UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Shifted community states in four marine ecosystems : some potential mechanisms Okey, Thomas Anthony 2004

You don't seem to have a PDF reader installed, try download the pdf

Item Metadata

Download

Media
[if-you-see-this-DO-NOT-CLICK]
[if-you-see-this-DO-NOT-CLICK]
ubc_2004-902447.pdf [ 17.44MB ]
Metadata
JSON: 1.0074877.json
JSON-LD: 1.0074877+ld.json
RDF/XML (Pretty): 1.0074877.xml
RDF/JSON: 1.0074877+rdf.json
Turtle: 1.0074877+rdf-turtle.txt
N-Triples: 1.0074877+rdf-ntriples.txt
Original Record: 1.0074877 +original-record.json
Full Text
1.0074877.txt
Citation
1.0074877.ris

Full Text

SHIFTED COMMUNITY STATES IN FOUR MARINE ECOSYSTEMS: SOME POTENTIAL MECHANISMS by THOMAS ANTHONY OKEY B.Sc, Saint Lawrence University, 1986 M.Sc, Moss Landing Marine Laboratories, San Jose State University, 1993 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES (Department of Zoology.) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA April 2004 © Thomas Anthony Okey, 2004 ABSTRACT The rigorous demonstration of truly stable alternate community states is elusive in marine ecosystems and might remain so for some time. Examples of marine community states that have shifted are nevertheless increasingly conspicuous. The growing concern over these altered community states is often related to questions of persistence and reversibility, especially when these shifted states are considered to be degraded. I used empirically-based trophic models and direct empirical field studies to explore the potential of particular hypothesized mechanisms to generate and maintain alternate community states in four marine ecosystems: a Galapagos rocky reef, Prince William Sound, Alaska, the West Florida Continental Shelf, and coral reefs of the Spermonde Archipelago or Southwest Sulawesi, Indonesia. Construction and analysis of an Ecopath with Ecosim (EwE) model of a Galapagos rocky reef indicated that the unsustainable fishery targeting the holothurian Stichopus fuscus can, by itself, trigger the replacement of previously diverse reef platform communities with Aiptasia sp. anemone barrens. Construction and analysis of a Prince William Sound, Alaska EwE model indicated that severe disturbances such as the Exxon Valdez oil spill can shift a marine biotic community to an alternate state that persists in a stable manner for decades, and that the character of such disturbances, in terms of their breadth and community-level life-history signatures, might strongly influence whether a community shifts to an alternate state. In addition, a search for keystone species provided a whole-system approach to identifying species or functional groups whose depletion or removal might have the most severe consequences for community structure, and the most potential for mediating shifts. Construction and analysis of a West Florida Shelf EwE model indicated that increased sea floor shading by coastal phytoplankton (i.e., resulting from nutrient runoff pollution) can cause broad shifts in this continental shelf community by shading benthic primary producers, which support much of the overall shelf community. An empirical field study of monsoons and runoff in Southwest Sulawesi, Indonesia indicates some of the mechanisms involved in shifting tropical reefs from coral-dominated to algae-dominated systems. These contributions feature some newly emerging approaches for gaining insights into marine communities and for developing hypotheses that can be more rigorously evaluated in the future. None of these examples, however, are comprehensive or strictly falsificationist by themselves. The continued integration of these emerging community/ecosystem modeling approaches with direct empirical studies should vastly increase the potential for distinguishing the relative roles of natural and anthropogenic forces in shaping marine communities. The goal of the first steps described here was to identify particular mechanisms in each example that have the potential to generate or maintain community shifts. 11 TABLE OF CONTENTS ABSTRACT II TABLE OF CONTENTS Ill LIST OF TABLES V LIST OF FIGURES VLIST OF EQUATIONS VII PREFACE VIIACKNOWLEDGEMENTS IX CHAPTER 1. OVERVIEW AND SUMMARY 1 Ecopath with Ecosim methodology 8 LITERATURE CITED 11 CHAPTER 2. TROPHIC MODEL OF A GALAPAGOS ROCKY REEF WITH SIMULATIONS OF FISHING IMPACTS 6 ABSTRACTINTRODUCTIONMETHODS 1 RESULTS 7 DISCUSSION : 33 LITERATURE CITED 3CHAPTER 3. DISCOVERY OF ANEMONE BARRENS IN THE GALAPAGOS AND POTENTIAL EXPLANATIONS 42 ABSTRACTINTRODUCTIONMETHODS 5 RESULTS : 47 DISCUSSION 49 CONCLUSIONS 5LITERATURE CITED 6 CHAPTER 4. A SEARCH FOR KEYSTONES IN PRINCE WILLIAM SOUND, ALASKA USING A MASS-CONTINUITY TROPHIC MODEL 60 ABSTRACT 6INTRODUCTIONMETHODS 4 RESULTS 9 DISCUSSION 73 CONCLUSIONS 80 LITERATURE CITEDCHAPTER 5. CAN OIL SPILLS SHIFT MARINE ECOSYSTEMS TO ALTERNATE STABLE STATES?: PRELIMINARY SIMULATIONS WITH AN ECOPA TH MODEL OF PRINCE WILLIAM SOUND, ALASKA 4 ABSTRACT .'. 8INTRODUCTION 8METHODS 7 RESULTS 91 DISCUSSION 2 iii CONCLUSIONS 99 LITERATURE CITEDCHAPTER 6. CAN SHADING BY PLANKTON BLOOMS CAUSE SHELF-WIDE COMMUNITY SHIFTS? 104 ABSTRACTINTRODUCTION 10> METHODS 9 RESULTS 112 DISCUSSION '. 114 LITERATURE CITED 123 CHAPTER 7. CORAL-TO-ALGAE SHIFTS IN INDONESIAN CORAL REEFS: SEDIMENTATION AND POLLUTION 130 ABSTRACTINTRODUCTIONMETHODS 1 RESULTS 136 DISCUSSION 145 LITERATURE CITED 154 CHAPTER 8. CONCLUSIONS 8 LITERATURE CITED 163 APPENDIX A. DIET COMPOSITION MATRIX FOR THE ECOPATH MODEL OF THE FLOREANA (GALAPAGOS) ROCKY REEF 5 APPENDIX B. LIST OF THE 57 SPECIES IN 27 FAMILIES ENCOUNTERED DURING THE 26 VISUAL FISH TRANSECTS AT ISABELA AND FERNANDINA ISLANDS, GALAPAGOS 167 APPENDIX C. REFINEMENT OF THE 1994-1996 PRINCE WILLIAM SOUND ECOPATH MODEL.... 168 APPENDIX D. DIET COMPOSITION MATRIX FOR THE ECOPATH MODEL OF THE PRINCE WILLIAM SOUND 170 APPENDIX E. DIET COMPOSITION MATRIX FOR THE ECOPATH MODEL OF THE WEST FLORIDA [CONTINENTAL] SHELF 172 iv LIST OF TABLES Table 2-1. Model space estimations within the 0-20 m isobath 22 Table 2-2. Basic parameters of the Ecopath model of the Floreana rocky reef, Galapagos 28 Table 2-3. Basic flows and indices in the Floreana rocky reef Ecopath model 29 Table 2-4. Flows from primary production and detritus...! 30 Table 2-5. Percentage of total annual catch comprised by the 10 functional groups targeted 30 Table 3-1. Coordinates and 'use zoning' of visual fish transects : 45 Table 3-2. Observations pertinent to the appearance of anemone barrens in the Galapagos Archipelago. 50 Table 3-3. Some predators of Aiptasia sp. or other Hexacorallia 53 Table 4-1. Basic parameters of the 51 compartment Ecopath model of Prince William Sound 68 Table 4-2. Basic flows and indices describing the 1994-1996 Prince William Sound Ecopath model 69 Table 4-3. Flows from primary production and detritus through the Prince William Sound model 70 Table 5-1. General sources of mortality estimates for the three catastrophic disturbance scenarios 89 Table 5-2. Specified direct mortalities imposed on each functional group 90 Table 5-3. Peer-reviewed Exxon Valdez oil spill biological effects information 96 Table 6-1. Basic parameters of the Ecopath model of the West Florida continental shelf Ill Table 6-2. Results of regressions of water column measurements against time 112 Table 6-3. Estimated production and biomass of the four primary producers on the WFS 112 Table 6-4. Flows from primary production and detritus 113 Table 7-1. Relative organic content of the vertical flux of non-carbonate suspended sediment 138 Table 7-2. Relationship between water clarity measures and distance (km) from land and rivers 138 Table 7-3. Summary of integrated GLM repeated measures ANOVAs for community comparisons 139 Table 7-4. Summary of separate GLM repeated measures ANOVAs for community comparisons 140 Table 7-5. Summary of log-linear analyses for community comparisons 141 Table 7-6. Summary of GLM integrated MANOVA for abiotic substrate comparisons 145 Table 7-7. Summary of log-linear analyses for abiotic substrate comparisons 146 Table 7-8. Percent cover of the five taxa with the highest areal dominance at each site 148 Table 7-9. Evidence of tolerance to low light, pollution, and sedimentation in corals 151 Table 8-1. Evaluation of the four marine community examples presented in this dissertation 162 v ( LIST OF FIGURES Figure 1-1. Response curve of a community that exhibits hysteresis 5 Figure 2-1. Map of the Galapagos Islands 17 Figure 2-2. A map of Floreana Island 8 Figure 2-3. Predicted changes resulting from the complete removal of sharks 31 Figure 2-4. Predicted catch and biomass curves for pepino sea cucumbers (Stichopus fuscus) 32 Figure 2-5. A spatial simulation of the effects of a fisheries exclusion zone on pepino (S. fuscus) 32 Figure 3-1. Isabela Island and Fernandina Island, Galapagos, Ecuador 44 Figure 3-2. Number of fish species observed 48 Figure 3-3. Predicted changes in biomass resulting from removal of predators of S. fuscus 49 Figure 4-1. Map of Prince William Sound, Alaska 64 Figure 4-2. Frequencies of the community importance and community longevity support scores 70 Figure 4-3. Community longevity support index rankings 71 Figure 4-4. Community longevity support of a species relative to its biomass 72 Figure 4-5. Frequencies of the trophic interaction strength index and keystone index scores 73 Figure 4-6. Interaction strength index and keystoneness of all groups in the PWS model 74 Figure 4-7. Keystone index ranking of a functional group versus its relative biomass 75 Figure 5-1. Map of Prince William Sound (PWS), Alaska 88 Figure 5-2. Simulations of three catastrophic disturbances in Prince William Sound, Alaska 93 Figure 6-1. Changes in estimates of phytoplankton production from 1966 until 1995 113 Figure 6-2. Predicted biomass changes after simulated shading over the West Florida shelf 116 Figure 6-3. Generalized predatory interactions in a benthic food web of the West Florida Shelf. 120 Figure 7-1. Locations of sampling stations at coral cay islands in the Spermonde Archipelago 132 Figure 7-2. Monthly flow measurements at ten river sampling stations in southwest Sulawesi 133 Figure 7-3. Sedimentation, water clarity, and percent cover of dominant categories 137 Figure 7-4. Number of species and dominance along a gradient of distance from rivers and land 142 Figure 7-5. Species-area curves of sessile benthic organisms 143 Figure 7-6. Reef substrate penetrability and types of abiotic substrate at the four reefs 144 vi LIST OF EQUATIONS Equation 1-1. Ecopath master equation 9 Equation 1-2. Ecosim dynamic formulationEquation 1-3. Consumption rate function 10 Equation 4-1. Community importance 66 Equation 4-2. Community longevity support 7 Equation 4-3. Trophic interaction strength index 6Equation 4-4. Keystone index 69 vii PREFACE This dissertation work was conducted and documented as individual manuscripts written for submittal to peer-reviewed scientific journals. Chapter 2 was reprinted from Ecological Modelling 172(2-4), T.A. Okey, S. Banks, A.F. Born, R.H. Bustamante, M. Calvopina, G.J. Edgar, E. Espinoza, J.M. Farina, L.E. Garske, G.K. Reck, S. Salazar, S. Shepherd, V. Toral-Granda, P. Wallem, A trophic model of a Galapagos subtidal rocky reef for evaluating fisheries and conservation strategies, Pages 383-401, Copyright (2004), with permission from Elsevier. Chapter 6 was reprinted from Ecological Modelling 172(2-4), T.A. Okey, G.A. Vargo, S. Mackinson, M. Vasconcellos, B. Mahmoudi, CA. Meyer, Simulating community effects of sea floor shading by plankton blooms over the West Florida shelf, Pages 339-359, Copyright (2004), also with permission from Elsevier. In some cases, the text from these manuscripts is updated in content and adapted to thesis formatting. For example, Ecopath with Ecosim modeling methodology descriptions—the main analytical method used herein—were removed from individual chapters and summarized at the end of Chapter 1. In other cases, certain paragraphs that I did not write or initiate were deleted for the present format. The cited literature is listed at the end of each chapter and the reference formats differ among chapters according to target journal. I generally approached each chapter with the goal of journal submittal in order to streamline the research communication process. Vlll ACKNOWLEDGEMENTS I thank my supervisor and mentor, Daniel Pauly, for his unwavering dedication and for input that is unfailingly insightful, expansive, and cheerful. I am similarly grateful for my other committee members: Charley Krebs, Rob DeWreede, Tony Sinclair, and Diane Srivastava. Their guidance, knowledge, and enthusiasm were equally crucial. I am indebted fundamentally to Jeffrey Polovina, Carl Walters and Villy Christensen for making complex trophic modeling accessible to ordinary ecologists in the form of Ecopath with Ecosim. Thanks also go to my external examiner, Charles H. Peterson, whose comments helped with the final stage of refinement, and to the other members of my examining committee for their time and insights: Royann Petrell, Les Lavkulich, and Al Lewis. I also extend deserving thanks to he staff and students at the UBC Fisheries Centre and Department of Zoology. Ultimately, I am most grateful for Kathleen Dobie who connects me with beauty, and for the unwavering encouragement of my loving parents Joan and Frank. Chapter 2 I thank the co-authors of the paper from which Chapter 2 was adapted (in alphabetical order): Stuart Banks, Abraham F. Born, Rodrigo H. Bustamante, Monica Calvopina, Graham J. Edgar, Eduardo Espinoza, Jose Miguel Farina, Lauren E. Garske, Giinther K. Reck, Sandie Salazar, Scoresby Shepherd, Veronica Toral-Granda, Petra Wallem. I also thank the staff and scientists of the Charles Darwin Research Station for general support and the Galapagos National Park Service, Puerto Ayora, Santa Cruz, Galapagos, Ecuador for granting the research permits and logistic support. In particular, I thank H. Snell for providing aerial photographs of the study area and V. Francisco and A. Herrera for information on echinoids and intertidal chitons, respectively. I also thank J. MacLean for editorial input, Daniel Pauly for jump-starting the cooperation that led to this contribution, the European Union Concerted Action Program (Contract # IC18-CT97-0175), and the Charles Darwin Research Station. Chapter 3 I thank Scoresby A. Shepherd, Graham J. Edgar, Priscilla C. Martinez, and all my co-authors on the Floreana Rocky Reef model paper including: S. Banks, A. F. Born, R. H. Bustamante, M. Calvopina, E. Espinoza, J. M. Farina, L. E. Garske, G. K. Reck, S. Salazar, V. Toral-Granda, and P. Wallem. I also thank C. Hickman, G. Muller-Parker, D.J. Ayre, J. Geller, I. Svane, A.J. Butler, G.M. Wellington, K. Fujiwara, Daphne Fautin, K. Bjorndal, J. Penaherrera, and F. Rivera for very helpful information on aspects of this subject, or for sharing their observations, or both. Finally, I thank the staff of the Charles Darwin Research Station for the opportunity to participate in the cruise to Fernandina Island, and D. Pauly for his ability to make things happen. ix Chapter 4 The Ecopath model of Prince William Sound was constructed for the purpose of ecosystem synthesis as part of the Exxon Valdez Oil Spill Restoration Program, and funding for that part of the study was provided by the Exxon Valdez Oil Spill Trustee Council (Restoration project 99330). The model was constructed by a broad collaboration of experts listed in Okey and Pauly (1999a), who are sincerely thanked. I thank S. Libralato, B. Ballachey, and B. Wright for additional helpful discussions and input. I am indebted to C. Walters and V. Christensen for developing Ecosim and its policy analysis routines. Chapter 5 The Ecopath model of Prince William Sound was constructed for the purpose of ecosystem synthesis as part of the Exxon Valdez Oil Spill Restoration Program, and funding for that part of the study was provided by the Exxon Valdez Oil Spill Trustee Council. I thank the numerous experts on the Prince William Sound ecosystem for their contributions to the Prince William Sound Ecopath model (listed in Okey and Pauly 1999a, b). I also wish to thank D. Pauly and B. Wright for their encouragement and S. Flake and T. Monk for insights on stochasticity and determinism. Chapter 6 I thank the coauthors of the paper from which Chapter 6 was adapted: Gabriel A. Vargo, Steven Mackinson, Marcelo Vasconcellos, Behzad Mahmoudi, Cynthia A. Meyer. Many thanks go to the experts at the Florida Marine Research Institute, the University of South Florida, and other institutes. L. Vidal-Hemandez played a central role in coordinating the collaborative construction of the West Florida Shelf model. T. Sutton, W. Arnold, S, Burghart, R. Caldwell, W. Graham, R. Matheson, M. Murphy, R. Muller, P. Houhoulis, A. Jackson, D. Marelli, J. Nance, S. Nesbitt, all contributed crucial expertise and knowledge. A special thanks to J. Browder and an anonymous reviewer for their careful comments, which improved the manuscript. This work was supported by the Florida Fish and Wildlife Conservation Commission, the European Union Concerted Action Program (Contract # IC18-CT97-0175), and the Charles Darwin Research Station, Galapagos. Chapter 7 T. Tomascik, D. Srivastava, R. DeWreede, D. Zeller, J. Oliver, J. Ruesink, D. Pauly, W. Moka, L. Thompson, E. Buchary, and B. Wright provided helpful guidance and feedback. Thanks also to Sa'id, Sepipa, D. Steller, M. Erdmann, and several others for providing field assistance. This work was supported by the Geological Society of America (GSA) through the Harold T. Sterns Fellowship for doctoral research and an additional GSA research grant. I am forever grateful for the hospitality of the Indonesian people I had the fortune to encounter during field expeditions, especially my friend Y. Yasir. x for Molly Rose McNulty-Finn Clara Riley McNulty-Finn and Maura Joan McNulty-Finn for the oceans xi CHAPTER 1. Overview and Summary The notion that biological communities are regulated at a balanced state (i.e., an equilibrium state) persists throughout society and modern ecology. The western roots of this equilibrium paradigm include Calvin's transcendental notion of divine order (Green 1995) and strict mechanistic determinism as construed by Laplace (1819). These ideas continue to permeate modern scientific inquiry (Botkin 1990) despite their metaphysical rather than empirical foundations (Dupre 1993). Classical ecological thought embraced the culturally engrained presumption of a 'balance of nature' (see Egerton 1973), or a nature that is self-regulating (sensu Hutchinson 1948, Margalef 1968), or one that tends towards 'maturity' through succession (Elton 1930, Odum 1969, 1971, Christensen 1995). However, natural variability and scale constraints makes the rigorous demonstration of community stability or 'balance' very challenging (Connell and Sousa 1983, Pickett et al. 1992, Dodd et al. 1995). Even defining ecological stability has been so challenging that some authors have proposed new vocabularies (Grimm and Wissel 1997). Despite a paucity of direct empirical evidence of stability per se, the existence of stabilizing feedback mechanisms in ecosystems (Hutchinson 1948, Margalef 1968) continues to tempt ecologists with the notion that biological communities are attracted to 'equilibria' on some scale, even if 'dynamic' (Huston 1979, O'Neill 2001). Ecological stability is, arguably, an entirely scale-dependent concept (Connell and Sousa 1983, Levin 1992, O'Neill 2001). From one year to the next, the world around us remains largely self-similar. Such experiential observations explain the origins of the stability idea and lend credence to the notion that stability is not a delusion. Such evidence cannot, however, transcend context and scale dependency. Although the reluctance to surrender a comforting world view might be a principle reason for the attraction of ecologists to the idea of stability, some find practical convenience in the equilibrium notion when applied to appropriately-useful scales and inference (Pauly and Christensen 1995, Walters et al. 1997, Pauly etal. 2000). The alternative notion that change, chance, or chaos pervades nature is also traditional and persistent (Karma-glin-pa 14th century, Milton 16th century, Hesiod 800 be, Engels 1882, Poincare 1914, Lorenz 1963, Kirthisinghe 1984, Prigogine and Stengers 1984, Laszlo 1991, Waldrop 1992). Thus, the maturing of ecological thought has proceeded from simple conceptual models of a neatly balanced nature (e.g., Odum 1953) to conceptual models in which disturbance and stochasticity, i.e., indeterminacy on the scale of an organism's experience (sensu Botkin 1990), dominate the shaping of biological communities (O'Neill et al. 1982, Sousa 1984, Pickett et al. 1992, Levin 1999). The burgeoning integration of these two realms of thought, i.e., balance versus chaos, or equilibrium versus stochastic variability and change, represents a conceptual threshold for understanding the basic rules of community organization. Notions of self-regulation of biological communities on one hand and external regulation of biological communities (e.g., via physical forces) are both given credence by theory and empirical evidence, but debates over particular ecological dilemmas still organize according to camps that advocate 1 either one view or the other (see Wilson 1998). The intellectual tribalism and isolationism underlying such polarization of discourse perpetuates false dichotomies, as in some debates over the primacy of biotic and physical forces as organizers of communities. Those that continue splitting non-equilibrium mechanisms from equilibrium mechanisms in an attempt to understand community regulation risk continued perpetuation of such false dichotomies (Beisner et al. 2003). In reality, biotic communities are regulated and organized by mixtures of biological and physical forces that vary continually in both time and space (e.g., Okey 2003). A corollary to the notion of equilibrium, stability, or 'balance' is the idea that a given biological community has global stability (Lewontin 1969); that is to say, a single 'stable' state to which the community is attracted. This 'global attractor' is illustrated by a marble at the bottom of a bowl (Hurd and Wolf 1974), or valley. A disturbance can push the marble from the global basin of attraction, but the marble always returns to the previous state after some period of fluctuations. A strictly deterministic 'equilibrium' system, in which the biotic community possesses a single-unique-stable-state, will presumably always return to this initial 'global attractor' position when a disturbance or stressor is removed, even when the community is perturbed beyond historical levels of fluctuation. Another view is that biological communities have multiple stable states, or more than one local attractor. This is envisioned as valleys in a dynamical landscape wherein a strong disturbance can push a marble from one local basin of attraction to another (Lewontin 1969, Holling 1973, Sutherland 1974, May 1977). In a system containing stochasticity, historical accidents (sensu Sutherland 1974) can direct a natural community to one of many possible attractors on the broader dynamical landscape. The effects of shape, height, and steepness of the bowl on the marble represent both a community's resistance to being changed by external forces (Boesch 1974) and the community's resilience, which is how much disturbance (e.g., magnitude, frequency, severity) a community can endure and still return to its previous state (definition based on Holling and Clark 1975). Resilience is also thought to incorporate the speed of return to the equilibrium state (Boesch 1974, Pimm 1991). Any system equilibria, whether global or local, develops within the context of natural environmental variability and disturbance regimes in addition to the ever-changing biotic forces within a community. Anthropogenic or exotic disturbances modify the conditions in which these supposed equilibria developed, especially when new disturbances are exotic in the sense that they differ in magnitude and quality from natural regimes of disturbance or-sources of variability (Sousa 1984). Such new disturbances can thus push communities away from normal community attractors. Thus, populations in a system with multiple community states will shift to new levels and persist there after a perturbation of adequate severity, or exotic enough character. The salient point is that if multiple persistent states exist, communities might not recover to a previous state after stressors are alleviated, but rather remain in an alternative basin of attraction (e.g., Krebs 2001, p. 509). 2 The detection of multiple persistent states has been constrained by inherent difficulties in conducting ecological studies at appropriate, i.e., broad enough, scales in time or space (Connell and Sousa 1983, Petraitis and Latham 1999). Nevertheless, some indications of multiple community states have recently emerged. Examples from non-marine systems include lakes (Blindow et al. 1993, Scheffer et al. 1993, Moss et al. 1996, Romo et al. 1996, Scheffer et al. 1997, Weisner et al. 1997, Bachmann et al. 1999), streams (Strange et al. 1993), rangelands (Walker et al. 1981, Laycock 1991, Perrings and Walker 1997), deer and forest systems (Stromayer and Warren 1997), savannah woodlands (Dublin et al. 1990), sand dunes (Adema et al. 2002), laboratory mesocosms of phytoplanktors and aquatic crustaceans (Drake et al. 1993), and yeast cultures (Zamamiri et al. 2001). Furthermore, empirically-based models indicate the potential for multiple stable states in settings that include Everglade swamps (Dong et al. 2002), rangelands (Noy-Meir 1975, Westoby et al. 1989, Laycock 1991, Lockwood and Lockwood 1993), island systems (Sinclair et al. 1998, Ward and Thornton 1998), a butterfly metapopulation (Boughton 1999), and grasslands/woodlands (Fuhlendorf et al. 1996). Evidence of multiple persistent states has likewise emerged from marine settings such as the temperate rocky intertidal (Barkai and McQuaid 1988, Petraitis and Dudgeon 1999, Dudgeon and Petraitis 2001), the sub-tropical rocky intertidal (Barry 1989), temperate hard bottoms (Elner and Vadas 1990), coral reefs (Hatcher 1984, Knowlton 1992, McManus et al. 2000), and marine soft bottoms (Herman et al. 2001, van de Koppel et al. 2001). Several notable reviews address multiple stable states (Connell and Sousa 1983, Done 1992, Knowlton 1992, Carpenter 2000, Muradian 2001, Scheffer et al. 2001). Sutherland (1974) concluded that, ".. .multiple stable points are an undeniable reality in space and time...," because history (timing of disturbance) could determine alternate communities of marine fouling invertebrates, which then persist. He also pointed out that in addition to different stable points for an intact community, alternate community states can take the form of 'non-trivial boundary points' in which one or some species are removed or missing from the system. That is to say, the removal of a strongly interacting species can lead to a system that becomes 'stuck' at a 'boundary point' away from the previous 'equilibrium point' or 'attractor' as illustrated in the following description. The alternation of kelp forest and urchin barren communities along rocky coastlines of the northeast Pacific Ocean is a classic marine example of alternative states that are persistent (Estes and Palmisano 1974, Dayton 1975, Estes et al. 1978, Simenstad et al. 1978, Estes and Duggms 1995, Konar and Estes 2003). Removal of sea otters (Enhydra lutris) leads to replacement of kelp forests with urchin barrens because otters normally control these voracious grazers of canopy-forming macrophytes. Effects cascade throughout the system because kelp canopies and understories provide production and biogenic habitat structure for a broad suite of organisms. This shift occurs in kelp forests where the otter-mediated trophic cascade is not naturally dampened by persistent wave disturbance or other factors (Foster 1990, Kvitek et al. 1998, Konar 2000). Rather than being an alternative stable point for an intact community, however, the urchin barrens presumably represent a non-trivial boundary point of the system's global equilibrium (Sutherland 1974); i.e., the system is stable and very different from a kelp forest at this boundary point because one or more strongly interacting species is missing. By definition then, non-trivial boundary points are achieved whenever keystone species (Paine 1969, Power et al. 1996) are removed. In this classic example, we have two alternative states that are stable, but only because of extrinsic forces that occasionally remove the keystone predator. Thus, the system exhibits alternative persistent states, but does not intrinsically manifest the maintenance of alternative stable states. Similar examples in other kelp forests are described by Steneck et al. (2002). Evidence of (intrinsic) alternative stable states must consist of the testable or observable attributes of hysteresis, which is a disparity between the levels of stressors that cause sudden (catastrophic) forward and backwards shifts of an observed state, or state variable (Figure 1-1). In hysteretic systems, the alleviation of a given stressor will fail to restore the community to the previous state until the stressor is reduced to somewhat lower levels than those that caused the catastrophic shift to the degraded state, and sometimes much lower levels. The observation of non-linear or catastrophic shifts indicates resistance (or persistence) of the alternative community states. Such resistance implies reinforcing feedbacks of a given community state and thus some degree of stability. Related attributes can be examined in terms of space or time rather than level of stressor(s). Thus, observations of spatially distinct boundaries of community types along smooth environmental gradients, or observations of sudden changes, indicates the potential for hysteresis, or alternative stable states. Catastrophe theory, initiated by Thorn (1972) and also described by Zeeman (1976) and Saunders (1980) for both biological and social systems, goes parallel to the theory of alternative community states in the sense that catastrophes are the sudden shifts to a radically different community state. A catastrophic state-and-threshold model developed for rangeland dynamics (Laycock 1991) led Lockwood and Lockwood (1993) to suggest that catastrophe theory could unify competing theoretical frameworks (i.e., continuous change versus discontinuous change models) for understanding rangeland dynamics. The first four of these authors' five essential symptoms of catastrophe systems match our conceptual model of multiple stable states (Figure 1-1): 1. "Modality (distinct conditions or states of existence); 2. Inaccessibility (conditions which are very unstable); 3. Sudden changes (relatively rapid movement between states); 4. Hysteresis (processes associated with degradation or recovery are not readily reversible by simply inverting the sequence of events); and 5. Divergence (relatively small changes in initial conditions can result in dramatically different outcomes with time)." Their fifth symptom is best illustrated by imagining a third axis on Figure 1-1 projecting out the back of the page. We can also imagine that differing environmental conditions along this hypothetical third 4 axis changes the response curve to a more linear shape. The result is a cusp in the dynamical landscape (Zeeman 1976). A resulting property of this cusp landscape is that very slight changes on such a landscape can cause dramatically different outcomes, as when a hiker misses a switchback while descending a steep trail. One stable point Two possible stable points One stable point Response curve (Basin 1) Disturbance Threshold Response curve (Basin 2) Stressor(s) Figure 1-1. Response curve of a community that exhibits hysteresis. Theory suggests that response curves of community state variables sometimes fold back underneath themselves to form two possible equilibrium points at levels of stress within a certain range. A community's resilience, or 'stability,' keeps it within a basin of attraction (Basin 1) until a threshold of stress is reached and a catastrophic shift occurs, or until a disturbance of adequate severity or character can force the system beyond the boundary of instability to an alternative basin of attraction (Basin 2). A considerable reduction of the stressor might be required to restore the system to the previous realm (Basin 1). Real systems contain multiple stressors with different shapes of response curves. Thus, multiple basins of attraction on a far more complex dynamical landscape are possible in real systems. Rigorous criteria for evaluating claims of multiple stable states were developed by Connell and Sousa (1983), who found no compelling evidence of the phenomenon in a number of empirical studies of the time. They found three classes of shortcomings: (1) differing physical environments in the different alternate states, (2) persistence of alternate states only when artificial controls are maintained, and (3) inadequate evidence of stability due, e.g., to inappropriate scales. However, Peterson (1984) argued that differences in physical environments between alternate stable states could be biogenic, and thus intrinsic to the community, and that such modifications were reinforcing feedbacks and would be a likely mechanism for the maintenance of multiple stable states. He also argued that their criteria for stability were too strict because persistence and self-replication of patches indicates stability. Sousa and Connell 5 (1985) agreed with "Peterson's feedback mechanism," but they disagreed that their criteria for stability were too strict. Scheffer et al. (2001) seemed to echo these debates when they concluded, ".. .case studies [can] suggest shifts between alternative stable states... [but] proof of multiplicity of stable states is usually far from trivial." They suggest that observations of non-linear shifts and demonstrations of positive feedback mechanisms are, by themselves, insufficient evidence to demonstrate the existence of alternate stable states. They conclude that, ".. .the strongest cases for the existence of alternate stable states are based on combinations of approaches." The present study was guided by this 'broad evidence' principle. Six types of evidence were chosen for evaluating apparent shifts to alternative persistent states (no order of evidence priority is implied): 1. Dynamic simulations showing plausibility of shifts; 2. Reinforcing feedback mechanisms that are intrinsic to a given community state; 3. Abrupt interfaces or shifts between alternative community states (in time or space); 4. Repeated shifts; 5. Persistence of a given state beyond one complete population turnover; 6. Discontinuity between forward and backwards shifts along an axis of stress (hysteresis). All these types of evidence require some degree of empirical information. The present study was focused on the first type of evidence, but the existence of other evidence became apparent to varying extents as the case studies unfolded. Some information was collected directly and some was distilled from available literature. I examined four apparent marine community shifts that were indicated from exploratory-level observations and analyses. These continental shelf examples range from tropical to sub polar settings. 1. The appearance of 'anemone barrens' on previously diverse shallow reefs in the Galapagos; 2. A possible alternative community state following the catastrophic Exxon Valdez Oil Spill; 3. A possible benthic to pelagic shift in West Florida Shelf community organization; 4. A shift from corals to ephemeral algae on polluted Indonesian reefs. After assembling the types of evidence needed to evaluate the apparent shifts, and the possibility of alternative stable states, my general approach was to evaluate the four examples using different combinations of whole-system trophic modeling, field sampling, and observation to address as many of the types of evidence as feasible within the scope of the current study. Different combinations of evidence were evaluated according to the unique constraints of each system. A related objective was to explore potential explanations (i.e., mechanisms) for the existence of alternative persistent states in these marine communities. An underlying pre-supposition of this investigation was that explanations for the existence of alternative community states are likely to emerge from the marriage of equilibrium and non-equilibrium theoretical views of community organization (e.g., Okey 2003). 6 Scheffer et al. (2001) suggested that physical forces are the primary the cause of hysteresis in ocean ecosystems, while both physical and biotic factors explain hysteresis in coral reef systems. The present study indicates that combinations of physical and biotic forces cause the apparent community shifts in at least three of the four marine communities examined herein: The replacement of a diverse reef community by 'anemone barrens' could have been initiated by recent strengthening of El Nino events, but the resulting alternative community state appears to be maintained by positive biological feedbacks (Chapters 2 and 3); A search for keystones in Prince William Sound provides a general approach that provides insights into how communities can shift (Chapter 4); Simulated alternate community states in Prince William Sound, Alaska indicate that the trophic character of disturbances may rival their magnitude and severity as shapers of communities, and that strictly biotic forces can maintain a radically changed system (Chapter 5); Simulations of seafloor shading by plankton blooms over the West Florida Shelf indicate the potential for shading to cause broad shifts of continental shelf biological communities due to benthic to pelagic shifts in energy flow patterns (Chapter 6); Finally, sedimentation or other types of pollution appear to have influenced shifts in the spatial and temporal patterns of coral reef flora and fauna in South Sulawesi, Indonesia, although modifications of fish communities might have shaped those apparent community shifts as well (Chapter 7). Mounting evidence that marine ecological systems are degraded and continuing to lose their biotic integrity has caused growing concern throughout the world (Vitousek et al. 1997, Jackson et al. 2001). This concern is particularly focused on continental shelf systems in which primary and secondary production is very high (Smith 1981, Duggins et al. 1989, Vetter 1995), and from which large quantities of food are collected by humans. A variety of human activities influence the organization of biological communities on, and overlying, continental shelves. Fisheries and various types of pollution are two conspicuous agents of change, or disturbance, in these systems (Goni 1998, Pauly et al. 1998a, Cloern 2001), but climatic (i.e., oceanographic) changes on various scales are also known to have profound effects on continental shelf biota, and marine biota in general (e.g., Glynn 1990, Anderson and Piatt 1999, Smith etal. 1999). The general goal of this dissertation is to scrutinize mechanisms by which human activities might fundamentally alter the natural biological communities of some continental shelves, and whether these changed states might persist within in local basins of attraction. Each of the case study examples include predicted or observed shifts in community organization and some preliminary attempts to scrutinize the natural and anthropogenic variables in order to explain the indicated shifts. A final introductory caveat must be understood when undertaking a study that fundamentally involves the notion of stability: it is unlikely that strict criteria for evaluating stability itself can be satisfied or even evaluated well in marine ecological settings. Indeed, natural populations or communities are variable at most, if not all, scales. The implications of finding compelling evidence of alternate stable states in natural communities would be far reaching because of the implications of irreversibility, as would the 7 implications of demonstrating alternate persistent states. Thus, my approach is to gather various types of evidence that might indicate alternative stable states, and otherwise to explore the possible reasons that some alternate community states appear persistent but are not indicated to be intrinsically and strictly stable. Demonstrating persistence of a community within stable limit cycles, or as Connell and Sousa (1983) put it, "stochastic boundedness without equilibria," would also imply similar cautions for human interactions with ecosystems. Indeed, the initial management implications would be similar with degraded communities that have no stability, persistence, or boundedness whatsoever. Restoring community or ecosystem services in degraded ecosystems usually requires the reduction of anthropogenic stressors or disturbance, whether or not the systems exhibit alternate domains of attraction or stability. Those systems that do exhibit hysteresis and alternate stable (or persistent) states will ultimately require special attention in terms of policy and management attention, and in terms of basic research. The present research was conceived to help avert ecological and economic catastrophes that are, by definition, characteristic of systems that exhibit hysteresis. The work was undertaken because it is likely that such precautionary investments will prove crucial in averting catastrophes that will be difficult and expensive, if not impossible, to reverse. Modeling and direct empirical studies The power and usefulness of analytical tools for characterizing biological communities and exploring ecological mechanisms has increased tremendously in recent years. Network analysis of food webs and dynamic simulation capabilities, such as those used in the mass-balanced trophic modeling approach Ecopath with Ecosim, exemplify this advancement (Christensen and Pauly 1992, Walters et al. 1997, Walters et al. 1999, Pauly et al. 2000, Walters et al. 2000). Such whole system modeling approaches are built on empirically based characterizations of food webs, and sometimes represent knowledge distilled from major scientific programs, or from many decades of empirical research. These new approaches to ecosystem synthesis and analysis can help provide unprecedented insights into how nature works and how humans influence nature. Such insights can, nevertheless, be critically limited without comparisons to independent empirical studies. My general approach, thus, is to compare the predictions and mechanisms indicated by trophic modeling and simulation with independently derived empirical evidence to gauge whether the models and simulations are acceptable for their intended uses (sensu Rykiel 1996) and to judge the overall usefulness of resulting predictions and insights. Ecopath with Ecosim methodology Ecopath trophic models describe the state of energy flows in a food web. They are designed to include all biotic components of an ecosystem, and the most typical currency is biomass wet-weight (used here). Polovina (1984) developed Ecopath to study coral reefs at French Frigate Shoals. A variety of 8 dynamic capabilities have since been added (e.g., Christensen and Pauly 1992, Walters et al. 1997, Walters et al. 1999, Christensen et al. 2000, Pauly et al. 2000). Scores of applications of Ecopath with Ecosim can be found at: http.V/www.ecopath.org/, along with the freely distributed software and documentation. Although the formulations and basic concepts are accessible in these venues, the general approach is summarized here. The master equation (Equation 1-1) expresses the law of conservation of mass or energy and it indicates the basic input parameters. This equation balances a group's net production (terms to the left of the equal sign) with all sources of mortality, migration, or change for that group (terms to the right). More specifically, it says that the net production of a functional group equals the sum of (1) the total mass (or energy) removed by predators and fisheries, (2) the net biomass accumulation of the group, (3) the net migration of the group's biomass, and (4) the mass flowing to detritus. Bj and Bj are biomasses of prey Q and predators Q respectively; P/Bj is the production/biomass ratio, equivalent to total mortality (Z) in most circumstances (Allen 1971); EEj is the ecotrophic efficiency; the fraction of the total production of a group utilized in the system; Yj is the fisheries catch per unit area and time (i.e., Y = F*B); Q/Bj is the food consumption per unit biomass of j; DCji is the contribution of i to the diet of j; BAj is the biomass accumulation of i (positive or negative); and NMj is the net migration of I (emigration less immigration). The implied thermodynamic constraints of this equation underscore the power of Ecopath models as a focal point for refinement of ecosystem information. The need to reconcile energy production and demand among components of the food web narrows the possible ranges of parameter estimates for particular groups. Inclusion of a biomass accumulation factor and migration factor in the general Ecopath equation distinguishes Ecopath modeling as an 'energy continuity' approach rather than a strictly 'steady state' approach. Conservation of energy {continuity) is assumed for every identified component of the ecosystem, and the whole system. This basic constraint enables representation of changes in populations (i.e., functional groups) when expressed in dynamic form. Ecopath was considerably refined with the dynamic simulation routines Ecosim and Ecospace (see Walters et al. 1997, Walters et al. 1999, Pauly et al. 2000, Walters et al. 2000). In Ecosim, information in the static Ecopath file is re-expressed in a dynamic formulation (Equation 1-2). Equation 1-1. B| • (P/B); - EE; = Yj + I Bj • (Q/B)j • DCjj + BAj + NM; dB, n Equation 1-2. 9 f (Bj) is a function of Bf if (i) is a primary producer, or f (Bj) = g;Z CJ; (Bj.Bj) if (i) is a consumer; gi is the net growth efficiency; Cjj(Bj.Bj) is the function used to predict consumption rates from Bj to Bj. Ecosim uses a function for Cy derived from assuming possible spatial/behavioral limitations in predation rates (Equation 1-3): v.aBB. Equation 1 -3. C„. = " 'J ' 1 11 v, + v'!l+a,lBl Cy is the trophic flow of biomass per time, between prey (i) and predator (j) pools; Bj and Bj are the biomasses of prey and predators, respectively; a;j is the rate of effective search for prey i by predator j; and Vjj and v'y are prey vulnerability parameters, with default setting V;J = v'y. Parameters v^ and v'y represent prey vulnerabilities, or the rate of exchange of biomass between two prey behavioral states: a state in which all predators have full access to prey and a state in which prey have full refuge from predators. Prey use refugia in real ecosystems. Thus, not all prey biomass is vulnerable to predation at any given time, and predator-prey relationships are limited by behavioral and physical mechanisms. Ecosim is designed so that the user can specify the type of trophic control (Lotka-Volterra type vs. donor control) that mediates any interaction in the food web. Maximum consumption rates are hypothesized, and thus the rate of exchange of biomass (Vy) that a predator normally exerts. For low predator biomass or high prey vulnerability (vy) the functional relationship approximates a mass-action flow, or Lotka-Volterra type interaction (c = aBjBj) implying a strong 'top-down' effect. For high predator biomass or low prey vulnerabilities the functional relationship approaches a donor-controlled (bottom-up) flow rate (c = VyBj), so Vy is the maximum possible instantaneous mortality rate that j can cause on i (see Walters et al. 1997). Prey vulnerabilities can be specified by adjusting the proportion of prey in vulnerable and invulnerable states (pools) via adjustment of the v values, which are scaled such that pure Lotka-Volterra type control = 1 and pure donor control = 0. In the real world, this mixture of trophic control is mediated by temporal or spatial refugia, or by the relative primacy of physical and biotic forces in regulating communities, i.e., predator-prey interactions. Although examples of biological communities with alternate stable states seem to be accumulating rapidly in the general ecological literature, compelling examples of truly alternate stable states in marine biological communities (wherein intrinsic mechanisms reinforce alternate states) are very rare (e.g., Barkai and McQuaid 1988). This is either because alternate stable states are truly rare in marine communities, or because the evidence required for demonstrating the existence of alternate stable states is 10 difficult to obtain from marine communities due to the unique constraints related to studying them. This dissertation is an examination of four marine communities, using empirically-based whole food web modelling approaches and empirical field studies, to explore the potential ability of particular mechanisms in shifting and maintaining alternate stable states. The results of these explorations are intended to be considered hypotheses that can be evaluated with iterations of experimentation (manipulative or "natural") and further simulations. LITERATURE CITED Adema, E. B., A. P. Grootjans, J. Petersen, and J. Grijpstra. 2002. Alternative stable states in a wet calcareous dune slack in The Netherlands. Journal of Vegetation Science 13:107-114. Allen, K. R. 1971. Relation between production and biomass. Journal of the Fisheries Research Board of Canada 28:1573-1581. Anderson, P. J., and J. F. Piatt. 1999. Community reorganization in the Gulf of Alaska following ocean climate regime shift. Marine Ecology Progress Series 189:117-123. Bachmann, R. W., M. V. Hoyer, and Canfield Daniel E, Jr. 1999. The restoration of Lake Apopka in relation to alternative stable states. Ffydrobiologia 394:219-232. Barkai, A., and C. McQuaid. 1988. Predator-prey role reversal in a marine benthic ecosystem. Science (Washington D C) 242:62-67. Barry, J. P. 1989. Pattern and process: Patch dynamics in a rocky intertidal community in southern California. Dissertation Abstracts International Part B: Science and Engineering 49:359. Beisner, B. E., D. T. Hanydon, and C. K. 2003. Alternate stable states in ecology. Frontiers in Ecology and the Environment 1:376-382. Blindow, I., G. Andersson, A. Hargeby, and S. Johansson. 1993. Long-term pattern of alternative stable states in 2 shallow eutrophic lakes. Freshwater Biology 30:159-167. Boesch, D. F. 1974. Diversity stability and response to human disturbance in estuarine ecosystems. Pages 109-114 in Proceedings of the First International Congress of Ecology. Structure, Functioning and Management of Ecosystems, Hague, Netherlands. International Scholarly Book Services: Portland, Oregon. Botkin, D. B. 1990. Discordant harmonies: a new ecology for the twenty-first century. Oxford University Press, New York. Boughton, D. A. 1999. Empirical evidence for complex source-sink dynamics with alternative states in a butterfly metapopulation. Ecology 80:2727-2739. Carpenter, S. R. 2000. Alternate states of ecosystems: Evidence and its implications for environmental decisions. Pages 357-383 in M. C. Press, N. Huntley, and S. Levin, editors. Ecology: achievement and challenge. Blackwell, London. Christensen, V. 1995. Ecosystem maturity: Towards quantification. Ecological Modelling 77:3-32. Christensen, V., and D. Pauly. 1992. Ecopath II: a software for balancing steady-state ecosystem models and calculating network characteristics. Ecological Modelling 61:169-185. Christensen, V., C. J. Walters, and D. Pauly. 2000. Ecopath with Ecosim: a user's guide. Univ. of British Columbia, Fisheries Centre, Vancouver, Canada and ICLARM, Penang, Malaysia. Cloern, J. E. 2001. Our evolving conceptual model of the coastal eutrophication problem. Marine Ecology Progress Series 210:223-253. Connell, J. H., and W. P. Sousa. 1983. On the evidence needed to judge ecological stability or persistence. The American Naturalist 121:789-824. Dayton, P. K. 1975. Experimental studies of algal canopy interactions in a sea otter dominated kelp community at Amchitka Island, Alaska, USA. Fishery Bulletin 73:230-237. Dodd, M., J. Silvertown, K. Mcconway, J. Potts, and M. Crawley. 1995. Community stability: A 60-year record of trends and outbreaks in the occurrence of species in the park grass experiment. Journal of Ecology 83:277-285. 11 Done, T. J. 1992. Phase-shifts in coral-reef communities and their ecological significance. Hydrobiologia 247:121-132. Dong, Q., P. V. McCormick, F. H. Sklar, and D. L. DeAngelis. 2002. Structural instability, multiple stable states, and hysteresis in periphyton driven by phosphorus enrichment in the Everglades. Theoretical Population Biology 61:1-13. Drake, J. A., T. E. Flum, G. J. Witteman, T. Voskuil, A. M. Hoylman, C. Creason, D. A. Kenny, G. R. Huxel, C. S. L. Duncan, and R. Jeffrey. 1993. The construction and assembly of an ecological landscape. Journal of Animal Ecology 62:117-130. Dublin, H. T., A. R. E. Sinclair, and J. McGlade. 1990. Elephants and fire as causes of multiple stable states in the Serengeti-Mara Tanzania woodlands. Journal of Animal Ecology 59:1147-1164. Dudgeon, S., and P. S. Petraitis. 2001. Scale-dependent recruitment and divergence of intertidal communities. Ecology 82:991-1006. Duggins, D. O., C. A. Simenstad, and J. A. Estes. 1989. Magnification of secondary production by kelp detritus in coastal marine ecosystems. Science 245:170-173. Dupre, J. 1993. The disorder of things : metaphysical foundations of the disunity of science. Harvard University Press, Cambridge, Mass. Egerton, F. N. 1973. Changing concepts of the balance of nature. Quarterly Review of Biology 48:322-350. Elner, R. W., and R. L. Vadas, Sr. 1990. Inference in ecology the sea urchin phenomenon in the northwestern Atlantic Ocean. American Naturalist 136:108-125. Elton, C. S. 1930. Animal ecology and evolution. Clarendon Press, Oxford. Engels, F. 1882. Dialectics of nature. International Publishers [1976], New York. Estes, J. A., and D. O. Duggins. 1995. Sea otters and kelp forests in Alaska: Generality and variation in a community ecological paradigm. Ecological Monographs 65:75-100. Estes, J. A., and J. F. Palmisano. 1974. Sea otters: their role in structuring nearshore communities. Science (Washington D C) 185:1058-1060. Estes, J. A., N. S. Smith, and J. F. Palmisano. 1978. Sea otter predation and community organization in the western Aleutian Islands, Alaska. Ecology 59:822-833. Foster, M. S. 1990. Organization of macroalgal assemblages in the northeast Pacific: the assumption of homogeneity and the illusion of generality. Hydrobiologia 192:21-33. Fuhlendorf, S. D., F. E. Smeins, and W. E. Grant. 1996. Simulation of a fire-sensitive ecological threshold: A case study of Ashe juniper on the Edwards Plateau of Texas, USA. Ecological Modelling 90:245-255. Glynn, P. W. 1990. Global ecological consequences of the 1982-83 El Nino-Southern Oscillation. Elsevier, Amsterdam ; New York. Goni, R. 1998. Ecosystem effects of marine fisheries: An overview. Ocean & Coastal Management 40:37-64. Green, R. 1995. The thwarting of Laplace's demon: arguments against the mechanistic world-view. Macmillan Press, Basingstoke. Grimm, V., and C. Wissel. 1997. Babel, or the ecological stability discussions: An inventory and analysis of terminology and a guide for avoiding confusion. Oecologia 109:323-334. Hatcher, B. G. 1984. A maritime accident provides evidence for alternate stable states in benthic communities on coral reefs. Coral Reefs 3:199-204. Herman, P. M. J., J. J. Middelburg, and C. H. R. Heip. 2001. Benthic community structure and sediment processes on an intertidal flat: results from the ECOFLAT project. Continental Shelf Research 21:2055-2071. Hesiod. 800 be. Theogony. M. L. West, editor. Clarendon Press [1966], Oxford. Holling, C. S. 1973. Resilience and stability of ecological systems. Annual Review of Ecology and Systematics, Vol. 4. Vii+424p. Illus. Map. 4:1-23. Holling, C. S., and W. C. Clark. 1975. Notes towards a science of ecological management. Pages 247-251 in W. H. Van Dobben and R. H. Lowe-McConnell, editors. Unifying Concepts in Ecology. First International Congress. Hague, Netherlands. Hague, Netherlands; Centre for Agricultural Publishing and Documentation: Wageningen, Netherlands. 12 Hurd, L. E., and L. L. Wolf. 1974. Stability in relation to nutrient enrichment in arthropod consumers of old-field successional ecosystems. Ecological Monographs 44:465-482. Huston, M. 1979. Resilience and stability in ecological systems. Annual Review of Ecology and Systematics 4:1-23. Hutchinson, G. E. 1948. Circular causal systems in ecology. Annals of the New York Academy of Science 50:221-246. Jackson, J. B. C, M. X. Kirby, W. H. Berger, K. A. Bjorndal, L. W. Botsford, B. J. Bourque, R. H. Bradbury, R. Cooke, J. Erlandson, J. A. Estes, T. P. Hughes, S. Kidwell, C. B. Lange, H. S. Lemhan, J. M. Pandolfi, C. H. Peterson, R. S. Steneck, M. J. Tegner, and R. R. Warner. 2001. Historical overfishing and the recent collapse of coastal ecosystems. Science 293:629-638. Karma-glin-pa. 14th century. The Tibetan book of the dead, or, The after-death experiences on the Bardo plane : according to Lama Kazi Dawa-Samdup's English rendering, 3rd edition. Oxford University Press [1960], New York. Kirthisinghe, B. P. 1984. Buddhism and science. Motilal Banarsidass, Delhi. Knowlton, N. 1992. Thresholds and multiple stable states in coral reef community dynamics. American Zoologist 32:674-682. Konar, B. 2000. Limited effects of a keystone species: trends of sea otters and kelp forests at the Semichi Islands, Alaska. Marine Ecology Progress Series 199:271-280. Konar, B., and J. A. Estes. 2003. The stability of boundary regions between kelp beds and deforested areas. Ecology 84:174-185. Krebs, C. J. 2001. Ecology: the experimental analysis of distribution and abundance, 5th edition. Benjamin Cummings, San Francisco. Kvitek, R. G., P. J. Iampietro, and C. E. Bowlby. 1998. Sea otters and benthic prey communities: A direct test of the sea otter as keystone predator in Washington State. Marine Mammal Science 14:895-902. Laplace, P. S. 1819. A Philosophical Essay on Probabilities, Student's edition. Dover Publications, New York, [1952]. Laszlo, E. 1991. The age of bifurcation : understanding the changing world. Gordon and Breach, Philadelphia. Laycock, W. A. 1991. Stable states and thresholds of range condition on north-American rangelands: A viewpoint. Journal of Range Management 44:427-433. Levin, S. A. 1992. The problem of pattern and scale in ecology. Ecology 73:1943-1967. Levin, S. A. 1999. Fragile Dominion: Complexity and the Commons. Perseus Books, Reading, Mass. Lewontin, R. C. 1969. The meaning of stability. Brookhaven Symposium of Biology 22:13-24. Lockwood, J. A., and D. R. Lockwood. 1993. Catastrophe theory: A unified paradigm for rangeland ecosystem dynamics. Journal of Range Management 46:282-288. Lorenz, E. N. 1963. Deterministic nonperiodic flow. Journal of the Atmospheric Sciences 20:130-148. Margalef, R. 1968. Perspectives in Ecological Theory. The University of Chicago Press, Chicago. May, R. M. 1977. Thresholds and breakpoints in ecosystems with a multiplicity of stable states. Nature 269:471-478. McManus, J. W., L. A. B. Menez, K. N. Kesner-Reyes, S. G. Vergara, and M. C. Ablan. 2000. Coral reef fishing and coral-algal phase shifts: implications for global reef status. Ices Journal of Marine Science 57:572-578. Milton, J. 16th century. Paradise lost. Macmillan [1993], New York. Moss, B., J. Stansfield, K. Irvine, M. Perrow, and G. Phillips. 1996. Progressive restoration of a shallow lake: A 12-year experiment in isolation, sediment removal and biomanipulation. Journal of Applied Ecology 33:71-86. Muradian, R. 2001. Ecological thresholds: a survey. Ecological Economics 38:7-24. Noy-Meir, I. 1975. Stability of grazing systems an application of predator prey graphs. Journal of Ecology 63:459-482. Odum, E. P. 1953. Fundamentals of Ecology. Saunders, Philadelphia. Odum, E. P. 1969. The strategy of ecosystem development. Science 164:262-270. Odum, E. P. 1971. Fundamentals of ecology, 3d ~ edition. Saunders, Philadelphia. 13 Okey, T. A. 2003. Macrobenthic colonist guilds and renegades in Monterey Canyon (USA) drift algae: Partitioning multidimensions. Ecological Monographs 73:415-440. O'Neill, R. V. 2001. Is it time to bury the ecosystem concept? (With full military honors of course!). Ecology 82:3275-3284. O'Neill, R. V., R. H. Gardner, and D. E. Weller. 1982. Chaotic models as representations of ecological systems. American Naturalist 120:259-263. Paine, R. T. 1969. A note on trophic complexity and community stability. American Naturalist 103:91-93. Pauly, D., and V. Christensen. 1995. Primary production required to sustain global fisheries. Nature 374:255-257. Pauly, D., V. Christensen, J. Dalsgaard, R. Froese, and F. Torres. 1998. Fishing down marine food webs. Science 279:860-863. Pauly, D., V. Christensen, and C. Walters. 2000. Ecopath, Ecosim, and Ecospace as tools for evaluating ecosystem impact of fisheries. Ices Journal of Marine Science 57:697-706. Perrings, C, and B. Walker. 1997. Biodiversity, resilience and the control of ecological-economic systems: The case of fire-driven rangelands. Ecological Economics 22:73-83. Peterson, C. H. 1984. Does a rigorous criterion for environmental identity preclude the existence of multiple stable points? American Naturalist 124:127-133. Petraitis, P. S., and S. R. Dudgeon. 1999. Experimental evidence for the origin of alternative communities on rocky intertidal shores. Oikos 84:239-245. Petraitis, P. S., and R. E. Latham. 1999. The importance of scale in testing the origins of alternative community states. Ecology 80:429-442. Pickett, S. T. A., V. A. Parker, and P. L. Fielder. 1992. The new paradigm in ecology: Implications for conservation above the species level. Pages 65-88 in P. L. Fielder and S. K. Jain, editors. Conservation biology: The theory and practice of nature conservation, preservation, and management. Chapman and Hall, New York. Pimm, S. L. 1991. The balance of nature? Ecological issues in the conservation of species and communities. University of Chicago Press, Chicago. Poincare, H. 1914. Science and method. T. Nelson, New York. Polovina, J. J. 1984. Model of a coral reef ecosystem 1. The Ecopath model and its application to French Frigate Shoals. Coral Reefs 3:1-12. Power, M. E., D. Tilman, J. A. Estes, B. A. Menge, W. J. Bond, L. S. Mills, G. Daily, J. C. Castilla, J. Lubchenco, and R. T. Paine. 1996. Challenges in the quest for keystones. Bioscience 46:609-620. Prigogine, I., and I. Stengers. 1984. Order out of chaos : man's new dialogue with nature. Bantam Books, New York, N.Y. Romo, S., E. VanDonk, R. Gylstra, and R. Gulati. 1996. A multivariate analysis of phytoplankton and food web changes in a shallow biomanipulated lake. Freshwater Biology 36:683-696. Rykiel, E. J. 1996. Testing ecological models: The meaning of validation. Ecological Modelling 90:229-244. Saunders, P. T. 1980. An introduction to catastrophe theory. Cambridge University Press, Cambridge [Eng.] ; New York. Scheffer, M., S. Carpenter, J. A. Foley, C. Folke, and B. Walker. 2001. Catastrophic shifts in ecosystems. Nature 413:591-596. Scheffer, M., S. H. Hosper, M.-L. Meijer, B. Moss, and E. Jeppesen. 1993. Alternative equilibria in shallow lakes. Trends in Ecology & Evolution 8:275-279. Scheffer, M., S. Rinaldi, A. Gragnani, L. R. Mur, and E. H. vanNes. 1997. On the dominance of filamentous cyanobacteria in shallow, turbid lakes. Ecology 78:272-282. Simenstad, C. A., J. A. Estes, and K. W. Kenyon. 1978. Aleuts, sea otters, and alternative stable-state communities. Science (Wash.) 200:403-411. Sinclair, A. R. E., R. P. Pech, C. R. Dickman, D. Hik, P. Mahon, and A. E. Newsome. 1998. Predicting effects of predation on conservation of endangered prey. Conservation Biology 12:564-575. Smith, R. C, D. Ainley, K. Baker, E. Domack, S. Emslie, B. Fraser, J. Kennett, A. Leventer, E. Mosley-Thompson, S. Stammerjohn, and M. Vernet. 1999. Marine ecosystem sensitivity to climate change. Bioscience 49:393-404. Smith, S. V. 1981. Marine macrophytes as a global carbon sink. Science 211:838-840. 14 Sousa, W. P. 1984. The role of disturbance in natural communities. Annual Review of Ecology and Systematics 15:353-392. Sousa, W. P., and J. H. Connell. 1985. Further comments on the evidence for multiple stable points in natural communities. American Naturalist 125:612-615. Steneck, R. S., M. H. Graham, B. J. Bourque, D. Corbett, J. M. Erlandson, J. A. Estes, and M. J. Tegner. 2002. Kelp forest ecosystems: biodiversity, stability, resilience and future. Environmental Conservation 29:436-459. Strange, E. M., P. B. Moyle, and T. C. Foin. 1993. Interactions between stochastic and deterministic processes in stream fish community assembly. Environmental biology of fishes. The Hague [ENVIRON. BIOL. FISH.] 36:1-15. Stromayer, K. A. K, and R. J. Warren. 1997. Are overabundant deer herds in the eastern United States creating alternate stable states in forest plant communities? Wildlife Society Bulletin 25:227-234. Sutherland, J. P. 1974. Multiple stable points in natural communities. The American Naturalist 108:859-873. Thorn, R. 1972. Stabilite Structurelle et Morphogenese: Essai d'une Theorie Generate des Modeles. New York, Benjamin (English translation by D.H. Fowler. 1975. Structural stability and morphogenesis: An outline of a general theory of models. Benjamin, Reading). Benjamin, New York. van de Koppel, J., P. M. J. Herman, P. Thoolen, and C. H. R. Heip. 2001. Do alternate stable states occur in natural ecosystems? Evidence from a tidal flat. Ecology 82:3449-3461. Vetter, E. W. 1995. Detritus-based patches of high secondary production in the nearshore benthos. Marine Ecology Progress Series 120:251-262. Vitousek, P. M., H. A. Mooney, J. Lubchenco, and J. M. Melillo. 1997. Human domination of Earth's ecosystems. Science 277:494-499. Waldrop, M. M. 1992. Complexity: The emerging science at the edge of order and chaos. Simon & Schuster, New York. Walker, B. H., D. Ludwig, C. S. Holling, and R. M. Peterman. 1981. Stability of semi-arid savanna grazing systems. Journal of Ecology 69:473-498. Walters, C, V. Christensen, and D. Pauly. 1997. Structuring dynamic models of exploited ecosystems from trophic mass-balance assessments. Reviews in Fish Biology and Fisheries 7:139-172. Walters, C, D. Pauly, and V. Christensen. 1999. Ecospace: prediction of mesoscale spatial patterns in trophic relationships of exploited ecosystems, with emphasis on the impacts of marine protected areas. Ecosystems 2:539-554. Walters, C, D. Pauly, V. Christensen, and J. F. Kitchell. 2000. Representing density dependent consequences of life history strategies in aquatic ecosystems: Ecosim II. Ecosystems 3:70-83. Ward, S. A., and I. W. B. Thornton. 1998. Equilibrium theory and alternative stable equilibria. Journal of Biogeography 25:615-622. Weisner, S. E. B., J. A. Strand, and H. Sandsten. 1997. Mechanisms regulating abundance of submerged vegetation in shallow eutrophic lakes. Oecologia 109:592-599. Westoby, M., B. Walker, and I. Noy-Meir. 1989. Opportunistic management of rangelands not at equilibrium. Journal of Range Management 42:266-274. Wilson, E. O. 1998. Consilience: The unity of knowledge, 1st edition. Knopf: Distributed by Random House, New York. Zamamiri, A. Q. M., G. Birol, and M. A. Hjortso. 2001. Multiple stable states and hysteresis in continuous, oscillating cultures of budding yeast. Biotechnology and Bioengineering 75:305-312. Zeeman, E. C. 1976. Catastrophe theory. Scientific American 234:65-83. 15 CHAPTER 2. Trophic model of a Galapagos rocky reef with simulations of fishing impacts ABSTRACT A balanced trophic model of a Galapagos rocky reef system was constructed using Ecopath and Ecosim. The Ecopath approach allowed characterization of food web structure through integration of disparate ecosystem information derived from many years of study of Galapagos shallow-water rocky reefs. Ecosim and Ecospace routines enabled us to explore various hypotheses about system dynamics.as well as potential solutions to conservation concerns about overfishing. A full series of functional group removal simulations resulted in estimations of trophic interaction strengths and 'keystone' potentials for the all the living functional groups in the 43-box model (the detritus group is non-living). Relative interaction strengths in a pristine unfished system are likely to be quite different from interaction strengths indicated by this present-day model. At present, humans extract food from very low trophic levels (mean trophic level = 2.3) in Galapagos rocky reef systems because sea cucumbers and detritivorous mullets comprised 71% and 15% respectively of the total fisheries catch. Catch rates of sea cucumbers (Stichopus fuscus; referred to here as 'pepinos') are shown to be unsustainable, and the population should be declining rapidly. The exclusion of fishing from 23% of the total reef area, representing a hypothetical non-extractive zone, prevented the functional extinction of pepinos that the present analysis predicted to occur with no areas protected (given 1999-2000 capture rates). Even with 23% of the hypothetical area protected, pepinos were predicted to decline overall to a stable 36% of their current estimated biomass. Pepino biomass was predicted to increase to 8 times that of current levels if pepino fishing were stopped altogether. INTRODUCTION The Galapagos Archipelago lies in the equatorial eastern sector of the Pacific Ocean, about 1000 km west of the coast of Ecuador, South America (between 01°40'N-0P25'S and 89°15'W-92°00'W) (Figure 2-1). The archipelago consists of 13 large and 6 small islands, 42 islets, and numerous exposed rocks (Snell et al., 1995). These current islands represent the tops of relatively young volcanoes that rose from the sea between 1 and 3 million years ago, though submerged islands in the archipelago have been dated at 9 million years old, and islands might have been produced over this mantle hotspot for much longer (Christie et al., 1992). The islands rise from a relatively shallow (<200m) Galapagos Platform surrounded by deep waters (>1000 m). The location of the Galapagos Archipelago, at the confluence of warm currents (26-29°C) from the north, cool waters (20-22 °C) from the southwest, and nutrient-rich upwelling waters from the west propagating eastward, has led to complex and poorly understood marine and coastal ecosystems (Houvenaghel, 1984; Wellington, 1984; James, 1991). Between three and five major biogeographic units 16 have been proposed for the archipelago; however, the number of units and their boundaries still require clarification (Abbott, 1966; Harris, 1969; Jennings et al., 1994; Banks, 1999; Wellington et al, 2001). The largest regional mix of marine species occurs on the central Galapagos shelf, including the study area located around the island of Floreana (Figure 2-2). This area is characterized by a particularly diverse combination of warm- and cool-water biota (Witman and Smith, 2003). 92° w 91° W 90° W 0°N' 1°S1 Zonation /\/2.1 No-Entry & No-Take /\/ 2.2 No-Take Tourism 23 Fishing & Conservation Management , 2.4 Locally Defined Pinta Marchena Genovesa / Fernandina , Santiago Isabela Santa Cruz • o 0 10 Kilometers Model Area • • Floreana San Cristobal Espaftola 0°N IS 92° W 91° W 90° W Figure 2-1. Map of the Galapagos Islands showing the coastal use zoning scheme that has been in effect since 2000. Floreana Island is shown in the lower central part of the map. Wolf and Darwin Islands are not pictured; they are smaller and they lie to the NNW of the main islands. The shallow water ecosystems around Floreana Island consist primarily of sloping lava fields interspersed with sandy pocket beaches, as do most of the coastal shores of the Galapagos. These beaches are composed of both biogenic material (white and brown sand mainly from corals and echinoid tests) and pulverized lava (black sand), while sporadic mangrove forests occur in sheltered inlets. Subtidal lava reefs surround all the Galapagos Islands, and they can be subdivided into categories that include bedrock, boulders, cobbles, and sand (Figure 2-2). 17 100% -I 90% • 80% • 70% -c 60% • 5 :| sodn 50% -o u 40% \ 30% \ 20% -10% -0% -• Sand • Small Rocks • Large Rocks Bedrock Waypoint reference Figure 2-2. A map of Floreana Island and the waypoint areas used to estimate proportions of habitat types around the island. The accompanying chart in the lower panel shows the estimated proportions of habitat types at each waypoint reference area. These relative proportions were used to estimating biomasses of various functional groups in the modeled area. 18 Cool nutrient-rich waters on the Galapagos platform support high biomasses of small pelagic fishes such as sardines, thread-herrings, anchovies, mackerel, scads, halfbeaks and lantern fishes, which in turn feed substantial populations of top predators such as sharks, tuna, wahoo, billfishes, jacks, barracuda, dolphinfish, seabirds and toothed cetaceans (Feldman, 1985, 1986), many of which visit and feed actively on Galapagos rocky reef habitats. This productive pelagic system surrounds and interfaces with the benthic rocky reef habitats fringing each island of the Galapagos Archipelago, and plankton carried by oceanic currents is a major source of primary production on the reefs. Planktivorous reef fishes inhabit the boulder-strewn reef and feed in the water column. They include gringo (Paranthias colonus)—the most abundant Galapagos reef-dwelling fish species. In addition to sustaining large numbers of pelagic and reef-associated fishes and seabirds, plankton supports a high biomass of suspension and filter feeding invertebrates. These include the barnacle Megabalanus peninsularis, the sessile mollusc Hipponix sp., brittle stars, and the black and stony corals Antipathes spp. and Pavona spp. Benthic primary production is the other major source of primary production in this system. Galapagos rocky reefs in some areas include an algal turf and foliose macroalgae that, along with diatoms and other microphytobenthos, provides a large bulk of the energy supporting high biomasses of whole suites of invertebrates, fishes, and other vertebrates, including the marine iguana (Amblyrhynchus cristatus)—a unique sea-going lizard and active marine grazer that symbolizes the system's strong dependence on benthic primary production. Three highly abundant species of sea urchins (Tripneustes depressus, Eucidaris thouarsii, and Lytechinus semituberculatus) exert intense grazing pressure on benthic primary producers (and corals), often forming extensive urchin barrens (Breen and Mann, 1976; Ayling, 1981; Himmelman and Lavergne, 1985). Herbivorous fish species include damselfish, surgeonfish and parrotfish. Herbivorous green sea turtles (Chelonia mydas) are also present, sometimes in relatively high numbers. Several species of sea cucumbers (holothurians, Spanish pepino del mar), which are traditionally categorized as 'detritivores' but likely consume meio-fauna and flora, are also very widespread and abundant, notably the slow growing Stichopus fuscus, hereafter pepino. Omnivorous reef fishes, including chubs, butterflyfish, and damselfish, consume algae and small benthic invertebrates. Small benthic invertebrate-eating fishes include grunts, small wrasses, and an angelfish. Predatory invertebrates include whelks, conch, spiny lobsters, and crabs. Upper trophic levels feature many species of piscivorous reef fishes such as groupers and snappers. Large benthic invertebrates are consumed by large wrasses and triggerfish. Other high-level predators include octopus, the Galapagos sea lion (Zalophus wollebaeki), and unique seabirds such as the Galapagos penguin (Spheniscus mendiculus) and the flightless cormorant (Nannopterum harrisi). A variety of shark species is present including the Galapagos shark (Carcharhinus galapagensis), the white-tipped reefshark (Triaenodon obesus), other reef sharks, and species that interface with pelagic systems. 19 Small-scale fisheries feed tourists and residents, but most of the catches in the archipelago (high-value sea cucumbers and lobsters) are exported to lucrative foreign markets (e.g., Japan, Taiwan, U.S.). As an example, in 1999 and 2000, local fishers were paid about US$ 0.90 for every landed pepino {Stichopus fuscus), although the price fell to $0.55 in 2001 (PLMPP, Programa de Investigacion y Monitoreo Pesquero Participativo, 2001). The international fisheries markets drive the growing population of local fishers to deplete exploitable marine invertebrates (as well as illegally captured sharks), potentially shifting the structure of these ecosystems directly and indirectly, and undermining their sustainability (Constant, 1993; Camhi, 1995; Merlen, 1995). In addition, water quality is jeopardized around urban areas, and overflows from rudimentary septic tanks increase nutrient levels near the expanding population centers. Large and small spills occur occasionally when fuel is brought to the islands for delivery to tourist vessels, and visitors directly stress coastal biota. A Special Law of Galapagos was enacted in 1998 to improve marine reserve management and enforcement, but these stresses nevertheless continue. Finally, oceanographic and climatic changes, including the El Nino/La Nina oscillations and the potential for global climate change can profoundly influence the structure of Galapagos marine communities (e.g., Colinvaux, 1972; Houvenaghel, 1984; Glynn, 1988; Bost and Le Maho, 1993). A provisional coastal use-zoning plan has been in place in the archipelago since 2000 (Bustamante et al., 2002; see Figure 2-1). In this scheme, fully-protected 'no-take' areas, i.e., areas where no entry or human uses other than scientific research are allowed, protect 8% of the island's coastlines (zones shaded in black in Figure 2-1); non-extractive use areas, i.e., areas where tourism, recreation and education are allowed, cover 10% of coastlines (zones shaded in dark grey in Figure 1); regulated extractive uses, i.e., recreational and fisheries uses, are allowed along 77% of the coastlines (zones shaded in light grey in Figure 2-1); and special zones nearby the inhabited port areas, i.e., areas where the local stakeholders will define their status through a participatory process, cover the remaining 5% of the island's coastline (zones shaded in thickened light grey in Figure 1). This zoning plan provides an opportunity to protect small and replicated portions of Galapagos coastlines, and to evaluate the potential consequences and benefits of small and large marine protected areas (Branch et al., 2002). Although some violations of this scheme occur, fishing in the 'fisheries exclusion' zones (about one fifth of the total coastlines) is lower than in the zones open to fishing. The Charles Darwin Research Station has an ongoing marine ecological monitoring program designed to gather baseline information about these unique marine systems and to reveal any biological changes that might be related to the zoning and associated changes in human use patterns throughout the islands. The rocky reefs around Floreana Island are divided among fisheries, tourism, and fully protected zones. The present modelling exercise complements this monitoring program in a way that allows refinement of knowledge and management through an iterative approach to learning and an adaptive (or experimental) approach to conservation and fisheries management. The purpose of the model is to 20 provide accessible 'views' of the whole system and to predict how it might respond to changes in human actions or other stresses. The model may also provide insights into the underlying ecological mechanisms operating in the system and explore possible solutions to conservation problems. In particular, this continually updated ecological synthesis can be used to generate hypotheses about the dynamics of this special system and to address questions such as: • Which functional groups currently exert large effects on the system? • What are the potential ecosystem consequences of removing particular species from the system? • Are any species in this system currently being fished at unsustainable levels? • To what extent will fisheries exclusion zones alleviate declines of overfished species or restore previous abundances? METHODS Modelling tools: Ecopath with Ecosim Ecopath trophic models are mass-balance models, or more accurately mass-continuity models, that account for the energy flows in a food web. The Ecosim routine expresses the mass balance constraint in a dynamic context to explore the direct and indirect ecological effects of fisheries, perturbations, and even physical forces. For example, the relative strengths of trophic interactions among species can be estimated, and the effects of changes in a particular fishery on various biotic components can be simulated. These models are continually refined and evaluated in an iterative process. See Chapter 1 for a summary of the formulation and basic approach of Ecopath and Ecosim (also see Polovina, 1984; Christensen and Pauly, 1992; Walters et al., 1997; Walters et al., 1999; Christensen et al., 2000; Pauly et al., 2000; Walters et al., 2000; Oritz and Wolff, 2002; Froese and Pauly 2003 [http://www.fishbase.orgl; Christensen and Walters, 2004). In Ecospace, the simulated interactions among organisms occur in a spatially explicit, and habitat-based, context rather than in a single 'reaction vat' (Walters et al., 1999). Ecosim, because of simulated refugia from predation, also does not work as a simple vat. The area of interest is represented by a spatial mosaic of cells that can be designated as land and a variable number of marine habitats. For each functional group, these habitats are specified as preferred or not preferred. The dynamic redistribution of organisms in the system is based on user-specified base dispersal rates, relative movement rates in bad habitat, relative feeding rates in bad habitat, and the constantly changing densities of predator and prey groups with which a given functional group interacts. The instantaneous dispersal rates across cell boundaries are determined by the specified base dispersal rate, the habitat type in the source cell, and the responses of functional groups to predation risk and feeding conditions in the source cell (Walters et al., 1999). Spatial variations in primary production and current advection fields can be specified, as can spatially relative fishing costs (i.e., effort) and the spatial distribution of marine 21 protected areas (i.e., fisheries exclusion zones). Grid number and sizes can be adjusted or scaled appropriately. Delineating Floreana rocky reefs The region characterized in this balanced trophic model includes reefs shallower than 20-m depth along the eastern, northern and western coasts of the Floreana Island - a relatively homogeneous area that is also representative of the central Galapagos shelf region, which includes the large islands of Santiago, Santa Cruz, Santa Fe, San Cristobal and northern Espanola, and the eastern coast of Isabela—the largest of the islands (Figure 2-1). Because the southern coast of Floreana is much more exposed and influenced by cooler waters associated with the southern equatorial current and equatorial undercurrent, and has been little studied, that region was excluded from the model described here. Spatially, the exclusion represents an estimated 42% of the 20-m isobath (all strata) and 64% of the total rocky reef model area for the island. Reefs in water depths >20 m, and soft-sediment habitat types, were also excluded from the current model. Exclusion of soft-sediment habitats from this model might have unrealistically simplified the model because potentially important trophic interactions between soft bottoms and overlying or adjacent reefs would be ignored. A combination of site aerial photography and chart bathymetric data, geo-referenced within an Arc View GIS system, was used to estimate the spatial extent of the 20-m isobath from the coast. Spatial estimations were further weighted against a modifier for estimated habitat coverage (see Figure 2-2). A series of 67 observations of substrate composition was taken around the coastal perimeter of the island at approximately 500-m intervals. The spatial localization of each observation was taken as a polygonal area roughly equidistant between adjacent sampling points extending from the coast to the reported 20-m isobath. Rocky and bedrock strata estimations were grouped as representative of the model space and weighted by localized area to give a final estimation of the entire model space (Table 2-1). Table 2-1. Model space estimations within the 0-20 m isobath Area Total area (km2) Weighted habitat modifier (%) Corrected area (km2) 0-20 m isobath 28.38 62.2 17.65 Southerly exclusion 11.91 94.2 11.22 Modelled rocky reef 16.47 39.1 6.44 Defining functional groups The 43 functional groups in the Floreana Island rocky reef model were the product of a collaborative process that defined the system. A number of experts, including the present authors, participated in several iterations of the list of functional groups. All the species in the system were aggregated into these functional groups based on similarity of ecological role, defined by similarities in diet, production and 22 consumption rates, life history, and habitat associations, but also sometimes on value-driven criteria such as commercial status or importance for tourism. In the final iteration, benthic invertebrates were represented by 19 functional groups; others were fishes, 13; primary producers, 3; zooplankton, 2; marine mammals, 2; marine reptiles, 2; birds, 1; and detritus, 1. Because of the nature of the Galapagos archipelago, i.e., a relatively small and narrow rocky platform surrounded by deep open-ocean waters, the model includes discrete, but interconnected, benthic-based and pelagic-based subsystems. Estimating Ecopath input parameters Biomass estimates were derived using methods specific to each functional group. Production/biomass (P/B), consumption/biomass (Q/B), and diet compositions for each species were derived mostly from the scientific literature and with the help of FishBase (Froese and Pauly 2003 [http://www.fishbase.org1). P/B was usually estimated by assuming that it equals total mortality (Z) under the assumption of population equilibrium (Allen, 1971). Q/B was most commonly estimated from the empirical relationship proposed by Palomares and Pauly (1999), and setting mean water temperatures at 22-25°C. Representative values for aggregated groups were derived as averages of species-specific estimates weighted by relative biomass (B) or consumption (Q) as appropriate. Input biomass estimates of benthic and demersal fish groups and large invertebrates (> 20 mm) were obtained directly from site-specific surveys of Floreana rocky reefs. These middle trophic level groups represent the strong core of the model, and confidence in these estimates is expected to increase further as the ecological monitoring program progresses. Site-specific data were also available for pelagic fish groups, birds, turtles, sharks, and marine mammals, but resulting biomass estimates (or estimates of dietary proportions originating from the rocky reef) are more uncertain for these more mobile groups because of the haphazard nature of the existing sightings data, or knowledge of foraging patterns. Biomasses of macro-invertebrates (0.5 to 20 mm) and lower trophic level groups were estimated by the model, but empirically-based estimates of primary production of benthic macroalgae and phytoplankton were used to structure the base of the food web. Missing input parameters were taken from the literature (e.g., production/biomass and consumption/biomass ratios, and diet composition) and adjusted proportionally as weighted estimates for all species in a functional group whenever possible (e.g., most of the fish groups). Diet compositions were the least certain type of input parameter because of the paucity of site-specific dietary data. Examination of the knowledge gaps revealed during model construction enabled adaptive refinements to strategies for the continuing monitoring program. 23 Primary producers Mean phytoplankton standing stock was estimated using multispectral image analysis software to extract SeaWiFS chlorophyll concentration estimates from geo-referenced 1.1 km2-resolution localities over the eight target sites group averaged to the surrounding 10 km2 at each point, accounting for coastal overlap and cloud cover. Data were collated over the year 2000 as available from local area coverage from the NASA-PODAAC distributed data archive. Time series plots at target sites were constructed to examine seasonal and geographic variability within the model area, and averages compared against in situ samples collected during trips on 26 May 2001 and 18 June 2001. An averaged value of 12 tkm"2 was derived from an estimation of 0.64 mgChl m"3 following conversion factors for phytoplankton standing stock from Durbin and Durbin (1998), Arreguin-Sanchez et al. (1993), and Pauly et al. (1993a). Macroalgal biomass on the rocky reef at Floreana Island was estimated based on measured standing wet biomass at two sites on Santa Cruz Island, and based on subtidal observations. Microphytobenthos biomass was left to be estimated by the Ecopath routine. SeaWiFS data estimate only chlorophyll concentration within a few centimeters of the surface, and macroalgal estimates were very rough in terms of extrapolation to broader reef areas. Invertebrate groups Biomass estimates for 10 of the 15 mega-invertebrate (>20 mm) groups in the model were derived from visual line transect surveys at 9 Floreana rocky reef sites sampled during 2000 and again in 2001. Twenty-five species were aggregated into these 10 sampled functional groups. At each site, a 50-m transect line was laid down along two selected depth contours, and the number of large invertebrates within one meter of each side of the line recorded as the diver moved along one side and then back along the other side of the transect. The wet masses of individuals of most mega-invertebrate species from Floreana were also measured to estimate mean wet mass. In a few cases, the maximum length (e.g., arm radius for seastars, shell for gastropods, or body for sea cucumbers) or diameters (for sea urchins) of up to 30 individuals of each invertebrate species were recorded in situ using a measuring tape for later estimation of mass using length-weight relationships. Mean densities were multiplied by mean weights to obtain biomass density estimations, and skeletal carbonate weights were subtracted as appropriate. Fish groups Biomass estimates for the 13 fish groups were derived from visual line transect surveys at 9 rocky reef sites around the Floreana coastline during 2000 and 2001. At each site, divers swam at a constant speed on each side of 50-m transect lines placed along the 6- and 15-m isobaths, while recording the numbers and sizes of fish species observed within a 500 m2 area (10 m total swathe) and 5 m above the transect line. The density of each size class for each species was transformed into biomass/area, using length-weight relationships in FishBase (www.fishbase.org). Conversion factors for related proxy species were 24 used when no conversion factor was available for a species. The biomass/area values of the size classes were summed for a total species biomass and these were summed for total functional group biomass estimates. Higher vertebrates (sea lions, sharks, turtles, iguanas) Biomass estimates for sharks and turtles were the products of the number of diver sightings and average mass of individuals divided by the approximate area surveyed per dive. A sightings correction factor based on the discrepancy between diver-observed sea lion biomass and sea lion biomass based on counts on rookeries was then applied to the shark and turtle visual estimates to calculate rough error-corrected biomass estimates. A corrected biomass estimate for sea lions that feed on reefs was derived by multiplying the haul-out survey biomass estimate by the proportion of the diet from reefs. The biomass estimate for marine iguanas was based on surveys on Floreana Island, and is likely an underestimate for most of the north side of the island. However, a correction factor for such a density discrepancy is not yet developed, and the density presently used is likely more representative of the archipelago in general than it is of the north side of the island. Fisheries information Fisheries catch and effort data for the Galapagos Archipelago have been collected since 1997 through a daily monitoring program that includes recordings of the landed catch, effort, and distributions of finfish, sea cucumbers and lobsters. Data are collected at the three main ports in Galapagos at the islands Santa Cruz, San Cristobal, Isabela, and occasionally in Floreana. For the calculation of the average annual catch rate for these species we divided the catches that are extracted from the Floreana area by an estimate of the modelled area. Catches were calculated for the monitored species with use of conversion factors that describe the relation between the state of the product at landing (dried, salted, gutted, etc.) and its actual fresh weight. Pepino catch data in numbers of individuals were converted into fresh weight with the assumption that the average length is 21.3 cm in the population around Floreana (Anon., 2001). Given the length-weight relationship for pepino estimated in 1999 (n = 4363, R2 = 0.355) this length corresponds to a 333 g fresh weight. The fresh weights were summed per functional group and an average was calculated for the period under investigation. The average monitored finfish landings were multiplied with a raising factor to convert the monitored landings into an estimate of the total landings. This conversion factor (1.66) expresses the 34% effectiveness of the monitoring program and was derived from Espinoza et al. (2001). Even with this conversion factor, some illegal fishing removals might have been missed. No conversion factor was used for invertebrate catches because these are fully covered (100% monitored) through a system of export certificates. Landings per km2 were calculated with a fixed area of 16.47 km2 because fishing activities around the total coast of Floreana were taken into account. The resulting estimates of fishing density are applicable to the more limited area that was modelled. 25 Analyses The majority of the missing parameters left to be estimated by the Ecopath software were ecotrophic efficiency (EE) values, as empirically based estimates were available for most basic parameters in the system. However, Ecopath estimated 10 missing biomass values by specifying a reasonable EE value and solving the Ecopath master equation. One missing P/B value was estimated by additionally specifying a reasonable rate of production to consumption (P/Q). Trophic levels were calculated as the biomass weighted average of food items plus 1, and the omnivory index was the variance of the trophic levels of the prey groups (Pauly et al., 1993b). The basic flows in the system and other indices were also summarized while characterizing the system (see Christensen and Pauly, 1992; Christensen et al., 2000). Using Ecosim, a full series of 'removal' simulations was conducted to evaluate the relative interaction strength of each species in the Floreana rocky reef food web. Additional mortality was imposed on one pepino group so that it declined to zero by year 10 of the 30-year simulation. System-wide changes in biomasses resulting from the removal of a species were recorded. Mortality rates were then reset to initial levels before the next removal simulation. An interaction strength index (ISI), defined as the sum of all resulting relative changes in the system (the total absolute relative changes in all but the removed group), was used to derive a 'keystone' index, which is the ISI divided by the relative biomass of the respective affecting groups (see Power et al., 1996, for definition of keystone species). Fisheries were analyzed in terms of the proportion of the total catch in the system contributed by each functional group as well as the proportion of each group's total mortality accounted for by fisheries. An analysis of the directed pepino fishery was conducted because it stood out as unsustainable. Biomasses and catches were plotted as a function of capture rate (annual catch/biomass) as predicted by surplus production models in hypothetical equilibrium conditions. A simple Ecospace simulation was used to explore the potential effects of fishery exclusion zones on pepinos in the Galapagos Islands. The Floreana rocky reef model was re-expressed spatially using a mosaic of cells scaled to simulate Floreana Island, but the simulation was set up as a hypothetical island within the Galapagos archipelago. The area of rocky reef was exaggerated on the Ecospace base map for diagrammatic purposes (the reefs in the spatial simulation are made much wider than the actual narrow band around most Galapagos Islands; Figure 2-2). The hypothetical no-fishing zone covers approximately 23% of the coastline in the Ecospace base map reflecting the current proportion of protected coastline in the Galapagos Archipelago (18%) plus the special port areas (5%) in which local communities specify uses (Figure 2-5). This would tend to provide overly optimistic predictions of population responses to protection since local communities choose fishing. This hypothetical zone takes the form of a single no-fishing zone at one island. The simulation assumes a base dispersal rate of 5 km year"1 for the pepino group, corresponding to 14 m per day. Additional simulations were performed to account for higher dispersal rates of pepino larvae. 26 RESULTS The Floreana rocky reef food web model is characterized by very high biomasses of fishes and invertebrates (Table 2-2). The model is unique among Ecopath mass-balanced models in that the primary and secondary production needed to support such high biomasses are specified as a net 'immigration' of phytoplankton and zooplankton delivered to these reefs by oceanic currents, assuming that oceanic islands are plankton sinks. This results in a strong system heterotrophy, as indicated in the descriptive statistics (Table 2-3) and visible when examining a summary of system flows (Table 2-4). The diet composition matrix for the Floreana rocky reef model is presented in Appendix A. The 10 groups with the highest estimated trophic interaction strengths (Table 2-2; see also Chapter 4 for index formulation) were (in decreasing order): pelagic predators, large benthic invertebrate-eating fishes, shrimps and small crabs, omnivorous reef fishes, benthic algae, microphytobenthos, small benthic invertebrate eaters, other herbivorous fishes, noncommercial reef predators, and herbivorous zooplankton. Sea lions and sharks ranked 12th and 15th, respectively. The 10 groups with the highest indicated 'keystone index' values (Table 2-2) were (in decreasing order): toothed cetaceans, birds, sharks, sea lions, octopus, Hexaplex gastropods, spiny lobsters, noncommercial reef predators, pelagic predators, and large benthic invertebrate eating fishes. Omnivorous reef fishes ranked 11th. Specific results of the first of 43 functional group removal simulations are shown in Figure 2-3. Toothed cetaceans, sea lions, and noncommercial reef predators are predicted to increase when sharks are removed, thus causing decreases in bacalao, i.e., the grouper Mycteroperca olfax, other commercial reef fishes, and small benthic invertebrate-eating fishes through increased predation or competition, or both. Sea turtles, marine iguanas, large benthic invertebrate-eating fishes, and parrotfish are also predicted to increase when sharks are removed. Some small benthic invertebrates are predicted to increase, while large benthic invertebrates are predicted to decrease, and other trophic cascades are apparent. The model also shows the mean trophic level of the fisheries catch to be particularly low (2.3; see Table 2-3). Humans fill an unusually low trophic position in the Galapagos because, for example, pepinos comprised 71% of the fisheries catch from Floreana Island during the late 1990s and detritivorous fishes (Mugilidae) comprised 15% (Table 2-5). Pepinos declined in every simulation that included status quo fishing rates, because estimated overall mortalities from fisheries, predators, and senescence exceeded this group's estimated production for the entire range of input parameters reasonable for this species. The pepino fishery, as executed at 1999-2000 levels, accounted for 88% of the total mortality of this species. The current capture rate far exceeds the optimum sustainable capture rate estimated by Ecosim (Figure 2-4) indicating highly unsustainable fishing pressure. On the other hand, six to eight-fold increases in pepino biomass were predicted when total fishing moratoriums were simulated. 27 Table 2-2. Basic parameters of the Ecopath model of the Floreana rocky reef, Galapagos Group name Trophic level OI Biomass (tkm2) P/B (year1) Q/B (year1) EE ISI % of biomass Keystone Index Sharks 4.4 0.40 0.75 0.24 4.90 0.030 8.6 0.03 286.7 Toothed cetaceans 4.4 0.50 0.02 0.08 14.60 0.000 1.6 0.001 1600.0 Bacalao grouper 4.2 0.34 7.14 0.35 4.50 0.649 1.0 0.27 3.6 Birds 4.1 0.32 0.01 5.40 80.00 0.340 0.2 0.0004 575.0 Sea lions 4.0 1.27 5.68 0.07 25.55 0.864 9.4 0.22 42.8 Pelagic predators 3.9 1.19 30.00 0.42 4.35 0.282 22.4 1.14 19.6 Non-commercial reef predators 3.8 0.23 14.86 1.03 11.07 0.877 11.5 0.57 20.2 Octopus 3.5 0.10 0.79 1.10 7.30 0.511 0.9 0.03 30.0 Pelagic planktivores 3.4 0.15 5.50 0.98 32.10 0.353 1.6 0.21 7.6 Other commercial reef predators 3.3 0.16 9.30 0.62 7.11 0.557 1.7 0.35 4.9 Large benthic invertebrate eaters 3.3 0.06 32.71 0.65 9.82 0.658 18.7 1.25 15.0 Planktivorous reef fish 3.3 0.31 281.13 1.50 45.07 0.260 7.0 10.73 0.7 Hexaplex gastropod 3.0 0.02 3.61 2.80 14.00 0.667 3.7 0.14 26.4 Small benthic invertebrate eaters 3.0 0.31 100.99 1.39 13.73 0.569 13.4 3.85 3.5 Carnivorous zooplankton 2.8 0.52 3.58 8.70 29.00 0.475 8.8 n/ab n/a Spiny lobsters 2.8 0.26 3.00 0.45 7.40 0.650 2.6 0.11 23.6 Slipper lobster 2.7 0.39 4.00 0.45 7.40 0.722 0.7 0.15 4.7 Omnivorous reef fishes 2.7 0.29 41.52 1.02 21.85 0.896 17.7 1.58 11.2 Shrimps and small crabs 2.6 0.33 55.13 3.60 20.45 0.950 18.0 2.10 8.6 Asteroids 2.5 0.36 10.49 0.49 3.24 0.105 0.4 0.40 1.0 Other herbivorous fish 2.4 0.42 200.60 0.88 25.83 0.265 13.4 7.66 1.7 Eucidaris urchin 2.2 0.23 104.43 1.40 2.81 0.830 8.4 3.99 2.1 Anemones 2.2 0.26 79.24 2.00 4.00 0.900 3.1 3.02 1.0 Worms and ophiuroids 2.2 0.25 84.67 4.14 61.60 0.950 10.3 3.23 3.2 Stony corals 2.2 0.22 91.16 1.09 15.00 0.900 2.6 3.48 0.7 Chitons 2.2 0.29 2.85 0.34 11.70 0.900 0.1 0.11 0.9 Detritivorous fish 2.1 0.12 39.95 1.37 13.70 0.095 0.6 1.52 0.4 Small gastropods 2.1 0.11 188.05 2.50 14.00 0.950 6.4 7.18 0.9 Sea turtles 2.1 0.15 3.02 0.15 3.50 0.162 0.2 0.12 1.7 Pepino sea cucumber 2.1 0.07 3.90 0.60 3.36 0.972 0.0 b 0.15 0.0b Other urchins 2.0 0.01 4.65 1.40 2.81 0.755 0.1 0.18 0.6 Parrotfishes 2.0 0.00 21.50 0.50 16.60 0.627 1.7 0.82 2.1 Marine iguana 2.0 0.00 0.80 0.11 15.00 0.376 0.1 0.03 3.3 Other sea cucumbers 2.0 0.00 3.55 0.60 3.36 0.166 0.1 0.14 0.7 Tripneustes urchin 2.0 0.00 48.74 1.40 9.71 0.350 3.9 1.86 2.1 Lytechinus urchin 2.0 0.00 8.72 1.40 2.81 0.903 0.5 0.33 1.5 Small crustaceans 2.0 0.03 91.41 9.00 125.50 0.950 0.5 3.49 0.1 Filter + suspension feeders 2.0 0.08 367.39 2.00 16.50 0.900 9.0 14.02 0.6 Herbivorous zooplankton 2.0 0.08 3.19 17.30 57.70 0.656 10.9 n/aa n/a Phytoplankton 1.0 0.00 12.00 70.00 - 0.946 3.7 n/aa n/a Microphytobenthos 1.0 0.00 393.59 23.70 - 0.990 16.1 15.02 1.1 Benthic algae 1.0 0.00 256.80 12.00 - 0.986 16.5 9.80 1.7 Detritus 1.0 0.29 500 - - 0.499 n/a n/a n/a Notes: Values in bold have been calculated by the Ecopath software; other values are empirically based inputs, or values that were adjusted from empirically based values during balancing. The omnivory index (OI) indicates dietary breadth; ecotrophic efficiency (EE) is the proportion of production not consumed or exported; P/B and Q/B are the ratios of production and consumption to biomass; ISI is the trophic interaction strength index, which is the sum of the predicted relative biomass change (of all affected groups) after removal of the indicated affecting group at the beginning of 30-year dynamic simulations. The keystone index is the ratio of the interaction strength index and the percent of the system's overall biomass that is represented by the group. See Chapter 4 for equations describing both of these indices. a. Keystone indices were not estimated for the three plankton groups in the system because of high specified immigration rates. b. Pepino almost entirely disappear during 30-year simulations due to unsustainable catch rates, so 'removing them' is redundant. 28 Table 2-3. Basic flows and indices in the Floreana rocky reef Ecopath model Flows (t*km~2*year_l) Calculated total net primary production 13,250 Net system production -14,388 Sum of all production 17,337 Sum of all consumption 51,600 Sum of all exports -5,412 Sum of all respiratory flows 27,638 Sum of all flows into detritus 21,024 Total system throughput 94,850 Total catches 4.15 Biomass (t-krrT) Total living biomass 2,620 Indices Total primary production/total biomass 5.06 year"' Total biomass/total throughput 0.03 year"1 Total primary production/total respiration 0.48 Proportion of flows originating from detritus 0.62 Connectance index 0.16 Mean trophic level of the catch 2.27 System omnivory index 0.25 TL units Notes: Flows and biomass are expressed in wet weight. Minus signs indicate net imports of production The biomass of pepinos was predicted to increase inside a hypothetical no-fishing zone (Figure 2-5), but the overall biomass of pepinos is predicted to decline and stabilize at 36% of the 2000-2001 levels by the end of the 10 year simulation. Although the fisheries exclusion zone does not prevent overall pepino biomass from declining, it does prevent these intense fisheries from completely eliminating this slow-growing species. Dispersal rates higher than 14 m day"1 result in a larger 'spillover' effect (catchable emigration), but a lower buildup (protection) of biomass in the no-fishing zone. 29 Table 2-4. Flows from primary production and detritus From primary production From detritus TL Consumed Export To detritus Respiration Throughput Consumed Export To detritus Respiration Throughput VI 0 0 0 1 1 0 0 0 1 1 V 1 0 5 15 20 1 0 4 12 16 IV 16 0 224 381 621 13 0 210 337 561 III 1099 -927 2125 3237 5533 1048 -927 1050 1684 2855 II 10276 -8843 9956 13058 24448 5086 -3648 7195 8912 17544 I 14668 -1600 155 0 13223 10490 10534 0 0 21124 Sum 26059 -11370 12465 16692 43846 16639 5959 8459 10946 42102 Note: Flows are expressed in (t-km"2-year"'). System imports and exports are not shown. Some flows reach trophic level VI because some organisms within some functional groups are supported by energy that has traversed five links from primary producers. Table 2-5. Percentage of total annual catch comprised by the 10 functional groups targeted in Floreana reef fisheries Functional group TL Catch (t-km^year1) % of total catch % of total mortality Bacalao grouper 4.2 0.031 0.8 . 1.4 Pelagic predators 3.9 0.221 5.4 1.9 Non-commercial reef predators 3.8 0.067 1.6 0.5 Pelagic planktivores 3.4 0.004 0.1 0.1 Other commercial reef predators 3.3 0.037 0.9 0.6 Large benthic invertebrate eaters 3.3 0.006 0.1 0.9 Spiny lobster 2.8 0.178 4.3 13.1 Slipper lobster 2.7 0.011 0.3 0.7 Detritivorous fishes 2.1 0.621 15.2 1.2 Pepino (S. fuscus sea cucumber) 2.1 2.922 71.3 87.5 Notes: Analysis based on data from 1997 to 2000. A great majority of the catch comprised pepinos and detritivorous fishes. TL is trophic level, and the final column is the percent of each group's total mortality that is directly caused by fisheries. 30 Biomass (End / Start) 1 2 Sharks Toothed cetaceans Bacalao Birds Sea lions Pelagic predators Non-commercial reef predators Octopods Pelagic planktivores Other commercial reef predators Large benthic invertebrate eaters Planktivorous reef fish Hexaplex gastropod Small benthic invertebrate eaters Carnivorous zooplankton Spiny lobsters Slipper lobster Omnivorous reef fishes Shrimps and small crabs Asteroids Other herbivorous fish Eucidaris urchin Anemones Worms and ophioroids Stony corals Chitons Detritivorous fish Small gastropods Sea turtles Pepino sea cucumber Other urchins Parrotfishes Marine iguana Other sea cucumbers Tripneustes urchin Lytechinus urchin Small crustaceans Filter + suspension feeders Herbivorous zooplankton Phytoplankton Mlcrophytobenthos Benthic algae Detritus Figure 2-3. Predicted changes resulting from the complete removal of sharks from the present day Floreana rocky reef trophic model. Results shown are the predicted relative change in biomasses at the end of a 30-year simulation in which sharks were removed by year 10 (V = 0.4). 31 25 SZ catc 20 ra 15 • E •D ra 5i 10 -M in Biom a 5 0 \ Pepino {Stichopus fuscus) * Biomass 1999-2000 capture rate Capture rate (annual catch / biomass) Figure 2-4. Predicted catch and biomass curves for pepino sea cucumbers (Stichopus fuscus) on the rocky reefs at Floreana Island, Galapagos. The 1999-2000 capture rate (annual catch/biomass) of this species was essentially twice that of sustainable levels. This figure represents the predicted states of the biomass and annual catch after the system reaches 'equilibrium' based on the specified biomass and production rate (P/B) of pepino and the combined effect of all sources of mortality in the system. It is possible for the actual capture rate to greatly exceed sustainable capture rates, but only if the population is rapidly collapsing. Figure 2-5. A simple diagrammatic representation of the potential effects of a fisheries exclusion zone on pepino (S. fuscus) biomass at the end of a 10-year Ecospace simulation at a hypothetical Galapagos island. Darker areas represent high biomasses and lighter areas represent low biomasses. Catchable emigration of pepinos can be seen as dark shading outside the dotted lines that demarcate the boundaries of the hypothetical fisheries exclusion zone. Pepinos still decline to a biomass lower than present, but the no-fishing zone prevents the intense fishery from extirpating them. 32 DISCUSSION It seems obvious after construction of the Floreana rocky reef model that the remarkably high biomasses of fishes, invertebrates, and other organisms on Galapagos rocky reefs is made possible not only by high in situ production of macroalgae and microphytobenthos, but also because these reefs are sinks for oceanic plankton. Stated another way, it is imported food (energy) that allows the high biomass observed on Galapagos reefs. The reefs must capture the primary and secondary production of large oceanic areas as currents continually flow past and around the islands. These large quantities of plankton are captured by high biomasses of filter- and suspension-feeding invertebrates and planktivorous reef fishes creating unusually rapid turnover of diversity and biomass, particularly in areas of continual upwelling (Witman and Smith, 2003). This might seem odd to terrestrial ecologists who are used to communities that are almost entirely supported by in situ primary production, but there are now many examples of marine or coastal communities that are strongly shaped by allochthonous subsidies (Okey, 1993; Vetter, 1994; Bustamante et al., 1995; Vetter, 1995; Polis and Hurd, 1996; Okey, 1997; Polis et al., 1997; Vetter and Dayton, 1999; Okey, 2003). The importation of large quantities of carbon to rocky reefs through this planktonic-benthic linkage is discussed by Bray et al. (1981) and Bray (1981). The necessity for such importation (to support the existing high biomasses on reefs) is, however, made clear only through construction of mass-balance trophic models. One implication of this import is that secondary production and tertiary production are strongly coupled and magnified by oceanographic conditions, as discussed by Menge et al. (1997). Although trophic connections, linkages, and cascades in nearshore rocky subtidal systems can be dampened by physical oceanographic forces (e.g., Kvitek et al., 1998), there is some evidence that biological oceanographic conditions (food inputs) can strengthen trophic connections along rocky shorelines (Polis and Hurd, 1996; Menge et al., 1997; also see Oksanen et al., 1981), just as kelp subsidies can increase competition and secondary production in the rocky intertidal zone (e.g., Bustamante et al., 1995). The changes predicted by the shark removal simulation (Figure 2-3) have presumably already taken place to a much larger extent than the present day simulation predicts. Sharks might be considerably reduced over Galapagos reefs due to unaccounted shark fisheries since the 1950s and which continue illegally today (Constant 1993, Camhi 1995). Indeed, the commercial reef predator groups (including groupers) make up only 1.8% of the present day Galapagos fisheries catch (present analysis), whereas these fishes were the main target in the past (Reck, 1984; Ruttenberg, 2001). The implication is that recovery of sharks could lead to increases in other reef predators by decreasing the biomass of their respective predators. Several functional groups in the system are likely to have lower than normal interaction strength than in the present day system because their biomasses or size distributions (or diets) have been 33 considerably reduced or modified. These now depleted functional groups with reduced biomass potentially include sharks, sea lions (Z. wollebaeki), birds, 'bacalao' grouper (Mycteroperca olfax), large benthic invertebrate eating fishes (e.g., Bodianus diplotaenia, Semicossyphys darwini), pepinos, spiny lobster (Panulirus gracilis and P. penicillatum), slipper lobster {Scyllarides astori) and stony and black corals. The situation of formerly important commercial large groupers (i.e., Epinephelus mystacinus and Epinephelus cifuentesi; Reck, 1986) is unclear, as they are confined to deeper waters, and their fishery has not been given much attention during the last decade. Such reductions of species and functional groups can severely modify marine ecosystems (Dayton et al., 1995; Dayton et al., 1998), especially because many of the groups removed are from upper trophic levels and have high 'keystone' values (Table 2-2). Local fisheries are now supported by lower trophic-level species instead of the upper trophic-level species that were preferred in the past (Table 2-5). Still other species may have become more abundant in response to reductions in biomasses of predators that structure the system. These species with net gain might include planktivorous reef fish (i.e., gringo, P. colonus), sea urchins (e.g., Eucidaris thouarsii, Tripneustes depressus, Lytechinus semituberculatus) and in some cases anemones (i.e., Aiptasia sp.). The central Galapagos rocky reefs appear to be a local example of the global pattern of 'fishing down marine food webs' (Pauly et al., 1998). Even if increases in pepino fishing pressure in the Galapagos is driven more by global increases in demand than local depletion of fishes like bacalao (M. olfax), the lucrative financial incentives for catching pepinos are arguably driven by global changes in coastal species composition (i.e., fishing down the food web) that have lead to increasing markets for holothurians and other low trophic level organisms. The fishing down effect is however reinforced by the fact that the bacalao grouper, a top predator, was in the past the main target for the local and mainland salt-dried market (Reck, 1984); today, two species of planktivorous mullets (Mugil spp.) dominate the salt-dried landings (Espinoza et al., 2001). Many of the changes might be exacerbated indirectly through trophic cascades. For example, Ayling (1981) suggested that the removal of large benthic invertebrate-eating fishes might have led to increases in sea urchin biomass in New Zealand. In the Galapagos, any such increase in densities of the urchin Eucidaris thouarsii could have contributed to the decline of stony corals and caused other changes in this benthic rocky reef system (Glynn et al., 1979). Wellington (1975) noted that a conspicuous urchin predator, the Mexican hogfish Bodianus diplotaenia, has declined locally. Recent data have suggested an increase in urchins and herbivorous fish resulting from the removal of such large predatory fishes during the 1970s (Ruttenberg, 2001). These ecological cascades are indicated even in the present day shark removal simulation shown in Figure 2-3. Indeed, modifications to the Galapagos marine ecosystem have shaped a present-day marine system that is probably more removed from its pristine state than we tend to think. This puts modelling exercises at a disadvantage when the working model is based on the present-day system, because organisms that might have played a strong structuring role in the past might now have only negligible 34 effects on the system. In the context of modified ecosystems, therefore, the only fair way to evaluate the potential role of organisms using whole trophic modeling is to construct a past system model (sensu Pitcher and Pauly, 1998; Pitcher, 2001; Pitcher et al., in press). This can be accomplished in relatively short order by using the present day Floreana rocky reef model as a template, but only if good information about the chosen past system is available. Luckily, some information is available on past abundances of some of the organisms in question throughout the Galapagos Archipelago. Heavy grazing by sea urchins is known to be the immediate cause of extensive 'barren grounds' where the bottom is dominated by crustose coralline algae and high abundances of urchins (Mann and Breen, 1972). Such an ecological phenomenon resembles a shift to an alternate stability domain (sensu Scheffer et al. 2001), though it is perhaps more accurately described as a shift to a non-trivial boundary point (Sutherland 1974) where one or more species in the system (i.e., predators of sea urchins) are removed. Regardless of how they are categorized, such shifts generally have negative implications for diversity and ecosystem integrity. Similar shifts that are linked to removal of urchin predators are documented worldwide (Estes and Palmisano, 1974; Elner, 1990; Levitan, 1992; Estes and Duggins, 1995; McClanahan et al., 1996; Sala, 1998; Sala et al., 1998). These barren grounds are now a common feature of the seascape of Galapagos reefs (Glynn et al., 1979; Ruttenberg, 2001). In some areas, 'anemone barrens' have begun to appear, in which a single species of anemone (Aiptasia sp.) has replaced previously diverse shallow reef habitats (Chapter 3; Okey et al., 2003). Questions surrounding the genesis of this Galapagos seascape can be explored using a model that features more pristine levels of urchin predators; i.e., lobsters and groupers. For example, why is the bacalao grouper indicated to have such low interaction strengths and such a low keystone index in the present day model, particularly when groupers are thought to be strong shapers of reef ecosystems (Hixon and Beets, 1993)? One hypothesis is that their biomass has been considerably reduced. Another is that their size class distribution shifted to smaller individuals and they simply do not interact like the big grouper predators they once were. Both trends might be true (Ruttenberg, 2001), but explicit specification of past information (e.g., from Reck, 1984) could provide insights into this group's past role in shaping the system relative to their current role. It is tempting to suggest that large bacalao groupers are size-overfished (Coello and Grimm, 1993) and functionally negligible in the Galapagos Archipelago, but more information is needed to evaluate that question. A 'past system' model will allow assessments of the roles of such strong interacting species, but moreover, it can be used to postulate and explore the trophic cascades and other mechanisms that changed a pristine system to a degraded system. Moreover, this approach can be used to provide potential 'roadmaps' to restoration (Pitcher, 2001), as well as helping to guide the continuing ecological monitoring of the Galapagos Marine Reserve. Unavoidable uncertainty in the predictions of pepino equilibrium catch and biomass in relation to capture rate is a function of the paucity of information on stock-recruitment relationships for pepinos, 35 which is probably nonlinear. Theoretically, stock recruitment relationships are implicit in the specified pepino production rate (P/B) of 0.6 year"', which is based on information in Pauly et al. (1993a). These authors assumed that total mortality (Z; and thus P/B) of holothurians is approximately twice that of natural mortality (M), like fishes targeted by a fishery. Opitz (1986) used a P/B of 0.29 year"1 as equivalent to the natural mortality of unfished Caribbean holothurians. One option for pursuing shortcomings in this analysis would be to specify split, but linked, pools for different life stages of pepinos. This can be done using Ecopath with Ecosim when enough information becomes available on the early life stages of this species. The simulation of the effects of a fishery exclusion zone on pepino is a simplistic representation of the real dynamics of the system. For example, the exaggeration of the width of the fringing reefs was necessary due to the resolution constraints of the Ecospace grid (e.g., a proportionally narrow band of reef around Galapagos islands could not be simulated with the current version of Ecospace). In the context of the spatial characteristics of these fringing reefs, there is considerable uncertainty with respect to dispersal and effort response effects, as implied above (also see Mangel, 2000). We expect that the resulting exaggerated area of the reef would overestimate the beneficial effects of the fishery exclusion zone (Figure 2-5), if anything. Size does matter when it comes to protected areas, pepinos or no pepinos (Walters et al., 1999; Martel et al. 2000; but see Halpern, 2003). Nevertheless, since overall distances across the cell matrix are scaled properly (with reference to Floreana Island), the simulation has at least made it clear that protecting small portions of reef areas (23%) is likely inadequate to prevent further overall declines of pepino biomass in the Galapagos Islands without considerable reductions in pepino capture rates. At the same time, the simulation indicated that pepino biomass increased in the hypothetical fishery exclusion zone, thus preventing extirpation of pepinos. This indicates a positive effect of the exclusion zone on the fishery —related to emigration (Walters et al., 1999) and consistent with empirical findings of Roberts et al. (2001)—despite the prediction of an overall catch decline even when implementing fishery exclusion zones. In spite of its simplicity and inherent uncertainties, these simulation results are remarkably similar to a recent empirical evaluation of the effects of a fishery and marine reserves on a closely related sea cucumber (Parastichopusparvimensis) in California's Channel Islands (Schroeter et al., 2001). This Ecospace simulation is also fully consistent with the conclusions of Allison et al. (1998) that marine reserves are necessary but not sufficient for marine conservation, and especially their conclusion that well-intentioned networks of marine reserves must be complemented with strong conservation efforts in the areas outside the reserves (also see Murray et al., 1999). A fisheries free-for-all justified by the establishment of a network of reserves is reasonable to expect, and would quite likely prevent achievement of conservation goals. Finally, we must stress that the present model does not represent some major areas of the archipelago. For example, the biotic communities of the central Galapagos shelf are markedly different 36 from the communities of the more northern Wolf and Darwin Islands and the western side of Isabela and Fernandina Islands (see Figure 2-1). Evidence is mounting that these latter two areas comprise discrete biogeographic zones, though separated by very short geographical distances (Abbott, 1966; Harris, 1969; Glynn and Wellington, 1984; Reck, 1986; Jennings et al., 1994; Bustamante et al., 2000; Wellington et al., 2001; Bustamante et al., 2002). This situation gives rise to several new questions: Should these differences be integrated into one model or should models of each biogeographic zone be constructed? What are the roles oceanographic forces in shaping these biotic communities relative to trophic forces? Such questions will inform the development of the present ecosystem research; the results presented here constitute a first step to explore and understand the nature and dynamics of the broader Galapagos marine ecosystems. My continuing strategy is to evaluate and refine the Floreana Island rocky reef ecosystem model iteratively in parallel with the ongoing baseline monitoring program. Only by combining such analytical approaches with ongoing empirical field investigations can the usefulness of ecological models be truly evaluated. This adaptive approach will help evaluate the potential effects of human activities and management policies such as the effectiveness of zone-based fisheries and conservation management in the Galapagos Marine Reserve. Examination of the knowledge gaps revealed during model construction has already enabled adaptive refinements to strategies for the continuing monitoring program in the sense that the focus of this program is shifting to less certain aspects of the Galapagos subtidal rocky reef. LITERATURE CITED Abbott, D.P., 1966. Factors influencing the zoogeographic affinities of the Galpagos inshore marine fauna. In R. Bowman (Editor), The Galapagos. Proceedings of the Symposia of the Galapagos International Scientific Project. Univ. Calif. Press, Los Angeles, pp. 108-122. Allen, R.R., 1971. Relation between production and biomass. J. Fish. Res. Board Can. 28, 1573-1581. Allison, G.W., Lubchenco, J., Carr, M.H., 1998. Marine reserves are necessary but not sufficient for marine conservation. Ecol. Appl. 8 (1), S79-S92. Anon., 2001. Analisis de la densidad poblacional y estructura de tallas del pepino de mar {Stichopus fuscus) en el archipielago de Galapagos durante Abril 2001, Informe preparado para la Autoridad Interinstitucional de Manejo. Sector Pesquero Artesanal de Galapagos, Servicio del Parque Nacional Galapagos, Estacion Cientifica Charles Darwin e Instituto Nacional de Pesca, 5 anexos. 19 p. Arreguin-Sanchez, F., Seijo, J.C., Valero-Pacheco, E., 1993. An application of ECOPATHII to the north continental shelf ecosystem of Yucatan, Mexico, In V. Christensen, D. Pauly (Editors), Trophic Models of Aquatic Ecosystems. ICLARM Conf. Proc. 26, pp. 269-278. Ayling, A.M., 1981. The role of biological disturbance in temperate subtidal encrusting communities. Ecology 62 (3), 830-847. Banks, S.A., 1999. The use of remotely sensed AVHRR data in determining SST variability and zonation across the Galapagos Marine Reserve. M.Sc. thesis in Oceanography at Southampton Oceanography Centre. 46 pp. Bost, C.A., Le Maho, Y., 1993. Seabirds as bio-indicators of changing marine ecosystems: new perspectives. Acta Oecol. 14 (3), 463-470. Branch, G.M., Witman, J., Bensted-Smith, R., Bustamante, R.H., Wellington, G.M., Smith, F., Edgar, G., 2002. Conservation Criteria for the Marine Biome. In E. Dinnerstein (Editor), A Biodiversity 37 Vision for the Galapagos Islands: An Exercise for Ecoregional Planning. WWF, Washington DC, USA. Bray, R.N., 1981. Influence of water currents and zooplankton densities on daily foraging movements of blacksmith, Chromis punctipinnis, a planktivorous reef fish. Fish. Bull. 78 (4), 829-841. Bray, R.N., Miller, A.C., Geesey, G.G., 1981. The fish connection: a trophic link between planktonic and rocky reef communities? Science 214 (4517), 204-205. Breen, P. A., Mann, K.H., 1976. Destructive grazing of kelp by sea urchins in eastern Canada. J. Fish. Res. Board Can. 33, 1278-1283. Bustamante, R., Collins, K.J., Bensted-Smith, R., 2000. Biodiversity conservation in the Galapagos Marine Reserve. Bull. Inst. R. Sci. Natl. Belg. 70 (Biologie Supplement), 31-38. Bustamante, R.H., Branch, G.M., Bensted-Smith, R., Edgar, G.J., 2002. The status of and threats to marine biodiversity. In E. Dinnerstein (Editor), A Biodiversity Vision for the Galapagos Islands: An Exercise for Ecoregional Planning. WWF, Washington DC, USA. Bustamante, R.H., Branch, G.M., Eekhout, S., 1995. Maintenance of an exceptional intertidal grazer biomass in South Africa: subsidy by subtidal kelps. Ecology 76, 2314-2329. Camhi, M., 1995. Industrial fisheries threaten ecological integrity of the Galapagos Islands. Conserv. Biol. 9(4), 715-719. Chavez, F.P., Brusca, R.C., 1991. The Galapagos Islands and their relation to oceanographic processes in the tropical Pacific. In MJ. James (Editor), Galapagos Marine Invertebrates, Plenum Press, N.Y., pp. 9-33. Christensen, V., Walters, C.J., 2004. Ecopath with Ecosim: methods, capabilities and limitations. Ecol. Model 172, 109-139. Christensen, V., Pauly, D., 1992. ECOPATH II - A system for balancing steady-state ecosystem models and calculating network characteristics. Ecol. Model. 61, 169-185. Christensen, V., Walters, C.J., Pauly, D., 2000. Ecopath with Ecosim - A User's Guide. Univ. of British Columbia, Fisheries Centre, Vancouver, Canada and ICLARM, Penang, Malaysia, 131 p. Christie, D.M., Duncan, R.A., McBirney, A.R., Richards, M.A., White, W.M., Harpp, K.S., Fox, C.G., 1992. Drowned islands downstream from the Galapagos hot spot imply extended speciation times. Nature 355, 246-248. Coello, S., Grimm, A.S., 1993. The reproductive biology of Mycteroperca olfax (Jenyns) (Pisces Serranidae): Protoginy and breeding season. Rev. Cien. Mar. Limn. 3, 115-128. Colinvaux, P.A., 1972. Climate and the Galapagos Islands. Nature 240, 17-20. Constant, P. 1993. New pirates of the Galapagos. Oceanorama. Inst. Oceanographique Paul Richard 21, 9-12. Dayton, P.K., Tegner, M.J., Edwards, P.B., Riser, K.L., 1998. Sliding baselines, ghosts, and reduced expectations in kelp forest communities. Ecol. Appl. 8 (2), 309-322. Dayton, P.K., Thrush, S.F., Agardy, M.T., R.J. Hoffman, 1995. Environmental effects of fishing. Aquat. Conserv. 5, 205-232. Durbin, A.G., Durbin, E.G., 1998. Effects of menhaden predation on plankton populations in Narragansett Bay, Rhode Island. Estuaries 21 (3), 449-465. Elner, R.W., 1990. Inference in Ecology: the sea urchin phenomenon in the northwestern Atlantic. American Naturalist 136, 108-125. Espinoza, E., Murillo, J.C., Toral, M.V., Bustamante, R.H., Nicolaides, F., Edgar, G.J., Moreno, J., Chasiluisa, C, Yepez, M., Barreno, J.C., Shepherd, S.A., Viscaino, J., Villalta, M., Andrade, R., Born, A.F., Figueroa, L., Guerrero, P., Piu, M., 2001. La pesca en Galapagos: Comparacion de las capturas entre 1997-2000. In Informe Galapagos 2000/2001, Fundacion Natura and WWF, Quito, Ecuador, pp. 55-64. Estes, J. A.-, Duggins, D.O., 1995. Sea otters and kelp forests in Alaska: generality and variation in a community ecological paradigm. Ecol. Monogr. 65(1), 75-100. Estes, J.A., Palmisano, J.F., 1974. Sea otters: their role in structuring nearshore communities. Science 185, 1058-1060. 38 Feldman, G.C., 1985. Satellites, seabirds and seals. In G. Robinson, E.M. del Pino (Editors), El Nino in the Galapagos Islands: The 1982-1983 Event. Quito, Ecuador: Charles Darwin Foundation, pp. 125-130. Feldman, G.C., 1986. Patterns of phytoplankton production around the Galapagos Islands. In M. Bowman, C. Yentsch, W. Peterson (Editors), Tidal Mixing and Plankton Dynamics, Lecture Notes on Coastal and Estuarine Studies, Vol. 17, Germany: Springer-Verlag, pp. 77-106. Froese, R. and D. Pauly. Editors. 2003. FishBase. World Wide Web electronic publication. www.fishbase.org. Glynn, P.W., 1988. El Nino-Southern Oscillation 1982-1983: Nearshore population, community, and ecosystem responses. Ann. Rev. Ecolog. Syst. 19, 309-345. Glynn, P.W., Wellington, G.M., 1984. Corals and Coral Reefs of the Galapagos Islands. Berkeley. University of California Press, 319 pp. Glynn, P.W., Wellington, G.M., Birkeland, C, 1979. Coral reefs growth in the Galapagos: limitations by sea urchin. Science 203, 47-49. Halpern, B.S., 2003. The impact of marine reserves: do reserves work and does reserve size matter? Ecol. Appl. 13(1), S117-S137. Harris, M.P., 1969. Breeding season of seabirds in the Galapagos Islands. J. Zool. (Lond) 159, 145-165. Himmelman, J.H., Lavergne, Y., 1985. Organization of rocky subtidal communities in the St. Lawrence estuary. Natur. Can. 112, 143-154. Hixon, M.A., Beets, J.P., 1993. Predation, prey refuges, and the structure of coral-reef fish assemblages. Ecol. Monogr. 63,(1)77-101. Houvenaghel, G. T., 1984. Oceanographic setting of the Galapagos Islands. In R. Perry (Editor), Key Environments: Galapagos. Pergamon Press, Oxford, pp.43-54. James, M.J. (Editor), 1991. Galapagos Marine Invertebrates - Taxonomy, Biogeography and Evolution in Darwin's Islands. Plenum Press, New York, 488 pp. Jennings, S., Brierley, A.S., Walker, J.W., 1994. The inshore fish assemblages of the Galapagos Archipelago. Biol. Conserv. 70, 49-57. Kvitek, R.G., Iampietro, P.J., Bowlby, C.E., 1998. Sea otters and benthic prey communities: a direct test of the sea otter as keystone predator in Washington state. Mar. Mammal Sci. 14 (4), 895-902. Levitan, D.R., 1992. Community structure in times past: influence of human fishing pressure on algal-urchin interactions. Ecology 73, 1597-1605. Mangel, M., 2000. Irreducible uncertainties, sustainable fisheries and marine reserves. Evol. Ecol. Res. 2 (4), 547-557. Mann, K.H., Breen, P.A., 1972. The relation between lobsters abundance, sea urchins and kelp beds. J. Fish. Res. Board Can. 29, 603-605. Martell, S.J.D., Walters, C.J., Wallace, S.S., 2000. The use of marine protected areas for conservation of lingcod {Ophiodon elongatus). Bulletin of Marine Science 66 (3), 729-743. McClanahan, T.R., Kamukuro, A.T., Muthiga, N., Gilagabher Yebio, M. and Obura, D., 1996. Effect of sea urchin reductions on algae, coral and fish populations. Conserv. Biol. 10, 136-154. McCosker, J.E., Rosenblatt, R.H., 1984. The inshore fish fauna of the Galapagos Islands. In R. Perry (Editor), Key Environments: Galapagos, Pergamon Press, Oxford, pp. 133-144. Menge, B.A., Daley, B.A., Wheeler, P.A., Dahlhoff, E., Sanford, E., Strub, P.T., 1997. Benthic-pelagic links and rocky intertidal communities: bottom-up effects or top-down control? Proc. Natl. Acad. Sci. USA. 94 (26), 14530-14535. Merlen, G., 1995. Use and misuse of the seas around the Galapagos Archipelago. Oryx, 29, 99-106. Murray, S.N., Ambrose, R.F., Bohnsack, J.A., Botsford, L.W., Carr, M.H., Davis, G.E., Dayton, P.K., Gotshall, D., Gunderson, D.R., Hixon, M.A., Lubchenco, J., Mangel, M., MacCall, A., McArdle, D.A., Ogden, J.C., Roughgarden, J., Starr, R.M., Tegner, M.J., Yoklavich, M.M., 1999. No-take reserve networks: sustaining fishery populations and marine ecosystems. Fisheries 24 (11), 11-25. Okey, T.A. 1993. Natural disturbances and benthic communities in Monterey Canyon Head. M.S. Moss Landing Marine Laboratories and San Jose State University, Moss Landing and San Jose. Okey, T.A. 1997. Sediment flushing observations, earthquake slumping, and benthic community changes in Monterey Canyon head. Continental Shelf Research 17, 877-897. 39 Okey, T.A. 2003. Macrobenthic colonist guilds and renegades in Monterey Canyon (USA) drift algae: Partitioning multidimensions. Ecological Monographs 73, 415-440. Okey, T.A., Shepherd, S.A., Martinez, PC, 2003. A new record of anemone barrens in the Galapagos. Noticias de Galapagos 62, 17-20. Oksanen, L., Fretwell, S., Arruda, J., Niemela, P., 1981. Exploitation ecosystems in gradients of primary productivity. Am. Nat. 118, 240-261. Opitz, S., 1986. Trophic interactions in Carribean coral reefs. ICLARM Tech. Rep. 43, 341 pp. Oritz, M., Wolff, M., 2002. dynamical simulation of mass-balance trophic models of benthic communities of north-central Chile: assessment of resilience time under alternative management scenarios. Ecol. Model. 148,227-291. Palomares, M.L.D., Pauly, D., 1999. Predicting the food consumption of fish populations as functions of mortality, food type, morphometries, temperature and salinity. Mar. Freshwater Res. 49, 447-453. Pauly, D., Christensen V., Walters C, 2000. Ecopath, Ecosim, and Ecospace as tools for evaluating ecosystem impacts of fisheries. ICES J. Mar. Sci. 57, 697-706. Pauly, D., Christensen, V., Dalsgaard, J., Froese, R., Torres Jr., F., 1998. Fishing down marine food webs. Science 279 (5352), 860-863. Pauly, D., Sambilay Jr., V., Opitz, S. 1993a. Estimates of relative food consumption by fish and invertebrate populations, required for modeling the Bolinao Reef ecosystem, Philippines. In V. Christensen, D. Pauly (Editors), Trophic Models of Aquatic Ecosystems. ICLARM Conf. Proc, 26, pp.236-251. Pauly, D., Soriano-Bartz, M., Palomares, M.L., 1993b. Improved construction, parameterization and interpretation of steady-state ecosystem mdoels. In V. Christensen, D. Pauly (Editors), Trophic Models of Aquatic Ecosystems. ICLARM Conf. Proc. 26, pp. 1-13. PFMPP, 2001. Informe tecnico final de la pesqueria del pepino de mar (Stichopus fuscus) en las Islas Galapagos 2001 (Analisis comparativo con las pesquerias de 1999 y 2000). Programa de Investigacion y Monitoreo Pesquero Participativo, Estacion Cientifica Charles Darwin, Puerto Ayora, Galapagos, 25 pp. Pitcher, T.J., 2001. Fisheries managed to rebuild ecosystems? Reconstructing the past to salvage the future. Ecol. Appl. 11 (2), 601-617. Pitcher, T.J., Heymans, S.J.J., Ainsworth, C, Buchary, E.A., Sumaila, U.R., Christensen, V., 2003. Opening the lost valley: implementing a Tjack to future' restoration policy for marine ecosystems and their fisheries. In: Knudsen, E.E., MacDonald, D.D., Muirhead, J.K. (Eds.), Fish in the Future? Perspectives on Fisheries Sustainability. American Fisheries Society, Bethesda, MD, USA., in press. Pitcher, T.J., Pauly, D., 1998. Rebuilding ecosystems, not sustainability, as the proper goal of fishery management. In: T.J. Pitcher, P.J.B. Hart and D. Pauly (Editors), Reinventing Fisheries Management. Kluwer Academic Publishers, London., pp. 311-329. Polis, G. A., Hurd, S. D. 1996. Allochthonous input across habitats, subsidized consumers, and apparent trophic cascades: Examples from the ocean-land interface. Pages 275-285 in G. A. Polis and K. O. Winemiller, editors. Food webs: Integration of patterns and dynamics. Chapman and Hall, New York. 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. Polovina, J.J., 1984. Model of a coral reef ecosystem I. The ECOPATH model and its applications to French Frigate Shoals. Coral Reefs 3, 1-11. Power, M. E., Tilman, D., Estes, J.A., Menge, B.A., Bond, W.J., Mills, L.S., Daily, G., Castilla, J.C., Lubchenco, J., Paine, R.T., 1996. Challenges in the quest for keystones. Bioscience 46 (8), 609-620. Reck, G. 1984. The coastal fisheries in the Galapagos Islands, Ecuador: description and consequences for management in the context of marine environmental protection and regional development. Doctoral thesis, Kiel, 231 p. Reck, G., 1986. Relaciones Biogeograficas y distribucion de algunas especies marinas de interes comercial en las islas Galapagos. Actas del Coloquio Ecuador. Cultura 3 (24), 241-254. 40 Roberts, CM., Bohnsack, J.A., Gell, F., Hawkins, J.P., Goodridge, R., 2001. Effects of marine reserves on adjacent fisheries. Science, 294(5548), 1920-1923. Ruttenberg, B., 2001. Effects of artisanal fishing on marine communities in the Galapagos Islands. Cons. Biol. 15, 1691-1699. Sala, E., 1998. Temporal viariability in abundance of the sea urchin Paracentrotus lividus and Arbacia lixula in the western Mediterranean: composition between a marine reserve and a non protected zone. Mar. Ecol. Prog. Ser. 168, 425-439. Sala, E., Boudouresque, C.F., Harmelin-Vivien, M., 1998. Fishing, trophic cascades, and the structure of algal assemblages: evaluation of an old but untested paradigm. Oikos 82, 425-239. Scheffer, M., Carpenter, S., Foley, J.A., Folke, C, Walker, B., 2001. Catastrophic shifts in ecosystems. Nature 413, 591-596. Schroeter, S.C., Reed, D.C., Kushner, D.J., Estes, J.A., Ono, D.S., 2001. The use of marine reserves in evaluating the dive fishery for the warty sea cucumber {Parastichopus parvimensis) in California, U.S.A. Can. J. Fish. Aquat. Sci. 58 (9), 1773-1781. Snell, H.M., Stone, P.A., Snell, H.L., 1995. Geographical characteristics of the Galapagos Islands. Noticias de Galapagos 55, 18-24. Sutherland, J.P., 1974. Multiple stable points in natural communities. Am. Nat. 108, 859-873. Vetter, EW. 1994. Hotspots of benthic production. Nature 372, 47-47. Vetter, E.W. 1995. Detritus-based patches of high secondary production in the nearshore benthos. Mar. Ecol. Prog. Ser. 120, 251-262. Vetter, E.W., Dayton, P.K. 1999. Organic enrichment by macrophyte detritus, and abundance patterns of megafaunal populations in submarine canyons. Mar. Ecol. Prog. Ser. 186, 137-148. Walters, C.J., Christensen, V., Pauly, D., 1997. Structuring dynamic models of exploited ecosystems from trophic mass balance assessments. Rev. Fish Biol. Fish. 7 (2), 139-172. Walters, C.J., Kitchell, J.F., Christensen, V., Pauly, D., 2000. Representing density dependent consequences of life history strategies in an ecosystem model. Ecosystems 3, 70-83. Walters, C.J., Pauly, D., Christensen, V., 1999. Ecospace: prediction of mesoscale spatial patterns in trophic relationships of exploited ecosystems, with emphasis on the impacts of marine protected areas. Ecosystems 2, 539-564. Wellington, G.M., 1984. Marine environment and protection. In R. Perry (Editor), Key Environments. Galapagos, Pergamon Press, pp. 247-263. Wellington, G.M., Strong, A.E., Merlen, G., 2001. Sea surface temperature variation in the Galapagos Archipelago: a comparison between AVHRR nighttime satellite data and in-situ instrumentation (1982-1988). Bull. Mar. Res. 69, 27-42. Wellington, G.R., 1975. The Galapagos coastal marine environments. Report to Ministry of Agriculture and Livestock, Ecuador. 341 p + refs. Witman, J. D., Smith, F., 2003. Rapid community change at a tropical upwelling site in the Galapagos Marine Reserve. Biodivers. Conserv. 12(1), 25-45. 41 CHAPTER 3. Discovery of anemone barrens in the Galapagos and potential explanations ABSTRACT Mono-specific carpets of the anemone Aiptasia sp. were observed for the first time after the 1997/98 El Nino event on shallow horizontal reef platforms at Fernandina and Isabela Islands in the Western Galapagos Archipelago, Ecuador. Detailed transects in this area during the middle 1970s (Wellington 1975), including a transect across one of these reef platforms, indicated that these recently observed 'anemone barrens' have replaced previous algal and invertebrate assemblages. Observations during the last 10 years also support the notion that these barrens emerged recently. Community surveys and whole community trophic modeling was conducted during the present study to evaluate the community impacts of the anemone barrens and the possible mechanisms for their emergence and maintenance. Surveyed sites at these two anemone barrens were found to have significantly fewer fish species than 11 other sites in the western portion of the archipelago. Declines in marine iguana (Amblyrhynchus cristatus) populations might be explained by the emergence of the anemones, which can preempt the subtidal reef areas that normally produce the marine algae food of adult marine iguanas. These anemone barrens might have emerged as the result of recent severe El Nino-Southern Oscillation events, but preliminary whole community trophic modeling indicates that depletion of predators by unsustainable fisheries might be sufficient, in itself, to trigger the emergence of anemone barrens. Anemone barrens might be a true alternate stable state because of their potential ability to resist invasion by any other species. INTRODUCTION Pre-emption of space is one mechanism that can lead to the domination of particular habitats (Sousa 1979b, a) and alternate stable states (Sutherland 1974), or at least alternate persistent states (Connell and Sousa 1983, but see discussion by Peterson 1984, Sousa and Connell 1985, Sutherland 1990). Such dominance by a single species that is competitively superior can result from the removal of predators (Paine 1966, Paine 1969, Knowlton 1992), or by changes to the natural disturbance regime (Dayton 1971, Sousa 1979b, a, 1984, Petraitis and Latham 1999, Dudgeon and Petraitis 2001), or both. Anemone barrens have recently appeared on some shallow horizontal reef platforms in the Western Galapagos Archipelago (Okey et al. 2003). I define anemone barrens as areas of reduced diversity of species or biogenic habitat structure caused by the unchecked spread of competitively dominant anemones. These carpets of Aiptasia sp. anemones (genus identified by D. Fautin; University of Kansas) were surveyed and quantified in the two locations at which they were observed in great abundance during the present study. Several lines of evidence indicate that this is a new phenomenon. For example, Wellington (1975) found a somewhat diverse assemblage of algae, invertebrate, and fish species at a transect location where Scoresby Shepherd (South Australian Research and Development Institute) 42 and I found 95% cover of Aiptasia anemones in 2000. One conservation concern relating to this phenomenon is that the anemones have invaded much of the space that was previously used for grazing on algae by the largest remaining population of marine iguanas (Amblyrhynchus cristatus) in the Galapagos. There are other examples of anemones and other Hexacorallia spreading over intertidal and subtidal areas and persisting for some time. Rhodactis rhodostoma, a corallimorpharian, invaded large areas of the inner reef flat at Eilat in the northern Red Sea (up to 69% of the area) after an unusually extreme low tide event in 1970, effectively replacing corals there (Chadwick-Furman and Spiegel 2000). Heteractis magnifica dominated the reef at Moorea, Society Islands, until cyclones disrupted the anemone fields (D. Fautin, pers. comm. with S. Shepherd). Anthopleura elegantissima can dominate large areas in sheltered intertidal habitat along North America's west coast (Dayton 1971, Ayre and Grosberg 1995, present author's pers. obs.). Metridium senile grows over large areas in the shallow subtidal of Europe and western North America (Purcell and Kitting 1982, Anthony and Svane 1995, present author's pers. obs.). A mechanism for the persistence of extensive anemone colonies was suggested by Dayton (1971): colonial anemones can successfully resist invasion and effectively pre-empt space by consuming any propagules of competitors that would otherwise settle. There is evidence that similar settlement interference by polychaetes plays a prominent role in shaping soft bottom benthic communities as well (Woodin 1974b, a, Peterson 1979). Two general classes of potential explanations for the appearance of the anemone barrens on the horizontal reef platforms in the Galapagos are (1) exotic magnitudes (sensu Sousa 1984) of physical oceanographic forces and (2) exotic changes in the trophic interactions shape these reef communities, or both. The 1997-1998 El Nino was a particularly extreme oceanographic event that heated the surface waters of this region to unusual levels for an unusual duration. It was barely second in magnitude to the strongest recorded El Nino event (1983-1984), but the 1997-1998 event had two maximum peaks rather than one (e.g., Wolter and Timlin 1998) thus causing widespread ecological changes throughout the tropical Pacific (Jimenez et al. 2001, Aronson et al. 2002) perhaps by causing longer-lasting maximum effects [still, the 1983-1984 El Nino event is known to have caused widespread ecological changes (Glynn 1990)]. As a whole, these increasing impacts might portend considerable changes to tropical reefs in the future (Wilkinson 1996, McClanahan 2002). Observations during benthic surveys in the western archipelago conducted during 1998-99 showed that this El Nino event caused considerable mortalities of cold-water sessile species on the shallow reef platforms along Eastern Fernandina Island (R. Bustamante, pers. comm.). Elevated regional water temperatures associated with El Nino likely combined with normally intense solar exposure of these shallow reef platforms. This might have opened space for the opportunistic and fast-spreading Aiptasia sp. anemones. Alternatively, the rapid serial depletion of reef predators in the Galapagos Archipelago, including the very intensely-fished 'pepino' sea cucumber (Stichopus fuscus) (Chapter 2, Okey et al. 2004a), might have released this competitively dominant sessile 43 reef cnidarian from its last constraining predation force and given it free reign on these reef platforms. Preliminary evidence from Bermeo-Sarmiento (1995) indicates that S. fuscus consume small cnidarian polyps in Galapagos, and this predation could be incidental small post-settlement or juvenile anemones as these holothunans use feeding tentacles with adhesive papillae to remove 'particles' from the sea floor (Barnes 1987). It is also plausible that both classes of changes have worked in concert. For example, the shift might have been initiated by a physical force (the El Nino event), but maintained by a biotic force (the depletion of predators) (see Petraitis and Latham 1999). The purpose of the present chapter is to document the sudden appearance, existence, and characteristics of the recently discovered anemone barrens in the Galapagos archipelago, and to begin exploring the possible explanations for their emergence and maintenance. The main operational question is, 'What caused the appearance of the anemone barrens?' To begin addressing this question, I conducted dynamic simulations using the trophic model of a Galapagos rocky reef, described in Chapter 2, in an attempt to recreate the initiation of the anemone barrens with strictly fishing/trophic derived forces (predation hypothesis) and no change in oceanographic conditions. Punta Caleta Iguana Figure 3-1. Isabela Island and Fernandina Island, Galapagos, Ecuador, with locations of Punta Espinosa and Punta Mangle, where the anemone barrens were observed, and indicating the locations of other areas that were surveyed between 3 and 16 December 2000 (see Table 3-1 for exact locations of 13 transects). Green coasts are fully protected (P), blue coasts are protected zones with tourist access (T), and red coastlines are zones in which extractive uses are allowed (E), though these new designations have no effect on any of the results presented here. 44 METHODS Study location The first confirmed observations of the anemone barrens were made in 1998 at Punta Espinosa and Punta Mangle during research expeditions to Fernandina Island in the western portion of the Galapagos Archipelago (R. Bustamante, pers. comm., Figure 3-1). Sampling and observations A research expedition on the R/V Beagle was conducted between 3 and 16 December 2000 as part of the Galapagos Marine Reserve biodiversity baseline ecological monitoring program (Danulat and Edgar 2002), which commenced that year at the Charles Darwin Research Station, Puerto Ayora, Galapagos. For this survey, fish community transect sampling was conducted using standard visual reef census techniques (specific to each group), in which divers recorded species occurrences, abundances, and sizes along a single 100 m-long weighted transect line deployed along the 6 m depth contour at each sampling location (Figure 3-1, Table 3-1). Table 3-1. Coordinates and 'use zoning' of visual fish transects. Site Latitude Longitude Use zone Caleta Iguana S 00° 58.748' W 91° 26.833' Extractive Punta Mangle Norte S 00° 21.881' W 91° 22.807' Tourist Punta Mangle Sur S 00° 26.472' W 91° 23.314' Extractive Bahia Elizabeth S 00° 38.547' W91°06.181' Tourist Las Marielas S 00° 35.989' W 91° 05.477' Protected fully Punta Espinosa Sur S00° 16.349' W 91° 26.276' Tourist Punta Espinosa Norte S00° 15.775' W 91° 26.523' Tourist Punta Tortuga S00° 15.336' W 91° 23.415' Protected fully Punta Vincente Roca S 00° 03.129' W 91° 34.076' Extractive Punta Vincente Roca S 00° 03.134' W 91° 33.106' Tourist Punta Vincente Roca S 00° 03.117' W 91° 56.833' Protected fully Cabo Marshal S 00° 00.529' W91° 12.986' Extractive Cabo Marshal S 00° 01.087' W91° 12.419' Tourist Note: Coordinates were plotted with a hand-held Global Positioning System just prior to deployment of transects. For the fish surveys, Scoresby Shepherd and I swam a 5 m-wide swathe on each side of the transect lines for a total combined swathe width of 10 m (each transect covered an area of 1,000 m2) including all fish to 5 m above the substratum. This information was recorded separately for each 10 meter increment on underwater paper. The transect length differs from that reported in Chapter 2 because of design changes made after this survey. Recorded fish identifications were verified by referring to Humann (1993), usually within an hour after each transect was completed. Data were computerized from the raw data sheets using a two-person verification technique. A one-tailed Mann-Whitney test with tied ranks was conducted on number of fish species data to test the hypothesis that there were fewer fish species at the sites with anemone barrens than there were at sites without anemone barrens. A total of 45 15,206 individual fish from 57 species and 27 families were identified and enumerated from 26 benthic line transects, each covering an area of 1,000 m2, for a total sampled area of 26,000 m2 along the coasts of Isabela and Fernandina Islands, Galapagos during this pilot study (see also Okey and Shepherd 2001). The list of fish species from these transects is shown in Appendix B. Bathymetry and habitat varied among locations; some transects were located along 25° to 40° boulder slopes, some on horizontal reef platforms, and some over sand or gravel. Visibility and habitat characteristics at each site were noted. Aiptasia sp. anemones were removed from two 10 m2 circular clearings at the Punta Espinosa anemone barren using marine rust scrapers, other types of scrapers, and wire brushes with the hopes of using these clearings to evaluate the resilience of the carpet. The Punta Espinosa anemone barren was also surveyed using a diver's benthic sled (a board attached to a rope towed behind a boat such that the pitch of the board can be controlled by a diver to adjust survey depth). The purpose of this diver's benthic sled survey was to evaluate the areal extent of this anemone barren. The modeling approach A whole community trophic model of a Galapagos rocky reef (Chapter 2, Okey et al. 2004a) was used to explore the plausibility of the predator limitation explanation for the Aiptasia sp. anemone barrens' emergence on shallow reef platforms at Punta Espinosa and Punta Mangle. This whole food web model was constructed using the Ecopath approach (Polovina 1984, Christensen and Pauly 1992, Pauly et al. 2000) through the coordination of a broad collaboration of scientists who helped design the model structure and who contributed the best available parameter estimates from the scientific literature on Galapagos marine biota and from the latest empirical information from the Galapagos Marine Reserve biodiversity baseline ecological monitoring program (Danulat and Edgar 2002). My main approach was to conduct a series of simulated predator removals using Ecosim (Walters et al. 1997) to try to reveal the combinations of predator removals that would generate and maintain anemone barrens. After finding a plausible trophic scenario, I repeated this simulation under a series of different assumptions for prey vulnerability settings for the purpose of testing the sensitivity and 'dynamic reliability' of the explanatory simulation. The thermodynamic continuity of the model (its mass-balanced starting point) cannot be achieved for the 'pepino' (S.fuscus) functional group, or for the Galapagos rocky reef model as a whole, without specifying a biomass accumulation value of negative one tonne-km"2-year"1, thereby specifying a declining biomass of S.fuscus as part of the baseline model. This is due to the enormous extraction rate of S.fuscus from this reef model (2.922 tonnes-km"2-year"'). Any simulation, then, would automatically include an intense pepino (S.fuscus) fishery from the beginning. For the purpose of simulating the onset of an S. fuscus fishery, therefore, the catch and the biomass accumulation for this group was set to zero at first, and then increased. In addition, because of the preliminary evidence that S. fuscus consume small cnidarian polyps (Bermeo-Sarmiento 1995) and because of the manner in which they forage, I split the 46 anemone group into an early juvenile stage (younger than 2 months) and an older stage using a multi-stanza modeling function in Ecosim, which employs a von Bertalanffy growth function to calculate reasonably consistent biomasses and rates for lesser-known stages. RESULTS Anemone barrens observations The Aptasia sp. anemone covered greater than 90% of the reef platform deeper than 4 m (and not beyond 12 m) on the shallow reef platform at Punta Mangle Sur, which slopes gently from the intertidal within a sheltered bay. The substratum between 0 m and 2 m was 30-40% covered with Ulva sp. and 10-20% covered with other algal species. At 3-4 m depth, cover of the fleshy brown alga Padina sp. was 40-50% with most of the other space occupied by the colonial Aptasia sp. anemone (a shallow, perpendicular transect was deployed here in December 2000 in addition to the 6 m isobath transect). The echinoid Eucidaris was common only at depths less than 2 m. The Aiptasia sp. anemone barrens covered 95% of the bottom between 5 and 13 m depths on the reef platform at Punta Espinosa Norte (Table 3.1). Here the anemones were interrupted by occasional clumps of Padina, which were partially covered with the anemones. This platform is also shallow and gradual. This area was described by Wellington (1975) by his 'Transect B' (which was perpendicular to the shore and extended from a depth of 4 m to beyond 8 m). We found an algal mat covering 100% of the area between 0 and 5 m depths (90% Ulva spp.), and a sharp transition to the anemone barrens. The benthic sled surveys indicated that this anemone barren was very large, but that it was restricted to open horizontal and non-structured reef flats (away from boulder habitats) on the protected east side of Punta Espinosa within sight of Mangroves. The removal of the Aiptasia sp. anemones from the two 10 m2 circular clearings at Punta Espinosa proved extremely difficult. Each clearing took two divers over an hour of very hard work. Scoresby Shepherd and I observed Aiptasia anemones under rocks at Playa Mansa, Tortuga Bay and Academy Bay, both on Santa Cruz Island, in January 2001. We also observed large patches of Aiptasia anemones near mangroves along the west coast of Isabela Island in March 2001. Fishes Significantly fewer fish species (62% fewer) were found at the two sampling locations with anemone barrens than at the other 11 sampling locations (those without anemone barrens) (one-tailed Mann-Whitney test with tied ranks; p < 0.05). The mean number of species at the anemone barrens location (± 1 SE) was 5 ± 1.0 species, whereas the locations without anemone barrens had 13 ± 1.7 species. The two locations that had the lowest number of demersal fish species out of 13 locations/transects were the same two locations that were almost entirely covered by Aiptasa sp. 47 anemones—Punta Mangle Sur and Punta Espinosa Norte (Figure 3-2). The number of demersal fish species was also relatively low at Las Marielas and Punta Vincente Roca sites, perhaps because these were the only two sites sampled in failing daylight (between 1700 and 1800 hours). At least 10 species per transect were found at all other sites. Punta Mangle Sur (E) Punta Espinosa Norte (T) Punta Mangle Norte (T) Punta Espinosa Sur (T) Las Marielas (P) Punta Vicente Roca(P) Caleta Iguana (E) Bahia Elizabeth(T) Punta V icente Roca (E) Cabo Marshal (E) Punta V icente Roca (T) Cabo Marshal (T) Punta Tortuga(P) 5 10 15 20 Number offish species per 1000 m2 Figure 3-2. Number of fish species observed in the December 2000 baseline monitoring program pilot study. Two sampling stations with the lowest number of species (black bars) were the two stations at which the colonial anemone Aiptasia sp. had covered 90-95 percent of the substratum. Anemone barrens were not present at stations with grey bars. (P) denotes fully protected zones; (T) denotes protected zones with tourist access; (E) denotes zones in which extractive uses (e.g., fishing) are allowed. These zones were being established at the time of the surveys and thus would have no effect on the data presented. Anemone barrens simulations Predicted changes in the biomass of anemones could be triggered in the model by specified changes in only two functional groups—pepinos (S. fuscus) and omnivorous reef fish. The omnivorous reef fish functional group is made up of damselfishes, including the territorial Stegastes leucorus beebei and S. arcifrons, which both consume and actively remove anemones and other invading biota from their territories and gardens (Branch et al. 1992). Hereafter, the omnivorous reef fish group will be referred to as Stegastes. Changes resulting from these removals are described here in sequence because of their informative trajectories. The biomass of juvenile anemones increased with the simulated removal of pepinos. This recruitment was followed by an increase of adult anemones, which then inhibited juvenile anemones (Figure 3-3). The increase in adult anemones was subsequently controlled by Stegastes in the Ecosim manifestation of the model, which has no explicit spatial structure. Subsequent specified removal of Stegastes lead to predicted dramatic increases in adult anemones. Juvenile anemones also began increasing with the onset of the Stegastes decline, but it then began declining after adult anemones had increased sufficiently to impose a detrimental competitive effect. 48 0 5 10 Time (years) Figure 3-3. Predicted changes in biomass resulting from the removal of Stichopus fuscus (upper panel) and omnivorous reef fish (Stegastes) and S. fuscus (lower panel). Aiptasia anemones were predicted to increase by a factor of five when both groups were removed. Results include all the groups in the 44 compartment Galapagos rocky reef model. Overall prey vulnerability (v) was set at 0.4, which represents a mixture of bottom-up and top-down forces structuring the system. DISCUSSION The emergence of the Aiptasia anemone barrens The first confirmed observations of the Aiptasia anemone barrens occurred in late 1998 and late 1999 at Punta Espinosa and Punta Mangle on the east coast of Fernandina Island (R. Bustamante, pers. comm.). Unconfirmed observations of Anemone 'patches' were made by Jimmy Penaherrera at Punta Espinosa in 1995. The anemone barrens were first described by Okey and colleagues (2003, 2004a). These are called 'anemone barrens' because of the virtual absence of other megafauna and flora where these carpets of Aiptasia occur. A number of observations and recollections provide further insights into the timing and character of the emergence of these anemone barrens (Table 3-2). Mr. Penaherrera's observations of smaller anemone patches in 1995 would be consistent with the notion that El Nino conditions, or reductions in predators such as S. fuscus, or both, can lead to the emergence of anemone barrens. El Nino conditions indeed predominated from 1991 to 1995 (though water temperatures did not reach the record levels of 1982/1983 and 1997/98). Also, the pepino (S. fuscus) fishery apparently arrived in the Galapagos from mainland Ecuador beginning in 1991 (Jenkins 49 and Mulliken 1999). This fishery generally increased in intensity through 1996, and probably beyond. The subsequent confirmed observations of the anemone barrens in 1998 and 1999 point to the strong 1997/1998 El Nino as the initiator of the broader-scale shift (Table 3-2), but reduction of predators cannot be ruled out as a contributor even to this conspicuous shift (or at least to its persistence). Table 3-2. Observations pertinent to the appearance of anemone barrens in the Galapagos Archipelago Year Observation Location Observer 1975 No anemones seen Punta Espinosa G.M. Wellington 1993 Does not recall seeing any anemones Punta Espinosa J. Penaherrera 1994 No anemones seen Punta Espinoza, Punta Mangle R. H. Bustamante 1993-1996 Ulva spp. and foliose red algae dominate Punta Espinosa & Mangle P.C. Martinez 1995 Recollections of anemone patches seen Punta Espinosa J. Penaherrera Aug 1997 Bleached coralline algae dominate; ophiuroids reduced Punta Espinosa & Mangle P.C. Martinez, R.H. Bustamante June 1998 Large and discrete patches of anemones found Punta Mangle F. Rivera, R.H. Bustamante Feb 1999 Extensive anemone barrens found Espinosa, Mangle, Priscilla F. Rivera, R.H. Bustamante Nov 1999 Barrens covered much of bottom; 1.5 - 10m Espinosa, Mangle, Priscilla P.C. Martinez June 1999 Anemones found in great abundances Punta Espinosa C. Hickman Dec 2000 Anemone barrens observed Punta Espinosa & Mangle Okeyetal. (2003) Jan 2001 Aiptasia anemones observed under rocks Playa Mansa, Tortuga Bay S.A. Shepherd Jan 2001 Observed only under rocks; Stegastes present Academy Bay, Santa Cruz Isl S.A. Shepherd Feb 2001 Anemones observed in abundance Cabo Douglas, Fernandina K. Fujiwara Mar 2001 Aiptasia patches observed near mangroves West coast of Isabela Island TAO & SAS Notes: Observations summarized from Okey et al. (2003) and further refined. Wellington (1975) found that Ulva, Amphiroa, and Codium covered almost 100% of the reef platform to 10 m depth at his transect B, which was located about 300 m south of Punta Espinosa. Scoresby Shepherd and I observed this same area of his transect B to be entirely carpeted with the anemones in December of 2000. Occasional surviving specimens of the alga Padina were somewhat smothered by anemones on this reef platform. At least the lower surfaces of the plants were covered by anemones and many upper surfaces had anemones as well. The ecological effect of Aiptasia anemone barrens The finding of significantly lower numbers of demersal fish species at the sampling locations at the Aiptasia anemone barrens, despite the low sample sizes and the unequal comparative design, indicates that the effect these barrens have on the ecology of Galapagos reefs is significant. The almost complete replacement of the diverse assemblage at Punta Espinosa described by Wellington (1975) by Aiptasia anemones indicates a dramatic and fundamental shift in the structure of this marine community, and its potential for supporting fish and invertebrate communities. Most of the apparent ecological effects of 50 these barrens have not been quantitatively evaluated, but one conspicuous conservation concern with global significance is that the largest remaining marine iguana (A. cristatus) colony in the world is located adjacent to the Aiptasia barren at Punta Espinosa (Laurie 1983, Snell and Marquez 2002). The adults of this herbivorous iguana species rely largely on the normally lush green macroalgae mats on tops of these reefs (juveniles feed in the intertidal and shallower subtidal). The anemone barrens have apparently reduced adult marine iguana foraging areas. The iguana population has experienced considerable famine mortalities, apparently related to recent El Nino-Southern Oscillation events (Romero and Wikelski 2001). An alternate explanation for the significantly lower number of demersal fish species is that reef topography influenced fish assemblages and the presence of anemones. That is to say, open reefs support few fish species due to a lack of topographic features rather than because of the assemblage of sessile fauna. Indeed, the anemones might owe part of their success on these open reefs to the preference of omnivorous reef fishes (i.e., Stegastes) to boulder habitat that provides refuge. Still, replacement of the sessile benthos that provides biogenic fish habitat from the open reefs likely has adverse effects on many fish and mega invertebrates. The overall community is shaped by a combination of biotic and abiotic habitat. Simulations of the emergence of Aiptasia anemone barrens The results of the pepino (S. fuscus) and Stegastes removal simulations are consistent with observed anemone distribution patterns. First, the Stegastes damselfishes stand out as the strongest existing predatory controllers of Aiptasia anemones in the present model, as indicated by the strong response of anemones to Stegastes removal (Figure 3-3). These fishes, however, are found only around boulder habitat, which they use as refuge from their predators. Anemone barrens are found only on exposed platform reefs, where Stegastes spp. was not observed to venture. In cobble habitat near boulders, Stegastes arcifrons quickly attacked anemones that Scoresby Shepherd and I exposed by turning rocks. In contrast to Stegastes, the sea cucumber S. fuscus inhabit exposed or moderately exposed platforms presumably because of its evolved defenses against predators (Francour 1997). S. fuscus is a suspected predator of Aiptasia because Bermeo-Sarmiento (1995) found 'cnidarian polyps' in S. fuscus stomachs that were collected from this same sub-region of the Galapagos. Being the only other predator of these anemones known to recently exist at functional levels (according to the specification of the Galapagos rocky reef model), it is quite possible that a recently emerged sea cucumber fishery has indirectly triggered the emergence of these anemone barrens by considerably reducing S. fuscus (Okey et al. 2004a)—which was perhaps the last remaining check on the anemone population on the open reef flats. This more spatially explicit hypothesis can be pursued further when Ecosim capabilities are combined in the future with the Ecospace routine, which is a habitat-based, spatially explicit routine. 51 Aiptasia is a well known aquarium pest that can take over all aquarium surfaces rapidly and harm preferred display species. Aquarists commonly use natural predators of Aiptasia, such as the hyppolytid shrimp Lysmata wurdemanni to control this anemone (Rhyne et al. 2001). Butterfly fish and nudibranchs (both discussed above) are less reliable for this purpose due to handling difficulties. Lysmata spp. are common in tropical areas of the eastern Pacific (Wicksten 1990) and at least three species in this genus are found in the Galapagos: L. argentopunctata, L. galapagensis, and L. chica (Hickman and Zimmerman 2000). Some species of Lysmata do not appear to prey on Aiptasia and, to my knowledge; it is not known whether these three do. Some members of the prawn genus Rhynchocinetes are also thought to feed on Aiptasia, and one species, Rhynchocinetes typus, occurs in the Galapagos (Hickman and Zimmerman 2000), but again it is not known to the present author whether this particular species feeds on Aiptasia. If any of these crustacean species do prey on, and help control, Aiptasia on Galapagos rocky reefs, their populations were likely reduced by the 1997/1998 El Nino event, at least on shallow reefs like much of the attached flora and fauna, thus enabling the anemone invasion. Established adult anemones might benefit from size or colonial refuge from such crustacean predators. Furthermore, most of these crustaceans are cryptic during the day when they can be found under boulders and active foragers at night. Thus, like Stegastes, these crustaceans seem more likely to have effects near boulder refuge habitats than on open reef flats where the anemone barrens are found. Notwithstanding these possibilities, crustaceans were not specified to consume anemones in the current model iteration because specific information on such feeding interactions was not available for the Galapagos reefs. It is possible that lobsters, whelks, and conch normally exert some predatory control over Aiptasia as well. All are regularly removed as "byproducts" during pepino (S.fuscus) diving by fishers. These casual non-target removals are very effective in removing this biota (R. Bustamante, pers. comm.). Mega-invertebrates such as sea urchins might also graze on anemones, either incidentally or intentionally. There is no information on the feeding of Aiptasia anemones by any of these organisms, however, so these are not considered here as main candidates for Aiptasia emergence. Many vertebrate species prey on anemones (Actinaria) and other orders in the anthozoan subclass Hexacorallia (Table 3-3). Some of these predators, or related species, were formerly more abundant in the Galapagos Archipelago than they are now. Chaetodon auriga can be seen only at the northern Darwin and Wolf Islands (Humann 1993), but it is rare, and C. unimaculatus has not been recorded, but two other Chaetodon species (Appendix B) are present in some locations. A. meleagris is uncommon, but has been recorded (Anon 2001). Hawksbill turtles were considerably more abundant in the past (NRC 1990, Bjorndal et al. 1993), and they might have shaped shallow reef ecosystems strongly (Jackson 1997, Bjorndal et al. 2000, Jackson 2001). The few predators that remain in the vicinity of these anemone barrens have been ineffective in controlling the spread of Aiptasia on these platform reefs. 52 The results of both predator removal simulations indicated competition for food between adult and juvenile (e.g., budding or settling) anemones, thus implying food limitation in the absence of predators. Although this is an expected result of using a mass-continuity Ecopath model, as a population will reach food limitation when all sources of mortality are removed, it indicates a reasonable result if these two predators are indeed the only remaining anemone predators in the system—an unresolved question. Nevertheless, these results are consistent with the pattern that anemone barrens, to the best of my knowledge, have appeared only on reef platforms that are protected from swell and that have minimal vertical profile or complexity that might offer shelter to cryptic predators. In addition, the Anemone barrens are often adjacent to mangroves, which are sources of considerable organic detritus to local marine ecosystems (Jennerjahn and Ittekkot 2002). Other anfhozoans are known to get large proportions of their nutrition from such terrigenous sources (Risk-et al. 1994). Manipulative field experiments should be conducted to explore these dynamics further, and to refine conceptual and quantitative models. Table 3-3. Some predators of Aiptasia sp. or other Hexacorallia Predator Common name Prey Location Source Stegastes arcifrons Yellow-tail damselfish Aiptasia sp. Galapagos Grove & Lavenberg (1997)a S. leucorus beebei White-tail damselfish Aiptasia sp. Galapagos Grove & Lavenberg (1997f Stichopus fuscus Pepino sea cucumber Cnidarian polyps Galapagos Bermeo-Sarmiento (1995) Lysmata wurdemanni Pepperment shrimp Aiptasia pallida Caribbean/Gulf Rhyneetal.(200T) Berghia verrucicornis Aeolid nudibranch A. pallida Caribbean, etc. Kempf(1991) B. major Aeolid nudibranch Aiptasia pulchella Hawaii Muller-Parker(1984) Arothron meleagris Guineafowl puffer A. pulchella Hawaii Muller-Parker(1984) Chaetodon fasciatus Butterflyfish Anemones The Red Sea Fricke(1975) Chaetodon auriga Threadfin butterflyfish A. pulchella Hawaii Muller-Parker(1984) C. unimaculatus Teardrop butterflyfish A. pulchella Hawaii Muller-Parker(1984) Clinocottus globiceps Mosshead sculpin Anthopleura spp. Washington, USA Hand (1996) Aeolidia papillosa Aeolid nudibranch Anthopleura spp. Washington, USA Waters (1973) Dermasterias imbricata Leather star Anthopleura spp. Washington, USA Sebens(1977) >50 fish species Fish Anemones The world Ates(1989) Eretmochelys imbricata Hawksbill sea turtle Anemones Den Hartog (1980) E. imbricata Hawksbill sea turtle Zoanthus sociatus St. Croix, USVI Mayor etal. (1998) E. imbricata Hawksbill sea turtle corallimorpharian Dom. Republic Leon (2000) Caretta caretta Loggerhead sea turtle Anemones Gulf of Mexico Plotkinetal. (1993) Erignathus barbatus Bearded seal Anemones Can. high arctic Finley and Evans (1983) Many bird species Birds Anemones The world Ates(1991) Notes: Some of these anemone predators, or closely related groups, are less common in the Galapagos now than they were in the past, and this depletion of predators might partially explain the emergence of Anemone barrens (see text), (a) and personal observations. In addition to the benefits these Aiptasia anemones might receive from adjacent mangroves, their gastrodermal cells contain the photosynfhetic symbiont zooxanthellae (a dinoflagellate), thus enhancing their growth and vigor in times of food scarcity (Clayton and Lasker 1985). The occurrence of Aiptasia on 53 these shallow reef platforms suggests that their distribution is influenced by this symbiosis. Anemones can reproduce sexually and asexually by cloning through longitudinal fission, inverse budding or marginal budding, but pedal laceration is the method used by Aiptasia (G. Muller-Parker, pers. comm.) This method of asexual reproduction entails very low reproductive effort (Hunter 1984), and it is considered a good strategy for colonizing space because clones inherit high site-specific fitness (Schick 1991, Ayre and Grosberg 1995) and because budded clones can colonize space rapidly (G. Muller-Parker pers. comm.). Cloning can thus result in extensive patches on reef bottoms. Pedal laceration also ensures the transfer of zooxanthellae (Muller-Parker and D'Elia 1997). The advantages of exploiting carbon subsidies from both mangroves and zooxanthellae, as well as the low reproductive effort required for cloning, are compatible with patterns observed in the field and in the present simulations. These characteristics are likely contributors to the recent emergence of Aiptasia anemone barrens. An explanation for the recent spread of the Aiptasia anemone in the Galapagos, other than the depletion of predators, is that the biological assemblages on the shallow reef flats of Fernandina were severely impacted by unusually high temperatures and other characteristics associated with the El Nino-Southern Oscillation events of the 1990s. Benthic surveys conducted during the 1997/98 ENSO event, have showed that nearly all fleshy macro algae and encrusting corallines were killed or bleached and up to 98% of the bed-forming sessile gastropods Hipponix grayanus and the giant barnacle Balanus spp. were dead by mid 1998 (R. Bustamante, unpubl. data). This mass mortality of the existing assemblage led to the availability of space for colonization by the opportunistic ("weedy") anemone species Aiptasia sp. The resulting physical pre-emptive presence of the anemone invaders, their consumption of settling propagules, and their nematocyst defenses all constitute reinforcing feedbacks that might stabilize the alternative community state of these anemone barrens as long as no successful predators or severe physical limiting mechanisms intervene. The reinforcing feedbacks related to the preemptive presence of these anemones is reasonable evidence that these anemone barrens constitute a true alternate stable state (Sutherland 1974, Connell and Sousa 1983, Carpenter 2000, Scheffer et al. 2001), but further scrutiny of this system will be required to evaluate this possibility. Several studies (Schick 1991, Anthony and Svane 1995, Chadwick-Furman and Spiegel 2000) lend support to the general hypothesis that catastrophic disturbances can lead to barrens dominated by anemones. The extreme low tide in the northern Red Sea (discussed in the Introduction) was presumably a natural, albeit rare, event. In this Red Sea example, it is possible that the anemone-coral assemblages have adapted to that rare disturbance, albeit on a broad temporal scale. Similarly, did the Galapagos fauna on these shallow reef platforms ever adapt to the apparent deleterious effects of El Nino events that are as severe as those recently observed? Observed increases in the frequency and intensity of El Nino events corresponding with sudden shifts in community states would lend weight to the notion that exotic disturbances or natural disturbances with exotic (new) characteristics may have led to an alternative community state that has persisted for several years. It would not be surprising if the severe 1997/98 El 54 Nino event was the proximate cause of the expansion of an anemone, probably already present in low abundance in the archipelago, but this would not rule out the possibility that a depletion of predators, by both ENSO-related disturbance and fishing, was a key reason for the spread of Aiptasia. Indeed, the simulations described here using a whole food web model of a Galapagos rocky reef indicates that anemones can spread over these reefs through trophic forces alone (i.e., the release from predator control), which could include the effects of unsustainable fisheries in the Galapagos Archipelago. Distinguishing the relative roles of these two general explanations is the task of the continuing investigation of the anemone barrens phenomenon. An iterative combination of manipulative experimentation (sensu Petraitis and Dudgeon 1999, Petraitis and Latham 1999) and continually refined whole food web dynamic simulations would be useful to this end. The emergence and persistence of these sessile benthic cnidarians resembles recent observations of increases in planktonic cnidarians (jellyfish) in various oceans (e.g., Mills 2001). Both seem to be alternative energy pathways in marine ecosystems that we can reasonably assume to have intrinsically-reinforced persistence. This similarity is especially worrisome when considering the notion that both types of cnidarian increases might be partially caused by fisheries (Pauly et al. 2002, the present chapter). Such observations beg the question of whether it is possible that further anthropogenic degradation of marine ecosystems could lead to an 'age of cnidarians,' like the Cambrian Period. CONCLUSIONS Extensive carpets of Aiptasia anemones were discovered at Punta Espinosa and Punta Mangle on the east coast of Fernandina Island, Galapagos, Ecuador. They appear to have emerged at these locations and in the vicinity during the 1990s, but particularly in 1998-1999. Fish community surveys indicated that these Aiptasia carpets were negatively associated with number of fish species, thereby helping to justify the term 'anemone barrens.' Although the increasing intensity and frequency of El Nino-Southern Oscillation events might have initiated the emergence of these anemone barrens, preliminary simulations using a whole community trophic model indicates that the depletion of predators, e.g., through unsustainable fisheries or environmental disturbance, is an adequate causal mechanism for their emergence. It is possible that predator removals and direct El Nino forces worked in concert. These anemone barrens might be a true alternate stable state because of their potential ability to resist invasion by any other species through preemptive mechanisms that reinforce their own state, but this possibility awaits further evaluation. 55 LITERATURE CITED Anon. 2001. Fish collection database of the National Museum of Natural History. Smithsonian Institution, Division of Fishes. Anthony, K. R. N., and I. Svane. 1995. Effects of substratum instability on locomotion and pedal laceration in Metridium senile (Anthozoa, Actiniaria). Marine Ecology Progress Series 124:171-180. Aronson, R. B., W. F. Precht, M. A. Toscano, and K. H. Koltes. 2002. The 1998 bleaching event and its aftermath on a coral reef in Belize. Marine Biology 141:435-447. Ates, R. M. L. 1989. Fishes that eat sea anemones: A review. Journal of Natural History 23:71-80. Ates, R. M. L. 1991. Predation on Cnidaria by vertebrates other than fishes. Hydrobiologia 216-217:305-308. Ayre, D. J., and R. K. Grosberg. 1995. Aggression, habituation and clonal existence in the sea anemone Anthopleura elegantissima. American Naturalist 146:427-453. Barnes, R. D. 1987. Invertebrate zoology, 5th edition. Saunders College Publishing, Philadelphia. Bermeo-Sarmiento, J. D. 1995. Informe sobre la evaluacion del recurso pepinos de mar Isostichopus fuscus (Ludwig 1875) en areas interiores del Canal Bolivar en las Islas Galapagos. Instituto nacional de Pesca, Informe. Bjorndal, K. A., A. B. Bolten, and M. Y. Chaloupka. 2000. Green turtle somatic growth model: Evidence for density dependence. Ecological Applications 10:269-282. Bjorndal, K. A., A. B. Bolten, and C. J. Lagueux. 1993. Decline of the nesting population of hawksbill turtles at Tortuguero, Costa-Rica. Conservation Biology 7:925-927. Branch, G. M., J. M. Harris, C. Parkins, R. H. Bustamante, and S. Eekhout. 1992. Algal 'gardening' by grazers: a comparison of the ecological effects of territorial fish and limpets. Pages 405-424 in D. John, editor. Plant-animal interactions in the marine benthos. Systematic Association Special Edition 46. Claredon press, Oxford. Carpenter, S. R. 2000. Alternate states of ecosystems: Evidence and its implications for environmental decisions. Pages 357-383 in M. C. Press, N. Huntley, and S. Levin, editors. Ecology: achievement and challenge. Blackwell, London. Chadwick-Furman, N. E., and M. Spiegel. 2000. Abundance and clonal replication in the tropical corallimorpharian Rhodactis rhodostoma. Invertebrate Biology 119:351-360. Christensen, V., and D. Pauly. 1992. Ecopath II: a software for balancing steady-state ecosystem models and calculating network characteristics. Ecological Modelling 61:169-185. Clayton, W. S., and H. R. Lasker. 1985. Individual and population growth in the asexually reproducing anemone Aiptasia pallida. Journal of Experimental Marine Biology and Ecology 90:249-258. Connell, J. H., and W. P. Sousa. 1983. On the evidence needed to judge ecological stability or persistence. The American Naturalist 121:789-824. Danulat, E., and G. Edgar, editors. 2002. Reserva Marina de Galapagos. Linea base de la Biodiversidad. Fundacion Charles Darwin/Servicio Parque Nacional Galapagos, Santa Cruz, Galapagos, Ecuador. Dayton, P. K. 1971. Competition disturbance and community organization the provision and subsequent utilization of space in a rocky intertidal community. Ecological Monographs 41:351-389. Den Hartog, J. C. 1980. Notes on the food of sea turtles: Eretmochelys imbricata (Linnaeus) and Dermochelys coriacea (Linnaeus). Netherlands Journal of Zoology 30:595-610. Dudgeon, S., and P. S. Petraitis. 2001. Scale-dependent recruitment and divergence of intertidal communities. Ecology 82:991-1006. Finley, K. J., and C. R. Evans. 1983. Summer diet of the bearded seal (Erignathus barbatus) in the Canadian high arctic. Arctic 36:82-89. Francour, P. 1997. Predation on holothurians: A literature review. Invertebrate Biology 116:52-60. Fricke, H. 1975. Selektives Feinderkennen bei dem Anemonenfisch Amphiprion bicinctus (Riippell). Journal of Experimental Marine Biology & Ecology 19:1-7. Glynn, P. W. 1990. Global ecological consequences of the 1982-83 El Nino-Southern Oscillation. Elsevier, Amsterdam, New York. 56 Grove, J. S., and R. J. Lavenberg. 1997. The fishes of the Galapagos Islands. Stanford University Press, Stanford, Calif. Hand, C. 1996. The alarm response and some predators of the sea anemone Anthopleura xanthogrammica (Brandt). Wash. J. Biol. 51:9-23. Hickman, C, and T. Zimmerman. 2000. A field guide to crustaceans of Galapagos. Sugar Spring Press, Lesington, VA. Humann, P. 1993. Reef fish identification: Galapagos. New World Publ., Jacksonville, Florida. Hunter, T. 1984. The energetics of asexual reproduction: Pedal laceration in the symbiotic sea anemone Aiptasia pulchella. Journal of Experimental Marine Biology and Ecology 83:127-148. Jackson, J. B. C. 1997. Reefs since Columbus. Coral Reefs 16:S23-S32. Jackson, J. B. C. 2001. What was natural in the coastal oceans? Proceedings of the National Academy of Sciences of the United States of America 98:5411-5418. Jenkins, M., and T. A. Mulliken. 1999. Evolution of exploitation in the Galapagos Islands: Ecuador's sea cucumber trade. Traffic Bulletin 17:[electronic pages]. Jennerjahn, T. C, and V. Ittekkot. 2002. Relevance of mangroves for the production and deposition of organic matter along tropical continental margins. Naturwissenschaften 89:23-30. Jimenez, C, J. Cortes, A. Leon, and E. Ruiz. 2001. Coral bleaching and mortality associated with the 1997-98 El Nino in an upwelling environment in the Eastern Pacific (Gulf of Papagayo, Costa Rica). Bulletin of Marine Science 69:151-169. Kempf, S. C. 1991. A primitive symbiosis between the aeolid nudibranch Berghia verrucicornis (a Costa, 1867) and a zooxanthella. Journal of Molluscan Studies 57:75-85. Knowlton, N. 1992. Thresholds and multiple stable states in coral reef community dynamics. American Zoologist 32:674-682. Laurie, W. A. 1983. Marine iguanas in Galapagos. Oryx 17:18-25. Leon, Y. M. 2000. Selective feeding by the hawksbill turtle, an important predator in coral reef ecosystems. M.S. University of Florida, Gainesville. Mayor, P. A., B. Phillips, and Z. M. Hillis-Starr. 1998. Results of the stomach content analysis on the juvenile hawksbill turtles of Buck Island Reef National Monument, USVI. Pages 230-232 in S. P. Epperly and J. Braun, editors. Proceedings of the seventeenth annual sea turtle symposium. U.S. Department of Commerce, NOAA Technical Memorandum, NMFS-SEFSC-415. McClanahan, T. R. 2002. The near future of coral reefs. Environmental Conservation 29:460-483. Mills, C. E. 2001. Jellyfish blooms: are populations increasing globally in response to changing ocean conditions? Hydrobiologia 451:55-68. Muller-Parker, G. 1984. Dispersal of zooxanthellae on coral reefs by predators on cnidarians. Biological Bulletin (Woods Hole) 167:159-167. Muller-Parker, G., and C. F. D'Elia. 1997. interactions between corals and their symbiotic algae. Pages 96-113 in C. Birkeland, editor. Life and death of coral reefs. Chapman and Hall Publishers, New York. NRC. 1990. Decline of the sea turtles: Causes and prevention. National Research Council (U.S.), Committee on Sea Turtle Conservation. National Academy Press, Washington, D.C. Okey, T. A., S. Banks, A. F. Born, R. H. Bustamante, M. Calvopina, G. J. Edgar, E. Espinoza, J. M. Farina, L. E. Garske, G. K. Reck, S. Salazar, S. A. Shepherd, V. Toral-Granda, and P. Wallem. 2004. A trophic model of a Galapagos subtidal rocky reef for evaluating fisheries and conservation strategies. Ecological Modelling 172:383-401. Okey, T. A., and S. A. Shepherd. 2001. Demersal fish assemblages at two depths (6 and 13 m) on Isabela and Fernandina Islands, Galapagos. Program BIOMAR, Charles Darwin Research Station, Santa Cruz Island, Galapagos, Ecuador. Okey, T. A., S. A. Shepherd, and P. C. Martinez. 2003. A new record of anemone barrens in the Galapagos. Noticias de Galapagos 62:17-20. Paine, R. T. 1966. Food web complexity and species diversity. American Naturalist 100:65-75. Paine, R. T. 1969. A note on trophic complexity and community stability. American Naturalist 103:91-93. Pauly, D., V. Christensen, S. Guenette, T. J. Pitcher, U. R. Sumaila, C. J. Walters, R. Watson, and D. Zeller. 2002. Towards sustainability in world fisheries. Nature 418:689-695. 57 Pauly, D., V. Christensen, and C. Walters. 2000. Ecopath, Ecosim, and Ecospace as tools for evaluating ecosystem impact of fisheries. Ices Journal of Marine Science 57:697-706. Peterson, C. H. 1979. Predation, competitive exclusion, and diversity in the soft-sediment benthic communities of estuaries and lagoons. Pages 233-264 in R. J. Livingston, editor. Ecological processes in coastal and marine ecosystems. Plenum, New York. Peterson, C. H. 1984. Does a rigorous criterion for environmental identity preclude the existence of multiple stable points? American Naturalist 124:127-133. Petraitis, P. S., and S. R. Dudgeon. 1999. Experimental evidence for the origin of alternative communities on rocky intertidal shores. Oikos 84:239-245. Petraitis, P. S., and R. E. Latham. 1999. The importance of scale in testing the origins of alternative community states. Ecology 80:429-442. Plotkin, P. T., M. K. Wicksten, and A. F. Amos. 1993. Feeding ecology of the loggerhead sea turtle Caretta caretta in the Northwestern Gulf of Mexico. Marine Biology 115:1-5. Polovina, J. J. 1984. Model of a coral reef ecosystem 1. The Ecopath model and its application to French Frigate Shoals. Coral Reefs 3:1-12. Purcell, J. E., and C. L. Kitting. 1982. Intraspecific aggression and population distributions of the sea-anemone Metridium-senile. Biological Bulletin (Woods Hole) 162:345-359. Rhyne, A., J. Lin, and V. Maxwell. 2001. Control of pest anemone Aiptasia pallida by ornamental shrimp Lysmata rathbunae and L. wurdemanni. in Marine Ornamentals 2001: Collection, Culture & Conservation. Florida Sea Grant College Program, Building 803 McCarty Drive PO Box 110400 Gainsville FL 32611 USA, Lake Buena Vista, FL (USA). Risk, M. J., P. W. Sammarco, and H. P. Schwarcz. 1994. Cross-continental shelf trends in 13C in coral on the Great Barrier Reef. Marine Ecology Progress Series 106:121-130. Romero, L. M., and M. Wikelski. 2001. Corticosterone levels predict survival probabilities of Galapagos marine iguanas during El Nino events. Proceedings of the National Academy of Sciences of the United States of America 98:7366-7370. Scheffer, M., S. Carpenter, J. A. Foley, C. Folke, and B. Walker. 2001. Catastrophic shifts in ecosystems. Nature 413:591-596. Schick, J. M. 1991. A functional biology of sea anemones, 1 st edition. Chapman & Hall, London ; New York. Sebens, K. 1977. Habitat suitability, reproductive ecology, and the plasticity of body size in two sea anemone populations (Anthopleura elegantissima and A. xanthogrammica). Ph.D. University of Washington, Seattle. Snell, H. L., and C. Marquez. 2002. Iguanas marinas. Pages 324-342 in E. Danulat and G. Edgar, editors. Reserva Marina de Galapagos. Linea base de la Biodiversidad. Fundacion Charles Darwin/Servicio Parque Nacional Galapagos, Santa Cruz, Galapagos, Ecuador. Sousa, W. P. 1979a. Disturbance in marine intertidal boulder fields the nonequilibrium maintenance of species diversity. Ecology 60:1225-1239. Sousa, W. P. 1979b. Experimental investigations of disturbance and ecological succession in a rocky intertidal algal community. Ecological Monographs 43:227-254. Sousa, W. P. 1984. The role of disturbance in natural communities. Annual Review of Ecology and Systematics 15:353-392. Sousa, W. P., and J. H. Connell. 1985. Further comments on the evidence for multiple stable points in natural communities. American Naturalist 125:612-615. Sutherland, J. P. 1974. Multiple stable points in natural communities. The American Naturalist 108:859-873. Sutherland, J. P. 1990. Perturbations, resistance, and alternative views of the existence of multiple stable points in nature. American Naturalist 136:270-275. Walters, C, V. Christensen, and D. Pauly. 1997. Structuring dynamic models of exploited ecosystems from trophic mass-balance assessments. Reviews in Fish Biology and Fisheries 7:139-172. Waters, V. 1973. Food-preference of the nudibranch Aeolidia papillosa, and the effect of the defenses of the prey on predation. Veliger 15:174-192. Wellington, G. M. 1975. The Galapagos coastal marine environments. A resource report to the Department of National Parks and Wildlife, Quito. 58 Wicksten, M. 1990. Key to the hippolytid shrimp of the eastern Pacific Ocean. Fishery Bulletin 88:587-598. Wilkinson, C. R. 1996. Global change and coral reefs: impacts on reefs, economies and human cultures. Global Change Biology 2:547-558. Wolter, K, and M. S. Timlin. 1998. Measuring the strength of ENSO: How does 1997/98 rank? Weather 53:315-324. Woodin, S. A. 1974a. Adult-larval interactions in dense infaunal assemblages: patterns of abundance. Journal of Marine Research 34:24-41. Woodin, S. A. 1974b. Polychaete abundance patterns in a marine soft sediment environment: the importance of biological interactions. Ecological Monographs 44:171-187. 59 CHAPTER 4. A search for keystones in Prince William Sound, Alaska using a mass-continuity trophic model ABSTRACT Keystone species play a central role in shaping at least some marine communities in the sense that system-wide phase shifts can be mediated by the presence, absence, or the relative abundance of these key interactors. Identifying keystones is considered crucial for understanding the resilience of ecosystems to exogenous forces because non-linear, or catastrophic, community changes might hinge on the populations of one or a few species. Removal experiments are traditionally the most reliable approach for identifying keystone species, but such experiments can be accomplished only on small scales and with only a few types of species in any given system. An alternate field approach to identifying keystones is the tracking of 'natural experiments', but this places scientists at the whim of chance and carries the intrinsic problem of confounding. Intensive study of a few famous keystone species has improved understanding of particular marine ecosystems and led to the development of keystone theory, but a comprehensive analytical approach to identifying and ranking keystones has not been proposed until now. I describe the results of a series of simulated (functional group) removal experiments using a whole-system food web model of Prince William Sound, Alaska for the period 1994-1996. Four indices were used to rank functional groups, based on the results of these simulated removal experiments: community importance, community longevity support, interaction strength index, and a keystone index. Here, interaction strength index (ISI) refers to trophic, not per capita, interaction strength, and the keystone index is a functional group's ISI relative to the percent of the system's biomass represented by the group. Transient orcas produced the highest keystone index followed by avian raptors. INTRODUCTION The notion that one or a small number of species can dominate a biological community (Strong 1977, Paine 1980, Dayton 1985) implies that large changes in the populations of dominant species (e.g., trees, mussels, kelp) would fundamentally alter the communities in which they dominate. Protection of dominants from too much degradation by exotic forces, while preserving natural disturbance regimes, is thus necessary for enabling the maintenance of such communities within 'natural' bounds of variation. In contrast, the realization that, in at least some ecosystems, one or a small number of species influence biological communities to an extent that is disproportionately large relative to their abundance (or biomass) (Dayton 1972, 1984) implies that small changes in certain populations can cause large, or catastrophic, shifts in the state of communities. This concept has fueled considerable interest since Paine (1966) described the phenomenon and used the term keystone species (1969) when referring to the asteroid predator Pisaster ochraceus that apparently structures a particular rocky intertidal community very strongly by preying mainly on 60 barnacles (Balanus glandula) and mussels (Mytilus californianus) thus preventing the competitively-dominant mussels from monopolizing a limiting resource; i.e., the open space in the intertidal zone at Mukkaw Bay, Washington, USA. Continued study has supported the notion that this intertidal community type is shaped strongly by keystone species, that interaction strengths vary greatly among species, and that indirect effects are common in marine rocky intertidal communities (e.g., Paine 1974, Menge et al. 1994, Menge 1995, Navarrete and Menge 1996). Paine's usage of the term keystone species—a species that maintains a community's structure, integrity, and stability by preventing a competitive dominant from monopolizing a resource—appeared to also fit sea otters (Enhydra lutris), which, when living in hard-bottom ecosystems, usually succeed in controlling herbivores that would otherwise control habitat-forming kelp canopies (Estes and Palmisano 1974, Estes and Duggins 1995, Estes et al. 1998). It also fit a number of other examples to which the term keystone species appears to have stuck (e.g., see Table 1 in Power etal. 1996). Because of the dramatic community shifts implied by, and inherent in, the loss of keystone species, Power and her colleagues (1996) suggested that searching for and identifying them is essential for understanding how biological communities will be affected by species losses or reductions. Their suggestion appears even more urgent in light of recent revelations about the remarkable rates of recent reductions of higher trophic level species in marine ecosystems (Pauly et al. 1998a, Myers and Worm 2003, Springer et al. 2003). Based on musings by Mills, Soule, and Doak (1993), Power et al. (1996) provided an operational definition for identifying and ranking keystone species that included an index called community importance; i.e., ".. .the change in a community or ecosystem trait per unit change in the abundance of the species', thus clarifying the keystone concept in terms very similar to Dayton's (1972) original discussion of foundation species. Finally, after presenting helpful approaches to revealing keystones of species in communities, they challenged ecologists to begin the important task of identifying keystones by measuring the relative effects of species or guilds in communities. An important aspect of the notion of keystone, or foundation, species is the observation that changes in the abundance or biomass of most species in a system have imperceptible effects on populations of other species or the overall community, while one or a few species (generally with low abundances or biomass) have large, or a very large, effects (Dayton 1984, Paine 1992, Power et al. 1996). Removal of these species from the system might well cause strong cascading effects (e.g., Carpenter and Kitchell 1993, Shiomoto et al. 1997) or the loss of natural cascades (Pace et al. 1999) that structure the system. The salient point is that catastrophic, or at least non-linear, community changes can occur if and when such ecologically special species are reduced or removed. Differences in relative interaction strengths, or the identity of keystone species, have been revealed in the past either through manipulative field experiments (e.g., Fowler 1981, Menge and Farrell 1989, Paine 1992) or by tracking 'natural experiments' (e.g., Estes and Duggins 1995, Estes et al. 1998) and the intensive study of such famous keystone examples has improved understanding of those particular 61 marine ecosystems. However, such approaches are inherently limited to either small scales or fortuitous opportunities to glimpse the roles of small subsets of species. In contrast, computer models of marine food webs are limited in the sense of their simplicity relative to the complexities of real systems, the uncertainties in their input parameters, and difficulties in verification and interpretation, but they do provide a framework in which all existing ecological information about a system can be integrated into dynamic simulations that can be designed to address particular questions. Computer modeling, in the present case, is an economical way to gain insights into such ecological questions as, 'which species are keystones?' and 'how keys tony are they?' The results of such dynamic simulations are especially useful when they are considered as testable hypotheses, rather than presented as reality. The present search for keystones in Prince William Sound, Alaska featured a full series of simulated removal experiments from a dynamic trophic model in which all direct and indirect trophic effects can be tracked, and the predicted effects of each of these removals on all other functional groups in the system were quantified and examined with four indices, including an index of community importance (Mills et al. 1993, Power et al. 1996) and three others described in the Methods: community longevity support, interaction strength index, and a keystone index. Paine (1992) suggested that, "Significantly, none of this information [for quantifying interaction strength] can be obtained without experimental manipulation..." Yodzis (2001) concluded that the only way to begin understanding such complex ecosystem dynamics is through multi-species models. Both authors are correct. The present chapter is an example of how empirically-based whole-system dynamic food web models can be used to identify and rank keystone species from the perspective of trophic interaction strength [the attribute preferred by Power et al. (1996) over per capita interaction strength as used by Paine (1992)]. As we shall see, the results of the present quantitative modeling approach match the theoretical model outlined by Power et al. (1996), which was inspired by per capita results from field manipulations (Paine 1992, Fagan and Hurd 1994). The present approach, because of its whole-system scope and its emergent properties that are keystone-like, addresses the criticism of Hurlbert (1997) that keystone-type patterns are, "..artifacts resulting from small sample sizes and the plotting of frequency distributions on arithmetic rather than logarithmic scales." With analytical approaches such as that presented here, combined with continued empirical refinement, the keystone species metaphor might no longer have such a, ".. .stultifying effect on ecological thought and argument." As implied by Paine (1966, 1969) the abundance (or biomass) of keystone species might influence the diversity and the stability of ecosystems strongly. He found that almost half the megafaunal species vanished from the plot from which Pisaster was removed. Although there has been some healthy debate about the effects of overall species diversity from this manipulation (e.g., Lohse 1993), several other keystone examples corroborate the positive effect of keystones on diversity (Bond 1993). Paine (1969) inferred, based on his results and other examples, that the effect of removing certain keystone species on stability or ecosystem structure might be much stronger than any effect that complexity might 62 have on stability. This is a difficult question, as complexity and interaction strengths are interrelated. May (1973) developed models and analyses to evaluate the relationship of complexity and stability, not to evaluate relative interaction strengths in the sense of any search for strong interactors or keystones in a particular system. His basic analyses indicated that stability is inversely related to complexity, thereby spawning a decades-long debate. It now appears that the complexity of interaction-webs generally adds to stability because of the dynamical attributes of real systems such as adaptive foraging (e.g., Fowler and Lindstrom 2002, Kondoh 2003) as well as stabilizing aspects of metapopulation structure (Caswell 1982, Hastings 2003, 2004), refugia (Sih 1987, Gonzalez-Olivares and Ramos-Jiliberto 2003), and density dependence (Murdoch 1994). Such attributes likely have great bearing on the interaction strengths of species or groups, and they can probably either increase or decrease community effects of keystones, depending on the specific context (e.g., Power et al. 1996, Estes et al. 1998, Konar 2000, Piraino et al. 2002, Springer et al. 2003). Wooton's (1994) combination of experiments and path analysis to predict direct and indirect effects was a step toward an approach that could rank interaction strengths without having to conduct separate manipulative experiments for each potential interaction. Berlow et al. (1999) made additional progress during their testing of the interaction strength indices provided by Paine (1992) and Power et al. (1996). The present chapter is presented as a next step toward identifying keystones and other strong interactors in biological communities, in this case by employing the mass continuity modeling approach Ecopath with Ecosim to examine trophic interaction strength and other indices on the scale of the whole system, thus applying whole system biological models to the problem, as advocated by Osenberg et al. (1997). Kitchell et al. (2002) performed functional group reduction simulations using an Ecopath model of the Central North Pacific, and their simulations allowed them to evaluate the keystone role of sharks, but their study was designed primarily to evaluate the community-wide effects of pelagic longline fisheries rather than as a systematic investigation of keystones and interaction strengths. Link (2002) expressed his uncertainty in the application of food web theory (e.g., connectivity and stability proxies) to marine ecosystems when presenting a 'topological food web' of the Northeast US Shelf. The present chapter is an example of the potential usefulness of food web theory when applied to a well refined and adequately complex mass-continuity model of a marine food web (see also Chapters 2 and 6, Okey et al. 2004b, Okey et al. 2004c) and when applied to a well-suited question. Nevertheless, inherent uncertainties in the model input parameters that characterize the Prince William Sound system inevitably lead to some uncertainty in the estimates and rankings in such keystone searches. 63 14ETW 146°W Figure 4-1. Map of Prince William Sound, Alaska, located at the northern apex of the Gulf of Alaska (North Pacific Ocean). Prince William Sound covers approximately 9,000 km2, and it contains both shallow and deep habitats (to almost 800 m depth). METHODS Study location Prince William Sound, Alaska (PWS) is a 9,000 km protected sub-polar embayment located at the northern apex of the Gulf of Alaska (Figure 4-1). It is 15 times the size of San Francisco Bay, and has a highly articulated coastline and a variety of rocky, cobbley, and sedimentary habitats. It is well defined geographically by islands that separate it from the Gulf of Alaska, but there are two major channels through which flow strong tidal and coastal currents. The sound is a submerged portion of the highest coastal mountain range in the world, thus partially explaining its variety of depths, which approach 800 64 m. This topography and coastal oceanography condenses the moisture from the saturated air that is pushed from the Gulf of Alaska via southerly winds to produce the world's highest rates of precipitation. Runoff from this precipitation and from the many glaciers surrounding the sound, in addition to an incoming freshwater lens atop the Alaska coastal current, and the steep terrain contributes to a productive and unique ecology. In terms of salmon, the sound naturally produces mostly pinks (Oncorhynchus gorbuscha) due to the steep terrain and short cobbly spawning reaches. The Prince William Sound model The present search for keystones in Prince William Sound's biological community was made possible by the existence of a trophic model of the system's whole food web (Okey and Pauly 1999a, b) that was constructed using the Ecopath with Ecosim modelling approach (Christensen and Pauly 1992, Walters et al. 1997, Christensen et al. 2000). I coordinated a broad collaboration of experts on the biological community of the Sound to construct this model, which characterizes the state biomass flows throughout the PWS food web between 1994 and 1996. This chosen time period was five to seven years after the catastrophic Exxon Valdez oil spill occurred in PWS, which in turn lead to a rich array of ecological information about the Sound mostly from projects funded by the Exxon Valdez Oil Spill Restoration Program. The resulting Ecopath model of PWS included a total of 48 functional groups, but this was increased to 51 groups for the present exercise by dis-aggregating the shark group into its 3 species and by adding an 'octopods' group. These changes were an attempt to ensure representation of apex predators to increase the usefulness of a search for keystones. The PWS model, at the time of the present analysis, included 6 marine mammal groups, 3 bird groups, 21 fish groups (including 6 'forage fish' groups), 9 benthic invertebrate groups, 1 squid group, 5 zooplankton groups (including gelatinous forms), 2 phytoplankton groups, 1 macrophyte group, and 3 detritus groups. Additional structural characteristics included the splitting of several lower trophic level groups, i.e., two zooplankton groups, phytoplankton, and one detritus group, into spatially defined sub-webs delineating nearshore and offshore habitats and the splitting of each of four fish species into two ontogenetic stages (resulting in 8 separate groups). Estimated mean fishery removals for that time period, including landings data, discard estimates, and other inputs, were also specified. The PWS model has been used to explore ecosystem-based policy alternatives with the goal of balancing disparate ecosystem objectives while restoring populations of concern (Okey and Wright in press). The collaborative approach to construction and the derivations of parameters is documented in accessible formats (Okey and Pauly 1999a, b). The shark and octopod group updates and the newly refined diet composition are presented in Appendices C and D, respectively. The updated basic input parameters are shown in Table 4-1, and summary statistics characterizing the PWS model are presented in subsequent tables. The detailed modelling methodologies for both the static Ecopath approach and the dynamic Ecosim approach are described in Chapter 1. 65 Simulated removal experiments As in Chapter 3 (also published as Okey et al. 2004b), the basic approach used here to search for keystones was a full series of functional group removal simulations and a quantitative analysis of the results. This approach is developed further in the present chapter, as the search for keystones is the central focus here. Ecosim's basic gaming interface (Walters et al. 1997), which is included in Ecopath with Ecosim (Christensen et al. 2000, Pauly et al. 2000, Christensen and Walters 2004) (see Chapter 1 for detailed methods), was used to simulate the complete removal of functional groups from the food web, and repeat these removal experiments for all functional groups. Enough mortality was imposed on a functional group to eliminate it from the system by the tenth year of 30-year simulations (increased mortality was imposed at year 2). Resulting system-wide changes in biomasses were recorded at year 30. The four indices described below were then applied to the estimated changes that each and all affected living functional groups experienced (i.e., the differences between the starting baseline biomasses specified for each functional group and the predicted ending biomasses [of all but the manipulated functional group]). Mortality rates were then reset to initial baseline levels before the next simulated removal. This was repeated for all 48 living functional groups (the three detritus groups were excluded from the analyses). All simulations were conducted using a universal prey vulnerability setting of v = 0.4, which represents a mix of top-down and bottom-up forces. This setting was shown to be a reliable and conservative overall setting in sensitivity tests with the PWS model (Okey and Wright in press) for purposes of scenario comparison. This setting also makes resulting analyses comparable to other analyses (Kitchell et al. 2002, Okey et al. 2004b). The 30-year time period for simulations was chosen simply because it was considered to be of adequate duration to examine emergent changes. In many cases functional groups appeared to reach equilibrium by 30 years, while in other cases the dynamics of some functional groups appeared to maintain some transient dynamics. Indices for identifying and ranking keystones Four indices were employed for evaluating the role of each functional group in the model: community importance, community longevity support, interaction strength index, and a keystone index. Community importance was proposed and operationalized by Mills et al. (1993) and Power et al. (1996), respectively, to provide an index of the impact of a species relative to its biomass. It is expressed here in a form suited to the present analysis (Equation 4-1). 47 ( Equation 4-1 CI, ML B: 66 Where Ba is the biomass of one of the 47 affected living groups and %Bj start is the percent of the overall system biomass made up by the one affecting group before it was removed. Community longevity support (CLS) is also an index of relative impact, but it is quite different from community importance in that the changes in affected groups are weighted by the estimated longevity of each affected group before being summed (Equation 4-2). Where 1/(P/B)a start is the longevity of each affected group. Longevity is the inverse of the ratio of production to biomass (P/B), which is a basic (production rate) input parameter for each group. The use of longevity as an 'ecological weighting factor' is discussed by Okey and Wright (in press), but here it is used in the CLS index as a measure of the degree to which a species or functional group supports long-lived biomass in the overall system. The underlying assumption of this index is that long-lived biomass in a system is an integrative representor of ecosystem integrity because these organisms tend to be apex predators (with the arguable exception of baleen whales) that require the production and biomass of the underlying biotic components. That is to say, long-lived organisms require some intactness of community structure, whereas short-lived organisms do not. Long-lived organisms also tend to be the most charasmatic organisms in the general view of society, thus having the highest social value. From this perspective, maximizing overall system longevity arguably maximizes both ecological and social values, whereas a degradation of system longevity means that these values are degraded. The Community Longevity Support index is used to indicate the degree to which a species (or functional group) in the system supports overall community longevity in a given circumstance. This index provides a very different view than community importance, [trophic] interaction strength index, and the keystone index. The interaction strength index (ISI) used here is a trophic formulation of the per capita function envisioned and used by Paine (1966, 1969, 1992). The ISI measure is similar also to 'functional importance' as discussed by Hurlbert (1997). Here it really indicates the estimated 'trophic interaction strength' of the manipulated group; it is equivalent to 'absolute value of a community impact' mentioned by Power et al. (1996) (Equation 4-3). Summing the absolute value of the resulting biomass changes prevents potential underestimations of system impacts, which can occur when using the CI. Equation 4-2 47 B end Equation 4-3 67 Table 4-1. Basic parameters of the 51 compartment Ecopath model of the 1994-1996 Prince William Sound food web arranged by trophic level and presented with 'community role' indices generated from a full series of functional-group removal simulations. See text for index formulations and notes below for descriptions. Group name Trophic level Biomass (t-km2) P/B (year1) Q/B (year1) EE Cl * CLSb ISIc Keystone index d Transient orca 5.41 0.001 0.05 6.04 0.00 -317.83 -46.2 14.6 66330.8 Salmon shark 5.10 0.221 0.10 7.30 0.14 -1.02 -31.9 7.9 161.5 Resident orca 4.92 0.015 0.05 8.67 0.00 -0.12 -0.2 0.2 54.4 Sleeper shark 4.88 0.110 0.07 3.65 0.00 -0.07 -14.2 6.5 265.3 Halibut 4.52 0.677 0.32 1.73 0.81 -0.02 -2.2 1.1 7.4 Pinnipeds 4.45 0.072 0.06 25.55 0.99 -0.96 -3.5 4.7 293.4 Porpoise 4.40 0.015 0.24 29.20 0.99 -4.26 0.5 4.3 1299.7 Lingcod 4.27 0.077 0.58 3.30 0.82 -0.07 -1.9 0.4 25.9 Adult an-owtooth 4.25 4.000 0.22 3.03 0.16 -0.14 -57.6 21.1 23.9 Adult salmon 4.18 3.410 1.31 13.00 0.94 0.02 22.6 5.9 7.8 Pacific cod 4.06 0.300 1.20 4.00 0.66 0.09 5.2 1.7 26.1 Sablefish 4.03 0.293 0.57 6.42 0.96 -0.05 0.7 1.8 27.9 Juv. arrowtooth 4.01 0.855 0.22 3.03 0.95 -0.63 -53.8 20.9 110.6 Spiny dogfish 3.96 0.110 0.09 4.77 0.83 -0.24 -3.3 0.9 38.2 Avian raptors 3.92 0.002 0.05 36.50 0.00 -108.59 -27.7 12.4 28064.8 Octopods 3.80 0.050 3.10 11.70 0.95 0.50 4.3 1.6 143.9 Seabirds 3.80 0.022 0.17 150.60 0.98 -5.15 -20.2 4.2 869.7 Deep demersal fishes 3.77 0.960 0.93 3.21 0.99 -0.17 -5.0 7.0 32.8 Adult Pollock (1+1 3.76 7.480 0.71 2.56 0.97 0.00 -12.6 9.0 5.4 Rockfish 3.74 1.016 0.17 3.44 0.99 -0.01 -0.2 1.1 4.9 Baleen Whales 3.65 0.149 0.05 10.90 0.54 -0.01 -0.4 2.2 66.3 Juvenile salmon (0-1) 3.51 0.072 3.91 62.80 0.99 0.59 18.7 5.8 367.1 Nearshore demersals 3.35 4.200 1.00 4.24 0.82 0.01 19.5 10.1 10.9 Squid 3.26 3.000 3.00 15.00 0.94 -0.03 -6.1 5.8 8.8 Eulachon 3.25 1.000 2.00 18.00 0.90 0.05 10.9 3.1 14.1 Sea otter 3.23 0.045 0.13 117.00 0.01 -1.65 -3.0 8.1 812.1 Deep epibenthos 3.16 30.000 3.00 10.00 0.96 -0.03 -42.7 29.4 4.4 Capelin 3.11 0.367 3.50 18.00 0.92 -0.02 1.5 0.8 10.3 Adult herring 3.10 2.810 1.54 18.00 0.98 -0.08 10.3 13.4 21.7 Juvenile pollock (0) 3.07 0.110 2.34 16.18 0.97 -0.52 -12.8. 9.3 385.0 Invert-eating birds 3.07 0.005 0.20 450.50 0.00 -3.08 0.2 0.8 689.2 Sandlance 3.06 0.595 2.00 18.00 0.95 -0.01 1.0 0.8 6.2 Shallow lg epibenthos 3.06 3.100 2.10 10.00 0.79 -0.06 -20.6 8.4 12.3 Juvenile herring 3.03 13.406 0.73 18.00 0.97 -0.02 9.7 13.4 4.5 Jellies 2.96 6.390 5.00 29.41 0.01 -0.03 -16.4 5.9 4.2 Deep sm infauna 2.25 49.400 3.00 23.00 0.92 -0.01 -53.3 15.2 1.4 Near omnivorous zoops 2.25 0.108 7.90 26.33 0.99 0.00 0.1 0.1 3.4 Omnivorous zooplank 2.25 24.635 11.06 22.13 1.00 0.00 31.1 11.6 2.1 Shallow sm infauna 2.18 51.500 3.80 23.00 0.94 -0.03 -77.5 35.9 3.2 Meiofauna 2.11 4.475 4.50 22.50 0.95 0.00 0.0 0.6 0.6 Deep lg infauna 2.10 28.350 0.60 23.00 0.93 0.00 4.6 3.7 0.6 Shallow sm epibenthos 2.05 26.100 2.30 10.00 0.98 0.01 24.9 11.0 1.9 Shallow lg infauna 2.00 12.500 0.60 23.00 0.53 -0.01 1.4 3.7 1.4 Near herbiv zooplank 2.00 0.136 27.00 90.00 0.98 -0.03 -0.2 0.2 6.0 Herbivorous zooplank 2.00 30.000 24.00 50.00 0.98 -0.05 -133.7 40.0 6.0 Near phytoplankton 1.00 5.327 190.00 - 0.95 0.05 15.2 6.7 5.7 Offshore phytoplankton 1.00 10.672 190.00 - 0.95 -0.14 83.0 82.2 34.9 Macrophytes 1.00 125.250 4.00 - 0.13 0.00 13.2 6.5 0.2 Nekton falls 1.00 2.000 - - 0.77 - - - -Inshore detritus 1.00 19.520 - - 0.54 - - - -Offshore detritus 1.00 114.480 - - 0.59 - - - -Notes: Values in bold have been calculated with the Ecopath software; other values in the first four columns are empirically based inputs, or values that were adjusted from empirically based estimates during balancing. P/B and Q/B are the ratios of production and consumption to biomass, respectively; ecotrophic efficiency (EE) is the proportion of production not consumed or exported. a. CI {community importance (i.e., Mills et al. 1993, Power et al. 1996)) is here the sum of the opposite (sign) of the real value of the predicted change of all affected living groups (at the end of a 30-year dynamic simulation in which group i was completely removed by year 10) divided by the percent of the overall biomass of living groups represented by species i before it was removed; b. CLS {community longevity support) is the sum of the opposite (sign) of the real value of the predicted change of each affected group multiplied by its longevity (i.e., the inverse of the P/B); c. ISI {interaction strength index or trophic interaction strength) is the sum of the absolute values of the predicted change of all affected groups; d. The keystone index is the ISI divided by the percent of the system's overall living biomass represented by group i before it was removed. 68 The keystone index is the ISI expressed in terms of the relative biomass of group i before it was removed (Equation 4-4) (see also Power et al. 1996 for definition of keystone species). ISI Equation 4-4. Keystoneness = — %Bsr Frequency histograms of the rankings of all living groups were produced to qualitatively evaluate the sensitivity and apparent usefulness of each index. Groups were also ranked according to the indices, and according to trophic level, and presented in tables and figures to provide a view of the roles of various functional groups in the PWS biological community. RESULTS The basic parameters of the refined and balanced Ecopath model of Prince William Sound are shown in Table 4-1 along with the results of the full series of removal simulations in the form of calculated rankings according to the four indices employed here. The functional groups in Table 4-1 are ranked by trophic level calculated as the biomass proportion weighted average of the trophic levels of diet items plus one. Table 4-2 shows the summarized community characteristics of Prince William Sound, Alaska, according to the present iteration of the model. Table 4-3 is a quantitative characterization of the flows of biomass from primary producer and detritus sources to consumption, detritus, and respiration at each trophic level in the PWS food web. Throughput is also displayed at each trophic level. Table 4-2. Basic flows and indices describing the 1994-1996 Prince William Sound, Alaska Ecopath model Flows (Mun^year1) Calculated total net primary production 3,541 Net system production 1,022 Sum of all production 5,158 Sum of all consumption 6,705 Sum of all exports 1,087 Sum of all respiratory flows 2,519 Sum of all flows into detritus 3,749 Total system throughput 14,059 Total catches 8.68 Biomass < tonnes^km"2) Total living biomass 453 Indices Total primary production/total biomass 7.81 vear"1 Total biomass/total throughput 0.03 year"1 Total primary production/total respiration 1.4057 Proportion of flows originating from detritus 0.44 Connectance index 0.16 Mean trophic level of the catch 3.81 System omnivory index 0.21 TL units Note: Flows and biomasses are expressed in wet weight 69 Table 4-3. Flows from primary production and detritus through the Prince William Sound, Alaska model TL From primary production From detritus To To Consumed detritus Respiration Throughput Consumed Export detritus Respiration Throughput VI 1 2 2 5 1 0 1 1 3 V 5 8 14 28 3 0 4 5 12 IV 24 54 107 186 12 0 18 20 50 III 147 381 539 1068 48 0 89 89 227 II 1060 1213 683 2955. 226 0 885 1058 2169 I 2955 585 0 3541 2169 1083 0 0 3760 Sum 4192 2244 1346 7784 2459 1084 998 1173 6221 Notes: Flows are expressed in (tonneskm"2-year"'). Some flows reach trophic level VI because some organisms within some functional groups are supported by energy that has traversed six links from primary producers. Export from primary production is not shown, as it was negligible (3). The frequency distribution of community importance indicated that 46 of the 48 living functional groups in the PWS model had almost no effect on the system when they were removed entirely, while transient orcas and avian raptors had strong effects (they scored far from zero) (Figure 4-2a). In contrast, the frequency distribution of community longevity support included at least six modes of 'influence' (Figure 4-2b). The six secondary modes for community longevity support seen on this figure include at least ten strongly negatively affecting species and at least two strongly positively affecting species for this index (Figure 4-3). 40 30 >> o c a §• 20 9> 10 Transient orca Avian raptors 12 10 o c <D 3 O" O b. -350 -250 -150 -50 50 150 Community importance 250 350 -150 -100 -50 0 50 100 Community longevity support 150 Figure 4-2. Frequencies of the community importance and community longevity support scores for all functional groups in the Prince William Sound model. Community importance is the sum of the opposite (sign) of the real value of the predicted change of all affected living groups (at the end of a 30-year dynamic simulation in which group i was completely removed by year 10) divided by the percent of the overall biomass of living groups represented by species i before it was removed (see Power et al. 1996). The two modes to the left of the central mode are transient orca and avian raptors. Community longevity support is the sum of the opposite (sign) of the real value of the predicted change of each affected group multiplied by its longevity (i.e., the inverse of the P/B). By way of comparison to Figure 4-3, commercial fisheries had a very strong negative effect on community longevity (-301), while subsistence fisheries produced only a small effect (-15) and recreational fisheries had virtually no net effect (-1). This index of the effects of fisheries on the whole system also corresponded with a pinniped support index (PSI) applied to the three fisheries categories, PSI is the estimated relative support provided to pinniped biomass. Commercial fisheries produced a PSI 70 Offshore herbivor zoops Shallow small infauna Adult arrowtooth Juvenile arrowtooth Deep small infauna Transient orca Deep epibenthos Salmon shark Avian predators Shallow large epibenthos Seabirds Jellies Sleeper shark Juvenile pollock Adult pollock Squid Deep demersal fishes Pinnipeds Spiny dogfish Sea otter Halibut Ling cod Baleen whales Rockfish Resident orca Nearshore herbivor. zoops Meiofauna Nearshore ormivor. zoops Invertebrate-eating birds Porpoise Sablefish Sand lance Shallow large infauna Capelin Octopods Deep large infauna Pacific cod Juvenile herring Adult herring Eulachon Macrophytes Nearshore phytoplankton Juvenile salmon Nearshore demersals Adult salmon Shallow small epibenthos Omnivorous zooplankton Offshore phytoplankton -150 -100 -50 0 50 100 Community longevity support 150 Figure 4-3. Community longevity support index rankings for all groups in the Prince William Sound model. The dashed line is the mean of the functional group values, and is negative (-7.7) as expected in predator prey-relationships (also shown by Paine 1992 for per capita interaction strength). 71 of -143, followed by subsistence (-98), and recreational fisheries (1.0), indicating that elimination of commercial fishing in PWS would allow a 143% positive increase from the baseline rate of pinniped change over the 30 year simulation; eliminating subsistence fishing would allow a 98% positive increase from the baseline rate of change; and eliminating the recreational fishery would cause a 1% decrease from the baseline rate of change (negligible effect on present trends). The pinniped functional group in PWS is comprised of 90% harbor seals (Phoca vitulina richardsi) and 10% Steller sea lions (Eumetopias jubatus). More insight into the community longevity support index is provided when these results are plotted against the percent of the system biomass represented by each manipulated group (Figure 4-4). o Q. Q. 3 (0 > Q> O C o c • E E o o 100 50 0 -50 -100 -150 Juvenile salmon Offshore phytoplankton + Omnivorous zooplankton Adult salmon \Small epibenthos yro a Nearshore demersals^ NearDhvro Macrophytes " i JuvpollocV * A Adult pollock • ^ Sleeper shark • ^Jellies • . Seabirds A Shallow large epibenthos Avian raptors 'Salmon shark ^Deep epibenthos Transient orca Juv. arrowtoot* * • Adult arrowtooth ueep Sl™" in,auna Shallow small infauna Herbivorous zooplankton^ 0.0001 0.001 0.01 0.1 1 10 Percent of system biomass 100 Figure 4-4. Community longevity support of a species (the sum of the longevity-weighted change in biomass of affected groups) versus the percent of the system's biomass it makes up before removal. Percent of system biomass was presented on a log scale. All functional groups with a change greater than 10 were labeled. The interaction strength index and the keystone index appear at first glance to produce multiple modes, but these are displayed on log scales. Each of theses distributions thus indicates a small number of primary interactors and a larger group of secondary interactors (right side and middle of each abscissa in Figure 4-5). Still, these indices appear more sensitive to revealing the identity of potential secondary keystones than does community importance (Figure 4-2a). The top interactors and keystones that stand out on the right sides of Histograms 4-5a and 4-5b are listed from the tops of Graphs 4-6a and 4-6b, respectively. 72 I II 0 11 31 49 67 85 0 20 250 3500 50000 Interaction strength index Keystone index Figure 4-5. Frequencies of the trophic interaction strength index and keystone index scores for all functional groups in the Prince William Sound, Alaska model. Here interaction strength index is the sum of the absolute values of the predicted change of all affected groups. And keystone index is the ISI divided by the percent of the system's overall living biomass represented by group i before it was removed. Both of these indices are displayed on a log scale with the maximum value displayed near the top end of each scale. Figure 4-6 shows all functional groups in the model ranked in order of trophic interaction strength (plotted on a normal scale) and keystoneness (displayed on a log scale). The term keystoneness was used by Hurlbert (1997). The contrast between these indices is clearly shown here in how the system's functional groups are ranked differently. This figure also shows the emergence of transient killer whales and avian raptors as the clear primary keystones in this iteration of the PWS model. Plotting the keystone index against the percent of the system biomass represented by each group (both on log scales) provides the most satisfactory view of keystones in the PWS system (Figure 4-7). Here, a secondary group of keystones (labeled groups above the diagonal) is apparent in addition to the two primary keystones. This report presents the first example of a distribution of community importance values generated by a whole system trophic model, which are matched by the educated prediction of leading thinkers on the subject (Dayton 1984, Power et al. 1996), but the distributions generated using the three other indices employed here were perhaps even more interesting because of their apparently greater sensitivity to the removal experiments in terms of revealing additional species or groups with strong interaction effects on the system—a secondary cluster of strong interactors. If the present analysis of the Prince William Sound biological community is reasonably correct, then the results represent an example that supports the suggestion of Power et al. (1996) that the distribution of community importance values for species or functional groups within communities should be commonly distributed such that fluctuations of the great majority of groups has almost no community-wide effects while a small number of groups (species) has very large effects (compare Figure 4-2a in the present contribution with Figure lb in Power et al. (1996)) (see also Dayton 1984). The whole-food-web DISCUSSION 73 3 Offshore phytopl - - - . - 1 h Transient orca Herbi zooplankt I Avian predators Shal sm infauna Porpoise Deep epibenthos *< r 1 Seabirds Adult arrow tooth 1 Sea otter Juv. arrow tooth 1 Invert-eating bird Deep sm infauna 1 Juvenile pollock Transient orca 1 Juvenile salmon Rnnipeds Juvenile herring 1 Sleeper shark Adult herring 1 Salmon shark Avian predators I Ormi-zooplankto Octopods Shal smepibent ZZ3 Juv. arrow tooth Near demersals Baleen w hales Juvenile pollock Resident orca Adult pollock ZZJ Spiny dogfish Shallow Ig epibent ZD Offshore phytopl Sea otter Deep demersals Salmon shark Sablefish Deep demersals m Racific cod Near phytoplktn si Lingcod Macrophytes m Adult arrow tooth Sleeper shark ZED Adult herring Adult salmon SI Eulachon Jellies Shallow Ig epibent Juvenile salmon S3 Near demersals Squid Capelin Pinnipeds 5| Squid Porpoise 3 Adult salmon Seabirds Halibut Shat Ig infauna Sandlance Deep Ig infauna 3 Herbi zooplankt Eutachon 1 Near herbi zoops Baleen w hales 0 Near phytoplktn Sablefish J Adult pollock Pacific cod J Rockfish Octopods ] Juvenile herring Halibut 1 Deep epibenthos Rockfish 1 Jellies Spiny dogfish 1 Near omni-zoo Capelin Shal sm infauna Sandlance Ormi-zooplankto Invert-eating bird Shal smepibent Meiofauna Deep sm infauna Lingcod Shal Ig infauna Resident orca Meiofauna Near herbi zoo Deep tg infauna Near omnr-zoo Macrophytes 0 30 60 Trophic interaction strength 10 100 1000 10000 100000 Keystone index Figure 4-6. Interaction strength index and keystoneness of all living groups in the 1994-1996 Prince William Sound, Alaska Ecopath model. See Methods and Table 4-1 for explanation of indices. 74 trophic modeling approach employed here enabled the estimation of such whole-system distributions and rankings because it allowed for sequential simulated removals of all functional groups in the system for a sufficiently long time period for all (direct and indirect) trophic effects to become evident. It also allowed for analyses of collective functional group responses. Many of the functional groups, especially at high and middle trophic levels, are individual species, but those at lower trophic levels are aggregated groupings of ecologically similar species. At the broad scale that this model is designed to examine, such aggregation does not detract from the information content of the simulation outputs. The notion of keystone guilds, for example, can be as useful as the notion of guilds for understanding ecosystems and for management purposes (Power et al. 1996). Still, this model can be re-structured in terms of disaggregating particular sub-webs (sensu Paine 1963) and aggregating other sub-webs, depending on the question at hand or the proclivities of particular researchers. 0.0001 0.001 0.01 0.1 1 10 100 Percent of system biomass Figure 4-7. Keystone index ranking of a functional group (sum of the absolute change in biomasses of affected groups divided by the percent biomass of the affecting group before removal) versus the percent of the system's biomass it makes up before removal. Here, both axes are presented on log scales. The diagonal line is not x = y, but it roughly bisects the full range of system spread in this view of the system components. Groups in the upper left hand corner display the most keystoneness. Transient killer whales (Orcinus orca) emerge from the removal simulations as super-predators. Some lines of evidence suggests that these transient orcas are dietary specialists on marine mammals (Bigg 1987, Matkin and Hobbes 1999, Saulitis et al. 2000), whereas resident orcas (also Orcinus orca) consume mostly fish (see also Pauly et al. 1998b on trophic levels of marine mammals). The estimated trophic level of transient orcas in the PWS model is 5.41, and the keystone index ranked them highest by far (Table 4-1, Figure 4-7). In addition, they were ranked eighth overall by the interaction strength index (Figure 4-6); they had the most negative community importance value (Table 4-1, Figure 4-2a); and the sixth most negative community longevity support value (Figures 4-3 and 4-4). 75 If, in the past, transient orcas fed much more on baleen whales, as suggested by Estes et al. (1998) and Springer et al. (2003), then their trophic position might be almost one full level lower than currently estimated. This lower trophic level would, perhaps, make more sense in the context of the dynamic constraints (sensu Pimm 1984) and energetic constraints (Connell and Orias 1964, Paine 1966) of food webs. The intelligence and adaptiveness of orcas might prevent dynamic instability even at this high trophic level, but if orcas are the main cause of the observed pinniped (and other) declines in the North Pacific in recent decades, as postulated by Springer et al. (2003), then the current trophic position of transient orcas might well be energetically unstable. That is to say, large biomasses of baleen whales with a trophic level of approximately 3.5 could support transient orcas in an energetically stable manner, but smaller populations of pinnipeds with trophic levels of approximately 4.5 might not be able to. Considered in the simplest of trophic terms, trophic level 3.5 should support ten times more killer whales than trophic level 4.5, assuming the transfer of energy from one level to the next is 10% efficient (Lindeman 1942, Pauly and Christensen 1995). Thus, the removal of baleen whales might, in theory, leave behind almost 10 times too many orcas for pinnipeds to support (depending on aspects of diet, prey switching, and other aspects of whole food web energy allocation). If these predators are forced in this way to shift to previously untargeted prey, it is plausible that they mediate considerable ecological change (Mangel and Clark 1988, Springer et al. 2003, present contribution). Avian raptors emerge from the removal simulations as super-predators in Prince William Sound as well. This group includes Bald Eagles (Haliaeetus leucocephalus) and Peregrine Falcons (Falco peregrinus). The emergence of avian raptors group as one of the two stand-out keystones in this iteration of the PWS model is surprising from a several standpoints. First, birds are simply not widely thought of as a group that strongly structures marine ecosystems. Although they stand out in their second position on the keystone index scale, they are ranked eleventh on the trophic interaction strength scale (Figure 4-6), and ninth, behind salmon sharks, in terms of negative community longevity support. Thus, avian raptors are indicated to have moderately large influences on the system. Their high keystone index ranking indicates their potentially strong influence at higher trophic levels if their populations were higher than in the present post oil spill model. However, the role of any given species is context dependent in terms of time, space, and community structure (Power et al. 1996) and so linear extrapolations of effects should not necessarily be expected (Piraino et al. 2002). Secondly, although the strongest indicated keystone group—transient orcas—also has the highest trophic level in the system, the avian raptor group ranks fifteenth in terms of trophic level (TL = 3.92) while it is indicated to be the second strongest keystone in the system (Figures 4-5b and 4-6). This is consistent with the observations by Power et al. (1996) that keystones need not be the highest trophic level species in a system and by Piriano et al. (2002) that the identification of keystone species is not always predictable a priori. Still, avian raptors are the apex predators of their own sub-web, and the lower trophic level ranking is partially due to the omnivory and scavenging habits of Bald Eagles. The invertebrate-eating bird and seabird groups scored relatively low in 76 terms of trophic interaction strength, but they were both positioned towards the top of the secondary keystone cluster (Figures 4-6 and 4-7). Furthermore, the negative effects of seabirds on community longevity are comparable to those of Avian raptors. Invertebrate-eating birds are estimated to have no effect on community longevity. The three shark groups, Pacific sleeper sharks (Somniosus pacificus), Salmon sharks (Lamna ditropis), and Spiny dogfish (Squalus acanthias) were ranked relatively high in terms of keystoneness (10th, 11th, and 16th, respectively) and a little lower in terms of trophic interaction strength (23rd, 19*, and 40th). They were respectively ranked 13 th, 8th, and 19th in terms of community longevity support. The moderate rankings for sharks in these simulations were surprising because a series of removal simulations using the previous iteration of the PWS model indicated that the aggregated PWS shark group was a very strong interactor, ranking second to Transient orcas in terms of keystoneness. Those previous high rankings for the shark group were because the shark group accounted for a very large proportion of the mortality of Arrowtooth flounder—a dominant fish group in the model, a very strong interactor, and a moderate keystone itself. Furthermore, shark diets and consumption rates were updated with new site-specific information (e.g., Hulbert and Rice 2002), and the decision was made to use conservative values for consumption rates. For example, a daily ration of 2% for salmon sharks was taken from Nagasawa (1998), but since salmon sharks have high metabolic rates (like some pelagic sharks), their real daily rations might be closer to 5%, which would increase the estimated impact rankings for salmon sharks in the PWS model. Similarly, predation of pinnipeds by sharks is not included in the present iteration of the model, as the distribution and extent of this predation is currently highly uncertain, and such information has not been provided for any iteration of the PWS model. Some anecdotal information indicates that sleeper sharks might feed on pinnipeds in Alaska, and this would increase the scores of sleeper sharks in terms of keystoneness and interaction strength, in addition to changing the results of pinniped simulations, etc. The power of the present analytical approach is that re-assessments can be made with relative ease as such new information becomes available. Simulations by Kitchell et al. (2002) indicated that neither reducing nor increasing shark biomasses through simulated fishery policies would have much effect on the pelagic fish community of the Central North Pacific, i.e., those sharks are not a keystone group in that setting. Pelagic systems are, however, fundamentally different from nearshore and coastal systems, and from any other kind of ecosystem (Dayton 1984). The central north Pacific (from the equator to 40° N latitude) is a pelagic system characterized mainly by a suite of fast fish predators, their fish prey, and the plankton communities supporting them, whereas PWS is a sub-polar coastal system with many benthic and demersal components and is characterized by a diverse spectrum of taxonomies and life histories. The negligible effect of removing the sharks group from the Central North Pacific model was due to the slow metabolism (and low consumption rates) of sharks relative to tunas and billfishes as well as emergent 77 compensatory responses within the overall apex predator guild related to the considerable overlap of diets between sharks and the other fish predators. Sea otters were ranked 5lh overall in terms of keystoneness, towards the top of the secondary cluster (Figures 4-6 and 4-7). They were ranked 18th in terms of trophic interaction strength (Figure 4-6) and they had a very small overall negative effect on community longevity support (Figure 4-3). There are two reasons that sea otters ranked as only 'moderate keystones' (in secondary keystone cluster) in the PWS model rather than as strong keystone species. First, much of the substratum of Prince William Sound as a whole is comprised of soft bottoms, and infaunal bivalves are thus the main prey of sea otters there. The well known keystone role of sea otters in hard-bottom habitats (Estes and Palmisano 1974, Estes and Duggins 1995) hence does not manifest in simulations using the largely soft bottom PWS model; i.e., any strong keystone effect over hard bottoms is diluted by the scale and average 'habitat' of the model when run in Ecosm. An Ecopath model of a kelp forest would undoubtedly reveal the strong sea otter mediated cascade in kelp forests, and their strong keystone role there. Secondly, the sea otter population in the whole of Prince William Sound was reduced by acute mortality of the Exxon Valdez oil spill probably by somewhere between 19 and 38% (Garrott et al. 1993, Degange et al. 1994) though some estimates suggest overall PWS mortality as low as 6% (Garshelis 1997). This reduction might partially explain the moderate trophic interaction strength rankings. Nevertheless, the removal of sea otters from the Prince William Sound model does reveal a sea otter mediated trophic cascade over soft bottoms wherein the biomass of their bivalve prey increases leading to a bottom-up cascade where decreases in available phytoplankton and detritus (the bivalve's prey) leads to decreases in available zooplankton and small epifauna, and in turn decreases the biomasses of nearshore demersal fishes and their predators— lingcod. This is a strictly trophic cascade that could occur only in areas of the Sound that are not exposed to rapid onshore advection of phytoplankton and zooplankton. But the effects of sea otters transcend direct trophic effects in soft bottom habitat with at least two forms of non-trophic mediation: (1) discarded bivalve shells serve as hard substrate for the settlement of habitat-forming macroalgae, anemones, and other sessile organisms (Kvitek et al. 1992), and (2) their excavations for clams help other predators such as the asteroid Pycnopodia helianthoides capture smaller clams, as well as changing the conditions and opportunities for many other benthic organisms (Kvitek and Oliver 1986). These and other non-trophic effects are not captured in the simulations presented here, but the Ecosim does allow for specification of such non-trophic facilitation (e.g., Okey et al. 2004c) or interference. Such relationships can be specified in future attempts to identify and rank keystone species. The large effects of commercial and subsistence fisheries on community longevity (and pinnipeds, for example) seems surprising at first considering that the fisheries catch makes up small proportions of net system production and system biomass (Table 4-2). Such a comparison is, however, misleading because the amount of primary production required to produce the annual 1994-1996 PWS catch of 8.68 tonnes of biomass (at an average trophic level of 3.81) was approximately 33 times that 78 amount (283 tonnes), which is about 8% of the calculated total annual net primary production and 28% of the calculated annual net system production (after Pauly and Christensen 1995). The roles or impacts of fisheries in PWS in general are underestimated by the 1994-1996 Ecopath model since effort and catches were considerably lower than normal at this time due to the effects of the Exxon Valdez oil spill and due to the failure of the Sound's Pacific herring (species) population and fishery. Fisheries (and pinnipeds for that matter) have a larger trophic impact than a simple analysis of absolute flow of biomass to them would suggest because the organisms they consume integrate much larger percentages of the systems overall biomass and production than suggested by the summary statistics. The comparison of fishery community longevity support (CLS) values to functional group CLS values calls for clarification of this particular index. It would be a mistake to confuse CLS with ecosystem, or community, integrity. It is natural for some predators to have an overall negative effect on community longevity support as formulated here, or similar impact indices (Paine 1992). Negative CLS values simply mean that these predators cause a net decrease the standing biomass of the relatively slower-growing and longer-lived organisms in the system. Such predation is akin to a natural disturbance that one could argue increases diversity and stability (sensu Levin and Paine 1974), and thus integrity. If a fishery has a strongly negative CLS value, this is likely to be an exotic disturbance (i.e., that the organisms in question did not co-evolve with) (sensu Sousa 1984), with the exception of (strictly) traditionally prosecuted subsistence fisheries that one could perhaps make a co-evolution argument for. This strongly negative fisheries CLS value represents use of slow-growing biomass in the system that could otherwise be used by, or become, higher trophic level organisms. Aside from model structure (i.e., aggregation issues) errors in estimating basic parameters can influence a model's overall behavior in important ways. For example, J. Kitchell (pers. comm.) points out that overestimation of production/biomass (P/B) and consumption/biomass (Q/B) values are common in Ecopath models, and this can cause exaggerations in the responses of components in a system to perturbations. The PWS model is the result of a collaboration of over 30 individuals representing broad teams of scientists with expertise in the biota of the region (Okey and Pauly 1999a, b). The model underwent at least two official reviews in the context of the Exxon Valdez Oil Spill Restoration Project, as well as several unofficial reviews. C. Walters checked the model's parameters for inflated estimates of P/B, and the values for three of the 48 groups were subsequently re-visited and adjusted. Nevertheless, models are never completely accurate by definition and parameters will continue to be refined to be more accurate. Ecosystems are complex and at the edge of our abilities to fully comprehend them. Human induced impacts to an ecosystem may well result in predictable outcomes, or they may result in unpredictable and possibly detrimental consequences. Computer modeling tools like Ecopath with Ecosim help us reveal possibilities of direct and indirect mechanisms that are initially elusive. Even the most complex ecosystem models, however, are simple and limited compared with real world dynamics. The 79 best models can indicate only the possibility of real dynamics, which can then be tested and further evaluated. In light of the present approach, its inherent uncertainties, and the possibilities for big surprises in natural systems, marine resource managers might wish to consider the relative interaction strengths of species or functional groups or their relative keystoneness when developing policies that manage human activities. Finally, one major caveat is called for, following a key point made by Piraino (2002). Ranking species as ecologically special automatically implies that other species are not important. This might be profoundly erroneous and might lead to unwise strategies for management and conservation. Clearly, strong keystone species are special. What happens to them and their populations, and what they are adapted to do, has important implications for the system. But those species that did not rank highly by any index we devise with our limited analytical approaches represents unknown potential and the very substance of biodiversity. Managers and policymakers are well advised to use caution when adjusting their strategies based on these types of analyses. If history and jellyfish are any guides, today's "unimportant" species will be tomorrow's important species. CONCLUSIONS The approach and results presented here are an example of how whole food web modeling approaches such as Ecopath with Ecosim can be used to identify and rank the species or functional groups in an ecosystem in terms of their community importance, overall trophic interaction strengths, keystoneness, and community longevity support—three indices adapted from previous work and one novel index for examining roles. Transient orcas and Avian raptors both stood out as keystone predators in the present iteration of the Ecopath model of Prince William Sound. The keystone and the community longevity support indices provided more sensitivity than the community importance to examine a secondary cluster of strong interactors. LITERATURE CITED Berlow, E. L., S. A. Navarrete, C. J. Briggs, M. E. Power, and B. A. Menge. 1999. Quantifying variation in the strengths of species interactions. Ecology 80:2206-2224. Bigg, M. A. 1987. Killer whales : a study of their identification, genealogy and natural history in British Columbia and Washington State. Phantom, Nanaimo, B.C. Bond, W. J. 1993. Keystone species. Pages 237-291 in E. D. Schulze and H. A. Mooney, editors. Biodiversity and ecosystem function. Springer, New York. Carpenter, S. R., and J. F. Kitchell. 1993. The trophic cascade in lakes. Cambridge University Press, Cambridge ; New York. Caswell, H. 1982. Life-history theory and the equilibrium status of populations. American Naturalist 120:317-339. Christensen, V., and D. Pauly. 1992. Ecopath II: a software for balancing steady-state ecosystem models and calculating network characteristics. Ecological Modelling 61:169-185. Christensen, V., and C. J. Walters. 2004. Ecopath with Ecosim: Methods, capabilities and limitations. Ecological Modelling 172:109-139. 80 Christensen, V., C. J. Walters, and D. Pauly. 2000. Ecopath with Ecosim: a user's guide. Univ. of British Columbia, Fisheries Centre, Vancouver, Canada and ICLARM, Penang, Malaysia. Connell, J. H., and E. Orias. 1964. The ecological regulation of species diversity. American Naturalist 98:399-414. Dayton, P. K. 1972. Toward an understanding of community resilience and the potential effects of enrichments to the benthos at McMurdo Sound, Antarctica. Pages 81-95 in B. C. Parker, editor. Proceedings of the Colloquium on Conservation Problems in Antarctica. Allen Press, Blacksburg, Virginia, USA. Dayton, P. K. 1984. Processes structuring some marine communities: are they general? Pages 181-198 in D. R. Strong, Jr, D. Simberloff, L. G. Abele, and A. B. Thistle, editors. Ecological communities: conceptual issues and the evidence. Princeton University Press, Princeton, N.J. Dayton, P. K. 1985. The structure and regulation of some South American kelp communities. Ecological Monographs 55:447-468. Degange, A. R., A. M. Dorff, and D. H. Monson. 1994. Experimental recovery of sea otter carcasses at Kodiak Island, Alaska, following the Exxon Valdez oil spill. Marine Mammal Science 10:492-496. Estes, J. A., and D. O. Duggins. 1995. Sea otters and kelp forests in Alaska: Generality and variation in a community ecological paradigm. Ecological Monographs 65:75-100. Estes, J. A., and J. F. Palmisano. 1974. Sea otters: their role in structuring nearshore communities. Science (Washington D C) 185:1058-1060. Estes, J. A., M. T. Tinker, T. M. Williams, and D. F. Doak. 1998. Killer whale predation on sea otters linking oceanic and nearshore ecosystems. Science (Washington D C) 282:473-476. Fagan, W. F., and L. E. Hurd. 1994. Hatch density variation of a generalist arthropod predator: population consequences and community impact. Ecology 75:2022-2032. Fowler, M. S., and J. Lindstrom. 2002. Extinctions in simple and complex communities. Oikos 99:511-517. Fowler, N. 1981. Competition and coexistence in a North Carolina USA grassland 2. The effects of the experimental removal of species. Journal of Ecology 69:843-854. Garrott, R. A., L. L. Eberhardt, and D. M. Burn. 1993. Mortality of sea otters in Prince William Sound following the Exxon Valdez oil spill. Marine Mammal Science 9:343-359. Garshelis, D. L. 1997. Sea otter mortality estimated from carcasses collected after the Exxon Valdez oil spill. Conservation Biology 11:905-916. Gonzalez-Olivares, E., and R. Ramos-Jiliberto. 2003. Dynamic consequences of prey refuges in a simple model system: more prey, fewer predators and enhanced stability. Ecological Modelling 166:135-146. Hastings, A. 2003. Metapopulation persistence with age-dependent disturbance or succession. Science 301:1525-1526. Hastings, A. 2004. Transients: the key to long-term ecological understanding? Trends in Ecology & Evolution 19:39-45. Hulbert, L. B., and S. D. Rice. 2002. Salmon shark, Lamna ditropis, movements, diet, and abundance in the eastern North Pacific Ocean and Prince William Sound, Alaska. 02396, Exxon Valdez Oil Spill Restoration Project 02396 Final Report, Anchorage, AK. Hurlbert, S. H. 1997. Functional importance vs keystoneness: Reformulating some questions in theoretical biocenology. Australian Journal of Ecology 22:369-382. Kitchell, J. F., T. E. Essington, C. H. Boggs, D. E. Schindler, and C. J. Walters. 2002. The role of sharks and longline fisheries in a pelagic ecosystem of the Central Pacific. Ecosystems 5:202-216. Konar, B. 2000. Limited effects of a keystone species: trends of sea otters and kelp forests at the Semichi Islands, Alaska. Marine Ecology Progress Series 199:271-280. Kondoh, M. 2003. Foraging adaptation and the relationship between food-web complexity and stability. Science 299:1388-1391. Kvitek, R. G., and J. S. Oliver. 1986. Sea otter foraging habits and effects on prey populations and comunities in soft-bottom environments. Pages 23-47 in G. R. Van Blaricom and J. A. Estes, editors. The community ecology of sea otters. Springer-Verlag, New York. 81 Kvitek, R. G., J. S. Oliver, A. R. Degange, and B. S. Anderson. 1992. Changes in Alaskan soft-bottom prey communities along a gradient in sea otter predation. Ecology 73:413-428. Levin, S. A., and R. T. Paine. 1974. Disturbance patch formation and community structure. Proceedings of the National Academy of Sciences of the United States of America 71:2744-2747. Lindeman, R. L. 1942. The trophic-dynamic aspect of ecology. Ecology 23:399-418. Link, J. 2002. Does food web theory work for marine ecosystems? Marine Ecology-Progress Series 230:1-9. Lohse, D. P. 1993. The importance of secondary substratum in a rocky intertidal community. Journal of Experimental Marine Biology and Ecology 166:1-17. Mangel, M., and C. W. Clark. 1988. Dynamic modeling in behavioral ecology. Princeton University Press. Matkin, C, and R. Hobbes. 1999. Orcas. Pages 59-60 in T. A. Okey and D. Pauly, editors. A trophic mass balance model of Alaska's Prince William Sound ecosystem, for the post spill period 1994 1996, second edition. Fisheries Centre Research Reports 7(4), University of British Columbia, Vancouver. May, R. M. 1973. Stability and complexity in model ecosystems. Princeton University Press, Princeton, N.J. Menge, B. A. 1995. Indirect effects in marine rocky intertidal interaction webs: Patterns and importance. Ecological Monographs 65:21-74. Menge, B. A., E. L. Berlow, C. A. Blanchette, S. A. Navarrete, and S. B. Yamada. 1994. The keystone species concept: Variation in interaction strength in a rocky intertidal habitat. Ecological Monographs 64:249-286. Menge, B. A., and T. M. Farrell. 1989. Community structure and interaction webs inshallow marine hard-bottom communities: tests of an environmental stress model. Advances in Ecological Research 19:189-262. Mills, L. S., M. E. Soule, and D. F. Doak. 1993. The keystone-species concept in ecology and conservation. Bioscience 43:219-224. Murdoch, W. W. 1994. Population regulation in theory and practice - the Robert H Macarthur Award lecture presented August 1991 in San Antonio, Texas, USA. Ecology 75:271-287. Myers, R. A., and B. Worm. 2003. Rapid worldwide depletion of predatory fish communities. Nature 423:280-283. Nagasawa, K. 1998. Predation by salmon sharks (Lamna ditropus) on Pacific salmon (Oncorhynchus spp.) in the North Pacific Ocean. North Pacific Anadromous Fish Commission Bulletin 1:419-433. Navarrete, S. A., and B. A. Menge. 1996. Keystone predation and interaction strength: Interactive effects of predators on their main prey. Ecological Monographs 66:409-429. Okey, T. A., S. Banks, A. R. Born, R. H. Bustamante, M. Calvopina, G. J. Edgar, E. Espinoza, J. M. Farina, L. E. Garske, G. K. Reck, S. Salazar, S. Shepherd, V. Toral-Granda, and P. Wallem. 2004a. A trophic model of a Galapagos subtidal rocky reef for evaluating fisheries and conservation strategies. Ecological Modelling 172:383-401. Okey, T. A., and D. Pauly. 1999a. A mass-balanced model of trophic flows in Prince William Sound: de-compartmentalizing ecosystem knowledge. Pages 621-635 in Ecosystem Approaches for Fisheries Management. University of Alaska Sea Grant, AK-SG-99-01, Fairbanks. Okey, T. A., and D. Pauly, editors. 1999b. A Trophic Mass-balance Model of Alaska's Prince William Sound Ecosystem, for the Post-spill Period 1994-1996, 2nd edition. Fisheries Centre Research Report 7(4), University of British Columbia, Vancouver. Okey, T. A., G. A. Vargo, S. Mackinson, M. Vasconcellos, B. Mahmoudi, and C. A. Meyer. 2004b. Simulating community effects of sea floor shading by plankton blooms over the West Florida Shelf. Ecological Modelling 172:339-359. Okey, T. A., and B. A. Wright, in press. Toward ecosystem-based extraction policies for Prince William Sound, Alaska: integrating conflicting objectives and rebuilding pinnipeds. Bulletin of Marine Science. Osenberg, C. W., O. Sarnelle, and S. D. Cooper. 1997. Effect size in ecological experiments: The application of biological models in meta-analysis. American Naturalist 150:798-812. 82 Pace, M. L., J. J. Cole, S. R. Carpenter, and J. F. Kitchell. 1999. Trophic cascades revealed in diverse ecosystems. Trends in Ecology & Evolution 14:483-488. Paine, R. T. 1966. Food web complexity and species diversity. American Naturalist 100:65-75. Paine, R. T. 1969. A note on trophic complexity and community stability. American Naturalist 103:91-93. Paine, R. T. 1974. Intertidal community structure: experimental studies on the relationship between a dominant competitor and its principal predator. Oecologia 15:93-120. Paine, R. T. 1980. Food webs: linkage, interaction strength and community infrastructure. Journal of Animal Ecology 49:667-685. Paine, R. T. 1992. Food-web analysis through field measurements of per capita interaction strength. Nature (London) 355:73-75. Pauly, D., and V. Christensen. 1995. Primary production required to sustain global fisheries. Nature 374:255-257. Pauly, D., V. Christensen, J. Dalsgaard, R. Froese, and F. Torres. 1998a. Fishing down marine food webs. Science 279:860-863. Pauly, D., V. Christensen, and C. Walters. 2000. Ecopath, Ecosim, and Ecospace as tools for evaluating ecosystem impact of fisheries. Ices Journal of Marine Science 57:697-706. Pauly, D., A. W. Trites, E. Capuli, and V. Christensen. 1998b. Diet composition and trophic levels of marine mammals. Ices Journal of Marine Science 55:467-481. Piraino, S., G. Fanelli, and F. Boero. 2002. Variability of species' roles in marine communities: change of paradigms for conservation priorities. Marine Biology 140:1067-1074. Power, M. E., D. Tilman, J. A. Estes, B. A. Menge, W. J. Bond, L. S. Mills, G. Daily, J. C. Castilla, J. Lubchenco, and R. T. Paine. 1996. Challenges in the quest for keystones. Bioscience 46:609-620. Saulitis, E., C. Matkin, L. Barrett-Lennard, K. Heise, and G. Ellis. 2000. Foraging strategies of sympatric killer whale (Orcinus orca) populations in Prince William Sound, Alaska. Marine Mammal Science 16:94-109. Shiomoto, A., K. Tadokoro, K. Nagasawa, and Y. Ishida. 1997. Trophic relations in the subArctic North Pacific ecosystem: Possible feeding effect from pink salmon. Marine Ecology-Progress Series 150:75-85. Sih, A. 1987. Prey Refuges and Predator Prey Stability. Theoretical Population Biology 31:1-12. Sousa, W. P. 1984. The role of disturbance in natural communities. Annual Review of Ecology and Systematics 15:353-392. Springer, A. M., J. A. Estes, G. B. van Vliet, T. M. Williams, D. F. Doak, E. M. Danner, K. A. Forney, and B. Pfister. 2003. Sequential megafaunal collapse in the North Pacific Ocean: an ongoing legacy of industrial whaling? Proceedings of the National Academy of Sciences of the United States of America 100:12223-12228. Strong, D. R. 1977. Epiphyte loads, tree falls, and perennial forest disruption: a mechanism for maintaining higher tree species richness in the tropics. Journal of Biogeography 4:215-218. Walters, C, V. Christensen, and D. Pauly. 1997. Structuring dynamic models of exploited ecosystems from trophic mass-balance assessments. Reviews in Fish Biology and Fisheries 7:139-172. Wootton, J. T. 1994. Predicting direct and indirect effects: an integrated approach using experiments and path-analysis. Ecology 75:151-165. Yodzis, P. 2001. Must top predators be culled for the sake of fisheries? Trends in Ecology & Evolution 16:78-84. 83 CHAPTER 5. Can oil spills shift marine ecosystems to alternate stable states?: Preliminary simulations with an Ecopath model of Prince William Sound, Alaska ABSTRACT Some of the adverse effects of the Exxon Valdez oil spill have persisted for 15 years in Prince William Sound (PWS), Alaska. The main challenge in detecting these effects is distinguishing lasting population impacts from background environmental fluctuations and ecological transients. Nevertheless, the consistent assumption is that the overall biological community is recovering, and that sooner or later it will be recovered. An alternate possibility is that this very large scale and severe spill shifted PWS's biological community to an alternative stable state from which it will not recover. I conducted a series of dynamic simulations that use a whole-food web mass-continuity model of PWS to approach this question. I simultaneously imposed different mortality trajectories on the various biotic components of the PWS ecosystem according to the character in which each of 48 living functional groups was affected by three different whole-system disturbances: the 1964 Great Alaska Earthquake (GAE), the Exxon Valdez oil spill (EVOS) using known effects information, and the EVOS including assumed, but unverified, effects. Overall system response patterns of the three scenarios differed characteristically. The GAE impacted mostly rapid-turnover species, and the corresponding simulation indicated rapid recovery. The impacts of EVOS on slow-turnover (charismatic) species are well documented, and this second simulation indicated much longer lasting impacts and a persistent shift of some functional groups to lower biomasses. Some suspected EVOS impacts went unverified due to inherent biases of a limited, albeit broad, science program. Enhancing the 'known' EVOS scenario with such 'assumed' effects led to a third scenario that more fully targeted both slow- and fast-turnover groups. This final scenario also indicated a persistent shift in the biological community to an alternate stable state from which the system did not recover during the 50-year simulation. If this community shift is not an analytical artifact, it indicates that the character of severe disturbances, in terms of the mix of direct impacts on slow versus fast organisms, has a strong bearing on the ecological destiny of Prince William Sound, Alaska. INTRODUCTION The notion that natural systems can persist in alternate stable states (Lewontin 1969, Sutherland 1974, May 1977) has fundamentally challenged classical ideas about ecosystem responses to anthropogenic and natural disturbances, and about nature in general. It implies that some systems, in some cases, will not recover from some perturbations. More specifically, it implies hysteresis, wherein a system is resilient to exogenous stressors up to a threshold level where a catastrophic shift (usually considered degradation) occurs, but shifting the system back to the previous state would require the exogenous 84 stressors to be decreased to levels much lower than those that triggered the forward shift. Hysteresis thus implies a lag between forward and backward catastrophic shifts in a system (Chapter 1). Such theoretical discussions of alternate stable states have captured the imaginations of many ecologists and triggered a flurry of claims of alternate stable states in different ecosystems. The discussion was focused by Connell and Sousa's (1983) critical review of these examples through the application of rigorous criteria for judging ecological stability. The ensuing debate further refined the search for examples of alternate stable states in ecosystems (e.g., Peterson 1984, Sousa and Connell 1985) with some maintaining that modern ecological literature (and nature) is replete with examples of alternate stable states, even though they are not always fully demonstrated using the strictest criteria (Sutherland 1990). Renewed interest has produced a new population of candidate examples (see Chapter 1), and the evidence for some of these claims is reasonably convincing (Knowlton 1992, Petraitis and Dudgeon 1999, Carpenter 2000, Jackson et al. 2001, Scheffer et al. 2001). Putative examples of alternate stable states in marine ecosystems are growing as well (see Chapter 1), and the anemone barrens described in Chapters 2 and 3 is yet another example. Although not every one of this new collection of examples will stand up to rigorous tests for alternate stable states, they all represent shifts in marine ecosystems that appear to be persistent. Most authors on the subject agree that multiple lines of evidence should be examined, if possible, when judging the existence of alternate stable states. The present chapter is a preliminary exploration of the effects of a large-scale and severe disturbance—the 1989 Exxon Valdez oil spill (EVOS)—on the whole marine community of Prince William Sound (PWS) Alaska. Two scenarios of the effects of the EVOS and one of the Great Alaska Earthquake (GAE) were simulated to compare their predicted marine community effects. The 1989 Exxon Valdez oil spill bespoiled much of the intertidal zone of western Prince William Sound in addition to areas along the Alaskan Peninsula and parts of Kodiak Island with 40 million liters of North Slope crude oil. Numerous publications and several compendiums and overviews on the biological and ecological impacts of the EVOS have arisen from the scientific programs surrounding the natural resource damage assessment and the protracted legal controversies (e.g., Tables 5-1 and 5-2). This body of science was largely funded by the Exxon Valdez Oil Spill Trustee Council, but also by other science programs and interests. Recent broad summaries of EVOS effects on Prince William Sound and adjacent marine ecosystems distinguished three major types of ecological effects: acute, chronic, and indirect (Peterson 2001) and concluded that both exposure to residual oil and adverse effects have persisted much longer than expected, that many effects are mediated indirectly through trophic cascades, and that modem toxicology and risk-assessment should be fundamentally changed to incorporate ecosystem-based considerations (Peterson et al. 2003). In addition to these syntheses, there are several volumes of collected work on the ecological impacts of the spill (Loughlin 1994, Wells et al. 1995, Rice et al. 1996, Warheit et al. 1997) as well as other notable reviews (e.g., Paine et al. 1996, Spies et al. 1996, Rice et al. 2001) and 85 more up-to-date summaries of the status of particular species or habitats (cited in Discussion). The ecological effects of the EVOS were both broad and severe, but they varied considerably depending on the habitat, physiology, diet, and life history of each species. Accordingly, the magnitude and persistence of exposure and effects varied, as did the recorded trajectories of population change after the EVOS. Another large scale and severe disturbance was the massive (magnitude 9.2) Great Alaska Earthquake (GAE), which occurred in the spring of 1964, 25 years before the EVOS. It resulted in a number of acute ecological effects throughout the coastal zone of South Central Alaska, some of which are still conspicuous. Fortunately, some of the biological impacts of the GAE in Prince William Sound were documented (NRC 1971, Wheelwright 1994). This earthquake tilted Prince William Sound such that some intertidal communities were lowered 2 meters and others were stranded above the tides by as much as nine meters. Most areas experienced intermediate changes in topographic/bathymetric heights, but much of PWS also moved many meters horizontally. A major tsunami and many local ones swept through the Sound and adjacent regions, and submarine slides, flows, and slumps likely occurred in many soft sediment areas (NRC 1971, Okey 1997). In general, the biological impacts of the earthquake were focused on relatively fast turnover (low trophic level) organisms inhabiting both rocky and soft bottom intertidal and subtidal habitats. In addition to benthic plants and invertebrates, small fishes and salmonids that depend on intertidal spawning habitats, or low-lying stream gravels, were affected. Some adult teleost fish mortality was also observed to occur through sudden depth displacements, though the extent of impacts on fishes is uncertain (NRC 1971). The earthquake disturbance simulation was based on documented information on impacts to specific groups of organisms as well as some assumptions about additional impacts that may not have been documented. The GAE was a natural experiment of a catastrophic disturbance and it became, "the best documented and most thoroughly studied earthquake in history" (NRC 1971). Classical equilibrium theory has detrimental effects on ecosystems because the presumption of inevitable recovery gives humans justified free-reign in their interactions with nature. The central point of Holling's (1973) resilience viewpoint is an applied one, in which disturbances, or extraction of biological resources, can erode a system's resilience leading to loss of structural integrity. A system can thus settle at an alternate domain of attraction, rather than recovering to a pre-disturbance state. The default expectation of ecological recovery after catastrophic disturbances like the Exxon Valdez oil spill might have assuaged ecologists into pursuing questions of when system recovery will take place, rather than more important questions of whether recovery will take place. In reality, biological communities are shaped by the interaction of deterministic forces with random events (i.e., stochasticity) (Holling 1973). Despite the apparent erroneousness and detrimental implications of a strict equilibrium viewpoint, it is still a useful convenience for ecologists because it can be employed to distinguish between deterministic biological forces and externally-imposed physical or 86 chemical forces, which could take on the stochastic character of'historical accidents.' Deterministic analytical tools can also be used to explore the nature of ecosystem structure and function. My initial question was, 'Is it plausible that a large oil spill, such as the EVOS, could shift the PWS biological community to an alternate stable state?' This chapter, in addition to addressing this question, is an exploration of the specific characteristics of a disturbance that might be required for such a shift to occur. I note that Carpenter's (2000) discussion of how alternating regimes can be created through feedbacks between slow and fast variables in a system is particularly germane to the present exercise because, as we shall see, the scenario that strongly impacts both of these types of components in the PWS model is the only one that leads to an apparent alternate stable state. METHODS Study location The location of Prince William Sound, Alaska is shown in Figure 5-1 and its physical and ecological setting is described in Chapter 4 as well as by Okey and Pauly (1999b). The modeling approach I used the dynamic simulation modeling approach Ecosim (Walters et al. 1997) to simulate and compare the broad ecological impact of three scenarios of severe perturbations to the Prince William Sound, Alaska biological community. These included the 1964 Great Alaska Earthquake (GAE), an Exxon Valdez oil spill (EVOS) scenario using known effects information, and an EVOS scenario that included assumed (but unverified) ecological effects information in addition to known effects information. Ecosim is a bundle of dynamic simulation routines in the ecological modeling suite Ecopath with Ecosim (Polovina 1984, Christensen and Pauly 1992, Christensen et al. 2000, Pauly et al. 2000, Christensen and Walters 2004) in which a whole-system food web model is constructed to describe a snapshot of biomass flows in an ecosystem. This empirically-based Ecopath model is then a starting baseline for temporal and spatial dynamic Ecosim simulations that are conducted to explore ecological questions or policy alternatives. The basic Ecosim utility allows the exploration of the potential temporal responses of the various biotic components in the model to simulated changes in other components uses by manipulating the matrix of simultaneous differential equations taken from a given Ecopath model. Functional groups respond to such manipulations because of the thermodynamic limitations expressed by each functional group and the whole system. An investigator can thus use Ecosim to specify changes in the mortality rates imposed on selected groups, either individually or simultaneously, to examine the simulated response of those groups, other groups, or the whole food web based on the direct and indirect trophic effects of the specified manipulation. We can thus estimate indirect effects given information on direct effects. Details of the Ecopath with Ecosim modeling approach are provided in Chapter 1. 87 The Prince William Sound model The Prince William Sound Ecopath model characterizes the average annual flows of biomass throughout the food web, based on the best available information from the time period 1994 to 1996, five to seven years after the EVOS. It was constructed through a collaborative process described in Okey and Pauly (1999a), wherein over thirty-five experts on the various components of the Prince William Sound ecosystem were assembled from numerous agencies and institutions for collaboration at workshops, through e-mail and telephone communications, and by contributing to an edited volume that describes the model and documents its construction and refinement through an iterative and interactive thermodynamic balancing process and continuous review (Okey and Pauly 1999b). The model is mass-balanced, in that some of the given parameters are adjusted such that the static Ecopath model is a thermodynamically likely scenario of trophic structure and flows. Its trophic level estimates (an output) were verified using a nitrogen stable isotope method described by Kline and Pauly (1998). Recent additional changes to the PWS model are described in Chapter 4. Figure 5-1. Map of Prince William Sound (PWS), Alaska (modified from Braddock et al. 1996). 88 Earthquake and oil spill disturbance simulations The three scenarios explored in this chapter were constructed to compare the ecological effects of the 1989 Exxon Valdez Oil Spill (EVOS) with those of another severe disturbance that affected Prince William Sound—the 1964 Great Alaska Earthquake (magnitude 9.2). This earthquake caused catastrophic alterations to Prince William Sound exactly 25 years before the catastrophic EVOS (NRC 1971, Wheelwright 1994). Thus, the first two simulations were constructed based generally on documented impacts of the GAE and the EVOS; the third simulation was based on assumptions of likely, but unverified, ecological impacts of EVOS, in addition to the documented impacts. Table 5-1 lists the main sources of the effects information used for these simulations. Table 5-1. General sources of mortality estimates and the general segment of the food web affected by the three catastrophic disturbance scenarios compared. Scenario Impacted organism turnover rates Sources of specified mortality trajectories GAE EVOS known EVOS enhanced Mostly fast Mostly slow Fast and slow (NRC 1971, Wheelwright 1994) (Loughlin 1994, Wells et al. 1995, Paine et al. 1996, Rice et al. 1996, Spies et al. 1996, EVOSTC 1997, Lord 1997, EVOSTC 1999, NOAA 1999, Okey and Pauly 1999b) 'EVOS known effects' specifications plus additional assumed impacts Mortalities of PWS marine organisms associated with the GAE were focused mainly on lower trophic level organisms that generally have a fast turnover. This scenario was specified accordingly by sketching sudden (brief) pulses of increased mortality rates into the mortality input interface for only the particular groups that were directly affected and in the ways they were affected, based on the documentation of direct effects (NRC 1971). The 'EVOS known effects' scenario focused more on impacts to mid and upper trophic level organisms that generally have a slower turnover, since much of the interest and early research was focused on those segments of the ecosystem (though impacts to low and relatively low trophic level species such as intertidal organisms were also included when they were documented). The 'EVOS enhanced' scenario included all the mortality profiles from the 'EVOS known effects' scenario, as well as mortality profiles on additional biotic components that could be reasonably assumed to have been affected by such a spill without being detected. This enhanced scenario thus included impacts on both low trophic level groups (fast turnover) and middle and high trophic level groups (slower turnover). Mortalities specified for each group are shown in Table 5-2. Although mortality trajectories associated with the GAE were specified as 'pulse' effects, some of those for the EVOS were specified to have an initial pulse followed by a tapering effect representing persistent, but declining, effects. These mortality trajectories (pulses and tapers) were specified individually for each affected functional group, according to the different character of effects experienced by each group. These tapering effects were rarely specified to last longer than 15 years. Simulations were run for 50 years, and universal prey vulnerabilities (for all trophic interactions) were wet at 0.4, which 89 represents an overall mixture of top down' and 'bottom up' forces in the system, and this level is known to be reliable based on sensitivity tests that evaluated model behavior (Okey and Wright in press). The 50-yer time frame was considered to be beyond the normal time horizon of interest, but it was used here in the interest of exploring the persistence of the community shifts indicated by the EVOS scenarios. Exploratory 100-year simulations also yielded essentially the same results. Table 5-2. Specified direct mortalities imposed on each functional group presented with the basic input parameters for each group of the 51 compartment Ecopath model of the 1994-1996 Prince William Sound food web. „ Trophic Group name ^ ^ Biomass (t-km2) P/B (year1) Q/B (year1) Specified percent direct mortality GAE EVOS#l EVOS #2 Transient orca 5.41 0.001 0.05 6.04 10 Salmon shark 5.10 0.221 0.10 7.30 10 Resident orca 4.92 0.015 0.05 8.67 20 20 Sleeper shark 4.88 0.110 0.07 3.65 10 Halibut 4.52 0.677 0.32 1.73 10 Pinnipeds 4.45 0.072 0.06 25.55 20 20 Porpoise 4.40 0.015 0.24 29.20 20 Lingcod 4.27 0.077 0.58 3.30 20 Adult arrowtooth 4.25 4.000 0.22 3.03 10 Adult salmon 4.18 3.410 1.31 13.00 Pacific cod 4.06 0.300 1.20 4.00 10 Sablefish 4.03 0.293 0.57 6.42 10 Juv. arrowtooth 4.01 0.855 0.22 3.03 20 Spiny dogfish 3.96 0.110 0.09 4.77 10 Avian raptors 3.92 0.002 0.05 36.50 10 10 Octopods 3.80 0.050 3.10 11.70 10 Seabirds 3.80 0.022 0.17 150.60 20 20 Deep demersal fishes 3.77 0.960 0.93 3.21 10 Adult Pollock (1+) 3.76 7.480 0.71 2.56 10 Rockfish 3.74 1.016 0.17 3.44 20 20 20 Baleen Whales 3.65 0.149 0.05 10.90 10 Juvenile salmon (0-1) 3.51 0.072 3.91 62.80 20 20 20 Nearshore demersals 3.35 4.200 1.00 4.24 20 20 20 Squid 3.26 3.000 3.00 15.00 10 Eulachon 3.25 1.000 2.00 18.00 10 Sea otter 3.23 0.045 0.13 117.00 20 20 Deep epibenthos 3.16 30.000 3.00 10.00 10 Capelin 3.11 0.367 3.50 18.00 10 Adult herring 3.10 2.810 1.54 18.00 10 Juvenile pollock (0) 3.07 0.110 2.34 16.18 10 Invert-eating birds 3.07 0.005 0.20 450.50 20 20 Sandlance 3.06 0.595 2.00 18.00 10 Shallow lg epibenthos 3.06 3.100 2.10 10.00 50 50 50 Juvenile herring 3.03 13.406 0.73 18.00 50 50 50 Jellies 2.96 6.390 5.00 29.41 10 Deep sm infauna 2.25 49.400 3.00 23.00 10 Near omnivorous zoops 2.25 0.108 7.90 26.33 50 50 Omnivorous zooplank 2.25 24.635 11.06 22.13 10 Shallow sm infauna 2.18 51.500 3.80 23.00 50 50 50 Meiofauna 2.11 4.475 4.50 22.50 20 20 Deep lg infauna 2.10 28.350 0.60 23.00 10 Shallow sm epibenthos 2.05 26.100 2.30 10.00 50 50 50 Shallow lg infauna 2.00 12.500 0.60 23.00 20 20 20 Near herbiv zooplank 2.00 0.136 27.00 90.00 50 50 Herbivorous zooplank 2.00 30.000 24.00 50.00 20 20 Near phytoplankton 1.00 5.327 190.00 - 50 50 Offshore phytoplankton 1.00 10.672 190.00 - 10 Macrophytes 1.00 125.250 4.00 - 30 30 30 Nekton falls 1.00 2.000 - - 10 Inshore detritus 1.00 19.520 - -Offshore detritus 1.00 114.480 - -Notes: Values in bold have been calculated with the Ecopath software; other values in the first four columns are empirically based inputs, or values that were adjusted from empirically based estimates during balancing. P/B and Q/B are the ratios of production and consumption to biomass, respectively. Many of the oil spill mortalities 'shapes' represented effects that taper off over time. 90 Each of the three disturbance scenarios specified for this exercise have a signature of initial impacts that is unique trophically, and in terms of life history strategies: the GAE affected mostly fast-turnover groups, EVOS 'known effects' scenario affected a mix of slow and fast turnover groups, but it was biased towards slow-turnover groups, and EVOS 'enriched effects' scenario affected both slow and fast turnover species. The GAE scenario resulted in relatively swift recovery while the EVOS scenarios resulted in a persistent shift, at least for some groups, to lower abundance levels. RESULTS The first simulation described here suggests that the biological community of Prince William Sound, Alaska, as a whole, can recover quickly from the known effects of a large earthquake such as the 1964 Great Alaska Earthquake (GAE). Even though this earthquake triggered devastating tsunamis and changed elevations; the simulation indicated a quick rebound of the marine biota because damage was generally limited to components with fast turnover rates. In contrast, the simulation of the Exxon Valdez oil spill (EVOS) based on documented information on the direct biotic impacts of the spill suggested much slower recovery trajectories than indicated for the GAE, probably because documented EVOS impacts were focused mostly on higher trophic level organisms with slower natural turnover rates. The existence of tapering effects in the EVOS simulations did not account for the characteristic differences as indicated by the long time scale of the simulation relative to the short time scale of the tapering. The simulation of an EVOS scenario in which the direct impacts of the oil spill were enhanced beyond documented impacts resulted in even more severe effects. Both of the EVOS scenarios suggested that such an oil spill might shift the overall PWS biotic community to an 'alternate stable state' from which the system does not recover. The results of these scenarios are summarized below. The Great Alaska Earthquake scenario—The biomass of many biotic components in the PWS model were responsive to the simulated 1964 earthquake, but projected recovery of these components to a pre-earthquake state was relatively rapid (Figure 5-2a). Several groups declined immediately, including intertidal and subtidal benthic groups, some small pelagic fishes, and salmon. Other groups, such as some plankton groups and squid, were predicted to undergo synchronal increases due to declines of competitors or predators. Many components were rapidly recovering as quickly as three years after the earthquake, and in some cases more quickly. Most components had recovered to near their pre-earthquake condition 10 years after the quake, and the model system had returned to the pre-earthquake state by 30 years. These patterns are generally consistent with the patterns of recovery from the GAE found in the real PWS (NRC 1971). EVOS 'known effects' scenario—The biomass of many PWS ecosystem components responded strongly to the EVOS 'known effects' scenario (Figure 5-2b). High and relatively high trophic level organisms, as 91 well as some intertidal and shallow subtidal benthic organisms, declined as a direct result of the oil spill, while some plankton groups and squid were predicted to increase concurrently with decreases in the directly affected organisms, as in the GAE scenario. The character of this EVOS simulation departs from the GAE simulation after approximately year 10 when the biomass of several less impacted competitors (e.g., Pacific cod, sablefish, lingcod, clams) of the impacted charismatic species became elevated for approximately 20 years. A conspicuous feature of this simulation is the persistent downward shift of a number of groups that do not appear recover to their previous levels. These groups include resident killer whale (Orcinus orca), porpoise (Dall's porpoise, Phocoenoid.es dalli, and harbor porpoise, Phocoena phocoena), pinnipeds (mostly harbor seal, Phoca vitulina richardsi), avian raptors (Bald Eagles, Haliaeetus leucocephalus, and Peregrine Falcons, Falco peregrinus), invertebrate-eating birds (e.g., sea ducks), shallow large infauna (clams), Pacific halibut (Hippoglossus stenolepis). The recovery trajectories of other functional groups appear to track actual recoveries of organisms documented since the EVOS, even though these simulations were conducted on a post-spill (1994-1996) model rather than a pre-spill model. EVOS 'enhanced effects' scenario —The biomasses of a number of functional groups in the system fluctuated more strongly to the EVOS 'enhanced effects' scenario than to the 'known effects' scenario. Like the EVOS 'known effects' scenario, several groups stabilized at levels lower than their initially-specified baselines for the duration of the 50-year simulation (Figure 5-2c). These include lingcod (Ophiodon elongatus) and Pacific cod (Gadus macrocephalus) in addition to the groups listed previously. Almost all groups declined initially due to specified impacts across all trophic levels. Many of the fast turnover groups rebounded to levels far above their initial state within two years of the disturbance. The stabilization of many groups at shifted levels occurred by 15 years after the simulated spill. By the end of the 50 year simulation, these shifted groups were not tending toward recovery from their shifted states. DISCUSSION The question of whether oil spills can shift marine ecosystems to alternate stable states is a different question than whether the Exxon Valdez oil spill shifted the Prince William Sound ecosystem to an alternate stable state. The simulations presented here illustrate the possibility that alternate stable states can be produced when a disturbance severely impacts both slow and fast turnover species (sensu Mac Arthur and Wilson 1967, Pianka 1970, 1972), or at least enough of the total number of biotic components in a system. This scenario was part of a preliminary attempt to simulate the effects of the EVOS on the PWS biological community using an empirically-based model (presently the only existing whole-PWS trophic model) and informed estimations of EVOS effects trajectories on the Sound's biological components. Although they ought to be explored further by updating specified mortality trajectories with the latest effects information, these simulations are currently the best existing estimates 92 of whole PWS ecological impacts of the EVOS because they are the only existing estimates of whole PWS ecological impacts of the EVOS. In this sense, the simulations presented here support the hypotb that alternate states can exist in the real PWS, and they provide some insights into the character of disturbance mechanisms that might cause such shifts. ' ' 1 1 r 0 10 20 30 40 50 Time (years) Figure 5-2. Simulations of three catastrophic disturbances in Prince William Sound, Alaska: (a) the Great Alaska Earthquake of 1964 (magnitude 9.2), which shook and tilted Prince William Sound causing tsunamis, and which mostly impacted fast turnover organisms; (b) An Exxon Valdez oil spill scenario based on documented impacts of the spill, which focused on impacts to slow turnover organisms (and some fast turnover organisms for which information was available); and (c) An Exxon Valdez oil spill scenario that was enriched with assumed impacts of the spill that went unverified, in addition to documented impacts of the spill. Both slow and fast turnover organisn were impacted in the EVOS scenarios. 93 The question, 'Did the EVOS actually shift the system to an alternate stable state from which some components will not recover' cannot be answered with the results of these preliminary simulations. To approach this, it will be necessary to not only refine the simulations with the latest effects information (Table 5-3), but to compare the resulting trajectories of each simulation to independently-estimated trend information to validate the various answers that the simulations are providing. Such a comparison of the simulation outputs with empirical data on the recovery of organisms in PWS ecosystem can inform us about whether the system actually has shifted to a new stable state, and more generally the mechanisms and indirect effects involved in the shifting or temporary effects experienced by the system. Still, some of the trajectories appeared to track known trajectories from the GAE and the EVOS, providing a preliminary indication of the reliability of the model and dynamic approach. The trajectories of the biotic components depicted in Figure 5-2 represent the response of the model components to simulations of qualitatively different catastrophic disturbances. Based on available information, and assumptions of unavailable information about the two disturbances, the suite of effects-mortality trajectories were applied to different ranges of trophic levels (and life history strategies): the GAE of 1964 mostly impacted lower trophic levels; documentation of Exxon Valdez oil spill impacts focused on higher trophic levels; and an enriched scenario of the Exxon Valdez oil spill (including possible unverified impacts) included impacts on both fast and slow biotic components. The main purpose of this exercise is to test the effects of these three characteristically distinct catastrophes on the general response of this representation of the PWS food web. Figure 5-2, therefore, should first be viewed in terms of the general patterns and differences among the three scenarios. These scenarios resulted in an ascending gradient of impacts to the initial (post spill) model system, ranging from rapid recovery from the GAE (and associated tsunamis) to a pattern representing an 'alternate stable state' resulting from the EVOS scenarios. Results from numerous studies presented at the 10th annual conference of the EVOS science program in March of 1999 (EVOSTC 1999) indicated that only two of numerous species of concern had recovered at that time—Bald Eagles and river otters (Lutra canadensis)—and that system-level impacts of EVOS were persisting after 10 years. More recently compiled information (e.g., Peterson 2001, Peterson et al. 2003) confirms that some EVOS exposure and effects have lasted much longer than previously suspected. Such results match either of the two oil spill scenarios presented here. The inputs and results of the two oil spill scenarios should be considered, at the present juncture, to have equal validity since it is not reasonable to assume that all EVOS impacts were documented. It is notable that even the scenario based only on early documented impacts—a conservative scenario—indicates that overall system recovery from the EVOS will take longer than 50 years. It is also worth noting that 94 assumed impacts of the spill in the enriched simulations were conservative (Table 5-2), and that specifying scenarios that are more severe might be useful for further exploring system behavior. One reason for caution when interpreting the results of the simulations is the 'bounded' nature of the output. For example, the flat lines observed in the two EVOS scenarios should not imply that the proper interpretation from this simulation is that these populations (groups) would be stable in time. The mortality trajectories imposed on each scenario represent the results of changes in only the imposed forces of interest, and the resulting trajectories are the result of these imposed forces combined with the trophic forces that are based on initial biomasses and specified flow rates among groups. Other physical forces in the system, such as seasonal fluctuations and oceanographic regime shifts and cycles, are not included in these preliminary simulations, though this can now be done explicitly, and multi-year trends can be accounted for in the master equation by setting a bioaccumulation factor. By experimenting with the various combinations of the various types of forces that exist in the system, future simulations will go much further in explaining observed changes and informing us about the best policy and management strategies for achieving particular goals. The stabilized shift in these two scenarios do indicate that specified trophic forces combined with specified mortality trajectories are adequate to produce an outcome that appears as an alternate stable state in the trophic model (not including exogenous oceanographic forces) given the specified time frame. Inclusion of all forces in the ecosystem into the model would either dampen or exaggerate this putative alternate state. True alternate stable state dynamics could emerge in these simulations if the juvenile stages of previously dominant groups exhibit 'cultivation-depensation effects' (Walters et al. 2000) at the hands of their predators or competitors after the 'dominant' predator was pushed to very low biomasses by the disturbance. That is to say, a prey species whose population was previously controlled by predators could switch roles and control the predators (which then become the prey) either by consuming young life stages of the now depleted predator, or simply by virtue of abundance (e.g., Barkai and McQuaid 1988). Such effects would imply a truly hysteretic dynamic. An alternate explanation to such a 'reinforced stable state' explanation, however, is that disturbances that are as severe and broad as large oil spills can cause a considerable lag in the return of functional groups to levels associated with the previous global equilibrium state. The indicated slow recovery rates after the oil spill scenarios relative to the more rapid recovery trajectories following the earthquake scenario indicates that high trophic level organisms are important for maintaining the structure of this community. This is likely due to the strong interaction effects of apex predators (Chapter 4). Stated another way, it is because the slow-turnover components of a community can control the dynamics of the faster components (Carpenter 2000, Carpenter and Turner 2000). The results of these simulations also indicate that the relatively low resiliency of high trophic level organisms to oil spills and their prominent role in maintaining community structure makes the broader system vulnerable when these key components are severely impacted. Ecosystem integrity is thus more eroded 95 when high trophic levels are impacted than when low trophic levels are impacted. However, the greatest erosion of ecosystem integrity appears to come about through forces that cause broad impacts across multiple trophic levels (Figure 5-2c). The ultimate application of this type of approach in ecology and management relates to the extent to which the model of PWS actually represents the real PWS ecosystem, and the extent to which dynamic modeling (e.g., Ecosim) can represent real ecological processes. These questions, the answers to them, as well as the confusion surrounding them, are at the core of discomfort that ecologists might have in employing modeling approaches to real systems. The key to optimizing the usefulness of such approaches lies in the interpretation of the simulations and the model itself. Although the PWS model is one of the most explicit and refined Ecopath models constructed to date, it is an inherently simplified representation of the food web and its biomass flows. The important question is whether the modeling approach can capture the character, or the core, of these processes at the scale of interest (the whole system) thereby revealing the potential operation of mechanisms, if not the details of the magnitude of those processes. In addition, some of the functional groups in the model represent an aggregation of species, and the character of that aggregation influences both the description of the system and the behavior of the model in simulations. Aggregation, therefore, must make ecological sense, and tailoring a model's aggregation to the system and questions of interest is crucial. Table 5-3. Peer-reviewed Exxon Valdez oil spill biological effects and population trajectory information. ; Species or functional group Source Harbor seals Marine birds Pigeon Guillemots Common Murres Nearshore demersal fishes Pink salmon Black Oystercatchers Harlequin Duck Sea ducks Developing fish Sea otters Pacific herring Shallow subtidal communities Deep benthos Soft-bottom benthos Protothaca staminea Urchins and kelp Intertidal community Mussel beds Eelgrass Fucus gardineri General reviews & syntheses Compendia Frost et al. (1999); Hoover-Miller et al. (2001) Irons et al. (2000); Lance et al. (2001); Wiens et al. (2001) Golet et al. (2002) Piatt and Anderson (1996) Jewett et al. (2002) Murphy et al. (2000); Heintz et al. (2000) Murphy and Mabee (2000) Lanctot et al. (1999); Esler et al. (2002) Trust et al. (2000) Short et al. (2003) Garrott et al. (1993); Degange et al. (1994); Eberhardt and Garrott (1997); Garshelis (1997); Monson et al. (2000); Bodkin et al. (2002); Dean et al. (2002) Pearson et al. (1999); Carls et al. (2002) Dean and Jewett (2001) Feder and Blanchard (1998) Jewett et al. (1999) Fukuyama et al. (2000) Dean et al. (2000) Driskell et al. (1996); Houghton et al. (1996); Lees et al. (1996); Driskell et al. (2001); Skalski et al. (2001) Carls et al. (2001) Deanet al. (1998) Stekoll and Deysher (2000); Driskell et al. (2001) Steiner et al. (1990); Paine et al. (1996); Spies et al. (1996); EVOSTC (1997); Lord (1997); Peterson et al. (2003) Loughlin (1994); Wells et al. (1995); Rice et al. (1996); Okey and Pauly (1999b); Peterson (2001) 96 The primary contribution that this modeling approach makes to our understanding of natural systems, and the effects of human activities, is that it provides a view of the potential effects of the trophic forces in a system relative to other forces (e.g., oceanographic) that can also be incorporated into the modeling approach. Researchers can thus begin understanding the overall effects of a given set of human actions by better understanding both the direct and indirect effects of those actions and other phenomena. The spectra of indirect effects that manifest in a given community, under different conditions, are the key to understanding how communities function within the purview of human stewardship and recklessness because these indirect effects are the key mediators of most of the biological changes that humans are concerned with. As suggested by Yodzis (2001), these indirect effects cannot be revealed adequately without such whole-system modeling approaches. The usefulness of such models, again, hinges on the availability of empirical information that will improve their performance. This debate between single possible stable states vs. alternate stable states and the corollary debate of strict determinism vs. the existence of indeterminism in nature is of central importance to ecosystem-based management and conservation because of the philosophical implications of these alternative views. The existence of a single stable state, or "global attractor", implies that the states of ecosystems are fundamentally pre-determined, and their fate cannot, ultimately, be changed. From this it would follow that ecosystems cannot be degraded, or if they are, they will always return to the pre determined state. This single-stable-state would affirm transcendental fatalism and inevitability, or a strict determinism whether externally-imposed or not. In contrast, a multiplicity of stable states implies that ecosystems can be changed or degraded to states from which they might not return, and that humans have real choices among actions that will influence the destiny of ecosystems. The latter view affirms existential liberty and responsibility in nature. The adherence of society to one view over another undoubtedly influences ecosystems because fundamental philosophical frameworks, as in religions, influence human-ecosystem interactions (Nash 1968, Botkin 1990, Sullivan 1997), especially in systems that are highly exploited or exposed to a high risk of catastrophic disturbance. Leading thinkers in modern ecology express similar views with Holling (1973) when suggesting that the solution to the question of strict determinism vs. stochasticity in the organization of natural communities can be found through examining the completeness of our knowledge of natural phenomena, and the scales at which they occur. Wilson (1998) argues that even if Laplace's mechanistic determinism is conceded, nature still might as well be influenced by stochastic events or historical accidents, simply because of the impossibility of understanding all causes and effects. Bodkin (1990) is stronger in his practical rejection of strict determinism, for the similar reason that individual organisms and assemblages (including scientists and camps) experience natural events on scales of time and space that are less than the whole. Stochasticity is therefore experienced by individual organisms and communities, effectively making the phenomenon real on scales that matter. Whether complex dynamics and stochasticity result 97 entirely from deterministic chaos, or have help from irreducible indeterminacy, recognition of complex dynamics in ecological systems can lead to hypothesis development and prediction that are less biased by classical deterministic views (Pimm 1991), in addition to improved management strategies (Holling 1973. Done and Reichelt 1998). It is possible that the news media, policy professionals, resource agency representatives, the general public, and scientists have all been asking the wrong, albeit hopeful, questions: "Has the ecosystem recovered yet?" or 'When will the ecosystem recover?" These questions assume a single possible stable state, and inevitable recovery to that state. This sentiment is a static, or "equilibrium", view of nature, which is deeply imbedded in western religion and science. Perhaps a more interesting and realistic question is "Will the system recover to a pre-oil spill state," or "Has the system been pushed to an alternate state from which it will not recover?" Two interesting details of the present approach must be highlighted. First, a shifted, or alternate, state emerged during this exercise using a linear model that is based on a global equilibrium point. This is surprising and it implies that the emergent dynamics resulting from this imposed broad and severe disturbance strongly resist the natural tendency of such a model to eventually recover to its base equilibrium point, at least on the time scale examined here. This alternate state might even reflect aspects of stability due the endogenous reinforcement of the cultivation-depensation mechanism discussed previously. Second, these simulations were conducted using a post-spill model, rather than a pre-spill model, to explore the impacts of a large oil spill. Thus, the explorations of oil spill impacts described here more strictly address the effects of a second severe oil spill in PWS, which is a distinct possibility—even a likelihood. There are two obvious next steps for refining the present exercise. First, these whole system simulations should be conducted using a variety of updated effects information that has recently become available (Table 5-3). Second, the predicted trajectories of particular species and functional groups should be compared to independently estimated time series of population changes, and this fitting to observed 'recovery' patterns should be repeated for each of the three scenarios. Ecosim includes a routine for fitting simulated trajectories against independently-estimated time series of biomass estimates to verify the model's dynamic behavior and to distinguish the relative importance of the various forces acting on the populations, in this case oil, fishing, and oceanographic changes. Although not all observed 'recovery' patterns will be strictly independent of sources of effects information, such time-series fitting would provide some degree of verification of the broad-system dynamic patterns discussed here. This would be an extremely worthwhile endeavor, especially since the PWS model is rigorously constructed and maintained. Such a synthesis would, however, require support and funding that is beyond the scope of this dissertation work. 98 CONCLUSIONS A general conclusion from the preliminary simulations presented here is that the resilience of biological communities to disturbance might well depend on the character of the disturbance in terms of its breadth of impacts across trophic levels, or more precisely its breadth of impacts across life history strategies, in addition to the disturbance's physical characteristics of magnitude, frequency, and severity. These simulations represent evidence that very severe disturbances that simultaneously affect both fast and slow turnover components can shift systems to alternate stable states (sensu Carpenter 2000, Carpenter and Turner 2000). This is, however, just one type of evidence, and particular situations should be subjected to multiple lines of evidence to properly judge whether alternate stables states exist in that situation. In addition, the simulations presented in this chapter are too preliminary to adequately judge the plausibility that the Exxon Valdez oil spill shifted the Prince William Sound biotic community to an alternate stable state from which it will not recover. A more in-depth program based on a whole-system dynamic simulation approach, and including the latest effects and trends information listed herein, should be developed to address this broader question and to better understand the specific mechanisms that might help or hamper recovery of damaged biological resources. LITERATURE CITED Barkai, A., and C. McQuaid. 1988. Predator-prey role reversal in a marine benthic ecosystem. Science (Washington D C) 242:62-67. Bodkin, J. L., B. E. Ballachey, T. A. Dean, A. K. Fukuyama, S. C. Jewett, L. McDonald, D. H. Monson, C. E. O'Clair, and G. R. VanBlaricom. 2002. Sea otter population status and the process of recovery from the 1989 'Exxon Valdez' oil spill. Marine Ecology Progress Series 241:237-253. Botkin, D. B. 1990. Discordant harmonies: a new ecology for the twenty-first century. Oxford University Press, New York. Braddock, J. F., J. E. Lindstrom, T. R. Yeager, and B. T. Brown. 1996. Patterns of microbial activity in oiled and unoiled sediments in Prince William Sound. Pages 94-108 in S. D. Rice, R. B. Spies, D. A. Wolfe, and B. A. Wright, editors. Proceedings of the Exxon Valdez Oil Spill Symposium. American Fisheries Society Symposium 18, Bethesda. Carls, M. G., M. M. Babcock, P. M. Harris, G. V. Irvine, J. A. Cusick, and S. D. Rice. 2001. Persistence of oiling in mussel beds after the Exxon Valdez oil spill. Marine Environmental Research 51:167-190. Carls, M. G., G. D. Marty, and J. E. Hose. 2002. Synthesis of the toxicological impacts of the Exxon Valdez oil spill on Pacific herring (Clupea pallasi) in Prince William Sound, Alaska, USA. Canadian Journal of Fisheries and Aquatic Sciences 59:153-172. Carpenter, S. R. 2000. Alternate states of ecosystems: Evidence and its implications for environmental decisions. Pages 357-383 in M. C. Press, N. Huntley, and S. Levin, editors. Ecology: achievement and challenge. Blackwell, London. Carpenter, S. R., and M. G. Turner. 2000. Hares and tortoises: Interactions of fast and slow variables in ecosystems. Ecosystems 3:495-497. Christensen, V., and D. Pauly. 1992. Ecopath II: a software for balancing steady-state ecosystem models and calculating network characteristics. Ecological Modelling 61:169-185. Christensen, V., and C. J. Walters. 2004. Ecopath with Ecosim: methods, capabilities and limitations. Ecological Modelling 172:109-139. 99 Christensen, V., C. J. Walters, and D. Pauly. 2000. Ecopath with Ecosim: a user's guide. Univ. of British Columbia, Fisheries Centre, Vancouver, Canada and ICLARM, Penang, Malaysia. Connell, J. H., and W. P. Sousa. 1983. On the evidence needed to judge ecological stability or persistence. The American Naturalist 121:789-824. Dean, T. A., J. L. Bodkin, A. K. Fukuyama, S. C. Jewett, D. H. Monson, C. E. O'Clair, and G. R. VanBlaricom. 2002. Food limitation and the recovery of sea otters following the 'Exxon Valdez' oil spill. Marine Ecology Progress Series 241:255-270. Dean, T. A., J. L. Bodkin, S. C. Jewett, D. H. Monson, and D. Jung. 2000. Changes in sea urchins and kelp following a reduction in sea otter density as a result of the Exxon Valdez oil spill. Marine Ecology Progress Series 199:281-291. Dean, T. A., and S. C. Jewett. 2001. Habitat-specific recovery of shallow subtidal communities following the Exxon Valdez oil spill. Ecological Applications 11:1456-1471. Dean, T. A., M. S. Stekoll, S. C. Jewett, R. O. Smith, and J. E. Hose. 1998. Eelgrass (Zostera marina L.) in Prince William Sound, Alaska: Effects of the Exxon Valdez oil spill. Marine Pollution Bulletin 36:201-210. Degange, A. R., A. M. Dorff, and D. H. Monson. 1994. Experimental recovery of sea otter carcasses at Kodiak Island, Alaska, following the Exxon Valdez oil spill. Marine Mammal Science 10:492-496. Done, T. J., and R. E. Reichelt. 1998. Integrated coastal zone and fisheries ecosystem management: Generic goals and performance indices. Ecological Applications 8:S110-S118. Driskell, W., A. Fukuyama, J. Houghton, D. Lees, A. Mearns, and G. Shigenaka. 1996. Recovery of Prince William Sound intertidal infauna from Exxon Valdez oiling and shoreline treatments, 1989 through 1992. Pages 362-378 in S. D. Rice, R. B. Spies, D. A. Wolfe, and B. A. Wright, editors. Proceedings of the Exxon Valdez Oil Spill Symposium 18. American Fisheries Society, Bathesda, MD (USA). Driskell, W. B., J. L. Ruesink, D. C. Lees, J. P. Houghton, and S. C. Lindstrom. 2001. Long-term signal of disturbance: Fucus gardneri after the Exxon Valdez oil spill. Ecological Applications 11:815-827. Eberhardt, L. L., and R. A. Garrott. 1997. Sea otter mortality from the Exxon Valdez oil spill: Evaluation of an estimate from boat-based surveys - Response. Marine Mammal Science 13:351-354. Esler, D., T. D. Bowman, K. A. Trust, B. E. Ballachey, T. A. Dean, S. C. Jewett, and C. E. O'Clair. 2002. Harlequin duck population recovery following the 'Exxon Valdez' oil spill: progress, process and constraints. Marine Ecology Progress Series 241:271-286. EVOSTC. 1997. Recovery of injured resources and services: 1996 update. Pages 121-128 in J. S. Picou, D. A. Gill, and M. J. Cohen, editors. The Exxon Valdez disaster: Readings on a modern social problem. Kendall/Hunt Publishing Company, Dubuque, Iowa. EVOSTC. 1999. Status report. Exxon Valdez Oil Spill Trustee Council, Anchorage, Alaska. Feder, H. M., and A. Blanchard. 1998. The deep benthos of Prince William sound, Alaska, 16 months after the Exxon Valdez oil spill. Marine Pollution Bulletin 36:118-130. Frost, K. J., L. F. Lowry, and J. M. VerHoef. 1999. Monitoring the trend of harbor seals in Prince William Sound, Alaska, after the Exxon Valdez oil spill. Marine Mammal Science 15:494-506. Fukuyama, A. K., G. Shigenaka, and R. Z. Hoff. 2000. Effects of residual Exxon Valdez oil on intertidal Protothaca staminea: Mortality, growth, and bioaccumulation of hydrocarbons in transplanted clams. Marine Pollution Bulletin 40:1042-1050. Garrott, R. A., L. L. Eberhardt, and D. M. Burn. 1993. Mortality of sea otters in Prince William Sound following the Exxon Valdez oil spill. Marine Mammal Science 9:343-359. Garshelis, D. L. 1997. Sea otter mortality estimated from carcasses collected after the Exxon Valdez oil spill. Conservation Biology 11:905-916. Golet, G. H., P. E. Seiser, A. D. McGuire, D. D. Roby, J. B. Fischer, K. J. Kuletz, D. B. Irons, T. A. Dean, S. C. Jewett, and S. H. Newman. 2002. Long-term direct and indirect effects of the 'Exxon Valdez' oil spill on pigeon guillemots in Prince William Sound, Alaska. Marine Ecology Progress Series 241:287-304. Heintz, R. A., S. D. Rice, A. C. Wertheimer, R. F. Bradshaw, F. P. Thrower, J. E. Joyce, and J. W. Short. 2000. Delayed effects on growth and marine survival of pink salmon Oncorhynchus gorbuscha 100 after exposure to crude oil during embryonic development. Marine Ecology-Progress Series 208:205-216. Holling, C. S. 1973. Resilience and stability of ecological systems. Annual Review of Ecology and Systematics, Vol. 4. Vii+424p. Illus. Map. 4:1-23. Hoover-Miller, A., K. R. Parker, and J. J. Burns. 2001. A reassessment of the impact of the Exxon Valdez oil spill on harbor seals (Phoca vitulina richardsi) in Prince William Sound, Alaska. Marine Mammal Science 17:111-135. Houghton, J., D. Lees, W. Driskell, S. Lindstrom, and A. Mearns. 1996. Recovery of Prince William Sound intertidal epibiota from .Exxon Valdez oiling and shoreline treatments, 1989 through 1992. Pages 379-411 in S. D. Rice, R. B. Spies, D. A. Wolfe, and B. A. Wright, editors. Proceedings of the Exxon Valdez Oil Spill Symposium 18. American Fisheries Society, Bethesda, MD (USA). Irons, D. B., S. J. Kendall, W. P. Erickson, L. L. McDonald, and B. K. Lance. 2000. Nine years after the Exxon Valdez oil spill: Effects on marine bird populations in Prince William Sound, Alaska. Condor 102:723-737. Jackson, J. B. C, M. X. Kirby, W. H. Berger, K. A. Bjorndal, L. W. Botsford, B. J. Bourque, R. H. Bradbury, R. Cooke, J. Erlandson, J. A. Estes, T. P. Hughes, S. Kidwell, C. B. Lange, H. S. Lenihan, J. M. Pandolfi, C. H. Peterson, R. S. Steneck, M. J. Tegner, and R. R. Warner. 2001. Historical overfishing and the recent collapse of coastal ecosystems. Science 293:629-638. Jewett, S. C, T. A. Dean, R. O. Smith, and A. Blanchard. 1999. 'Exxon Valdez' oil spill: impacts and recovery in the soft-bottom benthic community in and adjacent to eelgrass beds. Marine Ecology Progress Series 185:59-83. Jewett, S. C, T. A. Dean, B. R. Woodin, M. K. Hoberg, and J. J. Stegeman. 2002. Exposure to hydrocarbons 10 years after the Exxon Valdez oil spill: evidence from cytochrome P4501A expression and biliary FACs in nearshore demersal fishes. Marine Environmental Research 54:21-48. Kline, T. C, and D. Pauly. 1998. Cross-validation of trophic level estimates from a mass-balance model of Prince William Sound using 15N/14N data. Pages 693-702 in F. Funk, T. J. Quinn, II, J. Heifetz, J. N. Ianelli, J. E. Powers, J. F. Schweigert, P. J. Sullivan, and C. I. Zhang, editors. Fishery stock assessment models. University of Alaska Sea Grant, Fairbanks, AK. Knowlton, N. 1992. Thresholds and multiple stable states in coral reef community dynamics. American Zoologist 32:674-682. Lance, B. K., D. B. Irons, S. J. Kendall, and L. L. McDonald. 2001. An evaluation of marine bird population trends following the Exxon valdez oil spill, Prince William Sound, Alaska. Marine Pollution Bulletin 42:298-309. Lanctot, R., B. Goatcher, K. Scribner, S. Talbot, B. Pierson, D. Esler, and D. Zwiefelhofer. 1999. Harlequin Duck recovery from the Exxon Valdez oil spill: A population genetics perspective. Auk 116:781-791. Lees, D., J. Houghton, and W. Driskell. 1996. Short-term effects of several types of shoreline treatment on rocky intertidal biota in Prince William Sound. Pages 329-348 in S. D. Rice, R. B. Spies, D. A. Wolfe, and B. A. Wright, editors. Proceedings of the Exxon Valdez Oil Spill Symposium 18. American Fisheries Society, Bathesda, MD (USA). Lord, N. 1997. Oil in the sea: Initial biological impacts of the Exxon Valdez oil spill. Pages 95-105 in J. S. Picou, D. A. Gill, and M. J. Cohen, editors. The Exxon Valdez disaster: readings on a modern social problem. Kendall/Hunt Publishing Company, Dubuque, Iowa. Loughlin, T. R., editor. 1994. Marine mammals and the Exxon Valdez. Academic Press, San Diego. MacArthur, R. H., and E. O. Wilson. 1967. The theory of island biogeography. Princeton University Press, Princeton, N.J. Monson, D. H., D. F. Doak, B. E. Ballachey, A. Johnson, and J. L. Bodkin. 2000. Long-term impacts of the Exxon Valdez oil spill on sea otters, assessed through age-dependent mortality patterns. Proceedings of the National Academy of Sciences of the United States of America 97:6562-6567. Murphy, S. M., and T. J. Mabee. 2000. Status of Black Oystercatchers in Prince William Sound, Alaska, nine years after the Exxon Valdez oil spill. Waterbirds 23:204-213. Nash, R. 1968. The American environment: Readings in the history of conservation. Addison-Wesley Pub. Co., Reading, Mass. 101 NOAA. 1999. Monitoring of biological recovery of Prince William Sound intertidal sites impacted by the Exxon Valdez oil spill. NOAA Technical Memorandum, NOS OR&R 1, National Oceanic and Atmospheric Administration. NRC. 1971. The great Alaska earthquake of 1964: Biology. Publication 1603, National Research Council, National Academy of Sciences, Washington. Okey, T. A. 1997. Sediment flushing observations, earthquake slumping, and benthic community changes in Monterey Canyon head. Continental Shelf Research 17:877-897. Okey, T. A., and D. Pauly. 1999a. A mass-balanced model of trophic flows in Prince William Sound: de-compartmentalizing ecosystem knowledge. Pages 621-635 in Ecosystem Approaches for Fisheries Management. University of Alaska Sea Grant, AK-SG-99-01, Fairbanks. Okey, T. A., and D. Pauly, editors. 1999b. A Trophic Mass-balance Model of Alaska's Prince William Sound Ecosystem, for the Post-spill Period 1994-1996, 2nd edition. Fisheries Centre Research Report 7(4), University of British Columbia, Vancouver. Okey, T. A., and B. A. Wright, in press. Toward ecosystem-based extraction policies for Prince William Sound, Alaska: integrating conflicting objectives and rebuilding pinnipeds. Bulletin of Marine Science. Paine, R. T., J. L. Ruesink, A. Sun, E. L. Soulanille, M. J. Wonham, C. D. H. Harley, D. R. Brumbaugh, and D. L. Secord. 1996. Trouble on oiled waters: Lessons from the Exxon Valdez oil spill. Annual Review of Ecology and Systematics 27:197-235. Pauly, D., V. Christensen, and C. Walters. 2000. Ecopath, Ecosim, and Ecospace as tools for evaluating ecosystem impact of fisheries. Ices Journal of Marine Science 57:697-706. Pearson, W. FL, R. A. Elston, R. W. Bienert, A. S. Drum, and L. D. Antrim. 1999. Why did the Prince William Sound, Alaska, Pacific herring (Clupea pallasi) fisheries collapse in 1993 and 1994? Review of hypotheses. Canadian Journal of Fisheries and Aquatic Sciences 56:711-737. Peterson, C. H. 1984. Does a rigorous criterion for environmental identity preclude the existence of multiple stable points? American Naturalist 124:127-133. Peterson, C. H. 2001. The "Exxon Valdez" oil spill in Alaska: Acute, indirect and chronic effects on the ecosystem. Advances in Marine Biology 39:1-103. Peterson, C. H., S. D. Rice, J. W. Short, D. Esler, J. L. Bodkin, B. E. Ballachey, and D. B. Irons. 2003. Long-term ecosystem response to the Exxon Valdez oil spill. Science 302:2082-2086. Petraitis, P. S., and S. R. Dudgeon. 1999. Experimental evidence for the origin of alternative communities on rocky intertidal shores. Oikos 84:239-245. Pianka, E. R. 1970. On r and K selection. American Naturalist 104:592-597. Pianka, E. R. 1972. R and K Selection or B and D Selection. American Naturalist 106:581-588. Piatt, J. F., and P. Anderson. 1996. Response of common murres to the Exxon Valdez oil spill and long-term changes in the Gulf of Alaska Marine ecosystem. Pages 720-737 in S. D. Rice, R. B. Spies, D. A. Wolfe, and B. A. Wright, editors. Proceedings of the Exxon Valdez oil spill symposium. American Fisheries Society Symposium 18, Bethesda. Pimm, S. L. 1991. The balance of nature? Ecological issues in the conservation of species and communities. University of Chicago Press, Chicago. Polovina, J. J. 1984. Model of a coral reef ecosystem 1. The Ecopath model and its application to French Frigate Shoals. Coral Reefs 3:1-12. Rice, S., R. Thomas, G. Carls, R. Heintz, A. Werfheimer, M. Murphy, J. Short, and A. Moles. 2001. Impacts to pink salmon following the Exxon Valdez Oil Spill: Persistence, toxicity, sensitivity, and controversy. Reviews in Fisheries Science 9:165-211. Rice, S. D., R. B. Spies, D. A. Wolfe, and B. A. Wright. 1996. Proceedings of the Exxon Valdez oil spill symposium. American Fisheries Society Symposium 18, Bethesda, Maryland. Scheffer, M., S. Carpenter, J. A. Foley, C. Folke, and B. Walker. 2001. Catastrophic shifts in ecosystems. Nature 413:591-596. Short, J. W., S. D. Rice, R. A. Heintz, M. G. Carls, and A. Moles. 2003. Long-term effects of crude oil on developing fish: Lessons from the Exxon Valdez oil spill. Energy Sources 25:509-517. Skalski, J. R., D. A. Coats, and A. K. Fukuyama. 2001. Criteria for oil spill recovery: A case study of the intertidal community of Prince William Sound, Alaska, following the Exxon Valdez oil spill. Environmental Management 28:9-18. 102 Sousa, W. P., and J. H. Connell. 1985. Further comments on the evidence for multiple stable points in natural communities. American Naturalist 125:612-615. Spies, R. FJ., S. D. Rice, D. A. Wolfe, and B. A. Wright. 1996. The effects of the Exxon Valdez oil spill on the Alaskan coastal environment. Pages 1-16 in S. D. Rice, R. B. Spies, D. A. Wolfe, and B. A. Wright, editors. Proceedings of the Exxon Valdez Oil Spill Symposium. American Fisheries Society Symposium 18, Bethesda. Steiner, R., K. Byers, and S. Keller. 1990. Lessons of the Exxon Valdez. Alaska Sea Grant College Program University of Alaska Fairbanks, Fairbanks, Alaska. Stekoll, M. S., and L. Deysher. 2000. Response of the dominant alga Fucus gardneri (silva) (Phaeophyceae) to the Exxon Valdez oil spill and clean-up. Marine Pollution Bulletin 40:1028-1041. Sullivan, L. E. 1997. Preface, in M. E. Tucker and D. R. Williams, editors. Buddhism and ecology: the interconnection of dharma and deeds. Harvard University Press, Cambridge, Massachusetts, USA. Sutherland, J. P. 1990. Perturbations, resistance, and alternative views of the existence of multiple stable points in nature. American Naturalist 136:270-275. Trust, K. A., D. Esler, B. R. Woodin, and J. J. Stegeman. 2000. Cytochrome P450 1A induction in sea ducks inhabiting nearshore areas of Prince William Sound, Alaska. Marine Pollution Bulletin 40:397-403. Walters, C, V. Christensen, and D. Pauly. 1997. Structuring dynamic models of exploited ecosystems from trophic mass-balance assessments. Reviews in Fish Biology and Fisheries 7:139-172. Walters, C, D. Pauly, V. Christensen, and J. F. Kitchell. 2000. Representing density dependent consequences of life history strategies in aquatic ecosystems: Ecosim II. Ecosystems 3:70-83. Warheit, K. I., C. S. Harrison, and G. J. Divoky, editors. 1997. Exxon Valdez oil spill seabird restoration workshop. Pacific Seabird Group, Seattle, WA,. Wells, P. G., J. N. Butler, and J. S. Hughes, editors. 1995. Exxon Valdez oil spill: Fate and effects in Alaskan waters. American Society for Testing and Materials, STP 1219, Philadelphia. Wheelwright, J. 1994. Degrees of disaster: Prince William Sound: How nature reels and rebounds. Simon and Schuster, New York. Wiens, J. A., T. O. Crist, R. H. Day, S. M. Murphy, and G. D. Hayward. 2001. A canonical correspondence analysis of the effects of the Exxon Valdez oil spill on marine birds. Ecological Applications 11:828-839. Wilson, E. O. 1998. Consilience: The unity of knowledge, 1st edition. Knopf: Distributed by Random House, New York. Yodzis, P. 2001. Must top predators be culled for the sake of fisheries? Trends in Ecology & Evolution 16:78-84. 103 CHAPTER 6. Can shading by plankton blooms cause shelf-wide community shifts? ABSTRACT Phytoplankton blooms are increasingly conspicuous along the world's coastlines, and the toxic effects of these blooms have become a major concern. Nutrient enrichment often causes phytoplankton blooms, which decrease water transparency, but little is known about the effects of such light regime changes on whole communities of the continental shelf. A series of simulations designed to evaluate the potential effects of shading by phytoplankton blooms on community organization were conducted using a balanced trophic model of the West Florida Shelf ecosystem and the Ecopath with Ecosim modelling approach. Many functional groups in the system were predicted to decline as benthic primary production was inhibited through shading by phytoplankton, especially when associated biogenic habitat was lost. Groups that obtain most of their energy from planktonic pathways increased when shading impact and associated structural habitat degradation were complemented by enhanced phytoplankton production. Groups predicted to decline as the result of shading by plankton blooms include seabirds, manatees, and a variety of demersal and benthic fishes and invertebrates. Some counterintuitive predictions of declines (mackerel, seabirds, and surface pelagics) resulted because these groups are somewhat dependent on benthic primary production. The overall effect of the simulated bloom-associated shading of benthic primary producers resembled a trophic cascade where the number of full cycles of biomass gains and losses was about five, approximately equal to the number of trophic levels in the system. INTRODUCTION Benthic primary production can be a significant portion of the overall primary production in a system, but the community effects of sea floor shading by plankton blooms is underemphasized despite an early recognition of the key role of water clarity (Verwey 1931). Eutrophication and associated reductions of light penetration has long been known to decrease the maximum depth of macrophyte colonization in lakes (Maristo 1941, Spence 1982). Indeed, irradiance is the principal environmental characteristic determining depth distributions of marine and freshwater macrophytes (Sears and Cooper 1979, Chambers and Kalff 1985) as well as their seasonal and life history characteristics (e.g., Gomez 2001). Examples of light reduction impacts on marine macrophytes are becoming common (e.g., Dennison and Alberte 1985, Johansson and Snoeijs 2002, Nielsen et al. 2002, Hauxwell et al. 2003). Irradiance is presumably as crucial for 'microphytobenthic' primary producers as it is for macrophytes (e.g., Blanchard and Montagna 1992, Daviescolley et al. 1992). Microphytobenthos consist of photosynthetic microorganisms on the sea floor and on lake and stream bottoms includeing cyanobacteria, benthic diatoms, euglenoid flagellates, and dinoflagellates (which include zooxanthellae— the photosynthetic endosymbionts of anthozoans). Enrichment of marine surface waters now appears to 104 be the main mechanism inhibiting benthic photosynthesizers in coastal systems (e.g., Tomasko and LaPointe 1994, Hall et al. 1999, Hillman et al. 1995, Meyercordt and Meyer-Reil 1999). Organic enrichment has been recognized for some time as a major cause of faunal change in marine systems (Pearson and Rosenberg 1978), but it has been recognized for much longer that the penetration of light in water controls the depth of coral reefs (Verwey 1931). Still, whole marine ecosystems have been considered somewhat immune to enrichment by nutrients and organic material due to their large size and potential for dilution. This immunity view began to lose popular credence with cases such as Kaneohe Bay, Hawai'i, where organic pollution caused profound changes to coral and algal assemblages and the system as a whole (Banner 1974, Laws and Redalje 1979, Hunter and Evans, 1995). There are now many examples of adverse effects of eutrophication and shading on coral reef systems (e.g., Loya 1976, Rogers 1979, Walker and Ormond 1982, Tomascik and Sander 1985, 1987, van Woesik et al. 1999). Only recently, however, have sunlight and water transparency been labeled as cornerstones of coral research based on findings that Florida corals exist at depths near their respiration-production compensation point (Yentsch et al. 2002). This implies that small changes in water transparency might cause broad ecological changes on continental shelves. Fundamental ecological changes triggered by enrichment-related phytoplankton blooms are now evident in large coastal systems (e.g., Johansson and Snoeijs 2002, Rabalais et al. 2002). 'Harmful algal blooms' (HABs) are common and increasing along much of the world's coastlines, and various toxicological, economic, and ecological effects are recognized. The growing concern is underscored by the recent development of broad-based scientific assessments and research programs (e.g., Boesch et al. 1997, CENR 2000, NRC 2000, Anderson et al. 2002, Conley et al. 2002). One notable program— ECOHAB—is a multidisciplinary research project designed to study the ecology and oceanography of HABs (Anderson 1995). One result of this research is the realization that aeolian subsidies of iron from Saharan dust can trigger red tide events on the West Florida Shelf by stimulating nitrogen fixation by cyanophytes (Lenes et al. 2001, Walsh and Steidinger 2001). Concerns surrounding nutrient enrichment and HABs have focused on human health (e.g., shellfish poisoning), economic impact of fisheries losses, and toxicological impact on marine life as the toxic effects propagate through the food web. One example, given by Landsberg and Steidinger (1998), is that a large bloom of the dinoflagellate Karenia brevis (formerly Gymnodinium breve) caused the deaths of large numbers (about 10%) of the remaining endangered Florida manatees (Trichechus manatus latirostris) in 1996. Benthic primary production is underemphasized in marine ecosystems because phytoplankton is the main source of primary production in the world's oceans. In coastal settings however, macrophyte primary production has been recognized as an important component of overall primary production as well as serving as structural habitat (Mann 1972, Smith 1981, Duggins et al. 1989, Hacker and Steneck 1990, Duffy and Hay 1991, Vetter 1994, 1995, Okey 1997, Vetter and Dayton 1998, 1999, Okey 2003, Epifanio et al. 2003, Adams et al. 2004). A particularly underemphasized component of benthic primary 105 production is that accounted for by microphytobenthos. Recent work has demonstrated that this inconspicuous component can contribute a considerable portion of overall continental shelf primary production (Colijn and de Jonge 1984, Cahoon and Cooke 1992, Maclntyre and Cullen 1995, Maclntyre et al. 1996, Nelson et al. 1999). High rates of marine microphytobenthic primary production have been recognized for 30 years (e.g., Bunt et al. 1972, Sournia 1976, Hartwig 1978). High production of microphytobenthos occurs on tropical and subtropical shelves where overlying water is relatively clear. The rate of primary production by microphytobenthos, as well as its biomass, is strongly limited by the light reaching the sediment (Hartwig 1978). This light limitation is expressed in nature as a declining gradient of microphytobenthos production with increasing depth off Madagascar, for example (Plante-Cuny 1973 in Colijn and de Jonge 1984), but also on the West Florida Shelf (Vargo, unpublished data) where these primary producers likely occur below the 200 m depth contour. Tropical and subtropical 'clear water' shelf systems might in fact be particularly susceptible to the impacts of pollution that causes increased shading because sea floor primary production in these shelf areas can comprise a considerable portion of overall primary production, and because declines in the water transparency of these naturally oligotrophic systems might fundamentally shift patterns of energy flow as light fades below the minimum necessary thresholds of acclimated benthic primary producers. Many corals and foraminiferans rely on symbiotic zooxanfhellae to provide them with photosynthetic energy that is integral to the production and maintenance of tropical marine habitats and biodiversity. These organisms and habitats are degraded (or changed) by transparency declines related to nutrient enrichment (Hallock and Schlager 1986, Hallock 1987, 1988, Hallock et al. 1993, Hallock 2000). Microphytobenthic production is known to decrease when the biomass of overshading plankton increases, even though microphytobenthos can persist at low light levels (Cahoon and Cooke 1992, Meyercordt and Meyer-Reil 1999). This spatial and temporal primary production compensation by microphytobenthos should distribute primary production evenly across horizontal space and time, or at least moderate its overall variability. Similar compensation patterns are now emerging in other systems (e.g., Blanchard and Montagna 1992, Clavier and Garrigue 1999). Both horizontal and vertical distributions of micro-photosynthesizers within sub-systems undoubtedly mediate the character of secondary and higher level production. As phytoplankton increases across a shelf, microphytobenthos tends to decrease. Many organisms living over continental shelves utilize the primary production of microphytobenthos directly (Nelson et al. 1999) or indirectly. Shading by nutrient enrichment along coastal zones can, in theory then, shift communities from assemblages that rely on benthic primary production towards those that rely on planktonic primary production. This is consistent with simulations by Blanchard and Montagna (1995), which indicated that shading by phytoplankton blooms in Baffin Bay, Texas could profoundly decrease the production rate of microphytobenthos and the abundance of the benthic macrofauna supported by them. This shift along a benthos-based to plankton-based continuum 106 could have far reaching implications for community organization and benthic-pelagic coupling since a whole suite of organisms rely directly or indirectly on microphytobenthos. Continental shelves cover about 8% of the world's ocean area, but produce almost one quarter of the world's plankton (Pauly and Christensen 1995). Marine macrophytes on continental shelves add an additional 5% to the oceans' primary production, and make up fully two thirds of the oceans' plant biomass, despite inhabiting about one quarter of one percent of the total area inhabited by plankton (Smith 1981). Benthic microflora on continental shelves add even more to oceanic primary production. Shading related shifts in the assemblages of these continental shelf primary producers and dependent biological communities are especially important to humans because, for example, fisheries are concentrated near coasts. Of even greater concern than such pollution-induced shifts is the distinct possibility that such shifts might lead to degraded systems that are stable (i.e., do not recover). Not only could eutrophication and shading lead to large changes in flora and fauna, but certain changes in fauna (i.e., removal of filter feeders) could exacerbate cultural eutrophication in marine systems (Lenihan and Peterson 1998, Lenihan 1999, Jackson et al. 2001). Thus, positive feedbacks between nutrient enrichment and filter feeder degradation could potentially lead to 'alternate stability domains' in these marine systems (see Scheffer et al. 2001). In addition, Smith (1981) plausibly suggested that a good portion of the 910s tonnes of carbon captured annually by marine macrophytes could be entrained in marine carbon sinks because the fate of much of this material is unused detritus. Decreases in macrophyte production resulting from increases in shading by plankton could slow the flux to such a sink, thus slowing oceanic absorption of atmospheric carbon. The West Florida Shelf The West Florida continental shelf covers over 170 000 km2 extending more than 200 km west from the intertidal zone to the 200 m isobath across a very gentle slope of ancient limestone platforms (slope «1°). These platforms are overlain with a veneer of old carbonate sediment and more recent riverine sediment. To the north, the shelf extends to DeSoto Submarine Canyon near the Alabama-Florida border, and to the south the delineation is drawn at the transition to the Florida Keys. The nutrient content and optical properties of West Florida Shelf waters are strongly influenced by input from the Mississippi, Mobile, Apalachicola, Suwannee, and Caloosahatchee rivers (Muller-Karger et al. 1991, Gilbes et al. 1996, Del Castillo et al. 2000). Most Gulf waters are naturally poor in nutrients except adjacent to rivers and estuaries or when river flow is high, and except for seasonal patterns of upwelling and impingement of deeper nutrient-rich water from beyond the shelf break related to Loop Current frontal eddy intrusion and other forces (Paluszkiewicz et al 1983, Weisberg et al. 1996, Weisberg et al. 2000, Meyers 2001). Enrichment by riverine, estuarine, and upwelled nutrients is episodically complemented by aeolian deposits of iron from Sahara dust. 107 Florida Bay lies at the inner edge of the southwest Florida shelf. Documented recent events in the Bay provide clues to the types of changes that might be occurring on the broader shelf. Since the middle to late 1980s, Florida Bay has undergone massive die-offs of several species of seagrasses. A reasonable explanation for this die-off is that increased nutrient inputs led to shading of these plants by macroalgal epiphytes (Fong and Harwell 1994, Lapointe et al. 1994) and blooms of phytoplankton (Tomasko and LaPointe 1994, Boyer et al. 1999, Fourqurean and Robblee 1999, Hall et al. 1999). Phytoplankton blooms can increase water-column temperatures by absorbing solar radiation (Tomasko and LaPointe 1994). Changes in salinity from freshwater diversion can influence the oxygen saturation in the water column (Gunderson 2001). Furthermore, the die-off of seagrasses can increase sediment re-suspension in settings like Florida Bay, thereby causing more shading and seagrass die-off (Hall et al. 1999). Such a positive feedback can help explain the apparent shifts in stability domains in Florida Bay, in addition to the explanations presented by Gunderson (2001). Still another explanation that might have worked in concert with such scenarios is that sea turtle depletion led to outbreaks of epiphytes and diseases that kill seagrasses because naturally high densities of sea turtles cropped seagrass blades that would otherwise become susceptible (Jackson et al. 2001). Whatever the exact mechanisms, the ichthyofauna of Florida Bay appears to have undergone a shift from benthic species toward more planktonic-feeding species during the same period (Thayer et al. 1999), and sponges also died in parts of the area (Butler et al. 1995). Similar changes have been observed in seagrass areas of Western Australia for apparently similar reasons (Hillman et al. 1995). Livingston (2001) has described the mechanisms behind such processes in Gulf Coast estuaries. The widespread loss of seagrasses related to shading and related factors is a very conspicuous ecological change in Florida Bay, but analogous though less conspicuous changes might have occurred in deeper zones of the West Florida Shelf. An analysis of imagery from Advanced Very High Resolution Radiometry showed that reflectance and light attenuation increased between 1985 and 1997 in an area of the shelf west of Florida Bay (Stumpf et al. 1999). The present paper combines a new exploration of trends in water transparency over the West Florida Shelf with some preliminary dynamic simulations of broader community effects of shading using an up-to-date trophic model of the West Florida Shelf (Ecopath with Ecosim). Browder (1993) called for refinement of information about the Gulf of Mexico continental shelf after presenting a pilot model of the shelf using the Ecopath approach. Her preliminary modelling exercise indicated the existence of more benthos and higher benthic production than previously thought. The present exercise can be seen as following up on this, presenting results of a more recent synthesis of West Florida Shelf ecosystem information. The general questions addressed in this study were: (1) Are there multiyear trends in water transparency over the West Florida Shelf? (2) What proportion of the overall primary production on the West Florida Shelf is made up by microphytobenthos? (3) What broad community effects might result from nutrient enrichment and phytoplankton blooms? The preliminary simulations described below indicated that broad community 108 effects should be expected from a mechanism that decreases benthic primary production, such as shading by phytoplankton blooms. METHODS This investigation consisted of two components: (1) exploration of trends in water transparency over the West Florida Shelf, and (2) simulations of community-wide effects of seafloor shading by phytoplankton blooms using a recently constructed balanced trophic model of the West Florida Shelf. Trends in water transparency Changes in water transparency were investigated with two approaches. The first was a review of phytoplankton production estimates from the vicinity of the West Florida Shelf. The second approach was to examine trends in water quality parameters in the Southeast Area Monitoring and Assessment Program (SEAMAP) data for a chosen portion of the West Florida Shelf. The examined area extends between 24° to 26° N and between 83° to 85° W, and is located just to the northwest of the Dry Tortugas. It was named the 'Hemingway Quadrant,' and was chosen because it encompassed an area of the seafloor covered by living coralline algal nodules (e.g., Peyssonnelia rubra and P. simulans; Phillips and Thompson 1990) that can be used in the future as a proxy to document changes in shelf photosynthetic communities. The water quality data within this area were extracted from SEAMAP ichthyoplankton survey data collected during cruises in the Gulf of Mexico by the Florida Fish and Wildlife Conservation Commission (data held by C. Meyer and B. Mahmoudi, Florida Marine Research institute, St. Petersburg). Temporal trends of chlorophyll at three depths, turbidity at three depths, and Secchi depth at the surface were analyzed with simple linear regression. The West Florida Shelf model The Ecopath model of the West Florida Shelf was constructed at the Florida Marine Research Institute during 2000 as an initiative by the Florida Fish and Wildlife Conservation Commission to synthesize existing ecosystem information in a format that would allow simulations of ecosystem dynamics related to living resource management along the Gulf coast of Florida (Mahmoudi et al. 2002). This model was constructed by combining extensive literature reviews of local, regional, and global information with a coordinated collaboration of marine scientists with expertise in West Florida Shelf biota. The basic parameters for the 59 functional groups in the West Florida Shelf Ecopath model are shown in Table 6-1. Diet compositions and full documentation of sources for all parameter estimates are available from the Florida Marine Research Institute (Okey and Mahmoudi 2002). A slightly updated diet composition is presented here in Appendix E. Benthic primary producers were estimated to make up 35% of the overall primary production and 91%) of the primary producer biomass on the West Florida Shelf, with microphytobenthos making up 10% 109 and 11% of the respective totals (though the biomass and production of macroalgae might be underestimated in this analysis). Simulations Seven scenarios were developed to investigate community-wide effects of shading interference of benthic primary producers by phytoplankton blooms: 1. Shade microphytobenthos; 2. Shade macroalgae; 3. Shade seagrasses; 4. Shade all three benthic primary producer groups; 5. Shade all three and include prey protection effects of biogenic structure; 6. Shade all three and include enrichment of phytoplankton production; 7. Shade all three; include refuge effects; include phytoplankton enrichment. Shading mortality was evenly increased from 0 to 50% of the production rate (P/B) for each of the three benthic producer groups between 2.5 and 10 years after the beginning of of 30-year simulations. Scenario 5 added refuge effects of biogenic structure to the simulated shading of all three benthic primary producers. Seagrasses, macroalgae, drifting macrophytes, and sessile epibenthic fauna benefit many prey organisms through refuge effects. Small demersal fishes, for example, are less vulnerable to predators when they associate with this biogenic habitat structure, which is why they are found in higher abundances in such refugia (Holmquist 1994, Levin and Hay 1996). Ecosim allows specification of such a protective 'mediation,' such that prey organisms become more vulnerable to predators when their protective biogenic habitat declines. This is achieved by modifying their relationships within Ecosim by choosing or sketching a 'shape' of the relationship in the mediation interface representing how specified trophic interactions are modified by changes in the biomass of a 'mediating' group. This protective effect was specified to be qualitatively equal for the four biogenic structures listed above, and the same mediating relationship (biogenic habitat protection) was applied to the prey vulnerabilities for a variety of demersal and structure-associated fishes and other organisms in the system. This relationship was specified with a negative sigmoid function that was automatically scaled relative to Ecopath baseline inputs (see Christensen et al. 2000). Scenario 6 combines simulated shading of the three benthic primary producers with concomitant increases in surface phytoplankton at the approximate rate indicated by revealed trends in SEAMAP data (discussed later). Scenario 7 resembles bloom related shading best in that it combines overall shading with both refuge effects (loss of refuge) and phytoplankton enhancement. The system-wide prey vulnerability was set at 0.4 for all scenarios representing a mixture of top-down and bottom-up forces shaping the community. 110 Table 6-1. Basic parameters of the Ecopath model of the West Florida continental shelf. OI is the ornnivory index, which is the variance of prey trophic levels; P/B and Q/B are the ratios of production (P) and consumption (Q) to biomass; EE is the ecotrophic efficiency, or the proportion of production consumed by predators or exported. Values in bold have been calculated by the Ecopath algorithm; other values are empirically based inputs, or values that were adjusted from empirically based values during balancing. Documentation of the derivations of these estimates is available in Okey and Mahmoudi (2002). Group name Trophic level OI Biomass (t-km2) P/B (year1) Q/B (year1) EE Large oceanic piscivores . 4.7 0.313 0.070 0.680 7.400 0.845 Pelagic oceanic piscivores 4.5 0.266 0.150 1.057 8.500 0.829 Dolphins 4.4 0.283 0.038 0.099 40.439 0.082 Coastal sharks 4.3 0.732 0.090 0.410 3.290 0.909 Large groupers 4.3 0.252 0.119 0.458 4.103 0.880 Pelagic coastal piscivores 4.3 0.240 0.230 0.640 10.230 0.972 Benthic oceanic piscivores 4.2 0.331 0.045 0.450 7.940 0.961 Mackerel (adult) 4.2 0.097 0.183 0.384 8.000 0.938 Mackerel (juvenile) 4.2 0.051 0.126 0.769 9.000 0.970 Nearshore associated piscivores 4.2 0.263 0.013 1.057 7.670 0.900 Seabirds 4.2 0.289 0.001 0.100 80.000 0.000 Pelagic oceanic jelly eaters 4.1 0.045 2.200 1.560 8.071 0.674 Structure associated coastal piscivores 4.1 0.271 0.220 0.630 5.400 0.736 Benthic coastal piscivores 4.0 0.304 0.245 0.550 8.386 0.938 Demersal coastal piscivores 4.0 0.903 0.120 0.642 6.334 0.977 Squid 3.8 0.184 1.100 3.000 35.000 0.987 Large oceanic planktivores 3.7 0.305 0.043 0.110 1.800 0.500 Rays and skates 3.7 0.340 0.238 0.380 7.720 0.651 Octopods 3.6 0.400 0.074 3.100 11.700 0.950 Benthic coastal invertebrate eaters 3.5 0.129 0.860 0.860 10.110 0.991 Benthic oceanic invertebrate eaters 3.5 0.203 0.190 1.200 15.780 0.988 Demersal coastal invertebrate eaters 3.5 0.223 1.400 0.654 7.920 0.999 Structure associated coastal invertebrate eaters 3.5 0.169 1.200 0.748 7.330 1.000 Structure associated coastal planktovores 3.5 0.081 0.050 2.000 10.000 0.851 Carnivorous jellyfish 3.4 0.091 0.265 40.000 80.000 0.928 Demersal oceanic invertebrate eaters 3.4 0.069 0.045 1.200 15.760 0.971 Lobsters 3.4 0.246 0.028 0.900 8.200 0.858 Other fishes 3.4 0.225 3.870 1.300 7.040 0.950 Pelagic oceanic planktivores 3.4 0.509 1.500 0.872 11.710 0.949 Stomatopods 3.3 0.469 0.994 1.335 7.432 0.414 Turtles 3.3 0.639 0.007 0.192 3.500 0.417 Nearshore planktivores 3.2 0.262 2.215 2.000 15.920 0.990 Large crabs 3.1 0.189 0.705 2.800 8.500 0.990 Sardine and herring 3.1 0.471 2.400 1.050 12.106 1.000 Carnivorous zooplankton 3.0 0.171 21.600 8.700 20.000 0.250 Adult shrimps 2.9 0.443 0.550 5.380 19.200 0.987 Demersal coastal omnivores 2.9 0.490 0.700 1.340 15.130 0.784 Ichthyoplankton 2.9 0.427 0.048 50.448 132.130 0.748 Surface pelagics 2.9 0.859 0.099 2.600 11.700 0.950 Other mesozooplankton 2.6 0.277 6.700 17.300 50.000 0.851 Structure assoc. coastal omnivores 2.5 0.466 0.312 1.329 24.370 0.980 Echinoderms 2.4 0.347 19.246 1.200 3.700 0.277 Meiofauna 2.4 0.236 13.000 12.500 25.000 0.822 Sessile epibenthos 2.4 0.273 219.000 0.800 9.000 0.236 Small mobile epifauna 2.4 0.284 12.614 7.010 27.140 0.950 Small infauna 2.3 0.273 19.032 4.600 15.900 0.401 Small copepods 2.2 0.133 8.300 17.300 50.000 0.939 Bivalves 2.1 0.106 48.596 1.209 23.000 0.168 Mullets 2.1 0.101 0.329 0.701 11.030 0.512 Manatees 2.0 0.000 0.001 0.100 36.500 0.000 Microbial heterotrophs 2.0 0.000 60.000 100.000 215.000 0.235 Dead carcasses 1.0 0.421 1.000 - - 0.906 Drift macrophytes 1.0 0.000 2.659 - - 0.324 Macroalgae 1.0 0.000 36.050 4.000 - 0.396 Microphytobenthos 1.0 0.000 29.778 23.725 - 0.623 Phytoplankton 1.0 0.000 25.000 182.130 - 0.304 Seagrasses 1.0 0.000 175.617 9.014 - 0.017 Sediment detritus 1.0 0.274 390.000 - - 0.884 Watercolumn detritus 1.0 0.347 125.000 - - 0.910 111 RESULTS A general trend of increasing phytoplankton production from the early 1970s to the early 1990s in the vicinity of the West Florida Shelf emerged during a literature survey (Figure 6-1). This was consistent with the finding of significant increases in surface and middle-depth chlorophyll from the early 1980s to the late 1990s, significant increases in surface turbidity from 1993 to 1999, and the trend of declining surface clarity (Secchi depths) from 1983 to 1997 in SEAMAP data from the 'Hemingway Quadrant' (Table 6-2). Very small proportions of the total variations in these data are explained here, but sample sizes are high enough to reveal significant differences, or trends. Table 6-2. Results of regressions of water column measurements against time in the Hemingway Quadrant of the West Florida Shelf. Chlorophyll data spanned from early 1980s to late 1990s; turbidity from 1993-1999; and clarity from 1983-1997. Water column depths Intercept a Slope b ± s.e. n R2 P Chlorophyll Surface -66 0.03 ±0.01 210 0.05 0.001 Middle -517 0.26 ±0.11 33 0.16 0.02 Maximum -34 0.02 ± 0.04 34 0.01 0.65 Turbidity Surface -4663 2.37 ± 1.16 93 0.04 0.04 Middle 2572 -1.25 ± 1.28 93 0.01 0.33 Maximum 2181 -1.06 ± 1.34 93 0.01 0.43 Claritv (Secchi depth) Surface 395 -0.19 ±0.15 89 0.018 0.208 Benthic primary producers were estimated to comprise 35% of the overall primary production on the West Florida Shelf, with microphytobenthos accounting for 10%, macroalgae 2%, and seagrasses 23%, based on a literature review by Okey (2002). In contrast, benthic primary producers make up an estimated 91% of the shelf s overall primary producer biomass, with microphytobenthos accounting for 11%, macroalgae 14%, and seagrasses 66% (Table 6-3). The distribution of flows at each trophic level is shown in Table 6-4. Table 6-3. Estimated production and biomass of the four primary producers in the West Florida Shelf ecosystem Primary producer Production Biomass (t-km^year"1) (t-km2) Phytoplankton 4553 a 25.0" Microphytobenthos 706 29.8 Macroalgae 144 c 36.1c Seagrasses 1583 175.6 Notes: Estimates were chosen or derived based on a literature review by Okey (2000). Sources are indicated when chosen from a range of estimates, rather than derived; a. Steidinger (1973) and Tomas (1995); b. Steidinger (1973); c. Likely an underestimate for WFS. 112 Table 6-4. Flows from primary production and detritus (tkrn 2year~'). System imports and exports are not shown. Some flows reach trophic level six because some organisms within some functional groups are supported by energy that has traversed five links from primary producers. From primary production From detritus level Consumed To detritus Respiration Throughput Consumed To detritus Respiration Throughput VI 0 0 0 0 0 1 1 • 2 V 0 1 3 5 2 8 12 22 IV 5 15 18 38 22 91 71 184 III 38 138 83 259 184 952 393 1529 II 259 1109 540 1909 1529 8169 4854 14552 I 1907 5080 0 6987 14551 0 0 17164 Sum 2209 6344 645 9198 16289 9220 5332 33454 The biomass values of almost all functional groups in the system were predicted to decline when shading mortality on benthic primary producers was simulated without including concomitant refuge effects of biogenic structure or enhancement of phytoplankton (Figure 6-2) (the specified shading mortality of half the P/B values resulted in an approximate 50% reduction in the biomass of each or all of the benthic primary producers). Of the 55 living groups in the model, only phytoplankton, small copepods, ichthyoplankton, and carnivorous jellyfish never declined as the result of the specified simple seafloor shading mortalities. These were the only groups in the model whose food sources originated entirely with phytoplankton (and detritus). 15 CO - • 1 * ~ CM § E •c * 10 2 </> § i lo >. ra x: in CL 3 O 5 1965 1970 1975 1980 Year 1985 1990 1995 Figure 6-1. Changes in estimates of phytoplankton production from 1966 until 1995 (from Kondratyeva & Sosa 1966 in Vargo and Hopkins 1990, El-Sayed 1972, Steidinger 1973 in Vargo and Hopkins 1990, Yoder & Mahood 1983, Vargo, unpublished data from 1993, Brian Bendis, Florida Fish and Wildlife Conservation Commission, unpublished data from 1994-1996). Declines were generally more severe when refuge effects of benthic groups (biogenic habitat structure) were specified using the 'mediation' function in Ecosim. Most groups declined more as the now 'habitat-providing' benthic primary producers declined. This occurred whether or not a species directly benefited from these habitat structures. Five groups, however, increased as the habitat-providing benthic 113 primary producers declined. These were lobsters, large crabs; stomatopods, nearshore planktivores, and turtles, all clustered between trophic levels 3.1 and 3.4. The addition of simulated phytoplankton enrichment to the shading and habitat scenarios led to predicted increases in groups that obtain most of their energy from phytoplankton pathways and decreases in groups that obtain most of their energy from benthic primary production. The direction of effects alternated regularly as they propagated to higher trophic levels. The number of full cycles of alternating effects appears to match the number of trophic levels in the system (4.7 trophic levels) (Figure 6-2). DISCUSSION The mechanism of community change via shading The separate articulation of four primary producer groups in the West Florida Shelf Ecopath model (out of 59 functional groups) enabled exploration of the potential community effects of long-term declines of water-column transparency in this historically clear-water system. Three of these primary producers are benthic and one is pelagic (i.e., phytoplankton). The main assumption of this investigation is that benthic primary production by microphytobenthos, macrophytes, and seagrasses would decrease over this continental shelf as the result of shading by phytoplankton blooms. This assumption is supported by several empirical studies in marine and freshwater settings (e.g., Maristo 1941, Spence 1982, Meyercordt and Meyer-Reil 1999, Nelson et al. 1999). An alternative hypothesis is that shading would not limit seafloor primary production if the latter is limited by something other than light; i.e. nitrogen. Broad community changes should result from declines of biomasses or production rates of benthic primary producers if a considerable proportion of the energy used by species throughout the system originates with benthic primary production. Changes should propagate through the system when benthic primary production is decreased not only because of the degradation of food pathways, but also because of shifts in the character and distribution of food production. The main prediction was that changes in primary production patterns and other changes related to shading (e.g., detritus deposition and DO declines) would shift the continental shelf community to a different assemblage with a proportionally greater reliance on water-column primary production and detritus deposition, and less of a reliance on benthic primary production and refuge value provided by associated biogenic structure. The modelling simulations were generally consistent with this prediction. 114 0.5 Biomass change (end / start) 1 1.5 0 0.5 1 1.5 Large oceanic piscivores Pelagic oceanic piscivores Dolphins Coastal sharks Large groupers Pelagic coastal piscivores Benthic oceanic piscivores Mackerel (adult) Mackerel (juvenile) Nearshore piscivores Seabirds Pelagic oceanic jelly eaters Struc. assoc. coast, pisciv. Benthic coastal piscivores Demersal coastal pisciv. Squid Large oceanic planktivores Rays and skates Octopods Benth. coast, invert, eaters Benth. oceanic invert, eat. Dem. coast, invert, eaters Struc. ass. coast, invert, eat. Struc. ass. coast, planktiv. Carnivorous jellyfish Dem. oceanic invert, eaters Lobsters Other fishes Pelagic oceanic planktiv. Stomatopods ^Shade microphytobenthos • Shade macroalgae • Shade seagrasses • Shade all three • Shade all, habitat loss E3Shade all, enrich phyto. • Shade, habitat, enrich phyto. 115 Biomass change (end / start) 0-5 1 1.5 2 0 0.5 1 1.5 2 Turtles Nearshore planktiv. Large crabs Sardine and herring Carniv. zooplankton Adult shrimps Dem. coastal omniv. Ichthyoplankton Surface pelagics Other mesozooplank. Stru. ass. coast, omni. Echinoderms Meiofauna Sessile epibenthos Small mobile epifauna Small infauna Small copepods Bivalves Mullets Manatees Microb. heterotrophs Dead carcasses Drift macrophytes Macroalgae Microphytobenthos Phytoplankton Seagrasses Sediment detritus Watercolumn detritus iShade microphytobenthos I Shade seagrasses I Shade macroalgae I Shade all three • Shade all, habitat loss • Shade all, enrich phyto. • Shade, habitat, enrich phyto. Figure 6-2. Predicted biomass changes (proportional change relative to starting baseline) after the 30-year simulations for the 59 functional groups in the West Florida continental shelf ecosystem. Groups are organized by descending trophic level (from 4.7 to 1). The left panel shows the results of simulated shading of the three benthic primary producers (shading mortality equal to half the production rate). The right panel shows shading simulations with the refuge effect from biogenic habitat structure, with a concomitant enrichment of phytoplankton, and with all effects. The black line is a moving average (second period) of the predicted changes from these combined effects; it is added simply to show the general trend in the direction of biomass change at different trophic levels. 116 The West Florida Shelf is unique among continental shelves of North America. Historical water clarity has been so high that attached macroalgae in the region have been observed and collected at depths beyond the 200 m shelf break (Sylvia Earle, personal communication in Humm 1973; also see Littler et al. 1985). More recent data suggest that enough light reaches 75 m depths to enable substantial photosynthesis of West Florida Shelf benthic microflora. Tomas (1995) pointed out that conventionally measured estimates of overall primary production should be considerably higher to account for the benthic microflora whose contributions are likely profound, though not well known, on the West Florida Shelf. The present Ecopath model is among the first to account for recent high estimates of benthic microalgae in nearshore marine systems, and this could partially explain the high overall primary production estimation (6987 t ww-km^-year"1) relative to previous estimates from the region when less information about benthic primary production was included (26191 ww-km^-year"1 from Browder 1993 using conversions provided by Durbin and Durbin 1998 and Parsons et al. 1977). The inclusion of this previously underestimated component of primary production helps resolve a long-standing dilemma typically encountered while attempting to balance whole food web models—the apparent lack of adequate energy to support higher trophic levels. The estimate for overall primary production for the West Florida Shelf model is merely 2.7% lower than the estimate for phytoplankton primary production in the Mississippi delta bight (7182 t wwkm"2year"'; from Sklar 1976 using above conversion sources). Primary production is virtually equal in these two systems if phytoplankton in the Mississippi delta bight completely inhibits benthic primary production. Indicated community changes The most conspicuous pattern revealed in this series of simulations is a shift in community composition away from organisms that obtain most of their energy from benthic primary production pathways, and towards organisms that obtain most of their energy from planktonic primary production. This pattern is best revealed by examining the direction of change of particular species during the final simulation, which combines shading mortality, biogenic habitat structure effects, and enhancement of plankton production (Figure 6-2, right panel). Also, the general match between the number of trophic levels represented in this food web and the full cycles of alternation in the direction of biomass change in the final simulation is reminiscent of the known alternating effects of trophic cascades in freshwater food chains (e.g., Carpenter 1988, Power 1990, Wootton and Power 1993, Carpenter and Kitchell 1996) and marine systems (e.g., Paine 1966, Estes and Palmisano 1974, Estes et al. 1974, Estes and Duggins 1995, Steinberg et al. 1995). If this pattern reflects a true trophic cascade, it is a donor-controlled cascade in a complex aquatic food web, and is thus only partially consistent with Strong's (1992) premise that most true cascades are aquatic and donor controlled and that cascades occur only in simple systems. 117 The simulations that specified shading mortality of benthic primary producers without concomitant increases in phytoplankton and without refuge effects of biogenic structure were valuable because they indicated the relative support provided by benthic primary producers throughout the system (Figure 6-2; left panel). These simulations indicated that 51 of 55 living groups (or 93% of living groups) in the West Florida Shelf system depend on benthic primary producers to a considerable extent, whether directly or indirectly. The lack of a predicted biomass decline in the ichthyoplankton group could be partially explained by this pooled group's lack of specified ontogenetic linkage to the various fish groups in this model iteration. Nevertheless, fish groups usually did not decline by more than 50% during these simulations so that spawning stock would presumably be maintained. Some nearshore forage fish assemblages and higher trophic levels responded to simulated phytoplankton blooms and shading in unexpected ways (i.e., negatively rather than positively), but this paradox was resolved when a posteriori examinations of diet compositions revealed the dependence of these groups on benthic production pathways. For example, the considerable and apparently counterintuitive declines of'surface pelagics' to simulated enhancement of phytoplankton and inhibition of benthic primary production is explained by the group's species composition and specified diet. Halfbeaks (Hemirhamphidae) and flying fishes (Exocoetidae) make up this group and the specified diet is 47% seagrasses, 26% small mobile epifauna, 12 % nearshore planktivores, 9% other fishes, 6% small infauna, and 1% mesozooplankton (based on halfbeak diets in Randall 1967). This implies a near exclusive reliance on benthic production, and so the direction of their response is not surprising. However, flying fish diet and biomass information were not available, and it is possible that halfbeaks occasionally consume quantities of zooplankton surpassing 1% of their diet. The indicated declines of this group would be more moderate if phytoplankton food were a larger part of its diet. Seabirds are also specified in the Ecopath model as being somewhat reliant on benthic primary production; they consume a variety of demersal and benthic fish species in addition to plankton-dependent sardine, herring, and nearshore pelagics (Vidal-Hernandez and Nesbitt 2002). Adult and juvenile mackerel also rely on some demersal and benthic fishes in addition to planktivorous fishes. The alternating response of both of these species to the final two simulations (Figure 6-2) reveals the diverse production origins of their diets. Predicted responses of such species look surprising only when our immediate assumptions about a group's diet composition do not match with the diet composition specified in the model. In Ecosim, diet compositions do change dynamically during simulations as predators target prey at rates proportional to the prey's relative abundance. However, some level of consumption of a prey by a predator must be specified for such shifts to occur. "Apparent prey switching" by predators results from particular prey spending less time foraging (i.e., less time in the 'vulnerable' pool) as they become less abundant. This is complemented by predator 'satiation,' which is represented by decreases in feeding time at high prey encounter rates, and through direct handling time effects (C. Walters, pers. comm., 12 November 2001). 118 It was surprising that biomass increased in five assumed beneficiaries of biogenic habitat refuge after the biogenic refuge effect was added to the shading simulation (i.e., after available refuge decreased with shading). The five groups were large crabs, lobsters, stomatopods, nearshore planktivores, and turtles. This initial paradox was resolved by examining changes in the predicted biomasses of the main predators of these groups. Declines in these main predators (mostly benthic and demersal fish predators) were more pronounced when the biogenic structure effect was added to the combined shading simulation. The resulting release in predation pressure on the five groups in question allowed them to increase above baseline levels despite decreases in available refugia and benthic primary production. These opposing forces are integrated in the Ecopath model and the Ecosim simulation; the decline in predators was expressed as a relatively stronger effect than the decline in refugia (and benthic primary production). This result indicates only that these opposing forces exist. For application to lobster or crab policy development, the model should be refined to emphasize the articulation and relative strengths of protective effects of habitat and other related questions. For example, Butler et al. (1995) observed that the widespread decline of sponges in parts of Florida Bay (related to cyanobacteria-blooms) had a dramatic impact on the abundance and distribution of juvenile lobsters. Splitting lobsters into ontogenetically linked functional groups would better capture the energetic dynamics of the protective effects of sponges on juvenile lobsters, and thus the whole population, but the current West Florida Shelf model contains only one lobster group because the emphasis of the current exercise is on the whole system. Forces other than shading can, of course, shift systems from benthic to pelagic, or otherwise considerably restructure marine communities. These include fishing impacts that can preferentially remove benthic target organisms (e.g., Overholtz et al. 1999) or benthic nontarget organisms as in trawling (e.g., Jennings et al. 2001a, Jennings et al. 2001b, Koslow et al. 2001). Some potential consequences of benthic -pelagic shifts in community organization Shifts from benthic to pelagic modes of production might be occurring in coastal marine ecosystems worldwide. The example that follows helps reveal specific processes of change in coastal systems relating to changes in water column transparency and benthic-pelagic shifts of production pathways, in this case related to fisheries interests. Two questions regarding to the dynamics of forage fishes and other middle and high trophic level groups can be addressed by examining a benthic portion of the West Florida Shelf food web (Figure 6-3): (1) Does the fisheries removal of shrimps and crabs increase the number of forage fishes produced through benthic pathways, or would the habitat modifications of shrimp trawling offset such benefits? (2) Can the presence of a keystone predator, such as stomatopods, lead to more forage fish being produced through benthic pathways by controlling the crustacean energy pathway (thus favoring the alternative energy pathway of forage fishes)? 119 The third trophic level depicted in Figure 6-3 shows the alternate pathways of benthic production through forage fishes or higher crustaceans. Flow of benthic production to forage fishes might be facilitated by the presence of stomatopods—a possible keystone predator. Intense fisheries on one of the pathways might similarly facilitate the flow of benthic production up the alternate pathway, but water quality degradation (i.e., transparency declines) could decrease the flow of benthic primary production up both pathways. However, a shift to relatively higher planktonic primary production would lead to higher detritus fallout to the benthos (not shown) thereby increasing benthic detritus pathways. Such questions can be explicitly addressed using modelling approaches such as Ecopath with Ecosim when linkages between benthic and pelagic components are adequately represented. Higher predators Stomatopods Forage fishes Adult shrimp Crabs Microphytobenthos Benthic macrophytes & drift algae Microbial heterotrophs Benthic detritus Figure 6-3. Generalized predatory interactions in a simplified benthic food web of the West Florida Shelf (expanded from Walsh's example in Cushing 1984). Arrow (a) represents fishes feeding on plankton and arrow (b) represents crabs feeding on clams, which eat detritus and phytoplankton. Other major components of the West Florida Shelf benthos have been excluded here for simplicity. Notwithstanding the potential for such keystone mediation in community dynamics, seafloor shading should shift the composition of benthic communities rather than decrease the overall benthic biomass because plankton blooms should result in increased detritus deposition. Benthic primary consumers should thus shift away from herbivory and towards detritivory. Such compensation might partially explain the high resiliency of Gulf of Mexico shrimp populations to fishing pressure (Cushing 1984) because shrimp trawling can decrease the standing biomass of benthic primary producers but increase the input of detritus (e.g., discards), possibly providing a minor feedback that promotes shrimp production, especially in situations of moderate hypoxia (the direct removal of fish predators might be an even more beneficial effect on shrimp from trawling on shrimp). Moreover, trawling increases turbidity by resuspension of sediment thus further promoting shading and the shift to a detritus-based system. 120 Lowered dissolved oxygen would result not just from microbial decomposition of increased organic material reaching the marine sediment, but also from decreases in benthic photosynthesis. Even extreme examples of hypoxia in marine systems, such as the "dead zone" associated with Mississippi River inputs (e.g., Malakoff 1998), might have resulted, in part, from the loss of benthic boundary layer photosynthesis related to sea floor shading. Sources of underestimation of shading effects The West Florida Shelf contains considerable areas of hard substrate on which benthic macroalgae can attach and grow (hard bottom is -35% of the total shelf area; Phillips et al. 1990, Phillips and Thompson 1990). Some evidence suggests that the biomass and production of macroalgae on the West Florida Shelf as a whole are higher than the conservative estimates used in this analysis (Table 6-3). For example, Phillips and Thompson (1990) indirectly suggest that the macroalgal biomass on the overall West Florida Shelf is 45% higher than the value used here. Zieman et al. (1989) estimated that macroalgae biomass was equal to 118 t km"2 in Florida Bay, an area of high macroalgal biomass that does not necessarily represent this shelf as a whole. Information on macroalgae of the West Florida Shelf stands out as the most uncertain among other primary producers (notwithstanding work by Earle 1972), and underestimation of its biomass and production in this analysis would underestimate the effects of shading. Furthermore, Luning's (1990) ratio of production to biomass (P/B) that was used for macroalgae (4 year"1) might be conservative. The notion that benthic primary production could be approximately equivalent to water column primary production on the West Florida Shelf is consistent with several recent studies (e.g., Cahoon and Cooke 1992, Maclntyre et al. 1996, Miller et al. 1996) and indicates that the current analysis could greatly underestimate the impacts of shading. In some settings, sediment chlorophyll alone has been shown to exceed water column chlorophyll considerably (Cahoon et al. 1990, Nelson et al. 1999). Is it really happening? The trend of increasing phytoplankton production in Figure 6-1 might simply reflect an evolution of detectability from 1965 until the early 1990s, or it might reflect a chance sequence related to differences in sampling areas throughout the region. Alternatively, it could indicate widespread ecological changes on the West Florida Shelf that have gone unnoticed. El Sayed's (1972) measurements were made close to the shelf break in more oligotrophic waters than the later 'mid-shelf samples, which could have been more influenced by riverine enrichment (Gilbes et al. 1996, Del Castillo et al. 2000). Such measurements should be compared only for similar water types (Vargo and Myers 1991). Even when considering these sources of error, the El Sayed (1972) estimate is still low relative to current measurements. The trends in Table 6-2 support the notion that water-column transparency has been declining and that those declines have been caused by phytoplankton blooms in surface waters, but just barely since R2 values in the 121 present analyses are so low. These trends result from aggregation of sampling locations that shifted spatially among sampling dates throughout the Hemingway Quadrant (e.g., it is possible that a 'temporal' trend might simply represent spatial differences within the area because sites were not re-visited). However, sampling bias was minimized by the systematic sampling design, the large number of samples, and the effectively haphazard placement of the Hemingway Quadrant, which is equivalent to random placement when one cannot 'see' the system being sampled (Fager 1968). Phytoplankton production rates per unit area are very high in the region's estuaries where nutrients are more concentrated. Livingston (1984) found that phytoplankton productivity of the Apalachicola Bay estuary system during the 1970s and 1980s ranged from 63 to 1,694 mgC m~2-day"\ Eastbrook (1973 in Livingston 1984) estimated the annual phytoplankton production of the Apalachicola estuary to be 371 gC m"2, which is equivalent to 15,426 t-km~2-year~\ Thus, phytoplankton production for the region's estuaries can be over three times higher than the phytoplankton production (per unit area) for the overall West Florida Shelf. Increases in nutrients in the waters of the greater shelf can only increase the efficacy of the shading effect explored here. The broad shifts in benthic primary producers and the organisms they support might indeed be occurring on the West Florida Shelf, given the order-of-magnitude increases in terrestrial-origin nutrient loads in nearshore waters during the last two decades (Paul Carlson, Florida Marine Research Institute, pers. comm., 2000). A community shift from distinctly benthic-based production to one characterized by more pelagic primary production occurred in Florida Bay during the early 1990s (Hall et al. 1999, Zieman et al. 1999) when turbidity increased dramatically (Boyer et al. 1999). The results of the present analysis of the broader West Florida Shelf are consistent with these empirical findings. The ecology of the West Florida Shelf might well hinge on basic water quality since decreases in transparency could shift the predominant energy flow pathways along a benthic to pelagic (sea floor to water column) continuum (e.g., Thayer et al. 1999). Such a change in water quality might also shift and compress low-light-adapted photosynthetic organisms into shallower zones (in cases where appropriate substrate is available). Sears and Cooper (1979) recognized the potential importance of shading on continental shelf communities when they suggested that observed changes in the extinction depth of benthic macroalgal assemblages could be used as a bioassay for evaluating water conditions. The West Florida Shelf provides a good opportunity for applying this methodology because a unique 'deep-water' (70-100 m depth) algal nodule habitat was documented using geo-referenced video and photographs during MMS surveys during the mid 1980s (ESE et al. 1987). The question of whether microphytobenthos production or benthic primary production in general has declined over the West Florida Shelf during the last 20 years could be indirectly addressed by choosing such a measurable proxy, based on Littler et al.'s (1991) and Steller and Foster's (1995) descriptions of algal nodule habitats and their potential uses for exploration of long-term ecological changes. 122 LITERATURE CITED Adams, AJ., Locascio, J.V. and Robbins, B.D., 2004. Microhabitat use by a post-settlement stage estuarine fish: evidence from relative abundance and predation among habitats. J. Exp. Mar. Biol. Ecol., 299(1): 17-33. Allen, R.R., 1971. Relation between production and biomass. J. Fish. Res. Board Can., 28: 1573-1581. Anderson, D. M., Gilbert, P. M. and Burkholder, J. M., 2002. Harmful algal blooms and eutrophication: nutrient sources, composition, and consequences. Estuaries, 25: 704-726. Anderson, D.M. (Editor), 1995. ECOHAB, The Ecology and Oceanography of Harmful Algal Blooms: A National Research Agenda. Woods Hole Oceanographic Institution, Woods Hole, MA., 66 pp. Banner, A. H., 1974. Kaneohe Bay, Hawaii: Urban pollution and a coral reef ecosystem. Proc. 2nd Int. Coral Reef Symp., 2: 682-702. Blanchard, G.F. and Montagna, P.A., 1992. Photosynthetic response of natural assemblages of marine benthic microalgae to short-term and long-term variations of incident irradiance in Baffin Bay, Texas. J. Phycol., 28(1): 7-14. Blanchard, G.F. and Montagna, P.A., 1995. Assessment of a brown tide impact on microalgal benthic communities in Baffin Bay (Texas) in 1990 using a primary production simulation model. Oceanol. Acta, 18(3): 371-377. Boesch, D.F., Anderson, D.M., Horner, R.A., Shumway, S.E., Testor, P.A. and Whitledge, R.E., 1997. Harmful Algae Blooms in Coastal Waters: Options for Prevention, Control and Mitigation. NOAA Coastal Ocean Program Decision Analysis Series No. 10. NOAA Coastal Ocean Office, Silver Spring, MD, 46 pp. Boyer, J.N., Fourqurean, J.W. and Jones, R.D., 1999. Seasonal and long-term trends in the water quality of Florida Bay (1989-1997). Estuaries, 22(2B): 417-430. Browder, J.A., 1993. A pilot model of the Gulf of Mexico continental shelf, p. 279-284. In: V. Christensen and D. Pauly (Editors), Trophic models of aquatic ecosystems. ICLARM Conf. Proc, 26, 390 p. Bunt, J.S., Lee, C.C. and Lee, E., 1972. Primary production and related data from tropical and subtropical marine sediments. Mar. Biol., 16:28-36. Butler, M.J., Hunt, J.H., Herrnkind, W.F., Childress, M.J., Bertelsen, R., Sharp, W., Mathiews, T., Field, J.M. and Marshall, H.G., 1995. Cascading disturbances in Florida Bay, USA: Cyanobacteria blooms, sponge mortality, and implications for juvenile spiny lobsters Panulirus argus. Mar. Ecol. Prog. Ser., 129(1-3): 119-125. Cahoon, L.B. and Cooke, J.E., 1992. Benthic microalgal production in Onslow Bay, North Carolina, USA. Mar. Ecol. Prog. Ser., 84: 185-196. Cahoon, L.B., Redman, R.S. and Tronzo, C.R., 1990. Benthic microalgal biomass in sediments of Onslow Bay, North Carolina. Estuar. Coast. Shelf Sci., 31(6): 805-816. Carpenter, S.R. (Editor), 1988. Complex Interactions in Lake Communities. Springer-Verlag, New York, 283 p. Carpenter, S.R. and Kitchell, J.F., (Editors), 1996. The Trophic Cascade in Lakes. Cambridge University Press, New York, USA, 399 p. CENR., 2000. National assessment of harmful algal blooms in U.S. waters. National Science and Technology Council Committee on Environment and Natural Resources, Washington DC. Chambers, P.A. and Kalff, J., 1985. Depth distribution and biomass of submersed aquatic macrophyte communities in relation to Secchi depth. Can. J. Fish. Aquat. Sci., 42(4): 701-709. Christensen, V. and Pauly, D., 1992. ECOPATHll - A system for balancing steady-state ecosystem models and calculating network characteristics. Ecol. Modelling, 61: 169-185. Christensen, V., Walters, C.J. and Pauly, D., 2000. Ecopath with Ecosim - A User's Guide. Univ. of British Columbia, Fisheries Centre, Vancouver, Canada and ICLARM, Penang, Malaysia, 131 p. Clavier, J. and Garrigue, C, 1999. Annual sediment primary production and respiration in a large coral reef lagoon (SW New Caledonia). Mar. Ecol. Prog. Ser., 191: 79-89. Colijn, F., and de Jonge, V.N., 1984. Primary production of microphytobenthos in the Ems-Dollard Estuary. Mar. Ecol. Prog. Ser., 14: 185-196. 123 Conley, D. J., Markager, S., Andersen, J., Ellermann T. and Svendsen, L. M., 2002. Coastal eutrophication and the Danish National Aquatic Monitoring and Assessment Program. Estuaries, 25: 848-861. Cushing, D.H., 1984. Do discards affect the production of shrimps in the Gulf of Mexico? In: J.A. Gulland and B.J. Rothschild (Editors), Penaeid Shrimps—Their Biology and Management. Fishing News Books, Farnham, Surrey, England, pp. 254-257. Daviescolley, R.J., Hickey, C.W., Quinn, J.M. and Ryan, P.A., 1992. Effects of clay discharges on streams .1. Optical properties and epilithon. Hydrobiol., 248(3): 215-234. Del Castillo, C.E., Gilbes, F., Coble, P.G. and Mueller-Karger, F.E., 2000. On the dispersal of riverine colored dissolved organic matter over the West Florida Shelf. Limnol. Oceanogr., 45(6): 1425-1432. Dennison, W.C. and Alberte, R.S., 1985. Role of daily light period in the depth distribution of Zostera marina (eelgrass). Mar. Ecol. Prog. Ser., 25(1): 51-61. Duffy, J. E. and Hay, M.E., 1991. Food and shelter as determinants of food choice by an herbivorous marine amphipod. Ecology, 72(4): 1286-1298 Duggins, D.O., Simenstad, S.A. and Estes, J.A., 1989. Magnification of secondary production by kelp detritus in coastal marine ecosystems. Science, 245: 170-173. Durbin, A.G. and Durbin, E.G., 1998. Effects of menhaden predation on plankton populations in Narragansett Bay, Rhode Island. Estauries, 21(3): 449-465. Earle, S.A., 1972. Benthic algae and seagrasses. In: S.Z. El-Sayed,, W.M. Sackett, L.M. Jeffrey, A.D. Fredericks, R.P. Saunders, P.S. Conger, G.A. Fryxell, K.A. Steidinger, and S.A. Earle (Editors), Serial atlas of the marine environment: chemistry, primary productivity, and benthic algae of the Gulf of Mexico, American Geographical Society, New York. El-Sayed, S.Z., 1972. Introduction. In: S.Z. El-Sayed, W.M. Sackett, L.M. Jeffrey, A.D. Fredericks, R.P. Saunders, P.S. Conger, G.A. Fryxell, K.A. Steidinger, and S.A. Earle, Serial atlas of the marine environment: chemistry, primary productivity, and benthic algae of the Gulf of Mexico, American Geographical Society, New York. Epifanio, C.E., Dittel, A.I., Rodriguez, R.A. and Targett, T.E., 2003. The role of macroalgal beds as nursery habitat for juvenile blue crabs, Callinectes sapidus. J. Shellfish Res., 22(3): 881-886. ESE et al., 1987. Southwest Florida Shelf ecosystems study, Volume II - Data synthesis report. Environmental Science and Engineering, Inc. Gainsville, Florida, LGL Ecological Research Associates, Inc., Bryan, Texas, and Continental Shelf Associates, Inc. Tequesta, Florida. Prepared for Gulf of Mexico OCS Region, New Orleans, Louisiana, US Department of the Interior/Minerals Management Service, April 1987. Estes, J.A. and Duggins, D.O., 1995. Sea otters and kelp forests in Alaska: generality and variation in a community ecological paradigm. Ecol. Monogr., 65(1): 75-100. Estes, J.A. and Palmisano, J. F., 1974. Sea otters: their role in structuring nearshore communities. Science, 185: 1058-60. Estes, J.A., Smith, N.S. and Palmisano, J.F., 1974. Sea otter predation and community organization in the western Aleutian Islands. Ecology, 59: 822-833. Fager, E.W., 1968. A sand bottom epifaunal community of invertebrates in shallow water. Limnol. Oceanogr., 13:448-464. Fong, P. and Harwell, M.A. 1994. Modeling seagrass communities in tropical and subtropical bays and estuaries: A mathematical model synthesis of current hypotheses. Bull. Mar. Sci., 54(3): 757-781. Fourqurean, J.W. and Robblee, M.B., 1999. Florida Bay: A history of recent ecological changes. Estuaries, 22(2b): 345-357. Gilbes, F., Tomas, C, Walsh, J.J. and Muller-Karger, F.E., 1996. An episodic chlorophyll plume on the West Florida Shelf. Cont. Shelf Res., 16(9): 1201-1224, Gomez, I., 2001. Ecophysiology of Antarctic macroalgae: effects of environmental light conditions on photosynthetic metabolism. Rev. Chilena Hist. Nat. 74 (2), 251-271. Gunderson, L.H., 2001. Managing surprising ecosystems in southern Florida. Ecol. Econ., 37(3): 371-378. 124 Hacker, S.D. and Steneck, R.S., 1990. Habitat architecture and the abundance and body-size-dependant habitat selection of a phytal amphipod. Ecology, 71(6): 2269-2285. Hall, M.O., Durako, M.J., Fourqurean, J.W., and Zieman, J.C., 1999. Decadal changes in seagrass distribution and abundance in Florida Bay. Estuaries, 22(2B): 445-449. Hallock, P. and Schlager, W., 1986. Nutrient excess and the demise of coral reefs and carbonate platforms. Palaios, 1: 389-398. Hallock, P., 1987. Fluctuations in the trophic resource continuum: a factor in global diversity cycles? Paleoceanography, 2: 457-471. Hallock, P., 1988. The role of nutrient availability in bioerosion: consequences to carbonate buildups. Palaeo, 63: 275-291. Hallock, P., 2000. Symbiont-bearing foraminifera: harbingers of global change. Micropaleontology, 46(Suppl. 1): 95-104. Hallock, P., Muller-Karger, F.E. and Halas, J.C., 1993. Coral reef decline - anthropogenic nutrients and the degradation of western Atlantic and Caribbean coral reefs. Research & Exploration, 9(3): 358-378. Hartwig, E.O., 1978. Factors affecting respiration and photosynthesis by the benthic community of a subtidal siliceous sediment. Mar. Biol., 46: 283-293. Hauxwell, J., Cebrian, J. and Valiela, I., 2003. Eelgrass Zostera marina loss in temperate estuaries: relationship to land-derived nitrogen loads and effect of light limitation imposed by algae. Mar. Ecol. Prog. Ser., 247: 59-73. Hillman, K., McComb, A.J. and Walker, D.I., 1995. The distribution, biomass and primary producers of the seagrass Halophila ovalis in the Swan-Canning estuary, Western Australia. Aquat. Bot., 51(1-2): 1-54. Holmquist, J.G., 1994. Benthic macroalgae as a dispersal mechanism for fauna: Influence of a marine tumbleweed. J. Exp. Mar. Biol. Ecol., 180(2): 235-251. Humm, H.J., 1973. Seagrasses. In: J.I. Jones, R.E. Ring, M.O. Rinkel, R.E. Smith (Editors), A Summary of Knowledge of the Eastern Gulf of Mexico. The State University System of Florida, Institute of Oceanography, St. Petersburg, Florida. Hunter, CL. and Evans C.W., 1995. Coral reefs in Kaneohe Bay, Hawaii: Two centuries of western influence and two decades of data. Bull. Mar. Sci., 57: 501-15. Jackson, J.B.C., Kirby, M.X., Berger, W.H., Bjorndal, K.A., Botsford, L.W., Bourque, B.J., Bradbury, R.H., Cooke, R., Erlandson, J., Estes, J.A., Hughes, T.P., Kidwell, S., Lange, C.B., Lenihan, H.S., Pandolfi, J.M., Peterson, C.H., Steneck, R.S., Tegner, M.J., Warner, R.R., 2001. Historical overfishing and the recent collapse of coastal ecosystems. Science, 293: 629-38. Jennings, S., Dinmore, T.A., Duplisea, D.E., Warr, K.J., and Lancaster, J.E., 2001. Trawling disturbance can modify benthic production processes. J. Anim. Ecol., 70: 459-475. Jennings, S., Pinnegar, J.K., Polunin, N.V.C. and Warr, K.J., 2001. Impacts of trawling disturbance on the trophic structure of benthic invertebrate communities. Mar. Ecol. Prog. Ser., 213: 127-142. Johansson, G. and Snoeijs, P., 2002. Macroalgal photosynthetic responses to light in relation to thallus morphology and depth zonation. Mar. Ecol. Prog. Ser., 244: 63-72. Koslow J.A., Gowlett-Holmes, K., Lowry, J.K., O'Hara, T., Poore, G.C.B. and Williams, A., 2001. Seamount benthic macrofauna off southern Tasmania: community structure and impacts of trawling. Mar. Ecol. Prog. Ser., 213: 111-125. Landsberg, J. H. and Steidinger, K. A., 1998. A historical review of Gymnodinium breve red tides implicated in mass mortalities of the manatee (Trichechus manatus latirostris) in Florida, USA. Pp. 97-100 in B. Reguera, J. Blanco, M. L. Fernandez, and T. Wyatt. (eds.) Proc. 8th Int. Conf. on Harmful Algae. Vigo, Spain. Xunta de Galicia and Intergovernmental Oceanographic Commission of UNESCO. Lapointe, B.E., Tomasko, D.A. and Matzie, W.R., 1994. Eutrophication and trophic state classification of seagrass communities in the Florida Keys. Bull. Mar. Sci., 54: 696-717. Laws, E.A. and Redalje, D.G., 1979. Effect of sewage enrichment on the phytoplankton populations of a tropical estuary. Pac. Sci., 33: 129-144. 125 Lenes, J.M., Darrow, B.D., Cattrall, C, Heil, C.A., Callahan, M., Vargo, G.A., Byrne, R.H., Prospero, J.M., Bates, D.E., Fanning, K.A. and Walsh, J.J., 2001. Iron fertilization and the Trichodesmium response on the West Florida Shelf. Limnol. Oceanogr., 46(6): 1261-1277. Lenihan, H.S. 1999. Physical-biological coupling on oyster reefs: How habitat structure influences individual performance. Ecol. Monogr., 69(3): 251-275. Lenihan, H.S. and Peterson, C.H., 1998. How habitat degradation by fishery disturbances enhances impacts of hypoxia on oyster reefs. Ecol. Appl., 8(1): 128-140. Levin, P.S. and Hay, M.E., 1996. Responses of temperate reef fishes to alterations in algal structure and species composition. Mar. Ecol. Prog. Ser., 134(1-3): 37-47. Littler, M.M., Littler, D.S. and Hanisak, M.D., 1991. Deep-water rhodolith distribution, productivity, and growth history at sites of formation and subsequent degradation. J. Exp. Mar. Biol. Ecol., 150(2): 163-182. Littler, M.M., Littler, D.S., Blair, S.M. and Norris, J.M., 1985. Deepest known plant life discovered on uncharted seamount. Science, 227: 57-59. Livingston, R.J., 1984. The ecology of the Apalachicola Bay system: an estuarine profile. U.S. Fish and Wildlife Service, FWS/OBS-82/05, 145 p. Livingston, R.J., 2001. Eutrophication processes in coastal systems: origin and succession of plankton blooms and effects on secondary production in Gulf Coast estuaries. CRC Press, Boca Raton, Florida, 327 p. Loya, Y., 1976. Effects of water turbidity and sedimentation on the community structure of Puerto Rican Corals. Bull. Mar. Sci., 26: 450-466. Luning, K, 1990. Seaweeds: Their Environment, Biogeography, and Ecophysiology. John Wiley and Sons, Inc., New York, pp. 355-370. Macintyre, H.L. and Cullen, J.J., 1995. Fine-scale vertical resolution of chlorophyll and photosynthetic parameters in shallow-water benthos. Mar. Ecol. Prog. Ser., 122(1-3): 227-237. Maclntyre, H.L., Geider, R.J. and Miller, D.C., 1996. Microphytobenthos: The ecological role of the "secret garden" of unvegetated, shallow-water marine habitats. I. Distribution, abundance and primary production. Estuaries, 19(2A): 186-201. Mahmoudi, B., Mackinson, S., Vasconcellos, M., Vidal-Hernandez, L. and Okey, T.A., 2002. An ecosystem model of the West Florida Shelf for use in fisheries management and ecological research: Volume I. Summary and analyses. Florida Marine Research Institute-Fish and Wildlife Conservation Commission, St. Petersburg, 39p. Malakoff, D., 1998. Death by suffocation in the Gulf of Mexico. Science, 281(5374): 190-192. Mann, K.H., 1972. Ecological energetics of the seaweed zone in a marine bay on the Atlantic coast of Canada. II. Productivity of the seaweeds. Mar. Biol., 14: 199-209. Maristo, L., 1941. Die Seetypen Finnlands auf floristischer und vegetations-physiognomischer Grundlage. Ann. Bot. Soc. Vanamo. 15, 1-314. Meyercordt, J. and Meyer-Reil, L.A., 1999. Primary production of benthic microalgae in two shallow coastal lagoons of different trophic status in the southern Baltic Sea. Mar. Ecol. Prog. Ser., 178: 179-191. Meyers, S.D., Siegel, E.M. and Weisberg, R.H., 2001. Observations of currents on the West Florida Shelf break Geophysical Research Letters, 28(10): 2037-2040. Miller, D.C., Geider, R.J. and Maclntyre, H.L., 1996. Microphytobenthos: The ecological role of the "secret garden" of unvegetated, shallow-water marine habitats. II. Role in sediment stability and shallow-water food webs. Estuaries, 19(2A): 202-212. Muller-Karger, F.E., Walsh, J.J., Evans, R.H. and Meyers, M.B. 1991. On the seasonal phytoplankton concentration and sea surface temperature cycles of the Gulf of Mexico as determined by satellites. J. Geophys. Res., 96:2645-2665. National Research Council (NRC), 2000. Clean Coastal Waters: Understanding the Effects of Nutrient Pollution. National Academy Press, Washington, D.C. Nelson, J.R., Eckman, J.E., Robertson, C.Y., Marinelli, R.L. and Jahnke, R.A., 1999. Benthic microalgal biomass and irradiance at the sea floor on the continental shelf of the South Atlantic Bight: Spatial and temporal variability and storm effects. Cont. Shelf Res., 19(4): 477-505. 126 Nielsen, S. L., Sand-Jensen, K., Borum, J. and Geertz-Hansen, O., 2002. Depth colonization of eelgrass (Zostera marina) and macroalgae as determined by water transparency in Danish coastal waters. Estuaries, 25: 1025-1032. Okey, T.A. 2003. Macrobenthic colonist guilds and renegades in Monterey Canyon drift algae: partitioning multidimensions. Ecol. Monogr. 73 (3), 415-440. Okey, T.A., 2002. Primary producers In: T.A. Okey and B. Mahmoudi (Editors). An ecosystem model of the West Florida Shelf for use in fisheries management and ecological research: Volume II. Model construction. Fish and Wildlife Conservation Commission, Florida Marine Research Institute, St. Petersburg. Okey, T.A. and B. Mahmoudi (Editors)., 2002. An ecosystem model of the West Florida Shelf for use in fisheries management and ecological research: Volume II. Model construction. Fish and Wildlife Conservation Commission, Florida Marine Research Institute, St. Petersburg. Okey, T.A., 1997. Sediment flushing observations, earthquake slumping, and benthic community changes in Monterey Canyon head. Cont. Shelf Res., 17(8): 877-897. Overholtz, W.J., Link, J.S. and Suslowicz, L.E., 1999. Consumption and harvest of pelagic fishes and squids in the Gulf of Maine—Georges Bank ecosystem. In: Ecosystem Approaches for Fisheries Management. University of Alaska Sea Grant, AK-SG-99-01, Fairbanks, pp. 163-186. Paine, R.T., 1966. Food web complexity and species diversity. Am. Nat., 100: 65-75. Paluszkiewicz, T., Atkinson, L.P., Posmentier, E.S. and McClain, C.R., 1983. Observations of a Loop Current frontal eddy intrusion onto the West Florida Shelf. J. Geophys. Res. (C Oceans Atmos.), 88(C14): 9639-9651. Parsons, T.R., Takahashi M. and Hargrave B., 1977. Biological oceanographic processes. Permagon Press, New York, 332 p. Pauly, D. and Christensen, V., 1995. Primary production required to sustain global fisheries. Nature 374, 255-257. Pauly, D., Christensen, V. and Walters, C, 2000. Ecopath, Ecosim, and Ecospace as tools for evaluating ecosystem impacts of fisheries. ICES J. Mar. Sci., 57: 697-706. Pearson, T.H. and Rosenberg, R., 1978. Macrobenthic succession in relation to organic enrichment and pollution of the marine environment. Oceanogr. Mar. Biol. Ann. Rev., 16: 229-311. Phillips, N.W. and Thompson, M.J., 1990. Offshore benthic communities. In: N.W. Phillips, K.S. Larson, W.S. Alevison, S.P. Bannerot, P. Bogenreider (Editors), Synthesis of available biological, geological, chemical, socioeconomic, and cultural resources information for the South Florida area. Prepared by Continental Shelf Associates for Minerals Management Service, OCS Study -MMS 90-0019, May 1990, pp. 155-193. Phillips, N.W., Gettleson, D.A. and Spring K.D., 1990. Benthic biological studies of the southwest Florida shelf. Am. Zool., 30(1): 65-75. Polovina, J.J., 1984. Model of a coral reef ecosystem I. The ECOPATH model and its applications to French Frigate Shoals. Coral Reefs, 3: 1-11. Power, M.E., 1990. Effects offish in river food webs. Science, 250(4982): 811-814. Rabalais, N.N., Turner, R.E. and Wiseman, W.J., 2002. Gulf of Mexico hypoxia, aka "The dead zone". Annu. Rev. Ecol. Syst., 33: 235-263. Rogers, C.S., 1979. The effects of shading on coral reef structure and function. J. Exp. Mar. Biol. Ecol., 41: 269-288. Scheffer, M., Carpenter, S., Foley, J.A., Folke, C. and Walker, B., 2001. Catastrophic shifts in ecosystems. Nature, 413: 591-596. Sears, J.R. and Cooper, R.A., 1978. Descriptive ecology of offshore, deep-water, benthic algae in the temperate western North Atlantic Ocean. Mar. Biol., 44(4): 309-314. Sklar, F.H., 1976. Primary productivity in the Mississippi delta bight near a shallow bay estuarine system in Louisiana. Louisiana State University, Baton Rouge, M.S. Thesis, 96 p. Smith, S.V., 1981. Marine macrophytes as a global carbon sink. Science, 211: 838-840 Sournia, A., 1976. Primary production of sands in the lagoon of an atoll, and the role of foraminiferan symbionts. Mar. Biol., 37: 29-32. Spence, D.N.H., 1982. The zonation of plants in freshwater lakes. Adv. Ecol. Res. 12, 37-125. 127 Steidinger, K.A., 1973. Phytoplankton ecology: A conceptual review based on eastern Gulf of Mexico research. CRC Crit. Rev. Microbiol., 3: 49-68. Steinberg, P.D., Estes, J.A. and Winter, F.C., 1995. Evolutionary consequences of food chain length in kelp forest communities. Proc. Natl. Acad. Sci., 92: 8145-8148. Steller, D.L. and Foster M.S., 1995. Environmental factors influencing distribution and morphology of rhodoliths in Bahia Concepcion, BCS, Mexico. J. Exp. Mar. Biol. Ecol., 194(2): 201-212. Strong, D.R., 1992. Are trophic cascades all wet? Differentiation and donor-control in speciose ecosystems. Ecology, 73(3): 747-754. Stumpf, R.P., Frayer, M.L., Durako, M.J. and Brock J.C., 1999. Variations in water clarity and bottom albedo in Florida Bay from 1985-1997. Estuaries, 22(2B): 431-444. Thayer, G.W., Powell, A.B. and Hoss, D.E., 1999. Composition of larval, juvenile, and small adult fishes relative to changes in environmental conditions in Florida Bay. Estuaries, 22(2B): 518-533. Tomas, C.R., 1995. Dynamics of the early life history of recreational fishes on the Florida shelf. In: Coastal production and sportfish plankton dynamics on the Florida shelf. Prepared by the Florida Marine Research Institute for the U.S. Fish and Wildlife Service, Project F-65. Tomascik T, and Sander, F., 1987. Effects of eutrophication on reef-building corals: II. structure of scleractinian coral communities on fringing reefs, Barbados, West Indies. Mar. Biol., 94: 53-75. Tomascik, T. and Sander, F., 1985. Effects of eutrophication on reef-building corals: I. growth rate of the reef-building coral Monastrea annularis. Mar. Biol., 87:143-155. Tomasko, D.A. and LaPointe, B.E., 1994. An alternative hypothesis for the Florida Bay die-off. P. 1086 In: Symposium on Florida Keys Regional Ecosystem, November 1992. Bull. Mar. Sci., 54(3). van Woesik, R., Tomascik, T. and Blake, S., 1999. Coral assemblages and physico-chemical characteristics of the Whitsunday Islands: evidence of recent community changes. Mar. Freshw. Res., 50: 427-440. Vargo, G.A. and Hopkins, T.L., 1990. Plankton. In: N.W. Phillips and K.S. Larson (Editors), Synthesis of available biological, geological, chemical, socioeconomic, and cultural resource information for the South Florida area. OCS Study, MMS 90-0019, pp. 195-230. Vargo, G.A. and Meyers, M., 1991. A comparison of marine productivity among outer continental shelf planning areas. Sections on: Florida Straits, Eastern, Central and Western Gulf of Mexico. Prepared for Continental shelf Associates, Inc. USD!/ MMS Contract No. 14-35-0001-30487. Vargo, G.A., Carder, K.A., Gregg, W., Heil, C, Shanley, E., Steidinger, K.A. and Haddad, K.D., 1987. The potential contribution of primary production by red-tides to the West Florida Shelf ecosystem. Limnol. Oceanogr., 32: 762-767. Verwey, J., 1931. Coral reef studies. II. The depth of coral reefs in relation to their oxygen consumption and the penetration of light in water. Terubia 13: 169-198. Vetter, E.W. and Dayton, P.K., 1998. Macrofaunal communities within and adjacent to a detritus-rich submarine canyon system. Deep-Sea Research Part II- Tropical Studies in Oceanography, 45(1-3): 25-54. Vetter, E.W. and Dayton, P.K., 1999. Organic enrichment by macrophyte detritus, and abundance patterns of megafaunal populations in submarine canyons. Mar. Ecol. Prog. Ser., 186: 137-148. Vetter, E.W., 1994. Hotspots of benthic production. Nature, 327:47. Vetter, E.W., 1995. Detritus-based patches of high secondary production in the nearshore benthos. Mar. Ecol. Prog. Ser., 120:251-262. Vidal-Hernandez, L. and Nesbitt, S., 2002. Seabirds of the West Florida Shelf. In: Okey, T.A., Mackinson, S., Vasconcellos, M., Vidal-Hernandez, L., Mahmoudi, B. (Editors). An ecosystem model of the West Florida Shelf for use in fisheries management and ecological research: Model construction. Florida Marine Research Institute-Fish and Wildlife Conservation Commission, St. Petersburg. Walker, D.I. and Ormond, R.F.G., 1982. Coral death from sewage and phosphate pollution at Aqaba, Red Sea. Mar. Pollut. Bull., 13: 21-25. Walsh, J.J. and Steidinger, K.A., 2001. Saharan dust and Florida red tides: the cyanophyte connection. J. Geophys. Res. (C Oceans), 106(C6): 11597-11612. Walters, C.J., Christensen, V. and Pauly, D., 1997. Structuring dynamic models of exploited ecosystems from trophic mass balance assessments. Rev. Fish Biol. Fish., 7(2): 139-172. 128 Walters, C.J., Kitchell, J.F., Christensen, V. and Pauly, D., 2000. Representing density dependent consequences of life history strategies in an ecosystem model. Ecosystems 3: 70-83. Walters, C.J., Pauly, D. and Christensen, V., 1999. Ecospace: prediction of mesoscale spatial patterns in trophic relationships of exploited ecosystems, with emphasis on the impacts of marine protected areas. Ecosystems, 2: 539-564. Weisberg, R.H., Black, B.D. and Li, Z., 2000. An upwelling case study on Florida's west coast. J. Geophys. Res. (C Oceans), 105(C5): 11,459-11,469. Weisberg, R.H., Black, B.D. and Yang, H.J., 1996. Seasonal modulation of the west Florida continental shelf circulation. Geophys. Res. Lett., 23(17): 2247-2250. Wootton, J.T. and Power M.E., 1993. Productivity, consumers, and the structure of a river food chain. Proc. Natl. Acad. Sci., USA 90: 1384-1387. Yentsch, C. S., Yentsch, C. M., Cullen, J. J., Lapointe, B., Phinney, D. A. and Yentsch, S.W., 2002. Sunlight and water transparency: cornerstones in coral research. J. Exp. Mar. Biol. Ecol., 268: 171-183. Yoder, J. and A. Mahood. 1983. Primary production in Loop Current upwelling. In: Woodward-Clyde Consultants/Skidaway Institute of Oceanography, Southwest Florida Shelf Ecosystems Study -Year 2 Final Report. U.S. Dept. of Interior/Minerals Management Service, Vol.1 Zieman, J.C., Fourqurean, J.W. and Frankovich, T.A., 1999. Seagrass die-off in Florida Bay: Long-term trends in abundance and growth of turtle grass, Thalassia testudinium. Estuaries, 22(2B): 460-470. 129 CHAPTER 7. Coral-to-algae shifts in Indonesian coral reefs: sedimentation and pollution ABSTRACT Corals are being replaced by algae on many tropical reefs. The percent cover of hard corals in Southwest Sulawesi's Spermonde Archipelago (Indonesia) increased with distance from watersheds while abiotic substrate decreased. Water clarity also increased with distance. Conspicuous community shifts characterized by invasion of ephemeral algae occurred seasonally at stations in both watershed blocks examined, but the shift at each block occurred in opposite seasons and involved different colonizing flora. Fleshy algae covered 36% of the reef station near the 'agricultural' watershed during the wet monsoon (they covered only 2.5% of this station during the dry monsoon). Cyanobacteria covered 25% of the reef location near the more 'urban' watershed during the dry monsoon (they covered only 5% of this station during the wet monsoon). Differences in watershed runoff characteristics at the two watershed blocks will likely explain observed differences in ephemeral flora colonization and the distribution of more enduring reef characteristics such as hard corals. INTRODUCTION Phase shifts on tropical reefs—generally characterized by replacement of corals by fleshy algae (e.g., Done 1992; Knowlton 1992; McCook 1999)—likely result from combinations of increased sedimentation, nutrient enrichment, fishing, disease, and oceanographic changes (Hughes 1994; Hughes and Connell 1999; McManus and Polsenberg, in press). I considered that examining community changes in Southwest Sulawesi reefs relative to distance from coastal runoff sources and season would provide useful insights into the physical mechanisms that influence these community shifts, even though modification of trophic forces (e.g., reduction of herbivores) is probably a key mediator of community shifts on coral reefs (Hay 1984; Smith et al. 2001; Belliveau and Paul 2002; Jompa and McCook 2002). Indonesia is a principal hot spot of coral diversity (e.g., Hoeksema 1993; Wallace 1997; Roberts et al. 2002), but Indonesian reefs have degraded rapidly due to marginal socioeconomic conditions (Fauzi and Buchary 2002) and a severity of threats that was under-appreciated until recently (e.g., Edinger et al. 1998, 2000). Degradation of coral reefs is increasing worldwide due to human-caused stress and disturbance at all scales (Knowlton 2001; McClanahan 2002). A broad literature focuses on coastal sedimentation as the major stressor to reef-building organisms (e.g., Rogers 1983, 1990; Hodgson 1989; Wolanski et al. 2003). Poor agricultural practices; deforestation; overgrazing; coastal development; and destruction of mangroves all increase sedimentation on reefs. Bossher (1991) explained that sedimentation causes reef drowning by decreasing the ratio of accretion to erosion; reducing the number, coverage, and biomass of reef builders, and exacerbating the effects of other disturbances by altering reef assemblages. High rates of sedimentation are known to retard reef development by stressing reef-building corals (Stafford-Smith 1993; Reigl 1995). Several other 130 studies have shown that nutrient enrichment and associated water quality degradation (e.g., transparency) can considerably modify coral reef communities as well (e.g., Wellington 1982; Dubinsky and Stambler 1996; Yentsch etal. 2002). General declines of herbivores can also help shift coral communities (Hay 1984; Thacker et al. 2001; Belliveau and Paul 2002), but several authors have suggested that overall community changes are best explained by combined effects of nutrient enrichment and herbivore declines (e.g., McCook 1999; Stimson et al. 2001; Jompa and McCook 2002). Szmant (2002) refined this integrated view further when suggesting that observed patterns of reef degradation are best explained when effects of nutrient pollution and decreased fish abundance are combined with effects of additional stressors such as sedimentation (see also Hatcher 1997; Hughes and Connell 1999). Although the relative contributions of the various stressors remain uncertain and inherently variable, sedimentation, eutrophication, fishing, and other agents of disturbance and stress have been increasing in the Spermonde Archipelago of Southwest Sulawesi, Indonesia, and throughout the tropics as human populations increase and resources decline (Wijsman-Best et al. 1981; Neil et al. 2002). Deterioration of the rich coral areas of Indonesia has occurred at an alarming pace due to deforestation, agriculture intensification (including pesticide and nutrient runoff), and other land use changes (Soekarno 1989; Barber 1998). Substantial ecological and hydro-geological modification is widespread throughout the mountainous and alluvial terrain (see Whitten et al. 1987), and this has led to increased erosion as well as deadly flash floods. Blast and cyanide fishing, anchor damage, coral mining, fish traps, and direct collection of reef organisms also foretell a grim outlook for Indonesian coral reefs (e.g., Wijsman-Best et al. 1981; Wilkinson 1993; Tomascik et al. 1997; Barber and Pratt 1998; Pet-Soede and Erdmann 1999). In this study I addressed three general questions: (1) Do patterns of coral reef community structure differ consistently with distance from watersheds in Southwest Sulawesi, Indonesia? (2) Is there a relationship between sedimentation or water quality and Southwest Sulawesi reef community characteristics? (3) Do Southwest Sulawesi reef communities fluctuate between wet and dry monsoon seasons? METHODS Study location The morphology and ecology of the coral cay islands in the Spermonde Archipelago of Southwest Sulawesi, Indonesia (Figure 7-1) are influenced by two alternating monsoons: a December-February wet monsoon from the Southeast Asian peninsula and a June-September dry monsoon from Australia (Umbgrove 1930; Moll 1983). The regularity of these monsoons is indicated by the Southwest Sulawesi river flow oscillations shown in Figure 7-2. 131 100 110 120" 130 4"55' 51(M 119°15' 119"20" 119 25' 119" 30' 119 35' Figure 7-1. Locations of sampling stations at coral cay islands in the Spermonde Archipelago (Kepulauan Spermonde), South Sulawesi, Indonesia. Indicated kilometers are distances from chosen rivers—the Sangkarak or Telo Rivers—and nearest land (in parentheses). Detailed map is from JOG (G) SB 50-4 and SB 50-8, Ministry of Defence, United Kingdom. Broader scale map after Tomascik et al. (1997). Reef community characteristics at separate coral cay island stations near watersheds were compared to those at island stations further from watersheds. These comprised a near island station and a 132 far island station—at Balangcaddi and Pajenekeng—adjacent to one watershed in the northern portion of the study area (the Sankarak River), and a corresponding near and far pair of stations—at Laelae Besar and Kudingarenglompo—adjacent to a more southern watershed area (the Telo and Jeneberang Rivers) (Figure 7-1). Reef slopes on cay islands close to rivers receive riverine sedimentation (e.g., elastics and organic material)—especially during high flow—while those further offshore receive carbonate sediment that is resuspended from adjacent reef flats. The upper reaches of both watersheds are historically influenced by deforestation in the mountains 20-30 km to the east (Barber 1998), but agricultural and urban activities influence watershed runoff to varying extents (see Introduction). Current speeds vary around each island and across the shelf, but average ~14 cm /second (Moll 1983). Monthly flows in 10 southwest Sulawesi rivers 120 r 100 -CO E. 80 -o 60 ! / 40 j Wa1 20 -0 -Jan July Jan July Jan July Jan July Jan July Jan July 1977 1978 1979 1980 1981 1982 Years Figure 7-2. Monthly flow measurements at ten river sampling stations throughout southwest Sulawesi (1977-1982). Peak runoff occurs during the wet monsoon (December - March); lowest flows occur at the end of the dry season (September -November). Data are from RePPProT (1988); each line is the data from a single station. Sampling design and protocol Community structure of sessile reef organisms (percent cover, number of species, dominance), reef structural integrity (substrate penetrability, ratios of substrate types), downward flux of sediment particulate matter (DF-SPM), and water clarity (Secchi depth) were characterized at each of the four coral cay island stations. Twelve permanent sampling stations were distributed randomly along the 4 m (below MLLW) isobath of the NE facing side of each of the 4 cays by using random numbers along a 50 m transect line. This exposure direction and depth was the most comparable among cays since waves variably disturb NW and S sides during wet and dry monsoons and soft sediment predominates on direct E sides. The position of each transect was thus determined by depth and exposure direction (i.e., the northeastern-most arc of each reefs outline). Blind selection of transect positioning was ensured by surface chop by impairing visual assessment of reef details during deployment of each transect's origin. This stratum was consistently at -40° slope, and each transect covered a large proportion of that side of each respective cay. One 0.5 m2 photographic sample was taken at each of the 12 sampling locations along each transect using a Nikonos V underwater camera with strobe mounted on a framer made from PVC piping and fiberglass (Dodge et al. 1982). Optimum number of samples was determined during a 133 pilot study as the near asymptote of a curve of cumulative number of species per sample. Photo-samples were positioned with reference to a steel reinforcement rod and the transect line so that photo-quadrats could be re-located precisely during the alternate season. Sampling was conducted on 11-12 February 1993 (near the end of the wet season) and again on 14-15 September 1993 (near the end of the subsequent dry season). Photographic samples were analyzed by projecting each sample image onto a grid with fifty randomly selected points. This method is useful in describing species composition and cover (Foster et al. 1991; Sullivan and Chiappone 1992), but it is known to miss rare species (R. DeWreede, pers. comm., April 2003) especially when overstories are present. Underestimation of sponge biomass, for example, should be expected. The organism or substratum type intersected by each of the fifty sampling points was recorded for each of the 94 samples (12 samples x 2 distances x 2 watershed blocks x 2 monsoon seasons). Where possible, corals were identified to species by comparing to photographs in Veron (1986), but descriptors of convenience were used when confidence in species identification was limited. Sample and reference photos gave good impressions of micro-structure, but error was expected due to image resolution limits. These four cays were chosen because they extended WSW of the two watersheds. Currents flow in that general direction for three quarters of the year (March-November; Moll 1983) but other local rivers influence these cays as currents generally flow NE during the 3-month wet season (December-February) (Moll 1983; Erftemeijer 1994; T. Tomascik, pers comm., June 1999). To address this issue, samples were analyzed based on direct distances from the coast (land) in addition to direct distances from the chosen rivers. Irregular distribution of islands led to a staggering of distances of the sampled cays from rivers and land such that the northern stations were further than the southern stations from their respective watersheds. This natural staggering enabled a posteriori analyses along four-point gradients of proximity to rivers (8, 12, 20, and 24 km) and land (1.5, 9, 14, and 18 km) (when replication was sacrificed). Thus, this design allows the four stations to be considered in any of three ways: (1) as near vs. far from watersheds with twofold replication as originally intended, (2) as near vs. far within each area of watershed influence, with no true replication, and (3) as 4 sampling stations along an unreplicated gradient of distance from land or rivers. The substrate penetrability of reef locations was evaluated by recording the number of standard 'hits' with a 3 lb. hand sledge hammer necessary to drive a steel rod 0.5 m into the reef at 12 randomly-chosen locations along each transect. This measure was standardized by ensuring all hits were conducted by the same individual, and effort was taken to keep the force of the hits constant. I developed this method for the present study, and do not believe results of this penetration method can be compared to results by others. The structural state of reef locations was further evaluated by examining ratios of abiotic substrate types revealed by photographic samples. 134 One Secchi disk measurement was taken adjacent to each of the four transects at the end of the wet monsoon sampling (11-12 February 1993), and accuracy was assessed by repeating each measurement 3 times. Twelve sediment traps (25 cm long PVC pipe with a 2.6 cm opening diameter and sealed bottoms) were attached vertically to the twelve steel marker rods along the transects, i.e., one trap per rod along each transect on each sampling date to compare DF-SPM between reef locations and seasons. Trap openings were positioned 50 cm above the substratum (with their bottom ends 25 cm from the substratum). They were collected two weeks after deployment by sealing the tops with rubber plugs before snipping the zip tie attachments. Sediment samples were dried in the field on coffee filters (pore size 40 microns) and placed in envelopes for transport and storage. In the laboratory, sediment samples were dried further and weighed before and after removal of calcium carbonate fractions using repeated treatments with 10% hydrochloric acid. The relative organic content of the remaining non-carbonate sediment fraction was evaluated by qualitatively ranking the observed reaction strength when treated with 3% hydrogen peroxide solution into three categories of organic content: 'None', 'Medium', and 'High'. Weights of sediment fractions were presented graphically but not evaluated statistically because of unequal sample sizes resulting from the removal of some sediment tubes at Kudingarenglompo by protective locals (see also Edinger et al. 1998) and the corruption of 3 other samples during transport and drying. Resuspension and bioturbation were not explicitly assessed. Total fish abundance was estimated at each transect during dry-season sampling using a cine-transect method. A video recording, angled at 45 degrees downward and forward was made by a diver swimming one meter above the transect line at each site. Statistical analyses Repeated measures analysis of variance (ANOVA) was performed on biotic categories (and abiotic substrate) using the General Linear Model (GLM) in the statistical software SPSS 10.0.5 (SPSS Inc., Chicago, Illinois). These tests encompassed the entire sampling design by simultaneously evaluating the effects of proximity to watersheds (near vs. far) (between-subject factor), monsoon season (within-subject factor), and block (North vs. South) (between-subject factor). Levene's tests and Box's tests were employed with the GLM analyses to evaluate the assumptions of homogeneity of variances and equality of covariance matrices (where appropriate), respectively. Separate repeated measures ANOVAs were performed within individual watershed blocks in cases of severe violations of the assumptions of the fully integrated tests (incorporating both watershed blocks), but where assumptions were not violated for separate tests. These separate tests simultaneously evaluated effects of proximity and season thereby sacrificed the twofold replication of each environmental unit, which was provided by the twofold repetition of the watershed blocks. Prior to these analyses, raw proportion data of each faunal and sediment-type category (i.e., occurrence / 50 sampling points) were arcsin square root transformed to change their binomial distributions to near normal distributions (Zar 1984; p. 239), as indicated by 135 histograms. In addition, numbers of species data were square root (x + 0.1) transformed to negate the effects of zeros and to achieve reasonable homogeneity of variances. Multivariate analysis of variance (MANOVA) was performed to simultaneously evaluate the effects of proximity to watersheds and watershed block (2 fixed factors) on three abiotic substrate categories (3 dependent variables)—rock, rubble, and sediment—using data from the sampling date with the least ephemeral algae cover at each watershed block. This MANOVA was followed by three univariate ANOVAs to evaluate the effects of proximity to watersheds and watershed block (2 fixed factors) on each substrate type (1 dependent variable). An additional univariate ANOVA was performed on the substrate penetrability data. As a complement to this series of ANOVAs, and to aid in the interpretation of results, log-linear analyses were conducted using the backward hierarchical method to find the simplest models of main effects, or interaction effects, or both that best explained the distributions of each biotic and abiotic category tested with the ANOVAs described above. These analyses were also conducted using SPSS 10.0.5, but in this case using untransformed data. Species-area curves were calculated as the average accumulating sum of new species from 1000 Monte Carlo iterations that randomly resampled the array of new species occurrences in the 12 photographic samples, sorting each resampling-iteration in descending order. This was performed with PopTools 2.5 (G. Hood, Pest Animal Control Research Centre, Commonwealth Scientific and Industrial Research Organization, Australia). Dominance was calculated as the total cover of the dominant taxa / the total cover of all living organisms. Simple linear regressions were performed to compare relationships between Secchi disk extension depth and distance from land, Secchi depth and distance from rivers, light extinction coefficients (k = 1.7/Secchi depth) and distance from land, and k and distance from rivers. RESULTS Sedimentation and water clarity Downward flux of suspended non-carbonate particulate matter (i.e., of riverine origin) was always higher at near-river sites than at offshore sites (Figure 7-3). These differences were as large as a factor of 49 during the wet monsoon when downward flux of non-carbonate sediment exceeded 30 mg-cm"2day_1 at Laelae Besar. Conversely, downward flux of carbonate sediment was always higher at offshore reef stations than at nearshore stations. Sedimentation rates were always higher during the wet monsoon than during the dry monsoon. The organic content of suspended sediment was highest at the nearshore southern station—Laelae Besar (8 km from the Telo River; 1.5 km from Ujung Pandang city), lower at the nearshore northern station—Balangcaddi (12 km from the Sangkarak River and 9 km from land), and not detected at the offshore stations (Table 7-1). Organic content was highest during the wet monsoon at both nearshore stations. 136 40 V > ra E 30 _o "B) E. 20 s Q. V) • 10 u. o 0 16 JZ 12 Q. 0) TJ 8 !c u u <D 4 to n=10 I Sedimentation n=3 fish observed in tubes; fish scales in 3 of 6 sarrqles arrple: • Carbonate sediment H Non-carbonate sediment n=l n=10 x n=7 n=6 n=10 _1 fcgg'l n=9 100 > 80 o o 60 c d> 40 Q) Q. 20 dry wet dry wet dry wet dry wet Water clarity (wet season) o 100 o 80 > o o 60 Q_ 20 0 100 | 80 o ^ 60 c o 40 k_ V Q. 20 Hard corals • Dry ll Wet ± Abiotic substrate Fleshy algae and cyanobacteria Laelae Besar Balangcaddi 8(1.5) 12(9) Kudingareng Pajenekeng 20 (14) 24 (18) Distance from rivers (and land) (km) Figure 7-3. Sedimentation, water clarity, and percent cover of dominant categories at the four reef stations during both dry and wet monsoon seasons along a gradient of distance from river mouths (8, 12, 20, 24 km) and distance from land (1.5, 9, 14, 18 km). Error bars are standard error. 137 Table 7-1. Relative organic content of the vertical flux of non-carbonate suspended sediment onto reefs in the study area. Proximity of transects to rivers (and land) is indicated. Monsoon season South cays (i.e, Telo River) North cays (i.e., Sankarak River) Laelae Besar 8 (1.5) km Kudingarenglompo 20 (14) km Balangcaddi 12 (9) km Pajenekeng 24 (18) km Dry Wet Medium High None None Low Medium None None Note: Organic content is ranked based on the observed degree of chemical reaction to 3% hydrogen peroxide solution. Significant positive relationships were detected between Secchi depth and distance from both land and chosen rivers (Table 7-2; Figure 7-3). Negative relationships between k and distance from watersheds were slightly weaker. Distance from watershed is thus appears to be a good proxy for water clarity during the wet monsoon season. Table 7-2. Relationship (simple linear regressions) between water clarity measures and distance (km) from land and chosen rivers (n = 4). Variables Intercept a Slope b ± s.e. adj.** P Secchi disk extension depth (m) Distance from land 3.58 0.63 ± 0.08 0.95 0.015 Distance from chosen rivers 0.91 0.59 ±0.13 0.86 0.048 Light extinction coefficient" Distance from land 0.41 -0.02 ± 0.01 0.78 0.074 Distance from chosen rivers 0.48 -0.02 ± 0.01 0.59 0.147 a. The light extinction coefficient (k) is 1.7/Secchi depth Community structure Distance from watersheds had a consistently significant effect on community structure (Table 7-3; Table 7-4; Table 7-6). For example, scleractinian corals covered significantly higher proportions of the offshore stations than the nearshore stations, whereas this category did not vary significantly between seasons or watershed blocks at a = 0.05 (Table 7-3). In the northern block, percent cover of scleractinian corals was 9 times higher at the Pajenekeng station (24 km from the Sangkarak River) than at the Balangcaddi station .(12 km from that river) (18 vs. 9 km from land), and the separate