cycle. Ecology 79: 201–212. Abramsky, Z., M.L. Rosenzweig, and S. Brand. 1985. Habitat selection in Israeli desert rodents: Comparison of a traditional and a new method of analysis. Oikos 45: 79–88. Abramsky, Z., M.L. Rosenzweig, B. Pinshow, J.S. Brown, B. Kotler, and W.A. Mitchell. 1990. Habitat selection: An experimental ﬁeld test with two gerbil species. Ecology 71: 2358–2369. Abramsky, Z., M.L. Rosenzweig, and B. Pinshow. 1991. The shape of a gerbil isocline: An experimental ﬁeld study using principles of optimal habitat selection. Ecology 72: 329–340. Alkon, P.U., and L. Olsvig-Whittaker. 1989. Crested porcupine digs in the Negev Desert Highlands—patterns of density, size, and longevity. Journal of Arid Environments 17: 83–95. Anderson, W.B., and G.A. Polis. 1999. Nutrient ﬂuxes from water to land: seabird eﬀects on plant quality on Gulf of California islands. Oecologia 118: 324–332. Andrewartha, H.G., and L.C. Birch. 1954. The Distribution and Abundance of Animals. University of Chicago Press, Chicago. Andrewartha, H.G., and L.C. Birch. 1984. The Ecological Web. More on the Distribution and Abundance of Animals. University of Chicago Press, Chicago. Armstrong, R.A., and R. McGehee. 1980. Competitive exclusion. American Naturalist 115: 151–170. Ayal, Y., and O. Merkl. 1994. Spatial and temporal distribution of tenebrionid species (Coleoptera) in the Negev Highlands, Israel. Journal of Arid Environments 27: 347–361. Belovsky, G.E. 1987. Extinction models and mammalian persistence. Pp. 35–57 in Viable Populations for Conservation. Cambridge University Press, New York. Belyea, L.R., and J. Lancaster. 1999. Assembly rules within a contingent ecology. Oikos 86: 402–416. Benson, L., and R.S. Thompson. 1987. The physical record of lakes in the Great Basin. Pp. 241–260 in W.F. Ruddiman and H.E. Wright, Jr. (eds.), North America and Adjacent Oceans During the Last Deglaciation, Decade of North American Geology series. Geological Society of America. Bertness, M.D., and R. Callaway. 1994. Positive interactions in communities. Trends in Ecology and Evolution 9: 191–193. Boeken, B., M. Shachak, Y. Gutterman, and S. Brand. 1995. Patchiness and disturbance: Plant community responses to porcupine diggings in the central Negev. Ecography 18: 410–422. Brown, J.H. 1986. Two decades of interaction between the MacArthur-Wilson model and the complexities of mammalian distribution. Biological Journal of the Linnean Society 28: 231–251.

Unified Framework I 143 Brown, J.H. 1989. Patterns, mode and extent of invasions by vertebrates. In J.A. Drake, H.A. Mooney, F. di Castri, R.H. Grove, F.J. Kruger, M. Rejmanek, and M. Williamson (eds.), Biological Invasions, pp. 85–99. John Wiley and Sons, New York. Brown, J.H., and D. Davidson. 1977. Competition between seed-eating rodents and ants in desert ecosystems. Science 196: 880–882. Brown, J.H., and A. Kodric-Brown. 1979. Convergence, competition, and mimicry in a temperate community of hummingbird-pollinated ﬂowers. Ecology 60(5): 1022–1035. Brown, J.H., and G.A. Lieberman. 1973. Resource utilization and coexistence of seedeating desert rodents in sand dune habitats. Ecology 54(4): 788–797. Brown, J. H., O. Reichman, and D. Davidson. 1979. Granivory in desert ecosystems. Annual Review of Ecology and Systematics 10: 201–227. Callaway, R.M. 1995. Positive interactions among plants. The Botanical Review 61: 306–349. Callaway, R.M., and L.R. Walker. 1997. Competition and facilitation: A synthetic approach to interactions in plant communities. Ecology 78: 1958–1965. Caswell, H. 1978. Predator-mediated coexistence: A nonequilibruim model. The American Naturalist 112: 127–155. Chesson, P. 2000. Mechanisms of maintenance of species diversity. Annual Review Ecology and Systematics 31: 343–366. Cody, M.L. 1974. Competition and the Structure of Bird Communities. Princeton University Press, Princeton, NJ. Cody, M.L. 1993. Bird diversity components within and between habitats in Australia. Pp. 147–158 in R.E. Ricklefs and D. Schluter (eds.), Species Diversity in Ecological Communities. University of Chicago Press, Chicago. Connell, J.H. 1975. Some mechanisms producing structure in natural communities: a model and evidence from ﬁeld experiments. Pp. 460–490 in M.L. Cody and J.M. Diamond (eds.), Ecology and Evolution of Communities. The Belknap Press of Harvard University, Cambridge, MA. Connell, J.H. 1980. Diversity and the coevolution of competitors, or the ghost of competition past. Oikos 35:131–138. Connell, J.H. 1983. On the prevalence and relative importance of interspeciﬁc competition: evidence from ﬁeld experiments. American Naturalist 122: 661–696. Cornell, H.V. and J.H. Lawton 1992. Species interactions, local and regional processes, and limits to the richness of ecological communities: a theoretical perspective. Journal of Animal Ecology 61: 1–12. Crawford, C. 1986. The role of invertebrates in desert ecosystems. In W. Whitford (ed.), Pattern and Process in Desert Ecosystems. University of New Mexico, Albuquerque. Crawley, M.J. 1986. The structure of plant communities. Pp. 1–50 in M.J. Crawley (ed.), Plant Ecology. Blackwell Scientiﬁc Publications, Oxford. Danin, A. 1977. Desert Vegetation of Israel and Sinai. Cana Publishing House, Jerusalem. Darwin, C. 1859. The Origin of Species. Oxford University Press, Oxford. DeAngelis, D.L. 1992. Dynamics of Nutrient Cycling and Food Webs. Chapman and Hall, New York. Eissenstat, D.M., and M.M. Caldwell. 1988. Competitive ability is linked to rates of water extraction: a ﬁeld study of two aridland tussock grasses. Oecologia 75: 1–7.

144

Living Components of Biodiversity: Organisms

Fowler, N. 1986. The role of competition in plant communities in arid and semiarid regions. Annual Review of Ecology and Systematics 17: 89–110. Fuentes, E.R., A.J. Hoﬀmann, A. Poiani, and M.C. Alliende. 1986. Vegetation change in large clearings: patterns in the Chilean matorral. Oecologia (Berlin) 68: 358–366. Garcia-Moya, E., and C.M. McKell. 1970. Contribution of shrubs to the nitrogen economy of a desert-wash plant community. Ecology 51: 81–88. Giller, P.S., and J.H.R. Gee. 1987. The analysis of community organization: the inﬂuence of equilibrium, scale and terminology. Pp. 519–542 in J.H.R. Gee, and P.S. Giller (eds.), Organization of Communities. British Ecological Society Symposium. Blackwell Scientiﬁc Publications, Oxford. Goldberg, D.E., T. Rajaniemi, J. Gurevitch, and A. Stewart-Oate. 1999. Empirical approaches to quantifying interaction intensity: competition and facilitation along productivity gradients. Ecology 80: 1118–1131. Grime, J.P. 1973. Competitive exclusion in herbaceous vegetation. Nature 242: 344–347. Grime, J.P. 1974. Vegetation classiﬁcation by reference to strategies. Nature 250: 26–31. Grime, J.P. 1977. Evidence for the existence of three primary strategies in plants and its relevance to ecological and evolutionary theory. American Naturalist 111: 1169–1194. Groner, E., and Y. Ayal. 2001. The interaction between bird predation and plant cover in determining habitat occupancy of darkling beetles. Oikos 93: 22–31. Grover, J.P. 1997. Resource Competition. Chapman and Hall, London. Gurevitch, J. 1986. Competition and the local distribution of the grass Stipa neomexicana. Ecology 67: 46–57. Gurevitch, J., L.L. Morrow, et al. 1992. A meta-analysis of competition in ﬁeld experiments. The American Naturalist 140: 539–572. Gutterman, Y. 1982. Observations on the feeding habits of the Indian crested porcupine (Hystrix indica) and the distribution of some hemicryptophyes and geophytes in the Negev Desert Highlands. Journal of Arid Environments 5: 261–268. Heth, G. 1991. The environmental impact of subterranean mole rats (Spalax ehrenbergi) and their burrows. Symposium of the Zoological Society of London 63: 265–280. Hobbs, R.J. and H.A. Mooney. 1995. Spatial and temporal variability in a Californian annual grassland: results from a long-tern study. Journal of Vegetation Science 6: 43–56. Holmgren, M., M. Scheﬀer, et al. 1997. The interplay of facilitation and competition in plant communities. Ecology 78: 1966–1975. Holt, R.D. 1984. Spatial heterogeneity, indirect interactions, and the coexistence of prey species. The American Naturalist 124: 377–406. Holt, R.D. 1985. Density-independent mortality, non-linear competitive interactions, and species coexistence. Journal of Theoretical Biology 116: 479–493. Holt, R.D. 1993. Ecology at the mesoscale: The inﬂuence of regional processes on local communities. Pp. 77–88 in R. Ricklefs and D. Schluter (eds.), Species Diversity in Ecological Communities. University of Chicago Press, Chicago. Holt, R.D. 1997. Community modules. Pp. 333–349 in A.C. Gange and V.K. Brown (eds.), Multitrophic Interactions in Terrestrial Ecosystems, 36th Symposium of the British Ecological Society. Blackwell Science, Oxford.

Unified Framework I 145 Holt, R.D., and M. Barﬁeld. 2001. On the relationship between the ideal free distribution and the evolution of dispersal. Pp. 83–95 in J. Clober, E. Danchin, A.A. Dhondt, and J.D. Nichols (eds.), Dispersal, Oxford University Press, New York. Holt, R.D., and B.P. Kotler. 1987. Short-term apparent competition. The American Naturalist. 130: 412–430. Holt, R.D., and J.H. Lawton. 1994. The ecological consequences of shared natural enemies. Annual Review of Ecology and Systematics 25: 495–520. Holt, R.D., J. Grover, and D. Tilman. 1994. Simple rules for interspeciﬁc dominance in systems with exploitative and apparent competition. American Naturalist 144: 741–777. Holt, R.D., J.H. Lawton, G.A. Polis, and N. Martinez. 1999. Trophic rank and the species-area relation. Ecology 80: 1495–1504. Huston, M.A. 1999. Local processes and regional patterns: appropriate scales for understanding variation in the diversity of plants and animals. Oikos 86: 393–401. Inouye, R. 1991. Population biology of desert annual plants. Pp. 27–54 in G. Polis (ed.), The Ecology of Desert Communities. The University of Arizona Press, Tucson. Jones, C.G., J.H. Lawton, and M. Shachak. 1994. Organisms as ecosystem engineers. Oikos 69: 373–386. Jones, C.G., J.H. Lawton, and M. Shachak. 1997. Positive and negative eﬀects of organisms as physical ecosystem engineers. Ecology 78: 1946–1957. Keddy, P.A. 1990. Competition hierarchies and centrifugal organization in plant communities. Pp. 265–290 in J. Grace and D. Tilman (eds.), Perspectives on Plant Competition. Academic Press, New York. Kotler, B.P. 1984. Risk of predation and the structure of desert rodent communities. Ecology 65: 689–701. Kotler, B.P., and J.S. Brown. 1988. Environmental heterogeneity and the coexistence of desert rodents. Annual Review of Ecology and Systematics 19: 281–307. Kotler, B.P., J.S. Brown, R.J. Smith, and W.O. Wirtz II. 1988. The eﬀects of morphology and body size on the costs of predation in desert rodents. Oikos 53: 145–152. Kotler, B.P., J.S. Brown, and W.A. Mitchell. 1993. Environmental factors aﬀecting patch use in gerbilline rodents. Journal of Mammal 74: 614–620. Kruckeberg, A.R. 1969. Soil diversity and the distribution of plants, with examples from western North America. Madrono 20: 129–154. Latham, R.E., and R.E. Ricklefs. 1993. Continental comparisons of temperate-zone tree species diversity. Pp. 294–314 in R.E. Ricklefs and D. Schuluter (eds.), Species Diversity in Ecological Communities. University of Chicago Press, Chicago. Lieth, H., 1978. Primary productivity in ecosystems: comparative analysis of global patterns. Pp. 300–321 in H. Lieth (ed.), Patterns of Primary Production in the Biosphere. Volume 8 in Benchmark Papers in Ecology, Dowden, Hutchinson and Ross, Stroudsburg, PA. Liphschitz, N., and Y. Waisel. 1982. Adaptation of plants to saline environments: salt excretion and glandular structure. Pp. 197–214 in D.N. Sen and K.S. Rajpurohit (eds.), Tasks For Vegetation Science. Dr. W. Junk Publishers, The Hague. Louw, N., and M. Seely. 1982. Ecology of desert organisms. Longman, London. Ludwig, J.A. 1986. Primary production variability in desert ecosystems. Pp. 5–17 in W.G. Whitford (ed.), Patterns and Processes in Desert Ecosystems. University of New Mexico Press, Albuquerque.

146

Living Components of Biodiversity: Organisms

Ludwig, J.A. 1987. Primary production in arid lands: myths and realities. Journal of Arid Environments 13: 1–7. MacArthur, R.H. 1972. Geographical Ecology: Patterns in the Distribution of Species. Harper and Row, New York. MacArthur, R.H., and E.O. Wilson. 1967. The Theory of Island Biogeography. Princeton University Press, Princeton, NJ. MacKay, W. 1991. The role of ants and termites in desert communities. Pp. 113–150 in G. Polis (ed.), The Ecology of Desert Communities. The University of Arizona Press, Tucson. McNaughton, S.J., Oesterheld, M., Frank, D.A., and K.J. Williams. 1991. Primary and secondary production in terrestrial ecosystems. Pp. 120–140 in J. Cole, G. Lovett, and S. Findlay (eds.), Comparative Analysis of Ecosystems. SpringerVerlag, New York. Menge, B.A., and J.P. Sutherland. 1987. Community regulation: variation in disturbance, competition, and predation in relation to environmental stress and recruitment. The American Naturalist 130(5): 730–757. Miller, R.S. 1967. Pattern and process in competition. Advances in Ecological Research 4: 1–74. Mitchell, W.A., Abramsky, Z., Kotler, B.P., Pinshow, B., and J.S. Brown. 1990. The eﬀect of competition of foraging activity in desert rodents: theory and experiments. Ecology 71: 844–854. Newman, E.I. 1973. Competition and diversity in herbaceous vegetation. Nature 244: 310–311. Noy-Meir, I. 1973. Desert ecosystems: Environment and producers. Annual Review of Ecology and Systematics 4: 25–51. Noy-Meir, I. 1974. Desert ecosystems: higher trophic levels. Annual Review of Ecology and Systematics 5: 195–214. Noy-Meir, I. 1981. Theoretical dynamics of competitors under predation. Oecologia 50: 277–284. Noy-Meir, I. 1985. Desert ecosystem structure and function. Pp. 93–103 in M. Evenari, I. Noy-Meir, and D. Goodall (eds.), Hot Deserts and Arid Shrublands, A. Elsevier Scientiﬁc Publications, Amsterdam. Oksanen, L., S.D. Fretwell, et al. 1981. Exploitation ecosystems in gradients of primary productivity. The American Naturalist 118: 240–261. Parsons, P.A. 1996. Competition versus abiotic factors in variably stressful environments: evolutionary implications. Oikos 75: 129–132. Pimm, S.L., Russell, G.J., Gittleman, J.L., and T.M. Brooks. 1985. The future of biodiversity. Science 269: 347–350. Polis, G.A. 1980. The signiﬁcance of cannibalism on the population dynamics and surface activity of a natural population of desert scorpions. Behavioral Ecology and Sociobiology 7: 25–35. Polis, G.A. 1981. The evolution and dynamics of intraspeciﬁc predation. Annual Reviews of Ecology and Systematics 12: 225–251. Polis, G.A. 1991a. Desert communities: an overview of patterns and processes. Pp. 1–26 in G. Polis (ed.), The Ecology of Desert Communities. The University of Arizona Press, Tucson. Polis, G.A. 1991b. The Ecology of Desert Communities. The University of Arizona Press, Tucson. Polis, G.A. 1991c. Complex trophic interactions in deserts: An empirical critique of food web theory. American Naturalist 138: 123–155.

Unified Framework I 147 Polis, G.A. 1999. Why are parts of the world green? Multiple factors control productivity and the distribution of biomass. Oikos 86: 3–15. Polis, G.A., and R.D. Farley. 1980. Population biology of a desert scorpion— survivorship, microhabitat, and the evolution of life-history strategy. Ecology 61: 620–629. Polis, G.A. and T. Yamashita. 1991. The ecology and importance of predaceous arthropods in desert communities. Pp. 180–222 in G.A. Polis (ed.). The Ecology of Desert Communities. University of Arizona Press, Tucson. Polis, G.A., C.A. Myers, and R.D. Holt, 1989. The ecology and evolution of intraguild predation: Potential competitors that eat each other. Annual Review of Ecology and Systematics 20: 297–330. Polis, G.A., R.D. Holt, B.A. Menge, and K. Winemiller. 1996. Time, space and life history: Inﬂuences on food webs. Pp. 435–460 in G.A. Polis and K. Winemiller (eds.), Food Webs: Integration of Patterns and Dynamics. Chapman and Hall, New York. Polis, G.A., S.D. Hurd, C.T. Jackson, and F. Sanchez-Pinero. 1997a. El Nin˜o eﬀects on the dynamics and control of a terrestrial island ecosystem in the Gulf of California. Ecology 78: 1884–1897. Polis, G.A., W. Anderson, and R.D. Holt. 1997b. Toward and integration of landscape ecology and food web ecology: the dynamics of spatially subsidized food webs. Annual Review of Ecology and Systematics 28: 289–316. Polis, G.A., J.D. Barnes, M.K. Seely, J.R. Henschel, and M.M. Enders. 1998. Predation as a major cost of reproduction in Namib Desert tenebrionid beetles. Ecology 79: 2560–2566. Polis, G.A., M.E. Power, and G.R. Huxel (eds.). 2003. Food Webs at the Landscape Level. University of Chicago Press, Chicago. Pulliam, R.H. 1988. Sources, sinks, and population regulation. The American Naturalist 132: 652–661. Pulliam, R.H., and G.S. Mills. 1977. The use of space by wintering sparrows. Ecology 58: 1393–1399. Ricklefs, R.E., and D. Schluter. 1993. Species Diversity in Ecological Communities: Historical and Geographical Perspectives. The University of Chicago Press, Chicago. Riechert, S.E. 1991. Prey abundance versus diet breadth in a spider test system. Evolution Ecology 5: 327–338. Ritchie, M.E. 1999. Biodiversity and reduced extinction risks in spatially isolated rodent populations. Ecology Letters 2: 11–16. Robberecht, R., B.E. Mahall, and P.S. Nobel. 1983. Experimental removal of intraspeciﬁc competitors: Eﬀects on water relations and the productivity of a desert bunchgrass, Hilaria rigida. Oecologia 60: 21–24. Robertson, G.P., J.R. Crum, and B.G. Ellis. 1993. The spatial variability of soil resources following long-term disturbance. Oecologia 96: 451–456. Robertson, G.P., M.A. Huston, F.C. Evans, and J.M. Tiedje. 1988. Spatial variability in a successional plant community—patterns of nitrogen availability. Ecology 69: 1517–1524. Ronce, O., I. Olivieri, J. Clobert, and E. Danchin. 2001. Perspectives on the study of dispersal evolution. Pp. 341–357 in J. Clober, E. Danchin, A.A. Dhondt, and J.D. Nichols (eds.), Dispersal. Oxford University Press, New York. Rosenzweig, M.L. 1973. Evolution of the predator isocline. Evolution 27: 84–94. Rosenzweig, M.L. 1978. Competitive speciation. Biological Journal of the Linnean Society 10: 275–289.

148

Living Components of Biodiversity: Organisms

Rosenzweig, M. 1995. Species Diversity in Space and Time. Cambridge University Press, Cambridge. Rosenzweig, M.L., and Z. Abramsky 1986. Centrifugal community organization. Oikos 46: 339–348. Schaﬀer, W.M., D.B. Jensen, D.E. Hobbs, J. Gurevitch, J.R. Todd, and M.V. Schaﬀer. 1979. Competition, foraging energetics, and the cost of sociality in three species of bees. Ecology 60: 976–987. Schoener, T.W. 1983. Field experiments on interspeciﬁc competition. American Naturalist 122: 240–285. Seely, M. 1991. Sand dune communities. Pp. 348–382 in G. Polis (ed.), The Ecology of Desert Communities. University of Arizona Press, Tucson. Seely, M., and N. Louw. 1980. First approximation of the eﬀects of rainfall on the ecology and energetics of a Namib Desert dune ecosystem. Journal of Arid Environments 3: 25–54. Shachak, M., Sachs, M., and I. Moshe. 1998. Ecosystem management of desertiﬁed shrublands in Israel. Ecosystems 1: 475–483. Shmida, A., and S. Ellner. 1984. Coexistence of plant species with similar niches. Vegetatio 58: 29–55. Shmida, A., and M.V. Wilson. 1985. Biological determinants of species diversity. Journal of Biogeography 12: 1–20. Sih, A., P. Crowley, et al. 1985. Predation, competition, and prey communities: a review of ﬁeld experiments. Annual Review of Ecology and Systematics 16: 269– 311. Taylor, A.D. 1988. Large scale spatial structure of population dynamics in arthropod predator-prey systems. Annales Zoologici Fennici 25: 63–74. Tilman, D. 1988. Plant Strategies and the Dynamics and Structure of Plant Communities: Monographs in Population Biology 26. Princeton University Press, Princeton, NJ. Tilman, D. 1990. Mechanisms of plant competition for nutrients: the elements of a predictive theory of competition. Pp. 117–141 in J. Grace, and D. Tilman (eds.), Perspective on Plant Competition. Academic Press, New York. Tilman, D. and S. Pacala. 1993. The maintenance of species richness in plant communities. Pp.13–25 in R.E. Ricklefs and D. Schluter (eds.), Species Diversity in Ecological Communities: Historical and Geographical Perspectives. University of Chicago Press, Chicago. Whitford, W.G. 1991. Subterranean termites and long-term productivity of desert rangelands. Sociobiology 19: 235–243. Wiens, J.A. 1977. On competition and variable environments. American Scientist 65: 590–597. Wiens, J. 1991. The ecology of desert birds. Pp. 278–310 in G. Polis (ed.), The Ecology of Desert Communities. The University of Arizona Press, Tucson. Wiens, J.A., J.T. Rotenberry, and B. Van Horne. 1986. Lesson in the limitations of ﬁeld experiments: shrubsteppe birds and habitat alteration. Ecology 67: 365–376. Wisdom, C.S. 1991. Patterns of heterogeneity in desert herbivorous insect communities. Pp. 151–179 in G.A. Polis (ed.), The Ecology of Desert Communities. The University of Arizona Press, Tucson. Woodward, B., and S. Mitchell. 1991. The community ecology of desert anurans. Pp. 223–248 in G.A. Polis (ed.), The Ecology of Desert Communities. The University of Arizona Press, Tucson.

Unified Framework I 149 Yair, A. and M. Shachak. 1987. Studies in watershed ecology of an arid area. Pp. 146– 193 in M.O. Wurtele and L. Berkofsky (eds.), Progress in Desert Research. Rowman and Littleﬁeld, Totowa, NJ. Ziv, Y., Z. Abramsky, B.P. Kotler, and A. Subach. 1993. Interference competition and temporal and habitat partitioning in 2 gerbil species. Oikos 66: 237–246. Zobel, M. 1992. Plant species coexistence—the role of historical, evolutionary and ecological factors. Oikos 65: 314–320.

This page intentionally left blank

Part II

Ecological Complexes of Biodiversity, Ecosystems, and Landscapes

This page intentionally left blank

9 Species Diversity and Ecosystem Processes in Water-Limited Systems Moshe Shachak Steward T.A. Pickett James R. Gosz

T

here are many relationships between ecosystem properties and species (Jones and Lawton, 1995) with the potential links described by ﬁve hypotheses: 1. The null hypothesis claims that there is no eﬀect of species diversity on ecosystem processes. The following hypotheses imply biological mechanisms. 2. The diversity–stability hypothesis predicts that ecosystem productivity and recovery increase as the number of species increases (Johnson et al. 1996). 3. The rivet hypothesis predicts a threshold in species richness, below which ecosystem function declines steadily and above which changes in species richness are not reﬂected by changes in ecosystem function (Ehrlich and Ehrlich 1981; Vitousek and Hooper 1993). 4. The redundant species hypothesis states that species loss has little eﬀect on ecosystem processes if the losses are within the same functional group (Walker 1992) 5. The idiosyncratic response hypothesis suggests that as diversity changes so do ecocosystem processes (Lawton 1994, Lawton and Brown 1994). There have been both ﬁeld and laboratory attempts to test these hypotheses, (Naeem and Li 1998), however, the interpretation and the generality of the results remain contentious (Tilman 1999). A fundamental 153

154

Ecological Complexes of Biodiversity, Ecosystems, and Landscapes

reason for such uncertainty is that the hypotheses are not driven by a comprehensive theory of the relationship between species properties and ecosystem processes (Tilman et al., 1997). We propose that the foundations for the necessary theory are in models of the distribution of resources and their utilization by organisms. This is because ecosystem processes such as primary production, decomposition, mineralization, and evapotranspiration are dependent on the processing of resources by the species that are producers, consumers, and decomposers. A theory that links the direct participation of species in ecosystem processes may resolve diﬀerences among the various hypotheses or identify how they complement each other. From a community perspective, a theory of resource utilization is based on two alternative assumptions: 1. The rate of ecosystem processes is determined by the few species that are most eﬃcient in using and converting resources. For example, in a desert system, dominant species are those that are proﬁcient in using water for biomass production or in converting inorganic matter into organic materials. 2. The rate of ecosystem processes is determined by complementarity in resource use by diﬀerent species. Under this assumption, the rate of ecosystem processes of an assemblage of species should be higher than the rate determined by a few dominant species. There is some evidence to support each of the two contrasting assumptions. Ecological studies of natural communities provide some empirical evidence that productivity is determined by few species (Hooper and Vitousek 1997). In contrast, agricultural practice suggests that species complementarity can be of importance in determining ecosystem processes (Vandermeer 1989). However, there is no systematic body of studies that allows either assumption to be dismissed or limited in application. Therefore, both have to be considered in developing the theory of biodiversity and ecosystem function. Here, we present a conceptual model that links species properties and ecosystem processes for water-limited systems. The model is based on the diﬀerences among species in using water for biomass production and the spatial nonuniformity of soil moisture distribution. The model demonstrates that the four biological hypotheses on the relationship between species diversity and ecosystem function are not mutually exclusive. The model, which integrates dryland hydrology with the idea of individualistic species responses along environmental gradients (Whittaker 1967), enables us to show how and when to apply the diversity–stability, rivet, redundant species, and idiosyncratic response hypotheses when generating speciﬁc hypotheses on the relationship between species and ecosystem processes in water-limited systems.

Species Diversity and Ecosystem Processes in Water-Limited Systems

155

The Model Water, Species, and Productivity We take primary productivity as the focal ecosystem process. We suggest that the relationship between species diversity and primary productivity in drylands is intimately related to water ﬂow and place this relationship within a framework of a multiﬂow system (Shachak and Jones 1995) that integrates species, water, and energy ﬂows. In our model, we refer to water ﬂow (ﬁg. 9.1, ﬂow B) as the set of processes that determine the spatial and temporal distribution of soil moisture. The species ﬂow (ﬁg. 9.1, ﬂow A) refers to the colonization and extinction of species in a speciﬁc area. Colonization is possible due to the availability of a species pool in the region that is larger than the number of species in the speciﬁc local area. Similarly, the energy ﬂow (ﬁg. 9.1, ﬂow C) refers to biomass production by all species that occupy the speciﬁc area. Combining these three chains into a systems model shows how ecosystem productivity is controlled by interactions among soil moisture, species diversity, and biomass (ﬁg. 9.1). This model is the mechanistic foundation for deriving hypotheses that link species diversity to ecosystem function. We see the relationship between the ﬂow chains of species, soil moisture, and productivity at two scales. The coarse scale shows the inputs, which

Figure 9.1 The relationship among species diversity and the ecosystem processes of water ﬂow and primary production. A: species ﬂow; B: water ﬂow, C: energy ﬂow. O: ﬂow controller. (The numerical and alphabetic labels in the ﬁgure are used in the text to describe parts of the model.)

156

Ecological Complexes of Biodiversity, Ecosystems, and Landscapes

consist of a species pool, rainfall, run-oﬀ and solar energy. The species pool provides the species for the local assemblage, and rainfall and run-oﬀ provide the soil moisture. The ﬁner scale shows the feedbacks within the system, consisting of a web of interactions among soil moisture, species assemblage, and plant biomass. The model shows that the species assemblage that aﬀects ecosystem process biomass production is controlled by soil moisture (ﬁg. 9.1, 3A) and by plant biomass (ﬁg. 9.1, 4A). For example, high soil moisture can increase the species richness, while high plant biomass may increase species richness by reducing temperature or decrease richness by competition for reduced moisture. We can view these two controllers as the eﬀects of ecosystem processes on species diversity. However, the model demonstrates that species control ecosystem processes such as water (ﬁg. 9.1, 2B) and energy (ﬁg. 9.1, 2C) ﬂow. In the debate whether plant community diversity depends on productivity (Grime 1997, Huston 1997) or productivity depends on diversity (McNaughton 1993) our model adopted the approach that causation is in both directions (Tilman et al. 1996). The model demonstrates that the ecosystem eﬀects of the species assemblage are actually a system that links three ﬂows: species, energy, and materials. Species Assemblage Ecosystem processes are the product of a web of interactions. There are two steps in the creation of this web of interactions between species. The ﬁrst step is the generation of an assemblage of species in an area due to colonization and immigration. This is a process of ﬁltering a subset of species from the available species pool in the area. For plants, this stage is dependent on propagule arrival from the species pool and on-site environmental conditions for germination, establishment, and growth (Pickett et al., 1989). For fauna, processes of immigration, emigration and internal population controls are important. In water-limited systems, the environmental conditions are usually related to soil moisture availability. The second step is the formation of a web of interactions that controls ecosystem dynamics. The web of interactions should determine the ability of species to survive by producing, using, and sharing resources. The organization of the assemblage is the main controller over ecosystem processes (ﬁg. 9.1, controllers 2b and 2c). The organization and complementarity of resource use of the species assemblage in relation to ecosystem processes is determined by the number of species, and their abundance and composition (Hooper and Vitousek 1997). These three properties of the species assemblage can be aggregated in the concept of diversity if we adopt the deﬁnition that diversity encompasses the number of entities and the diﬀerences and similarities among them (Gaston 1996). Species number and abundance captures the number of components of diversity, while composition captures the diﬀerences and

Species Diversity and Ecosystem Processes in Water-Limited Systems

157

similarities among species. Composition implies diﬀerences or complementarity in species traits that drives ecosystem processes. In a water-limited system, diﬀerences in traits such as water-use eﬃciency or drought resistance can have a profound eﬀect on ecosystem processes related to biomass production. Species Diversity Along A Soil Moisture Gradient We adopt the individualistic approach to describe the distribution and abundance of species along a soil moisture gradient (Whittaker 1967, Keddy 1991). Adopting Whittaker’s view implies that the abundance of a species along a moisture gradient is unimodal with a maximum at intermediate levels of moisture availability. Using this approach enables us to describe species diversity of a species assemblage, in terms of both species numbers and the diﬀerences among species in abundance, on various ranges of soil moisture. The diﬀerence among species over a given range of moisture is fundamental to the expression of diversity in arid lands. To evaluate the eﬀect of a given assemblage on an ecosystem process (e.g., primary production), we combine the species distribution with a property that controls its ecosystem function. For a species in a water-limited system we assume that primary production will be higher in areas with species assemblages that are more eﬃcient in using water for biomass production (controller 2B). Thus, we add their water eﬃciency use along a water availability gradient to the distribution and abundance of the species (ﬁg. 9.2). Here we can assume that biomass production per unit available water for a species is unimodal with a maximum at intermediate levels of moisture availability (Pastor and Bridgham 1999). We can draw a species diversity diagram that combines species responses in terms of abundance and water use eﬃciency along a soil moisture gradient (ﬁg. 9.2). The diagram is constructed to expose diﬀerences among species, not just species number. In this diagram the diﬀerences among species are (1) their ability to use water for converting solar energy into biomass (biomass production per unit water under a given soil moisture availability), and (2) their distribution range and abundance along the soil moisture gradient. Formation of the Soil Moisture Gradient Our expression of plant species diversity is deﬁned in relation to soil moisture heterogeneity (ﬁg. 9.2). Thus, factors controlling the spatial and temporal distribution of soil moisture are essential to our conceptual scheme of species–ecosystem relationships. Specifying diversity in this way is a necessary step in testing hypotheses about the relationship of diversity and ecosystem function. Hydrological studies show that the spatial and temporal distribution of soil moisture is controlled by several factors: (1) spatiotemporal variability in

158

Ecological Complexes of Biodiversity, Ecosystems, and Landscapes

Figure 9.2 Hypothetical map of species relationships along a soil moisture gradient. The scheme shows three species that diﬀer in their distribution range, abundance, and ecosystem-related properties. This representation shows both the resource use diﬀerence and the richness components of diversity. Each species has maximum density and water-use eﬃciency at an intermediate level of soil moisture for its range.

rainfall on various scales; (2) physical patchiness on various scales that controls input and output of water into and from the soil; and (3) biological patchiness at various scales and its eﬀects on input and output of water into and from the soil (Shachak et al. 1999). Each of these controllers is determined by additional factors. For instance, variability in rainfall is patchy on a scale of kilometers (Morin et al. 1999) Rainfall can also be patchy on a single slope (Sharon 1980). Rainfall distribution changes through time during a single rainfall event as a storm moves across a landscape. The distribution of rain varies during a given rainy season, and also varies between years in relation to the physics of the atmosphere. The physical patchiness of input and output of water on the land can be controlled by the rock to soil ratios, and slope directions within watersheds (Yair and Shachak 1987). On a coarser scale, the spatial arrangement and size of watersheds aﬀects water input (Yair, 1999). Each of these parameters controls the frequency and magnitude of evaporation and run-oﬀ, as well as source–sink relationships at

Species Diversity and Ecosystem Processes in Water-Limited Systems

159

various spatial scales. In drylands, the mosaic of vegetation caused by individual or aggregated shrubs, with herbaceous plants in the intershrub space, exposed rock surfaces, and the crusted soil are main controllers of soil moisture heterogeneity on a small scale (Le Roux et al. 1995). Spatial organization of woody plants in relation to herbaceous plants is a major determinant of soil moisture distribution on larger scales such as slopes, wadis, and watersheds (Ludwig and Tongway 1995). Species Overlap and Ecosystem Processes The degree of specialization and overlap in water utilization by plant species along a soil moisture gradient have implications for the species diversity– ecosystem processes hypotheses. We illustrate the simplest relationship by analyzing two species and their performances (ﬁg. 9.3). We analyze the eﬀect of the two species on productivity with respect to spatial scale of their distribution and in relation to their combined eﬀect in the range in which the two species overlap. To separate between the eﬀect of spatial scale of observation and species overlap on species–ecosystem processes hypothesis we ﬁrst analyze a case of two species with no overlap. When the two species of the assemblage specialize on two diﬀerent ranges of the soil moisture gradient there is no overlap (ﬁg. 9.3A). In this case, the concept of species diversity is applicable if the scale of observation includes the soil moisture range of the two species. Here species diversity is two species that diﬀer in their distribution along a soil moisture gradient and their physiological trait–water use. Under these conditions, the hypothesis that predicts that ecosystem productivity increases as the number of species increases is applicable (the rivet hypothesis). When there is no species overlap there is no redundancy and also no insurance that one species will replace another when a species becomes extinct. The situation is more complex when the two species overlap in part of their distribution range (ﬁg. 9.3B). Looking over the scale of all their distribution ranges we see that ecosystem productivity increases as the number of species increases because high productivity occurs over a broader range of soil moisture. However, if we look only at the soil moisture range of species overlap there are several ways to view the relationship between species diversity and ecosystem function. We can focus on soil moisture ranges in which one species is superior to the other in water-use eﬃciency. In a spatial scale which is smaller then the entire range of the two species we ﬁnd a soil moisture range over which species a is better than species b and a soil moisture range over which species b is better than species a in water utilization for biomass production. We suggest that in the range over which species a is better than species b the losses of species b have a positive eﬀect on ecosystem productivity. This is because the water used by species b can be now used more eﬃciently for production by species a. We suggest the opposite situation in the soil moisture range over which species b is better than species a in water utilization for biomass production.

160

Ecological Complexes of Biodiversity, Ecosystems, and Landscapes

Figure 9.3 Scale dependence of ecosystem eﬀect of species diversity (two annual species; no complementarity in water use). A: Species with no overlap in distribution along soil moisture gradient. B: Species with overlap in distribution along soil moisture gradient. sp. a: species a; sp. b: species b; S: scale of species diversity–ecosystem processes analysis (S1: analysis over the whole range of species distribution; S2: analysis over the range of species overlap; S3: analysis over the range that species a is more eﬃcient in water use; S4: analysis over the range that species b is more eﬃcient than species a in water use).

This analysis demonstrates that in areas in which species overlap and they are not complementary in water utilization, there is a cost to ecosystem functioning for higher species number. This implies that ecosystem production decreases as the number of species increases. This cost derives from the tendency of species to diﬀer in their eﬃciency in water use in areas where they overlap in their distribution range. In this range the losses of species that are inferior in water use have a positive eﬀect on ecosystem productivity.

Species Diversity and Ecosystem Processes in Water-Limited Systems

161

The area of overlap can also be viewed from a temporal perspective. Assuming a ﬂuctuating environment that causes an extinction of species a in the overlap area species b provides a guarantee that some production will be maintained even if species a becomes extinct. This view of species overlap is in accordance with the insurance hypothesis. This hypothesis claims that species diversity insures ecosystems against declines in their functioning because many species provide greater guarantees that some will maintain functioning even if others fail. We see that the area along the soil moisture gradient over which the two species overlap can be viewed from present and future perspectives. At present, given a speciﬁc set of environmental parameters, there is a cost of species overlap. A species less eﬃcient in water utilization is part of the assemblage and reduces the potential rate of primary production by the more proﬁcient species. However, a future-oriented view that takes into account species extinction from an assemblage as a part of ecosystem dynamics sees species overlap as an insurance of a continuation of ecosystem function in case of species extinction. In a very short range over their distribution, where the two curves of biomass production per unit water of the two species intersect, we refer to another hypothesis. In this range, the performance of the two species in relation to biomass production is equal. Therefore, we claim that here, there is a redundancy in species performance. Implications for Research Recently, ecologists have begun to ask how organismal diversity aﬀects ecosystem performance (Schlapfer and Schmid 1999, Tilman 1999). Studies on this subject lack clear deﬁnitions of diversity that are relevant to ecosystem processes and procedures for generating and testing hypotheses important to those relationships. Our simpliﬁed conceptual model has implications for deﬁning species diversity and for generating and testing hypotheses in the context of species and ecosystem interactions. In particular, our model can help in generating and analyzing hypotheses on the relationship between productivity and species diversity in water-limited systems. The traditional concept of species diversity, which was developed as a community ecology concept, refers either to the number of species (species richness) or to indices that combine the number of species and the diﬀerences among them in the number of individuals per species. If we adopt the idea that the concept of diversity encompasses a variety of diﬀerences among entities (Gaston 1996) then we should ask what types of diﬀerences should be included in our deﬁnition of species diversity. We suggest including diﬀerences in traits that are relevant to the variables in ecological system that respond to species diversity. In studies of ecosystem eﬀects of species diversity, diﬀerences among species such as diﬀerences in litter chemistry or nutrient uptake eﬃciency can be included in the deﬁnition of species diversity. The inclusion of diﬀerences in

162

Ecological Complexes of Biodiversity, Ecosystems, and Landscapes

physiological traits in relation to resource acquisition can help in identiﬁcation of the mechanistic relationships between species properties and ecosystem processes. To generate and test hypotheses on the inﬂuences of species on ecosystem processes, investigators are motivated by questions such as: Is there a relationship between plant species diversity and ecosystem function? Is the relationship positive or negative for a given assemblage? (Schlapfer and Schmid 1999.) Our analysis shows that we expect a relationship if the deﬁnition of species diversity takes into account the diﬀerence in species distribution range along a gradient of soil moisture and their diﬀerence in water utilization. This deﬁnition of species diversity integrates physiological traits (water utilization) with community (species distribution) and ecosystem attributes (soil moisture heterogeneity). We suggest that the chance of ﬁnding a diversity–ecosystem relationship is dependent on how many of the diﬀerences among species are presented in the deﬁnition of species diversity and how many of those diﬀerences are related to ecosystem processes. This approach of considering the diﬀerences in ecosystem-related properties of species is common in studies aimed at comparing the role of the number of species versus their composition in ecosystem processes. An example is the study of the ecosystem consequences of species assemblages on nitrogen ﬂux with and without legumes (Hooper and Vitousek 1997, Tilman 1999). To generate speciﬁc hypotheses on whether relationships between species diversity and productivity are positive or negative is a more diﬃcult task. The hypotheses are dependent on the assumptions of water accessibility and utilization by a diverse plant assemblage and scale of analysis. Our model assumes partitioning in space among species along a soil moisture gradient and no complementarity in water use in the areas of species overlap. It shows negative relationships in areas of species overlap, but a positive relationship if the scale of observation is over the whole range of the species distribution. Our model is applicable to studies of desert annuals because annuals use the same moisture layer and there is little complementarity in water use. The model on positive or negative relations will change if the species diversity is related to two functional types such as shrubs and annuals (ﬁg. 9.4). In this case, complementarity in water use is expected for a given spatial area. Shrubs extract a substantial proportion of their moisture from deeper soil layers, while annuals extract mainly from shallower layers (Breshears et al. 1999). In this case a positive relation between diversity and productivity is expected. In conclusion, we suggest that before generating a speciﬁc hypothesis to be experimentally tested, we need a detailed analysis of the diversity deﬁnition that includes traits related to ecosystem properties, as well as diﬀerences in those traits at the organism and assemblage scales. This is in addition to specifying ecosystem response variables (Schlapfer and Schmid 1999). The experimental approach for testing hypotheses should be related to the nature of the hypotheses. The prevailing experimental approach for testing hypotheses on the relationship between species and ecosystem processes is to

Species Diversity and Ecosystem Processes in Water-Limited Systems

163

Figure 9.4 Scale dependence of ecosystem eﬀect of species diversity (for two annual species and one shrub species). S: scale of species diversity–ecosystem processes analysis (S1–S4: scales of species ecosystem processes analysis for two annual species (a and b) with no complementarity in water use (see ﬁg. 8.3 for details); S5–S8: new scales of analysis when a new shrub species (c) is added). S5: the shrub uses deeper water than the annuals. (In this range there is a positive relationship between species number and biomass production analysis over the whole range of species distribution. S2: analysis over the range of species overlap; S3: analysis over the range that species a is more eﬃcient in water use; S4: analysis over the range that species b is more eﬃcient than species a in water use.) S6, S7, S8: the shrub uses the same water layer as the annuals. (In all these ranges the relationships between species number and biomass production are negative.)

manipulate the number of species or functional groups as an independent variable and to measure ecosystem processes, such as productivity, decomposition, mineralization etc., as response variables. The results from such studies show how species richness and composition are related to the rate of ecosystem processes. The approach is an attempt to predict ecosystem responses by counting the number of species and/or their functional traits that are involved in the processes. However, when the properties of the species in relation to resource processing are of importance, then manipulation should take into account the differences among species traits in resource processing as well as the number of species. A step toward the integration of the diﬀerences in resource processing among species into species-ecosystem studies are the experimental studies of the relationship between functional groups and ecosystem processes (Collins and Benning 1996). Our approach requires an expansion of the usual experimental approach to studying species–ecosystem relationships. The experimental approach for

164

Ecological Complexes of Biodiversity, Ecosystems, and Landscapes

testing species richness–ecosystem process relationships is basically a comparative study of the diﬀerences in the rates of ecosystem processes among species assemblages over a uniform environment. Our model requires an extension of the standard experimental approach into new directions. The ﬁrst is the study of ecosystem eﬀects related to individual species properties. In drylands this implies experimentally studying water use and biomass production of individual species along a soil moisture gradient. The second extension requires the study of the degree of species overlap along moisture gradients. This will enable us to test hypotheses on spatial scales in which species overlap in scale and where there is no species overlap. The third extension requires the study of the reciprocal relationship between soil moisture and species diversity (i.e., how soil moisture controls the number, abundance, and composition of the assemblage and how the assemblage uses the moisture to drive ecosystem processes). The conceptual model we have outlined here combines (1) a system model that identiﬁes the crucial ﬂows that connect species to the larger ecological system, (2) the identity of the species that are a part of the ecosystem, (3) a speciﬁcation of how the diﬀerent species contribute to the processes that govern ecosystem function, and (4) an assessment of whether the assemblage of species present process resources in a positive compensatory or a negatively synergistic way. The approach is an attempt to explicitly deﬁne the diﬀerences among species in an assemblage in terms that mechanistically link resource utilization by species with its ecosystem consequences. We have illustrated the approach by exploiting the clear relationships among water availability, water use, and biomass production in arid lands. This introduces properties of the individual species that are relevant to an ecosystem process into the speciﬁcation of species diversity. The concept of species diversity we use can be speciﬁed in particular systems to include: the number of species, the diﬀerences among them in resource utilization, spatial distribution and local abundance along a soil moisture gradient. Rather than relying only on numbers of species, hypotheses generation and experiments must also account for diﬀerences between species.

References Breshears, D., D. Barnes, and J. Fairley. 1999. Interrelationships between plant functional types and soil moisture heterogeneity for semiarid landscapes within the grassland/forest continuum: A uniﬁed conceptual model. Landscape Ecology 14(5): 465–478. Collins, S.T., and T.L. Benning. 1996. Spatial and temporal patterns in functional diversity. Pp. 253–280 in K.J. Gaston (ed.), Biodiversity—A Biology of Numbers and Diﬀerences. Blackwell Science, Oxford. Ehrlich, P.R., and A.H. Ehrlich. 1981. Extinction: The causes and Consequence of the Disappearance of Species. Random House, New York. Gaston, K.J. 1996. What is biodiversity? Pp. 1–5 in K.J. Gaston (ed.), Biodiversity—A Biology of Numbers and Diﬀerences. Blackwell Science, Oxford.

Species Diversity and Ecosystem Processes in Water-Limited Systems

165

Grime, J.P. 1997. Biodiversity and ecosystem function: the debate deepens. Science 277: 1260–1261. Hooper, D.U., and P.M. Vitousek. 1997. The eﬀect of plant composition and diversity on ecosystem processes. Science 277: 1302–1305. Huston, M.A. 1997. Hidden treatments in ecological experiments: re–evaluating the ecosystem function of biodiversity Oecologia 110: 449–460. Johnson, K.H., K.A. Vogt,. H.J. Clark, O.J. Schmitz, D.J. Vogt. 1996. Trends in Ecology and Evolution 11(9): 372–377. Jones, C.G. and J.H. Lawton (eds). 1995. Linking Species and Ecosystems. Chapman and Hall, New York. Keddy, P.A. 1991. Working with heterogeneity: an operators guide to environmental gradients. Pp. 181–201 in J. Kolasa and S.T.A Pickett (eds)., Ecological Heterogeneity. Springer-Verlag, New York. Lawton, J.H. 1994. What do species do in ecosystems? Oikos 71: 367–374. Lawton, J.H., and V.K. Brown. 1994. Redundancy in ecosystems. Pp. 255–270 in E.D. Schulze and H.A. Mooney (eds.), Biodiversity and Ecosystem Function. SpringerVerlag, Berlin. Le Roux X., T. Bariac, and A. Mariotti. 1995. Spatial partitioning of the soil water resource between grass and shrub components in West African humid savanna. Oecologia 104: 147–155. Ludwig, J.A., and D.J. Tongway. 1995. Spatial organization of landscapes and its function in semi arid woodlands, Australia. Landscape Ecology 10: 51–63. McNaughton, S.J. 1993. Biodiversity and function of grazing ecosystem. Pp. 361–383 in E.D. Schulze and H.A. Mooney (eds.), Biodiversity and Ecosystem Function. Springer-Verlag, Berlin. Morin, J., D. Rosenfeld, and E. Amitai. 1999. Radar rainﬁeld evaluation and use. In T.W. Hoekstra, and M. Shachak (eds.), Arid Lands Management—Toward Ecological Sustainability. Illinois University Press, Urbana. Naeem, S., and S. Li. 1998. A more reliable design for biodiversity study? Nature 394: 30. Pastor, J., and S.D. Bridgham. 1999. Nutrient eﬃciency along nutrient availability gradients. Oecologia (Berlin) 118(1): 50–58. Pickett, S.T.A., J. Kolasa, J.J. Armesto, and S. L. Collins. 1989. The ecological concept of disturbance and its expression at various hierarchical levels. Oikos 54: 129–136. Schlapfer, F., and B. Schmid. 1999. Ecosystem eﬀects on biodiversity: a classiﬁcation of hypotheses and exploration of empirical results. Ecological Applications 9(3): 893–912. Shachak, M., and C. G. Jones. 1995. Ecological ﬂow chains and ecological systems: concepts for linking species and ecosystem perspectives. Pp. 280–294 in C.G. Jones, and J.H. Lawton (eds.), Linking Species and Ecosystems. Chapman and Hall, New York. Shachak, M., S.T.A. Pickett, B. Boeken, and E. Zaady. 1999. Managing patchiness, ecological ﬂows, productivity and diversity in drylands: concepts and applications in the Negev Desert. Pp. 254–264 in T.W. Hoekstra and M. Shachak (eds.), Arid Lands Management—Toward Ecological Sustainability. Illinois University Press, Urbana. Sharon, D. 1980. The distribution of hydrologically eﬀective rainfall incident on sloping ground. Journal of Hydrology 46: 165–188.

166

Ecological Complexes of Biodiversity, Ecosystems, and Landscapes

Tilman, D. 1999. The ecological consequences of changes in biodiversity: a search for general principles. Ecology 80(3): 1455–1474. Tilman, D., C. Lehman, and K. Thompson. 1997. Plant diversity and ecosystem productivity: theoretical consideration. Proceedings of the National Academy of Sciences (USA) 94: 1857–1861. Tilman, D., D. Wedin, and J. Knops. 1996. Productivity and sustainability inﬂuenced by biodiversity in grassland ecosystem. Nature 379: 718–720. Vandermeer, J.H. 1989. The Ecology of Intercropping. Cambridge University Press, Cambridge. Vitousek, P.M., and D.U. Hooper. 1993. Biological diversity and terrestrial biogeochemistry. Pp 3–14 in E.D. Schulze and H.A. Mooney (eds.), Biodiversity and Ecosystem Function. Springer-Verlag, New York. Walker, B.H. 1992. Biodiversity and ecological redundancy. Conservation Biology 6: 18–23. Whittaker, R.H. 1967. Gradient analysis of vegetation. Biological Reviews 42: 207– 264. Yair, A. 1999. Spatial variability in runoﬀ generated in small arid watershed: implications for water harvesting. Pp. 212–222 in T.W. Hoekstra and M. Shachak (eds.), Arid Lands Management—Toward Ecological Sustainability. Illinois University Press, Urbana. Yair, A., and M. Shachak. 1987. Studies in watershed ecology of an arid area. Pp. 146–193 in M.O. Wurtele, and L. Berkofsky (eds.), Progress in Desert Research. Rowman and Littleﬁeld, Totowa, NJ.

10 Linking Species Diversity and Landscape Diversity Bertrand Boeken Yarden Oren Shlomo Brandwine Sol Brand

E

cologists generally agree that species diversity is linked to landscape features (Pickett and White 1985, Glenn et al. 1992, Wiens et al. 1993, Rosenzweig 1995, Hoagland and Collins 1997, Ritchie and Olﬀ 1999). We present a conceptual framework for connecting species diversity and landscapes by showing how changes in species assemblages and changes in landscape structure coincide. We focus on the dynamics of the mutual relationship between (1) the frequency of occurrence of the various landscape mosaic components (patches) and their properties in terms of abiotic conditions, resource availabilities, and structural features, and (2) the occurrence and abundance of the species of an assemblage within and among these components. Although we use examples of assemblages of annual plants in semiarid shrubland, we stress the generality of our approach and its applicability to many other groups of organisms and landscapes. Most ecologists would also agree that there are connections between the observations that (1) individuals and populations of organisms are aﬀected by environmental heterogeneity in the landscape, (2) species assemblages (or communities) consist of populations (or parts of them), and (3) changes in the landscape aﬀect species assemblages, and vice versa. In this chapter we explore this often intuitive relationship explicitly. Our basic premise is that species assemblages are collections of populations interacting with the heterogeneity of the landscape. We use the term ‘‘assemblage’’ to preclude assumptions about interactions and proximity or encounters among the organisms. Simple presence in the sampled landscape is the criterion for belonging to an assemblage; the landscape mosaic is an assemblage of patches, which, like species, may or may not interact. We assume that the landscape is hetero167

168

Ecological Complexes of Biodiversity, Ecosystems, and Landscapes

geneous, comprising a mosaic of distinct patches, which can be distinguished by some patch property. Our approach does not require a particular size or kind of landscape, but its scale and structure and the deﬁnition of patches have to be relevant for the distribution of the organisms whose diversity we study. In this chapter we discuss the functional connection between the dynamics of landscapes and of species assemblages. We then deﬁne diversity of a) species within an assemblage and b) patches within a landscape in a way that preserves this dynamic link and highlights the processes involved. We will demonstrate how the relationship between species and landscape diversity can be mapped using experimental studies with annual plants in a semiarid shrubland in the Negev. Finally, we will discuss how this mapping of the relationships between landscape and species diversity can help in developing a theory of biodiversity.

Assemblage and Landscape Dynamics Species assemblages change through time by means of two separate pathways, internal and external to the assemblage (ﬁg. 10.1A). Some of the species in the assemblage at time t1 pass on to the assemblage at time t2 (the internal pathway), and new species may enter from the larger species pool outside the deﬁned assemblage (the external pathway). The dynamics of the assemblage are the result of demographic processes of all the species as they either colonize the patch, remain resident, or become extinct. The dichotomy between internal and external pathways of assemblage dynamics applies to the entire landscape as well as to single patches. At any time, the assemblage of the entire landscape considered consists of species that were already present (internal dynamics) and those that colonized from the ‘‘regional’’ species pool external to it (Pa¨rtel et al. 1996). Likewise, the local subassemblage of a patch is the result of internal dynamics of resident species and the external dynamics of colonization of species from other patches. Loss of species from the landscape implies assemblage-wide extinction, while loss from a patch implies local extinction only. In contrast to species, new patches originate from existing ones (or parts of them) (ﬁg. 10.1A). The external controls of landscape dynamics are biotic and abiotic agents of patch formation (control 1 in ﬁg. 10.1A) that create new patches, and sources of resource supply (control 2). Patch creation is the result of long-term processes such as mound formation by accumulation of soil and litter under shrubs (Boeken and Shachak 1994), or of sudden disturbance such as diggings made by foraging animals (Platt 1975, Boeken et al. 1995). Both landscape and assemblage dynamics are connected by a number of control processes (ﬁg. 10.1A and B). Control 3 shows the eﬀects of patch properties and conﬁguration on species composition within each patch and of the entire assemblage through ﬁltering colonization by incoming propagules.

Linking Species Diversity and Landscape Diversity

169

Figure 10.1 Relationships between landscape (A) and species assemblage (B) dynamics. Patches in the landscape mosaic are formed (control 1) and change in quality (2), which includes disappearance. Species of an assemblage colonize landscape patches and grow in density (or dwindle and become extinct). Landscape and species assemblage dynamics are connected by: (i) patch properties (which include conditions, resources, and interactions with organisms of other assemblages) controlling colonization of species (3) and extinctions, recolonizations, and dynamics of populations (4), in addition to within-patch biological interactions within the assemblage (5); (ii) species controlling formation, maintenance, and modiﬁcation of patches (6).

Control 4 indicates that local conditions, resources, and interactions with other organisms determine species-speciﬁc recruitment, survivorship, growth, and fecundity. Control 5 represents the control of these processes by the assemblage itself through biological interactions among its members, such as site or resource competition and facilitation. Control 6 represents the opposite eﬀect of the species assemblage on the landscape, as particular species or species combinations diﬀerentially alter the properties of patches, as is common in autogenic succession for instance. This can happen by means of shading or litter production by high-density species, acidiﬁcation of the soil, and so on. In our model landscape, a gentle hill-slope with semiarid shrubland in the northern Negev, the patches are soil and litter mounds at the foot of (sub-) shrubs, similar to the ‘‘resource islands’’ described in North and South American semiarid shrublands (Garner and Steinberger 1989, Reynolds et al. 1999, Aguiar and Sala 1999), scattered within a matrix of soil surface covered with a microbial crust (Zaady and Shachak 1994). The patch-forming agent is a shrub seedling that has successfully established on the crust, and is accumulating a mound of loose matter at its base (control 1). This changes the state of the landscape at time t1 to a new state at t2, by gradually converting the immediate surroundings of the shrub from a crusted patch to

170

Ecological Complexes of Biodiversity, Ecosystems, and Landscapes

a shrub patch. The properties of this patch vary with resource inputs (control 2), for it can absorb a large quantity of run-oﬀ (with soil, nutrients, organic matter) or only a little (Shachak et al. 1998), depending on the rainfall and on the landscape itself (size and surface texture of the area contributing run-oﬀ, the density of other patches, etc.). These state changes of patch properties occur simultaneously at the scale of the individual patch as it is formed, changes, and eventually disappears, and at the scale of the landscape, as its patchiness (conﬁguration, numbers, and size distribution of patches of all types) is altered. Focusing on the shrub patch that has taken the place of a crusted patch (or a bit of matrix), a diﬀerent assemblage of species is found under the shrub than on the remaining crusted surface; it typically has more and diﬀerent species of annual plants (Boeken and Shachak 1994, 1998a, 1998b). This is due to the inﬂuence patch properties have on colonization of previously locally absent species (control 3). The patch properties also change the relative abundances and extinctions among species that were already there (control 4). However, these changes in abundance are controlled by the assemblage itself as well, in the form of biological interactions among the annual plants in the shrub patch (control 5). These controls on the species assemblage operate both during and after shrub-patch formation. At the same time, the plants growing in the shrub patch produce and capture plant litter, and their roots penetrate the soil, which aﬀects the properties of the patch (control 6). Its size is one property that may change, when the mound at the base of the shrub expands as plants tend to colonize the interface between shrub patch and crust, where run-oﬀ water is absorbed. Shifting focus to the entire landscape, with its altered patchiness compared with the landscape without the shrub patch, there are more (or diﬀerent) opportunities for colonization and conditions for persistence of species from outside the assemblage of the landscape (controls 3, 4, and 6). At this scale, the dynamics of the landscape and the internal and external assemblage dynamics within all the patches of the landscape are determined by formation, change, and disappearance of all patches, not just one patch as in the patch-scale situation. In our approach of biodiversity of organisms, the structure of the landscape is central, as deﬁned by the patch types in the mosaic. Patch types should be chosen by properties and scales relevant for the group of organisms studied, in terms of accessibility, resource availability and/or protection from biotic or abiotic causes of mortality. Patch types are based on properties such as topography, texture, architecture, microclimate, and resource concentrations or supply rates. Our approach is appropriate for organisms that spend a signiﬁcant portion of their lives within distinguishable landscape patches. Using criteria based on general relevance to the presence of species of the assemblage to deﬁne patch types, but ignoring the various scales of movement of the individual species turns the landscape and assemblage into independently measured realms. The ecological relationships between patches and assemblages can be studied, focusing on the presence and abundance of (sub)populations in and among patches.

Linking Species Diversity and Landscape Diversity

171

Mapping Species and Landscape Diversity We propose to quantify species and landscape diversity in a way that expresses the link between assemblage and landscape dynamics by population-level processes of all species within (local) and among patches (‘‘spatial’’ or ‘‘landscape-wide’’). We do this by applying the general deﬁnition of diversity as proposed by Gaston (1996), based on (1) the number of entities and (2) the diﬀerences between them. We apply this deﬁnition to both species and landscape patches as entities. We describe species diversity of an assemblage as the number of species as points in a phase plane of incidence and abundance (ﬁg. 10.2A) (Boeken and Shachak 1998a). The diﬀerences between species as entities are: (1) their local population abundances within patches and (2) their landscape-wide incidence over all patches. Abundance represents properties of the species’ populations, such as density, biomass, cover, and so on averaged over all sampled patches where present. Incidence represents the frequency of occurrence of the species in the landscape, as a fraction of all patches. Classical approaches to species diversity, in contrast, tend either to transform numbers and diﬀerences of many species to a single index (e.g., Simpson’s or Shannon’s and Weaver’s diversity index), or to use only one dimension (richness and its derivatives). This limits diversity to aggregate assemblage parameters without the ability to link it either to population-level processes or to changes in the landscape. In our approach, abundance and incidence reﬂect diﬀerences between species density and where and how often they occur in the landscape. Without incidence, abundance does not provide any spatial information on the species. Usually, when using abundance without incidence, it is calculated as the overall mean population density per patch over all patches. Including unoccupied patches in the abundance causes a highly signiﬁcant but spurious correlation with incidence (Wright 1991, Hanski et al. 1993). In our case, null-populations are excluded from abundance, since incidence already takes the spatial aspect into account.

Figure 10.2 Mapping species and landscape diversity in a phase plane with species and patches as entities. (A) Species assemblage in a landscape in the incidence– abundance phase plane (each point is a species). (B) Patch assemblage in a landscape in the property–frequency phase plane (a; b; and c are patch types).

172

Ecological Complexes of Biodiversity, Ecosystems, and Landscapes

The positions of species in the incidence–abundance phase plane reﬂect their individualistic behavior in the landscape, which explicitly links population dynamics of all species to the patchiness of the landscape. Although incidence of a species expresses its landscape-wide occurrence, and abundance its local occurrence, incidence is also aﬀected by local population processes (within patches), and abundance by spatial processes (among patches). Abundance of a species focuses on local population growth and persistence, augmented or colonized by input from outside the patch. The latter include spatial processes in the form of immigration of individuals, which are partially controlled by local properties aﬀecting arrival. Incidence reﬂects colonization, persistence, and extinction of the species in patches over the entire landscape, thus focusing more on spatial processes, with less emphasis on local processes. The spatial processes are based on dispersal of organisms among patches, which include emigration, movement, and arrival. In addition to the species’ positions in the incidence–abundance phase plane, the relationship between them has also informative value. Most studies have found a signiﬁcant positive correlation between the two (Wright 1991, Hanski et al. 1993, Gaston 1994, Brown 1995), though not always (Obeso 1992). Brown (1995) explained this ‘‘background’’ relationship mechanistically by pointing out that locally abundant species are widespread because they are generalists utilizing abundant resources or microsites, and that their high reproductive output can maintain dense local populations and high rates of immigration into other patches. Variation in the slope of the line or its statistical signiﬁcance may reﬂect important diﬀerences or changes in landscape-wide species composition and diversity of the assemblage. A steeper line signiﬁes high abundance of common species when they become dominant in the patch communities, while a ﬂatter slope indicates greater equitability. Landscape diversity can be deﬁned similarly to species, as patches are landscape-level entities that can be assigned to a number of patch types. Again we stress that the distinction between patch types should be based on properties considered important for the group of organisms whose diversity is being studied. Analogous to species, patch types also have local and landscape-wide diﬀerences. Patch types diﬀer locally in properties such as the availability of a particular resource (water, nutrients, organic matter, food or host plants, and prey), composition of the substrate (soil, vegetation, etc.), or any other relevant diﬀerence. At the scale of the entire landscape, patch types diﬀer in the frequency of occurrence in the landscape (ﬁg. 10.2B). Patch type frequency is the net result of patch formation (controlled by biotic or abiotic disturbance or gradual long-term processes, ﬁg. 10.1B), while patch properties are the result of resource availabilities, conditions, and biological interactions prevailing in the patches of diﬀerent types. The positions of patch types in the frequency–property phase plane express landscape diversity by showing how many patch types there are, how often these occur in the landscape, and how diﬀerent they are. The terms patch type and patch property are deﬁned by diﬀerent criteria, while properties often vary less within patch types than among them. In general, patch types diﬀer in

Linking Species Diversity and Landscape Diversity

173

some major way in overall structure. For instance, a forest stand diﬀers from a treefall gap in the amount of light, and in other variables like soil moisture, soil organic matter content, and so on, while within the same type, stands or gaps may also vary among themselves in light regimes or soil properties. In dryland systems, landscape patch types can be recognized at three scales: (1) patches of ‘‘bare’’ or crusted soil, and patches associated with large plants (shrubs, cacti, grass tussocks, etc.), litter-covered soil, diggings and mounds made by animals or humans, or bare rock outcrops (Garner and Steinberger 1989, Yair and Shachak 1982, Shachak et al. 1999, Aguiar and Sala 1999, Aguiar et al. 1992, Boeken and Shachak 1994, 1998a, 1998b, Boeken et al. 1995, Zaady and Shachak 1994, Shachak et al. 1998), and (2) watersheds and their components (water courses, slopes, hilltops, and plains), in addition to (3) larger geographical patches with diﬀerent properties, including rainfall, temperature, remoteness from major water bodies, and other gradients. Patches and species each have their own dynamics and deﬁning criteria; therefore the analogy between them should not be overemphasized. For instance, species abundance is a single diﬀerentiating property measured by density, biomass or cover, and so on, while patch types have numerous properties, such as size, shape, topography, surface texture, soil structure, soil moisture, light regime, nutrient availabilities, microclimatic conditions, presence of herbivores and pathogens, proneness to disturbance, position within the landscape, distance from other patches, and so on.

Trajectories of Species and Patches As both patches and species change through time, both kinds of entities describe trajectories in their respective phase plane (ﬁg. 10.3). These reveal the local and spatial processes taking place in the species assemblage and the

Figure 10.3 A joint mapping of species and landscape diversity and the relationship between their dynamics. In response to increase in quality of patch a the abundance and incidence of species A increase. In response to increase in the frequency of patch b the incidence of species B increases). See text for explanation.

174

Ecological Complexes of Biodiversity, Ecosystems, and Landscapes

landscape mosaic from time t1 to time t2. Trajectories of patch types (patch type vectors) to the right show how patches are formed, changes to the left show how they disappear in time, while changes in some measurable properties are shown by vertical changes along the property axis. Because we use frequency of patch types, each change in one patch type is reﬂected in the others, either just because they add up to 1, or also because a patch of one type is converted to one of another type, as when an animal digs a pit in a ﬂat undisturbed surface. If completely new patch types are created or all patches of one type disappear, the number of entities (patch types) changes in addition. Species trajectories show how species change in their incidence and abundance as a response to changes in the landscape mosaic. As new patches form and disappear (changes in frequency), and as they change in quality (gradually, by disturbance or due to annual climatic variation), some species increase and others decrease in incidence and abundance. Changes in incidence signify (1) new invasions into the assemblage, (2) colonization into more patches, (3) extinctions in some of the patches, or (4) landscape-wide extinction. Changes in abundance signify positive, negative, or zero local population growth rates. Population growth can be densityindependent (only limited by seed production and arrival) or density-dependent (resource or site limitation). Thus, assemblage dynamics reﬂect a multispecies network of the local and spatial demographic behaviors of all species of the assemblage. These behaviors occur at the population level, and reﬂect the probabilities (transition rates) of individuals of the diﬀerent species to pass from one state to another. A change in the position of a patch type from time t1 to t2 (as in ﬁg. 10.1) and the resulting trajectory of a species can be viewed using incidence–frequency as a common axis (ﬁg. 10.3). Figure 10.3 shows that as patch type a changes in quality from one year to the next (i.e., soil moisture due to annual rainfall), species A only increases in abundance (from A(t1) to A(t2)) as a result of a local increase in the rate of recruitment, or also in incidence (A0 (t2)) as the recruitment rate also increases in unoccupied patches of the same type from the few propagules that had arrived. In another example, patch type b changes in frequency as more undisturbed patches are disturbed. As a result, we see that species B increases in incidence alone, apparently due to easy dispersal into the new patches.

Case Studies We present two examples to demonstrate how species assemblages can be mapped in relation to experimental landscape changes. In both cases an incidence–abundance phase plane describes the local within-patch and landscape-wide population behavior of the species of an annual plant species assemblage. Diﬀerences in patch types are represented by a phase plane of patch properties and frequency in the landscape. The experiments alter the landscape diﬀerently, either by adding new patches within an undisturbed

Linking Species Diversity and Landscape Diversity

175

matrix (Boeken and Shachak 1994) or by disturbing parts of natural patches (Oren 2001). Both studies compare phase planes of undisturbed patches with both undisturbed and disturbed patches together. The studies took place on a south and a north facing slope (both ca. 10%) in semiarid shrubland on loessial soil in Sayeret Shaked park in the northern Negev desert of Israel. The landscape consists of roundish shrub patches of 0.30–1.50 m diameter with a 20–50 cm high shrub and a 3–10 cm high soil mound, scattered within a matrix of ﬂat crusted soil surface with intershrub distances between 0.5 and 5.0 m (Boeken and Shachak 1994, 1998a, 1998b). The crust is the product of the interactions among cyanobacteria, bacteria, algae, fungi, mosses, lichens, and the soil particles (Zaady and Shachak 1994). In addition, there are other patches with sizes comparable to the shrub patches, such as disturbances made by humans (ditches, tracks, etc.) and animals (ant nests, porcupine diggings, mole–rat mounds, livestock-trampled areas). Patch types diﬀer in species richness, equitability, and biomass production, and densities of nearly all of the annual plant species (Boeken and Shachak 1994, 1998a, 1998b). The patch addition experiment (Boeken and Shachak 1994) consisted of 20 units on a single slope, each unit containing three adjacent patches 1 m 0.3 m in the form of undisturbed crusted soil surface upslope, and a pit 30 cm deep and a mound downslope 20 cm high. The patches were contiguous with their long side, oriented along the contour lines. The pits and mounds were constructed in 1987. In 1992 we counted the plants per species in the undisturbed and disturbed patches at peak biomass (March–April). Annual rainfall totaled 164 mm during the winter of 1991/92, below the long-term average of 200 mm (Stern et al. 1986). The analysis compares undisturbed crusted patches with a mix of disturbed and undisturbed ones. In both cases (crust and three patch types together) 18 patches are used. Of the 20 units sampled, 18 crust patches were taken randomly, while six random units provided data for all three patch types per unit, also giving 18 patches. This way we compare undisturbed and disturbed landscapes without sampling intensity artifacts. The randomization procedure was repeated a number of times, giving essentially the same results. The assemblage composition of the patch addition experiment (ﬁg. 10.4) as depicted in the incidence–abundance phase plane showed that most of the species are rare and sparse with few common and dense species. The creation of new patches increased species number of the assemblage by additional rare species, while causing an increase in incidence and abundance of both rare and common species. Two of the common species on undisturbed crust (Stipa capensis and Plantago coronopus) decreased, while two (Bromus fasciculatus and Rostraria cristata) increased both in incidence in the landscape and in abundance of local populations. In spite of these changes, the slope of the incidence and abundance relationship did not change when new patches were included. The shift of the species along the abundance–incidence line was caused by greater arrival and retention of seeds in the new patches (Boeken and

176

Ecological Complexes of Biodiversity, Ecosystems, and Landscapes

Figure 10.4 The eﬀect of landscape diversiﬁcation on annual plant species diversity. Incidence and abundance of all species of annual plants in spring 1992 in (A) 18 crust patches (1 0:30 m) and (B) 18 crust patches + experimental pits and mounds created in 1987 (six of each patch type). Brofas: Bromus fasciculatus, Placor: Plantago coronopus, Roscri: Rostraria cristata, Sticap: Stipa capensis. The slope of the line changed from 1.3463 to 1.4372, its intercept from 0.5413 to 0.4604.

Shachak 1994). This was due to capture of larger seeds on the rough soil surface of the new patches, which lack the dense ﬂat crust cover, but was independent of diﬀerences in soil moisture in the patches. In the patch disturbance experiment, crust and shrub patches were disturbed in spring 1994 (March) by above-ground vegetation clipping and/or surface trampling (ﬁg. 10.5). The manipulations mimic the eﬀects of grazing by small livestock (sheep and goats). Of a total of 20 patches of around 0.25 m2 on a slope (10 shrub patches and 10 crusted areas), eight were left undisturbed, four were clipped, four were trampled and four were both clipped and trampled. Plants were identiﬁed and counted in March 1995, a year with above-average rainfall (283 mm). In the analysis we compared (a) a landscape with only undisturbed patches (four shrub and four crust patches) with (b) a landscape with both undisturbed and disturbed patches (two undisturbed shrub and two undisturbed crust patches and four disturbed shrub and crust patches). The latter were either clipped, trampled, or both. Neither clipping nor trampling increased assemblage species richness (ﬁg. 10.6), but their combination did (ﬁg. 10.6D). In the clipping and trampling treatments some rare species increased in incidence but not in abundance, others in abundance and not in incidence, with some diﬀerences between the treatments. Most species, however, did not change at all. In the three treatments alike, the dominant species, S. capensis, R. cristata, and P. coronopus decreased in abundance but maintained 100% incidence. Only trampling, alone and together with clipping, caused two species to increase in abundance (Anagallis arvensis and Ononis reclinata) without aﬀecting their incidence. Trampling increased incidence of one species (Iﬂoga spicata), while clipping increased incidence of three species

Linking Species Diversity and Landscape Diversity

177

Figure 10.5 Undisturbed crust and shrub patches characteristic of the landscape mosaic of semiarid slopes.

including I. spicata, and in combination with trampling six species. Only one species, Pteranthus brevis, decreased in incidence due to clipping. Trampled patches provide a rough surface instead of ﬂat crust, while clipped patches lack standing vegetation (last year’s dead annuals in crust patches, and shrub plus annuals in shrub patches). Intact vegetation may inhibit arrival of seeds at the soil surface, while ﬂat dense surface structure may not provide suitable sites for seeds to remain in place until germination or to complete germination. The experiments illustrate how plotting species in the incidence–abundance phase plane exposes important links between landscape and assemblage diversity, and between assemblage dynamics and the demographic behavior of the species. Comparing the results of the two experiments (ﬁg. 10.7, table 10.1) shows that diﬀerent changes in landscape diversity, as were made in the experiments, have some similar and some diﬀerent eﬀects on species diversity. In both cases the number of patch types and their contrast increased, causing a shift to higher incidence of the most common species, and an addition of newly colonizing rare species. Creating two new patches (ﬁg. 10.7A) added more new species to the assemblage than adding two modiﬁed patches (ﬁg. 10.7B). Therefore, the number of new rare species colonizing patches not previously occupied by them is not a function of simply adding patches,

178

Ecological Complexes of Biodiversity, Ecosystems, and Landscapes

Figure 10.6 Phase planes of abundance and incidence of annual plant species assemblages in 1995 in undisturbed, trampled, clipped, and clipped and trampled shrub and crust patches (eight patches in all cases).

but of the relative increase in patch type number, and the contrast between the patch types. Pits and mounds diﬀer in structure and soil texture more from the undisturbed crusted surface than clipped or trampled shrub and crust patches diﬀer from their undisturbed equivalents. In the second experiment (ﬁg. 10.7B), greater contrast by applying both kinds of disturbance increased species number more than either one alone. On the other hand, the abundance–incidence relationship, which acts as a background pattern for assemblage behavior, did not change with the addition of very diﬀerent patches nor by disturbing existing ones (ﬁg. 10.7).

General Predictions Based on the case studies and on the underlying model of assemblage and landscape dynamics (ﬁg. 10.1), we can make a number of predictions in the

Linking Species Diversity and Landscape Diversity

179

Figure 10.7 The relationship between changes in patch diversity and species diversity of annual plant assemblages in semiarid shrubland landscapes in (A) a patch addition experiment and (B) a patch disturbance experiment. The changes in species diversity are represented by species trajectories (arrows) in the abundance– incidence phase planes, and changes in landscape diversity are represented by patchtype vectors in the property–frequency phase plane. The properties are disturbance of patch topography/architecture and surface structure in pits (horizontal hatches) and mounds (vertical hatches) made in crusted soil (from state a to b in A) and by clipping and/or trampling of crust and shrub patches (undisturbed: dark gray shade and vertical hatches, respectively; and disturbed: dots pattern and vertical hatches) (from state a to b, c, or d in B).

form of hypotheses on annual species diversity responses in various situations where landscapes change. Two hypothetical examples of annual plant responses in semiarid shrubland are represented in table 10.2 and ﬁg. 10.8, describing assemblage dynamics as a result of changes in patch properties

180

Ecological Complexes of Biodiversity, Ecosystems, and Landscapes

Table 10.1 Comparisons of the effect of changes in landscape structure on annual plant assemblages (two ﬁeld experiments in a semiarid shrubland in the Negev) Expt. 1: Addition of Patches

Expt. 2: Disturbance of Patches

Landscape changes 1. Action

New patches created

Existing patches modiﬁed by clipping, trampling or both

2. Comparison

Undisturbed crusted landscape vs. a landscape with three patch types (crust, pits and mounds)

Undisturbed landscape with crust and shrub patches vs. landscapes with undisturbed and disturbed shrub and crust patches

3. Patch property contrast

Large

Small

1. Species number

Increased

Increased (with double disturbance)

2. Slope line in the I/A plane

Constant

Constant

3. Density of points in the I/A plane

Increased at medium incidence

Increased at low incidence

4. Species trajectories in the I/A plane

Moved to higher incidence and abundance

Shifted in position at low incidence

Assemblage responses

(soil moisture due to drought, ﬁg. 10.8A) and in patch type frequency (increased shrub cover). In both cases we assume sampling is done during the growing season before and after the change, with comparable amounts of rainfall. The predictions are based on the local and spatial demographic processes aﬀecting abundance and incidence of many species. The property in question is soil moisture, while the patch types are deﬁned by topography, surface structure, canopy architecture, and microclimate. We had to make some assumptions about the species’ performance in patches, based on the speciﬁcs of the semiarid shrubland system; most species respond positively to better soil moisture conditions and surface disturbance, while a few are specialized in low-quality or undisturbed patches. The reduction in water availability, caused in both patch types by very low amounts of rainfall, is hypothesized to cause a reduction both of species number and of the slope of the abundance–incidence line (ﬁg. 10.8A). This is the result of increased local (per patch) extinction of rare to intermediate species in both patch types, and assemblage-wide extinctions of rare species. The extinctions are, in turn, caused by combinations of site limitation, demographic stochasticity in small populations (Forney and Gilpin 1989), and decreasing opportunities for rescue eﬀects (Hanski et al. 1996). Common species will only decrease in abundance. During increasing

Linking Species Diversity and Landscape Diversity

181

Table 10.2 Predicted assemblage dynamics of annual plants in semiarid shrubland due to changes in climate or in landscape structure Scenario

Drought

Increased Shrub Cover

1. Action

Reduction in patch property (soil moisture)

Addition of (shrub) patches

2. Comparison

Before and after action

Before and after action

3. Patch property contrast (shrub and crust patch)

Small

Large

1. Species number

Decreases

Increases

2. Slope line in the I/A plane

Less steep

Constant

3. Density of points in the I/A plane

Increases at low, decreases at intermediate and constant at high incidence

Increases at all incidences

4. Species trajectoriesin the I/A plane

Moving to lower incidence and abundance

Moving to higher incidence and abundance

Landscape changes

Assemblage responses

Figure 10.8 Predicted changes in incidence and abundance of a species assemblage in response to (A) general impoverishment of all landscape patches and (B) an increase in density of high-quality patches. Arrows represent trajectories from the original to the new state of patches and species. Circles depict groups of species moving in the direction of the arrows, lines represent abundance–incidence relationships (dark: initial; light: predicted). Squares depict patch types changing in a property (A) or in their frequency (B).

182

Ecological Complexes of Biodiversity, Ecosystems, and Landscapes

shrub cover (ﬁg. 10.8B), all low to intermediate incidence and abundance species will increase at the same rate, while some of the most common species decrease (presumably due to changes in surface structure aﬀecting arrival or recruitment of crust specialists). These predictions can also apply to similar scenarios with diﬀerent organisms and landscapes. If overall patch quality decreases due to a climatic change, for instance when hotter summers or colder winters reduce insect abundance, an assemblage of insectivorous breeding birds in a geographical area with a mosaic of agricultural ﬁelds and scattered patches of natural vegetation may respond according to the same dynamics as in the drought case (ﬁg. 10.8A, table 10.2). Likewise, we hypothesize that an increase of the frequency of vegetation patches by oldﬁeld succession will result in dynamics similar to those of annual plants after shrub encroachment.

Applications of the Approach The proposed methodology for studying species and landscape diversity can be applied to a large number of ecological questions arising from changes in landscape structure. Empirical studies of the responses of species assemblages of many groups of organisms can beneﬁt from the explicit connection with local and spatial population processes. One way to further develop the approach besides accumulating empirical evidence is to use numerical simulation to predict the dynamics of a species assemblage in response to particular landscape changes, under various assumptions about distributions of species attributes (patch type preferences, dispersal, breeding system, and biological interactions). Theoretical questions that can be addressed, at least conceptually, are those related to species composition and species interactions (from competitive exclusions to facilitation). In this context, assemblage dynamics diﬀer from more classical views of community dynamics, where the community is usually deﬁned at a large scale with smaller-scale replicate samples. In our view a species assemblage is deﬁned at the scale of the landscape with (replicate) samples of the constituent landscape patches. Because of the proximity and potential for interactions of the plants, subassemblages in samples (patches) come much closer to individual communities than the entire assemblage (which is therefore sometimes called metacommunity (Wilson 1992)). What this view adds is that local species composition (within patches) is not only aﬀected by local population dynamics depending on in situ conditions, resources, and other organisms, but also by spatial population processes of colonization and extinction, determining and maintaining community structure at the scale of the patch and of the whole landscape. Species composition and scalars like species number, species richness (correcting number for density or area, as in rarefaction’s and Fisher’s a), equitability, and other diversity indices (e.g., Simpson’s and Shannon’s and Weaver’s) are then viewed as net results of these processes. Species co-occurrences are also net results of

Linking Species Diversity and Landscape Diversity

183

local and spatial processes. In addition to local biological interactions among the species, which should lead to equilibrium species associations, cooccurrence (or coincidence) is strongly inﬂuenced by spatial stochasticity, causing nonequilibrium species combinations. Because our approach relates the diversity of an assemblage explicitly to the structure of the landscape, it can be helpful in the formulation of hypotheses and predictions about assemblage responses to scenarios of future change in landscapes and climate. The practical reason to link species assemblages and landscapes is that manipulating the former may be most eﬃciently done by modifying the latter. This eﬀect of ecological landscape management has been demonstrated in semiarid shrublands and elsewhere (Ludwig and Tongway 1995, Brussaard et al. 1996, Shachak et al. 1998, 1999, Pickett et al. 1999), and can be a powerful tool in conservation and restoration of biodiversity in general. The approach can be applied to numerous questions related to human-induced changes in landscape structure at various scales. Desertiﬁcation of semiarid landscapes (which we deﬁne as a decline in productivity and/or diversity, Boeken and Shachak 1994, Shachak et al. 1998), and its counterpart, sustainable land use, can be understood if we study the eﬀects of grazing or of clear-cutting of woody plants on landscape patchiness and on species incidence and abundance.

Contrasts with Other Perspectives The proposed approach to studying species diversity diﬀers from a number of other approaches of diversity because of its explicit incorporation of landscape dynamics. Besides the use of patches deﬁned by criteria and scales partially independent of species’ perspectives, the main diﬀerences are in keeping account of all individual species of an assemblage, and in linking with population-level phenomena with local and spatial components, instead of only with local phenomena or general species attributes. Unlike other approaches linking species with assemblages, we do not concentrate on life history or other traits of individual organisms (Wiens et al. 1993, Wiens 1991, Keddy 1992), but on presence and abundance of species populations. We diﬀerentiate between two kinds of population-level processes: within the components of the landscape mosaic and among them. The former are local population dynamics, while the latter are processes that determine the spatial distribution of the species across the landscape. These spatial processes are based on dispersal and immigration, determining colonization of populations in individual landscape patches, and their extinction in these patches. Spatial dynamics of organisms across landscapes are essential features of single-species metapopulation studies, but have rarely been linked explicitly to species diversity (though see Wilson 1992, Hanski et al. 1993), nor have they been incorporated in studies relating communities to population densities along gradients (Whittaker and Niering 1965, Weiher and Keddy 1995, Hoagland and Collins 1997). Because both colonization and

184

Ecological Complexes of Biodiversity, Ecosystems, and Landscapes

extinction have a considerable stochastic element (Gleason 1917, 1926, MacArthur and Wilson 1967, Glenn and Collins 1990, Gutzwiller and Anderson 1992), variation in presence and abundance among patches is often viewed as error among replicate samples. However, this ‘‘noise’’ contains valuable information on the relations between colonization, persistence, and extinction of species populations in the landscape and the diversity of the assemblage (e.g., Tilman 1993, Thorhallsdottir 1990). The approach we suggest contrasts with organism-centered approaches, where the focus of research is on the behavior of particular species. All species perceive the same landscape diﬀerently (Kotliar and Wiens 1990); this implies detailed information of populations and individuals of a species of animal or plant according to its particular perception. Depending on the size of the organism, how mobile it is and what resources it needs, it responds to diﬀerent topographic, physical, chemical, and biological features of the same landscape, at diﬀerent scales from other organisms. From the organism’s point of view, the relevant scales are (1) the distances of dispersal, (2) the size of the individual’s site, territory or home range, (3) the sizes and distribution of foraging patches, and (4) the size and aggregation of individual resource items. For evolutionary questions, about organisms’ adaptations to their environment and how they manage to live together, the landscape must be deﬁned from the organism’s point of view. Our central question, however, is about the relationship between the diversity of an entire assemblage of species belonging to a particular life form, taxon, functional group, or guild, and the diversity of the spatial mosaic of the landscape they inhabit. Applying an organism-centered deﬁnition of landscape heterogeneity would not only miss the point, but would be highly impractical, especially in very speciesrich assemblages, since this would entail describing nearly as many landscapes as species. The diﬀerence between organism- and landscape-centered approaches is one of relative scale, as organisms’ perceptions overlie the actual landscape, which is made up of a hierarchical (nested) structure of patches (Kotliar and Wiens 1990). Only at the ‘‘lower levels’’ of the hierarchy do these patches correspond to the ‘‘traditional’’ landscape patches, belonging to the landscape mosaic, and recognized because of a contrast between diﬀerent types. Known examples are the shrub patches and intershrub areas in our model system (Zaady and Shachak 1994), trees in savannas (Belsky 1994), herbivore diggings on rocky desert slopes (Gutterman and Herr 1981, Shachak et al. 1991, Boeken et al. 1995), tree canopy gaps in forests (Canham 1988), and so on. For plants and animals with home ranges (including microsites) smaller than the landscape patches, single landscape patches are relevant, since many population-level processes occur within them. Examples are herbs under shrubs (Boeken and Shachak 1994) and trees (Belsky 1994), trees in forest gaps (Canham 1988), mosses in uprooting pits (Jonsson and Esseen 1990), arthropods on animal cadavers (Hanski 1987), or marine benthos on boulders (Sousa 1984). For many organisms, however, single landscape patches are too small to be appropriate for the scales of the

Linking Species Diversity and Landscape Diversity

185

home, foraging, and dispersal ranges. In these cases, the relevant patches are aggregates of the lower-level ones, sometimes even deﬁned by diﬀerent criteria altogether. Most animals, even if they choose shelter or nesting sites within particular landscape patches, use larger parts of the landscape, consisting of many patches, for foraging (MacArthur and Pianka 1966, Kotler 1992). For large plants such as trees, actual landscape patches are only relevant for the initial stages of the life cycle (Canham 1988), but not for adults. Aggregate patches are inherently more internally heterogeneous than the component patches, and often have arbitrary boundaries (Kotliar and Wiens 1990). Often, however, there is a larger scale where aggregates of basic landscape patches form a high contrast with other patches at that scale. This corresponds to watersheds and their parts (channel, slopes, runnels, hilltops), and to the anthropocentric ‘‘landscape scale’’ of urban and rural land use and fragments of natural systems such as forests, grasslands, shrublands, wetlands, lakes, and so on. Although metapopulation studies are also organism centered, they are often done within an explicit landscape context (Hanski and Gilpin 1997) on organisms whose (sub)populations occur in recognizable landscape patches. We simply expand on the metapopulation approach, by considering an assemblage to be made up of (sub)populations of many species cooccurring in local within-patch assemblages. One implication of this approach is that no prior information on the metapopulation structure of the organisms is required. On the contrary, information on metapopulation dynamics in a patchy landscape is the result of these studies (Hanski and Gilpin 1997). The incidences of the species help to distinguish whether organisms occur in patches as subpopulations within a continuous population, populations within a metapopulation, or isolated populations (Hanski and Gilpin 1997). Boeken and Shachak (1998a) showed that very common species (basically the dominant herbs) with incidence approaching 1, behave like single populations in most of the landscape by virtue of rapid widespread dispersal. The rarer a species was (the lower its incidence), the more likely it was that the species had nearly isolated populations (Boeken and Shachak 1998a). Most species, however, have low to intermediate incidence (and abundance), and their spatial behavior in the speciﬁed landscape can appropriately be labeled metapopulation dynamics, including local extinctions and rescue events (recolonization). The assemblage–landscape dynamics described here highlight the use of abundance and incidence of species and the use of the frequency and properties of landscape patches as a promising research tool for studying biodiversity, for generating hypotheses and for designing and analyzing surveys and experiments. Our approach is applicable in a wide variety of types and scales of landscapes and organisms, provided patches can be recognized, and a number of species spend a signiﬁcant portion of their lives in these patches. The approach may also prove to be useful in identifying broad macroecological relationships (Brown 1995) and their mechanisms, and in dealing with larger-scale problems, such as habitat fragmentation, degradation,

186

Ecological Complexes of Biodiversity, Ecosystems, and Landscapes

and destruction, and for evaluating management eﬀorts of restoration and rehabilitation.

References Aguiar, M.R., and Sala O.E. 1999. Patch structure, dynamics and implications for the functioning of arid ecosystems. Trends in Ecological Evolution 14: 273–277. Aguiar, M.R., Soriano, A., and Sala, O.E. 1992. Competition and facilitation in the recruitment of seedlings in Patagonian steppe. Functional Ecology 6: 66–70. Belsky, A.J. 1994. Inﬂuences of trees on savanna productivity: Tests of shade, nutrients, and tree-grass competition. Ecology 75(4): 922–932. Boeken, B., and Shachak, M. 1994. Desert plant communities in human-made patches—implications for management. Ecological Applications 4: 702–716. Boeken, B., and Shachak, M. 1998a. Colonization by annual plants of an experimentally altered desert landscape: source-sink relationships. Journal of Ecology 86: 804–814. Boeken, B., and Shachak, M. 1998b. The dynamics of abundance and incidence of annual plant species during colonization in a desert. Ecography 21(1): 63–73. Boeken, B., Shachak, M., Gutterman, Y., and Brand, S. 1995. Patchiness and disturbance: plant community responses to porcupine diggings in the Central Negev. Ecography 18: 410–422. Brown, J. 1995. Macroecology. Chicago, Chicago University Press. Brussaard, L., Bakker, J.P., and Olﬀ, H. 1996. Biodiversity of soil biota and plants in abandoned arable ﬁelds and grasslands under restoration management. Biodiversity and Conservation 5(2): 211–221. Canham, C.D. 1988. Growth and canopy architecture of shade-tolerant trees: response to canopy gaps. Ecology 69: 786–795. Forney, K.A., and Gilpin, M. E. 1989. Spatial structure and population extinction: A study with Drosophila ﬂies. Conservation Biology 3(1): 45–51. Garner, W., and Steinberger, Y. 1989. A proposed mechanism for the formation of ‘fertile islands’ in desert ecosystems. Journal of Arid Environments 16: 257–262. Gaston, K.J. 1994. Rarity. London, Chapman and Hall. Gaston, K.J. 1996. What is biodiversity? In Biodiversity—A Biology of Numbers and Diﬀerence, K.J. Gaston (ed.). Oxford, Blackwell Science: 1–9. Gleason, H.A. 1917. The structure and development of the plant association. Bulletin of the Torrey Botanical Club 44: 463–481. Gleason, H.A. 1926. The individualistic concept of the plant association. Bulletin of the Torrey Botanical Club 53: 7–26. Glenn, S.M., and Collins, S.L. 1990. Patch structure in tallgrass prairies: Dynamics of satellite species. Oikos 57(2): 229–236. Glenn, S.M., Collins, S.L., and Gibson, D.J. 1992. Disturbances in tallgrass prairie: Local and regional eﬀects on community heterogeneity. Landscape Ecology 7(4): 243–251. Gutzwiller, K.J., and Anderson, S.H. 1992. Interception of moving organisms: Inﬂuences of patch shape, size, and orientation on community structure. Landscape Ecology 6(4): 293–303. Hanski, I. (1987). Colonization of ephemeral habitats. In Colonization, Succession and Stability, A.J. Gray, M.J. Crawley, and P.J. Edwards (eds.). London, Blackwell Scientiﬁc Publications: 155–185.

Linking Species Diversity and Landscape Diversity

187

Hanski, I.A., and Gilpin, M.E. (eds.). 1997. Metapopulation Biology. San Diego, Academic Press. Hanski, I., Kouki, J., and Halkka, A. (1993). Three explanations of the positive relationship between distribution and abundance. In Species Diversity in Ecological Communities, R.E. Ricklefs and D. Schluter (eds.). Chicago, University of Chicago Press: 108–116. Hanski, I., Moilanen, A., and Gyllenberg, M. 1996. Minimum viable metapopulation size. American Naturalist 147(4): 527–541. Hoagland, B.W., and Collins, S.L. 1997. Gradient models, gradient analysis, and hierarchical structure in plant communities. Oikos 78(1): 23–30. Jonsson, B.G., and Esseen, P. A. 1990. Treefall disturbance maintains high bryophyte diversity in a boreal spruce forest. Journal of Ecology 78(4): 924–936. Keddy, P.A. 1992. Assembly and response rules: Two goals for predictive community ecology. Journal of Vegetation Science 3(2): 157–164. Kotler, B.P. 1992. Behavioral resource depression and decaying perceived risk of predation in two species of coexisting gerbils. Behavioral Ecology and Sociobiology 30: 239–244. Kotliar, N.B., and Wiens, J.A. 1990. Multiple scales of patchiness and patch structure: A hierarchical framework for the study of heterogeneity. Oikos 59: 253–260. Ludwig, J. A., and Tongway, D.J. 1995. Desertiﬁcation in Australia: An eye to grass roots and landscapes. Environmental Monitoring and Assessment 37(1–3): 231–237. MacArthur, R.H., and Pianka, E.R. 1966. On optimal use of a patchy environment. American Naturalist 100: 603–609. MacArthur, R.H., and Wilson, E.O. (1967). The Theory of Island Biogeography. Princeton, NJ, Princeton University Press. Obeso, J.R. 1992. Geographic distribution and community structure of bumblebees in the northern Iberian peninsula. Oecologia 89: 244–252. Oren, Y. 2001. Patchiness, disturbances, and the ﬂows of matter and organisms in an arid landscape: A multiscale experimental approach. Ph.D. thesis, Ben-Gurion University of the Negev. Pa¨rtel, M., Zobel, M., Zobel, K., and Van der Maarel, E. 1996. The species pool and its relation to species richness: Evidence from Estonian plant communities. Oikos 75: 111–117. Pickett, S.T.A., and White, P.S. 1985. Patch dynamics: a synthesis. In The Ecology of Natural Disturbance and Patch Dynamics, S.T.A. Pickett and P.S. White (eds.). Orlando, Academic Press: 371–384. Pickett, S.T.A., Shachak, M., Boeken, B., and Armesto, J.J. 1999. The management of ecological systems. In Arid Lands Management—Toward Ecological Sustainability, T. Hoekstra and M. Shachak (eds.). Urbana, University of Illinois Press: 8–17. Platt, W.J. 1975. The colonization and formation of equilibrium plant species associations on badger disturbances in a tall-grass prairie. Ecol. Monogr. 45: 285–305. Reynolds, J.F., Virginia, R.A., Kemp, P.R., De Soyza, A.G., and Tremmel, D.C. 1999. Impact of drought on desert shrubs: eﬀects of seasonality and degree of resource island development. Ecological Monographs 69(1): 69–106. Ritchie, M.E., and Olﬀ, H. 1999. Herbivore diversity and plant dynamics: compensatory vs. additive eﬀects. In Herbivores: Betwen Plants and Predators, H. Olﬀ, V.K. Brown and R. Drent (eds.). Blackwell, Oxford: 175–204.

188

Ecological Complexes of Biodiversity, Ecosystems, and Landscapes

Rosenzweig, M.L. 1995. Species Diversity in Space and Time. Cambridge, Cambridge University Press. Shachak, M., Sachs, M., and Moshe, I. 1998. Ecosystem management of desertiﬁed shrublands in Israel. Ecosystems 1: 475–483. Shachak, M., Pickett, S.T.A., Boeken, B., and Zaady, E. 1999. Managing patchiness, ecological ﬂows, productivity, and diversity in drylands: Concepts and applications in the Negev desert. In Arid Lands Management—Toward Ecological Sustainability, T. Hoekstra and M. Shachak (eds.). Urbana, University of Illinois Press: 254–263. Sousa, W.P. 1984. The role of disturbance in natural communities. Annual Review of Ecology and Systematics 15: 353–391. Stern, E., Gradus, Y., Meir, A., Krakover, S., and Tsoar, H. 1986. Atlas of the Negev. Beer-Sheva, Ben-Gurion University of the Negev Press. Thorhallsdottir, T.E. 1990. The dynamics of a grassland community: A simultaneous investigation of spatial and temporal heterogeneity at various scales. Journal of Ecology 78(4): 884–908. Tilman, D. 1993. Species richness of experimental productivity gradients: How important is colonization limitation? Ecology 74(8): 2179–2191. Weiher, E., and Keddy, P.A. 1995. The assembly of experimental wetland plant communities. Oikos 73(3): 323–335. Whittaker, R.H., and Niering, W.A. 1965. Vegetation of the Santa Catalina Mountains, Arizona: A Gradient Analysis of the South Slope. Ecology 46(4): 429–452. Wiens, J.A. 1991. Ecological similarity of shrub-desert avifaunas of Australia and North America. Ecology 72(2): 479–495. Wiens, J.A., Stenseth, N.C., Van Horne, B., and Ims, R.A. 1993. Ecological mechanisms and landscape ecology. Oikos 66(3): 369–380. Wilson, D. S. 1992. Complex interactions in metacommunities, with implications for biodiversity and higher levels of selection. Ecology 73(6): 1984–2000. Wright, D. H. 1991. Correlations between incidence and abundance are expected by chance. Journal of Biogeography 18: 463–466. Yair, A., and Shachak, M. 1982. A case study of energy, water and soil ﬂow chains in an arid ecosystem. Oecologia 54: 389–397. Zaady, E., and Shachak, M. 1994. Microphytic soil crust and ecosystem leakage in the Negev desert. American Journal of Botany 81: 109.

11 The Impact of Animals on Species Diversity in Arid-Land Plant Communities Andrew Wilby Bertrand Boeken Moshe Shachak

T

here are many mechanisms whereby animal activity can directly or indirectly inﬂuence the species diversity of plant communities. Most obviously, herbivory can inﬂuence the species composition directly through plant mortality or indirectly by changing the outcome of interspeciﬁc competition (Hulme 1996). Animals may also aﬀect plant species composition by modifying the physical structure of the environment such that the ﬂow of resources required for plant growth is altered. Such physical eﬀects mediated by the physical structure of the environment have been termed ‘ecosystem engineering’ (deﬁned in table 11.1; Jones et al. 1994, 1997, Lawton 1994, Lawton and Jones 1995). Animals may have other functions that inﬂuence the persistence of populations (e.g., pollination) or the colonization of new sites (seed dispersal). In this chapter we aim to provide an overview of how these diverse eﬀects of animals inﬂuence plant species diversity, and to this end, we propose community assembly theory as a conceptual framework. Community assembly theory provides us with a schematic representation of the vital steps involved in the determination of species presence or absence at a particular site. By asking how might animal activity inﬂuence each of the steps of assembly, we ensure a comprehensive outlook on how animals aﬀect plant species diversity. That said, we restrict ourselves in this chapter to ecological mechanisms and we do not consider evolutionary eﬀects that are of undoubted importance at higher levels of spatial and temporal scales. Following our discussion of animal eﬀects on community assembly, we highlight two case studies of herbivores arising from recent research in Israel. We use insights provided by these examples to suggest mechanisms that are likely to be of particular importance in arid ecosystems. For example, we suggest 189

190

Ecological Complexes of Biodiversity, Ecosystems, and Landscapes

Table 11.1 Glossary of terms used in the text Term

Deﬁnition

Physical ecosystem engineering

The direct or indirect control of the availability of resources to other organisms by physical state changes in biotic or abiotic materials (Jones et al. 1997)

Transport ecosystem engineering

Control of the availability of resources to other organisms by directly changing the spatial distribution of abiotic or biotic materials (Wilby et al. 2001)

Total species pool (TSP)

All those species that occur in the region of a target site (Belyea and Lancaster 1999)

Geographic species pool (GSP)

The subset of species from the TSP that are able to reach the target site by dispersal (Belyea and Lancaster 1999)

Habitat species pool (HSP)

The subset of the TSP that are able to live under the abiotic conditions of the target site (Belyea and Lancaster 1999)

Ecological species pool (ESP)

Those species in both the GSP and ESP (Belyea and Lancaster 1999)

Actual species pool (ASP)

Those species in the ESP that persist despite biological interactions with other species, i.e., the observable community (Belyea and Lancaster 1999)

that since the ﬂow of water, the primary limiting factor in these systems, is so easily inﬂuenced by structural changes in the environment, physical ecosystem engineering may be a relatively important, though previously neglected, interaction type in arid ecosystems.

Animal Interactions with Plant Community Assembly The processes governing which species occur in a particular plant community can be summarized very simply. For species to occupy a site they must be capable of dispersing to the site, they must have the correct physiological attributes to survive at the site, and they must be capable of coexisting with other species at the site. These three broad processes connect between a set of species pools that have been labeled the total (TSP), geographical (GSP), habitat (HSP), ecological (ESP), and actual (ASP) species pools (see ﬁg. 11.1 and table 11.1 for deﬁnitions). Previously, much attention has been given to the biotic ﬁltration mechanisms in the determination of species diversity, in particular the ways in which species are able to avoid competitive exclusion (e.g., Tilman and Pacala 1993). Increasingly, however, evidence amasses that emphasizes the importance of abiotic and dispersal constraints in community assembly (Pa¨rtel et al. 1996, Zobel 1997, Lawton 1999, Tofts and Silvertown 2002). Our contention is that the diverse activities of animals can potentially inﬂuence each of the processes of community assembly,

The Impact of Animals on Species Diversity

191

Figure 11.1 The process of plant community assembly showing subprocesses that connect the total species pool to the actual species pool (terminology follows Belyea and Lancaster, 1999). Animal inﬂuences are labeled: TI, trophic interaction; PE, physical ecosystem engineering; TE, transport ecosystem engineering.

through changing the expression of biotic, abiotic, and dispersal constraints, thereby inﬂuencing plant species diversity (ﬁg. 11.1). Animals inﬂuence dispersal constraints by changing the way that plant propagules move within ecosystems. Many plant species have seeds or fruits, which have adaptations to promote dispersal by animals (Howe and Westley 1997). Thus, animals can inﬂuence species arrival by dispersing seeds following consumption of the seeds (endozoochory), by carrying seeds on the external surface of their bodies (ectozoochory), or by changing the structure of the environment such that dispersal vectors are inﬂuenced (e.g., Boeken et al. 1995). Animals can also have a large impact on abiotic conditions at a site, and therefore abiotic ﬁltration of the GSP, by changing the rate or the site at which nutrients are cycled (Jeﬀeries et al. 1994, Polis et al. 1997, Jeﬀeries 1999), or by modifying the structure of the landscape such that resource ﬂow is modiﬁed (e.g., Gibson 1989, Boeken et al. 1995, Whitford 1999). Finally, animals can alter biotic ﬁltering of plant commu-

192

Ecological Complexes of Biodiversity, Ecosystems, and Landscapes

nity composition by consuming live plant material (herbivory) or by nontrophic plant–animal interactions such as pollination (Howe and Westley 1997). Despite this broad array of ecological interactions between animals and plants, a large proportion of research concerning the inﬂuence of animals on plant communities has focused on trophic interactions between plants and animals, i.e., the consequences of herbivory (deﬁned as the consumption of plant biomass by animals). Consequently, our understanding of how animals aﬀect plant community assembly mostly encompasses animal inﬂuences on biotic ﬁltering as opposed to their eﬀect of abiotic ﬁltering and species arrival processes. Herbivory has often been treated in theoretical studies as either a form of disturbance, or as a modiﬁer of competitive relationships between plants (Hulme 1996). Many theoretical studies relate the inﬂuence of herbivores on plant diversity to diﬀerential impacts on dominant and inferior competitors (Louda et al. 1990, Pacala and Crawley 1992, Tilman and Pacala 1993). More recently, other direct eﬀects of herbivores on plant survivorship have been shown to have a strong inﬂuence on species diversity (Hulme 1996). While mature plants rarely die as a consequence of herbivory, juvenile stages often do (Crawley 1983), and seed and seedling predation has been shown to be an important factor structuring plant communities (Brown and Davidson 1977, Brown et al. 1979, Stinson 1983, Wilby 1996, Wilby and Brown 2001). Relative to the inﬂuence of herbivory on plant species diversity, the inﬂuence of nontrophic interactions has been under-studied. Prior to the introduction of the concept of ecosystem engineering (Jones et al. 1994, 1997, Lawton 1994, Lawton and Jones 1995), interactions occurring as a result of modiﬁcation of the functional structure of the landscape by organisms were not formalized. This was despite extensive study of speciﬁc examples, such as soil disturbance by fossorial mammals (Platt 1975, Inouye et al. 1987, Huntly and Inouye 1988, Hobbs and Mooney 1991) or the more extensive ecosystem eﬀects of beavers (e.g., Thayer 1979, Naiman et al. 1988). We believe that the lack of formalization into theory of these types of eﬀects hindered the interpretation of animal roles in ecosystems. The problem was highlighted by Huntly (1995), who pointed out that responses of vegetation to experimental exclosure/enclosure techniques are often assumed to be due to herbivory alone, ignoring the range of other potential interaction types. Olﬀ and Ritchie (1998) integrated trophic and ecosystem engineering eﬀects in their assessment of the role of herbivores in maintaining plant species diversity in grasslands. In our view, such a comprehensive approach to animal–plant interactions, connected with an appreciation of the important processes governing plant community assembly, is necessary if we are to fully understand the eﬀects of animals on species diversity in plant communities. The following case studies employ a comprehensive approach, inclusive of trophic and ecosystem engineering eﬀects, to highlight mechanisms of animal inﬂuence on plant community assembly and plant species diversity, which are of particular importance in arid systems.

The Impact of Animals on Species Diversity

193

Case Studies The Indian Crested Porcupine Porcupines (Hystrix indica) are widely distributed in South and Central Asia and the Middle East. In the arid and semiarid areas of Israel they occur in all habitat types (Alkon 1999). Porcupines are relatively large mammals (ca. 14 kg), which feed principally on below-ground storage organs of plants including tubers, rhizomes, corms, and bulbs (Gutterman 1982, 1987, Alkon and Saltz 1986). Since their principal food is below ground, porcupines are important agents of soil disturbance (Gutterman and Herr 1981, Gutterman 1982, 1987, 1988, Alkon and Olsvig Whittaker 1989, Gutterman et al. 1990, Shachak et al. 1991, Boeken et al. 1995). Typically, porcupines excavate soil pits (15 cm deep, 15–20 cm diameter), which persist for several to 20 years depending on the substrate (Alkon 1999). In a 17-year study on a hill in the Central Negev highlands, an area receiving an average of circa 90 mm rainfall per year, porcupine diggings were found to occur predominantly in midslope areas where plant biomass was maximal due to run-oﬀ water accumulation (Shachak et al. 1991, Boeken et al. 1995). Each year, between 1.4% and 3.5% of the soil surface area was disturbed by porcupine digging. We can identify two major routes whereby porcupines may inﬂuence plant species diversity. First, they consume the perennating organs of geophytic plants, therefore aﬀecting the survival of individuals, potentially limiting population size of the consumed species. Herbivory may also aﬀect competitive relationships between the consumed geophytes and other plant species (eﬀects on biotic ﬁltration, ﬁg. 11.1). Second, their digging may inﬂuence the ﬂow of abiotic and biotic materials in the landscape and thereby plant community structure (eﬀects on species arrival and abiotic ﬁltration, ﬁg. 11.1). The relatively little data that exist concerning the trophic eﬀect of bulb consumption by porcupines suggest that the eﬀect of herbivory on geophyte populations is low. In a 5-year study of a population of Tulipa systola, in the Central Negev highlands (Boeken 1986), it was found that predation rates were extremely variable, but generally very low, while recruitment, although equally variable, was consistently greater. These and similar observations of a number of other geophyte species (Boeken 1986, 1989, 1990, Gutterman 1988) suggest that porcupine consumption does not limit geophyte populations. There have been a number of studies of the eﬀect of porcupine diggings on the plant assemblage (reviewed in Alkon 1999), showing the eﬀect of porcupine consumption of bulbs by means of structural changes in the landscape. Pits dug by porcupines accumulate organic material, including seeds, and water (Gutterman and Herr 1981, Yair and Shachak 1982, 1987, Boeken et al. 1995, 1998). Accumulation occurs as run-oﬀ water, after a rain event, tends to ﬂow across the surface of the soil and collects in the porcupine pits. As a result of this accumulation of materials, plant abundance tends to be much higher in porcupine pits than in undisturbed areas (Gutterman

194

Ecological Complexes of Biodiversity, Ecosystems, and Landscapes

and Herr 1981, Gutterman 1982, 1987, 1988, Alkon and Olsvig Whittaker 1989, Gutterman et al. 1990, Shachak et al. 1991). Plant community response to porcupine digging was documented in detail by Boeken et al. (1995). Plant density, biomass, and species richness were found to be considerably higher in pits than in areas of undisturbed soil (ﬁg. 11.2). A total of 49 species were recorded in a survey of 150 porcupine pits, compared with only 28 in the same number of matched samples from undisturbed soil. There was a large overlap in the species recorded in pits and on undisturbed soil (25 of the 28 species recorded on the undisturbed soil were also found in pits) and those species found exclusively in either the pits or on the crusted soil were recorded very infrequently. It seems that the altered abiotic conditions in the pits and increased seed capture increase local species richness. The species that occurred exclusively in the pits perhaps require more mesic conditions, but few of those species found on the crusted soil do not also occur in the pits. Following a complex analysis involving an estimation of seed source strength and water source strength at various areas of the study site, Boeken et al. (1995) showed that plant density and diversity on the undisturbed soil crust were limited by microsite availability due to a lack of water inﬁltration. In contrast, the pits remained moist throughout the growing season and plant density and diversity were limited by seed arrival.

Harvester Ants Several species of harvester ant of the genus Messor inhabit the arid regions of Israel. All feed principally on seeds and fruits, although they also incorporate animal and vegetative plant parts in the diet, particularly during periods when seeds are in short supply. Having processed food items, all species discard inedible parts of food items (e.g., seed coats, awns) on ‘‘chaﬀ piles’’

Figure 11.2 Comparison of (a) plant density, (b) total biomass, and (c) species richness of the plant assemblages occupying porcupine pits and undisturbed soil crust. Means are of 150 porcupine pits and paired undisturbed soil plots. Pits and undisturbed plots averaged 270 cm2 in area.

The Impact of Animals on Species Diversity

195

close to the nest entrance (Steinberger et al. 1991). Occasionally, viable seeds are discarded on chaﬀ piles, particularly if they are protected from seed predation by a hardened seed coat. Two common ant species, Messor arenarius and M. ebeninus, are widely distributed in the Negev, often coexisting at the same sites. Messor ebeninus constructs large, permanent nest mounds ranging in size from 0.5 to 3 m in diameter where vegetation is removed and chaﬀ is discarded. In contrast, M. arenarius constructs small mounds and chaﬀ piles (< 0.5 m diameter) around the nest entrance. The mounds of M. arenarius are used for short periods, new entrances being constructed up to several times each year (A. Wilby pers. obs.). As in the porcupine case, we can identify several pathways whereby Messor species are likely to inﬂuence plant community composition. Both M. ebeninus and M. arenarius collect large numbers of seeds, and exhibit strong selectivity between plant species (Steinberger et al. 1991). Selective feeding by granivores has been shown to have large eﬀects on plant diversity and species composition in arid ecosystems (Brown et al. 1979, Samson et al. 1992, Davidson 1993, Valone et al. 1994), so there is evidence that granivory may have an important trophic impact on the relative abundance of plant species (biotic ﬁltration—direct eﬀects of herbivory, ﬁg. 11.1). In addition to the eﬀect of granivory, construction of nest mounds by ants is also known to aﬀect the composition of plant communities. The composition of nest-mound vegetation is often diﬀerent from the surrounding vegetation due to the combined eﬀects of soil modiﬁcation (abiotic ﬁltration, ﬁg. 11.1), seed dispersal (species arrival, ﬁg. 11.1) and seed predation (biotic ﬁltration, ﬁg. 11.1) (Danin and Yom Tov 1990, Dean and Yeaton 1993, Whitford and Dimarco 1995). The inﬂuence on plant species diversity of nest mounds of M. ebeninus and the feeding activities of M. ebeninus and M. arenarius were studied over ﬁve years in a semiarid shrubland in the northern Negev Desert. Physical exclusion techniques were used to manipulate harvester ant access to a series of experimental plots (1 m 0.5 m), each enclosing an undisturbed soil area and a shrub (Wilby et al. 2001). The experiment commenced in the late spring of 1997, following pretreatment surveys of the plant community composition. After four experimental seasons, ant exclusion had no signiﬁcant eﬀect on plant species richness, though there were eﬀects on the abundance of individual species and on the stability of total plant abundance (Wilby and Shachak, unpublished work). On the same research site, vegetation on nest mounds of M. ebeninus was compared with that on undisturbed soil (Wilby et al. 2001). Plant density, biomass and species richness were found to be signiﬁcantly higher on nest mounds than on the undisturbed soil crust (ﬁg. 11.3). A total of 55 species were recorded on the nest mounds, compared with only 25 on the undisturbed soil. Similarly to the porcupine pit example, there was a large overlap in the species recorded on the mounds and on undisturbed soil (23 of the 25 species recorded on the undisturbed soil), but many additional species occurred on the nest mounds, compared with the crusted soil. Almost all plant species

196

Ecological Complexes of Biodiversity, Ecosystems, and Landscapes

Figure 11.3 Comparison of (a) plant density, (b) total biomass, and (c) species richness of the plant assemblages occupying nest mounds of Messor ebeninus and undisturbed soil crust. Twenty vegetation samples 20 cm 20 cm were taken from each of 20 nest mounds and compared with vegetation samples of equal size taken from the undisturbed soil surface approximately 1 m from the edge of the nest mound.

occurred with approximately equal or higher plant density on the nest mounds than on the undisturbed crust, and with higher incidence, that is, they were present in more samples on the nest mounds than in the undisturbed soil, and where they occurred, they were at higher density (ﬁg. 11.4). Two plant species exhibited responses that were indicative of more direct control by the ants. The favored food items of M. ebeninus are seeds of the annual grass, Stipa capensis. This species is the most common species, occurring in all samples of both mound vegetation and undisturbed soil vegetation. However, in contrast with all other species, it occurred at much lower density on the nest mounds, reﬂecting more intensive seed collection close to the nest entrance (ﬁg. 11.4b). The incidence of another common species, the crucifer Reboudia pinnata, increased from circa 10% of samples on the undisturbed soil to 85% of mound samples (ﬁg. 11.4a), with only a small increase in local abundance (ﬁg. 11.4b). This diﬀerence occurred because the terminal seeds in each pod of R. pinnata are surrounded by a hardened fruit wall, which protects the seeds from predation (Gutterman 1993). Consequently, seeds of R. pinnata are frequently discarded on chaﬀ piles, thus increasing this species’ incidence on nest mounds.

Animal Roles in Arid Ecosystems Dispersal Effects Both of our case studies highlight important mechanisms whereby animal activity inﬂuences species arrival and consequently local species diversity (summarized in table 11.2). Porcupines engineer the surface of the soil, creat-

The Impact of Animals on Species Diversity

197

Figure 11.4 Comparison of (a) incidence (proportion of samples occupied) and (b) abundance (mean number of individuals per occupied sample) of species in the plant assemblages on nest mounds of Messor ebeninus and on undisturbed soil. Each point represents a single species: Sc, Stipa capensis; Rp, Reboudia pinnata.

ing pits, which act as sinks for dispersing seeds. Harvester ant mounds probably function in a similar way, as the mound increases the topographic complexity of the soil surface, and the chaﬀ pile provides structural complexity, which may halt the secondary dispersal of some species. The harvester ant study also provided evidence for more direct control of seed dispersal, at least for one species, R. pinnata (Wilby et al. 2001). Thus, we can identify two broad categories of inﬂuence by animals on species arrival: (1) physical ecosystem engineering of landscape structure, such that the ﬂow of dispersal vectors is intercepted, and (2) zoochory, or animal-mediated seed dispersal. Taking the physical engineering mechanisms ﬁrst, we know that many seeds have structural adaptation to facilitate dispersal by wind or water and tend to accumulate where structures in the landscape intercept these vectors. Surface ﬂow or ﬂooding of water has been shown to redistribute seeds and organic

198

Ecological Complexes of Biodiversity, Ecosystems, and Landscapes

Table 11.2 Summary of the apparent effects of porcupines and harvester ants on the plant communities of the Negev Desert, Israel Assembly Process

Porcupine Eﬀects

Harvester Ant Eﬀects

Species Arrival

Feeding-induced pits trap seeds— increased local species richness

Seeds are collected and discarded on nest mounds—increased frequency of some species Mounds trap windborne seeds— increased local species diversity

Abiotic ﬁltration

Feeding-induced pits accumulate run-oﬀ water—increased local productivity and species richness

Nutrient enrichment of nestmound soils—increased local productivity and species richness Nest mounds increase rainfall and run-oﬀ water inﬁltration— increased local productivity and species richness

Biotic ﬁltration

Consumption of geophyte bulbs— little apparent eﬀect on geophyte or annual populations

Seed predation is known to impact plant species composition in other systems—no evidence of eﬀect in our studies

matter and deposit them on high features in the landscape in an arid region of Australia (Mott and McComb 1974). Wind acts in a similar way in the Sonoran Desert of the United States (Reichman 1984). Hence, if animals alter the structure of the environment such that the ﬂow of a dispersal vector is altered, seed distribution is also likely to be altered. An example of physical ecosystem engineering similar to that of porcupines (also see Gutterman and Herr 1981, Boeken et al. 1995, Boeken et al. 1998) is shown by goannas in Australia (Whitford 1999). We suggest that physical ecosystem engineering eﬀects on seed dispersal may be relatively more important in arid ecosytems than mesic ecosystems. The two principal dispersal vectors, wind and water, tend to ﬂow across the soil surface in arid systems due to low vegetation or litter cover, and soil surfaces with low water permeability. In contrast to soil disturbance, the inﬂuence of zoochory on seed dispersal is well studied. Two major classes of zoochory have been identiﬁed in the literature. Endozoochory involves the dispersal of propagules following ingestion of seeds or fruit. Endozoochory is known to be an important process controlling plant species composition in successional oldﬁelds (McDonnell and Stiles 1983), mature tropical forests (Whitney et al. 1998), semiarid savanna (Hilger and Schultka 1988) and Mediterranean grasslands (Malo and Suarez 1998). A similar process, epizoochory, involves the dispersal of seeds on the external surface of animals; many plant species have seeds with distinct structures that facilitate transport by animals. Although epizoochory is rarely studied directly, it has been shown that epizoochorous plants dominate the vegetation at sites where animals congregate, such as

The Impact of Animals on Species Diversity

199

under shade trees and around waterholes (Hilger and Schultka 1988, Ernst et al. 1992). This may mean that zoochory is an important factor that determines the spatial heterogeneity of vegetation in arid ecosystems. Abiotic Effects Both of our case studies showed marked eﬀects of animal activity on abiotic ﬁltration of the species pool (table 11.2). We suggest that soil disturbance has a relatively stronger impact on local plant diversity in arid ecosystems than on more mesic ecosystems. This is because plants in arid ecosystems are primarily limited by water, the redistribution of which is sensitive to the physical structure of the landscape. Redistribution of water is responsible for marked patchiness of vegetation in arid systems (Yair and Danin 1980, Yair and Shachak 1982, Tongway and Ludwig 1994, Aguiar and Sala 1999). As we have seen, structural modiﬁcation of soils by fossorial animals often results in increases in the local availability of water in arid ecosystems, and consequently has a marked eﬀect on productivity and diversity of vegetation. It is notable that animals may also have the opposite eﬀect on vegetation. The activity of larger animals, such as large ungulates, compacts soil, therefore reducing water inﬁltration and resulting in lower productivity of vegetation (Fuls 1992, Holt et al. 1996). Nutrient enrichment due to animal activity may also be of particular importance to plant community assembly in arid ecosystems. Spatial redistribution of nutrients is an inevitable part of trophic interactions that involve mobile animals. The eﬀect on plant communities is most evident when the animals habitually return to the same area, as is the case for central-place foragers or visitation of water sources by arid-land animals. Nutrient transfer by animals from one area to another may drive production in some unproductive environments, providing links between seemingly separate food webs (Polis et al. 1997). On a smaller scale, snails in the Negev desert are important consumers of plant material and their habitual ‘roosting’ on shrubs (and estivating under shrubs) is an important mechanism of fertilization of the soils under shrubs (Zaady et al. 1996). Although areas of high animal activity are usually also areas of nutrient accumulation, animal activity can also decrease nutrient availability to plants. For example, in the arid areas of the southwestern United States, termites consume a large proportion of primary productivity, but the resulting nutrients are transferred deep into the subterranean nests out of reach of most plants (Nash and Whitford 1995). While water is the primary limiting factor in arid ecosystems, during wet periods, nutrients can become limiting to plant growth. Hence, in arid ecosystems, animals may have a marked inﬂuence on plant communities by changing the concentration and spatial pattern of soil nutrients. Animals are known to increase the rates of nitrogen cycling (Jeﬀeries 1999) by digesting cellulose-rich plant material and defecating material that is much more amenable to mineralization. Since nitrogen mineralization is extremely sensitive to drying, animal facilitation of nitrogen mineralization is probably par-

200

Ecological Complexes of Biodiversity, Ecosystems, and Landscapes

ticularly important in arid ecosystems. Exactly what eﬀect increases in nutrient supply rate have on plant diversity in arid ecosystems is unclear. On poor soils in arid regions vegetation is often extremely sparse, so increases in productivity do not necessarily lead to increased intensity of competitive interactions and to competitive exclusion. Indeed, on the resource-rich nest mounds of M. ebeninus, plant biomass was increased approximately eightfold, while small- and medium-scale species richness increased considerably (ﬁg. 11.3). Animal inﬂuences on plant diversity, driven by alterations in local abiotic conditions are likely to be important in areas with soils of low fertility generally, and particularly in arid regions where lack of moisture limits nitrogen mineralization rate. Biotic Effects Animals can inﬂuence plant diversity by modiﬁcation of, or involvement in interspeciﬁc interactions among plants. Herbivore eﬀects on plant communities are often considered to result from changes in the rates of competitive exclusion between plant species (Brown 1990, Frank and McNaughton 1992, McInnes et al. 1992, Ritchie and Tilman 1995). Changes in the abiotic environment as a result of animal activity, as discussed in the previous section, may also inﬂuence plant competition. Alternatively, or in addition, herbivory may have an impact on plant diversity directly by inﬂuencing plant survivorship (Brown and Davidson 1977, Brown et al. 1979, Stinson 1983, Wilby and Brown 2001). As presented by Hulme (1996), theoreticians have often concentrated on the indirect eﬀect of herbivory, mediated by plant competition, rather than on the direct eﬀect on plant mortality. Our case studies support the contention that in arid ecosystems, direct eﬀects of herbivory on plant survival may be more important in controlling plant species diversity than indirect eﬀects mediated via plant competition. The porcupine case study suggested that herbivory did not have a large impact on plant species diversity by direct or indirect mechanisms. However, we did observe a trophic eﬀect of granivory by harvester ants in reducing plant density. An explanation for the general absence of eﬀects mediated through competitive interactions may be found in evolutionary rather than ecological interactions. Seasonal patterns in water availability in arid lands have favored ephemeral life histories among most plants. Those plants that exhibit nonephemeral life histories (this includes annuals and geophytes) tend to be heavily protected from herbivory (Olﬀ and Ritchie 1998), at least on infertile soils. Therefore, while herbivores have probably been a strong inﬂuence on evolutionary processes in arid plants, at ecological scales, herbivory of vegetative plant parts is not a major feature of arid-land food webs on infertile soils. Specialized herbivores, such as granivores or those of perennating organs such as porcupines, may, however, have important inﬂuences on plant community composition and potentially on plant diversity, but their eﬀect is likely to be via changes in plant species survivorship rather than via modiﬁcation of interactions between plant species.

The Impact of Animals on Species Diversity

201

Conclusions and Future Directions Our studies indicate that animal inﬂuences on each of the three processes of community assembly, species arrival, and abiotic and biotic ﬁltration, play a role in plant community assembly. Animals are able to inﬂuence all of these processes by a trophic and ecosystem engineering interactions. Evidence from our case studies and the literature suggest that certain types of engineering and trophic interactions may be more important in arid systems than in mesic systems. Soil disturbance has a great inﬂuence in arid systems because it causes redistribution of water, the primary limiting resource. Soil disturbance also inﬂuences species diversity by interception of wind- and run-oﬀ-borne seeds. Trophic eﬀects on plant recruitment (e.g., seed or seedling predation) are particularly important to the composition of the annual component of arid-land plant communities, as are animal eﬀects on nutrient cycling, since microbial activity is often severely retarded by arid conditions. Above all, the aim of this chapter is to communicate the broad range of potential interactions between animals and plants that can inﬂuence plant species diversity. We call for further integrative studies of animal eﬀects on plant communities in a wide range of ecosystem types. It is possible that the interaction of plant life history and the physical structure of ecosystems will mean that diﬀerent interaction types predominate in diﬀerent biomes. Our preliminary analysis points to the importance of physical ecosystem engineering in arid ecosystems. However, our data are based on eﬀects at small spatial scales. An important challenge for the future will be to scale up to look at the eﬀect of these activities on the species diversity and species abundance distributions at landscape scales. A second area of interest is the relationships and feedbacks between animal movement and the spatial structure of plant communities. By transporting nutrients and by altering seed ﬂow, animals seem to have strong eﬀects on the distribution of plants in arid systems, but are these processes equally important in mesic environments? Thirdly, if arid environments are particularly sensitive to physical disturbance, this will have important implications for the conservation of arid areas and in particular, issue of human access to sites of conservation importance. The resolution of these issues will require further comparative work on the functioning of ecosystems in diﬀerent biomes and an integrative approach to the functional consequences of animal activities in ecosystems.

References Aguiar, M.R., and O.E. Sala. 1999. Patch structure, dynamics and implications for the functioning of arid ecosystems. Trends in Ecology and Evolution 14: 273–277. Alkon, P.U. 1999. Microhabitat to landscape impacts: Crested porcupine digs in the Negev Desert highlands. Journal of Arid Environments 41: 183–202. Alkon, P., and D. Saltz. 1986. Crested porcupine activity patterns: ecological and management implications. In: Dubinsky, Z., and Steinberger, Y. (eds.). Environ-

202

Ecological Complexes of Biodiversity, Ecosystems, and Landscapes

mental Quality and Ecosystem Stability. Bar Ilan University Press, Ramat Gan, Israel. Alkon, P.U., and L.O. Whittaker. 1989. Crested porcupine digs in the Negev desert highlands (Israel): Patterns of density, size, and longevity. Journal of Arid Environments 17: 83–96. Belyea, L.R., and J. Lancaster. 1999. Assembly rules within a contingent ecology. Oikos 86: 402–416. Boeken, B. 1986. Utilization of reserves in some desert geophytes. PhD thesis, BenGurion University of the Negev. Boeken, B. 1989. Life histories of desert geophytes: The demographic consequences of reproductive biomass partitioning patterns. Oecologia 80: 278–283. Boeken, B. 1990. Life histories of two desert species of the bulbous genus Bellevalia: The relation between biomass partitioning and water availability. Oecologia 82: 172–179. Boeken, B., M. Shachak, Y. Gutterman, and S. Brand. 1995. Patchiness and disturbance: Plant community responses to porcupine diggings in the central Negev. Ecography 18: 410–422. Boeken, B., C. Lipchin, Y. Gutterman, and N. van Rooyen. 1998. Annual plant community responses to density of small-scale disturbances in the Negev Desert of Israel. Oecologia 114: 106–117. Brown, J.H. and D.W. Davidson, 1977. Competition between seed-eating rodents and ants in desert ecosystems. Science 196: 880–882. Brown, J.H., O.J. Reichman, and D.W. Davidson. 1979. Granivory in desert ecosystems. Annual Review of Ecology and Systematics 10: 201–227. Brown, V.K. 1990. Insect herbivory and its eﬀect on plant succession. In: Burdon, J.J., and Leather, S.R. (eds). Pests, Pathogens and Plant Communities. Blackwell Scientiﬁc Publications, Oxford, 275–287. Crawley, M.J. 1983. Herbivory: the Dynamics of Animal-Plant Interactions. Blackwell Science, Oxford. Danin, A., and Y. Yom Tov. 1990. Ant nests as primary habitats of Silybum marianum (Compositae). Plant Systematics and Evolution 169: 209–218. Davidson, D.W. 1993. The eﬀects of herbivory and granivory on terrestrial plant succession. Oikos 68: 23–35. Dean, W.R.J., and R.I. Yeaton. 1993. The inﬂuence of harvester ant Messor capensis nest mounds on the productivity and distribution of some plant species in the Southern Karoo, South Africa. Vegetatio 106: 21–35. Ernst, W.H.O., E.M. Veendendaal, and M.M. Kebakile. 1992. Possibilities for dispersal in annual and perennial grasses in a savanna in Botswana. Vegetatio 102: 1–11. Frank, D.A., and S.J. McNaughton. 1992. The ecology of plants, large mammalian herbivores, and drought in Yellowstone National Park. Ecology 73: 2043–2058. Fuls, E.R. 1992. Ecosystem modiﬁcation created by patch-overgrazing in semi-arid grassland. Journal of Arid Environments 23: 59–69. Gibson, D.J. 1989. Eﬀects of animal disturbance on tallgrass prairie vegetation. American Midland Naturalist 121: 144–154. Gutterman, Y. 1982. Observations on the feeding habits of the Indian crested porcupine (Hystrix indica) and the distribution of some hemicryptophytes and geophytes in the Negev desert highlands. Journal of Arid Environments 5: 261–268.

The Impact of Animals on Species Diversity

203

Gutterman, Y. 1987. Dynamics of porcupine (Hystrix indica kerr) diggings: their role in the survival and renewal of geophytes and hemicryotophyes in the Negev desert Highlands. Israel Journal of Botany 6: 133–143. Gutterman, Y. 1988. An ecological assessment of porcupine activity in a desert biome. In: Ghosh, P.K. and Prakash, I. (eds.). Ecophysiology of Desert Vertebrates. Scientiﬁc Publishers, Jodhpur, India, 289–372. Gutterman, Y. 1993. Seed Germination in Desert Plants. Springer-Verlag, Berlin, Heidelberg, New York. Gutterman, Y., and N. Herr. 1981. Inﬂuences of porcupine (Hystrix indica) activity in the slopes of the northern Negev mountains—germination and vegetation renewal in diﬀerent geomorphological types and slope directions. Oecologia 51: 332–334. Gutterman, Y., T. Golan, and M. Garsani. 1990. Porcupine diggings as a unique ecological system in a desert environment. Oecologia 85: 122–127. Hilger, H.H., and W. Schultka. 1988. Plant dispersal in a semiarid thornbush savanna in northern Kenya. Botanische Jahrbuecher fuer Systematik Pﬂanzengeschichte und Pﬂanzengeographie 109: 385–406. Hobbs, R.J., and H.A. Mooney. 1991. Eﬀects of rainfall variability and gopher disturbance on serpentine annual grassland dynamics. Ecology 72: 59–68. Holt, J.A., K.L. Bristow, and J.G. McIvor. 1996. The eﬀects of grazing pressure on soil animals and hydraulic properties of two soils in semi-arid tropical Queensland. Australian Journal of Soil Research 34: 69–79. Howe, H.F., and L.C. Westley. 1997. Ecology of pollination and seed dispersal. In: Crawley, M.J. (ed). Plant Ecology. Blackwell Science, Oxford, 262–283. Hulme, P.E. 1996. Herbivory, plant regeneration and species coexistence. Journal of Ecology 84: 609–616. Huntly, N. 1995. How important are consumer species to ecosystem functioning? In: Jones, C.G. and Lawton, J.H. (eds.). Linking Species and Ecosystems. Chapman and Hall, New York, 72–83. Huntly, N. and R. Inouye, 1988. Pocket gophers in ecosystems: patterns and mechanisms. Bioscience 38:786–793. Inouye, R., N. Huntly, D. Tilman, and J.R. Tester. 1987. Pocket gophers (Geomys bursarius), vegetation, and soil nitrogen along a successional sere in east-central Minnesota. Oecologia 72: 178–184. Jeﬀeries, R.L. 1999. Herbivores, nutrients and trophic cascades in terrestrial environments. In: Olﬀ, H., et al. (eds.). Herbivores: Between Plants and Predators. Blackwell Science, Oxford, 301–330. Jeﬀeries, R.L., D.R. Klein, and G.R. Shaver. 1994. Vertebrate herbivores and northern plant communities: Reciprocal inﬂuences and responses. Oikos 71: 193–206. Jones, C.G., J.H. Lawton, and M. Shachak. 1994. Organisms as ecosystem engineers. Oikos 69: 373–386. Jones, C.G., J.H. Lawton, and M. Shachak. 1997. Positive and negative eﬀects of organisms as physical ecosystem engineers. Ecology 78: 1946–1957. Lawton, J.H. 1994. What do species do in ecosystems? Oikos 71: 367–374. Lawton, J.H. 1999. Are there general laws in ecology? Oikos 84: 177–192. Lawton, J.H. and C.G. Jones. 1995. Linking species and ecosystems: organisms as ecosystem engineers. In: Jones, C.G., and Lawton, J.H. (eds.). Linking Species and Ecosystems. Chapman and Hall, New York, 141–150. Louda, S.M., K.H. Keeler, and R.D. Nolt. 1990. Herbivore inﬂuences on plant performance and competitive interactions. In: Grace, J.B., and Tilman,

204

Ecological Complexes of Biodiversity, Ecosystems, and Landscapes

D. (eds.). Perspectives on Plant Competition. Academic Press, San Diego, 445–474. Malo, J.E., and F. Suarez. 1998. The dispersal of a dry-fruited shrub by red deer in a Mediterranean ecosystem. Ecography 21: 204–211. McDonnell, M.J., and E.W. Stiles. 1983. The structural complexity of old ﬁeld vegetation and the recruitment of bird-dispersed plant species. Oecologia 56: 109–116. McInnes, P.F., R.J. Naiman, J. Pastor, and Y. Cohen. 1992. Eﬀect of moose browsing on vegetation and litter of the boreal forest, Isle Royal, Michigan, USA. Ecology 73: 2059–2075. Mott, J., and A.J. McComb. 1974. Patterns in annual vegetation and soil microrelief in an arid region of western Australia. Journal of Ecology 62: 114–126. Naiman, R.J., C.A. Johnston, and J.C. Kelley. 1988. Alteration of North American streams by beaver. Bioscience 38: 753–762. Nash, M.N., and W.G. Whitford. 1995. Subterranean termites: Regulators of soil organic matter in the Chihuahuan Desert. Biology and Fertility of Soils 19: 15–18. Olﬀ, H., and M.E. Ritchie. 1998. Eﬀects of herbivores on grassland plant diversity. Trends in Ecology and Evolution 13: 261–265. Pacala, S.W., and M.J. Crawley. 1992. Herbivores and plant diversity. American Naturalist 140: 243–260. Pa¨rtel, M., M. Zobel, K. Zobel, and E. van der Maarel, 1996. The species pool and its relation to species richness—evidence from Estonian plant communities. Oikos 75: 111–117. Platt, W.J. 1975. The colonization and formation of equilibrium plant species associations on badger disturbances in a tall-grass prairie. Ecological Monographs 45: 285–305. Polis, G.A., W.B. Anderson, and R.D. Holt. 1997. Toward an integration of landscape and food web ecology: The dynamics of spatially subsidized food webs. Annual Review of Ecology and Systematics 28: 289–316. Reichman, O.J. 1984. Spatial and temporal variation of seed distributions in Sonoran Desert soils. Journal of Biogeography 11: 1–11. Ritchie, M.E., and D. Tilman. 1995. Responses of legumes to herbivores and nutrients during succession on a nitrogen-poor soil. Ecology 76: 2648–2655. Samson, D.A., T.E. Philippi, and D.W. Davidson. 1992. Granivory and competition as determinants of annual plant diversity in the Chihuahuan desert. Oikos 65: 61–80. Shachak, M., S. Brand, and Y. Gutterman. 1991. Porcupine disturbances and vegetation pattern along a resource gradient in an desert. Oecologia (Heidelberg) 88: 141–147. Steinberger, Y, H. Leschner, and A. Schmida. 1991. Chaﬀ piles of harvester ant (Messor spp.) nests in a desert ecosystem. Insectes Sociaux 38: 241–250. Stinson, C.S.A. 1983. Eﬀects of insect herbivores on early successional habitats. Ph.D. thesis, University of London. Thayer, C.W. 1979. Biological bulldozers and the evolution of marine benthic communities. Science 203: 458–461. Tilman, D., and S. Pacala. 1993. The maintenance of species richness in plant communities. In: Ricklefs, R.E. and D. Schluter (eds). Species Diversity in Ecological Communities. University of Chicago Press, Chicago, 13–25. Tofts, R., and J. Silvertown. 2002. Community assembly from the local species pool: an experimental study using congeneric pairs. Journal of Ecology 90: 385–393.

The Impact of Animals on Species Diversity

205

Tongway, D.J., and J.A. Ludwig. 1994. Small-scale resource heterogeneity in semiarid landscapes. Paciﬁc Conservation Biology 1: 201–208. Valone, T.J., J.H. Brown, and E.J. Heske. 1994. Interactions between rodents and ants in the Chihuahuan Desert: an update. Ecology 75: 252–255. Whitford, W.G. 1999. Contribution of pits dug by goannas (Varanus gouldii) to the dynamics of banded mulga landscapes in eastern Australia. Journal of Arid Environments 40: 453–457. Whitford, W.G., and R. Dimarco. 1995. Variability in soils and vegetation associated with harvester ant (Pogonomyrmex rugosus) nests on a Chihuahuan Desert watershed. Biology and Fertility of Soils 20: 169–173. Whitney, K.D., M.F. Fogiel, A.M. Lamperti, K.M. Holbrook, D.J. Stauﬀer, B.D. Hardesty, V.T. Parker, and T.B. Smith. 1998. Seed dispersal by Ceratogymna hornbills in the Dja Reserve, Cameroon. Journal of Tropical Ecology 14: 351–371. Wilby, A. 1996. Vegetation development on setaside arable land: the role of animals. PhD thesis, University of London. Wilby, A., and V.K. Brown. 2001. Herbivory, litter and soil disturbance as determinants of vegetation dynamics during early old-ﬁeld succession under set-aside. Oecologia 127: 259–265. Wilby, A., M. Shachak, and B. Boeken. 2001. Integration of ecosystem engineering and trophic eﬀects of herbivores. Oikos 92: 436–444. Yair, A., and A. Danin. 1980. Spatial variations in vegetation as related to the soil moisture regime over an arid limestone hillside, northern Negev, Israel. Oecologia 54: 389–397. Yair, A., and M. Shachak. 1982. A case study of energy, water and soil ﬂow chains in an arid ecosystem. Oecologia 54: 389–397. Yair, A., and M. Shachak. 1987. Studies in watershed ecology of an arid area. In: Berkofsky, L., and Wurtele, G. (eds.). Progress in Desert Research. Rowman and Littleﬁeld, Totowa, NJ, 145–193. Zaady, E., P.M. Groﬀman, and M. Shachak. 1996. Release and consumption of nitrogen by snail feces in Negev Desert soils. Biology and Fertility of Soils 23: 399–404. Zobel, M. 1997. The relative role of species pools in determining plant species richness: an alternative explanation of species coexistence. Trends in Ecology and Evolution 12: 266–269.

12 Resource Partitioning and Biodiversity in Fractal Environments with Applications to Dryland Communities Mark E. Ritchie Han Olff

A

rid and semiarid ecosystems (drylands) often contain a higher diversity of animals and plants than would be expected from their low productivity. High spatial heterogeneity of resources and physical habitats, exhibited at a wide range of spatial scales (Rundel 1996, Holling 1992, Peterson et al. 1998), may be a major factor explaining such high diversity. For example, at extremely small scales (<10 cm), branched plant material and various soil physical processes can create spatial niches for invertebrates, cyanobacteria, and other cryptogamic organisms (Lightfoot and Whitford 1991). At somewhat larger scales (<10 m), desert shrubs may aggregate water and organic material in ‘‘islands of fertility,’’ yielding a highly patchy heterogeneous distribution of resources (e.g., seeds, water) for other plants and animals (Gibbens and Beck 1988, Halvorson et al. 1997, chapter 13 this volume, chapter 11 this volume). At even larger scales (>100 m), soil erosion patterns create topographic variation that locally concentrates available water and nutrients, yielding a marked heterogeneity in the distribution of productivity across the landscape (Milne 1992). These heterogeneous distributions of physical environments, biotic material, and resources are likely to have strong eﬀects on biodiversity. Ecologists have long associated greater spatial heterogeneity with higher species diversity (MacArthur 1964; Brown 1981; May 1988). Within a particular physical environment (habitat), this association exists presumably because collections of species that use similar resources, or ‘‘guilds,’’ can coexist whenever they can more ﬁnely divide up space and diﬀerent-sized resource ‘‘packages’’ (Hutchinson and MacArthur 1959, Brown 1981, 1995, Morse et al. 1985, Peterson et al. 1998). The partitioning of space and diﬀerent resource patches 206

Resource Partitioning and Biodiversity in Fractal Environments

207

may be constrained by the diﬀerent body sizes of species within guilds (Hutchinson and MacArthur 1959, Morse et al. 1985, Belovsky 1986, 1997, Brown 1995, Siemann et al. 1996). However, the mechanism by which body size and spatial heterogeneity of habitats and resources determine species diversity remains unclear (May 1988, Brown 1995, Siemann et al. 1996, Belovsky 1997). Resource partitioning and spatial heterogeneity therefore may strongly inﬂuence diversity in drylands, where, for example, wellknown guilds of granivorous vertebrates and invertebrates are structured by competition for diﬀerent sizes of seeds and seed patches (Brown et al. 1979, Davidson et al. 1980, 1985).

A Scale-Dependent Model of Resource Partitioning and Species Diversity In this chapter, we employ a new approach based on fractal geometry to describe how species within guilds ﬁnd resources in space, and how limits to the similarity in body size between any two species predicts the potential number of species in a community (Ritchie and Olﬀ 1999). We then discuss how this model can help understand diversity patterns in drylands. Two key aspects of spatial heterogeneity are variation in suitable physical/chemical conditions (e.g., suitable soil texture and pH, thermal conditions), and the nonuniform spatial distribution of resources. Few general hypotheses exist to predict how such spatial variation aﬀects species’ persistence and coexistence with other species (Kareiva and Wennergren 1995). One tool that may help generate such hypotheses is fractal geometry. Fractal geometry oﬀers the advantage of being able to describe the amount of some material as a function of the spatial pattern of its distribution and scale of observation. This property makes fractal geometry ideal for incorporating spatial heterogeneity in analytical models of consumer–resource interactions, a traditional tool of community ecologists (Ritchie 1998, Ritchie and Olﬀ 1999). It has been used successfully to describe many spatially complex physical, geological, and biological distributions (Milne 1997, Kunin 1998, as well as many chapters in this volume) with simple power laws (scaling laws) of the form M ¼ aLD

ð1Þ

in which M is the ‘‘mass’’ or amount of material and L is a length of the area or volume being observed. D is the fractal dimension or degree to which additional material is encountered as the environment is sampled at increasing spatial scales around existing material. D varies from 0 (a single point) to 3 (a fully ﬁlled cube with sides of length L). The coeﬃcient a is the ‘‘lacunarity’’ (Mandelbrot 1983) and reﬂects the local density or aggregation of material in the immediate vicinity of other material. The scaling law implies that the distribution is statistically self-similar (or self-aﬃne) across ecologically relevant ranges of scales, typically three to four orders of magnitude (Milne

208

Ecological Complexes of Biodiversity, Ecosystems, and Landscapes

1997, Kunin 1998, this volume). Under the assumptions of self-similarity or self-aﬃnity, any distribution within a given amount of space of length L can potentially be described with three parameters: the proportion of space occupied f , the lacunarity a, and the fractal dimension D (ﬁg. 12.1). The proportion f constrains the potential values of a and D. The application of such scaling laws helps us understand species diversity by showing how spatial heterogeneity inﬂuences resource acquisition (Ritchie 1998). Individual organisms must search within a space containing complex distributions of suitable physical/chemical conditions (habitat), for resources contained within other material (food) that itself is complexly distributed (ﬁg. 12.2). Therefore, resources available to organisms are nested within food, and available food is nested within habitat. These complex distributions present habitat and food ‘‘patches’’ of diﬀerent size, and food patches vary in their concentration of resources. For example, rodents search beneath shrub cover (habitat) for seeds (food) that contain diﬀerent concentrations of protein (resources). Insect herbivores move through suitable microclimates on terrestrial plants (habitat) to eat plant tissue (food), which contains digestible carbohydrates (resources). Terrestrial plants ‘‘forage’’ within rock-free soil for nutrients (resources) contained in soil solution (food). Therefore, within the same habitat, diﬀerent species of similar trophic position may harvest the same resources by consuming diﬀerent sizes or types of food. This diﬀerential use of patches of habitat and food is therefore a potential mechanism of coexistence that contributes to the diversity of species limited by the same resource (Ritchie and Olﬀ 1999).

Figure 12.1 Hypothetical landscapes of extent x, showing the distribution of habitat (black) and its description with a spatial scaling law of the form H ¼ axD . The lacunarity coeﬃcient a describes the average local density or contagion of habitat in the immediate vicinity of other habitat. The fractal dimension D describes the rate at which organisms will encounter additional habitat as they sample larger spatial scales around existing habitat. The two parameters together can successfully describe a wide variety of habitat conﬁgurations for a given fraction of the map occupied by habitat (15%).

Resource Partitioning and Biodiversity in Fractal Environments

209

Figure 12.2 (a) Hypothetical space of extent x used by species of diﬀerent size, including habitat, the food in which resources are contained, and resources contained in food. Each element is fractal with the following mass fractal dimensions [17]: habitat: D ¼ 1:91; food: F ¼ 1:67; resource: Q ¼ 1:34. Larger species (b) exclusively use large patches with low resource concentration, while smaller species (c) exclusively use small patches with high resource concentration. Note the ‘‘microhabitat’’ separation in space of the two species’ exclusive resources. (From Ritchie and Olﬀ 1999.)

Complex distributions of habitat, food, and resources may be easily described with fractal geometry, i.e., spatial scaling laws. Within a landscape of extent (length) x, the total amount of habitat is hxD , where D is the fractal dimension of the habitat and h is the lacunarity coeﬃcient for habitat (Milne 1997). Likewise, food patches occupy a volume mxF and resources occupy a volume rxQ . The fractal dimensions D, F, and Q represent the rate at which additional habitat, food, and resources are detected by sampling increasing spatial scales. However, the assumption that habitat, food, and resources are nested distributions (Milne 1992) requires that D F Q. Greater lacunarity coeﬃcients h, m, and r reﬂect higher local density, aggregation and contagion of habitat, food, and resources, respectively (Mandelbrot 1983, Milne 1992, 1997, Ritchie and Olﬀ 1999).

210

Ecological Complexes of Biodiversity, Ecosystems, and Landscapes

In a fractal environment, body size critically determines the abundance of food and resources a species perceives (O’Neill et al. 1988, Ritchie 1998, Ritchie and Olﬀ 1999) (ﬁg. 12.2). Individuals can be assumed to instantaneously search a volume of space at a particular scale of resolution, i.e., the length, w, of the ‘‘ruler’’ with which they perceive or sample the environment. This scale of resolution is presumably proportional to body size. If so, a species will subdivide its habitat into subvolumes of size wD . For example, search volume will be a box of area w2 in an unfragmented habitat (D ¼ 2), but will be a volume less than w2 if the habitat is fragmented (D < 2). The search volume therefore reﬂects the fragmentation of the habitat in the entire landscape, only expressed at a smaller scale. Individuals will consume food in all subvolumes where the amount of food exceeds a particular minimum volume, or ‘‘patch’’ size, P. The total amount of food available is therefore the food contained in all subvolumes with more than P amount of food. Large aggregations of food are perceived as contiguous collections of these minimum-sized patches. Applying optimal foraging principles (box 12.1), we now show that larger species, with larger search volumes, require a greater amount of food, P, within their search volume. They will thus ignore small, scattered food patches and detect less total volume of food than smaller species. However, because they search a greater total volume per unit time, they can tolerate lower resource concentrations within each patch. Smaller species require a lower P, and thus more frequently encounter search volumes with food amounts exceeding P. They thus perceive a food distribution consisting of both individual small patches and large aggregations of small patches, but require higher resource concentrations within these patches (Belovsky 1986, 1997, Ritchie 1998, Ritchie and Olﬀ 1999).

Box 12.1 Determination of minimum patch size and resource concentration and limiting similarity in size of species. To persist under the widest range of conditions, individuals of a species with resolution w should simultaneously minimize patch size P and resource concentration R within the persistence conditions P ¼ LwDF =ðmkR) (eq. (3)). The choice of P can be expressed as a fraction fP of the maximum possible patch size, that is, a search volume ﬁlled with food: Pmax ¼ wD . Likewise, R can be expressed as a fraction fR of the maximum occurring resource concentration, that is, a particular volume of food ﬁlled with resource: Rmax ¼ wQF . Optimal P and R are found by simultaneously minimizing fP and fR subject to the condition fP fR ¼ LwDF =ðmkRmax Pmax ), using the standard method of Lagrangian multipliers. Substituting for Pmax and Rmax yields fP ¼ fR ¼ ðL=ðmkÞÞ1=2 wQ=2 and

Resource Partitioning and Biodiversity in Fractal Environments

P ¼ fP Pmax ¼ ðL=mkÞ1=2 wDQ=2 R ¼ fR Rmax ¼ ðL=mkÞ1=2 wQ=2F

211

ðB1Þ

The limit to how similar in size species can be, or the size ratio g of larger to smaller species of two adjacent-sized species, is determined by the rate of consumption of exclusive resources E relative to resource losses L : E L. For three species i, j, k, ranked in order of increasing size, that sample k search volumes per unit time, the rate of acquisition of exclusive resources E is ð Pk ð Ri E ¼k P f ðPÞ R f ðRÞ dP dR ðB2Þ Pj

Rj

where f ðPÞ and f ðRÞ are the frequency of occurrence of food patches of size P or larger and resource concentrations of R or higher, respectively. For any fractal image subdivided into subvolumes of a certain length (scale of resolution w), the fraction of all subvolumes of the image that are occupied is awD (Milne 1992, Hastings and Sugihara 1993), where a is the lacunarity and D is the fractal dimension of the image. It follows then that for food within habitat, a ¼ m=h and P is the subvolume sampled. For resources within food, a ¼ r=m and R is the lowest resource concentration at which a subvolume is considered to be occupied by resource. Therefore, f ðPÞ ¼ ðm=hÞ=P and f ðRÞ ¼ ðr=mÞ=R. Substituting for f ðPÞ and f ðRÞ, and evaluating the integral yields E ¼ kr ðPk Pj ÞðRi Rj Þ L

ðB3Þ

where r is the constant r=h. We can now ﬁnd the size ratio g by assuming that the size ratios for each of the two adjacent species pairs are equal (g ¼ wk =wj ¼ wj =wi Þ, and replacing wk with gwj and wi with ð1=gÞwj in so that Pk ¼ ðL=mkÞ1=2 ðgwj ÞDQ=2 and Ri ¼ ðL=mkÞ1=2 ðwj =gÞQ=2F . The ratio g likely will deviate from 1 by less than an order of magnitude, so gDQ=2 ﬃ gFQ=2 . With this assumption, we can substitute functions of g for Pk and Ri , and, using for Pj and Rj , we can solve approximately for g as a function of species’ size: n o1=ðDQ=2Þ gðwÞ ﬃ 1 þ ðm=rÞ1=2 wðFDÞ=2 ðB4Þ

To prove these assertions, suppose that individuals of a species search k sub-volumes of size wD in a time period dt. If resources are instantaneously replaced following consumption, the population growth rate of that species can be described as dN=dt ¼ qNðkBPR LÞ

ð2Þ

212

Ecological Complexes of Biodiversity, Ecosystems, and Landscapes

(Schoener 1973), in which N is population size, q converts resources into individuals, and L is resource loss rate per individual. Food is encountered within a search volume with probability B ¼ mwF =wD . The minimum patch size P and resource concentration R reﬂect which food patches the individuals of a species selects. Resources losses, L, reﬂect both metabolic rate and mortality, so, for simplicity we assume that L is size invariant. A species can persist if kBPR L, which requires that a minimum food patch size (P) and minimum resource concentration (R) are exceeded. Choosing a larger P allows a choice of lower R to meet resource losses, which can be seen by rearranging the persistence conditions to yield a tradeoﬀ relationship (Curve T in ﬁg. 12.3a), P ¼ LwDF =ðmkRÞ

ð3Þ

Equation (3) implies that individuals can choose P and R that will satisfy conditions for the species’ persistence. Because D > F, larger species must choose a greater P for a given resource concentration R. For a species to persist under the greatest range of conditions, P and R should be simultaneously minimized relative to their potential maximum values in the landscape: Pmax ¼ wD and Rmax ¼ wQF . This optimization results in unique thresholds P , R for a species of a given size and scale of resolution w (box 12.1). P ¼ ðL=mkÞ1=2 wDQ=2 R ¼ ðL=mkÞ1=2 wQ=2F

ð4Þ

By deﬁnition (ﬁg. 12.2), D F Q, so P scales positively with size, while R scales negatively (ﬁg. 12.3b). Applying these scaling laws to a group of species using similar resources, we ﬁnd a ‘‘packing’’ rule for how close in size species can be, that is, the size ratio g between species of adjacent size. For a set of species i; j; k, ranked by increasing species’ size, the P and R of all three species deﬁne an exclusive niche for species j (ﬁg. 12.2b,c, Fig. 12.3c), that is, both Pi < Pj < Pk and Rk < Rj < Ri . For species j, there is a unique set of patches, ranging from size Pk to Pj and resource concentration Ri to Rj , that are both too small for species k and too low in resource concentration for species i (ﬁg. 12.3c). A species can persist on just these exclusive resources, regardless of their competitive dynamics with other species, if its consumption rate of exclusive resources exceeds resource losses. The conditions under which this occurs (box 12.1) yields an approximate solution for the size ratio between two adjacent-sized species in a community: n o1=ðDQ=2Þ gðwÞ ﬃ 1 þ m1=2 wðFDÞ=2 ð5Þ The size ratio gðwÞ should decline with increasing organism size (ﬁg. 12.4a) because D F Q, and because small resource-rich patches needed by

Resource Partitioning and Biodiversity in Fractal Environments

213

Figure 12.3 (a) Minimum threshold patch size P and resource concentration R for a species derived by minimizing P and R simultaneously but still meeting resource requirements (constraint T). (b) Power law relationships for minimum food patch size P and resource concentration R as a function of a species’ size (eq (4)). (c) Thresholds (R , P ) for three species i; j; k that deﬁne the exclusive niche for species j (shaded area). Thresholds for species i and k will be positioned so that Pj < Pk for every Rj > Ri and to ensure that the consumption rate of exclusive resources kr(Pk Pj ÞðRi Rj ) is greater than resource loss rate. (From Ritchie and Olﬀ (1999).)

smaller species occupy proportionately less total volume than larger, resource-poor patches used by larger species. The functional equation for size ratios (eq. (5)) dictates the number of species ranging in size from wmin to wmax that can be ‘‘packed’’ into an environment. The maximum size (wmax ) is determined by whether there is at least one suitable patch of size P and resource concentration R in a ﬁnite space of extent x. The number of suitable patches is found by dividing the total volume of resources rxQ by the resource volume contained within suitable patches (P R ). Since P and R depend on w, however, the actual number of patches in the ﬁnite space is weighted by the probability of the

214

Ecological Complexes of Biodiversity, Ecosystems, and Landscapes

Figure 12.4 (a) Predicted relationship between the size ratio (larger/smaller) of pairs of adjacent-sized species and ln(size) of the larger species of each pair. Observed relationships for (b) granivorous ants and (c) granivorous vertebrates (birds and mammals) at the Cave Creek Bajada site in the Chihuahuan desert of Arizona, and (d) vascular plants at Deseret Land and Livestock in the Great Basin. Curves in (b)–(d) are ﬁts (showing R2 ) of the data to the nonlinear function ð1 þ b0 wb1 Þb2 using least-squares nonlinear regression.

occurrence of a patch with the length w speciﬁed by P and R . This probability is (F 1)wF (Milne 1992, Ritchie 1998). Therefore, (F 1ÞwF rxQ / (P R Þ ¼ 1. Substituting for P and R (eq. (1)) and solving for w yields 1=D wmax ¼ ðF 1ÞxQ krm=L

ð6Þ

Likewise a minimum size is set by whether at least one patch of suﬃcient resource concentration occurs. This is found by dividing the expected resource concentration within food rxQF by the resource concentration contained within suitable patches (R ). Since R scales with w, the actual number of patches in the ﬁnite space of extent x must be weighted by the probability of the occurrence of resources within a patch of length w. This probability is (Q 1)wQ (Milne 1992, Ritchie 1998). Therefore, (Q 1ÞwQ rxQF = ðR Þ ¼ 1. Substituting for R (eq. (4)) and solving for w yields 1=ð3Q=2DÞ wmin ¼ ðQ 1ÞxQF krm=L

ð7Þ

Resource Partitioning and Biodiversity in Fractal Environments

215

Equations (6) and (7) suggest that landscapes with larger extents support both larger and smaller species because this increases the probability of patches that are suﬃciently large or resource-rich. This prediction matches patterns observed for mammals, birds, and ﬁsh on islands of diﬀerent size, including mammals on Sonoran desert islands in the Gulf of California (Marquet and Taper 1998). Species richness (S) is then the number of exclusive size niches allowed between wmin and wmax , which yields the solution S Y i¼1

gðwi Þ ¼

wmax wmin

ð8Þ

An approximate solution for S is S ﬃ lnðwmax Þ=½lnðgðwmax ÞÞ lnðwmin Þ

ð9Þ

The equation for S predicts a left-skewed, unimodal distribution of species richness versus organism size (ﬁg. 12.5a). This distribution reﬂects the larger size ratios and thus looser species packing required for smaller species (eq. (5))

Figure 12.5 Left-skewed frequency distribution of species in diﬀerent body size classes (a) predicted by the scaling model, and observed patterns for (b) Chihuahuan desert vertebrate granivores, and (c) Great Basin vascular plants. There were insuﬃcient species to construct a reliable size distribution for ants.

216

Ecological Complexes of Biodiversity, Ecosystems, and Landscapes

and the limitation of the largest species by the maximum patch size in the environment and the smallest species by the maximum resource concentration. The functions implicit in wmax (eq. (3)) and gðwmax ) mean that the model also predicts the eﬀects on species richness of sampling area (x2 ), habitat fragmentation (D), and the amount and distribution of food and resources (m; r; F; Q). However, we will not explore these predictions here.

Community Structure and Diversity in Dry Environments Do these theoretical predictions apply to species assemblages in drylands? We tested the model predictions for three guilds of species that use similar resources in North American deserts. These included two guilds, granivorous ants and vertebrates, at Cave Creek Bajada in the Chihuahuan desert (Brown et al. 1979, Davidson et al. 1985) and vascular plants at Deseret Ranch in the Great Basin desert (Ritchie et al. 1994). We tested for both declining size ratios with increasing body size and right-skewed distributions of diversity versus body size. We used body length as a measure of size for ants and vertebrates, which all consume seeds of various annual and perennial plants at the soil surface. We measured mean crown width of 10 randomly selected individual plants of each of 20 common species growing in a sagebrush–rabbitbrush–western wheatgrass community (Artemisia tridentata, Chrysothamnus viscidiﬂorus, Pascopyrum smithii). We then examined size patterns for species growing within a single 6.25 m2 plot. Instead of being constant (Hutchinson 1959, Brown 1981), size ratios in these very diﬀerent assemblages declined signiﬁcantly with increasing size (ﬁg. 12.4b,c), and the relationships ﬁt the shape predicted by the model (eq. (5), ﬁg. 12.4a). These patterns are consistent with the hypothesis that these granivore communities are structured by resource partitioning to avoid competition, as has been shown in numerous ﬁeld experiments (Brown et al. 1979, Davidson et al. 1980, 1985). The size diﬀerences among species may lead them to use diﬀerent-sized seeds or seed patches (Brown et al. 1979), thus enabling coexistence despite considerable diet overlap. The species richness–size distributions of both the vertebrate granivores in the Chihuahuan desert and the plants from the Great Basin desert are both signiﬁcantly left-skewed (ﬁg. 12.5b,c). At small scales, close species packing of similar-sized species may help explain the clusters of similar-sized birds and heteromyid rodent species and clusters of harvester ant species that occur in the Chihuahuan desert (Brown et al. 1979, Davidson et al. 1980, 1985). These distributions diﬀer from the log-normal or right-skewed distributions most commonly reported (Hutchinson and MacArthur 1959, Morse et al. 1985, May 1988, Brown et al. 1993, Siemann et al. 1996). Virtually all these previously published distributions combine diversity–size distributions of separate guilds (e.g., nectarivores, granivores, herbivores, carnivores) or species that use similar resources but diﬀerent habitats. Our model may not apply to such communities that include species that use diﬀerent resources or

Resource Partitioning and Biodiversity in Fractal Environments

217

habitats. In addition, communities assembled through diﬀerent mechanisms (e.g., colonization limitation) may show diﬀerent patterns (Hubbell 1997). This discrepancy in patterns suggests that diversity–size patterns may be scale-dependent (Brown and Nicoletto 1991). Patterns at smaller scales may be driven by competition, predation, and resource partitioning, while patterns at larger spatial extents may be driven by geographical isolation and speciation coupled with colonization limitation. The application of spatial scaling laws suggests that a theory of biodiversity may emerge from ﬁrst principles of how organisms ﬁnd resources in space. The analysis formalizes earlier ideas that diversity depends on the number of spatial niches (Hutchinson 1959, MacArthur 1964, Brown 1981, May 1988) and suggests that coexisting species cannot inﬁnitely partition space (Rosenzweig and Abramsky 1994, Rosenzweig 1995). In addition, the model synthesizes recent ideas about how resource acquisition (Wright 1983, Belovsky 1986, 1997, Brown et al. 1993) and spatial characteristics of habitat (MacArthur 1964, Palmer 1992) inﬂuence diversity. Clearly, other factors, including diversity of resource types (Tilman 1982, Jones and Lawton 1991, Huisman and Weissing 1999), disturbance (Huston 1994), colonization limitation (Tilman 1997, Hubbell 1997), and biogeographical history (Schluter and Ricklefs 1993, Rosenzweig 1995, Hubbell 1997) are also important in explaining diversity patterns. Nevertheless, the spatial scaling of resource use by species of diﬀerent body size may explain many species diversity patterns across a range of spatial scales and taxa. Drylands, which have visible and easily measurable heterogeneity in physical habitats and resources at a variety of scales, may be especially good for applying this scaling approach. The well-studied communities of ants and granivorous mammals and birds in southern Arizona ﬁt the model’s prediction well. Plants from a Great Basin desert site also support the model’s predictions. These results suggest that diﬀerentiation in size (scale of resolution) may be a major mechanism structuring local desert communities. However, the discrepancy between left-skewed diversity-size distributions at local scales of observation and right-skewed or log-normal distributions at larger scales (Brown and Nicoletto 1991), suggests that other mechanisms may be more important in structuring communities assessed at larger scales of observation. Likely mechanisms at larger scales include geographic isolation and speciation coupled with colonization limitation (Ricklefs and Schluter 1994, Tilman 1997). The scale-dependent approach we introduce makes these scale distinctions explicit and suggests new hypotheses for assessing how species diversity patterns change across spatial scales.

Acknowledgments We thank B.T. Milne, J.H. Brown, H. DeKroon, J.M. Emlen, S.A. Gripne, A. Guss, N. Haddad, W.A. Kunin, L. Li, S. Naeem, and W.C. Pitt, for comments. The U.S. National Science Foundation, Santa Fe Institute, Netherlands NWO, USU Ecology Center, and the Utah Agricultural Experiment Station supported this work.

218

Ecological Complexes of Biodiversity, Ecosystems, and Landscapes

References Belovsky, G.E. 1986. Generalist herbivore foraging and its role in competitive interactions. Am. Zool. 26, 51–69. Belovsky, G.E. 1997. Optimal foraging and community structure: the allometry of herbivore food selection and competition. Evol. Ecol. 11, 641–672. Brown, J.H. 1981. Two decades of homage to Santa Rosalia: toward a general theory of diversity. Am. Zool. 21, 877–888. Brown J.H. 1995. Macroecology. University of Chicago Press, Chicago. Brown, J.H., and P.F. Nicoletto. 1991. Spatial scaling of species composition: body masses of North American land mammals. Am. Nat. 138, 1478–1512. Brown, J.H., D.W. Davidson, and O.J. Reichman. 1979. An experimental study of competition between seed-eating desert rodents and ants. Am. Zool. 19, 1129–1143. Brown, J.H., P. Marquet, and M. Taper. 1993. Evolution of body size: consequences of an energetic deﬁnition of ﬁtness. Am. Nat. 142, 573–584. Davidson, D.W., J.H. Brown, and R.S. Inouye. 1980. Competition and the structure of granivore communities. BioScience 30, 233–238. Davidson, D.W., D.A. Samson, and R.S. Inouye. 1985. Granivory in the Chihuahuan desert: interactions within and between trophic levels. Ecology 66, 486–502. Gibbens, R.P., and R.F. Beck. 1988. Changes in grass basal area and forb densities over a 64-year period on grassland types of the Jornada Experimental Range. J. Range Manage. 41, 186–192. Halvorson, J.J., H. Bolton, and J.L. Smith. 1997. The pattern of soil variables related to Artemisia tridentata in a burned shrub-steppe site. Soil Sci Soc. Am. J. 61, 287–294. Holling, C.S. 1992. Cross-scale morphology, geometry, and dynamics of ecosystems. Ecol. Monogr. 62, 447–502. Hubbell, S.P. 1997. A uniﬁed theory of biogeography and relative species abundance and its application to tropical rainforests and coral reefs. Coral Reefs 16, S9-S21. Huisman, J., and F.J. Weissing. 1999. Biodiversity of plankton by species oscillations and chaos. Nature 402, 407–410. Huston, M.A. 1994. Biological Diversity: the Coexistence of Species on Changing Landscapes. Cambridge University Press, Cambridge. Hutchinson, G.E. 1959. Homage to Santa Rosalia, or why are there so many kinds of animals? Am. Nat. 93, 145–159. Hutchinson, G.E., and R.H. MacArthur. 1959. A theoretical model of size distributions among species of animals. Am. Nat. 93, 117–125. Jones, C.G., and J.H. Lawton. 1991. Plant chemistry and insect species richness of British umbellifers. J. Anim. Ecol. 60, 767–777. Kareiva, P., and U. Wennergren. 1995. Connecting landscape patterns to ecosystem and population processes. Nature 373, 299–302. Kunin, W.E. 1998. Extrapolating species abundance across spatial scales. Science 281, 1513–1515. Lightfoot, D.C., and W.G. Whitford. 1991. Productivity of creosotebush forage and associated canopy arthropods along a desert roadside. Am. Midl. Nat. 125, 310–322. MacArthur, R.H. 1964. Environmental factors aﬀecting bird species diversity. Am. Nat. 98, 387–397. Mandelbrot, B.B. 1983. The Fractal Geometry of Nature. Freeman, New York.

Resource Partitioning and Biodiversity in Fractal Environments

219

Marquet, P.A., and M.L. Taper. 1998. On size and area: patterns of mammalian body size extremes across landmasses. Evol. Ecol. 12, 127–139. May, R.M. 1988. How many species are there on earth? Science 241, 1441–1449. Milne, B.T. 1992. Spatial aggregation and neutral models in fractal landscapes. Am. Nat. 139, 32–57. Milne, B.T. 1997. Applications of fractal geometry in wildlife biology. In Wildlife and Landscape Ecology: Eﬀects of Pattern and Scale (ed. Bissonette, J.). Springer, New York, pp. 32–69. Morse, D., N.E. Stork, and J.H. Lawton. 1985. Fractal dimension of vegetation and the distribution of arthropod body lengths. Nature 314, 731–732. O’Neill, R.V., B.T. Milne, M.G. Turner, and R.H Gardner. 1988. Resource utilization scales and landscape pattern. Land. Ecol. 2, 63–69. Palmer, M.W. 1992. The coexistence of species in fractal landscapes. Am. Nat. 139, 375–397. Peterson, G., C.R. Allen, and C.S. Holling. 1998. Ecological resilience, biodiversity, and scale. Ecosystems 1, 6–18. Ritchie, M.E. 1998. Scale-dependent foraging and patch choice in fractal environments. Evol. Ecol. 12, 309–330. Ritchie, M.E., and H. Olﬀ. 1999. Spatial scaling laws yield a synthetic theory of diversity. Nature 400, 557–560. Ritchie, M.E., M.L. Wolfe, and R. Danvir. 1994. Predation of artiﬁcial sage grouse nests in treated and untreated sagebrush. Great Basin Nat. 54, 122–129. Rosenzweig, M.L. 1995. Species Diversity in Space and Time. Cambridge University Press, Cambridge. Rosenzweig, M.L., and Z. Abramsky. 1994. How are diversity and productivity related? In Species Diversity in Ecological Communities: Historical and Geographical Perspectives (ed. Schluter, D., and R. Ricklefs). University of Chicago Press, Chicago, pp. 52–65. Rundel, P.W. 1996. Ecological Communities and Processes in a Mojave Desert Ecosystem. Cambridge University Press, Cambridge. Schluter, D., and R. Ricklefs. 1993. Species diversity: regional and historical inﬂuences. In Species Diversity in Ecological Communities: Historical and Geographical Perspectivs (eds. Ricklefts, R., and D. Schluter). University of Chicago Press, Chicago, pp. 350–364. Schoener, T.W. 1973. Population growth regulated by intraspeciﬁc competition for energy or time: some simple representations. Theor. Pop. Biol. 6, 265–307. Siemann, E., D. Tilman, and J. Haarstad. 1996. Insect species diversity, abundance and body size relationships. Nature 380, 704–706. Tilman, D. 1982. Resource Competition and Community Structure. Princeton University Press, Princeton, NJ. Tilman, D. 1997. Community invasibility, recruitment limitation, and grassland biodiversity. Ecology 78, 81–92. Wright, D.H. 1983. Species-energy theory: an extension of species-area theory. Oikos 41, 496–506.

13 Uniﬁed Framework II Ecosystem Processes: A Link Between Species and Landscape Diversity Moshe Shachak Robert Waide Peter M. Groffman

T

he discipline of ecology can be subdivided into several subdisciplines, including community, ecosystem, and landscape ecology. While all the subdisciplines are important to the study of biodiversity, there is great variation in the extent to which their contributions have been analyzed. For example, the role of community ecology in biodiversity studies is well established. In community ecology, the entities of study are species that diﬀer in their properties and generate a web of interactions that, in turn, organize the species into a community. Similar to community ecology, the contribution of landscape ecology to biodiversity is apparent. The entities of study, deﬁnable ‘‘patches,’’ are tangible. They diﬀer in their properties and generate a web of interactions that organize the patches into a landscape mosaic. In contrast to community and landscape ecology, the role of ecosystem ecology in biodiversity is less apparent. In ecosystem ecology, it often is not clear what the entities are, and how they are organized. To the extent that ecosystem ecology focuses on energy ﬂow and nutrient cycling, we can deﬁne fundamental entities as compartments and vectors in models that depict the ﬂows of water, energy, and nutrients through communities. If we apply diversity criteria to these entities, we can use the term ecosystem diversity to refer to the number of compartments and vectors, the diﬀerences among them in type and size, and their organization in promoting energy ﬂow or nutrient cycling. To our knowledge, ecosystem scientists have not yet developed criteria for ecosystem diversity similar to those used for species and landscape diversity. There has been some use of the term ‘‘ecosystem diversity’’ to refer to a diversity of ecosystems, implying a variety of habitats, landscapes, or biomes. 220

Unified Framework II

221

As discussed above, we suggest that to deﬁne the role of ecosystem ecology in biodiversity studies, the approach should be to study the relationships among species, landscape, and ecosystem diversities (chapters 1 and 13). However, since the concept of ecosystem diversity awaits further development, we adopt a diﬀerent approach for understanding the role of ecosystem science in biodiversity studies. In this chapter, we examine relationships among ecosystem processes, species diversity, and landscape diversity. Our thesis is that ecosystem processes related to the ﬂuxes of energy, water, soil, and nutrients are important links between species and landscape diversity and are fundamental to the development and maintenance of biodiversity. We start with an example of the web of interactions among ecosystem processes, organisms, and landscape mosaic that control the development of landscape and species diversity in the Negev. We then discuss in a more general way the various ways that ecosystem processes link species and landscape diversity.

Ecosystem Processes and Biodiversity in the Negev To illustrate the relationship between ecosystem processes and biodiversity in drylands, we discuss the role of ecosystem processes in the development of species and landscape diversity in the Negev. Figure 13.1 shows how ecosystem processes related to the ﬂuxes of energy, water, soil, and nutrients are important links between species and landscape diversity and are fundamental to the development and maintenance of biodiversity in the Negev. The ﬁrst stage in the development of ecological systems in the Negev is the formation of a rocky watershed by geological processes. At this stage, landscape diversity (one patch type rocks) and species diversity (only a few species of algae and lichens) are low. This is because the only substrate available for colonization by organisms is rocks and the main resource for sustaining species diversity, water, leaks out of the system as surface run-oﬀ (ﬁg. 13.1 (II)). Over time, species diversity increases as a result of a set of ecosystem processes that increase landscape diversity, that is, the formation of soil patches in the rocky watershed. This process of landscape diversiﬁcation is controlled by the ecosystem processes of dust deposition, the redistribution of water among landscape patches and the growth of cyanobacteria (Eldridge et al. 2000). The formation of soil patches depends on dust deposition and accumulation (Oﬀer et al. 1997, Shachak and Lovett 1998). Soil accumulation in the Negev is often slow, because the rate of erosion by run-oﬀ water is high on bare bedrock. The process of soil accumulation is greatly facilitated by the colonization of cyanobacteria on the soil surface. The cyanobacteria, by secretion of polysaccharides, form a soil crust, which is highly resistant to erosion (Yair and Shachak 1987). As dust accumulates, soil forms and the landscape becomes composed of two patch types, rocks and soils that exist in

222

Ecological Complexes of Biodiversity, Ecosystems, and Landscapes

Figure 13.1 Species diversity, landscape diversity, and ecosystem processes. (I) Ecosystem processes as a link between species and landscape diversity. (II)–(IV) Changes in trajectories, through time, of the relationships among ecosystem processes, species, and landscape diversity in the Negev desert. (See text for explanation.)

a source–sink relationship. The rocky patches (source) generate run-oﬀ and the soil patches (sink) absorb the run-oﬀ water. This new landscape mosaic opens new opportunities for colonization and establishment of higher plants. High numbers of annual plant species colonize and grow on the soil patches, increasing species diversity. These links between landscape diversity, ecosystem processes, and species diversity are shown in ﬁg. 13.1(III). Over time, the process of changing landscape and organism diversity continues as shrubs colonize the soil patches. Shrubs increase the interception of surface run-oﬀ and the accumulation of soil materials in these patches, resulting in the formation of soil mounds (Shachak et al. 1998). As in the case of the formation of soil patches on the rocky substrate, the formation of an additional patch type (soil mounds) by ecosystem processes increases the complexity of source–sink interactions in the landscape. The crusted soil, due to its low inﬁltration capacity, is a source for run-oﬀ

Unified Framework II

223

water, and associated sediments and nutrients. The soil mound is a sink for water, sediments, and nutrients. The consequence of landscape diversiﬁcation by soil mound formation is a signiﬁcant increase in plant species diversity (ﬁgs. 13.1(I–IV)). This example shows how an organism (shrub) can inﬂuence landscape diversity (forming a soil mound) aﬀecting the ecosystem processes (water, nutrient availability) that control species diversity. Over time, landscape diversiﬁcation in the Negev continues with the formation of mounds and pits by animals such as porcupines, isopods, and ants, further increasing landscape diversity. This leads to further increases in species diversity (Boeken and Shachak 1994). Our example from the Negev shows how the development of biodiversity is a complex process, involving interactions between two types of entities: organisms and patches. Interactions between the two types of entities are controlled by ecosystem processes. The organisms (cyanobacteria, shrubs, and animals) control the number of patch types and their distribution in space and time. The landscape mosaic controls the ecosystem processes of redistribution of water, soil, and nutrients. The resulting increase in the heterogeneity of resource distribution promotes high species diversity.

Ecosystem Processes and Biodiversity We suggest that there is a set of basic interactions that link species with landscape diversity through ecosystem processes. The link shown in the Negev example occurs when the landscape mosaic aﬀects the ﬂow of resources among patches. This ﬂow creates heterogeneity in the distribution of resources and allows for coexistence of a diversity of species. The ﬁrst link is landscape–ecosystem–species. The Negev example becomes more complex via two additional linkages involving species eﬀects on ecosystem and landscape processes. The second link (species–ecosystem–species) is driven by the eﬀects of speciﬁc species on water, energy, and material ﬂuxes in the ecosystem that lead to increases in species diversity. The third link (species– landscape–ecosystem–species) is driven by the ability of speciﬁc species to alter landscape diversity (e.g., ecological engineers), which alters ecosystem processes and ultimately species diversity. Below we elaborate on these links, giving examples from ecosystems other than the Negev, to emphasize the role of ecosystem processes in the production and maintenance of biodiversity. Landscape–Ecosystem–Species The landscape–ecosystem–species link encompasses the redistribution of resources in space by the landscape mosaic and its eﬀect on species diversity. It is rare that resources will remain in a single landscape patch. Rather, ecosystem processes move water, soil, organic matter, and nutrients across the landscape from one patch to another. These processes control the distribution of resources in space and thus are fundamental controllers of species

224

Ecological Complexes of Biodiversity, Ecosystems, and Landscapes

diversity. The creation of landscape and species diversity by ecosystem processes can be characterized by a web of source–sink interactions among landscape patches. Materials and propagules move from one patch and are then absorbed by another. This movement is controlled by patch properties and the vectors (e.g., water, wind, animals) that transport materials across the landscape. While the landscape–ecosystem–species link is obvious in drylands like the Negev, it can also be seen in other ecosystems. Didier and Poudevigne (2000) identiﬁed three community types along a landscape gradient in northwestern France. The species diversity of the communities was correlated with physical patchiness of the landscape: calcicolous communities on chalk slopes, mesophilous communities on colluvium, and hydrophilous communities on alluviums. These diﬀerences in ecosystem parent material inﬂuenced ecosystem processes related to water and nutrient availability, which ultimately controlled species diversity. In forests, landscape diversity is increased by the structural complexity formed by trees. This complexity arises from the presence of trees of varying sizes, conditions, and species, as well as standing dead trees, logs, and woody debris on the forest ﬂoor, multiple canopy levels, and canopy gaps. This complexity inﬂuences the pattern and rate of ecosystem processes related to nutrient and water availability, which increases the diversity of habitats for a wide range of organisms (Baskent and Glenwood 1996). Wardell and Horwitz (1996) found that the southwestern Australia landscape harbors pockets of habitats that show that ﬁne-scale hydrological patterns, persisting at local level patchiness, support high species diversity. The web of interactions among landscape, ecosystem, and species diversities is illustrated by the long-term research on Blackhawk Island in Wisconsin, USA. Varied glacial deposits on this island created landscape diversity through the production of soil types of diﬀerent texture (Pastor et al. 1982). Driven primarily by the eﬀects of soil texture on water availability, diﬀerent vegetation assemblages developed on the varied soil types (Pastor and Post 1986). The vegetation diﬀerences created variation in ecosystem properties and processes related to soil organic matter quality and nitrogen availability (Pastor et al. 1982, McClaugherty et al. 1985). The diﬀerences in nitrogen availability function as a positive feedback, amplifying the vegetation and ecosystem diﬀerences. Blackhawk Island is an example of how landscape diversity aﬀects ecosystem processes and organism diversity and how organisms and ecosystem processes feedback to amplify landscape diversity. In a more general sense, much basic ecosystem and landscape ecology research addresses how spatial heterogeneity at the landscape level aﬀects ecosystem processes (Forman and Godron 1986, Turner 1989, Huston 1994, Pickett and Cadenasso 1995), but links between landscape diversity, ecosystem processes, and species diversity have rarely been explicitly addressed. The eﬀects of this heterogeneity can be summarized in a ‘‘state factor’’ model of ecosystems where ecosystem processes are viewed as a function of climate, organisms, topography, parent material, time and humans

Unified Framework II

225

(Amundson and Jenny 1997). The climate, topography, and parent material state factors have been useful for conceptualizing the landscape–ecosystem– species trajectory (Pastor et al. 1982, Schimel et al. 1985, Groﬀman and Tiedje 1989, Zak et al. 1991). The evolution of landscapes is well studied in geomorphology (Huggett 1975, Hole 1978) and there have been eﬀorts to develop quantitative indices of landscape and soil diversity (Ibanez et al. 1995). However, there has been little analysis of relationships between landscape and soil diversity and organism and ecosystem diversity. Given the strong controls of landscape on ecosystem processes, these relationships are likely quite important, and should be a focus of future research. Species–Ecosystem–Species Analysis of species eﬀects on ecosystem processes has been a major theme of ecological research over the last 10 years or so. Species inﬂuence on organic matter quality, nutrient and water use, and physical structure have been studied in numerous ecosystems (Jones and Lawton 1995). The ‘‘species– ecosystem–species’’ link relates to the ecosystem processes controlled by organisms and their eﬀect on species diversity. In the state factor model of ecosystems and soil development described above (Amundson and Jenny 1997), the ‘‘organism’’ state factor represents the eﬀects organisms have on fundamental soil and ecosystem properties through the accumulation of organic matter, stimulation of chemical and physical weathering, and alteration of water and air ﬂow (Groﬀman and Bohlen 1999). There has been a great recent increase in interest in species–ecosystem links. This interest has been motivated by concern that human-induced loss of species diversity will inﬂuence important ecosystem processes related to nutrient retention, decomposition and production, and more fundamentally, to the stability of these processes, that is, their ability to persist through time in the face of environmental change. Species control ecosystem processes because they diﬀer in the rates and pathways by which they process resources. Thus, changes in species composition are likely to alter ecosystem processes through changes in the functional traits of biota. Traits with profound eﬀects are those that modify the availability, capture, and use of soil resources such as water and nutrients (Vitousek 1990, Chapin et al. 1995, Lawton and Jones 1995). For example, litter production by plants aﬀects soil temperature and moisture (Scholes and Walker 1993, Hobbie 1995), which aﬀects nutrient mineralization. Species inﬂuence the stability (resistance and resilience) of ecosystem processes, via diﬀerential environmental sensitivity among functionally similar species. The more functionally similar species there are in a community—that is, the greater the diversity within a functional group—the greater is its resilience in responding to environmental change (McNaughton 1977, Chapin

226

Ecological Complexes of Biodiversity, Ecosystems, and Landscapes

and Shaver 1985, Walker 1992, 1995, Lawton and Brown 1993). For example, the presence of drought-tolerant species allowed diverse grasslands to maintain higher productivity in response to drought than grasslands whose diversity had been reduced by experimental addition of nutrients (Tilman and Downing 1994). Although species–ecosystem links have been well studied, the feedbacks where alteration of ecosystem processes by speciﬁc species inﬂuence species diversity, that is, the full species–ecosystem–species link, have not been well developed (Chapin et al. 1997). For example, while Tilman and Downing (1994) show that species diversity can inﬂuence productivity, we know little about the eﬀects of productivity on diversity (Tilman et al. 1996). Given that ecosystem processes fundamentally control the availability of critical resources, the full species–ecosystem–species link is clearly worthy of further study. Species–Landscape–Ecosystem–Species In the former links we show how landscape diversity and species diversity inﬂuence ecosystem processes, and how these processes feedback to inﬂuence species diversity. However, there is a more complex set of interactions that is initiated by organism activity that aﬀects both landscape and species diversity. Organisms can aﬀect species diversity by changing the landscape mosaic, which in turn changes the distribution of resources. The concept of organisms as ‘‘ecosystem engineers’’ (Jones et al. 1994) suggests that species can control ecosystem processes that underlie resource distribution and availability by modiﬁcation and creation of patches in the landscape. The species–landscape–ecosystem–species trajectory is frequently driven by patch formation by plants and burrowing animals. Many studies have shown how vegetation pattern strongly inﬂuences water redistribution, biomass production, and nutrient dynamics (Wilson and Agnew 1992, Bach 1988a,b, Maubon et al. 1995, McCoy and Bell 1991, Doak et al. 1992, Roland 1993, Kareiva 1985, Turner 1989, Shachak et al. 1998). As described above in our example for the Negev, the diversity of vegetation patterns is particularly pronounced in drylands due to water scarcity, competition, and herbivory. Desert shrubs are examples of plants as ecosystem engineers. Shrublands can be viewed as landscape mosaics composed of shrub and intershrub patches. In a natural state the arrangement of the landscape mosaic is such that resources generated by the intershrub patch are intercepted by and accumulate in the shrub patches. The intershrub patches are sources for soil materials and run-oﬀ water (Abrahams et al. 1994, 1995, Eldridge 1993, Snow and McClelland 1990, Stockton and Gillette 1990, Rostango 1989, Yair and Shachak 1987). Shrub patches are sinks and therefore the principal loci of plant species diversity. This is mainly because of the accumulation of soil, water, and nutrients, which promote the growth of rich herbaceous vegetation under the shrubs (Allen 1991, Boeken and Shachak 1994, Garner and Steinberger 1989, Noy-Meir 1985, Schlesinger

Unified Framework II

227

et al. 1990, Shachak and Lovett 1998, Weinstein 1975, West 1989, Zaady et al. 1996). Under natural conditions the source–sink relationship between the two patch types is critical to the functioning of the system (Ludwig and Tongway 1995). Burrowing animals in deserts, such as porcupines (Alkon 1999), isopods (Shachak and Yair 1984), and ants, are important as links among species diversity, ecosystem processes, and landscape diversity. They modify landscape diversity by adding new patch types, such as pits and mounds. The new patches are capable of accumulating water and nutrients and thus enhance the increase in annual plant diversity. The burrowing animals act as ecosystem engineers which aﬀect biodiversity (Wilby et al. 2001).

Summary This chapter attempts to integrate concepts of ecosystem science with the study of biodiversity. We propose a framework that shows how ecosystem processes link landscape and species diversities and how the development of biodiversity is inexorably intertwined with these processes. Our example from the Negev and other systems demonstrates that these linkages are complex. There are not clear, unidirectional relationships among landscape diversity, ecosystem processes, and species diversity. However, there is a ﬁnite suite of linkages among these elements that can be applied to many diﬀerent ecosystems. We suggest that this framework and set of linkages can be of great value in assessment and management of biodiversity in many areas. If we recognize the inherent linkages between ecosystem processes and species and landscape diversity, we will have a better understanding of the production and maintenance of biodiversity. This understanding will be of great value as we attempt to maintain biodiversity in the face of global climate and land-use change and the demand for a diverse set of ecosystem services by an expanding human population.

References Abrahams, A.D., A.J. Parsons, and J. Wainwright. 1994. Resistance to overland ﬂow on semiarid grassland and shrubland hillslopes, Walnut Gulch, southern Arizona. Journal of Hydrology 156: 431–446. Abrahams, A.D., A.J. Parsons, and J. Wainwright. 1995. Eﬀects of vegetation change on interrill run-oﬀ and erosion, Walnut Gulch, southern Arizona. Geomorphology 13: 37–48. Alkon, P.U. 1999. Microhabitat to landscape impacts: crested porcupine digs in the Negev Desert highlands. Journal of Arid Environments 41: 183–202. Allen, E.B. 1991. Temporal and spatial organization of desert plant communities. Pp. 295–332 in J. Skujins, ed., Semiarid Lands and Desert: Soil Resource and Reclamation. Dekker, New York.

228

Ecological Complexes of Biodiversity, Ecosystems, and Landscapes

Amundson, R., and H. Jenny. 1997. On a state factor model of ecosystems. BioScience 47: 536–543. Bach, C.B. 1988a. Eﬀects of host plant patch size on herbivore density: patterns. Ecology 69: 1090–1102. Bach, C.B. 1988b. Eﬀects of host plant patch size on herbivore density: underlying mechanisms. Ecology 69: 1103–1117. Baskent, E.Z., and J. Glenwood. 1996. Designing forest management to control spatial structure of landscapes. Landscape and Urban-Planning 34: 55–74. Boeken, B., and M. Shachak. 1994. Desert plant communities in human made patches—implications for management. Ecological Applications 4: 702–716. Chapin, F.S., III, and G.R. Shaver. 1985. Individualistic growth response of tundra plant species to manipulation of light, temperature, and nutrients in a ﬁeld experiment. Ecology 66: 564–576. Chapin, F.S., J. Lubchenco, and H.L. Reynolds. 1995. Biodiversity eﬀects on patterns and processes of communities and ecosystems. Pp. 23–45 in V.H. Heywood, ed., Global Biodiversity Assessment. Cambridge University Press, Cambridge. Chapin, F.S. III, B.H. Walker, R.J. Hobbs, D.U. Hooper, J.H. Lawton, O.E. Sala, and D. Tilman. 1997. Biotic control over the functioning of ecosystems. Science 277: 500–504. Didier, A., and I. Poudevigne. 2000. Diversity patterns in grasslands along a landscape gradient in northwestern France. Journal of Vegetation Science 2: 287–294. Doak, D.F., P.C. Marino, and P.M. Kareiva. 1992. Spatial scale mediates the inﬂuence of habitat fragmentation on dispersal success: implications for conservation. Theoretical Population Biology 41: 315–336. Eldridge, D.J. 1993. Cryptogram cover and soil surface condition: eﬀects on hydrology on a semi arid and woodland soil. Arid Soil Research and Rehabilitation 7: 203–217. Eldridge, J.D., E. Zaady, and M. Shachak. 2000. Inﬁltration through three contrasting soil crusts in patterned Negev Desert landscapes. Catena 40: 323–336. Forman, R.T.T., and M. Godron. 1986. Landscape Ecology. John Wiley and Sons, New York. Garner, W., and Y. Steinberger. 1989. A proposed mechanisms for the formation of ‘fertile islands’ in desert ecosystems. Journal of Arid Environments 16: 257–262. Groﬀman, P.M. and P.J. Bohlen. 1999. Soil and sediment biodiversity: cross-system comparisons and large scale eﬀects. BioScience 49: 139–148. Groﬀman, P.M. and J.M. Tiedje. 1989. Denitriﬁcation in north temperate forest soils: spatial and temporal patterns at the landscape and seasonal scales. Soil Biology and Biochemistry 21: 613–620. Hobbie, S.E. 1995. Direct and indirect species eﬀects on biogeochemical processes in arctic ecosystems. Pp. 213–224 in F.S. Chapin, III, and C. Ko¨rner, eds., Arctic and Alpine Biodiversity: Patterns, Causes, and Ecosystem Consequences. SpringerVerlag, Berlin. Hole, F.D. 1978. An approach to landscape analysis with an emphasis on soils. Geoderma 21: 1–23. Huggett, R. 1975. Soil landscape systems: a model of soil genesis. Geoderma 13: 1–22. Huston, M.A. 1994. Biological Diversity: The Coexistence of Species in Changing Landscapes. Cambridge University Press, New York. Ibanez, J.J., S. De-Alba, F.F. Bermudez, and A. Garcia-Alvarez. 1995. Pedodiversity: concepts and measures. Catena 24: 215–232.

Unified Framework II

229

Jones, C.G., and J.H. Lawton. 1995. Linking Species and Ecosystems. Chapman and Hall, New York. Jones, C.G., J.H. Lawton, and M. Shachak. 1994. Organisms as ecosystem engineers. Oikos 69: 373–386. Kareiva, P.M. 1985. Finding and losing host plants by Phylotreta: Patch size and surrounding habitat. Ecology 66: 1809–1816. Lawton, B.H., and Brown, V.K. 1993. Redundancy in ecosystems. Pp. 255–270 in E.D. Schulze, and H.A. Mooney, eds., Biodiversity and Ecosystem Function. Springer-Verlag, New York. Lawton, J.H. and C.G. Jones. 1995. Linking species and ecosystems: Organisms as ecosystem engineers. Pp. 141–150 in C.G. Jones, and J.H. Lawton, eds. Linking Species and Ecosystems. Chapman and Hall, New York. Ludwig, J., and D.T. Tongway. 1995. Spatial organization of landscapes and its function in semiarid woodlands, Australia. Landscape Ecology 10: 51–63. Maubon, M., J.F. Ponge, and J. Andre. 1995. The dynamics of Vaccinium myrtillus patches in mountain forests. Journal of Vegetation Science 6: 343–348. McClaugherty, C.A., J. Pastor, J.D. Aber, and J.M. Melillo. 1985. Forest litter decomposition in relation to soil nitrogen dynamics and litter quality. Ecology 66: 266–275. McCoy, E.D., and S.S. Bell. 1991. Habitat structure: the evolution and diversiﬁcation of a complex topic. Pp. 3–27 in S.S. Bell, E.D. McCoy, and H.R. Musbinsky, eds., Habitat Structure. Chapman and Hall, London. McNaughton, S.J. 1977. Diversity and stability of ecological communities: a comment on the role of empiricism in ecology. The American Naturalist 111: 515–525. Noy-Meir, I. 1985. Desert ecosystem structure and function. Pp. 93–103 in M. Evenari, et al., eds., Hot Deserts and Arid Shrublands. Elsevier Science, Amsterdam. Oﬀer Z.Y., E. Zaady, and M. Shachak. 1997. Aeolian particles input to the soil surface at the northern limit of the Negev desert. Arid Soil Research and Rehabilitation 12: 55–62. Pastor, J., and W.M. Post. 1986. Inﬂuence of climate, soil moisture, and succession on forest carbon and nitrogen cycles. Biogeochemistry 2: 3–27. Pastor, J., J.D. Aber, C.A. McClaughertry, and J.M. Melillo. 1982. Geology, soils and vegetation of Blackhawk Island, Wisconsin. American Midland Naturalist 108: 266–277. Pickett, S.T.A., and M.L. Cadenasso. 1995. Landscape ecology: spatial heterogeneity in ecological systems. Science 269: 331–334. Roland, J. 1993. Large scale forest fragmentation increases the duration of tent caterpillar outbreaks. Journal of Animal Ecology 63: 392–398. Rostango, C.M. 1989. Inﬁltration and sediment production as aﬀected by soil surface conditions in a shrubland of Patagonia, Argentina. Journal of Range Management 42: 382–385. Schimel, D.S., MA. Stillwell, and R.G. Woodmansee. 1985. Biogeochemistry of C, N, and P in a soil catena of the shortgrass steppe. Ecology 66: 276–282. Schlesinger, W.H., J.F. Reynolds, G.L. Cunningham, L.F. Huenneke, W.M. Jarrell, R.A. Virginia, and W.G. Whitford. 1990. Biological feedbacks in global desertiﬁcation. Science 247: 1043–1048. Scholes R.J., and B.H. Walker. 1993. An African Savannah. Cambridge University Press, Cambridge.

230

Ecological Complexes of Biodiversity, Ecosystems, and Landscapes

Shachak, M., and G.M. Lovett. 1998. Atmospheric deposition to a desert ecosystem and its implication for management. Ecological Applications 8: 455–463. Shachak, M., and A. Yair. 1984. Population dynamics and the role of Hemilepistus reaumuri in a desert ecosystem. Symposium Zoological Society of London 53: 295–314. Shachak, M., M. Sachs, and I. Moshe. 1998. Ecosystem management of desertiﬁed shrublands in Israel. Ecosystems 1: 475–483. Snow, J.T., and T.M. McClelland. 1990. Dust devils at White Sands Missile Range, New Mexico. I. Temporal and spatial distributions. Journal of Geophysical Research 95: 13707–13721. Stockton, P.H., and D.A. Gillette. 1990. Field measurement of the sheltering eﬀect of vegetation on erodible land surfaces. Land Degradation and Rehabilitation 2: 77–85. Tilman, D., and J.A. Downing. 1994. Biodiversity and stability in grasslands. Nature 367: 363–365. Tilman, D., D. Wedin, and J. Knops. 1996. Productivity and sustainability inﬂuenced by biodiversity in grassland ecosystems. Nature 379: 718–720. Turner, M.G. 1989. Landscape ecology: the eﬀect of pattern on process. Annual Review of Ecology and Systematics 20: 171–197. Vitousek, P.M. 1990. Biological invasions and ecosystem processes: toward an integration of population biology and ecosystem studies. Oikos 57: 7–13. Walker, B.H. 1992. Biodiversity and ecological redundancy. Conservation Biology 6(1): 18–23. Walker, B. 1995. Conserving biological diversity through ecosystem resilience. Conservation Biology 9: 1–7. Wardell J.G., and P. Horwitz. 1996. Conserving biodiversity and the recognition of heterogeneity in ancient landscapes: A case study from south-western Australia. Forest Ecology and Management 85: 219–238. Weinstein, N. 1975. The eﬀects of a desert shrub on its micro-environment and the herbaceous plants. M.Sc. Thesis. The Hebrew University, Jerusalem. West, N.E. 1989. Spatial pattern—functional interactions in shrub dominated plant communities. Pp. 283–305 in C.M. McKell, ed. The Biology and Utilization of Shrubs. Academic Press, London. Wilby, A., M. Shachak, and B. Boeken. 2001. Integration of ecosystem engineering and trophic eﬀects of herbivores. Oikos 92: 436–44. Wilson, J.B., and A.D.Q. Agnew. 1992. Positive-feedback switches in plant communities. Advances in Ecological Research 23: 263–336. Yair, A., and M. Shachak. 1987. Studies in watershed ecology of an arid area. Pp. 146–193 in M.O. Wurtele, and L. Berkofsky, eds., Progress in Desert Research. Rowman and Littleﬁeld, Totowa, NJ. Zaady, E., P.M. Groﬀman, and M. Shachak. 1996. Release and consumption of nitrogen by snail faeces in Negev Desert soils. Biology and Fertility of Soils 23: 399–404. Zak, D.R., A. Hairston, and D.F. Grigal. 1991. Topographic inﬂuences on nitrogen cycling within an upland pin oak ecosystem. Forest Science 37: 45–53.

Part III

Biodiversity, Conservation, and Management

This page intentionally left blank

14 The Effects of Grazing on Plant Biodiversity in Arid Ecosystems David Ward

C

onventional wisdom views heavy grazing as the major cause of desertiﬁcation in semiarid and arid areas of Africa, Asia, and Australia (see, e.g., Acocks 1953, Jarman and Bosch 1973, Sinclair and Fryxell 1985, Middleton and Thomas 1997) (table 14.1). Nowhere is the eﬀect of heavy grazing more apparent than in the Sahel of Africa (Sinclair and Fryxell 1985). This land denudation has resulted in a negative feedback loop via decreased soil nutrient status and increased soil albedo (due to lower vegetation cover), causing increased evaporation and decreased precipitation, which in turn reduces the stocking capacity of the land, further exacerbating the negative eﬀects of grazing (Schlesinger et al. 1990). A less dramatic result of overgrazing is a long-term decline in agricultural productivity. For example, the arid Karoo region of South Africa has experienced no climatic change over the last two centuries, yet there has been a 50% decline in stocking rates in seven of eight magisterial districts from 1911 to 1981 (Dean and McDonald 1994). These authors ascribe this decline to heavy grazing that reduced palatable plant populations and hence the carrying capacity of the vegetation in the long term. These examples of the negative eﬀects of grazing in arid ecosystems lie in stark contrast with a large number of African studies that compared the eﬀects of commercial (privately owned) and communal (subsistence, no private ownership) ranching on vegetation and soils (e.g., Archer et al. 1989, Tapson 1993, Scoones 1995, Ward et al. 1999a,b, reviewed by Behnke and Abel 1996). In spite of 5–10-fold higher stocking rates on communal ranches, few studies have shown diﬀerences in eﬀects on biodiversity, plant species composition and soil quality between these ranching types (Archer et al. 1989, 233

234

Biodiversity, Conservation, and Management

Table 14.1 The main causes of soil degradation in susceptible drylands. Results expressed in millions of hectares of degraded soil (data from Middleton and Thomas 1997) Region

Overgrazing Deforestation Agricultural

Overexploitation

Bioindustrial

Total Degraded

Africa

184.6

18.6

62.2

54.0

0

319.4

Asia

118.8

111.5

96.7

42.3

1.0

370.3

Australasia

78.5

4.2

4.8

0

0

87.5

Europe

41.3

38.9

18.3

0

0.9

99.4

North America

27.7

4.3

41.4

6.1

0

79.5

South America

26.2

32.2

11.6

9.1

0

79.1

Tapson 1993, Scoones 1995, Ward et al. 1999a,b—ﬁg. 14.1). Similarly, studies of grazing in Mediterranean semiarid grasslands (reviewed by Seligman 1996) and Middle Eastern arid rangelands (Ward et al. 1999b) show that the eﬀects of grazing on biodiversity are relatively small.

Parameters of Consensus on Effects of Grazing in Arid Ecosystems A consensus has developed in recent years that arid grazing ecosystems are nonequilibrial, event-driven systems (see, e.g., O’Connor 1985, Venter et al. 1989, Milchunas et al. 1989, Parsons et al. 1997). Ellis and Swift (1988), Tapson (1993), Werner (1994), and Sullivan (1996) contend that rainfall in arid regions is the major driving variable and has the ability to ‘‘recharge’’ a system that suﬀers heavy grazing pressure. Indeed, it is generally agreed that where pastoralists are able to maintain their activities on a large spatial scale by migrating to areas where key rich resources can be exploited, allowing previously used resources time to recover, negative eﬀects of grazing on plant biodiversity do not develop (Sinclair and Fryxell 1985, Ellis and Swift 1988, Behnke and Abel 1996). Moreover, even where pastoralists are forced to settle in small areas, abiotic variables such as precipitation may be of such overriding importance that these negative eﬀects of grazing on plant cover, plant species richness, and diversity cannot be detected (see, e.g., O’Connor 1985, Ward et al. 1999a,b). An additional factor limiting the eﬀects of grazing in arid ecosystems is the high level of plant resistance to herbivory (Ward and Olsvig-Whittaker 1993, Rohner and Ward 1997, Ward et al. 1997). The cost of regrowth subsequent to herbivory in these ecosystems is high due to low levels of precipitation and low soil nutrient availability (Coley et al. 1985). This has resulted in selection

The Effects of Grazing on Plant Biodiversity in Arid Ecosystems

235

Figure 14.1 Perennial plant percentage cover, species richness, and species diversity (Shannon–Wiener measure) on communal and commercial ranches in arid Otjimbingwe, Namibia. None of the comparisons was signiﬁcant in spite of the 10fold higher stocking densities on communal ranches (data from Ward et al. 1999a).

in plants for resistance (or tolerance) to herbivory, thereby minimizing the impacts of organ (e.g., branches, leaves) removal on ﬁtness (lifetime reproductive success and survival). Thus, low herbivore numbers and high resistance in plants in arid ecosystems work in concert to reduce the eﬀects of grazing in these ecosystems. Empirical data appear to bear this out (see Ward et al. 2000a for examples from arid and semiarid Namibia). In a global review, Milchunas and Lauenroth (1993) showed that the eﬀects of grazing on species composition and plant biomass increase with increasing precipitation (ﬁg. 14.2). The high spatial and temporal variance in plant cover, biodiversity, and soil quality in arid ecosystems makes comparisons of the eﬀects of grazing on plant biodiversity across habitats and ecosystems rather complex. For this reason, many researchers have emphasized comparisons in the piosphere, the region of heavy grazing around watering points. Clearly, animals in arid ecosystems are limited by the availability of water and hence tend to collect there. As a result, these are usually the most heavily damaged areas in a ranching ecosystem. James et al. (1999) summarized piosphere eﬀects in arid Australian ecosystems as follows: 1. The area near a watering point is usually bare, but supports shortlived, often unpalatable, trample-resistant species after rain. 2. A dense zone of unpalatable woody shrubs usually occurs immediately beyond the denuded area.

236

Biodiversity, Conservation, and Management

Figure 14.2 Diﬀerences in species composition between grazed and ungrazed lands in arid ecosystems of Americas and Africa plus Asia (data from Milchunas and Lauenroth 1993). There was no signiﬁcant relationship between species dissimilarity and mean annual rainfall for the American comparison, while there was a signiﬁcant relationship for the African/Asian comparison ðp < 0:001Þ.

3. Palatable perennial plants decline in abundance and species richness within zones 1 and 2 above. 4. Species richness does not change consistently with increasing distance from watering points.

Effects of Evolutionary History of Grazing I focus here on those aspects of the eﬀects of grazing on plant biodiversity of arid lands that are least understood and that should prove to be the most interesting and exciting new ﬁelds of research. Milchunas et al. (1988) predicted that a long evolutionary history of grazing results in selection for regrowth following herbivory and for prostrate growth forms. In such communities, grazing causes rapid shifts between suites of species adapted to either grazing avoidance/tolerance or competition. In their global review, Milchunas and Lauenroth (1993) found that increasing evolutionary history of grazing produced increasing dissimilarity in species composition between grazed and ungrazed sites regardless of the level of precipitation. However, in the Middle East and North Africa, where heavy grazing has occurred for thousands of years, grazing has seldom been shown to aﬀect species composition (Noy-Meir et al. 1989, Perevolotsky 1994, Ward et al. 1999b). A possible reason for this lack of grazing impact is the Narcissus

The Effects of Grazing on Plant Biodiversity in Arid Ecosystems

237

eﬀect (Colwell and Winkler 1984), because selection in the past has resulted in the extinction of all nonresistant/tolerant genotypes. Thus, all extant species are similarly resistant to herbivores, resulting in the absence of an eﬀect of current herbivory on biodiversity (Ward and Olsvig-Whittaker 1993, Perevolotsky 1994). Presumably, in such ecosystems, conditions seldom favor growth-dominated genotypes. Thus, only one (resistant/tolerant) genotype exists in these populations. The conditions under which selection should favor multiple or single genotypes in a population of plants in arid ecosystems has not been adequately investigated. A combination of theoretical and experimental approaches should be followed with this aim.

Period of Rest After Drought and Heavy Grazing There are conﬂicting opinions and data on the eﬀects of the timing of grazing and the period of rest after drought on plant biodiversity. It has widely been observed that plants require a period of rest after drought and heavy grazing (e.g., Danckwerts 1993, Danckwerts and Stuart-Hill 1987, Van der Heyden and Stock 1995). In ecosystems where plants are not allowed to recover, land degradation may occur as quickly as in ecosystems where stocking rates are 5–10 times as high but plants are allowed to recover. This may be one reason why comparisons of the eﬀects of commercial and communal ranching seldom show large diﬀerences in impacts on vegetation in spite of far higher mean stocking rates on communal ranches (Archer et al. 1989, Tapson 1993, Scoones 1995, Ward et al. 1999a,b). Commercial ranchers, who can usually aﬀord to restock their lands soon after drought-induced mortality (or supplement forage to minimize mortality), do not allow the vegetation time to recover lost resources. Communal ranchers, on the other hand, can seldom aﬀord to restock and thus must allow their herds to enlarge through natural reproduction, much in the way that wild herbivore populations recover after drought-induced mortality (Ellis and Swift 1988). During herd regrowth, stocking rates are well below capacity, allowing the plants to recover lost resources. Indeed, the degradation of the Sahel is now considered to be a result of the settlement of pastoralists and the provision of supplementary feed during drought periods, resulting in increased stock survival and greater depletion of plant resources during the postdrought recovery phase (Mainguet 1991, Sinclair and Fryxell 1985). In marked contrast to these studies, many studies comparing continuous and rotational grazing have found that the period of rest after grazing had little independent eﬀect on biodiversity and plant cover in semiarid and arid systems above that induced by diﬀerences in stocking density (e.g., Denny and Steyn 1978, O’Connor 1985). Bryant et al. (1989) compared the purported beneﬁts of a rotational short-duration grazing strategy developed by Savory (1978) with conventional continuous grazing. They found that this stocking system produced no positive inﬂuence on germination or establishment of plants in arid and semiarid regions of North America. Furthermore, this stocking regimen

238

Biodiversity, Conservation, and Management

did not improve range condition at the same or higher stocking rates compared with continuous grazing, and did not increase grass or forb standing crop. They considered stocking density to be the only important factor aﬀecting plant cover and biodiversity. Nonetheless, they recognized that grazing systems that provide plants with rest periods are the only way to restore root reserves and replenish photosynthetic material (Bryant et al. 1989). Controlled studies of the eﬀects of rates of postdrought restocking and herd regrowth on plant biodiversity have not been conducted thus far. It would be particularly useful to examine the eﬀects of grazing on the total nonstructural carbohydrate reserves of plants as a functional index of recovery rate (Van der Heyden and Stock 1995). Another potentially useful tool that bears further investigation in this regard is the traditional range-management classiﬁcation of grass species according to their abilities to withstand grazing (see, e.g., Stoddart et al. 1975). Species are classiﬁed either as increasers (i.e., species that increase in cover and abundance after grazing or disturbance) or decreasers (species that decrease in cover or abundance with grazing). This classiﬁcation, and its many modiﬁcations, was an important ﬁrst step towards understanding and predicting the eﬀects of grazing on biodiversity. These descriptions also represent an important early attempt at ﬁnding assembly rules for grazed plant communities. However, increaser–decreaser species descriptions are rooted in Clements’ (1916) notion that for any given soil/habitat type there is a speciﬁc type of community that will prevail at the end of the successional process (i.e., the climax). Thus, the presence of any species or combinations thereof that are inconsistent with the preconceived notion of the climax was used to indicate that the range was not in good condition. Conversely, the presence of a single species (e.g., Themeda triandra in southern Africa) was universally considered to be a sign of a range in good condition regardless of the associated biodiversity or stocking conditions (O’Connor and Bredenkamp 1997). More recently, it has been shown that such classiﬁcations depend on genotype-by-environment interactions and on the particular competitors in the environment (see Kirkman 1988, Milchunas and Lauenroth 1993, O’Connor and Bredenkamp 1997, among others). In addition to the need for jettisoning the Clementsian notions involved, there is a need for a more mechanistic understanding of what makes a species either increase or decrease with grazing. Ecophysiological studies of the eﬀects of grazing on regrowth, competitive ability, and ﬁtness of key species under a variety of abiotic conditions and with diﬀerent suites of competitors are needed. Genotypic variation in widespread species may result in diﬀerent responses to grazing within the same species (Kirkman 1988). Thus, studies of genotype by environment interactions under diﬀerent grazing intensities will be particularly important. Furthermore, because grazing may aﬀect plant quality and the induction of physical and chemical defenses in plants (see, e.g., McNaughton and Tarrants 1983, Rohner and Ward 1997), which will aﬀect the probability of subsequent grazing, more studies of the eﬀects of grazing on palatability of plants under natural conditions are required.

The Effects of Grazing on Plant Biodiversity in Arid Ecosystems

239

Rate of Degradation Several studies have shown that degradation as a result of overgrazing may take 80–100 years to become manifest, particularly in shrub-dominated habitats (Dean and MacDonald 1994, Wiegand and Milton 1996, Ward and Ngairorue 2000). Examples from arid Namibia are illustrative at two spatial scales. If one compares grass production at the end of the wet season near watering points that have been in use for at least 150 years with those that have been in use for only 10 years, one can see that there is a diﬀerence in grass production. However, one new watering point with particularly high current grazing is in as poor a condition as the old sites (ﬁg. 14.3). At a larger spatial scale at 31 sites along a rainfall gradient from 100–450 mm per annum, we found that there was no correlation between the residuals of grass production (regressed against mean annual rainfall) and stocking density in the current season or averaged over the past 11 years (Ward and Ngairorue 2000). However, when we compare data along the same gradient between 1939 and 1997, grass production in 1997 was approximately 50% less than what it was in the earlier period (Ward and Ngairorue 2000). Thus, some claims that grazing may have little eﬀect on ecosystems may depend on the

Figure 14.3 Comparison of mean slopes of grass height against distance from watering points in March 1998 (end of wet season) at old (>150 years of use) and new (<10 years of use) communal and commercial watering points in arid Otjimbingwe, Namibia (data from Ward et al. 1999a). Commercial watering points have been in use for 80–100 years and have about 10 times fewer stock than at communal watering points. Thus, the diﬀerences in slope shown here indicate that duration of use of watering points has a greater eﬀect on grass production than stocking density.

240

Biodiversity, Conservation, and Management

short time scale of most ecological studies. Appropriate monitoring schemes, coupled with historical studies (e.g., Hennessey et al. 1983, Dick-Peddie 1993, Rohde 1997, Ward et al. 2000b), are necessary to determine whether degradation is occurring.

Bush Encroachment and Denudation Heavy grazing causes denudation in some habitats and bush encroachment in others. Bush encroachment is ascribed to the outcome of competition between tree and grass layers. That is, grasses usually outcompete trees for water and soil nutrients, thereby limiting the opportunity for tree seedling germination and recruitment. Once heavy grazing of grasses occurs, more water, nutrients, and space become available for the trees to germinate, which often occurs en masse, resulting in bush encroachment (Skarpe 1990). The switch between denudation following heavy grazing and bush encroachment usually occurs at the higher end of the rainfall gradient in semiarid/arid ecosystems because trees can only germinate and survive at higher rainfall. However, examples exist where heavy grazing results in bush encroachment in one habitat and not in an adjacent one, in spite of similar rainfall (Ward et al. 1999a). In these cases, it seems that soil type is important (Stock et al. 1997). In deep sandy soils, water percolates deeper into the ground following rainfall than in rocky habitats, where water is trapped close to the soil surface. Thus, less rainfall is required on the latter substrate to facilitate germination of tree seeds en masse following heavy grazing. In the Middle East and in arid parts of Asia, bush encroachment is rare— the exception being encroachment by Sarcopoterium spinosum in some semiarid areas of the Middle East (Noy-Meir et al. 1989)—in spite of appropriate rainfall and soil conditions. In these regions, heavy grazing frequently results in an increase in the density of geophytes, which are chemically well defended (Noy-Meir et al. 1989, Ward et al. 1997). A possible reason for the absence of bush encroachment may be the species pool capable of encroaching. Greater understanding of the interactions between rainfall, soil nutrients and species pool eﬀects is necessary to understand this phenomenon. An interesting hypothesis regarding bush encroachment suggests that increases in atmospheric CO2 since the industrial revolution (ca. 30%) have beneﬁted C3 over C4 plants (Idso 1992); that is, the historical displacement of C4 grasses by C3 woody plants may be due to the diﬀerential responses of their photosynthetic physiologies to increases in CO2 rather than (or in addition to) the eﬀects of grazing (Archer 1996). Archer (1996) notes that there are a number of factors related to life form and growth form that confound results of the eﬀects of CO2 increase on bush encroachment. Nonetheless, this is a most interesting hypothesis deserving of more detailed multifactorial experiments with CO2, grazing, rainfall, and ﬁre treatments.

The Effects of Grazing on Plant Biodiversity in Arid Ecosystems

241

Impact of Grazing on Soil Resources The popular concept of arid ecosystems as fragile has frequently been exaggerated (Forse 1989). Some arid ecosystems are capable of recovering after years of heavy grazing and apparently suﬀer no long-term eﬀects (Venter et al. 1989, Ward et al. 1999a—ﬁg. 14.4; reviewed by Shackleton 1993). A key factor in the recovery of ecosystems following heavy grazing may lie in the diﬀerential ability of certain soils to withstand vegetation removal without suﬀering undue losses of nutrients (clearly, recovery is also contingent upon the availability of seeds in either the seed bank or seed rain (O’Connor and Pickett 1992). It will be particularly important to survey the eﬀects of grazing on soil organic carbon (which is closely associated with the amount of soil nitrogen and soil water-holding capacity; Ward et al. 1999a,b). Interannual variation in vegetation cover may be diﬃcult to ascribe to either grazing or rainfall ﬂuctuation because of the high spatiotemporal variance in rainfall in arid systems. However, sustained heavy grazing will cause long-term declines in vegetation cover, which will reduce detrital contributions to the soil and, thus, soil organic carbon (Ward et al. 1999a). Soil organic carbon is a far more reliable measure of ecosystem decline and long-term potential to sustain biodiversity than is plant biomass or standing crop. Also, there is a strong positive correlation between soil organic carbon and soil nitrogen (the most important nutrient for plants) and soil water-holding capacity (Ward et al. 1999a).

Figure 14.4 Diﬀerences in soil nutrients between communal and commercial ranches at Otjimbingwe, Namibia. None of the comparisons was signiﬁcant in spite of the 10fold higher stocking densities on communal ranches (data from Ward et al. 1999a). Organic carbon is measured in percent dry soil mass, nitrogen and phosphorus are measured in mg/kg, while the bioassay is the dry mass in grams of radish plants grown in pots containing the two soil types after 30 days (Ward et al. 1999a).

242

Biodiversity, Conservation, and Management

Another important factor in ecosystem change and change in biodiversity following grazing is the spatial redistribution of soil nutrients (Schlesinger et al. 1996). Schlesinger et al. (1996) found that encroachment of arid grasslands by trees (esp. Larrea tridentata) in the American Southwest led to a redistribution of soil nutrients from a relatively even distribution in grassland to a concentration of nutrients under the trees. They consider this to lead to a net loss of productivity of these systems. However, Stock et al. (1999) have recently shown that such soil patterns can occur in the absence of grazing and purported desertiﬁcation. Also, if the trees that encroach are nitrogen ﬁxers (as is the case in most African bush encroaching species), total soil quality may increase rather than decrease. More emphasis needs to be placed on the development of appropriate indicators of soil degradation and on the study of nutrient cycling and redistribution under diﬀerent levels of grazing in arid ecosystems before any general pattern can be claimed.

Replacement of Perennials by Annuals In many arid ecosystems, annual species replace perennials following heavy grazing owing to their ability to quickly invade open spaces and utilize soil resources (e.g., Kelly and Walker 1977, Cheal 1993, Freeman and Emlen 1995). The transient nature of the annual lifestyle means that herbivores are less likely to encounter them and thus annuals increase in abundance, while perennials decrease with grazing. This eﬀect seems obvious and has achieved the status of dogma. However, Ellis and Swift (1988) and Tainton et al. (1996) indicate that such an eﬀect may only occur in semiarid and arid ecosystems where supplementary livestock feed is provided during drought or where animals are introduced from outside the system (see comparison of commercial vs. communal ranching above), so that animals may be held at artiﬁcially high levels in relation to available forage at certain times. This practice has been shown to lead to a progressive damping of postdrought recovery of perennial species in semiarid grasslands, and the ultimate loss of the perennial component (Hatch 1994). I note, as have Kelt and Valone (1995) and Tainton et al. (1996), that many studies have reported the coassociation of heavy grazing and a prevalence of annual species in arid systems and assumed that there is a cause-and-eﬀect relationship between them (i.e., perennials have been removed by grazing) without ascertaining that this is indeed the case. Annual species predominate in arid ecosystems, due to the patchy nature of water availability (Rutherford 1980) regardless of the level of grazing. Indeed, in the Middle East and in parts of Africa where heavy grazing has been prevalent for thousands of years, the prevalence of annual species in the herb layer in arid ecosystems cannot be unequivocally ascribed to the eﬀects of grazing (Noy-Meir et al. 1989, Ward et al. 1999a,b). In North America, Kelt and Valone (1995) found that the removal of herbivores (cattle) had little impact on the abundance and diversity of annual plants in a Chihuahuan desert site in south-eastern Arizona. The presence of

The Effects of Grazing on Plant Biodiversity in Arid Ecosystems

243

many species that are facultatively biennial or perennial (depending on the amount and persistence of rainfall) further clouds an already blurred picture. There has been an almost complete absence of integration of grazing studies with life history strategy theory (with the notable exception of O’Connor 1991)—models of adaptive phenotypic plasticity and facultative perenniality have a lot to oﬀer rangeland science as they predict the conditions under which certain strategies are optimal. Several optimization models produced in the 1960s and 1970s (e.g., Cohen 1966, Gadgil and Bossert 1970, Charnov and Schaﬀer 1973, Bell 1976, Stearns 1976) indicated that an annual lifestyle may be an optimal one in highly variable systems (e.g., in arid ecosystems below 200 mm rainfall regardless of the level of grazing). For example, if the shape of the tradeoﬀ curve between optimal reproductive eﬀort (ordinate) and adult survival (abscissa) is concave downwards, there will be selection for perenniality (ﬁg. 14.5a). If the tradeoﬀ curve is convex downwards, the optimal lifespan is the minimal one (i.e., annual) regardless of the source of mortality (ﬁg. 14.5b—Stearns 1976). Unfortunately, such models have been all but abandoned and experimental tests of their predictions ignored. An important point to consider here, which bears on transcontinental compar-

Figure 14.5 Eﬀects of the shape of the tradeoﬀ curve between optimal reproductive eﬀort and adult survival on annual and perennial lifestyles (after Stearns 1976). (a) A tradeoﬀ curve that is concave downwards results in selection for iteroparity and thus perenniality in species with a single reproductive season per year because the ﬁtness isocline subtends the tradeoﬀ curve at an intermediate value of adult survival. (b) A tradeoﬀ curve that is convex downwards results in selection for semelparity and thus an annual lifestyle in species with a single reproductive season per year.

244

Biodiversity, Conservation, and Management

isons of the eﬀects of grazing, is that many studies that report shifts to an annual lifestyle following grazing have been in North America and Australia, where deserts are typically wetter (but hotter) than those elsewhere (Louw and Seely 1982). Thus, the lower annual rainfall in the Middle Eastern and African deserts may select for an annual lifestyle regardless of the level of grazing. Appropriate tests of these life history models may help to understand the probability of invasion by annuals following grazing.

Positive Effects of Herbivores on Biodiversity Herbivores can alter the competitive interactions among plant species, sometimes removing a small number of dominant species and thereby allowing a larger number of subordinate species to become more abundant (Crawley 1983, Landsberg et al. 1999). Another instance of the promotion of greater species diversity by herbivores occurs in areas where they frequently defecate, creating nutrient-rich patches for plant establishment, or where they disturb the ground by rolling and sandbathing, thereby facilitating the invasion of disturbance-tolerant species (Midgley and Musil 1990, James et al. 1999). Also, herbivores may increase the germination of many plant species with hard seed coats such as Acacia species via passage through their digestive tracts and subsequent defecation (Miller and Coe 1993, Rohner and Ward 1999). Where such plant species are keystone species promoting the biodiversity of other plant species, herbivores may be particularly important (Belsky et al. 1993, Dean et al. 1999, Munzbergova and Ward 1999). In ecosystems where mammalian herbivores have been removed or have declined, the absence of recruitment facilitation by herbivores may have serious impacts on the maintenance of biodiversity (Reid and Ellis 1995, Ward and Rohner 1997, Rohner and Ward 1999). Further examination of the relative importance of facilitation by herbivores on ecosystem biodiversity, as opposed to a local-scale description of the phenomena involved, is necessary.

Acknowledgments This study was partially ﬁnanced by grants from U.S. AID grant no. TAU-MOU-94-C13–149 and Keren Keyemet L’Israel.

References Acocks, J.P.H. 1953. Veld types of South Africa. Memoirs of the Botanical Survey of South Africa 28, 1–192. Archer, F.M., M.T. Hoﬀman, and J.E. Danckwerts. 1989. How economic are the farming units of Leliefontein, Namaqualand? Journal of the Grassland Society of Southern Africa 6, 211–215.

The Effects of Grazing on Plant Biodiversity in Arid Ecosystems

245

Archer, S. 1996. Assessing and interpreting grass-woody plant dynamics. Pp. 101–134 in J. Hodgson, and A.W. Illius, eds., The Ecology and Management of Grazing Systems. CAB International, Wallingford. Behnke, R. and N. Abel. 1996. Revisited: the overstocking controversy in semiarid Africa. World Animal Review 87, 4–27. Bell, G. 1976. On breeding more than once. American Naturalist 110, 57–77. Belsky, A.J., S.M. Mwonga, R.G Amundson, J.M. Duxbury, and A.R. Ali. 1993. Comparative eﬀects of isolated trees on their undercanopy environments in high- and low-rainfall savannas. Journal of Applied Ecology 26, 1005–10024. Bryant, F.C., B.E. Dahl, R.D. Pettit, and C.M. Britton. 1989. Does short-duration grazing work in arid and semi-arid regions? Journal of Soil and Water Conservation 44, 290–296. Charnov, E.L., and W.M. Schaﬀer. 1973. Life history consequences of natural selection: Cole’s result revisited. American Naturalist 107, 791–793. Cheal, D.C. 1993. Eﬀects of stock grazing on the plants of semi-arid woodlands and grasslands. Proceedings of the Royal Society of Victoria 105, 57–65. Clements, F.E. 1916. Plant succession: an analysis of the development of vegetation. Carnegie Institute Washington Publication 242, 1–512. Cohen, D. 1966. Optimizing reproduction in a randomly varying environment. Journal of Theoretical Biology 12, 119–129. Coley, P.D., J.P. Bryant, and F.S. Chapin. 1985. Resource availability and plant antiherbivore defense. Science 230, 895–899. Colwell, R.K., and D.W. Winkler. 1984. A null model for null models in biogeography. Pp. 344–359 in D.R. Strong, D. Simberloﬀ, L.G. Abele, and A.B. Thistle, eds., Ecological Communities: Conceptual Issues and the Evidence. Princeton University Press, Princeton, NJ. Crawley, M.J. 1983. Herbivory. Blackwell, Oxford. Danckwerts, J.E. 1993. Reserve carbon and photosynthesis: their role in regrowth of Themeda triandra Forsk., a widely distributed subtropical graminaceous species. Functional Ecology 7, 634–641. Danckwerts, J.E., and G.C. Stuart-Hill. 1987. Adaptation of a decreaser and an increaser grass species to defoliation in semi-arid grassveld. Journal of the Grassland Society of Southern Africa 5, 218–222. Dean, W.R.J., and I.A.W. MacDonald. 1994. Historical changes in stocking rates of domestic livestock as a measure of semi-arid and arid rangeland degradation in the Cape Province, South Africa. Journal of Arid Environments 26, 281–298. Dean, W.R.J., S.J. Milton, and F. Jeltsch. 1999. Large trees, fertile islands, and birds in arid savanna. Journal of Arid Environments 41, 61–78. Denny, R.P., and J.H.S. Steyn. 1978. A comparison of continuous and rotational grazing on open granite sandveld. Annual Report of the Department of Agriculture. Harare, Zimbabwe. Dick-Peddie, W.A. 1993. New Mexico Vegetation: Past, Present and Future. University of New Mexico Press, Albuquerque, NM. Ellis, J.E., and D.M. Swift. 1988. Stability of African pastoral ecosystems: alternate paradigms and implications for development. Journal of Range Management 41, 450–459. Forse, W. 1989. The myth of the marching desert. New Scientist 121, 31–32. Freeman, D.C., and J.M. Emlen. 1995. Assessment of interspeciﬁc interactions in plant communities: an illustration of the cold desert saltbush grasslands of North America. Journal of Arid Environments 31, 179–198.

246

Biodiversity, Conservation, and Management

Gadgil, M., and Bossert, W. 1970. Life history consequences of natural selection. American Naturalist 104, 1–24. Hatch, G.P. 1994. The bioeconomic implications of various stocking strategies in the semi-arid savanna of Natal. Unpubl. PhD thesis, University of Natal, Pietermaritzburg, South Africa. Hennessey, J.T., R.P. Gibbens, J.M. Tromble, and M. Cardenas. 1983. Vegetation changes from 1935 to 1980 in mesquite dunelands and former grasslands of southern New Mexico. Journal of Range Management 36, 370–374. Idso, S.B. 1992. Shrubland expansion in the American Southwest. Climate Change 22, 85–86. James, C.D., J. Landsberg, and S.R. Morton. 1999. Provision of watering points in the Australian arid zone: a review of eﬀects on biota. Journal of Arid Environments 41, 87–121. Jarman, N.G. and O.J.H. Bosch. 1973. The identiﬁcation and mapping of extensive secondary invasive and degraded ecological types (test site D). Pp. 77–80 in O.G. Malan, ed., To Assess the Value of Satellite Imagery in Resource Evaluation on a National Scale. CSIR, Pretoria, South Africa. Kelly, R.D. and B.H. Walker. 1977. The eﬀects of diﬀerent forms of land use on the ecology of a semi-arid region in south-eastern Rhodesia. Journal of Ecology 62, 553–574. Kelt, D.A. and T.J. Valone. 1995. Eﬀects of grazing on the abundance and diversity of annual plants in Chihuahuan desert scrub habitat. Oecologia 103, 191–195. Kirkman, K.P. 1988. Factors aﬀecting the seasonal variation of veld quality in South Africa. Unpubl. MSc thesis, University of Natal, Pietermaritzburg, South Africa. Landsberg, J., T. O’Connor, and D. Freudenberger. 1999. The impacts of livestock grazing on biodiversity in natural ecosystems. Pp. 752–777 in H-J.G. Jung, and G.C. Fahey, eds., Nutritional Ecology of Herbivores. Proceedings of the Vth International Symposium on the Nutrition of Herbivores, American Society of Animal Science, Savoy, IL. Louw, G.N., and M.K. Seely, 1982. Ecology of Desert Organisms. Longman, Harlow, Essex, UK. Mainguet, M. 1991. Desertiﬁcation: Natural Background and Human Mismanagement. Springer-Verlag, Berlin. McNaughton, S.J., and J.L. Tarrants. 1983. Grass leaf siliciﬁcation: natural selection for an inducible defense against herbivores. Proceedings of the National Academy of Sciences, USA 80, 790–791. Middleton, N.J., and D.S.G. Thomas. 1997. World Atlas of Desertiﬁcation, 2nd edn. Edward Arnold, London. Midgley, G.F., and C.F. Musil. 1990. Substrate eﬀects of zoogenic mounds on vegetation composition in the Worcester-Robertson valley, Cape province. South African Journal of Botany 56, 158–166. Milchunas, D.G., O.E. Sala, and W.K. Lauenroth. 1988. A generalized model of the eﬀects of grazing by large herbivores on grassland community structure. American Naturalist 132, 87–106. Milchunas, D.G., W.K.Lauenroth, P.L. Chapman, and K. Kazempour. 1989. Eﬀects of grazing, topography, and precipitation on the structure of a semi-arid grassland. Vegetatio 80, 11–23. Milchunas, D.G., and W.K. Lauenroth. 1993. Quantitative eﬀects of grazing on vegetation and soils over a global range of environments. Ecological Monographs 63, 327–366.

The Effects of Grazing on Plant Biodiversity in Arid Ecosystems

247

Miller, M.F., and M. Coe. 1993. Is it advantageous for Acacia seeds to be eaten by ungulates? Oikos 66, 364–368. Munzbergova, Z., and D. Ward. 1999. Acacia trees as keystone species in Negev desert ecosystems. Journal of Vegetation Science 13, 227–236. Noy-Meir, I., M. Gutman, and Y. Kaplan. 1989. Responses of mediterranean grassland plants to grazing and protection. Journal of Ecology 77, 290–310. O’Connor, T.G. 1985. A synthesis of ﬁeld experiments concerning the grass layer in the savanna regions of southern Africa. South African National Scientiﬁc Programmes Report 114. CSIR, Pretoria. O’Connor, T.G. 1991. Local extinction in perennial grasslands: a life-history approach. American Naturalist 137, 753–773. O’Connor, T.G., and G.J. Bredenkamp. 1997. Grassland. Pp. 215–257 in R.M. Cowling, D.M. Richardson, and S.M. Pierce, eds., Vegetation of Southern Africa. Cambridge University Press, Cambridge. O’Connor, T.G., and G.A. Pickett, 1992. The inﬂuence of grazing on seed production and seed banks of some African savanna grasslands. Journal of Applied Ecology 29, 247–260. Parsons, D.A.B., C.M. Shackleton, and R.J. Scholes. 1997. Changes in herbaceous layer condition under contrasting land use systems in the semi-arid lowveld, South Africa. Journal of Arid Environments 37, 319–329. Perevolotsky, A. 1994. Tannins in Mediterranean oak woodland species: lack of response to browsing and thinning. Oikos 71, 333–340. Reid, R.S., and J.E. Ellis. 1995. Impacts of pastoralists on woodlands in south Turkana, Kenya: livestock-mediated tree recruitment. Ecological Applications 5, 978–992. Rohde, R.F. 1997. Looking into the past: interpretations of vegetation change in western Namibia based on matched photography. Dinteria 25, 121–149. Rohner, C., and D. Ward. 1997. Chemical and mechanical defense against herbivory in two sympatric species of desert Acacia. Journal of Vegetation Science 8, 717–726. Rohner, C., and D. Ward. 1999. Mammalian herbivores and the conservation of arid Acacia stands in the Middle East. Conservation Biology 13, 1162–1171. Rutherford, M. 1980. Annual plant production-precipitation relations in arid and semi-arid regions. South African Journal of Science 76, 53–56. Savory, A. 1978. A holistic approach to ranch management using short duration grazing. Pp. 555–557 in Proceedings of the 1st International Rangeland Congress, Denver, CO. Schlesinger, W., J.F. Reynolds, G.L. Cunningham, L.F. Huenneke, W.M. Jarrell, R.A. Virginia, and W.G. Whitford. 1990. Biological feedbacks in global desertiﬁcation. Science 247, 1043–1048. Schlesinger, W., J.A. Raikes, A.E. Hartley, and A.F. Cross. 1996. On the spatial pattern of soil nutrients in desert ecosystems. Ecology 77, 364–374. Scoones, I. 1995. Contrasting savanna dynamics: implications for livestock populations in Zimbabwe’s dryland communal areas. Pp. 353–360 in G.F. Campbell, ed., Sustainable Land Management in Semi-arid and Subhumid Regions. Proceedings of the SCOPE Workshop, Dakar, Senegal. CIRAD, Montpellier, France. Seligman, N.G. 1996. Management of Mediterranean grasslands. Pp. 359–381 in J. Hodgson, and A.W. Illius, eds., The Ecology and Management of Grazing Systems. CAB International, Wallingford, UK.

248

Biodiversity, Conservation, and Management

Shackleton, C.M. 1993. Are the communal grazing lands in need of saving? Development Southern Africa 10, 65–78. Sinclair, A.R.E., and J.M. Fryxell. 1985. The Sahel of Africa: ecology of a disaster. Canadian Journal of Zoology 63, 987–994. Skarpe, C. 1990. Shrub layer dynamics under diﬀerent herbivore densities in an arid savanna, Botswana. Journal of Applied Ecology 27, 873–885. Stearns, S.C. 1976. Life history tactics: a review of the ideas. Quarterly Review of Biology 51, 3–47. Stock, W.D., N. Allsopp, F. Van der Heyden, and E.T.F. Witkowski, 1997. Plant form and function. Pp. 376–396 in R.M. Cowling, D.M. Richardson, and S.M. Pierce, eds., Vegetation of Southern Africa. Cambridge University Press, Cambridge. Stock, W.D., T.S. Dlamini, and R.M. Cowling. 1999. Plant-induced fertile islands as possible indicators of desertiﬁcation in a succulent desert ecosystem in northern Namaqualand, South Africa. Plant Ecology 142, 161–167. Stoddart, L.A., A.D. Smith, and T.W. Box. 1975. Range management. McGraw-Hill, New York. Sullivan, S. 1996. Towards a non-equilibrium ecology: perspectives from an arid land. Journal of Biogeography 23, 1–5. Tainton, N.M., C.D. Morris, and M.B. Hardy. 1996. Complexity and stability in grazing systems. Pp. 275–299 in J. Hodgson, and A.W. Illius, eds., The Ecology and Management of Grazing Systems. CAB International, Wallingford, UK. Tapson, D.R. 1993. Biological sustainability in pastoral systems: The kwazulu case. Pp. 118–135 in R.H. Behnke, I. Scoones, and C. Kerven, eds., Range Ecology at Disequilibrium. New Models of Natural Variability and Pastoral Adaptation in African Savannas. ODI/IIED/Commonwealth Secretariat, London. Van der Heyden, F., and W.D. Stock. 1995. Non-structural carbohydrate allocation following diﬀerent frequencies of simulation browsing in three semi-arid shrubs. Oecologia 102, 238–245. Venter, J., B. Liggitt, N.M. Tainton, and G.P.M. Clarke, 1989. The inﬂuence of diﬀerent land use practices on soil erosion, herbage production, and grass species richness and diversity. Journal of the Grassland Society of Southern Africa 6, 89–98. Ward, D., and B.T. Ngairorue. 2000. Are Namibia’s grasslands desertifying? Journal of Range Management 53, 138–144. Ward, D., and L. Olsvig-Whittaker. 1993. Plant species diversity at the junction of two desert biogeographic zones. Biodiversity Letters 1, 172–185. Ward, D., and C. Rohner. 1997. Anthropogenic causes of high mortality and low recruitment in three Acacia tree taxa in the Negev desert, Israel. Biodiversity and Conservation 6, 877–893. Ward, D., M. Spiegel, and D. Saltz. 1997. Gazelle herbivory and interpopulation diﬀerences in calcium oxalate content of leaves of a desert lily. Journal of Chemical Ecology 23, 333–346. Ward, D., B.T. Ngairorue, J. Kathena, R. Samuels, and Y. Ofran. 1999a. Land degradation is not a necessary outcome of communal pastoralism in arid Namibia. Journal of Arid Environments 40, 357–371. Ward, D., D. Saltz, M. Rowen, and I. Schmidt, 1999b. Eﬀects of grazing by reintroduced Equus hemionus on the vegetation in a Negev desert erosion cirque. Journal of Vegetation Science 10, 579–586. Ward, D., B.T. Ngairorue, J. Karamata, and I. Kapoﬁ, 2000a. The eﬀects of communal and commercial pastoralism on vegetation and soils in an arid and a semi-arid

The Effects of Grazing on Plant Biodiversity in Arid Ecosystems

249

region of Namibia. Pp. 344–347 in P.S. White, L. Mucina, J. Leps, and E. Van Der Maarel, eds., Vegetation Science in Retrospect and Perspective. Opulus Press, Uppsala, Sweden. Ward, D., B.T. Ngairorue, A. Apollus, and H. Tjiveze. 2000b. Perceptions and realities of land degradation in arid Otjimbingwe, Namibia. Journal of Arid Environments 45, 337–356. Werner, W. 1994. Land tenure issues in Namibia: interactions with sustainable resource management. Pp. 27–41 in S. Wolters, ed., Proceedings of Namibia’s National Workshop to Combat Desertiﬁcation. Desert Research Foundation of Namibia, Windhoek, Namibia. Wiegand, T., and S.J. Milton. 1996. Vegetation change in semiarid communities: simulating probabilities and time scales. Vegetatio 125, 169–183.

15 Sustainability in Arid Grasslands New Technology Applications for Management Arian Pregenzer Robert R. Parmenter Howard Passell John R. Vande Castle Thomas K. Budge Gregory Michael Bonito

A

s biodiversity, ecosystem function, and ecosystem services become more closely linked with human well-being at all scales, the study of ecology takes on increasing social, economic, and political importance. However, when compared with other disciplines long linked with human well-being, such as medicine, chemistry, and physics, the technical tools and instruments of the ecologist have generally lagged behind those of the others. This disparity is beginning to be overcome with the increasing use of biotelemetric techniques, microtechnologies, satellite and airborne imagery, geographic information systems (GIS), and both regional and global data networks. We believe that the value and eﬃciency of ecosystem studies can advance signiﬁcantly with more widespread use of existing technologies, and with the adaptation of technologies currently used in other disciplines to ecosystem studies. More importantly, the broader use of these technologies is critical for contributing to the preservation of biodiversity and the development of sustainable natural resource use by humans. The concept of human management of biodiversity and natural systems is a contentious one. However, we assert that as human population and resource consumption continue to increase, biodiversity and resource sustainability will only be preserved by increasing management eﬀorts—if not of the biodiversity and resources themselves, then of human impacts on them. The technologies described in this chapter will help enable better management eﬀorts. In this context, biodiversity refers not only to numbers of species (i.e., richness) in an arbitrarily deﬁned area, but also to species abundances 250

Sustainability in Arid Grasslands

251

within that area. Sustainability refers to the maintenance of natural systems, biodiversity, and resources for the beneﬁt of future generations. Arid-land grazing systems support human social systems and economies in regions all over the world, and can be expected to play increasingly critical roles as human populations increase. Further, grazing systems represent a nexus of natural and domesticated systems. In these systems, native biodiversity exists side by side with introduced species and populations, and in fact can beneﬁt from them. The economic and political implications of grazing in arid lands, the controversy over its impacts, and the pressing need for resolving the controversies and for creating sustainable grazing systems make arid-land grazing systems a prime focal point for the application of new technologies (Milchunas and Lauenroth 1993, Fleischner 1994, Bich et al. 1995, NyakoLartey and Baxter 1995, Memmott et al. 1998, Belsky et al. 1999, Hiernaux et al. 1999, Jones and Longland 1999, Rohner and Ward 1999). We frame the chapter around the hypothetical management of an area of semiarid and arid grasslands, shrublands, and woodlands for sustainable grazing. However, the approach described here could equally well be applied to pure conservation and preservation of the same areas. For example, the same approaches could be applied to the management of mixed ungulate herds in an African game park, as well as to a cattle ranch in Texas. Considering these approaches in the context of arid lands, grazing is particularly appropriate, because in most regions of the world, arid lands are being used for grazing. We have organized the discussion with the following questions: (1) What kinds of data about arid-land grazing systems, biodiversity, and sustainability are currently needed by researchers and land managers? (2) How are data collection and analysis needs met currently? (3) How can technological innovations ﬁll gaps or improve eﬃciency? To characterize ecosystem function and biodiversity, or to develop a plan for the sustainable yield of economically valuable resources and the sustainable maintenance of natural systems, data will be needed on climate, soil, vegetation, domestic livestock, and/or wildlife. In the next sections we will discuss particular data needs and how technology can play a role in data collection. Then we will present a scenario that shows how data on all these variables can be collected, analyzed, and integrated into management strategies using currently or soon-to-be available technologies. Our discussion of technologies is necessarily brief, but numerous references provide more detailed information. The ideas expressed here about the application of technologies to arid-land grazing systems transfer easily to other ecosystems and other research and management issues.

Acquiring Data on Vegetation Maintaining sustainable arid-land grazing systems requires extensive data on vegetation. Important parameters include: (1) the rates of increase or decrease in vegetation biomass and/or net primary productivity due to biotic and

252

Biodiversity, Conservation, and Management

abiotic drivers (including grazing); (2) the changes in species composition; (3) the eﬀect of changes in species composition on wildlife and livestock. Better understanding of these critical parameters will shed light on potential longterm anthropogenic impacts on ecosystem function and native biodiversity. Field methods for obtaining such data include clipping, drying, and weighing vegetation samples from line transects or study plots, measuring plant volumes, and other labor-intensive techniques. However, in the last two decades, optical and multispectral remote sensing data from aircraft and satellites have become a signiﬁcant source of spatially large-scale data on soils, biomass, vegetation patchiness, distribution of microbiotic crusts, leaf area index, vegetation indices, changes in species composition, and net primary production (Anderson 1996, Reed et al. 1996, Karnieli 1997, Todd et al. 1998, Zhou 1998). Remote sensing is also being used to collect data on water quality and availability, mineral composition of soils and soil erosion patterns (Rao et al. 1996, Vande Castle 1998). It should be noted that these approaches, and many others we describe, are large-scale approaches that must complement smaller scale studies on the ground. Once properly calibrated by on-the-ground studies, the eﬃcacy of remote sensing technologies is greatly enhanced. The primary use of remote sensing, given the coarse pixel sizes available today, is to identify portions of a landscape with characteristics that may support particular species or assemblages of interest (Muldavin et al. 2001). As pixel sizes become smaller and as multispectral imagery becomes more sensitive, we would optimistically predict that remote sensing could be used to identify individuals or small populations of particular species within a matrix of other species and populations. Remote sensing can be divided into the two broad categories of passive and active sensing. Passive sensing refers to the collection of optical and multispectral data from the reﬂection of the sun’s electromagnetic radiation from an object, or from thermal emission from an object. ‘‘Active sensing’’ refers to technologies that project radiation—such as a laser or radar energy—at the object, and then measure the amount of reﬂected energy. Synthetic aperture radar (SAR) is a common example of active sensing. Both active and passive sensing can be employed from aircraft or satellites, and both, along with image processing and GIS software applications that manage the data, have increasingly important uses in ecological studies (Anderson 1996, Vande Castle 1998). Three kinds of sensor resolution are important in matching a particular remote sensing technology to a particular study: spatial, spectral and temporal (Anderson 1996, Vande Castle 1998). The ﬁrst kind, spatial resolution, refers to the smallest object that can be detected by a remote sensing instrument. An instrument with a 30 m resolution, such as that available from the Landsat Thematic Mapper (TM), for example, will average the spectral reﬂections for each 30 m pixel in its view. This resolution is suﬃcient for providing data on coarse-scaled vegetation diﬀerences (grassland to shrubland to forest), and on geographic and topographic features. A TM image covers an area of 170 km 185 km. Data available from the Advanced Very High Resolution Radiometer (AVHRR)

Sustainability in Arid Grasslands

253

instruments cover areas wider than 2000 km at a spatial resolution of 1 km. Other commercial data achieve spatial resolutions as high as 1 m. The second kind of resolution important to remote sensing is spectral resolution, which refers both to the total range of wavelengths and the number of spectral bands into which that range of wavelengths is subdivided. Instruments with higher spectral resolution (including more bands and a larger range of wavelengths) can provide data that better discriminate between shades, colors, and emitted heat, which can often be associated with diﬀerent kinds of soils, diﬀerent species, or diﬀerent quantities of foliage. Landsat 7 TM images, for instance, include data on seven bands spanning wavelengths from the visible to thermal infrared spectra. Current airborne and future satellite-based hyperspectral sensing technologies will provide data in more than 300 bands, greatly increasing data resolution. The third kind of resolution is temporal resolution, which refers to how often a scene can be revisited by the same instrument, or the time between successive images. The revisit time for Landsat 7, for instance, is 16 days. Other satellites are able to revisit a scene within one day due to pointable optics; that is, they can record data from a scene even if their orbit does not take them directly over the scene. The revisit time for visible AVHRR data is twice a day using two satellites. Common metrics used in remote sensing include the Normalized Diﬀerence Vegetation Index (NDVI) and other reﬂectance indices (Karnieli 1997, Vande Castle 1998). NDVI is a measure of ‘‘greenness’’ and is described by the following equation: NDVI ¼ (Near IR Red)/(Near IR + Red). The NDVI is calculated for each pixel in a landscape with data values from the respective bands, and then landscapes can be described by the mean NDVI, the distribution of NDVI values, or other statistics for the entire data set comprising all pixels across the landscape. The NDVI values typically show changes following periods of precipitation, such as the onset of summer monsoons, or following drought or winter senescence of vegetation; the NDVI has also been used to estimate net primary production (NPP) (Karnieli 1997, Vande Castle 1998). Satellite-borne radar sensors are becoming increasingly important for detecting topography, erosion, and vegetation types. Radar systems are active systems, projecting their own electromagnetic radiation at an object. Unlike passive remote sensing instruments, they are not dependent on solar reﬂectance, allowing them to operate at night. Clouds are transparent to the microwave energy employed by radar systems, so these systems are not limited by weather, as are passive remote systems (Forster 1996). A constant challenge facing all these remote sensing technologies is the development of eﬃcient and reliable methods for ‘‘ground-truthing,’’ or the classiﬁcation of spectral information in the data to the actual species or species assemblages on the ground. Classiﬁcation is currently time, energy, and labor intensive, and seriously retards the speed at which the huge data sets provided by remote sensing can be processed. In order to accurately link spectral information with actual ground information, ground level ﬁeld, lab, and in situ measurements are critical. A few technologies currently available,

254

Biodiversity, Conservation, and Management

which may enhance ground level productivity and biodiversity measurements, include ground-penetrating radar (GPR), minirhizotrons, and DNA microarrays. Ground-penetrating radar emits electromagnetic waves and can map patterns of below ground soil and substrate characteristics (soil depth, depth to a clay layer, etc.). Minirhizotrons use small video cameras equipped with a magnifying lens to image below-ground systems. The minirhizotrons are lowered into a clear PVC pipe previously buried in the soil, and can be used to record time-series of soil proﬁles, root systems, and below-ground organism interactions in situ. DNA microarrays allow rapid automated genetic analysis of prepared samples (Shi 2002), and have great potential for accelerating the speed of biodiversity discovery. The capacity of DNA microarrays has the potential to increase the eﬃciency with which remotely gathered data can be correlated with actual ground species. Once the initial classiﬁcation of a landscape is completed, some analysis of data collected later on the same landscape can be conducted with less intensive classiﬁcation, but also with less precision. Using GIS systems, landscape classiﬁcations can be integrated with other data sets to search for patterns and potential ecosystem feedbacks.

Acquiring Data on Domestic Livestock and Wildlife Understanding the impacts of grazing by wild or domestic ungulates on ecosystem function, biodiversity, and sustainability requires data on: (1) routes and patterns of movement used by grazing animals, and their eﬀect on vegetation, soils and water; (2) modiﬁcation of wildlife species movements and behavior as a result of interaction with grazing animals; (3) eﬀects of grazing on other species population dynamics; and (4) the eﬀects of grazing systems on wildlife populations in neighboring regions. The current standard for addressing these questions is through ﬁeld observations. Though ﬁeld studies ultimately provide the foundation for all ecosystem studies by linking theory with actual ecosystem function, they are constrained by limitations on personnel, time, and the spatial scale of study. The rapidly advancing technologies of wildlife biotelemetry address these limitations by allowing the collection of data over great spatial scales on a wide range of animal taxa, including small mammals, birds, bats, and even insects, allowing greater insight into animal habitat preferences, short- and long-term movements, activities, and migratory routes (U.S. Geological Survey 1997). These technologies include sensors attached to animals that record, store and/or transmit information on animal location, movement, and physiology (heart rate, body temperature, etc.). Researchers can access the data in diﬀerent ways: sensors can store data ‘‘on board’’ for periodic retrieval, transmitters can send data to ground-based receivers, or they can transmit data to a satellite from which the data can be downloaded anywhere in the world. The development and miniaturization of global positioning system (GPS) devices to under 1 kg has revolutionized researchers’

Sustainability in Arid Grasslands

255

abilities to collect data on locations of large animals (i.e., elk, caribou, bear) via satellite technology (U.S. Geological Survey 1997). The range of taxa for which the full gamut of telemetric technologies has been used includes animals of land, sea, and air, and spans an animal size range from passerine birds to whales (Lewis 1994, Marzluﬀ et al. 1994, King et al. 1995, Parker et al. 1996, Reading and Davies 1996, Sone and Kohno 1996, Shivik et al. 1996, Waller and Mace 1997, Adams et al. 1998, Lariviere and Messier 1998). Other telemetric innovations include harmonic radar, which can record the positions over multiple time steps of ﬂying insects mounted with metallic tags that weigh less than 1 mg. This has special importance in monitoring the behavior of pollinators or outbreaks of pest species (Riley et al. 1996, 1998, Gewin 2002). In addition, new technologies using GPS, GIS, and electrical pulse stimulation to control livestock movements are being deployed in the southwestern United States. ‘‘Virtual fences’’ are being mapped into GIS data layers and loaded into microprocessors attached to livestock equipped with miniature GPS units; when an animal approaches a ‘‘virtual GIS fence’’ (e.g., delineating a property line, riparian zone, etc.), an electric shock of increasing magnitude is generated by a device attached to the animal’s neck, causing it to move away from the ‘‘fence’’ (Anderson et al. 2003, Rango et al. in press). The GIS maps can be changed and updated via radio transmissions, allowing livestock movements to be ‘‘controlled’’ remotely.

Acquiring Data on Soils Researchers and managers alike require soil data on: (1) soil structures, textures, and chemistries; (2) the relationships between soil temperatures, soil moisture and evapotranspiration, and the distribution of soil patches across landscapes; (3) the eﬀect of grazing intensity on soil characteristics and soil microbial communities; (4) the eﬀect of grazing on erosion rates. Characterization of soil organic matter, soil texture, structure, and chemistry (especially salinity and nutrient availability) currently involves laboratory analysis or laborious (and sometimes inaccurate) ﬁeld techniques. In addition, laboratory analysis entails sample collection, storage, and transportation, all of which can introduce errors and remove site-speciﬁc contextual information from the analysis process. Optical remote sensing is now used extensively to measure gully formation and other eﬀects of erosion, surface albedo, barren areas, waterlogged soils, nutrient availability, and water-holding capacity. Synthetic aperture radar (SAR) is used for mapping soil moisture and other soil and near-subsurface geological and soil characteristics (Forster 1996) and burgeoning GIS technologies allow the organization and analysis of multiple layers of data for any given region (Anderson 1996). To assist in on-site analyses, new and developing microtechnologies oﬀer novel approaches for chemical detection and analysis of soils, water, and air.

256

Biodiversity, Conservation, and Management

These include ‘‘electronic noses,’’ ‘‘electronic tongues,’’ and microsensor systems on a computer chip. ‘‘Electronic noses’’ quantify chemical vapors using mathematical tools such as neural networks and principal component analysis (PCA). ‘‘Electronic noses’’ detect and quantify mixtures of complex gases and use surface acoustic wave (SAW) sensors to provide vapor chemical signatures for a given sample. Currently used by the food and chemical manufacturing industries for quality control, electronic nose technology could be used for measuring gases emitted from soils, waters, and organisms (Staples 2000). Currently, ‘‘electronic noses’’ are unable to detect compounds with low molecular weights such as CO2 and O2. ‘‘Electronic tongues’’ detect chemicals in liquid rather than in vapor phase, and are able to detect hundreds of organic and inorganic chemicals simultaneously. One version of an ‘‘electronic tongue’’ is based upon microwells spaced across a ﬂexible strip in which a speciﬁc chemical sensor is placed (Lavigne 1998). When a speciﬁc chemical is detected by the sensor, a reaction occurs, changing the color of that particular well. The color change, detected by a ‘‘camera on a chip,’’ represents the presence of a speciﬁc chemical, and the color intensity is proportional to the concentration of the speciﬁc chemical detected. An ‘‘electronic tongue’’ chemical test-strip may contain over 100 wells, each designed to test for a speciﬁc chemical. Electronic tongues have been used for analyzing soils and water quality. ‘‘Systems-on-a-chip’’ are sensitive analytical devices developed that utilize microfabricated substrates. These small devices, originally developed for rapid chemical and biological weapons detection, provide fast wet and dry chemical analysis. Microtechnologies make such systems-on-a chip possible, including preconcentrators and microscale-gas chromatography (GC) columns, which separate gas-phase particles according to size. Surface acoustic wave sensors are also used to identify and quantify chemicals in the concentrated sample. Microscale high-performance liquid chromatography (HPLC) systems take advantage of microﬂuidic technologies. Examples of systems-ona-chip technology include the ‘‘Chem-Lab-on-a-Chip’’ developed by Sandia National Laboratory, and ‘‘HPLC-on-a-Chip’’ developed by Eksigent Technologies LLC (http://www.eksigent.com/) (Ho et al. 2001). These palm-size instruments can rapidly detect biological and chemical substances in gas or liquid phase, and may eventually provide rapid DNA and protein analysis in situ. The small size, and detection speed, sensitivity, and reproducibility of these technologies make them especially useful. Electronic noses, electronic tongues, and systems on a chip could be adapted for assessing soil and water nutrient availability, soil organic matter, soil chemistry, soil microorganism DNA, and other ecological data needs for assessing the eﬀects of grazing on biodiversity in drylands. Such in situ soil characterization could also be an important ﬁrst step in developing signatures for remote sensing. Some researchers have even suggested the use of mobile robotic technology for ecological data collection, just as NASA’s Mars Rover roamed the surface of Mars collecting, analyzing, and transmitting geological data.

Sustainability in Arid Grasslands

257

Acquiring Data on Climate Climate has an enormous impact on ecosystem function, biodiversity, and sustainability in arid regions. In addition to historical records on climate, data are required for: (1) the amount, frequency and spatial scale of precipitation events; (2) the correlation between net primary production and precipitation in local and regional landscapes; (3) solar radiation levels; (4) wind speed and direction. Meteorological data collection is already widely automated by meteorological stations with radio transmitters to local receivers or to satellite uplinks. Stations commonly monitor precipitation amounts, maximum, minimum, and mean temperatures, relative humidity, vapor pressure, barometric pressure, wind speeds and directions, and solar radiation, and some are outﬁtted to monitor gamma radiation as well. However, meteorological station coverage is still sparse, with less than 3000 permanent meteorological data collection sites in the United States, a density of less than one station per 10,000 km2 (Goodin and Bonito 2002). Installation of large numbers of meteorological stations may be prohibitively expensive. However, broader spatial climate data can be obtained using networks of microsensor clusters, such as ‘‘sensor webs.’’ A sensor web instrument is a network of intracommunicating, spatially distributed sensor pods. Each sensor web pod contains clusters of oﬀ-the-shelf microsensors that are relatively small and inexpensive. Information gathered at each pod of a sensor web is ‘‘hopped’’ or radioed between, and used by other pods in the network, creating a virtual presence in the environment. A ‘‘portal’’ node transmits data by radio from the sensor web to a receiver and can be accessed by computer. Parameters measured by the various pods in prototype instruments include: soil and air temperature, soil moisture and pH, relative humidity, solar radiation, O2, N2O, CH4, and H2S (Delin 2002). NEXRAD (‘‘NEXt Generation RADar’’) is a nationwide U.S. National Weather Service system of ground-based radar stations that can track the movements, speed, and intensity of precipitation events. NEXRAD data in conjunction with remote lightning locator systems have been adapted to measure the spatial and temporal distributions of summer storms in arid lands. Correlations between the number and locations of lightning strikes and total precipitation amounts per strike can be established, allowing estimates of patchily distributed precipitation over large areas. This information can then be used to estimate the amounts and locations of thunderstorm rainfall events as storms move across the landscape; and by pooling data from longer time periods, site-speciﬁc seasonal, annual, or multiyear totals can be calculated. Similarly, NEXRAD radar images can indicate the spatial distribution of storms, although studies aimed at predicting precipitation amounts are still in the development stage. One of the great advances already made in climate studies is the availability of data at a wide range of spatial and temporal scales through the World Wide Web. The availability of these data through these various networks (described below) should be a model for the distribution of other kinds of ecological data

258

Biodiversity, Conservation, and Management

as well. The eﬃcient integration of all these data on wildlife, vegetation, soils, and climate into timely resource management is the pressing challenge.

New Applications for Technology: A Resource Management Scenario The technologies described above constitute an extensive ‘‘tool kit’’ of both conventional and innovative technologies for measuring and evaluating physical and biological properties of arid ecosystems. How might these technologies be incorporated into an integrated ecosystem monitoring scheme to provide real-time or near-real-time information for management decisions? The following example presents a hypothetical scenario for an ecosystem management scheme that utilizes an array of technological devices and approaches to provide real-time and near-real-time information for management decision-making. For the purpose of this exercise, the primary goal of the hypothetical managers will be to reintroduce a sustainable population of elk (Cervus elaphus) to their historic range in an arid to semiarid grassland habitat, without negatively aﬀecting regional biodiversity and ecosystem structure and function. As a template for our hypothetical scenario, we will use the Sevilleta National Wildlife Refuge in central New Mexico, USA, which is also part of the Long Term Ecological Research Network (LTER). The site is 100,000 ha in size, and encompasses many biome types found in the American Southwest—Chihuahuan Desert shrublands, Great Plains grasslands, Colorado Plateau shrub-steppe, juniper savannas, and pin˜on-juniper woodlands. At present, the Sevilleta NWR is not being grazed by domestic livestock; rather, it is a conservation area for native species protection, research, and environmental education. Prior to the implementation of a management plan for a region such as the Sevilleta, a series of baseline data sets should be acquired and put into a GIS database, including data on topography, drainages, soils, geologic formations, vegetation, the locations of political divisions and property boundaries, dwellings, transportation routes, agricultural ﬁelds, and other anthropogenic features. Examples of such GIS maps made for the Sevilleta are available on the Sevilleta LTER website: http://sev.lternet.edu/data/archive/gis/. Climatic data (both historical and current), as described above, are also needed and can be collected remotely at meteorological stations. At the Sevilleta, in order to account for spatial variation in weather, meteorological data are collected at ten stations spread across the reserve. In addition to the collection of these baseline data sets, managers must calculate estimates of forage production necessary for the nutritional requirements of the reintroduced elk. From these data, the managers could estimate the maximum carrying capacity within each type of vegetation included in the GIS, given varying climate patterns. These estimates require knowledge of trends in NPP relative to grazing pressure and climatic variables; and both

Sustainability in Arid Grasslands

259

historical and baseline data sets will help managers understand develop that knowledge. Estimates of NPP can be derived from various ecosystem parameters, including plant cover or density estimates obtained from aerial or satellite images, or from ground-sampled plots. Landscape-scale NPP can be estimated from NDVI data generated with satellite data, as described above. Values of NPP will vary through space and time: spatially, as one moves across soil types and elevations; and temporally, through seasonal or annual shifts of precipitation and temperature. The manager’s task will be to observe and quantify these changes and adjust the grazing regimes to optimally utilize the available forage. This kind of adaptive management requires (1) a constantly updated database of the spatial distribution of NPP values, and (2) a detailed knowledge of the location of the elk on the landscape. Temporal and spatial changes in NPP can be evaluated with frequent updates from satellite images and can be expected to respond to periods and locations of precipitation, which can be identiﬁed from an array of meteorological stations. In the southwestern United States, summer precipitation often occurs during intense, highly localized, short-duration thunderstorms that move across the landscape with a high degree of unpredictability. Using NEXRAD and lightning locator systems described above, Sevilleta LTER researchers have determined that 36,190 m3 of water (or approximately 1.3 mm of precipitation) falls within a 3 km radius of every cloud-to-ground lightning strike (Gosz et al. 1995). This type of accuracy is essential to establish the relationship between NPP and precipitation. These data can be used to estimate the amounts and locations of thunderstorm rainfall events as storms move across the landscape; and by pooling data from longer time periods, site-speciﬁc seasonal, annual, or multiyear totals can be calculated. Similarly, NEXRAD images can indicate the spatial distribution of storms, although studies aimed at predicting precipitation amounts are still in the development stage. Data on the location and amounts of precipitation can be integrated with remotely sensed satellite data to determine the spatiotemporal patterns of NPP (Oesterheld et al. 1998). By analyzing areas of precipitation and NDVI on a weekly or monthly time step, managers can determine which sections of the land have experienced increases in NPP, and thus may support a higher level of herbivore activity. With such knowledge, managers can make decisions on modifying elk movements by providing water or forage and to address other considerations such as soil conservation, and preservation of biodiversity and/or species of special concern. This task of modifying animal densities and movements across the landscape requires detailed knowledge of the locations, at any point in time, of the animals of interest (whether they are livestock or important wildlife species). The telemetric techniques, such as those described above, have historically been used on wild ungulates such as deer, elk, and moose, but can be adapted to such domesticated animals as cattle, sheep, goats, horses, oxen, camels or llamas. Thus far, in this hypothetical scenario, the managers have used technological applications to establish baseline GIS databases, track the locations and

260

Biodiversity, Conservation, and Management

amounts of precipitation on the landscape, determine levels of plant production, and monitor the movements of elk and other species of concern. From these databases, the managers can plan interventions to ensure habitat preservation and biodiversity levels. For example, grazing by large herds of wildlife or livestock typically is ‘‘uneven’’ on the landscape, and creates areas of severe disturbance (e.g., around waterholes or streamcourses) that may lead to erosional loss of topsoil and a decline in watershed quality; also habitat disturbances in critical areas may have detrimental eﬀects on other native plant and animal species, causing local or regional reductions or shifts in biodiversity. Ungulate activities also may interfere with activities of other native game animals, particularly during critical breeding periods. Application of modern technologies can assist with the management of these problems as well. In the case of overgrazing in local patches, remotely sensed images can be used to estimate plant cover at pixel sizes now approaching 1 m. Estimates of leaf area index (LAI) also may be derived from imagery using mathematical modeling techniques. Soil characteristics, including mineral compositions and moisture levels, can be derived from spectral reﬂectance signatures and soil-penetrating synthetic-aperture radar images. Surface-reﬂecting radar, mounted in either aircraft or satellites, can be used to produce updates to topographic databases such as digital elevation models (Forster 1996), and identify surface erosion patterns, channel changes in perennial streams, or the formation of drainage channels cut into the landscape by the intermittent streams known locally as arroyos. As areas with low vegetation cover or high erosional rates are identiﬁed, management decisions can be made to augment water or forage supplies. Management strategies for preserving biodiversity also can be enhanced with new technological applications. Areas of critical habitats, such as springs or marshes, can be identiﬁed, quantiﬁed, and monitored remotely for size, productivity, water quantity, and water quality. If changes through time are noted, teams can be dispatched for on-site inspection of the problems. In the case of protected/regulated wildlife or predators, the GPS/radiotelemetry instruments described above can be used to measure the movements and activities of individual animals or family groups, and permit assessments of management practices and interactions with other species and/or livestock (which can be monitored simultaneously). Pest species of grasshoppers and locusts can increase markedly in grazed ecosystems, presenting potential pest invasion problems for nearby agricultural regions (Gangwere et al. 1997). Similarly, grazing activities have been associated with changes in the densities of rodent and rabbit species (Bich et al. 1995, Nyakolartey and Baxter 1995, Jones and Longland 1999), which carry diseases (plague, hantavirus, tularemia, rabies) transmissible to humans and livestock (Grenfell and Dobson 1995). As conditions on the landscape become more favorable to populations of these pests and disease vectors, satellite imagery would be able to detect the changing variables in vegetation and moisture, and provide data for predictive models of insect or disease outbreaks. Harmonic radar technology can be used to track swarms of locusts across the landscape (Riley et al. 1998), and satellite data have been

Sustainability in Arid Grasslands

261

used to construct GIS models of patterns of hantavirus infection in humans (Glass et al. 2000), and grasshopper outbreaks in rangelands (Burt et al. 1995). Presumably, these tools are equally valuable for monitoring nonpest invertebrate populations of ecological signiﬁcance.

Data Sharing and Communications Networks Various data sharing and communications networks that could contribute to land-use management scenarios like the one described above are already in place and operating. The Argos Data Collection and Location Satellite System is operated under a joint partnership agreement with the U.S. National Oceanographic and Atmospheric Administration (NOAA) and the French Centre National d’Etudes Spatiales (CNES) to provide worldwide data collection on radio-tagged animals. The United States Geological Survey (USGS) Geostationary Operational Environmental Satellite (GOES) system downloads to the World Wide Web hydrology and water quality data in near real time from stations all over the United States. The U.N. Environmental Programme’s Global Environment Monitoring System (GEMS) assembles global water quality data, and the Global Climate Observing System (GCOS), sponsored by the World Meteorological Organization (WMO), two U.N. organizations and the International Council for Science (ICSU) assembles data on long-term precipitation, atmospheric temperature, sea surface temperature, El Nin˜o–Southern Oscillation (ENSO) data, and snow and ice data. Also, the World Data Center for Greenhouse Gases posts past and current concentrations of individual greenhouse gases from hundreds of locations globally at its website. A system that integrates all the technologies above, and others, is the Gap Analysis Program (GAP). This is a nationwide program in the United States administered by the Biological Resources Division of the USGS. It uses GIS data on abundance and distribution of vertebrate species overlaid on data for land cover types to aid in land-use and endangered species management decisions. Lastly, at least one commercial company, Agri ImaGIS, provides processed satellite imagery to agriculturalists for large-scale analysis of farmland topography, yield variability and soil analysis (http://home.netscape.com/ misc/snf/popup_aol4.html); and the Boeing Commercial Space Company has plans to launch a satellite system called Resource21, which will oﬀer similar kinds of commercial satellite products for agricultural applications (http://www.boeing.com/). Global Terrestrial Observing System A pioneering international eﬀort to create a global ‘‘network of networks’’ for the applications of technology to large-scale land-use questions can be found in the Global Terrestrial Observing System (GTOS). This was established in

262

Biodiversity, Conservation, and Management

1996 by ﬁve organizations, including the World Meteorological Organization (WMO), the International Council of Scientiﬁc Unions (ICSU), and three from the United Nations, as well as participation from ﬁve international space agencies. The system is intended to provide data for detecting, quantifying, locating, and giving early warning of changes in the capacity of the earth to sustain development and improvements in human welfare. The GTOS focuses on changes in land quality, freshwater resources availability, pollution and toxicity, loss of biodiversity, and climate change. The U.S. Long Term Ecological Research network (LTER), one of the participating scientiﬁc networks, has established an aﬃliated international network (ILTER) that can help scientiﬁc organizations around the world establish ground-based ecological monitoring systems that can be coordinated with remote sensing systems accessed through GTOS. The GTOS program facilitates access to information on terrestrial ecosystems so that researchers and policy makers can detect and manage global and regional environmental change. It maintains a Terrestrial Ecosystem Monitoring database known as TEMS—an international directory of sites and networks that carry out long-term terrestrial monitoring and research activities. The GTOS also initiates observation activities on primary productivity, terrestrial carbon, land dynamics, terrestrial coastal environments, and climate. The program accomplishes these activities by maintaining linkages between sites and terrestrial networks.

Conclusion Any attempt to enhance our understanding of ecosystem function, the delivery of ecosystem services, and the impacts made upon ecosystems by humans must be able to accurately identify and measure, over time, the most inﬂuential biotic and abiotic processes in the system. The ability to collect and analyze data on large spatial and temporal scales on these subjects becomes more crucial as human population and consumption increases, and as human needs for the products of ecosystem services increase. Historically, ecosystem studies have been labor- and time-intensive, such that a large amount of time and energy invested by researchers yielded a relatively small amount of useful data. We believe that new applications of the technological advances of the last half-century, used in conjunction with conventional techniques, can signiﬁcantly increase the quality and quantity of data that can be collected and analyzed in ecosystem studies. With increasing emphasis on the use of these technologies in ecological science, students will need education and training on the science and applications of the technologies. In addition, funding organizations should develop plans for integrating these technologies, and others, into ecological research. These plans would guide increased budgets for technological research and development and for incorporating technology into ongoing eﬀorts. The high cost of landscape-scale data collection and regional networking will

Sustainability in Arid Grasslands

263

present one of the greatest challenges for integrating those technologies into day-to-day management routines; any research that will show ways to streamline systems and reduce those costs will be extremely valuable.

Acknowledgments This project was completed with funding from the Cooperative Monitoring Center at Sandia National Laboratories and the University of New Mexico’s Sevilleta Long Term Ecological Research Program (NSF Grant DEB0080529). Sandia National Laboratories consist of multiprogram laboratories operated by Sandia Corporation, a Lockheed Martin Company, for the United States Department of Energy under contract DE-AC04-94AL85000. We wish to thank our colleagues David Betsill, Sally Laundre-Woerner, Gaurav Rajen, Reynolds Salerno, Moshe Shachak, Michael Vannoni, Bob Waide, Andy Wilby, and Patty Sprott for their assistance in this project.

References Adams, N.S., D.W. Rondorf, S.D. Evans, and J.E. Kelly. 1998. Eﬀects of surgically and gastrically implanted radio transmitters on growth and feeding behavior of juvenile chinook salmon. Transactions of the American Fisheries Society 127(1): 128–136. Anderson, D.M., C.S. Hale, R. Libeau, and B. Nolen, 2003. Managing stocking density in real time. Pp. 840–843 in Proceedings of the 7th International Rangeland Congress, H. Allsopp, A.R. Palmer, S.J. Milton, K.P. Kirkman, G.I.H. Kerley, C.R. Hurt, and C.J. Brown (eds.). Anderson, G.L. 1996. The application of spatial technologies for rangeland research and management. Geocarto International 11(3): 5–11. Belsky, A.J., A. Matzke, and S. Uselman. 1999. Survey of livestock inﬂuences on stream and riparian ecosystems in the western United States. Journal of Soil and Water Conservation 54(1): 419–431. Bich, B.S., J.L. Butler, and C.A. Schmidt. 1995. Eﬀects of diﬀerential livestock use on key plant species and rodent populations within selected Oryzopsis hymenoidesHilaria jamesii communities of Glen Canyon National Recreation Area. Southwestern Naturalist 40(3): 281–287. Burt, P.J.A., J. Colvin, and S.M. Smith. 1995. Remote-sensing of rainfall by satellite as an aid to Oedaleus senegalensis (Orthoptera, Acrididae) control in the Sahel. Bulletin of Entomological Research 85(4): 455–462. Delin, K.A. 2002. The Sensor Web: A Macro-Instrument for Coordinated Sensing. Sensors 2: 270–285. Fleischner, T.L. 1994. Ecological costs of livestock grazing in Western North America. Conservation Biology 8(3): 629–644. Forster, B. 1996. Current advances in radar remote sensing research and its application in South East Asia. Geocarto International 11(4): 3–10. Gangwere, S.K., M.C. Muralirangan, and M. Muralirangan. 1997. The Bionomics of Grasshoppers, Katydids and Their Kin. New York: CAB International. Gewin, V. 2002. Radio tracks crickets. Nature. Available online at http://www. nature.com/nsu/020603/020603–21.html. Glass, G.E., J.E. Cheek, J.A. Patz, T.M. Shields, T.J. Doyle, D.A. Thoroughman, D.K. Hunt, R.E. Enscove, K.L. Gage, C. Irland, C.J. Peters, and R. Bryan. 2000.

264

Biodiversity, Conservation, and Management

Using remotely sensed data to identify areas at risk for hantavirus pulmonary syndrome. Emerging Infectious Diseases 6(3): 238–247. Goodin, D., and G. Bonito. 2002. Sensor networks: Emerging sensor technologies. Pp. 17–21 in Scalable Information Networks for the Environment (SINE), A. Withey, W. Michener, and P. Tooby (eds.). San Diego: San Diego Supercomputer Center. Gosz, J.R., D.I. Moore, G.A. Shore, H.D. Grover, W. Rison, and C. Rison. 1995. Lightning estimates of precipitation location and quantity on the Sevilleta LTER, New Mexico. Ecological Applications 5(4): 1141–1150. Grenfell, B.T., and A.P. Dobson. 1995. Ecology of Infectious Diseases in Natural Populations. New York: Cambridge University Press. Hiernaux, P., C.L. Bielders, C. Valentin, A. Bationo, and S. Fernandez-Rivera. 1999. Eﬀects of livestock grazing on physical and chemical properties of sandy soils in Sahelian rangelands. Journal of Arid Environments 41(3): 231–245. Ho, C.K., M. Kelley, M.T. Itamura, and R.C. Hughes. 2001. Review of Chemical Sensors for In-Situ Monitoring of Volatile Contaminants. Albuquerque, NM: Sandia National Laboratories. Jones, A.L., and W.S. Longland. 1999. Eﬀects of cattle grazing on salt desert rodent communities. American Midland Naturalist 141(1): 1–11. Karnieli, A. 1997. Development and implementation of spectral crust index over dune sands. International Journal of Remote Sensing 18(6): 1207–1220. King, D.T., J.F. Glahn, and K.J. Andrews. 1995. Daily activity budgets and movements of winter roosting double-crested cormorants determined by biotelemetry in the Delta Region of Mississippi. Colonial Waterbirds 18(S11): 152–157. Lariviere, S., and F. Messier. 1998. The inﬂuence of close-range radio-tracking on the behavior of free-ranging striped skunks, Mephitis mephitis. Canadian Field Naturalist 112(4): 657–660. Lavigne J.J., S. Savoy, M.B. Clevenger, J.E. Ritchie, B. McDoniel, S.-J. Yoo, E.V. Anslyn, O.T. McDevitt, J.B. Shear, and D. Neikirk. 1998. Solution-based analysis of multiple analytes by a sensor array: toward the development of an electronic tongue. Journal of the American Chemical Society 120: 6429–6430. Lewis, S.E. 1994. Night roosting ecology of pallid bats (Antrozous pallidus) in Oregon. American Midland Naturalist 132(2): 219–226. Marzluﬀ, J.M., M.S. Vekasy, and C. Coody. 1994. Comparative accuracy of aerial and ground telemetry locations of foraging raptors. Condor 96(2): 447–454. Memmott, K.L., V.J. Anderson, and S.B. Monsen. 1998. Seasonal grazing impact on cryptogamic crusts in a cold desert ecosystem. Journal of Range Management 51(5): 547–550. Milchunas, D.G., and W.K. Lauenroth. 1993. Quantitative eﬀects of grazing on vegetation and soils over a global range of environments. Ecological Monographs 63(4): 327–366. Muldavin, E.H., P. Neville, and G. Harper. 2001. Indices of grassland biodiversity in the Chihuahuan desert ecoregion derived from remote sensing. Conservation Biology 15(4): 844–855. Nyako-Lartey, Q., and R.M. Baxter. 1995. The eﬀects of diﬀerent grazing regimes on the population dynamics of small mammals in the eastern cape. Transactions of the Royal Society of South Africa 50(2): 143–151. Oesterheld, M., C.M. DiBella, and H. Kerdiles. 1998. Relation between NOAAAVHRR satellite data and stocking rate of rangelands. Ecological Applications 8(1): 207–212.

Sustainability in Arid Grasslands

265

Parker, N., A. Pascoe, H. Moller, and R. Maloney. 1996. Inaccuracy of a radiotracking system for small mammals: The eﬀect of electromagnetic interference. Journal of Zoology 239(2): 401–406. Rango, A., D.M. Anderson, C.S. Hale, and K.M. Havstad. 2004. A developing method for Directional Virtual Fencing (DVFTM) and real time range management using remote sensing, GIS and GPS. Proceedings of the 30th International Symposium on Remote Sensing of the Environment (in press). Rao, B.R., R.S. Dwivedi, L. Venkataratnam, and A.N. Singh. 1996. Monitoring saltaﬀected soils using remote sensing data. Geocarto International 11(4): 41–46. Reading, C.J., and J.L. Davies. 1996. Predation by grass snakes (Natrix natrix) at a site in southern England. Journal of Zoology 239(1): 73–82. Reed, B.C., T.R. Loveland, and L.L. Tieszen. 1996. An approach for using AVHRR data to monitor U.S. Great Plains grasslands. Geocarto International 11(3): 13–22. Riley, J.R., A.D. Smith, D.R. Reynolds, A. S. Edwards, J.L. Osborne, I.H. Williams, N.L. Carreck, and G.M. Poppy. 1996. Tracking bees with harmonic radar. Nature 379(6560): 29–30. Riley, J.R., P. Valeur, A.D. Smith, D.R. Reynolds, G.M. Poppy, and C. Lofstedt. 1998. Harmonic radar as a means of tracking the pheromone-ﬁnding and pheromone-following ﬂight of male moths. Journal of Insect Behavior 11(2): 287–296. Rohner, C., and D. Ward. 1999. Large mammalian herbivores and the conservation of arid Acacia stands in the Middle East. Conservation Biology 13(5): 1162–1171. Shi, L. 2002. DNA microarray (genome chip): monitoring the genome on a chip. Available online at http://www.gene-chips.com. Shivik, J.A., M.M. Jaeger, and R.H. Barrett. 1996. Coyote movements in relation to the spatial distribution of sheep. Journal of Wildlife Management 60(2): 422–430. Sone, K., and A. Kohno. 1996. Application of radiotelemetry to the survey of acorn dispersal by Apodemus mice. Ecological Research 11(2): 187–192. Staples, E.J. 2000. The ﬁrst qualitatively validated electronic nose for environmental testing of air, water, and soil. In Conference Proceedings, American Chemical Society, San Francisco, 26–30 March 2000. Available online at http://www. estcal.com/TechPapers/ACS2000C.pdf. Todd, S.W., R.M. Hoﬀer, and D.G. Milchunas. 1998. Biomass estimation on grazed and ungrazed rangelands using spectral indices. International Journal of Remote Sensing 19(3): 427–438. U.S. Geological Survey. 1997. Forum on Wildlife Telemetry: Innovations, Evaluations, and Research Needs; 21–23 September 1997, Snowmass Village, Colorado. Program and Abstracts. U.S. Geological Survey and The Wildlife Society. Vande Castle, J. 1998. Field studies of large mobile organisms: scale, movement, and habitat utilization. Pp. 271–298. In Ecological scale: theory and applications, D.L. Peterson and V.T. Parker (eds.). New York: Columbia University Press. Waller, J.S., and R.D. Mace. 1997. Grizzly bear habitat selection in the Swan Mountains, Montana. Journal of Wildlife Management 61(4): 1032–1039. Zhou, Q. 1998. Use of GIS technology for land resource inventories and modelling for sustainable regional development. Ambio 27(6): 444–450.

16 Reconciliation Ecology and the Future of Species Diversity Michael L. Rosenzweig

A

lexander von Humboldt (1807) provided the ﬁrst hint of one of ecology’s most pervasive rules: larger areas contain more species than do small ones. Many ecologists see that rule—the species–area relationship—as one of ecology’s very few general laws (e.g., Lawton 1999, Rosenzweig and Ziv 1999). In the past two centuries, ecologists have learned a lot about species–area relationships. I will explore that knowledge and show that we can already use it in the struggle to minimize extinction losses. It teaches us what proportion of diversity is truly threatened and how to prevent most losses by applying a new strategy of conservation biology.

Species–Area Equations Olaf Arrhenius (1921) and Frank Preston (1960) formalized the species–area pattern by ﬁtting it with a power equation: S ¼ Caz

ð1Þ

where S is the number of species, A is the area, and C and z are constants. For convenience, ecologists generally employ the logarithmic form of this equation: log S ¼ c þ z log A

ð2Þ

where c ¼ log C. (Note that I do not use the jargon term ‘‘species richness.’’ To understand why, see Rosenzweig et al. 2003.) The species–area power 266

Reconciliation Ecology and the Future of Species Diversity

267

equation, or SPAR, can be ﬁtted to an immense amount of data (Rosenzweig 1995). Ecologists are not sure why a power equation ﬁts islands or continents. But we do have a successful mathematical theory for areas within a province. Brian McGill (personal communication) has deduced the species–area curve within provinces from four assumptions: The geographical range of each species is independently located with respect to all others. Species vary in abundance with respect to each other. Species have a minimum abundance. Each species’ abundance varies signiﬁcantly across its own range, being relatively scarce more often than relatively common. (‘‘Relatively’’ means with respect to its own average abundance.)

Data support all four assumptions. From them, McGill shows that there is a species–area curve and that it approximates a power equation whose z-value ranges between 0.05 and 0.25 with a mean of about 0.15. McGill’s theory supplants those of many others (e.g., Leitner and Rosenzweig 1997, May 1975, Preston 1962) because of its generality and realism, and because it embeds SPARs in a set of seven macroecological patterns, four of which it predicts. Other ingenious attempts at a SPAR theory exist, but are weakened by problems outside the scope of this chapter (Durrett and Levin 1996, Harte et al. 1999, Hubbell 2001, Wissel and Maier 1992). One frequent problem: the assumption of a single habitat type for all places. No model can possibly account for a pattern known to be caused by a variable that the model does not use. (This is true no matter how well such a model might ﬁt the data.) Yet we certainly know that the species–area relationship within a province depends on the inclusion of more habitats in larger areas (Williams 1964). Several of the theories, including that of McGill, show that the power equation is merely a good ﬁt rather than an exact description of the species–area curve. Since power equations are very plastic curves and ﬁt a host of monotonic relationships, one should not be too surprised.

Scales of the Species–Area Curve Four diﬀerent processes determine S as a function of area (Rosenzweig 1995). Each of these produces SPARs with z-values in a restricted range, and the ranges abut, covering virtually the entire unit interval. Let us review all four SPAR puzzles and the processes that produce them. Sample-Size SPARs Determining the number of species in an area requires sampling. Sampling comes with a bias, that is, the larger the number of individuals identiﬁed, the

268

Biodiversity, Conservation, and Management

greater the number of species in the sample. Usually, more individuals will be identiﬁed from a larger area than a smaller one. A SPAR generated by statistical sampling artifacts holds no biological interest. Fisher et al. (1943) showed us one powerful way to recognize such SPARs and eliminate them from further consideration. Fisher devised a statistic called Fisher’s a that is almost insensitive to sample size but does vary with number of species, and S. Burnham and Overton (1979) and Lee and Chao (1994) have introduced other successful bias-reducing statistics. These have been made available in modern, free software packages (WS2M at http://eebweb.arizona.edu/diversity, and EstimateS at http://viceroy.eeb. uconn.edu/EstimateS). Sample-size SPARs tend to have very small z-values. All those I have encountered have z 0:12. Sample-Area SPARs All species have habitat requirements that restrict them in space. A larger area will tend to include more habitat types than a smaller one (Williams 1943). Thus SPARs emerge from diﬀerent-size samples within the same biological region. They tend to have z-values between 0.1 and 0.2. Archipelagic SPARs The islands of an archipelago combine the habitat-sampling process with another process. Their SPARs have z-values ranging from 0.25 to 0.55. I shall review this other process, but ﬁrst I shall mention the fourth type of SPAR. Interprovincial SPARs Biogeographical provinces of similar environment, such as wet tropical forests, have diversities in proportion to their areas. A biogeographical province is a region whose species have evolved within it, rather than immigrating from somewhere else. Although the concept is merely an ideal—every place has at least a few species that arrived as immigrants—it is very close to true in many real places, such as diﬀerent continents or well-separated periods in the history of life. The z-values of interprovincial SPARs begin at 0.6 and range upwards, with most about 0.9 to 1. Interprovincial z-values will turn out to be the most useful to conservation biologists because they predict diversity far into the future.

What Makes z-Values Different? To understand the three SPARs of ﬁg. 16.1 (and their diﬀerences), we need to connect the state variable we call species diversity to its derivatives.

Reconciliation Ecology and the Future of Species Diversity

269

Figure 16.1 Three scales of species–area curves taken from bird censuses. The steepest curve connects points from diﬀerent biogeographical provinces. The least steep curve connects points from diﬀerent-sized samples within the same province. The archipelagic curve lies in between.

Recognizing that, and beginning the journey to accomplish it, was one of the greatest achievements of Robert H. MacArthur. MacArthur teamed with Edward O. Wilson to apply dynamic analysis to the problem of island diversities (MacArthur and Wilson 1967). They carefully deﬁned the two derivatives that should matter most: the rate at which species not on an island arrive on it; the rate at which species on an island become extinct there.

Then, in search of a self-regulating system, they asked how these rates should vary with diversity itself. Their result was powerful and robust: an island with no species can suﬀer no extinctions. An island with all the species of its mainland source pool can experience no further immigrations. In contrast, immigrants must be arriving at some positive rate on an island with no species. And an island with as many species as possible must suﬀer extinctions at some positive rate. Thus there has to be at least one intermediate diversity at which the two rates neutralize each other. There has to be at least one steady-state diversity. But MacArthur and Wilson went beyond a demonstration of self-regulating diversity on islands. Their theory also predicts the existence of archipelagic SPARs: larger islands should tend to contain more habitats and, at any particular S, larger populations of species. Those inﬂuences would depress the extinction rate curve of a large island compared with a smaller one. The result: the steady state should increase with island size. The theory of island biogeography does not predict the shape or the zvalue of archipelagic SPARs. However, it does predict that their z-values should vary. At any particular diversity, islands farther from the source of

270

Biodiversity, Conservation, and Management

colonization should receive immigrant species at a reduced rate compared with nearby islands. So two islands at diﬀerent distances, but of the same area, will have diﬀerent steady states. The farther island will have the lower S. This is mathematically equivalent to a higher z-value. The theory that predicts interprovincial SPARs also depends on dynamics (Rosenzweig 1975). In fact, its extinction curves are the same shape as those from island biogeography. This is not so with the curves describing the rate at which it receives new species. These diﬀer considerably from island curves. Recall the deﬁnition of provinces: provinces are areas whose species originate by speciation from within. Hence, provincial dynamics depend on speciation rates and extinction rates, rather than immigration rates and extinction rates. The cumulative diﬀerence between the creative process of speciation and the destructive process of extinction determines the number of species alive in a biological province. But existing species are the nurseries for new species, so speciation rate should rise as diversity does. Hence, the slope of the speciation rate curve for a province should be positive, whereas it is negative for an island. Do speciation rate curves with positive slopes prevent steady states in provinces? No. Provincial steady states emerge once one considers the biogeographical ranges of species and how those ranges respond to diversity. As diversity rises in a province, competition and predation tend to restrict individual species to smaller geographical ranges. Smaller geographical ranges decrease speciation rate (for a variety of reasons). So, as diversity grows, the speciation rate of the average species declines. That decline imposes a negative second derivative on the curve of total speciation rate. Meanwhile, as diversity grows, the total rate of extinction accelerates. The two curves intersect, producing a provincial steady-state diversity. In sum, even with no sample-size bias, the ecologist expects to see three types of SPAR. But why should they have dissimilar z-values? Why should SPARs representing samples of one province have the gentlest slopes, those between provinces have the steepest, and those of archipelagos have slopes that ﬁll in between the other two? Again we turn to rates for our answer. A piece of a province will contain those species which the habitats of that piece can sustain. But it will have other species, too. The piece will be good enough to support individuals of these other species. It may even be suﬃcient to support some reproduction by them. But it will not be suﬃcient for those species to maintain their populations. What then keeps them in the piece? Their dispersal into it. If their dispersal rate plus the rate at which they reproduce is (together) suﬃcient to counterbalance their death rate, they will occur in the piece again and again. Inspired by desert plants, Shmida and Ellner (1984) recognized and gave a special name to populations of species that require dispersal contributions from elsewhere. They called them sink populations. In contrast, they called sustainable populations source populations. A piece of a province has both source and sink populations, which, by extension of the metaphor, I call

Reconciliation Ecology and the Future of Species Diversity

271

‘‘source and sink species.’’ Shmida and Ellner called the extra species conferred by sink populations, the ‘‘mass eﬀect.’’ Now we move upscale to islands. We take a great, imaginary blade and cut the piece of mainland free of its provincial moorings. We set it adrift in the sea. Its source species will remain, but its sink species will not. The high regular dispersal rates that they require for their maintenance will have been replaced by much lower rates of origination by rare colonization events. So the island will have fewer species than the provincial piece. A line connecting it to the area–diversity point of the entire province will be steeper than one connecting the piece to the province. Consequently, the z-value of the island will be larger than that of the piece. As time unfolds, the source species of the island will suﬀer occasional extinctions. During most years those species will have healthy reproduction, but not during every year. All species encounter stochastic disasters now and again. The island’s diversity therefore tends to decay. But accidents are rare, and the extinctions will be counterbalanced by immigrations. Thus the island will reach its steady state S. As we push our imaginary island farther from the province, immigrations will occur less and less frequently. The rate of ﬂux of species on the island will slow down to reﬂect the lower immigration rate. The steady state declines. The line connecting the island to the area–diversity point of the entire province will grow steeper, sending its z-value higher. If we push the island far enough away, immigrations will occur so rarely that speciation rates will match them. Our island will change into a new small province with a very steep z. So, our attempts at theory have met with considerable success. Theory predicts the existence of four distinct SPARs, and it identiﬁes the processes controlling each one. It tells us that SPARs within a province should and do ﬁt power equations, albeit imprecisely. It predicts that the exponent of those equations should hover around 0.15. It also predicts that the slopes of island SPARs should exceed those within a province, and those between provinces should exceed those of islands. True, we cannot yet predict the exponents (z-values) of island or interprovincial SPARs. But meanwhile, data tell us that the latter range from 0.6 to 1.0 and the former from 0.25 to 0.55, and that is enough for practical purposes.

Steady-State Diversities in the Fossil Record Benton (1995) and some others claim that steady-state theories are irrelevant to diversity because diversity has risen fairly steadily throughout the Phanerozoic Eon (the last 550 my or so). He ﬁts a single explosive exponential equation to animal diversities throughout the Phanerozoic. But mounting evidence now suggests that diversity has been near a steady state during most of that time. First, the deviations from Benton’s equation are heroic. Moreover, the data themselves are suspect on two grounds. They report generic or familial

272

Biodiversity, Conservation, and Management

diversities, not species diversities. We know little about the processes that lead to the origination or extinction of such higher taxa; we cannot yet build any theory of them; and we have no basis to believe they should behave like species diversities. Moreover, all the data sets that lead to the conclusion of a steadily rising diversity share a troubling deﬁciency. They are uncorrected for sample-size problems. David Raup (1976) warned that, usually, older fossils are scarcer than younger ones, prejudicing us to the conclusion that older times had fewer species. The worst bias comes in the last 65 my, the Cenozoic (or Tertiary). Rocks of the last 65 my are 10 times as abundant as those of 200, 300, or 400 my ago. And not only do they cover much more of the earth, they are often easier to work with. They tend to be unconsolidated, which means they have not turned into hard rock. So, their fossils can be recovered simply by washing away the rock matrix in water. With considerably less paleontological eﬀort, they produce far more fossils of much better quality than consolidated rock. In my view, it is not a coincidence that most of the explosiveness of Benton’s equation derives from the apparent huge increase in diversity during the Cenozoic. Even the raw record of familial or generic diversities during the Phanerozoic does little to support a single exponential equation. In fact, from the raw data, Sepkoski (1978) recognized three diversity eras during this time. The earliest, during the Cambrian of 500 my, had the poorest diversities, but appeared to have reached a steady state. It was replaced by the second, a richer time, which also reached a steady state and lasted until roughly 100 my. The third, beginning in the Cretaceous, is the one in which we now live. Its raw data do seem to indicate a rapid rise in diversity and also seem to show no signs of leveling oﬀ. But, of course, the last 100 my is the time of maximum bias, and recent, very careful studies of that bias (discussed below) suggest that it may be the principal or even the sole source of the apparent rise in diversity during the Cretaceous–Tertiary. Sample-size bias pervades work in diversity. The most successful statistical tools to reduce this problem belong to a family of jackknife methods (Burnham and Overton 1979, Chao and Lee 1992, Chao et al. 1992, Lee and Chao 1994). Paleobiologists are beginning to explore these methods. Meanwhile, many paleobiologists have approached their data sets with older tools such as rarefaction analysis (e.g., Miller and Foote, 1996). That does not eliminate the bias; rather it equalizes bias among samples. So rarefaction can be used in comparing diﬀerent samples, albeit somewhat crudely. More reﬁned examinations of fossil diversities leave no doubt that life has often ﬂuctuated about a steady state. Boucot (1975) showed this to be true of long time periods he called ecological-evolutionary units (EEUs), which persist for tens of millions of years. Brett et al. (1996) have demonstrated the existence of temporal subunits within EEUs that are even more stable. Focusing on the dynamics, I and my students began analyzing a particularly well-structured set of samples from the Nicolet River Valley of Quebec (Bretsky and Bretsky 1976). The strata represent a period of some 5 my at

Reconciliation Ecology and the Future of Species Diversity

273

the end of the Ordovician (440 my ago), and for the latter 3.5 my of this time (at least), species diversity ﬂuctuated about a steady state (ﬁg. 16.2) (Rosenzweig 1995). Subsequently, others have seen steady states within the Cenozoic itself (Allmon et al. 1993, Alroy 1998, Alroy et al. 2001, Nichols and Pollock 1983, Van Valkenburgh and Janis 1993). This is particularly damaging to Benton’s interpretation, because the Cenozoic ought instead to exhibit the sharpest, easiest-to-document increases in diversity.

What Human Impact Will Do to Diversity Man, the most ecologically adaptable of species, can compete with almost all other animals in almost all abiotic milieux. As a consequence, humans have been reducing the area available to most other species. Science and society have to be aware of this reduction because the number of species at diversity’s steady state depends on available area. A number of estimates of this reduction exist. Vitousek et al. (1997) say 40% to 50% of the ice-free, terrestrial surface has been degraded for wild species by human use. Norman Myers and his colleagues estimate the degradation of speciﬁc habitats: 75% of the forests (which once covered some 40% of the world’s terrestrial surface) (Myers 1999); 88% of the world’s most diverse habitats (Myers et al. 2000). Certainly the loss of temperate grasslands to wildlife is now close to 100%. And Huston (1993) estimates the average loss of ice-free, terrestrial surface area at 95%. Marine environments have suﬀered much the same fate (Jackson 2001).

Figure 16.2 An example of a steady state in species diversity from the fossil record. Species come from the latest Upper Ordovician muddy benthos of the Nicolet River Valley, Quebec (ca. 440 my ago). Time is indicated in meters above the oldest strata. The jackknife estimator reduced the sample size bias of the raw diversities. (Adapted from Rosenzweig 1995.)

274

Biodiversity, Conservation, and Management

Faced with the continuing loss and degradation of natural habitat, society battles to save diversity by setting aside some natural areas. Because extinction is a process that often requires many generations, this strategy has helped so far. But area constitutes a basic inherent property of every biome, a property crucial to the dynamical functioning of its components. So it is an oxymoron to imagine a pristine biome that retains only 2% or 5% or even 10% of its original size. Instead, because of the severe loss of natural habitat, ecologists predict a new mass extinction on a scale that has not visited the earth for 65 my. The optimists among us base their quantitative prediction on the archipelagic z-value (e.g., Pimm et al. 1995). The shrunken natural part of the earth, they say, has become an island. It will maintain species only according to the equation: S ¼ A 0:3

ð3Þ

where S and A are the proportions of diversity and natural area that will remain. (Notice that C, the constant of eq. (1), equals unity in eq. 3 because, in eq. (3), S and A are proportions.) According to eq. (3), 5% of the area will sustain about 41% of species diversity. As diversity relaxes to satisfy the island equation, the ﬁrst species to go will be the endemics—those species whose habitat gets entirely expropriated (Harte and Kinzig 1997). These extinctions will be deterministic and virtually instantaneous. Following them will be the sink species—those that get restricted to marginal habitats, that is, habitats in which their death rates exceed their birth rates. These extinctions will also be deterministic in the sense that we should be able to point out the victims unambiguously by separating sink from source species (Patterson, 1990; Patterson and Atmar 2000). Conservation’s two strategies of reservation and restoration have stayed the imposition of the island equation, and may even have reduced its severity (ﬁg. 16.3). Instead of saving natural areas randomly, we diversiﬁed what we saved, deliberately focusing on preserving or restoring the habitats most likely to vanish entirely. Hence, we considerably reduced the likelihood that any species would lose everything, even including its sink habitats. That transferred some extinction of endemics to the category of extinction of sink species, slowing down the course of mass extinction because some individuals of sink species not only survive, they also reproduce. Thus the deterministic extinction of a sink species takes more time than the instantaneous extinction of an endemic. Nevertheless, diversity’s decline will not cease after our small, new natural world reaches levels of island-like diversity. The world of nature reserves is not an island but a shrunken province. Its source pool is the past. Species that become extinct in it cannot immigrate from the past to recolonize the world of the future. So, like any evolutionarily independent province, our miniaturized natural world must seek its future steady state along the interprovincial SPAR, not the island SPAR.

Reconciliation Ecology and the Future of Species Diversity

275

Figure 16.3 Conservation’s two dominant strategies, reservation and restoration, view the world as divided into two sorts of areas: natural set-asides and places ruined by the activities of people. Reservation prevents further areas from becoming degraded. Restoration returns areas to the high quality pool.

The z-value of interprovincial SPARs is approximately unity, so its governing equation is approximately S¼A

ð4Þ

Thus our losses of species should be approximately linear. Lose 10% of the natural world’s surface and we save about 90% of its species. Lose 95% and save only 5% of the species. Diversity in provinces appears to have been following such a law for hundreds of millions of years (Rosenzweig and Ziv 1999). We have no evidence to indicate that law has been repealed. Once our mini-world has dwindled to island-like diversity, the remaining species left will all begin with at least one source population. So how could further deterioration occur? Part of the answer is accidents. Source species can vanish merely because they encounter a series of poor years. In addition, global warming may change them into sink species by pushing their remaining source habitats out of all reserves and into cornﬁelds or the sea (Peters and Darling 1985). New parasites and diseases will also emerge to take their toll. Thus, relaxation below the island-like steady state will come from inﬂated extinction rates. Misfortunes that eradicate successful species have always accompanied life. In ordinary times, life replaces such losses by speciation. However, this time, because constricted geographic ranges produce fewer isolates, the loss of area will also be depressing the speciation rate curve. Many species are being restricted to a single reserve—with no chance of further allopatric speciation. The loss of ecological theater is changing the evolutionary play.1 1

Thanks to G.E. Hutchinson for the image.

276

Biodiversity, Conservation, and Management

Previous mass extinctions were a violent interruption and perturbation of steady-state diversities rather than a change in speciation and extinction rate curves. Afterward, background conditions returned and life gradually recovered its steady states under the inﬂuence of more or less the same speciation rate curves as had obtained before the catastrophe. But the process of reachieving a steady state after the current biotic crisis will not resemble any previous recovery from a mass extinction (Rosenzweig 2001). New speciations will not keep up with the losses. Yes, eventually, a balance will be restored. However, the mass extinctions of our era represent a gradual relaxation to a new steady state. This new state is dictated by the shrunken area available to nature and by the shrunken speciation rates that must characterize such a shrunken area. For restoration of a steady state, enough species must vanish so that the total extinction rate of those that remain declines to the level of their total speciation rate (ﬁg. 16.4). Recovery of steady-state dynamics will occur as soon as the mass extinction is over, that is, after complete relaxation. To a large extent, the trajectory of stochastic extinctions—not the trajectory of originations—will determine how long the process will take. Furthermore, at the steady-state of the future, when life is again replacing its losses by speciation, its richness will be decimated. Human pressure may greatly accelerate the relaxation process by increasing extinction rates. Various human activities suggest this. We increasingly commingle evolutionarily separate provincial biotas, creating the New Pangaea and introducing predatory and competitive threats from exotic species (Mooney and Cleland 2001). We rapidly transport novel diseases

Figure 16.4 Because it was much larger than today’s, the natural world of 2000 years ago (—) had both a higher speciation-rate curve (convex up) and a lower extinction-rate curve (concave up). The circle indicates its steady state. The new curves (-:::-) predict the steady state of today’s smaller natural world (œ). It will be achieved only after excess extinction has reduced diversity to the point where the total rate of extinction again equals the total rate of speciation.

Reconciliation Ecology and the Future of Species Diversity

277

and parasites around the world. We simplify biotic temporal regimes (e.g., by limiting disturbances such as ﬁre). And we are warming the globe. The National Research Council (1995) implicates exotic species (pp. 37, 38) or lack of adequate disturbance (p. 105) as the root cause in endangering a signiﬁcant proportion of threatened U.S. species. But global warming may constitute the worst threat of all; by altering the basic abiotic conditions of reserves, global warming can destroy their ability to do much of their job. When the earth was covered with contiguous tracts of natural habitat, species could track such changes, moving to keep up with the shifts in location of their favored habitats and so avoiding extinction (Brett 1998, Coope 1987, Davis 1983). But today, with natural habitats restricted to patches of reserves, this is not possible. Meanwhile, we show little sign of abandoning the thoughtless destruction of whatever unprotected natural habitat remains. We stand at the edge of an abyss as deep as the greatest known catastrophe in the history of life, the Permo-Triassic mass extinction, which, some 225 my ago, exterminated more than 95% of the earth’s species. Life eventually recovered from that catastrophe but life will never recover from this one. The Permo-Triassic catastrophe occurred because of a temporary disaster. Recovery commenced as soon as environmental conditions returned to their more usual states. But this time, we are the disaster and we have no intention of going away. But we can do something to step away from the precipice. Today, conservation biology battles to save species by using two dominant tactics: reservation ecology and restoration ecology. Unhappily, owing to the power of area, these two cannot do much by themselves. No conservationist seriously believes that we can reserve much more than the 5% or 10% that now remains. And many will admit that the human population is likely to continue expanding. If we use only reservation ecology and restoration ecology, it would seem that we are doomed to lose nearly every species alive today. Reservation and restoration ecology must be supplemented. ‘‘Conservation philosophy, science, and practice must be framed against the reality of human-dominated ecosystems, rather than the separation of humanity and nature underlying the modern conservation movement’’ (Western 2001). Fortunately, some people have begun this work.

Reconciliation Ecology Ecologists are pioneering a new area of research, which will make long-term diversity conservation possible. I call it reconciliation ecology. Reconciliation ecology discovers how to modify and diversify anthropogenic habitats so that they harbor a wide variety of wild species. In essence, it seeks techniques to give many species back their geographical ranges without taking away ours (ﬁg. 16.5). Thus it is trying to expand the area available to nature. That will establish a steady-state diversity far greater than would be available if conservation biologists restrict themselves to the areas aﬀorded by set-asides alone.

278

Biodiversity, Conservation, and Management

Figure 16.5 The capacity of land to support diversity. Reconciliation ecology treats this as a continuous variable. It seeks techniques to move land to the right along that continuum. This it accomplishes by redesigning human habitats to give some species back their geographical ranges without taking away ours.

A growing number of examples demonstrate that reconciliation ecology can work (Rosenzweig 2003). I will describe a few in the rest of the chapter.

Reconciliation in Agricultural Sites Because agricultural uses dominate most of the land areas that people have taken for themselves, perhaps the most important cases of reconciliation ecology are those associated with agriculture. Led by Gretchen Daily (Daily et al. 2001), John Vandermeer and Yvette Perfecto (Vandermeer and Perfecto 1995), and Russell Greenberg (Greenberg et al. 1997), ‘‘Countryside Biogeography’’ is showing that some styles of land use, especially those of traditional agriculture, are already compatible with the needs of many species. Sometimes the compatibility of diversity and agriculture occurs quite accidentally, sometimes it is deliberate. Such compatibility exists in pasturelands, croplands, plantations, and timberlands. It comes from rich and poor countries, sponsored by private agencies or by governmental agencies. The following examples illustrate the variety.

Cardamom Growers maintain many tree species in their cardamom (Elettaria cardamomum) plantations, both to provide shade for the herb and a steady supply of nectar for its pollinators, principally honey bees. Bees visit 37 tree species in the plantations, of which 10 supply both nectar and pollen, three nectar only and the rest pollen only. However, from May to September, ﬂowers are not

Reconciliation Ecology and the Future of Species Diversity

279

very abundant. Biologists are looking for additional plant species to provide a steadier nectar supply for the bees (Kuruvilla et al. 1995). This is deliberate reconciliation ecology. Pest Control Reconciliation seems to have considerable potential for minimizing losses to agricultural pests. For example, California viticulturists rely on a parasitoid wasp for biological control of leafhoppers in their vineyards (Doutt and Nataka 1973). Grape growers planted patches of native blackberries in shady spots near vineyards to maintain wasp populations during the winter. Each spring, wasps reinvade the vineyards from the blackberry patches, thus keeping pest populations down. The mosaic of habitats designed by the growers includes patches suitable for all the native species associated with blackberries. Pasturelands Pasturelands, too, provide some encouraging examples of reconciliation. For example, the Ocosingo Valley in Chiapas, Mexico, has extensive pastureland, as well as patches of managed and unmanaged woodlands, and patches of Acacia pennatula, a species whose stems and branches are too spiny for cattle to eat (Greenberg et al. 1997). Yet, the acacia pods contain high amounts of protein and make a valuable, seasonal cattle food. The ranchers also need the acacia for fence posts. Therefore ranchers maintain small woodlots composed almost exclusively of acacia. Greenberg et al. (1997) censused birds in 18 diﬀerent Chiapas habitats, including the acacia woodlots. The woodlots have more species of overwintering songbirds (18) and more individuals than any other habitat—including ‘‘natural’’ forest at low elevation. Thus, the management of these woodlots is a case of reconciliation by accident.

Reconciliation of Disappearing Ecosystems Longleaf Pine Forest A joint project of the U.S. Air Force and The Nature Conservancy produced an exciting example of reconciliation in timberlands. To save longleaf pine forest and its endangered species such as the red-cockaded woodpecker, they undertook novel, carefully studied, and continuous management in Eglin Air Force Base, Florida (McWhite et al. 1993). In 1992, the base retained only 693 hectares of old-growth longleaf pine, and even these were not reproducing. In 1993 the Air Force began to remove large numbers of other species of pines. They planted more than three million longleaf seedlings. And, annually, they are burning substantial fractions

280

Biodiversity, Conservation, and Management

of the forests’ understory. Longleaf pine now dominates more than 80,000 hectares of Eglin Air Force Base. The Air Force has also helped the rare animals of the pineland, especially, the red-cockaded woodpecker. The red-cockaded woodpecker excavates its nest holes in living longleaf pine trees. To supply them, Air Force crews drill artiﬁcial nest cavities in the trunks of the healthy young longleaf pines. Redcockaded woodpeckers nest in 30% of these holes, and their population has begun to grow. Meanwhile the base continues to develop and test various explosive weapons. Many thousands of people live on the base, and thousands of others buy permits to camp, ﬁsh, and hunt in its pineland. Timbering has actually increased and is proﬁtable. Never before has there been a longleaf pine forest like this one. Eilat Salt Marsh Until 30 years ago, a 12 km2 natural salt marsh in Eilat, Israel, provided a critical feeding stop on the migratory route of 257 bird species, perhaps a third of all the bird individuals in Europe and western Asia. But the marsh was totally destroyed by resort development. To replace it, Ruven Yosef created a single patch of non-natural salt marsh, carefully built up, contoured, and planted on a refuse dump (Cherrington 1999). Its soil came from the excavations of the hotel-construction industry itself. It is regularly irrigated with treated, nutrient-rich sewage water and has roughly four times the productivity of the natural marsh it replaced. The marsh little resembles its predecessor or any natural habitat. But it shows that reconciliation can be practical in ecosystems that produce no direct beneﬁts to people. Backyard Wildlife Habitat Residential areas oﬀer important opportunities for reconciliation ecology although they are much less extensive than agricultural lands. Recognizing this opportunity, the U.S. National Wildlife Federation has sponsored a campaign called ‘‘Backyard Wildlife Habitat.’’ It encourages people to bring nature to their own homes. So far, it has enrolled more than 20,000 private sites that vary from a few hectares to a single balcony. All participants try to create a modiﬁed human habitat that speaks to the needs of at least some wildlife (Tufts and Loewer 1995).

Reconciliation for Disappearing Species In a number of the previous examples, reconciliation ecology was associated with reinventing a scarce or even an endangered ecosystem. But reconciliation may also focus on preserving a single rare species. This section will look at a few such cases.

Reconciliation Ecology and the Future of Species Diversity

281

Bluebirds Eastern bluebirds (Sialia sialis) once thrived in human habitats of North America. Today, house sparrows and European starlings have outcompeted them for nest holes near people. Starlings also eat the berries that the bluebirds need to survive during the winter. Bluebirds have almost vanished. However, some designs of nestbox suit bluebirds but not house sparrows or starlings. In 1979, the North American Bluebird Society was founded and began to encourage people to deploy appropriate nestboxes on their property. Bluebird numbers are building back. Natterjack Toads Reconciliation eﬀorts in England on behalf of Bufo calamita, the natterjack toad, constitute a detailed, multifaceted, and sustained eﬀort by some 50 researchers over the course of a quarter of a century. They culminated in the development and installation of habitats to save this species in the United Kingdom (Denton et al. 1997). Having ﬁrst characterized the natterjack’s niche, this team cleared dense vegetation and reintroduced grazing to maintain the early stages of succession. They fought acidiﬁcation by adding Ca(OH)2 to natterjack ponds every year or two, or scraping the sulfate-rich silt from the bottom of the ponds. They removed some B. bufo competitors. They also built some 200 new ponds, often using old bomb craters and active golf courses. The new ponds rescued or increased B. calamita populations in two-thirds of the sites, and the grazing has turned a small proﬁt. Shrikes Many species of shrikes are endangered (Yosef and Lohrer 1995). Ruven Yosef found that loggerhead shrikes (Lanius ludovicianus), like most shrikes, prefer to hunt by sitting on a post or branch, scanning the ground around them, and then pouncing on an insect. Cattle ranches often have quite suitable ﬁelds full of insects, but few perches. Yosef installed cheap fence posts in a working cattle ranch and greatly improved its value for shrikes (Yosef and Grubb 1994). Shrike territories with the extra fence posts shrank an average of 77% and the shrike population increased 60%. Parent birds in smaller territories had 33% more successful clutches than controls, and raised 29% more chicks per successful clutch. Other species of shrikes are also being helped by similar methods of reconciliation (van Nieuwenhuyse 1998, Scho¨n 1998, Devereux 1998).

Afterword Reconciliation ecology addresses the new, sterile habitats in which most species cannot function at all. If this new strategy of conservation biology

282

Biodiversity, Conservation, and Management

spreads and inﬂuences a substantial proportion of the earth’s area, it can halt the current mass extinction. The path that leads to reconciled human habitats calls for new attitudes and new public and private institutions. An immense amount of research lies ahead as we accumulate a library of the habitat requirements of myriad species, and learn how to combine them (National Research Council 2001). We will even need to alter the way we manage our reserves. But, evidence indicates that we cannot preserve the large-scale at the tiny scale. If the area available to wild species remains very low or declines even further, even our biotic preserves will not be able to maintain their diversities for very long.

Acknowledgments At ﬁrst I was afraid that the interprovincial SPAR meant that nothing could be done to save diversity. But Ron Pulliam rescued me from despair. Arnie Miller taught me the signiﬁcance of the fact that Cenozoic fossils very often come from unconsolidated deposits. Gordon Orians, David Policansky, Margaret Mayﬁeld and Caleb Gordon suggested some of the examples of reconciliation. Thanks to Ruven Yosef for information, inspiration, and hospitality. A slightly expanded version of this chapter originally appeared in Oryx, vol. 37, no. 2.

References Allmon, W.D., Rosenberg, G., Portell, R.W., and Schindler, K.S. 1993. Diversity of Atlantic coastal plain mollusks since the Pliocene. Science 260: 1626–1629. Alroy, J. 1998. Equilibrial diversity dynamics in North American mammals. In Biodiversity Dynamics: Turnover of Populations, Taxa and Communities, ed. M.L. McKinney and J.A. Drake, 232–287. New York: Columbia University Press. Alroy, J., Marshall, C.R., Bambach, R.K., Bezusko, K., Foote, M., Fuersich, F.T., Hansen, T.A., Holland, S.M., Ivany, L.C., Jablonski, D., Jacobs, D.K., Jones, D.C., Kosnik, M.A., Lidgard, S., Low, S., Miller, A.I., Novack-Gottshalle, P.M., Olszewski, T.D., Patzkowsky, M.E., Raupf, D.M., Roy, K., Sepkoski, J.J., Jr., Sommers, M.G., Wagner, P.J., and Webber, A. 2001. Eﬀects of sampling standardization on estimates of phanerozoic marine diversiﬁcation. Proceedings National Academy of Sciences (USA) 98: 6261–6266. Arrhenius, O. 1921. Species and area. Journal of Ecology 9: 95–99. Benton, M.J. 1995. Diversiﬁcation and extinction in the history of life. Science 268: 52–58. Boucot, A.J. 1975. Evolution and Extinction Rate Controls. Amsterdam: Elsevier Scientiﬁc. Bretsky, P.W., and Bretsky, S.W. 1976. The maintenance of evolutionary equilibrium in Late Ordovician benthic marine invertebrate faunas. Lethaia 9: 223–233. Brett, C.E. 1998. Sequence stratigraphy, palaeoecology, and evolution: biotic clues and responses to sea-level ﬂuctuations. Palaios 13: 241–262. Brett, C.E., Ivany, L.C., and Schopf, K.M. 1996. Coordinated stasis: an overview. Palaeogeography, Palaeoclimatology, Palaeoecology 127: 1–20.

Reconciliation Ecology and the Future of Species Diversity

283

Burnham, K.P., and Overton, W.S. 1979. Robust estimation of population size when capture probabilities vary among animals. Ecology 60: 927–936. Chao, A., and Lee, S.-M. 1992. Estimating the number of classes via sample coverage. Journal of the American Statistical Association 87: 210–217. Chao, A., Lee, S.-M., and Jeng, S.-L. 1992. Estimating population size for capturerecapture data when capture probabilities vary by time and individual animal. Biometrics 48: 201–216. Cherrington, M. 1999. The falcon and the ﬁrebrand. Earthwatch May/June: 39–47. Coope, G.R. 1987. The response of late Quaternary insect communities to sudden climatic changes. In Organization of Communities Past and Present, ed. J.H.R. Gee and P.S. Giller, 421–438. Oxford: Blackwell Scientiﬁc. Daily, G., Ehrlich, P.R., and Sa´nchez-Azofeifa, G.A. 2001. Countryside biogeography: use of human-dominated habitats by the avifauna of southern Costa Rica. Ecological Applications 11: 1–13. Davis, M.B. 1983. Holocene vegetational history of the eastern United States. In LateQuaternary Environments of the United States, ed. H.E. Wright, Jr., 166–181. Minneapolis: University of Minnesota Press. Denton, J.S., Hitchings, S.P., Beebee, T.J.C., and Gent, A. 1997. A recovery program for the natterjack toad (Bufo calamita) in Britain. Conservation Biology 11: 1329– 1338. Devereux, C. 1998. The ﬁscal shrike. Africa—Birds and Birding 3: 52–57. Doutt, R.L., and Nataka, J. 1973. The Rubus leafhopper and its egg parasitoid: an endemic biotic system useful in grape pest management. Environmental Entomology 2: 381–386. Durrett, R., and Levin, S. 1996. Spatial models for species-area curves. Journal of the Theoretical Biology 179: 119–127. Fisher, R.A., Corbet, A.S., and Williams, C.B. 1943. The relation between the number of species and the number of individuals in a random sample of an animal population. Journal of Animal Ecology 12: 42–58. Greenberg, R., Bichier, P., and Sterling, J. 1997. Acacia, cattle and migratory birds in southeastern Mexico. Biological Conservation 80: 235–247. Harte, J., and Kinzig, A. 1997. On the implication of species-area relationships for endemism, spatial turnover, and food web patterns. Oikos 80: 417–427. Harte, J., Kinzig, A., and Green, J. 1999. Self-similarity in the distribution and abundance of species. Science 284: 334–336. Hubbell, S.P. 2001. The Uniﬁed Neutral Theory of Biodiversity and Biogeography. Princeton, NJ: Princeton University Press. Huston, M. 1993. Biological diversity, soils, and economics. Science 262: 1676–1680. Jackson, J.B.C. 2001. What was natural in the coastal oceans? Proceedings National Academy of Sciences (USA) 98: 5411–5418. Kuruvilla, K.M., Radhakrishnan, V.V., and Madhusoodanan, K.J. 1995. Small cardamom plantations—ﬂoristic calendar and bee pasturage trees. Journal of Environmental Resources 3: 32–33. Lawton, J. 1999. Are there general laws in ecology? Oikos 84: 177–192. Lee, S.-M., and Chao, A. 1994. Estimating population size via sample coverage for closed capture-recapture models. Biometrics 50: 88–97. Leitner, W.A., and Rosenzweig, M.L. 1997. Nested species-area curves and stochastic sampling: a new theory. Oikos 79: 503–512. MacArthur, R.H., and Wilson, E.O. 1967. The Theory of Island Biogeography. Princeton, NJ: Princeton University Press.

284

Biodiversity, Conservation, and Management

May, R.M. 1975. Patterns of species abundance and diversity. In Ecology and Evolution of Communities, ed. M.L. Cody and J.M. Diamond, 81–120. Cambridge, MA: Belknap Press of Harvard University Press. McWhite, R.W., Green, D.R., Petrick, C.J., Seiber, S.M., and Hardesty, J.L. 1993. Natural Resources Management Plan, Eglin Air Force Base, Florida. Eglin Air Force Base, FL: U.S. Department of the Air Force. Miller, A.I., and Foote, M. 1996. Calibrating the Ordovician radiation of marine life: implications for Phanerozoic diversity trends. Paleobiology 22: 304–309. Mooney, H.A., and Cleland, E.E. 2001. The evolutionary impact of invasive species. Proceedings National Academy of Sciences (USA) 98: 5446–5451. Myers, N. 1999. Everyday extinctions in embattled forests. Nature 400: 231–232. Myers, N., Mittermeier, R.A., Mittermeier, C.G., da Fonseca, G.A.B., and Kent, J. 2000. Biodiversity hotspots for conservation priorities. Nature 403: 853–858. National Research Council. 1995. Science and the Endangered Species Act. Washington, DC: National Academy Press. National Research Council. 2001. Grand Challenges in Environmental Sciences. Washington, DC: National Academy Press. Nichols, J.D., and Pollock, K.H. 1983. Estimating taxonomic diversity, extinction rates, and speciation rates from fossil data using capture-recapture models. Paleobiology 9: 150–163. Patterson, B.D. 1990. On the temporal development of nested subset patterns of species composition. Oikos 59: 330–342. Patterson, B.D., and Atmar, W. 2000. Analyzing species composition in fragments. In Isolated Vertebrate Communities in the Tropics, ed. G. Rheinwald, 9–24. Proceedings 4th International Symposium, Bonn, Bonn Zoology Monograph 46. Peters, R.L., and Darling, J.D.S. 1985. The greenhouse eﬀect and nature reserves. Bioscience 35: 707–717. Pimm, S.L., Russell, G.J., Gittelman, J.L., and Brooks, T.M. 1995. The future of biodiversity. Science 269: 347–350. Preston, F.W. 1960. Time and space and the variation of species. Ecology 41: 785–790. Preston, F.W. 1962. The canonical distribution of commonness and rarity. Ecology 43: 185–215; 410–432. Raup, D.M. 1976. Species diversity in the Phanerozoic: an interpretation. Paleobiology 2: 289–297. Rosenzweig, M.L. 1975. On continental steady states of species diversity. In Ecology and Evolution of Communities, ed. M.L. Cody and J.M. Diamond, 121–140. Cambridge, MA: Belknap Press of Harvard University Press. Rosenzweig, M.L. 1995. Species Diversity in Space and Time. Cambridge, UK: Cambridge University Press. Rosenzweig, M.L. 2001. Loss of speciation rate will impoverish future diversity. Proceedings National Academy of Sciences (USA) 98: 5404–5410. Rosenzweig, M.L. 2003. Win-Win Ecology: How Earth’s Species Can Survive in the Midst of Human Enterprise. New York: Oxford University Press. Rosenzweig, M.L. and Ziv, Y. 1999. The echo pattern in species diversity: pattern and process. Ecography 22: 614–628. Rosenzweig, M.L., Turner, W.R., Cox, J.G., and Ricketts, T.H. 2003. Estimating diversity in unsampled habitats of a biogeographical province. Conservation Biology 17(3): 864–874. Scho¨n, M. 1998. Conservation measures and implementation for the great grey shrike (Lanius excubitor) in the southwestern Schwa¨bische Alb of southwestern

Reconciliation Ecology and the Future of Species Diversity

285

Germany. In Shrikes of the World—II: Conservation Implementation, ed. R. Yosef and F. Lohrer, 68–73. Eilat: International Birdwatching Center in Eilat. Sepkoski, J.J., Jr. 1978. A kinetic model of Phanerozoic taxonomic diversity I. Analysis of marine orders. Paleobiology 4: 223–251. Shmida, A., and Ellner, S. 1984. Coexistence of plant species with similar niches. Vegetatio 58: 29–55. Tufts, C., and Loewer, P. 1995. Gardening for Wildlife. Emmaus, PA: Rodale Press. van Nieuwenhuyse, D. 1998. Conservation opportunities for the red-backed shrike (Lanius collurio). In Shrikes of the World—II: Conservation Implementation, ed. R. Yosef and F. Lohrer, 79–82. Eilat: International Birdwatching Center in Eilat. Van Valkenburgh, B., and Janis, C.M. 1993. Historical diversity patterns in North American large herbivores and carnivores. In Species Diversity in Ecological Communities: Historical and Geographical Perspectives, ed. R. Ricklefs and D. Schluter, 330–340. Chicago: University of Chicago Press. Vandermeer, J., and Perfecto, I. 1995. Breakfast of Biodiversity: the Truth About Rain Forest Destruction. Oakland: Institute for Food and Development Policy. Vitousek, P., Mooney, H., Lubchenco, J., and Melillo, J. 1997. Human domination of Earth’s ecosystems. Science 277: 494–499. von Humboldt, F.H.A. 1807. Essai sur la geographie des plantes. Paris: von Humboldt. Western, D. 2001. Human-modiﬁed ecosystems and future evolution. Proceedings National Academy of Sciences (USA) 98: 5458–5465. Williams, C.B. 1943. Area and the number of species. Nature 152: 264–267. Williams, C.B. 1964. Patterns in the Balance of Nature. London: Academic Press. Wissel, C., and Maier, B. 1992. A stochastic model for the species-area relationship. Journal of Biogeography 19: 355–361. Yosef, R., and Grubb, T.C., Jr. 1994. Resource dependence and territory size in loggerhead shrikes (Lanius ludovicianus). Auk 111: 465–469. Yosef, R., and Lohrer, F.E., eds. 1995. Shrikes (Laniidae) of the World: Biology and Conservation. Proceedings Western Foundation of Vertebrate Zoology 6: 1–343.

17 Management for Biodiversity Human and Landscape Effects on Dry Environments Avi Perevolotsky Moshe Shachak Steward T.A. Pickett

B

iodiversity is one of the principal pillars of natural ecosystems. In fact, biodiversity can be interpreted as a manifestation of the various biotic and abiotic components of the ecological systems and their mutual interactions or as the totality or variation (chapter 19). Biodiversity applies at diﬀerent realms of ecological criteria: organism (genetic/phenological), species, habitat, and landscape (Loidi 1999, Noss 1990). Historically, it was the speciﬁc assemblage of organisms—the species diversity—that attracted the attention of scientists. Later, the eﬀect of landscape structure on biological diversity, through habitat and niche properties, became an additional focus of biodiversity research (Malanson and Cramer 1999). The impact of diﬀerent disturbances on the ecosystem and community structure has also become part of the study of landscape–biodiversity interrelationships (Moloney and Levin 1996; Trabaud and Galtie 1996). In this chapter we present a third dimension that aﬀects biodiversity: human intervention through management and land-use patterns. One may consider this dimension as another source of disturbance, but we believe that such an approach is narrow. In contrast to disturbance, management is intentional, directional, goal-oriented, and, in some cases, scientiﬁcally or professionally guided. Human societies have modiﬁed the biodiversity of their environments since prehistoric time. Traditional land use that has evolved from ancient practice usually produces highly diverse landscapes based on knowledge of old systems of land exploitation (Loidi 1999). In Scotland, for example, biodiversity was enhanced by the interactions between farmers and the woodlands surrounding the agricultural ﬁelds (Tipping et al. 1999). Modern aﬀorestation schemes fail to create diverse woodlands similar to the ancient ones. 286

Management for Biodiversity

287

Maintenance of biodiversity through active management has recently become an important challenge for modern conservation (Monkkonen 1999). In this chapter, we use a conceptual model of human–biodiversity relationships (ﬁg. 17.1) and apply it to water-limited systems. The model describes how ecosystem services provided to traditional and modern societies are enhanced by management actions. The ecosystem services discussed in this chapter are water accumulation, food production (mainly through primary production), and recreation potential. The essence of the model is that without external input of water, ecosystem services are controlled by relationship between landscape mosaic, ecosystem processes, and organisms. However, in many cases humans have actively managed landscape diversity to direct ecosystem processes, mostly through water redistribution, for enhancing the distribution and abundance of organisms that provide ecosystem services. Theoretically, there are two directions that this model may take. The ﬁrst implies that all three ecological variables—landscape diversity, species diversity and ecosystem processes—are site-speciﬁc and vary according to a speciﬁc set of dependent and independent interactions that characterize the site. The second avenue suggests that only the diversity variables are sitespeciﬁc, while the ecosystem processes remain similar in principle. In other words, landscape and species may diﬀer to a great deal from one site to the other, but the principal ecological processes remain the same. The model is tested against data from ﬁve case studies located along a rainfall gradient (70–600 mm/year) in Israel and the Sinai. First, the changes in landscape along the 500 km gradient are described. Second, we relate landscape mosaic to the process of water ﬂow. Third, we demonstrate how the ﬂow of water aﬀects the diversity of organisms that provide ecosystem

Figure 17.1 Simpliﬁed relationship among biodiversity (species and landscape diversity), ecosystem processes, human activities, and ecosystem services.

288

Biodiversity, Conservation, and Management

services. We conclude with an analysis of human–landscape relationships and their eﬀect on ecosystem services. In three of the case studies (Mt. Sinai, Sede Boqer, and Lehavim) the emphasis is on human–biodiversity relationships within the context of traditional subsistence societies where contemporary activities carry on ancient practices. The other two case studies (Park Sayeret Shaked and Park Ramat Hanadiv) deal with modern society and one of its growing needs: recreation in open areas. At all ﬁve locations the natural environment provides a basic patchy structure, but human activities have promoted, developed, and modiﬁed this spatial pattern into a more complex and diverse one, hence aﬀecting the biodiversity at all levels.

Landscape Diversity The ﬁve study sites represent diversity of landscapes along a moisture (south to north) and rainfall gradient ranging from 70 to 600 mm (ﬁg. 17.2). The most southern site—Mount Sinai area—is located at the center of a mountainous region in the southern part of the Sinai peninsula with an average annual rainfall of 70–100 mm, some of it as snow. Mount Sinai is one of many neighboring peaks rising up to more than 2000 m above sea level. The area is composed mainly of two types of rocks—‘‘black’’ (volcanic and dark plutonic rocks) and ‘‘red’’ (mostly red granite). The cracked ‘‘black’’ rocks are characterized by large slopes in which the ratio between exposed rock and soil patches is low. In other words, most of the slope is covered by large soil patches. On the other hand, the adjacent red granite produces huge exposed rocky slopes with soil mainly in the valleys. The red granite landscape is characterized therefore by high rock-to-soil ratio on the watershed scale (rocky slopes and soil-covered valleys). The Sede Boqer site is located in the central Negev Highlands in Israel, where average annual rainfall is about 100 mm. As in the Sinai site, the area is a rocky desert. The landscape mosaic is a matrix of exposed limestone bedrock with patches of airborne loess soil. Most of the area is composed of small watersheds. In a watershed, rock-to-soil ratio can be observed on three spatial scales. On the whole watershed scale, landscape diversity is determined by the ratio of rock cover on the slopes to soil cover in the valley. On the slope scale, landscape mosaic is related to the ratio of rocky extensions over the upper slope and the colluvial soil of the lower slope. On a smaller scale, the bedrock, which usually forms a stepped topography, exhibits a high rock-tosoil ratio. Bedrock is exposed over most of the area while soil material is found in rock crevices and at the base of bedrock steps (Yair and Shachak 1987). From a landscape perspective, the Sinai and Sede Boqer sites represent rocky deserts where landscape diversity is determined mainly by physical processes. Such processes are related to geological, pedological, and geomorphological phenomena. These two sites represent a landscape mosaic in which

Management for Biodiversity

289

Figure 17.2 Map of Israel and the Sinai Peninsula, emphasizing the ﬁve research areas (black squares).

rock and soil types, and their spatial arrangement, deﬁne landscape diversity on various scales. Park Sayeret Shaked, which is located in the northern Negev, represents a transition from landscape dominated by physical patchiness to a landscape dominated by biological patchiness (Shachak et al. 1999). This is mainly due to the increase in rainfall amount (200 mm) and the fact that the bedrock is almost completely covered by loess soil. The landscape mosaic is characterized by a matrix of soil crust, covered with a microphytic community, and small patches of dwarf bushes. The microphytic soil matrix is a ﬂat compacted soil surface composed of cyanobacteria, bacteria, algae, bryophytes, and lichens (Zaady and Shachak 1994). Within the crust matrix, shrub patches are interspersed. Shrub patches consist of a shrub with a loose soil mound underneath. Herbaceous vegetation is spread over the intershrub space and on the mound. At this site, landscape diversity is deﬁned by the ratio of crust to shrub patches, their spatial conﬁguration, and the properties

290

Biodiversity, Conservation, and Management

of the two patch types. The properties of the crusted soil are related to the soil grain size distribution and the composition of the microphytic community. The properties of the shrub patches are related to the architecture of the shrub and the soil mound underneath. The Lehavim site is also located in the northern Negev but in a semiarid area (300 mm average annual rainfall). This site contains the same landscape elements as Sede Boqer and Sayeret Shaked sites but with diﬀerent organization. These elements are exposed rocks, crusted soil, and shrub patches, but in Lehavim they are adjacent to each other. Herbaceous vegetation is spread heterogeneously in patches in all of these landscape units. The Lehavim site represents a complex network of physical and biological patchiness, with higher rainfall than the previous three locations. The Ramat Hanadiv Park is located outside of the desert. This park extends over the southern tip of Mount Carmel in central Israel, and enjoys a Mediterranean climate (600 mm rainfall/year). Consequently, there is a signiﬁcant shift in the biological patchiness and the factors determining landscape diversity. The soil crust is not a landscape component in this area. On the other hand, Ramat Hanadiv is almost completely covered by evergreen, woody vegetation. The network of patches within this site is controlled, to a great extent, by the woody vegetation cover and related biological processes (Hadar et al. 1999). Typical patch types are in accordance with vegetation types: low shrubs, moderate-sized shrubs that create dense stands, and patches of planted trees. Within each of these patches, diverse herbaceous vegetation develops in open areas (Hadar et al. 2000). Thus, diversity of vegetation architecture is a principal factor determining landscape mosaic in Ramat Hanadiv. However, physical patchiness expressed as a variety of rock and soil patches is also present in this park, and plays a signiﬁcant ecological role. Dense vegetation is highly correlated with the redistribution of subterranean water as detailed in the following section. Low cover of woody vegetation may be associated with the local water regime or with speciﬁc land use history. The Sinai–Ramat Hanadiv transect demonstrates a high landscape diversity within and among locations. It is expected that an increase in landscape diversity would diversify ecosystem processes and enhance organism diversity. Therefore, one may assume that the strategy of management and utilization in each site would be speciﬁc, based on its speciﬁc features of diversity. If this was the case, the uniﬁed principles for human–landscape relationships in these water-limited systems would be diﬃcult to ascertain. However, if higher landscape diversity does not signiﬁcantly modify ecosystem processes and organism diversity, then a single model may be developed for all sites. We believe that the latter is true for the locations discussed in this chapter. In order to address the issue of general versus speciﬁc human–landscape relationships we will discuss the relationships among landscape diversity, a principal ecosystem process (water ﬂow), a measure of organism diversity (the diversity of vegetation functional groups), and services provided by the ecosystem.

Management for Biodiversity

291

Landscape Diversity and Water Redistribution Typical to all ﬁve sites is the process of redistribution of rainfall water by overland or subterranean ﬂow. This ﬂow is controlled by landscape components and aﬀects soil moisture distribution across the area. Water redistribution is especially signiﬁcant in areas with less than 100 mm of rainfall, where productivity is highly dependent on the existence of water-enriched landscape patches (Noy-Meir 1973). Such enrichment occurs only when the amount of water received and inﬁltrated into the soil is much higher than precipitation. It happens when in addition to direct rainfall the soil receives substantial runoﬀ water. In the Sinai the cracked ‘‘black’’ rock creates a low soil moisture environment. This is because only a small amount of run-oﬀ is shed from the cracked rock and is available to the soil patches. The black rocks also provide small run-oﬀ contributing areas in relation to relatively large soil patches. Therefore, soil moisture increases only slightly due to run-oﬀ generated on these rocky slopes. Contrary to the ‘‘black rocks’’ in the red granite area of the Sinai, the interaction between rainfall and surface properties creates a water-enriched environment, much higher in soil moisture content than can be attained by direct precipitation. The huge exposed rocky slopes generate considerable amounts of run-oﬀ during each rainfall event. The large continuous areas of bare bedrock direct the run-oﬀ into the valleys and dry river beds. Consequently, internal valleys and dry river beds in the red granite area enjoy a highly enriched water regime. The main factor controlling biological activity at Sede Boqer, as in the Sinai, is the concentration of run-oﬀ water shed by the rocky area into soil patches (Yair and Shachak, 1987). Water redistribution at Sede Boqer was found to be related to the interaction between rainfall properties and physical patchiness on various scales. During rainfall events of low intensity and amount, redistribution of water is mainly on the small scale of the rocky part of the slope. Run-oﬀ generated over the exposed bedrock inﬁltrates the soil patches in the rock crevices and at the base of bedrock layers. During rainfall events of intermediate intensity and amount, redistribution of water mainly occurs between the rocky areas of the upper slope and the colluvial soil located on the lower part of the slope. Only during rainfall events of high intensity and amount, water redistribution occurs between the whole slope and the valley. At Sayeret Shaked, redistribution of water also controls productivity and diversity (Shachak et al. 1998). However, at this location, water redistribution is not controlled by physical patchiness but by biological processes. The landscape mosaic is arranged in such a form that resources generated by the crust patches are intercepted by shrub patches and are accumulated in them. Shrub patches function as sinks for the run-oﬀ water and therefore become the principal loci for productivity and diversity (Garner and Steinberger 1989; Shachak and Pickett 1997). This is mainly due to the accumulation of soil,

292

Biodiversity, Conservation, and Management

water, and nutrients, which promotes the growth of rich herbaceous vegetation under the shrubs. In Lehavim, the landscape consists of two patch types that generate runoﬀ: exposed bedrock and bare soil crust, and one patch type—the shrub patch—that absorbs water. The rock and crusty soil function as sources for run-oﬀ, while the shrub understorey acts as a sink for most of this water. This spatial structure creates a complex network of soil moisture. From the perspective of conversion of rainfall to soil moisture, Lehavim exhibits redistribution of water by physical patchiness as in Mount Sinai and Sede Boqer. But, at the same time, biological patchiness is also involved in the process, as in Sayeret Shaked. At Ramat Hanadiv, water redistribution is not an outcome of lateral processes that determine water ﬂow origin and target, but results chieﬂy from vertical movement. High cover of relatively deep soil and dense woody vegetation prevent the development of run-oﬀ and surface ﬂow of water and material, as is the case in rocky or crust-covered desert environments. In Ramat Hanadiv, water inﬁltrates through the karstic cavities and cervices that characterize the upper hard limestone rock layers, and accumulates in the lower, soft, water-holding layer of tuﬃc-marl rock. The redistribution of water occurs when this layer is exposed in springs or when roots of woody plants reach it through cracks or shallow cover. In other words, water at Ramat Hanadiv is channeled through the geological layers and becomes available in certain sites. These sites can be easily identiﬁed by the welldeveloped, mesophilic vegetation.

Organism Diversity In this section we refer only to the diversity of organisms that contribute to primary production, which is one of the principal services provided by the ecosystem discussed in this chapter. Primary production is the basis for livestock and wildlife grazing and recreation systems that are common along the Sinai–Ramat Hanadiv gradient. We focus on the relationship between landscape diversity and plant species diversity and the relationship between functional diversity and ecosystem function. In all dry locations, solar energy is converted into biomass mainly by a diverse annual plant community (Evenari et al. 1971). However, at Ramat Hanadiv, as at other Mediterranean sites, most of the biomass is produced by perennial woody plants. Nevertheless, at Ramat Hanadiv, as at all other sites along the gradient, the diversity of the ﬂora is found mainly in the annual, herbaceous component. On a local scale, annual species diversity is dependent on soil moisture distribution in water-enriched patches via the above mentioned mechanisms. Usually, high species diversity is positively correlated in arid habitats with the degree of water enrichment in the soil patches (OlsvigWhittaker et al. 1983).

Management for Biodiversity

293

This is, most probably, the reason for the higher species richness in the ‘‘red’’ granite landscape of Mount Sinai than in the ‘‘black’’ rocks (Shmida and Arbel 1979), and in the rocky environment of the central Negev than in the loess plains (Danin 1977). Biological patchiness is also of importance in determining annual species diversity. Boeken and Shachak (1994) found in Sayeret Shaked higher species richness on the loose soil mound underneath the shrub than on the crusted soil in the intershrub area. At Ramat Hanadiv the distribution of annual species in relation to woody vegetation is reversed. Species richness is much higher in the open space than under or in the vicinity of the woody vegetation (Hadar et al. 1999). In the desert, shrubs provide improved microenvironments (enrichment of soil moisture, shade, refuge from grazing) while in the Mediterranean region, the woody vegetation is so dense that it competes with the herbaceous vegetation for resources (light, water). Physical patchiness can also control annual species diversity on a larger scale. Along the transect, in addition to landscape diversity, other factors, such as land-use history, rainfall regime, and soil properties, could aﬀect annual species diversity (Cowling et al. 1996). There is much evidence to show that traditional human activity, mainly cultivation and livestock husbandry, increased biodiversity (Komas 1983). Humans, it is suggested, enhanced landscape diversity (b diversity) and thus the diversity of vegetation patterns increased (Tipping et al. 1999). Loss of biodiversity under such circumstances (traditional land use) is mostly due to increasing dominance of competitive species, and implies that the impact of human activity has been decreased (Alard et al. 1994). Folke et al. (1996) examined the functional diversity while considering the signiﬁcance of biological diversity in relation to complex and dynamic ecological systems. They identiﬁed two categories of species functional groups related to ecosystem processes: keystone-process species and ecosystem-resilience species. Keystone-process species are those that control, for example, ecosystem production. Ecosystem-resilience species are those that maintain ecosystem sustainability. Typical of all locations along the gradient is high species richness of annuals accompanied by a much lower species richness of woody plants. On the average, annual species richness is at least an order of magnitude higher (hundreds of species per location) than woody species richness (tens of species per location at the most). We suggest that both these groups are essential for the functioning of the ecosystems along the Sinai–Ramat Hanadiv gradient. In the Sinai–Ramat Hanadiv gradient, keystone-process species are the annual plants. They are very eﬃcient in converting available moisture into biomass and can respond to the variability and unpredictability of rainfall (Shmida et al. 1986, Ungar et al. 1999). Most of these species are edible and they serve as the main energy and nutrient source for many herbivores, granivores, and detrivores. They also play a major role in organic matter dynamics, decomposition, and nutrient cycling. In essence, annuals are the

294

Biodiversity, Conservation, and Management

main factor in controlling trophic dynamics and therefore energy ﬂow and nutrient ﬂux (Walker 1988). The functional group essential for ecosystem resilience is the woody plants, mostly shrubs. The shrubs are a principal factor in the development of biological patchiness (Shachak and Lovett 1998). Water, soil, litter, and nutrients accumulate under the shrub patch. These resources are available to the annual plants, thus ensuring their sustainability. We assume that the high resource concentration under the shrub, together with shrub protection against high solar radiation and herbivory, are responsible for the system resilience. The fact that most of the shrubs, at least in the Old World dry environments, are unpalatable and well defended against grazing helps them perform their ecosystem role. It is not clear why species diversity of the annual plants, which are the keystone-process species, is much higher than species diversity of the ecosystem-resilience component—the shrubs. It may be related to the selective pressure on the vegetation due to thousands of years of human activity and livestock grazing (Perevolotsky 1999a). However, we suggest that richness of both annuals and shrubs is of importance in assuring the long-term existence of the ecosystem in relation to changes in rainfall amount, patchiness, and water redistribution along the gradient.

Ecosystem Services Humans have played a signiﬁcant role in all ﬁve ecological systems for thousands of years. They have exploited and managed the landscape and have aﬀected organism diversity to obtain ecosystem services. In our case the major ecosystem services sought are primary productivity, run-oﬀ water, and recreation. In the following discussion, we refer to both natural and managed ecosystem services. Natural Services Natural services are powered by solar energy and are provided without management. Primary Productivity Traditionally, grazing animals have been a major source of meat, milk, wool, and leather in water-limited systems. Therefore, these systems rely on primary production, which provides livestock and wildlife with forage, as an ecosystem service. The utilization of primary production by humans through its consumption by domestic animals is a multiscale process. On a small scale, the interactions between livestock and vegetation are on the watershed level and are based on daily foraging activity. On a larger scale, the interactions are based on seasonal and yearly movements between

Management for Biodiversity

295

diﬀerent ecological regions, one of the characteristic features of nomadic pastoral societies (Dyson-Hudson and Dyson-Hudson 1980). We suggest that there are uniﬁed principles for small- and large-scale interactions. These interactions are independent of location-speciﬁc species diversity, but are dependent on the relationships between landscape diversity and functional group diversity, namely the herbaceous and woody components of rangeland vegetation. Livestock impact on the system is through annual vegetation consumption and soil trampling. In dry environments, livestock feed mainly on green annual plants in the intershrub area before seed-set. This may reduce seed production and therefore lower annual plant production in the following years (O’Connor and Pickett 1992, Bertiller 1996). It has been suggested that the patchy nature of the local environment helps to mitigate the grazing impact on annual seed production (Russell and Schupp 1998). Shrubs act as refuges for herbaceous plants, thus helping them escape grazing before seedset (Osem et al. 1999). Later, seeds disperse from the herbaceous vegetation under the shrubs into the intershrub space. The relative invulnerability of both shrubs and understorey annual species—the former through defense mechanisms (unpalatability by means of secondary metabolites, low nutritional value, and structural deterrence) and the latter by being inaccessible for most livestock—are essential for providing sustainable ecosystem services for the grazing system. The damage caused to the soil by trampling is also mitigated by the two functional groups that create the landscape mosaic. Shrub patches counteract soil erosion caused by trampling by intercepting run-oﬀ ﬂow and trapping eroded sediments. Some species, for example Poa bulbosa, produce a thick, prostrate cover, inaccessable to grazing animals, that also helps to lower erosion from the grazed slopes. Small-scale herding in the intershrub area is a common practice in traditional pastoral societies. It was also introduced to the newly established parks in Sayeret Shaked and Ramat Hanadiv as a management tool (Perevolotsky 1999a). However, high intensity grazing, especially in drought years, may destroy the shrub and intershrub mosaic. This may cause degradation of the system. The solution of traditional societies was to exploit regional heterogeneity by migration, sometimes movement over hundreds of km. In Sinai, goat herding, the traditional subsistence source, has taken advantage of the heterogeneous environment by using the elevation–geological spatial changes in an adaptive temporal sequence that provides forage resources almost all year long (Perevolotsky et al. 1989). The spatial distribution of the rainfall cells introduces another dimension, albeit seasonal and unpredictable, into the landscape diversity (Sharon, 1972). It determines the exact location of abundant seasonal pasture in the desert each year and hence—the spatial movements of the Bedouin herds (Perevolotsky 1987). In the central and northern Negev we do not know enough about historical and traditional grazing practices. Undoubtedly, the herders coordinated the movement of their herds spatially and temporally in accordance with the

296

Biodiversity, Conservation, and Management

spatial and temporal distribution of productivity of natural vegetation in the area (Noy-Meir and Seligman 1979). Grazing systems along the gradient demonstrate human–landscape relationships and their eﬀects on ecosystem services. The chain of relationships is as follows: Landscape diversity ! distribution of soil moisture ! primary production/forage distribution ! livestock movement and performance

However, human–landscape relationships are not unidirectional, which implies only human utilization of ecosystem services. Grazing feeds back on landscape structure and ecosystem processes by soil trampling and imposing selective pressure on vegetation quantity and quality. Elsewhere we introduced the concept of dry, shrub-dominated grazing systems (Perevolotsky and Shachak 1999). Such systems are the end-product of the long process of human impact on landscape and natural ecosystems in the Old World (Perevolotsky 1999b). These systems are well adapted to livestock impact and one may consider grazing as already incorporated into the natural ecosystem rather than being an external source of disturbance to it.

Managed Services Water In an arid area covered by sand dunes most of the rainfall inﬁltrates into the soil and is transferred into soil moisture. Run-oﬀ water management for farming under these conditions is not possible. However, when the landscape is rocky and generates overland ﬂow as surface run-oﬀ, water management for agriculture is possible. Run-oﬀ water management is a good example of human–landscape interactions through management of an ecosystem service for food production. Basically, run-oﬀ agriculture intimates the intensiﬁcation of natural processes of source–sink relationships in diversiﬁed landscapes. It engages with attempts to increase, through management, the frequency and magnitude of run-oﬀ harvesting from the slopes and to divert water to human-made productive sinks in the valleys (Evenari et al. 1971). In Sinai, the high and rich water table in the ‘‘red’’ rock region enables local Bedouin to practice orchard agriculture, mostly of Rosaceae fruit trees (Perevolotsky 1981). It is, most likely, the agricultural potential of the ‘‘red’’ granite that enabled human settlement in the Mount Sinai area during the Byzantine period (A.D. 400–600). These settlements (chapels, farms, and monasteries) are located almost exclusively on the ‘‘red’’ granite as are most of the orchards (Perevolotsky and Finkelstein 1985). In the Negev, in order to provide water for humans and livestock, cisterns were dug in the limestone substrate. Channels diverted run-oﬀ water into

Management for Biodiversity

297

them and into agricultural ﬁelds. To increase the sink potential for water in the agricultural ﬁelds, a complex pattern of dams and terraces was constructed in the valleys. Deep soil between the dams and the terraces was capable of storing the run-oﬀ water generated on the rocky slopes. These activities modiﬁed landscape diversity. Human-made patchiness in the form of cultivated ﬁelds was added to the natural physical and biological patchiness (Evenari et al. 1971). Ancient run-oﬀ agriculture has been intensively studied in the Negev with the help of numerous archeological sites distributed in this region. The central Negev is interspersed with the remains of ancient channels and dams for diverting run-oﬀ water to agricultural ﬁelds (Evenari et al. 1971). When studying the complex network of channels and dams it is clear that the ancient farmers understood the relationships among landscape mosaic, runoﬀ, and soil moisture. They skillfully exploited the ecological mechanisms of the dry, patchy environment and converted the natural sinks into agricultural ﬁelds. The remains of these ancient agricultural systems are no longer functional, but they still have a positive eﬀect on natural species diversity by providing more mesic habitats within a very dry context. An experimental farm was founded in the 1950s on the basis of an ancient farm. Within a decade, this farm produced fruits and crops in quantity equivalent to modern desert agriculture at that time (Evenari et al. 1971). Local dwellers of the Ramat Hanadiv Park area also developed, more than 1000 years ago, a complex water system. They exposed the underground water stored in the tuﬃc layer and created a collecting system that fed agriculture ﬁelds and a public bath, and that provided drinking water to a local village. The remains of this system were excavated and reconstructed, and serve as a recreational attraction of the park. Recreation In countries such as Israel, where outdoor activities are on the rise, a large number of people beneﬁt from recreational services. Water-limited systems provide numerous recreation services including sightseeing, hiking, aesthetic beauty, wildlife observation, and cultural, intellectual, and spiritual inspiration. These services are highly dependent on landscape, species, and functional group diversity. Along the Sinai–Ramat Handiv gradient there are extensively managed (nature reserves) and more intensively managed (parks) recreation systems. The nature reserves are valuable recreation areas in which landscape diversities were generated by physical and biological processes. The combination of high landscape diversity and organism diversity is the foundation for a successful recreation system. In the parks, landscape and organism diversity are enhanced through management. The Forestry Department of the Jewish National Fund (JNF) initiated management activities in the heavily exploited areas of the northern Negev, in an attempt to convert them into recreational areas. Ecosystem management objectives of the JNF are to develop practices to reduce the leakage of

298

Biodiversity, Conservation, and Management

resources (mainly water and soil) in order to increase productivity and diversity in degraded areas. The main technique used concerns the establishment of contour-line catchments (shikim) which collect water from 10 to 15 m upslope of the catchment (Shachak et al. 1998). This management helps decrease leakage of resources by creating sink patches that are characterized by two features. One is a structure that prevents the ﬂow of run-oﬀ water, soil, organic matter, and nutrients outside the landscape unit. The other feature is a storage unit, which absorbs and maintains water and nutrients, thus developing a more productive and diverse patch (Boeken and Shachak 1994, 1998). In order to assure the ﬂow of water and matter from upslope to the productive patches, the run-oﬀ regime must be maintained. A dense cover of herbaceous vegetation that has been developed over time on the runoﬀ contributing area may absorb the ﬂowing material. Consequently, a Bedouin sheep herd was introduced to these areas in the spring in order to graze the herbaceous vegetation (Perevolotsky 1999b). Management activities increase the value of the land for the Negev inhabitants. These modiﬁed, patchy landscape systems, located near urban settlements, serve as green belts. Scenic roads, walking trails, and observation points fulﬁll needs of the nearby urban population for open landscapes. According to management schemes, some areas will also serve as rangeland for herders. At Ramat Hanadiv, long-term human activities including ﬁre, woodcutting, and land-clearing for cultivation, have enhanced the patchiness of the environment and in some cases even determined the patch nature (e.g., aﬀorestation). The ﬁnal outcome is a complex of patches with diverse vegetation formations. The fauna and ﬂora of Ramat Hanadiv have responded to the network of patches, making the maintenance of this spatial structure a principal management issue. Developed stands of Mediterranean scrubland and woodland are the favored habitat for nocturnal fauna—wild boar, small carnivores, and the newly introduced roe deer (Rosenfeld 1999; Woodley 2001). On the other hand, dense vegetation creates a ﬁre hazard that may destroy the park (as occurred in 1980). Open landscapes are also essential for the diversity of birds (Adar 1997) and plants (Hadar et al. 1999). In conclusion, park management should emphasize the maintenance of a patchy environment as its principal goal. The current ﬁre prevention management policy partly produces such patchiness (Perevolotsky 2001). This management scheme is composed of seasonal grazing by a cattle herd in large patches of the park, characterized by high standing biomass of dry herbaceous vegetation (valleys, tree plantations) and a system of fuel breaks established in the perimeter of the park (Gutman et al. 2001; Perevolotsky et al. 2001). The fuel breaks are strips from which most woody vegetation was manually removed, and very heavy grazing once a year controls regrowth (Perevolotsky et al. 1996). Both components of the ﬁre prevention management help to maximize biodiversity through the optimization of habitat complexity (grazed vs. ungrazed patches, dense vs. opened-up patches).

Management for Biodiversity

299

Human-Biodiversity Interactions—A Conceptual Model Table 17.1 provides an overview of the ﬁve case studies discussed in this chapter, emphasizing both the principal natural and human parameters involved. Based on this analysis we propose a uniﬁed conceptual scheme that captures the essence of the relationships among biodiversity (landscape and species diversity), ecosystem processes, management, and principal ecosystem services (ﬁg. 17.3). Landscape diversity, in this scheme, is deﬁned as a complex network of physical and biological patchiness, functionally interrelated through water ﬂow, and organism responses to geomorphological, hydrological, and climatic conditions. In addition to the natural processes shaping landscape diversity, human-introduced grazing, agricultural, and recreation systems also feed back on landscape diversity.

Table 17.1 An overview of the ﬁve case studies discussed in the text

Case Study

Patchiness

Landscape Function

Landscape Modiﬁcation by Humans

Processes Controlling Landscape Diversity

Processes Controlling Species Diversity

Mount Sinai

Physical

Source–sink relationshjips between slope and valley

Creation of sink patches for water

Geological and geomorphological

Water redistribution

Sede Boqer

Physical

Source–sink relationships beetween upper and lower slope

Creation of source and sink patches for water

Geological and geomorphological

Water redistribution

Lehavim

Physical & biological

Source–sink relationships on small scale

Shrub clearcutting

Geological and geomorphological; Shrub–crust relationships

Water redistribution; Seed redistribution

Sayeret Shaked

Biological

Source–sink relationships on small scale

Creation of sink patches for water

Geological and geomorphological; Shrub–crust relationships

Trampling/ crust breaking

Ramat Hanadiv

Physical & biological

Fire; Shrub clear-cutting

Geological

Fire woodcutting, grazing

300

Biodiversity, Conservation, and Management

Figure 17.3 A conceptual model of the relationship among biodiversity (species and landscape diversity), ecosystem processes, human activities, and ecosystem services. Bold arrows indicate direct and indirect eﬀects of human activities. Regular arrows indicate direct and indirect natural processes.

In water-limited systems, rainfall redistribution by natural and humanmodiﬁed landscape mosaics makes hydrological processes a major determinant of ecosystem services. Organism diversity patterns of natural and managed systems stem from soil moisture heterogeneity. Grazing, agricultural, and recreation activity involve managing interactions between landscape diversity, water ﬂow, and the resulting vegetation, to produce beneﬁts over time. By selecting techniques for water-enriched patch formation and water harvesting, overall system productivity and diversity become higher than by natural water redistribution alone. These techniques are suitable for integrating agricultural plots into the landscape (Mount Sinai, Sede Boqer) and for the introduction of tree groves into the recreation system (Sayeret Shaked). The model (ﬁg. 17.3) considers the signiﬁcance of biological diversity in relation to ecosystem processes, economic (grazing and run-oﬀ agriculture) and recreation activities. It focuses on landscape, species, and functional

Management for Biodiversity

301

diversity, and their relation to production and maintenance of ecological services that underpin human societies. We suggest that within landscape diversity, the important categories are physical and biological patchiness. The two categories aﬀect organism diversity in a similar manner. They provide, by water redistribution, a greater diversity of potential niches for plants and animal species. This is the main factor controlling species diversity of annual plants, within and between locations along the gradient. This diversity is essential for grazing and recreation systems. The model also emphasizes the essential role of functional diversity in the functioning of water-limited systems. We also suggest that within the functional groups of organisms common along the gradient, the two most signiﬁcant categories of species are the herbaceous vegetation that functions as keystone-process species and the shrubs that are essential for ecosystem resilience. Ecosystem management for grazing and recreation should ensure that functional diversity, at least at the level of these two broad categories, is maintained. The relationship between species diversity and functional group categories is of importance for the development of uniﬁed principles for ecosystem management. Species composition is location-speciﬁc, while organization of the ecosystem into annuals and shrubs is a common phenomenon in waterlimited systems. We suggest that species turnover in relation to changes in rainfall amount, patchiness, and water redistribution is a major factor that enables stability in preserving the patches of annuals and woody plants. While in dry habitats natural water redistribution and grazing maintain the spatial relations between the two functional groups, in more mesic habitats (wadis in the desert, Ramat Hanadiv) human intervention is required to prevent the shrubs from dominating the entire landscape (Perevolotsky 2001). The model illustrates that the complexity of the system due to landscape and species diversity can be reduced once the focus is on key processes that are dependent on diversity. The process of water redistribution is the outcome of landscape diversity while being the driving force of primary production. The latter is also aﬀected by annual species diversity. In relation to management, the model shows the synergetic operation between the key processes and anthropogenic factors that integrate grazing, agricultural and recreation systems into the landscape structure. Particular attention should be paid to the development of spatial structures by human activities that feed back to landscape and species diversity (e.g., Komas 1983). The development of spatial structures is a basic tool in biodiversity management for ecosystem services, as was shown along the Sinai–Ramat Hanadiv transect. Therefore, the development and maintenance of the adequate spatial structure or ‘‘patch management’’ could be considered as a uniﬁed principle in human–landscape relationships in drylands. In conclusion, we think that new approaches are necessary to deal with the complexity of biological diversity in managed water-limited systems. Our model, which was motivated by human–landscape relations along a geographical gradient, provides an example of a framework that reduces the com-

302

Biodiversity, Conservation, and Management

plexity associated with landscape and species diversity. Our approach is an attempt to deal with key ecosystem processes controlled by biodiversity. In this approach, the integration of human-made systems into the natural environment can be achieved only by preserving the key processes ensuring the functioning of the ecosystem.

Acknowledgment The authors wish to thank Yehoshua Shkedy for his valuable comments on a draft of this chapter.

References Adar, M. 1997. Ecological considerations in landscape management—the breeding birds of Ramat Hanadiv as a case study. Ecology and Environment 4: 15–24. (In Hebrew with English summary.) Alard, D., Bance, J.F., and Frileux, P.N. 1994. Grassland vegetation as an indicator of the main agro-ecological factors in a rural landscape—consequences for biodiversity and wildlife conservation in central Normandy (France). Journal of Environmental Management 42: 91–109. Bertiller, M.B. 1996. Grazing eﬀects on sustainable semiarid rangelands in Patagonia: the state and dynamics of the soil seed bank. Environmental Management 20: 123–132. Boeken, B., and Shachak, M. 1994. Desert plant communities in human-made patches—implications for management. Ecological Applications 4: 702–716. Boeken, B., and Shachak, M. 1998. Colonization by annual plants of an experimentally altered desert landscape: source-sink relationships. Journal of Ecology 86: 804–814. Cowling, R.M., Rundel, P.W., Lamont, B.B., Arroyo, M.K., and Arianoutsou, M. 1996. Plant diversity in mediterranean-climate regions. Trends in Ecology and Evolution 11: 362–366. Danin, A. 1977. The Vegetation of the Negev. Tel Aviv: Sifriat H’apoalim (in Hebrew). Dyson-Hudson, R., and Dyson-Hudson, N. 1980. Nomadic pastoralism. Annual Reviews in Anthropology 9: 15–61. Evenari, M., Shanan, L., and Tadmor, N.H. 1971. The Negev: The Challenge of a Desert. Cambridge, MA: Harvard University Press. Folke, C., Holling, C.S., and Perrings, C. 1996. Biological diversity, ecosystems and the human scale. Ecological Applications 6: 1018–1024. Garner, W., and Steinberger, Y. 1989. A proposed mechanism for the formation of ‘fertile islands’ in desert ecosystems. Journal of Arid Environments 16: 257–262. Gutman, M., Perevolotsky, A., Yonatan, R., and Gutman, R. 2001. Grazing as a management tool for prevention of ﬁre in open areas: Ramat Hanadiv Park. Ecology and Environment 6: 239–247. (In Hebrew, with English abstract.) Hadar, L., Noy-Meir, I., and Perevolotsky, A. 1999. The eﬀect of shrub clearing and grazing on the composition of a Mediterranean plant community: functional groups versus species. Journal of Vegetation Science 10: 673–682. Hadar, L., Noy-Meir, I., and Perevolotsky, A. 2000. Scale dependent eﬀects of fuelbreaks management on species diversity in Mediterranean garrigue. Journal of Mediterranean Ecology 1: 237–248.

Management for Biodiversity

303

Komas, J. 1983. Man’s impact upon the ﬂora and vegetation in central Europe. In: Holzner, W., Werger, M.J.A., and Kusima, I. (eds.), Man’s Impact on Vegetation. Dordrecht, The Netherlands: Dr. W. Junk Publishers. Loidi, J. 1999. Preserving biodiversity in the European Union: the habitats directive and its application in Spain. Plant Biosystems 133: 99–106. Malanson, G.P., and Cramer, B.E. 1999. Landscape heterogeneity, connectivity, and critical landscapes for conservation. Diversity and Distributions 5: 27–39. Moloney, K.A., and Levin, S.A. 1996. The eﬀects of disturbance architecture on landscape-level population dynamics. Ecology 77: 375–394. Monkkonen, M. 1999. Managing Nordic boreal forest landscape for biodiversity: ecological and economic perspectives. Biodiversity and Conservation 8: 85–99. Noss, R.F. 1990. Indicators for monitoring biodiversity: a hierarchical approach. Conservation Biology 4: 355–364. Noy-Meir, I. 1973. Desert ecosystems: environment and producers. Annual Review of Ecology and Systematics 4: 25–51. Noy-Meir, I., and Seligman, N.G. 1979. Management of semi-arid ecosystems in Israel. In: Walker, B.H. (ed.), Management of Semi-Arid Ecosystems. Amsterdam: Elsevier, pp. 113–160. O’Connor, T.G., and Pickett, G.A. 1992. The inﬂuence of grazing on seed production and seed banks of some African savanna grasslands. Journal of Applied Ecology 29: 247–260. Olsvig-Whittaker, L., Shachak, M., and Yair, A. 1983. Vegetation patterns related to environmental factors in a Negev Desert watershed. Vegetatio 54: 153–165. Osem, Y., Sternberg, M., Perevolotsky, A., and Kigel, J. 1999. Species composition and abundance of annual plants in the seed-bank and in the vegetation in a semiarid region in the northern Negev as aﬀected by sheep grazing. Ecology and Environment 5: 180–189. (In Hebrew, with English abstract.) Perevolotsky, A. 1981. Orchard agriculture in the high mountain region of southern Sinai. Human Ecology 9: 331–357. Perevolotsky, A. 1987. Territoriality and resource sharing among the Bedouin of southern Sinai: A socio-ecological interpretation. Journal of Arid Environments 13: 153–161. Perevolotsky, A. 1999a. Natural conservation, reclamation and livestock grazing in the northern Negev: contradictory or complementary concepts. In: Hoekstra, T.W. and Shachak M. (eds.), Arid Lands Management: Towards Ecological Sustainability. Urbana: University of Illinois Press. Perevolotsky, A. 1999b. Interrelationships between conservation, landscape development and grazing in the northern Negev. Ecology and Environment 5: 190–199. (In Hebrew, with English abstract.) Perevolotsky, A. 2001.The management of the Ramat Hanadiv Park: an overview. Ecology and Environment 6: 287–289. (In Hebrew, with English abstract.) Perevolotsky, A., and Finkelstein, I. 1985. The southern Sinai exodus route in ecological perspective. Biblical Archaeology Review 11: 26–41. Perevolotsky, A., and Shachak, M. 1999. Epilog. Ecology and Environment 5: 200–202. Perevolotsky, A., Perevolotsky, Ayelet and Noy-Meir, I. 1989. Environmental adaptation and economic change in a pastoral nomadic society: The case of the Jebaliyah Bedouin of the Mt. Sinai region. Mountain Research and Development 9: 153–164. Perevolotsky, A., Ettinger, E., Shwartz, R., Yonatan, R., Gutman, M., and Altshuler, I. 1996. Fuel breaks for ﬁre prevention: an evaluation of a manage-

304

Biodiversity, Conservation, and Management

ment model. Ecology and Environment 3: 103–112. (In Hebrew, with English abstract.) Perevolotsky, A., Ettinger, E., Shwartz, R., and Yonatan, R. 2001. Management of fuel breaks in the Israeli Mediterranean ecosystem: the case of Ramat Hanadiv Park. Ecology and Environment 6: 248–251. (In Hebrew, with English abstract.) Rosenfeld, A. 1999. Wild Boars in Ramat Hanadiv and Alona Hills. Tel Aviv: Society for the Protection of Nature in Israel. (In Hebrew, with English abstract.) Russell, S.K., and Schupp, E.W. 1998. Eﬀects of microhabitats patchiness on patterns of seed dispersal and seed predation of Cercocarpus ledifolius (Rosaceae). Oikos 81: 434–443. Shachak, M., and Lovett, G.M. 1998. Atmospheric particle deposition to a desert ecosystem and its implications for conservation and management. Ecological Applications 8: 455–463. Shachak, M., and Pickett, S.T.A. 1997. Linking ecological understanding and application: patchiness in a dryland system. In: Pickett, S.T.A., et al. (eds.), The Ecological Basis of Conservation. New York: Chapman and Hall, pp. 108–119. Shachak, M., Sachs, M., and Moshe, I. 1998. Ecosystem management of desertiﬁed shrublands in Israel. Ecosystems 1: 475–483. Shachak, M., Pickett, S.T.A., Boeken, B., and Zaady, E. 1999. Managing patchiness, ecological ﬂows, productivity and diversity in drylands: concepts and applications in the Negev Desert. In: Hoekstra, T.W. and Shachak M. (eds.), Arid Lands Management: Towards Ecological Sustainability. Urbana: University of Illinois Press, pp. 254–264. Sharon, D. 1972. The spottiness of rainfall in a desert area. Journal of Hydrology 17: 161–175. Shmida, A., and Arbel, O. 1979. The Vegetation of the High Mountains of Southern Sinai. Tel Aviv: SPNI (in Hebrew). Shmida, A., Evenari, M., and Noy-Meir, I. 1986. Hot desert ecosystems: an integrated view. In: Evenari, M., Noy-Meir, I., and Goodel, D. (eds.), Hot Desert Ecosystems, Amsterdam: Elsevier. Tipping, R., Buchanan, J., Davies, A., and Tisdall, E. 1999. Woodland biodiversity, palaeo-human ecology and some implications for conservation management. Journal of Biogeography 26: 33–43. Trabaud, L., and Galtie, J.F. 1996. Eﬀects of ﬁre frequency on plant communities and landscape pattern in the Massif des Aspres (southern France). Landscape Ecology 4: 215–224. Ungar, E.D., Perevolotsky, A., Yonatan, R., Barkai, D., Hefetz, Y., and Baram, H. 1999. Primary production of natural pasture of the hilly northern Negev—contributory factors and management implications. Ecology and Environment 5: 130–139. (In Hebrew, with English abstract.) Walker, B.H. 1988. Autecology, synecology, climate and livestock as agents of rangeland dynamics. Australian Rangeland Journal 10: 69–75. Woodley, B. 2001. Reintroduction of roe deer Capreolus capreolus to the Carmel Ridge: a study of acclimatization and dispersal. Ecology and Environment 6: 267–272. (In Hebrew, with English abstract.) Yair, A., and Shachak, M. 1987. Studies in watershed ecology of an arid area. In: Wurtele, M.O. and Berkofsky, L. (eds.), Progress in Desert Research. Totowa, NJ: Rowman and Littleﬁeld, pp. 146–193. Zaady, E., and Shachak, M., 1994. Microphytic soil crust and ecosystem leakage in the Negev Desert. American Journal of Botany 81: 109.

18 Uniﬁed Framework III Human Interactions with Biodiversity Anna A. Sher Bruce M. Kahn Christopher R. Dickman

C

onsidering humans as components of ecosystems is not new; geographers have been doing it in human ecology departments for decades (see Field and Burch 1988). There have also been many volumes dedicated to the subject (McDonnel and Pickett 1993, Schnaiberg and Gould 1994, Catton 1982, Wilson 1988). Recently there has been a development in the ﬁeld of ecology to consider humans as a part of ecosystems, rather than simply agents of destruction, including the full complexity of human interactions (including social, cultural, and economic) with the environment (chapter 17 this volume, Folke et al. 1996, Turner and Carpenter 1999, Pickett et al. 1999, Haeuber and Ringold 1998). The goal of this chapter is to provide a framework for the types of interactions between humans and biodiversity. We use biodiversity as an umbrella term encompassing genetic, species, and landscape diversities (chapter 1 this volume). In particular, we emphasize human–biodiversity interactions in the context of arid and semiarid ecosystems. In part I, we analyze the various types of human–biodiversity interactions. In part II we suggest a framework for the study of these interactions.

Part I: Types of Interactions Not only do humans have the power to aﬀect biodiversity, but biodiversity impacts humans as well. The nature of this reciprocal relationship can be positive and/or negative. Our reference to positive and negative impacts on 305

306

Biodiversity, Conservation, and Management

biodiversity will usually be in the mathematical sense, that is, an increase or decrease in species and landscape diversity. However, we must be careful not to put a value judgment on such numbers. Increases in species or habitat diversity are not necessarily desirable for all ecosystems or management goals. All the elements of biodiversity are not equal in terms of ecological and economical value. For example, restoration eﬀorts for a few endemic species may be detrimental to other, nonendemic species, resulting in less species diversity. This may especially be true when the diversity of weedy species that have taken over a disturbed area is threatened by restoring historical conditions. In this case, a lower level of biodiversity that includes native species may be more desirable than a higher level of nonendemic species diversity. That is, promoting gamma diversity on a global scale may sometimes mean sacriﬁcing alpha (i.e., local) diversity. Biodiversity management goals must be clearly deﬁned in both mathematical terms and value to ecosystems and/or humans (Hobbs 1998). A speciﬁc example of the potential conﬂict between these two approaches can be found in the case of a plant invasion in arid American Southwest riparian systems. Invasive Tamarix can sometimes increase species diversity while destroying the native ecosystem; Tamarix thickets have high diversity of invertebrates (Stevens 1990) and may have higher bird species diversity (Ellis 1995) than native tree communities, including supporting the endangered willow ﬂycatcher. However, Tamarix trees have also been associated with the decline in ﬂoral and faunal endemics, changing local hydrology and soil characters, and reducing human land-use options, all of which are considered negative (Crawford et al. 1993). Therefore, deﬁning positive aspects of biodiversity in terms of numerical values may not reﬂect our goals of management or our values of ecosystems, for instance, endemism (Perlman and Adelson 1997). Another problem that can occur when considering human interactions with biodiversity is in characterizing any agent of impact as universally positive or negative. For example, trampling by goats increases landscape diversity of soil crusts and vegetation on sand dunes (Eldridge et al. 2000, Olsvig-Whittaker et al. 1993, West 1986). The number of species in the community of the disturbed habitat increases. However, if the whole region is disturbed, we create an undesirable situation because the unique stabilized sand dune community is destroyed (Abramsky 1988). Therefore, the most desirable condition both mathematically and in human values is a landscape mosaic of protected and disturbed units, instead of universal inclusion or exclusion of goats (Seligman and Perevolotsky 1994). Furthermore, eﬀects may not be geographically generalized; the result of crust disturbance on biodiversity appears to be diﬀerent in the American Southwest, where even low disturbance levels have been shown to be detrimental (West 1990). Even more controversial and unresolved are the eﬀects of livestock grazing (Homewood and Rodgers 1987). Thus, we will take care to avoid gross generalizations, especially in areas where the scientiﬁc evidence is mixed.

Unified Framework III

307

Human Impacts on Biodiversity: Negative The most discussed aspect of human interactions with biodiversity is the negative impact humans have on species diversity. Road building, agricultural tilling, chemical inputs, deforestation, and urbanization all change the environment, reduce resource availability or heterogeneity, and therefore, species diversity (Wilson 1988). As this aspect has been well covered elsewhere (e.g., McDonnell and Pickett 1993) we will not attempt to provide an exhaustive review here. The dramatic negative eﬀects on species diversity are best illustrated by current extinction rates, estimated to be 100–1000 times higher than prehuman levels and projected to reach up to 10,000 times higher, based on extinction of currently endangered species (Pimm et al. 1995). There are broad negative ecosystem consequences of lost species diversity, including reductions of net primary productivity, nutrient cycling, and resistance to drought and disturbance (Vitousek et al. 1986, Chapin et al. 1998). Examples of negative eﬀects of humans on species diversity in arid lands have focused primarily on livestock grazing (Noy-Meir 1973, Seligman and Perevolotsky 1994, Ellis and Swift 1988), ﬁre disturbance (Hadar et al. 1999, Hodgkinson et al. 1984, Flannery 1994, Perevolotsky 1996), and modiﬁed water regimes (Dregne 1996, Le Houerou et al. 1988). However, other factors such as hunting of ‘‘undesirable’’ species also may negatively aﬀect species diversity. For example, extermination of desert snakes directly aﬀects biodiversity, not only by removing those species, but also by cascading trophic eﬀects (Thomas and Shaw 1991, Huenneke and Noble 1996). Humans also negatively aﬀect biodiversity in arid lands by homogenizing the landscape. This includes a range of activities, such as habitat destruction and fragmentation, which lead to desertiﬁcation and the disruption of ecological system functioning (Morton 1990, Dale et al. 2000, Shachak et al. 1998, Turner 1989). We must strive to reduce misuse of the environment that results in the reduction of biodiversity. In addition to the ethical reasons to do so, we depend on many ecosystem services and goods that result directly or indirectly from biodiversity (Christensen 1996). While management practices should strive to reduce our footprint on the environment, the view that all human activities are negative limits the alternatives for biodiversity management (Hobbs 1998, chapters 16, 17 this volume). In the perspective we present below, we hope to provide a more balanced view of human interactions with biodiversity than is often argued. We consider positive human impacts on the environment, as well as the reciprocal aspect of the relationship between humans and ecosystems.

Human Impacts on Biodiversity: Positive There are both inadvertent and purposeful human impacts on biodiversity, including conservation, reclamation, and reconciliation (chapter 16). Conservation encompasses the maintenance of species and landscape diver-

308

Biodiversity, Conservation, and Management

sities and ecosystem processes (Hobbs 1998). In cases in which biodiversity is currently high and stable, conservation of current conditions may be the best approach. Such conservation can be passive or active. Passive conservation simply entails monitoring biodiversity and limiting the initiation of threatening processes or activities. In contrast, active conservation may depend on maintaining time-honored methods of management. In central Australia, for example, small-scale patch burning has been carried out by Aboriginal peoples for millennia to improve access to game and also to ‘‘look after country’’ (Baker et al. 1993). Small ﬁres are widely believed to increase biodiversity, both by triggering regeneration of biotic and nutrient resources in the burned sites and by maintaining mosaics of habitats at diﬀerent seral stages. Where humans have reduced biodiversity, our goals may shift from the conservation approach to reclamation. Reclamation entails restoration of ecosystem processes by manipulating both landscape and species diversity. This can be done by assessing the role that biodiversity has on ecosystem dynamics that leads to management strategies for modifying biodiversity (Dickman 1999, Perlman and Adelson 1997). However, re-creating historical conditions often may not be the best option for either biodiversity or human needs, and in many cases is impossible to implement (chapter 16 this volume). Reconciliation (sensu Rosenzweig 2003) is a management approach that seeks to design habitats to reﬂect the realities and needs of humans while also promoting species diversity. In arid lands, conservation and reclamation have been implemented mainly by removing human-associated activities such as livestock grazing and farming from areas that have been designated as nature reserves (Little 1996). The cessation of grazing is based on the assumption that livestock grazing has negative impacts on species diversity. Although overgrazing, particularly in more arid zones, is arguably a major force in biodiversity loss, recent studies suggest that it is not always true that livestock grazing will negatively aﬀect the ecosystem (chapter 14 this volume). The creation of exclosures to protect lands from herbivory has had mixed results. Grazing eﬀects are not straightforward or without exceptions (Homewood and Rodgers 1987). There are a surprising number of cases in which grazing by human-tended or introduced herbivores had a negligible or even a positive eﬀect on species diversity. This has been shown in the semiarid African savanna (Western and Gichohi 1994), the American Southwest (Milchunas et al. 1988), and the Middle East (Noy-Meir 1990, Zaady et al. 2001). However, in a recent literature review, a pattern was found that in low productivity habitats, increases in grazing intensity was generally associated with decreases in biomass and species diversity (Proulx and Mazumder 1998). Overgrazing is still widely considered an agent of desertiﬁcation and biodiversity loss. Species diversity generally may be dependent on management approaches that provide spatial and temporal heterogeneity of grazing pressure, thereby making the system less vulnerable to drought and shrub encroachment (Perrings and Walker 1995). The upper bounds of tolerable

Unified Framework III

309

grazing intensities and patterns will exist, even in ecosystems where plant communities have evolved as grazing resilient systems (Noy-Meir 1975). There are dangers, therefore, in making broad generalizations, especially across arid and mesic systems, on the impact of human activities such as livestock rearing or species introductions on biodiversity. Depending on scale and how the eﬀect is measured and valued, the same human activities may have either negative or positive impacts on biodiversity.

Biodiversity Impacts on Humans: Negative Although not immediately obvious, there are several ways in which biodiversity can negatively aﬀect humans, with regards to health, economics, aesthetics, and other aspects of human existence on this planet. Because we are deﬁning biodiversity in its broadest sense, it is important to acknowledge that not all humans ﬁnd all species or habitats desirable. Direct negative eﬀects of species diversity can be through agricultural pests, weeds, and predators. Also, there are direct negative eﬀects on humans of increased diversity of disease-causing microorganisms (Ford 1971, Freeland 1993). The direct negative eﬀects of increased emergence of virulent diseases may be linked to increased diversity of disease-causing microorganisms (Mack et al. 2000). The World Health Organization (1996) reports that since 1976, 30 diseases have emerged that are new to medicine. Indirect negative impacts of biodiversity generally involve conﬂict of interest between a human group and its natural environment. Increasing protected areas for biodiversity may conﬂict with the needs of humans, especially when land where a human group lives or draws income is taken away or threatened by animals encroaching from protected areas. Large predators may enter human settlements or other animals may become agricultural pests. It is also possible that increases in biodiversity that are positive for one human group may be negative for another human group. Increases in elephant or rhino abundance in African savannas may be positive for ecotourism and those who beneﬁt from it, but are negative for farmers whose crops are destroyed by the animals (Owen-Smith 1998, Murphree 1995). In the Negev desert of Israel, Bedouin tribes once depended on large tracts of land for livestock grazing. It remains an important activity today, both culturally and economically (Ginguld et al. 1997), but the establishment of nature reserves decreased rangeland availability to the Bedouin (Briggs 1995). The indigenous people of arid and semiarid areas are often poor, and there may be a tendency for the needs of these groups to be marginalized, as they are less likely to be in positions to aﬀect management of the lands on which they live (Behnke and Scoones 1993; Ellis and Swift 1988). In contrast, conservation values are of interest to other, wealthier groups, therefore imposing a need for a public-political process to resolve the conﬂict. To what extent human needs should or can be considered is a delicate and complicated issue that should not be ignored. Any management plan must consider not only

310

Biodiversity, Conservation, and Management

biodiversity, but also the social, economic and human health aspects of land use (Dale et al. 2000). Biodiversity Impacts on Humans: Positive We can consider the positive eﬀects of biodiversity on humans in terms of the way it inﬂuences ecosystem processes. These processes can be of positive value by directly providing goods and services to humans, or by indirectly stabilizing ecosystem functioning (chapter 10). These two types of positive eﬀects have gained more attention and clariﬁcation recently in the context of ‘‘ecosystem goods and services’’ and ‘‘ecosystem resilience’’ (Daily 1997, Christensen et al. 1996, Holling 1986). Ecosystem goods and services are the ways in which biodiversity provides humans with economic, aesthetic, health, and other beneﬁts (Christensen et al. 1996). Many human uses of the environment depend, directly or indirectly, on biodiversity, particularly the role it plays in buﬀering disturbances to ecosystems (Folke et al. 1996). An example of biodiversity–ecosystem services interactions is pollination of agricultural crops by wild pollinators (Nahban and Buchman 1997). A biodiversity–resilience interaction provides an ecosystem with alternative species within functional groups, or landscape diversity to maintain ecosystem processes (Frost et al. 1995, Peterson et al. 1998). For example, in semiarid grasslands in east and southern Africa, ecosystem resilience to drought disturbance has been shown to heavily depend on the maintenance of grass functional group diversity (Westoby 1989, Ellis and Swift 1988, Dodd 1994). In addition to the importance of biodiversity for maintaining the ecosystems from which humans beneﬁt, species diversity itself can be a direct service. Hidden reserves of desert biodiversity, such as genetic material for developing drought resistance in crops or discovering new medicines are biodiversity–dependent ecosystem services (Folke et al. 1996). Diversity of habitat types and species also increases the aesthetic value of an area; the diverse and beautiful ‘‘desert blooms’’ that occur after a good rain season are only one example. Finally, there exists the elusive but signiﬁcant psychological and spiritual beneﬁt that humans derive from biodiversity. This is a nonmeasurable eﬀect, but should not be downplayed in importance. John Muir (1913) asserted that the rights of nature are on a par with human rights. If one subscribes to this view, it can be argued that the maintenance of biodiversity is an important indicator of the general moral state of a society.

Part II: A Framework for Human–Biodiversity Interactions In the ﬁrst part of this chapter, we elucidated the various types of human– biodiversity interactions. The emphasis was on diﬀerent aspects of biodiversity. However, we generally treated humans as one entity.

Unified Framework III

311

In the second part, we intend to emphasize the hierarchical nature of human grouping and will suggest a general framework for human–biodiversity interaction studies. Individual and Group Interactions with Biodiversity The strength, nature, and direction of human interactions with biodiversity depend on a multitude of factors, including not only biodiversity and its ecological eﬀects, but also the complexities of human biology, culture, psychology, and sociology. Several conceptual models have been presented in the literature to organize and quantify human behavior in response to ecosystems. An increasing number of models focus on ecosystem dynamics in response to human activity (for a review of human ecology models see Machlis et al. 1997). Such frameworks are necessary for making sense of the list of possible important factors, the goals including: to understand patterns (pure knowledge), to gain predictive power, and to create management plans. Here, we will highlight those aspects and challenges of human interactions with biodiversity that focus on the human component of the interactions. We use a hierarchical approach to understand the interactions between humans and biodiversity, and thereby address individual and socioeconomic levels of interactions with biodiversity. Just as ecological systems can be hierarchically organized (Allen and Starr 1982), so are intrahuman processes and their interactions with biodiversity. Every level of biodiversity has the potential of interacting with each level of human organization (ﬁg. 18.1). The hierarchical nesting of these levels of organization can provide a framework to determine the direction and strength of human interactions with biodiversity. Considering human eﬀects in terms of a hierarchy is not new (Ahl and Allen 1996). If we recognize the hierarchical nature of our interactions with biodiversity, we may be able to better understand the nature, strength, and direction of those interactions. Human behavior that will impact biodiversity must be considered on both individual and hierarchical group levels. Individual human behavior will be determined by values, attitudes, education, and experience (Heberlein 1974). In addition to the role an individual’s behavior has in forming groups, group behavior will be strongly driven by the institutional structures in place and the policies and programs enacted by local, regional, and global governing bodies (Edelman 1964). Community and societal factors that inﬂuence human interactions with biodiversity include: education and media (Molotch 1979), value systems and power structures (Gaventa 1980, Stone 1980), and economics (Goodland and Daly 1996). The hierarchical nature of human biodiversity interactions in arid lands can be demonstrated by examining the Bedouin society in the Middle East. Bedouin livestock grazing is identiﬁed as a potentially important means of interactions between humans and biodiversity (Briggs 1995). Ginguld et al. (1997) described livelihood strategies of Bedouin herd managers in the Negev and how livestock rearing varied among individuals, corresponding to var-

312

Biodiversity, Conservation, and Management

Figure 18.1 The hierarchical nature of both biodiversity and human activities and their interactions. Each level of the two hierarchies has the potential of interacting with any other level.

ious levels of social and political constraints. This interaction aﬀects both the people and biodiversity, and can do so in both positive and negative ways. The most obvious way is a negative eﬀect of human behavior on biodiversity by promoting overgrazing. However, how and whether this occurs will depend on several factors that can be understood by considering the issue in a hierarchical way. On the lowest level, an individual Bedouin herder will make choices about the size of the herd, as well as where and for how long the herd grazes. These choices are based on tradition, values, the herder’s understanding of resource availability, and what patterns of movement the landscape allows. Perceived species and landscape diversities can inﬂuence an individual’s behavior. There is evidence that Bedouin herders do make choices that help to minimize negative impact on plant communities, and Bedouin sheep grazing may have less of an impact on community productivity than previously thought (Zaady et al. 2001). On the next level of human organization, individuals and biodiversity will be inﬂuenced by community-level decisions and actions, such as what varieties of sheep are available for sale, where resources like water are available, and education. Diminishing biodiversity on all levels can force changes on this community level. Finally, society and government have a strong impact on the Bedouin community in ways that aﬀect individuals’ interactions with the environment. This is done by compelling settlement (therefore concentrating grazing pressure in one area), increasing interaction with non-Bedouins (which has both positive and negative potential for inﬂuencing grazing impact on biodiversity, e.g., education programs), and providing economic alternatives. Thus, both research on the eﬀects of grazing and policy to manage these

Unified Framework III

313

arid ecosystems are well advised to consider the hierarchical nature of the interaction before making broad generalizations.

The Role of Economics in Human–Biodiversity Interactions in Arid Lands In general, the ways in which humans interact with biodiversity in arid lands are similar to the ways we interact with biodiversity in temperate or tropical zones; considerations of biological processes, social structure, values, economics, and institutions are relevant regardless of location. We will argue that arid systems will have some unifying key issues and relative emphasis on diﬀerent aspects of human–biodiversity interactions than more mesic zones (chapter 14). Likewise, we cannot discount the variation of interactions within arid lands—both between geographical locations on a large scale and variation within a single area. Arid lands have many characteristics that bear upon how human activity interacts with biodiversity: 1. Human populations are historically less concentrated in arid zones and often include nomadic peoples. 2. Water is a limiting ecosystem service for humans, both directly and by making other ecosystem services possible. To say that arid and semiarid ecosystems have common issues with regard to human interactions with biodiversity is to say that water limitation is an important unifying factor. Because water limitation aﬀects renewable resources (e.g., forest, agricultural, and horticultural products), arid areas are less likely to be highly productive, both in the biomass sense, and in the economic sense (Fisher et al. 1999). 3. Some arid lands are economically productive through their nonrenewable geological goods and services, usually mineral resources such as oil or diamonds, because of the very high capital value of such resources (Beaumont 1993). We might divide arid lands into two categories, those with mineral resources (rich) and those without (poor). Most arid lands fall in the second category, and even in places rich with mineral resources, wealth tends to be concentrated in a few individuals, with poverty being a signiﬁcant problem (Sahn 1999). The poverty of arid regions has signiﬁcant implications for how humans interact with biodiversity. If an arid zone is not buﬀered within a larger political boundary that contains richer areas, the people of that area will have fewer management choices, as many management approaches require money or at least a stable economy. Some argue that ecological sustainability in poorer areas will only be possible through redistribution of capital from richer to poorer countries or regions (Goodland and Daly 1996). Furthermore, natural resource poverty in arid zones means that peoples in these areas have to rely on other ecosystem services that are not resource-

314

Biodiversity, Conservation, and Management

based—an issue that even those rich countries that rely on nonrenewable resources will eventually also have to consider. A summary of the role of economics in human–biodiversity interactions in arid lands is shown in table 18.1. When we consider a biodiversity–economy relationship in arid lands we have to relate it to the source of capital. As a framework, we suggest four sources of capital: ecosystem goods and services, geological goods and services, space use, and information. Which sources of capital are chosen can have diﬀerent implications for biodiversity. When considering ecosystem goods and services, implications are dependent on management. For example, as we have already discussed, the eﬀect of grazing may be either positive or negative (Homewood and Rodgers 1987). Sometimes the most proﬁtable income developments in arid lands disrupt ecosystem processes and/or are not sustainable (e.g., mining, some uses of open space). All of these activities emphasize the conﬂict between economy and biodiversity in arid lands. A bright aspect for biodiversity in arid lands is the development of information goods and services. The high-technology industry could be a good alternative for desert development and conservation in drylands. This is because the information industry does not need large tracts of land, does not pollute, and is proﬁtable. Using information technology as a capital would shift economic reliance away from ecosystem services. This would provide direct beneﬁt to biodiversity by reducing impact on the environment.

Table 18.1 Categories of income development in arid lands, as relates to both economics and biodiversity. Classiﬁcation as good or bad for economics or biodiversity is based on the assumptions that (a) water is limiting and (b) sale of mineral resources is proﬁtable Economically Strong

Good for Biodiversity

Food Water Aesthetics

No

Yesa Nob

Nonrenewable

Oil and other mineral resources

Yes

No

Renewable

Military activity Recreation Agriculture

Yes

No

Nonrenewable

Waste disposal Human settlement

Yes

No

Renewable

Hi-tech

Yes

Yes

Source of Capital

Renewable?

Examples

Ecosystem goods and services

Renewable

Geological goods and services Space

Information a

With sustainable ecosystem management.

b

With mismanagement.

Unified Framework III

315

Conclusions Based on our analysis of the human–biodiversity interactions and the general framework we have suggested, we propose that future studies concentrate on the following questions: 1. What is the role of the source of capital (table 18.1) in conserving biodiversity? 2. How can we consolidate the needs of human residents with biodiversity? What are the particular concerns of nomadic peoples in these areas (e.g., Bedouin, Maasai)? 3. How can we link the hierarchical natures of human society and of biodiversity? 4. What is the eﬀect of social inequality on biodiversity in arid lands? 5. What is the eﬀect of changing the water regime by humans on biodiversity? 6. What are the eﬀects of increased access (e.g., roads) to arid lands on biodiversity? 7. What are eﬀects of exotic species on biodiversity? We assume that to answer the questions raised, there must be an interdisciplinary approach. Cooperative research among sociologists, economists, ecologists, stakeholder groups, and politicians is a prerequisite. In summary, we have attempted in this chapter to bring attention to the fact that our relationship with biodiversity is hierarchical, diverse, and reciprocal. By moving beyond the dogmas that all human interactions with the environment are negative for biodiversity and that humans are a monolithic entity, we hope to improve biodiversity studies and the ability to communicate with public and economic sectors.

References Abramsky, Z. 1988. The role of habitat and productivity in structuring desert rodent communities. Oikos 52: 107–114. Ahl, V., and T.F.H. Allen. 1996. Hierarchy Theory: A Vision, Vocabulary, and Epistemology. Columbia University Press, New York. Allen, T.F.H., and T.B. Starr. 1982. Hierarchy: Perspectives for Ecological Complexity. University of Chicago Press, Chicago. Baker, L., S. Woenne-Green, and the Mutitjulu Community. 1993. Anangu knowledge of vertebrates and the environment. Pp. 79–132 in Uluru Fauna: The Distribution and Abundance of Vertebrate Fauna of Uluru (Ayers Rock-Mount Olga) National Park, N.T. (J.R.W. Reid, J.A. Kerle, and S.R. Morton, eds.). Australian National Parks and Wildlife Service, Canberra. Beaumont, P. 1993. Drylands. Routledge Publishers, London. Behnke, R., and I. Scoones. 1993. Rethinking range ecology: Implications for rangeland management in Africa. Pp. 1–30 in Range Ecology at Disequilibrium

316

Biodiversity, Conservation, and Management

(R. Behnke, I. Scoones, and C. Kervan, eds.). Overseas Development Institute, London. Briggs, J. 1995. Environmental resources: their use and management by the Bedouin of the Nubian Desert of southern Egypt. Pp. 61–68 in People and Environment in Africa (T. Binns, ed.). John Wiley and Sons, Chichester. Catton, W.R., Jr. 1982. Overshoot: The Ecological Basis of Revolutionary Change. Illinois Press, Urbana. Chapin, F.S., O.E. Sala, I.C. Burke, J.P. Grime, D.U. Hooper, W.K. Lauenroth, A. Lombard, H.A. Mooney, A.R. Mosier, S. Naeem, S.W. Pacala, J. Roy, W.L. Steﬀen, and D. Tilman. 1998. Ecosystem consequences of changing biodiversity: Experimental evidence and a research agenda for the future. BioScience 48: 45–52. Christensen, N.L., A.M. Bartuska, J.H. Brown, S.R. Carpenter, C.D’Antonio, R. Francis, J.F. Franklin, J.A. MacMahon, R.F. Noss, D.J. Parsons, C.H. Peterson, M.G. Turner, and R.G. Woodmansee. 1996. The report of the Ecological Society of America committee on the scientiﬁc basis for ecosystem management. Ecological Applications 6: 665–691. Crawford, C.S., A.C. Cully, R. Leutheuser, M.S. Sifuentes, L.H. White, and J.P. Wilber. 1993. Middle Rio Grande ecosystem: Bosque biological management plan. Report submitted to U.S. Fish and Wildlife Service, Albuquerque, NM. Daily, G.C. 1997. Introduction: What are ecosystem services? Pp. 1–10 in Nature’s Services: Societal Dependence on Natural Ecosystems (G. Daily, ed.). Island Press, Washington, D.C. Dale, V.H., S. Brown, R.A. Haeuber, N.T. Hobbs, N. Huntly, R.J. Naiman, W.E. Riebsame, M.G. Turner, and T.J. Valone. 2000. Ecological principles and guidelines for managing the use of land. Ecological Applications 10: 639–670. Dickman, C.R. 1999. Rodent-ecosystem relationships: a review. Pp. 113–133 in Ecologically-Based Rodent Management (G.R. Singleton, L.A. Hinds, H. Leirs and Z. Zhang, eds.). Australian Centre for International Agricultural Research, Canberra, Australia. Dodd, J.L. 1994. Desertiﬁcation and degradation in Sub-Saharan Africa. The role of livestock. BioScience 44(1): 28–34. Dregne, H.E. 1996. Desertiﬁcation: People Problem. Forum for Applied Research and Public Policy. Edelman, M. 1964. Symbols and political quiescence. Pp. 22–43 in The Symbolic Uses of Politics. University of Illinois Press, Urbana. Eldridge, D.J., E. Zaady, and M. Shachak. 2000. Inﬁltration through three contrasting biological soil crusts in patterned landscapes in the Negev, Israel. Catena 40: 323–336. Ellis, J.E., and D.M. Swift. 1988. Stability of African pastoral ecosystems: Alternate paradigms and implications for development. Journal of Range Management 41: 450–459. Ellis, L. M. 1995. Bird use of saltcedar and cottonwood vegetation in the Middle Rio Grande Valley of New Mexico, U.S.A. Journal of Arid Environments 30: 339–349. Field, D.R., and W.R. Burch Jr. 1988. Rural Sociology and the Environment. Social Ecology Press, Middleton, WI. Fisher, J.T, M.S. Smith, R. Cavazos, H. Manzamilla, P.T. Ffolliott, D. Saltz, M. Irwin, T.W. Sammis, D. Swietlik, I. Moshe, and M. Sachs. 1999. Land use and management: research implications from three arid and semiarid regions of the world. Pp. 143–170 in Arid Lands Management: Toward Ecological Sustainability (T.W. Hoekstra, and M. Shachak, eds.). University of Illinois Press, Urbana.

Unified Framework III

317

Flannery, T. 1994. The Future Eaters. Reed Books, Sydney, Australia. Folke, C., C.S. Holling, and C. Perrings. 1996. Biological diversity, ecosystems, and the human scale. Ecological Applications 6: 1018–1024. Ford, J. 1971. The Role of Trypanosomiases in African Ecology. Clarendon Press, Oxford. Freeland, W.J. 1993. Parasites, pathogens and the impacts of introduced organisms on the balance of nature in Australia. Pp. 171–180 in Conservation Biology in Australia and Oceania (C. Moritz and J. Kikkawa, eds.). Surrey Beatty and Sons, Sydney, Australia. Frost, T.M., S.R. Carpenter, A.R. Ives, and T.K. Kratz. 1995. Species compensation and complementarity in ecosystem function. Pp. 224–239 in Linking Species and Ecosystems (C.G. Jones and J.H. Lawton, eds.). Chapman and Hall, New York. Gaventa, J. 1980. Power and Powerlessness: Quiescence and Rebellion in an Appalachian Valley. Part I. ‘‘Power and Participation’’ and ‘‘The Case of a Central Appalachian Valley’’ and Conclusions, pp. 3–44 and 252–261. University of Illinois Press, Urbana. Ginguld, M., A. Perevolotsky, and E.D. Ungar, 1997. Living on the margins: Livelihood strategies of Bedouin herd-owners in Northern Negev, Israel. Human Ecology 25(4): 567–591. Goodland, R., and H. Daly. 1996. Environmental sustainability: universal and nonnegotiable. Ecological Applications 6: 1002–1017. Hadar, L., I. Noy-Meir, and A. Perevolotsky. 1999. The eﬀect of shrub clearing and grazing on the composition of a Mediterranean plant community: functional groups versus species. Journal of Vegetation Science 10: 673–682. Haueber, R., and P. Ringold. 1998. Ecology, the social sciences and environmental policy. Ecological Applications 8(2): 330–331. Heberlein, T.A. 1974. The three ﬁxes: technological, cognitive and structural. In Water and Community Development: Social and Economic Perspectives (D. Field, J.C. Barren, and B.F. Long, eds.). Ann Arbor Science Publishers, Ann Arbor, MI. Hobbs, R.J. 1998. Managing ecological systems and processes. Pp. 459–483 in Ecological Scale: Theory and Applications (D.L. Peterson and V.T. Parker, eds.). Columbia University Press, New York. Hodgkinson, K.C., G.N. Harrington, G.F. Griﬃn, J.C. Noble, and M.D. Young. 1984. Management of vegetation with ﬁre. Pp. 141–156 in Management of Australia’s Rangelands (G.N. Harrington, A.D. Wilson, and M.D. Young, eds.). CSIRO Publications, Melbourne, Australia. Holling, C.S. 1986. The resilience of ecosystems: local surprise and global change. Pp. 292–317 in Sustainable Development of the Biosphere (W.C. Clark and R.E. Munn, eds.). Cambridge University Press, Cambridge, UK Homewood, K.M., and W. Rodgers. 1987. Pastoralism, conservation and the overgrazing controversy. Pp. 111–128 in Conservation in Africa: People, Policies and Practice (D. Anderson and R. Grove, eds.). Cambridge University Press, Cambridge, UK. Huenneke, L.F., and I. Noble. 1996. Ecosystem function of biodiversity in arid ecosystems. In Functional Roles of Biodiversity: A Global Perspective (H.A. Mooney, J.H. Cushman, E. Medina, O.E. Sala, and E.D. Schulze, eds.) John Wiley and Sons, New York. Le Houerou, H.N., R.L. Bingahm, and W. Skerbeck. 1988. Relationships between the variability of primary production and the variability of annual precipitation in world arid lands. Journal of Arid Environments 15: 1–8.

318

Biodiversity, Conservation, and Management

Little, P.D. 1996. Pastoralism, biodiversity, and the shaping of savanna landscapes in East Africa. Africa 66: 1–15. Machlis, G.E., J.E. Force, and W.R. Burch. 1997. The human ecosystem part I: The human ecosystem as an organizing concept in ecosystem management. Society and Natural Resources 10: 347–367. Mack, R., D. Simberloﬀ, W. M. Lonsdale, H. Evans, M. Clout, and F. Bazzaz. 2000. Biotic invasions: causes, epidemiology, global consequences and control. Issues in Ecology No. 5. McDonnell, M.J., and S.T.A. Pickett, eds. 1993. Humans as Components of Ecosystems: The Ecology of Subtle Human Eﬀects and Populated Areas. Springer-Verlag, New York (Paperback edition 1997.) Milchunas, D.G., O.E. Sala, and W.K. Laurenroth. 1988. A generalized model of the eﬀects of grazing by large herbivore on grassland community structure. American Naturalist 132: 87–106. Molotch, H., 1979. Media movements. Pp. 71–93 in The Dynamics of Social Movements (M.L. Zald and J.D. McCarthy, eds.) Winthrop, Cambridge, MA. Morton, S.R. 1990. The impact of European settlement on the vertebrate animals of arid Australia: A conceptual model. Proceedings of the Ecological Society of Australia 16: 201–213. Muir, J. 1913. The Story of My Boyhood and Youth. Hougton Miﬂin, New York. Murphree, M. 1995. Summary report on human elephant conﬂict in areas adjacent to Maputo Reserve. IUCN, DNFFB, CBNRM Advisor. Typescript, 5 pp. Nabhan, G.P., and S.L. Buchmann. 1997. Pollination services: Biodiversity’s direct link to world food stability. Pp. 133–150 in Nature’s Services: Societal Dependence on Natural Ecosystems (G. Daily, ed.). Island Press, Washington, D.C. Noy-Meir, I. 1973. Desert ecosystems: environment and producers. Annual Review of Ecology and Systematics 4: 25–51. Noy-Meir, I. 1975. Stability of grazing systems: an application of predator-prey graphs. Journal of Ecology 63: 459–481. Noy-Meir, A. 1990. The response of two-semi-arid plant communities to grazing. Israel Journal of Botany 39: 431–442. Olsvig-Whittaker, L.S., P.E. Hosten, I. Marcus, and E. Shochat. 1993. Inﬂuence of grazing on sand ﬁeld vegetation in the Negev desert. Journal of Arid Environments 24: 81–93. Owen-Smith, R.N. 1988. Megaherbivores: the Inﬂuence of Very Large Body Size on Ecology. Cambridge University Press, Cambridge, UK. Perevolotsky, A. 1996. Fuel breaks for ﬁre prevention: a practical evaluation of a management model. Ecology and Environment 3: 103–112. (In Hebrew.) Perlman, D.L., and G. Adelson. 1997. Biodiversity: Exploring Values and Priorities in Conservation. Blackwell Science, Malden, MA. Perrings, C., and B.H. Walker. 1995. Biodiversity loss and the economics of discontinuous change in semi-arid rangelands. Pp. 190–210 in Biodiversity Loss: Ecological and Economic Issues (C.A. Perrings, K.-G. Maler, C. Folke, C.S. Holling, and B.-O. Janson, eds.). Cambridge University Press, Cambridge, UK. Peterson, G., C.R. Allen, and C.S. Holling. 1998. Ecological resilience, biodiversity and scale. Ecosystems 1: 6–18. Pickett, S.T.A., W.R. Burch, Jr., and J.M. Grove. 1999. Interdisciplinary research: maintaining the constructive impulse in a culture of criticism. Ecosystems 2: 302–307.

Unified Framework III

319

Pimm, S.L, G.J. Russell, J.L. Gittleman, and T.M. Brooks. 1995. The future of biodiversity. Science 269: 347–350. Proulx, M. and A. Mazumder, 1998. Reversal of grazing impact on plant species richness in nutrient-poor vs. nutrient-rich ecosystems. Ecology 79(8): 2581–2592. Rosenzweig, M.L. 2003. Reconciliation ecology and the future of species diversity. Oryx 37(2): 194–205. Sahn, D.E. 1999. Structural Adjustment Reconsidered. Cambridge University Press, Cambridge, UK. Schnaiberg, A., and K.A. Gould. 1994. Environment and Society: The Enduring Conﬂict. St. Martins Press, New York. Seligman, N.G., and A. Perevolotsky. 1994. Has intensive grazing by domestic livestock degraded Mediterranean Basin rangelands. Pp. 93–103 in Plant-Animal Interactions in Mediterranean-Type Ecosystems (M. Arianoutsou and R.H. Groves, eds.). Kluwer, Amsterdam. Shachak, M., S.T.A. Pickett, B. Boeken, and E. Zaady. 1998. Managing patchiness, ecological ﬂows, productivity and diversity in the Negev Desert. In Managing Drylands: Towards Ecological Sustainability. (T. Hoekstra, and M. Shachak, eds.). University of Illinois Press, Urbana. Stevens, L.E. 1990. Pp. 99–105. Tamarisk control in southwestern United States. In Proceedings of Tamarisk Conference, University of Arizona, Tucson, AZ, 23–30 September 1987. (M.R. Kunzmann, R.R. Johnson, and P.S. Bennett, eds.). Special Report No. 9. National Park Service, Cooperative National Park Resources Studies Unit, School of Renewable Natural Resources, University of Arizona, Tucson, AZ. Stone, C.N. 1980. Systemic Power in community decision making: A restatement of stratiﬁcation theory. American Political Science Review 74: 978–90. Thomas, D.S.G., and P.A. Shaw. 1991. The Kalahari Environment. Cambridge University Press, Cambridge, UK. Turner, M.G. 1989. Landscape ecology: the eﬀect of pattern on process. Annual Review of Ecology and Systematics 20: 171–197. Turner, M.G. and S.R. Carpenter. 1999. Tips and traps in interdisciplinary research. Ecosystems 2: 275–276. Vitousek P.M., P.R. Ehrlich, A.H. Ehrlich, and P.A. Matson. 1986. Human appropriation of the products of photosynthesis. BioScience 36: 368–373. West, N.E. 1986. Desertiﬁcation or xeriﬁcation? Nature 32: 562–563. West, N.E. 1990. Structure and function of microphytic soil crusts in wildland ecosystems of arid to semi-arid regions. Advances in Ecological Research 20: 179–223. Western, D., and H. Gichohi. 1994. Segregation eﬀects and the impoverishment of savanna parks: the case for ecosystem viability analysis. African Journal of Ecology 31: 269–281. Westoby, M, B. Walker, and I. Noy-Meir. 1989. Opportunistic management for rangelands not at equilibrium. Journal of Range Management 42: 266–274. Wilson, E.O. 1988. Biodiversity. National Academy Press, Washington, D.C. World Health Organization. 1996. The World Health Report 1996: Fighting Disease, Fostering Development. WHO, Geneva, Switzerland. Zaady, E., R. Yonatan, M. Shachak, and A. Perevolotsky. 2001. The eﬀects of grazing on abiotic and biotic parameters in a semi-arid ecosystem: a case study from the northern Negev Desert, Israel. Arid Soil Research and Management. 15: 245–261.

19 Toward a Uniﬁed Framework in Biodiversity Studies Moshe Shachak James R. Gosz Avi Perevolotsky Steward T.A. Pickett

B

iodiversity has been deﬁned as the ‘‘full variety of life on Earth’’ (Takacs 1996). Because biodiversity refers to a fundamental property of ecological systems, it continues to develop as a scientiﬁc frontier for exploration and discovery. Ecological systems are a mixture, or diversity, of living and nonliving entities interconnected by a web of interactions. Biodiversity is a concept used to describe the number, variety, and organization of entities in the biosphere, or a unit of the biosphere, and their relationships to each other (Gaston 1996). More importantly, biodiversity includes consideration of processes that create and maintain variation in ecological systems (Groombridge 1992). Therefore, biodiversity is not simply about entities and taxonomy. Rather, biodiversity is concerned with the diversity of species within communities, the range of ecological processes within ecosystems, and the diversity of ecosystem processes across landscape mosaics (Heywood 1994, Levin 1997). A similar perspective on biodiversity is provided by Noss (1990), who applies the concept over hierarchical levels of organization ranging from the gene to the entire biosphere, and recognizes compositional, structural, and functional approaches to biodiversity. In contrast to this broad and comprehensive scope of biodiversity suggested above, researchers and conservationists often employ a narrow deﬁnition of biodiversity shaped by their values, interests, and goals. A more precise and more widely recognized interpretation of the scope of biodiversity is required in order to promote scientiﬁc research and to develop management programs (Christensen et al. 1996). The objective of this chapter is to build on the comprehensive framework for biodiversity proposed in the introduction (chapter 1, ﬁg.1.1) and combine 320

Toward a Unified Framework in Biodiversity Studies

321

it with additional insights from the other chapters. The initial framework recognizes biodiversity to include four components, and one complex output. The components are the roster of entities, the number of each kind of entity, the nature and degree of diﬀerence between the entities, and the spatial or functional organization of the entities. The outcome is the eﬀect of biodiversity on ecosystem processes. This framework enables us to (1) map, in a more uniﬁed way, the present achievements of research in the ﬁeld of biodiversity, (2) understand the complexity of biodiversity, and (3) advance the scientiﬁc basis for biodiversity management. We develop the conceptual framework by introducing the concept of biodiversity as a complex-state variable. We show how the entities of biodiversity, for instance, species and landscape patches, could be integrated into simple-state variables and how together they form the complex variable. The simple variables in our scheme are numbers, diﬀerences, and the organization of the organisms, resources, and landscape mosaics (chapter 1, ﬁg. 1.1). We will use the expanded framework to show the value of integrative studies for biodiversity, as well as indicate how it enables biodiversity research within the traditional subdisciplines of ecology. We also present circumstances in biodiversity studies that require integration of Earth and social sciences with ecology. We conclude by demonstrating how the complex variable of biodiversity and integration can be used as a background for the management of biodiversity.

Biodiversity as a Complex-State Variable The Concept We are able to detect and study processes in the biosphere because the multiple forms in the biosphere are in constant change. We study, for example, the changes in the number of individuals of a given species over time, which represents a change in the state of a population. The rate of the state change characterizes the rate of the population dynamics. The controllers over the rate of change are factors aﬀecting population dynamics. The case of population dynamics is an example of a change in a unidimensional state (i.e., number), which is a simple-state variable. By analogy with such aggregate state variables as population density, we introduce the concept of‘‘biodiversity state.’’ ‘‘Biodiversity state’’ is an aggregation of all the components involved in biodiversity. A change in a biodiversity state is a more complex change than those for individual populations or other individual ecological entities. This is because by deﬁnition, biodiversity is multidimensional (see chapter 1). A change in biodiversity may include changes in one of three dimensions: the number of ecological entities, the diﬀerences among them, and their organization. When we discuss a change in biodiversity of a given biosphere unit we may refer to changes in the number of species due to

322

Biodiversity, Conservation, and Management

disturbance (Dunstan and Fox 1996, Williams and Ashton 1987) or invasion by exotic species (Vitousek 1990). We may also imply changes in patch structure that modify new organization of the landscape mosaic and ecosystem processes (O’Neill et al. 1988), or the creation of new food web organization due to new species composition. A biodiversity complex state can be decomposed into a simpler state (ﬁg. 19.1). The decomposition of biodiversity into simple states is hierarchical. In the ﬁrst stage, we decompose the biodiversity complex state into three types of less complex states (ﬁg. 19.1(II)): numbers, diﬀerences, and organization. The ﬁrst state refers to the number of ecological entities. A variety of speciﬁc state changes can be depicted as a change in numbers. For example, change in the numerical state can be seen in species, genes, functional groups,

Figure 19.1 A change in biodiversity depicted as a change in a complex state variable: decomposition of the complex biodiversity state from I into components of component states of number (n,N), diﬀerence (d,D), and organization (o,O). Internal relationships within the complex state: œ, a complex state variable; , less complex state variable; , change in state; letters are simple state variables.

Toward a Unified Framework in Biodiversity Studies

323

or patches in a landscape (e.g., Noss 1990). The second of the simple biodiversity states addresses the diﬀerences among the entities (ﬁg. 19.1). This kind of state can be resolved as the changes in the attributes, such as diﬀerences in body mass (chapter 5), water use for biomass production (chapter 9), or diﬀerences in abundance for a given species (chapter 10). The diﬀerences among patches could be determined in terms of size (chapter 12) resource level (chapter 13), or fractal dimensions (chapter 12). The third type of change in biodiversity state can be in the state of organization of the entities (ﬁg.19.1). Organization implies grouping biodiversity entities by interactions. We can visualize biodiversity organization as a web of interactions among ecological entities such as water-enriched patches in a watershed and annual plant species (Shachak et al. 1998). The changes in the state of biodiversity organization could indicate a change in the number and/or strength of interactions among ecological entities. An example of a speciﬁc simple change in organization could be a change in the number of links in a food web or change in the pattern of movement of species among patches in a landscape mosaic (O’Connor 1995). State Change Drivers Internal and external factors can change a biodiversity complex state. Internal changes in a complex biodiversity state indicate the eﬀect of changes of one or more internal, less complex, states on the transition of the complex state. External changes refer to the eﬀect of factors arising outside of a particular set of entities and their spatial neighborhood (ﬁg. 19.2(I)). There are compelling examples of internal and external controls on transitions in the complex state of biodiversity in drylands. A principal internal factor controlling state transition of biodiversity in drylands is the process of biological patch formation by shrubs (Shachak et al. 1999). The number of shrub species, the diﬀerences among them in physical features or in genetic relationship, and their organization in space are the main controllers of biodiversity state change. Shrub-induced patchiness aﬀects the system through water ﬂow, soil movement (Shachak et al. 1998) and nutrient ﬂux (Zaady et al. 1998), which in turn control species composition and interactions (Boeken and Shachak 1994). Shrub invasion is a driver of changes in many components of biodiversity. Shrub invasion changes include (l) landscape diversity by creating new patch types (i.e., number and diﬀerence of patches), (2) the source–sink organization of the functioning landscape—the between-shrub area functions as a source of water and soil nutrients, and the shrub as a sink (i.e., organization of patches), (3) the species composition (i.e., number and diﬀerence of species), and (4) food web structure (i.e., organization of species). The main external drivers of changes in biodiversity state in drylands are factors associated with changes in resource base, disturbances, and human activities. Drought and rainfall variability are external factors that change the resource base of drylands and have an eﬀect on biodiversity transition. Severe

324

Biodiversity, Conservation, and Management

Figure 19.2 Changes in biodiversity state variable controlled by internal and external drivers. (I) , biodiversity complex state; , controller of state change. (II) An example of the eﬀect of an internal driver within a complex variable: , less complex biodiversity state. (III) An example of the eﬀect of an external driver on the biodiversity state change.

drought can transform grasslands of predominantly palatable, perennial grasses to grasslands dominated by unpalatable perennial shrubs, annual grasses, and forbs (O’Connor 1995). Drought thus controls state transition changes in species and functional group diversities that aﬀect the number, diﬀerence, and organization of grasslands. Wondzell and Ludwig (1995) demonstrate that variations in resource input aﬀect state changes in biodiversity. They found that grassland species diversity and its landscape mosaic change due to invasion by shrubs, which are sensitive to multiple-year periods of above- or below-average precipitation. This demonstrates how changes in resource base aﬀect two components of the complex system, namely species and landscape diversity, and their interactions.

Toward a Unified Framework in Biodiversity Studies

325

Fire as disturbance is another external driver of the complex state of biodiversity in drylands (Frost et al. 1986). Fire, together with rainfall, drought, and grazing, create high patch diversity, which is characteristic of savanna ecosystems (Mentis and Bailey 1990). Fire can maintain spatial and temporal heterogeneity across the landscape, thus controlling landscape diversity. Experiments with patch mosaic burning systems conducted in Australia and South Africa (Saxon 1984, Russell-Smith 1994, Russell-Smith et al. 1997) showed how ﬁre creates and maintains landscape diversity. Fireinduced patch diversity can also change the abundance and diversity of small mammals (Rowe and Lowry 1982) and ground-nesting birds (Mentis and Bigalke 1981) and therefore, community organization. Human activities such as land use for agriculture, forestry, and housing (Burgess and Sharpe 1981) are main external causes for biodiversity state changes. These activities induce habitat fragmentation, that is, the division of continuous habitats into smaller, more isolated areas (Saunders et al. 1991, Andren 1994). Habitat fragmentation determines patch size, habitat richness within patches, physical ﬂuxes (e.g., wind and water) and chemical ﬂuxes (e.g., nutrients) across the landscape (Bennett 1987, 1990a,b, 1993, Hobbs 1993). Fragmentation changes the state of patches in relation to number, size diﬀerences among patches and their interactions by ﬂows of energy, materials, and organisms across the landscape. The changes in landscape diversity through fragmentation can change the number, diﬀerences, and organization of species. Habitat fragmentation has been found to alter the guild structure of termites (de Souza and Brown 1994), proportions of insect predators and parasites to herbivores (Kruess and Tscharntke 1994), and proportions of arachnid habitat specialists to generalists (Webb and Hopkins 1984). It has also changed insect-mediated processes such as pollination, seed predation, dispersal, nutrient recycling (Janzen 1987) and rates of herbivory on native trees (Nuckols and Connor 1995).

Implications of Biodiversity as a Complex Variable Biodiversity and Complexity The framework of biodiversity as a change in a complex-state variable that is controlled by internal and external drivers enables us to assess how speciﬁc biodiversity studies address the gradient of complexity. If a biodiversity study attempts to understand only the changes in one simple-state variable, it represents a low degree of complexity. The study of the changes in species number with changes of area is an example of a biodiversity study that has a low degree of complexity. In such a study, for simplicity, researchers ignore the diﬀerences in species properties and their community organization. In addition, species–area studies neglect the eﬀect of landscape diversity and its organization on species number. However, many biodiversity problems are not tractable by low-complexity studies. When a change in biodiversity state

326

Biodiversity, Conservation, and Management

is controlled by external drivers such as ﬁre, fragmentation, or grazing, we have to tackle biodiversity change as a complex phenomenon. In drylands, the relationship between grazing system dynamics and biodiversity is an example of a high-complexity problem. The complexity emerges when the dynamics of rangelands exhibit multiple stable states (Westoby et al. 1989), or when multiple causes drive change in rangeland biodiversity. The case of multiple stable states in grassland biodiversity requires complex biodiversity assessment. A multiple stable state community in grazing systems is the presence of more than one stable community structure. Each state is mutually exclusive of the others but can exist in similar physical conditions. Grazing systems in Australia and North America have alternate states of woody vegetation, grassland, and lack of vegetation (Laycock 1991). Grazing can shift the communities from a grass state to a woody shrub state or to a lack-of-vegetation state. An additional example of multiple stable states appears in Africa. The Sahel exhibits state changes under the inﬂuences of rainfall variation and grazing pressure. The community states include perennial grasses, annual grasses, and annual herbs. There is evidence for transition through the three states (Rietkerk et al. 1996). The transition of states is controlled by heavy grazing (Waser and Price 1981), evolutionary history of the site, climatic regimes (Milchunas et al. 1988), ﬁre (Mueggler 1984), and removal of trees and shrubs. The second dimension of complexity in dryland biodiversity changes appears when multiple causes act on the system. A state transition in the system is complex when a transition appears in all components of biodiversity. The number of species, the diﬀerence among them, and their organization is aﬀected by rangeland state transition (Johnson et al. 1980, Belsky et al. 1989). Similarly, patch structure, properties, and organization in relation to water and nutrients change due to system state transition (Carlson et al. 1990, West 1991). As studies are ranked higher along a complexity gradient, the number, types, and interactions among the simple states increase. The position on the complexity gradient axis is a function of the question. For example, when dealing with the question of how diﬀerences in body mass aﬀect species number in a given habitat (chapter 5), we are dealing with a more complex question than if we are just dealing with a change in species number. The question becomes even more complex when studying the relationships among species number, body mass, and organization within a food web (chapter 4). Additional complexity emerges when the spatial disposition of various participants in the food web are of interest (Cadenasso et al. 2002). An example of a question that addresses increased complexity is how organisms that function as ecosystem engineers alter the number of patches in a landscape and how that aﬀects the number of species and the diﬀerences among them (Jones et al. 1994 and chapter 14). We may also ask how changes in the landscape generated by the engineers aﬀect the source–sink relationships for resource distribution in the landscape and, in turn, how that aﬀects food web organi-

Toward a Unified Framework in Biodiversity Studies

327

zation (Boeken et al. 1995). For scientiﬁc research we may ask questions at any point along the complexity gradient. However, for understanding the eﬀect of global changes and disturbances, and for management purposes, we must deal with changes in biodiversity as a complex-state variable. This is because any large-scale change in the ecosystem aﬀects many of the biodiversity states (chapters 6 and 17). Integration in Biodiversity Studies There are many diﬀerent approaches to the study of biodiversity. They range from speciﬁc studies of organisms to integrative studies that combine ecological, earth, and social sciences (ﬁg. 19.3). Common to biodiversity studies, either speciﬁc or integrative, is the study of processes that induce changes in a complex state (ﬁg. 19.2). We deﬁne domain-speciﬁc, or less integrative biodiversity studies as the study of a change in biodiversity state that refers to one type of ecological entity. In this book, we classiﬁed three domains of ecological entities: entities that are associated with organisms (genes, species, functional groups), entities related to resources (chemical components, water, energy) and entities that are associated with landscapes (soil types, patches, human exploitation). In most ecosystems the three entity types, organisms, resources, and landscape, interact (ﬁg. 19.3(II)). From a biodiversity perspective this implies that the transition of a complex biodiversity state is a combined eﬀect of changes in one entity type (e.g., number of species), which controls the transition of another entity type (e.g., number of patches). For example, a transition of the numbers, variety, and organization of patches in the landscape, due to disturbance generated by burrowing animals may cause a change in the number, diﬀerences, and organization of species and functional groups (Boeken et al. 1995). To study these interactions, integration of community and landscape ecology is required. We deﬁne the integration needed for the study of interactions among complex biodiversity states as ﬁrst-order integration. The essence of ﬁrst-order integration is the integration of subdisciplines of ecology (Jones and Lawton 1995). This implies integrating ecological principles, hypotheses, and methods that relate to simultaneous changes in a web of interactions among a set of complex biodiversity states. Examples for ﬁrst-order integration appear throughout the book. Chapters 2, 3, 5, and 10 integrate organism and landscape domains. Chapters 9 and 11 integrate organisms, ecosystem processes, and landscape patchiness. In many cases, ecology alone does not capture the richness of factors that aﬀect the dynamics of biodiversity states. This is because many factors that aﬀect biodiversity are the objects of study by other disciplines, often earth sciences. This occurs when in addition to internal controllers of the transition of the biodiversity state, external factors cause a change in a speciﬁc biodiversity state or in the web of interactions (ﬁg. 19.3(III)). External factors are

328

Biodiversity, Conservation, and Management

Figure 19.3 Degrees of integration in biodiversity studies. (I) Domain-speciﬁc studies. The studies refer to changes in a complex biodiversity state (see ﬁg. 19.1) occurring within the domain of either the organism, resource or landscape entity. ! change in a complex biodiversity state (N, numbers; D, diﬀerences; O, organization). (II) First-order integration. The studies refer to changes in a complex biodiversity state within a web of interactions between the organisms, resources, and landscape entities. (III) Second-order integration. The studies refer to changes in a complex biodiversity state which consider in addition to the ﬁrst order integration also the eﬀect of meteorological, geological, geomorphological, and pedological factors (earth sciences). (IV) Third-order integration, which adds the impact of human activity on the biodiversity parameters. The studies refer to changes in a complex biodiversity state that integrates ecology, earth, and social sciences.

Toward a Unified Framework in Biodiversity Studies

329

usually disturbances and changes in resource input. In drylands, external controllers are frequently associated with meteorology and hydrology. Rainfall amount and its redistribution as surface and subsurface runoﬀ can shift the biodiversity state. It may change landscape mosaic, community organization, and the number and distribution of functional groups. In addition, geological, geomorphological and pedological processes can aﬀect the biodiversity state changes (chapter 17). For example, soil erosion and sedimentation over geological and geomorphological, substrates can change landscape organization that, in turn, may change the structure and organization of a community of organisms (Yair and Shachak 1987). This second type of integration, which includes earth sciences, requires adding principles, hypotheses, and methods from disciplines beyond the natural sciences. Studies that include social science increase the degree of integration and generate a third type of integration (ﬁg. 19.3(III)) (chapters 14 and 17). Social sciences should be included, since human society, values, and activities inﬂuence biodiversity (Pickett et al. 1997, Pickett 1999, Pickett et al. 2001). Social sciences that interface with biodiversity are sociology, anthropology, and economics (Berkes and Folke 1992). Integration of social sciences is crucial in human-dominated regions. In these regions, landscape diversity, which aﬀects other variables of diversity, is created by a mixture of cultural practices, traditions, myths, and institutions. Patch dynamics is ultimately determined by laws, regulations, taxation, technologies, cultural values, and beliefs (Pickett et al. 1997, Dale et al. 2000), as well as by the ecological and physical processes in settled landscapes. Therefore, human-induced patch dynamics and its eﬀect on biodiversity demand a high level of integration and an interdisciplinary perspective (Pickett et al. 1999). The ability of humans to signiﬁcantly alter local, regional, and global biodiversity requires an integrative and interdisciplinary perspective for the understanding and management of biodiversity. This is one of the greatest challenges facing the future of biodiversity research. The LandUse Change Analysis System (LUCAS) provides a good example in which economists, sociologists, and ecologists together built an integrated model (Berry et al. 1996, Turner et al. 1996). Other examples are the urban Long Term Ecological Research (LTER) projects in Phoenix, Arizona and Baltimore, Maryland, USA (Grimm et al. 2000) The implementation of the concept of biodiversity as integrative science is a great challenge. This is because the study of higher level integration in biodiversity requires an interdisciplinary approach. This approach has been hampered in the past by the lack of a conceptual framework (Turner and Carpenter 1999). We hope that the framework here, which recognizes biodiversity as a complex, multidimensional concept, as well as a powerful tool for linking disparate disciplines, can promote this important integration. Such complex, multidisciplinary, multiscalar, and multidimensional integrations will require a longer time to achieve success than disciplinary research (Naiman 1999). They also demand establishing teamwork with trust and understanding (Likens 1998).

330

Biodiversity, Conservation, and Management

Biodiversity Management Management of biodiversity is an active operation by humans that is aimed at preventing undesired changes in the biodiversity state or at controlling the transition between states or to maximize or increase biodiversity. Biodiversity management should take into consideration the nature of biodiversity as a complex state and as a web of interactions between various types of entities (ﬁg. 19.4(I)). In our framework, management can control the transition related to organisms, resources, and landscape entities. Management can control each of the transitions or all of them. However, a change in a state of one entity type could feed back to another. Thus, we suggest that when considering management, we should distinguish between the target, the manipulation, and the trajectory of management (ﬁg. 19.4(II)). The target is the speciﬁc transition in the state change at which we are aiming. For example, our target might be increasing the number of annual

Figure 19.4 Biodiversity management as a controller on a complex state transition of number (n to N), diﬀerence (d to D), and organization (o to O) of three types of interacting ecological entities: organisms (OR), resources (RE), and patches (LA). (I) Options of biodiversity management (bold arrows) and alternative indirect responses (light arrows). (II) Decomposing biodiversity management into target (the objective), the actual action (manipulation), and the pathway from the manipulation to the target (trajectory).

Toward a Unified Framework in Biodiversity Studies

331

plant species in a water-limited system. It is impractical, and perhaps impossible to directly manipulate the number of species, that is, add species. However, we may increase the number of annual plants by adding pits to the landscape (Shachak et al. 1998). In this case, the applied management treatment is construction of pits, a landscape feature. This implies that we are controlling a state transition in the landscape mosaic. We change the number, variety, and organization of the patches in the landscape by adding pits. We deﬁne the actual modiﬁcation of a component of biodiversity by using management as a means of manipulation. In addition, we deﬁne the pathway of processes starting with manipulation and terminating by the target, as management trajectory. In our example, the management trajectory is the pathway starting with the construction of pits. It continues with the creation of a water-enriched patch, due to the absorption of surface run-oﬀ water, and terminating with the increase in number of annual plant species, due to high soil moisture. In biodiversity management the most common target is a change in organism-related entities: preserving or changing the numbers, variety, and organization of species, functional groups, or genes. However, the most common manipulation is landscape modiﬁcation. Thus, most biodiversity management involves trajectories. Ecosystem processes link manipulations and targets (chapter 13). The theoretical framework needed for biodiversity management is determined by the nature of its trajectory. The theoretical background should provide knowledge about ecosystem processes as links between organisms and landscape diversity (chapter 13). Biodiversity managers require an integrative framework for biodiversity (ﬁg. 19.3). The essence of this framework is the relationship between landscape and organism diversity. Boeken et al. (chapter 10) suggest an approach for the integrative study of landscape and organism diversity. They suggest a method of mapping trajectories of species and patches as a tool for linking organism and landscape diversities. Changes in both landscape and organism diversity lead to changes in complex state variables and require theoretical and experimental studies. The theoretical aspect should cope with the interface between landscape and community ecology (chapter 17), for example, how patch formation, development, and extinction aﬀect the number, variety, and organization of species. The experimental eﬀorts require studies of patch manipulations and their eﬀects on communities (Boeken and Shachak 1994). The uncertainties involved in the manipulation–target pathway (ﬁg. 19.4 (II)) require adaptive management strategies. Adaptive management is a methodology for implementing policies as experiments in management and improve it according to their performance (Holling 1973, Walters 1986). Adaptive management is based on the premise that we do not have enough knowledge to predict changes of complex-state variables. Adaptive management is management-as-experiment for complex, dynamic situations where controls and strict replications are not possible.

332

Biodiversity, Conservation, and Management

Biodiversity management takes into consideration the complexity of biodiversity and the degree of integration needed to cope with complex biodiversity problems. In conducting management as experiments we aim to learn about the trajectory of biodiversity processes (Crossley 1996). The complexity suggests that even simple steps of landscape manipulation may yield surprising outcomes in terms of organism diversity. In practice, through adaptive management, we should learn, over time, how the manipulation is related to the overall target (Grumbine 1997). Drylands are especially valuable as a model system for evaluating the application of the concepts of complex-state variables, the degree of integration, and adaptive management for biodiversity management. This is because the number, variation, and organization of organisms and landscape entities are controlled, to a high degree, by the processes of redistribution of rainfall over space by abiotic and biotic factors (Yair and Shachak 1987, Shachak et al. 1998). Therefore, the main management option is to manage landscape diversity in order to direct ecosystem processes. This is done by manipulating water redistribution, for enhancing the distribution and abundance of organisms (chapters 17 and 18). In conclusion, the contributions to this volume have demonstrated that new approaches are necessary to deal with the complexity of biological diversity in managed water-limited systems. The model presented by Perevolotsky et al. (chapter 17) provides an example of a framework that reduces the complexity associated with human impact and landscape and species diversity. The approach is to concentrate on key ecosystem processes that control biodiversity and, in turn, are controlled by it.

References Andren, H. 1994. Eﬀects of habitat fragmentation on birds and mammals in landscapes with diﬀerent proportions of suitable habitat: a review. Oikos 71: 355–366. Belsky, A.J., R.G. Amundson, J.M. Duxberry, S.J. Richa, A.R. Ali, and S.M. Mwonga. 1989. The eﬀects of trees on their physical, chemical and biological environments in a semiarid savanna in Kenya. Journal of Applied Ecology 26: 105–124. Bennett, A.F. 1987. Conservation of mammals within a fragmented forest environment: the contributions of insular biogeography and autecology. Pp. 41–52 in D.A. Saunders, G.W. Arnold, A.A. Burbidge, and A.J.M. Hopkins (eds.), Nature Conservation: the Role of Remnants of Native Vegetation. Surrey Beatty and Sons, Sydney. Bennett, A.F. 1990a. Land use, forest fragmentation and the mammalian fauna at Naringal, south-western Victoria. Australian Wildlife Research 17: 325–347. Bennett, A.F. 1990b. Habitat corridors and the conservation of small mammals in a fragmented forest environment. Landscape Ecology 4:109–122. Bennett, A.F. 1993. Microhabitat use by the long-nosed potoroo, Potorous tridactylus, and other small mammals in remnant forest vegetation of south-western Victoria. Wildlife Research 20: 267–285.

Toward a Unified Framework in Biodiversity Studies

333

Berkes, F., and C. Folke. 1992. A systems perspective on the interrelations between natural, human-made and cultural capital. Ecological Economics 5: 1–8. Berry, M., R. Flamm, B. Hazen, and R. MacIntyre. 1996. LUCAS: a system for modeling land-use change. IEEE Computational Science and Engineering 3: 24–35. Boeken, B., and M. Shachak. 1994. Desert plant communities in human made patches—implications for management. Ecological Applications 4: 702–716. Boeken, B., M. Shachak, Y. Gutterman, and S. Brand. 1995. Patchiness and disturbance: plant community responses to porcupine diggings in the Central Negev. Ecography 18: 410–422. Burgess, R.L., and D.M. Sharpe (eds.). 1981. Forest Island Dynamics in ManDominated Landscapes. Springer-Verlag, New York. Cadenasso, M.L., S.T.A. Pickett, and P.J. Morin. 2002. Experimental test of the role of mammalian herbivores on old ﬁeld succession: community structure and seedling survival. Journal of the Torrey Botanical Society 129: 228–237. Carlson, D.H., T.L. Thurow, R.W. Knight, and R.K. Heitschmidt. 1990. Eﬀect of honey mesquite on the water balance of Texas Rolling Plains rangeland. Journal of Range Management 43: 491–496. Christensen, N.L, A.M. Bartuska, J.H. Brown, S. Carpenter, C. D’Antonio, R. Francis, J.F. Franklin, J.A. MacMahon, R.F. Noss, D.J. Parsons, C.H. Peterson, M.G. Turner, and R.G. Woodmansee, 1996. The report of the Ecological Society of America committee on the scientiﬁc basis for ecosystem management. Ecological Applications 6: 665–691. Crossley, J.W. 1996 Managing Ecosystems for integrity, theoretical considerations for resource and environmental managers. Society and Natural Resources 9: 465–481. Dale V.H., S. Brown, R. Haeuber, N.T. Hobbs, N. Huntly, R.J. Naiman, W.E. Riebsame, M.G. Turner, and T.J. Valone. 2000. Ecological principles and guidelines for managing the use of land. Ecological Applications 10: 639–670. de Souza, O.F.F., and V.K. Brown. 1994. Eﬀects of habitat fragmentation on Amazonian termite communities. Journal of Tropical Ecology 10: 197–206. Dunstan, C.E., and B.J. Fox. 1996. The eﬀects of fragmentation and disturbance of rainforest on ground-dwelling small mammals on the Robertson Plateau, New South Wales, Australia. Journal of Biogeography 23: 187–201. Frost P., J.C. Menaut, B.H. Walker, E. Medina, O. Solbrig, and M. Swift. 1986. Responses of savannas to stress and disturbance. Biology International Special Issue 10, International Union of Biological Sciences, Paris. Gaston, K.J (ed.). 1996. Biodiversity—A Biology of Numbers and Diﬀerences. Blackwell Science, Oxford. Grimm, N.B., J.M. Grove, S.T.A. Pickett, and C.L. Redman. 2000. Integrated approaches to long-term studies of urban ecological systems. BioScience 50(7): 571–584. Groombridge, B. 1992. Global Biodiversity: Status of the Earth’s Living Resources. Chapman and Hall, London. Grumbine, R.E. 1997. Reﬂections on what is ecosystem management. Conservation Biology 11: 41–47. Heywood, V.H. 1994. The measurement of biodiversity and the politcs of implementation. Pp 15–22 in P.L. Forey, C.J. Humphries, and R.I. Vane-Wright (eds.), Systematics and Conservation Evaluation. Oxford University Press, Oxford. Hobbs, R.J. 1993. Eﬀects of landscape fragmentation on ecosystem processes in the Western Australian wheatbelt. Biological Conservation 64: 193–201.

334

Biodiversity, Conservation, and Management

Holling, C.S. 1973. Resilience and stability of ecological systems. Annual Review of Ecology and Systematics 4: 1–23. Janzen, D.H. 1987. Insect diversity of a Costa Rican dry forest: why keep it, and how? Biological Journal of the Linnean Society 30: 343–356. Johnson, R.R., L.T. Haight, M.M. Riﬀey, and J.M. Simpson. 1980. Brushland/ steppe land communities. Pp. 91–112 in Management of Western Forest and Grasslands for Nongame Birds. USDA, Forest Service, Intermount. Forest and Range Experimental Station, Ogden, Utah General Technical Report. INT-86. Jones, C.G. and J.H. Lawton (eds.). 1995 Linking Species and Ecosystem. Chapman and Hall, New York. Jones, C.G., J.H. Lawton, and M. Shachak. 1994. Organisms as Ecosystem engineers. Oikos 69: 373–386. Kruess, A., and T. Tscharntke. 1994. Habitat fragmentation, species loss, and biological control. Science 264: 1581–1584. Laycock, W.A. 1991. Stable states and thresholds of range condition on North American rangelands: A viewpoint. Journal of Range Management 44: 422–426. Levin, S.A. 1997. Biodiversity: Interfacing Populations and Ecosystems. Pp. 277–287. In: Takuya, A., Levin, S.A. and M. Higashi (eds.), Biodiversity: An Ecological Perspective. Springer-Verlag, New York. Likens, G.E. 1998. Limitations to intellectual progress in ecosystem science. Pp. 247–271 in M.L. Pace and P.M. Groﬀman (eds.), Successes, Limitations, and Frontiers in Ecosystem Science. Springer-Verlag, New York. Mentis, M.T., and A.W. Bailey. 1990. Changing perceptions of ﬁre management in savanna parks. Journal of Grassland Society of South Africa 7: 81–85. Mentis, M.T., and R.C. Bigalke. 1981. The eﬀect of scale of burn on thedensities of grassland francolins in the Natal Drakensberg. Biological Conservation 21: 247–261. Milchunas, D.G., O.E. Sala, and W.K. Lauenroth. 1988. A generalized model of the eﬀects of grazing by large herbivores on grassland community structure. American Naturalist 132: 87–106 Mueggler, W.F. 1984. Diversity of western rangelands. Pp. 211–217 n J.L. Cooley and J.H. Cooley (eds), Natural Diversity if Forest Ecosystems. Proceedings Workshop, University of Georgia, Athens, GA. Naiman, R.J. 1999. A Perspective on Interdisciplinary Science. Ecosystems 2: 292–295. Noss, R. 1990. Indicators for monitoring biodiversity: A hierarchical approach. Conservation Biology 4: 355–364. Nuckols, M.S., and E.F. Connor. 1995. Do trees in urban or ornamental plantings receive more damage by insects than trees in natural forests. Ecological Entomology 20: 253–260. O’Connor, T.G. 1995. Transformation of a savanna grassland by drought and grazing. African Journal of Range and Forage Science 12: 53–60. O’Neill, R.V., B.T. Milne, M.G. Turner, and R.H. Gardner. 1988. Resource utilization scales and landscape pattern. Landscape Ecology 2: 63–69. Pickett, S.T.A. 1999. The culture of synthesis: habits of mind in novel ecological integration. Oikos 87: 479–487. Pickett, S.T.A., R.S. Ostfeld, M. Shachak, and G.E. Likens, eds. 1997. The Ecological Basis for Conservation—Heterogeneity, Ecosystems, and Biodiversity. Chapman and Hall, New York.

Toward a Unified Framework in Biodiversity Studies

335

Pickett, S.T.A., W.R. Burch, Jr., and J.M. Grove. 1999. Interdisciplinary research: maintaining the constructive impulse in a culture of criticism. Ecosystems 2: 302–307. Pickett, S.T.A., M.L. Cadenasso, J.M. Grove, C.H. Nilon, R.V. Pouyat, W.C. Zipperer, and R. Costanza. 2001. Urban Ecological Systems: Linking terrestrial ecological, physical, and socioeconomic components of metropolitan areas. Annual Review of Ecology and Systematics 32: 127–157. Rietkerk, M., P. Ketner, L. Stroosnijder, and H.H.T. Prins. 1996. Sahelian rangeland development; a catastrophe? Journal of Range Management 49: 512–519. Rowe, D.T., and P.B. Lowry. 1982. Inﬂuence of ﬁre on small-mammal populations in the Natal Drakensberg. South African Journal of Wildlife Research 16: 32–35. Russell-Smith, J. 1994. Fire management in Kakadu National Park: an overview. Pp. 89–94 in K.R. McDonald and D. Batt (eds.). Workshop Proceedings on Fire Management on Conservation Reserves in Tropical Australia. Queensland Department Environment and Heritage. Russell-Smith, J., P.G. Ryan, and R. Durieu. 1997. A LANDSAT MSS-derived ﬁre history of Kakadu National Park, monsoonal northern Australia, 1980–94: seasonal extent, frequency and patchiness. Journal of Applied Ecology 34: 748– 766. Saunders, D.A., R.J. Hobbs, and C.R. Margules. 1991. Biological consequences of ecosystem fragmentation: a review. Conservation Biology 5: 18–32. Saxon, E.C. 1984. Anticipating the inevitable: a patch-burn strategy for ﬁre management at Uluru (Ayers Rock-Mt Olga) National Park. CSIRO Division of Wildlife and Rangelands Research, Melbourne. Shachak, M., M. Sachs, and I. Moshe. 1998. Ecosystem management of desertiﬁed shrublands in Israel. Ecosystems 1: 475–483. Shachak, M., S.T.A. Pickett, B. Boeken, and E. Zaady. 1999. Managing patchiness, ecological ﬂows, productivity and diversity in the Negev. Pp. 254–263 in T.W. Hoekstra, and M. Shachak (eds.), Arid Lands Management—Toward Ecological Sustainability. Illinois University Press, Urbana. Takacs, D. 1996. The Idea of Biodiversity: Philosophies of Paradise. John Hopkins University Press, Baltimore, MD. Turner, M.G., and S.R. Carpenter. 1999 Tips and traps in interdisciplinary research. Ecosystems 2: 275–276. Turner, M.G., D.N. Wear, and R.O. Flamm. 1996. Land ownership and land-cover change in the Southern Appalachian Highlands and the Olympic Peninsula. Ecological Applications 6: 1150–1172. Vitousek, P.M. 1990. Biological invasions and ecosystem processes: Towards an integration of population biology and ecosystem studies. Oikos 57: 7–13. Walters, C.J. 1986. Adaptive Management of Renewable Resources. Macmillan, New York. Waser, N.M., and M.V. Price. 1981. Eﬀects of grazing on the diversity of annual plants in the Sonoran Desert. Oecologia 50: 407–411. Webb, N.R., and P.J. Hopkins. 1984 Invertebrate diversity on fragmented Calluna heathland. Journal of Applied Ecology 21: 921–933. West, N.E. 1991. Nutrient cycling in soils of semiarid and arid regions. Pp. 295–332 in J. Skujins (ed.), Semiarid Lands and Deserts: Soil Resource and Reclamation. Marcel Dekker, New York.

336

Biodiversity, Conservation, and Management

Westoby, M., Walker, B., and I. Noy-Meir. 1989. Opportunistic management for rangelands not at equilibrium. Journal of Range Management 42: 266–274. Williams R.J., and D.H. Ashton. 1987. Eﬀects of disturbance and grazing by cattle on the dynamics of heathland and grassland communities on the Bogong High Plains, Victoria. Australian Journal of Botany 35: 413–431. Wondzell, S., and J.A Ludwig. 1995. Community dynamics of desert grasslands: Inﬂuence of climate, landforms, and soils. Journal of Vegetation Science 6: 377–390. Yair, A., and M. Shachak. 1987. Studies in watershed ecology of an arid area. Pp. 146–193 in M.O. Wurtele and L. Berkofsky (eds.), Progress in Desert Research. Rowman and Littleﬁeld, Totowa, NJ. Zaady, E., Groﬀman, P., and M. Shachak. 1998. Nitrogen ﬁxation in macro- and microphytic patches in the Negev desert. Soil Biology and Biochemistry 30: 449–454.

Index

Aboriginal people 308 Adaptive landscape 63–67 Adaptive management 259, 331, 332 Advanced Very High Resolution Radiometer (AVHRR) 252–253 African savannas. See Savannas (Africa) AgriImaGIS (satellite imagery for agricultural land use) 261 Agricultural pests. See Pests Agricultural systems 278 Agriculture ancient 297 and biodiversity 286, 313, 314, 325 and ecosystem processes 154 and grazing 260 and landscape diversity 4, 126, 299, 300, 301 orchard 296 and reconciliation 278–280 and species diversity 307, 309, 310 and substrate types 296–297 and water management 296, 297 airborne imagery 250 Alectoris chukar (chukar partridge) 42–44, 46, 48, 49 American Southwest 124, 199, 255, 258, 259, 308

and invasive species 242, 306 Anagallis arvensis 176 Ant nest mounds 175, 195–196, 197, 198, 200 Ants. See Harvester ants Argos Data Collection and Location Satellite System 261 Asia 40, 46, 233, 234, 236, 240, 280 Assemblage dynamics 174, 177, 179, 181, 182, 224 eﬀects of climate 181, 183 and landscape dynamics 168–170 Australia 34, 233, 244, 311, 325 ecosystem engineering 198 grazing 326 piosphere eﬀects 235 seed and organic matter distribution 198 small-scale patch burning 307–308 species diversity 224 Bedouins 295, 296, 298, 309, 311–312, 315 Biodiversity. See also Species diversity deﬁnition 3, 10, 162, 184, 190, 306, 320–321 ‘‘ﬁsher’s alpha’’ 80, 306

337

338

Index

Biodiversity (continued) gamma 306 hierarchical 184, 311–313, 315, 322 human impact 38, 250, 273, 296, 307–308 human values and 10, 306, 309–313, 320, 329 maintenance 227, 251, 287, 298, 301 herbivores 244 human values 310 importance to ecosystems 4 in the Negev 221, 223 management 9, 285, 286–304, 316 as a controller on a complex state 330–332 and ecosystem services 301 and human values 306 role of economics 300, 306, 309–314 value to ecosystems 298, 306, 310 Biodiversity–ecosystem function hypotheses 153, 154, 159, 161, 159 Biogeographical province 267–269, 270, 271, 274, 275 Biotelemetry (ﬁeld technology for acquiring data on desert fauna) 254–255 Body size 8, 207, 210, 215, 216, 217 as a component of modeling 71–86 rodent 58–59, 139 and spatial scale 49 tenebrionid beetles 24 Bromus ectorum (cheatgrass) 33 Bromus fasciculatus 175, 176 Byzantine period 296 C3 plants 240 C4 plants 240 Cantor set transect 99–100 Capital (value of dryland resources) 313–314, 315 Cell, landscape. See Landscape cell Cell, single-cell anomaly 102, 103–105 Centre National d’Etudes Spatiales (CNES) 261 Cervus elaphus (elk reintroduction management scenarios) 258–259 Chaﬀ, discarded by ants 194–197 Chemical inputs 307

Chem-lab-on-a-chip 256 Chihuahuan Desert. See Desert, Chihuahuan Chrysothamnus viscidiﬂorus, Pascopyrum smithii, Artemisia tridentata (species assemblage) 216–217 Chukar partridge. See Alectoris chukar Climate 67, 181, 183, 227, 251, 262 acquiring data on 257–258 micro- 170, 180, 208 modeling 57–61 as a state factor 224–225 Colonization augmentation of population density 169, 183–184 emigration, movement and arrival 128–132, 189, 217, 270, 271 ﬁltering 156, 168 and landscape diversity 168–185 and soil formation 221–222 and species diversity 127, 130–133, 134 Colonization–extinction dynamics 70, 93, 94, 155, 182 Community assembly theory 189–192, 199, 201 Community ecology 20, 124, 220, 331 deﬁnition of 161 and regional species dynamics 129 Community invasibility 4, 63–66 Community-level saturation 97–98, 102, 103, 105 modeling 75, 77, 79 Competition 169, 226, 236, 240, 270 among desert granivores 207, 216, 217 interspeciﬁc 70, 80–86, 128, 189 intraspeciﬁc 81, 82 modeling 57, 62, 65, 67 plant 200 and resource limitation 16, 17, 22, 24, 127 and species richness 127–141, 156 Complex state variable 306, 311, 321–327, 330, 331, 332 Conservation, active and passive 308 Conservation biology 30, 31, 41, 50 and human interactions 305–319 and reconciliation 266, 277, 281 Core–periphery 30–56

Index Dactylis glomerata (clonal grass) 42, 44–46 Decomposition 25, 47, 126, 154, 163, 225 and annual plants 293 low rates in deserts 18–19 Deforestation 234, 307 Degradation 234, 237, 239, 240–242 See also Soil degradation Denitriﬁcation 110, 111, 117 Density-dependence 57, 138, 140, 174 and ﬁtness 61–64, 66 and local scale processes 75, 77, 78 Desert Chihuahuan 124, 214, 215, 216–217, 242, 258 Great Basin 58, 214, 216–217 Namib 18, 23 Desert, Negev 18, 113–118, 309, 311 and annual plants 168 and biodiversity management 288, 289, 290, 293, 295, 297 and ecosystem processes 221–224, 226, 227 and eﬀects of desert fauna 126, 138, 189–201 and microbial communities 110–118 and shrub patches 169, 175, 180 and species–area relationships (SAR) 98–104 Desert, Sonoran 58 Sonoran Desert Islands 215 Desertiﬁcation 183, 233, 242, 307–308 Detritus 17, 18, 22, 24, 25, 112, 128 Diamonds 313. See also Mineral resources Diseases, new 260, 275, 276, 309 Dispersal 67, 110, 270–271, 295. See also Seed dispersal core and peripheral populations 32, 38, 40 modeling 72, 73–74, 78–81, 84–86 and patterns of spatiotemporal variability 130–132, 139, 172, 174, 182–185 Distribution range 77, 157–158, 159, 160, 162 range, ‘‘core and periphery’’ 31–50 Diversity. See also Species diversity annual 293, 294, 301 functional (see Functional diversity)

339

Gaston’s deﬁnition 3, 4, 30, 156, 161, 186, 320 hotspots 31, 39, 116 DNA microarrays 254 Drought and assemblage dynamics 180–182 and grazing 237–238, 242, 295 modeling 78 rainfall variability 323–324, 325 resistance 157, 226, 307, 308, 310 and spatial patchiness 127 Ecological (ecosystem) engineers 4, 5, 9, 133, 189–201 controlling ecosystem processes 226, 227 and landscape diversity 223 modeling 70–86 Ecosystem function 251, 252, 250, 254, 257 and biodiversity 153–155, 157, 159, 160–162, 164 and functional diversity 5, 292 and microbial processes 109, 111, 115, 117 and remote sensing 262 Ecosystem management 10, 116, 258, 297, 301, 314 Ecosystem predictability 4, 5 Ecosystem processes 4, 5, 7, 9, 10 and ecosystem services 287 eﬀect of biodiversity on 320, 321, 322, 327, 331, 332 and human impact 310, 314 and landscape diversity 290, 293, 296, 308 and microbial communities 109, 110, 111, 118 and species diversity 153–164, 220–227 Ecosystem resilience 225, 293, 294, 301, 310 Ecosystem services 250, 262, 294–298 and food production 287, 296, 299–300, 302 nature reserve 297 primary production 289, 292, 294–296, 301 recreational 297–298 water accumulation 287

340

Index

Ecotone 39–51 Ectozoochory. See Seed dispersal El Nin˜o-Southern Oscillation (ENSO) 261 Endemism 31, 306 Endozoochory. See Seed dispersal Energy ﬂow 9, 19, 155, 220 294, Entities 3–10, 327–28, 330–331 deﬁning and identifying for study purposes 220, 223 deﬁnition of diversity 156, 161, 171, 173–174 interactions between 220–223 Environmental gradient 100, 183, 224 modeling 154 precipitation (moisture) 8, 164, 173, 239, 240, 287, 288 and patterns of diversity 123 and perennial plants 138 and soils 110, 116, 117, 157–162 productivity 22, 135, 137 and species distribution 39, 40, 43, 48–51 Environmental heterogeneity, spatial 36, 85, 199, 224, 308, 325 and predation intensity 17, 21 and productivity 22–23 and resource partitioning 206, 207, 208 and species diversity 122, 124, 126–128, 131, 139, 140 Environmental heterogeneity, temporal 47, 308, 325 and species diversity 122, 124, 129, 131, 140 Event-driven systems 234 See also Stochasticity Evolutionarily stable strategies 57, 59, 63–67 Exotic species 276, 277, 322, 315 See also Invasive species Extinction assemblage-wide 168–169, 170, 172, 174, 180–185 species 31–32, 60, 271–272, 274, 276–277, 315 Fertilization, of soils by animals 199 Filtration, between species pools 190–199, 201

Fire disturbance 298, 299, 307, 308, 325, 326 Fitness 57–61, 62–67, 78, 130–132, 235, 238, 243 Food chain, homeotherm 20 Food web 141, 199, 322, 323, 326 and eﬃciency in deserts 17–26 poikilotherm-based 21 and primary productivity 125–127 reticulate 135, 140 Foraging costs 58–61, 139 Forests 224, 286, 298, 313, 325 canopy gaps 173, 184 habitat fragmentation 185, 273 longleaf pine 279–280 remote sensing 252 Fractals 91–102 geometry and species spatial variation 207–217 summation 95–96, 105–106 Functional diversity 4, 5, 9, 290 deﬁning 4, 6, 9, 16 and ecosystem function 292–295, 297, 299, 300–301 of microbial communities 109, 111, 118 Functional groups and biodiversity management 322, 324, 327, 329, 331 deﬁning 4, 6, 9, 16 and ecosystem processes 163, 290, 293, 295, 310 ecosystem-resilience species 293, 294 keystone-process species 293, 294, 301 and microbial communities 111, 112, 118 Gap Analysis Program (GAP) 261 Gape-limitation hypothesis 20 Genetic diversity and core-periphery clines 30–50 Geographic information systems (GIS). 250, 252, 254, 255, 258, 259, 261. See also Satellite technology Geomorphology. See also Rock formations eﬀects on diversity 221, 224–225, 288–297 Global change 313, 327 Global warming 275, 277

Index Goats 176, 306. See also Livestock grazing Granivorous ants 137, 214, 216–217 Granivorous vertebrates 137, 139, 207, 214–215, 216–217 Granivory. See Seed predation Grazing 176, 281, 325, 326 in Australia 326 biodiversity management 298–300 and changes in species composition 252 eﬀect on species diversity 300–301, 306, 308 humans and biodiversity 306, 311, 312, 314 intensity 238, 255, 295, 308, 309, 312 livestock 260, 292, 294, 306–309, 311–312 North America 326 and plant diversity 231–244 and sustainability 251–252, 254–256, 258–260 and woody vegetation 292–296 Great Basin Desert. See Desert, Great Basin Ground-penetrating radar 254 Ground-truthing 253

Habitat colonization (soil formation) 221–222 destruction 307 diversity 70, 306 fragmentation 185, 216, 311, 312, 325, 326 overlap 65–66, 67 patches 42, 63, 70, 277 patchiness 70 (see also Patch) poor 23 rich 23 segregation 21, 24, 25 size 70 Harmonic radar 260 Harvester ant mounds 197 Harvester ants (Messor spp.) 194–200, 216 and seed dispersal 197 Herbivory 193, 195, 200, 233–249, 325 and annual plants 293, 242, 237 eﬀects on biodiversity 192, 244, 308

341

eﬀects on plant communities 19, 134, 200, 242, 244 and evolutionary processes 200, 235, 237 inﬂuence on species composition/ diversity 189, 192, 193, 195, 200, 226, 237 insect 18, 208 and productivity 16, 17, 18, 19, 135, 140, 259 Heteromyidae 58, 216 Hordeum spontaneum. See also Wild barley Human activities 92, 286–302, 311, 312, 328, 329 Human dominated regions 280, 281, 282, 315, 329 Human-related disturbance 38, 173, 175, 183, 224, 286, 301 Humans as ecosystem components 10, 201, 305, 311, 327 Humans, biodiversity impacts 273–274, 276–278, 287–288, 305–315, 323, 325, 328–330, 332 Humans, nomadic 313, 315 Hunting (predation) 23, 280, 281, 307 Hydrology 118, 154, 157, 224, 261, 329 eﬀects on landscape 229, 230, 306 Hystrix indica. See porcupine

Ideal-free distribution 78 Iﬂoga spicata 176 Incidence and abundance interpretation of relationships 168, 169, 170, 181 local and spatial processes 172, 173, 183 phase plane 171–175, 177–179 spurious correlation 171 Incidence, colonization, persistence and extinction 170, 172, 184 Incidence curves 90–91 International Council of Scientiﬁc Unions (ICSU) 261 International Long Term Ecological Research network (ILTER) 262 Invasive species 306. See also Exotic species Island biogeography 129, 269–270

342

Index

Israel 30, 33–35, 39–50, 91, 110 animal eﬀects on community assembly 189–201 Eilat 280 management scenarios 287, 289, 290, 297 See also Desert, Negev Karoo region, South Africa 233 Lacunarity 207–209, 211 Land Use 10, 183, 185, 227, 261, 306, 310, 311, 325. See also Agriculture Land use traditional 278, 286, 293, 294, 295 urban 185, 298, 307 Land Use Change Analysis System (LUCAS) 315, 329 Landscape cell 71, 72, 73, 79, 80, 81, 93, 95, 96, 97–98 Landscape diversity 287, 311 changes in patch properties 168, 179 and ecosystem processes 220–227 forests 224 framework for study 323–332 management 288–290 modeling 70–88 patch formation 168, 169, 170, 172 patch frequency and properties 167, 171–174, 179–182, 185 trajectories of patch types 173–174 See also Patch Landscape, human impact 296 Landscape matrix 72 Landscape mosaic 223, 226 anthropocentric approach 185 components 167, 173 and ecosystem processes 287–300 experimental patch addition 175–176 experimental patch disturbance 176–177 metapopulation approach 183–185 organism-centered approach 184 relevant scale 184 Landscape scale. See Scale, landscape Landscape structure 286, 288, 292, 296, 298, 301 Leaf area index 252, 260 Lehavim 288, 290, 292, 299

Life history strategy 127, 139, 183, 201, 243–244 Litter 125, 161, 225, 294 and desert food webs 18–23 and microbes 112–114 mounds 168, 169 and patch dynamics 170, 173 and seed dispersal 198 Livestock grazing 259, 295 husbandry 293 impact 293, 295, 296 trampling 175, 176, 177, 179–180 and management 295, 296 Long Term Ecological Research network (LTER) 258, 263, 315, 329 Maasai 315 Macrodetritivores 15–26 Map saturation 97–98, 103, 105 Mediterranean climate 290 Mediterranean region 33, 40–44, 46–51, 292, 293 and semi-arid plant communities 101, 198, 234, 298 Messor spp. See Harvester ants Metabolic maintenance cost 60, 62–67, 73, 139, 212 Metabolic rate as part of tradeoﬀ 58 Metabolic rate, modeling 73–75, 81 Metabolic rate, poikilotherms 17, 20 Metapopulations 73, 80, 93, 131, 183, 185 modeling 131 Microbes and biodiversity 109–121 Microbial activity in drylands 125, 201 communities in soils 255 crust 169 (see also Soil crust) decomposers 18, 19, 25, 255 Microbiotic crusts, determining distribution 252 Microphytic crust 115–117, 289–290 Microtechnologies 255 ‘‘electronic noses’’ 256 Mineral resources 313, 314 Mineralization 18, 154, 163, 199, 200, 225 Mount Sinai 288–301

Index Muir, John 310 Multiple stable state 312, 326 Namibia 235, 239, 241. See also Desert, Namib Narcissus eﬀect 236–237 Nature Conservancy, The 279 Negev Desert. See Desert, Negev Nest mounds of ants. See Ant nest mounds Net primary production (NPP) 20, 123, 251, 252, 253, 257 human impact on 307 NEXRAD 257, 259 Niche 36, 126, 138, 206, 286, 301 identifying 212–217 maps 92–94, modeling 73, 75, 76 requirements 129–131 Nitrogen ﬁxation 116 Nonequilibrial systems 127, 183, 234 Normalized diﬀerence vegetation index (NDVI) 253, 259 North America 124, 216, 281 food web studies 17 grazing 237, 242, 244, 326 rodent diversity 59, 60 Nutrient accumulation 133, 170, 173, 199, 224 acquiring data on 255, 256 and microbial diversity 109, 111–112 Nutrient cycling 201 220, 242, 293, 307 and microbial diversity 109, 111–112, 114–118 Object-Oriented Programming 71 Oil 313 Ononis reclinata 176 Optimal foraging principals, determining 210–212 Organic matter dynamics 115, 172, 223, 224, 225, 293, 298 Pastoralists, 237, 234 Patch biological 158, 289- 301, 323 burning 308 crust 170–184 dynamics 8, 118, 329 human-made 297

343

management 301 network 290, 298, 299, physical 158, 224, 289–301 properties, eﬀect of climatic variation 181 properties, eﬀect of disturbance 168, 172, 173–176, 178–180 shrub 170–184 size and thresholds 210–213, 212, 213, 216 soil 6, 221–222, 255, 288, 290–292 Patches and ecosystem processes 220–227 Patches, eﬀect on species diversity 206–217 Patchiness habitat 224 landscape 224 Pests, agricultural 260, 279, 309 Phenotypic diversity 35, 40, 42, 44, 46–48, 243 Piosphere 235 Plant cover 9, 17, 125, 126, 134, 140, 259, 260 eﬀects of grazing on 233–244 role in predator–prey interactions 21–25 Plantago coronopus 175, 176 Plants annual 22, 301, 323, 324, 326 and diversity in deserts 123, 131, 132, 138, 139, 293–295 impact of animals on 196, 198, 200, 202 and management 330–331 mapping species and landscapes 174–182 replacing perennials 242–243 and soil patches 222, 227 in semi-arid shrublands 167, 168, 170 species diversity and ecosystem processes 162–163 Trifulium purpureum and Hordeum spontaneum 44–50 Plants, perennial 49, 139, 216, 235–236, 242–243, 326 Poikilotherms 18, 21 Pollination/pollinators 189, 278–279, 310, 325 Population dynamics 126, 130, 182, 183

344

Index

Population dynamics (continued) eﬀects of grazing on 254 and estimating diversity 38, 43, 76, 77–79 growth, persistence and augmentation 172 rate of state change and 321 resource and site limitation 182 seed limitation 174, 175, 177 Porcupine 175, 193–196, 198–200, 227 Poverty 313 Power law 207–213, 217. See also Scaling law Predation and diversity 58, 63, 217, 270 and diversity, inﬂuences on 127–128, 133, 135, 137–141 role of in deserts 21–25 Prey density 17 Prey type 17 Primary producer 16 Primary productivity. See Net primary productivity Process landscape-scale 71, 72, 74, 78–80 local-scale 71, 72, 73, 75–79 Productivity 16, 18, 19, 20, 22, 23, 26 low 64–65, 123, 124, 206, 308 low, and high animal diversity 15–26 low, and species interactions 130–131, 135–136, 140 secondary 17, 125 Pteranthus brevis 177 Rainfall 220–222, 225–227, 229–230 annual 168, 170, 176, 181, 182 Rainfall gradient 239, 240, 287, 288 and genetic diversity 48–49 and soils 116–117 Rainfall variability 96, 293 309, 323 Ramat Hanadiv Park 288–301 Ranching commercial 233, 235, 237, 239, 241, 242 communal 233, 235, 237, 239, 241, 242 Rate birth 73, 7, 77, 81 death 73, 7, 77, 81 metabolic 20, 58, 73–75, 81, 212

Reboudia pinnata, dispersal of 104, 196, 197 Reclamation 307–308 Reconciliation 307–308 Reconciliation ecology 266–282 Refuge 22, 25 Refuge, plants as refuge from grazing 21–25, 125–140, 293, 295 Remote sensing 126, 252–253, 255, 256, 262 Resource concentration, eﬀect on species diversity (‘‘packing’’) 212–213, 216–217 heterogeneity 223 limitation 15, 17, 127 management, appropriate technology for 258–261 partitioning 16, 81, 82, 206–217 utilization theory 154, 164 Road building 307, 315 Rock formations, eﬀects on diversity 221, 224–225, 288–297. See also Geomorphology Rodent communities 58 Root : shoot ratio 19 Roots 18, 170, 238, 254, 292 Rostraria cristata 175, 176 Run-oﬀ. See Water Sahel 326 Sand dunes 306 Satellite imagery 250, 252, 253–254, 255, 259, 260, 261 AgriImaGIS 261 GAP analysis program 261 Landsat Thematic Mapper 252 Satellite technology Argos Data Collection and Location Satellite System 261 Geostationary Operational Environmental Satellite (GOES) 261 Global Climate Observing System (GCOS) 261 Global Environmental Monitoring System (GEMS) 261 Global Positioning System (GPS) 261 Global Terrestrial Observing System (GTOS) 261–262 Savannas (Africa) 308, 309

Index Sayeret Shaked Park 288, 289, 290, 291, 292, 293, 295, 300 Sayeret Shaked Park LTER 175 Scale landscape 70–88, 115, 185, 201 landscape, collecting data on 262–259 local 70–88 Scale-area curves 92–104 Scaling 73, 77, 91, 92, 95, 98, 103, 107 Scaling law 207–213, 217. See also Power law Sede Boqer 45, 100–101, 113, 141, 288–300 Seed dispersal 189–191, 195, 196–199, 325 animal mediated 189, 195, 197–198 Seed predation 195, 198, 325 Sevilleta National Wildlife Refuge 258, 259, 263 SHALOM (simulation model) 70–88 Sheep 176, 259, 298, 312 Shrub encroachment 240, 308 invasion 323–324 ‘‘islands of fertility’’ 206 Shrub patches 170, 226 and grazing 289–292, 294–295 in the Negev 175–177, 179–180, 184 Sinai 290–292 Single-cell anomaly 97–98, 102, 105 Snakes 23, 307 Social inequality 315 Social sciences 321, 327 328, 329, 327, 328 Soil crust (cyanobacteria) 221–223, 289, 290, 292, 306. See also Microbial, crust Soil degradation 19, 109, 185, 239–240, 273, 274, 295 disturbance, by animal activity 192–193, 198 erosion 92, 206, 221, 252, 253, 255, 260, 295, 329 ﬁeld technology for acquiring data – minirhizotrons 254 ﬁeld technology for acquiring data – multispectral remote sensing 252 formation, parent material 221, 224–225

345

mounds 222, 223, 227, 289, 290, 293 nitrogen 241–242 nutrients 111, 170, 199, 323 and grazing 233, 234, 240, 241–242, 244 organic carbon 241 pits, eﬀect on water, seed and organic matter ﬂow 193, 194, 197, 198 texture 207, 224, 255 trampling 295, 296 (see also Livestock trampling) water-holding capacity 241, 255, 292 Soil moisture 8, 92, 111, 115 acquiring data on 255, 257 and landscape diversity 173, 174, 176, 180, 291–300 modeling 154–162, 164 Sonoran Desert. See Desert, Sonoran Source–sink populations 93, 270–274, 275, 309, 323 Source–sink relationships 222–224, 226–227, 323, 326 Spatial heterogeneity and distribution of resources-eﬀects on diversity 206–207, 224 Spatial heterogeneity plants 18 Spatial scales 16, 206, 207, 209, 210, 216–217 Speciation 43, 128, 217, 275, 276 in adaptive landscapes 64–66 and conservation 31 rate curve 270, 275, 276 and species-area power equation (SPAR) 270–276 Species assemblages 157, 162, 164 assumptions 167 Chrysothamnus viscidiﬂorus, Pascopyrum smithii, Artemisia tridentata 216–217 connection to the landscape 167, 168 eﬀect on patch properties 169 internal and external pathways of change 168 links with population dynamics 182 mapping 171–173 predictions for future scenarios 183, 216 sagebrush–rabbitbrush–western wheatgrass 216–217

346

Index

Species body size aﬀects patterns of species diversity 207–210, 215, 216–217 Species coexistence mechanisms 57–59 Species composition and ecosystem management 301, 322, 323 eﬀects of grazing on 233, 235, 236, 252 and food webs 141 impact of animals on 189, 195, 198 mapping 172, 182 modeling 80, 82, 86, 155, 157, 168 Species dispersal 78, 86, 271 Species diversity on ant mounds 197 climate and 60, 64–66 and ecosystem processes 222–224, 227 in the fossil record 272–273 incidence and abundance 171 and landscape diversity 220–227 latitudinal gradient in 60 local vs. regional inﬂuences on 57, 67–68 and management 286–302, 306–310 modeling 80, 82, 84, 85, 86 ‘‘more individuals’’ hypothesis 59 productivity and 59, 64–66, 226 single indices 171 Species guilds and resource availability 206–207 Species persistence 132, 170, 172, 184, 189, 207, 210, 212 Species pools 128–129, 155–156, 168, 190–191 Species population growth rate, determining 211 Species richness 215 and disturbance 175–176, 194–196, 198, 200 and grazing 234–236 local and spatial patterns 5, 123, 128–129, 131, 132, 182 modeling 156, 161, 215–216 and predator–prey interactions 141 and productivity 59 and rate of ecosystem processes 163–164 and substrate types 293 and thresholds 153

Species–area curve 95, 267–269 Species-area power equation 267–271, 274–275. See also Speciation Species–area relationship 87–107, 266 Species–habitat match 73 State change (biodiversity state change) 170, 190, 306, 309, 320–336 State factor model of ecosystems and soil development 224–225 Steady-state diversity 269, 271–273, 276–277 Stipa capensis 175, 176 seed predation 196–197 Stochastic events 127, 183–184, 271, 276 Stochasticity, catastrophic 72, 78–86, 271, 276 Stochasticity, demographic 43, 76–86, 130–131, 180 Surface reﬂecting radar 260 Sustainability 250–265, 293, 294, 313 Synthetic aperture radar (SAR) 252, 255, 260 Systems-on-a-chip 256 Tamarix 306 Tenebrionid beetles 18, 19, 21–26 Tenebrionid beetles in the Namib desert 123, 125, 128 Termites 18, 19, 21, 126, 199, 325 Terrestrial communities 18, 19, 26 Tilling 307 Transport, of seeds and nutrients 190, 191, 198, 201 Trifolium purpureum (purple clover) 42, 44, 45 Trophic levels 6, 15–25, 134–137 Tulipa systola, predation by porcupines 193 Turnover zone 39,40, 42, 44,47, 51 Ungulates 199, 251, 254, 259, 260 United Nations Environmental Program 261 U.S. Air Force 279 U.S. Geological Survey 261 U.S. National Wildlife Federation 280 Utilization habitat 73,74 resource 73 154, 164, 294 water 154, 159, 161, 162

Index Vegetation techniques for acquiring data 251–254 unpalatable 294 Vegetation herbaceous and competition with woody vegetation 293 and grazing 298, 301 in the Negev Desert 289–301 and shrub patches 226, 284, 290, 292, 295 woody and grazing 183, 295, 298, 301, 326 and patchiness 159, 294 encroachment 240 in Ramat Hanadiv Park 290, 292, 293

347

Water 307, 312, 213, 314, 315 accumulation 287, 291 contour-line catchment 298 enrichment 291, 292, 293 ﬂow 226, 287, 290–300 inﬁltration 19, 125, 194, 198, 199, 222 redistribution 226, , 291–292 run-oﬀ 221–222, 226, 291–300, 315, 329 storage 298 use eﬃciency 157–158, 159 Weeds 306, 309 Wild barley 34–49 World Health Organization 309 World Meteorological Organization (WMO) 261 Zoochory. See Seed dispersal