Dedication Donald Perrin De Sylva

Dedication

The Proceedings of the First International Symposium on Mangroves as Fish Habitat are dedicated to the memory of University of Miami Professors Samuel C. Snedaker and Donald Perrin de Sylva.

Samuel C. Snedaker Donald Perrin de Sylva (1938–2005) (1929–2004)

Professor Samuel Curry Snedaker Our longtime collaborator and dear passed away on March 21, 2005 in friend, University of Miami Professor Yakima, Washington, after an eminent Donald P. de Sylva, passed away in career on the faculty of the University Brooksville, Florida on January 28, of Florida and the University of Miami. 2004. Over the course of his diverse A world authority on mangrove eco- and productive career, he worked systems, he authored numerous books closely with mangrove expert and and publications on topics as diverse colleague Professor Samuel Snedaker as tropical ecology, global climate on relationships between mangrove change, and wetlands and fish communities. Don pollutants made major scientific contributions in marine to this area of research close to home organisms in south and sedi- Florida ments. One and as far of his most afield as enduring Southeast contributions Asia. He to marine sci- was the ences was the world’s publication leading authority on one of the most in 1974 of ecologically important inhabitants of “The ecology coastal mangrove habitats—the great of mangroves” (coauthored with Ariel barracuda. His 1963 book Systematics Lugo), a paper that set the high stan- and Life History of the Great Barracuda dard by which contemporary mangrove continues to be an essential reference ecology continues to be measured. for those interested in the taxonomy, Sam’s studies laid the scientific bases biology, and ecology of this species. for the protection of mangroves. Within Don shared his extensive knowledge, the Rosenstiel School property, a small experience and insight with great mangrove stand containing all species enthusiasm and unique humor. He of Florida mangroves is dedicated is sorely missed by all those he be- to his memory. A commemorative friended, educated, and inspired. inscription marks the place and it is a testament to Sam’s legacy of research, teaching, and scholarship. BULLETIN OF MARINE SCIENCE, 80(3): 451, 2007

Obituary Dail Woodward Brown (1942–2006)

ail W. Brown played a key role in making the First International Symposium on Mangroves as Fish HabitatD a reality. Unfortunately, his life ended prematurely on September 27, 2006. Dail was Division Chief of NOAA’s National Marine Fisheries Service Ecosystem Assessment Division and had retired in August of 2006. He dedicated 38 years of service to the National Oceanic and Atmospheric Association and Smithsonian Institute and was an enthusiastic supporter of coastal zone management and coral reef ecosystem research and conservation. Dail Brown was born on August 11, 1942 in Columbus, Ohio and graduated from Upper Arlington High School in 1960. He double-majored in chemistry and biology at Denison University in Gran- ville, Ohio where he earned his Bachelor of Science degree. In 1965, he married his high school classmate, Patti Rea, before moving to California to continue his studies in Biological Oceanography at the University of California Santa Barbara. His doctorate work focused on midwater-fish communities of three contiguous oceanic environments, which led to his Ph.D. in 1969. He and his wife moved to Virginia for Dail’s first appointment under Dr. William Aron at the Smithsonian Institute in Washington D.C. While at the Smithsonian, Dail worked on U.S. preparations for the 1972 United Nations Conference on the Human Environment in Stockholm, the meeting which established the multilateral framework for addressing global environmental problems. Later, Dail joined NOAA, where he served in a number of roles in NOAA’s National Ocean Service and National Marine Fisheries Service, helping to develop the national coastal zone management program and working on endangered species, watershed management, and wetlands policy. In the 1980s, he also participated in salvaging efforts of the Battleship Monitor off the coast of North Carolina. He took a sabbatical in the early 1990s to work with the Chesapeake Bay Foundation. From 1996, Dail served as Division Chief of the Ecosystem Assessment Division focusing on coral reef conservation, deep sea corals, habitat classification, and invasive species. Dail was a mentor to many people, a warm and gentle colleague, and truly cared for his team. For those of us who worked with him closely, we will miss his “open-door” policy, his approach to problem-solving, and his subtle reminders to not take ourselves too seriously. He had a way of making everyone laugh and despite the various demands of Washington D.C., he remained focused on the big picture, committed to investing in cutting-edge research and finding innovative conservation solutions. Dail loved reading about science, fisheries management, physics, history and poetry (he could recite T. S. Eliot’s The Waste Land at one time). In the early 1990s, he also wrote creative articles for the Baltimore Sun promoting wetlands conservation. Dail enjoyed cooking for his family and was an avid tennis player before he suffered from post-polio syndrome. His love of science was demonstrated by his enthusiasm and dedication for judging at the Fairfax County Regional Science Fair for over 35 years. He is survived by his wife Patti, daughter Bonnie, and son Dail Jr. We honor his life and are grateful for his many contributions including his ardent support for the success of the First International Symposium on Mangroves as Fish Habitat. In loving memory…—Contributed by Shauna N. Slingsby.

Preface

Mangroves are conspicuous components of tropical and subtropical ecosystems around the globe. Because they occur along marine, estuarine, and riverine shorelines, a wide variety of fishes can be found among their inundated root systems. Mangroves undoubtedly play a variety of roles in the lives of associated fishes— feeding areas for some, daytime refugia for others, and nesting or nursery areas for yet more. Clearly, variation is broad in terms of mangrove habitat species composition, location, and configuration, as well as in terms of the fishes, invertebrates, and life stages that occupy them. Consequently, questions regarding the contribution of a given mangrove forest or mangrove-lined shoreline to the diversity, productivity, and stability of broader fish communities—and their exploited components—must be carefully qualified. Unfortunately, mangroves have only recently received focused attention from the international scientific and conservation communities as potentially important fish habitats—habitats which continue to rapidly shrink worldwide. While the body of scientific literature on mangrove-fish relationships is relatively small, it is growing. Recently, there has been a surge of correlative—and to a lesser extent, manipulative— investigations that address the basis for fish (and decapod) utilization (i.e., food vs shelter) as well as connectivity to fisheries and adjacent ecosystems. Many of these new studies are taking advantage of recent advances in fish tracking technology, active acoustic systems, chemical analysis of otoliths, underwater digital video photog- raphy, remote sensing, and geographic information systems. In addition to these new technologies, is the emerging science of creating or restoring mangrove habitat that is optimal for fish diversity, abundance, and/or production.

Early in 2005, the idea of an international meeting dedicated to the topic of linkages among mangrove habitats, fishes, fisheries, and adjacent systems seemed long overdue. Sound science is vital for the wise management and conservation of mangroves and fisheries, especially given that pressures on these resources continue to intensify. Our particular location for such a conference also seemed fitting. The State of Florida’s three southernmost counties contain virtually all of the estimated 2 million square kilometers of mangrove forest in the United States and Florida’s mangrove ecosystems were among the first in the world to receive serious scientific attention, especially from the standpoint of the role that red mangroves play in marine food webs. From the outset, we knew that the symposium’s Miami venue had potential for local, national, and international participation as well as strong historical connections. Nevertheless, we were taken aback by the scope, geographic extent, and sheer number of enthusiastic responses to our first announcement and call for papers of a mangrove-fish symposium to “…assess the state of the science, identify critical information gaps, and chart a course for future research, conservation, and management.” This issue of the Bulletin of Marine Science constitutes the “Proceedings of the First International Symposium on Mangroves as Fish Habitat.” The symposium convened April 19–21, 2006, at the University of Miami’s Rosenstiel School of Ma- rine and Atmospheric Science, Miami, Florida, USA. The three-day event attracted over 200 researchers, resource managers, educators, and students from 25 nations. The program featured 55 oral- and 25 poster-presentations, which were organized around the following themes: (1) Nursery and trophic function; (2) Conservation, management, and socio-economics; (3) Community ecology and connectivity; (4) Mangrove-fishery linkages; and (5) Disturbance and restoration. The symposium prompted the submission of 40 written manuscripts for this special issue of the Bulletin. Of these, 25 papers ultimately emerged from the peer-review process and have been arranged in this issue following the themes of the symposium. In addition to the published papers, we include all contributed abstracts to permanently capture this remarkable event’s content, character, and active participation. No doubt, this information will be critical for planning the second symposium on this fascinating, rapidly-evolving topic. The symposium and this publication were made possible by several key sponsors and a large number of very special and talented individuals. Sponsors included the National Oceanic and Atmospheric Administration’s (NOAA) Coral Reef Conserva- tion Program and its Southeast Fisheries Science Center; the University of Miami’s Rosenstiel School of Marine and Atmospheric Science (RSMAS) and its Pew Institute for Ocean Science (PIOS); the United States Geological Survey’s Center for Coastal & Watershed Studies; Environmental Defense; The Nature Conservancy; Lewis En- vironmental Services, Inc.; the International Society for Mangrove Ecosystems; and Florida Atlantic University’s Center for Environmental Studies. We thank RSMAS Dean Otis Brown, PIOS Director Ellen Pikitch, and NOAA’s Dail Brown for their warm welcome and cogent introductory remarks at the symposium’s start. We deeply appreciate the insightful and inspiring keynote addresses made by Stephen J. M. Blaber, Ivan Negelkerken, and Jurgenne Primavera. We are indebted to session chairs Samuel Gruber, Carole McIvor, Craig Faunce, and Robin Lewis. Our deepest gratitude to all those who contributed papers and/or abstracts for the symposium as well as to all manuscript reviewers who donated their expertise PREFACE 455

and precious time. Also, we would like to thank Bulletin of Marine Science Editor, Su Sponaugle, for supporting this endeavor and offering her advice and expert opinion throughout the peer-review process. Many other individuals helped in multiple ways throughout the planning and execution stages of the symposium and beyond. From NOAA, we are grateful to Shauna Slingsby, Dail Brown, Nancy Thompson, Peter Thompson, Theo Brainard, Eric Prince, James Bohnsack, Eric Orbesen, Derke Snodgrass, and Jennifer Schull. From the University of Miami, we thank Robert Cowen, Beth Babcock, Brian Teare, Neil Hammerschlag, Xaymara Serrano, Caroline Peyer, Evan D’Alessandro, Jiangang Luo, Rose Mann, Monica Valle, Monica Lara, and Dave Jones. From the NGO community, we thank Ken Lindeman and Phil Kramer. From the International Society for Mangrove Ecosystems, we thank Shigeyuki Baba and Keiko Ogura. Our sincere gratitude goes to Doreen DiCarlo, Florida Center for Environmental Studies, for approaching us initially and ultimately masterminding all symposium logistics—we were lucky she and Serena Edic found us. We would like to acknowledge the work of Juan Camilo Jaramillo, Luis Gomez, and Angel Li in the design and maintenance of the symposium’s web page. In retrospect, we have only one major regret—we know that had Sam Snedaker and Don de Sylva been alive to attend, contribute, and enjoy this event, this symposium and the resulting proceedings would have been further enriched. They both are greatly missed.

Joseph E. Serafy Rafael J. Araújo Guest Editor Guest Editor NOAA Fisheries / University of Miami University of Miami

List of Reviewers

Aaron J. Adams, Mote Marine Lab, USA Dennis M. Allen, University of South Carolina, USA Mário Barletta, Universidade Federal de Pernambuco, Brazil Stephen J. M. Blaber, CSIRO Marine & Atmospheric Research, Australia Steven J. Bouillon, Free University of Brussels, Belgium Joseph A. Butler, University of North Florida, USA Paulo de Tarso Chaves, Universidade Federal do Paraná, Brazil Ving Ching Chong, University of Malaya, Malaysia Felicia C. Coleman, Florida State University Coastal and Marine Laboratory, USA Guillermo Duque, INVEMAR, Colombia Ginny L. Eckert, University of Alaska Fairbanks, USA David B. Eggleston, North Carolina State University, USA Colin Field, University of Technology Sydney, Australia Sarah Frias-Torres, NOAA—Southeast Fisheries Science Center, USA Elise Granek, Portland State University, USA David Green, National Audubon Society, USA 456 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Marin F. D. Greenwood, Florida Fish and Wildlife Conservation Commission, USA Mark Harmon, Oregon State University, USA Kristen M. Hart, US Geological Survey Florida Integrated Science Center, USA Euan Harvey, University of Western Australia, Australia Peter J. Hogarth, University of York, UK Thomas L. Jackson, NOAA—Southeast Fisheries Science Center, USA Dave Jones, NOAA—Southeast Fisheries Science Center, USA Darryl Jory, University of Miami, USA Todd Kellison, NOAA—Southeast Fisheries Science Center, USA Christopher C. Koenig, Florida State University Coastal and Marine Laboratory, USA Uwe Krumme, Center for Tropical Marine Ecology (ZMT), Germany Craig A. Layman, Florida International University, USA Roy R. “Robin” Lewis III, Lewis Environmental Services, USA Janet A. Ley, Australian Maritime College, Australia Peter N. G. Kee Ling, National University of Singapore, Singapore William F. Loftus, U. S. Geological Survey Florida Integrated Science Center, USA Jorge López-Portillo, Instituto de Ecología, A.C., Mexico Jerome J. Lorenz, National Audubon Society, USA Richard A. MacKenzie, USDA Forest Service, USA Carole C. McIvor, US Geological Survey Center for Coastal & Watershed Studies, USA Liana Talaue McManus, University of Miami, USA Ivan Nagelkerken, Radboud University, The Netherlands George Morara Ogendi, Arkansas State University, USA Peter B. Ortner, NOAA—Atlantic Oceanographic and Meteorological Laboratory, USA Bruce Pease, NSW Department of Primary Industries, Australia Jurgenne H. Primavera, SEAFDEC Aquaculture Department, Philippines Jennifer S. Rehage, Nova Southeastern University, USA Eric A. Reyier, Dynamac Corporation, USA Paul M. Richards, NOAA—National Marine Fisheries Service, USA Lawrence P. Rozas, NOAA—Southeast Fisheries Science Center, USA Neil Saintilan, Department of Environment and Conservation, Australia Peter F. Sale, University of Windsor, Canada Christopher R. Sasso, NOAA—Southeast Fisheries Science Center, USA Emily Schmitt, Nova Southeastern University, USA Richard F. Shaw, Louisiana State University, USA Marcus Sheaves, James Cook University, Australia William F. Smith-Vaniz, U. S. Geological Survey Florida Integrated Science Center, USA Leonel O. Sternberg, University of Miami, USA D. Scott Taylor, Brevard County Environmentally Endangered Lands Program, USA R. Eugene Turner, Louisiana State University, USA Alan K. Whitfield, South African Institute for Aquatic Biodiversity, South Africa Robert J. Williams, NSW Department of Primary Industries, Australia BULLETIN OF MARINE SCIENCE, 80(3): 457–472, 2007 FIRST INTERNATIONAL SYMPOSIUM ON MANGROVES AS FISH HABITAT KEYNOTE ADDRESS MANGROVES AND FISHES: ISSUES OF DIVERSITY, DEPENDENCE, AND DOGMA

Stephen J. M. Blaber

ABSTRACT Tropical estuarine fishes are inextricably linked with mangroves, which are the dominant vegetation of tropical and subtropical estuaries. Among the most productive of aquatic areas and heavily exploited, their future may depend upon ecosystem understanding. This paper reviews diversity, dependence, and connectivity between mangroves and fisheries in the light of data from previously unstudied systems in developing countries and new approaches in developed countries. Fish diversity in mangroves varies at global, latitudinal, regional, local and habitat scales, and species composition in any one system represents the combined influences of factors operating at each of these scales. Mangrove dependence paradigms require critical evaluation as new data become available and as catches and mangrove areas decline. Although it is a widely held dogma that mangroves are essential for fish populations, most evidence is circumstantial. Therefore experimental and quantitative studies are needed to support arguments that the value of retaining mangroves exceeds that of their destruction.

Mangroves are the dominant vegetation and one of the most characteristic habitats of tropical and subtropical estuaries. Their approximate distribution is bounded north and south of the equator by the 20 °C isotherm. The more than 80 species of mangroves form forests and stands varying in complexity from single species fringes in most subtropical estuaries to multi-species forests covering many square kilometers in tropical systems. Some mangrove forests in the tropics have a complex zonation, may contain upwards of 17 species of trees (Hutchings and Saenger, 1987), and are very dense with many prop roots. The largest mangrove forests with greatest diversity occur in southeast Asia, northern Australia, and northeast South America. Any discussion of fishes in tropical inshore and estuarine areas is inextricably linked with mangroves, areas of high fish species diversity and complex interrelationships. More than 200 species of fish can occur in any one mangrove estuary of the tropical Indo-West Pacific, with somewhat fewer, although still more than 100, in tropical East Atlantic and neotropical estuaries (Blaber, 2000). These systems are among the most productive on the planet (Costanza et al., 1997) and hence are heavily exploited by man. The present total worldwide area of mangroves has been estimated at < 150,000 km2 (FAO, 2003), but this represents only about 40% of the original cover. Many countries, such as Thailand, Philippines, and Ecuador, have lost far more than 50% of their mangroves; Singapore only has 0.5% of their mangroves remaining (Sasekumar and Wilkinson, 1994). One of the major factors in the loss of estuarine habitat in the subtropics and tropics has been the destruction of huge areas of mangroves for aquaculture, timber, and industrial developments. Most mangrove systems in the tropics are in developing countries where fisheries are heavily exploited and issues of food security usually override management and conservation considerations. There is now the realization that for sustainable management to be successful humans must be included as part of the tropical estuarine ecosystem (Fig. 1) (Blaber, 2005).

Bulletin of Marine Science 457 © 2007 Rosenstiel School of Marine and Atmospheric Science of the University of Miami 458 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 by Chilean artist Monica Mangrove Forest Wetland sed with permission from the artist and courtesy of the Mangrove Action Project (www. . U Gutierrez-Quarto (www.gutierrez-quartogallery.com) mangroveactionproject.org). Figure 1. Issues of dependence,diversity, and dogma in mangrove ecosystems are shown in the painting BLABER: MANGROVES AND FISHES—DIVERSITY, DEPENDENCE, AND DOGMA 459

Against this background there is increasing concern about the future integrity and sustainable management of mangrove systems and their associated fishes. The need to understand the interrelationships between mangroves and fishes is now greater than ever, not only for sustainable fisheries management reasons, but for overall management of the exploitation or conservation of mangrove forests. Such understanding involves a number of concepts, particularly those concerning (1), the varying diversity of mangrove-associated fishes and the major influences causing differences among fish faunas of mangrove systems; (2), the various paradigms about estuarine or mangrove dependence of fishes and their validity; and (3), the relationships between mangroves and fisheries production. Each of these three points will be re-examined in this paper in the light of new data, especially from previously unstudied systems in developing countries and new approaches in developed countries. Relationships Between Mangroves and Fish Species Diversity.—The diversity of fishes in mangroves varies significantly at a number of scales; global, latitudinal, regional, local, and among habitats within individual systems. The number of species in a system represents the combined influences of factors operating at each of these scales. It is necessary to examine the role of each of these scales because the species composition in any one place is the result of their combined influences, and results in every system being unique. At a global scale, in terms of the world’s major zoogeographic regions, the highest diversities are found in the tropical Indo-West Pacific, a huge area stretching from the coast of East Africa to the central Pacific, where upwards of 600 species have been recorded in mangrove estuaries and many individual systems may have as many as 200 species (Table 1). The high diversity decreases latitudinally away from the equatorial “core area” (sensu Blaber, 2000) in southeast Asia, but larger sub-tropical mangrove systems still contain at least 100 species. The mangroves of the Tropical East Atlantic region along the West African coast have somewhat fewer species, but are still relatively rich, with larger estuaries such as the Senegal, having more than 130 species and smaller systems such as the Fatala in Guinea about 100 species. The Tropical West Atlantic region from the Gulf of Mexico to northern South America has similar numbers of species, with most open estuaries in the equatorial region containing at least 100 species. There is some attenuation away from the equator, although species numbers are more influenced by size and type of system rather than by latitude (Table 1). At a regional level, species composition is influenced mainly by habitat diversity, structure, and hydrology. For example, a comparison of the large Embley and Norman estuaries, both in the Gulf of Carpentaria, Australia, reveals that the former has twice as many species. The Norman has a narrow fringe of mangroves, soft muddy substrata, a large tidal range (> 4 m), extremes of salinity, and high current speeds, whereas the Embley has a broad range of habitats, including extensive mangrove forests, a variety of substrata, and less extreme hydrological conditions (Blaber et al., 1994). The possible influences on species composition of extensive mangrove forests vs smaller fringing mangroves is illustrated by the Matang and Merbok estuaries in West Malaysia. The former has large and diverse mangrove forests covering many square kilometers, while in the latter, large tracts of mangroves have been removed and there are now only fringing stands of mangroves. The Matang system has at least 117 species while about half this number have been recorded in the Merbok (Khoo, 1989; Chong et al., 1990; Sasekumar et al., 1994; Hayase and Fadzil, 1999). Although the numbers of species in 460 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Table 1. Numbers of species recorded from different types of mangrove systems (O = open, B = blind, CL = coastal lake) in the four major zoogeographic regions (modified from Blaber, 2000).

System Country Type and size No. of spp. Source

Indo-West Pacific Alligator Creek Australia O small 150 Robertson and Duke (1990) Trinity Australia O medium 91 Blaber (1980) Embley Australia O large 197 Blaber et al. (1989) Vellar Coleroon India O large 195 Krishnamurthy and Jeyaseelam (1981) Chilka India CL large 152 Jones and Sunjansingani (1954) Chuwei Taiwan O small 30 Lin and Shao (1999) Ponggol Singapore O medium 78 Chua (1973) Matang W. Malaysia O medium 117 Sasekumar et al. (1994) Kretam Kechil E. Malaysia O small 44 Inger (1955) Purari New Guinea O large 140 Haines (1979) Pagbilao Philippines O medium 128 Pinto (1988) Solomon Islands Solomon Islands* O small 136* Blaber and Milton (1990) Morrumbene Mozambique O medium 113 Day (1974) Tudor Creek Kenya O small 83 Little et al. (1988) Kosi South Africa CL large 163 Blaber and Cyrus (1981) Mhlanga South Africa B Small 47 Harrison and Whitfield (1995)

East Atlantic Lagos Nigeria CL large 79 Fagade and Olaniyan (1972) Fatala Guinea O medium 102 Baran (1995) Ebrié Ivory Coast CL large 153 Albaret (1994) Niger Nigeria O large 52 Boeseman (1963) Gambia Gambia O large 89 Baran (2000) Sine-Saloum Senegal O large 123 Baran (2000)

West Atlantic Caeté Brazil O medium 82 Barletta et al. (2005) Guaratuba Brazil CL medium 61 Chaves and Bouchereau (1999) Itamaraca Brazil O large 81 Paranagua and Eskinazi-Leca (1985) Orinoco Venezuela O large 87 Cervigon (1985) Ciénaga Grande Colombia CL large 114 Leon and Racedo (1985) Tortuguero Costa Rica O small 70 Gilbert and Kelso (1971) Sinnamary French Guiana O medium 83 Boujard and Rojas-Beltran (1988) Everglades USA O large 53 Thayer et al. (1987)

East Pacific Río Claro Costa Rica O small 22 Lyons and Schneider (1990) Guerrero Lakes Mexico** CL small 105 Yáñez-Arancibia (1978) * incorporates 13 small estuaries—no one system more than 50 species. ** Sum of species for 10 small coastal lakes. these two systems are different, the dominant families in both are the Ariidae and Sci- aenidiae. However, the order of abundance of species in these families is quite different in the two estuaries, especially among the Sciaenidae. BLABER: MANGROVES AND FISHES—DIVERSITY, DEPENDENCE, AND DOGMA 461

At a more local level, it was shown that various mangrove habitats in Biscayne Bay Florida harbor different fish assemblages depending upon a suite of factors, the most important of which are (1) proximity to offshore reef habitats, (2) the salinity regime along the shoreline, and (3) the water depths within the mangrove forest interior (Serafy et al., 2003). Fish assemblages also vary in small Solomon Island estuaries (Blaber and Milton, 1990), but here the major influences are (1) the species of mangrove (i.e., type of structure), (2) whether the channels are blocked or choked by fallen mangrove branches, and (3) the type of substratum (ranging from soft mud to hard sand and rock). In the Pagbilao mangroves in the Philippines, Rönnbäck et al. (1999) found that fish species composition and abundance were influenced by (1) the dominant mangrove species and the structural complexity of the root system, (2) proximity to open water habitat, and (3) water depth. All of these examples serve to illustrate that the composition of mangrove fish assemblages is determined by an interplay of factors that include structural diversity of the habitat, hydrological features of current speed, tidal range, turbidity and salinity, and the nature of adjacent waters. The fish assemblage in any one mangrove system is the result of the combined effects of all of these factors and the relative importance of each. One factor in the interrelationship between mangroves and fishes that appears to have received little attention is the age of a mangrove. Most mangrove forests are dynamic and undergo constant change with various successions of tree species and tree sizes that are relatively well understood in a growing mangrove forest (Hutchings and Saenger, 1987). In the Matang mangroves of Malaysia, density and biomass of invertebrate epifauna was greatest in mature forest, intermediate in 15 yr old forest, and lowest in recently cleared forest, but the reverse was the case for invertebrate infauna, with highest densities in recently cleared sites (Sasekumar and Chong, 1998). These results are supported by an investigation of the fishes of a restored mangrove area in Florida where fish diversity among the oldest mangrove habitats was significantly higher than among more recently planted habitats (Barimo and Serafy, 2003). These patterns likely significantly influence the distribution patterns and species composition of fishes in mangroves, and furthermore, suggest that there may be a succession pattern of fishes colonizing mangroves. The question of scale is important when discussing diversity of mangrove fishes and their interrelationships with adjacent habitats. The very large areas of estuarine waters and their associated mangroves extending over 1000s of km2 of sea in Southeast Asia, West Africa, and Northeast South America contrast markedly with the smaller mangrove systems of estuaries along the East African, East Australian, and Caribbean coasts. The marked contrast between clear reef waters and turbid mangrove waters so characteristic of the latter are not evident in the Indo-West Pa- cific where the fishes live in a more uniformly turbid and low salinity environment (Blaber, 2000) and where linkages between mangrove structure, depth, and current speed may be more influential. On a smaller scale, further indication of the importance of linkages has been demonstrated recently for fishes in sub-tropical Moreton Bay in Queensland using a novel hierarchical landscape approach (Pittman et al., 2004). They quantified substratum structure at three spatial scales: whole landscape mosaic (10s of ha); habitat type (100s m2 to 1 ha); and within patch scale (cm2 to m2) and using fish densities, showed that more species used mangroves with adjacent continuous seagrass beds than used mangroves with adjacent patchy seagrass or 462 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 unvegetated mudflats. Highest assemblage densities were within spatially heterogeneous mangroves, but larger bodied and schooling species (e.g., Sillago spp.) exhibited a preference for fragmented mangroves with open water areas with high edge, high light penetration, and low shoot density that allowed unimpeded swimming and foraging. In addition, the proportion of mud in surficial sediments was the most important within-patch variable for both density and number of species using mangroves, particularly for gobiids and pleuronectids. Dependence.—The dependence or otherwise of tropical fishes on mangroves is a question that is not only of importance for understanding ecosystem functioning, but is one that is increasingly being asked by fisheries and environmental managers as fish catches decline and the worldwide area of mangroves decreases. Estuarine dependence of fishes has been extensively reviewed and discussed and detailed information is available, for example, in Lenanton and Potter (1987), Day et al. (1989), Pot- ter et al. (1990), Whitfield (1994, 1998), Blaber (1997, 2000), and Nordlie (2003). The estuarine dependence paradigm was first developed for fishes of the southeastern USA and southern Africa, primarily from studies of temperate and warm temperate estuaries. In these non-tropical, mainly non-mangrove areas, many species have been shown to be dependent on the estuarine environment during their juvenile phase (Day et al., 1981; Deegan and Thompson, 1985). Unfortunately, the paradigm gained too ready acceptance by workers everywhere, and in the case of tropical mangrove- fish associations, often without sufficient critical evidence. Longhurst and Pauly (1987) questioned the degree of estuarine dependence of sub-tropical and tropical coastal fishes and concluded that out of 20 areas examined, estuarine dependence be demonstrated in only two areas: in South Africa and southern Brazil, both in non-mangrove systems. Blaber and Blaber (1980) and Blaber (1981) postulated that most of the estuarine fishes of the tropics are not estuarine per se, but are characteristic of shallow turbid areas, often with variable salinity. These areas occur over very large areas of the sea in South and Southeast Asia, West Africa, and northern South America as well as in estuaries, but are restricted to estuaries in East Africa, eastern Australia, and parts of the USA. All of these areas and their fishes coincide with regions where mangroves are a primary habitat type. Variations in the complexity of mangrove habitats and the high species diversity of fishes makes any generaliza- tions about dependency difficult. Nevertheless, many species are dependent on the mangrove environment for all or part of their life cycle. Three main hypotheses have been advanced to explain such dependency: (1) reduced predation linked to depth, structure, and turbidity; (2) increased food supply for post-larvae and juveniles; and (3) shelter in quiet (non-turbulent) waters for post-larvae and juveniles. Numbers of fish species that are dependent on the mangrove environment varies both taxonomically and zoogeographically. For example, in most tropical regions, juveniles of most species of Mugilidae are found mainly in sheltered estuarine/mangrove waters, whereas among the Carangidae, juveniles of only some species are estuarine- dependent and most are not. Comparing regions, about 30% of the species in the Embley estuary in northern Australia and a similar percentage in the Lupar estuary in Sarawak, Malaysia, are estuarine/mangrove dependent (Blaber et al., 1989). Al- though similar percentages of estuarine-dependent species are also recorded in the tropical east Atlantic in West African mangrove areas (Baran, 1995), the situation in the tropical west Atlantic appears to be different. In the Caete estuary in northeast Brazil, at least 70% of species are estuarine-dependent (Barletta et al., 2003, 2005). BLABER: MANGROVES AND FISHES—DIVERSITY, DEPENDENCE, AND DOGMA 463

It is important, to note however, that some tropical species that are estuarine-dependent, are apparently not mangrove-dependent. A good example here would be the economically important tropical shads (genus Tenualosa) of south and southeast Asia (Blaber et al., 2003). There has been much discussion about the role of mangrove habitats as nurseries or feeding grounds for fish from adjacent habitats. As with species composition, the role of mangroves in this regard is apparently highly variable. In a series of studies in the Caribbean it has been shown unequivocally that many reef fish species utilize mangroves as nurseries. These studies are particularly interesting because they refer mainly to non-estuarine mangroves and give some insight into whether fishes also depend on mangroves in non-estuarine conditions. These non-estuarine mangroves can have higher densities of juvenile fishes than adjacent seagrass beds, channels, and mudflats, but similar densities to those on coral reefs (Nagelkerken and van der Vel- de, 2002). Different patterns of abundance of juveniles in the mangroves are thought to be related to the degree of structural complexity. The juveniles of at least 17 species of Caribbean reef fish are highly associated with bays containing mangroves and seagrass beds and are absent in bays without these habitats (Nagelkerken et al., 2000, 2002, 2004a; Dorenbosch et al., 2004). The situation is not straightforward, however, and interestingly, Nagelkerken et al. (2004b) demonstrated using stable carbon and nitrogen isotope analyses that mangroves in the Caribbean are not important feeding grounds for juvenile reef fish from adjacent seagrass beds. They analyzed 23 species of which only five showed signatures indicative of food intake from the mangroves. The patterns of habitat use in five habitats in a marine embayment in Zanzibar showed a high similarity to those in the Caribbean (Lugendo et al., 2005). However, caution is needed in equating any lack of mangrove isotope signals in fishes to a lack of food utilization from mangrove habitats. Epiphytes growing on mangroves or their pneumatophores may be an important food source, both directly and indirectly, for fishes entering mangrove environments and this carbon would have a different signal from mangrove carbon. Results from Biscayne Bay, Florida lend only partial support to an obligatory ontogenetic movement between reef and mangrove, with only two of five species examined conforming to this paradigm (Serafy et al., 2003). However, working in Belize and Mexico, Mumby et al. (2004) showed that mangroves were unexpectedly important as intermediate nursery habitats for reef fish and strongly influenced the community structure of fish on neighboring coral reefs. They showed that the biomass of four commercially important species more than doubled when the adult habitat was connected to mangroves, and that one species, Scarus guacamaia Cuvier, 1829 is functionally dependent on mangroves as a juvenile and has suffered local extinctions after mangrove removal. Recent experimental results from the Caribbean emphasize the different responses by juveniles of various reef species to the structure, food, and shade offered by mangroves, but clarify that the presence of these features contributes significantly to their attractiveness to juvenile reef fish (Verweij et al., 2006). In the large estuarine areas of south and southeast Asia, extensive mangrove forests are not often in close proximity to coral reefs, as they are in the Caribbean, so most ontogenetic habitat shifts relate to movements between coastal ocean and estuary or shallow to deeper water (Blaber, 1997). There are exceptions, however: Chong et al. (2005) found that the high diversity of fishes in the mangroves of Lang- kawi Island in northwest Malaysia is strongly influenced by the presence of coral 464 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 6 7 – n 10 17 2 r 0.58 0.48 0.92 0.79 0.66 ) 2 t = 1, . 1, = i D, metric i at time t (t) Penaeus merguiensis i at time Y function Fish capture (t) Percentage brown shrimps of Penaeid production (×1000 t) log10 of penaeid of catch (t) log10 Catch of 2 t M, hectare of fringing mangrove area; and of mangrove area 10 function tonnes of fish and invertebrates per hectare mangrove area. K, metric tons of fish and invertebratesmangrove area; if unexploited in the total For mangrove related species (2 shrimps, 1 crab and 1 fish) mangrove fish) 1 and crab 1 shrimps, (2 species related mangrove For area and perimeter accounted for most variation; for non-mangrove estuarine species (2 shrimps, 1 crab and 1 fish) mangrove perimeter was significant, but latitude dominant variable. X MSY = Maximum sustainable yield of penaeids MSY = degrees of latitude Area of mangroves; L AM = Percent of saline vegetation in a hydrological unit Mangrove area (×10,000 ha) log = 1, . . . , 11 yrs (1983–1993) . , 5 zones, t = 1, 11 HDit = demersal fish harvest (kg) for zone EDLMit = demersal fisheries effort (hrs) × log mangrove area (km Pooled-time series cross-sectional analysis of the relationship between relationship the of analysis cross-sectional series Pooled-time where fisheries, demersal for area mangrove and effort harvest, Coastal marshes in km for zone i at time (hrs) squared for zone ED 2 it = demersal fisheries effort Mangrove shoreline in km AM − 0.0212 L + 2.41 10 + 6.07 X + 0.0991 4.39 X −

X MSY = 0.4875 log MSY = 0.496 ln = 0.496 10 = 1.96 = 0.8648 HDit = b 0 + 1 EDLMit 2 ED it μit K ( t + 1) = ) − [ M 1)] D ln Y Relationships between mangrove environment and Multiple regressions and CPUE data. PCA log Formula Y X + 5.473 Y = 0.1128 Y Y = 1.074 X + 218.3 Gulf of Thailand Philippines— leiognathids Mexico Queensland, Australia Tropics Area USA Indonesia Asia Southeast Australia Barbier et al. (2002) Nickerson (1999) Yañez-Arancibia et al. Yañez-Arancibia (1985) Manson et al. (2005) Pauly and Ingles (1986) Table 2. Relationships between mangroves and fisheries production that have been quantified (modified from Blaber, 2000) 2. Relationships between mangroves and fisheries production that have been quantified (modified from Blaber, Table Source Turner (1977) Turner Martosubroto and Naamin (1977) Paw and Chua (1989) Staples et al. (1985) BLABER: MANGROVES AND FISHES—DIVERSITY, DEPENDENCE, AND DOGMA 465 reefs, limestone karst reefs, and seagrass meadows adjacent to the mangroves, that the fish fauna contains many species not usually found in Indo-Pacific mangroves, and that there is considerable movement among the habitats; a situation that mirrors that of the Caribbean. In the South Pacific, however, where relatively small areas of mangroves often occur adjacent to coral reefs, there appears to be little connectivity between the two habitats although the mangroves may be used as feeding grounds by visiting mobile piscivorous species. Quinn and Kojis (1985) showed that mangroves in northeast Papua New Guinea are nursery areas for a few species of coral reef fish. Similarly, Blaber and Milton (1990) in a study of 13 small Solomon Islands mangrove estuaries showed that these systems had an insignificant role as nurseries for coral reef fishes, Dorenbosch et al. (2005) obtained similar results in east Africa. Mangrove-Fisheries Connectivity.—It is in the area of fisheries production, including commercial, artisanal, and recreational, that the dogma concerning mangroves and fish has really taken hold. It is now a widely held view by fisheries managers, conservationists, and the general public that mangroves are essential for the maintenance of fish populations. There is no doubt that this has had very positive benefits, and given the popularity of recreational fishing in developed countries such as Australia and the USA, has immeasurably helped with the conservation of mangrove environments. However, in order to defend the value of mangroves, such as in debates about alternative land-use, good data are required, and unfortunately, the degree to which many fisheries in the tropics are dependent on mangroves is often unclear. It is important to distinguish between “fisheries within mangrove systems”, usually of an artisanal or subsistence nature in developing countries, and “offshore (of mangroves) fisheries” that are usually commercial or industrial concerns. In the former case, the activities by traditional or artisanal fishermen may be long-established and are totally dependent on the existence of the mangrove system. Many of these fisheries have been studied in detail and examples are given in Blaber (2000), Jhin- gran (2002), Islam and Haque (2004), and Kathiresan and Qasim (2005). Although the major threats to mangrove fishes are usually linked to environmental degradation, such as removal of mangroves, there is also evidence to suggest that many fish species in tropical estuaries, particularly in south and southeast Asia, are declining in abundance primarily as a result of overfishing. Such overfishing is strongly linked to issues of food security, but unlike many other human activities in and around tropical estuaries, fisheries are almost completely dependent on the maintenance of ecosystem integrity (Blaber, 2005). Links between offshore fisheries production and mangrove loss are much harder to quantify, if only because the large scale removal of mangroves for aquaculture and development purposes has coincided with increased fishing pressure and more efficient fishing technologies. Much of the evidence about the importance of mangroves to fisheries production has come from studies of penaeid prawns. Research in the Gulf of Mexico (Turner, 1977), Indonesia (Martosubroto and Naamin, 1977), India (Kathiresan and Rajen- dran (2002), Australia (Staples et al., 1985; Vance et al., 1996), and the Philippines (Primavera, 1998), provides good evidence that there is a correlation between commercial offshore prawn catches and the total area of adjacent mangroves. Pauly and Ingles (1986) concluded that most of the variance in the catches of penaeids could be explained by a combination of mangrove area and latitude (Table 2). 466 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

The significance of the relationships between fishes and mangroves is rather more equivocal and Robertson and Blaber (1992) concluded that in spite of the correlations between mangrove area and commercial fish catches, a causal link has not been established experimentally. In the Gulf of Mexico, Yanez-Arancibia et al. (1985) showed a positive correlation between commercial fish catches and mangrove area, and Barbier and Strand (1998) determined the effects of changes in mangrove area in the Laguna de Terminos on the production and value of prawn harvests in Campeche from 1980 to 1990. In Vietnam, De Graaf and Xuan (1998) demonstrated a similar relationship and indicated that 1 ha of mangrove forest supports a marine catch of –1 450 kg yr . The few studies that have quantified relationships between mangroves and coastal resources were summarized by Baran (1999) and Manson et al. (2005), but it must be reiterated that these are correlations and that causal links have not been established experimentally (see Table 2). In recent reviews of the subject, Baran (1999) and Baran and Hambrey (1999), showed that all of the studies to date suffer from problems of auto-correlation, with many factors other than just mangrove area, such as river discharge, area of shallow coastal water, intertidal area, and food availability, contributing to the relationships. These reviews also show thatfifinding nding a rela-relationship between mangrove area and fish production is not straightforward, because (i) closely related fish species can have very different ecological requirements, which blurs possible global relationships; (ii) results drawn from site-specific studies cannot be generalized to large areas in different geomorphic and climatic settings; and (iii) fishery statistics, in most cases, cannot be disaggregated enough to link catches to specific mangrove zones. More recently, in a review of the role of mangroves as nursery habitats for transient fishes and decapods, Sheridan and Hays (2003) concluded that the case for identifying flooded mangrove forests as critical nursery habitat remains equivocal until sufficient further experimental and quantitative studies have been carried out. Furthermore, in Queensland, Australia, Manson et al. (2005) demonstrated an empirical link between the extent of mangrove habitat and fishery production (mainly Crustacea) for three mangrove-related species, but such links for four non-mangrove estuarine species were less significant, and for these species, latitude was the dominant variable. As indicated earlier, not all mangroves or areas in mangrove systems, have the same relationships with fishes. For example, Vance et al., 1996 demonstrated that in northern Australia, the mangrove forests fringing deeper water contain much of the functionality compared with the more inland shallower, intertidal areas. There are other studies that showed that the inland shallower area of the mangrove forests are preferred areas by shrimps and small fishes (Rönnbäck et al., 1999; Affendy and Chong, 2007) presumably to avoid predation. Hence, if much of the loss of mangroves could be confined to the inland side, and deeper fringing areas left intact, perhaps much of the functional value could be retained. Kapetsky (1985) suggested that much of the functional value of mangroves might be retained from a smaller area of mangroves, for example, 75% of nursery function from 50% of original area. However, it is the deeper fringing areas of mangroves adjacent to the ocean that have been most attractive and suitable for aquaculture pond development throughout the tropics and have hence suffered greater proportional losses than more inland mangroves. This is not to say that shallower inland mangroves may not be important, particularly with regard to small fish species diversity and conservation, as demonstrated by Taylor et al. (1995) in Florida and Rönnbäck et al. (1999) in the Philippines. Loneragan et al. BLABER: MANGROVES AND FISHES—DIVERSITY, DEPENDENCE, AND DOGMA 467

(2005) suggested that landings of penaeids in west peninsular Malaysia have been maintained or increased despite large losses of mangroves. However, a recent review by Chong (2006) indicated that landings of penaeids on the west coast of peninsular Malaysia had actually declined from 60,967 t in 1989 to 39,296 t in 2003 (35% reduction) despite a reduction in fishing effort, as their main nursery habitat (mangroves) shrunk in area (23% loss). In the tropical Atlantic, recent quantitative data from the Caribbean suggest that the current rates of mangrove deforestation will have serious consequences for coral reef fish and fisheries (Mumby et al., 2004). Costanza et al. (1997) calculated that the economic value of estuaries in terms of services and natural capital per hectare was the highest of all ecosystems. Tropical mangrove systems, in particular, are zones of high productivity (Blaber, 2000), and assuming that most of the fisheries productivity is closely linked to mangroves, a number of recent studies have emphasized the economic value of mangroves, especially in the developing world (Hamilton et al., 1989; Barbier and Strand, 1998; Nick- erson, 1999; Barbier, 2000). Rönnbäck (2001) stated that “one major driving force behind the loss of more than 50% of the world’s mangroves during the last decade is the inability among economists to recognize and value all goods and services produced by this ecosystem.” In estimating the welfare effects of mangrove-fisheries linkages in Thailand, Barbier et al. (2002) commented that the fisheries most likely to be affected by habitat losses and impacts are those containing a high proportion of artisanal fishers. Evidence of the value of mangroves to fish and fisheries is rapidly increasing and appears almost overwhelming, but the case at the moment is not proven because much of the evidence is circumstantial. Therefore, despite the fact that the relationship has gained widespread scientific and public acceptance, there is still an urgent need for experimental and quantitative studies [such as that of Verweij et al. (2006)] to support the dogma, and to lend weight to the economic arguments that the value of retaining mangroves far exceeds the value of their destruction for whatever purpose. Mangroves are still being lost at an unacceptable rate from all points of view— ecological, economic, conservation, and human safety (e.g., Tsunami protection).

Acknowledgments

I am very grateful to the organizers of the “First International Symposium on Mangroves as Fish Habitat” at the Rosenthiel School of Marine and Atmospheric Science, University of Miami, for providing the stimulus for writing this paper and the funds for presenting it at the conference.

Literature Cited

Affendy, N. and V. C. Chong. 2007. Shrimp ingress into mangrove forests of different age stands, Matang Mangrove Forest Reserve, Malaysia. Bull. Mar. Sci. 80: 915. Albaret, J. J. 1994. Les poissons: biologie et peuplements. Pages 239–279 in J. R. Durand, Du- four, D. Guiral, and Zabi, eds. Environnement et ressources aquatiques de Côte d’Ivoire, tome II - les mileux lagunaire. Editions de L’ORSTOM, Paris. Baran, E. 1995. Dynamique spatio-temporelle des peuplements de poissons estuariens en Gui- née - relations avec les milieu abiotique. PhD Thesis, Université de Bretagne Occidentale, France. 468 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

______. 1999. A review of quantified relationships between mangrovesand coastal resources. Phuket Mar. Biol. Centre, Res. Bull. 62: 57–64. ______. 2000. Biodiversity of estuarine fish faunas in West Africa. Naga, The ICLARM Quar- terly 23: 4–9. ______and J. Hambrey. 1999. Mangrove conservation and coastal management in Southeast Asia: What impact on fisheries resources. Mar. Pollut. Bull. 37: 431–440. Barbier, E. B. 2000. Valuing the environment as input: review of applications to mangrove fishery linkages. Ecol. Econ. 35: 47–61. ______and I. Strand. 1998. Valuing mangrove-fishery linkages. Environ. Resour. Econ. 12: 151–166. ______, ______, and S. Sathirathai. 2002. Do open access conditions affect the valuation of an externality? Estimating the welfare effects of mangrove-fishery linkages in Thai- land. Environ. Resour. Econ. 21: 343–367. Barimo, J. F. and J. E. Serafy. 2003. Fishes of a restored mangrove habitat on Key Biscayne, Florida. Fla. Sci. 66: 12–22. Barletta, M., A. Barletta-Bergen, U. Saint-Paul, and G. Hubold. 2003. Seasonal changes in density, biomass, and diversity of estuarine fishes in tidal mangrove creeks of the lower Caeté Estuary (northern Brazilian coast, east Amazon). Mar. Ecol. Prog. Ser. 256: 217–228. ______, ______, ______, and ______. 2005. The role of salinity in structuring the fish assemblages in a tropical estuary. J. Fish Biol. 66: 45–72. Blaber, S. J. M. 1980. Fishes of the Trinity Inlet system of North Queensland with notes on the ecology of fish faunas of tropical Indo-Pacific estuaries. Aus. J. Mar. Freshw. Res. 31: 137–146. ______. 1981. The zoogeographical affinities of estuarine fishes in southeast Africa. S. Afr. J. Sci. 77: 305–307. ______. 1997. Fish and Fisheries of Tropical Estuaries. Chapman & Hall, London. 367 p. ______. 2000. Tropical Estuarine Fishes: Ecology, Exploitation and Conservation, Black- well, Oxford, 372 p. ______. 2007. Fisheries management and conservation in tropical estuaries–is a balance possible? Proc. 4th Word Fisheries Congress, Vancouver. ______and T. G. Blaber. 1980. Factors affecting the distribution of juvenile estuarine and inshore fish. J. Fish Biol. 17: 143–162. ______and D. P. Cyrus. 1981. A revised checklist and further notes on the fishes of the Kosi system. Lammergeyer 31: 5–14. ______and D. A. Milton. 1990. Species composition, community structure and zoogeog- raphy of fishes of mangroves in the Solomon Islands. Mar. Biol. 105: 259–268. ______, D. T. Brewer, and J. P. Salini. 1989. Species composition and biomasses of fishes in different habitats of a tropical northern Australian estuary: their occurrence in the adjoining sea and estuarine dependence. Estuar. Coast. Shelf Sci. 29: 509–531. ______, ______, and ______. 1994. Comparisons of the fish communities of tropical estuarine and inshore habitats in the Gulf of Carpentaria, northern Australia. Pages 363–372 in K. R. Dyer and R. J. Orth, eds. Int. Symp. Series, ECSA 22, Olsen and Olsen, Denmark. ______, D. A. Milton, D. T. Brewer, and J. P. Salini. 2003. Biology, fisheries and status of tropical shads (Tenualosa) in south and southeast Asia. Am. Fish. Soc. Symp. 35: 49–58. Boeseman, M. 1963. An annotated list of fishes from the Niger Delta. Zool. Verh. 61: 1–54. Boujard, T. and R. Rojas-Beltran. 1988. Longitudinal zonation of the fish population of the Sin- namary River (French Guiana). Revue d’hydrobiologie tropicale 21: 47–61. Cervigón, F. 1985. The ichthyofauna of the Orinoco estuarine water delta in the west Atlantic coast, Caribbean. Pages 57–78 in A. Yáñez-Arancibia, ed. Fish community ecology in estuaries and coastal lagoons: towards an ecosystem integration. UNAM Press, Mexico City. Chaves, P. and J. L. Bouchereau. 1999. Biodiversité et dynamique des peuplements ichtyiques de la mangrove de Guaratuba, Bresil. Oceanol. Acta 22: 353–364. BLABER: MANGROVES AND FISHES—DIVERSITY, DEPENDENCE, AND DOGMA 469

Chong, V. C. 2006. Importance of coastal habitats in sustaining the fisheries industry. Pages 1–28 in National Fisheries Symp. on “Advancing R and D Towards Fisheries Business Op- portunities”, 26–28 June 2006, Crown Plaza Riverside Hotel, Kuching, Sarawak, Malaysia. ______, A. Sasekumar, M. U. C. Leh, and R. D’Cruz. 1990. TheTh e fifish sh and prawn communi-communities of a Malaysian coastal mangrove system, with comparisons to adjacent mud flats and inshore waters. Estuar. Coast. Shelf Sci. 31: 703–722. ______, N. Affendy, A. L. Ooi, and L. L. Chew. 2005. The fishes and invertebrates of -man grove estuaries in northeastern Langkawi Island. Malay. Nat. J. 57: 193–208. Chua, Thia-Eng. 1973. An ecological study of the Ponggol estuary in Singapore. Hydrobiologia 43: 505–533. Costanza, R., R. d’Arge, R. de Groot, S. Farber, M. Grasso, B. Hannon, K. Limburg, S. Naeem, R. V. O’Neill, J. Paruelo, R. G. Raskin, P. Sutton, and M. van den Belt. 1997. The value of the world’s ecosystem services and natural capital. Nature 387: 253–260. Day, J. H. 1974. The ecology of Morrumbene estuary, Moçambique. Trans. Roy. Soc. S. Afr. 41: 43–97. ______, S. J. M. Blaber, and J. H. Wallace. 1981. Estuarine Fishes. Pages 97–221 in J. H. Day, ed. Estuarine Ecology with particular reference to southern Africa A. A. Balkema, Cape Town, South Africa. Day, J. W., Jr., C. A. S. Hall, W. M. Kemp, and A. Yáñez-Arancibia. 1989. Estuarine Ecology. John Wiley and Sons, New York. 576 p. De Graaf, G. J. and T. T. Xuan. 1998. Extensive shrimp farming, mangrove clearance and marine fisheries in the southern provinces of Vietnam. Mangroves and Saltmarshes 2: 159–166. Deegan, L. A. and B. A. Thompson. 1985. The ecology of fish communities in the Mississippi River deltaic plain. Pages 35–56 in A. Yáñez-Arancibia, ed. Fish Community Ecology in Estuaries and Coastal Lagoons: Towards an Ecosystem Integration. UNAM Press, Mexico City. Dorenbosch, M., M. C. van Riel, I. Nagelkerken, and G. van der Velde. 2004. The relationship of fish densities to the proximity of mangrove and fish nurseries. Estuar. Coast. Shelf Sci. 60: 37–48. ______, M. G. G. Grol, J. A. Christianen, I. Nagelkerken, and G. van der Velde. 2005. Indo-Pacific seagrass beds and mangroves contribute to fish density and diversity on adjacent coral reefs. Mar. Ecol. Prog. Ser. 302: 63–76. Fagade, S. O. and C. I. O. Olaniyan. 1972. The biology of the West African shad, Ethmalosa fimbriata (Bowdich) in the Lagos Lagoon. J. Fish Biol. 4: 519–533. FAO. 2003. Status and trends in mangrove area extent worldwide. By Wilkie, M. L. and For- tuna, S. Forest Resources Assessment Working Paper No. 63. Forest Resources Division. FAO, Rome. (Unpublished) Gilbert, C. R. and D. P. Kelso. 1971. Fishes of the Tortuguero area, Carribbean Costa Rica. Bull. Fl State Mus. Biol. Sci. 16: 1–54. Haines, A. K. 1979. An ecological survey of fish of the lower Purari River system, Papua New Guinea. Department of Minerals and Energy, Papua New Guinea, Purari River Hydroelec- tric Scheme Environmental Studies 6: 1–102. Hamilton, L., J. Dixon, and G. Miller. 1989. Mangroves: an undervalued resource of the land and sea. Ocean Yearb. 8: 254–288. Harrison, T. D. and A. K. Whitfield. 1995. Fish community structure in three temporarily open/ closed estuaries on the natal coast. Ichthyol. Bull. J. L. B. Smith Inst. Ichthyol. 64: 1–80. Hayase, S. and M. Fadzil. 1999. Fish distribution and abundance in Matang and Merbok mangrove waters, west coast of peninsula Malaysia: II. Preliminary results for 1996–1997. Pages 59–91 in V. C. Chong and P. S. Choo, eds. Productivity and sustainable utilization of brackish water mangrove ecosystems. Japan International Research Center for Agricultural Sci- ences, Tokyo, Japan. Hutchings, P. and P. Saenger. 1987. Ecology of Mangroves, University of Queensland Press, St Lucia, Queensland, Australia. 470 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Inger, R. E. 1955. Ecological notes on the fish fauna of a coastal drainage of North Borneo. Fieldiana; Zoology 37: 47–90. Islam, M. S. and M. Haque. 2004. The mangrove-based coastal and nearshore fisheries of Ban- gladesh: ecology, exploitation and management. Rev. Fish Biol. Fish. 14: 153–180. Jhingran, V. G. 2002. Fish and fisheries of India (3rd ed). Hindustan Publishing Corporation, New Delhi. 728 p. Jones, S. and K. H. Sujansingani. 1954. The fish and fisheries of the Chilka Lake with statistics of fish catches for the years 1948–1950. Ind. J. Fish. 1: 256–344. Kapetsky, J. M. 1985. Mangroves, fisheries and aquaculture. FAO Fisheries Report 338: 17–36. Kathiresan, K. and N. Rajendran. 2002. Fishery resources and economic gain in three mangrove areas on the south-east coast of India. Fish. Manage. Ecol. 9: 277–283. ______and S. Z. Qasim. 2005.Biodiversity of Mangrove Ecosystems. Hindustan Publish- ing Corporation, New Delhi. Khoo, H. K. 1989. The fisheries in the Matang and Merbok mangrove ecosystems. Pages 147– 169 in S. M. Phang, A. Sasekumar, and S. Vickineswary, eds. Proc. 12th Annual Seminar of the Malaysian Society of Marine Sciences, Kuala Lumpur, Malaysia. Krishnamurthy, K. and M. J. Jeyaseelan. 1981. Early life history of fifishes shes from Pichavaram man-mangrove ecosystem of India. Rapports et procés verbaux des Réunions, Conseil Permanent International pour l’exploration de la Mer 178: 416–423. Lenanton, R. C. J. and I. C. Potter. 1987. Contribution of estuaries to commercial fisheries in temperate Western Australia and the concept of estuarine dependance. Estuaries 10: 28–35. León, R. A. and J. B. Racedo. 1985. Composition of fish communities in the lagoon and estuarine complex of Cartagena Bay, Ciénaga de Tesca and Ciénaga Grande de Santa Marta, Colombian Caribbean. Pages 536–556 in A. Yáñez-Arancibia, ed. Fish community ecology in estuaries and coastal lagoons: towards an ecosystem integration. UNAM Press, Mexico. Lin, H. J. and K. T. Shao. 1999. Seasonal and diel changes in a subtropical mangrove fish assemblage. Bull. Mar. Sci. 65: 775–794. Little, M. C., P. J. Reay, and S. J. Grove. 1988. Distribution gradients of ichthyoplankton in an East African Mangrove creek. Estuar. Coast. Shelf Sci. 26: 669–677. Loneragan, N. R., N. Ahmad Adnan, R. M. Connolly, and F. J. Manson. 2005. Prawn landings and their relationship with the extent of mangroves and shallow waters in western peninsula Malaysia. Estuar. Coast. Shelf Sci. 63: 187–200. Longhurst, A. R. and D. Pauly. 1987. Ecology of Tropical Oceans, Academic Press, San Diego. Lugendo, B. R., A. Pronker, I. Cornelissen, A. de Groene, I. Nagelkerken, M. Dorenbosch, G. van der Velde, and Y. D. Mgaya. 2005. Habitat utilisation by juveniles of commercially important fish species in a marine embayment in Zanzibar, Tanzania. Aquat. Living Resour. 18: 149–158. Lyons, J. and D. W. Schneider. 1990. Factors influencing fish distribution and community structure in a small coastal river in southwestern Costa Rica. Hydrobiologia 203: 1–14. Manson, F. J., N. R. Loneragan, B. D. Harch, G. A. Skilleter, and L. Williams. 2005. A broad- scale analysis of links between coastal fisheries production and mangrove extent: A case study for northeastern Australia. Fish. Res. 74: 69–86. Martosubroto, P. and N. Naamin. 1977. Relationship between tidal forests (mangroves) and commercial shrimp production in Indonesia. Mar. Res. Indones. 18: 81–86. Mumby, P. J., A. J. Edwards, J. E. Arias-González, K. C. Lindeman, P. G. Blackwell, A. Gall, M. I. Gorczynska, A. R. Harborne, C. L. Pescod, H. Renken, C. C. C. Wabnitz, and G. Llewellyn. 2004. Mangroves enhance the biomass of coral reef fish communities in the Caribbean. Nature 427: 533–536. Nagelkerken, I. and G. van der Velde. 2002. Do non-estuarine mangroves harbour higher densities of juvenile fish than adjacent shallow-water and coral reef habitats in Curacao (Nether- lands Antilles)? Mar. Ecol. Prog. Ser. 245: 191–204. BLABER: MANGROVES AND FISHES—DIVERSITY, DEPENDENCE, AND DOGMA 471

______and ______. 2004a. Relative importance of interlinked mangroves and seagrass beds as feeding habitats for juvenile reef fish on a Caribbean island. Mar. Ecol. Prog. Ser. 274: 153–159. ______and ______. 2004b. Are Caribbean mangroves important feeding grounds for juvenile reef fish from adjacent seagrass beds. Mar. Ecol. Prog. Ser. 274: 143–151. ______, G. van der Velde, M. W. Gorissen, G. J. Meijer, T. van’t Hof, and C. den Hartog. 2000. Importance of mangroves, seagrass beds and the shallow coral ref. as a nursery for important coral reef fishes, using a visual census technique. Estuar. Coast. Shelf Sci. 51: 31–44. ______, C. M. Roberts, G. van der Velde, M. Dorenbosch, M. C. van Riel, E. Cocheret de la Morinière, and P. H. Nienhuis. 2002. Mar. Ecol. Prog. Ser. 244: 299–305. Nickerson, D. J. 1999. Trade-offs of mangrove area development in the Philippines. Ecol. Econ. 28: 279–298. Nordlie, F. G. 2003. Fish communities of estuarine saltmarshes of eastern North America, and comparisons with temperate estuaries of other continents. Rev. Fish Biol. Fish. 13: 281– 325. Paranagua, M. N. and E. Eskinazi-Leça. 1985. Ecology of a northern tropical estuary in Brazil and technological perspectives in fish culture. Pages 595–614 in A. Yáñez-Arancibia, ed. Fish community ecology in estuaries and coastal lagoons: towards an ecosystem integration. UNAM Press, Mexico. Pauly, D. and J. Ingles. 1986. The relationship between shrimp yields and intertidal vegetation (mangrove) areas: a reassessment. Pages 227–284 in A. Yáñez-Arancibia and D. Pauly, eds. IOC/FAO Workshop on Recruitment in Tropical Coastal Demersal Communities. Ciudad de Carmen, Mexico. IOC, UNESCO, Paris. Pinto, L. 1988. Population dynamics and community structure of fish in the mangroves of Pag- bilao, Philippines. J. Fish Biol. 33: 35–43. Pittman, S. J., C. A. McAlpine, and K. M. Pittman. 2004. Linking fish and prawns to their environment: a hierarchical landscape approach. Mar. Ecol. Prog. Ser. 283: 233–254. Potter, I. C., L. E. Beckley, A. K. Whitfield, and R. C. J. Lenanton. 1990. Comparisons between the roles played by estuaries in the life cycles of fishes in temperate western Australia and southern Africa. Environ. Biol. Fish. 28: 143–178. Primavera, J. H. 1998. Mangroves as nurseries: shrimp populations in mangrove and non-mangrove habitats. Estuar. Coast. Shelf Sci. 46: 457–464. Quinn, N. J. and B. L. Kojis. 1987. The influence of diel cycle, tidal direction and trawl align- ment on beam trawl catches in an equatorial estuary. Environ. Biol. Fish. 19: 297–308. Robertson, A. I. and S. J. M. Blaber. 1992. Plankton, epibenthos and fish communities. Pages 173–224 in A. I. Robertson and D. M. Alongi, eds. Tropical mangrove ecosystems. Ameri- can Geophysical Union, Washington DC, USA. ______and N. C. Duke. 1990. Mangrove fish communities in tropical Australia: spatial and temporal patterns in densities, biomass and community structure. Mar. Biol. 104: 369–379. Rönnbäck, P. 2001. Ecological economics of fisheries supported by mangrove ecosystems: economic efficiency, mangrove dependence, sustainability and benefit transfer.nd 2 Western Indian Ocean Science Association Scientific Symp.–Book of Abstracts. 41 p. ______, M. Troell, N. Kautsky, and J. H. Primavera. 1999. Distribution pattern of shrimps and fish amongAvicennia and Rhizophora microhabitats in the Pagbilao mangroves, Philip- pines. Estuar. Coast. Shelf Sci. 48: 223–234. Sasekumar, A. and V. C. Chong. 1998. Faunal diversity in Malaysian mangroves. Global Ecol. Biogeogr. Lett. 7: 57–60. ______and C. R. Wilkinson. 1994. Compatible and incompatible uses of mangroves in ASEAN. Pages 77–83 in C. R. Wilkinson, ed. Proc. Third ASEAN-Australia Symp. on Liv- ing Coastal Resources, Bangkok, Thailand. Australian Institute of Marine Science, Towns- ville, Australia. 472 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

______, V. C. Chong, K. H. Lim, and H. Singh. 1994. The fish community of Matang waters. Pages 457–464 in S. Sudara, C. R. Wilkinson, and L. M. Chou, eds. Proc. Third ASEAN-Australia Symp. on Coastal Living Resources, vol. 1: Status reviews. Chulalong- korn University, Bangkok, Thailand. Serafy, J. E., C. H. Faunce, and J. J. Lorenz. 2003. Mangrove shoreline fishes of Biscayne Bay, Florida. Bull. Mar. Sci. 72: 161–180. Sheridan, P. and C. Hays. 2003. Are mangroves nursery habitat for transient fishes and decapods. Wetlands 23: 449–458. Staples, D. J., D. J. Vance, and D. S. Heales. 1985. Habitat requirements of juvenile penaeid prawns and their relationship to offshore fisheries. Pages 47–54 in P. C. Rothlisberg, B. J. Hill, and D. J. Staples, eds. Second Australian National Prawn Seminar, Kooralbyn, Aus- tralia. CSIRO, Cleveland, Queensland, Australia. Taylor, D. S., W. P. Davis, and B. J. Turner. 1995. Rivulus marmoratus: ecology of distributional patterns in Florida and the central Indian River Lagoon. Bull. Mar. Sci. 57: 202–207. Thayer, G. W., D. R. Colby, and W. F. Hettler, Jr. 1987. Utilization of the red mangrove prop root habitiat by fishes in south Florida. Mar. Ecol. Prog. Ser. 35: 25–38. Turner, R. E. 1977. Intertidal vegetation and commercial yields of penaeid shrimps. Trans. Am. Fish. Soc. 106: 411–416. Vance, D. J., M. D. E. Haywood, D. S. Heales, R. A. Kenyon, N. R. Loneragan, and R. C. Pendrey. 1996. How far do prawns and fish move into mangroves? Distribution of juvenile banana prawns Penaeus merguiensis and fish in a tropical mangrove forest in northern Australia. Mar. Ecol. Prog. Ser. 131: 115–124. Verweij, M. C., I. Nagelkerken, D. De Graaf, M. Peeters, E. J. Bakker, and G. van der Velde. 2006. Structure, food and shade attract juvenile coral reef fish to mangrove and seagrass habitats: a field experiment. Mar. Ecol. Prog. Ser. 306: 257–268. Whitfield, A. K. 1994. Fish species diversity in southern African estuarine systems: an evolutionary perspective. Environ. Biol. Fish. 40: 37–48. ______. 1998. Biology and ecology of fishes in southern African estuaries. Ichthyol. Bull. J. L. B. Smith Inst. Ichthyol., Grahamstown, South Africa. 223 p. Yáñez-Arancibia, A. 1978. Taxonomy, ecology and structure of fish communities in coastal lagoons with ephemeral inlets on the Pacific coast of Mexico. Instituto de Ciencias del Mar y Limnologia, Universidad Nacional Autónoma de México, Publicaciones Especiales 2: 1–306. ______, A. L. Lara-Dominguez, P. Sanchez-Gil, I. Vargas Maldonado, M. De La C. García Abad, H. Álvarez-Guillén, M. Tapia García, D. Flores Hernández, and F. Amezcua Linares. 1985. Ecology and evaluation of fish community in coastal ecosystems: estuary- shelf interrelationships in the southern Gulf of Mexico. Pages 475–498 in A. Yáñez-Aran- cibia, ed. Fish community ecology in estuaries and coastal lagoons: towards an ecosystem integration. UNAM Press, Mexico City.

Address: Australian Commonwealth Scientific and Industrial Research Organisation (CSIRO) Marine and Atmospheric Research, Cleveland Laboratories, 233 Middle Street, PO Box 120, Cleveland QLD 4163, Australia. E-mail: . BULLETIN OF MARINE SCIENCE, 80(3): 473–495, 2007

nearshore habitat use by gray snapper (lutjanus griseus) and bluestriped grunt (Haemulon sciurus): environmental gradients and ontogenetic shifts

Craig H. Faunce and Joseph E. Serafy

ABSTRACT Fringing mangrove forests and seagrass beds harbor high densities of juvenile snappers and grunts compared to bare substrates, but their occupancy of these habitats is not homogeneous at ecologically meaningful scales, thus limiting our ability to compare habitat value. Here, density and size information were used to determine how gray snapper, Lutjanus griseus (Linnaeus, 1758) and bluestriped grunt, Haemulon sciurus (Shaw, 1803), use vegetated habitats during their ontogeny, and how their use of mangrove forests varied with season across broad spatial scales and physicochemical conditions. Both species exhibited a three-stage ontogenetic strategy: (1) settlement and grow-out (8–10 mo) within seagrass beds, (2) expansion to mangrove habitats at 10–12 cm total length, and (3) increasing utilization of inland mangroves during the dry season and with increasing body size. For fishes inhabiting mangroves, multivariate tests revealed that the factors distance from oceanic inlet and water depth were stronger predictors of reef fish utilization than the factors latitude, temperature, or habitat width. These findings highlight that the nursery function of mangrove shorelines is likely limited to the area of immediately accessible habitat, and that more expansive forests may contain a substantial number of larger adult individuals.

The importance of coastal vegetated seascapes to the early life history stages of diadromous nekton has been well documented, and has led to the conclusion that many temperate and warm-water marine fishes are “estuarine dependent” G( unter, 1967; McHugh 1967; Day Jr. et al., 1989; Blaber, 2002). However, the spatial separation of juveniles from adults and their utilization of coastal vegetated habitats is not limited to estuarine waters. Day (1951) was among the first to postulate that the fauna observed within a coastal embayment were drawn to “quiet water” conditions rather than estuarine conditions per se. Expanding this idea further, Nagelkerken and van der Velde (2002) offered the idea of “bay habitat dependence”. Despite attempts to provide guidelines for assessing the comparative “importance value” among fish habitats (USDOC, 1996; Beck et al., 2001), the relative utilization of multiple habitats by fishes has been difficult to quantify. Although numerous studies of animal-environment relationships have been conducted, these studies largely have been limited by a focus on a single habitat type or life-history stage, and/or by sampling at spatial scales that are too small to adequately capture the full suite of habitat use by fishes during their ontogeny (Pittman and McAlpine, 2003). At latitudes that support coral reefs, natural seascapes are vegetated primarily by seagrass beds and coastal mangrove forests. Here, many exploited reef fishes occupy inshore regions as juveniles before migrating offshore to reproduce thereby undergoing an ontogenetic pattern of habitat utilization similar to that exhibited by diadromous fishes in estuarine waters (Ogden, 1997). It has been postulated that the presence of mangroves and seagrass beds serve as extra “waiting room” habitats for juvenile coral reef fishes, and that adopting such a life-history strategy may buffer

against poor recruitment years (Bardach, 1959; Parrish, 1989). Recently, there have been renewed investigations into whether the presence of seagrass and mangrove habitats in the Caribbean enhance populations of fishes inhabiting adjacent coral reefs (e.g., Nagelkerken et al., 2002; Dorenbosch et al., 2004; Halpern, 2004; Mumby et al., 2004). In all of these studies, mangrove shorelines have been considered as homogenous habitats. However, mangrove shoreline habitats with different geographic locations, physical structure, and associated physicochemical environments may not contribute equally to future adult populations of reef fishes. Despite these management implications, few field studies have investigated the factors responsible for the utilization of mangrove shoreline habitats by reef fishes during their ontogeny, and those few have limited analyses to simple correlations performed at the entire fish assemblage level (Faunce and Serafy, 2006). The preceding is important because there is growing evidence that the processes or factors underpinning observed patterns of habitat utilization are species- and scale-dependent. For example, Martino and Able (2003) suggested that patterns in fish assemblage structure across large spatial scales (i.e., > 10 km) are due to individual species responses to dominant environmental gradients, whereas smaller-scale (1 km) patterns are due to habitat selection, competition, and/or predator avoidance strategies. In addition, recent studies in the Netherlands Antilles have reported patterns in the ontogenetic utilization of seagrass and mangrove habitats by reef fishes as a function of geographic scale (Nagelkerken et al., 2000b; Cocheret de la Morinière et al., 2002). However, their results apply to a 3 km wide bay. Determining whether similar patterns of ontogenetic utilization occur across larger spatial scales in other regions of the western Atlantic is important, especially if reef fish can move seasonally over distances more than a few kilometers (Gillanders et al., 2003; Pittman and McAlpine, 2003). An expansive (ca. 1000 km2) series of lagoons and embayments are found in southeastern Florida bordered to the west by the Florida mainland and to the east by the Florida Keys. In general, the influence of fresh water increases to the west-southwest, whereas oceanic and tidal influences are greater to the east and northeast. Across this diverse hydrographic seascape, the bays of southeastern Florida are primarily vegetated by the same seagrass (i.e., Thalassia testudinumBanks ex. König, 1805) and mangrove (i.e., Rhizophora mangle Linnaeus) species. Therefore, the region provides the opportunity to examine how reef fish utilization of seagrass and mangrove habitats changes with ontogeny and how utilization of a single habitat type (i.e., mangroves) varies across broad spatial scales and physicochemical conditions. This study aims to address these issues for two reef fishes common to the region, gray snapper,L utjanus griseus (Linnaeus, 1758) and bluestriped grunt, Haemulon sciurus (Shaw, 1803).

Methods

Data Collection.—The southeastern Florida coast was defined as the area between Key Biscayne to the northeast (80°09Ń, 25°43.8´W), Russell Key to the southwest (25°4.2Ń, 80°38.4´W), and Tavernier Creek to the southeast (25°06Ń, 80°32.4´W). The mangrove-lined shorelines of southeastern Florida were divided into five spatially-distinct strata. In order of their general position from west to east these strata include: (1) the mainland (ML); (2) islands within bays (IS); (3) landbridges (LB) that connect portions of the ML and the Florida Keys; (4) the western (i.e., leeward) shorelines of the Keys FAUNCE AND SERAFY: NEARSHORE ONTOGENIC HABITAT SHIFTS BY TWO REEF FISHES 475

(LK); and (5) the eastern (i.e., windward) shorelines of the Keys (WK; Fig. 1A). These strata were sampled using an iterative stratified random design during the wet (June–Au- gust) and dry (January–March) seasons of 2000 and 2001 following Cochran (1977) and detailed by Ault et al. (1999a; Fig. 1B). Sample locations, including five alternates, were selected each season by first dividing each stratum into sequentially numbered 30 m latitudinal sections, and using a random number generator to select the latitude of each sample. Samples consisted of (30 × 2 m) 60 m2 visual belt transects wherein a trained underwater observer identified, enumerated, and estimated the minimum, average, and maximum total length (TL) of each fish species to the nearest 2.5 cm. Only one individual conducted all fish surveys, thus eliminating inter-observer bias (St. John, 1990).P rior to the study, the observer’s ability to identify and rapidly enumerate fishes accurately was tested according to three criteria. First, the observer must have correctly identified all of the species common to the region (as defined from table 1 of Serafy et al., 2003) from digital photographs, including ontogenetic color variations where appropriate (e.g., parrotfishes). The observer must also have correctly estimated the identification of all species and number of fish within 10% from a 30 s digital video clip of a 30 m mangrove transect. Finally, following Bell et al. (1985), accurate and precise length assessment of fishes underwater was achieved by having the observer repeatedly estimate the sizes of plastic pipes of various lengths. To ensure fishes could be effectively observed within each transect, horizontal visibility was determined prior to surveys using a vertically mounted secchi disk and a measuring line. If visibility was < 2 m, the survey location was abandoned and an alternative location was visited. If no suitable alternative was available, locations were re-visited on a later date. Because fish observations in mangroves may be influenced by tidal stage and light levels, we attempted to control for this factor by conducting surveys between 0900 and 1700 on days that maximized the hours of daytime flood tide. A suite of environmental variables was also measured at each sample location. The factors water temperature and salinity were obtained using a Hydrolab® multi-probe instrument. Water depth and the width of mangrove canopy (WOC) were used as metrics of the vertical and horizontal components of the mangrove habitat, respectively. Depth was averaged from measurements at the mid- and end-points of each transect, while width of canopy was defined as the distance from the outer submerged mangrove drop- or prop-root to the upland edge. Latitude and distance to nearest oceanic inlet were used as landscape metrics. Distance to inlet (DI) was approximated with straight line measurements using Geographic Information System software (ArcView ver. 3.2). Because these distances sometimes crossed land barriers, they are considered here as indices and not potential routes of travel. Data Manipulations.—Length-frequency distributions from mangrove samples were constructed for L. griseus and H. sciurus following Meester et al. (1999). Briefly, this technique assumes the length-frequency distribution per sample has a mode at the mean length estimate with the two tails defined by the minimum and maximum lengths estimates. Length information was used to assign fish to one of three life-history components (LHC). Individuals smaller than the size at age 1 from published age-and-growth studies were designated as juveniles. Published size at sexual maturity (i.e., the size at which > 50% of individuals are sexually mature) was used to define the two remaining groups; individuals larger than their respective size at maturity were classified as adults, and remaining fish were classified as sub-adults. The literature sources, length ranges, and data manipulations used to define each LHC are given in Table 1. Data Analyses.—Average size and density information were generated for L. griseus and H. sciurus inhabiting seagrass beds and mangroves of southeastern Florida. Data 476 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 95.1 119.0 253.3 247.1 (mm TL) (mm Final size

Source (L-L conversions) Source (L-L Billings and Munro, 1974 Billings and Munro, 1974 Haemulon plumeieri; Manooch III and Matheson III, 1981 95.1 TL 95.1 103.5 FL 220.0 FL 195.0 SL Input size none Edits Mean 2 and 3 Mean 6 and 7 Source (sizes) Manooch III and Matheson III, 1981 Billings and Munro, 1974 Males: Domeier et al., 1996 García-Arteaga, 1992 Billings and Munro, 1974 Females: Domeier et al., 1996 Mean 4 and 5 Both sexes: Stark II, 1970 Size (mm) 95.1 TL 95.1 105 FL 182 SL 102 FL 220 FL 198 SL 190 SL 200 SL 1 2 4 3 8 5 6 7 ID Source Table 1. Table Information used to determine length; cut-off fork sizes FL: for conversions; life-history length-length components L-L: (LHC) sub-adults. used defined in were analyses. in-between Individuals those less and than adults, the defined size were at maturity age-1 at were size definedthe as than larger those juveniles, TL: total length. SL: standard length; Species Age-1 Lutjanus griseus Haemulon sciurus Size at maturity Lutjanus griseus Haemulon sciurus FAUNCE AND SERAFY: NEARSHORE ONTOGENIC HABITAT SHIFTS BY TWO REEF FISHES 477

from seagrass beds were obtained from raw data used by Serafy et al. (1997), who sampled fishes at eight sites within Biscayne Bay over 14 consecutive months using paired “rollerframe” trawls (3-m wide × 1-m high with 10-mm mesh) and identified and measured all captured fishes for TL. Average density and size data were plotted for each month, compared to published spawning seasons for L. griseus (Domeier et al., 1996; García-Cagide et al., 2001) and H. sciurus (Munro et al., 1973) and used to elucidate the timing of first recruitment, defined as the month when a minimum average length was observed during the year. Mean densities of each L. griseus and H. sciurus LHC, were determined for each mangrove stratum / season combination using a delta-distribution mean estimator; a measure of fish abundance (hereafter density, or fish 60 m–2) that separately considers the proportion of occupied sites and the number of individuals within them (Pennington, 1983). Patterns in these density values were subsequently compared to determine the relative contribution of each LHC to the overall distribution of each species. In addition, the average size of L. griseus and H. sciurus was generated per stratum per season. These data were plotted against the average distance (to inlet) index. Multivariate analyses were performed on the LHC assemblage with the statistical package PC-ORD (ver. 4.3) following the general recommendations of McCune and Grace (2002). The average number of juveniles, sub-adults, and adults ofL . griseus and H. sciurus observed within each stratum per year, was calculated and ln(x+1) transformed prior to analyses. A graphical representation of the underlying structure of these data was performed using Non-metric Multidimensional Scaling (NMS). A six-dimensional solution to a Bray-Curtis ordination was used to seed the initial geometry for NMS. This ordination was calculated using Sorensen (Bray and Curtis) distances, minimum deviation endpoint selection, and Euclidean geometry. A six-dimensional NMS solution was iteratively reduced (500 maximum) using an instability criterion of 0.0001. One hundred runs were performed on real data and 50 runs were performed on data that had been randomized within columns (i.e., each LHC). These provided the basis for a Monte Carlo test between the stress of each dimension from real and randomized data (i.e., that expected by chance). The appropriate number of dimensions for the final NMS solution and the significance of Monte Carlo tests for each dimension were determined from visual inspection of the resulting scree-plot. One final run on real data was performed using the appropriate number of dimensions (axes) and the same number of iterations and instability criterion as for prior runs (McCune and Grace, 2002). Whereas NMS was used to elucidate relationships among species and life-history stages at broad spatial scales (i.e., among strata), Canonical Correspondence Analysis (CCA) was performed to graphically express each LHC in terms of the environmental variables measured at sampling (i.e., direct gradient analysis, ter Braak, 1986). This method constrains the ordination of the LHC matrix by a linear multiple regression on the environmental variable matrix, so that each successive canonical axis accounts for a smaller proportion of the total variance. Fish data were ln(x+1) transformed and standardized to maximum (Minchin, 1987), whereas environmental data were screened for outliers and standardized to z-scores. To control for the influence of “double-zero” matches in the data, only samples with at least one individual belonging to an LHC were included in the analyses (Legendre and Legendre, 1983). The CCA was run with one hundred iterations with randomized site locations to facilitate Monte Carlo tests between (a) the eigenvalues and species-environment correlations for each axis that resulted from CCA and (b) those expected by chance. Canonical Correspondence Analysis produces a biplot where environmental variables are represented as arrows (vectors) radiating from the origin of the ordination. For analy- 478 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Figure 1. Map of Southeastern Florida depicting (A) landscape features and strata used in sampling design and (B) locations of visual transect samples. FAUNCE AND SERAFY: NEARSHORE ONTOGENIC HABITAT SHIFTS BY TWO REEF FISHES 479

Figure 2. Monthly average (± 1 SE) densities (top panels), and sizes (lower panels) of Lutja- nus griseus and Haemulon sciurus within Biscayne Bay seagrass beds. Horizontal arrows depict spawning seasons (for sources see Materials and Methods). Vertical arrows depict timing of initial recruitment.

ses, row (i.e., site) and column (i.e., environmental factor or LHC) scores were centered by unit variance, and optimized so that LHC scores (i.e., biplot projection points) were weighted mean sample location scores. In this way distances between each LHC facilitated direct spatial interpretation with environmental variables in the biplot (McCune and Grace, 2002). The length of an environmental vector is related to strength of the relationship between (a) the environmental variable that vector represents, and (b) the species community (i.e., each LHC belonging to L. griseus and H. sciurus). The order of LHC projection points with respect to an environmental vector (measured by “drawing” a perpendicular line from the projection point to the vector) corresponds to the rank- ing of LHC weighted averages with respect to that environmental factor (Jongman et al., 1995).

Results

Utilization of Seagrass Beds.—Recruitment of L. griseus and H. sciurus into Biscayne Bay seagrass beds was determined from temporal patterns in density and size (Fig. 2). First recruitment occurred during September–October (at the end of the spawning season) for L. griseus, when individuals averaging 7.8 mm TL were col- 480 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Figure 3. Seasonal delta-distribution mean estimators (DME ± 1 SE) belonging to Lutjanus griseus and Haemulon sciurus observed within five mangrove shoreline strata of southeastern Flor- ida. Strata abbreviations: IS: Islands; LB: Landbridges; LK: Leeward Key; ML: Mainland; WK: Windward Key. Note: y-axes are not of equal scale. FAUNCE AND SERAFY: NEARSHORE ONTOGENIC HABITAT SHIFTS BY TWO REEF FISHES 481

Figure 4. Seasonal proportion of juvenile, sub-adult, and adult Lutjanus griseus and Haemulon sciurus observed within mangrove shoreline strata within southeastern Florida. Strata abbreviations follow Figure 3. lected in densities of 1.90 1000 m–2. For H. sciurus, first recruitment occurred during February (at the middle of the spawning season), when the smallest individuals of the year (averaging 5.5 mm TL) were collected in densities of 9.93 1000 m–2. For both species, monthly average density declined and average length increased during the period between first recruitment and approximately 7–8 mo afterward. The greatest average length collected from seagrass beds was 11.4 mm TL for L. griseus and 10.8 mm TL for H. sciurus. Utilization of Mangrove Shorelines.—Density patterns among mangrove strata differed by season and life-history stage for both species. Relatively high densities of all L. griseus were observed among the IS and LB during the dry season, and within the LK and WK during the wet season (Fig. 3). Although juvenile L. griseus were largely restricted to Keys strata (i.e., the LK and WK) during both seasons, sub- adults and adults exhibited seasonal patterns similar to those observed for all fish. For example, relatively high densities of sub-adult and adult L. griseus were observed during the dry season within the center-strata, where IS > LB for sub-adults, and IS >> LB for adults. During the wet season the greatest density of sub-adults was observed within the LK, and that of adults was observed within the LB. Compared to patterns in L. griseus, relatively high densities of all H. sciurus were observed within Keys-strata during both seasons (Fig. 3). This trend was also observed for juveniles and sub-adults of this species. However, whereas juveniles were rarely observed within the ML, IS, and LB, densities of sub-adults within these strata were more comparable to those within Keys-strata. Spatial differences in H. sciurus seasonal density patterns were most evident for adults. For example, relatively high 482 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Figure 5. Size distributions of Lutjanus griseus and Haemulon sciurus observed within seagrass and mangrove shoreline habitats of southeastern Florida. Hashed areas depict range of monthly average size data for fishes collected from seagrass beds (depicted in Figure 3). Average (± 1 SE) sizes of fishes from each mangrove strata during the wet and dry seasons are plotted as a function of the average (± 1 SE) distance from the nearest oceanic inlet. Strata abbreviations follow Figure 3. For L. griseus, average size = 15.09 + 6.01[1 − e–0.36(distance from inlet)] (Adj. r2 = 0.33, P = 0.1002). For H. sciurus, average size = 10.26 + 11.05[1 − e–0.55(distance from inlet)] (Adj. r2 = 0.57, P = 0.0216). densities of adults were observed within the IS during the dry season and within the LK during the wet season (Fig. 3). These results demonstrate that utilization of western mangrove shorelines by L. griseus and H. sciurus was greater during the dry season compared to the wet, and was related to the ontogenetic stage of the individual: utilization of western mangrove shorelines was greatest for adults and least for juveniles. Because their patterns did not match, it is unlikely that juveniles were responsible for observed overall patterns of either L. griseus or H. sciurus among mangrove shorelines. Indeed, juveniles represented only a small portion of total fish observed during both seasons (Fig. 4). FAUNCE AND SERAFY: NEARSHORE ONTOGENIC HABITAT SHIFTS BY TWO REEF FISHES 483

Figure 6. Scree plot of non-metric multidimensional scaling eigenvalues and Monte Carlo test results. A lack of overlap between randomized data and real data allows rejection of the null hypothesis that the eigenvalues resulted from pure chance. From this plot a final solution with two axes was selected (point of “elbow”). Similar trends in the mean sizes of L. griseus and H. sciurus were observed. With the exception of the WK for H. sciurus, the average size of these species was always greater along mangroves than within seagrass beds (Fig. 5). Among mangrove shorelines, average size was always lowest within the WK and increased with average distance from oceanic inlet, where: L. griseus average size = 15.09 + 6.01[1 − e–0.36(DI)] (Adj. r2 = 0.33, P = 0.10); and H. sciurus average size = 10.26 + 11.05[1 − e–0.55(DI)] (Adj. r2 = 0.57, P = 0.02). Structure of LHC Assemblage.—The NMS scree plot revealed a reduction in the amount of final stress accounted for by each dimension after dimension 2, and that the amount of stress for the initial dimensions was significantly different than expected by chance (P = 0.02, Fig. 6). Therefore, a two-dimensional solution was selected for final NMS. This solution reached tolerance after 12 iterations, had a stress value of 6.76, and accounted for 96.5% of the variance in the original data set. The first dimension accounted for 80.1% of the original variance and separated the LHC assemblage among strata, whereas the second dimension accounted for 15.6% of the variance and separated the assemblage according to LHC (Fig. 7). Relationships Between LHC Structure and Environmental Factors.— Canonical correspondence analysis generated three canonical axes from continuous environmental variables that accounted for 21.2% of the variation in the species community; axis 1 accounted for 81.6% of the explained variance, axis 2 accounted for an additional 16.7%, and axis 3 accounted for 1.7% (Table 2). Monte Carlo tests revealed that the eigenvalues and species-environment correlations were significantly greater than expected by chance for the first two axes P( = 0.01). Because axis 3 accounted for a low proportion of the cumulative variance explained by environmental vari- 484 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Figure 7. Non-metric multidimensional scaling biplot depicting spatiotemporal relationships among species and life stages. Axis 1 separates strata (open circles), while the second axis separates life-history components (solid circles). Stratum labels are in the form season_strata_year, where D: dry season; W: wet season; ML: mainland, IS: island; LB: landbridge; LK: leeward key; WK: windward key; 00: 2000; 01: 2001. Life-history component labels are in the form life-history stage_species, where JUV: Juvenile; SUB: Sub-adult; MAT: Adult; G: Lutjanus griseus; S: Haemulon sciurus.

Table 2. Results of canonical correspondence analysis (CCA). * = significant. ns: non-significant

Axis 1 2 3 Total inertia 1.4408 Eigenvalue Real data 0.250 0.051 0.005 Monte Carlo randomized data (100 iterations) Minimum 0.008 0.004 0.001 Mean 0.021 0.009 0.004 Maximum 0.052 0.020 0.011 P (0.01) .* .* .ns Cumulative percentage variance of species data 17.4 20.9 21.2 of species-environment relation 81.6 98.3 100 Species-environment correlations Pearson 0.660 0.416 0.158 Monte Carlo randomized data (100 iterations) Minimum 0.131 0.105 0.072 Mean 0.226 0.177 0.132 Maximum 0.352 0.267 0.205 P (0.01) .* .* .ns FAUNCE AND SERAFY: NEARSHORE ONTOGENIC HABITAT SHIFTS BY TWO REEF FISHES 485

Figure 8. Biplot of canonical correspondence analysis ordination depicting relationships between environmental factors and three life-history stages of Lutjanus griseus and Haemulon sciurus. WOC: Width of canopy. Life-history stages are separated by factors strongly correlated with axis 1 whereas the species are separated by factors strongly correlated with axis 2 (see Table 2 for details). Correlations between individual environmental vectors and each axis is related to the length and direction of the vector; the length of vector is related to the strength of correlation, whereas the direction of the vector depicts which axis that vector is most related to.

ables and failed the Monte Carlo test, further consideration of results will be limited to the first two axes. The relationships between each LHC and the environmental variables are depicted in a species-environment biplot (Fig. 8). Whereas the LHC of L. griseus and H. sciurus were separated along axis 1, the origin effectively separated these species from one another along axis 2. The factors distance index and water depth were important correlates of the first two axes (Table 3), and LHC of both species were ordered juveniles < sub-adults < adults along corresponding vectors. Differences in the mean position of each LHC were less influenced by salinity, whereas separation of species was less influenced by latitude and width of canopy (Table 3, Fig. 8).

Discussion

The ontogenetic utilization of seagrass beds and mangrove shorelines byL . griseus and H. sciurus consist of three phases: (1) settlement into seagrass beds, (2) movement or expansion to mangrove habitats, and (3) increasing inland utilization with size. 486 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Table 3. Weighted inter-set correlations between environmental variables and CCA axes. Correla- tions > 0.50 are underlined.

Canonical axis Environmental variable 1 2 3 Temperature 0.185 0.281 −0.291 Salinity 0.321 −0.159 −0.873 Depth −0.630 −0.555 −0.367 Width of canopy −0.134 −0.373 0.030 Distance index −0.711 0.579 0.336 Latitude 0.104 0.416 −0.090

Utilization of Seagrass Beds.—Peak recruitment of L. griseus and H. sciurus into seagrass beds was observed during the fall (September–October) and winter (February), respectively, and residence by these fishes was indicated by a steady increase in average size during the subsequent 8 mo. The timing and size that these species recruit to seagrass beds are corroborated by other available data from the region. Working on the Atlantic Ocean side of Lower Matecumbe Key (24˚51´N, 88˚44´W), Springer and McErlean (1962) observed the smallest L. griseus and H. sciurus in monthly samples during August–September and December, respectively. In addition to an earlier timing of recruitment, these authors observed much smaller individuals than in the present study. Gear differences between studies are likely at play here; Springer and McErlean (1962) utilized push nets (~1 mm mesh) in addition to seines (10 mm mesh), but in the present study trawls (10 mm mesh) were utilized. The size of individuals (~15 mm SL) captured in push nets by Springer and McErlean (1962) closely matched the proposed size of ingress and settlement for these species (Lindeman et al., 2001; Tzeng et al., 2003). From their data, offshore spawning and seagrass recruitment occurred 1–2 mo earlier than indicated in the present study, i.e., during August–September for L. griseus, and during December–January for H. sciurus. Therefore, although the residence time of new recruits within Biscayne Bay seagrass beds was estimated at approximately 8 mo (as suggested from monthly average size plots), the age of these fishes was likely closer to 10 mo; thus exclusive use of seagrass beds is probably restricted to the first year of life for the fishes studied. Utilization of Mangrove Shorelines.—Lutjanus griseus and H. sciurus above a certain size also utilize mangrove shoreline habitats. The maximum size of these fishes within seagrass beds matched the smallest average size of individuals within mangrove shorelines, which strongly suggests that mangroves are utilized as a secondary or sequential habitat by these species. Data from prior studies in south Florida substantiate that larger snappers and grunts inhabit mangroves than within seagrass beds (Serafy et al., 1997, 2003; Ley et al., 1999). Data from the present study indicate that a shift from seagrass to mangroves occurs after approximately 8–10 mo and at a size of 10.5–12 cm TL. Because the “fingerbars” used on rollerframe trawls exclude fishes above a certain size, it is alternatively possible that gear selectivity was responsible for observed size differences between habitats. However, previous studies using identical gear types in both mangroves and seagrass habitats have reported size differences of L. griseus and H. sciurus similar to those reported here (Cocheret de la Morinière et al., 2002). FAUNCE AND SERAFY: NEARSHORE ONTOGENIC HABITAT SHIFTS BY TWO REEF FISHES 487

That L. griseus and H. sciurus within mangrove shorelines are larger than those within seagrass beds does not necessarily indicate that mangroves are exclusively utilized by fish above a certain size. Rather, these data likely represent an expansion of habitat breadth with ontogeny (Beck et al., 2001; Cocheret de la Morinière et al., 2002; Serafy et al., 2003). In particular, mangroves appear to supplant the role of predator refugia for reef fishes first provided by seagrasses. The shade and structurally heterogeneous root habitat provided by mangroves reduces the visual effectiveness of larger-bodied predators as well as their ability to capture prey (Helfman, 1981; Laegdsgaard and Johnson, 2001; Cocheret de la Morinière et al., 2004). The importance of mangroves as predator refugia is further evidenced by behavioral and stable isotope studies that demonstrated mangroves are primarily utilized as “daytime resting sites” while seagrass beds are utilized as nocturnal foraging grounds (Rooker and Dennis, 1991; Vega-Cendejas et al., 1994; Nagelkerken et al., 2000a; Val- dés-Muñoz and Mochek, 2001). Expansion of habitat utilization with ontogeny not only occurs between seagrass and mangrove habitats, but among spatially separated mangrove shorelines as well. These patterns of habitat utilization can be remarkably similar between species. In the present study, the NMS analysis of the LHC community separated the ML from other strata along the first axis, and this axis accounted for 80% of the variance. Analysis of stratum-specific densities revealed that L. griseus and H. sciurus were much lower in the ML stratum compared to the other mangrove shorelines studied. Therefore, the first NMS axis represented a “filter” for the presence of these fishes, whereas the second axis separated the data points based on their densities. Separa- tion along the second axis resulted in groupings by LHC and not by species, indicating that these two fishes exhibited similar ontogenetic habitat utilization patterns. That L. griseus and H. sciurus utilize more westward and inland mangrove shorelines with ontogeny was supported by two sources of data. First, juvenile density and contribution to the total number of both species declined from the WK (the easternmost strata) to the LK (the adjacent strata to the west), and additional western and inland strata were almost exclusively utilized by sub-adults and adults. Second, average size increased among mangrove shoreline strata in a westward and inland direction during both seasons. Only one other study has examined the relationship between the size of reef fish within a coastal embayment and geographic distance to the bay/ocean interface in the western Atlantic. Nagelkerken et al. (2000b) found a positive relationship between these metrics for Acanthurus chirurgus (Bloch, 1787), Scarus iserti (Bloch, 1789), and Haemulon flavolineatum (Desmarest, 1823) inhabiting seagrass beds at scales up to 3 km in Curaçao, and concluded that the stability of such patterns must be tested in further studies. The present study demonstrates that positive relationships between body size and distance to inlet are also evident for L. griseus and H. sciurus inhabiting mangrove shorelines at scales > 10 km. Relationships Among LHC and Environmental Factors.—Martino and Able (2003) suggested that patterns in estuarine fish assemblage structure across large spatial scales (i.e., > 10 km) are due to individual species responses to dominant environmental gradients, whereas smaller-scale (≤ 1 km) patterns are due to individual behaviors. Previous analyses of coastal fish assemblages using CCA have found that distance from inlet and either salinity (Kupschus and Tremain, 2001; Martino and Able, 2003), temperature (Hajisamae and Chou, 2003), or depth (Araujo et al., 2002) controlled species composition and structure across broad spatial scales. In the 488 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

present study, CCA revealed that distance from inlet and water depth were primarily responsible for the separation of LHC of L. griseus and H. sciurus, and that these species were separated to a lesser extent by latitude, salinity, temperature, and width of canopy. Following Martino and Able (2003), the present CCA results demonstrate a hierarchy of factors associated with the distribution of fishes. Whereas prior studies have shown that dominant environmental gradients influence the composition of entire fish assemblages over broad spatial scales, this study is the first to demonstrate that such gradients also result in the geographic sorting of the ontogenetic stages of two reef fishes. Water depth, rather than temperature or salinity, represented the dominant environmental spatial gradient in the present study. Temperatures may not have reached extremes to exclude fishes from mangroves, and the development of the South Dade Conveyance System has resulted in a reduced salinity gradient from west to east in the region studied (Light and Dineen, 1994). With the exception of the WK during the dry season, water depth tended to decrease among strata from east to west during both seasons (i.e., exhibited a gradient at large spatial scales), but was not significantly correlated with distance from inlet at the level of individual samples (weighted Pearson Correlation Coefficient −0.02). Water depth has an influence on the amount of coastal vegetated habitat available to fishes at numerous scales. At large scales, depth may be correlated with more meaningful metrics related to wetland area. For example, Lee (2004) demonstrated that the extent of intertidal areas and organic matter availability (as represented by tidal amplitude) have a stronger influence on prawn catch from mangroves than mangrove area per se. At smaller scales, water depth dictates the vertical area available to fishes. For example, Faunce et al. (2004) noted a positive relationship between fish size and water depth, and it has been suggested that shallow depths regulate predation on small fishes by excluding their potential predators (Vance et al., 1996; Rönnbäck et al., 1999; but see Sheaves, 2001). Corroborating evidence that water depth sorts fishes by body size comes from the present CCA that showed adults resided in the deepest waters, sub-adults in moderately deep waters, and juveniles in the shallowest waters. Latitudinal gradients also played a part in separating the species in the CCA biplot. This analysis suggested thatL . griseus had greater associated latitudes than H. sciurus. The waters of Biscayne Bay comprise the higher latitudes of the southeastern Florida study area. Expressing the number of L. griseus and H. sciurus observed within Bis- cayne Bay as a proportion of the total number of these species (within southeastern Florida) revealed that proportions within Biscayne Bay were substantially greater for L. griseus than for H. sciurus (Table 4). Juvenile L. griseus and H. sciurus within the LK of Biscayne Bay accounted for a respective 0.83–1.00 and 0.70–0.86 of the total number of juveniles within southeastern Florida. Furthermore, the proportion of total fish observed within the LK of Biscayne Bay was greater for juveniles than for sub-adults and least for adults, and this pattern was observed for both species during both seasons. These data lend additional support to the earlier finding that more eastern located mangrove strata are utilized earlier in fish ontogeny. Because the ML and LK strata within Biscayne Bay represent 34.7% and 42.1% of the total 2-m wide shoreline area of their respective strata within the study area, the proportion of reef fishes, especially juveniles, observed within Biscayne Bay is substantial. It is likely that the comparatively large opening south of Key Biscayne (i.e., the “Safety Valve”) led to increased water exchange, westward net flow, and ulti- FAUNCE AND SERAFY: NEARSHORE ONTOGENIC HABITAT SHIFTS BY TWO REEF FISHES 489

Table 4. Proportion of the total Lutjanus griseus and Haemulon sciurus life-history components of southeastern Florida observed within the mainland (ML) and leeward key (LK) portions of Biscayne Bay.

Strata ML LK Species Lutjanus griseus Haemulon sciurus Lutjanus griseus Haemulon sciurus Dry season Juvenile 0.00 No fish 0.83 0.70 Sub-adult 0.04 0.00 0.59 0.28 Mature 0.00 0.00 0.56 0.06

Wet season Juvenile 1.00 No fish 1.00 0.86 Sub-adult 0.92 0.90 0.59 0.60 Mature 0.77 1.00 0.58 0.25

mately increased utilization of Biscayne Bay mangrove shoreline habitats by juvenile reef fishes. This topic warrants further research in the region, such as the biophysical simulation modeling approach used by Ault et al. (1999b) and Wang et al. (2003) for pink shrimp (Farfantepenaeus duorarum Burkenroad, 1939) and spotted seatrout (Cynoscion nebulosus Cuvier, 1830), and highlights the influence of bay-ocean connectivity on local recruitment patterns. In addition to latitude, width of canopy influenced separation the two reef species in the CCA. Width of canopy was expected to be an important habitat metric because it reflects the structural complexity and extent of shelter provided by mangrove root systems. This complexity in turn has been shown to provide shading and visual advantages for forage fishes H( elfman, 1981), and to reduce predator efficacy (Primavera, 1997; Laegdsgaard and Johnson, 2001). Ley and McIvor (2001) used CCA to examine habitat relationships between several species of reef fishes, including L. griseus and H. sciurus within the LB stratum of the present study. They found that for H. sciurus, mangrove tree height and water depth were significantly correlated with the first axis, whereas for L. griseus, mangrove fringe width, height, prop-root density, and water depth were significantly correlated with the second axis. Results of the present study closely resemble those of Ley and McIvor (2001) with the exception that we examined different life-history components. Both studies demonstrated a separation of species according to gradients in habitat complexity including mangrove fringe width. In addition, results from these field surveys agree with the experimental manipulations of Cocheret de la Morinière et al. (2004) and Verweij et al. (2006) that show abundances of fishes including H. sciurus increase as a function of both habitat complexity and shading.

Summary and Implications

The structure and composition of coastal fish assemblages are thought to be a function of individual species’ responses to environmental gradients that operate at multiple spatial scales. Although several studies have examined the relationship between the entire fish assemblage and environmental gradients, such an approach is less likely to capture individual species’ responses as they change with ontogeny. In the present study, two ontogenetic patterns of habitat utilization were exhibited by L. 490 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

griseus and H. sciurus within southeastern Florida. First, size-distributions averaging > 13 cm TL were exclusively observed during daytime within mangrove habitats, whereas smaller fish were commonly collected within seagrass beds. Such size discrepancies between habitats are likely due to the inability of seagrasses to effectively shelter comparatively large fish from predation (Jackson et al., 2001). Second, an increased utilization of inland mangrove shoreline strata was observed with ontogenic development. Such patterns are likely due to an expansion of home range and mobil- ity with body size (e.g., Kramer and Chapman, 1999). As such, the increase in mean size observed here as a function of distance from inlet likely represents the exclusion of smaller bodied fishes, rather than habitat resource selection by larger individuals. One application of these data is that they may explain observed fish distributions within Florida estuarine “no-take” reserves. Two such reserves exist on the east coast of Florida, the Crocodile Sanctuary of Everglades National Park (CS-ENP), and the No Take Zone of Merritt Island National Wildlife Refuge (NTZ-MINWR). Stud- ies conducted within these reserves have noted a high proportion of large-bodied individuals relative to surrounding waters (Johnson et al., 1999; Faunce et al., 2002). Recently, Tremain et al. (2004) evaluated tag-return data from the Indian River La- goon System and the NTZ-MINWR and concluded that a greater percentage of fish immigrated to the no-take zone than emigrated from it. Distances between these reserves and permanently accessible oceanic inlets range from 27–29 km for the CS- ENP to 75–100 km for the NTZ-MINWR, and are substantially larger than the 10 km scale commonly cited for fishes that have spatial separation of juvenile and adult populations (Gillanders et al., 2003; Pittman and McAlpine, 2003). For these reasons, the relatively large size of fishes observed within these coastal no-take reserves relative to surrounding areas may partially be the result of differences in motility, i.e., only larger-bodied fishes with greater home ranges and migratory capabilities can access the remotely located estuarine reserves, or similarly, the time it takes to slowly reach these locations begets a larger body size upon arrival. While the mechanism by which larger fish occur within remote inland waters remains unclear, it seems necessary to assess the contribution of these fish towards the overall spawning population. Whether larger fishes present in remote bays or estuaries seasonally emigrate to join the spawning population elsewhere, spawn within remote areas without emigration, or merely subsist without spawning or emigration should be considered by decision makers as to whether such areas remain no-take and/or no-access. A second application of the present study is that it provides the basis for prioritization of areas containing the same habitat type (seagrass or mangrove) for preservation/protection, as in the United States, there are inadequate resources to fully manage and protect all habitats all of the time. From our results, habitat value (as defined by daytime density values) for L. griseus and H. sciurus during ontogeny is a function of distance from oceanic inlet, with seagrass beds being the primary habitat for fish during their first year, and mangroves thereafter. The expansion of inland habitat utilization with ontogeny observed in the present study may be either accompanied by an equivalent range expansion eastward and offshore, or be followed by a rapid long-distance offshore migration facilitated by a large body size. A steady increase in average size with distance from inlet eastward towards the coral reef would support the first scenario. Such data are lacking for L. griseus. However, larger mean sizes of H. sciurus have been collected from inshore reefs compared to mangrove shorelines (Acosta, 1997), suggesting that this fish likely FAUNCE AND SERAFY: NEARSHORE ONTOGENIC HABITAT SHIFTS BY TWO REEF FISHES 491 undergoes regular seasonal movements into and out of coastal embayments. In contrast, L. griseus appear to remain within mangrove habitats for an extended period, as a greater proportion of mature-sized individuals have been recorded within mangrove habitats than within adjacent coral reefs (Nagelkerken et al., 2000b; Cocheret de la Morinière et al., 2002; Serafy et al., 2003). Indeed, new tagging efforts using acoustic telemetry show that circular movements of gray snapper between offshore reefs and inshore mangroves can occur in a matter of days (Luo et al., 2006). Evi- dence from these studies demonstrates that although H. sciurus and L. griseus utilize inshore coastal vegetated habitats, their occupancy is not homogeneous, but is location-, season-, and life-history stage-dependent. In addition, seasonal forays between inshore and offshore locations may be of widely different durations between species, lasting longer for H. sciurus than for L. griseus, which may be considered more of a coastal species than a reef species.

Acknowledgments

An earlier version of this manuscript was originally submitted in partial fulfillment of a doctoral dissertation by C. Faunce (RSMAS, University of Miami). We thank O. Becerio, J. Luo, G. Schmalzle, and D. Jones for help with analyses, and J. Kool for his GIS assistance and valuable discussions regarding landscape ecology. Primary funding was provided by the Perry Institute for Marine Science, Environmental Defense, the United States Geological Survey (Biological Resources Division, Restoration Ecology Branch) and the National Park Service (Biscayne National Park). Additional funding was provided by the American Institute for Ma- rine Studies, the Florida Keys Audubon Society, and the Captain Bob Lewis and Paul Kopp Memorial scholarships awarded through the University of Miami.

Literature Cited

Acosta, A. 1997. Use of multi-mesh gillnets and trammel nets to estimate fish species composition in coral reef and mangroves in the southwest coast of Puerto Rico. Caribb. J. Sci. 33: 45–57. Araujo, F. G., M. C. C. De Azevedo, M. A. Silva, A. L. M. Pessanha, I. D. Gomes, and A. G. Da Cruz-Filho. 2002. Environmental influences on the demersal fish assemblages in the Sepe- tiba Bay, Brazil. Estuaries 25: 441–450. Ault, J. S., G. A. Diaz, S. G. Smith, J. Luo, and J.E. Serafy. 1999a. An efficient sampling survey design to estimate pink shrimp population abundance in Biscayne Bay, Florida. North Am. J. Fish. Manage. 19: 696–712. ______, J. G. Luo, S. G. Smith, J. E. Serafy, J. D. Wang, R. Humston, and G. A. Diaz. 1999b. A spatial dynamic multistock production model. Can. J. Fish. Aquat. Sci. 56: 4–25. Bardach, J. E. 1959. The summer standing crop of fish on a shallow Bermuda reef. Limnol. Oceanogr. 4: 77–85. Beck, M. W., K. L. Heck, K. W. Able, D. L. Childers, D. B. Eggleston, B. M. Gillanders, B. Halp- ern, C. G. Hays, K. Hoshino, T. J. Minello, R. J. Orth, P. F. Sheridan, and M. R. Weinstein. 2001. The identification, conservation, and management of estuarine and marine nurseries for fish and invertebrates. BioScience 51: 633–641. Bell, J. D., G. C. S. Craik, D. A. Pollard, and B. C. Russell. 1985. Estimating length frequency- distributions of large reef fish underwater. Coral Reefs 4: 41–44. Billings, V. C. and J. L. Munro. 1974. The biology, ecology, exploitation and management of Caribbean reef fishes. Scientific report of the ODA/UWI fisheries ecology research project, Port Royal Marine Laboratory, Jamaica, 1969–1973. Part 5e. The biology, ecology and bio- 492 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

nomics of Caribbean reef fishes: Pomadasyidae (grunts). Res. Rep. Zool. Dept., Univ. West Indies. 128 p. Blaber, S. J. M. 2002. ‘Fish in hot water’: the challenges facing fish and fisheries research in tropical estuaries. J. Fish. Biol. 61: 1–20. Cocheret de la Morinière, E., I. Nagelkerken, H. van der Meij, and G. van der Velde. 2004. What attracts juvenile coral reef fish to mangroves: habitat complexity or shade? Mar. Biol. 144: 139–145. ______, B. J. A. Pollux, I. Nagelkerken, and G. van der Velde. 2002. Post- settlement life cycle migration patterns and habitat preference of coral reef fish that use seagrass and mangrove habitats as nurseries. Estuar. Coast. Shelf Sci. 55: 309–321. Cochran, W. G. 1977. Sampling Techniques. Wiley, New York. 448 p. Day, J. H. 1951. The ecology of South African estuaries Part I. General considerations. Trans. Royal Soc. S. Afr. 33: 53–91. Day Jr., J. W., C. A. S. Hall, W. M. Kemp, and A. Yáñez-Arancibia. 1989. Estuarine ecology. Wiley Interscience, New York. 558 p. Domeier, M. L., C. Koenig, and F. Coleman. 1996. Reproductive biology of the gray snapper (Lutjanus griseus), with notes on spawning for other western Atlantic snappers (Lutjani- dae). Pages 189–201 in F. Arreguin-Sanchez, J. L. Munro, M. C. Blagos, and D. Pauly, eds. Biology, fisheries and culture of tropical snappers and groupers. ICLARM Conf. Proc. Intl. Ctr. Liv. Aquat. Res. Mgt., Manila. 48. Dorenbosch, M., M. C. van Riel, I. Nagelkerken, and G. van der Velde. 2004. The relationship of reef fish densities to the proximity of mangrove and seagrass nurseries. Estuar. Coast. Shelf Sci. 60: 37–48. Faunce, C. H. and J. E. Serafy. 2006. Mangroves as fish habitat: 50 years of field studies. Mar. Ecol. Prog. Ser. 318: 1–18. ______, J. E. Serafy, and J. J. Lorenz. 2004. Density-habitat relationships of mangrove creek fishes within the southeastern saline Everglades (USA), with reference to managed freshwater releases. Wetl. Ecol. Mgt. 12: 377–394. ______, J. J. Lorenz, J. A. Ley, and J. E. Serafy. 2002. Size structure of gray snapper (Lut- janus griseus) within a mangrove ‘no-take’ sanctuary. Bull. Mar. Sci. 70: 211–216. García-Arteaga, J. P. 1992. Edad y crecimiento del ronco amarillo, Haemulon sciurus (Shaw) (Pisces: Haemulidae) en la plataforma suroccidental de Cuba. Cienc. Biol. 25: 104–116. García-Cagide, A., R. Claro, and B. V. Koshelev. 2001. Reproductive patterns of fishes of the Cuban shelf. Pages 73–114 in R. Claro, K. C. Lindeman, and L. R. Parenti, eds. Ecology of the marine fishes of Cuba. Smithsonian Institution Press, Washington. Gillanders, B. M., K. W. Able, J. A. Brown, D. B. Eggleston, and P. F. Sheridan. 2003. Evidence of connectivity between juvenile and adult habitats for mobile marine fauna: an important component of nurseries. Mar. Ecol. Prog. Ser. 247: 281–295. Gunter, G. 1967. Some relationships of estuaries to the fisheries of the Gulf of Mexico. Pages 621–638 in G. H. Lauff, ed. Estuaries. Am. Assoc. Adv. Sci., Washington, D.C. Hajisamae, S. and L. M. Chou. 2003. Do shallow water habitats of an impacted coastal strait serve as nursery grounds for fish? Estuar. Coast. Shelf Sci. 56: 281–290. Halpern, B. 2004. Are mangroves a limiting resource for two coral reef fishes? Mar. Ecol. Prog. Ser. 272: 93–98. Helfman, G. S. 1981. The advantage to fishes of hovering in shade. Copeia 198: 392–400. Jackson, E. L., A. A. Rowden, M. J. Attrill, S. J. Bossey, and M. B. Jones. 2001. The importance of seagrass beds as a habitat for fishery species. Oceanogr. Mar. Biol. 39: 269–303. Johnson, D. R., N. A. Funicelli, and J. A. Bohnsack. 1999. Effectiveness of an existing estuarine no-take fish sanctuary within the Kennedy Space Center, Florida. North Am. J. Fish. Man- age. 19: 436–453. Jongman, R. H. G., C. J. F. ter Braak, and O. F. R. Van Tongeren. 1995. Data Analysis in Com- munity and Landscape Ecology. Cambridge University Press, Cambridge. 299 p. FAUNCE AND SERAFY: NEARSHORE ONTOGENIC HABITAT SHIFTS BY TWO REEF FISHES 493

Kramer, D. L. and M. R. Chapman. 1999. Implications of fish home range size and relocation for marine reserve function. Environ. Biol. Fish. 55: 65–79. Kupschus, S. and D. Tremain. 2001. Associations between fish assemblages and environmental factors in nearshore habitats of a subtropical estuary. J. Fish. Biol. 58: 1383–1403. Laegdsgaard, P. and C. Johnson. 2001. Why do juvenile fish utilize mangrove habitats? J. Exp. Mar. Biol. Ecol. 257: 229–253. Lee, S. Y. 2004. Relationship between mangrove abundance and tropical prawn production: a re-evaluation. Mar. Biol. 145: 943–949. Legendre, L. and P. Legendre 1983. Numerical Ecology. Elsevier, New York. 853 p. Ley, J. A. and C. C. McIvor. 2001. Linkages between estuarine and reef fish assemblages: enhancement by the presence of well-developed mangrove shorelines. Pages 539–562 in J. W. Porter and K. G. Porter, eds. The Everglades, Florida Bay, and coral reefs of the Florida Keys: an ecosystem sourcebook. CRC Press, Boca Raton. ______, ______, and C. L. Montague. 1999. Fishes in mangrove prop-root habitats of northeastern Florida Bay: distinct assemblages across an estuarine gradient. Estuar. Coast. Shelf Sci. 48: 701–723. Light, S. S. and J. W. Dineen. 1994. Water control in the Everglades: a historical perspective. Pages 47–84 in S. M. Davis and J. C. Ogden, eds. Everglades: the ecosystem and its restoration. St. Lucie Press, Boca Raton. Lindeman, K. C., T. N. Lee, W. D. Wilson, R. Claro, and J. S. Ault. 2001. Transport of larvae originating in southwest Cuba and the Dry Tortugas: evidence for partial retention in grunts and snappers. Proc. Gulf Caribb. Fish. Inst. 52: 732–747. Luo, J., S. Sponaugle, J. E. Serafy, and J. J. Lorenz. 2006. Multiple habitat utilization by a coastal fish: diel, seasonal, and ontogenetic movement of gray snapper (Lutjanus griseus), Florida Sea Grant Final Report, Project Number R/C-RE-48. 96 p. Manooch III, C. S. and R. H. Matheson III. 1981. Age, growth, and mortality of gray snapper collected from Florida waters. Pages 331–344 in Proc. of the 35th Ann. Conf. Southeastern Assoc. Fish and Wildlife Agencies, Tulsa. Martino, E. J. and K. W. Able. 2003. Fish assemblages across the marine to low salinity transition zone of a temperate estuary. Estuar. Coast. Shelf Sci. 56: 969–987. McCune, B and J. B. Grace. 2002. Analysis of Ecological Communities. MjM Software Design, Gleneden Beach. 300 p. McHugh, J. L. 1967. Estuarine Nekton. Pages 581–620 in G. H. Lauff, ed. Estuaries. American Association for the Advancement of Science, Washington, D. C. Meester, G. A., J. S. Ault, and J. A. Bohnsack. 1999. Visual censusing and the extraction of average length as an indicator of stock health. Naturalista Siciliana 23: 205–222. Minchin, P. R. 1987. An evaluation of the relative robustness of techniques for ecological ordination. Vegetatio 69: 89–107. Mumby, P. J., A. J. Edwards, J. E. Arias-Gonzales, K. C. Lindeman, P. G. Blackwell, A. Gall, M. I. Gorczynska, A. R. Harborne, C. L. Pescod, H. Renken, C. C. C. Wabnitz, and G. Llewellyn. 2004. Mangroves enhance the biomass of coral reef fish communities in the Caribbean. Nature 427: 533–536. Munro, J. L., V. C. Gaut, R. Thompson, andP . H. Reeson. 1973. The spawning seasons of Carib- bean reef fishes. J. Fish Biol. 5: 69–84. Nagelkerken, I. and G. van der Velde. 2002. Do non-estuarine mangroves harbour higher densities of juvenile fish than adjacent shallow-water and coral reef habitats in Curacao (Nether- lands Antilles)? Mar. Ecol. Prog. Ser. 245: 191–204. ______, M. Dorenbosch, W. C. E. P. Verberk, E. Cocheret de la Morinière, and G. van der Velde. 2000a. Day-night shifts of fishes between shallow-water biotopes of a Caribbean bay, with emphasis on the nocturnal feeding of Haemulidae and Lutjanidae. Mar. Ecol. Prog. Ser. 194: 55–64. ______, ______, ______, ______, and ______. 2000b. Importance of shallow-water biotopes of a Caribbean bay for juvenile coral 494 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

reef fishes: patterns in biotope association, community structure and spatial distribution. Mar. Ecol. Prog. Ser. 202: 175–192 ______, C. M. Roberts, G. van der Velde, M. Dorenbosch, M. C. van Riel, E. Cocheret de la Morinière, and P. H. Nienhuis. 2002. How important are mangroves and seagrass beds for coral-reef fish? The nursery hypothesis tested on an island scale. Mar. Ecol. Prog. Ser. 244: 299–305. Ogden, J. C. 1997. Ecosystem interactions in the tropical coastal seascape. Pages 288–297 in C. Birkeland, ed. Life and Death of coral reefs. Chapman & Hall, New York. Parrish, J. D. 1989. Fish communities of interacting shallow-water habitats in tropical oceanic regions. Mar. Ecol. Prog. Ser. 58: 143–160. Pennington, M. 1983. Efficient estimators of abundance, for fish and plankton surveys. Biomet- rics 39: 281–286. Pittman, S. J. and C. A. McAlpine. 2003. Movements of marine fish and decapod crustaceans: process, theory and application. Adv. Mar. Biol. 44: 205–294. Primavera, J. H. 1997. Fish predation on mangrove-associated penaeids: the role of structures and substrate. J. Exp. Mar. Biol. Ecol. 215: 205–216. Rönnbäck, P., M. Troell, N. Kautsky, and J. H. Primavera. 1999. Distribution pattern of shrimps and fish among Avicennia and Rhizophora microhabitats in the Pagbilao Mangroves, Phil- ippines. Estuar. Coast. Shelf Sci. 48: 223–234. Rooker, J. R. and G. D. Dennis. 1991. Diel, lunar and seasonal changes in a mangrove fish assemblage off southwestern Puerto Rico. Bull. Mar. Sci. 49: 684–698. Serafy, J. E., C. H. Faunce, and J. J. Lorenz. 2003. Mangrove shoreline fishes of Biscayne Bay, Florida. Bull. Mar. Sci. 72: 161–180. ______, K. C. Lindeman, T. E. Hopkins, and J. S. Ault. 1997. Effects of freshwater canal discharge on fish assemblages in a subtropical bay: field and laboratory observations. Mar. Ecol. Prog. Ser. 160: 161–172. Sheaves, M. J. 2001. Are there really few piscivorous fishes in shallow estuarine habitats? Mar. Ecol. Prog. Ser. 222: 279–290. Springer, V. G. and A. J. McErlean. 1962. Seasonality of fishes on a south Florida shore. Bull. Mar. Sci. 12: 39–60. St. John, J., G. R. Russ, and W. Gladstone. 1990. Accuracy and bias of visual estimates of numbers, size structure and biomass of a coral reef fish. Mar. Ecol. Prog. Ser. 64: 253–262. Stark II, W. E. 1970. Biology of the gray snapper, Lutjanus griseus (Linnaeus), in the Florida Keys. Pages 11–150 in W. E. Stark II and R. E. Schroeder, eds. Investigations on the gray snapper, Lutjanus griseus. University of Miami Press, Coral Gables. ter Braak, C. J. F. 1986. Canonical correspondence analysis: a new eigenvector technique for multivariate direct gradient analysis. Ecology 67: 1167–1179. Tremain, D. M., C. W. Harnden, and D. H. Adams. 2004. Multidirectional movements of sportfish species between an estuarine no-take zone and surrounding waters of the Indian River Lagoon, Florida. Fish. Bull. 102: 533–544. Tzeng, M. W., J. A. Hare, and D. G. Lindquist. 2003. Ingress of transformation stage gray snapper, Lutjanus griseus (Pisces: Lutjanidae) through Beaufort Inlet, North Carolina. Bull. Mar. Sci. 72: 891–908. United States Department of Commerce (USDOC). 1996. Magnusen-Stevens Fishery Conser- vation and Management Act as amended through October 11, 1996., National Oceanic and Atmospheric Administration Technical Memorandum NMFS-F/SPO-23. 121 p. Valdés-Muñoz, E, and A. D. Mochek. 2001. Behavior of marine fishes of the Cuban shelf.P ages 73–114 in R. Claro, K. C. Lindeman, L. R. Parenti, eds. Ecology of the marine fishes of Cuba. Smithsonian Institution Press, Washington. Vance, D. J., M. D. E. Haywood, D. S. Heales, R. A. Kenyon, N. R. Loneragan, and R. C. Pendrey. 1996. How far do prawns and fish move into mangroves? Distribution of juvenile banana prawns Penaeus merguiensis and fish in a tropical mangrove forest in northern Australia. Mar. Ecol. Prog. Ser. 131: 115–124. FAUNCE AND SERAFY: NEARSHORE ONTOGENIC HABITAT SHIFTS BY TWO REEF FISHES 495

Vega-Cendejas, M. E., U. Ordóñez, and M. Hernández. 1994. Day–night variation of fish population in the mangrove of Celestun lagoon, Mexico. Intl. J. Ecol. Env. Sci. 20: 99–108. Verweij, M. C., I. Nagelkerken, D. de Graaff, M.P eeters, E. J. Bakker, and G. van der Velde. 2006. Structure, food and shade attract juvenile coral reef fish to mangrove and seagrass habitats: a field experiment. Mar. Ecol. Prog. Ser. 306: 257–268. Wang, J. D., J. G. Luo, and J. S. Ault. 2003. Flows, salinity, and some implications for larval transport in South Biscayne Bay, Florida. Bull. Mar. Sci. 72: 695–723.

Addresses: (C.H.F.) Florida Fish and Wildlife Research Institute, Tequesta Field Labora- tory, 19100 SE Federal Highway, Tequesta, Florida 33469. (J.E.S.) Southeast Fisheries Science Center, National Marine Fisheries Service,75 Virginia Beach Drive, Miami, Florida 33149. Corresponding Author: (C.H.F.) E-mail: . BULLETIN OF MARINE SCIENCE, 80(3): 497–512, 2007

RELATIVE IMPORTANCE OF MANGROVES AS FEEDING HABITATS FOR FISHES: A COMPARISON BETWEEN MANGROVE HABITATS WITH DIFFERENT SETTINGS

Blandina R. Lugendo, Ivan Nagelkerken, Guus Kruitwagen, Gerard van der Velde, and Yunus D. Mgaya

ABSTRACT The importance of mangroves as feeding grounds for fish and other macrozoobenthos in the Indian Ocean and elsewhere has been a subject of debate. This could partly be due to the fact that studies describing this role have been conducted in mangrove systems that differed in their settings. By using stable isotope analysis of carbon and nitrogen, we investigated two different settings of mangroves along the Tanzanian coast, to establish if mangrove setting influences the extent to which this habitat is utilized as a potential feeding ground by fish. The two mangrove settings were: mangrove-lined creeks which retain water during low tides and fringing mangroves that drain completely during low tides. The δ13C signatures of most fishes from the mangrove-lined�� creeks�� were similar to those of food items from the mangrove habitat, which suggests that these fishes feed from the mangrove habitats. In contrast, the overlap in δ13C of some food items from the fringing mangroves with those from adjacent habitats, and the more enriched δ13C signatures of fishes from the fringing mangroves with respect to most typical food items from the mangrove habitat could be an indication that these fishes feed from both habitats but to a lower extent from the fringing mangroves. The results suggest that fishes feed more from the mangrove-lined creeks as compared to fringing mangroves which is probably related to differences in the degree of mangrove inundation. TheTh e more or less concon-- tinuous access provided more time for fishes to stay and feed in the mangrove-lined creeks compared to fishes from the fringing mangroves, which have access to these mangroves only during high tide and have to migrate to adjacent habitats with the ebbing tide.

Mangrove ecosystems are widely recognized as potential nursery grounds for juvenile fishes of exploited populations (Blaber et al., 1989; Parrish, 1989; Lugendo et al., 2005). Although most mangrove studies have found a limited role for mangrove detritus in estuarine food webs (Fry and Ewel, 2003), mangrove habitats are assumed to provide abundant food sources, such as algae, crustaceans, and other macrofauna, to resident and transient animals. However, the importance of food from the mangrove habitats is not limited to the potential contribution to fisheries production, rather, to functioning of the whole system that supports fish populations (Fry and Ewel, 2003; Layman, 2007). The importance of mangrove habitats as feeding grounds for fish and other macrozoobenthos is a subject of debate. While some mangrove habitats are reported to form a major feeding habitat for fish and macrozoobenthos in some parts of the Indo-Pacific (e.g., Rodelli et al., 1984; Marguillier et al., 1997; Sheaves and Molony, 2000; Chong et al., 2001; Guest and Connolly, 2004) this importance is refuted in other parts of the Indo-Pacific (e.g., Bouillon et al., 2002a,b) and elsewhere (e.g., Philippines: Primavera, 1996; and the Caribbean: Nagelkerken and van der Velde, 2004a). The reason for this discrepancy could be due to the fact that these studies have been conducted in mangrove systems that differed in their settings. Lee (1995, 1999)

pointed out that the degree of “outwelling” of mangrove carbon to adjacent aquatic environments depends to a large degree on the geomorphology and tidal characteristics of the ecosystem.�� Large�� tidal ranges that characterize most mangrove ecosystems of the Indo-Pacific influence their functioning as major feeding habitats for fish due to movement of large quantities of seawater and fishes between neighboring habitats showing connectivity by tidally migrating fish (Nagelkerken and van der Velde, 2004b). However, mangrove swamps of the Caribbean, a region characterized in contrast by low tidal differences and slow tidal currents, have also been reported to be major feeding grounds for fish which reside permanently there or always have access to them (Odum and Heald, 1975; Thayer et al., 1987; Ley et al., 1994). More factors than tidal difference alone are involved with respect to the functioning of mangrove habitats as feeding grounds for fish. Recently, Nagelkerken and van der Velde (2004b) found that fringing mangroves (as opposed to large mangrove forests) in the Caribbean are not an important feeding habitat for most fish species occurring in adjacent habitats. We propose here that the mangrove setting could be one of the important factors influencing the functioning of mangrove habitats as an important feeding area. We investigated two different settings of mangrove habitats along the Tanzanian coast, to determine whether the mangrove setting influences the use of this habitat as a feeding ground by fish. These two -man grove settings were: mangrove-lined creeks which retain water during low spring tides and fringing mangroves which drain completely during low spring tides. We used stable isotope analysis of carbon and nitrogen since the use of conventional methods such as gut content analysis alone may provide inadequate results because of the following reasons: (1) differences in digestion rates of ingested material, (2) gut contents may be hard to identify, (3) not all contents are digested, (4) gut content analyses provide just a snap-shot of the true diet, and (5) gut content does not reveal from where the food originates (MacDonald et al., 1982; Gearing, 1991; Polis and Strong, 1996). Analysis of the stable carbon and nitrogen isotopes can provide a clearer understanding of diets because they reflect the actual assimilation of organic matter into consumer tissue rather than merely its consumption (Gearing, 1991). The power of stable isotope analysis as a tool in the investigation of aquatic food web structures and dietary patterns is based on the significant and consistent differences in isotopic composition of different types of primary producers due to different photosynthetic pathways or different inorganic carbon sources (Fry and Sherr, 1984; Deegan and Garritt, 1997). The stable isotopic composition of an animal reflects that of its diet with up to 1.0‰ enrichment of 13C and an average of 3.5‰ enrichment of 15N occurring between consumer and its food source (DeNiro and Epstein, 1978; Fry and Sherr, 1984; Minagawa and Wada, 1984) due to the discrimination against lighter isotopes during assimilatory and excretory functions within consumers (Minagawa and Wada, 1984). However, the actual degree of fractionation varies as a function of taxonomy, food quality, type of analyzed tissue, sample treatment, and environmental factors (Vander Zanden and Rasmussen, 2001; McCutchan et al., 2003; Vanderklift and Ponsard, 2003). The following questions were asked in the present study: (1) is it possible to discriminate potential food items for fishes from the mangrove habitats with those from adjacent habitats in both mangrove settings? and (2) To what degree do fishes present in the mangrove and adjacent habitats utilize food items associated with mangrove habitats? LUGENDO ET AL.: MANGROVES AS FEEDING GROUNDS FOR FISHES 499

Materials and Methods

Study Area.—The study was carried out in Tanzanian coastal waters for two different mangrove settings which differ in their degree/duration of inundation (Table 1): 1. Mangrove-lined creeks which retain water during low spring tides. These include the Ma- popwe mangrove creek in Chwaka Bay, Zanzibar, and Mbegani mangrove creek in Bagamoyo, Tanzania mainland coast (Fig. 1). Chwaka Bay is a shallow marine bay located on the east coast of Unguja Island, Zanzibar. The bay consists of a large intertidal flat partly covered with mixed assemblages of algae and seagrass beds. On the landward side, the bay is fringed by a dense mangrove forest dominated by Rhizophora mucronata Lamarck, 1804 and Avicennia marina (Forsskål) Vierh., 1907 with an approximate area of 3000 ha (Mohammed et al., 1995). The mangrove forest has a number of tidal creeks, with Mapopwe Creek (approximately 2 m deep) being the largest and the main water exchange route between the forest and the bay. The mangrove creeks are intertidal in nature with water remaining during low spring tide, and none have any significant freshwater input other than rain. Mbegani mangroves are charac- terised by a band of mangroves (about 420 m wide) of mainly Sonneratia alba J. E. Smith, 1819, but mixed with R. mucronata, A. marina, and Bruguiera gymnorhiza (Linnaeus) Lamarck, 1797. The mangrove forest has a big tidal creek which never falls dry, even during low spring tide. The channel ends into a landward stream (Nyanza river) which can be a potential source of freshwater into the mangrove forest during the rainy season (data were collected here during the dry season). The mudflats adjacent to the mangroves are dominated by heaps of the small bivalve Arcuatula arcuatula (Hanley, 1844) imbedded within the mud. There are patches of seagrasses on the mudflat area and a more or less continuous seagrass bed at the mouth of the channel. On the landward side, these mangroves are partly sheltered from the open ocean by the small island Mapopo. 2. Fringing mangroves which drain completely during low spring tides. These are represented by Kaole and Nunge mangroves in Bagamoyo, and Mtoni estuary located south of Dar es Salaam, Tanzanian mainland coast (Fig. 1). Kaole mangroves are characterized by a band of mangroves of about 450 m wide with small intertidal creeks which fall dry during low tide. The mangroves (mainly S. alba) are bordered by intertidal seagrass beds that consist of sparse, but mixed (about 8) seagrass species. The Nunge area is characterized by a narrow band (approximately 160 m wide) of mainly S. alba. It is also located close to two reefs (Mwamba Pwani and Mwamba Mjini). Between the mangrove fringe and the reefs lies a seagrass bed that contains almost all species of seagrasses found in Tanzania. Topographically, the area is flat and shallow but behind the reef there is a drastic change in slope where the depth increases abruptly. No channels or creeks are found in this area and the mangroves fall completely dry during low tide. Covering an area of 378.4 ha, Mtoni estuary is located south of Dar es Salaam (Fig. 1).

Table 1. Location, setting, characteristics, and type of adjacent habitats of the studied mangrove habitats along the Tanzanian coast.

Location Site Mangrove setting Estuarine vs non- Nature of the estuarine adjacent habitats Chwaka Bay, Mapopwe Mangrove-lined creek Non-estuarine Mud/sand flats and Zanzibar seagrass beds

Bagamoyo Mbegani Mangrove-lined creek Non-estuarine Mudflats, seagrass beds, and reef habitats

Bagamoyo Kaole Fringing mangrove Non-estuarine Seagrass beds

Bagamoyo Nunge Fringing mangrove Non-estuarine Seagrass beds and reef habitats Dar es Salaam Mtoni estuary Fringing mangrove Estuarine Mudflats with sparse seagrasses 500 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Figure 1. Map of the coast of Tanzania showing the location of Chwaka Bay on Unguja Island (Zanzibar), Kaole, Mbegani and Nunge in Bagamoyo, and Mtoni estuary in Dar es Salaam. Dark grey areas in Bagamoyo, Chwaka Bay and Mtoni indicate mangrove forests, open squares indicate the study sites.

The estuary receives freshwater from two creeks: the Mzinga creek and the Kizinga creek. The estuary is fringed by various species of mangroves with two dominating species, viz. S. alba and A. marina. Adjacent to the mangroves a large mudflat area with sparse seagrass cover is located. Despite the two creeks that pour freshwater into the estuary, Mtoni mangroves fall completely dry during low tide. Sampling Design.—Sample collection was carried out between November 2001 and Oc- tober 2002 in Chwaka Bay, between January and February 2004 in the Mtoni estuary, and between August 2004 and January 2005 in the Bagamoyo area. Fish samples were collected in the mangroves and in the adjacent habitats using a seine net, while a range of benthic macro-invertebrates, seagrasses and algae were collected at the fishing sites by hand. In the continuously inundated mangrove creeks, fish were collected at low tide, whereas in the fringing mangroves the fishes were collected at high tide (when inundated). In all habitats adjacent to the mangroves, fish were collected at low tide.Particulate organic matter was sampled only from Mtoni estuary and was sampled by filtering several liters of seawater over a Whatman GF/C glass fiber filter. In the field, samples were put in a cool box and later frozen at −20 °C LUGENDO ET AL.: MANGROVES AS FEEDING GROUNDS FOR FISHES 501 pending analysis. All collected fishes were identified, counted, and their fork length measured to the nearest 0.1 cm. In the majority of the cases, a minimum of three samples were used to include a species in the analysis. Stable Isotope Analysis.—Muscle tissue was removed from the fish, while molluscs (gastropods and bivalves) and crustaceans (crabs and shrimps) were dissected from their exoskeleton and shell prior to drying. Samples were dried at 70 °C for 48 hrs and ground to powder (homogeneous mixture). For samples rich in carbonates such as whole individuals of small hermit crabs and detritus, sub-samples were acid-washed and oven-dried. These sub-samples were used for stable carbon isotope analysis while the remaining untreated sub- samples were used for stable nitrogen isotope analysis since acid could interfere with stable nitrogen isotopes (Pinnegar and Polunin, 1999). A pre-determined sample of known weight was placed in ultra-pure tin capsules and combusted in a CHN Elemental Analyser from Car- lo Erba® (Thermo group), interfaced with a continuous flow isotope ratio mass-spectrometer, the DeltaPlus from Thermo Finnigan, Bremen, Germany. The reference gasses used were cali- brated with the IAEA reference standards, IAEA-N-2 and IAEA-CH-6. The potential feeding habitats and food items for fish were estimated in view of the average enrichment in isotope signatures between animals and their potential food items of 1‰ and 3.5‰, for carbon and nitrogen, respectively (DeNiro and Epstein, 1978; Minagawa and Wada, 1984; Vander Zanden and Rasmussen, 2001; McCutchan et al., 2003). Statistical Analysis.—A Levene’s test was used to check the data for homogeneity of variances (Field, 2000). In cases where variances were homogeneous, a one-way ANOVA or Student’s t-test was done to test for differences in stable isotope signatures of carbon for fish among different habitats. Since sample sizes were very different, a Hochberg’s GT2 was used as a post-hoc test due to its greater statistical power in such data compared to other tests (Field, 2000). Either a Kruskal-Wallis test or a Mann-Whitney U-test was used as a non-parametric test equivalent when variances were not homogeneous even after log- transformation. A Games-Howell post-hoc test was used following the Kruskal-Wallis tests because it is more powerful and specifically designed for lack of homogeneity of variances (Field, 2000). A significance level of P = 0.05 was used in all tests. All analyses were performed using the programme SPSS 11.5 for Windows (Field, 2000).

Results

Mangrove-lined Creeks.—A clear distinction in δ13C could be discerned for macrofauna/flora and fishes from the mangrove creeks (Mapopwe and Mbegani) and from the adjacent habitats (Fig. 2). In Mapopwe, the average δ13C values of macrofauna/flora from the mangrove creeks ranged between −28.4‰ and −20.4‰ (except seagrass leaves) while those from the mud/sand flats ranged between −17.1‰ and −13.8‰ (except detritus; Fig. 2A). Considering the fishes, those from the mangrove habitats were more δ13C depleted than those from the mud/sand flats (Fig. 2B). The highest enrichment in mean δ13C values between fish from the mangrove creeks and mud/sand flats was 6.3‰ for individuals of Lutjanus fulviflamma (Forsskål, 1775). Three fish species showed deviating values (Fig. 2B). Siganus sutor (Valenciennes, 1835) from the mud/sand flats had a highly depleted13 δ C value compared to the other species from the mud/sand flats. Gerres oyena (Forsskål, 1775) and Lutjanus ehrenbergi (Peters, 1869) from the large mangrove creek showed enriched δ13C values compared to other species from the mangrove creeks. However, since δ13C values of G. oyena from the mud/sand flats were much more enriched than those from the mangrove creek, we consider that this species also showed the separation in δ13C between the mangrove creeks and the mud/sand flats. The δ13C of individuals of Gerres filamentosus (Cuvier, 1829), G. oyena, Lethrinus lentjan (Lacépède, 1802), L. 502 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Figure 2. Mean δ13C and δ15N values of macrofauna/flora and fishes from mangrove habitats and from adjacent habitats in continuously inundated Mapopwe creek (A–B) and Mbegani creek (C–D) (opposite page). The dotted lines indicate the main separation in δ13C values between the mangrove and adjacent habitats. fulviflamma, and Sphyraena barracuda (Walbaum, 1792) differed significantlyP ( < 0.01) between the mangrove creeks and the mud/sand flats. The δ13C values of the other species were not tested because these species occurred in one habitat only. The δ15N values of food items from Mapopwe ranged from −1.5‰ (detritus) to 6.1‰ (insects) while those of fishes ranged from 5.6‰ S.( sutor) to 9.8‰ (S. barracuda). These δ15N values clearly show the trophic position of fishes belonging to different feeding guilds, with lowest values for herbivores, intermediate values for zoobenthivores, and highest values for piscivores. LUGENDO ET AL.: MANGROVES AS FEEDING GROUNDS FOR FISHES 503

In Mbegani, average δ13C values of macrofauna/flora from the mangrove creek ranged between −23.0‰ and −11.4‰ (Fig. 2C). Average δ13C values of macrofauna/ flora from the seagrass bed overlapped in several cases with those from the mangrove and reef habitats and they ranged between −18.4‰ and −9.3‰ for seagrass beds and between −15.0‰ and −5.5‰ for the reef habitat. Considering specific groups of macrofauna/flora, however, there were clear differences in mean13 δ C values for polychaetes, gastropods, macroalgae, and seagrass leaves between mangrove and seagrass/reef habitats (Fig. 2C). With respect to the fish, a very clear separation 13δ C was present between specimens from the mangrove creek and those from the adjacent seagrass beds/reef habitats (Fig. 2D). Fish from the seagrass beds and reef habitats were similar in δ13C but more enriched (by on average 3.1‰) than those from the 504 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

mangrove creek (Fig. 2D). A species of Atherinidae and G. oyena from the mangrove creek were significantly depleted (t-test: P = 0.005 and P = 0.013, for Atherinidae and G. oyena, respectively) as compared to those from the reef habitat. Statistical tests could not be performed for the other species as they occurred in only a single habitat. For the whole Mbegani area, the δ15N values of food items ranged between −0.3‰ (seagrass leaves) and 7.2‰ (gastropods) while those of fishes ranged between 7.1‰ (Lutjanus argentimaculatus (Forsskål, 1775)) and 10.3‰ (S. barracuda), reflecting the trophic positions of the studied fish species (Fig. 2C, D). Fringing Mangroves.—Although a clear difference in 13δ C values could be discerned for macrofauna/flora between the mangrove habitats and the adjacent habitats in Kaole, Nunge, and Mtoni, this could not be discerned for the fishes in these habitats (Fig. 3). In Kaole, the mean δ13C values of macrofauna/flora from the mangrove creek ranged between −19.4‰ and −16.0‰ (except shrimps; Fig. 3A). Those from the intertidal seagrass beds ranged between −15.2‰ and −11.3‰, while those from the subtidal seagrass beds overlapped partly with those from the intertidal seagrass beds and ranged between −13.4‰ and −7.5‰. The 13δ C values of fishes from the mangrove habitat of Kaole (mean ± SE = 14.9 ± 0.7‰) overlapped with those from the intertidal seagrass beds (14.9 ± 0.4‰; Fig. 3B). However, fishes from the subtidal seagrass beds were much more enriched in δ13C (10.9 ± 0.1‰). At species level, G. oyena, Pelates quadrilineatus (Bloch, 1790), and L. fulviflamma from the subtidal seagrass beds all differed significantly (t-test: P < 0.013) from those from the mangrove habitat or intertidal seagrass bed. The δ15N values of food items ranged between 0.6‰ (Lit- toraria sp.) and 5.4‰ (insects) while those of fishes ranged from 4.9‰ Leptoscarus vaigiensis (Quoy and Gaimard, 1824) to 9.4‰ (Atherinidae). In Nunge, the mean δ13C values of macrofauna/flora from the mangrove habitat ranged between −21.0‰ and −15.2‰ (except shrimps), those from the seagrass beds ranged between −15.7‰ and −13.2‰, and those from the reef habitat ranged between −15.1‰ and −10.0‰ (except for the gastropod Terebralia sp.; Fig. 3C). The δ13C values of fishes in Nunge did not show a clear distinction between the different habitats, rather their mean δ13C values were clustered between −16.0‰ and −10.3‰ and overlapped between habitats at species level (Fig. 3D). The δ15N values of food items ranged from 1.9‰ (Littoraria sp.) to 8.3‰ (crabs) while those of fishes ranged from 6.8‰ (G. oyena) to 11.7‰ (P. quadrilineatus) indicating that different fish species belong to different trophic positions. In Mtoni, the average δ13C values of mangrove macrofauna/flora ranged between −25.9‰ and −17.8‰, with the exception of shrimps, swimming crabs, and barnacles (Fig. 3E). Those from the estuary basin ranged between −17.7‰ and −13.5‰, except particulate organic matter, polychaetes, and macroalgae. The 13δ C values of fishes from Mtoni did not show a clear distinction in 13δ C signature between the mangrove habitat and the estuary basin, and did not show a large variation within species (Fig. 3F). Only the herbivore S. sutor and one species of Carangidae devi- ated considerably in their δ13C and/or δ15N values from the cluster of other species (Fig. 3F). No significant difference was found between the13 δ C values of fishes from the mangrove habitat and the estuary basin for Albula glossodonta (Forsskål, 1775), Carangidae, G. filamentosus, G. oyena, Sillago sihama (Forsskål, 1775), and Terapon jarbua (Forsskål, 1775) (t-test: P > 0.144). The remaining species could not be tested statistically as they occurred in only a single habitat. In the whole estuary, the δ15N values of food items ranged between 3.4‰ (detritus) to 12.6‰ (shrimps) while those LUGENDO ET AL.: MANGROVES AS FEEDING GROUNDS FOR FISHES 505

Figure 3. Mean δ13C and δ15N values of macrofauna/flora and fishes from the mangrove habitats and from adjacent habitats, in Kaole (A–B), Nunge (C–D) (on page 506) and Mtoni (E–F) (on page 507). The fish species of Mtoni are represented by the following numbers, 1: Leiognathus equulus, 2: Gobiidae sp. 1, 3: Terapon jarbua, 4: Gerres acinaces (Bleeker, 1854), 5: Gerres oyena, 6: Gerres filamentosus, 7: Sillago sihama, 8: Albula glossodonta, 9: Gerres oyena, 10: Albula glossodonta, 11: Arothron hispidus (Linnaeus, 1758), 12: Scarus russelii (Valenciennes, 1840), 13: Lethrinus lentjan, 14: Arothron immaculatus (Bloch and Schneider, 1801), 15: Sar- dinella albella, 16: Leiognathus equulus, 17: Apogon sp. 1, 18: Apogon sp. 2, 19: Apogon sp. 3, 20: Terapon jarbua (Forsskål, 1775), 21: Atherinomorus duodecimalis (Valenciennes, 1835), 22: Gerres filamentosus, 23: Atherinidae, 24: Lutjanus fulviflamma, 25: Gobiidae sp. 2, 26: Sillago sihama, 27: Gerres acinaces, 28: Gobiidae sp. 3, 29: Platycephalus indicus (Linnaeus, 1758), 30: Mugilidae sp. 1, 31: Saurida gracilis (Quoy and Gaimard, 1824), 32: Mugilidae sp. 2, 33: Calli- onymus sp., 34: Pelates quadrilineatus, 35: Gobiidae sp. 4, 36: Siganus sutor, 37: Carangidae, 38: Carangidae. The dotted lines indicate the main separation in δ13C values of food items between the mangroves and adjacent habitats. 506 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Figure 3. Continued.

of fishes ranged from 8.2‰ (Carangidae) to 15.2‰ (Apogon sp.). These δ15N values are relatively higher compared to δ15N values of the food items and fishes from the other four studied sites, probably due to the inflow of Karibu textile effluents into the area (eutrophication).

Discussion

On the basis of the differences in 13δ C signatures of most of the potential food items in the various habitats, the most likely feeding habitats of the fish can be deduced. The food items (macrofauna/flora) from the mangrove habitats of both man- LUGENDO ET AL.: MANGROVES AS FEEDING GROUNDS FOR FISHES 507

Figure 3. Continued.

grove settings were generally more δ13C depleted compared to similar food items from their respective adjacent habitats. While the stable isotope values of fishes from the mangrove-lined creeks followed a similar trend, those from the fringing mangrove habitats did not. Fishes from the mangrove-lined creeks displayed a δ13C signature similar or close to food items from the mangrove creeks, which could be an indication that they feed mostly on food items from the mangrove habitats. Another possibility is that the fishes caught from the mangrove creeks feed elsewhere (possibly during high tide), but on food items with a similar δ13C signature as those of food items from the mangrove habitats; especially those food items which were not covered in the sampling for the present study. However, this is unlikely since food 508 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

items from other habitats were always more enriched than those from the mangrove habitats, a situation which has widely been observed in other studies (e.g., Rodelli et al., 1984; Bouillon et al., 2002b; Kieckbusch et al., 2004; Nagelkerken and van der Velde 2004a). In contrast, the much more enriched δ13C signature of the fishes from the fringing mangrove habitats as compared to those of food items from the mangrove habitats suggests that these fishes obtained most of their food from the adjacent habitats. On the other hand, the overlap in the δ13C signatures of some of the food items (e.g., motile crabs, shrimps, and hermit crabs) between the fringing mangrove habitats and the adjacent habitats, suggest that possibly those fishes may feed from all the studied habitats. The above observations suggest that the utilization of mangrove habitats as feeding grounds for fishes is different in the two mangrove settings. Both settings of mangroves are flooded completely only during high tide. However, fishes from the mangrove-lined creeks can remain in the water held within the creeks during low tide, and hence have a more or less permanent access to the mangrove habitat, where they apparently also feed and hence acquire more depleted δ13C signatures similar to those of food items from the mangrove habitat. In contrast, fishesfi shes �� fromfrom thethe fringfring-- ing mangroves have access to the mangrove habitats only during high tide, and have to migrate to adjacent habitats (seagrasses, reef or mud/sand flats) with the ebbing tide where they probably also feed. Although these fish could feed in the fringing mangroves at high tide, their more enriched δ13C signatures (similar to those of food items from the adjacent habitats) suggest that they probably obtained more food from the adjacent habitats than from the mangrove habitats. Since these fringing mangrove habitats are not continuously accessible, this suggests that fishes may partly use these mangrove habitats for feeding at high tide, but largely use adjacent habitats as feeding habitats at low tide. As pointed out by Fry and Ewel (2003), the access and residency (time) of animals within the mangrove ecosystems could influence the utilization of mangrove resources. Additionally, the fact that fishes caught in the mangrove-lined creeks at low tide did not include fishes with a more enriched δ13C signature (except possibly L. ehrenbergi from the mud/sand flats of Chwaka Bay) suggests that fishes from the mangrove-lined creeks (with a depleted δ13C signature) and those from adjacent habitats (with an enriched δ13C signature) form two different feeding populations. This was also observed for permanently inundated fringing mangroves in the Caribbean (Nagelkerken and van der Velde, 2004b). The species that appeared to feed primarily from the mangrove-lined creeks in the present study have mostly been reported to be mangrove residents, especially at juvenile and young adult stages (i.e., Caranx sexfasciatus Quoy and Gaimard, 1825, G. filamentosus, G. oyena, Leiognathus equulus (Forsskål, 1775), Lutjanus argentimaculatus (Forsskål, 1775), L. fulviflamma, L. lentjan, S. barracuda, and Zenarchopterus dispar (Valenciennes, 1847); Froese and Pauly, 2005; Lugendo et al., 2005). Again, the δ13C signatures of their respective species in the fringing mangroves were far more enriched, consistent with the fact that fishes from the fringing mangroves probably obtained their food largely from the adjacent habitats. This is another indication that the degree of mangrove inundation is important in determining the extent to which a mangrove habitat is utilized as a feeding ground by fishes. The size of the mangrove forest itself is believed to be another critical factor that determines the importance of mangroves as major feeding grounds for fishes. Stud- LUGENDO ET AL.: MANGROVES AS FEEDING GROUNDS FOR FISHES 509

ies from the Caribbean report large mangrove swamps to be major feeding grounds for fishes (Odum and Heald, 1975; Thayer et al., 1987; Ley et al., 1994) as opposed to small fringing mangroves (Nagelkerken and van der Velde, 2004b). This is appli- cable to the Chwaka Bay mangrove forest (where the Mapopwe creek is found) due to its large size, but not to the Mbegani mangrove creek which is lined by a narrow mangrove forest. We do not rule out the possible influence of mangrove forest size in selection of feeding habitats by fishes in our study. However, the smaller sizes of the studied mangroves of Bagamoyo (e.g., Mbegani and Kaole mangrove forests together cover 179.1 ha; Semesi, 1991) and the differences in the way they were utilized by fishes, as we observed, is consistent with our hypothesis that degree of mangrove inundation is a more important factor influencing the extent to which mangrove habitats are utilized as feeding grounds by fishes. Although most fishes did not seem to use the fringing mangroves during low tide as major feeding habitats, some macrofauna did. Some macrofauna species were restricted to the mangrove habitats as suggested by their highly depleted δ13C signatures (< −19.0‰). This was the case for mangrove crabs (mostly Sesarminae and Uca spp.), gastropods (mostly Littoraria spp. and Terebralia sp.) and insects, which are all known to feed on either mangrove leaves, detritus, micro-epiphytes, diatoms, or mud from the mangrove habitats (Slim et al., 1997; Bouillon et al., 2002a; Ólafsson et al., 2002; Skov and Hartnoll, 2002; Fratini et al., 2004). These species that likely feed solely from the mangrove habitat remained in the mangroves during low tide even when all water drained away. Mangrove crabs and gastropods are adapted to living in intertidal areas for extended periods by retreating in their burrows (crabs) and in their shells (gastropods) during low tide. In contrast, fishes depend solely on the presence of water and must move with the ebbing tide; therefore, degree of mangrove inundation is a crucial determinant of the extent to which fishes can use a mangrove habitat as a feeding ground. Motile macrofauna (e.g., shrimps, swimming crabs, and hermit crabs) from the fringing mangroves, on the other hand, showed enriched δ13C values (> 16.8‰) typical for macrofauna from the adjacent habitats, indicating an interaction or connectivity between mangroves and their adjacent habitats. These species probably use the mangrove only as a refuge and feed on food items from the adjacent habitats. The same trend was also evident for sessile filter-feeding barnacles within the mangroves at Mtoni estuary. Since the mangroves drain completely during low tide the barnacles depend on suspended food sources brought in with the incoming tide from the estuary. Furthermore, the δ13C values of particulate organic matter from the estuary were in-between those of seagrass and mangrove leaves, indicating that it consisted of a mixture of plant material from the terrestrial mangrove habitat and from the estuary basin (e.g., seagrass leaves). In contrast to the above, in continuously inundated mangroves of Mapopwe Creek (Chwaka Bay) all macrofauna (δ13C values < −20.3‰) appeared to feed largely from the mangrove habitats. However, in the case of the continuously inundated Mbe- gani mangrove creek, the mangrove gastropods Littoraria spp. and polychaetes (δ13C values < −21.1‰) appeared to feed largely from the mangrove habitats, whereas the more motile fauna (crabs, shrimps, and hermit crabs) appear to feed largely from adjacent habitats. Interestingly, the δ13C value of the bivalve A. arcuatula from the seagrass habitat of Mbegani (mean δ13C value of −18.4‰) suggests incorporation of a mixture of material from both the mangrove and seagrass habitats. 510 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

In summary, in almost all cases, food items from both mangroves settings could be discriminated by δ13C from those of adjacent habitats. Exceptions were always found for the motile swimming and hermit crabs and shrimps, and for sessile filter- feeding organisms. Fishes from all habitats adjacent to mangroves, and those from the fringing mangroves showed a more enriched δ13C signature, indicating that they probably obtained their food largely from the adjacent habitats. In contrast, mangroves that allowed a more or less continuous utilization of the mangrove habitats by fishes appeared to contribute to a larger degree to the food of fishes as compared to fringing mangroves that drain completely during low tide. This study revealed that the potential of a mangrove habitat in functioning as a feeding area for fish can to be influenced by its configuration or setting. Fringing mangrove habitats seem to contribute less to the food items of fishes as compared to continuously accessible mangrove-lined creeks. Ecosystems showing more similarity in configuration may function in a similar way despite geographical differences.

Acknowledgments

We are thankful to M. Manzi,�� S. Simgeni, M. Makame, A. Ramadhani, H. Kauka, S. Mu- kama, and Y. Mhonda for assistance with the field work, and J. Eygensteyn of the Radboud University Nijmegen for analyzing the stable isotope samples. The administration and staff of the Institute of Marine Sciences in Zanzibar provided logistical support and research fa- cilities. We also appreciate the logistical support we received from the personnel of Mbegani Fisheries Institute and Bagamoyo District Fisheries Office. This study was financially supported by a UNESCO-Loreal for Women in Science Fellowship, NUFFIC through the ENVI- RONS-MHO Project, Schure-Beijerinck-Popping Foundation and Quo Vadis Fonds (Radboud University Nijmegen). I. N. was supported by a VIDI grant from the Netherlands Organisation for Scientific Research (NWO). The manuscript also benefited from the comments of two anonymous reviewers. This is Centre for Wetland Ecology publication no. 440.

Literature Cited

Blaber, S. J. M., D. T. Brewer, and J. P. Salini. 1989. Species composition and biomasses of fishes in different habitats of a tropical northern Australian estuary: Their occurrence in the adjoining sea and estuarine dependence. Estuar. Coast. Shelf Sci. 29: 509–531. Bouillon, S., N. Koedam, and F. Dehairs. 2002a. Primary producers sustaining macro-invertebrate communities in intertidal mangrove forests. Oecologia 130: 441–448. ______, A. V. Raman, P. Dauby, and F. Dehairs. 2002b. Carbon and nitrogen stable isotope ratios of subtidal benthic invertebrates in an estuarine mangrove ecosystem (Andhra Pradesh, India). Estuar. Coast. Shelf Sci. 54: 901–913. Chong, V. C., C. B. Low, and T. Ichikawa. 2001. Contribution of mangrove detritus to juvenile prawn nutrition: a dual stable isotope study in a Malaysian mangrove forest. Mar. Biol. 138: 77–86. Deegan, L. A. and R. H. Garritt. 1997. Evidence for spatial variability in estuarine food webs. Mar. Ecol. Prog. Ser. 147: 31–47. DeNiro, M. J. and S. Epstein. 1978. Influence of diet on the distribution of carbon isotopes in animals. Geochem. Cosmochim. Acta 45: 341–351. Field, A. 2000. Discovering statistics using SPSS for Windows. Sage Publications, London. 496 p. Fratini, S., V. Vigiani, M. Vannini, and S. Cannicci. 2004. Terebralia palustris (Gastropoda; Potamididae) in a Kenyan mangal: size structure, distribution and impact on the consumption of leaf litter. Mar. Biol. 144: 1173–1182. LUGENDO ET AL.: MANGROVES AS FEEDING GROUNDS FOR FISHES 511

Froese, R. and D. Pauly. (eds.) 2005. FishBase. Available from: www.fishbase.org Accessed De- cember, 2005 Fry, B. and K. C. Ewel. 2003. Using stable isotopes in mangrove fisheries research - a530 review and outlook. Isotopes in Environmental and Health Studies 39: 191–196. _____ and E. B. Sherr. 1984. δ13C measurements as indicators of carbon flow on marine and freshwater ecosystems. Contrib. Mar. Sci. 27: 13–47. Gearing, J. N. 1991. The study of diet and trophic relationships through natural abundance13 C. Pages 201–218 in D. D. Coleman and B. Fry, eds. Carbon isotope techniques. Academic Press, San Diego. Guest, M. A. and R. M. Connolly. 2004. Fine-scale movement and assimilation of carbon in saltmarsh and mangrove habitats by resident animals. Aquat. Ecol. 38: 599–609. Kieckbusch, D. K., M. S. Margurite, J. E. Serafy, and W. T. Anderson. 2004. Trophic linkages among primary producers and consumers in fringing mangroves of subtropical lagoons. Bull. Mar. Sci. 74: 271–285. Layman, C. A. 2007. What can stable isotope ratios reveal about mangroves as fish habitat? Bull. Mar. Sci. 80: 513–527. Lee, S. Y. 1995. Mangrove outwelling: a review. Hydrobiologia 295: 203–212. ______. 1999. Tropical mangrove ecology: physical and biotic factors influencing ecosystem structure and function. Austr. J. Ecol. 24: 355–366. Ley, J. A., L. Montague, and C. C. McIvor. 1994. Food habits of mangrove fishes: a comparison along estuarine gradients in northeastern Florida Bay. Bull. Mar. Sci. 54: 881–899. Lugendo, B. R., A. Pronker, I. Cornelissen, A. de Groene, I. Nagelkerken, M. Dorenbosch, G. van der Velde, and Y. D. Mgaya. ��2005. Habitat utilization by juveniles of commercially important fish species in Chwaka Bay, Zanzibar, Tanzania.Aquat. Living Resour. 18: 149–158. MacDonald, J. S., K. G. Waiwood, and R. H. Green. 1982. Rates of different digestion of different prey in Atlantic cod (Gadus morhua), ocean pout (Macrozoarces americanus), winter flounder (Pseudopleuronectes americanus), and American plaice (Hippoglossoides plates- soides). Can. J. Fish. Aquat. Sci. 39: 651–659. McCutchan, J. H., W. M. Lewis, C. Kendall, and C. C McGrath. 2003. Variation in trophic shift for stable isotope ratios of carbon, nitrogen, and sulfur. Oikos 102: 378–390. Minagawa, W. and E. Wada. 1984. Stepwise enrichment of 15N along food chains: further evidence and the relation between δ15N and animal age. Geochim. et Cosmochim. Acta 48: 1135–1140. Mohammed, S. M., R. W. Johnstone, B. Widen, and E. Jordelius. 1995. The role of mangroves in the nutrient cycling and productivity of adjacent seagrass communities, Chwaka Bay, Zanzibar. Zoology Department, Stockholm University, Sweden, and University of Dar es Salaam, Tanzania. Nagelkerken, I. and G. van der Velde. 2004a. Are Caribbean mangroves important feeding grounds for juvenile reef fish from adjacent seagrass? Mar. Ecol. Prog. Ser. 274: 143–151. ______and ______. 2004b. Relative importance of interlinked mangroves and seagrass beds as feeding habitats for juvenile reef fish on a Caribbean island. Mar. Ecol. Prog. Ser. 274: 153–159. Odum, W. E. and E. J. Heald. 1975. The detritus-based food web of an estuarine mangrove community. Pages 265–286 in L. E. Cronin, ed. Estuarine research. Academic Press, New York. Ólafsson, E., S. Buchmayer, and M. W. Skov. 2002. The East African decapod crab Neosarma- tium meinerti (de Man) sweeps mangrove floors clean of leaf litter. Ambio 31: 569–573. Parrish, J. D. 1989. Fish communities of interacting shallow-water habitats in tropical oceanic regions. Mar. Ecol. Prog. Ser. 58: 143–160. Pinnegar, J. K. and N. V. C. Polunin. 1999. Differential fractionation of 13δ C and δ15N among fish tissues: implication for the study of trophic interactions. Funct. Ecol. 13: 225–231. Polis, G. and D. R. Strong. 1996. Food web complexity and community dynamics. Am. Nat. 147: 813–846. 512 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Primavera, J. H. 1996. Stable carbon and nitrogen isotope ratios of penaeid juveniles and primary producers in a riverine mangrove in Guimaras, Philippines. Bull. Mar. Sci. 58: 675–683. Rodelli, M. R., J. N. Gearing, P. J. Gearing, N. Marshall, and A. Sasekumar. 1984. Stable isotope ratios as a tracer of mangrove carbon in Malaysian ecosystems. Oecologia 61: 326–333. Semesi, A. K. 1991. Management plan for the mangrove ecosystem of mainland Tanzania, Vol. 11: Mangrove Management plan of all coastal districts. Ministry of Tourism, Natural Re- sources and Environment (MTNRE), Forestry and Beekeeping Division, Catchment For- estry Project. Sheaves, M. and B. Molony. 2000. Short-circuit in the mangrove food chain. Mar. Ecol. Prog. Ser. 199: 97–109. Skov, M. W. and R. G. Hartnoll. 2002. Paradoxical selective feeding on a low-nutrient diet: Why do mangrove crabs eat leaves? Oecologia 131: 1–7. Slim, F. J., M. A. Hemminga, C. Ochieng, N. T. Jannink, E. Cocheret de la Morinière, and G. van der Velde. 1997. Leaf litter removal by the snail Terebralia palustris (Linnaeus) and sesarmid crabs in an East African mangrove forest (Gazi Bay, Kenya). J. Exp. Mar. Biol. Ecol. 215: 35–48. Thayer, G. W., D. R. Colby, and W. F. Hettler. 1987. Utilization of the red mangrove prop root habitat by fishes in south Florida. Mar. Ecol. Prog. Ser. 35: 25–38. Vander Zanden, M. J. and J. B. Rasmussen. 2001. Variation in δ15N and δ13C trophic fractionation: implications for aquatic food web studies. Limnol. Oceanogr. 46: 2061–2066 Vanderklift, M. A. and S. Ponsard. 2003. Sources of variation in consumer-diet δ15N enrichment: a meta-analysis. Oecologia 136: 169–182.

Addresses: (B.R.L., Y.D.M) Faculty of Aquatic Sciences and Technology, University of Dar es Salaam, P.O. Box 35064, Dar es Salaam, Tanzania. (I.N., G.K.) Department of Animal Ecol- ogy and Ecophysiology, Institute for Water and Wetland Research, Faculty of Science, Radboud University Nijmegen, Toernooiveld 1, 6525 ED Nijmegen, The Netherlands. (G.vdV.) National Museum of Natural History, P.O. Box 9519, 2300 RA Leiden, The Netherlands. Correspond- ing Author: (I.N.) E-mail: . BULLETIN OF MARINE SCIENCE, 80(3): 513–527, 2007

What can stable isotope ratios reveal about mangroves as fish habitat?

Craig A. Layman

abstract Stable isotope ratios are applied in many ways to explore the relationship between mangroves and fishes. Here I summarize information pertinent to three central questions regarding the mangrove-fishes link: Can stable isotope ratios be used to (1) identify basal resource pools that support fishes associated with mangrove habitat?, (2) reveal aspects of consumer trophic structure in mangrove fish communities?, and (3) describe fish movement patterns with respect to mangroves? I review recent research developments that pertain to each question, including a discussion of the limitations of stable isotope ratios in understanding the mangrove-fishes link. I emphasize promising avenues for future research, including labeled isotope tracer experiments, probabilistic and spatial analyses of potential basal resource pools, novel community-wide and intraspecific metrics, and utilization of differential tissue turnover rates within and among organisms. Stable isotope ratios alone provide numerous insights into the role of mangroves as fish habitat, but the most thorough understanding of the mangrove-fisheries link is likely to stem from integrative approaches that combine stable isotope ratio analyses and other methodologies into a broad research framework.

Analysis of stable isotope ratios is a common tool employed in many sub-disci- plines of biological and ecological research, and applications to deduce potential links between mangrove habitat and fishes are no exception. Through examination of natural variation in stable isotopes ratios of three elements (carbon, nitrogen, and sulfur), and the underlying reasons that drive this variation, many aspects of the mangrove-fishes link have been inferred. For nitrogen, the ratio of 15N to 14N (expressed as δ15N) exhibits stepwise enrichment with trophic transfers, and is a powerful tool to estimate trophic position of organisms (Minagawa and Wada, 1984; Peterson and Fry, 1987; Post, 2002b). The ratio of carbon isotopes (δ13C) varies substantially among primary producers with different photosynthetic pathways (e.g., C3 vs C4 plants), but changes little with trophic transfers (DeNiro and Epstein, 1981; Peterson and Fry, 1987; Post, 2002b). Therefore, 13δ C can be used to determine ultimate sources of dietary carbon. Similarly, the ratio of sulfur isotopes (δ34S) changes relatively little with progression through a food web, and thus can be used to identify important resource pools (Peterson et al., 1986; Connolly et al., 2004). Because stable isotope applications have become commonplace, it is germane to this symposium volume to review the question: what can stable isotopes reveal about the role of mangroves as fish habitat? Numerous reviews have addressed aspects of stable isotope application to mangrove/estuarine research (e.g., Lee, 1995; Fry and Ewel, 2003; Sheridan and Hays, 2003; Herzka, 2005), and I do not endeavor to review all of that information here. Further, I do not delve into the myriad issues regarding preparation of marine plant and animal samples for stable isotope analysis (i.e., acidification, isolation of appropriate basal resource pools, etc.). Instead, my purpose is to summarize emergent themes regarding the role of mangroves as fish habitat as suggested through field-based inquiry using stable isotope ratios. Specifically, I -ad dress three questions:

(I) Can stable isotope ratios be used to identify basal resource pools that support fishes associated with mangrove habitat?, (II) Can stable isotope ratios reveal particular aspects of consumer trophic structure (i.e., food web structure) in mangrove habitats?, and (III) Can stable isotope ratios be used to describe fishmovement patterns in relation to mangrove habitat? I place specific emphasis on future directions within each of these thematic areas, and highlight applications of stable isotopes that have been under-utilized to address longstanding questions in the field (“Future Directions” sections). Throughout the text, I use the phrase “mangrove community” to refer to fish assemblages (and the food webs in which they are embedded) associated with a coastal water body that contains a significant amount of mangrove vegetation. I attempt to identify broad themes that span substantial variation in the structure and function of ecosystems in which mangroves are a conspicuous component (Ewel et al., 1998; Lovelock et al., 2005), and hope to provide a foundation from which to develop future scientific inquires in the field.

Can Stable Isotope Ratios Be Used to Identify Basal Resource Pools that Support Fisheries Associated with Mangrove Habitats?

In the 1970s, Bill Odum and colleagues introduced evidence that sub-tropical and tropical estuarine food webs were supported primarily by a mangrove detritus-based resource pool (Odum and Heald, 1975), and this paradigm served as the model for estuarine food web research for almost two decades. Over the last decade, however, stable isotope ratios have served as the primary tool to overturn this paradigm (Fry and Ewel, 2003). Although results vary among systems and geographic regions, the balance of evidence now suggests that other basal resource pools are the primary energy sources in sub-tropical or tropical mangrove fish communities (e.g., Lee, 1995; Newell et al., 1995; Primavera, 1996; Lonergan et al., 1997; Bouillon et al., 2002a; Hsieh et al., 2002; Kieckbusch et al., 2004; Benstead et al., 2006). Only in specific circumstances, for example, when particular species are restricted to mangrove- dominated portions of a wetland system, has mangrove detritus been suggested as a critical nutritional source (Rodelli et al., 1984; Newell et al., 1995; Sheaves and Molony, 2000; Christensen et al., 2001; Bouillon et al., 2004). Yet if not mangrove detritus, what basal resource(s) support the productive fisheries typically associated with mangroves? Potentially important resource bases include epiphytic algae, macroalgae, microphytobenthos, particulate organic matter, phytoplankton, and seagrass. In specific cases, stable isotopes have provided valuable insights into the likely importance of one or more of these resources in supporting particular consumers of interest (e.g., Lee, 1995; Newell et al., 1995; Primavera, 1996; Lonergan et al., 1997; Bouillon et al., 2002a; Hsieh et al., 2002; Kieckbusch et al., 2004). Yet determination of the exact contribution of each of these resources to a particular consumer is usually not feasible because the number of major basal source pools (often 6 or more) exceeds the number of stable isotope ratios (typically 2 or 3). More generally, for any number (n) of isotopes employed, an exact combination for only n+1 potential sources can be determined (Phillips and Gregg, 2003). As such, in most mangrove fish communities, without a priori aggregating potential resource LAYMAN: MANGROVES, FISH, AND ISOTOPES 515

pools (Phillips et al., 2005), it is impossible to quantify the exact contribution of each basal resource to a focal species, population, or individual. Two general approaches have been proposed to deal with situations in which the number of resource pools exceeds the number of isotope ratios: the Isosource model developed by Phillips and Gregg (2003), and a spatially-based approach that relies on correlations among isotopic signatures of a consumer of interest and signatures of potential resource pools (Melville and Connolly, 2003). The Isosource model de- rives all feasible combinations of basal resource pools that can account for observed isotope values of a consumer. Data typically are presented in the form of frequency histograms that illustrate the complete range of possible percent contributions of each basal resource. This model has been widely applied in ecosystems in which mangroves are one of many potential resource pools (e.g., Melville and Connolly, 2003, 2005; Abed-Navandi and Dworschak, 2005; Connolly et al., 2005b; Benstead et al., 2006). Although more informative than qualitative inference based on δ13C, δ15N, and δ34S, the model remains a probabilistic approach that cannot reveal a unique solution as to the actual contribution of each source pool. As such, caution is ad- vised when applying this approach, especially when the derived range of potential source contributions is wide (Phillips and Gregg, 2003), as is often the case in mangrove fish communities (e.g., Melville and Connolly, 2003, 2005; Abed-Navandi and Dworschak, 2005; Connolly et al., 2005b; Benstead et al., 2006). The spatially-based analysis proposed by Melville and Connolly (2003) takes advantage of the large-scale (e.g., among estuary) variation in isotopic signatures within a basal resource pool. For example, it is well-documented that anthropogenic inputs to estuarine ecosystems can affect isotopic signatures of primary producers (McClel- land et al., 1997; McClelland and Valiela, 1998; Fry, 1999; deBruyn and Rasmussen, 2002; Tewfik et al., 2005), and natural sources of variation in primary producer signatures also can be substantial (Marguillier et al., 1997; Bouillon et al., 2000; Fry and Smith, 2002). Melville and Connolly (2003) propose that the degree of correlation between a basal resource pool and a consumer of interest among spatially distinct areas provides an indication of the importance of that basal resource pool to the consumer. For example, Connolly et al. (2005b) used this model to demonstrate the importance of seagrass in supporting commercially valuable yellowfin whiting Sil( - lago schomburgkii Peters, 1864) in temperate Australian waters. This contribution would have been unclear from standard interpretation of the natural variation in isotope ratios or with an Isosource analysis. Future Directions.—Future applications of stable isotope techniques in mangrove fish communities should avoid a narrow focus on assessing potential contributions of mangrove detritus to fisheries production (Fry and Ewel, 2003), and instead should work toward a broader understanding of the energy flow through food webs associated with mangroves. When seeking to determine important basal resource pools contributing to mangrove fish communities, sulfur isotope ratios are likely to reveal important insights not possible with use of δ13C and δ15N alone (Connolly et al., 2004). Multivariate statistical analysis derived from δ13C, δ15N, and δ34S values, such as the analysis conducted by Litvin and Weinstein (2004), is one powerful way to identify critical basal resource pools. The Isosource model P( hillips and Gregg, 2003) and spatial-correlation model (Melville and Connolly, 2003) warrant further application and testing. Also, there remain interesting avenues for future research 516 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

regarding potential indirect contributions of mangroves to fishery production, such as the role of mangroves in nutrient cycling (Fry and Ewel, 2003). Ultimately, however, documentation of critical basal resource pools in mangrove fish communities may require use of labeled isotopic tracers to directly track the flow of energy through food webs. In such experiments, a compound enriched in the heavy isotope of an element (typically 15N or 13C) is added to an aquatic system over a defined period of time, and extensive collections of basal resources and consumers are carried out frequently during and after the addition. Energy flow can then be tracked from basal resources to consumers, including quantification of the most important pathways, without significantly altering nutrient availability or rates of biological production. Labeled isotope tracer experiments are common in freshwater streams (e.g., Peterson et al., 1997; Hall et al., 1998; Mulholland et al., 2000; Tank et al., 2000; Hamilton et al., 2001), and also have been utilized in temperate estuaries (Holmes et al., 2000; Hughes et al., 2000; Deegan et al., 2002; Tobias et al., 2003). As an example of the latter, in the oligohaline reach of a Massachusetts estuary, Hughes et al. (2000) were able to use the labeled isotope approach to clearly show that centric diatoms dominated nitrogen uptake and served as the primary vector of nitrogen to upper trophic levels. One central reason there have been no whole-system isotope tracer additions in mangrove ecosystems is the high connectivity between mangroves and other marine and coastal areas, i.e., these systems are relatively “open” (especially fringing mangroves along coastlines or in large estuaries) and characterized by a high degree of material and/or organism movement. However, relatively small, mangrove-lined tidal creeks (e.g., Layman et al., 2004; Valentine et al., 2007) may serve as ideal systems to conduct “whole-ecosystem” stable isotope tracer experiments. Such tidal creeks are more “well-bounded” (sensu Post et al., 2006) than other mangrove fish communities, yet often contain many of the diverse suite of habitat types and organisms that characterize back-reef ecosystems (Dahlgren and Marr, 2004; Adams et al., 2006). Such an approach was applied to track uptake and cycling of nitrogen in a relatively small salt marsh tidal creek (Gribsholt et al., 2005). Alternatively, localized additions of labeled isotopes may help resolve ambiguities among importance of selected basal resources to particular consumers of interest (Mutchler et al., 2004). Although logistically challenging and costly, these labeled isotope additions have the potential to substantially advance our knowledge of the most critical basal resource pools that support mangrove fishes and the food webs of which they are a part.

Can Stable Isotope Ratios Reveal Aspects of Consumer Trophic Structure (i.e., Food Web Structure) in Mangrove Habitats?

Since the earliest applications of stable isotope ratios to study trophic structure, a common approach has been to present bi-plots with species (or individuals, populations, etc.) depicted based on δ13C-δ15N (or δ34S) values. Relative position of species in these plots typically is used to make qualitative inferences about the general structure of the food web, for example, inferring likely dietary sources of focal consumers. Recent studies in freshwater ecosystems, however, have emphasized quantification of particular aspects of tropic structure, such as the trophic position of consumers, food chain length, or ontogenetic diet shifts (e.g., Vander Zanden et al., 1999 ; Post, 2002a, 2003; Layman et al., 2005). For example, (Post, 2002a) proposed a detailed LAYMAN: MANGROVES, FISH, AND ISOTOPES 517

framework for estimating food chain length, a methodology which has allowed for rigorous tests of long-standing questions in ecology (Post et al., 2000). In contrast, isotope studies of trophic structure in mangrove fish communities remain rather qualitative in nature (Marguillier et al., 1997; Bouillon et al., 2002b; Cocheret de la Morinière et al., 2003; Nagelkerken and van der Velde, 2004; Nagelkerken et al., 2006). In this section, I discuss the underlying reasons why even ostensibly straightforward questions regarding food web structure (e.g., trophic position of organisms) can be especially problematic to answer in mangrove fish communities. Many of the difficulties in quantifying aspects of food web structure in these communities arise from the difficulty in establishing a standard isotopic baseline. There can be substantial variation in isotope ratios of resources among habitats or ecosystems, and an isotopic baseline is required to standardize values to a common level to allow for appropriate comparisons (Post, 2002b). In freshwater ecosystems, basal resources are relatively easily isolated at a coarse level (e.g., seston and periphy- ton/detritus in northern U.S. lakes, Post, 2002b), or can be aggregated into distinct, ecologically meaningful, categories (e.g., autochthonous vs allochthonous pools in rivers, Layman et al., 2005). If a primary consumer taxa (e.g., snails and mussels in northern U.S. lakes) provides a spatially- and temporally-integrated measure of isotope ratios of the respective basal resource pools, it then can serve as an appropriate baseline taxa (Post, 2002b). Subsequently, such baseline taxa can be used for calculation of specific aspects of food web structure, such as food chain length, and across- system comparisons are possible (Vander Zanden et al., 1999; Post et al., 2000). In mangrove fish communities, and more generally in estuarine ecosystems, establishing an isotopic baseline is extremely problematic. First, it is still unclear in many mangrove fish communities which of the many potential basal resource pools are most important (see Section I above). Second, these systems typically are inhabited by a very diverse suite of primary consumer taxa, many of which have highly plastic feeding behaviors and may utilize multiple basal resource pools. Even if diet of assumed primary consumers can be well-characterized (based on stomach contents or observations of feeding behavior), it may not reflect what is actually assimilated into tissue. All of these factors, as well as the degree of isotopic discrimination (i.e., processes which lead to different isotope ratios in the tissue of an organism from its diet), affects the realized isotopic signatures of the consumers. In a small tidal creek on Abaco Island, Bahamas, 18 macro-invertebrates in the resident food web (Layman et al., 2007) could be classified as primary consumers. Mean 15δ N values of these taxa range from −3.5‰ to 3.0‰. Since calculation of trophic position is derived directly from a specific primary consumer taxa and the assumed discrimination factor P( ost, 2002b), estimated trophic positions of other taxa in the web could vary substantially depending on the initial choice of “the” baseline primary consumer. Assuming a mean δ15N fractionation of 2.54‰ with trophic transfers (Vanderklift and Ponsard, 2003), a δ15N range of 6.5‰ would span more than two entire trophic levels. As such, trophic position calculations can be especially sensitive to the isotopic baseline and variation at the base of a food web in δ15N values—as is the case in most mangrove communities—rendering precise estimates of trophic positions of species extremely difficult to obtain. Other well-acknowledged caveats associated with application of stable isotope ratios also pertain to the study of trophic structure in mangrove fish communities. Isotopic discrimination is dependent on many factors, and can be quite variable both 518 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

among tissues within a species and among species. Because a single value typically is assumed in a given study, significant error addition can be introduced into trophic position calculations (DeNiro and Epstein, 1981; Macko et al., 1982; Hobson and Clark, 1992; Post, 2002b; Vanderklift and Ponsard, 2003). This may be especially problematic in mangrove communities because organisms vary substantially in taxonomy, body size, metabolic rates, elemental concentrations, and diet—all of which may affect discrimination. The degree to which such sources of error influence characterizations of trophic structure depend on the specificity of the question being asked. If detailed diet information or movement patterns of a particular organism are being sought, extensive laboratory and field studies to quantify isotopic discrimination, tissue turnover, etc. are a necessity (e.g., Fry et al., 2003). Future Directions.—Although characteristics of targeted species (e.g., trophic position) are often difficult to quantify in mangrove ecosystems, stable isotopes may be applied at a different level to reflect characteristics of the food webs in which these species are imbedded. For example, δ13C and δ15N (or δ34S) values provide a representation of the trophic niche of a particular species, as they are the result of all the trophic pathways that have culminated in that individual. As such, position of a species in δ13C-δ15N bi-plot space (or 3-dimensional space with inclusion of δ34S) is one representation of the trophic niche of that species. When all species are plotted into this “niche space”, their dispersion (spacing relative to one another) can be used to examine properties of the entire food web. The total extent of niche space occupied can be quantified as a measure of the overall degree of trophic diversity within the web. Bias at the species level due to the caveats discussed above (e.g., variable discrimination rates) have critical effects on calculations of species-level characteristics, but these effects are less critical on metrics derived across all species in a web—especially when comparing across gradients that have significant effects on food web structure. This may be viewed as one of the promising aspects of the community-wide approach, i.e., by looking for overall patterns in food web structure, intricacies of determining every factor affecting a single species’ isotope ratios are not essential to elucidate. In this manner, quantitative metrics based on species dispersion in niche space can provide a robust means with which to compare food webs across environmental gradients of interest or to evaluate responses of food webs to specific perturbations (Layman et al., 2007). In Figure 1, I provide two (obviously over-simplified) hypothetical food webs where each circle represents a single species in δ13C-δ15N niche space. In each plot, both filled and open circles represent species in the food web before a specific perturbation, and only open circles represent those species that remain following a perturbation. Three metrics could be used to describe trophic structure in these webs: total area of the convex hull encompassing all species (TA), the δ15N range (NR), and δ13C range (CR) (for more detailed description of these and additional metrics, see Layman et al., 2007). In both food web A and B, TA is 4 (the larger triangle in each plot), but NR in A is twice that in B, reflecting real underlying differences in food web structure. Since δ15N typically increases with trophic transfers, a larger NR will likely reflect additional trophic levels within a food web (Minagawa and Wada, 1984; Peterson and Fry, 1987; Post, 2002b). As such, larger NR in A could signify presence of top predators that are not found in B. Alternatively, larger CR in B suggests a more diversified set of basal resources that can support a more diverse set of primary consumers. Post-perturbation (shaded triangles) TA values are 2 in both LAYMAN: MANGROVES, FISH, AND ISOTOPES 519

Figure 1. Two hypothetical food webs as represented by positions of species in δ13C-δ15N bi-plot space. Each circle represents a different species. Filled and open circles are species in an initial food web, with open circles the only species remaining in a food web following a perturbation. The larger triangle indicates the total area (TA) occupied by species in bi-plot space before the perturbation, and the shaded triangle is TA after the perturbation. The shape of the triangles represent underlying differences in food web structure, and how perturbations can affect food webs in different fashions. webs, but each provides a different conceptual scenario illustrating how the metrics may indicate very specific shifts in food web structure. In A, smaller NR drives the reduction in TA. This example could simulate a change in food web structure following over-exploitation of top predators, i.e., “fishing down” food webs (Pauly et al., 1998). In B, the perturbation has primarily affected the diversity of primary consumers, i.e., loss of species that were at the base of the triangle. This quantitative approach also can be used to examine feeding ecology at the population level, as variation in individuals’ δ13C-δ15N signatures in a population provides information as to the degree of intraspecific niche variation (Bearhop et al., 2004). For example, the TA metric provides a direct measure of the amount of overall niche space occupied by a group of individuals, and thus can be used as one representation of the population’s niche width (Layman et al., 2007). In an example from mangrove-lined tidal creeks in the Bahamas (sites described by Valentine et al., 2007), niche width of gray snapper Lutjanus griseus (Linnaeus, 1758) as measured by TA appears to contract in the two creek systems in which tidal connectivity is restricted (fragmentation categories based on Layman et al., 2004) (Fig. 2). This is likely driven by a decrease in prey diversity in fragmented tidal creeks, and an associated decrease in the between-individual variation in resource usage (Van Valen, 1965; Bolnick et al., 2003). Building from extensive applications of stable isotope ratios in mangrove communities, these community-wide and population-level isotope 520 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Figure 2. Plot representing δ13C and δ15N values of nine gray snapper (Lutjanus griseus) individuals from each of three mangrove-lined tidal creeks on Andros Island, Bahamas. One creek was classified as fragmented, one partially fragmented, and one unfragmented, following the categories from Layman et al. (2004). Values given below the plot are the total areas of the convex hull encompassing the nine individuals from each creek. This population-level “total area” metric (TA) provides for a quantitative measure of niche width of each population. metrics may provide new perspectives from which to assess the role of fishes in these food webs.

Can Stable Isotope Ratios be Used to Describe Fish Movement Patterns in Relation to Mangrove Habitat?

Natural variation in stable isotope ratios can be used to infer movement patterns because organisms acquire the signature of their diet which is often habitat-specific (Fry et al., 2003). Thus, if a fish is collected in habitat A but has an isotopic signature reflective of habitat B (which is distinct from A), movement between habitats could be inferred. An obvious requirement for this approach is that habitat types (and prey items in these habitats) differ significantly in at least one stable isotope ratio. Carbon ratios are useful to distinguish movements between habitat types dominated by basal resources pools with different photosynthetic pathways, such as between seagrass and mangroves. Nutrient enrichment can lead to elevated δ15N values (McClelland et al., 1997; McClelland and Valiela, 1998; Fry, 1999; deBruyn and Rasmussen, 2002; Tewfik et al., 2005), and difference in the resulting isotopic baseline may allow for identification of movements between areas with high and low anthropogenic nitrogen inputs (Hansson et al., 1997). Sulfur isotopes are especially useful to distinguish pathways of benthic and pelagic production (Peterson et al., 1986; MacAvoy et al., 1998), and thus can be used to identify movements between shallow and open water areas of an ecosystem. In any movement study based on stable isotope ratios, it is critical that organisms have remained resident in a habitat long enough (or feed consistently in a given area) to become equilibrated to the local food sources, and that additional sources of variability (e.g., variation in the isotopic contents of prey) do not exceed the differences among focal habitat types. Any such study will rely directly on detailed knowledge of tissue turnover times, and thus the rate at which a LAYMAN: MANGROVES, FISH, AND ISOTOPES 521

signature from one site will be lost following a feeding shift in the new habitat (Dale- rum and Angerbjörn, 2005; Herzka, 2005). Ontogenetic and seasonal habitat shifts are the movement patterns that are most likely to be detected with stable isotope ratios. For example, using multivariate analyses based on δ13C, δ15N, and δ34S, Litvin and Weinsetin (2004) were able to provide convincing evidence of ontogenetic movements of juvenile weakfishCynoscion regalis (Bloch and Schneider, 1801) from the Upper Delaware Bay and adjacent marsh areas to open waters near the mouth of the Bay. Fry and colleagues demonstrated that shrimp caught in offshore waters had spent a portion of their life cycle in estuarine habitats (Fry, 1983), and distinguished among individuals that inhabited seagrass beds and mangrove areas (Fry et al., 1999). Such studies provide important insight into aspects of connectivity among coastal habitat types, without the need to individually mark and re-capture numerous individuals of the focal population (Ruben- stein and Hobson, 2004). Attempts also have been made to infer whether organisms undertake daily/tidal feeding migrations, or feed on local resources and thus are resident to that habitat type. Nagelkerken and van de Velde (2004) reported that δ13C values of haemulids (grunts) and lutjanids (snappers) that shelter in mangroves during the day were significantly enriched compared to available food items in the mangrove habitat. They concluded that fishes must move to seagrass areas to feed, as the consumer isotope signatures fell between values of mangrove- and seagrass-associated food items. However, the extent to which this pattern can be attributed to underlying movement patterns is severely limited by the number of basal resource pools, with distinctly overlapping isotope ratio ranges, that may be supporting prey. That is, assuming mangrove and seagrass (or any other subset) of resources are the only possible end- points creates a false dichotomy that may a priori ignore basal resource pools critical to the food web of interest. Isotope ratios also have been used to demonstrate that organisms draw the majority of their energy from local (i.e., a scale of meters) basal resources, and thus used to infer that these organisms undertake little or no among- habitat feeding movements (Weinstein and Litvin, 2000; Hsieh et al., 2002; Guest and Connolly, 2004; Guest et al., 2004, 2006). However, conclusions from all of these studies are tenuous because many basal resource pools can be readily transported at small spatial scales within mangrove ecosystems (Connolly et al., 2005a; Melville and Connolly, 2005), and there may be many other distinct sources of environmental variation that affect local resource isotope ratios H( erzka, 2005). Future Directions.—Simultaneous use of δ13C, δ15N, and δ34S will provide the most comprehensive foundation to infer fish movement patterns (Litvin and Wein- stein, 2004; Herzka, 2005). Within each study system, detailed research is necessary to explicitly document species-specific tissue turnover times, and the time it takes organisms to equilibrate to local food sources. Recent research (MacAvoy et al., 2005; Perga and Gerdeaux, 2005; Sakano et al., 2005; Trueman et al., 2005; MacNeil et al., 2006) continues to highlight the importance of understanding effects of both growth rate and metabolism in determining tissue turnover rates (Fry and Arnold, 1982; Hesslein et al., 1993; Herzka, 2005). Yet more research is direly needed to provide for a general conceptual framework that can be applied broadly across study organisms and systems. Species-specific experiments are a necessity if different tissue types within an organism are to be used to infer movement patterns (Hesslein et al., 1993; Dalerum and Angerbjörn, 2005; Phillips and Eldridge, 2006). For example, Fry 522 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 et al. (2003) conducted extensive field enclosure experiments to examine differential turnover rates among brown shrimp Farfantepenaeus aztecus (Ives, 1891) tissue types, and then applied these rates to assess movement patterns based on the degree to which the different tissues had equilibrated to the local diet signature. For a broader view of the relationship between fishes and mangroves, stable isotopes are most likely to provide useful insights when integrated with other tools and approaches. For example, a combination of isotopic analyses and direct tracking of fish movement patterns would be a powerful approach to detail complex intricacies of fish trophic ecology (Cunjak et al., 2005), especially with recent advances in tagging methodologies and acoustic telemetry technologies that allow for detailed descriptions of movements of individuals. Coupled with stable isotope analyses, feeding migrations of individuals can be identified at multiple temporal and spatial scales. Especially interesting would be study on the incidence of individual trophic special- ists within a fish population (Bearhop et al., 2004). For example, in the Miramichi River in New Brunswick, Canada, a combination of passive integrated transponder (PIT) tags and stable isotope ratio analyses were used to distinguish three apparently distinct trophic groupings within a population of Atlantic salmon. Consistent data from PIT tag monitoring and stable isotope analysis suggested some individuals that remained in tributaries to feed, a second group that resided and fed in the main river channel, and a third group that fed in both the main stem and tributary (Cunjak et al., 2005). Such multi-faceted studies could provide valuable insight into trophic ecology of mangrove fishes, and help resolve ambiguities with respect to the specific role that mangroves fill with regard to fish habitat.

Acknowledgments

Friends of the Environment (a Bahamian NGO) provided invaluable logistical assistance on Abaco Island. Individuals who have provided support on Abaco include R. and A. Ball, T. Berry, D. Haines, K. Rennirt, and J. O’Daniel. Stable isotope data were compiled by C. Mon- taña and J. Quattrochi through work at the Yale University Institute for Biospheric Studies, Center for Stable Isotopic Studies of the Environment, with the support of G. Olack. D. Albrey Arrington, J. Quattrochi, and two reviewers provided comments on an earlier manuscript draft. This work was funded by The Nature Conservancy, Friends of the Environment, and the Acorn Alcinda Foundation.

Literature Cited

Abed-Navandi, D. and P. C. Dworschak. 2005. Food sources of tropical thalassinidean shrimps: A stable-isotope study. Mar. Ecol. Prog. Ser. 291: 159–168. Adams, A. J., C. P. Dahlgren, G. T. Kellison, M. S. Kendall, C. A. Layman, J. A. Ley, I. Nagelkerken, and J. E. Serafy. 2006. The juvenile contribution function of tropical backreef systems. Mar. Ecol. Prog. Ser. 318: 287–301. Bearhop, S., C. E. Adams, S. Waldron, R. A. Fuller, and H. Macleod. 2004. Determining trophic niche width: a novel approach using stable isotope analysis. J. Anim. Ecol. 73: 1007–1012. Benstead, J. P., J. G. March, B. Fry, K. C. Ewel, and C. M. Pringle. 2006. Testing Isosource: stable isotope analysis of a tropical fishery with diverse organic matter sources. Ecology 87: 326–333. Bolnick, D. I., R. Svanbäck, J. A. Fordyce, L. H. Yang, J. M. Davis, C. D. Hulsey, and M. L. Foris- ter. 2003. The ecology of individuals: incidence and implications of individual specialization. Am. Nat. 161: 1–28. LAYMAN: MANGROVES, FISH, AND ISOTOPES 523

Bouillon, S., P. Chandra Mohan, N. Sreenivas, and F. Dehairs. 2000. Sources of suspended organic matter and selective feeding by zooplankton in an estuarine mangrove ecosystem as traced by stable isotopes. Mar. Ecol. Prog. Ser. 208: 79–92. ______, N. Koedam, A. V. Raman, and F. Dehairs. 2002a. Primary producers sustaining macro-invertebrate communities in intertidal mangrove forests. Oecologia 130: 441–448. ______, A. V. Raman, P. Dauby, and F. Dehairs. 2002b. Carbon and nitrogen stable isotope ratios of subtidal benthic invertebrates in an estuarine mangrove ecosystem (Andhra Pradesh, India). Estuar. Coast. Shelf Sci. 54: 901–913. ______, T. Moens, I. Overmeer, N. Koedam, and F. Dehairs. 2004. Resource utilization patterns of epifauna from mangrove forests with contrasting inputs of local versus imported organic matter. Mar. Ecol. Prog. Ser. 278: 77–88. Christensen, J. T., P. G. Sauriau, P. Richard, and P. D. Jensen. 2001. Diet in mangrove snails: Pre- liminary data on gut contents and stable isotope analysis. J. Shellfish Res. 20: 423–426. Cocheret de la Morinière, E., B. J. A. Pollux, I. Nagelkerken, M. A. Hemminga, A. H. L. Huiskes, and G. van der Velde. 2003. Ontogenetic dietary changes of coral reef fishes in the mangrove-seagrass-reef continuum: stable isotopes and gut-content analysis. Mar. Ecol. Prog. Ser. 246: 279–289. Connolly, R. M., D. Gorman, and M. A. Guest. 2005a. Movement of carbon among estuarine habitats and its assimilation by invertebrates. Oecologia 144: 684–691. ______, J. S. Hindell, and D. Gorman. 2005b. Seagrass and epiphytic algae support nutrition of a fisheries species, Sillago schomburgkii, in adjacent intertidal habitats. Mar. Ecol. Prog. Ser. 286: 69–79. ______, M. A. Guest, A. J. Melville, and J. M. Oakes. 2004. Sulfur stable isotopes separate producers in marine food-web analysis. Oecologia 138: 161–167. Cunjak, R. A., J. -M. Roussel, M. A. Gray, J. P. Dietrich, D. F. Cartwright, K. R. Munkittrick, and T. D. Jardine. 2005. Using stable isotope analysis with telemetry or mark-recapture date to identify fish movement and foraging. Oecologia 144: 636–646. Dahlgren, C. and J. Marr. 2004. Back reef systems: Important but overlooked components of tropical marine ecosystems. Bull. Mar. Sci. 75: 145–152. Dalerum, F. and A. Angerjörn. 2005. Resolving temporal variation in vertebrate diets using naturally occurring stable isotopes. Oecologia 144: 647–658. deBruyn, A. M. H. and J. B. Rasmussen. 2002. Quantifying assimilation of sewage-derived organic matter by riverine benthos. Ecol. Appl. 12: 511–520. Deegan, L. A., A. Wright, S. G. Ayvazian, J. T. Finn, H. Golden, R. R. Merson, and J. Harrison. 2002. Nitrogen loading alters seagrass ecosystem structure and support of higher trophic levels. Aquat. Conserv.: Mar. Fresh. Ecosyst. 12: 193–212. DeNiro, M. J. and S. Epstein. 1981. Influence of diet on the distribution of nitrogen isotopes in animals. Geochim. Cosmochim. Acta 45: 341–351. Ewel, K. C., R. R. Twilley, and J. E. Ong. 1998. Different kinds of mangrove forests provide different goods and services. Global Ecol. Biogeography 7: 83–94. Fry, B. 1983. Fish and shrimp migrations in the northern Gulf of Mexico analyzed using stable C, N, and S isotope ratios. Fish. Bull. 81: 789–801. ____. 1999. Using stable isotopes to monitor watershed influences on aquatic trophodynamics. Can. J. Fish. Aquat. Sci. 56: 2167–2171. ____ and C. Arnold. 1982. Rapid 13C/12C turnover during growth of brown shrimp (Penaeus aztecus). Oecologia 54: 200–204. ____ and K. C. Ewel. 2003. Using stable isotopes in mangrove fisheries research - a review and outlook. Isot. Environ. Health Stud. 39: 191–196. ____ and T. J. Smith. 2002. Stable isotope studies of red mangrove and filter feeders from the Shark River Estuary, Florida. Bull. Mar. Sci. 70: 871–890. ____, P. L. Mumford, and M. B. Robblee. 1999. Stable isotope studies of pink shrimp (Farfan- tepenaeus duorarum Burkenroad) migrations on the southwestern Florida shelf. Bull. Mar. Sci. 65: 419–430. 524 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

____, D. M. Baltz, M. C. Benfield, J. W. Fleeger, A.G ace, H. L. Haas, and Z. J. Quinones-Rivera. 2003. Stable isotope indicators of movement and residency for brown shrimp (Farfante- penaeus aztecus) in coastal Louisiana marshscapes. Estuaries 26: 82–97. Gribsholt, B., H. T. S. Boschker, E. Struyf, M. Andersson, A. Tramper, L. De Brabanndre, S. van Damme, N. Brion, P. Meire, F. Dehairs, J. J. Middelburg, and C. H. R. Heip. 2005. Nitro- gen processing in a tidal freshwater marsh: A whole-ecosystem 15N labeling study. Limnol. Oceanogr. 50: 1945–1959. Guest, M. A. and R. M. Connolly. 2004. Fine-scale movement and assimilation of carbon in saltmarsh and mangrove habitat by resident animals. Aquat. Ecol. 38: 599–609. ______, R. M. Connolly, and N. R. Loneragan. 2004. Carbon movement and assimilation by invertebrates in estuarine habitats at a scale of meters. Mar. Ecol. Prog. Ser. 278: 27–34. ______, ______, S. Y. Lee, N. R. Loneragan, and M. J. Breitfaus. 2006. Mechanism for the small-scale movement of carbon among estuarine habitats: organic matter transfer not crab movement. Oecologia 148: 88–96. Hall, R. O., B. J. Peterson, and J. L. Meyer. 1998. Testing a nitrogen-cycling model of a forest stream by using a 15N tracer addition. Ecosystems 1: 283–298. Hamilton, S. K., J. L. Tank, D. F. Raikow, W. M. Wollheim, B. J. Peterson, and J. R. Webster. 2001. Nitrogen uptake and transformation in a midwestern U.S. stream: A stable isotope enrichment study. Biochemistry 54: 297–340. Hansson, S., J. E. Hobbie, R. Elmgren, U. Larsson, B. Fry, and S. Johansson. 1997. The stable nitrogen isotope ratio as a marker of food-web interactions and fish migration. Ecology 78: 2249–2257. Herzka, S. Z. 2005. Assessing connectivity of estuarine fishes based on stable isotope ratio analysis. Estuar. Coast. Shelf Sci. 64: 58–69. Hesslein, R. H., K. A. Hallard, and P. Ramlal. 1993. Replacement of sulfur, carbon, and nitrogen, in tissue of growing broad whitefish (Coregonus nasus) in response to a change in diet traced by 34S, 13C, and 15N. Can. J. Fish. Aquat. Sci. 50: 2071–2076. Hobson, K. A. and R. G. Clark. 1992. Assessing avian diets using stable isotopes II: factors influencing diet-tissue fractionation. Condor 94: 189–197. Holmes, R. M., B. J. Peterson, L. A. Deegan, J. E. Hughes, and B. Fry. 2000. Nitrogen biogeo- chemistry in the oligohaline zone of a new England estuary. Ecology 81: 416–432. Hsieh, H.-L., C.-P. Chen, Y.-G. Chen, and H. H. Yang. 2002. Diversity and benthic organic matter flows through polychaetes and crabs in a mangrove estuary: 13C and 34S signals. Mar. Ecol. Prog. Ser. 227: 145–155. Hughes, J. E., L. A. Deegan, B. J. Peterson, R. M. Holmes, and B. Fry. 2000. Nitrogen flow through the food web in the oligohaline zone of a New England Estuary. Ecology 81: 433– 452. Kieckbusch, D. K., M. S. Koch, J. E. Serafy, and W. T. Anderson. 2004. Trophic linkages among primary producers and consumers in fringing mangroves of subtropical lagoons. Bull. Mar. Sci. 74: 271–285. Layman, C. A., K. O. Winemiller, D. A. Arrington, and D. B. Jepsen. 2005. Body size and trophic position in a diverse tropical food web. Ecology 86: 2530–2535. ______, D. A. Arrington, R. B. Langerhans, and B. R. Silliman. 2004. Degree of fragmentation affects fish assemblage structure in Andros Island (Bahamas) estuaries. Caribbean J. Sci. 40: 232–244. ______, ______, D. M. Post, and C. G. Montaña. 2007. Can stable isotope ratios provide quantitative, community-wide measures of trophic structure? Ecology 88: 42–48. Lee, S. Y. 1995. Mangrove Outwelling - a review. Hydrobiologia 295: 203–212. Litvin, S. Y. and M. P. Weinstein. 2004. Multivariate analysis of stable-isotope ratios to infer movements and utilization of estuarine organic matter by juvenile weakfish (Cynoscion regalis). Can. J. Fish. Aquat. Sci. 61: 1851–1861. LAYMAN: MANGROVES, FISH, AND ISOTOPES 525

Lonergan, N. R., S. E. Bunn, and D. M. Kellaway. 1997. Are mangroves and seagrasses sources of organic carbon for panaeid prawns in a tropical Australian estuary? A multiple stable isotope study. Mar. Biol. 130: 289–300. Lovelock, C. E., I. C. Feller, K. L. McKee, and R. Thompson. 2005. Variation in mangrove forest structure and sediment characteristics in Bocas del Toro, Panama. Caribbean J. Sci. 41: 456–464. MacAvoy, S. E., S. A. Macko, and G. C. Garman. 1998. Tracing marine biomass into tidal freshwater ecosystems using stable sulfur isotopes. Naturwissenschaften 85: 544–546. ______, S. A. Macko, and L. S. Arneson. 2005. Growth versus metabolic tissue replacement in mouse tissues determined by stable carbon and nitrogen isotope analysis. Can. J. Zool. 83: 631–641. Macko, S. A., W. Y. Lee, and P. L. Parker. 1982. Nitrogen and carbon isotope fractionation by two species of marine amphipods: laboratory and field studies. J. Exp. Mar. Biol. Ecol. 63: 145–149. MacNeil, M. A., K. G. Drouillard, and A. T. Fisk. 2006. Variable uptake and elimination of stable nitrogen isotopes between tissues in fish. Can. J. Fish. Aquat. Sci. 63: 345–353. Marguillier, S., G. Van der Velde, F. Dehairs, M. A. Hemminga, and S. Rajagopal. 1997. Trophic relationships in an interlinked mangrove-seagrass ecosystem as traced by 13C and 15N. Mar. Ecol. Prog. Ser. 151: 115–121. McClelland, J. W. and I. Valiela. 1998. Linking nitrogen in estuarine producers to land-derived sources. Limnol. Oceanogr. 43: 577–585. ______, I. Valiela, and R. H. Michener. 1997. Nitrogen-stable isotope signatures in estuarine food webs: A record of increasing urbanization in coastal watersheds. Limnol. Oceanogr. 42: 930–937. Melville, A. J. and R. M. Connolly. 2003. Spatial analysis of stable isotope data to determine primary sources of nutrition for fish. Oecologia 136: 499–507. ______and ______. 2005. Food webs supporting fish over subtropical mudflats are based on transported organic matter not in situ microalgae. Mar. Biol. 148: 363–371. Minagawa, M. and E. Wada. 1984. Stepwise enrichment of 15N along food chains: further evidence and the relation between 15N and animal age. Geochim. Cosmochim. Acta 48: 1135–1140. Mulholland, P. J., J. L. Tank, D. M. Sanzone, W. M. Wollheim, B. J. Peterson, J. R. Webster, and J. L. Meyer. 2000. Nitrogen cycling in a forest stream determined by a 15N tracer addition. Ecol. Monogr. 70: 471–493. Mutchler, T., M. J. Sullivan, and B. Fry. 2004. Potential of 14N isotope enrichment to resolve ambiguities in coastal trophic relationships. Mar. Ecol. Prog. Ser. 266: 27–33. Nagelkerken, I. and G. van der Velde. 2004. Relative importance of interlinked mangroves and seagrass beds as feeding habitats for juvenile reef fish on a Caribbean island. Mar. Ecol. Prog. Ser. 274: 153–159. ______, G. van der Velde, W. C. E. P. Verberk, and M. Dorenbosch. 2006. Segregation along multiple resource axes in tropical seagrass fish community. Mar. Ecol.P rog. Ser. 308: 79–89. Newell, R. I. E., N. Marshall, A. Sasekumar, and V. C. Chong. 1995. Relative importance of benthic microalgae, phytoplankton, and mangroves as sources of nutrition for penaeid prawns and other coastal invertebrates from Malaysia. Mar. Biol. 123: 595–606. Odum, W. E. and E. J. Heald. 1975. The detritus food-web of an estuarine mangrove community. Pages 265–286 in L. Cronin, ed. Estuarine research. Academic Press, New York. Pauly, D., V. Christensen, J. Dalsgaard, R. Froese, and F. Torres. 1998. Fishing down marine food webs. Science 279: 860–863. Perga, M. E. and D. Gerdeaux. 2005. ‘Are fish what they eat’ all year round? Oecologia 144: 598–606. Peterson, B. J. and B. Fry. 1987. Stable isotopes in ecosystem studies. Annu. Rev. Ecol. Syst. 18: 293–320. 526 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

______, M. Bahr, and G. W. Kling. 1997. A tracer investigation of nitrogen cycling in a pristine tundra river. Can. J. Fish. Aquat. Sci. 54: 2361–2367. ______, R. W. Howarth, and R. H. Garritt. 1986. Sulfur and carbon isotopes as tracers of salt-marsh organic matter flow. Ecology 67: 865–874. Phillips, D. L. and P. M. Eldridge. 2006. Estimating the timing of diet shifts using stable isotopes. Oecologia 147: 195–203. ______and J. W. Gregg. 2003. Source partitioning using stable isotopes: coping with too many sources. Oecologia 136: 261–269. ______, S. D. Newsome, and J. W. Gregg. 2005. Combining sources in stable isotope mixing models: alternative methods. Oecologia 144: 520–527. Post, D. M. 2002a. The long and short of food-chain length. Trends Ecol. Evol. 17: 269–277. ______. 2002b. Using stable isotopes to estimate trophic position: Models, methods, and assumptions. Ecology 83: 703–718. ______. 2003. Individual variation in the timing of ontogenetic niche shifts in largemouth bass. Ecology 84: 1298–1310. ______, M. L. Pace, and N. G. J. Hairston. 2000. Ecosystem size determines food-chain length in lakes. Nature 405: 1047–1049. ______, M. W. Doyle, J. L. Sabo, and J. C. Finlay. 2006. The problem of boundaries in defining ecosystems: a potential landmine for uniting geomorphology and ecology. Geomorphol- ogy. In press. Primavera, J. H. 1996. Stable carbon and nitrogen isotope ratios of penaeid juveniles and primary producers in a riverine mangrove in Guimaras, Philippines. Bull. Mar. Sci. 58: 675–683. Rodelli, M. R., J. N. Gearing, P. J. Gearing, N. Marshall, and A. Sasekumar. 1984. Stable isotope ratio as a tracer of mangrove carbon in Malaysian ecosystems. Oecologia 61: 326–333. Rubenstein, D. R. and K. A. Hobson. 2004. From birds to butterflies: animal movement patterns and stable isotopes. Trends Ecol. Evol. 19: 256–263. Sakano, H., E. Fujiwara, S. Nohara, and H. Ueda. 2005. Estimation of nitrogen stable isotope turnover rate of Oncorhynchus nerka. Environ. Biol. Fish. 72: 13–18. Sheaves, M. and B. Molony. 2000. Short-circuit in the mangrove food chain. Mar. Ecol. Prog. Ser. 199: 97–109. Sheridan, P. and C. Hays. 2003. Are mangroves nursery habitat for transient fishes and decapods? Wetlands 23: 449–458. Tank, J. L., J. L. Meyer, D. M. Sanzone, P. J. Mulholland, J. R. Webster, B. J. Peterson, W. M. Wollheim, and N. E. Leonard. 2000. Analysis of nitrogen cycling in a forest stream during autumn using a 15N-tracer addition. Limnol. Oceanogr. 45: 1013–1029. Tewfik, A., J. B. Rasmussen, and K. S. McCann. 2005. Anthropogenic enrichment alters a marine benthic food web. Ecology 86: 2726–2736. Tobias, C. R., M. Cieri, B. J. Peterson, L. A. Deegan, J. Vallino, and J. Hughes. 2003. Processing watershed-derived nitrogen in a well-flushed New England estuary. Limnol. Oceanogr. 48: 1766–1778. Trueman, C. N., R. A. R. McGill, and P. H. Guyard. 2005. The effect of growth rate on tissue-diet isotopic spacing in rapidly growing animals. An experimental study with Atlantic salmon (Salmo salar). Rapid Commun. Mass Spectrom. 19: 3239–3247. Valentine-Rose, L., C. A. Layman, and D. A. Arrington. 2007. Habitat fragmentation decreases fish secondary production in Bahamian tidal creeks. Bull. Mar. Sci. 80: 863–878. Van Valen, L. 1965. Morphological variation and width of the ecological niche. Am. Nat. 99: 377–389. Vander Zanden, M. J., J. M. Casselman, and J. B. Rasmussen. 1999. Stable isotope evidence for the food web consequences of species invasions in lakes. Nature 401: 464–467. Vanderklift, M. A. and S. Ponsard. 2003. Sources of isotopic variation in consumer-diet 15N enrichment: a meta-analysis. Oecologia 136: 169–182. LAYMAN: MANGROVES, FISH, AND ISOTOPES 527

Weinstein, M. P. and S. Y. Litvin. 2000. The role of tidal salt marsh as an energy source for marine transient and resident finfishes: a stable isotope approach. Trans. Am. Fish. Soc. 129: 797–810.

Address: (C.A.L.) Florida International University, Department of Biological Sciences, Ma- rine Sciences Program, 3000 NE 151st St., North Miami, Florida 33181. E-mail: , Telephone: 786-390-0578. BULLETIN OF MARINE SCIENCE, 80(3): 529–553, 2007

SPATIAL AND TEMPORAL VARIATIONS IN MANGROVE AND SEAGRASS FAUNAL COMMUNITIES AT BIMINI, BAHAMAS

Steven P. Newman, Richard D. Handy, and Samuel H. Gruber

ABSTRACT Mangrove and nearshore seagrass macrofaunal communities were concurrently sampled in two areas of contrasting primary productivity (North Sound: low; South Bimini: high) off Bimini, Bahamas. Over 200,000 individuals, comprising 175 species, were identified from catches of block nets, seines, and trawls between March 2000 and March 2003. The Index of Relative Importance (IRI), which is typically used for dietary analysis and combines percentage weight, abundance and occurrence, was applied to catch data to enable easy spatial and temporal variations in community composition. Cluster-analysis revealed distinct mangrove and seagrass communities, with Morisita’s index indicating a greater degree of spatial and temporal homogeneity in the North Sound. Catch diversity and biomass were significantly greater in the mangroves than over seagrass in both locations, and highest off South Bimini. Low productivity, faunal diversity, and abundance in the North Sound were probably due to extreme abiotic variables. Juveniles of most species were present in mangroves and seagrass beds around Bimini, and therefore the protection of mangroves in the Bahamas should be an issue of immediate concern.

Mangroves and seagrass beds have long been recognized as important habitats for fishes (Austin, 1971; Odum et al., 1982; Bell et al., 1984; Thayer et al., 1987; Nagelkerken et al., 2000a,b; Dorenborsch et al, 2004; Mumby et al., 2004), supporting a high diversity of juvenile estuarine and coral reef fish. Mangroves and seagrass beds are important habitats due to the structural complexity, food, shelter, and the protection from predators that they provide (Odum and Heald, 1972; Bell et al., 1984; Blaber et al., 1985; Arrivillaga and Baltz, 1999; Nagelkerken et al., 2000a). Seagrass beds are among the most productive of marine ecosystems, supporting highly diverse fish and macroinvertebrate communities in both temperate and tropical seas. Seagrass beds provide important nursery and feeding grounds for coral reef fishes (Robblee and Zieman, 1984, Arrivillaga and Baltz, 1999), reef-associated predators (Weinstein and Heck, 1979), and commercially important finfish (Arrivillaga and Baltz, 1999; Nagelkerken et al., 2000a). Mangroves are also essential for the growth and development of many marine fishes and there is a high dependence of juveniles on mangroves as nursery areas (Baelde, 1990; Rooker and Dennis, 1991; Nagelkerken et al., 2000a; Mumby et al., 2004). Mangrove fish assemblages have been sampled and described, but a large proportion of the work conducted to date has been qualitative or limited to semi-quantitative visual surveys due to the difficulties with quantitatively sampling the prop root habitat. Studies have identified distinct differences in mangrove and seagrass fish populations, in terms of species composition, numbers and biomass (Robertson and Duke, 1987; Baelde, 1990; Nagelkerken et al., 2000a). The importance of mangroves as nursery habitats is still in question, and there is undoubtedly a need for more quantitative studies in this area (Sheridan and Hays, 2003; Mumby et al., 2004). Faunal communities of the Great Bahama Bank and the Bimini area have previously been described by Newell and Imbrie (1955), Newell et al. (1959), Voss and

Voss (1960), and Böhlke and Chaplin (1968). However, few studies have investigated mangrove and seagrass communities in the Bahamas (Layman and Silliman, 2002; Newman and Gruber, 2002). Despite worldwide evidence indicating the importance of mangroves, they are still unprotected in the Bahamas and the lack of quantitative data on their faunal communities means their role and importance in the life cycle of many marine organisms is not yet fully appreciated. The continual destruction of mangroves by anthropogenic activities could seriously impact commercially important fish stocks and the diversity and replenishment of coral reefs. The primary aim of this study was to quantitatively describe and compare the diversity and biomass of mangrove and seagrass faunal communities in two areas of contrasting standing crop. This study is apparently the first to describe faunal communities using the index of relative importance, which is typically used for dietary analysis.

Materials and Methods

Study Sites.—Mangrove and seagrass faunal communities were sampled in two contrasting lemon shark nurseries at Bimini, Bahamas (25°44´ N 79°16´ W, Fig. 1): the North Sound, an area of low productivity with dwarf mangrove habitat, and South Bimini, an area of high productivity with dense fringing mangroves (Newman, unpubl. data). The North Sound is a 3 km2 semi- enclosed lagoon, while the study site off South Bimini is an open coast line 4 km long (Fig. 1). At the time of this study both areas were undeveloped, but a development has subsequently begun near the North Sound resulting in destruction of some of the mangroves along the western shore. Previous studies have shown the water quality in the North Sound to be variable, displaying relatively high salinities and temperatures (Voss and Voss, 1960). Low tidal flow combined with shallow depth in the North Sound means that water temperature is greatly influenced by cold fronts in the winter and by hot weather in the summer, causing wide and rapid fluctuations (Newell et al., 1959), in a similar manner seen in southern Florida (Zieman, 1982) and across the Great Bahama Bank (Roberts et al., 1983). Salinities in the sheltered reaches of the North Sound fluctuate greatly with rainfall, exceeding 40 during times of low rainfall, and decreasing rapidly to below 31 during times of heavy rainfall (Turekian, 1957). Environmental Conditions.—Abiotic environmental conditions were measured during the course of this study to provide quantification of each study area. Temperature (mercu- ry thermometer, Aqua Chem), and salinity (Sper scientific salt refractometer with automatic temperature compensation) were recorded concurrently with each sampling event. Faunal Collection Techniques.—A combination of block nets, seine nets, and bottom trawls were used to sample macrofauna. Sampling took place between 0900 and 1800 from March 2000 to March 2003. The seine net and block net were both 75 m long and 2 m deep with a 5 mm mesh. Block nets enclosed an area of mangroves in a U-shape, with each end extending above the high tide mark, using natural openings in the mangroves where possible. Block nets were supported by steel rods and weighted to the bottom, allowing organisms to initially move in and out of the area to be enclosed. Block nets were left for 3 d to allow the area to settle from the disturbance of setting, after which they were lifted and tied to the steel rods on the high tide, preventing ingress and egress of all organisms. The seine net was deployed parallel to shore, and closed shorewards with four people acting as “scarers” to concentrate the catch in the area encircled by the net, with the net sampling the entire water column. The trawl had an opening of 1 m × 0.5 m with a 5 mm mesh bag and was pulled perpendicular to shore for a distance of 50 m. Seine nets and bottom trawls complemented each other, with seine nets sampling the water column but under-sampling benthic organisms that were targeted with the bottom trawls. Sampling was conducted with emphasis on quantifying prey items of juvenile lemon sharks; therefore shelled organisms (e.g., gastropods and NEWMAN ET AL.: VARIATION IN MANGROVE AND SEAGRASS MACROFAUNAL COMMUNITIES 531

Figure 1. Location of Bimini, Bahamas. Insert shows location (rectangle) with respect to the southeast United States. Study area off South Bimini was between points A and B, and north of point C in the North Sound. hermit crabs) were excluded. Stratified random sampling locations were triangulated with a rangefinder (DME Rangefinder, Laseroptronix AB) using prepositioned marker poles and the mangrove fringe, as well as with a WAAS GPS (Garmin, GPS 72), enabling spatial analysis of catch from nearshore habitats. All fish and macroinvertebrates were identified to species, measured [in mm, fork length

(LF) and total length (LT) in fish, carapace width (CW) in crabs, and dorsal disc (DD) width in rays] and weighed (± 0.1 g, Ohaus Scout digital scales). Large organisms were weighed to the nearest 10 g using either a 10 kg or 60 kg handheld dial scale. All organisms were released at the site of capture, with in situ observations indicating very low mortality rates. To reduce potential mortality of large catches, species-specific length measurements were taken from a random sample of 30 specimens, which were then weighed and released, with remaining fish counted and weighed only. In extreme cases (≥ 1000 individuals) several hundred were counted and weighed and their average weight used to estimate abundance from the total biomass. Data Analyses.—To analyze spatial variation across mangroves and near shore habitats, seagrass catches were divided into three habitat categories with respect to distance from the mangrove fringe: 0–50 m, 50–100 m, and 100–200 m. All near shore habitats will be referred to as “seagrass” despite seagrass coverage being patchy and variable in density. Organisms were identified to species level, but for clarity and to facilitate comparison were pooled to family for some comparisons. Biomass and abundance data were transformed and expressed 532 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 as densities to account for differences in area and sample size with each sampling technique. Seine nets and trawls targeted different components of the faunal communities. Therefore, densities were complimentary and summed prior to analysis of seagrass communities. Sea- sonal variations in faunal communities were analyzed using May–October for the wet season and November–April for the dry season. Compositions of faunal communities were compared using the index of relative importance (IRI, Pinkas et al., 1971), which is typically used for dietary analysis. Species were pooled into families for ease of analysis. This index incorporates the percentage of each family present by number (%N), wet weight (%W), and frequency of occurrence (%O): IRI =+%%ON# %W ^h. Percentage number and weight were calculated for each sample and then averaged, as each sample was not independent and pooling resource items results in sacrificial pseudoreplica- tion (Krebs, 1999). The index of relative importance was then expressed as a percentage:

n %IRIii= 100 IRI/ IRIi i = 1 .

Cluster analysis (unweighted paired group mean averages and squared Euclidean distances) was used to compare taxonomic composition (% IRI) with respect to location, habitat, and season. Overlap in faunal communities with respect to location and season was calculated using Spearman’s rank correlation coefficient (rs) and the simplified Morisita index (Horn,

1966; CH) calculated as:

S 2/ xyii t i = 1 Cm = S S 2 2 //xi + yi i ==1 i 1

where out of a total of S families in both samples, family i is represented xi times in population

X and yi times in population Y. Spearman’s rank correlation provides a parametric measure of overlap using % IRI values, with rs values ranging from –1 (no similarity) to +1 (identical). The simplified Morisita index is a quantitative measure that is not influenced by the number of categories or sample size because it uses the relative proportions of each species rather than assigning ranks. Values vary between 0 (no categories in common) and 1 (identical categories), with values > 0.60 indicating significant overlap (Zaret and Rand, 1971). In addition to numerical community composition, each habitat was further characterized by catch biomass, the number of species caught, and the size of organisms captured. Catch biomass and sizes were pooled by family. Catch biomass data were transformed prior to analysis (log10 + 1), but these data and species per catch data were still heavily skewed and therefore variations in median catch were compared between the North Sound and South Bimini or between wet and dry seasons using the Mann Whitney U test and Kruskal-Wallis tests. Spatial and temporal variations in the distribution of fish and invertebrate sizes were analyzed using the non-parametric Kolmogorov-Smirnov due to highly skewed data with outliers reducing the effectiveness of testing differences in median sizes despite data transformation. All statistical analyses were performed using Statgraphics Plus (Statistical Graphics Corp.).

Results

Environmental Conditions.—Mean temperatures and salinities were significantly greater in the North Sound than South Bimini throughout the course of this NEWMAN ET AL.: VARIATION IN MANGROVE AND SEAGRASS MACROFAUNAL COMMUNITIES 533

study (t-test: temperature: P = 0.002, using log10 transformed data; salinity: P < 0.0001 using log10 transformed data). Water temperature in the North Sound averaged 27.1 ± 0.4 °C (n = 102), with a low of 13 °C in January 2003 and a high of 36 °C in July 2002. Temperatures off South Bimini averaged 25.8 ± 0.4 °C (n = 66), with a low of 15 °C recorded in January 2003 and a high of 32 °C in July 2002. Mean temperatures were significantly greater during the wet season in both the North Sound (t-test: P <

0.0001, using log10 transformed data) and off South Bimini (t-test: P < 0.0001 using log10 transformed data). The North Sound was generally hypersaline (mean 37.4 ± 0.5) with mean salinities significantly lower in the wet season than in the dry (t-test: P = 0.007 using log transformed data). Salinities off South Bimini averaged 35.0 ± 0.4 and were also significantly lower in the wet season than in the dry season (t-test: P <

0.001 using log10 transformed data). Catch Summary.—In total, 216,150 fishes and macroinvertebrates, weighing 2.2 t, were captured and identified from the mangroves and adjacent seagrass beds off -Bi mini using block nets, seines, and trawls. Twenty-four block nets in the North Sound (enclosing 11,751 m2) and 19 off South Bimini (enclosing 8494 m2). Seagrass faunal communities were sampled with 270 seine nets (enclosing a total area of 120,690 m2) in each study area, in addition to 285 trawls in the North Sound (total area of 14,250 m2) and 213 trawls off South Bimini (10,650 m2). Catches swept 128 species of fish from 49 families, 40 species of crustaceans from 17 families, six species of elasmobranch from four families, and one species of cephalopod. Mangrove Communities.—Mangrove communities in the North Sound consisted of 19 species of teleosts from 11 families and 11 species of crustaceans from five families. Block nets in the North Sound yielded 3322 organisms, with mojarras (Gerreidae), swimming crabs (Portunidae), needlefish (Belonidae), killifish (Cyprin- odontidae), silversides (Atherinidae), and mud crabs (Xanthidae) accounting for 97% of the total numbers caught (Table 1). Mojarras and swimming crabs occurred in 96% of all block nets and were the most prevalent organisms, followed by needlefish (92%) and killifish (63%; Table 1). All families had post-larval juveniles present in the mangroves except for shrimp (Penaeidae, see Table 1), which were probably able to escape through the net mesh. Crustaceans formed a larger proportion of mangrove communities in the North Sound (22% of total abundance, 18% of total biomass) in comparison to mangroves off South Bimini (2% of total abundance and biomass). However, South Bimini mangrove communities were more diverse and consisted of 45 species of teleosts from 20 families and 26 species of crustaceans from 13 families. Block nets off South Bimini captured 25,742 organisms, with mojarras, silversides, grunts (Haemulidae), snappers (Lutjanidae), gobies (Gobiidae), and parrotfish (Scari- dae) accounting for 96% of the mangrove faunal community off South Bimini (see Table 1), with most families present as juveniles. Mangrove catches averaged significantly more species per net haul off South Bimini than in the North Sound in both seasons (Mann Whitney U tests: wet: P = 0.015, dry: P < 0.001). The median number of species caught per block net in the North Sound mangroves did not demonstrate any significant seasonal variation (Mann Whitney U test: P > 0.05), although almost all families were present in greater numbers during the dry season (Table 1). More species were caught per block net set during the wet season off South Bimini, but this was also not significant (Mann Whitney U test: P > 0.05). Silversides, herrings (Clupeidae), swimming crabs, prawns (Palemonidae), and 534 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Table 1. Total unadjusted number (N), weight (W, in g), and occurrence (O) of teleost and crustacean families caught within the mangroves in wet and dry seasons at two locations at Bimini, Bahamas. Sample sizes (number of nets: area sampled, m2) are provided in parentheses. Juve- niles present in *North Sound, † South Bimini. Wet season: May–October, Dry season: November –April.

North Sound South Bimini Wet Dry Wet Dry (7: 3,058 m2) (17: 8,693 m2) (6: 2,038 m2) (13: 6,456 m2) Family N W O N W O N W O N W O Teleosts Acanthuridae 1 34.1 1 1 42.8 1 (surgeonfishes) *†Atherinidae 432 929.0 1 14,122 2,449.6 2 8,093 3,526.2 4 (silversides) †Batrachoididae 13 93.6 2 (toadfishes) *†Belonidae 47 486.5 6 306 5,409.0 16 43 847.2 3 51 607.9 8 (needlefishes) Bothidae 1 161.2 1 (lefteye flounders) †Brotulidae 1 1.7 1 7 6.7 3 (brotulas) †Carangidae 1 4.1 1 (jacks) †Chaetodontidae 3 8.7 2 5 15.7 2 (butterflyfishes) *†Clupeidae 1 4.5 1 148 51.4 3 2 0.9 1 (herrings) *Cyprinodontidae 45 25.2 4 249 94.6 11 (killifishes) *†Gerreidae 218 4,962.8 7 1,507 13,610.5 16 452 4,818.8 6 672 29,068.8 12 (mojarras) *†Gobiidae 1 4.5 1 17 57.5 5 119 173.8 6 230 351.6 12 (gobies) †Haemulidae 36 1,931.6 5 349 9,196.4 9 (grunts) Labridae 2 28.1 1 (wrasses) *†Lutjanidae 7 56.9 4 35 224.8 7 70 984.9 5 293 23,781.3 10 (snappers) †Mullidae 7 13.1 2 (goatfishes) *†Peociliidae 1 0.6 1 2 0.6 2 (livebearers) †Pomacentridae 11 83.3 3 62 213.0 9 (damselfishes) †Scaridae 37 690.8 6 188 2,768.4 11 (parrotfishes) *†Soleidae 1 0.1 1 3 0.4 1 (soles) *†Sphyraenidae 6 124.4 2 20 149.0 6 19 528.7 4 53 2,081.4 9 (barracudas) *†Tetraodontidae 2 24.9 1 29 877.9 9 5 848.7 2 12 1,919.9 6 (puffers) Xenocongridae 1 8.4 1 (false morays) NEWMAN ET AL.: VARIATION IN MANGROVE AND SEAGRASS MACROFAUNAL COMMUNITIES 535

Table 1. Continued.

North Sound South Bimini Wet Dry Wet Dry (7: 3,058 m2) (17: 8,693 m2) (6: 2,038 m2) (13: 6,456 m2) Family N W O N W O N W O N W O Crustaceans †Alphaeidae 11 10.0 4 81 149.0 5 (snapping shrimps) †Calappidae 1 16.0 1 (box crabs) *Dromiidae 1 5.2 1 2 169.6 2 (sponge crabs) †Grapsidae 2 0.5 1 (shore crabs) †Hippolytidae 3 3.1 1 1 0.9 1 (broken back shrimps) †Majidae 2 3.9 1 14 34.4 5 (spider crabs) *†Ocypodidae 1 0.1 1 1 0.3 1 (fiddler crabs) †Palaemonidae 153 4.5 1 143 41.1 4 (prawns) †Palinuridae 1 0.1 1 7 386.9 2 (lobsters) †Penaeidae 1 0.4 1 7 3.6 2 23 8.9 3 4 6.0 4 (shrimps) *†Portunidae 91 919.6 7 646 4,983.9 16 37 632.4 4 37 719.7 9 (swimming crabs) †Pseudosquillidae 15 32.4 6 (false mantis shrimps) *†Xanthidae 20 61.4 4 63 110.2 6 43 147.4 3 46 85.2 9 (mud crabs) shrimp were more abundant in the mangroves during the wet season off South Bi- mini, while all other families were more abundant during the dry season (Table 1). Seagrass Communities.—Seagrass communities in the North Sound comprised 71 species of teleosts from 28 families, 26 species of crustaceans from 11 families, two species of elasmobranchs from two families, and one species of cephalopod. Seine and trawl samples over seagrass in the North Sound yielded 50,190 organisms, weighing 321.9 kg, with silversides (56%), mojarras (33%), penaeid shrimps (3%), needlefish (2%), and swimming crabs (2%) accounting for 96% of all organisms captured (Table 2). South Bimini seagrass communities, like the mangroves, were more diverse than those in the North Sound and comprised of 106 species of teleosts from 40 families, 29 species of crustaceans from 15 families, and three species of elasmobranchs from three families. Catches from seagrass beds off South Bimini yielded 124,553 organisms weighing 738.1 kg, of which herrings and silversides accounted for 94% of the catch numerically, whereas southern stingrays (Dasyatidae) accounted for 59% of the biomass (Table 3). Other major families off South Bimini were mojarras, grunts, parrotfish, needlefish, snappers, and barracuda, which combined represented 82% of the remaining catch (Table 3). Seagrass crustacean catches were almost half those caught in the mangroves numerically, with snapping shrimps (Alphaeidae), prawns, lobsters (Palinuridae), mud 536 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 1 4 7 3 8 8 1 3 1 4 1 3 O . 30 41 12 10 W 1.1 1.2 9.1 37.6 25.6 22.0 35.5 18.4 380.7 512.6 263.7 374.7 1,040.0 1,502.6 7,174.4 2,003.1 1 8 5 1 4 1 9 1 N . Dry (45, 50) 42 28 14 16 36 14 116 998 3 1 2 6 8 8 1 1 1 5 8 1 O . 21 41 12 15 W 100–200 m 3.3 1.4 5.5 2.9 38.5 47.5 23.6 14.9 218.7 530.5 625.7 646.1 636.1 617.7 4,907.5 16,452.4 4 2 8 1 2 1 5 1 N . Wet (45, 45) Wet 46 36 20 51 49 319 146 4,131 5 1 3 6 3 1 2 3 3 1 O . 32 45 12 10 W 4.3 4.5 4.96 1.4 26.7 30.9 67.6 14.2 416.6 228.3 528.5 3,733.3 4,757.0 28,163.2 Dry (45, 50) 1 5 5 2 3 3 1 N . 15 17 17 17 46 206 1,965 3 6 4 4 5 2 6 1 8 2 O . 26 43 22 50–100 m W 1.9 10.2 22.5 23.6 35.0 23.3 269.6 483.1 179.9 692.3 2,536.1 1,632.6 22,220.9 5 8 5 6 2 6 1 7 N . Wet (45, 45) Wet 81 54 51 588 4,538 2 2 4 1 1 4 5 2 1 1 4 O . 35 43 10 W 0.1 5.1 2.1 1.9 16.6 13.4 94.2 492.0 281.2 790.2 185.2 3,912.3 14,725.4 21,405.7 2 5 1 1 4 2 7 2 N . Dry (45, 50) 11 18 14 249 476 1,375 4 5 1 5 2 1 1 3 5 3 1 1 O . 25 46 17 0–50 m W 4.1 2.0 2.1 1.5 3.9 61.8 37.8 . *Juveniles present. 2 348.0 135.0 845.1 875.1 4,134.7 30,789.2 84,130.2 40,169.6 1 2 1 2 9 3 3 1 N . 20 13 49 34 Wet (45, 45) Wet 218 3,623 26,852 and each trawl 50 m 2 Teleosts Table 2. Total unadjusted number (N), weight (W, in g), and unadjusted occurrence number (O) (N), of weight teleost, (W, Total 2. crustacean, elasmobranch, and cephalopod families Table caught within 200 m Each of the mangrove fringe in the North Sound at Bimini, Bahamas. Catches were separated parentheses. by distance from the mangrove fringe (0–50 m, 50–100 m, 100–200 m) in provided are trawls) of number nets, seine of (number sizes Sample November–April). season: dry May–October, season: (wet season by and seine covered an area of 447 m *Albulidae (bonefishes) *Atherinidae (silversides) *Balistidae (triggerfish and filefishes) *Batrachoididae (toadfishes) *Belonidae (needlefishes) *Bothidae (lefteye flounders) *Carangidae (jacks) *Chaetodontidae (butterflyfishes) *Clinidae (clinids) *Clupeidae (herrings) *Cynoglossidae (tonguefishes) Cyprinodontidae (killifishes) Fistulariidae (cornetfishes) *Gerreidae (mojarras) *Gobiidae (gobies) *Haemulidae (grunts) *Hemiramphidae (halfbeaks) *Labridae (wrasses) *Lutjanidae (snappers) *Ostraciidae (trunkfishes) *Scaridae (parrotfishes) *Scorpaenidae (scorpionfishes) NEWMAN ET AL.: VARIATION IN MANGROVE AND SEAGRASS MACROFAUNAL COMMUNITIES 537 5 1 1 2 2 4 3 7 1 1 O . 15 13 40 39 28 W 9.2 1.5 3.8 3.1 0.4 16.1 49.6 16.4 115.8 327.2 101.3 136.7 721.5 1,103.5 1,247.4 5 1 1 2 2 5 3 1 2 N . Dry (45, 50) 23 24 77 12 181 109 8 4 4 2 1 9 1 O . 11 11 11 41 45 W 100–200 m 2.2 0.7 25.7 13.1 633.4 143.0 141.3 647.6 417.0 199.9 138.4 275.6 4 2 1 1 N . 11 Wet (45, 45) Wet 14 35 16 15 15 193 165 7 1 2 4 1 2 1 1 1 O . 17 15 38 49 21 10 W 7.5 0.6 6.3 0.6 1.4 76.6 12.4 76.7 39.5 23.6 167.7 372.8 3,931.0 1,633.8 6,618.0 Dry (45, 50) 8 1 2 4 1 3 1 1 1 N . 51 30 99 50 16 198 2 1 2 1 1 2 8 O . 14 16 10 37 44 13 50–100 m W 0.0 4.3 36.6 25.2 90.1 88.6 19.9 102.9 289.5 507.5 140.4 1,332.3 1,900.0 2 1 2 1 1 2 N . Wet (45, 45) Wet 33 30 13 10 19 382 267 3 1 2 1 3 1 1 1 2 1 7 O . 25 12 18 32 41 W 1.2 0.6 1.2 0.0 7.2 2.0 11.8 60.3 53.0 88.3 20.4 204.0 233.2 650.4 1,500.0 16,916.6 3 1 5 1 8 1 1 1 2 1 N . Dry (45, 50) 27 43 99 13 131 167 8 5 1 2 1 8 4 O . 23 15 43 47 0–50 m W 9.0 69.1 30.3 38.9 49.1 105.6 121.4 179.3 182.5 9,405.1 10,650.5 9 7 1 2 1 5 N . 35 10 Wet (45, 45) Wet 114 354 230 Teleosts Table 2. Continued. Table *Soleidae (soles) *Sparidae (porgies) *Sphyraenidae (barracudas) *Syngnathidae (pipefishes) *Synodontidae (lizardfishes) (puffers) *Tetraodontidae (searobins) Triglidae Crustaceans Alphaeidae (snapping shrimps) *Dromiidae (sponge crabs) *Hippolytidae (broken back shrimps) *Majidae (spider crabs) *Neogonodactylidae (mantis shrimps) *Palinuridae (spiny lobsters) Parthenopidae crabs) (elbow *Penaeidae (penaeid shrimps) *Portunidae (swimming crabs) *Pseudosquillidae (mantis shrimps) *Xanthidae (mud crabs) Elasmobranchs *Carcharhinidae (requiem sharks) *Dasyatidae (whiptail stingrays) Cephalopods Loliginidae (inshore squid) 538 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

2 - 1 1 4 2 4 8 4 1 1 1 1 8 1 O . 33 10 23 12 16 32 W 7.1 4.1 6.1 5.2 7.6 0.5 0.4 11.0 55.1 27.1 15.6 410.3 107.7 6,404.2 4,396.3 1,626.6 2,148.4 1,982.8 2,600.0 1 8 1 4 3 1 8 1 1 1 4 N . Dry (45, 36) 17 74 62 34 14 196 237 6,657 6 4 6 2 7 9 2 9 2 2 1 1 O . 14 12 22 29 39 100–200 m W 1.2 3.0 0.6 30.9 35.1 35.7 21.0 104.7 621.5 122.8 562.5 185.0 3,437.6 2,675.7 4,465.2 2,496.1 34,700.3 6 7 2 7 2 2 1 1 N . Wet (45, 38) Wet 17 24 10 14 748 109 182 480 90,620 5 5 4 1 2 2 8 2 1 O . 26 22 25 21 W 9.1 3.2 7.8 1.7 12.8 60.0 801.7 794.3 431.0 6,113.4 1,006.1 3,424.3 16,546.3 8 4 1 2 2 1 N . Dry (45, 36) 11 89 32 938 173 129 1,282 5 8 2 5 2 7 3 1 1 1 4 O . 11 16 19 32 32 50–100 m W 5.9 1.7 4.9 31.7 48.1 28.4 34.0 49.2 808.7 742.0 1,045.9 1,781.0 7,558.8 2,499.3 3,153.6 1,950.0 7 3 3 8 2 1 1 4 N . Wet (45, 32) Wet 12 23 24 471 205 446 1,256 6,105 2 4 1 3 1 1 1 1 O . 30 38 23 16 W 1.3 1.3 0.3 7.2 13.7 117.0 150.3 961.7 3,873.3 2,367.6 2,398.3 11,906.4 4 1 3 1 1 1 1 Dry (45, 40) N . 72 319 298 145 2,955 1 1 1 4 2 1 2 4 O . 0–50 m 16 25 33 17 22 W 1.6 1.4 0.7 15.3 16.2 88.9 270.4 987.5 924.2 3,172.3 1,197.1 3,981.1 11,166.5 1 1 1 4 6 5 4 N . Wet (45, 31) Wet 35 36 114 134 694 5,879 . *Juveniles present. 2 Echeneidae (remoras) Table 3. Total unadjusted in number g), (N), 3. Total and weight occurrence (W, (O) Table of teleost, crustacean, and elasmobranch families caught within 200 m of the man and each trawl 50 m * Acanthuridae (surgeonfishes) Albulidae (bonefishes) * Antennariidae (frogfishes) * Apogonidae (cardinalfishes) * Atherinidae (silversides) Aulostomidae (trumpetfishes) filefishes) and * Balistidae (triggerfish * Batrachoididae (toadfishes) * Belonidae (needlefishes) * Bothidae (lefteye flounders) * Brotulidae (brotulas) * Carangidae (jacks) * Chaetodontidae (butterflyfishes) * Clinidae (clinids) * Clupeidae (herrings) Dactylopteridae (flying gurnards) Dactyloscopidae (stargazers) * Fistulariidae (cornetfishes) * Gerreidae (mojarras) * Gobiidae (gobies) * Haemulidae (grunts) * Hemiramphidae (halfbeaks) grove fringe off South Bimini, Bahamas. Catches are separated by distance from the mangrove dry: November–April). Sample sizes (number of seine nets, number of trawls) are provided in parentheses. Each seine covered an area of 447 m May–October, fringe (0–50 m, 50–100 m, 100–200 m) and by season (wet: Teleosts NEWMAN ET AL.: VARIATION IN MANGROVE AND SEAGRASS MACROFAUNAL COMMUNITIES 539 5 1 5 4 8 1 1 1 2 O . 11 13 18 41 18 W 3.5 2.4 1.4 2.8 4.7 11.2 47.2 63.8 102.4 675.5 576.0 680.6 8,327.7 2,313.5 7 1 6 3 4 8 1 2 2 N . Dry (45, 36) 35 28 39 28 151 2 4 4 3 5 1 9 1 2 O . 17 12 54 13 22 31 100–200 m W 6.7 1.9 6.2 0.4 14.5 36.9 19.1 189.3 475.8 121.1 2,996.1 1,673.3 2,398.9 2,022.9 2,305.4 5 6 3 8 1 1 2 N . Wet (45, 38) Wet 44 38 15 37 34 12 342 140 4 6 3 1 2 8 1 1 5 1 8 1 1 O . 14 14 37 14 W 7.2 0.4 1.2 0.3 0.1 0.2 7.5 69.8 42.4 23.3 27.8 476.7 798.7 6,999.6 4,348.6 2,885.0 1,003.6 3 1 2 1 1 5 1 9 1 1 N . Dry (45, 36) 17 12 22 20 31 16 154 8 2 8 4 2 1 2 3 2 O . 11 44 16 28 10 50–100 m W 2.4 0.4 3.6 0.7 3.8 44.7 25.7 12.6 46.2 7,496.2 2,029.7 3,127.4 6,526.3 1,993.4 2 4 2 2 2 3 2 N . Wet (45, 32) Wet 13 20 24 54 12 112 247 7 1 1 4 4 3 7 2 1 9 O . 16 15 37 31 W 4.8 6.9 7.8 1.4 35.7 12.4 50.4 58.7 1,894.6 9,545.9 1,085.1 1,027.3 1,527.7 20,837.3 7 1 1 5 6 3 8 3 1 Dry (45, 40) N . 27 23 18 148 156 5 1 1 1 3 2 1 2 6 3 9 O . 0–50 m 19 13 37 33 W 2.8 4.8 1.7 6.6 5.2 3.9 67.3 16.9 59.8 971.1 320.0 2,288.8 7,547.7 1,456.1 10,540.2 1 1 1 4 2 2 3 6 3 N . Wet (45, 31) Wet 11 12 27 88 111 252 Table 3. Continued. Table * Labridae (wrasses) * Lutjanidae (snappers) * Mullidae (goatfishes) Ophichthidae (snake eels) * Ostraciidae (trunkfishes) * Pomacentridae (damselfishes) * Scaridae (parrotfishes) * Scorpaenidae (scorpionfishes) Serranidae (sea basses) Soleidae (soles) * Sparidae (porgies) * Sphyraenidae (barracudas) * Syngnathidae (pipefishes) * Synodontidae (lizardfishes) (puffers) * Tetraodontidae Crustaceans * Alphaeidae (snapping shrimps) * Calappidae (box crabs) * Dromiidae (sponge crabs) (broken back shrimps) * Hippolytidae Holocentridae (squirrelfishes) Teleosts 540 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 1 9 3 5 3 1 O . 12 W 4.6 4.6 0.3 11.5 22.1 468.1 10,2293.6 1 5 6 3 1 N . Dry (45, 36) 14 24 1 1 9 5 1 1 1 O . 13 100–200 m W 0.1 3.1 6.1 2.0 0.8 13.6 13.9 33,500.0 1 1 5 1 1 1 N . Wet (45, 38) Wet 18 10 7 9 1 2 3 5 1 1 1 O . W 9.9 0.1 1.6 1.6 3.3 12.5 400.0 1,900.0 105,630.8 8 1 2 4 6 1 2 1 N . Dry (45, 36) 13 2 2 1 9 1 6 O . 11 12 50–100 m W 5.9 7.8 1.6 14.6 14.0 25.6 374.9 42,500.0 2 2 1 2 N . Wet (45, 32) Wet 11 16 12 12 7 4 8 2 2 6 5 O . 10 W 6.1 9.3 3.9 21.4 17.1 204.8 2,920.1 103,773.0 8 4 6 2 2 8 Dry (45, 40) N . 11 10 3 2 7 8 5 1 3 3 O . 0–50 m 17 W 7.4 5.2 3.7 14.0 19.8 30.9 183.2 9,942.7 49,348.0 3 4 9 7 1 3 3 N . Wet (45, 31) Wet 17 23 Table 3. Continued. Table * Majidae (spider crabs) shrimps) * Neogonodactylidae (mantis shrimps) * Palaemonidae (commensal * Palinuridae (spiny lobsters) * Penaeidae (penaeid shrimps) * Portunidae (swimming crabs) * Pseudosquillidae (mantis shrimps) * Squillidae (mantis shrimps) * Xanthidae (mud crabs) Elasmobranchs * Carcharhinidae (requiem sharks) Dasyatidae (whiptail stingrays) (nurse sharks) * Ginglymostomatidae Crustaceans NEWMAN ET AL.: VARIATION IN MANGROVE AND SEAGRASS MACROFAUNAL COMMUNITIES 541 crabs, and swimming crabs more abundant in the mangroves. In both the North Sound and South Bimini, catches of mojarra consisted largely of two species, with yellowfin mojarra Gerres ( cinereus, Walbaum 1792) the dominant species in the mangroves and slender mojarra (Eucinostomus jonesii, Günther 1879) more abundant over adjacent sand and seagrass. Index of Relative Importance.—Index of relative importance values are presented for each family captured in the North Sound (Table 4) and South Bimini (Ta- ble 5). These values provide a combination of abundance, biomass, occurrence, and facilitate comparison of faunal communities within each area. Index of relative importance values revealed mojarra, swimming crabs, shrimps, and needlefish were the dominant organisms throughout the North Sound. Barra- cuda were most commonly found within 50 m of the mangrove fringe, while the importance of grunts, parrotfish, halfbeaks (Hemiramphidae), and trunkfish (Ostra- ciidae) increased with distance from the mangroves (Table 4). Community composition varied seasonally, with mojarra forming a larger proportion (over 70% IRI) of the community over 50 m from the mangrove fringe during the wet season, whereas communities in or near the mangroves were dominated by silversides, mojarras, and swimming crabs (Table 4). During the dry season, swimming crabs and shrimps formed a large proportion of the communities farthest from the mangroves. Faunal communities off South Bimini, in contrast to those in the North Sound, were not dominated by just a few families (see Table 5). Mojarra were not as important off South Bimini, but they still formed a significant proportion of mangrove communities in both seasons (wet: 33% IRI, dry: 35% IRI), with importance decreasing with increasing distance from the mangroves (Table 5). Snapper were important in the mangrove community during the dry season, but during the wet season they were more important in the seagrass. Seagrass communities off South Bimini were dominated by parrotfish, grunts, stingrays, and trunkfish (Table 5). Like the North Sound, barracuda formed a large proportion of faunal communities near the mangrove fringe, and silversides were highly abundant nearshore during the wet season, but were farther out during the dry season (Table 5). Reef fishes such as surgeonfish (Acanthuridae), jacks (Carangidae), butterflyfish (Chaetodontidae), and damselfish (Pomacentridae) were more common off South Bimini than in the North Sound, although none of these families comprised a significant proportion of faunal communities at any distance (Table 5). Spatial and Temporal Variation in Faunal Communities.—Cluster analysis using index of relative importance values for the 64 families collected revealed distinct mangrove and seagrass communities for North Sound and South Bimini (Fig. 2). Seagrass faunal communities in the North Sound differed from communities off South Bimini, and were seasonally more uniform (Fig. 2, note Euclidean distances). While North Sound seagrass communities were closely grouped by season, off South Bimini they were more closely grouped by distance from the mangrove fringe (Fig. 2). Spearman’s rank correlations indicated a high degree of overlap in community structure with respect to habitat, season, and location (Table 6). Morisita’s index revealed a greater degree of homogeneity in faunal communities in the North Sound than off South Bimini (Table 6). South Bimini mangrove communities were more similar to mangrove and seagrass communities in the North Sound than their own fringing seagrass communities. 542 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Table 4. Index of Relative Importance (IRI) values for families captured in the North Sound at Bimini, Bahamas. Values are divided by distance from mangrove fringe and season (wet: May– October, dry: November–April). Index of relative importance was calculated as (% number + % weight) * % occurrence and then expressed as a percentage. Values > 1% IRI are highlighted in bold. Blank cells indicate species not present, with values of 0.000 indicating species present. See Tables 1 and 2 for raw abundance, biomass, and occurrence data from the North Sound.

Seagrass Seagrass Seagrass Mangroves 0–50 m 50–100 m 100–200 m Family Wet Dry Wet Dry Wet Dry Wet Dry Acanthuridae (surgeonfishes) Albulidae (bonefishes) 2.423 0.197 0.004 0.003 0.092 Atherinidae (silversides) 0.587 22.740 0.227 0.772 0.039 0.084 0.210 Balistidae (triggerfish and filefishes) 0.000 0.010 0.000 0.005 0.102 Batrachoididae (toadfishes) 0.158 Belonidae (needlefishes) 9.885 16.850 2.652 19.705 1.026 2.184 0.778 5.546 Carangidae (jacks) 0.013 0.001 0.010 0.005 0.025 0.014 Chaetodontidae (butterflyfishes) 0.002 0.033 0.001 Clupeidae (herrings) 0.002 0.023 0.000 0.000 0.024 Cyprinodontidae (killifishes) 5.928 3.073 0.001 Gerreidae (mojarras) 57.595 55.026 54.420 50.065 78.928 32.276 72.616 49.195 Gobiidae (gobies) 0.034 0.130 0.000 0.003 Haemulidae (grunts) 0.023 0.313 0.002 0.443 0.104 Hemiramphidae (halfbeaks) 0.143 0.040 0.005 0.044 0.966 0.534 Labridae (wrasses) 0.003 0.001 0.002 0.022 Lutjanidae (snappers) 0.967 0.473 0.421 0.209 3.062 1.067 0.990 0.776 Mullidae (goatfishes) Ostraciidae (trunkfishes) 0.052 0.085 0.278 0.088 3.224 2.323 Pomacentridae (damselfishes) Scaridae (parrotfishes) 0.001 0.141 0.044 1.005 0.693 0.623 Soleidae (soles) 0.033 0.281 0.374 0.589 1.273 0.290 0.583 Sparidae (porgies) 0.016 0.008 0.093 0.026 Sphyraenidae (barracudas) 0.448 0.248 5.085 11.864 0.875 0.308 0.383 0.696 Tetraodontidae (puffers) 0.071 1.339 0.011 0.406 0.063 1.337 0.157 3.443 Palinuridae (lobsters) 0.000 0.000 Penaeidae (shrimps) 0.033 0.016 5.383 7.410 5.684 18.129 5.967 23.007 Portunidae (swimming crabs) 22.296 21.763 5.413 8.803 7.689 40.813 12.788 12.024 Pseudosquillidae (false mantis shrimps) 0.004 0.005 0.004 0.010 0.041 Xanthidae (mud crabs) 2.678 0.493 0.100 0.314 0.415 0.866 0.276 0.300 Carcharhinidae (requiem sharks) 0.740 0.038 0.080 0.015 0.110 Dasyatidae (whiptail stingrays) 0.014 Other 0.033 0.000 0.096 0.066 0.114 0.386 0.190 0.206

Spatial and Temporal Variation in Species Diversity and Catch Bio- mass.—No significant seasonal variation was observed in the median number of species caught in the North Sound in the mangroves or seagrass (Kruskal-Wallis test: P > 0.05; Fig. 3A). However, significantly more species were caught per net set in the mangroves off South Bimini than over the seagrass beds, in both the wet and dry season (Kruskal-Wallis test: P < 0.001; Fig. 3B). Catch of species over the seagrass beyond 50 m from the mangroves off South Bimini was significantly greater than that in the North Sound (Mann Whitney U NEWMAN ET AL.: VARIATION IN MANGROVE AND SEAGRASS MACROFAUNAL COMMUNITIES 543

Table 5. Index of Relative Importance values for organism families captured off South Bimini, Bahamas. Values are divided up by distance from mangrove fringe and season (wet: May–October, dry: November–April). Index of relative importance was calculated as (% number + % weight) * % occurrence and then expressed as a percentage. Values > 1% IRI are highlighted in bold. Blank cells indicate species not present, with values of 0.000 indicating species present. See Tables 1 and 3 for raw abundance, biomass, and occurrence data from South Bimini.

Seagrass Seagrass Seagrass Mangroves 0–50 m 50–100 m 100–200 m Family Wet Dry Wet Dry Wet Dry Wet Dry Acanthuridae (surgeonfishes) 0.037 0.004 0.005 0.005 0.044 0.213 0.000 Albulidae (bonefishes) 0.078 0.049 Atherinidae (silversides) 33.019 21.390 35.770 8.207 7.373 25.473 1.408 59.490 Balistidae (triggerfish and filefishes) 0.001 0.001 0.048 0.138 0.216 0.243 Batrachoididae (toadfishes) 0.032 0.007 0.004 0.008 0.000 0.004 Belonidae (needlefishes) 2.795 0.672 3.275 17.303 11.855 13.075 9.438 3.018 Carangidae (jacks) 0.005 0.137 0.159 0.061 0.589 0.239 0.167 Chaetodontidae (butterflyfishes) 0.024 0.009 0.005 0.128 0.037 0.118 0.002 Clupeidae (herrings) 0.600 0.001 0.002 6.482 2.243 18.706 0.358 Cyprinodontidae (killifishes) Gerreidae (mojarras) 33.024 35.023 11.558 7.755 4.118 2.039 5.460 0.575 Gobiidae (gobies) 1.799 2.083 0.963 6.969 0.547 0.504 0.112 0.563 Haemulidae (grunts) 10.315 9.075 3.418 2.600 13.313 6.282 15.519 6.571 Hemiramphidae (halfbeaks) 0.251 0.004 0.394 0.513 0.757 1.280 Labridae (wrasses) 0.004 0.165 0.110 0.049 1.452 0.544 0.471 Lutjanidae (snappers) 5.514 22.277 2.448 1.211 4.652 0.329 7.546 0.437 Mullidae (goatfishes) 0.011 0.005 0.001 0.003 0.066 0.072 0.013 Ostraciidae (trunkfishes) 2.807 3.291 3.361 2.526 2.522 2.899 Pomacentridae (damselfishes) 0.295 0.513 0.143 0.425 0.171 0.097 0.034 0.002 Scaridae (parrotfishes) 4.569 3.908 21.736 19.563 37.662 28.473 28.660 14.911 Soleidae (soles) 0.002 0.003 Sparidae (porgies) 0.084 0.100 1.062 0.698 2.213 Sphyraenidae (barracudas) 2.295 1.908 10.494 16.967 4.453 1.248 3.162 1.030 Tetraodontidae (puffers) 1.792 1.035 0.009 0.120 0.009 0.534 0.122 0.148 Palinuridae (lobsters) 0.001 0.075 0.010 0.000 Penaeidae (shrimps) 0.096 0.012 0.276 0.052 0.595 0.520 0.293 0.265 Portunidae (swimming crabs) 2.801 0.765 0.164 0.248 1.023 0.031 0.061 1.408 Pseudosquillidae (false mantis shrimps) 0.073 0.162 0.182 0.383 0.016 0.002 0.027 Xanthidae (mud crabs) 0.591 0.324 0.993 0.439 0.248 0.091 0.029 0.063 Carcharhinidae (requiem sharks) 0.821 0.046 Dasyatidae (whiptail stingrays) 4.064 13.482 1.694 12.517 2.337 5.691 Other 0.428 0.803 0.244 0.797 0.190 0.446 0.213 0.315 tests: P < 0.05, Figs. 3A,B). However, there was no significant difference in species diversity of nets hauled in each location at all other distances (Mann Whitney U tests: all P > 0.05). No significant seasonal variation was observed in the number of species captured per seine haul over seagrass in the North Sound (Mann Whitney U tests: all P > 0.470; Fig. 3A). However, off South Bimini, median catch of species was significantly greater at distances beyond 50 m from the mangroves during the wet season (Mann Whitney U tests: P < 0.003), although there was no seasonal variation 544 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Figure 2. Dendrogram from cluster analysis (unweighted paired group mean averages and squared Euclidean distances) based on the index of relative importance of 64 families collected from mangrove and seagrass beds in the North Sound and off South Bimini at Bimini, Bahamas. Index of relative importance is calculated as [(% number + % weight) × % occurrence] and expressed as a percentage. Location codes: NS, North Sound; SB, South Bimini. Habitat codes: 0, mangroves; 50, seagrass within 50 m of mangrove fringe; 100, seagrass 50–100 m; 200, seagrass 100–200 m from mangrove fringe. Season codes: D, dry season (November–April); W, wet season (May–Oc- tober). For example, the code NS0W is North Sound mangrove community in the wet season. Dotted line highlights the difference between mangrove and seagrass communities, and seagrass communities in the North Sound and South Bimini, as determined by the cluster analysis. in the number of species caught within 50 m of the mangroves (Mann Whitney U test: P = 0.788; Fig. 3B). Mangrove catch biomass did not vary significantly with season within each location (Mann-Whitney U tests: North Sound P = 0.949, South Bimini P = 0.557; Figs. 3C,D). Median mangrove catch biomasses were significantly greater than seagrass catches in both seasons off South Bimini (Kruskal-Wallis test: dry season: P = 0.001; wet season: P = 0.006; Fig. 3D). Catch biomass in the North Sound was significantly greater within 50 m of the mangroves than all other habitats during the wet season (Kruskal-Wallis test: P = 0.012; Fig. 3C), and significantly less in seagrass over 100 m from the mangroves in the dry season (Kruskal-Wallis test: P = 0.007; Fig. 3C). Man- grove and seagrass catch biomasses were significantly greater off South Bimini than in the North Sound during the dry season at all distances except 50–100 m from shore (Mann Whitney U tests: P < 0.05, with P = 0.524 at 5–100 m, Figs. 3C,D), but there was no significant difference in median catch biomass in each area during the wet season (Mann Whitney U tests, P > 0.05; Fig. 3). Seagrass catch biomasses off South Bimini increased at all distances during the dry season, although none of these were significantly greater than catches during the wet season (Mann Whitney U tests: P > 0.05, Fig. 3D). A significant increase in catch biomass was observed within 50 m of the mangroves during the wet season in the North Sound (Mann Whitney U test: P = 0.006), while there were no significant differences in catch biomass at all other distances (Mann Whitney U tests: P > 0.05; Fig. 3C). Silversides were caught in large numbers within 50 m of the mangroves dur- NEWMAN ET AL.: VARIATION IN MANGROVE AND SEAGRASS MACROFAUNAL COMMUNITIES 545 – 200 0.02 0.25 0.02 0.40 0.38 0.03 0.86 0.04 0.74 0.02 0.03 0.46 0.69 0.04 0.37 – 100 0.05 0.68 0.05 0.83 0.80 0.27 0.09 0.87 0.16 0.09 0.06 0.39 0.54 0.12 0.77 Dry – 50 0.18 0.62 0.17 0.73 0.63 0.30 0.23 0.66 0.42 0.71 0.24 0.15 0.35 0.40 0.29 0 – 0.67 0.34 0.66 0.33 0.44 0.82 0.65 0.54 0.63 0.66 0.49 0.66 0.45 0.91 0.65 – South Bimini 200 0.13 0.12 0.92 0.77 0.14 0.15 0.67 0.48 0.21 0.84 0.15 0.10 0.45 0.30 0.17 – 100 0.10 0.83 0.09 0.82 0.17 0.13 0.75 0.65 0.20 0.81 0.13 0.10 0.53 0.36 0.16 Wet – 50 0.23 0.76 0.23 0.80 0.72 0.54 0.24 0.80 0.30 0.50 0.24 0.17 0.53 0.79 0.25 0 – 0.62 0.60 0.62 0.58 0.51 0.87 0.62 0.55 0.59 0.64 0.56 0.63 0.44 0.77 0.63 – 200 0.90 0.61 0.87 0.60 0.63 0.86 0.47 0.60 0.91 0.63 0.95 0.81 0.42 0.88 0.33** – 100 0.70 0.54 0.63 0.53 0.63 0.63 0.81 0.56 0.49 0.69 0.61 0.80 0.50 0.63 0.79 Dry – 50 0.88 0.60 0.86 0.58 0.58 0.87 0.83 0.49 0.60 0.62 0.92 0.77 0.35 0.48 0.94 0 – 0.92 0.89 0.45 0.44 0.86 0.48 0.44 0.58 0.99 0.50 0.56 0.57 0.37** 0.37** 0.39** – North Sound 200 0.69 0.99 0.69 0.66 0.90 0.85 0.60 0.61 0.73 0.65 0.95 0.73 0.36 0.40 0.42 – 100 0.90 0.60 0.63 0.56 0.89 0.84 0.56 0.64 0.78 0.63 0.92 0.81 0.44 0.48 0.46 Wet – 50 0.78 0.47 0.89 0.57 0.54 0.78 0.45 0.59 0.74 0.52 0.87 0.78 0.42 0.47 0.53 – 0 0.48 0.51 0.53 0.51 0.47 0.47 0.60 0.49 0.78 0.31* 0.31* 0.31* 0.25* 0.36** 0.33** 0

200 200 100 100 100 200 50 200 50 50 50 100 0 0 0

Dry Dry Wet Wet

South Bimini South North Sound North Table 6. Spearman’s rank correlations and simplified Morisita’s index of similarity comparing community composition between two habitats in different locations different in habitats two between composition community comparing similarity of index Morisita’s simplified and correlations rank Spearman’s 6. Table rank Spearman’s locations. two between Bimini at identified families 64 all of (IRI) importance the in (overlap) similarity the identify indices These seasons. and Significant 0.01. < P ** 0.05, < P * stated: unless 0.001) < (P significant were rank Spearman’s All above. Morisita’s and line diagonal the below given are values index (> 0.60) are highlighted values in of bold. Morisita’s Habitat codes: 0, mangroves; 50, seagrass within 50 m of mangrove fringe; 100, seagrass 50–100 m; 200, seagrass 100–200 m from mangrove fringe. For example, overlap between mangrove communities in rank). the and 0.33 (Spearman’s North season is: 0.63 (Morisita’s) Sound and South Bimini during the wet 546 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Figure 3. Spatial and temporal variation in the number of species caught (A, B) and biomass density (C, D) at different distances from shore in the North Sound (A, C) and off South Bimini (B, D). Solid line and empty bars represent dry season (November–April) and dashed line and solid bars represent the wet season (May–October). * indicates significant difference with distance from shore and † indicates significant seasonal difference (Kruskal-Wallis tests: P < 0.05). Man- groves are 0 m, seagrasses 0–50 m, 50–100 m, and 100–200 m from mangrove fringe.

ing the wet season in the North Sound (see Table 2). However, even with silversides removed, there was still a significant increase in the median catch biomass within 50 m of the mangroves in the North Sound (Mann Whitney U test: P = 0.015), suggesting an overall increase in catch at this distance during the wet season. Community Size Structure.—Most organisms captured in the North Sound and off South Bimini were < 15 cm in size L( T for fish and shrimp, CW for crabs; Fig.

4). Few organisms in the North Sound exceeded 40 cm, while fish up to 60 cmL T were regularly captured over seagrass off South Bimini (Fig. 4B). Significant spatial and temporal variations in community size structure were observed in both the North Sound and South Bimini. Spatial comparison of median organism size revealed that mangrove communities off South Bimini were comprised of significantly larger -or ganisms during the dry season, and significantly smaller organisms during the wet season than catches over seagrass (K-S tests: P < 0.001; Fig. 4B). This was probably largely due to the high abundance of silversides during the wet season (see Tables 1 and 2). Within the North Sound, catches over seagrass within 50 m of the mangrove fringe were comprised of significantly larger organisms during the wet season (K-S test: P < 0.001; Fig. 4A). During the dry season, faunal communities within in the mangroves and farthest from shore (100–200 m) were comprised of significantly smaller organisms (K-S tests: P < 0.001; Fig. 4A), with larger organisms found within 100 m of the mangrove fringe. Significant seasonal variations in fish and invertebrate sizes were observed in both North Sound and South Bimini mangrove and seagrass communities. Mangrove faunal communities consisted of significantly larger organisms during the dry season in the North Sound and South Bimini, and this trend was also observed during the NEWMAN ET AL.: VARIATION IN MANGROVE AND SEAGRASS MACROFAUNAL COMMUNITIES 547

Figure 4. Spatial and temporal variation in organism sizes caught in (A) North Sound and (B) South Bimini, Bahamas. Sizes are total length in fish and shrimp, carapace width in crabs, and dorsal disc with in rays, all measured in centimeters. Location codes: NS, North Sound; SB, South Bimini. Habitat codes: 0, mangroves; 50, seagrass within 50 m of mangrove fringe; 100, seagrass 50–100 m; 200, seagrass 100–200 m from mangrove fringe. Season codes: D, dry season (November–April); W, wet season (May–October). + symbol within each box plot indicate mean organism size, with median notch also indicated. + symbols outside box plots indicate outliers. dry season within 50 m of the mangroves in the North Sound (K-S tests: P < 0.001). Seagrass catches during the wet season consisted of significantly larger organisms at all distances off South Bimini, and over 50 m from shore in the North Sound in comparison to catches during the dry season (K-S tests: P < 0.001).

Discussion

Sampling of mangrove and seagrass faunal communities off Bimini resulted in the identification of 128 species of teleost fish, 40 species of crustacean, six species of elasmobranch, and one species of cephalopod. Block nets enabled the accurate quantification of mangrove communities, whereas seine nets and bottom trawls complemented each other to provide a comprehensive description of the seagrass faunal communities off Bimini. The large sample size enabled comparisons of spatial and temporal variations in mangrove and seagrass communities, species diversity, and catch densities in high and low productivity areas. 548 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Index of relative importance values were amenable to catch data and facilitated spatial and temporal comparison of faunal communities. They also had the strength of using density, abundance, and weight as well as occurrence values, rather than having inherent bias from using just one index—a shortcoming of many qualitative studies (Sheridan and Hays, 2003). Mangrove and seagrass habitats differed significantly in terms of community composition, species diversity, and catch biomass. Mangroves typically supported significantly more diverse faunal communities and significantly greater catch biomasses than seagrass beds. Sampling in the present study was restricted to the daytime, but preliminary analysis of nocturnal catches off Bimini indicates a decline in mangrove community abundance, diversity, and biomass corresponding with an increase in seagrass communities (Newman, 2003). This supports earlier studies (Rooker and Dennis, 1991; Griffiths, 2001), although contrasting trends have also been reported (Thayer et al., 1987; Lin and Shao, 1999). Mangroves appeared to influence adjacent seagrass habitats in Bimini with a sharing of organisms and an extension of nursery habitat, a trend that has been previously reported for mangrove communities in Guadeloupe (Baelde, 1990), but has rarely been supported elsewhere (Sheridan and Hays, 2003). North Sound communities were significantly less diverse than those off South -Bi mini, and although cluster analysis revealed distinct mangrove and seagrass communities, there was a greater degree of homogeneity in the North Sound. North Sound communities demonstrated greater seasonal variation, with little spatial variation in community composition and catch diversity in either season. Generally, catch abundances of all taxa were greatest in the wet season off South Bimini, but the opposite was observed in North Sound, with catches highest in the dry season. Temperature is one of the most important physico-chemical factors influencing fish distributions (Moyle and Cech, 1988). The upper thermal limit of fish is approximately 38° C (Brett, 1979), with high temperatures also reducing dissolved oxygen. Temperature reached a near-lethal maximum of 36 °C during the wet season in the North Sound, and probably influenced community diversity and biomass. In addition, decreased mangrove catch biomass combined with the increase in catch biomass over adjacent seagrass beds within 100 m of the mangroves during the wet season suggests that many species migrated out of the mangroves during the wet season to avoid these near-lethal temperatures, although further study is required to confirm this. Fau- nal communities have been shown to be significantly lower in species diversity and abundance in thermally impacted seagrass beds in Florida (McLaughlin et al., 1983). During the dry season when temperatures were lower, catch diversity and biomasses in the North Sound indicated a greater association with the mangroves, similar to observations off South Bimini. This is consistent with the hypothesis that temperature limits diversity and catch biomass in the North Sound. Maximum temperatures (36 °C) and salinities (49) in the North Sound exceeded the tolerances for turtle grass, Thalassia testudinum Banks and Soland. ex Koenig, probably due to low tidal flow, resulting in lower standing crop than off South -Bi mini. Seagrass beds were significantly more widespread and dense in the more favorable environment off South Bimini and contained a greater diversity and biomass of macroalgae (Newman, unpubl. data). It has been demonstrated that there are more species in vegetated habitats than over bare sand (e.g., Lewis, 1984) and species richness, abundance, and biomass have all been correlated with increasing plant biomass (Heck and Wetstone, 1977; Stoner, 1980; NEWMAN ET AL.: VARIATION IN MANGROVE AND SEAGRASS MACROFAUNAL COMMUNITIES 549

Lewis and Stoner, 1983; De Troch et al., 1996). The North Sound was typified by a greater abundance of swimming crabs and shrimps than South Bimini habitats, probably due to extreme temperatures and salinities increasing seagrass blade turnover, resulting in more food for invertebrate detritus feeders. Therefore, greater primary productivity (as a result of the more favorable environmental conditions) was probably the principal reason behind the greater species diversity and catch biomass off South Bimini, with the high temperatures in the North Sound limiting primary productivity and thus indirectly reducing species diversity. Higher abundances of fish and invertebrates within the mangroves off South Bimini are probably due to the structural complexity and high organic content in the sediments of mangrove habitats. Faunal communities off Bimini were separated into distinct mangrove and seagrass assemblages as has been reported in Australian and West Indian waters (Rob- ertson and Duke, 1987; Louis et al., 1995). Mangrove communities off South Bimini are comparable to other mangrove habitats in the Caribbean region, but those in the North Sound were atypical, with significantly lower species diversity and catch biomass. More species were identified in the mangroves and seagrass beds in the more salubrious environment off South Bimini, despite a greater sampling effort in the North Sound. Forty-five species of fish (71 species total including invertebrates) were identified in the mangroves off South Bimini, in comparison to only 19 species of fish (30 species total) in the North Sound mangroves. The number of species caught in the mangroves in the North Sound is one of the lowest values reported in the literature and both mangrove communities have fewer species than mangroves in nearby Florida (87 species: Thayer et al., 1987; 76 species: Ley et al., 1999). Catch biomass within the mangroves was significantly lower in the North Sound than off South Bimini, with the mean catch biomass lowest in the wet season (2.5 ± 0.8 g m–2, mean ± SE, n = 7; area sampled: 3058 m2). Catch biomasses in the mangroves off South Bimini, even at their peak during the dry season (10.3 g m–2), were lower than values reported for the mangroves in nearby Florida (15.0 g m–2; Thayer et al., 1987) but greater than values reported in Australia (8.2 g m–2; Blaber et al., 1985). Species counts over seagrass beds in both areas (North Sound: 71 species of teleosts, 100 species total; South Bimini: 106 species of teleosts, 153 species total) were greater than values reported in the West Indies (61 species: Baelde, 1990), US Virgin Islands (34 species: Robblee and Zieman, 1984) and Puerto Rico (38 species: Rooker and Dennis, 1991) and were comparable to Florida seagrass communities (108 species: Tremain and Adams, 1995). However, this may be in part due to the increased sampling effort in this study, resulting in identification of more rare species in the environment. Mangroves provide shelter for juveniles and adults off South Bimini, but shallow water depth and unfavorable conditions may preclude larger organisms from the mangroves in the North Sound. Larger organisms, of comparable size to juvenile lemon sharks, were caught in greatest numbers within 100 m of the mangrove fringe. Catch data suggest that many large predatory fish such as barracuda and snappers may orientate to the mangroves, with catch rates greatest in and near the mangrove fringe. Juvenile lemon sharks demonstrate a positive correlation with the shoreline (Morrissey and Gruber, 1993), usually swimming within 50 m of the mangroves, probably due to high prey abundance and predator avoidance. Similar behavior has been demonstrated in tiger sharks, Galeocerdo cuvier (Péron and Lesueur in Lesueur, 550 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

1822), which were observed to spend more time over seagrass habitats with abundant prey (Heithaus et al., 2002). The tendency of juvenile lemon sharks and similar sized predators to spend most of their time within this area is probably due to the high abundance, diversity, and biomass of prey in the mangroves and over nearby seagrass. The patchy distribution of prey within the nursery indicates that predators probably increase their foraging efficiency by spending more time within this area. Mangroves supported a wide range of species at Bimini of which most were present as both juveniles and adults. The importance of mangroves in this study was not only demonstrated by an increase in catch diversity and biomass in comparison to seagrass sampling, but also by an extension of faunal communities onto adjacent seagrass habitats. The South Bimini environment was more comparable to other areas in the Caribbean; therefore the trends observed off South Bimini probably occur elsewhere in this region. The lack of significant seasonal variations in community composition, species diversity, and catch biomass off South Bimini indicate the relative stability of the environment in comparison to the North Sound. K-selective organisms, such as sharks, should do better in areas where prey communities are less impacted by seasonal variations (Margalef, 1968). A more stable environment should promote a greater degree of selective feeding resulting in greater growth rates in K- selective organisms. Although South Bimini has a more favorable environment with more diverse and abundant food, the North Sound supports the largest population of juvenile lemon sharks on the entire western edge of the Bahamas (Gruber, pers comm.). This suggests that a high degree of shelter may be an important selection criterion for nurseries by adult sharks to reduce potential predation rates, while juveniles may seek the mangroves due to the high abundance of prey. The high diversity and abundance of juveniles of many species in the mangroves means continual destruction of these habitats will seriously impact the diversity and health of nearby waters. It is therefore important to not underestimate the essential role that mangroves play in the life cycle for many species in this region, and their protection in the Bahamas should be an issue of immediate concern.

Acknowledgments

This study was financed with the aid of a studentship from the University of Plymouth as well as grants from PADI project AWARE. This work was supported in part by the Florida Department of Education (Grant 874-97030-00001), the Bimini Biological Field Station and its supporters, the Hoover Foundation, B. Newman, N. and E. Phillips, T. and S. Daniels, and T. and T. Fujino. Many thanks to the staff and volunteers of the Bimini Biological Field Sta- tion with special thanks to K. Parsons, D. McElroy, and C. Hofmann for their assistance in the field. This work was carried out under S. Gruber’s Bahamian research permit (MAF/LIA/22), and we would like to thank the Commonwealth of the Bahamas, Department of Fisheries (M. Braynen, Director) for issuing permits to allow us work in the controlled waters of the Com- monwealth of the Bahamas.

Literature Cited

Arrivillaga, A. and D. M. Baltz. 1999. Comparison of fishes and macroinvertebrates on seagrass and bare-sand sites on Guatemala’s Atlantic coast. Bull. Mar. Sci. 652: 301–319. Austin, H. M. 1971. A survey of the ichthyofauna of the mangroves of western Puerto Rico during December, 1967-August, 1968. Caribb. J. Sci. 11: 27–39. NEWMAN ET AL.: VARIATION IN MANGROVE AND SEAGRASS MACROFAUNAL COMMUNITIES 551

Baelde, P. 1990. Differences in the structures of fish assemblages in Thalassia testudinum beds in Guadeloupe, French West Indies, and their ecological significance. Mar. Biol. 105: 163– 173. Bell, J. D., D. A. Pollard, J. J. Burchmore, B. C. Pease, and M. J. Middleton. 1984. Structure of fish community in a temperate tidal mangrove creek in Botany Bay, New South Wales. Aust. J. Mar. Freshw. Res. 351: 33–46. Blaber, S. J. M., J. M. Young, and M. C. Dunning. 1985. Community structure and zoogeographic affinities of the coastal fishes of the Dampier region of North-western Australia. Aust. J. Mar. Freshw. Res. 36: 257–266. Böhlke, J. E. and C. C. G. Chaplin. 1968. Fishes of the Bahamas and adjacent tropical waters. Academy of Natural Sciences of Philapdelphia. 771 p. Brett, J. R. 1979. Energetic factors and growth. Pages 599–675 in W. S. Hoar, D. J. Randall, and J. R. Brett, eds. Fish Physiology. Academic Press, New York. De Troch, M., J. Mees, I. Papadopoulos, and E. O. Wakwabi. 1996. Fish communities in a tropical bay Gazi Bay, Kenya, Seagrass beds vs. unvegetated areas. Neth. J. Zool. 463-4: 236– 252. Dorenborsch, M, M. C. van Riel, I. Nagelkerken, and G. van der Velde. 2004. The relationship of reef fish densities to the proximity of mangrove and seagrass nurseries. Estuar. Coast. Shelf Sci. 60: 37–48. Griffiths, S. P. 2001. Diel variation in the seagrass ichthyofaunas of three intermittently open estuaries in south-eastern Australia, implications for improving fish diversity assessments. Fish. Manage. Ecol. 8: 123–140. Gruber, S. H., J. R. de Marignac, and J. M. Hoenig. 2001. Survival of juvenile lemon sharks Negaprion brevirostris at Bimini, Bahamas, estimated by mark-depletion experiments. Trans. Am. Fish. Soc. 130: 376–384. Heck, K. L., Jr. and G. S. Wetstone. 1977. Habitat complexity and invertebrate species richness and abundance in tropical seagrass meadows. J. Biogeogr. 4: 135–142. Heithaus, M. R., L. M. Dill, G. J. Marshall, and B. Buhleier. 2002. Habitat use and foraging behaviour of tiger sharks Galeocerdo cuvier in a seagrass ecosystem. Mar. Biol. 1402: 237– 248. Horn, H. S. 1966. Measurement of “overlap” in comparative ecological studies. Am. Nat. 100: 419–424. Krebs, C. J. 1999. Ecological Methodology 2nd ed. Benjamin/Cummings California. 620 p. Layman, C. A. and B. R. Silliman. 2002. Preliminary survey and diet analysis of juvenile fishes of an estuarine creek on Andros Island, Bahamas. Bull. Mar. Sci. 701: 199–210. Lewis, F. G. 1984. Distribution of macrobenthic crustaceans associated with Thalassia, Halod- ule and bare sand substrata. Mar. Ecol. Prog. Ser. 19: 101–113. ______and A. W. Stoner. 1983. Distribution of macrofauna within seagrass beds, An explanation for patterns of abundance. Bull. Mar. Sci. 332: 296–304. Ley, J. A., C. C. McIvor, and C. L. Montague. 1999. Fishes in mangrove prop-root habitats of North-eastern Florida Bay, Distinct Assemblages across an estuarine gradient. Estuar. Coast. Shelf Sci. 48: 701–723. Lin, H. J. and K. T. Shao 1999. Seasonal and diel changes in a subtropical mangrove fifish sh assem-assemblage. Bull. Mar. Sci. 65: 775–794. Louis, M., C. Bouchon, Y. Bouchon-Navaro, Y. -S. Wong, and N. F. Y. Tam. 1995. Spatial and temporal variations of mangrove fish assemblages in Martinique French West Indies. Hyd- robiologia 2951: 275–284. Margalef, R. 1968. Perspectives in ecological theory. University of Chicago Press, Chicago. 111 p. McLauglin, P. A., S. A. Treat, A. Thorhaug, and R. Lematril. 1983. Restored seagrass and its animal communities. Environ. Conserv. 10: 247–252. Morrissey, J. F. and S. H. Gruber. 1993. Habitat selection by juvenile lemon sharks, Negaprion brevirostris. Environ. Biol. Fish. 38: 311–319. 552 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Moyle, P. B. and J. J. Cech. 1988. Fishes, an introduction to ichthyology. Prentice Hall, Engle- wood Cliffs. 559 p. Mumby, P. J., A. J. Edwards, J. E. Arias-Gonzalez, K. C. Lindeman, P. G. Blackwell, A. Gall, M. I. Gorczynska, A. R. Harborne, C. L. Pescod, H. Renken, C. C. C. Wabnitz, and G. Llewellyn. 2004. Mangroves enhance the biomass of coral reef fish communities in the Caribbean. Nature 427: 533–536. Nagelkerken, I., M. Dorenbosch, W. C. E. P. Verberk, E. Cocheret de la Morinière, and G. van der Velde. 2000a. Day-night shifts of fishes between shallow-water biotopes of a Carib- bean bay, with emphasis on the nocturnal feeding of Haemulidae and Lutjanidae. Mar. Ecol. Prog. Ser. 194: 55–64. ______, G. Van der Velde, M. W. Gorissen, G. J. Meijer, T. van’t Hof, and C. den Hartog. 2000b. Importance of Mangroves, Seagrass beds and the shallow coral reef as a nursery for important coral reef fishes, using a visual census technique. Estuar. Coast. Shelf Sci. 51: 31–44. Newell, N. D. and J. Imbrie. 1955. Biogeological reconnaissance in the Bimini area, Great Ba- hama Bank. Trans. N.Y. Acad. Sci. 2: 3–14. ______,______, E. G. Purdy, and D. L. Thurber. 1959. Organism communities and bottom facies, Great Bahama Bank. Bull. Am. Mus. Nat. Hist. 1174: 183–228. Newman, S. P. 2003. Spatial and temporal variation in diet and prey preference of nursery- bound juvenile lemon sharks, Negaprion brevirostris, at Bimini, Bahamas. University of Plymouth, England. 268 p. ______and S. H. Gruber. 2002. Comparison of mangrove and seagrass fish and macroinvertebrate communities in Bimini, Bahamas. Bah. J. Sci. 9: 19–27. Odum, W. E. and E. J. Heald. 1972. Trophic analyses of an estuarine mangrove community. Bull. Mar. Sci. 22: 671–738. ______, C. C. McIvor, and T. J. Smith III. 1982. The ecology of the mangroves of south Florida, a community profile. Washington DC, US Fish and Wildlife Service, Office of Bio- logical Sciences FWS/OBS-81-24. 144 p. Pinkas, L. M., S. Oliphant, and I. L. K. Iverson. 1971. Food habits of albacore, bluefin tuna, and bonito in California waters. California Dept. Fish Game, Fish. Bull. 152: 1–105. Robblee, M. B. and J. C. Zieman. 1984. Diel variation in the fish fauna of a tropical seagrass feeding ground. Bull. Mar. Sci. 343: 335–345. Roberts, H. H., L. J. Rouse, Jr., and N. D. Walker. 1983. Evolution of cold water stress conditions in high latitude reef systems, Florida reef track and the Bahama Banks. Caribb. J. Sci. 191: 55–60. Robertson, A. I. and N. C. Duke. 1987 Mangroves as nursery sites, comparisons of the abundance and species composition of fish and crustaceans in mangroves and other near shore habitats in tropical Australia. Mar. Biol. 96: 193–205. Rooker, J. R. and G. D. Dennis. 1991. Diel, Lunar and seasonal changes in a mangrove fish assemblage off southwestern Puerto Rico. Bull. Mar. Sci. 493: 684–698. Sheridan, P. and C. Hays. 2003. Are mangroves nursery habitat for transient fishes and decapods? Wetlands 23: 449–458. Stoner, A. W. 1980. The role of seagrass biomass in the organisation of benthic macrofaunal assemblages. Bull. Mar. Sci. 303: 537–551. Thayer G. W., D. R. Colby, and W. F. Hettler, Jr. 1987. Utilization of the red mangrove prop root habitat by fishes in south Florida. Mar. Ecol. Prog. Ser. 35: 25–38. Tremain, D. M. and D. H. Adams. 1995. Seasonal variations in species diversity, abundance and composition of fish communities in the Northern Indian River Lagoon, Florida. Bull. Mar Sci. 571: 171–192. Turekian, K. K. 1957. Salinity variations in sea water in the vicinity of Bimini, Bahamas, British West Indies. Am. Mus. Nov. 1822: 1–12. NEWMAN ET AL.: VARIATION IN MANGROVE AND SEAGRASS MACROFAUNAL COMMUNITIES 553

Voss, G. L. and N. A. Voss. 1960. An ecological survey of the marine invertebrates of Bimini, Bahamas, with a consideration of the zoogeographical relationships. Bull. Mar. Sci. Gulf Caribb. 101: 96–116. Weinstein, M. P. and K. L. Heck. 1979. Ichthyofauna of seagrass meadows along the Caribbean coast of Panama and in the Gulf of Mexico, composition, structure and community ecology. Mar. Biol. 50: 97–107. Zaret, T. M. and A. S. Rand. 1971. Competition in tropical stream fishes, support for the competitive exclusion principal. Ecology 522: 336–342. Zieman, J. C. 1982. The ecology of seagrasses of south Florida, a community profile. U.S. Fish and Wildlife Service Biol. Serv. Prog. FWS/OBS-82/85. 124. 26 p.

Addresses: (S.P.N., R.D.H.) School of Biological Sciences, University of Plymouth, Plymouth, Devon, PL4 8AA, United Kingdom. (S.P.N., S.H.G.) Bimini Biological Field Station, Bimini, Bahamas. (S.H.G.) Rosenstiel School of Marine and Atmospheric Sciences, University of Mi- ami, Miami, Florida 33149-1098. Current Address: (S.P.N.) School for Field Studies, Center for Marine Resource Studies, South Caicos, Turks and Caicos Islands, British West Indies. E- mail: . BULLETIN OF MARINE SCIENCE, 80(3): 555–565, 2007

NOTE

Juvenile blue crab abundances in natural and man-made tidal channels in mangrove habitat, Tampa Bay, Florida (USA)

Lauren A. Yeager, Justin M. Krebs, Carole C. McIvor, and Adam B. Brame

The blue crab Callinectes( sapidus M. J. Rathbun, 1896) is an ecologically and commercially important species common in the Gulf of Mexico and Western Atlantic Ocean (Virnstein, 1977; Hines et al., 1990; Guillory et al., 1998). The life cycle of the blue crab is considered to be classically estuarine-dependent with females mating within the estuary then migrating offshore to spawn. Larvae are transported back into estuarine habitats where they settle and grow to adult sizes (Tagatz, 1968). Studies of habitat utilization by juvenile blue crabs have focused largely on seagrass, saltmarsh, and unvegetated habitats. Previous studies indicate that seagrass and saltmarsh habitats are important to juvenile blue crabs: researchers report higher juvenile densities, decreased predation, and/or faster growth rates compared to unvegetated habitats (Orth and van Montfrans, 1987, 2002; Thomas et al., 1990; Williams et al., 1990; Perkins-Visser et al., 1996; Rozas and Minello, 1998; Rozas and Zimmerman, 2000). While the use of saltmarsh and seagrass by juvenile blue crabs is relatively well understood, knowledge of mangrove wetlands as habitat for blue crabs is limited. In Rookery Bay, Florida, blue crab densities were lowest in intertidal mangrove habitat, and higher in seagrass and open water (Sheridan, 1992). However, a study comparing mangrove and saltmarsh habitats in Caminada Bay, Louisiana found that blue crabs were more closely associated with mangrove habitat and suggested that this relationship was due to the greater structural complexity offered by mangroves (Caudill, 2005). However, juvenile densities were not reported in these two studies. Coastal wetlands, including mangrove habitats, are being increasingly altered or destroyed by human development (Dahl, 1990; Valiela et al., 2001). An area of increasing interest to scientists and habitat managers is the ecological equivalency of altered and restored habitats relative to natural habitats. Contrasting results on the functioning of non-natural wetlands have been presented from studies of blue crab habitat use (Minello, 2000; Peterson et al., 2000; Jivoff and Able, 2003), suggesting that the effects of habitat alteration (both the type and degree) on wetland communities, and specifically for blue crabs, have yet to be fully explained. Within Tampa Bay, Florida, common forms of wetland alteration have been ditching for mosquito control and stormwater drainage. Mosquito-control ditches were constructed in the 1960s and 1970s to inundate poorly flushed coastal wetlands and allow access for fishes that consume mosquito larvae. Stormwater-drainage ditches were constructed to convey rainwater runoff from residential areas to Tampa Bay and thereby prevent urban flooding.D irect comparison of fauna in altered vs natural wetlands provides insight into how human activities affect mangrove wetland function and is particularly important for assessing restoration need and/or efficacy. The objective of this study was to determine if juvenile blue crab abundance differed between naturally formed tidal creeks and man-made ditches in mangrove wetlands.

To address this objective, we first estimated the natural spatial variability of crab abundance among three naturally formed tidal creeks. This allowed establishment of a baseline by which comparisons could be made with man-made ditches. Second, we compared crab abundance between tidal creeks and man-made mosquito-control ditches within two of the wetland regions. Third, in addition to alterations caused by mosquito ditching, one of our sample regions (i.e., Mobbly Bayou) had been altered for stormwater drainage, allowing us to compare tidal creeks with two types of man- made habitat within a wetland. The null hypothesis for each objective was one of no difference in juvenile blue crab abundance between sample regions and/or habitat type (i.e., naturally formed and man-made tidal channels).

Materials and Methods

Sample Regions.—Sampling was conducted in State and County preserves within three wetland regions along a north-south transect within the Tampa Bay estuary, Florida (USA) (Fig. 1). Sample sites in the naturally formed tidal creek in the Terra Ceia region (lower bay) were located closest to the mouth of the estuary. Weedon Island (mid bay) is located along the Tampa Bay shoreline near the middle of the estuary and is the location of extensive gridded mosquito ditching. Feather Sound (mid bay), approximately 8 km north of Weedon Island, contains a tidal creek that empties directly into Tampa Bay. Mobbly Bayou (upper bay) is located at the northern-most extent of Tampa Bay. The tidal creek at Mobbly Bayou runs through the center of the County preserve, while three stormwater-drainage ditches were constructed parallel to the creek at its upstream, central, and downstream reaches. Mos- quito-control ditches are located in the central portion of the Mobbly Bayou wetland approximately 200 m northeast of the creek, as well as in the fringing mangroves along the shoreline of Tampa Bay. Unlike mosquito-control ditches, stormwater-drainage ditches are less common features in the wetland landscape and differ from mosquito-control ditches in their linearity and their conductance of larger water volumes. These differences are related to their intended purpose and have resulted in contrasting hydrology, channel widths, water depths, and bottom substrates (Krebs, unpubl. data). Sample Design.—For each sample region (e.g., upper bay), random points were selected within each habitat type. Fixed sites were installed as close as possible to these randomly- generated points. Each of the 49 fixed sites was sampled twice from December 2003 to May 2004 (i.e., once each season during recruitment). To estimate the natural spatial variability of juvenile blue crab abundances in naturally formed channels, we sampled tidal creeks in the upper bay (n = 9 fixed sites), mid bay (n = 6), and lower bay (n = 6) regions. To compare crab abundance in natural and man-made habitat types, we sampled sites within mosquito- ditched wetlands. This comparison of natural and man-made habitat was limited to mosquito-control ditches in upper bay (n = 9 fixed sites) and mid bay (n = 10) regions because of the absence of mosquito-control ditches in the lower bay sample region. Finally, abundance of juvenile blue crabs was compared among sites in tidal creeks and two types of man-made habitat: mosquito-control ditches and stormwater-drainage ditches (n = 9 fixed sites) in the upper bay region. Crabs were sampled by isolating a 9.1-m-long section of the creek or ditch using two block nets (3-mm mesh) and seining through the site using a center bag haul seine (1.2 m deep with 3-mm mesh) stretched across the channel from bank to bank. Size of block nets (4.6 m, 6.1 m, 9.1 m, 12.2 m, or 15.2 m) and seines used depended on channel width. Blue crabs were enumerated and carapace width (CW) was measured to the nearest 1-mm for the first 20 individuals from each sample. Statistical Analysis.—For analysis, juvenile crabs were defined as those < 80 mm CW. Size frequency distributions from this study (Yeager et al., unpubl. data) and previous studies (Fitz and Wiegert, 1991; Hsueh, 1992; McClintock et al., 1993) suggest that this is a conserva- NOTES 557

Figure 1. Location of sample sites within lower (Terra Ceia), mid (Weedon Island), and upper bay (Mobbly Bayou) regions of Tampa Bay along the west coast of Florida (USA). tive estimate of maximum size for young-of-the-year (YOY) juveniles. For all analyses, abundance of juvenile crabs was measured as the number of crabs 100 m–2. To test for a statistical difference in the abundance of juvenile blue crabs among three naturally formed tidal creeks, we used a repeated measures analysis of variance (RMANOVA) with maximum likelihood estimation. Season (i.e., winter and spring) was the repeated variable. Because samples from fixed sites were not independent of one another, spatial autocorrelation resulting from repeated sampling of fixed sites was accounted for using a covariance structure that was first order heterogeneous autoregressive whereby autocorrelation diminishes through time. We then compared the abundance of juvenile blue crabs between tidal creeks and mosquito-control ditches in the upper and mid bay regions using a similar RMANOVA to test the null hypothesis of no difference between naturally formed and man-made habitat types within both regions. Finally, we compared juvenile crab abundance among tidal creeks, mosquito-control and stormwater-drainage ditches in the upper bay region using RMANOVA as above. Tukey multiple comparison tests were used to identify the source of the difference, when present. In all cases, data were log (x + 1) transformed prior to analysis. Data met the assumption of normality (Shapiro-Wilk; P > 0.05) with the exception of the comparison of tidal creeks and 558 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 mosquito-control ditches (w = 0.9877, P = 0.0217). However, given the sample size (n = 68 samples) and the robustness of ANOVA to departures from normality (Zar, 1999), we feel this use of a parametric test was justified. All analyses were performed using SAS version 9.1 (SAS Institute, 2003).

Results

Young blue crabs recruited to our study sites in Tampa Bay wetlands during winter and spring (December 2003–May 2004). Abundances of juvenile crabs in mangrove wetlands (natural and man-made) during the winter season (December–February) ranged from 0–222.2 crabs 100 m–2. The modal size of juvenile blue crabs during this time was 10–20 mm CW within all three sample regions. During the spring season (March–May) abundances ranged from 0–225.4 crabs 100 m–2. Modal size remained constant from winter until the end of spring at which time the modal size was observed to have increased to 30–50 mm CW. A concurrent decrease in the range of abundances to 0–34.2 crabs 100 m–2 during the summer (June–August) was observed with the increase in modal size. Mean abundance of juvenile crabs differed among tidal creeks in Tampa Bay during recruitment and early post-settlement (P = 0.028; Fig. 2), but only because of a significantly lower density of juvenile crabs in the mid bay during late spring. The greatest abundance was observed in the lower bay creek and was approximately three times that of the mid bay creek during early post-settlement (P = 0.026; Fig. 2). How- ever, mean abundance in the upper bay creek did not differ from that of creeks in the mid or lower bay (P = 0.608 and P = 0.101, respectively). Regionally, abundance of juvenile crabs was significantly greater at the upper bay wetland than the mid bay wetland (P = 0.0004), probably the result of high abundances in mosquito-control ditches in the upper bay. Juvenile blue crabs were significantly more abundant in mosquito-control ditches than in creeks, in only the upper bay region (P = 0.010; Fig. 2). There was no difference in blue crab abundance between the tidal creek and mosquito-control ditches at the mid bay wetland (P = 0.782). Mean abundance of juvenile blue crabs differed significantly among the natural and altered habitats in the upper bay wetland (P < 0.0001). Juvenile crab abundances were greatest in the mosquito-control ditches; moderate in the tidal creek, and lowest abundances in stormwater-drainage ditches (P < 0.0001–0.002; Fig. 2).

Discussion

Abundance of juvenile blue crabs differed among natural and man-made tidal channels though not consistently at each sampled wetland, suggesting that spatial location in the estuary may be equally as important in shaping habitat use patterns as habitat type (i.e., natural vs man-made). Mosquito-control ditches have, in many cases, been shown to have a detrimental effect on wetland-associated fauna H( arrington and Harrington, 1961; Gilmore et al., 1982; Brockmeyer et al., 1997), and as a result, are frequently viewed as inferior habitat for organisms of fishery importance. Yet the highest densities of juvenile blue crabs in three types of tidal channels at 49 sites in three Tampa Bay wetlands occurred in mosquito-control ditches in the upper bay. All mosquito-control ditches sampled at both mid and upper bay wetlands were open to tidal flushing and had consolidated substrates that allowed unimpeded seine NOTES 559 - Figure 2. Mean ± SE abundance of juvenile blue crabs 80 (< mm CW) in tidal creeks mosquito-controlbar), (white ditches (grey age ditches andbar), (black bar) instormwater-drain mangroveFL. wetlands, ResultsBay, Tampa of the repeated comparingmeasures creeks ANOVAs (A) in mid,the andlower, upper creeks(B) bay, and mosquito-control ditches in the mid and andupper creeks,(C) bay, mosquito-control ditches, and stormwater ditches in the upper bay areindicated the by letters each Bars bar. above with the same letter are not significantly different0.05. at P < 560 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

sampling. There are, however, mosquito-control ditches in Tampa Bay where neither of these conditions is true and mosquito-control ditches sampled during this study may not be representative of the range of such ditches around Tampa Bay. During the time since their construction, ditches that received minimal tidal influence have accumulated sediment and were colonized by mangroves. Haul seining constraints limited sampling in ditches that possessed deep mud or dense mangrove structure (i.e., prop roots, pneumatophores). Our results clearly can not be extended to these mosquito-control ditches. Nonetheless, we found that juvenile blue crab abundances in tidally flushed mosquito-control ditches with relatively firm substrates did not differ from that of naturally formed tidal creeks in one of two bay sample regions. The fact that the abundance of juvenile blue crabs in mosquito-control ditches differed between the mid and upper bay wetlands is cause for further consideration. The widths (mid bay = 4.5 ± 0.9 m; upper bay = 5.2 ± 0.5 m) and water depths (mid bay = 0.44 ± 0.02 m; upper bay = 0.35 ± 0.02 m) of mosquito-control ditches in these two wetlands were similar. However, at a landscape scale, the wetlands are likely positioned differently with regard to blue crab recruits transported up-estuary in bay waters, assuming passive transport of larvae (McConaugha, 1988). Circulation at the mouth of the bay is such that the lower bay creek would presumably receive a greater supply of larvae as they enter the bay from the Gulf of Mexico. The net circulation in Old Tampa Bay (i.e., upper bay in our study) is north along the eastern shoreline, and south along the western shoreline to the mid bay region (Weisberg and Zheng, 2006). This pattern is consistent with more juveniles at the lower bay than the upper bay wetland and mid bay wetland on the western shore of Old Tampa Bay, based on order of arrival of bay waters. Of the 1715 juvenile crabs collected in this study, 279 could be considered early juveniles (J3–J5, sensu Pile et al., 1996). Most of the early juveniles were found in the lower bay region (57.3%) followed by upper bay region (29.0%), further suggesting that bay circulation and transport of crab larvae may play an important role in determining patterns of habitat use. In contrast, Buchanan and Stoner (1988) found the highest proportion of small crabs at the head of the Laguna Joyuda estuary, farthest from the inlet indicating transport of juvenile crabs towards the uppermost reaches of the estuary. A similar argument can be made at a smaller scale regarding the orientation of mosquito-control ditches in a given wetland. At the wetland in the upper bay, six of nine sampled mosquito-control ditches are located directly along the bay perimeter and all receive consistent tidal flushing (Krebs, unpubl. data). In contrast, most mosquito-control ditches at sample sites in the mid bay are located behind a series of overwash mangrove islands and receive less direct tidal input. Only three of ten mosquito-control ditches at the mid bay wetland are directly open to tidal input; the others receive tidal waters via intermediate mosquito-control ditches. Patterns of early juvenile habitat are consistent with the hypothesis that spatial configuration of habitats within a wetland may affect access to recruits. In the upper bay region, mosquito-control ditches received more early juveniles than both other habitats (76 of 82 total early juveniles at upper bay). In contrast, only five of the 27 observed early juveniles in the mid bay region occurred in mosquito-control ditches. Differences in the abundance of juvenile blue crabs in the same habitat type (i.e., mosquito-control ditches) at different spatial locations suggest that location relative to estuarine circulation must be considered as a factor affecting larval dispersal and consequently patterns of habitat use. NOTES 561

Abundance of recruits among habitats can also be influenced by prey availability (Seitz et al., 2005) and mortality from predation (Etherington et al., 2003). Seitz et al. (2005) suggested that unvegetated subtidal habitats adjacent to salt marshes may rival seagrasses as habitat for juvenile blue crabs because of higher prey abundances in the unvegetated areas. Although the relative quality of food resources (i.e., composition, caloric value) is not known for our sample sites, differences in sediment grain size and accumulation of detritus among tidal creeks, mosquito-control ditches, and stormwater-drainage ditches suggest differences in food availability [(i. e., infaunal community (Snelgrove, 1999) and detritus (Krebs, unpubl. data)] that may affect habitat use by juvenile blue crabs. Whether predation mortality differs among the habitat types examined in our study is unknown. However, several species which were collected in considerable abundance in our study sites (Krebs et al., 2007) have been implicated as potential predators on juvenile blue crabs (Guil- lory et al., 2001). Of these predators, fewer sheepshead Archosargus probatocephalus (Walbaum, 1792), pinfish Lagodon rhomboids (Linnaeus, 1766), spot Leiostomus xanthurus Lacépède, 1802, and red drum Sciaenops ocellatus (Linnaeus, 1766) were collected in mosquito-control ditches than in tidal creeks and stormwater-drainage ditches (Krebs et al., 2007), consistent with a hypothesis of lower predation pressure in mosquito-control ditches. Regardless of the mechanism(s) underlying habitat use by juvenile blue crabs, comparison of densities among studies is important, but complicated by gear efficiency differences. Recent studies conducted in a variety of habitats report a wide range of densities for juvenile blue crabs (Table 1) and, predictably, researchers have reached differing conclusions. Using lift nets, Caudill (2005) observed that blue crabs were more closely associated with mangroves than saltmarshes in Caminada Bay, Loui- siana. Alternatively, studies utilizing drop samplers or suction samplers (Orth and van Montfrans, 1987; Thomas et al., 1990; Williams et al., 1990; Wilson et al., 1990; Sheridan, 1992) reported densities that were generally one order of magnitude greater in saltmarsh, seagrass, and openwater habitats than the densities observed along mangroves by Caudill (2005) and during our study. The lower observed densities in our study could be related to gear such as seines being less efficient at catching macroinvertebrates than drop samplers (Freeman et al., 1984). Some of the lowest reported densities of juvenile blue crabs are from an intertidal marsh at Sapelo Island, Georgia, where block nets were used to capture blue crabs exiting tidally drained creeks (Fitz and Wiegert, 1991). Clearly, comparative studies of habitat use em- ploying similar methods to sample mangrove, saltmarsh, and seagrass habitats are needed to better understand the relative value of mangrove wetlands in supporting juvenile blue crabs. The results of our study have important implications for management of mangrove wetlands and proposed restoration projects. The removal of stormwater-drainage ditches and restoration of more natural flow to the creek at Mobbly Bayou (upper bay) should have no negative impact on local blue crab numbers. Densities of juvenile blue crabs in stormwater-drainage ditches in the upper bay were consistently lowest compared to the other two channel types there. The features of these stormwater-drainage ditches (coarse, shell-hash substrates, relatively little detritus, relatively high densities of potential predators, relatively high current velocities) are apparently less favorable for juvenile blue crabs than conditions in tidal creeks or mosquito-control ditches. However, filling in mosquito-control ditches at that site may reduce the 562 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 Study Thomas et al., 1990 et al., 1990 Wilson et al., 1990 Williams 1991 Wiegert, Fitz and Thomas et al., 1990 Orth and van Montfrans, 1987 et al., 1990 Wilson Thomas et al., 1990 Orth and van Montfrans, 1987 et al., 1990 Wilson et al., 1990 Williams Sheridan, 1992 Caudill, 2005 et al., 2007, this study Yeager Caudill, 2005 et al., 1990 Wilson suction sampler block net suction sampler suction sampler Sampling gear drop sampler suction sampler drop sampler suction sampler suction sampler drop sampler suction sampler drop sampler lift net seine with block nets lift net suction sampler (mm) ≤ 80 < 40 < 50 < 40 < 40 < 40 < 30 < 80 < 80 < 70 ≤103 ≤103 ≤ 103 ≤ 103 < 102 < 102 Carapace width ) 2 70 80 60 70 28* 10–560 10–800 15–81* 33–45* Density 0.2–47* 0 – > 0.5 0 – < 100 130–2,210 280–5,060 < 300–9,000 < 200–3,650 (crabs/100 m Unvegetated Unvegetated Unvegetated Saltmarsh Saltmarsh Saltmarsh Saltmarsh SAV SAV SAV SAV Mangrove Mangrove Mangrove Mangrove/saltmarsh transition Macroalgae Table 1. Table Literature-reported estimates of juvenile blue crab density in various habitat types. Density values were aquatic vegetation. = submerged reported as monthly noted. * = Seasonal mean. SAV means unless otherwise Habitat NOTES 563 extent of an important juvenile habitat. In contrast, restoration of mosquito-control ditches at Weedon Island (mid bay) may have little effect on the juvenile blue crab numbers because densities in this habitat were consistently low. Differential habitat use by juvenile blue crabs is likely due to an interaction of habitat characteristics, spatial arrangement of habitat types relative to estuarine circulation patterns, and possibly to differences in survival among habitats. Our results emphasize the importance of considering both habitat characteristics and spatial location when determining relative value of natural and altered mangrove wetlands.

Acknowledgments

We thank the managers, K. Hill, L. Rives (Mobbly Bayou Wilderness Preserve), R. Runnels (Terra Ceia State Aquatic BufferP reserve), P. Leasure, and K. Thompson (Weedon IslandP reserve) for their support and cooperation throughout this project. We also thank D. Neilsen, N. Hansen, J. Sanford, N. Silverman, R. Krebs, R. Poyner, and G. Craig for their help with field sampling and J. Rehage, G. Garrison, and two anonymous reviewers for comments whose incorporation has greatly improved this manuscript. This project was funded by the U.S.G eo- logical Survey’s Tampa Bay Integrated Science Program.

Literature Cited

Brockmeyer, Jr., R. E., J. R. Rey, R. W. Virnstein, R. G. Gilmore, and L. Earnest. 1997. Reha- bilitation of impounded estuarine wetlands by hydrologic reconnection to the Indian River Lagoon, Florida (USA). Wetl. Ecol. Manage. 4: 93–109. Buchanan, B. A. and A. W. Stoner. 1988. Distributional patterns of blue crabs (Callinectes sp.) in a tropical estuarine lagoon. Estuaries 11: 231–239. Caudill, M. C. 2005. Nekton utilization of black mangrove (Avicennia germinans) and smooth cordgrass (Spartina alterniflora) sites in southwestern Caminada Bay, Louisiana. M.Sc. Thesis, Louisiana State University. 71 p. Dahl, T. E. 1990. Wetlands losses in the United States 1780’s to 1980’s. U.S. Department of the Interior, Fish and Wildlife Service, Washington, D.C. 21 p. Etherington, L. L., D. B. Eggleston, and W. T. Stockhausen. 2003. Partitioning loss rates of early juvenile blue crabs from seagrass habitats into mortality and emigration. Bull. Mar. Sci. 72: 371–391. Fitz, H. C. and R. G. Wiegert. 1991. Utilization of the intertidal zone of salt marsh by the blue crab Callinectes sapidus: density, return frequency, and feeding habits. Mar. Ecol. Prog. Ser. 76: 249–260. Freeman, B. J., H. S. Greening, and J. Douglas Oliver. 1984. Comparison of three methods for sampling fishes and macroinvertebrates in a vegetated freshwater wetland. J. Freshw. Ecol. 2: 603-609. Gilmore, R. G., D. W. Cooke, and C. J. Donohoe. 1982. A comparison of the fish populations and habitat in open and closed salt marsh impoundments in east-central Florida. Northeast Gulf Sci. 5: 25–37. Guillory, V., H. Perry, P. Steele, T. Wagner, P. Hammerschmidt, S. Heath, and C. Moss. 1998. The Gulf of Mexico blue crab fishery: historical trends, status, management, and recommendations. J. Shellfish Res. 17: 395–403. ______, H. Perry, P. Steele, T. Wagner, W. Keithly, B. Pellegrin, J. Petterson, T. Floyd, B. Buckson, L. Hartman, E. Holder, and C. Moss. 2001. The blue crab fishery of the Gulf of Mexico, United States: a regional management plan. Gulf States Marine Fisheries Commis- sion. 301 p. 564 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Harrington R. W. and E. S. Harrington. 1961. Food selection among fishes invading a high subtropical salt marsh: from onset of flooding through the progress of a mosquito brood. Ecology 42: 642–666. Hines, A. H., A. M. Haddon, and L. A. Wiechert. 1990. Guild structure and foraging impact of blue crabs and epibenthic fish in a subestuary of Chesapeake Bay. Mar. Ecol. Prog. Ser. 67: 105–126. Hsueh, P. 1992. A comparative study of the population dynamics, life history characteristics, and physiological ecology of Callinectes similis and C. sapidus in estuarine environments of the northern Gulf of Mexico. Ph.D. Diss., University of Alabama. 54 p. Jivoff, P. R and K. W. Able. 2003. Evaluating salt marsh restoration in Delaware Bay: The response of blue crabs, Callinectes sapidus, at former salt hay farms. Estuaries 26: 709–719. Krebs, J. M., A. B. Brame, and C. C. McIvor. 2007. Altered mangrove wetlands as habitat for estuarine nekton: are dredged channels and tidal creeks equivalent? Bull. Mar. Sci. 80: 839– 861. McClintock, J. B., K. R. Marion, J. Dindo, P. Hsueh, and R. A. Angus. 1993. Population studies of blue crabs in soft-bottom, unvegetated habitats of a subestuary in the northern Gulf of Mexico. J. Crust. Biol. 13: 551–563. McConaugha, J. R. 1988. Export and reinvasion of larvae as regulators of estuarine decapod populations. Am. Fish. Soc. Symp. 3: 90–103. Minello, T. J. 2000. Temporal development of salt marsh value for nekton and epifauna: utilization of dredged material marshes in Galveston Bay, Texas, USA. Wetl. Ecol. Manage. 8: 327–341. Orth, R. J. and J. van Montfrans. 1987. Utilization of a seagrass meadow and tidal marsh creek by blue crabs Callinectes sapidus. I. Seasonal and annual variations in abundance with emphasis on post-settlement juveniles. Mar. Ecol. Prog. Ser. 41: 283–294. ______and ______. 2002. Habitat quality and prey size as determinants of survival in post-larval and early juvenile instars of the blue crab Callinectes sapidus. Mar. Ecol. Prog. Ser. 231: 205–213. Perkins-Visser, E., T. G. Wolcott, and D. L. Wolcott. 1996. Nursery role of seagrass beds: enhanced growth of juvenile blue crabs (Callinectes sapidus Rathbun). J. Exp. Mar. Biol. Ecol. 198: 155–173. Peterson, M. S., B. H. Comyns, J. R. Hendon, P. J. Bond, and G. A. Duff. 2000. Habitat use by early life-history stages of fishes and crustaceans along a changing estuarine landscape: differences between natural and altered shoreline sites. Wetl. Ecol. Manage. 8: 209–219. Pile, A. J., R. N. Lipcius, J. van Montfrans, and R. J. Orth. 1996. Density-dependent settler-recruit-juvenile relationships in blue crabs. Ecol. Mono. 66: 277–300. Rozas, L. P. and T. J. Minello. 1998. Nekton use of salt marsh, seagrass, and nonvegetated habitats in a south Texas (USA) estuary. Bull. Mar. Sci. 63: 481–501. ______and R. J. Zimmerman. 2000. Small-scale patterns of nekton use among marsh and adjacent shallow nonvegetated areas of the Galveston Bay estuary, Texas (USA). Mar. Ecol. Prog. Ser. 193: 217–239. SAS Institute. 2003. SAS User’s Guide: Statistics (version 9.1). Cary, North Carolina: SAS In- stitute, Inc. Seitz, R. D., R. N. Lipcius, and M. S. Seebo. 2005. Food availability and growth of the blue crab in seagrass and unvegetated nurseries of Chesapeake Bay. J. Exp. Mar. Biol. Ecol. 319: 57–68. Sheridan, P. 1992. Comparative habitat utilization by estuarine macrofauna within the mangrove ecosystem of Rookery Bay, Florida. Bull. Mar. Sci. 50: 21–39. Snelgrove, P. V. R. 1999. Getting to the bottom of marine biodiversity: sedimentary habitats. BioScience 42: 129–148. Tagatz, M. E. 1968. Biology of the blue crab, Callinectes sapidus Rathbun, in the St. John’s River, Florida. Fish. Bull. 67: 17–33. NOTES 565

Thomas, J. L., R. J. Zimmerman, and T. J. Minello. 1990. Abundance patterns of juvenile blue crabs (Callinectes sapidus) in nursery habitats of two Texas bays. Bull. Mar. Sci. 46: 115– 125. Valiela, I., J. L. Bowen, and J. K. York. 2001. Mangrove forests: one of the world’s most threatened major tropical environments. Bioscience 51: 807–815. Virnstein, R. W. 1977. The importance of predation by crabs and fishes on benthic infauna in Chesapeake Bay. Ecology 58: 1199–1217. Weisberg, R. H. and L. Zheng. 2006. Circulation of Tampa Bay driven by buoyancy, tides, and winds, as simulated using a finite volume coastal ocean model. J. Geophys. Res. C. 111: 1–20. Williams, A. H., L. D. Coen, and M. S. Stoelting. 1990. Seasonal abundance, distribution, and habitat selection of juvenile Callinectes sapidus (Rathbun) in the northern Gulf of Mexico. J. Exp. Mar. Biol. Ecol. 137: 165–183. Wilson, K. A., K. W. Able, and K. L. Heck Jr. 1990. Habitat use by juvenile blue crabs: a comparison among habitats in southern New Jersey. Bull. Mar. Sci. 46: 105–114. Zar, J. H. 1999. Biostatistical Analysis 4th ed. Prentice Hall, New Jersey. 185 p.

Addresses: (L.A.Y., J.M.K., A.B.B., C.C.M.) U.S. Geological Survey, 600 4th St. S., St. Peters- burg, Florida 33701. Corresponding Author: (L.A.Y.) Current Address: Florida Inter- national University, Department of Biological Sciences, 11200 S.W. 8th Street, Miami, Florida 33199. E-mail: . BULLETIN OF MARINE SCIENCE, 80(3): 567–586, 2007

MANGROVES AS ESSENTIAL NURSERY HABITAT FOR GOLIATH GROUPER (EPINEPHELUS ITAJARA)

Christopher C. Koenig, Felicia C. Coleman, Anne-Marie Eklund, Jennifer Schull, and Jeffrey Ueland

Abstract We evaluated goliath grouper’s [Epinephelus itajara (Lichtenstein, 1822)] use of mangroves as essential nursery habitat by estimating absolute abundance, density, survival, age structure, home range, mangrove habitat association, habitat quality, and recruitment to the adult population. Densities (numbers km–1 mangrove shoreline) were calculated using Jolly-Seber mark-recapture methods for mangrove-lined rivers and mangrove islands of the Ten Thousand Islands (TTI) and Everglades Na- tional Park, which includes Florida Bay, Florida, USA. Juveniles had smaller home ranges around islands (170 m) than in rivers (586 m), as determined from observations on telemetered fish. Goliath grouper remained in mangrove habitats for 5–6 yrs (validated ages from dorsal spine sections), then emigrated from mangroves at about 1.0 m total length. In the TTI, juvenile densities around mangrove islands were higher (mean = 25 km–1, SE = 6.2, CV = 0.5) and less variable than those in rivers (mean = 11 km–1, SE = 4.2, CV = 1.2). Density was negatively correlated with the frequency of dissolved oxygen and salinity minima. Mean growth rate of recaptured fish around mangrove islands (0.358 mm –1d , 95% CL = 0.317–0.398) was significantly higher than that in rivers (0.289 mm d–1, 95% CL = 0.269–0.308). The annual survival rate, as estimated by the Kaplan-Meier method on telemetered fish, was 0.947 (95% CL = 0.834–1.0). Very low densities in Florida Bay were probably related to other water-quality variables in this human-altered system. The offshore abundance of adults was largely explained by abundance of mangrove, but not seagrass habitat. Mangrove habitat with suitable water conditions, which appears essential to the recovery and sustainability of goliath-grouper populations, should be protected and/or restored.

Nearshore estuarine habitats are thought to provide “essential” 1 nursery habitat for many fish species (Kapetsky, 1985; Koenig and Coleman, 1998; Thorrold et al., 1998; Fitzhugh et al., 2005), but definitive tests of the nursery role rarely occur (Beck et al., 2001). Confirmation of the nursery role requires an understanding of the focal species’ life history and a thorough evaluation that culminates in demonstration of the habitat’s greater contribution per unit area to the production of the adult population (Beck, 1995; Beck et al., 2001; Laegdsgaard and Johnson, 2001; Nagelkerken et al., 2002; Halpern, 2004a,b). This confirmation is critical from a conservation perspective because it provides a distinction between source and sink habitats; source habitats are currently the primary focus of efforts to protect fishery production (Crowder et al., 2000). In this study, we tested the hypothesis that mangroves are integral and essential to the early life stages of goliath grouper [Epinephelus itajara (Lichtenstein, 1822)] by pursuing four objectives: (1) estimating juvenile density, survival, growth, and move-

1 Throughout the present paper, “essential” is used in the sense of the Sustainable Fisheries Act, wherein “essential fish habitat” is defined as “those waters and substrate necessary to fish for spawning, feeding, or growth to maturity” and “necessary” means “required to support a sustainable fishery and the managed species’ contribution to a healthy ecosystem.”

ment patterns in the Ten Thousand Islands (TTI), southwest Florida; (2) comparing our estimates of density in the TTI with abundance estimates from published and unpublished sampling records from other estuarine and offshore habitats; (3) comparing the effects of water condition on densities in mangrove habitats of the TTI, in rivers of the Everglades National Park (ENP), and in Florida Bay (FB), a system altered by human activities; and (4) investigating the relative importance of juvenile habitat to adult production by correlating regional near-shore habitat coverage with offshore adult abundance.

Study Animal

The goliath grouper is the largest grouper in the western North Atlantic. It ma- tures in 5–7 yrs, is long-lived (surviving up to 37 yrs), and reaches sizes exceeding 400 kg and 3.0 m total length (TL) (Robins et al., 1986; Bullock et al., 1992). Its life cycle includes ontogenetic shifts in habitat use: larvae move from the pelagic environment to shallow coastal areas, benthic juveniles live in estuaries, and adults occur on offshore reefs (< 50 m deep) (Eklund and Schull, 2001). Goliath grouper is considered by the World Conservation Union (IUCN) to be critically endangered throughout its range which extends from the west coast of Af- rica to the west coast of Central America including the Caribbean, and from Brazil to the southeastern U.S. (Heemstra and Randall, 1993; IUCN Red List, 2006, [http:// www.iucnredlist.org/search/search-basic]). A fishing ban was put into effect in the southeastern U.S. in 1990 by the South Atlantic and the Gulf of Mexico fishery management councils and in the Caribbean in 1993 by Caribbean Fishery Management Council because it was considered to be severely overfished (GMFMC, 1990; Sadovy and Eklund, 1999). Although the National Marine Fisheries Service (NMFS) continues to list this species as overfished (NMFS, 2004), the actual status of the population remains unknown and difficult to assess (Porch and Eklund, 2004), as is often the case with fully protected species for which fisheries data collection has not been replaced with reliable population monitoring. This situation hinders development of management measures aimed at rebuilding the population, ending overfishing, or both, as required by the Sustainable Fisheries Act (Public Law 104-297).

Study Sites

Study sites in the TTI and ENP (Fig. 1A) occurred in red mangrove (Rhizopho- ra mangle Linnaeus) habitat along the shorelines of islands, tidal rivers and passes, and shallow bays. The TTI system covers 1924 km2 between latitudes 25°40´N and 26°00´N and longitudes 081°20´W and 081°40´W. The ENP covers the southern tip of the Florida peninsula and extends through FB, the largest (2072 km2) subtropical estuary in North America (Loftus, 2000). In both systems, we haphazardly chose widely dispersed study locations primarily on the basis of accessibility (sites were often accessible only at high tides). In the TTI, we sampled seven rivers (Littlewood, Palm, Blackwater, Pumpkin, Wood, Whitney, and Ferguson), three man-made canals (92 Canal East, 92 Canal West, and Faka Union Canal), and four mangrove-island systems (Remuda Ranch, Russell Key, Fakahatchee, and Rabbit Key) (Fig. 1B). Water depths averaged < 0.1 to 2.0 m (MLW), reaching 5 m in passes. Mixed diurnal tides range from 2.9 ft at Naples to 4.5 ft KOENIG ET AL.: JUVENILE GOLIATH GROUPER HABITAT 569

Figure 1. Study sites in southwest Florida, USA. A. Areas covered by the present study and Flor- ida coastal regions as delimited for Reef Environmental Education Foundation surveys. B. Ten Thousand Islands system, showing sites used in the present study (those marked with asterisks are man-made). River and canal sites: (A) 92 Canal West,* (B) 92 Canal East,* (C) Palm River, (D) Blackwater River, (E) Whitney River, (F) Pumpkin River, (G) Little Wood River, (H) Faka Union Canal,* (I) Ferguson River. Passes: (1) Remuda Ranch Channel, (2) Fakahatchee Pass, (3) Russell Pass, (4) Rabbit Key Pass. C. Everglades National Park system. Rivers: (A) Lostmans River, (B) Broad River, (C) Harney River, (D) North Prong River, (E) Shark River, (F) Roberts River. Bay sites: (1) Bob Allen Key, (2) Black Betsey Key, (3) Little Buttonwood, (4) Blackwater Sound. Gray = land; white = water. 570 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 at the Shark River (NOAA tide tables). Undercuts (= scoured shorelines) formed a structurally complex habitat in sampled areas, especially around mangrove islands where mangrove roots formed “curtains” and submerged dead trees provided additional habitat structure. The ENP study sites included six rivers (Harney, Broad, Shark, Roberts, Lostman’s, and North Prong) entering the Gulf of Mexico and four mangrove-island sites (Blackwater Sound, Little Buttonwood Sound, E. Bob Allen Key, and Black Betsy Key) in FB (Fig. 1C). The river sites in the TTI and ENP differed little in appearance, whereas the TTI mangrove island sites differed from those in the ENP in being shallower and more widely spaced and having fewer undercuts.

Materials and Methods

Juveniles Capture.— In the rivers, we captured fish using blue-crab traps and commercial grouper fish traps baited with dead striped mullet (Mugil cephalus Linnaeus, 1758). Blue-crab traps were 61 cm wide by 61 cm long by 46 cm high and made of 3.8-cm, 17.5-gauge plastic-coated wire mesh; two funnels, each 13 cm long (proximal opening 19 × 12.7 cm; distal opening, 18 × 7.6 cm), led from the outside into a lower chamber, and two funnels, each 10 cm long (proximal and distal openings, 18 × 7.6-cm) led from the lower chamber into an upper chamber. The fish traps were 61 cm wide by 107 cm long by 46 cm high. The top and bottom panels were 3.8-cm, 12.5-gauge plastic-coated wire mesh, and the side panels 2.5 × 5.2-cm, 14-gauge wire mesh. The single entry funnel was 53 cm long (proximal opening, 45.7 × 30.5 cm; distal opening, 30.5 × 10.2 cm). We deployed 50 traps in each river at 93-m (0.05 NM) intervals in a linear sequence 4.6 km (2.5 NM) long (Figs. 2A,B). The line of traps extended up each river from the mouth, except in rivers shorter than 4.6 km, such as the Pumpkin, Little Wood (Fig. 2B), and Whitney, where it extended past the river mouth into the bay. Trap spacing was arbitrary, but small enough to maximize catch and large enough to sample most of the river or canal. Each deployment lasted 21 d (sampled at 7 d intervals), with inter-deployment intervals about 42 d. Capture-efficiency studies during the first year identified the peak capture period for subsequent -sam pling: captures per set from June through November (105/1365 = 0.077) were significantly greater (binomial test, P < 0.0001) than those from January through May (6/460 = 0.013). We sampled mangrove-island shorelines monthly during the summer and fall, mostly at neap tides, using setlines (Fig. 2C). Each setline was composed of a baited 14-0 circle hook attached by a 10-cm piece of 600-lb. test monofilament (which minimized hook electrolysis) to a 500-lb. test stainless steel cable attached to a 4-cm long section of 20 cm diam PVC pipe (as a reel and an aid in relocating the setlines), which was attached to a mangrove limb. At least three setlines were used per site (24–27 lines per sampling area). Each sampling event included at least two bait-and-sample rounds, each 4–12 hrs in duration. We tagged each newly captured fish ventrally with an individually numbered stainless- steel-core internal-anchor tag (Floy Tag Company) and removed the third dorsal-fin spine and second and third dorsal-fin rays for age determination (Fig. 3). Records for each fish captured or recaptured included the time, date, tag number, weight (g), size (mm total length, TL), capture location (Trimble GeoExplorer GPS receiver in Arcview 8 GIS), and evidence of previous spine and ray removal. We recaptured fish ourselves, but we included a toll-free telephone number on each tag so that others could also report captures. Age, Growth, and Tag Loss.—Collected spines were cleaned and dried, and several trans- verse 0.5 mm sections were cut from each (Beuhler low speed isomet saw) 5–10 mm above the spine base. The sections were mounted on slides, and two readers independently determined age by counting translucent bands (= annuli) under compound dissecting microscopes. The spine-aging method was validated by comparison with ages determined from otoliths and by determination of incremental change over time from additional spines collected from re- KOENIG ET AL.: JUVENILE GOLIATH GROUPER HABITAT 571

Figure 2. Examples of capture sites for juvenile goliath grouper, Epinephelus itajara in the Ten Thousand Islands. (A) trap sampling in the Palm River. (B) trap sampling in the Little Wood River. (C) setline sampling in Russell Pass. Traps were set at 0.093-km intervals. Sites of capture (small closed circles) and no capture (small open circles) are indicated along with home range (larger open circles) around capture sites. Gray = land; white = water. 572 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Figure 3. Juvenile goliath grouper Epinephelus itajara (38 cm TL) caught by fish trap in the Ten Thousand Islands. Note tag on abdomen. Photograph by M. Finn. captured fish that were either marked with oxytetracycline (50 IU kg–1) or not (Brusher and Schull, unpubl. data). Growth rates (average daily increase in length, mm day–1) were estimated from recaptured fish. To minimize the effects of measurement error on growth rate we used only recaptured fish at liberty for > 30 d (n = 259). Growth was assumed to be linear over the size range of juveniles captured in the study (following Bullock et al., 1992). Growth rate differences among sites were tested for significance with the Kruskal-Wallis non-parametric test (no normalizing transformation was found). Where data were normal (Shapiro-Wilk test), ANOVA was used to test for differences. Tag loss was indicated by evidence of previous spine and ray removal in untagged fish. Absolute Abundance and Survival.—To estimate juvenile absolute abundance and survival, we used Jolly-Seber (J-S) mark-recapture methods (Krebs, 1999), the assumptions of which are (1) that every individual in the sampled population has the same capture probability (equal catchability); (2) that every marked individual has the same probability of surviving between sampling times; (3) that no tag loss occurs during the sampling period. It also requires that sampling time be short relative to intervals between samples (a feature built into our sampling design), that at least three sampling times occur for each site, and that a relatively high proportion (~ 0.50) of fish be marked. We knew a priori that satisfying the equal catchability assumption would be difficult because no single gear type uniformly sampled the entire juvenile size range (20–1000 mm TL). (We collected newly settled juvenile goliath grouper at about 20 mm TL in multiple locations in mangrove leaf litter, but could not include them in the J-S abundance estimate due to collection gear limitations. However, they were included in the estimates of overall absolute abundance and density.) Traps worked in rivers but not around mangrove island undercuts, where they were dislodged by swift currents and rapidly fouled. Despite this bias and the potential difficulty of distinguishing habitat-specific size (or age) distributions from gear selectivity, we pursued multigear sampling to ensure capture of a greater size range (and therefore age spectrum) of fish. We tested for equal catchability within the selection limits of the gear with the zero-truncated Poisson test and then compared the observed and expected frequency distributions using a chi-squared goodness-of-fit test (Krebs, 1999). KOENIG ET AL.: JUVENILE GOLIATH GROUPER HABITAT 573

We estimated absolute abundance for all sizes, 20–1000 mm TL and for all ages, 0–6, present in the mangroves (NA) (1) for trap sampling, by determining site-specific abundances of

2-yr-olds, the age most often selected by traps (= mean proportion of 2-yr-olds caught × NJ-S in each river), and (2) for setline sampling, by determining the site-specific abundances of 4-yr- olds, the age most often selected by setlines (= mean proportion of 4-yr-olds caught × NJ-S at each mangrove island site). We calculated the number in each age group by reconstructing all age groups (ages 0–5) based on M = 0.05, our Kaplan-Meier estimate of survival (see below), then summing to obtain the absolute abundance of all age groups at each sampling site during the sampling period. We did not include age 6 in the calculation because most fish appeared to leave the mangroves in their fifth year (Table 1). To ensure equal probability of survival from one sampling time to the next we made sure that all released fish were in good condition. We estimated survival in two ways: (1) with mark-recapture and the J-S method, and (2) with telemetry and the Kaplan-Meier method (Krebs, 1999). J-S survival estimates, which depend only on the marked population, can be confounded by many factors related to the behavior of the fish and/or the sampling design. For example, if marked individuals emigrate, or for some reason cannot be recaptured (e.g., they outgrow the sampling gear or actively avoid recapture), they are assumed by the model to have died. The Kaplan-Meier survival estimate does not have such limitations and makes use of all the relocation data until the fish either dies or is lost from the sample. Density, Habitat Association, and Home Range.—We estimated home range by tagging 22 juveniles (size range 430–940 mm TL) intraperitoneally with individually coded transmitters (Sonotronics; battery life 48 mo) and relocating them during our normal sampling activities over the period of the study. We assumed that tracking fish telemetrically left them relatively undisturbed and therefore provided more reliable information on movement patterns and home ranges than recaptures with traps or setlines, which could affect their behavior. So that telemetered fish could be recognized externally, a numbered internal-anchor tag with external streamer was placed on either side of the incision, which was closed with stainless steel medical staples and covered with antibiotic salve before the fish was released. We relocated telemetered fish with a directional hydrophone (Sonotronics), recorded positions, and confirmed that detected fish were still alive by direct observation using SCUBA or by prodding. We then estimated home range (= distance traversed by an individual during the sampling period) by averaging the greatest straight-line distances between relocation sites (White and Garrott, 1990). Half of the home range was equivalent to the radius of a circle centered on the sampling site (Seber, 1982). Home ranges were estimated separately for rivers and islands. Densities (D) were determined from Dn= H-1 A (1)

Table 1. Size and age distribution of juvenile goliath grouper, Epinephelus itajara, in mangrove- lined rivers and mangrove islands of the Ten Thousand Islands, southwest Florida, USA, 1998– 2000. All fish caught in rivers were taken in traps; those caught near mangrove islands were taken with setlines. Data are numbers of individuals; TL = total length.

TL (mm) Age 0 Age 1 Age 2 Age 3 Age 4 Age 5 Age 6 Total 101–200 1 8 3 25 37 201–300 3 90 107 108 3 311 301–400 52 239 100 9 400 401–500 1 81 48 17 2 149 501–600 8 21 36 3 68 601–700 4 8 34 11 1 58 701–800 3 22 17 2 44 801–900 15 17 2 34 901–1,000 6 7 2 15 Total n 4 151 442 313 142 57 7 1,116 574 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

where nA = estimated absolute abundance for all size classes (20–1000 mm TL) and H = kilometers of mangrove shoreline within the home range about each sampling area. Patterns of dispersion were estimated by use of the coefficient of dispersion (CD, variance to mean ratio), where means and variances were calculated from the number of fish captured at each sampling site over the 2 yr sampling period. Water Condition.—We evaluated the relationship between dissolved oxygen (DO, mg L–1) and salinity and fish density using 5 yrs of water quality data (1997–2001) collected monthly near or at our sampling sites from the surface and bottom (total samples = 116–120 per site) by the South Florida Water Management District and archived on websites [http://www.sfwmd.gov/org/ema/envmon/wqm/coastal/tti.html and http://www.sfwmd.gov/org/ema/envmon/wqm/flab/flabindex.html]. We assumed that salinity and DO were among the most important variables affecting fish distribution in these systems. We did not include temperature minima in this analysis—even though goliath grouper are sensitive to low temperatures (Sadovy and Eklund, 1999)—because any cold fronts moving through the area would affect the entire SW Florida system, not just individual sites within the system. The State of Florida minimum DO water quality standard for marine waters is 4 mg L–1, but we used 3 mg L–1 to reflect the dynamic diurnal and seasonal conditions in estuarine habitats. The State of Florida has no salinity standards for estuaries, so we used lower (salinity of 5) and upper (salinity of 40) limits known for other serranid fishes found in estuaries (e.g., gag, black sea bass, and red grouper) (Koenig and Coleman, 1998; Atwood et al., 2001).

Adults Abundance.—We used goliath grouper sighting frequency data as a proxy for adult population abundance rather than population-density data to minimize the effects of feeding or reproductive aggregations on regional estimates. We determined sighting frequency based on dive surveys conducted by The Reef Environmental Education Foundation (REEF) on Florida reefs from 1993 to 2004 (REEF database = http://www.reef.org/data/twa/index.htm). REEF surveyors use a standardized “roving diver method” in which divers swim freely through sites, recording every observed fish species that can be positively identified. Because adult goliath grouper are distinct in terms of size, shape, and color pattern, we are confident that both novice and expert surveyors made correct identifications. We compared goliath grouper sighting frequency off Florida with the abundance of two major estuarine habitat types in coastal Florida, mangrove habitat (data from the Florida Nat- ural Areas Inventory, available at http://www.fnai.org/about.cfm) and seagrass habitat (data from Florida Fish and Wildlife Conservation Commission’s Fish and Wildlife Research Insti- tute), using a logistic regression model to determine the significance (P < 0.05), direction, and strength of the relationship (R2) between the variables.

Results

Age, Growth, and Tag Loss.—Juvenile goliath grouper (n = 1116) were found in mangrove habitat up to age 6 and 1000 mm TL (Table 1). The overall mean growth rate for all recaptured fish was 0.300 mm –1d (95% CL = 0.282–0.318, n = 259). In the TTI, growth rates were not significantly different among rivers (Kruskal-Wallis, P > 0.05) nor among mangrove-island sites (Kruskal-Wallis, P > 0.05), but growth rates around islands (mean = 0.358 mm d–1, 95% CL = 0.317–0.398, n = 41) were significantly greater (Kruskal-Wallis, P < 0.01) than growth rates in rivers (mean = 0.289 mm d-1, 95% CL = 0.269–0.308, n = 218). We found no significant (Kruskal-Wallis, P > 0.05) relationship between growth rate and population density among rivers, nor among island sites. Tag loss was negligible (n = 1), all fish were released in good condition, and recaptured fish did not have infected tagging wounds. KOENIG ET AL.: JUVENILE GOLIATH GROUPER HABITAT 575

Figure 4. Percent frequency of size classes of juvenile goliath grouper Epinephelus itajara captured by each of the three gear types in the Ten Thousand Islands. TL = total length. Absolute Abundance and Survival Equal Catchability.—Although TTI goliath grouper met the general assumption of equal catchability (within the selective limits of the gear; zero-truncated Poisson test; χ2, P > 0.05), all three gears showed some size (= age) selectivity (ANOVA, P < 0.05; Fig. 4). Crab traps selected primarily fish 300–400 mm TL because the funnel opening excluded larger individuals and the mesh size allowed smaller ones to escape during trap retrieval. Fish traps were less selective at the upper size range but also allowed fish < 200 mm TL to escape. Setlines captured primarily fish 500–1000 mm TL and appeared to exclude fish < 400 mm TL. Absolute Abundance.—Because of gear selectivity, the J-S method only estimated the absolute abundance of the selected sizes for each gear type. These were calculated directly by the J-S method and indirectly with a predictive model (Table 2; Fig. 5) based on catch per unit effort (CPUE; one unit of effort = one set of traps 4.6 km long), such that:

YX=-79..+ 51 (2)

where Y is the absolute abundance and X is the mean CPUE for the sampling site (Linear Regression, P < 0.05, R2 = 0.89). The CPUE model was developed from annual J-S estimates of absolute abundance of trap catches for 1999 and 2000 for six rivers. [We did not include Little Wood River because the fish in that river were unusually abundant and highly clumped (Fig. 2B); such a situation produced a low CPUE relative to absolute abundance because traps saturated in high density areas around the river mouth]. The regression model provided an efficient means of estimating abundance outside of the TTI system and in areas where capture frequencies were too low for direct use of the J-S method. Estimating the abundance over all sizes (and therefore all gears) required accept- ing three assumptions: (1) that all sizes were present in river and island systems, (2) that year-class strength was constant among sampling years, and (3) that mortality was uniform across all age groups. The first assumption is justified because in our collections, fish sizes ranged from 20 to 800 mm TL in rivers and 20 to 930 mm TL 576 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Table 2. Mean estimates of absolute abundance (n), survival probability (PHI), and density (D) of juvenile goliath grouper, Epinephelus itajara, sampled in mangrove shoreline of the Ten Thou- sand Islands (TTI), Everglades National Park (ENP), and Florida Bay (FB), southwest Florida, USA, 1998–2001. Density estimates include all fish 20–1,000 mm SL (ages 0–5). S = number of sampling times. α = maximum proportion of population marked. nJ-S = age-specific absolute abundance estimate based on Jolly-Seber model. PHI = finite probability of survival standardized to 30 d. nA = absolute abundance based on all ages, with annual mortality (M) of 0.05. H = area sampled in kilometers of mangrove shoreline.

System S α nJ-S PHI nA H D (95% CL) (95% CL) (M = 0.05) (km) (n km−1) TTI rivers Little Wood River 6 0.39 202 (163–242) 0.84 (0.48–1.0) 571 13.1 44 Palm River 6 0.58 70 (58–81) 0.69 (0.42–0.93) 196 9.2 21 Blackwater River 6 0.67 42 (1–83) 0.64 (0.28–0.87) 118 9.2 13 92 Canal East 6 0.79 41 (28–55) 0.86 (0.60–1.0) 116 10.4 11 Pumpkin River 5 0.43 38 (14–62) 0.81 (0.52–1.0) 106 8.5 12 Faka Union Canal 5 0.26 12 (2–22) 0.32 (0.16–0.60) 24 10.6 2 92 Canal West 5 0.72 10 (8–36) 0.64 (0.23–0.88) 27 10.4 3 Wood River* 5 – 5 (0–20) – 14 5.4 3 Whitney River* 6 – 3 (0–20) – 8 8.2 1 Ferguson River* 3 – 3 (0–20) – 8 10.7 1 ENP rivers Harney River* 2 – 26 (18–36) – 74 15.9 5 Broad River* 2 – 18 (7–29) – 52 15.8 3 Shark River* 2 – 3 (0–20) – 8 12.2 1 Roberts River* 2 – 0 – 0 13.7 0 Lostmans River* 2 – 0 – 0 7.8 0 North Prong* 2 – 0 – 0 12.1 0 TTI mangrove islands Remuda Ranch Pass 4 0.42 31 (13–158) 1.0 (0.57–1.0) 82 2.0 41 Russell Key Pass 7 0.80 31 (15–117) 0.78 (0.32–1.0) 82 3.0 27 Fakahatchee Pass 5 0.75 18 (12–58) 0.61 (0.22–0.93) 49 3.2 15 Rabbit Key Pass 7 0.86 23 (13–70) 0.95 (0.53–0.97) 60 4.0 15 FB mangrove islands Blackwater Sound* 2 – 3 (0–20) – 8 3.6 2 Little Buttonwood S* 2 – 0 – 0 4.3 0 E. Bob Allen Key* 2 – 0 – 0 2.3 0 Black Betsey Key* 2 – 0 – 0 5.3 0 *Absolute abundance estimate based on catch-per-unit-effort linear-regression model (Y = −7.9 + 5.1X). around islands. That, combined with strong site fidelity (see below), suggests that juveniles remained near settlement sites in both systems. The second assumption is justified because between sampling years at each site, we observed similar J-S absolute abundances, suggesting a similar year-class size among years, especially in rivers where highly selective crab-trap gear was used. For rivers, the mean absolute abundances for 1999 and 2000 were: 217 and 193 for Littlewood; 64 and 73 for Palm; 33 and 42 for Blackwater; 37 and 39 for Pumpkin; and 48 and 37 for 92 Canal East. For island sites, the mean absolute abundances for 1998 and 1999 were: 28 and 33 for Remuda Ranch; 18 and 18 for Fakahatchee; 34 and 17 for Rabbit Key; and 36 and 28 for Russell Key. The third assumption is probably unjustified because young (small) fish likely experience higher mortality rates than older (larger) fish, which is a general KOENIG ET AL.: JUVENILE GOLIATH GROUPER HABITAT 577

Figure 5. Linear regression of absolute abundance on catch per unit effort (CPUE; individuals per trap set) for trap-caught goliath grouper, Epinephelus itajara, in rivers of the Ten Thousand Islands: Blackwater River, Palm River, Pumpkin River, Faka Union Canal, 92 Canal West, and 92 Canal East. Mean values for 1999 and 2000. Regression model: Y = −7.9 + 5.1X (R2 = 0.89). Dashed lines represent 95% confidence boundaries. pattern (Roff, 1992). Because we have no way of estimating size-specific mortality, we used the Kaplan-Meier estimate of annual mortality, M = 0.05, which was observed on telemetered fish from about 400 to 1000 mm TL. Survival.—The proportion (α) of marked juvenile goliath grouper increased over the course of the study to a maximum that usually exceeded 0.5 (Table 2), suggesting high survival of marked fish, which was further supported by monitoring telemetered fish. The survival probability of TTI mangrove island fish (mean = 0.835) and river fish (mean = 0.727) did not differ significantly in J-S 30-d survival probability (ANOVA, P > 0.05, Table 2). Two of the transmitters implanted in fish malfunctioned (could not be detected upon release), and three others were never detected after the initial check. The remaining 17 fish were monitored for 34 mo, and Kaplan-Meier survival was estimated on those fish. The 30-d survival rate was 0.995 (95% CL = 0.985–1.00); annual survival rate was 0.947 (95% CL = 0.834–1.00). Habitat Association, Home Range, and Density.—Observations of telemetered juveniles showed that they were strongly associated with structure (mangrove undercuts and root curtains, submerged mangrove branches or trees, and rocky outcrops). They also showed patchy distributions in both rivers and around islands (mean CD for islands = 5.3 and rivers = 3.0). Because structure is associated with shoreline, we used a linear measure (number km–1 shoreline) as the basis for density estimation. Recaptured juveniles showed high site fidelity; over 90% moved no farther than 0.5 km during the entire sampling period. Straight-line home ranges for telemetered fish in rivers averaged 586 m (radius = 293 m) and those around mangrove islands, 170 m (radius = 85 m; Table 3). Home range estimates served as the basis for density estimation. Mean densities around mangrove islands (25 km–1, SE = 6.2, CV = 0.5) were significantly greater and less variable than those in rivers (11 km–1, SE = 4.2, CV = 1.2) (Table 2). 578 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Table 3. Home-range estimates of telemetered juvenile goliath grouper, Epinephelus itajara, in the Ten Thousand Islands rivers and mangrove islands, southwest Florida, USA. Ranges are measured linearly, in kilometers of mangrove shoreline.

Location Fish size at Times observed Observation period Maximum distance tagging (mm TL) (months) traversed (m) Rivers Little Wood R. 555 22 20 259 543 20 18 611 Palm R. 438 18 12 926 620 10 19 315 579 24 20 818 Mean home range 586 Mangrove islands Rabbitt Key 610 6 26 340 640 9 16 350 Russell Key 909 4 10 197 825 2 9 86 580 4 16 100 Fakahatchee 695 9 25 50 805 4 15 67 580 5 25 86 Pumpkin Bay 508 8 13 100 559 16 31 320 Mean home range 170

Water Condition.—Fish densities in all TTI rivers and mangrove islands were negatively correlated with low salinity (Correlation, P < 0.05) and with low salinity in concert with low DO (Correlation, P < 0.01) (Fig. 6), but not with low DO alone (P = 0.07). Low densities in FB were not correlated with low DO or salinity (Table 4). Adult Abundance.—Offshore adult abundance (from REEF surveys) was positively correlated with the quantity of mangrove shoreline along the Florida coast (Logistic Regression, P < 0.001, R2 = 0.83; Table 5, Fig. 7). A stronger relationship (R2 = 0.92) resulted if the Florida Keys were excluded—exclusion was justified by the low proportion of goliath-grouper sightings (4% of 14,386 observations) and the poor water quality around the Keys (Kruczynski, 1999). Adult abundance was negatively related to seagrass abundance (Logistic Regression, P < 0.001, R2 =0.75).

Discussion

Beck et al. (2001) refined the definition of nursery habitat in such a way that demonstrating nursery function for a particular species requires that researchers evaluate density, survival, growth, and above all, the contribution of the habitat to the production of the adult population. Other than this study, there are virtually no published studies that have unequivocally demonstrated a nursery function of mangroves for fishes according to this definition (Faunce and Serafy, 2006). This study clearly demonstrates that mangroves can only provide important nursery habitat for juvenile goliath grouper when local environmental conditions are suitable. Rivers are more likely than mangrove islands to experience frequent hypoxic (DO < 3 mg L–1) or low-salinity (< 5) conditions because they drain upland ar- KOENIG ET AL.: JUVENILE GOLIATH GROUPER HABITAT 579

Figure 6. Relationship between density (number km–1 of mangrove shoreline) of juvenile goliath grouper, Epinephelus itajara, and dissolved oxygen and salinity minima (percent of monthly sampled dissolved oxygen and salinity measures collected top and bottom from 1997 to 2001, total of about 120 samples, below 3.0 mg L–1 and 5, respectively), in the Ten Thousand Islands and Everglades National Park rivers and mangrove islands. Negative correlations for dissolved oxygen and salinity minima are P = 0.07 and P = 0.03, respectively. eas, include man-made canals, and are impacted more acutely by upland freshwater management projects that occur throughout this area (Ogden et al., 2005). Thus, it is not surprising to find in rivers that juveniles occur at lower densities, experience slower growth rates, and have greater home ranges than island fish—keeping in mind that density may be influenced by the availability of structure (Frias-Torres, 2006), and growth rates may be affected by food availability. The near lack of juveniles in FB is best explained by the bay’s long-standing degraded state, characterized by periods of hypersaline and high-nutrient conditions; periodic toxic phytoplankton blooms; and significant seagrass, sponge, and mangrove die-offs (Boyer et al., 1999; Fourqurean and Robblee, 1999). These conditions and events probably affected nursery function (Butler et al., 1995), although other factors, such as limited larval delivery and lack of microhabitat features, cannot be ruled out. Nursery habitat is expected to promote survival in part by providing refuge from predation. While mangrove habitat is rife with possible predators of juvenile goliath grouper—including gray snapper (Lutjanus griseus Linnaeus, 1758), snook [Centro- pomus undecimalis (Bloch, 1792)], red drum [Sciaenops ocellatus (Linnaeus, 1766)], and gag [Mycteroperca microlepis (Goode and Bean, 1879)], as well as bull sharks [Carcharhinus leucas (Müller and Henle, 1839)], and lemon sharks [Negaprion brevirostris (Poey, 1868)]—it also is structurally complex, providing significant refugia. Measuring survival is problematic. We found that the J-S method routinely underestimated survival in goliath grouper because fish loss due to emigration or es- capement was indistinguishable from fish loss due to mortality. The Kaplan-Meier method minimizes these biases, and therefore provides a more realistic estimate. Indeed, the Kaplan-Meier survival estimate is quite high (annual rate = 0.947, 95% 580 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Table 4. Dissolved oxygen, salinity, and the density of juvenile goliath grouper, Epinephelus itajara, in mangroves at Ten Thousand Islands (TTI), Everglades National Park (ENP), and Florida Bay (FB) water-quality stations. Density estimated for all juveniles 20–1,000 mm TL (ages 0–5). DO min = percentage of monthly dissolved oxygen observations below 3 mg L–1. Salinity min = percentage of monthly salinity observations below 5; max = percentage of monthly salinity observations above 40. Density was measured as individuals per km of shoreline.

Salinity Location Station DO min Minimum Maximum Density Rivers Littlewood River TTI 72 2.5 0.0 3.4 44 Palm River TTI 75 15.0 2.6 0.0 21 Blackwater River TTI 75 15.0 2.6 0.0 13 Pumpkin River TTI 72 2.5 0.0 3.4 12 92 canal E TTI 75 15.0 2.6 0.0 11 Harney River FLAB 36 10.0 0.0 0.0 5 92 Canal W TTI 75 15.0 2.6 0.0 3 Wood River TTI 70 8.6 15.5 3.4 3 Faka Union Canal TTI 70 8.6 15.5 3.4 3 Broad River FLAB 35 5.1 13.0 0.0 3 Whitney River TTI 75 15.0 2.6 0.0 1 Ferguson River TTI 64 16.0 12.0 1.7 1 Shark River FLAB 39 28.0 37.0 0.0 1 Lostmans River FLAB 29 0.0 20.0 0.0 0 Roberts River FLAB 47 0.0 39.0 0.0 0 North Prong River FLAB 38 11.2 74.3 0.0 0 TTI mangrove islands Remuda Ranch TTI 74 2.1 0.0 0.0 41 Russell Key TTI 65 0.0 0.0 2.5 27 Fakahatchee TTI 69 1.7 1.7 2.5 15 Rabbit Key TTI 52 0.0 0.0 2.6 15 FB mangrove islands Blackwater Sound FLAB 05 0.0 0.0 0.0 2 E. Bob Allen Keys FLAB 21 0.0 0.0 17.1 0 Little Buttonwood Sound FLAB 05 0.0 0.0 0.0 0 Black Betsey Key FLAB 23 0.0 0.0 0.0 0

CL = 0.834–1.00), and thus strengthens confidence that the marked (= telemetered) population adequately represents the unmarked population, an important consideration in any mark-recapture study. The most important criterion for determining the nursery status of a particular habitat is its contribution to the adult population relative to that of other juvenile habitats occupied by the particular species (Beck et al., 2001). The fact is that juvenile goliath grouper only rarely occur in any other habitats besides mangroves. No juveniles were collected in intensive inshore surveys of seagrass beds (> 2500 trawl tows and gillnet sets by Coleman et al., 1993, 1994; Thayer et al., 1999; Fitzhugh et al., 2005) and mud bottom (~500 trawl tows by Coleman et al., 1993, 1994) from the Florida panhandle to Florida Bay. Only three of 143 juvenile sightings reported to the FWC hotline occurred in habitat other than mangroves, including one offshore, one in seagrass about 100 m from a mangrove coastline, and one in concrete rubble near a river mouth. Further, anecdotal information from interviews with long-time local KOENIG ET AL.: JUVENILE GOLIATH GROUPER HABITAT 581

Table 5. Distributions of adult goliath grouper Epinephelus itajara sightings offshore (1993–2004) and the extent of mangrove shoreline and seagrass area inshore in all Florida coastal regions, from Reef Environmental Education Foundation surveys. % = (no. sightings/no. surveys) × 100.

Location (as marked in Fig. 1A) No. No. % Mangrove Seagrass sightings surveys shoreline (km) (km2) Gulf of Mexico Coast Northwest Florida 36 322 11 48 2,778 Central West Florida 105 332 32 2,778 307 Southwest Florida (with Florida Bay) 365 535 68 8,752 1,538 Florida Keys 640 14,386 4 2,584 2,862 Atlantic Coast NE Florida to Canaveral 33 170 19 1,065 217 Canaveral to Jupiter 61 765 8 545 71 Jupiter to Biscayne 170 4,498 4 532 2,018

fishermen revealed that mangrove-associated goliath grouper are exclusively juveniles (i.e., < 1000 mm TL) (L. Demere, Chokoloskee, FL fishing guide and third generation fisherman), and that offshore reef-associated goliath grouper are virtually all adults (i.e., > 1000 mm TL) (D. DeMaria, Key West, FL commercial spearfisher who targeted goliath grouper before 1990; Porch and Eklund, 2004). Indeed, the cover provided by habitat types other than mangroves is insufficient for a fish that reaches a meter in length. Demonstrating connectivity between nursery and adult habitats for a protected species such as goliath grouper is more difficult in the absence of a regular monitoring program. In this study, we opted to compare regional abundance of adult goliath grouper offshore with the quantity of mangrove shoreline immediately inshore of sampled populations. (We assumed that the local population, still being in the recovery phase, had not reached the carrying capacity that would force population expansion to neighboring lower-density areas.) The relationship between adult abundance and quantity of mangrove shoreline is strongly positive and corroborates the contention that southwest Florida is the center of goliath grouper abundance in the southeastern United States (see Sadovy and Eklund, 1999; Porch and Eklund, 2004). Emigration of juvenile goliath grouper occurred in 5- to 6-yr-olds at about 1000 mm TL, just under the size of maturity estimated by Bullock et al. (1992): 1100–1150 mm TL for males; 1200–1350 mm TL for females. Emigration is more likely to be size rather than age dependent, given that size-at-age became increasingly more variable over time and that fish left the estuary at about one meter TL regardless of age (Table 1). A seasonal cue such as temperature is unlikely to trigger egress in goliath grouper as it appears to in gag (see, e.g., Koenig and Coleman, 1998) because juvenile goliath grouper remain in mangroves for years, whereas gag leave their nursery habitat during the first fall season of residency. Mangrove habitat suitable as goliath grouper nursery in the South Florida Ecosys- tem was probably much more abundant and widespread a century ago than it is now, extending at least from Indian River Lagoon on the east coast to the Caloosahatchee estuary on the west coast, including Biscayne and Florida bays (Ogden et al., 2005). Human activity has had an enormous impact, including that caused by water management, agriculture, industrial and housing developments, roadways, and mosquito abatement, resulting in major changes in the timing, quantity, and quality of freshwater flows into the estuary. Goliath grouper is only one of a diverse array of species 582 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Figure 7. Logistic regression of proportion of REEF surveys with adult goliath grouper (proxy for offshore adult abundance) on kilometers of mangrove shoreline for coastal regions of Florida (P < 0.001, R2 = 0.83). affected by these hydrologic and water quality alterations (Sklar and Browder, 1998; Ogden et al., 2005). In mangrove habitat, goliath grouper appear to be intolerant of large and/or highly variable flows of fresh water, but conditions of low DO, other water quality factors, and episodic events such as red tide, possibly stimulated by enhanced nutrient loads, undoubtedly affect the suitability of the mangrove nursery. Additional studies are needed to determine the productivity of other South Florida mangrove habitat for goliath grouper (and other species) and to define the various factors affecting density, growth, and home range including interactions of water quality, food availability, microhabitat features. In addition to reduced mangrove habitat quality in the South Florida system, direct mangrove loss is occurring. Current mangrove coverage (~2174 km2, based on FWRI data) represents a significant reduction from coverage that existed 100 yrs ago, before significant human-induced changes in the system (Johnson, 1974; Gilmore and Snedaker, 1993; Sklar and Browder, 1998; Ogden et al., 2005). The most recent losses (between 1986 and 1997) include 3.3% (from 366 to 354 km2) for FB and 28% (from 259 km2 to 186 km2) for TTI (Fig. 8). Worldwide, mangrove ecosystems have declined by ~35% since the 1980s, primarily because of mariculture in third-world countries (Mendelssohn and Mckee, 2000; Valiela et al., 2001). Our data suggest that availability of suitable mangrove habitat is the primary bottleneck to goliath grouper production, making suitable mangrove habitat not only essential, as defined by the Sustainable Fisheries Act, but critical in the sense defined by the U.S. Endangered Species Act (16 U.S.C. Section 1531 et seq.). Given that nearly 75% of Florida’s human population lives in coastal counties, that the population is expected to increase by 9 million people in the next 25 yrs (Florida Office of Eco- nomic and Demographic Research, http://edr.state.fl.us/population.htm), and that mangrove losses of 20% are expected in the next 50 yrs (J.U., unpubl. data), we contend that to be effective, any recovery plan for this species should include protection and restoration of mangrove nursery habitat, which in turn depends in large part on restoration of “defining characteristics” (Ogden et al., 2005) of the South Florida ecosystem. KOENIG ET AL.: JUVENILE GOLIATH GROUPER HABITAT 583

Figure 8. Satellite images of Florida Bay and Ten Thousand Islands mangrove coverage. Upper panels, before mangrove die-back; lower panels, after mangrove die-back. Black = mangrove, gray = land and other wetlands; white = water. Data are from Landsat TM imagery through an ISO data classification at 30-m resolution. Fish and Wildlife Research Institute (Florida Fish and Wildlife Conservation Commission), St. Petersburg, FL, USA.

Acknowledgments

We thank L. Demere (Chokoloskee, FL), J. and B. White (Chokoloskee, FL), and T. Schmidt (National Park Service) for habitat information. M. Finn (Goodland, FL), D. DeMaria (Key West, FL), T. Bevis (Florida State University), J. Brusher (NMFS, Panama City) provided technical assistance; H. Norris and N. Morton (FWC, FWRI) provided data on mangrove distribution. J. Dodrill, B. Horn, and K. Mille (FWC) provided data on goliath grouper abundance. The FWC Florida Wildlife Research Institute supported the tagging hotline. The Ocean Conser- vancy provided posters. We acknowledge grants from the NMFS Office of Protected Species, the NMFS Office of Habitat, the Curtis and Edith Munson Foundation, the National Fish and Wildlife Foundation, and The Pew Fellows Program in Marine Conservation (to F.C.C.). We also thank A. B. Thistle for formatting and editing the manuscript. This research was conducted under the guidelines of the Florida State University Animal Care and Use Commit- tee and under permits from the Florida Fish and Wildlife Conservation Commission. This is Contribution No. 1109 of the Florida State University Coastal and Marine Laboratory.

Literature Cited

Atwood, H. L., S. P. Young, and J. R. J. Tomasso. 2001. Salinity and temperature tolerance of black seabass juveniles. N. Am. J. Fish. Manage. 63: 285–288. Beck, M. W. 1995. Size-specific shelter limitation in stone crabs: a test of the demographic bottleneck hypothesis. Ecology 76: 968–980. 584 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

______, K. L. Heck, K. W. Able, D. L. Childers, D. B. Eggleston, B. M. Gillanders, B. Halpern, C. G. Hays, K. Hoshino, T. J. Minello, R. J. Orth, P. F. Sheridan, and M. Weinstein. 2001. The identification, conservation, and management of estuarine and marine nurseries for fish and invertebrates. BioScience 51: 633–641. Boyer, J. N., J. W. Fourqurean, and R. D. Jones. 1999. Seasonal and long-term trends in the water quality of Florida Bay (1989–1997). Estuaries 22: 417–430. Bullock, L. H., M. D. Murphy, M. F. Godcharles, and M. E. Mitchell. 1992. Age, growth, and reproduction of jewfish Epinephelus itajara in the eastern Gulf of Mexico. Fish. Bull. 90: 243–249. Butler, M. J., J. H. Hunt, W. F. Herrnkind, M. J. Childress, R. Bertelsen, W. Sharp, T. Mathews, J. M. Field, and H. G. Marshall. 1995. Cascading disturbances in Florida Bay, USA: cyanobacteria blooms, sponge mortality, and implications for juvenile spiny lobsters Panulirus argus. Mar. Ecol. Prog. Ser. 129: 119–125. Coleman, F. C., C. C. Koenig, and W. F. Herrnkind. 1993. Annual report: survey of Florida inshore shrimp trawling by-catch and preliminary tests of by-catch reduction devices. 1992. Florida Department of Natural Resources, Tallahassee, Florida. 28 p + appendices. ______, ______, and ______. 1994. Second annual report: survey of Florida inshore shrimp trawlers bycatch. DNR Contract #C-7779. Florida Department of Natural Resources, Florida Marine Research Institute, Tallahassee, Florida. 60 p. Crowder, L. B., S. J. Lyman, W. F. Figueria, and J. Priddy. 2000. Source-sink population dynamics and the problem of siting marine reserves. Bull. Mar. Sci. 66: 799–820. Eklund, A. M. and J. Schull. 2001. A stepwise approach to investigating the movement patterns and habitat utilization of Goliath grouper, Epinephelus itajara, using conventional tagging, acoustic telemetry, and satellite tracking. Pages 189–216 in J. R. Sibert and J. L. Nielsen, eds. Electronic tagging and tracking in marine fisheries. Kluwer, Dordrecht. Fitzhugh, G. R., C. C. Koenig, F. C. Coleman, C. B. Grimes, and W. A. Sturges. 2005. Spatial and temporal patterns in fertilization and settlement of young gag (Mycteroperca microlepis) along the West Florida Shelf. Bull. Mar. Sci. 77: 377–396. Fourqurean, J. W. and M. B. Robblee. 1999. Florida Bay: a history of recent ecological changes. Estuaries 66: 345–357. Frias-Torres, S. 2006. Habitat use of juvenile goliath grouper (Epinephelus itajara) in the Flor- ida Keys, USA. Endang. Species Res. 1: 1–6. Gilmore, R. G. and S. C. Snedaker. 1993. Mangrove forests. Pages 165–198 in W. H. Martin, S. G. Boyce, and A. C. Echternacht, eds. Biodiversity of the southeastern United States: lowland terrestrial communities. Wiley, New York. GMFMC (Gulf of Mexico Fishery Management Council). 1990. Amendment 2 to the fishery management plan for reef fish, includes regulatory impact review, regulatory flexibility analysis and environmental assessment. Gulf of Mexico Fishery Management Council, Tampa, Florida. 20 p. Faunce, C. H. and J. E. Serafy. 2006. Mangroves as fish habitat: 50 years of field studies. Mar. Ecol. Prog. Ser. 318: 1–18. Halpern, B. S. 2004a. Are mangroves a limiting resource for two coral reef fishes? Mar. Ecol. Prog. Ser. 272: 93–98. ______. 2004b. Habitat bottlenecks in stage-structured species: hermit crabs as a model system. Mar. Ecol. Prog. Ser. 276: 197–207. Heemstra, P. C. and J. E. Randall. 1993. FAO species catalogue. Vol. 16. Groupers of the world. (Family Serranidae, Subfamily Epinephelinae). An annotated and illustrated catalogue of the grouper, rockcod, hind, coral grouper and lyretail species known to date. Food and Agriculture Organization of the United Nations, Rome. Johnson, L. 1974. Beyond the fourth generation. University Presses of Florida, Gainesville. 230 p. Kapetsky, J. M. 1985. Mangroves, fisheries, and aquaculture. Pages 17–36 in FAO Fish. Rep. No. 338 (Suppl.): 87 p. (Doc FIRD/R338 Suppl.). KOENIG ET AL.: JUVENILE GOLIATH GROUPER HABITAT 585

Koenig, C. C. and F. C. Coleman. 1998. Absolute abundance and survival of juvenile gags in sea grass beds of the northeastern Gulf of Mexico. Trans. Am. Fish. Soc. 127: 44–55. Krebs, C. J. 1999. Ecological methodology. Benjamin/Cummings, Menlo Park, California. 620 p. Kruczynski, W. L. 1999. Water quality concerns in the Florida Keys: sources, effects, and solutions. Report to the FKNMS Water Quality Protection Program. Florida Keys National Marine Sanctuary, Marathon, Florida. 65 p. Laegdsgaard, P. and C. Johnson. 2001. Why do juvenile fish utilize mangrove habitats? J. Exp. Mar. Biol. Ecol. 257: 229–253. Loftus, W. F. 2000. Inventory of fishes of Everglades National Park. Fla. Sci. 63: 29–26. Mendelssohn, I. A. and K. L. Mckee. 2000. Saltmarshes and mangroves. Pages 501–536 in M. G. Barbour and W. D. Billings, eds. North American terrestrial vegetation, 2nd ed. Cam- bridge University Press, Cambridge. Nagelkerken, I., C. M. Roberts, G. van der Velde, M. Dorenbosch, M. C. van Riel, E. Cocheret del la Morinière, and P. H. Hienhuis. 2002. How important are mangroves and seagrass beds for coral-reef fish? The nursery hypothesis tested on an island scale. Mar. Ecol. Prog. Ser. 244: 299–305. NMFS (National Marine Fisheries Service). 2004. Annual Report to Congress on the status of U.S. Fisheries—2003. U.S. Department of Commerce, NOAA, National Marine Fisheries Service, Silver Spring, Maryland. 24 p. Ogden, J. C., S. M. Davis, T. K. Barnes, K. J. Jacobs, and J. H. Gentile. 2005. Total system conceptual ecological model. Wetlands 25: 955–979. Porch, C. E. and A. M. Eklund. 2004. Standardized visual counts of goliath grouper off south Florida and their possible use as indices of abundance. Gulf Mexico Sci. 2: 155–163. Robins, C. R., G. C. Ray, J. Douglass, and R. Freund. 1986. A field guide to Atlantic coast fishes of North America. Houghton Mifflin, Boston. 354 p. Roff, D. A. 1992. The evolution of life histories, theory and analysis. Chapman & Hall, New York. 535 p. Sadovy, Y. and A. M. Eklund. 1999. Synopsis of biological information on the Nassau grouper, Epinephelus striatus (Bloch 1792), and the jewfish, E. itajara (Lichtenstein 1822). NOAA Tech. Rep. NMFS 146. 65 p. Seber, G. A. F. 1982. The estimation of animal abundance, 2nd ed. Griffin, London. 654 p. Sklar, F. H. and J. A. Browder. 1998. Coastal environmental impacts brought about by alterations to freshwater flow in the Gulf of Mexico. Environ. Manage. 23: 347–358. Thayer, G. W., A. B. Powell, and D. E. Hoss. 1999. Composition of larval, juvenile, and small adult fishes relative to changes in environmental conditions in Florida Bay. Estuaries 22: 518–533. Thorrold, S. R., C. M. Jones, P. K. Swart, and T. E. Targett. 1998. Accurate classification of juvenile weakfish Cynoscion regalis to estuarine nursery areas based on chemical signatures in otoliths. Mar. Ecol. Prog. Ser. 173: 253–265. Valiela, I., J. L. Bowen, and J. K. York. 2001. Mangrove forests: one of the world’s threatened major tropical environments. BioScience 51: 807–815. White, G. C. and R. A. Garrott. 1990. Analysis of wildlife radio-tracking data. Academic Press, New York. 383 p.

Addresses: (C.C.K., F.C.C.) Florida State University Coastal and Marine Laboratory, St. Te- resa, Florida 32858-2702. (A.M.E., J.S.) NOAA-Fisheries, Southeast Fisheries Science Center, Division of Protected Resources and Biodiversity, 75 Virginia Beach Drive, Miami, Florida 33149. (J.U.) Department of Geography, Clippinger Lab 122, Ohio University, Athens, Ohio 45701. Corresponding Author: (C.C.K.): E-mail: . BULLETIN OF MARINE SCIENCE, 80(3): 587–594, 2007

NOTE

Activity patterns of three juvenile goliath grouper, Epinephelus itajara, in a mangrove nursery

Sarah Frias-Torres, Pedro Barroso, Anne-Marie Eklund, Jennifer Schull, and Joseph E. Serafy

Goliath grouper, Epinephelus itajara (Lichtenstein, 1822), was a protected species historically distributed throughout tropical and subtropical coastal waters of the western and eastern Atlantic Ocean. It also occurs at similar latitudes in eastern Pacific coastal waters H( eemstra and Randall, 1993). The species—listed as critically endangered by the World Conservation Union (Hudson and Mace, 1996)—has been protected through a total fishing ban in U.S. federal and state waters since 1990 (Sadovy and Eklund, 1999) and in Puerto Rico and U.S. Virgin Islands waters since 2004 (NMFS, 2006). As one of the few grouper species that shows a high affinity for mangrove-dominated areas, E. itajara typically spends its first 5–8 yrs in that nearshore habitat before migrating to adult habitats (i.e., coral reefs, rock ledges, isolated patch reefs, and artificial structures) (Sadovy and Eklund, 1999). Visual underwater surveys in the Florida Keys reveal that juvenile E. itajara prefer well-developed fringing red mangrove Rhizophora mangle Linnaeus shorelines with high structural complexity (undercuts, extensive overhangs) and soft sediment (Frias-Torres, 2006). To date, knowledge of the movement of goliath grouper juveniles within their mangrove nursery areas is based mainly on conventional tagging studies. However, these have provided only limited insight into the fine-scale activity patterns, especially those occurring on a daily or subdaily basis (Eklund and Schull, 2001). In the present study, we used electronic acoustic tags and stationary hydrophone receiver stations to examine the activity patterns of three tagged juveniles inhabiting the fringing red mangrove shorelines of Ten Thousand Islands, southwest Florida, U.S. (Fig. 1). Bullock and Smith (1991) described this region as the historical center of abundance for E. itajara in U.S. waters. Our primary objective was to examine the extent to which juvenile activity was associated with time of day and tidal stage. Previous investigations demonstrated a high degree of juvenile site fidelity as well as resilience to, and rapid recovery after, handling (Eklund and Schull, 2001). To test our ability to detect movement at fine timescales over periods of weeks to months, as well as cues for fish movement, we examined the relationship between fish presence- absence, tidal cycle, and time of day in juvenile goliath grouper. Our results provide insight into multiple habitat use by this species and are important to consider when designing and implementing population abundance surveys.

Methods

In October 2000, three goliath grouper (Table 1) were captured and tagged with acoustic transmitters in Rabbit Key Pass (25°78´N, 81°36´W), Ten Thousand Islands, southwest Florida (Fig. 1). Temperature ranged from 18 to 31.7 °C, salinity from 1.8 to 40.6, dissolved oxygen from 2.4 to 7.7 mg L–1, water depth from 2 to 10 m, and tidal amplitudes ranged up to 2 m (Eklund and Schull, 2001; SFWMD, 2006). Acoustic transmitters (VEMCO®, 16 mm diameter,

Table 1. Juvenile goliath grouper size, capture, and acoustic monitoring details. Abbreviations: ID = fish identification number; TL = total length.

ID Capture date TL (mm) at Capture Dataset: Dataset: Days of capture site start date end date data 11 October 5, 2000 410 1 December 12, 2000 August 27, 2002 615 30 October 4, 2000 660 1 December 12, 2000 December 31, 2000 20 31 October 4, 2000 840 2 December 12, 2000 August 7, 2001 237

92 mm long, 16 g in water) were implanted within the intraperitoneal cavity; each tag emitted a unique signal every 2 min at 69 kHz. Manufacturer specifications estimate a battery life of 643 d. Two Vemco VR1 data-logging acoustic receivers were moored within 20 m of each capture location. Range tests demonstrated an approximate 100 m radius range for each receiver, and there was overlap between both receivers’ range (Fig. 1). A prior conventional tagging and manual tracking experiment (Eklund and Schull, 2001) revealed that juveniles at this study site were frequently recaptured at their original tagging locations, and therefore, would be readily detected by the receivers. Initially, we attached additional pingers to each receiver for equipment relocation purposes, but later documented interference between relocation pingers and acoustic fish tags. We subsequently removed the relocation pingers in December 2000, accounting for the gap between capture dates and dataset start dates. Due to non-normality and lack of independence of errors (i.e., autocorrelation), variation in fish detection in relation to tidal stage and time of day was examined using a nonparametric analysis of variance equivalent as well as spectrum (Fourier) analysis. In the former, Scheirer- Ray-Hare extension of the Kruskal-Wallis tests (followed by Mann-Whitney U-tests for post- hoc comparisons) were used (Sokal and Rohlf, 1995) whereby the dependent variable was the proportion of (positive) fish detections (of 32 possible per hour) and the independent variables were tidal stage (four levels), time-of-day (four levels), and the interaction term. The four tidal stages were low, high, flood, and ebb; the former stages were defined as 2-hr periods of least tidal flow. The four times of day were day, night, dawn, and dusk; the latter were defined as 3- hr intervals beginning 1.5 hrs before and ending 1.5 hrs after sunrise or sunset, respectively. Samples sizes for each case and fish are shown in Table 2. To reveal cyclical components of each fish detection time series, single spectrum analyses using the fast Fourier transform

Figure 1. Map of study area: (A) southern Florida, USA; (B) Ten Thousand Islands; and (C) Rabbit Key Pass. Shown are fish capture sites (1 and 2), and hydrophone receiver locations (A and B). NOTES 589

Table 2. Sample sizes (number of hourly presence proportions) for each data point in Figure 2 and used in the nonparametric analyses of variance (Scheirer-Ray-Hare extension of the Kruskal- Wallis test) as well as post hoc mean comparisons (Mann Whitney U-tests). Times of day indicated as dawn, day, dusk, and night.

Tide Low Flood High Ebb Fish 11 Dawn 155 264 215 614 Day 656 1,151 577 1,328 Dusk 137 250 243 621 Night 662 1,057 575 1,495 Fish 30 Dawn 6 8 4 39 Day 28 49 15 41 Dusk 6 14 17 20 Night 34 54 38 84 Fish 31 Dawn 90 172 126 326 Day 391 714 381 748 Dusk 79 115 134 386 Night 358 558 279 849 were performed. These results were examined via periodograms (i.e., plots of spectral density vs period).

Results

Tagged fish measured 410–840 mm total length (TL) and were detected for 20– 615 d after interfering pingers were removed from the listening stations (Table 1). For the longest detected fish (fish 11), signs of tag failure were evident during the last 32% of its record. Specifically, near linear reduction in detections began on day 416 and continued thereafter. Therefore, data after day 416 were omitted from analyses. Over the entire sampling period, fifish sh 11 was only detected at hydrophone A (cap-(capture site 1); fish 30 was also detected at hydrophone A, and for 20 hrs at hydrophone B; fish 31 was detected at hydrophone B (capture site 2), and for 5 hrs at hydrophone A. Although cross-detections were rare, overall the data indicate that hydrophone locations resulted in range detection overlap, and were appropriate to detect fish movement within the site. Most absences were < 24 hrs, and the majority of movements out of the receiver’s range were within 12 hrs (98% for fish 11, 89% in fish 30, and 70% in fish 31). For all three fish, the time of day × tide interaction terms were highly significant (Table 3). While activity levels differed by fish and time of day, for the most part, activity patterns were similar in that the highest proportion of detections occurred during the low and ebb tidal stages and the lowest during flood and high tides (Fig. 2). Spectrum analysis tended to corroborate that fish activity was synchronized with the tidal cycle. For all three fish, spectral density peaks were evident at 12 and 24 hr periods, which roughly correspond to the semidiurnal and diurnal tidal periods (Fig. 3). 590 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Table 3. Results of nonparametric analyses of variance (i.e., Scheirer-Ray-Hare tests) whereby the proportion of positive fish detections (presence) was the dependent variable and time of day (TOD) 2 and tidal stage (Tide) were the independent variables. Critical χ 9, 0.05 = 16.919.

Source SS SS/MStotal df P Fish 11 TOD 217,221,944.8 26.51291 3 < 0.001 Tide 9,767,117,728 1,192.12 3 < 0.001 TOD × Tide 804,418,809.5 98.1829 9 < 0.001 Fish 30 TOD 176,345.468 10.26977 3 0.0164 Tide 446,900.487 26.02599 3 < 0.001 TOD × Tide 414,874.14 24.16088 9 0.004 Fish 31 TOD 299,571,114.9 168.3022 3 < 0.001 Tide 158,628,691.2 89.11924 3 < 0.001 TOD × Tide 83,915,063.55 47.14435 9 < 0.001

Discussion

Site Fidelity.—Juvenile goliath grouper showed high site-fidelity. Absence ap-appeared size-related, with the longest absence periods in the largest of the three fish (fish 31, 840 mm TL at capture). The relation between longer absences and increased fish size may be due to higher metabolic demands which require greater time foraging, in concert with a reduction in vulnerability to predators (Levin and Grimes, 2002). Several grouper species show site-fidelity as juveniles (Mycteroperca microlepis (Goode and Bean, 1879) Koenig and Coleman, 1998; Epinephelus marginatus (Lowe, 1834) Lembo et al., 1999; E. itajara and Epinephelus striatus (Bloch, 1792) Sadovy and Eklund, 1999). Low water levels may limit fish movements. At low tide juvenile goliath grouper may be restricted to deeper sites, such as eroded ledges under the mangroves (undercuts) or other deep microhabitats until the next tidal rise. In this study, the sites where the fish were monitored had deep undercuts and soft sediment, a preferred habitat for juvenile E. itajara (Frias-Torres, 2006). If a juvenile fails to return before low tide, it may effectively be separated from its refuge. Fish 11, the smallest of the three fish, was detected for almost 2 yrs, although transmitter interference and apparent malfunction reduced the time series to about 13.5 mo. The signal from fish 30 barely lasted one month after the start date of data collection. Whether its acoustic tag failed, the fish moved to a different location, or it was lost due to predation or removal is unknown. The signal from fish 31 lasted 10 mo after capture and yielded 7 mo of useful data. As the largest of the fish, it is possible that the lack of detection of fish 31 later in the study reflected the onset of the ontogenetic migration. Goliath grouper show ontogenetic habitat migrations and move from the nursery mangrove habitat to the adult habitat (reefs, rock ledges, coral reefs) as they reach sexual maturity, within the 1100–1200 mm TL range (Sadovy and Eklund, 1999; Eklund and Schull, 2001). Assuming fish 31 was growing normally, it would have attained about 1000 mm TL by the end of its detection record (Bullock et al., 1992). Daylight/Tidal Phase.—Both the Scheirer-Ray-Hare tests and spectrum analyses suggested fish activity was correlated mainly with tidal cycle with a tendency to NOTES 591

Figure 2. Presence (mean ± SE) of tagged juvenile Epinephelus itajara juveniles for each time of day and tidal stage in mangrove habitats of Rabbit Key Pass. Means sharing the same lower case letters are not statistically different (P > 0.05); panels without letters contain means that are not significantly different. Abbreviations: L = low tide, F = flood tide, H = high tide, E = ebb tide. move out of the receiver’s range on flood and high tides, and remain within range on ebb and low tides. While there were a few exceptions to this general pattern, its consistency suggests high and increasing water levels prompt fish movement regardless of time-of-day. Without a larger array of receivers, we cannot determine where juveniles go during their tidal excursions. Juvenile gag (M. microlepis), another grouper species that utilizes estuarine habitats as nursery, appears to forage in seagrass meadows at high tide (Ross and Moser, 1995) and use 2-m deep channels that meander through the much shallower marsh-seagrass landscape when the tide ebbs. Twilight (dusk) migrations also occur in mangrove habitats for a variety of families, such as Haemulidae, Lutjanidae, Acanthuridae, Scaridae, Chaetodontidae, and Pomacentridae (Rooker and Dennis, 1991; Nagelkerken et al., 2000). Although fish movement was correlated with tide, in this study we could not ascertain what type of activity was associated with such movement, but foraging seems likely. Excursions during high water levels allow for expanded exploration of foraging 592 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Figure 3. Juvenile Epinephelus itajara activity and tidal cycle. Plots of spectral density vs period (periodograms) based on spectrum analyses using the fast Fourier transform.

grounds, including those otherwise too shallow to access. Potthoff and Allen (2003) found that pinfish L[ agodon rhomboides (Linnaeus, 1766)] which expanded their foraging range with rising water levels consumed more shrimp than those that did not. Crustaceans, mostly shrimps and xanthid crabs (Odum, 1971; Bullock and Smith, 1991) are important in the diet of juvenile goliath grouper. Several observations support the hypothesis that juvenile goliath grouper might move during flood and high tides for foraging purposes. Odum et al. (1982), proposed that juvenile goliath grouper invaded tidal streams primarily to feed. During underwater visual surveys in the mangroves of the Florida Keys, juvenile goliath grouper were found hovering over the bottom at sites recently flooded by high tides (Frias-Torres, 2006). In conclusion, we provide empirical evidence suggesting that short-term movement patterns of three juvenile E. itajara were strongly related to tide. Future population monitoring of juvenile E. itajara abundance in natural mangrove habitats should take into consideration the movement patterns indicated here, and adjust sampling efforts (e.g., among shoreline and non-shoreline habitats) accordingly. Comparing NOTES 593 stomach contents of juvenile E. itajara using non-lethal methods (Eklund and Schull, 2001), at various tidal stages, might provide further evidence of foraging activity during tidal migrations.

Acknowledgments

We thank J. A. Bohnsack, C. Koenig, and two anonymous reviewers who provided comments that improved the manuscript; and T. Bevis, J. Contillo, L. Demere, C. Koenig, D. Mc- Clellan, J. White, and B. White who helped in fieldwork and data logistics. Thanks also to J. Luo for help with statistics. This study was funded by a National Research Council Post- doctoral Fellowship to SFT, a European Union Scholarship to PB, the Office of Protected Resources, and the Marine Fisheries Initiative from the Southeast Fisheries Science Center, National Oceanic and Atmospheric Administration. The National Marine Fisheries Service, National Oceanic and Atmospheric Administration does not endorse any proprietary product mentioned here.

Literature Cited

Bullock L. H. and G. B. Smith. 1991. Seabasses (Pisces: Serranidae). Memoirs of the Hourglass Cruises. Florida Marine Research Institute, St. Petersburg, Florida. 243 p. ______, M. D. Murphy, M. F. Godcharles, and M. E. Mitchell. 1992. Age, growth and reproduction of jewfish Epinephelus itajara in the eastern Gulf of Mexico. Fish. Bull. 90: 243–249. Eklund, A. M. and J. Schull. 2001. A stepwise approach to investigate the movement patterns and habitat utilization of goliath grouper, Epinephelus itajara, using conventional tagging, acoustic telemetry and satellite tracking. Pages 189–216 in J. R. Sibert and J. L. Nielsen, eds. Electronic tagging and tracking in marine fisheries. Springer-Verlag, New York. Frias-Torres, S. 2006. Habitat use of juvenile goliath grouper Epinephelus itajara in the Florida Keys, USA. Endangered Species Research 1: 1–6. Heemstra, P. C. and J. E. Randall. 1993. FAO Species Catalogue. Groupers of the World (Fam- ily Serranidae, Subfamily Epinephelinae). An annotated and illustrated catalogue of the grouper, rockcod, hind, coral grouper and lyretail species known to date. FAO Fisheries Synopsis 16 (125), 382 p. Hudson, E. J. and G. M. Mace. 1996. Marine fish and the IUCN Red List of Threatened Ani- mals. Pages 1–26 in E. J. Hudson and G. M. Mace, eds. Report of the IUCN-WWF Work- shop at the Zoological Society of London, April 29–May 1, 1996. Zoological Society of London, London. Koenig, C. C. and F. C. Coleman. 1998. Absolute abundance and survival of juvenile gags in seagrass beds of the northeastern Gulf of Mexico. Trans. Am. Fish. Soc. 127: 44–55. Lembo, G., I. A. Fleming, F. Ǿkland, P. Carbonara, and M. T. Spedicato. 1999. Homing behavior and site fidelity of Epinephelus marginatus (Lowe, 1834) around the island of Ustica: Pre- liminary results from a telemetry study. Biologia Marina Mediterranea 6: 90–99. Levin, P. S. and C. B. Grimes. 2002. Reef fish ecology and grouper conservation and management. In Peter F. Sale, ed. Coral reef fishes–dynamics and diversity in a complex ecosystem. Academic Press. San Diego. 549 p. Nagelkerken, I., M. Dorenbosch, W. C. E. P. Verberk, E. Cocheret de la Morinière, and G. van der Velde. 2000. Day-night shifts of fishes between shallow-water biotopes of a Caribbean bay, with emphasis on the nocturnal feeding of Haemulidae and Lutjanidae. Mar. Ecol. Prog. Ser. 194: 55–64. NMFS. 2006. Status report on the continental United States distinct population segment of the goliath grouper (Epinephelus itajara). January 12, 2006. 49 p. [online report]. National Ma- 594 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

rine Fisheries Service, Southeast Regional Office, St. Petersburg, FL. Available from http:// www.sefsc.noaa.gov/sedar Odum, W. E. 1971. Pathways of energy flow in a southern Florida estuary. Univ. Miami Sea Grant Progr. Sea Grant Tech. Bull. 7. 162 p. ______, C. C. McIvor, and T. J. Smith. 1982. The ecology of the mangroves of South Florida: A community profile. Fish and Wildlife Service, Office of Biological Services, Washington DC, FWS/OBS-81/24. 144 p. Pothoff,M. T. and D. M. Allen. 2003. Site-fidelity, home range, and tidal migrations of juvenile pinfish,L agodon rhomboides, in salt marsh creeks. Environ. Biol. Fish. 67: 231–240. Rooker, J. R. and G. D. Dennis. 1991. Diel, lunar and seasonal changes in a mangrove fish assemblage of southwestern Puerto Rico. Bull. Mar. Sci. 49: 684–698. Ross, S. W. and M. L. Moser. 1995. Life-history of juvenile gag, Mycteroperca microlepis, in North Carolina estuaries. Bull. Mar. Sci. 56: 222–237. Sadovy, Y and A. M. Eklund. 1999. Synopsis of biological information on the Nassau grouper, Epinephelus striatus, (Bloch 1792), and the jewfish, E. itajara (Lichtenstein 1822). NOAA Technical Report, NMFS 146, and FAO Fisheries Synopsis 157. 65 p. SFWMD. 2006. South Florida Water Management District [Internet], West Palm Beach, Flor- ida, USA: Water Quality Monitoring Database: c 2006. Available from: http://www.sfwmd. gov/org/ema/envmon/wqm. Sokal, R. R. and F. J. Rohlf. 1995. Biometry. Third Edition. W. H. Freeman and Company, New York. 859 p.

Addresses: (S.F.T.) Cooperative Institute for Marine and Atmospheric Studies (RSMAS/UM), NOAA-NMFS Southeast Fisheries Science Center, 75 Virginia Beach Drive, Miami, Florida, 33149-1003. (P.B.) Faculdade de Ciencias da Universidade de Lisboa, Edificio do Campo Grande, C1, Lisboa, 1149-016, Portugal. (A.M.E, J.S., J.E.S.) NOAA-NMFS Southeast Fisher- ies Science Center, 75 Virginia Beach Drive, Miami, Florida, 33149-1003. Corresponding Author: (S.F.T.) E-mail: . BULLETIN OF MARINE SCIENCE, 80(3): 595–607, 2007 FIRST INTERNATIONAL SYMPOSIUM ON MANGROVES AS FISH HABITAT KEYNOTE ADDRESS ARE NON-ESTUARINE MANGROVES CONNECTED TO CORAL REEFS THROUGH FISH MIGRATION?

Ivan Nagelkerken

ABSTRACT Mangroves are an important fish habitat, but little is known of their nursery function and connectivity to other habitats such as coral reefs. Here, the present status of knowledge on connectivity between non-estuarine mangroves and coral reefs by postlarval coral reef fishes is reviewed. Only since the year 2000 has more indirect evidence been obtained for such connectivity, largely based on studies: (1) quantifying juvenile/adult fish abundances in these habitats to deduce ontogenetic migrations, (2) investigating the effect of absence of mangroves on reef fish assemblages, (3) investigating the effect of mangrove forest size on reef fish abundances, and (4) investigating the effect of distance from mangroves on reef fish abundances. Almost all studies have been done in the Caribbean, and they are practically absent for the much larger Indo-Pacific region. So far, it appears that coral reef fish species do not show an obligate dependence on mangroves as a juvenile habitat, except perhaps for the vulnerable Caribbean parrotfish speciesScarus guacamaia Cuvier, 1829. Six Caribbean species of Haemulidae and Lutjanidae show high dependence on mangroves/seagrass beds as juvenile habitats, and may be the most vulnerable to loss of these habitats. A study on otolith microchemistry has provided some evidence for one species that mangroves may indeed contribute to coral reef fish populations.

Mangroves are among the few terrestrial tropical plant species which can withstand continuous contact with seawater. Not all mangrove species grow along the shoreline, and a clear zonation of species from the shore inwards is often present. Species typically present along the shoreline and growing in direct contact with seawater or subject to daily tidal influences of marine water are mostly Rhizophora mangle Lin- naeus, 1753 and sometimes Avicennia germinans (Linnaeus) Stearn, 1958 in the Ca- ribbean, and Sonneratia alba J. Smith 1819, Rhizophora mucronata Lamarck, 1804 and Avicennia marina (Forsskål) Vierhapper, 1907 in the Indo-Pacific. The mangrove surface area subject to (periodic) salt water influence can range from a few square meters (in fringing mangroves) to several hectares (in large estuarine areas). Mangroves can harbor a variety of fish species, depending on their setting within the coastal landscape and the influence of marine waters and rivers (i.e., influence of salinity). Species include freshwater, estuarine, and marine species, which visit the mangroves for spawning, foraging, shelter, or as a nursery habitat (Blaber, 2000). Mangroves are ideal habitats for many fish species because they provide a high abundance of food, and shelter against predation (Nagelkerken et al., 2000a; Laegdsgaard and Johnson, 2001; Cocheret de la Morinière et al., 2004; Verweij et al., 2006). The latter is the result of the high structural complexity of the mangrove prop-roots which can be used as shelter, the often high turbidity of the water, the dark mangrove environment which makes predation on fish less successful, and the distance away from coral reefs or offshore areas where densities of the largest fish predators are often highest (Shulman, 1985; Parrish, 1989; Manson et al., 2005). The high abundance of juvenile fish (and prawns) in mangroves has been used for several decades as an argument that mangroves serve as important nursery areas. However, hardly any proof exists of true migration from mangrove nurseries to the

Table 1. Search results (on “topic”) from ISI Web of Science® (Thomson Corporation™) (15 March 2006). Asterisk (*) denotes a wildcard to specify zero or more alphanumeric characters (e.g., “fish*” includes the search terms “fish,” “fishes,” “fisheries,” etc.).

Search term # hits Reef* and Fish* 4,860 Mangrove* and Fish* 609 Mangrove* and Reef* and Fish* 147 Mangrove* and Reef* and Fish* and Link* 16 Mangrove* and Reef* and Fish* and Connect* 10 adult habitats on coral reefs or offshore areas, and mangroves may merely function as sinks for juvenile fish. According to Beck et al. (2001), habitats only serve as true nurseries when their contribution to adult fish populations is larger, on average, than that of other juvenile habitats. This should be supported by a higher density, growth, or survival of animals in the nursery habitat, and by ontogenetic migration between the nursery habitat and the adult habitat (Beck et al., 2001). By far the majority of studies have focused on fish densities in a variety of shallowwater habitats, whereas studies on growth, survival, and movement are largely lacking (Beck et al., 2001; Heck et al., 2003; Sheridan and Hays, 2003; Faunce and Serafy, 2006). Although mangrove fish communities have been studied for several decades, little is known of their nursery function and their connectivity with surrounding habitats (Table 1). This is firstly the result of many studies focusing on only the mangrove habitat (e.g., prop-root system, creeks, channels) and not on other habitats as well. Secondly, in cases where multiple habitats have been surveyed, the fish communities of mangroves and their adjoining habitats mostly have been studied with different methods, making comparisons difficult. Thirdly, in very few cases did studies of -(ju venile) mangrove fish communities include the adult fish habitat (i.e., offshore areas or coral reefs). Fourthly, in most cases only total densities of species are reported, without distinguishing between juvenile and adult fish so as to determine the distribution of size classes in different habitats. As a result, relatively little is known about connectivity between mangroves and other habitats, and only 12 studies could be used to investigate ontogenetic habitat utilization of juvenile and adult fish in mangroves and coral reefs (Table 2). The focus of this review is on the present status of knowledge on connectivity between non-estuarine mangroves and coral reefs for postlarval coral reef fishes. The reason for this selection is based on the fact that most studies of mangrove fish communities have been done in estuaries. This is probably because many of the largest mangrove ecosystems are located there, many of which within a reasonable distance of marine laboratories. However, there are probably hundreds to thousands of (fairly isolated) islands around the world which harbor marine mangrove systems (i.e., not influenced by rivers) close to coral reefs. Very little is known of these systems, and although the size of the mostly small fringing mangrove stands on these islands is not comparable to the large mangrove forests in estuaries, on a local scale they possibly serve as important nursery or juvenile habitats that sustain and/or enhance the islands’ coral reef fish populations. This review will not deal directly with the nursery function of mangroves, since several reviews on the potential mechanisms of the nursery function of (estuarine) mangroves and other habitats already exist (e.g., Parrish, 1989; Sheridan and Hays, NAGELKERKEN: CONNECTIVITY BETWEEN MANGROVES AND CORAL REEFS 597

Table 2. Overview of studies including at least the mangrove and coral reef habitat, and distinguishing between juvenile and adult densities (or between size-classes), so as to deduce ontogenetic migration between habitats at species level (first section), followed by other studies related to mangrove–reef connectivity. M = mangroves, S = seagrass beds, SB = soft bottom, HB = hard bottom, N = subtidal notches in fossil coral reef terrace, C = channel, CR = coral reefs.

Reference Habitats studied Location Geographic area Mangrove-reef ontogenetic studies Thollot, 1992 M, CR New Caledonia Indo-Pacific Rooker, 1995 M, CR Puerto Rico Caribbean Nagelkerken et al., 2000b M, S, SB, C, CR Curaçao Caribbean Nagelkerken et al., 2000c M, S, CR Bonaire Caribbean Cocheret de la Morinière et al., 2002 M, S, CR Curaçao Caribbean Nagelkerken and van der Velde, 2002 M, S, SB, C, CR Curaçao Caribbean Christensen et al., 2003 M, S, CR Puerto Rico Caribbean Nagelkerken and van der Velde, 2003 M, S, SB, C, CR Bonaire/Curaçao Caribbean Serafy et al., 2003 M, CR Florida, USA Caribbean Eggleston et al., 2004 M, S, HB, C, CR Florida, USA Caribbean Mumby et al., 2004 M, S, CR Belize/Mexico Caribbean Dorenbosch et al., 2006 M, S, N, CR Aruba Caribbean

Mangrove-absence studies Ogden et al., 1985 CR Sombrero Island Caribbean Miller and Gerstner, 2002 CR Navassa Island Caribbean Mumby et al., 2004 CR Belize/Mexico Caribbean Dennis et al., 2005 CR Mona Island Caribbean Dorenbosch et al., 2005 CR Tanzania Indo-Pacific

Distance-to-mangrove studies Thollot, 1992 CR New Caledonia Indo-Pacific Nagelkerken et al., 2000b CR Curaçao Caribbean Appeldoorn et al., 2003 CR Old Providence/ Caribbean Santa Catalina Halpern, 2004 CR Virgin Islands Caribbean Dorenbosch et al., 2004 CR Curaçao Caribbean Dorenbosch et al., 2005 CR Tanzania Indo-Pacific Dorenbosch et al., 2006 CR Aruba Caribbean Dorenbosch et al., 2007 CR Aruba Caribbean

Mangrove-size studies Halpern, 2004 CR Virgin Islands Caribbean Mumby et al., 2004 CR Belize/Mexico Caribbean

Otolith studies Chittaro et al., 2004 M, CR Belize Caribbean

2003; Manson et al., 2005; Adams et al., 2006), while studies are still lacking which directly test the factors related to nursery function as proposed by Beck et al. (2001). Instead, studies providing (indirect) evidence for the existence of any connectivity between non-estuarine mangroves and coral reefs by fishes will be reviewed, whether or not the mangroves serve as true nurseries. 598 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Mangrove–Reef Connectivity Based on Size-frequency Data

Non-estuarine mangroves mostly co-occur with seagrass beds; hence, there are few studies which have been done in seascapes where only mangroves and adjacent coral reefs are present. The presence of seagrass beds in this seascape makes the habitat interactions more complex, since ontogenetic migrations can occur between mangroves and seagrass beds before fish potentially move to the coral reef (e.g., Rooker and Dennis, 1991; Nagelkerken et al., 2000c). Also, the combined presence of mangroves and seagrass beds may cause differences (i.e., enhancement) in fish density, species richness, and species presence (Robertson and Blaber, 1992; Nagelkerken et al., 2001). As a result, relatively little is known of the connectivity between solely mangroves and coral reefs. In situations where seagrasses are absent, some indirect evidence has been obtained for mangrove–reef connectivity. In New Caledonia, Thollot and Kulbicki (1988) found an overlap of 13 species between the two habitats, and suggested that Epinephelus caeruleopunctatus (Bloch, 1790) and Epinephelus howlandi (Günther, 1873) utilize the mangroves as a nursery habitat. However, due to the lack of density and size-frequency data in the study, this conclusion is difficult to evaluate. In a more detailed study in the same bay, Thollot (1992) found an overlap of 43 species between the two habitat types, of which 13 species were found as juveniles in the mangroves and as adults on the coral reefs. These included species of Lutjanidae (Lutjanus argentimaculatus (Forsskål, 1775), Lutjanus fulviflamma (Forsskål, 1775), Lutjanus fulvus (Forester in Bloch and Schneider, 1801), Lutjanus russelli (Bleeker, 1849)), Acanthuridae (Acanthurus blochii Valenciennes, 1835, Acanthurus dussumieri Va- lenciennes, 1835), Carangidae (Caranx ignobilis (Forsskål, 1775), Caranx melampygus (Cuvier, 1833), Caranx papuensis Alleyne and MacLeay, 1877), and species from four other families (Hyporhamphus dussumieri (Valenciennes, 1846), Lethrinus harak (Forsskål, 1775), Parupeneus indicus (Shaw, 1803), Siganus lineatus Valenciennes, 1835). However, juveniles of several of the above species were also found in other habitats (i.e., soft bottom, seagrass beds, and coral reefs), and the adults of these species also occurred in areas lacking extensive mangrove habitats (Table 3), suggesting that for most species mangroves are not obligate juvenile habitats of New Caledonia. In Papua New Guinea, coral reefs located adjacent to mangroves did not show a higher density of adult fish than isolated coral reefs, suggesting that the mangroves did not contribute significantly to the adult reef fish population (Birkeland and Amesbury, 1988). However, only total fish density for all species pooled was provided, masking possible effects at the species level. Off southeastern Florida, Serafy et al. (2003) found juvenile as well as adult Hae- mulon sciurus (Shaw, 1803), Haemulon parra (Desmarest, 1823), Lutjanus apodus (Walbaum, 1792), Lutjanus griseus (Linnaeus, 1758), and Sphyraena barracuda (Walbaum, 1792) in mangroves and on adjacent coral reefs showing that mangroves are not exclusively juvenile habitats. For two species (L. apodus and S. barracuda) a clear mangrove-to-reef ontogenetic migration was inferred from length-frequency data, suggesting enhancement of local reefs. This is supported by Mumby et al. (2004), who found a significant increase in fish biomass (Haemulon flavolineatum (Desmarest, 1823), Haemulon plumierii (Lacépède, 1801), H. sciurus, L. apodus, Ocy- urus chrysurus (Bloch, 1791), Scarus iserti Bloch, 1790) on shallow fore-reefs and deeper Montastrea spp. reefs located close to mangroves compared to reefs isolated NAGELKERKEN: CONNECTIVITY BETWEEN MANGROVES AND CORAL REEFS 599

Table 3. Habitat utilization by juvenile fishes for species observed in non-estuarine mangroves, indicated for each study separately (+ = species common: density > 10% of highest density in any habitat, P = present (no density data provided), L = low density (< 10% of highest density in any habitat), - = absent). For the Indo-Pacific (New Caledonia, Tanzania) data are from: Thollot (1992) and Dorenbosch et al. (2005); for the Caribbean (Aruba, Bahamas, Belize, Bonaire, Curaçao, Florida, Mexico, Puerto Rico) data are from: Rooker (1995), Nagelkerken et al. (2000b,c), Cocheret de la Morinière et al. (2002), Nagelkerken and van der Velde (2002), Serafy et al. (2003), Eggleston et al. (2004), Mumby et al. (2004), Chittaro et al. (2005) and Dorenbosch et al. (2006). For habitat abbreviations see Table 2. EA = early juvenile habitat (< 5 cm FL, habitat contribution > 20% of total density of all habitats), ? = unknown; data are from Nagelkerken et al. (2000b) and Nagelkerken and van der Velde (2003). Isolat. = occurrence of adult fish on isolated reefs far away from mangroves/ seagrass beds (P = species present–no density data provided, = = similar fish density as reefs close to mangroves/seagrass beds, R = reduced fish density compared to reefs close to mangroves/seagrass beds, - = absent); data are from Ogden et al. (1985), Nagelkerken et al. (2002), Miller and Gerstner (2002), Mumby et al. (2004), and Dennis et al. (2005). Dist. = decrease of reef fish density with distance from mangrove/seagrass beds (+ = decrease present, ± = potential decrease, - = decrease absent); data are from Thollot (1992), Nagelkerken et al. (2000b), Halpern (2004), and Dorenbosch et al. (2004, 2005, 2007).

Juvenile habitat Species EA M S SB HB C CR Isolat. Dist. Indo-Pacific Acanthurus blochii P P P ± Acanthurus dussumieri P P ± Caranx ignobilis P P - Caranx melampygus P P - Caranx papuensis P P - Hyporhamphus dussumieri P P + Lethrinus harak P P R +- Lutjanus argentimaculatus P - +± Lutjanus fulviflamma P P R +- Lutjanus fulvus P P P - Lutjanus russelli P - Parupeneus indicus P P - +- Siganus lineatus P -

Caribbean Acanthurus chirurgus S, CR ++L +++ - + ++L =P- +- Chaetodon capistratus M, S, C ++ ++ L + +- =P- +- Gerres cinereus M ++ LL L L- -- R-- - Haemulon chrysargyreum C, CR +- -- - + +- R-- - Haemulon flavolineatum M, S, CR ++++ ++++ L L ++LL =PR-- -- Haemulon parra M ++ L - - +- --- Haemulon plumierii S + + - + - R---- Haemulon sciurus M, S +++++++ +++++L L L L- +LLLL-- RRR-- +- Hypoplectrus unicolor ? + L - + - R- Lutjanus apodus M ++++++ +LLL L LL ++LLL PRRR ++-- Lutjanus griseus M +++++ +LLL L + +L +LL-- R-- Lutjanus mahogoni S, CR ++- LL- - L +L- =P- ++- Mulloidichthys martinicus ? + L - - + PP - Ocyurus chrysurus S, C +++L ++++ L + ++++ RRRR- ++- Scarus guacamaia ? +++ L-- L - --- R--- - Scarus iserti S, C +++ +++ L + +++ PPRR -- Sparisoma chrysopterum C + + - + + =PP +- Sparisoma viride S, CR +L +L - + ++ PP - Sphyraena barracuda M ++++ +LL L L- +LL- PR- +- 600 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

from mangroves. However, these species were not completely absent on isolated reefs, possibly due to the presence of extensive seagrass beds that could have been used as an alternative juvenile habitat (e.g., Nagelkerken et al., 2000b). In the Carib- bean area, therefore, it seems that mangroves alone are not obligate nursery habitats for juveniles of most reef fish species. The only exception may be the parrotfish Scarus guacamaia Cuvier, 1829, which is listed as a vulnerable species on the IUCN Red List of Threatened Species (International Union for Conservation of Nature and Natural Resources, 2007). Juveniles of these species are only found in mangroves (Nagelkerken and van der Velde, 2002; Mumby et al., 2004; Dorenbosch et al., 2006), and adults are (largely) absent on reefs isolated from mangroves (Nagelkerken et al., 2002; Mumby et al., 2004).

Mangrove/Seagrass–Reef Connectivity Based on Size-frequency Data

Studies related to mangrove–reef connectivity have mostly been conducted in seascapes with ample presence of seagrass beds. In this case it is difficult to distinguish the specific role of mangroves. Nevertheless, the results of these studies are reviewed here because at many oceanic islands, mangroves co-occur with seagrass beds in lagoons, embayments, or back-reef areas. In the Indo-Pacific, very few studies have reported juvenile and adult densities in non-estuarine mangroves, seagrass beds, and coral reefs. Dorenbosch et al. (2005) examined 76 reef fish species in Tanzania and identified eight species for which the juveniles were predominantly (> 70% of total juvenile density) found in mangroves and/or seagrass beds, and adults predominantly on coral reefs. Of these, six species (Cheilinus undulatus Rüppell, 1835, Hipposcarus harid (Forsskål, 1775), Lutjanus monostigma (Cuvier, 1828), Monodactylus argenteus (Linnaeus, 1758), Plectorhin- chus flavomaculatus (Cuvier, 1830), and Scolopsis ghanam (Forsskål, 1775) showed complete absence or a much lower adult density on coral reefs of an island (Grande Comoros) completely lacking mangroves/seagrass beds compared to reefs adjacent to bays harbouring mangroves/seagrass beds (Dorenbosch et al., 2005). Also some other species, which were only to some degree (< 70% of total juvenile density) associated with mangroves/seagrass beds during their juvenile stage, showed reduced densities on reefs isolated from these two habitats, possibly as a result of their partial dependence on mangroves/seagrass beds. On Curaçao, Nagelkerken et al. (2000b) found an overlap of 38 species between mangroves and the adjacent coral reef. The most in-depth study on the distribution of juvenile and adult fish species in Caribbean shallow-water habitats, including mangroves and coral reefs, is by Nagelkerken and van der Velde (2002). They investigated habitat utilization of 50 coral reef species in four different bay habitats and at four different depth zones on the coral reef. Their results suggested a strong mangrove-to- reef ontogenetic migration for at least 10 species [Chaetodon capistratus Linnaeus, 1758, H. flavolineatum, H. parra, H. sciurus, L. apodus, Lutjanus mahogoni (Cuvier in Cuvier and Valenciennes, 1828), S. guacamaia, Sparisoma chrysopterum (Bloch and Schneider, 1801), S. barracuda, and partially Gerres cinereus (Walbaum, 1792)], and a mangrove- and/or channel-to-reef ontogenetic migration for two species (L. griseus, Ocyurus chrysurus). The latter may also be true forH. plumierii (Nagelkerken and van der Velde, 2003). The above results are supported by other studies which were based on a smaller selection of fish species (Rooker, 1995; Nagelkerken et al., NAGELKERKEN: CONNECTIVITY BETWEEN MANGROVES AND CORAL REEFS 601

2000c; Cocheret de la Morinière et al., 2002; Christensen et al., 2003; Eggleston et al., 2004). When interpreting these data, however, the distinction between early and late stage juveniles should be considered. Juveniles may settle in one habitat (e.g., seagrass beds) and move to another habitat (e.g., mangroves) with increasing size, before making a final migration to the coral reef (Rooker and Dennis, 1991; Nagelkerken et al., 2000c; Christensen et al., 2003). The above results for mangrove-to-reef ontogeny are based on the late stage juveniles, but may be different for early stage juveniles. In Curaçao and Bonaire, the main settlement habitat for H. plumierii, L. mahogoni, and O. chrysurus is seagrass (and channel/reef) instead of mangroves, for C. capistratus, H. flavolineatum, and H. sciurus, mangroves as well as seagrass beds are used as settlement habitats, and only G. cinereus, H. parra, L. apodus, L. griseus, and S. barracuda settle predominantly in mangroves (Table 3). These data also apply to individual species on other islands (Rooker, 1995; Christensen et al., 2003; Eggleston et al., 2004). On the coral reef, early juveniles of most of the above 13 species are either absent or found in very low densities (except for the occasional late stage juvenile), suggesting that in contrast to mangroves/seagrass beds, the reef is not an important juvenile habitat for these species in terms of fish density (Nagelkerken et al., 2000b; Nagelkerken and van der Velde, 2002). This probably explains why adult densities of these species are enhanced on coral reefs near bays harboring mangroves and seagrass beds (Table 3). In the complete absence of such bays, densities of H. sciurus, L. apodus, and O. chrysurus were highly reduced on the coral reef, while those of G. cinereus, H. plumierii, L. griseus, S. guacamaia, and S. barracuda were reduced in most cases (Nagelkerken et al., 2002). Furthermore, Miller and Gerstner (2002) found highly reduced densities of O. chrysurus and a complete absence of Haemulidae on the reefs of Navassa Island, while Ogden et al. (1985) found complete absence of six of the above eight species (the exceptions were L. apodus and S. barracuda) on the reefs of Sombrero Island; both islands lack mangroves and seagrass beds. Also on Mona Island with no mangroves and restricted seagrass areas, most of the above species were absent or showed reduced abundances (Dennis et al., 2005). These findings suggest that mangroves/seagrass beds contribute significantly to reef fish populations, although it does not appear that this dependence is obligate (i.e., that the species as such will not persist when these potential nursery habitats are absent). Survival of a few juveniles settling directly on the coral reef may potentially be sufficient to sustain the adult reef populations, merely because the total surface area of coral reefs is often much larger than that of fringing mangroves on small oceanic islands. Reef fish densities ofC. capistratus, H. flavolineatum, L. mahogoni, and S. chrysopterum were not reduced on coral reefs when bays harbouring mangroves/seagrass beds were absent, suggesting that these species show little dependence on these two habitats (Nagelkerken et al., 2002). On Mona Island, all four species were present (Dennis et al., 2005), however, three of the four species were absent from two other islands with little or no mangroves/seagrass beds: H. flavolineatum was absent from Navassa Island (Miller and Gerstner, 2002), while C. capistratus, H. flavolineatum, and L. mahogoni were absent from Sombrero Island (Ogden et al., 1985). Clearly, there is variability in the possible nursery roles of shallow-water habitats. Some other coral reef species are also commonly found as juveniles in the mangroves (Acanthurus chirurgus (Bloch, 1787), Haemulon chrysargyreum Günther, 1859, Hypoplectrus unicolor (Walbaum, 1792), Mulloidichthys martinicus (Cuvier, 602 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

1829), S. iserti, Sparisoma viride (Bonnaterre, 1788); Table 3), but their densities are lower there compared to other shallow-water habitats (Nagelkerken and van der Velde, 2003), and mangroves probably play only a minor or negligible role in the life- cycle of these fish species. Combining all of the above data, it appears that for all studied Caribbean islands at least seven species associated with mangrove/seagrass beds as juveniles are not present or have reduced adult densities on the reef when embayments harboring mangrove/seagrass beds are absent: G. cinereus, H. parra, H. plumierii, H. sciurus, L griseus, O. chrysurus, and S. guacamaia (Table 3). These species may be the most vulnerable to loss of mangrove/seagrass habitats. This is also supported by Nagelkerken et al. (2001), who showed that juveniles of at least H. parra, H. sciurus, L. griseus, and O. chrysurus were absent or scarce in embayments lacking mangroves. An exception is G. cinereus which can use soft muddy substratum as an alternative habitat in the absence of mangroves (Nagelkerken et al., 2001, 2002). Although other species may depend less on the presence of mangroves/seagrass beds, these habitats could still enhance adult fish densities on adjacent coral reefs. For H. flavolineatum, L. apodus, S. iserti, and S. barracuda, the results are contradictory: some studies found absence or reduced adult densities on isolated reefs whereas other studies found no effect (Ta- ble 3), indicating a possible variability in habitat dependence of these species based on local difference in some combination of biotic and/or abiotic factors.

Connectivity Based on Distance to Mangroves/Seagrass Beds

Studies relating adult fish densities on the coral reef to proximity of mangrove/ seagrass habitats have produced mixed results (Tables 2, 3). Halpern (2004) did not find a relationship between distance from mangroves and density of G. cinereus and L. apodus on the reefs of the Virgin Islands. Lack of a relationship with distance to mangroves/seagrass beds was also found off Aruba for 14 of the 15 reef species associated with mangroves/seagrass beds as juveniles (Dorenbosch et al., 2006, 2007), and for C. capistratus, H. flavolineatum, and S. iserti off Curaçao (Nagelkerken et al., 2000b). Other studies in the Caribbean as well as in the Indo-Pacific have reported relationships with distance to embayments or areas harboring a combination of mangroves and seagrass beds. For example, off Curaçao, reef densities ofA. chirurgus, L. apodus, L. mahogoni, O. chrysurus, S. chrysopterum, and Stegastes dorsopu- nicans (Troschel, 1865) decreased exponentially with distance from an embayment (Nagelkerken et al., 2000b), while densities of H. sciurus, L. apodus, L. mahogoni, O. chrysurus, Scarus coeruleus (Bloch, 1786), and S. barracuda were significantly higher on reefs adjacent to embayments harboring mangroves and seagrass beds than on reefs adjacent to embayments without these two habitats and on reefs located away from mangroves/seagrass beds (Dorenbosch et al., 2004). In Old Providence/Santa Catalina, Appeldoorn et al. (2003) found that the biomass of Lutjanidae and Hae- mulidae was four times greater on patch reefs near mangrove/seagrass areas than on patch reefs farther away (> 9 km) from these areas. In Tanzania, Dorenbosch et al. (2005) showed that adult densities of species which occurred commonly as juveniles in mangrove/seagrass beds (> 70% of total juvenile density from all habitats) were higher on reefs adjacent to an embayment harbouring mangrove/seagrass beds than on reefs located further away (17 km on average) from these two habitats. This was the case for 14 out of 18 species, with significant differences for fiveCheilio ( NAGELKERKEN: CONNECTIVITY BETWEEN MANGROVES AND CORAL REEFS 603

inermis (Forsskål, 1775), C. undulatus, P. flavomaculatus, S. ghanam, Scarus ghoban Forsskål, 1775). From the data of Thollot (1992) in New Caledonia, it can be deduced that Hyporhamphus dussumieri and possibly three other species (L. argentimaculatus, A. blochii, A. dussumieri) show a decrease in density on the reef with distance away from the mangroves. It is not clear why the above studies obtained different results, but one of the possible explanations is that most species are not obligately dependent on mangroves because they can use seagrass beds as alternative juvenile habitats. In this case, relating fish densities to distance from mangroves may give ambiguous results because presence of nearby seagrass beds is not considered. Secondly, variation in species-specific dispersal of adults on the reefs, and the specific habitat configuration and habitat size within the marine seascape may also interact with this relationship.

Connectivity Based on Mangrove Size

Studies of the linkage between fish stocks on the reef and non-estuarine mangrove size are almost non-existent (Table 2). Mumby et al. (2004) found a significant effect of mangrove extent on reef fish community structure, but did not elaborate to species level. Halpern (2004) found a positive relationship for G. cinereus between total island mangrove size and reef fish density, but not forL. apodus. The latter may have been caused by an interaction with fishing pressure (Halpern, 2004).

Connectivity Based on Otolith Studies

Elements from seawater are continuously incorporated into the otoliths of fishes during their growth, thus otoliths can provide a history of different habitats (with different water quality) in which a fish has lived during its ontogeny (e.g., Gilland- ers and Kingsford, 2000; Gillanders, 2002). While many studies of otolith microchemistry have been conducted in temperate areas or tropical estuaries, only one published study was done in non-estuarine mangroves and coral reefs. Using otolith microchemistry, Chittaro et al. (2004) determined that 36% of 39 individuals of H. flavolineatum from a coral reef had probably passed through a mangrove nursery. If we calculate binomial 95% confidence limits about their percentage (i.e., 23%–52%), the upper limit suggests substantial connectivity, and thus enhancement, of reef fish populations, whereas the lower does not.

Concluding Remarks

This review shows that studies of mangrove-reef connectivity are dominated by those from the Caribbean region. Studies conducted in the much larger Indo-Pacific region are mostly concentrated on estuarine mangroves, which often lack adjacent coral reefs. It is likely that due to the absence of directly adjacent reefs and/or very different environmental variables (e.g., turbidity, salinity) in estuaries compared to non-estuarine lagoons and embayments, studies in the Indo-Pacific have concluded that mangroves play a minor role in this region as a potential nursery habitat for coral reef fishes compared to the Caribbean. Of the very few studies done in non-estuarine mangroves with adjacent reefs in the Indo-Pacific (Thollot, 1992; Dorenbosch et al., 2005), evidence suggests that these two habitats are possibly connected. I argue here 604 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 that when more studies are done on Indo-Pacific islands with non-estuarine mangroves and adjacent coral reefs, the importance of mangroves (possibly in combination with seagrass beds) may prove to be greater than assumed thus far. Most studies have provided indirect evidence of mangrove-reef connectivity based on size-frequency studies. Direct evidence of this connectivity is still lacking (except for a small study by Chittaro et al., 2004). New or advanced techniques such as otolith microchemistry, stable isotope analysis, DNA analysis, and use of internal micro-tags or transponders have recently become more readily available at lower costs, providing new opportunities for investigating the connectivity between tropical coastal habitats by fishes. The present need for studies related to habitat connectivity has already been amply discussed in other reviews (Parrish, 1989; Beck et al., 2001; Gillanders et al., 2003; Heck et al., 2003; Sheridan and Hays, 2003; Manson et al., 2005; Adams et al., 2006; Dahlgren et al., 2006) and applies to the issue of mangrove—reef connectivity. Among the most urgent questions are whether mangroves contribute to coral reef fish populations, the magnitude of the contribution, the distance fish disperse from their nursery areas and the extent of spatio-temporal variability in nursery contribution.

Acknowledgments

This study was funded by a Vidi grant to the author from the Netherlands Organisation of Scientific Research (NWO). J. Serafy and the sponsors of the First International Symposium on Mangroves as Fish Habitat (Miami, 2006) are thanked for inviting me to this symposium, which inspired for the current review. This is Centre for Wetland Studies publication no. 448.

Literature Cited

Adams, A. J., C. P. Dahlgren, G. T. Kellison, M. S. Kendall, C. A. Layman, J. A. Ley, I. Nagelkerken, and J. E. Serafy. 2006. Nursery function of tropical back-reef systems. Mar. Ecol. Prog. Ser. 318: 287–301. Appeldoorn, R. S., A. Friedlander, J. Sladek Nowlis, P. Usseglio, and A. Mitchell-Chui. 2003. Habitat connectivity in reef fish communities and marine reserve design in Old Provi- dence–Santa Catalina, Colombia. Gulf Caribb Res. 14: 61–77 Beck, M. W., K. L. Heck Jr., K. W. Able, D. L. Childers, D. B. Eggleston, B. M. Gillanders, B. Halpern, C. G. Hays, K. Hoshino, T. J. Minello, R. J. Orth, P. F. Sheridan, and M. P. Wein- stein. 2001. The identification, conservation, and management of estuarine and marine nurseries for fish and invertebrates. BioScience 51: 633–641. Birkeland, C. and S. S. Amesbury. 1988. Fish-transect surveys to determine the influence of neighboring habitats on fish community structure in the tropical Pacific. Pages 195–202in A. L. Dahl, ed. Co-operation for environmental protection in the Pacific. UNEP Regional Seas Reports and Studies No. 97. Blaber, S. J. M. 2000. Tropical estuarine fishes. Ecology, exploitation and conservation. Fish and aquatic resources series 7. Blackwell Science, Oxford. 372 p. Chittaro, P. M., B. J. Fryer, and P. F. Sale. 2004. Discrimination of French grunts (Haemulon flavolineatum, Desmarest, 1823) from mangrove and coral reef habitats using otolith microchemistry. J. Exp. Mar. Biol. Ecol. 308: 169–183. ______, P. Usseglio, and P. F. Sale. 2005. Variation in fish density, assemblage composition and relative rates of predation among mangrove, seagrass and coral reef habitats. Environ. Biol. Fish. 72: 175–187. NAGELKERKEN: CONNECTIVITY BETWEEN MANGROVES AND CORAL REEFS 605

Cocheret de la Morinière, E., B. J. A. Pollux, I. Nagelkerken, and G. van der Velde. 2002. Post- settlement life cycle migration patterns and habitat preference of coral reef fish that use seagrass and mangrove habitats as nurseries. Estuar. Coast. Shelf Sci. 55: 309–321. ______, I. Nagelkerken, H. van der Meij, and G. van der Velde. 2004. What attracts juvenile coral reef fish to mangroves: habitat complexity or shade? Mar. Biol. 144: 139–145. Christensen, J. D., C. F. G. Jeffrey, C. Caldow, M. E. Monaco, M. S. Kendall, and R. S. Appel- doorn. 2003. Cross-shelf habitat utilization patterns of reef fishes in southwestern Puerto Rico. Gulf Caribb Res. 14: 9–27. Dahlgren, C. P., G. T. Kellison, A. J. Adams, B. M. Gillanders, M. S. Kendall, C. A. Layman, J. A. Ley, I. Nagelkerken, and J. E. Serafy. 2006. Marine nurseries and effective juvenile habitats: concepts and applications. Mar. Ecol. Prog. Ser. 312: 291–295. Dennis, G. D., W. F. Smith-Vaniz, P. L. Colin, D. A. Hensley, and M. A. McGehee. 2005. Shore fishes from islands of the Mona passage, greater Antilles with comments on their zoogeog- raphy. Caribb. J. Sci. 41: 716–743. Dorenbosch, M., M. G. G. Grol, I. Nagelkerken, and G. van der Velde. 2006. Seagrass beds and mangroves as potential nurseries for the threatened Indo-Pacific humphead wrasse, Cheilinus undulatus and Caribbean rainbow parrotfish, Scarus guacamaia. Biol. Conserv. 129: 277–282. ______, M. C. van Riel, I. Nagelkerken, and G. van der Velde. 2004. The relationship of reef fish densities to the proximity of mangrove and seagrass nurseries. Estuar. Coast. Shelf Sci. 60: 37–48. ______, C. E. P. Verberk, I. Nagelkerken, and G. van der Velde. 2007. Influence of habitat configuration on connectivity between fish assemblages of Caribbean seagrass beds, mangroves and coral reefs. Mar. Ecol. Prog. Ser. 334: 103–106. ______, M. G. G. Grol, M. J. A. Christianen, I. Nagelkerken, and G. van der Velde. 2005. Indo-Pacific seagrass beds and mangroves contribute to fish density and diversity on adjacent coral reefs. Mar. Ecol. Prog. Ser. 302: 63–76. Eggleston, D. B., C. P. Dahlgren, and E. G. Johnson. 2004. Fish density, diversity, and size-structure within multiple back reef habitats of Key West national wildlife refuge. Bull. Mar. Sci. 75: 175–204. Faunce, C. H. and J. E. Serafy. 2006. Mangroves as fish habitat: 50 years of field studies. Mar. Ecol. Prog. Ser. 318: 1–18. Gillanders, B. M. 2002. Connectivity between juvenile and adult fish populations: do adults remain near their recruitment estuaries? Mar. Ecol. Prog. Ser. 240: 215–223. ______and M. J. Kingsford. 2000. Elemental fingerprints of otoliths of fish may distinguish estuarine ‘nursery’ habitats. Mar. Ecol. Prog. Ser. 201: 273–286. ______, K. W. Able, J. A. Brown, D. B. Eggleston, and P. F. Sheridan. 2003. Evidence of connectivity between juvenile and adult habitats for mobile marine fauna: an important component of nurseries. Mar. Ecol. Prog. Ser. 247: 281–295. Halpern, B. S. 2004. Are mangroves a limiting resource for two coral reef fishes? Mar. Ecol. Prog. Ser. 272: 93–98. Heck, K. L., G. Hays, and R. J. Orth. 2003. Critical evaluation of the nursery role hypothesis for seagrass meadows. Mar. Ecol. Prog. Ser. 253: 123–136. International Union for Conservation of Nature and Natural Resources [internet]. 2007. 2006 IUCN Red List of Threatened Species. Available from: http://www.redlist.org. Accessed 22 May 2007. Laegdsgaard, P. and C. Johnson. 2001. Why do juvenile fish utilise mangrove habitats? J. Exp. Mar. Biol. Ecol. 257: 229–253. Manson, F. J., N. R. Loneragan, G. A. Skilleter, and S. R. Phinn. 2005. An evaluation of the evidence for linkages between mangroves and fisheries: a synthesis of the literature and identification of research directions. Oceanogr. Mar. Biol. Ann. Rev. 43: 483–513. 606 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Miller, M. W. and C. L. Gerstner. 2002. Reefs of an uninhabited Caribbean island: fishes, benthic habitat, and opportunities to discern reef fishery impact. Biol. Conserv. 106: 37–44. Mumby, P. J., A. J. Edwards, J. E. Arias-Gonzalez, K. C. Lindeman, P. G. Blackwell, A. Gall, M. I. Gorczynska, A. R. Harborne, C. L. Pescod, H. Renken, C. C. C. Wabnitz, and G. Llewellyn. 2004. Mangroves enhance the biomass of coral reef fish communities in the Caribbean. Nature 427: 533–536. Nagelkerken, I. and G. van der Velde. 2002. Do non-estuarine mangroves harbour higher densities of juvenile fish than adjacent shallow-water and coral reef habitats in Curaçao (Nether- lands Antilles)? Mar. Ecol. Prog. Ser. 245: 191–204. ______and G. van der Velde. 2003. Connectivity between coastal habitats of two oceanic Caribbean islands as inferred from ontogenetic shifts by coral reef fishes. Gulf Caribb. Res. 14: 43–59. ______, M. Dorenbosch, W. C. E. P. Verberk, E. Cocheret de la Morinière, and G. van der Velde. 2000a. Day-night shifts of fishes between shallow-water biotopes of a Caribbean bay, with emphasis on the nocturnal feeding of Haemulidae and Lutjanidae. Mar. Ecol. Prog. Ser. 194: 55–64. ______, ______, ______, ______, and ______. 2000b. Im- portance of shallow-water biotopes of a Caribbean bay for juvenile coral reef fishes: patterns in biotope association, community structure and spatial distribution. Mar. Ecol. Prog. Ser. 202: 175–192. ______, G. van der Velde, M. W. Gorissen, G. J. Meijer, T. van‘t Hof, and C. den Hartog. 2000c. Importance of mangroves, seagrass beds and the shallow coral reef as a nursery for important coral reef fishes, using a visual census technique. Estuar. Coast. Shelf Sci. 51: 31–44. ______, S. Kleijnen, T. Klop, R. A. C. J. van den Brand, E. Cocheret de la Morinière, and G. van der Velde. 2001. Dependence of Caribbean reef fishes on mangroves and seagrass beds as nursery habitats: a comparison of fish faunas between bays with and without mangroves/seagrass beds. Mar. Ecol. Prog. Ser. 214: 225–235. ______, C. M. Roberts, G. van der Velde, M. Dorenbosch, M. C. van Riel, E. Cocheret de la Morinière, and P. H. Nienhuis. 2002. How important are mangroves and seagrass beds for coral-reef fish? The nursery hypothesis tested on an island scale. Mar. Ecol. Prog. Ser. 244: 299–305. Ogden, N. B., W. G. Gladfelter, J. C. Ogden, and E. H. Gladfelter. 1985. Marine and terrestrial flora and fauna notes on Sombrero Island in the Caribbean. Atoll Res. Bull. 292: 61–74. Parrish, J. D. 1989. Fish communities of interacting shallow-water habitats in tropical oceanic regions. Mar. Ecol. Prog. Ser. 58: 143–160. Robertson, A. I. and S. J. M. Blaber. 1992. Plankton, epibenthos and fish communities. In A. I. Robertson and D. M. Alongi, eds. Tropical mangrove ecosystems. Coastal and Estuarine Studies 41: 173–224. Rooker, J. R. 1995. Feeding ecology of the schoolmaster snapper Lutjanus apodus (Walbaum), from southwestern Puerto Rico. Bull. Mar. Sci. 56: 881–894. ______and G. D. Dennis. 1991. Diel, lunar and seasonal changes in a mangrove fish assemblage off southwestern Puerto Rico. Bull. Mar. Sci. 49: 684–698. Serafy, J. E., C. H. Faunce, and J. L. Lorenz. 2003. Mangrove shoreline fishes of Biscayne Bay, Florida. Bull. Mar. Sci. 72: 161–180. Sheridan, P. and C. Hays. 2003. Are mangroves nursery habitat for transient fishes and decapods? Wetlands 23: 449–458. Shulman, M. J. 1985. Recruitment of coral reef fishes: effects of distribution of predators and shelter. Ecology 66: 1056–1066. Thollot, P. 1992. Importance of mangroves for Pacific reef fish species, myth or reality? Proc.th 7 Int. Coral Reef Symp. 2: 934–941. NAGELKERKEN: CONNECTIVITY BETWEEN MANGROVES AND CORAL REEFS 607

______and M. Kulbicki. 1988 Overlap between the fish fauna inventories of coral reefs, soft bottoms and mangroves in Saint-Vincent Bay (New Caledonia). Proc. 6th Int. Coral Reef Symp. 2: 613–618. Verweij, M. C., I. Nagelkerken, D. de Graaff, M. Peeters, E. J. Bakker, and G. van der Velde. 2006. Structure, food and shade attract juvenile coral reef fish to mangrove and seagrass habitats: a field experiment. Mar. Ecol. Prog. Ser. 306: 257–268.

Address: Department of Animal Ecology and Ecophysiology, Institute for Wetland and Wa- ter Research, Faculty of Science, Radboud University, Toernooiveld 1, 6525 ED Nijmegen, The Netherlands. E-mail: . BULLETIN OF MARINE SCIENCE, 80(3): 609–624, 2007

species-specific patterns of fish abundance and size along a subtropical mangrove shoreline: an application of the DELTA APPROACH

Joseph E. Serafy, Monica Valle, Craig H. Faunce, and Jiangang Luo

ABSTRACT From 1998 to 2005, 537 visual fish surveys were conducted along a 50-km stretch of mangrove-lined shoreline in the vicinity of Biscayne Bay (southeastern Florida, USA). The shoreline lies directly downstream of a major wetlands restoration project that aims to return more natural salinity regimes to the western margin of the region’s coastal bays. As part of a “baseline” ecological assessment, we applied the delta approach to examine spatial and temporal patterns of mangrove habitat use by three fishes:Lutjanus griseus (Linnaeus, 1758), Sphyraena barracuda (Walbaum, 1792), and Floridichthys carpio (Gunther, 1866). Along a north-south gradient, seasonal variation in their size-composition and three abundance metrics (occurrence, concentration, and density) was quantified and overall correlations with water salinity, depth, and temperature were evaluated using multiple regression. Results indicated that the shoreline is used by subadult and adult L. griseus and by mostly juvenile S. barracuda; for both species, highest abundance metric values occurred during the wet season, generally increasing southwards. In contrast, F. carpio were almost exclusively of adult (mature) sizes, with greatest values during the dry season at the shoreline’s northern extent. For all species, water depth and/or temperature had a significant effect on abundance metrics. Positive correlations were found for L. griseus and S. barracuda abundance metrics, whereas the reverse was true for F. carpio. To date, most mangrove-fish studies have evaluated a single measure of fish abundance, usually density. We suggest consideration of occurrence and concentration, in addition to density, has value in an assessment and monitoring context as well as for gaining insight into how a given fish species disperses and clusters within habitats and across gradients.

A series of three subtropical bays are located on Florida’s southeastern coast. From north to south, these are Biscayne Bay, Card Sound, and Barnes Sound (Fig. 1). To varying degrees, all three are bordered by mangroves (primarily Rhizophora mangle Linneaus) and their substrates vegetated with marine seagrasses (mostly Thalassia testudinum Banks and Soland. ex Koenig). In general, anthropogenic impacts on the three bays, their watersheds, and their shorelines decrease from north to south (Snedaker and Biber, 1996; Browder et al., 2005). Encompassed by the highly urbanized metropolis of Miami, northern Biscayne Bay’s mangroves have been almost entirely replaced with concrete seawalls and limestone boulders (Teas et al., 1976). However, from central Biscayne Bay southward, human influence tends to lessen as the dominant watershed use grades from urban to suburban to agricultural, and finally to serve primarily as parkland (Roessler and Beardsley, 1974). Areal coverage by mangroves increases about 24-fold from central Biscayne Bay to the southwestern boundary of Barnes Sound (Teas, 1974; Ross et al., 2001). The western margins of central and southern Biscayne Bay, Card Sound, and Barnes Sound form a semi-continuous, 50-km stretch of mangrove-lined shoreline, which is interspersed with natural creeks, artificial channels, and freshwater canal mouths (Fig. 1). Restoring more natural freshwater flows and salinity regimes along the west-

ern margins of the three bays is a major objective of the Biscayne Bay Coastal Wet- lands Project (BBCWP), a $300 million component of the Comprehensive Everglades Restoration Plan (USACE and SFWMD, 2002). The BBCWP comprises numerous large-scale engineering projects designed to re-hydrate existing coastal wetlands that are now drained by the canal system, and to redistribute freshwater flow to the Bay across a broad front using pump stations, spreader swales, culverts, and other means. These efforts will modify the timing, location, and volume of freshwater -in flows to the three bays, and likely change salinity regimes, contaminant loads, and nutrient dynamics, which, in turn, could have consequences for the bays’ habitats, fishes, and fisheries. Therefore, an important element of the BBCWP is to gauge its impact on the fishes of the area, before, during, and after implementation, which is planned to begin during or before 2008. This paper represents part of a “baseline” assessment that focuses on fishes inhabiting one of the habitats mostly likely to be affected by BBCWP activities—i.e., the 50-km stretch of mangrove shoreline along the western perimeter of central Bis- cayne Bay, Card Sound, and Barnes Sound. To reveal seasonal and spatial patterns of fish abundance and size-composition along this shoreline, we draw on a portion of the data collected in an ongoing visual belt-transect fish survey. Our focus is on the analysis of abundance and size data for three species: gray snapper, Lutjanus griseus (Linnaeus, 1758), great barracuda Sphyraena barracuda (Walbaum, 1792), and goldspotted killifish,Floridichthys carpio (Gunther, 1866). These fishes were selected because: (1) they are among the most abundant fishes that are readily identified (visually) to the species level; (2) two have economic importance (i.e, gray snapper and great barracuda); and (3) together, they span at least two trophic levels. Conducting species-by-species analyses can be highly informative, but difficulties can arise. At the species-specific level, fish abundances are often positively skewed and dominated by zeros (Lo et al., 1992), reflecting patchiness in the environment and in the distribution of the species concerned (Gaston, 1994). Consequently, such data are usually inappropriate for conventional parametric statistical analyses. A recent paper by Fletcher et al. (2005) promoted wider application of the delta approach to ecological problems, which in our case was to gauge pre-restoration levels, patterns, and variability in fish abundance at the species-specific level. The delta approach is not new (Aitchison and Brown, 1957; Seber, 1982; Pennington, 1983) and is often used in fisheries applications (e.g., Lo et al., 1992) and in ecological studies on niche overlap (Krebs, 1999). It is virtually absent, however, in the mangrove-fish literature (Faunce and Serafy, 2006). In our study, the delta approach entailed analyzing a species’ frequency of occurrence over the transects surveyed (hereafter termed “occurrence”) and concentration (i.e., density where present, exclusive of zeros) separately and then in combination (as delta-densities). As an extension of the delta approach, we also examined intra-specific concentration-occurrence relationships G( aston et al., 2000) to determine their utility for ecological assessment and monitoring.

Methods

Data Collection.—The present study focused on fish use of the mainland mangrove fringe between latitudes 25°16´N and 25°39´N. Portions of this stretch of habitat have been surveyed visually for fishes during daylight hours following a biannual, random stratified sampling design (Serafy et al., 2003; Faunce, 2005) for over 7 yrs. Specifically, from 1998 to SERAFY ET AL.: PATTERNS OF MANGROVE-FISH ABUNDANCE AND SIZE 611

2005, a total of 537 60-m2 visual belt-transect fish surveys were conducted over 15 consecutive wet (July–September) and dry (January–March) seasons (Table 1). Details of the visual fish survey and microhabitat measurement techniques employed are provided by Serafy et al. (2003). Briefly, the former entailed a trained observer, using mask, snorkel, and waterproof paper, recording the identity, number, and size-structure (minimum, mean, and maximum total length) of fishes observed within a 30 × 2 m belt of inundated prop-root habitat. Belt width was measured landward from the prop-root edge and belt length was measured by means of spooling pre-measured length of parachute cord during each fish survey. Care was taken to record all fishes from the top of the water column to the substrate. Upon completion of each fish survey, the microhabitat measurements of water temperature and salinity were made using a multi-probe water quality instrument and water depth using a 2 m-long polyvinyl chloride pole marked off every 2 cm. Survey data were subsequently transferred to computer spreadsheets prior to analysis using SAS (1990) statistical software. Study Design and Data Analysis.—The purpose of this analysis was to conduct a seasonally-resolved, spatial characterization of mangrove shoreline use by three fish species prior to implementation of the BBCWP. Because human impacts and mangrove coverage vary along a latitudinal gradient, we: (1) subdivided the mainland mangrove fringe into a north-south series of seven shoreline segments, each measuring approximately seven km in length (Fig. 1); (2) assigned each transect to its appropriate segment; and then (3) analyzed for seasonal abundance and size trends along north-south and microhabitat gradients. We arrived at the seven-segment scheme to balance the conflicting needs of having: (1) adequate sample sizes per season-segment combination; and (2) a sufficient number of segments to expose patterns (if any) of mangrove utilization along the north-south gradient. Data analysis consisted of several steps. Abundance-frequency plots were constructed as a basic data diagnostic step to judge the appropriateness of taking the delta approach, a method useful in situations where data are positively skewed and zero values predominate. As noted by Fletcher et al. (2005), the delta approach involves generating two data sets from the original: one indicating the species presence (occurrence, proportion of transects positive for the species in question) and the other abundance when present (concentration, numbers observed per transect, when present). The product of occurrence O( ) and mean concentration values (C) yields an index of relative density, hereafter termed “delta-density” (D): DO= ) C (1) var DO= 2 ) var C ] g ] g (2) where occurrence (O) is a constant variable. The variance of density can be estimated via the delta method, which uses a Taylor approximation to estimate variances of functions of random variables (Rice, 1995). This index is considered more representative of the data than a mean density estimate calculated in the conventional way, which generally has large variance due to the large number of zeros (Seber, 1982). Separate analysis of the two components results in more robust estimation and understanding of the variance associated with each, and often results in a greatly reduced estimate of the variability around the (composite) delta- density value (Lo et al., 1992; Ortiz et al., 2000, 2001). Species-specific patterns in occurrence, concentration, and delta-density were examined at two levels. First, mean and standard error values of all three abundance metrics were generated for each season and shoreline segment. This was achieved by pooling across years (1998– 2005), and in the case of occurrence values, weighting each proportion by sampling effort (i.e., the number of visual transects). The year time-series was pooled because this project seeks for single reference, or “baseline”, values with variability. Results were then plotted against each shoreline segment’s (central) latitude and regression analysis was used to describe general north-south trends along the 50-km stretch. We also constructed concentration-occurrence relationship plots (termed “abundance-occupancy relationships” by Gaston et al., 2000) as a possible framework for describing pre-restoration abundance metric levels and variabil- 612 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 - 89 60 64 68 55 100 537 101 Grand total 44 29 48 31 34 46 26 258 Total dry Total 45 31 52 33 34 55 29 279 Total wet Total 8 W 11 18 16 10 96 20 13 ‘05 8 D 11 18 14 10 94 20 13 8 9 4 4 7 6 – W 38 ‘04 7 9 4 3 5 – D 38 10 7 9 4 3 5 – W 38 10 ‘03 6 9 4 4 4 1 – D 28 7 9 4 4 4 2 – W 30 ‘02 – 7 4 4 2 2 D 10 29 Year-Season 2 3 5 6 3 1 – W 20 ‘01 × 2 m belt transect. 3 8 3 3 3 2 4 D 26 3 2 1 5 1 3 W 12 27 ‘00 7 3 4 4 5 1 3 D 27 – – 4 3 3 3 2 W 15 ‘99 – – – 4 4 5 3 D 16 – 3 2 5 5 – – W ‘98 15 Table 1. Sampling (number effort of Table visual fish transects) conducted along the 50-km stretch of mainland mangrove habitat season, per and year, shoreline seg ment. Shoreline segments are numbered from = north wet W to season south. (July–September); D = dry season (January–March). Each visual transect entailed recording the identity and size structure of fishes within a 30 Shoreline segment 7 1 Totals: 2 3 4 5 6 SERAFY ET AL.: PATTERNS OF MANGROVE-FISH ABUNDANCE AND SIZE 613

Figure 1. Map depicting Biscayne Bay, Card Sound, and Barnes Sound and the 50-km stretch of mangrove-lined shoreline subdivided into seven segments for analysis purposes. The gray, shaded area indicates coverage of adjacent mangrove-dominated wetlands. Black lines across the watershed indicate coastal canal system. ity. This involved plotting the mean concentration of each species per segment-year-season against its respective occurrence value and expressing variability in this space as an 80% confidence ellipse. The second level at which species-specific occurrences, concentrations, and delta-densities were examined aimed to reveal overall correlations, if any, with water salinity, depth and temperature. Following Hocking (1976) and Draper and Smith (1981) stepwise multiple regression analysis was performed. Specifically, a backwards elimination approach was taken whereby factors were removed sequentially if their P-values were > 0.05 and final model fit was judged from adjusted R2-values. Prior to these regression analysis, occurrence data were arcsine-transformed; concentration and density data loge-transformed. Throughout, concentration and delta-density values are expressed on a per 60 m2 basis. Finally, following Serafy et al. (2003), species-specific (percent) length-frequency distribution plots were constructed using the method of Meester et al. (1999). These were generated with 1, 5 or 10 cm intervals (length bins) to infer life stage (mature/immature) and to examine for possible differences in size composition between northern (segments 1–4) vs southern (segments 5–7) shoreline segments. Kolmogorov-Smirnov two sample tests (Sokal and Rohlf, 614 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

1981) were applied separately to dry and wet season data to determine whether north-south differences in cumulative size frequency distributions were statistically significant (i.e., P < 0.05). In the event that within-season, along-shore differences were found to be non-significant, northern and southern segment data were pooled and overall seasonal differences analyzed using the same Kolmogorov-Smirnov two sample tests.

Results

General.—The realized distribution of sampling effort (i.e., number of belt-transect fish surveys) within each year, season, and shoreline segment reflects increasing effort along the entire 50-km stretch, as manpower and other resources became available, peaking in final year (2005) at 94 and 96 visual transects during the dry and wet seasons, respectively (Table 1). Pooling across years, effort per shoreline segment—season ranged from 26 to 52 visual transects. The water salinities, temperatures, and depths measured during fish surveys ranged 0–42, 15–36 °C, and 9–175 cm, respectively. Forty-seven fish taxa were observed along the 50-km stretch over the period of record. The 10 most abundant fish taxa were (in decreasing order): small, water column fishes (Atherinidae, Clupeidae, and Engraulidae), goldspotted killifish, small mojarras (Eucinostomus spp.), gray snapper, yellowfin mojarra, Gerres cinerus (Walbaum, 1792), sailfin molly, Poecilia latipinna (Lesueur, 1821) pinfish, Lagodon rhomboides (Linnaeus, 1766), rainwater killifish, Lucania parva (Baird and Girard, 1855), great barracuda, and bluestriped grunt, Haemulon sciurus (Shaw, 1803). Gray snapper, great barracuda, and goldspotted killifish were the 6th, 3rd, and 2nd most frequently- observed fishes, respectively. Of the total number of transects (537), 138 (26%) were positive for gray snapper, 194 (36%) were positive for great barracuda, and 257 (47%) were positive for goldspotted killifish. Abundance and size analyses were based on sightings of 2640 gray snapper ranging 5–51 cm (TL), 395 great barracuda ranging 5–76 cm and 12,507 goldspotted killifish ranging 1–6.5 cm. Spatiotemporal Abundance Patterns.—Spatial and seasonal similarities in the abundance metric values of gray snapper and great barracuda were evident along the 50-km stretch. Gray snapper mean occurrence values tended to be highest during the wet season, with peak values along the central shoreline segments (Fig. 2A). During the dry season, however, gray snapper occurrence was generally lower, increasing linearly from north to south. The mean concentration values of gray snapper also increased linearly from north to south during both seasons, with only minor differences in magnitude between seasons (Fig. 2B). Mean delta-density values for gray snapper (Fig. 2C) were consistently higher during the wet season vs the dry, increasing exponentially from north to south during both seasons, although one value (i.e., the southernmost shoreline segment 7) strayed from this general trend. Mean occurrence values for great barracuda showed no clear pattern along the north-south gradient during the dry season (Fig. 2D). However, as with gray snapper, mean values were higher during the wet season and peaked along the central shoreline segments. A slight linear increase in mean concentration was detected from north to south during the dry season, but not during the wet (Fig. 2E). Great barracuda delta-density means were all elevated during the wet season, but no clear north-south trends were evident (Fig. 2F). SERAFY ET AL.: PATTERNS OF MANGROVE-FISH ABUNDANCE AND SIZE 615

Figure 2. Seasonal abundance patterns of three species observed along the 50-km mainland mangrove fringe: (A–C) gray snapper, (D–F) great barracuda, and (G–I) goldspotted killifish. Shore- line segments 1–7 on the y-axis are numbered consecutively from north to south. Solid symbols indicate wet season and open symbols dry. Values are means per (± 1 standard error) per shoreline segment. Occurrence values are the proportion of transects positive for the species in question. Concentration and density values expressed on a per 60 m−2 basis (i.e., 30 × 2 m belt transect). Lines and their associated R2-values indicate significant (P < 0.05) north-south trends.

Mean abundance metric levels and trends of goldspotted killifish were the most distinct from the other two species. Mean occurrence values were uniformly low with no north-south trend during the wet season (Fig. 2G), but during the dry season were very high along the northernmost shoreline segments, and decreased sharply towards the south. Mean concentration values for goldspotted killifish followed the same north-south pattern as their occurrence values, with a sharp linear decline evident during the dry season, but not during the wet (Fig. 2H). Mean delta-density 616 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Figure 3. Concentration-occurrence plots with 80% confidence ellipses for each species by season. Shown are values and ellipses for (A) gray snapper, (B) great barracuda, and (C) goldspotted killifish. Thick lines and solid symbols indicate wet season; thin lines and open symbols indicate dry season. Note that fish concentration axis is in natural logarithm units. See text for details.

values for goldspotted killifish (Fig. 2I) increased exponentially from south to north during the dry season only. During the wet season, these values were uniformly low and no clear spatial trend was evident. Concentration-occurrence Confidence Ellipses.—Seasonal variability concentration-occurrence space is depicted in Figure 3 for each species along the entire 50-km stretch. Most ellipses indicated weakly positive correlations between occurrence and concentration. For gray snapper and great barracuda, dry season el- SERAFY ET AL.: PATTERNS OF MANGROVE-FISH ABUNDANCE AND SIZE 617

Table 2. Results of backwards elimination multiple regression analysis indicating strength and direction of relationships between species-specific abundance metrics and three habitat measures. Final model intercept values and coefficient estimates (± 1 standard error) associated with water salinity, depth, and temperature are shown with adjusted R2 values.

2 (A) Gray snapper B0 BSAL BDEP BTEMP Adj. R Occurrence (S.E.) −1.559 (0.327) 0.011 (0.005) 0.014 (0.002) 0.0372 (0.009) 0.42 Concentration (S.E.) −0.441 (0.427) ns 0.038 (0.008) ns 0.19 Density (S.E.) −3.381 (0.901) 0.041 (0.015) 0.029 (0.006) 0.0648 (0.025) 0.28

2 (B) Great barracuda B0 BSAL BDEP BTEMP Adj. R Occurrence (S.E.) −1.046 (0.303) ns 0.012 (0.012) 0.0026 (0.010) 0.32 Concentration (SE) −0.054 (0.157) ns 0.016 (0.003) ns 0.24 Density (S.E.) −0.996 (0.282) ns 0.012 (0.002) 0.0324 (0.010) 0.29

2 (C) Goldspotted killifish B0 BSAL BDEP BTEMP Adj. R Occurrence (S.E.) 7.973 (1.157) ns −0.015 (0.003) −0.0532 (0.011) 0.37 Concentration (SE) 6.583 (1.085) ns ns −0.1484 (0.039) 0.13 Density (S.E.) −0.023 (0.011) ns −0.023 (0.011) −0.1726 (0.039) 0.22

lipses were smaller and more laterally compressed than wet season ellipses. The opposite was true for goldspotted killifish. Microhabitat Correlations.—Final regression models describe the strength and direction of correlations between each of the species-specific abundance metrics and water salinity, temperature and depth (Table 2). Adjusted R2 values of the final models ranged from 0.13 to 0.42. Best fits were for occurrences and the poorest for fish concentrations. In final models pertaining to gray snapper and great barracuda, all factor coefficients were positive, whereas those pertaining to goldspotted killifish were negative. Gray snapper occurrence and delta-density increased with salinity, depth, and temperature, whereas great barracuda occurrence and density increased with only depth and temperature. Concentrations of these species appeared unrelated to salinity or temperature, with a tendency for increase with depth alone. Goldspot- ted killifish occurrence and delta-density both decreased with increasing depth and temperature; however, their concentration values were only related to temperature, declining as it increased. Size-composition.—Given that minimum size at maturity for gray snapper is about 20 cm TL (Claro et al., 2001), a roughly equal percentage of immature (47%) and mature fish was observed. According to de Sylva (1963) minimum size at maturity for great barracuda is about 60 cm TL, suggesting 89% of individuals observed were immature. Published size-maturity information is lacking for goldspotted killifish, but assuming the same minimum size at maturity for sheepshead minnow, Cyprinodon variegates Lacépède, 1803, applies (i.e., 2 cm TL; Hardy, 1978), 95% of the individuals observed were sexually mature. Within the wet season, north-south differences in size composition were not statistically significant for any of the species examined (Fig. 4). During the dry season, however, significant north-south differences were detected for gray snapper (Fig. 4A,B) and goldspotted killifish (Fig. 4D,E). For gray snapper the greatest difference was in the predominance of 10–15 cm size class in the north versus the south. For goldspotted killifish, size distribution was unimodal in the north and bimodal in the 618 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

south, with the greatest difference being in the predominance of the 4–5 cm size class in the north. Because no spatial differences in the size composition of great barracuda were detected during either season (data not shown), segments were pooled to test for strictly seasonal size composition differences. This analysis revealed that great barracuda size distribution was unimodal during the wet season and bimodal during the dry (Fig. 4C). During the wet season the 10–20 cm size class predominated, whereas during the dry season, the 30–40 and 60–70 cm size classes were best represented.

Discussion

The present study reports baseline information pertaining to the abundances and sizes of three commonly observed fishes along the mangrove-lined, western perimeter of central and southern Biscayne Bay, Barnes Sound, and Card Sound. Similar analyses have been performed for other fish species and at the entire assemblage level (Serafy et al., 2005); however, those results are not detailed here. A recent review of over 100 mangrove-fish studies (Faunce and Serafy, 2006) indicated most analyses to date have been conducted at the assemblage- rather than the species-level. We suggest both assemblage- and species-level analyses are required to improve our understanding of the roles that mangroves play in the lives of fishes. Our species selection (gray snapper, great barracuda, and goldspotted killifish) represents a balancing of practical, ecological, and economic considerations. From a practical standpoint, given our underwater visual methodology, these fishes pose far less of a challenge to accurately identify to species, count, and measure than, for example, the occasionally more numerous anchovies (Engraulidae), silversides (Atherinidae), herrings (Clupeidae), and mojarras (Gerridae), that share this habitat in highly variable, often mixed-species schools. Other, typically larger, mangrove fishes, such as the common snook, Centropomis undulates (Bloch, 1792), are readily identified and quantified visually, however, the relative rarity of their detection along the 50-km stretch precludes meaningful analyses, at least for this pre-restoration time period. An ecological basis for focusing on these species is that they span at least two trophic levels, as inferred from previous gut content, stable isotope, and energetic modeling studies conducted in mangrove habitats in the region (de Sylva, 1963; Starck and Schroeder, 1970; Odum, 1970; Harrigan et al,. 1989; Hettler, 1989; Schmidt, 1989). At the size ranges in which they occur, goldspotted killifish represent secondary consumers in the food web, whereas gray snapper and great barracuda, both predators of goldspotted killifish, represent tertiary consumers. Finally, gray snapper and great barracuda are targeted by both recreational and commercial fisheries; therefore, results pertaining to these species resonate with fishers and the greater public. Unlike the first quantitative mangrove-fish study conducted along this shoreline (see Serafy et al., 2003), here we incorporated more years of sampling, extended our view farther south, and provided greater detail, especially in terms of spatial variation in fish abundances. Our use of the delta approach also represents a substantial information increase for the species examined here. This was prompted by the pre- ponderance of zero abundance values, a condition that appears to be more the rule than the exception when analyzing species-specific fish abundance data from field surveys. As noted by Fletcher et al. (2005), by separately considering the likelihood SERAFY ET AL.: PATTERNS OF MANGROVE-FISH ABUNDANCE AND SIZE 619

Figure 4. Size-compositions of the three focal species along the 50-km mainland mangrove fringe during the dry and wet seasons. Gray snapper (A–B), great barracuda (C) and goldspotted killifish (D–E). In all panels but (C), plots show within-season percent length-frequency distributions for northern shoreline segments (segments 1–4, thin lines) versus southern shoreline segments (segments 5-7, thick lines). In (C), plots show percent length-frequency distributions by season (i.e., spatial differences not significant, thus segments were pooled) with the thick line indicating the wet season and dotted line the dry. In all plots, the D- and P-values indicate Kolmogorov-Smirnov two sample test results: D = maximum difference between cumulative size frequency distributions; P = probability of obtaining the observed difference by chance alone. Vertical lines (with arrows) indicate minimum size at maturity values obtained from the literature. 620 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 of presence (occurrence) and the abundance given presence (concentration), we have gained more insight into the system than had we only considered fish densities, computed either conventionally or as delta-densities (i.e., occurrence × concentration). From a monitoring perspective, the delta approach is somewhat precautionary in that it furnishes additional, biologically-meaningful abundance measurements that can each be tracked and evaluated as potential indicators of habitat status or change. Moreover, the concentration-occurrence framework has potential for visualizing and detecting simultaneous change in these abundance metrics. Certainly, as restoration and monitoring proceed, new fish data can be plotted in occurrence-concentration space to: (1) determine whether they fall inside or outside the pre-restoration (i.e., “baseline”) confidence ellipses; and (2) examine the degree to which density changes are driven by changes in concentration, occurrence, or both. With additional data, this framework may be applied at finer spatial levels than demonstrated here (e.g., shoreline segment). The numerous differences among the three species revealed here undoubtedly reflect their very different life history strategies, trophic positions, habitat requirements, and physiological tolerances as well as the disparate time-scales and mechanisms underlying the observed abundances and sizes of each species. Whereas gray snapper and great barracuda larvae originate from offshore waters, goldspotted killifish spend their entire life cycle within shallow, wetland and nearshore seagrass habitats. A valid argument against use of gray snapper and great barracuda abundances as indicators of wetland restoration impacts is that their “pathways” to mangrove shorelines depends on a series of unrelated, prior conditions and stochastic events. These include fishing pressure on their parental stocks (Ault et al., 1998; Harper et al., 2000) and the many biophysical processes and behavioral factors behind successful navigation from pelagic, offshore waters to the littoral zone. However, it is important to reiterate the primary objective of this paper: to characterize mean levels, broad patterns, and inter-annual variability of habitat use by three species that presently occur within our sampling domain. This objective differs from attempting to find and/or justify the monitoring of one species over another Z( acharias and Roff, 2001; Rice, 2003) as a “sentinel” of condition or change. Our results show that fish utilization of this 50-km stretch of red mangrove differs according to the latitude, season, species, and abundance metric under scrutiny. Based on our size composition analysis, the shoreline serves to varying degrees as habitat for mature and immature gray snapper, and mostly immature great barracuda. Conversely, the goldspotted killifish we observed were almost exclusively of adult (mature) sizes, with greater usage during the dry season as compared to the wet. Low detection of juvenile and subadult stages of this species (i.e., individuals < 2 cm TL), however, may have contributed to this result. Regardless of detection problems, our size composition results for all three species provide baseline (pre-restoration) size distributions by shoreline and season against which future comparisons can be made. Several species-specific differences were evident among the abundance metrics examined, along a north-south gradient, seasonally, and in terms of overall relationships with water quality and depth. Results for gray snapper and great barracuda were similar in that their occurrences were consistently elevated during the wet season, with no significant north-south trend. This may reflect a broadly spread initial ingress by both species to this shoreline during the wet season, a pattern that great barracuda retains during the dry season when they tend to be larger, and pre- SERAFY ET AL.: PATTERNS OF MANGROVE-FISH ABUNDANCE AND SIZE 621 sumably, older. The dry season trend of southerly increase for gray snapper, however, may reflect subsequent movement or poor habitat quality that only becomes clear when water depths and temperatures tend to be lower and salinity higher. Gray snapper and great barracuda also diverged in that the former displayed a strong tendency for increased concentrations from north to south, but the latter did not. The consistently low great barracuda concentration values may reflect the preference of individuals of this species to distance themselves from one another, possibly to reduce hunting interference and/or cannibalism among these ambush predators (de Sylva, 1963). In contrast, the seemingly more gregarious, gray snapper and goldspotted killifish, which tend to feed on benthic invertebrates H( ettler, 1989; Ley et al., 1994), occurred in much higher concentrations, depending on location and season. In keeping with the large southerly increase in mangrove-dominated wetlands along the 50-km stretch, the delta-density values for gray snapper tended to increase exponentially along this gradient. This was not the case for great barracuda, which showed either no, or very little, variation attributable to mangrove coverage. Spatially, temporally, and for all three of its abundance metrics, goldspotted killifish displayed opposite trends relative to the other species, with the tendency for sharp southerly decrease during the dry season. As noted by Serafy et al. (2003), this exclusively dry season pattern may reflect that wetland drainage, which serves to force fishes from the forest interior to the shoreline, declines in a southerly direction. Restoration and Conservation Considerations.—The BBCWP is at the beginning of its implementation period and so are several associated ecological monitoring projects put in place to gauge its impacts. The primary purpose of most of the monitoring efforts, including ours, is to establish present “baseline” levels and/or patterns of one or more biotic and abiotic measurements and then track changes in them as project-related modifications ensue. Under ideal circumstances, we would have reliable, prior knowledge regarding: (1) precisely when, where, and by how much freshwater flows and salinity regimes will change due to restoration activities; (2) what the relevant ecological metrics are; and (3) how they should be measured and interpreted. Unfortunately, due to the lack of quantitative, spatially-explicit historical data, the many uncertainties regarding the timing, magnitude, and form of restoration effects, and the limited understanding of community dynamics in general, it seems prudent to “hedge our bets” by monitoring multiple community components in multiple ways. While a recent surge of studies is revealing important and overwhelmingly positive connections among mangrove habitats, fishes, fisheries, and wider ecosystems (Faunce and Serafy, 2006), they are also demonstrating important differences in the roles that mangroves play, which vary over time, by location and on a species-by-species basis. Mangrove habitats continue to be destroyed or degraded in a “piecemeal” fashion. Therefore, from a conservation standpoint, the need for detailed and spatially-focused faunal utilization studies is particularly pressing. In practice, efforts to preserve or enhance the ecological integrity of mangrove habitats may require numerous, small scale mangrove-fish assessment studies, strategically placed in, or directly downstream of, areas slated for alteration. By virtue of its relatively narrow spatial focus, the present study provides such information. Although funded for their relevance to the BBCWP, the data may also have value as numerous other changes, ultimately driven by human population increase, ensue. In our particular case, these 622 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 include threats of further development of the area’s watershed for housing and additional increases in fish exploitation, especially by the recreational fishing sector. Our results, therefore, may have utility beyond evaluating restoration impacts. In conclusion, we recommend that researchers complement their assemblage-level analyses with detailed, species-specific analyses of fish abundance and size. For the latter, we suggest consideration of the delta approach as a potentially useful and informative way of examining species-specific abundances. The approach may provide insight into how a given fish species disperses and clusters within habitats and across gradients that would not be gained by considering fish density (or biomass) alone (Fletcher et al., 2005). Moreover, each of the “triad” of abundance metrics, along with size-composition data, is simple to convey to scientists and non-scientists alike, and certainly more straightforward to compare among studies than, for example, most multivariate statistical results. We contend that the delta approach holds promise as means of refining our understanding of mangroves as fish habitat, and therefore, may be an important tool for conservation and management.

Acknowledgments

Financial support for this work was provided by the U.S. Army Corps of Engineers, South Florida Water Management District, National Oceanic and Atmospheric Administration, National Park Service, National Audubon Society, and U.S. Geographical Survey, Perry In- stitute for Marine Science Caribbean Marine Research Center, and the University of Miami’s Rosenstiel School of Marine and Atmospheric Science. We are indebted to the technical support provided by J. Barimo, E. D’Alessandro, D. Keickbusch, C. Peyer, E. Orbesen, D. Snod- grass, M. South, and P. B. Teare. This paper is Protected Resources Division Contribution PRD-06/07-05.

literature Cited

Aitchison, J. and J. A. C. Brown. 1957. The lognormal distribution, with special reference to its uses in economics. Cambridge University Press, Cambridge. 176 p. Ault, J. S., J. A. Bohnsack, and G. A. Meester. 1998. A retrospective (1979–1996) multispecies assessment of coral reef fish stocks in the Florida Keys. Fish. Bull. 96: 395–414. Browder, J. A., R. Alleman, S. Markley, P. Ortner, and P. A. Pitts. 2005. Biscayne Bay conceptual ecological model. Wetlands 25: 854–869. Claro, R., K. C. Lindeman, and L. R. Parenti. 2001. Ecology of the Marine Fishes of Cuba. Smithsonian Instituion Press, Washington D.C. 253 p. de Sylva, D. P. 1963. Systematics and life history of the great barracuda Sphyraena barracuda (Wal- baum). Studies in Tropical Oceanography No. 1. University of Miami, Coral Gables. 179 p. Draper, N. and H. Smith. 1981. Applied regression analysis. Wiley Interscience, New York. 709 p. Faunce, C. H. 2005. Reef fish utilization of mangrove shoreline habitats within southeastern Florida. Ph.D. Diss. University of Miami, Coral Gables, Florida. 146 p. ______and J. E. Serafy. 2006. Mangroves as fish habitat: fifty years of field studies. Mar. Ecol. Prog. Ser. 318: 1–18. Fletcher. D., D. Mackenzie, and E. Villouta. 2005. Modelling skewed data with many zeros: a simple approach combining ordinary and logistic regression. Environmental and Ecological Statistics 12: 45–54. Gaston, K. J. 1994. Rarity. Chapman & Hall. London. 205 p. ______, T. M. Blackburn, J. D. Greenwood, R. D. Gregory, R. M. Quinn, and J. H. Lawton. 2000. Abundance-occupancy relationship. J. Appl. Ecol. 37: 39–59. SERAFY ET AL.: PATTERNS OF MANGROVE-FISH ABUNDANCE AND SIZE 623

Hardy, J. D. 1978. Development of the fishes of the mid-Atlantic Bight: an atlas of egg, larval and juvenile stages, vol. II, Anguillidae through Sygnathidae. Biological Services Program, Fish and Wildlife Service (FWS/OBS-78/12), U.S. Department of the Interior. 458 p. Harper, D. E., J. A. Bohnsack, and B. R. Lockwood. 2000. Recreational fisheries in Biscayne National Park, Florida, 1976–1991. Mar. Fish. Rev. 62: 8–26. Harrigan, P., J. C. Zieman, and S. A. Macko. 1989. The base of nutritional support for the gray snapper (Lutjanus griseus): an evaluation based on a combined stomach content and stable isotope analysis. Bull. Mar. Sci. 44: 65–77. Hettler, W. F. 1989. Food habits of spotted seatrout and gray snapper in western Florida Bay. Bull. Mar. Sci. 44:155–162. Hocking, R. R. 1976. The analysis and selection of variables in linear regression. Biometrics 32: 1–50. Krebs, C. J. 1999. Ecological methodology. 2nd Edition. Addison Wesley Longman Inc. Menlo Park, California. 620 p. Ley, J. A., C. L. Montague, and C. C. McIvor. 1994. Food habits of mangrove fishes: a comparison along estuarine gradients in northeastern Florida Bay. Bull. Mar. Sci. 54: 881–889. Lo, N. C., L. D. Jacobson, and J. L. Squire. 1992. Indices of relative abundance from fish spotter data based on delta-lognormal models. Can. J. Fish. Aquat. Sci. 49: 2515–2526. Meester, G. A., J. S. Ault, and J. A. Bohnsack. 1999. Visual censusing and the extraction of average length as an indicator of stock health. Naturalista Siciliana 23: 205–222. Odum, W. E. 1970. Pathways of energy flow in a south Florida estuary: Ph.D. Diss. University of Miami, Coral Gables, Florida. 180 p. Ortiz, M. and G. P. Scott. 2001. Standardized catch rates for white marlin (Tretapturus albidus) and blue marlin (Makaira nigricans) from the pelagic longline fishery in the Northwest Atlantic and the Gulf of Mexico. Col. Vol. Sci. Pap. ICCAT, 53: 231–248. ______, C. M. Legault, and G. P. Scott. 2000. Variance component estimation for standardized catch rates of king mackerel (Scomberomorus cavalle) from U.S. Gulf of Mexico recreational fisheries useful for inverse variance weighting techniques. NMFS SEFSC Sustainable Fisheries Division Contribution SFD-99/00-86. Mackerel Stock Assessment Panel Report MSAP/00/03. Pennington, M. 1983. Efficient estimators of abundance, for fish and plankton surveys. Biomet- rics 39: 281–286 Rice, J. 1995. Mathematical statistics and data analysis. 2nd ed. Duxbury. 597 p. Rice, J. 2003. Environmental health indicators. Ocean Coast. Manage. 46: 235–259. Roessler, M. A. and G. L. Beardsley. 1974. Biscayne Bay: its environment and problems. Fla. Sci. 37: 286–204. Ross, M. S., P. L. Ruiz, G. J. Telesnicki, and J. F. Meeder. 2001. Estimating above-ground biomass and production in mangrove communities of Biscayne National Park, Florida (U.S.A.) Wetl. Ecol. Manage. 9: 27–37. SAS. 1990. User’s Guide, Ver. 6., 4th ed. SAS Institute, Cary N.C. 584 p. Schmidt, T. W. 1989. Food habits, length-weight relationship and condition factor of young great barracuda, Sphyraena barracuda (Walbaum), from Florida Bay, Everglades National Park, Florida. Bull Mar. Sci. 44: 163–170. Seber, G. A. F. 1982. The estimation of animal abundance. Edward Arnold. London. 205 p. Serafy, J. E., C. H. Faunce, and J. J. Lorenz. 2003. Mangrove shoreline fishes of Biscayne Bay, Florida. Bull. Mar. Sci. 72: 161–180. ______, J. Luo, M. Valle, and C. H. Faunce. 2005. Shoreline fifish sh community visual assess-assessment. CERP Monitoring and Assessment Plan Component. Activity Number 3.2.3.6. First Cumulative Report. 48 p. Sokal, R. R. and F. J. Rohlf. 1981. Biometry. 2nd Edition. W. H. Freeman and Company. New York. 859 p. 624 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Snedaker, S. C. and P. D. Biber. 1996. Restoration of mangroves in the United States of America: a case study in Florida. Pages 170–188 in C. D. Field, ed. 1996. Restoration of mangrove ecosystems. ISME, Okinawa, Japan. 304 p. Starck, W. A. and R. E. Schroeder. 1970. Investigations on the gray snapper, Lutjanus griseus. Stud. Trop. Oceanogr. (Miami) 10. 224 p. Teas, H. J. 1974. Mangroves of Biscayne Bay. Mimeo, Date County. 107 p. ______, H. R. Wanless, and R. Chardon. 1976. Effects of man on the shore vegetation of Bis- cayne Bay. Pages 133–156 in A. Thorhaug and A. Volker, eds. Biscayne Bay: past/present/ future: Papers presented for Biscayne Bay Symposium I, Coral Gables. Univ. Miami Sea Grant Program. 315 p. USACE and SFWMD (U.S. Army Corps of Engineers and South Florida Water Management District) 2002. Central and Southern Florida project, Comprehensive Everglades Restora- tion Plan, Project Management Plan, Biscayne Bay Coastal Wetlands. 126 p. with Appen- dices. Zacharias, M. A. and J. C. Roff. 2001. Use of focal species in marine conservation and management: a review and critique. Aquatic Conserv: Mar. Freshw. Ecosyst. 11: 59–76.

Addresses: (J.E.S.) Southeast Fisheries Science Center, National Marine Fisheries Service, 75 Virginia Beach Drive, Miami, Florida 33149. (M.V., J.L.) Rosenstiel School of Marine and Atmospheric Science, 4600 Rickenbacker Causeway, Miami Florida, 33149. (C.H.F.) Florida Fish and Wildlife Conservation Commission, Florida Fish and Wildlife Research Institute, Te- questa Field Laboratory, P.O. Box 3478, Tequesta, Florida 33469. Corresponding Author: (J.E.S.) E-mail: . BULLETIN OF MARINE SCIENCE, 80(3): 625–645, 2007

Seasonal fish community variation in headwater mangrove creeks in the southwestern Everglades: an examination of their role as dry-down refuges

Jennifer S. Rehage and William F. Loftus

Abstract The connectivity between the fish community of estuarine mangroves and that of freshwater habitats upstream remains poorly understood. In the Florida Everglades, mangrove-lined creeks link freshwater marshes to estuarine habitats downstream and may act as dry-season refuges for freshwater fishes. We examined seasonal dynamics in the fish community of ecotonal creeks in the southwestern region of Ev- erglades National Park, specifically Rookery Branch and the North and Watson rivers. Twelve low-order creeks were sampled via electrofishing, gill nets, and minnow traps during the wet season, transition period, and dry season in 2004–2005. Catch- es were greater in Rookery Branch than in the North and Watson rivers, particularly during the transition period. Community composition varied seasonally in Rookery Branch, and to a greater extent for the larger species, reflecting a pulse of freshwater taxa into creeks as marshes upstream dried periodically. The pulse was short-lived, a later sample showed substantial decreases in freshwater fish numbers. No evidence of a similar influx was seen in the North and Watson rivers, which drain shorter hydroperiod marshes and exhibit higher salinities. These results suggest that headwater creeks can serve as important dry-season refugia. Increased freshwater flow resulting from Everglades restoration may enhance this connectivity.

Biological connectivity between fish communities of mangrove regions and those of other marine and coastal habitats (e.g., coral reefs, seagrass beds, sand, and mudflats), although deserving further attention, has been explored in a number of recent studies (see reviews by Beck et al., 2001, Gillanders et al., 2003; Sheridan and Hays, 2003; Mason et al., 2005; Sheaves, 2005; Faunce and Serafy, 2006). The presence of mangroves along coastal areas enhances the richness, abundance, and biomass of fishes in marine habitats (e.g., coral reefs; Nagelkerken et al., 2002; Dorenbosch et al., 2004; Mumby et al., 2004). Mangroves provide nursery grounds for larval and juvenile marine fishes and crustaceans (Robertson andD uke, 1987; Laegdsgaard and Johnson, 1995; Nagelkerken et al., 2000) due to their high prey abundance (Robert- son et al., 1988; Sheridan, 1997) and their role as a predation refuge (Primavera, 1998; Acosta and Butler, 1999). Juvenile survival may be enhanced in shallow mangrove habitats where structural complexity, shading, and turbidity are relatively high (Lae- gdsgaard and Johnson, 2001; Ellis and Bell, 2004). Mobile marine fishes use mangrove habitats transiently as foraging grounds (Blaber and Milton, 1990; Chong et al., 1990), reproductive grounds (Chaves and Bouchereau, 2000), or move in when environmental conditions in diel or seasonal cycles are favorable (e.g., at high tide or with increased salinity or temperature; Ley et al., 1999; Barletta et al., 2005). By comparison, the connectivity between fish communities in mangrove regions and upstream freshwater habitats has received much less attention. A reason for this is that in many mangrove systems, the freshwater influence is small, and the contribution of freshwater fishes to the estuarine community is limited (Pinto and Punchi- hewa, 1996; Laroche et al., 1997; Kuo et al., 1999; Nordlie, 2003; Hindell and Jenkins,

2004). In the Greater Everglades Ecosystem, shallow vegetated freshwater marshes transition into an extensive region of tidal mangrove forests (up to 15 km in width, over 60,000 ha of mangroves), which dominates the landscape along the southwest Florida coast (Smith et al., 1994). At the ecotone, mangrove-lined creeks drain upland marshes into a network of interconnecting estuarine rivers, bays, and mangrove forests. The ecosystem is rainfall-driven, marked by strong seasonality (high rainfall in the summer and fall, low in the winter and spring), which greatly influences the spatial extent of inundation of freshwater marshes, as well as the salinity regime of this broad estuarine region (Gunderson and Loftus, 1993). As in other estuarine systems, salinity levels play an important role in structuring the plant and animal communities of the Greater Everglades Ecosystem (Montague and Ley, 1993; Serafy et al., 1997; Ley et al., 1999; Lorenz, 1999; Faunce et al., 2004). Historically, large volumes of freshwater reached estuarine areas, particularly during the wet season (Fennema et al., 1994). Today, drainage, channelization, and impoundment of marshes have greatly diminished the freshwater inflow into estuarine areas, resulting in substantially higher and more variable salinity regimes (Smith et al., 1989; Montague and Ley, 1993; Light and Dineen, 1994; McIvor et al., 1994). Fish community response to the natural and the anthropogenically-derived variation in freshwater inflow and salinity has been relatively well-studied in the southern and eastern parts of the ecosystem, namely, Florida Bay and Biscayne Bay (Thayer et al., 1987; Montague and Ley, 1993; Serafy et al., 1997; Ley et al., 1999; Lorenz, 1999; Serafy et al., 2003; Faunce et al., 2004; Lorenz and Serafy, 2006); but remains under- studied along the southwest region (but see Green et al., 2006), where the mangrove zone is substantially more extensive than in the southern and eastern parts (Smith et al., 1994). Mangrove creeks along this area also drain generally longer hydroperiod marshes than the southern and eastern regions (Fenema et al., 1994). These marshes (Shark Slough) support more diverse and abundant fish assemblages than southern marshes (Taylor Slough) (Trexler et al., 2001; Chick et al., 2004; Green et al., 2006); thus high connectivity between the mangrove and freshwater fish communities may be expected. In this study, we examined variation in the fish community of headwater mangrove creeks in response to seasonal fluctuations in freshwater flow and salinity in the southwestern region of Everglades National Park (ENP). In particular, we explored the role of low-order, ecotonal mangrove creeks as dry-season refuges for freshwater fishes. As marsh water levels drop, fishes are forced into deeper habitats such as -al ligator holes, solution holes, canals (Kushlan, 1974; Nelson and Loftus, 1996; Chick et al., 2004; Kobza et al., 2004; Rehage and Trexler, 2006), and presumably headwater creeks. We sampled the fish community in the uppermost stretches of creeks, where habitat may be most suitable for freshwater species because of proximity to marshes and low salinity regimes. A secondary goal of this study was to compare sampling efficiency among gears. Sampling with electrofishing and gill nets targeted large -fish es, whereas minnow traps targeted small fishes. Sampling focused on two regions: Rookery Branch (RB) and the North and Watson rivers (NW) (Fig. 1). Headwater creeks in the RB region link the main freshwater drainage of the southern Everglades (Shark Slough) to Tarpon Bay, and the Shark and Harney rivers. Creeks in the NW area are headwaters of the North and Watson rivers which flow into Whitewater Bay. In neither system have the fish communities in the oligohaline reaches received enough attention to describe their seasonal and long-term dynamics beyond surveys REHAGE AND LOFTUS: EVERGLADES MANGROVE CREEKS AS DRY-SEASON REFUGES 627 that provided inventory data (Tabb and Manning, 1961; Odum, 1971; Loftus and Kushlan, 1987), despite their historical importance as a prey source for wading birds (Ogden, 1994) and their key role in the mangrove food web (Odum, 1971).

Methods

Site Description.—We sampled the large and small fish community in the oligohaline to mesohaline headwater reaches of six creeks in the RB region and six creeks in the NW region (Fig. 1). RB sites included four creeks in the main stem of Rookery Branch, as well as Squawk and Otter creeks (RB7 and RB12, respectively). NW sites were located along three Watson River creeks and three North River creeks. Creeks in NW drain shorter hydroperiod marshes than RB creeks. By hydroperiod, we refer to the number of days the marsh is flooded in a yearly cycle. Marshes are typically considered dry if water levels drop below 5 cm, at which depth little standing water remains (Loftus and Eklund, 1994). According to data from hydrologic stations P35 and P38 (Fig. 1), over the past 20-yr period (1986–2006), the hydroperiod averages 332 d of flooding (± 8.3 d) in marshes upstream of RB creeks, and 305 d (± 13.5 d) in marshes upstream of NW creeks. Nutrient concentrations are similar between regions: relatively high nitrogen (approximately 1 mg L–1) and low phosphorus concentrations (below 0.02 mg L–1) are characteristic of both RB and NW waters (Levesque, 2004). Sampling Effort.—All sampling was conducted in the main channel of headwater creeks and in the uppermost boat-accessible 600 m reach of each creek. Sampling included only first and second order creeks (Strahler, 1957). Creek shorelines were vegetated by riverine mangrove forests dominated by red mangrove, Rhizophora mangle Linnaeus, 1753 (Lugo and Snedaker, 1974). Creek depth at sampling locations averaged 1.37 m (± 0.03 m, n = 108); width averaged 10.8 m (± 1.0 m, n = 107). Sampling was conducted during November 2004, February 2005, and April 2005, corresponding to the wet season, the transition between wet and dry

Figure 1. Map of southwestern Everglades National Park showing location of headwater creeks included in this study. Twelve creeks (filled circles) were sampled: six in the North and Watson rivers (NR1-3 and WR4-6) and six in the Rookery Branch region (RB7-12). Location of four reference NPS hydrological stations is indicated by open triangles (see Figure 2 for data from these stations). Distance from sampling sites to stations varies. CN and NR stations are located in creek channels downstream of sites sampled (CN is 4200 m downstream from RB7 and NR is 900 m downstream from NR3). P35 and P38 are located in freshwater marshes upstream of creeks (P35 is 300 m from RB9 and P38 is 6300 m from NR2). 628 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Figure 2. (A) Mean daily salinity and (B) water levels collected by the nearest four NPS monitoring stations to study headwater creeks. See Figure 1 for approximate location of stations in reference to sites. Bold lettering indicates sampling months. Dotted line indicates 5-cm water depth cutoff used for calculation of marsh hydroperiod (see text for explanation). seasons (hereafter “transition”), and dry season. Daily marsh water level and creek salinity measurements were obtained from the nearest National Park Service (NPS) hydrologic stations (Figs. 1 and 2). While sampling, we measured salinity at creek sites with a YSI® 85 unit (Fig. 3). Large-fish Sampling.—Large fishes (55–750 mm standard length, SL) were sampled using a boat-mounted electrofishing unit (two-anode, one-cathode Smith-Root® generator-pow- ered pulsator 9.0 unit rated to a maximum salinity of 15). Electrofishing has been shown to be an effective method for sampling larger fishes in other Everglades habitats (Nelson and Loftus, 1996; Chick et al., 1999). At each creek, sampling was conducted in three 5-min (pedal time) bouts (three bouts × six creeks × two regions × three seasons = 108 electrofishing samples). For all bouts, electrofishing power was standardized to 1500 watts according to temperature and salinity conditions (Burkhardt and Gutreuter, 1995). On average, each bout sampled 122.6 m (± 2.8 m, n = 108) of creek shoreline. Bouts were distributed evenly over the 600-m segment of creek, so that each bout was considered an independent sampling unit. For each bout, we randomly selected a creek shoreline and made a single pass with the electrofish- REHAGE AND LOFTUS: EVERGLADES MANGROVE CREEKS AS DRY-SEASON REFUGES 629

Figure 3. Salinities over the three sampling seasons in the two study regions: the North and Wat- son rivers and Rookery Branch. Shown are means ± 1 standard error (SE). ing boat. All fish captured were identified to species, measured to nearest mm SL, weighed to nearest g, and released after full recovery. Non-indigenous species were collected and brought to the laboratory for processing. We sampled the upper 100-m reaches of each creek with two passive techniques—experimental gill nets and minnow traps. Gill nets are commonly used to monitor fish populations in a wide range of habitats, typically targeting highly mobile and large-bodied species (e.g., Hubert and O’Shea, 1992). Experimental gill nets have panels of several mesh sizes, thus reducing the potential for size or single-species selectivity (Argent and Kimmel, 2005). Nets were 38 m long, with six mesh sizes (25.4, 38.1, 50.8, 63.5, and 76.2 mm). One net was set in the upper 100 m of each creek (one gill net × six creeks × two regions × three seasons = 36 gill net samples). Logistic constrains prevented us from obtaining greater gill-net sample sizes that would be comparable to electrofishing sample sizes. To comply with NPS regulations, gill nets were set mid-channel, parallel to the direction of current flow, and for only 30-min periods. All fishes captured were identified, measured, and weighed in the field, then released. Small-fish Sampling.—Small fishes (< 50 mm SL) were sampled with 3-mm, metal-mesh minnow traps (25.4 mm opening) deployed unbaited, overnight along creek banks. Minnow traps are a commonly used and easily replicable sampling device, but it suffers from several sampling biases (Rozas and Minello, 1997), one of which is trap placement. Minnow traps are typically set on the substrate, where they are unlikely to be encountered by water-column or surface dwellers (Layman and Smith, 2001). In this study, we deployed minnow traps in pairs; one set on the substrate and a second suspended just beneath the water surface, secured to mangrove prop roots. In each creek, we deployed three pairs of traps during the November 2005 sampling event, but increased effort to five pairs for subsequent sampling events (No- vember: six traps × six creeks × two regions = 72 samples, February and April: 10 traps × six creeks × two regions x two seasons = 240 samples; total sample size is 312). Fish captures from minnow traps were preserved in 10% formalin and brought to the laboratory for processing. Statistical Analyses.—We examined variation in the abundance of fishes among regions and creeks and as a function of season with nested, repeated-measures ANOVA or AN- COVA models. Season was the repeated measure in our analyses, and nesting allowed us to 630 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 account for spatial variation among regions (RB and NW); creeks were nested within regions. Focal response variables included: CPUE for the large fishes caught in electrofishing (number 5-min–1 pedal time) and gill nets (number 30-min–1 soak time), CPUE for the small fishes caught in minnow traps (number 24 h−1), and the proportion of CPUE that was freshwater in electrofishing and minnow trap samples (CPUE was too low in gill net samples). Species were classified as either marine, estuarine, or freshwater (Table 1) based upon their habitat occurrence (per Loftus and Kushlan, 1987; Loftus, 2000). Preliminary analyses examined seasonal and spatial variation in the number of species caught in all gears, but results were indistinguishable from analyses of CPUE; and thus, are not presented here. A two-way ANOVA was used to examine seasonal and spatial variation in salinity levels. Salinity was used as a covariate in analyses of the large fish data; no salinity measurements were made at the time of minnow trap deployment. To better satisfy assumptions of parametric tests, CPUEs were ln (observed value + 1)-transformed and proportions were subject to angular transformations prior to analyses. Post-hoc pairwise comparisons were performed using Tukey-corrected contrasts. If salinity was a significant covariate, simple linear regressions were used to examine the relationship between response variables and salinity. All analyses were performed using Proc Mixed in SAS Version 9.1.3®. We used analyses of similarity (ANOSIM) based on Bray-Curtis similarity matrices to test for effects of region, season, and gear (electrofishing vs gill net, and top vs bottom minnow trap) on fish community structure (Clarke and Warwick, 2001). Dissimilarity matrices were constructed based on ln (observed value + 1)-transformed estimates of the relative abundance of all taxa in samples, except for the gear comparison of gillnets and electrofishing, where a presence/absence matrix was used. Analyses included 28 taxa from electrofishing samples, 10 taxa from gill nets, and 22 taxa from minnow traps (Table 1). We followed ANOSIM analyses with percentage of similarity analyses (SIMPER) to determine which taxa contributed most to groupings observed among samples. We constructed non-metric multi-dimensional scaling (NMDS) plots to illustrate dissimilarity among groups. In these plots, the distance between data points is proportional to the degree of similarity between samples. All community structure analyses were conducted using Primer© Version 5.2.9.

Results

Freshwater Flow and Salinity.—Salinity levels near our study sites increased substantially with the yearly onset of the dry season as flows from upstream freshwater marshes decreased (Fig. 2). In 2005, these increases occurred earlier in NW creeks than in RB creeks (Fig. 2A). Maximum salinities were higher in the vicinity of our NW sites than near RB sites (29 and 19, respectively). Hydroperiod was shorter in marshes upstream of NW sites than in those upstream of RB sites. Marshes upstream of NW sites were flooded for 357 d in 2004–2005; whereas marshes upstream of RB creeks were flooded for 324 d (Fig. 2B). Marshes upstream of RB also dried more frequently, but for short periods of time. In contrast, marshes upstream of the North River dried less frequently, but once dry, remained dry for a longer period of time. At our study creeks, salinity varied both among sites and across seasons (ANOVA: significant season by region interaction, F2, 102 = 18.6, P = 0.0001; Fig. 3). Salinities were comparable between regions during the wet season sample, but diverged as the dry season progressed, reaching 10 in NW but < 5 in RB (Tukey pairwise comparisons of RB vs NW, P < 0.0001 for both the transition period and dry season). Estimates of Abundance.—Fish abundance in the oligohaline to mesohaline reaches of mangrove creeks, as estimated by electrofishing, gill net, and minnow trap CPUE, was consistently higher in RB creeks than in NW creeks (Figs. 4,5). CPUE varied as a function of season (significant season by region interactions for all three REHAGE AND LOFTUS: EVERGLADES MANGROVE CREEKS AS DRY-SEASON REFUGES 631 - 2 4 1 2 D 97 44 119 357 1 11 49 25 T 390 468 102 1,016 2 4 1 Minnow traps 13 15 21 W 137 1 1 D 23 8 1 T 31 CPUE Gill nets 1 W 5 2 3 2 1 19 14 D 246 2 2 6 6 1 42 52 56 T 351 102 5 8 1 9 7 1 Electrofishing 17 33 W FW FW E, M EF, E EF, EF, E EF, FW, EF FW, FW, EF FW, EF FW, FW, EF FW, FW, EF FW, FW, EF FW, EF FW, EF, E, M EF, EF, E, M EF, EF, E, M EF, EF, E, M EF, EF, E, M EF, E, M EF, EF, E, M EF, FW, EF, E EF, FW, Occurrence FW, EF, E, M EF, FW, E, M EF, FW, FW, EF, E, M EF, FW, rainwater killifish eastern mosquitofish least killifish sailfin molly golden topminnow bluefin killifish American eel coastal shiner Tarpon redfin pickerel gulf toadfish inland silverside unidentified needlefish bluespotted sunfish hardhead catfish Florida gar bowfin striped mullet Atlantic needlefish sheepshead minnow Snook Warmouth Bluegill Common name Species Lacépède, 1803 Girard, 1859 Valenciennes, 1847 Valenciennes, (Lesueur, 1821) (Lesueur, (Cope, 1867) Jordan, 1880 Linnaeus, 1758 (Linnaeus, 1766) Linnaeus, 1766 Lucania parva (Baird and Girard, 1855) Gambusia holbrooki Agassiz, 1855 Heterandria formosa Poecilia latipinna (Günther, 1866) Fundulus chrysotus (Günther, Lucania goodei (Lesueur, 1817) (Lesueur, Anguilla rostrata 1942 petersoni Fowler, Notropis Megalops atlanticus Ariopsis felis Esox americanus Gmelin, 1789 Opsanus beta (Goode and Bean, 1880) Menidia beryllina sp. Strongylura Enneacanthus gloriosus (Holbrook, 1855) DeKay, 1842 DeKay, Lepisosteus platyrhincus Amia calva Mugil cephalus 1860) notata (Poey, Strongylura Cyprinodon variegatus Centropomus undecimalis (Bloch, 1792) Centropomus 1829) Lepomis gulosus (Cuvier, Rafinesque, 1819 Lepomis macrochirus water habitats, EF = oligohaline to mesohaline reaches of estuarine habitats, E = mesohaline to polyhaline reaches of estuarine habitats, and M = marine habitats marine = M and habitats, estuarine of reaches polyhaline to mesohaline = E habitats, estuarine of reaches mesohaline to oligohaline = EF habitats, water (Loftus and Kushlan, 1987; Loftus, 2000). Table 1. Number of fishesTable caught (catch per unit CPUE) effort, by the three sampling methods inWatson headwater rivers creeks and in Rookery the North and Branch over three sampling events (W = wet, T = transition, and D = Dry) in 2004–2005. Species were classified according to their occurrence in FW = fresh Scientific name 632 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 1 2 4 2 11 D 646 3 1 2 3 22 T 216 312 2,287 3,143 2 2 2 3 1 4 3 Minnow traps W 210 3 1 1 1 1 D 31 1 3 T 47 36 84 10 CPUE Gill nets 2 2 6 W 7 2 2 5 1 2 11 16 10 D 348 6 2 8 12 38 31 21 15 27 T 100 213 108 1,066 1,782 2 5 7 5 2 1 5 3 4 Electrofishing 11 17 W 111 113 368 Occurrence EF, E, M EF, E, M E, M EF, EF FW, EF, E, M EF, E, M EF, E, M EF, E, M EF, FW, EF FW, EF FW, FW, EF FW, FW, EF FW, EF, E EF, FW FW, EF FW, EF FW, EF FW, EF, E, M EF, E, M EF, E, M EF, Sheepshead spotted seatrout Redfish pygmy sunfish Everglades crevalle jack gray snapper tidewater mojarra striped mojarra unidentified sunfishes bass largemouth spotted sunfish Redear naked goby dollar sunfish Mayan cichlid Blue tilapia spotted tilapia Common name crested goby clown goby Hogchoker a a Species a (Günther, 1859) (Günther, (Linnaeus, 1766) Non-indigenous species (Günther, 1862) (Günther, Cichlasoma urophthalmus (Walbaum, 1792) (Walbaum, probatocephalus Archosargus (Cuvier, 1830) Cynoscion nebulosus (Cuvier, Jordan, 1884 Elassoma evergladei Sciaenops ocellatus Caranx hippos (Linnaeus, 1766) Lutjanus griseus (Linnaeus, 1758) Goode and Bean, 1879 Eucinostomus harengulus 1830) plumieri (Cuvier, Eugerres Lepomis sp. salmoides (Lacépède, 1802) Micropterus (Valenciennes, 1831) Lepomis punctatus (Valenciennes, Lepomis microlophus (Steindachner, 1864) (Steindachner, aureus Oreochromis 1899 mariae Boulenger, Tilapia Gobiosoma bosc (Lacépède, 1800) (Holbrook, 1855) Lepomis marginatus Table 1. Continued. Table a Scientific name Lophogobius cyprinoides (Pallas, 1770) gulosus (Girard, 1858) Microgobius 1801) maculatus (Bloch and Schneider, Trinectes CPUE total by season Total number of samples Total Total CPUE by gear Total Total taxa Total REHAGE AND LOFTUS: EVERGLADES MANGROVE CREEKS AS DRY-SEASON REFUGES 633

Figure 4. (A) Electrofishing and (B) gill net catch-per-unit effort (CPUE) over the three sampling seasons in the two study regions. Shown are means ± 1 SE.

CPUEs, Table 2), but season had a different effect in the two sampled regions. Across sampling gears, CPUE was highest in RB samples during the transition period. CPUE increased four-fold in electrofishing samples, eight-fold in gill nets, and nine-fold in minnow traps between the wet and transition samples (electrofishing, P = 0.009; gill nets, P = 0.002; minnow traps, P = 0.0001). The abundance of large species, such as Florida gar, bowfin, snook, largemouth bass, Mayan cichlid, and sunfishes peaked in the transition period. Among the small fishes, catches of bluefin killifish, eastern mosquitofish, coastal shiners, and smaller-bodied sunfishes also peaked during the transition period (Table 1). RB CPUE decreased significantly, returning to wet-season levels, in the dry season for electrofishing and minnow traps, but not for gill nets (electrofishing, P = 0.003; minnow traps, P = 0.0001). In NW creeks, electrofishing CPUE was highest in the wet season (wet vs dry, P = 0.03; wet vs transition, P = 0.01; Fig. 4A), whereas no seasonal variation was detected in minnow trap nor gill net CPUE (Figs. 4B, 5). There was a trend for electrofishing CPUE to be negatively related to salinity (Table 2). The relationship had a relatively better fit in NW than in RB creeks (NW, P = 0.0001, r2 = 0.27; RB, P = 0.045, r2 = 0.08; Fig. 6). We detected no relationship between salinity and gill net CPUE (Table 2). All gears varied significantly among creeks within the two study regions, and this variation was affected by season (Table 2). In NW, electrofishing CPUE was higher in North River creeks than in Watson River creeks (P = 0.003), although CPUE in gill nets and minnow traps did not differ. Seasonally, electrofishing CPUE was higher in North River headwaters in the transition and dry-season samples (P = 0.06 and P = 0.0006, respectively), but not in the wet season. In RB, electrofishing CPUE was lower in Squawk Creek (RB 7, Fig. 1) than in other creeks, particularly in the transition sample (P < 0.05). Minnow trap and gill net CPUE were significantly higher in Otter Creek (RB 12, Fig. 1) than in Squawk Creek (P = 0.0006 and P = 0.02, respectively), while CPUE in other RB creeks was intermediate. 634 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Figure 5. CPUE in minnow traps shown separately by trap placement: (A) top vs (B) bottom of water column) over the three sampling seasons and in the two study regions. Shown are means ± 1 SE.

Large-fish Community Structure.—The composition of electrofishing catches varied equally between regions and among seasons (Table 3). Community structure was similar between NR and RB creeks in the wet season, but diverged considerably during the transition and dry seasons (wet, R = 0.12, P = 0.14; transition, R = 0.54, P = 0.002; dry, R = 0.74, P = 0.002; Fig. 7A). This divergence can be explained by increases in the relative contribution of freshwater taxa to the creek community. The contribution of freshwater species to CPUE was comparable between regions in the wet season (5% in NW vs 20% in RB; P = 0.225), but differed significantly in later samples (Fig. 8A). NW catches remained < 10% freshwater, whereas in RB, 80%–90% of the catch was composed of freshwater taxa during the transition and dry seasons (transition, P = 0.0011; dry season, P = 0.0001). During these drier samples, Florida gar, largemouth bass, bowfin, Mayan cichlid, and several sunfish species were almost

Table 2. Summary of results of nested, repeated-measures ANOVA and ANCOVAs testing the effects of season, region, creek, placement (for minnow traps only), and salinity on catch-per-unit- effort (CPUE) from electrofishing, gill nets, and minnow traps.

Electrofishing CPUE Gill-net CPUE Minnow-trap CPUE Source of variation df F P df F P df F P Season 2, 42 6.7 0.0031 2, 14 3.7 0.0502 2, 200 26.9 0.0001 Region 1, 55 60.4 0.0001 1, 11 9.3 0.0113 1, 220 46.8 0.0001 Placement 1, 112 2.6 0.1077 Season × region 2, 45 14.5 0.0001 2, 16 4.1 0.0368 2, 200 31.7 0.0001 Season × placement 2, 198 21 0.0001 Region × placement 1, 112 19.1 0.0001 Region × season × placement 2, 198 9.5 0.0001 Creek (region) 3, 38 8.4 0.0002 3, 6 6.9 0.0202 4, 140 4.8 0.0012 Creek (region) × season 6, 37 2.3 0.0558 6, 14 0.8 0.5973 6, 201 4.1 0.0006 Salinity 1, 41 3.8 0.058 1, 13 0.01 0.9453 REHAGE AND LOFTUS: EVERGLADES MANGROVE CREEKS AS DRY-SEASON REFUGES 635

Figure 6. Estimates of large fish abundance (CPUE) in electrofishing samples (Log-transformed) plotted as a function of salinity. Separate least-squares regressions were fitted to the two regions: Rookery Branch (RB, solid line) and the North and Watson rivers (NW, dotted line).

exclusively caught in RB creeks. The relative abundance of snook was also higher in RB, whereas needlefishes and tidewater mojarras were exclusively caught in NW creeks. Composition of gill net samples was similar between regions during the wet season, but tended to differ in the dry season sample (wet, R = 0.29, P = 0.20; dry, R = 0.65, P = 0.10; Fig. 7B). Florida gar was dominant in RB gill net samples, whereas NW gill nets were dominated by a small number of striped mojarras. Small-fish Community Structure.—Variation in the small community structure was higher between regions than among seasons (Table 3, Fig. 9A), yet the proportion of freshwater species in traps varied as a function of both season and region (Table 4). The contribution of freshwater species to the RB small fish fauna showed no seasonal variation, averaging 96% throughout the study (Fig. 8B). In NW, however, the contribution of freshwater species decreased significantly between the wet and dry seasons, from 24% to 2% (P = 0.024). Minnow trap CPUE in NW primarily contained estuarine species (rainwater killifish, tidewater mojarra, and clown goby), whereas the RB community primarily contained freshwater species (eastern mosquitofish, sailfin molly, bluefin killifish, least killifish, dollar sunfish, and bluespotted sunfish; Table 1). Gear Comparison.—Large fish catches averaged 16.5 fish 5-min–1 bout in electrofishing samples, whereas gill nets only averaged 2.3 fish per 30-min set (Fig. 4). Gill nets failed to detect the marked seasonal variation in the numbers of large fresh- Table 3. Summary of ANOSIM results testing variation in fish-community structure as a function of region, season, and gear in electrofishing, gill net, and minnow trap samples.

Region Season Gear comparison Target community Sampling method Global R P Global R P Global R P Large fishes (> 50 mm) Electrofishing 0.446 0.001 0.416 0.001 0.279a 0.002 Large fishes (> 50 mm) Gill nets 0.511 0.019 0.275 0.012 Large fishes (> 50 mm) Minnow traps 0.641 0.001 0.1 0.001 0.135b 0.002 a Comparison of electrofishing and gill nets b Comparison of minnow trap placement: top vs bottom of water column 636 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Figure 7. Two-dimensional non-metric MDS ordinations illustrating large fish (50–750 mm SL) community structure in (A) electrofishing samples, (B) gill net samples, and (C) gill net vs electrofishing samples based on Bray-Curtis similarities of log-transformed, standardized CPUE. water fishes present in creeks seen in the electrofishing data, particularly in RB (Fig. 4A). Electrofishing and gill-net samples also differed significantly in composition, although this dissimilarity was less than that observed as a function of spatial or seasonal factors (Table 3, Fig. 7C). Florida gar was the most abundant species caught using both methods, but numbers caught by electrofishing were higher than those caught using gill nets. CPUE of snook, striped mullet, largemouth bass, mojarras, and largemouth bass were also higher in electrofishing samples than in gill nets. In minnow traps, mean CPUE was similar between the top and bottom trap, but placement affected the magnitude of CPUE variation across seasons and between regions (Table 2). For instance, CPUE in RB doubled between the wet and transition REHAGE AND LOFTUS: EVERGLADES MANGROVE CREEKS AS DRY-SEASON REFUGES 637

Figure 8. Average proportion of (A) electrofishing and (B) minnow trap CPUE composed of freshwater species over the three sampling seasons in the two study regions. Only species known to occur in freshwater marshes (listed as FW in Table 1) were included in these analyses. Shown are means ± 1 SE. samples in the top trap, but increased by 50 times in the bottom trap (P = 0.0001 for both cases, Fig. 5). Large numbers of sunfishes, bluefin killifish, and coastal shiners accounted for this increase in the bottom trap (Table 1). In NW headwaters, seasonal variation in CPUE was detected only in the top trap. In spite of very low catches, CPUE increased between the transition period and the dry season (P = 0.0001; Fig. 5A). Dissimilarity between minnow trap samples as a function of trap placement was lower than the separation observed when comparing gill nets and electrofishing samples (Table 3, Fig. 9B). The contribution of freshwater species was higher in the top than in the bottom trap (80.6% and 54.2%, respectively; Table 4). Eastern mosquitofish and least killifish were more abundant in traps placed at the top of the water column, whereas bluefin killifish, rainwater killifish, clown gobies, and dollar sunfish were more abundant in traps placed at the bottom of the water column. Table 4. Summary of nested, repeated-measures ANCOVA and ANOVA analyses testing variation in the proportion of CPUE that is composed of freshwater taxa (FW in Table 1) per sample as a function of region, season, and placement (for minnow traps only) in electrofishing and minnow trap samples.

Proportion of freshwater CPUE Electrofishing samples Minnow trap samples Source of variation df F P df F P Season 2, 61 5.8 0.0049 2, 129 5.7 0.0044 Region 1, 58 38.7 0.0001 1, 128 199.7 0.0001 Placement 1, 103 18.1 0.0001 Season × region 2, 63 16.9 0.0001 2, 130 4.1 0.0191 Season × placement 2, 129 5.1 0.007 Region × placement 1, 103 0.1 0.7619 Region × season × placement 2, 129 7.9 0.0006 Creek (region) 3, 29 3.4 0.0319 4, 105 0.9 0.4609 Creek (region) × season 6, 53 1.1 0.3533 6, 115 1.1 0.3586 Salinity 1, 61 0.04 0.8399 638 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Figure 9. Two-dimensional non-metric MDS ordinations illustrating variation in small fish community structure (< 5 cm SL): (A) among seasons and sites and (B) as a function of trap placement. MDS plots are based on Bray-Curtis similarities of standardized minnow traps CPUE.

Discussion

Disturbance from recurrent, seasonal dry-down events has a strong structuring effect on freshwater fish communities inhabiting Everglades marshesK ( ushlan, 1974; Loftus and Eklund, 1994; Trexler et al., 2001, 2005; Chick et al., 2004). In response to dry-down, fish move from marshes to deeper habitats such as alligator holes, solution holes, and canals (Nelson and Loftus, 1996; Trexler et al., 2001; Kobza et al., 2004; Rehage and Trexler, 2006). Thus, access to dry-season refuges is a key element underlying long-term population dynamics in freshwater fishes D ( eAngelis et al., 1997). Our results indicate that mangrove creek headwaters in the southern part of the ecosystem can also serve as important dry-season refugia, particularly for large- bodied species, whose abundance is strongly limited by seasonal dry-down (Trexler et al., 2001, 2005; Chick et al., 2004). A pulse of freshwater fishes was detected in RB creeks in February as marshes upstream began to dry periodically, which resulted in marked seasonal variation in patterns of abundance and composition in RB headwaters. The pulse was composed of both predatory species, such as Florida REHAGE AND LOFTUS: EVERGLADES MANGROVE CREEKS AS DRY-SEASON REFUGES 639 gar, bowfin, and centrarchids; and of the small cyprinodontoids, although some species (e.g., mosquitofish) appeared to reside in creek all-year around. No evidence of a similar pulse was noted in the North and Watson rivers, where the contribution of freshwater taxa to the community was consistently small (0%–24%) and showed lower seasonal variation. Mangrove fish communities are highly variable in both short (tidal) and longer time scales (seasonal) because of pronounced environmental fluctuations K( upschus and Tremain, 2001). Seasonal changes in the abundance and composition of tropical and subtropical fish communities have been reported in mangrove systems throughout the world, including Madagascar (Laroche et al., 1997), Brazil (Barletta et al., 2005), Australia (Loneragan et al., 1986), the Solomon Islands (Blaber and Milton, 1990), Taiwan (Lin and Shao, 1999), and Mexico (Yanez-Arancibia et al., 1988). Of these examples, mangrove creeks of the Caeté River estuary in Brazil (Barletta et al., 2005) exhibit the greatest freshwater inflow and may closely resemble the mangrove creeks at our Everglades sites. However, the directionality of seasonal variation in RB and NW creeks is opposite that of Caeté, where the influx of freshwater species occurs during the wet season when salinities are low. Seasonal community dynamics have also been shown in Everglades mangrove regions (Thayer et al., 1987; Ley et al., 1999; Lorenz et al., 1999; Faunce et al., 2004). In Florida Bay and Whitewater Bay, Thayer et al. (1987) reported increases in both fish numbers and biomass during the wet season. Our results from NW fit their findings. Fish abundances varied monthly in mangrove creeks of southeastern Florida Bay, with those of freshwater species increasing during February and March (Faunce et al., 2004), as seen in our RB sites. Several factors may be responsible for the lack of a freshwater species influx in NW headwaters. Marshes upstream of NW creeks have shorter hydroperiods than those upstream of RB sites, and consequently may contain lower densities of fishes, particularly of the large species (Lorenz, 1999; Trexler et al., 2001, 2005; Chick et al., 2004). Freshwater fishes may be absent from NW creeks simply because the pool of potential marsh migrants is small. Secondly, salinity levels are higher in NW than RB headwaters, and may approach or exceed the physiological tolerances or preferences for some of the freshwater species such as centrarchids (Loftus and Kushlan, 1987). However, other marsh inhabitants exhibit high salinity tolerances (Lorenz and Serafy, 2006; Nordlie, 2006), and should find suitable salinity conditions in NW, despite the fact that there were rarely caught there. Thirdly, the pattern of marsh dry-down differed between regions, and marshes upstream of NW remained flooded beyond our dry-season sample. A pulse of freshwater species could have possibly occurred later in the season, and would have been missed by our sampling. Other studies, however, suggest that marsh fishes move into deep-water refugia well in advance of low-water conditions (Chick et al., 2004; Re- hage and Trexler, 2006). Even in long-hydroperiod marshes that rarely dry, and where direct mortality due to dry-down conditions is unlikely, large-fish densities decrease significantly in the open marsh during the dry season and concentrate in deep-water refuges. Marsh water-levels upstream of NW were low (close to 5 cm); therefore, a pulse of migrants should have occurred by our April sample. Furthermore, salinities in later months exceeded 15 in NW, which is too stressful for many of the potential migrants. Another explanation may be a higher abundance of alternative dry-down refuges (e.g., solution and alligator holes) in marshes upstream of NW relative to RB. However, abiotic (high ammonia and low oxygen) and biotic (high predation) condi- 640 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

tions in these alternative refuges are often stressful (Nelson and Loftus, 1996; Kobza et al., 2004), and could make these refuges less preferred relative to creeks. Lastly, small differences in local topography (e.g., the presence of berms along creeks) could limit fish movement in and out of creeks, perhaps to a greater extent in NW than RB sites (Green et al., 2006). The pulse of freshwater species in RB occurred early in the dry season, and despite the fact that RB assemblages remained dominated by freshwater species in the later sample, their abundances decreased considerably. This decrease could be explained by a large-scale return of the freshwater taxa to marshes, if marshes had reflooded. However, this was unlikely in 2005 since water levels in upstream marshes at the time of our transition and dry season samples were relatively low and similar (11 and 13 cm, respectively). Alternatively, the increase in salinity between the transition and dry samples in RB (from < 1 to 5) could explain the decline in freshwater species through mortality or movement to more suitable salinity environments. An important source of mortality could also be predation. Mangrove habitats provide important foraging grounds for marine and estuarine piscivores (Blaber and Milton, 1990; Chong et al., 1990). Even though mangroves can provide a refuge from predation because of their high habitat complexity (Primavera, 1998; Acosta and Butler, 1999), the abundance of predators is not necessarily lower in these shallow coastal habitats (Sheaves, 2001). In RB creeks, piscine predators such as snook, Florida gar, largemouth bass, and bowfin were abundant early in the dry season and could account for the declines in the cyprinodontoids between February and April. Top predators, such as alligators, wading birds, and bull sharks could possibly account for the decreases in the abundance of the larger freshwater species. More extensive, paired sampling in creeks, marshes (including other dry-down refuges), and downstream portions of the estuary is needed to discriminate between these and other plausible explanations for the timing and extent of pulsing of freshwater taxa into headwater creeks. Results from studies that rely on a single gear type to sample mangrove fishes may be restricted in their applicability because of gear selectivity (Rozas and Minello, 1997). CPUE in gill nets was appreciably lower than that for electrofishing, and catch composition differed between gears. This suggests that gill nets with the 30-min soak times used in our study do not provide a reliable index of abundance, nor detect seasonal variation in community structure, even if previous studies have shown that gill nets may be better at capturing certain aspects of target fish populations, such as size structure, relative to electrofishing (Colvin, 2002). Comparison of the small- fish CPUE showed that minnow-trap placement in the water column strongly affects catch numbers and species composition. This is likely explained by variation in microhabitat use among small-fish species; a factor that needs to be considered if sampling is targeted to multiple species. The influx of freshwater species into RB headwater creeks may enhance estuarine fish abundance and richness, and should provide an important prey source for marine and estuarine piscine predators, as well as for avian predators. Interannual variation in drying patterns may create circumstances in which other creek headwater habitats (including NW) may serve as dry-season refugia. In other parts of the ecotone, factors such as high salinity and local topography could limit the connectivity between mangrove and upstream marsh habitats (i.e., Green et al., 2006). Ongoing restoration of the Greater Everglades ecosystem aimed at re-establishing historical freshwater flows (CERP, 1999), could greatly enhance this connectivity. REHAGE AND LOFTUS: EVERGLADES MANGROVE CREEKS AS DRY-SEASON REFUGES 641

Increased freshwater flows are expected to result in reduced salinities, prolonged pooling of freshwater, and a spatially-expanded and seasonally-extended oligohaline zone at that marsh-mangrove ecotone, including our study creeks (Davis et al., 2005). These conditions should make large portions of the mangrove region suitable for freshwater species. In northern parts of Florida Bay, increased freshwater flow has resulted in higher abundance and biomass of small-bodied freshwater taxa, and thus the recovery of the demersal forage fish community (Lorenz, 1999; Lorenz and Se- rafy, 2006). Similar effects could occur in our study area in southwestern Everglades, but overall responses of freshwater fishes are somewhat uncertain. Increased freshwater flow and decrease salinity are expected the influence multiple components of marine, estuarine, and freshwater food webs and how these interact over a complex and heterogeneous ecotonal landscape. Further research is needed to develop predictions and gain a better understanding of the net effects of hydrological restoration on the freshwater fish community of mangrove headwaters.

Acknowledgments

This study was funded by the Army Corps of Engineers (ACOE), Jacksonville District, as part of the Comprehensive Everglades Restoration Plan Monitoring and Assessment Program (Work Order 12) and supported by NSF grant No. DEB-9910514. We thank E. Kurzbach and K. Luisi of the ACOE, P. Telis, D. Elswick, and C. Fadeley of USGS, and J. Lorenz and J. Wolkowsky of Audubon of Florida for their support. We thank B. Zepp and P. Teague of ENP for their assistance with permitting and GIS. We are grateful to B. Dunker and B. Shamblin for their assistance with field work, and K. Dunker for her assistance with sampling processing.

Literature Cited

Acosta, C. A. and M. J. Butler IV. 1999. Adaptive strategies that reduce predation on spiny lobster postlarvae during onshore transport. Limnol. Oceanogr. 44: 494–501. Argent, D. G. and W. G. Kimmel. 2005. Efficiency and selectivity of gill nets for assessing fish community composition of large rivers. North Am. J. Fish. Manage. 25: 1315–1320. Barletta, M., A. Barletta-Bergan, U. Saint-Paul, and G. Hubold. 2005. The role of salinity in structuring the fish assemblages in a tropical estuary. J. Fish Biol. 66: 45–72. Beck, M. W., K. L. Heck Jr., K. W. Able, D. L. Childers, D. B. Eggleston, B. M. Gillanders, B. Halpern, C. G. Hays, K. Hoshino, T. J. Minello, R. J. Orth, P. F. Sheridan, and M. P. Weinstein. 2001. The identification, conservation and management of estuarine and marine nurseries for fish and invertebrates. BioScience 51: 633–641. Blaber, S. J. M. and D. A. Milton. 1990. Species composition, community structure and zooge- ography of fishes of mangrove estuaries in the Solomon Islands. Mar. Biol. 105: 259–267. Burkhardt, R. W. and S. Gutreuter. 1995. Improving electrofishing catch consistency by stan- dardizing power. North Am. J. Fish. Manage. 15: 375–381. Comprehensive Everglades Restoration Plan (CERP). 1999. U.S. Army Corps of Engineers and South Florida Water Management District. Jacksonville FL. Available from: http://www. evergladesplan.org/pub/restudy_eis.aspx. Chaves, P. and J. L. Bouchereau. 2000. Use of mangrove habitat for reproductive activity by the fish assemblage in the Guaratuba Bay, Brazil. Oceanol. Acta 23: 273–280. Chick, J. H., S. Coyne, and J. C. Trexler. 1999. Effectiveness of airboat electrofishing for sampling fishes in shallow vegetated habitats. North Am. J. Fish. Manage. 19: 957–967. ______, C. R. Ruetz, and J. C. Trexler. 2004. Spatial scale and abundance patterns of large fish communities in freshwater marshes of the Florida Everglades. Wetlands 24: 652–664. 642 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Chong, V. C., A. Sasekumar, M. U. C. Leh, and R. D’ Cruz. 1990. The fish and prawn communities of a Malaysian coastal mangrove system, with comparison to adjacent mud flats and inshore waters. Estuar. Coast. Shelf Sci. 31: 703–722. Clarke, K. R. and R. M. Warwick. 2001. Changes in marine communities: an approach to statistical analyses and interpretation, 2nd edition. National Environmental Research Council, Plymouth Marine Laboratory, Plymouth, United Kingdom. 124 p. Colvin, M. 2002. A comparison of gill netting and electrofishing as sampling techniques for white bass in Missouri’s large reservoirs. North Am. J. Fish. Manage. 22: 690–702. Davis, S. M., D. L. Childers, J. J. Lorenz, H. R. Wanless, and T .E. Hopkins. 2005. A concep- tualmodel of ecological interactions in the mangrove estuaries of the Florida Everglades. Wetlands 25: 832–842. DeAngelis, D. L., W. F. Loftus, J. C. Trexler, and R. E. Ulanowicz. 1997. Modeling fish dynamics in a hydrologically pulsed ecosystem. J. Aquatic Ecosyst. Stress and Recov. 6: 1–13. Dorenbosch, M., M. C. van Riel, I. Nagelkerken, and G. van der Velde. 2004. The relationship of reef fish densities to the proximity of mangrove and seagrass nurseries. Estuar. Coast. Shelf Sci. 60: 37–48. Ellis, W. L. and S. S. Bell. 2004. Conditional use of mangrove habitats by fishes: depth as a cue to avoid predators. Estuaries 27: 966–976. Faunce, C. H. and J. E. Serafy. 2006. Mangroves as fish habitat: 50 years of field studies. Mar. Ecol. Prog. Ser. 318: 1–18. ______, ______, and J. J. Lorenz. 2004. Density-habitat relationships of mangrove creek fishes within the southeastern saline Everglades (USA), with reference to managed water releases. Wetland Ecol. Manag. 12: 377–394. Fennema, R. J., C. J. Neidrauer, R. A. Johnson, T. K. MacVicar, and W. A. Perkins. 1994. A computer model to simulate natural Everglades hydrology. Pages 249–289 in S. M. Davis and J. C. Ogden, eds. Everglades: the ecosystem and its restoration. St. Lucie Press, Delray Beach. Green, D. P. J., J. C. Trexler, J. J. Lorenz, C. C. McIvor, and T. Philippi. 2006. Spatial patterns of fish communities along two estuarine gradients in southern Florida. Hydrobiologia 569: 387–399. Gillanders, B. M., K. W. Able, J. A. Brown, D. B. Eggleston, and P. F. Sheridan. 2003. Evidence of connectivity between juvenile and adult habitats for mobile marine fauna: an important component of nurseries. Mar. Ecol. Prog. Ser. 247: 281–295. Gunderson, L. H. and W. F. Loftus. 1993. The Everglades. Pages 199–255 in W. H. Martin, S. G. Boyce, and A. C. Echternacht, eds. Biodiversity of the Southeastern United States. John Wiley and Sons, New York. Hindell, J. S. and G. P. Jenkins. 2004. Spatial and temporal variability in the assemblage structure of fishes associated with mangroves,Avicennia marina, and intertidal mudflats in temperate Australian embayments. Mar. Biol. 144: 385–395. Hubert, W. A. and D. T. O’Shea. 1992. Use of spatial resources by fishes inG rayrocks Reservoir, WY. J. Freshw. Ecol. 7: 219–225. Kobza, R. M, J. C. Trexler, W. F. Loftus, and S. A. Perry. 2004. Community structure of fishes inhabiting aquatic refuges in a threatened Karst wetland and its implications for ecosystem management. Biol. Conserv. 13: 898–911. Kuo, S. R., H. J. Lin, and K. T. Shao. 1999. Fish assemblages in the mangrove creeks of northern and southern Taiwan. Estuaries 22: 1004–1015. Kupschus, S. and D. Tremain. 2001. Associations between fish assemblages and environmental factors in nearshore habitats of a subtropical estuary. J. Fish Biol. 58: 1383–1403. Kushlan, J. A. 1974. Effects of a natural fish kill on the water quality, plankton, and fish population of a pond in the Big Cypress Swamp, Florida. Trans. Am. Fish. Soc. 103: 235–243. Laegdsgaard, P. and C. R. Johnson. 1995. Mangrove habitats as nurseries: unique assemblages of juvenile fish in subtropical mangroves in eastern Australia. Mar. Ecol. Prog. Ser. 126: 67–81. REHAGE AND LOFTUS: EVERGLADES MANGROVE CREEKS AS DRY-SEASON REFUGES 643

______and ______. 2001. Why do juvenile fish utilize mangrove habitats? J. Exp. Mar. Biol. Ecol. 257: 229–253. Laroche, J., E. Baran, and N. B. Rasoanandrasana. 1997. Temporal patterns in fish assemblage of a semiarid mangrove zone in Madagascar. J. Fish. Biol. 51: 3–20. Layman, C. A. and D. E. Smith. 2001. Sampling bias of minnow traps in shallow aquatic habitats on the eastern shore of Virginia. Wetlands 21: 145–154. Levesque, V. A. 2004. Water flow and nutrient influx from five estuarine rives along the southwest coast of the Everglades National Park, Florida, 1997–2001. U.S. Geological Survey Scientific Investigations Report 2004-5142. Ley, J. A., C. C. McIvor, and C. L. Montague. 1999. Fishes in mangrove prop-root habitats of northeastern Florida Bay: distinct assemblages across an estuarine gradient. Estuar. Coast. Shelf Sci. 48: 701–723. Light, S. S. and J. W. Dineen. 1994. Water control in the Everglades: a historical perspective. Pages 47–84 in S. M. Davis and J. C. Ogden, eds. Everglades: the system and its restoration. St. Lucie Press, Delray Beach. Lin, H. J. and K. T. Shao. 1999. Seasonal and diel changes in a subtropical mangrove fish assemblage. Bull. Mar. Sci. 65: 775–794. Loftus, W. F. 2000. Inventory of fishes of Everglades National Park. Fla. Sci. 63: 27–47. ______and A. M. Eklund. 1994. Long-term dynamics of an Everglades fish community. Pages 461–483 in S. M. Davis and J. C. Ogden, eds. Everglades: the system and its restoration. St. Lucie Press, Delray Beach. ______and J. A. Kushlan. 1987. Freshwater fishes of southern Florida. Bull. Fla. Mus. Nat. Hist. Biol. Sci. 31: 147–344. Loneragan, N. R., I. C. Potter, R. C. J. Lenanton, and N. Caputi. 1986. Spatial and seasonal differences in the fish fauna of the shallows in a large Australian estuary. Mar. Biol. 92: 575–586. Lorenz, J. J. 1999. The response of fishes to physiochemical changes in the mangroves of northeast Florida Bay. Estuaries 22: 500–517. ______and J. E. Serafy. 2006. Subtropical wetland fish assemblages and changing salinity regimes: implications for Everglades restoration. Hydrobiologia 569: 401–422. Lugo, A. and S. Snedaker. 1974. The ecology of mangroves. Ann. Rev. Ecol. System. 5: 39–64. Mason, F. J., N. R. Loneragan, G. A. Skilleter, and S. R. Phinn. 2005. An evaluation of the evidence for linkages between mangroves and fisheries: a synthesis of the literature and identification of research directions. Oceanogr. Mar. Biol. Annu. Rev. 43: 485–515. McIvor, C. C., J. A. Ley, and R. D. Bjork. 1994. Changes in freshwater inflow from the Everglades to Florida Bay including effects on biota and biotic processes: a review. Pages 533–570 in S. M. Davis and J. C. Ogden, eds. Everglades: the system and its restoration. St. Lucie Press, Delray Beach. Montague, C. L. and J. A. Ley. 1993. A possible effect of salinity fluctuation on abundance of benthic vegetation and associated fauna in northeastern Florida Bay. Estuaries 16: 703– 717. Mumby, P. J., A. J. Edwards, J. E. Arlas Gonzales, K. C. Lindeman, P. G. Blackwell, A. Gall, M. I. Gorczynska, A. R. Harborne, C. L. Pescod, H. Renken, C. C. C. Wabnitz, and G. Llewellyn. 2004. Mangroves enhance the biomass of coral reef fish communities in the Caribbean. Nature 427: 533–536. Nagelkerken, I., G. van der Velde, M. W. Gorrissen, G. J. Meijer, T. van’t Hof, and C. den Har- tog. 2000. Importance of mangroves, seagrass beds, and shallow coral reef as a nursery for important coral reef fishes, using a visual census technique. Estuar. Coast. Shelf Sci. 51: 31–44. ______, C. M. Roberts, G. van der Velde, M. Dorenbosch, M. C. van Riel, E. Cocheret de la Morinière, and P. H. Nienhuis. 2002. How important are mangrove and seagrass beds for coral-reef fish? The nursery hypothesis tested on an island scale. Mar. Ecol. Prog. Ser. 244: 299–305. 644 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Nelson, C. M. and W. F. Loftus. 1996. Effects of high-water conditions on fish communities in Everglades alligator ponds. Pages 89–101 in T. V. Armentano, ed. Proc. 1996 conference: ecological assessment of the 1994–1995 high water conditions in the southern Everglades. Florida International University, Miami, FL. Nordlie, F. G. 2003. Fish communities of estuarine salt marshes of eastern North America, and comparisons with temperate estuaries of other continents. Rev. Fish Biol. Fish. 13: 281–325. ______. 2006. Physicochemical environments and tolerances of cyprinodontoid fishes found in estuaries and salt marshes of eastern North America. 2006. Rev. Fish. Biol. Fish. 16: 51–106. Odum, W. E. 1971. Pathways of energy flow in a south Florida estuary. Sea Grant Tech. Bull. No. 7. Univ. of Miami Sea Grant Program, Miami, FL. 162 p. Ogden, J. C. 1994. A comparison of wading bird nesting colony dynamics (1931–1946 and 1974–1989) as an indication of ecosystem conditions in the southern Everglades. Pages 533–570 in S. M. Davis and J. C. Ogden, eds. Everglades: the system and its restoration. St. Lucie Press, Delray Beach. Pinto, L. and N. N. Punchihewa. 1996. Utilization of mangroves and seagrasses by fishes in the Negombo estuary, Sri Lanka. Mar. Biol. 126: 333–345. Primavera, J. H. 1998. Mangroves�� as nurseries: shrimp populations in mangrove and non-mangrove habitats. Estuar. Coast. Shelf Sci. 46: 457–464. Rehage, J. S. and J. C. Trexler. 2006. Assessing the net effect of anthropogenic disturbance on aquatic communities in wetlands: community structure relative to distance from canals. Hydrobiologia 569: 359–373. Robertson, A. I. and N. C. Duke. 1987. Mangroves as nursery sites: comparisons of the abundance and species composition of fish and crustaceans in mangroves and other nearshore habitats in tropical Australia. Mar. Biol. 96: 193–205. ______, P. Dixon, and P. A. Daniel. 1988. Zooplankton dynamics in mangrove and other nearshore habitats in tropical Australia. Mar. Ecol. Prog. Ser. 43: 139–150. Rozas, L. P. and T. J. Minello. 1997. Estimating densities of small fishes and decapod crustaceans in shallow estuarine habitats: A review of sampling design with focus on gear selection. Estuaries 20: 199–213. Serafy, J. E., C. H. Faunce, and J. J. Lorenz. 2003. Mangrove shoreline fishes of Biscayne Bay, Florida. Bull. Mar. Sci. 72: 161–180. ______, K. C. Lindeman, T. E. Hopkins, and J. S. Ault. 1997. Effects of freshwater canal discharge on fish assemblages in a subtropical bay: field and laboratory observations. Mar. Ecol. Prog. Ser. 160: 161–172. Sheaves, M. J. 2001. Are there really few piscivorous fishes in shallow estuarine habitats? Mar. Ecol. Prog. Ser. 222: 272–290. ______. 2005. Nature and consequences of biological connectivity in mangrove systems. Mar. Ecol. Prog. Ser. 302: 293–305. Sheridan, P. F. 1997. Benthos of adjacent mangrove, seagrass, and non-vegetated habitats in Rookery Bay, Florida, U.S.A. Estuar. Coast. Shelf Sci. 44: 455–469. ______and C. Hays. 2003. Are mangroves nursery habitat for transient fishes and decapods? Wetlands 23: 449–458. Smith, T. J. III, M. B. Robblee, H. R. Wanless, and T. W. Doyle. 1994. Mangroves, hurricanes, and lighting strikes. Bioscience 44: 256–262. ______, J. H. Hudson, M. B. Robblee, G. V. N. Powell, and P. J. Isdale. 1989. Freshwater flow from the Everglades to Florida Bay: a historical reconstruction based on fluorescent banding in the coral Solenastrea bournoni. Bull. Mar. Sci. 44: 274–282. Strahler, A. N. 1957. Quantitative analysis of watershed geomorphology. Amer. Geophys. Union T. 38: 913–920. Tabb, D. C. and R. B. Manning. 1961. Checklist of flora and fauna of north Florida Bay. Bull. Mar. Sci. 11: 552–640. REHAGE AND LOFTUS: EVERGLADES MANGROVE CREEKS AS DRY-SEASON REFUGES 645

Thayer, G. W., D. R. Colby, and W. F. Hettler. 1987. Utilization of the red mangrove prop root habitat by fishes in South Florida. Mar. Ecol. Prog. Ser. 35: 25–38. Trexler, J. C., Ww.. F. Loftus, and S. Perry. 2005. Hhydrologicalydrological limitation of Everglades fishfi sh comcom-- munities by a twenty-five year intervention study. Oecologia 145: 140–152. ______, W. F. Loftus, C. F. Jordan, J. H. Chick, K. L. Kandl, T. C. McElroy, and O. L. Bass. 2001. Ecological scale and its implications for freshwater fishes in the Florida Everglades. Pages 153–181 in J. W. Porter and K. G. Porter, eds. The Everglades, Florida Bay, and coral reefs of the Florida Keys: An ecosystem sourcebook. CRC Press, Boca Raton. Yañez-Arancibia, A., A. L. Lara-Domínguez, J. L. Rojas-Galaviz, P. Sánchez-Gil, J. W. Day, and C. J. Madden. 1988. Seasonal biomass and diversity of estuarine fishes coupled with tropical habitat heterogeneity (southern Gulf of Mexico). J. Fish. Biol. 33: 191–200.

Addresses: (J.S.R.) Nova Southeastern University, Oceanographic Center, 8000 North Ocean Drive, Dania Beach, Florida 33004-3078. (W.F.L.) U.S. Geological Survey, Florida Integrat- ed Science Center, Everglades National Park Field Station, 40001 State Road 9336, Florida 33034. Corresponding Author: (J.S.R.) Email: . BULLETIN OF MARINE SCIENCE, 80(3): 647–680, 2007

Comparison of fish assemblages and guilds in tropical habitats of the Embley (Indo-West Pacific) and Caeté (Western Atlantic) estuaries

Mário Barletta and Stephen J. M. Blaber

ABSTRACT Fish assemblages of two tropical estuaries are compared with regard to taxonomic structure and functional guilds using biomass data from different habitats of the Embley Estuary (north Australia) and the Caeté Estuary (north Brazil). First, the similarity of the habitats of the two estuaries was compared using taxonomic classification of their respective ichthyofauna. Important taxonomic differences between the two estuaries emerged: the Neotropical fish species of the upper Caeté Estuary have no equivalents in the Embley Estuary, and the diverse chondrichthyan fauna of the Embley have no equivalents in the Caeté Estuary. On the other hand, the more ubiquitous families Engraulidae, Sciaenidae, Ariidae, Carangidae, Haemulidae, and Clupeidae, which were characteristic of the main channel and mangrove tidal creeks of the Caeté Estuary, showed 70% similarity with the main channel of the Embley Estuary. Second, the similarity of the fish assemblages was compared in terms of ecological guilds (estuarine use and feeding mode). For the Embley Estuary, marine immigrants (mainly piscivorous and benthophagous) were the dominant functional guild, whereas in the Caeté Estuary, the estuarine functional guild (mainly benthophagous) contributed to more than 50% of the biomass in each habitat. Factorial analysis of species functional guilds shows that the first factorial axis was formed positively by the dominant fish species which spend part of their life cycle in Embley Estuary habitats, principally in the main channel, tidal channel (anadromous/piscivorous, marine immigrants/piscivorous, benthophagous, detritivorous, and marine stragglers/piscivorous), and negatively by dominant estuarine (planktivorous and benthophagous) fish species which spend their life in the seagrass beds and in the mangrove intertidal creeks. The second factorial axis best represented the distribution of fish species in the estuary (upper, middle, and lower), which was evident in the Caeté than in the Embley Estuary, where the freshwater and marine straggler guilds were most important in determining this second factorial axis.

Organisms, organic matter, and nutrients are transferred between freshwater and the ocean via estuaries and nearshore coastal waters (Lee, 1995; Wolanski, 1995; Bar- letta-Bergan et al., 2002a,b; Barletta et al., 2003, 2005). Fish landings around the world are composed of species that spend part of their lives in estuarine waters (Barthem, 1985; Pauly and Yáñes-Arancibia, 1994; Barthem and Goulding, 1997; Barletta et al., 1998; Blaber, 2000; Islam and Haque, 2004). Like estuaries in temperate regions (Thiel et al., 1995), tropical (Blaber and Blaber, 1980; Barletta-Bergan et al., 2002a,b; Barletta et al., 2003, 2005; Dorenbosch et al., 2005) and sub-tropical estuaries (Jau- reguizar et al., 2004) act as nursery areas for many invertebrate and fish species. The species composition of estuarine fish assemblages is dictated by a combination of biotic and abiotic variables, particularly competition for space and food, tolerance of diel and seasonal changes in salinity, turbidity, and temperature (Cyrus and Blaber, 1987a,b; Blaber et al., 1989, 1990; Cyrus and Blaber, 1992; Barletta et al., 2003, 2005; Krumme et al., 2005). An understanding of the roles and relative importance to fish of different habitats within estuaries is necessary for both management and conser-

Bulletin of Marine Science 647 © 2007 Rosenstiel School of Marine and Atmospheric Science of the University of Miami 648 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 vation (Elliott and Hemingway, 2002). Although there have been many comparative studies of the fish assemblages of estuaries and their relationships with habitat types and physical conditions, they have generally been restricted to one zoogeographic region. Such studies have demonstrated that factors such as geology, geomorphology, and more immediate environmental factors such as salinity and temperature, are associated with fish distributions, species richness, and fisheries catch (Vieira and Musik, 1993, 1994; Mahon et al., 1998; Mathieson et al., 2000; Araújo and Azevedo, 2001; Roy et al., 2001; Thiel et al., 2003). In addition, broad-scale comparisons of tropical estuaries across zoogeographic regions have indicated possible significant differences in the ways in which their fish assemblages use the various estuarine habitats, as well as differences in their patterns of species composition (Blaber, 2000). For example, it has been suggested that the relative proportions of freshwater and marine species using estuaries may be different (Blaber, 2000; Barletta et al., 2003, 2005). To explore possible fundamental similarities and differences in tropical estuarine fishes from different zoogeographic realms, it is necessary to compare their fish assemblages, not only taxonomically, but also in terms of their ecological structure and resource use in different habitats of the estuary. In this paper, the fish assemblages of two tropical estuaries, one in the Indo-West Pacific and one in the Tropical West Atlantic, are compared with regard to taxonomic structure, estuarine functional groupings, and habitat use.

Material and Methods

Source of Data and Study Sites Previously unpublished and published data sets of the authors were used for this study. Analyses based on biomass percentage for different habitats of each estuary were used. The source of each data set is described below. Data were available for a total of 10 individual estuarine habitats of Embley Estuary (North Australia) and Caeté Estuary (North Brazil). Location characteristics and sampling methods are described below. Embley Estuary.—The Embley Estuary (12˚37´S; 141˚50´E) is approximately 59 km long, almost entirely fringed by a well developed intertidal zone of mangrove forest (2962 ha), and with a freshwater catchment of 2076 km2. It has a maximum tidal range of 2.6 m. The mangrove forest is drained by tidal channels and mangrove tidal creeks. Seagrass beds occur on the south side of the lower reaches. Samples were taken in the main channel (upper–MCUe, middle–MCMe, and lower–MCLe; monofilament gill nets–66 m of 50, 75, 100, 125, and 150 mm stretch mesh), tidal channel (middle–TCMe and upper–TCUe; multifilament block net– 50 mm stretch mesh), mangrove tidal creeks (middle–MTCe) (rotenone), intertidal sandy mud beaches (lower–BLe; seine net–60 m × 2 m of 25 mm stretch mesh fitted with a 25 mm cod-end), seagrass beds (lower–SLe; beam trawl–2 m × 1 m with 28 mm stretch mesh fitted with a 12 mm mesh cod-end was towed at 5 km hr–1 and rotenone–20 m2 sampled area) and intertidal mud flats (lower–IMLe; stake net–240 m × 2 m × 50 mm stretch mesh). Sampling methods and study area details are available in Blaber et al. (1989), Blaber et al. (1990), Blaber (2000), and Cyrus and Blaber (1992). Caeté Estuary.—Caeté River and its estuary (1˚07´S; 46˚40´E) has a length of approximately 100 km, a freshwater catchment of 3000 km2, and a inundation area (mangrove forest and wetland) of approximately 100 km2. Salinity is higher during the dry season. During the rainy season, freshwater run-off contributes to the salinity decrease throughout the estuary, and even in adjacent coastal waters. The coastal plain is a macro tidal (4–5 m) depositional system and is characterized by sand-mud and mud sediments. The mangrove forest is drained by tidal channels, mangrove tidal creeks, and mangrove intertidal creeks. In this region the mangrove forest is flooded twice daily at high tide.D uring low tide the mangrove is exposed. BARLETTA AND BLABER: FISH ASSEMBLAGES AND ESTUARINE GUILDS COMPARISON 649

Fish samples were collected in the main channel (upper–MCUc, middle–MCMc, and lower–MCLc; otter trawl net–7.72 m long and consisted of 35, 22, and 7.2 mm stretch mesh in the body, cod-end and cover, respectively), in the mangrove tidal creeks (lower–MTCc; block net–30 m × 5 m and stretch mesh 10 mm) and in the mangrove intertidal creeks during low tide (upper–MICUc, middle–MICMc, and lower–MICLc; ichthyocide–cunabí–250 g 10 m2). The tidal channel (lower–TCLc) was sampled using a seine net (15 m × 2 m and stretch mesh 10 mm) during low tide. Sampling methods and the study area details are available in Barletta (1999) and Barletta et al. (2000), (2003), and (2005).

Data Classification and Analysis The fish species were classified taxonomically and by the estuarine use and feeding mode functional groups guilds described by Elliott et al. (in press) who proposed eight estuarine use functional groups: marine stragglers, marine immigrants, estuarine species, anadromous, catadromous, aphidromous, freshwater stragglers, freshwater immigrants; and seven feeding mode functional groups: planktivorous, detritivorous, herbivorous, piscivorous, benthophagous, hyperbenthophagous, opportunistic (see Appendix). Fish nomenclature followed Nelson (1994), Eschemeyer (2006), Marceniuk and Menezes (2007), and Froese and Pauly (2006). Data were available as species biomass per unit area or catch per unit effort (CPUE), depending on the sampling methods employed. Species biomass or CPUE (in grams) for each individual habitat were summed for the entire time series and the percentage was calculated for each species per habitat. A similarity matrix (presence/absence) using Bray-Curtis index was computed for the R analysis; where the fish species (for each estuary separated) and families (both estuaries together) were considered as attributes (Clarke and Warwick, 1994). Preceding the analyses, the original data matrix for species was reduced to remove any undue effects of rare species on the analysis (Gauch, 1982). The similarity matrix using the Bray-Curtis index was computed using Primer 5 (Plymouth routines in multivariate ecological analysis: Clarke and Warwick, 1994). Multidimensional scaling analysis (MDS) was performed on a Bray-Curtis (presence/ absence) similarity index matrix where each habitat from each estuary (species level) and for both estuaries (family level) (described above) were considered as attributes. Factorial analysis of biomass data (percentage) for each estuarine functional group and for feeding mode groups for each habitat of both estuaries was conducted after log10 (x+1)-transformation (Legendre and Legendre, 1998). Statistical associations of the estuarine functional groups and feeding mode patterns were thereby quantified and compared between estuaries.

Results

Taxonomic Comparisons Patterns in the Fish Species Structure and Species-environment Associations.— Embley Estuary.—The cluster analysis distinguished two main groups among the 65 dominant species in six habitats of this estuary (Fig. 1A). Group I consisted of two sub-groups: The first sub-group was represented by species which occurred in shallow (< 2 m) intertidal beach areas of the lower estuary (sand/mud) (Leiognathus equulus, Ambassis nalua, Gerres filamentosus, Leiognathus decorus, Arrhamphus sclerolepis, Anodontostoma chacunda, Acanthopagrus berda, Pseudorhombus elevatus, Sillago analis, and Trixiphichthys weberi) and species which were common in beach (sand/mud-lower estuary), seagrass (lower estuary), and mangrove intertidal creeks (middle estuary) (Terapon jarbua, Pseudomugil gertrudae, Zenarchopterus buffonis, Sardinella albella, Tetraodon erythrotaenia, and Omobranchus rotundiceps). Sub-group 2 was represented by species characteristic of the seagrass habitat (Arothron immaculatus, Monacanthus chinensis, Acreichthys tomentosus, Siganus 650 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Figure 1. For the Embley Estuary, (A) Cluster dendrogram based on similarities of the most important species distribution in different habitats, and (B) MDS plot of the habitats [MICM, mangrove intertidal creeks middle-; SBL, beach (sand/mud) lower-; SGL, seagrass lower-; MTCU, mangrove tidal creek upper-; MTCM, mangrove tidal creek middle-; MCU, main channel upper-; MCM, main channel middle-; MCL, main channel lower- and IML, intertidal mud flat lower estuary], in which these species occur. Samples were clustered by group average of Bray-Curtis similarity index (presence/absence). Species codes are listed in the appendix. BARLETTA AND BLABER: FISH ASSEMBLAGES AND ESTUARINE GUILDS COMPARISON 651

canaliculatus, Lethrinus lentjan, Gerres longirostris, Lutjanus russellii, Secutor ruconius, Apogon rueppellii, Pelates quadrilineatus, Amniataba caudavittata, Centrog- enys vaigiensis, Epinephelus coioides, Ambassis dussumieri, and Chelonodon patoca). Group II contained three sub-groups: The first sub-group was represented by species which occur principally in the middle and upper tidal channel and mangrove intertidal creeks (Toxotes chatareus, Stolephorus carpentariae, Tylosurus punctulatus, and Sphyraena putnamae), and by species, except Carcharhinus leucas, which occur principally in the middle and upper main channel and tidal channel (Pomadasys argenteus, Gnathanodon speciosus, Epinephelus malabaricus, Megalops cyprinoides, Arius sp. 2, Nematalosa erebi, Arius proximus, Arius macrocephalus, Valamugil buchanani, Scomberoides commersonnianus, Liza subviridis, Pomadasys kaakan, Lates calcarifer, Rhynchobatus australiae, Negaprion acutidens, Drepane punctata, and Polydactylus macrochir). Sub-group 2 included species that occur mainly in the intertidal mudflats (lower estuary) Gerres( erythrourus, Tylosurus crocodilus, Himantura uarnak, Pastinachus sephen, and Taeniura lymna). The third sub-group contained species which were more frequent in the main channel (lower estuary) (Lactarius lactarius, Sciades mastersi, Rhizoprionodon acutus, Carcharhinus dussumieri, Carcha- rhinus cautus, Carcharhinus amblyrhynchoides, and Eleutheronema tetradactylum). The resulting MDS plot, based on the importance of each estuarine habitat for the dominant species shows that, in the Embley Estuary, the habitats divided in two groups (Fig. 1B). The first group (I) was the shallow water habitats located in the lower and middle estuary [seagrass beds, beach (sand/mud)] and mangrove intertidal creeks). All of these habitats are adjacent to the tidal channels of the middle and lower estuary. The second group (II) contained the deeper water habitats of the main channels of the upper, middle and lower reaches, the tidal channels of the upper and middle reaches and the intertidal mud flats of the lower reaches. Caeté Estuary.—Two main groups were also identified from this estuary (Fig. 2A). Group I can be divided in two sub-groups: Sub-group 1 contained species which stay in the mangrove creeks at low tide, mainly in crab burrows (Myrophis punctatus and Guavina guavina), or buried or attached to the Rhizophora mangle roots (Gobionel- lus oceanicus, Evortodus lyricus, Ctenogobius stigmaticus, and Ctenogobius smaragdus). Poecilia sp. and Gambusia sp. stay in open water during low tide. Sub-group 2-A, consisted of species, both adults and juveniles, which occur in the mangrove creeks throughout the year (Sciades herzbergii, Cathorops agassizii, Anchovia clupeoides, Pterengraulis atherinoides, Anableps anableps, Stellifer naso, Genyatremus luteus, Cynoscion acoupa, Colomesus psitacus, Aspistor parkeri, and Sciades proops). Sub-group 2-B contained species which are more common in the mangrove tidal creeks during high tide (lower estuary) (Caranx latus, Selene vomer, Epinephelus itajara, Strongylura timucu, Batrachoides surinamensis, Cetengraulis edentulus, Rhi- nosardinia amazonica, Megalops atlanticus, Scomberomorus maculatus, Trichiurus lepturus, Lutjanus jocu, and Caranx crysus). Group II consisted of species characteristic of the main channels of the upper estuary (II-1-A) (Hypostomus plecostomus, Plagioscion squamosissimus, Eigenma- nia virescens, Distocyclus conirostris, and Gymnotus carapo), and upper and middle estuary (II-1-B) (Aspredo aspredo, Stellifer microps, Pseudauchenipterus nodosus, Brachyplatystoma vaillanti, and Aspredinichthys filamentosus), and middle and lower estuary (II-2) (Stellifer rastrifer, Cathorops spixii, and Conodon nobilis). 652 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Figure 2. For the Caeté Estuary, (A) Cluster dendrogram based on similarities of the most important species distribution in the different habitats, and (B) MDS plot of the habitats (MICU, mangrove intertidal creek upper-; MICM, mangrove intertidal creek middle-; MICL, mangrove intertidal creek lower-; MTCL, mangrove tidal creek lower-; TC, tidal channel lower-; MCU, main channel upper-; MCM, main channel middle; MCL, main channel lower estuary). Samples were clustered by group average of Bray-Curtis similarity index (presence/absence). Species codes are listed in the appendix. BARLETTA AND BLABER: FISH ASSEMBLAGES AND ESTUARINE GUILDS COMPARISON 653

Results of MDS analysis of habitat groups based on the distribution of dominant species in the Caeté Estuary indicate that as in the Embley Estuary, the habitats in the Caeté Estuary could be divided in two groups (Fig. 2B). The first group included the habitats of the main channel of the upper, middle, and lower estuary. Group two consisted of the intertidal mangrove creeks (upper, middle, and lower estuary), and the mangrove intertidal channel and tidal channel habitats of the lower estuary. All of these habitats are located in shallow waters adjacent to the main channels.

Analysis of Taxonomic Similarity Many families were common to both estuaries. Embley Estuary had the highest diversity in both numbers of species (203) and families (65). For this estuary, the families Carcharhinidae (12 spp.—main channel, tidal channel, and intertidal mudflats), Carangidae (10 spp.—main channel, tidal channel, beaches, intertidal mudflat, and mangrove intertidal creeks), and Gobiidae (16 spp.—beaches, seagrass, and mangrove intertidal creeks) had the highest number of species (Table 1). In Caeté Estuary, the families Ariidae (10 spp.—main channel, mangroves tidal creeks, and tidal channels), Carangidae (12 spp.—main channel and mangroves tidal creeks), Sciaenidae (15 spp.—main channel am mangrove tidal creeks), and Gobiidae (10 spp.—main channel, mangrove intertidal, and tidal creeks) were the families with the highest number of species (Table 2). Species diversity was highest in the main channel, beaches, seagrass, and mangrove intertidal creeks in the Embley Estuary and in the main channel and mangrove tidal creeks of the Caeté Estuary. Assessment of the similarity of estuarine habitats of two estuaries based on taxonomic classification was undertaken using presence/absence of common families. This analysis showed, at 60% similarity, a clear major division into three separate groups of habitats for each estuary (Fig. 3), but overall, these two estuaries were clearly separated at only a relatively low level of similarity (40%). Based on this information, further cluster analyses were undertaken at the family level, based on the habitats clustered in each estuary. Comparisons were conducted between each cluster from one estuary and each from the other estuary. Family composition comparisons among mangrove intertidal creeks (MICc–Caeté Estuary) and the groups of habitats from Embley Estuary showed very low similarity (< 30%) (Fig. 4B 1 and 2). Even when compared with others habitats of Caeté Estuary, this habitat showed low similarity (40%) (Fig. 4A). The main channel of Caeté Estu- ary (MCc) and the main channel (MCe), tidal channel (TCe), and intertidal mudflats (IMe) from the Embley Estuary showed also low similarity (< 20%) (Fig. 4D 1). This group consisted of families which are characteristic of the upper reaches of Caeté Estuary. They include Auchnipteridae, Pimelodidae, Loricaridae, Auchnipteridae, Characidae, Sternopygidae, and Gymnotidae, families that are characteristic of the neotropical region. On the other hand, the more ubiquitous families Engraulidae, Sciaenidae, Ariidae, Carangidae, Haemulidae, and Clupeidae which are characteristic of the main channel (lower and middle reaches) of Caeté Estuary showed 70% similarity with the main channel of the Embley Estuary (Fig. 4D 3-b-2). Comparisons between the main channel (MCc–Caeté Estuary) and seagrass bed habitats (SLe– Embley Estuary) revealed two main groups (Fig. 4E). Group I consisted of the families characteristic of seagrass habitats, and group II (at the level of 70% of similarity) was divided in two sub-groups: Sub-group “a” consisted of the common families of 654 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 9 1 1 5 1 1 1 7 1 1 1 7 1 1 1 3 2 6 2 1 4 2 2 6 12 Total 1 1 2 5 1 2 1 1 1 3 Middle Mangroves intertidal creeks 1 1 6 1 2 1 2 1 Lower Intertidal mudflats 1 1 1 1 2 2 Lower Seagrass 4 3 1 2 2 1 2 2 1 2 3 Lower Beaches 2 1 1 1 2 1 4 Middle 2 1 1 1 2 1 Tidal channel Tidal Upper 1 1 3 2 1 1 1 2 1 8 Lower

1 3 1 1 1 1 4 Middle Main channel 1 1 3 1 1 1 1 1 1 7 Upper Plotosidae Batrachoididae Hemirhamphidae Chirocentridae Chanidae Synodontidae Ariidae Pristigasteridae Engraulidae Ophicthidae Clupeidae Myliobatidae Rhinopteridae Elopidae Megalopidae Dasyatidae Atherinidae Pseudomugilidae Syngnathidae Platycephalidae Sphyrnidae Pristidae Rhinobatidae Taxonomic groups Taxonomic Belonidae Carcharhinidae 18 19 20 14 15 16 17 12 13 10 11 6 7 8 9 5 22 23 24 25 2 3 4 21 Table 1. Number of fish species per family in each habitat Embley Estuary. Table 1 BARLETTA AND BLABER: FISH ASSEMBLAGES AND ESTUARINE GUILDS COMPARISON 655 3 1 5 4 9 1 2 2 1 1 2 1 1 1 2 3 3 1 5 4 6 1 1 1 10 Total 1 2 2 4 1 2 1 1 1 3 2 1 4 Middle Mangroves intertidal creeks 2 1 2 2 1 1 1 1 1 1 1 Lower Intertidal mudflats 1 4 1 6 3 1 1 3 3 1 2 Lower Seagrass 5 1 7 1 2 2 1 1 1 1 2 3 1 1 Lower Beaches 1 1 1 1 1 1 1 1 1 1 Middle 2 1 2 1 1 2 1 1 Tidal channel Tidal Upper 2 1 1 1 3 1 Lower 2 1 1 1 1 2 1 Middle Main channel

2 1 1 1 1 2 2 Upper Lethrinidae Sparidae Haemulidae Gerreidae Lutjanidae Ephippidae Scatophagidae Stromateidae Leiognathidae Sciaenidae Mullidae Monodactylidae Toxotidae Drepaneidae Ambassidae Serranidae Centrogeniidae Terapontidae Apogonidae Sillaginidae Lactariidae Rachycentridae Echeneidae Carangidae Taxonomic groups Taxonomic Centropomidae 42 43 41 40 39 49 50 37 38 44 45 46 47 48 27 28 29 30 31 32 33 34 35 36 Table 1. Continued. Table 26 656 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

5 2 2 2 3 1 4 4 2 1 1 3 5 2 16 203 Total 4 1 1 1 5 11 29 66 Middle Mangroves intertidal creeks 1 1 1 1 4 24 37 Lower Intertidal mudflats 3 3 1 1 4 3 2 24 51 Lower Seagrass 1 2 1 3 2 1 2 2 5 34 71 Lower Beaches 1 1 1 3 21 28 Middle 1 2 1 2 Tidal channel Tidal 18 25 Upper

1 2 2 19 35 Lower

2 3 16 26 Middle Main channel

1 1 1 3 21 34 Upper No. of families No. of species Soleidae Triacanthidae Monacanthidae Tetraodontidae Paralichthyidae Eleotridae Signatidae Scombridae Blenniidae Callionymidae Gobiidae Polynemidae Sphyraenidae Taxonomic groups Taxonomic Chaetodontidae Mugilidae 62 63 64 65 61 58 59 60 55 56 57 54 53 Table 1. Continued. Table 51 52 BARLETTA AND BLABER: FISH ASSEMBLAGES AND ESTUARINE GUILDS COMPARISON 657 1 1 1 2 1 1 1 1 1 6 1 1 2 3 1 1 1 1 1 1 3 15 12 10 10 Total 1 2 1 4 2 1 Tidal channels Tidal 1 1 1 9 1 8 1 1 1 1 1 3 5 1 1 1 1 1 1 1 Lower Mangroves Tidal creeks Tidal 1 3 Lower 6 1 1 Middle Mangroves Intertidal creeks 4 1 1 1 2 Upper 3 1 1 9 2 5 1 8 4 1 1 2 2 1 1 1 1 Lower 2 1 1 1 6 1 7 5 1 1 2 2 1 1 1 11 Middle Main channel 5 1 1 9 1 3 1 9 4 1 1 2 1 1 1 Upper Gobiidae Polynemidae Mugilidae Ephippidae Sciaenidae Haemulidae Carangidae Gerreidae Serranidae Centropomidae Belonidae Batrachoididae Ariidae Engraulidae Cynoglossidae Pristigasteridae Achiridae Clupeidae Paralichthyidae Ophicthidae Scombridae Megalopidae Trichiuridae Gymnuridae Taxonomic groups Taxonomic Eleotridae 19 18 17 16 15 14 13 12 11 10 9 8 7 6 25 5 24 4 23 3 22 2 21 1 20 Table 2. Number of fish species per family in each habitat Caeté Estuary. Table 658 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 2 1 2 2 1 2 1 3 5 1 3 1 Total 103 8 1 1 13 Tidal channels Tidal 1 1 1 1 23 37 Lower Mangroves Tidal creeks Tidal 2 4 Lower 4 2 11 Middle Mangroves Intertidal creeks 7 1 1 11 Upper 1 5 1 19 46 Lower 1 1 2 5 1 2 20 50 Middle Main channel 2 2 1 2 1 3 5 1 22 55 Upper No. of families No. of species Anablepidae Poecilidae Characiidae Sternopygidae Gymnotidae Loricariidae Auchnopteridae Pimelodidae Aspredinidae Diodontidae Ogcocephalidae Tetraodontidae Taxonomic groups Taxonomic 37 36 35 34 33 32 31 30 29 27 28 26 Table 2. Continued. Table BARLETTA AND BLABER: FISH ASSEMBLAGES AND ESTUARINE GUILDS COMPARISON 659

Figure 3. Cluster dendrogram based on similarities in the distribution of fish families in the habitats of the Caeté Estuary (CE) (MICc, mangrove intertidal creeks; MCc, main channel; MTCc, mangrove tidal creeks; TCc, tidal channel) and the Embley Estuary (EE) (MCe, main channel; TCe, tidal channel; IMe, intertidal mudflats; SGe, seagrass beds; SBe, sand/mud beaches; MICe, mangrove intertidal creeks). Samples were clustered by group average of Bray-Curtis similarity index (presence/absence). both habitats, and sub-group “b”, the families characteristic of the upper, middle, and lower reaches of the Caeté Estuary. Comparisons of the main channel of the Caeté Estuary (MCc) with the beach, sand/ mud, (Be), and mangrove intertidal creeks (MICe) of the Embley Estuary revealed four groups at the level of about 60% similarity (Fig. 4F). The first group comprised families characteristic of the upper reaches of the main channel of Caeté Estuary. The second group was represented by the families common to the main channel of Caeté Estuary (MCc) and to the sand/mud beach habitats. Group 3 contained families characteristic of the beach (sand/mud) habitat of the Embley Estuary, sub-group 4-a, those common to the mangrove intertidal creeks and beach (sand mud) (sub- group 3-a), and sub-group 4-b, those characteristic of the mangrove intertidal creeks of the Embley Estuary (Fig. 4F 4-b). Mangrove tidal creeks and tidal channels (MTCc and TCc–Caeté Estuary), and main channel, tidal channel, and intertidal mudflats (MCe, TCe, and IMe–Embley Estuary) comparisons (Fig. 4G) revealed four groups at the level of 60% similarity. The first group consisted of families characteristic of the mangrove tidal creeks and tidal channel of the Caeté Estuary. Groups 2 and 3 were formed by families characteristic of tidal channels and intertidal mudflats, respectively, of the Embley Estuary. Group 4, at 80% of similarity, consisted of sub-group 4a, which contained the families common in the main channel and tidal channel of the Embley Estu- ary, and sub-group 4b were families common to these habitats in both estuaries. The comparison of mangrove tidal creeks (MTCc) and tidal channels (TCc) of the Caeté Estuary with the seagrass beds (SLe) of the Embley Estuary revealed that the 660 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Figure 4. Cluster dendrogram based on similarities in the distribution of fish families in the different habitats of both estuaries: (A) mangrove intertidal creeks (MICc–Caeté Estuary) and main channel (MCe-), tidal channel (TCe-), and intertidal mudflats (IMe–Embley Estuary); (B) mangrove intertidal creeks (MICc–Caeté Estuary) and seagrass beds (SGe–Embley Estuary); (C) MICc and sand/mud beaches (SBe), mangrove intertidal creeks (MICe–Embley Estuary); (D) (see page 661) main channel (MCc–Caeté Estuary) and MCe, TCe, and IMe; (E) MCc and SGe; (F) MCc and SBe, and MICe; (G) (see page 662) mangrove tidal creeks (MTCc), and tidal channel (TCc–Caeté Estuary) and MCe, TCe, and IMe; (H) MTCc, and TCc and SGe; (I) MTCc, and TCc and SBe, and MICe. Samples were clustered by group average of Bray-Curtis similarity index (presence/absence). The common families (CF) among habitats are indicated. BARLETTA AND BLABER: FISH ASSEMBLAGES AND ESTUARINE GUILDS COMPARISON 661

Figure 4. Continued. 662 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Figure 4. Continued. BARLETTA AND BLABER: FISH ASSEMBLAGES AND ESTUARINE GUILDS COMPARISON 663 estuaries had only two families in common (Group 2-a-1)—the ubiquitous Engrau- lidae and Tetraodontidae. The comparison between the mangrove tidal creeks (MTCc) and tidal channels (TCc) of the Caeté Estuary, and the beach (sand/mud) (BLe), and mangrove intertidal creeks (MICe) of the Embley Estuary revealed four groups at a 60% level of similarity (Fig. 4I). The first group comprised the families characteristic of mangrove tidal creeks and tidal channels of the Caeté Estuary. The common families for all habitats from both estuaries were represented in group 2. Groups 3 and 4 were represented by families’ characteristic of the beach and mangrove intertidal creeks of the Embley Estuary, respectively.

Ecological Comparisons Analysis by Estuarine Use and Feeding Mode Functional Groups (Guilds).—The proportions of each guild in terms of biomass varied among habitats within and between estuaries (Table 3). For the Embley Estuary, marine immigrants (66%), marine stragglers (17%), estuarine (12%), and anadromous (10%) guilds were dominant. Marine immigrants (piscivorous–main channel and tidal channel; benthophagous–main channel upper, beach (sand/mud), and intertidal mudflats) were the dominant functional guild in the Embley Estuary, but marine stragglers (piscivorous) were also important in the main channel and tidal channels. Mangrove intertidal creeks (MICMe–38.4%) and seagrass (Sle–20.1%) habitats were important habitats for the estuarine guild (Table 3). In contrast, in the Caeté Estuary the estuarine guild (82%), principally benthophagous fish (81%), was dominant and contributed to > 50% of biomass in each habitat (Table 3). Marine immigrants (planktivorous and hyperbenthophagous) were only important in the habitats of the lower reaches of this estuary. Cluster analysis of the percentage of biomass composition of functional guilds (Fig. 5) indicated a major division at low similarity levels among all habitats of both estuaries. However, at around 60%– 70% similarity, it was possible to separate the habitats of each estuary into two main groups. In the Embley Estuary group 1 was characterized by the sand/mud beaches (Ble) located in the lower estuary. In this habitat the marine immigrant—benthophagous (41%) was the most important functional guild (Table 3). Group 2 consisted of intertidal mudflat habitats also located in the lower estuary. This habitat was used principally by marine immigrant-benthophagous fish species (54%). The third group was formed by the main channel (upper, middle, and lower estuary), tidal channel (upper and middle estuary) where the predominant functional groups were the marine straggler-piscivorous, marine immigrant-piscivorous, and anadromous-piscivorous (Table 3). Seagrass bed habitat (Fig. 5-A, group 4; Table 3) was characterized principally by the presence of marine immigrant-planktivorous (33%) and hyperbenthophagous groups (35%), and the estuarine-planktivorous functional group (11%). The fifth group, mangrove intertidal creeks, was characterized principally by the functional groups marine immigrant-detritivorous (26%) and estuarine-herbivorous (23%). In the Caeté Estuary, the first group consisted of the mangrove intertidal habitats (upper, middle, and lower), where most of the species were estuarine (Table 3; Fig. 5B). Group 2 was formed principally by freshwater immigrants and stragglers, characteristic of the main channel of the upper estuary (MCUc; Fig. 5B), while group 3 contained estuarine and marine immigrant guilds characteristic of the upper (during the end of dry season) middle and lower main channel of Caeté Estuary (Fig. 5-3a). The sub-group 3b contained the mangrove tidal creeks and the tidal channel habitats. In those habitats, the most im- 664 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 0.21 9.29 2.23 0.38 0.27 3.67 9.68 9.55 0.24 2.89 18.07 73.97 13.13 20.78 12.87 26.59 18.67 20.19 Average Average

0.02 1.16 0.01 0.74 0.01 0.05 0.03 2.60 3.53 60.45 14.90 14.56 23.51 25.56 38.45 12.27 MICMe

0.79 8.71 9.50 0.01 8.96 0.08 4.20 IMLe 19.19 86.72 54.25

Sle 0.48 0.04 0.17 0.34 1.04 3.67 5.07 0.83 5.90 3.71 11.36 35.28 78.70 32.98 20.11 TCLc, tidal channel lower).

Ble 0.11 1.42 0.57 0.01 0.45 1.14 0.85 0.05 0.01 0.07 76.74 12.68 10.60 10.04 41.11

2.78 6.37 0.01 7.33 0.04 0.04 22.01 55.36 28.39 30.63 14.59 TCMe Habitats

5.37 9.06 0.88 0.34 0.88 27.36 60.50 27.36 31.36 14.35 TCUe

0.03 2.24 4.95 6.90 41.80 46.85 41.83 32.74 MCLe

0.26 5.01 4.53 15.10 61.38 15.36 39.68 12.15 MCMe

1.72 9.86 6.92 MCUe 19.38 73.14 21.10 32.08 24.27 1. Marine straggler 1. Planktivorous 2. Piscivorous Table 3. Biomass (%) of each functional guild for each habitat of (A) Embley Estuary—(BLe, sand/mud beach; IMLe, intertidal mudflats lower estuary; MCUe, estuary; lower mudflats intertidal IMLe, beach; sand/mud Estuary—(BLe, Embley (A) of habitat each for guild functional each of (%) Biomass 3. Table channel tidal TCMe, estuary; upper channel tidal TCUe, estuary; lower channel main MCLe, estuary; middle channel main MCMe, estuary; upper channel main estu - upper creeks intertidal mangrove Estuary—(MICUc, Caeté (B) and creeks), intertidal Mangrove MICMe, estuary; lower bed seagrass SGe, estuary; middle ary; MICMc, mangrove intertidal creeks middle estuary; MICLc, mangrove intertidal creeks lower estuary; MCUc, main channel upper estuary; MCMc, main channel middle estuary; MCLc, main lower MTCLc, mangrove tidal creeks (A) Embley Estuary Functional guilds 3. Benthophagous 7. Lepidophagous 4. Hyperbenthophagous Total Total 3. Estuarine 1. Planktivorous 2. Herbivorous 2. Marine immigrant 1. Planktivorous 3. Benthophagous 2. Detritivorous 4. Omnivorous 3. Herbivorous Total 4. Piscivorous 5. Benthophagous 6. Hyperbenthophagous BARLETTA AND BLABER: FISH ASSEMBLAGES AND ESTUARINE GUILDS COMPARISON 665

0.14 0.53 0.24 0.94 1.47 3.54 9.18 0.07 1.32 0.14 0.12 0.10 0.43 5.26 11.47 11.47 81.91 15.01 98.24 Average Average

1.03 1.04 MICMe

3.77 3.77 IMLe 100 100

Sle 0.13 0.13 97.36 97.36

Ble 0.01 0.01 22.00 22.00 98.49 98.49

0.56 7.40 1.25 1.25 6.83 16.20 16.20 90.09 92.59 TCMe Habitats

0.25 0.47 1.62 3.54 0.34 0.16 4.10 0.52 11.25 11.26 14.92 74.98 82.63 12.57 TCUe

0.29 8.24 0.09 0.29 0.09 0.09 0.09 0.40 1.59 0.09 11.31 11.31 10.23 68.07 20.80 89.18 MCLe

0.09 1.62 6.40 1.62 0.09 0.19 0.09 0.09 0.30 5.22 0.09 11.81 21.63 21.63 51.59 35.41 87.20 MCMe

5.48 5.48 0.09 0.26 0.86 0.26 0.01 1.15 0.01 0.02 0.09 0.09 MCUe 74.67 13.44 88.12 4. Anadromous 4. 1. Piscivorous Total Functional guilds 5. Fresh water immigrant 1. Planktivorous Table 3. Continued. Table 3. Benthophagous 2. Benthophagous 4. Hyperbenthophagous Total (B) Caeté Estuary 1. Marine straggler 1. Planktivorous Total 3. Estuarine 1. Piscivorous 2. Herbivorous 2. Benthophagous 3. Piscivorous 3. Hyperbhenthophagous 4. Benthophagous Total Total 2. Marine immigrant 1. Planktivorous 2. Piscivorous 666 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

0.31 4.19 0.06 0.37 0.66 2.44 0.10 0.10 3.30 1.69 0.60 1.90 Average Average

MICMe

IMLe

Sle 2.63 2.63

Ble 0.29 0.08 0.37 0.36 0.36 0.76 0.76

TCMe Habitats

0.94 0.03 0.98 0.94 0.95 TCUe

0.09 0.10 0.09 0.10 MCLe

0.09 0.10 0.19 0.19 0.29 0.09 0.39 MCMe

0.09 0.10 0.97 4.67 0.09 0.09 5.84 1.07 3.70 4.77 MCUe 4. Anadromous 4. 1. Detritivous 2. Piscivorous Functional guilds Table 3. Continued. Table Total 5. Fresh water straggler 1. Herbivorous 2. Benthophagous 3. Hyperbenthophagous 4. Lepidophagous Total 6. Fresh water immigrant 1. Planktivorous 2. Detritivorous 3. Hyperbenthophagous Total BARLETTA AND BLABER: FISH ASSEMBLAGES AND ESTUARINE GUILDS COMPARISON 667

Figure 5. Cluster dendrogram of habitats from (A) Embley Estuary (BLe, sand/mud beach; IMLe, intertidal mudflats lower estuary; MCUe, main channel upper estuary; MCMe, main channel middle estuary; MCLe, main channel lower estuary; TCUe, tidal channel upper estuary; TCMe, tidal channel middle estuary; SGe, seagrass bed lower estuary; MICMe, Mangrove intertidal creeks) and (B) Caeté Estuary (MICUc, mangrove intertidal creeks upper estuary; MICMc, mangrove intertidal creeks middle estuary; MICLc, mangrove intertidal creeks lower estuary; MCUc, main channel upper estuary; MCMc, main channel middle estuary; MCLc, main channel lower estuary; MTCLc, mangrove tidal creeks lower estuary; TCLc, tidal channel lower). Habitats were clustered by group average of Bray-Curtis Similarity index based on (log x+1) transformed biomass (%) of ecological guilds. portant functional guild was the estuarine-benthophagous, which was responsible for 75% and 90% of total biomass, respectively (Table 3). Factorial analysis ordination biplot diagrams of species functional guild scores (Fig. 6), as well as regression statistics (Table 4) and cluster analysis (Fig. 5) permitted an interpretation of the distribution of the estuarine functional guilds in different habitats of both estuaries. The first factorial axis explained 40% of the total variability in the Embley Estuary and in the Caeté Estuary. This axis was formed positively by the dominant fish species which spend part 668 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Figure 6. Factorial analysis ordination biplot showing the ecological guilds (see Table 4 for codes) centroids in relation to the different habitats in (A) Embley Estuary and (B) Caeté Estuary. of their life cycle in Embley Estuary habitats, principally in the main channel and tidal channels (Anadromous/piscivorous, marine immigrants/piscivorous, benthophagous and detritivorous, and marine stragglers/piscivorous), and negatively by dominant estuarine (planktivorous and benthophagous) fish species which spend their life in the seagrass beds and in the mangrove intertidal creeks (Fig. 6A). The second factorial axis explained only around 20%–30% and best represented the distribution of fish species along the estuary (upper, middle, and lower). This was more evident in the Caeté than in the Embley Estuary (Fig. 6A,B). The freshwater and marine stragglers were most important in determining this second factorial axis in the Caeté Estuary (Fig. 6B). The freshwater Neotropical species were the most important in negatively determining the second factorial axis, and defined the main gradient structuring the fish species into ecological groups in the upper and lower Caeté Estuary. BARLETTA AND BLABER: FISH ASSEMBLAGES AND ESTUARINE GUILDS COMPARISON 669

Table 4. Results of factorial analysis for fuctional guilds from (A) Embley Estuary and (B) Caeté Estuary.

(A) Embley Estuary Code Variables Axis 1 Axis 2 11 Marine straggler/Planktivorous −0.73** −0.62** 12 Marine straggler/Piscivorous 0.79** −0.16 13 Marine straggler/Benthophagous 0.34 * 0.05 14 Marine straggler/Hyperbenthophagous −0.50* −0.32* 21 Marine immigrant/Planktivorous −0.98** −0.10 22 Marine immigrant/Detritivorous 0.25 0.88** 23 Marine immigrant/Herbivorous −0.47* 0.35* 24 Marine immigrant/Piscivorous 0.73** −0.27 25 Marine immigrant/Benthophagous 0.35* 0.50* 26 Marine immigrant/Hyperbenthophagous −0.17 −0.77** 27 Marine immigrant/Lepidophagous −0.60* 0.73** 31 Estuarine/Planktivorous −0.96** 0.06 32 Estuarine/Herbivorous −0.68* −0.72** 33 Estuarine/Benthophagous −0.90** 0.30* 34 Estuarine/Omnivorous 0.23 0.03 41 Anadromous/Piscivorous 0.87** −0.05 51 Fresh water immigrant/Planktivorous −0.60* 0.73** 52 Fresh water immigrant/Benthophagous 0.34* −0.07 (B) Caeté Estuary Code Variables Axis 1 Axis 2 11 Marine straggler/Planktivorous 0.53* 0.01 12 Marine straggler/Piscivorous 0.89** 0.15 13 Marine straggler/Benthophagous 0.96** 0.19 14 Marine straggler/Hyperbenthophagous 0.35* 0.1 21 Marine immigrant/Planktivorous 0.78** 0.27 22 Marine immigrant/Detritivorous 0.91** −0.21 23 Marine immigrant/Herbivorous 0.91** −0.01 24 Marine immigrant/Piscivorous 0.79** −0.05 31 Estuarine/Planktivorous 0.67* 0.27 32 Estuarine/Herbivorous −0.76** 0.26 33 Estuarine/Benthophagous 0.72** −0.45* 41 Anadromous/Detritivorous 0.64* 0.09 42 Anadromous/Piscivorous −0.12 0.21 51 Fresh water stragler/Herbivorous −0.25 −0.89** 52 Fresh water stragler/Benthophagous -0.05 −1** 53 Fresh water stragler/Hyperbenthophagous -0.09 −0.99** 54 Fresh water stragler/Lepidophagous -0.09 −0.99** 61 Fresh water immigrant/Planktivorous −0.56* 0.27 62 Fresh water immigrant/Detritivorous 0.573* −0.69** 63 Fresh water immigrant/Hyperbenthophagous −0.07 −0.99** * P < 0.05; **P < 0.01 670 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Discussion

The present study focused on comparisons of the taxonomy, ecological structure, and resource use of fish assemblages in distinct habitats of two tropical estuaries located in different zoogeographic realms. Both estuaries were subdivided according to their geomorphology and salinity gradients into upper, middle, and lower reaches. These divisions were reflected in the distribution of fish assemblages and their use of the estuaries. The main channel, the flood plain tidal channels, and mangrove tidal and intertidal creek habitats each had characteristic fish assemblages. Another important habitat was the seagrass beds, characteristic of the lower reaches of the Embley Estuary. Zoogeographical Analysis of the Taxonomic Composition in Both Es- tuaries.—No single species occurred in the data sets of both estuaries, although C. leucas, Aetobatus narinari, T. lepturus, and Chilomycterus spinosus are known to inhabit both regions (Blaber et al., 1989; Cyrus and Blaber, 1992; Barletta et al., 1998, 2003). The numbers of species and families are greater in the Embley Estuary (Indo- West Pacific) than in the Caeté Estuary. Similarly, other Indo-West Pacific estuaries, such as those of the Solomon Islands (Blaber and Milton, 1990) and Moreton Bay, Australia (Morton, 1990) have higher species diversity than the Caeté Estuary, Brazil (Barletta et al., 2000, 2003, 2005). This difference in species numbers between the Indo-West Pacific and Tropical West Atlantic owes its origins to evolutionary history and tectonics. In the Indo-West Pacific region there is a concentration of species in the triangle bordered by the Philippines, Malay Peninsula, and New Guinea (Briggs, 1995; Blaber, 1997, 2000). This area was considered by Briggs (1995) as the East Indies center of evolutionary radiation (Indo-Polynesian Province) of the Indo-West Pacific region. This center of high diversity originated in the Tethys Sea of the Cretaceous to early Miocene Ocean, which was situated between the northern and southern continents (Briggs, 1995). The Indo-WestP acific Region (and within it, the East Indies Tri- angle) represents an evolutionary and distributional center for much of the world’s tropical fauna and flora and there is a notable decrease in species diversity correlated with distance from this center of high diversity (East Indies Triangle). This is true for almost every tropical fish family for which the systematics are well known G( obie- socidae, Lutjanidae, Singnathidae, Chaetodontidae, Ariidae, Clupeidae, Engraulidae, and Sciaenidae) (Allen, 1975; Burgess, 1978, 1989; Allen and Talbot, 1985; McDow- all, 1988; Blum, 1989; Blaber, 2000; Marceniuk and Menezes, 2007). It is important to note, however, that many taxa reached the Atlantic region from the Indo-West Pacific. For example, Ricklefs and Latham (1993) and Briggs (1995) suggested, that the mangrove genus Rhizophora, a very important component of tropical estuarine ecosystems, probably originated in the East Indies, spread west to East Africa and thence east to the New World, reaching the Atlantic through the Central American archipelago. This genus probably reached the Eastern Atlantic from the western At- lantic, since the three West African mangrove species are identical to those of the Western Atlantic (Ricklefs and Latham, 1993). The Western Atlantic Region, which includes three main provinces (Caribbean, West Indian, and Brazilian), constitutes a secondary center of evolutionary radiation (Briggs, 1995). Many species that evolved in this area migrated eastwards to colonize the Tropical Eastern Atlantic. BARLETTA AND BLABER: FISH ASSEMBLAGES AND ESTUARINE GUILDS COMPARISON 671

A significant difference between the Indo-West Pacific and Tropical Western Atlantic estuarine fish faunas is the much higher diversity of freshwater species in tropical South America. Caeté Estuary, located in the Brazilian Province, is strongly influenced, especially in its upper reaches, by freshwater species (Characiforms, Siluriforms, Gymnotiforms, and Perciforms) characteristic of the Neotropic Zoo- geographic Region. According to Nelson (1994) and Fink and Fink (1997), the Os- tariophysan fish (Characiforms, Siluriforms, andG ymnotiforms) are a monophyletic primary freshwater group, which arose in the Upper Jurassic. The most primitive orders of this group are the Characiforms and Siluriforms. Representatives of these two orders were able to colonize Asian, African and South American waters before these continents were separated. Aside from two catfish families (Ariidae and Plotosidae) that became secondarily adapted to salt water, Ostariophysan fish have not been able to reach Madagascar, the West Indies, New Zealand, and Australia. This suggests that those land masses have not been connected to any of the larger continents since the upper Jurassic (Briggs, 1995). This also explains why, despite seemingly suitable habitats and hydrological conditions, there are no freshwater Ostariophysan fish species in the Embley Estuary. Ecological Functional Groups.—Certain functional guilds have a clear and demonstrated value for describing the use of different estuarine habitats by fish and for comparing the different estuarine use and feeding mode functional groups in the estuary. In particular, the ecological guilds proved useful for highlighting the predominance of estuarine species and marine juveniles, and the strong effects of salinity on assemblages in the different reaches of the two systems, particularly in the Caeté Estuary (Barletta-Bergan et al., 2002a,b; Barletta et al., 2003, 2005). Based on studies of the Orinoco Delta, Venezuela (Cervigón, 1985), in the Cayenne River Estuary, French Guiana (Morais and Morais, 1994) and in the coastal areas of Guy- ana (Lowe-McConnell, 1962) and Caeté Estuary, north Brazil (Barletta et al., 2003, 2005) it may be concluded that the fish species distribution and movement patterns induced by salinity gradients and seasonal fluctuations could be generalized for northern South America. Also the similarity of the fish assemblages and habitats among these estuaries suggests that estuarine use and feeding mode functional groups could be generalized for all estuaries in northern South America. Similarly, the Embley Estuary fish assemblages and ecological information (Blaber et al., 1989; Blaber, 1990, 1997, 2000; Cyrus and Blaber, 1992) indicate that the ecological characteristics of these guilds may be common to many estuaries of the Indo-West Pacific, and particularly for those of northern Australia. In estuaries such as the Embley, where salinities are relatively uniform in all reaches throughout the year, marine immigrants (piscivorous—main and tidal channel; benthophagous—main tidal channels, beaches, and intertidal mud flats; hyperbenthophagous—seagrass beds) and marine stragglers (piscivorous—main and tidal channels) were the dominant guilds. The semi-catadromous (piscivorous) functional group (L.. calcarifer) was also an important component of all habitats, except the seagrass beds of the lower reaches of Embley Estuary. Plots of functional groups of cluster and factorial analysis centroids for both estuaries show distinct patterns in the structure of the guilds in both estuaries. In the Caeté Estuary, the distribution of guilds reflected the salinity gradient in the estuary, principally for freshwater groups (straggler and immigrant guilds) and marine stragglers, whereas, in the Embley Estuary, the stability of the salinity even in the 672 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 upper reaches (Cyrus and Blaber, 1992) allowed the marine immigrant and marine straggler guilds to use most of the habitats. In this case, the dominant fish species which spend part of their life cycle in Embley habitats (principally the main channel and tidal channels)–anadromous/piscivorous, marine immigrant/piscivorous, benthophagous, and detritivorous, and marine straggler/piscivorous–and the dominant estuarine (planktivorous and benthophagous) fish species which spend their life in the seagrass beds and mangrove intertidal creeks explain the patterns of fish distribution in Embley Estuary.

Main Conclusions and Recommendations

The Neotropical fish species of the upper Caeté Estuary and the chondrichthyan fauna of the Embley underlie the most important taxonomic differences between these estuaries. On the other hand, the families Engraulidae, Sciaenidae, Ariidae, Carangidae, Haemulidae, and Clupeidae, which are characteristic of the main channel and mangrove tidal creeks of Caeté Estuary showed 70% similarity with the main channel of the Embley Estuary. Moreover, for the Embley Estuary, marine immigrants (mainly piscivorous and benthophagous) were the dominant functional guilds, whereas in Caeté Estuary the estuarine functional guilds (benthophagous) contributed to > 50% of the biomass in each habitat. The fish communities inhabiting estuaries have been studied worldwide and there have been several attempts to indicate common features of those communities. In addition to supporting their own resident fish community, estuaries are nursery grounds, migration corridors, and refuge areas for a variety of fish species at different life stages. Given the increase in available data, it is possible to begin determining similarities and differences among biogeographical areas and thus to examine the features of fish community structure of estuaries on a global basis.

Acknowledgments

The authors would like to thank CSIRO Marine Laboratories, Cleveland, Australia for the 2-mo stay during which this manuscript was prepared. The data from Caeté Estuary was undertaken within the framework of the “Mangrove Dynamics and Management (MADAM)”, supported by the Ministry for Education, Science, Research and Technology (BMBf) (Proj- ect number: 03F0154A)–Germany, and Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq–Brazil).

Literature Cited

Allen, G. R. 1975. Damselfishes of the South Seas. TFH Publications, Neptune City. 400 p. ______and F. H. Talbot. 1985. Review of the snappers of the genus Lutjanus (Pisces: Lut- janidae) from the Indo-Pacific, with the description of a new species. Indo-Pac. Fishes 11: 1–87. Araújo, F. G. and M. C. C. Azevedo. 2001. Assemblages of southeast-south Brazilian coastal systems based on distribution of Fishes. Estuar. Coast. Shelf Sci. 52: 729–738. Barletta-Bergan, A., M. Barletta, and U. Saint-Paul. 2002a. Structure and seasonal dynamics of larval fish in the Caeté River Estuary in North Brazil. Estuar. Coast. Shelf Sci. 54: 193–206. ______, ______, and ______. 2002b. Community structure and temporal variability of ichthyoplankton in North Brazilian mangrove creeks. J. Fish Biol. 61: 33–51. BARLETTA AND BLABER: FISH ASSEMBLAGES AND ESTUARINE GUILDS COMPARISON 673

Barletta, M. 1999. Seasonal changes of density, biomass and species composition of fishes in different habitats of Caeté Estuary (North Brazilian Coast – East Amazon) ZMT Contribu- tion 7. 115 p. ______, A. Barletta-Bergan, and U. Saint-Paul. 1998. Description of the fisheries structure in the mangrove-dominated region of Bragança (State of Pará, North Brazil). Ecotropica 4: 41–53. ______, ______, ______, and G. Hubold. 2003. Seasonal changes in density, biomass, and diversity of estuarine fishes in tidal mangrove creeks of the lower Ca- eté Estuary (northern Brazilian coast, east Amazon). Mar. Ecol. Prog. Ser. 256: 217–228. ______, ______, ______, and ______. 2005. The role of salinity in structuring the fish assemblages in a tropical estuary.J. Fish Biol. 66: 45–72. ______, U. Saint-Paul, A. Barletta-Bergan, W. Ekau, and D. Schories. 2000. Spatial and temporal distribution of Myrophis punctatus (Ophichthidae) and associated fish fauna in a northern Brazilian intertidal mangrove forest. Hydrobiologia 426: 65–74. Barthem, R. B. 1985. Ocorrência, distribuição e biologia dos peixes da Baía de Marajó, Estuário Amazônico. Boletim do Museu Paraense Emilio Goeldi 2: 49-69. ______and M. golding.Golding. 1997. ThThe e catficatfish sh connection: ecology, migration, and conserva-conservation�� of Amazon predators. Biology and resource management in the tropics series. Colum- bia University Press. New York. 210 p. Blaber, S. J. M. 1997. Fish and fisheries of tropical Estuaries. Fish and fisheries series 22. Chap- man & Hall, London. 367 p. ______. 2000. Tropical Estuarine Fishes: Ecology, Exploitation and Conservation. The Blackwell, Oxford. 372 p. ______and T. G. Blaber. 1980. Factors affecting the distribution of juvenile estuarine and inshore fishes. J. Fish Biol. 17: 143–162. ______and D. A. Milton. 1990. Species composition, community structure and zoogeog- raphy of fishes of mangrove estuaries in the Salomon Islands. Mar. Biol. 105: 259–268. ______, D. T. Brewer, and J. P. Salini. 1989. Species composition and biomass of fishes in different habitats of a tropical northern Australian estuary: Their occurrence in the adjoining sea and estuarine dependence. Estuar. Coast. Shelf Sci. 29: 509–531. ______, ______, and ______. 1990. Biomasses, catch rates and abundances of demersal fishes, particularly predators of prawns, in a tropical bay in theG ulf of Carpentaria, Australia. Mar. Biol. 107: 397–408. Blum, S. D. 1989. Biogeography of the Chaetodontidae: an analysis of allopatry among closely related species. Environ. Biol. Fish. 25: 9–31. Briggs, J. C. 1995. Global Biogeography. Elsevier, The Netherlands. 452 p. Burgess, W. E. 1978. Butterfly fishes of the world. TFH Publications, Neptune City. 654 p. ______. 1989. An atlas of freshwater and marine Catfishes. A preliminary survey of the Siluriformes. TFH Publications, Neptune City. 784 p. Cervigón, F. 1985. The ichthyofauna of the Orinoco estuarine water delta in the West Atlantic coast, Caribbean. Pages 57–78 in A. Yáñes-Arancibia, ed. Fish community ecology in estuaries and coastal lagoons: towards an ecosystem integration. DR (R) UNAM Press, Mexico City. Clarke, K. R. and R. M. Warwick. 1994. Change in communities: an approach to statistical analysis and interpretation. Natural Environment Research Council (NERC), Plymouth. 144 p. Cyrus, D. P. and S. J. M. Blaber. 1987a. The influence of turbidity on juvenile marine fishes in estuaries. Part 1, field studies. J. Exp. Mar. Biol. Ecol. 109: 53–70. ______and ______. 1987b. The influence of turbidity on juvenile marine fishes in estuaries. Part 2, laboratory studies. J. Exp. Mar. Biol. Ecol. 109: 71–91. ______and ______. 1992. Turbidity and salinity in a tropical Northern Australian Estuary and their influence on fish distribution. Estuar. Coast. Shelf Sci. 35: 545–563. 674 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Dorenbosch, M., M. G. G. Grol, I. Nagelkerken, and van der Veld. 2005. Distribution of coral reef fishes along a coral reef-seagrass gradient: edge effects and habitat segregation. Mar. Ecol. Prog. Ser. 299: 277–288. Elliot, M. and K. L. Hemingway. 2002. Fishes in Estuaries. Blackwell Science, Oxford. 600 p. ______, A. K. Whitfield, I. C. Potter, S. J. M. Blaber, D. P. Cyrus, F. G. Nordlie, and T. D. Har- rison. In press. A global review of the categorization of major fish guilds utilizing estuaries. Fish Fish. Eschmeyer, W. N. 2006. Catalog of fishes. California Academic of Sciences. San Francisco. 2905 p. Fink, S. V. and W. L. Fink. 1997. Interrelationships of Ostariophysan fishes (Teleostei). Pages 209–245 in M. Stiassney, L. Parenti, and G. D. Johnson, eds. Interrelationships of fishes. Academic Press. New York. Froese, R. and D. Pauly. 2006. Fish base [World Wide Web] electronic publication. Available from: www.fishbase.org. Accessed 15 May 2007. Gauch, H. G. 1982. Multivariate analysis in community ecology. Cambridge University Press, New York. 150 p. Islam, M. S. and M. Haque. 2004. The mangrove-based coastal and nearshore fisheries of Bang- ladesh: ecology, exploitation and management. Rev. Fish Biol. Fish. 14: 153–180. Jaureguizar, A. J., R. Menni, R. Guerrero, and C. Lasta. 2004. Environmental factors structuring fish communities of Río de la Plata Estuary. Fish. Res. 66: 195–211. Krumme, U., H. Heuthen, M. Barletta, W. Villwock, and U. Saint-Paul. 2005. Contribution to the feeding ecology of predatory wingfin anchovy Pterengraulis atherinoides (L.) in north Brazilian mangrove creeks. J. Appl. Ichthyol. 21: 469–477. Lee, S. Y. 1995. Mangrove outwelling: a review. Hydrobiologia 295: 203–212. Legendre, P. and L. Legendre. 1998. Numerical ecology. 2nd edition, Elsevier. 853 p. Lowe-McConnel, R. H. 1962. The fishes of the British Guiana continental shelf, Atlantic coast of South America, with notes on their natural history. J. Linn. Soc. Lond. - Zoology 44: 669–697. Mahon, R., S. K. Brown, K. C. T. Zwanenburg, D. B. Atkinson, K. R. Buja, L. Clafin, G. D. Howell, M. E. Monaco, R. N. O’Boyle, and M. Sinclair. 1998. Assemblage and biogeography of demersal fishes of the east coast of North America. Can. J. Fish. Aquat. Sci. 55: 1704–1738. Marceniuk, A. P. and N. A. Menezes. 2007. Systematics of family Ariidae (Ostariophysi, Siluri- formes), with a redefinition of the genera. Zootaxa 1416: 1–126. Mathieson, S., A. Cattrijsse, M. J. Costa, P. Drake, M. Elliott, J. Gardner, and J. Marchand. 2000. Fish assemblages of European tidal marshes: a comparison based on species, families and functional guilds. Mar. Ecol. Prog. Ser. 204: 225–242. McDowall, R. M. 1988. Diadromy in fishes: migrations between freshwater and marine environments. Kluwer Academic Publishers, London. 250 p. Morais, A. T. and L. T. Morais. 1994. The abundance and diversity of larval and juvenile fish in a tropical estuary. Estuaries 17: 216–225. Morton, R. M. 1990. Community structure, density and standing crop of fishes in a subtropical Australian mangrove area. Mar. Biol. 105: 385–394. Nelson. J. 1994. Fishes of the world. 3rd ed. Wiley and Sons. New York. 606 p. Pauly, D. and A. Yáñes-Arancibia. 1994. Fisheries in coastal lagoons in B. Kjerfve, ed. Coastal lagoon processes. Elsevier, Oceanography Series 60, Amsterdam. 600 p. Ricklets, E. R. and R. E. Latham. 1993. Global Patterns of diversity in Mangroves Floras. Chap. 20 in E. R. Ricklefs and D. Schluter, eds. Species Diversity in Ecological Communities. The University of Chicago Press, Chicago. 414 p. Roy, P. S., R. J. Willians, A. R. Jones, I. Yassini, P. J. Gibbs, B. Coats, R. J. West, P. R. Scanes, J. P. Hudson, and S. Nichol. 2001. Structure and Function of South-east Australian Estuaries. Estuar. Coast. Shelf Sci. 53: 351–384. BARLETTA AND BLABER: FISH ASSEMBLAGES AND ESTUARINE GUILDS COMPARISON 675

Thiel, R.,H . Cabral, and M. J. Costa. 2003. Composition, temporal changes and ecological guild classification of the ichthyofaunas of large European estuaries – a comparison between the Tagus (Portugal) and the Elbe (Germany). J. Appl. Ichthyol. 19: 330–342. ______, A. Sepulveda, R. Kafemann, and W. Nellen. 1995. Environmental factors as forces structuring the fish community of the Elbe Estuary. J. Fish Biol. 46: 47–69. Vieira, J. P. and J. A. Musik. 1993. Latitudinal patterns in diversity of fishes in warm-temperate and tropical estuarine waters of the Western Atlantic. Atlântica 15: 115–133. ______and ______. 1994. Fish faunal composition in warm-temperate and tropical estuaries of Western Atlantic. Atlântica 16: 31–53. Wolanski, E. 1995. Transport of sediment in mangrove in mangrove swamps. Hydrobiologia 295: 31–42.

Addresses: (M.B.) Laboratory of Ecology and Management of Estuarine and Coastal Ecosys- tems, Departamento de OceanografiOceanografia, a, UFPE, Cidade Universitária,Univ��ersitária, ��50740�-��550��, Recife, Per- nambuco, Brazil. (S.J.M.B.) CSIRO Marine Laboratories, P.O. Box 120, Cleveland, Queensland 4163, Australia. Corresponding Author: (M.B.) E-mail: . 676 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Appendix 1. List of families and species with their codes, ecological guilds (EG; 1, marine straggler; 2, marine immigrants; 3, estuarine species; 4, anadromous; 5, catadromous; 6, amphidromous; 7, freshwater stragglers; 8, freshwater immigrants), trophic guilds (TG; 1, planktivorous; 2, detritivorous; 3, herbivorous; 4, piscivorous; 5, benthophagous; 6, hyperbenthophagous; 7, opportunistic), and life stage (A, adult; J, juvenile) found in each estuary (E, Embley; C, Caeté) considered in this study.

EG TG Life Stage E No. Code Taxonomic groups A J 1 CARCH Carcharhinidae 2 CAMBL Carcharhinus amblyrhynchoides (Whitley, 1934) 1 4 1 1 E 3 CCAUT Carcharhinus cautus (Whitley, 1945) 2 4 1 1 E 4 CDUSS Carcharhinus dussumieri (Müller and Henle, 1839) 1 4 1 1 E 5 CLEUC Carcharhinus leucas (Müller and Henle, 1839) 2 4 1 1 E 6 NACUT Negaprion acutidens (Rüppell,1837) 2 4 1 1 E 7 RACUT Rhizoprionodon acutus (Rüppell, 1837) 1 4 1 1 E 8 SPHYR Sphyrnidae 9 SLEWI Sphyrna lewini (Griffith and Smith, 1834) 1 4 0 1 E 10 SMOKA Sphyrna mokarran (Rüppell, 1837) 1 4 0 1 E 11 PRIST Pristidae 12 PPECT Pristis pectinata Lathan, 1794 2 4 1 1 E 13 PPRIS Pristis pristis (Linnaeus, 1758) 2 4 1 1 E 14 RHINO Rhinobatidae 15 RDJID Rhynchobatus australiae Whitley, 1939 1 5 0 1 E 16 DASYA Dasyatidae 17 PSEPH Pastinachus sephen (Forsskål, 1775) 2 5 1 1 E 18 HUARN Himantura uarnak (Forsskål, 1775) 2 5 1 1 E 19 TLYMN Taeniura lymma (Forsskål, 1775) 2 5 1 1 E 20 RHINOP Rhinopteridae 21 RADSP Rhinoptera cf. adspersa Müller and Henle, 1841 1 5 1 0 E 22 ELOPI Elopidae 23 EMACH Elops machnata (Forsskål, 1775) 2 4 1 1 E 24 MEGAL Megalopidae 25 MCYPR Megalops cyprinoides (Broussonet, 1782) 2 4 1 1 E 26 MATLA Megalops atlanticus Valenciennes, 1847 2 4 0 1 C 27 OPHIC Ophicthidae 28 MPUNC Myrophis punctatus Lütken, 1852 3 5 1 1 C 29 ENGRA Engraulidae 30 SANDH Stolephorus andhraensis Babu Rao, 1966 1 1 1 1 E 31 SCARP Stolephorus carpentariae (De Vis, 1882) 2 1 1 1 E 32 PATHE Pterengraulis atherinoides (Linnaeus, 1766) 3 4 1 1 C 33 CEDEN Cetengraulis edentulus (Cuvier, 1829) 2 1 0 1 C 34 ACLUP Anchovia clupeoides (Swainson, 1839) 2 1 1 1 C 35 CLUPE Clupeidae 36 ACHAC Anodontostoma chacunda (Hamilton, 1822) 2 2 1 1 E 37 NEREB Nematalosa erebi (Günter, 1868) 8 2 1 1 E 38 SALBE Sardinella albella (Valenciennes, 1847) 2 1 1 1 E 39 RAMAZ Rhinosardinia amazonica (Steindachner, 1879) 2 1 1 1 C 40 CHANI Chanidae 41 CCHAN Chanos chanos (Forskäl, 1775) 2 2 1 1 E BARLETTA AND BLABER: FISH ASSEMBLAGES AND ESTUARINE GUILDS COMPARISON 677

Appendix 1. Continued.

EG TG Life Stage E No. Code Taxonomic groups A J 42 ASPRE Aspredinidae 43 AASPR Aspredo aspredo (Linnaeus, 1758) 3 5 1 1 C 44 AFILA Aspredinichthys filamentosus (Valenciennes, 1840) 3 5 1 1 C 45 PIMEL Pimelodidae 46 BVAIL Brachyplatystoma vaillanti (Valenciennes, 1840) 8 6 1 1 C 47 AUCH Auchnopteridae 48 PNODO Pseudauchenipterus nodosus (Bloch, 1794) 8 2 1 1 C 49 ARIID Ariidae 50 AMAST Sciades mastersi Ogilby, 1898* 2 5 1 1 E 51 APROX Arius proximus (Ogilby, 1898) 2 5 1 0 E 52 AMACR Arius macrocephalus (Bleeker, 1846) 2 5 1 1 E 53 CSPIX Cathorops spixii (Agassiz, 1829) 3 5 1 1 C 54 CAGAS Cathorops agassizii (Eigenmann and Eigenmann, 1888) 3 5 1 1 C 55 HPARK Aspistor parkeri (Traill, 1832)** 1 5 0 1 C 56 SHERT Sciades hertzbergii (Bloch, 1794) 3 5 1 1 C 57 HPROO Sciades proops (Valenciennes, 1840)*** 2 5 0 1 C 58 PLOTO Plotosidae 59 ENUDI Euristhmus nudiceps (Günther, 1880) 2 5 1 1 E 60 LORICA Loricariidae 61 HPLEC Hypostomus plecostomus (Linnaeus, 1758) 7 3 0 1 C 62 STERN Sternopygidae 63 DCONI Distocyclus conirostris (Eigenmann and Allen, 1942) 7 5 1 0 C 64 EVIRE Eigenmania virescens (Valenciennes, 1836) 7 5 1 1 C 65 GYMNO Gymnotidae 66 GCARA Gymnotus carapo Linnaeus, 1758 7 5 1 1 C 67 SYNOD Synodontidae 68 SAUSP Saurida sp. 2 1 4 1 0 E 69 BATR Batrachoididae 70 BSURI Batrachoides surinamensis (Bloch and Schneider, 1801) 3 6 1 1 C 71 MUGIL Mugilidae 72 LSUBV Liza subviridis (Valenciennes, 1836) 2 2 1 1 E 73 VBUCH Valamugil buchanani (Bleeker, 1854) 2 2 1 1 E 74 PSEUDM Pseudomugilidae 75 PGERT Pseudomugil gertrudae Weber, 1911 8 1 1 1 E 76 ATHER Atherinidae 77 ADUOD Atherinomorus duodecimalis (Valenciennes, 1835) 2 1 1 1 E 78 ANDRA Atherinomorus endrachtensis (Quoy and Gaimard, 1825) 2 1 1 1 E 79 BELON Belonidae 80 TCROC Tylosurus crocodilus (Péron and Lesueur, 1821) 2 4 1 1 E 81 TPUNC Tylosurus punctulatus (Günther, 1872) 2 4 1 1 E 82 STIMO Strongylura timucu (Walbaum, 1792) 2 4 1 1 C 83 HEMI Hemiramphidae 84 ASCLE Arrhamphus sclerolepis Günther, 1866 2 3 1 1 E 85 ZBUFF Zenarchopterus buffonis (Valenciennes, 1847) 3 1 1 1 E 86 ANABL Anablepidae 87 AANAB Anableps anableps (Linnaeus, 1758) 3 5 1 1 C 678 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Appendix 1. Continued.

EG TG Life Stage E No. Code Taxonomic groups A J 88 POECI Poeciliidae 89 POECSP Poecilia sp. 8 1 1 1 C 90 GAMSP Gambusia sp. 8 1 1 1 C 91 SYNGN Syngnathidae 92 HHEPT Hippichthys heptagonus Bleeker, 1849 2 1 1 1 E 93 HKUDA Hippocampus kuda Bleeker, 1852 2 1 1 1 E 94 HWHIT Hippocampus borboniensis Duméril, 1870 2 1 1 1 E 95 HSP Hippocampus sp. 2 1 1 1 E 96 PLATY Platycephalidae 97 CNEMA Cymbacephalus nematophthalmus (Günther, 1860) 2 4 1 0 E 98 PENDR Platycephalus endrachtensis Quoy and Gaimard, 1825 2 4 1 1 E 99 PINDI Platycephalus indicus (Linaeus, 1758) 2 4 1 1 E 100 CENTR Centropomidae 101 LCALC Lates calcarifer (Bloch, 1790) 4 4 1 1 E 102 AMBAS Ambassidae 103 ADUSS Ambassis dussumieri (Cuvier, 1828) 2 1 1 1 E 104 ANALU Ambassis nalua (Hamilton, 1822) 2 1 1 1 E 105 SERRA Serranidae 106 EMALA Epinephelus malabaricus (Bloch and Schneider, 1801) 2 4 1 0 E 107 ECOIO Epinephelus coioides (Hamilton, 1822) 2 4 1 1 E 108 EITAJ Epinephelus itajara (Lichtenstein, 1822) 2 5 0 1 C 109 CENTRG Centrogeniidae 110 CVAIG Centrogenys vaigiensis (Quoy and Gaimard, 1824) 2 5 1 1 E 111 APOGO Apogonidae 112 AHYAL Apogon hyalosoma Bleeker, 1852 3 1 1 1 E 113 ARUPP Apogon rueppellii Günther, 1859 2 1 1 0 E 114 ASANG Apogon sangiensis Bleeker, 1857 2 1 1 0 E 115 SROSE Siphamia roseigaster (Ramsay and Ogilby, 1887) 3 1 1 1 E 116 SILLA Sillaginidae 117 SANAL Sillago analis Whitley, 1943 2 5 1 1 E 118 SINGE Sillago ingenuua McKay, 1985 2 5 1 1 E 119 SMACUL Sillago maculata Quoy and Gaimard, 1824 2 5 0 1 E 120 SSIHA Sillago sihama (Forsskål, 1775) 2 5 0 1 E 121 SLUTE Sillago lutea McKay, 1985 2 5 0 1 E 122 LACT Lactariidae 123 LLACT Lactarius lactarius (Bloch and Scheneider, 1801) 2 4 1 0 E 124 RACHI Rachicentridae 125 RCANA Rachycentron canadum (Linnaeus, 1766) 1 4 1 0 E 126 CARAN Carangidae 127 SCOMM Scomberoides commersonnianus (Lacépède, 1801) 2 4 1 1 E 128 CCRYS Caranx crysos (Mitchill, 1815) 2 6 0 1 C 129 SVOME Selene vomer (Linnaeus, 1758) 2 6 0 1 C 130 CLATU Caranx latus Agassiz, 1831 2 6 0 1 C 131 GSPEC Gnathanodon speciosus (Forsskål, 1775) 2 6 0 1 E BARLETTA AND BLABER: FISH ASSEMBLAGES AND ESTUARINE GUILDS COMPARISON 679

Appendix 1. Continued.

EG TG Life Stage E No. Code Taxonomic groups A J 132 LEIOG Leiognathidae 133 LDECO Leiognathus decorus (De Vis, 1884) 2 1 0 1 E 134 LEQUU Leiognathus equulus (Forsskål, 1775) 2 1 0 1 E 135 SRUCO Secutor ruconius (Hamiltton, 1822) 1 1 0 1 E 136 LUTJA Lutjanidae 137 LRUSS Lutjanus russellii (Bleeker, 1849) 2 6 0 1 E 138 LJOCU Lutjanus jocu (Bloch and Schneider, 1801) 2 6 0 1 C 139 GERRE Gerreidae 140 GEURY Gerres erythrourus (Bloch, 1791) 2 5 0 1 E 141 GFILA Gerres filamentosus Cuvier, 1829 2 5 0 1 E 142 GLONG Gerres longirostris (Lacépède, 1801) 2 5 0 1 E 143 HAEMU Haemulidae 144 PARGE Pomadasys argenteus (Forsskål, 1775) 2 5 0 1 E 145 PKAAN Pomadasys kaakan (Cuvier, 1830) 2 5 0 1 E 146 GLUTE Genyatremus luteus (Bloch, 1790) 3 5 1 1 C 147 CNOBI Conodon nobilis (Linnaeus, 1758) 2 5 1 1 C 148 SPARI Sparidae 149 ABERD Acanthopagrus berda (Forsskål, 1775) 2 5 1 1 E 150 LETHR Lethrinidae 151 LLENT Lethrinus lentjan (Lacépède, 1802) 2 6 0 1 E 152 POLYN Polynemidae 153 ETETRA Eleutheronema tetradactylum (Shaw, 1804) 2 4 1 1 E 154 PMACR Polydactylus macrochir (Günther, 1867) 2 6 1 1 E 155 PVIRG Polydactylus virginicus (Linnaeus, 1758) 2 6 1 0 C 156 SCIAE Sciaenidae 157 SRAST Stellifer rastrifer (Jordan, 1889) 2 6 1 1 C 158 SMICR Stellifer microps (Steindachner, 1864) 3 6 1 1 C 159 CACOU Cynoscion acoupa (Lacépède, 1801) 2 6 0 1 C 160 PSQUA Plagioscion squamosissimus (Heckel, 1840) 7 6 1 1 C 161 SNASO Stellifer naso (Jordan, 1889) 3 6 1 1 C 162 MULLI Mullidae 163 UTRAG Upeneus tragula Richardson, 1846 2 5 0 1 E 164 MONOD Monodactylidae 165 MARGE Monodactylus argenteus (Linnaus, 1758) 3 1 1 1 E 166 TOXOT Toxotidae 167 TCHAT Toxotes chatareus (Hamilton, 1822) 3 7 1 1 E 168 DREPA Drepaneidae 169 DPUNC Drepane punctata (Linnaeus, 1758) 2 6 0 1 E 170 CHAET Chaetodontidae 171 CMUELL Chelmon muelleri Klunzinger, 1879 1 6 0 1 E 172 PECEL Parachaetodon ocellatus (Cuvier, 1831) 1 6 0 1 E 173 TERAP Teraponidae 174 ACAUD Amniataba caudavittata (Richardson, 1845) 2 6 1 1 E 175 PQUAD Pelates quadrilineatus (Bloch, 1790) 2 6 0 1 E 176 TJARB Terapon jarbua (Forsskål, 1775) 2 8 0 1 E 177 TPUTA Terapon puta Cuvier, 1829 1 6 0 1 E 178 BLENN Blenniidae 179 OROTU Omobranchus rotundiceps (Macleay, 1881) 2 5 1 0 E 680 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Appendix 1. Continued.

EG TG Life Stage E No. Code Taxonomic groups A J 180 CALLI Callionymidae 181 CSP Callionymus sp. 1 5 0 1 E 182 ELEO Eleotridae 183 GGUAV Guavina guavina (Valenciennes, 1837) 3 5 1 1 C 184 GOBII Gobiidae 185 ELYRI Evorthodus lyricus (Girard, 1858) 3 5 1 1 C 186 GOCEA Gobionellus oceanicus (Pallas,1770) 3 6 1 1 C 187 CSMAR Ctenogobius smaragdus (Valenciennes, 1837) 3 5 1 1 C 188 CSTIG Ctenogobius stigmaticus (Poey, 1860) 3 5 1 1 C 189 EPHIP Ephippidae 190 ZNOVE Zabidius novemaculeatus (McCulloch, 1916) 1 3 1 0 E 191 CFABE Chaetodipterus faber (Broussonet, 1782) 1 3 0 1 C 192 SCATO Scatophagidae 193 SARGUS Scatophagus argus (Linaeus, 1766) 2 3 1 1 E 194 SMULT Selenotoca multifasciata (Richardson, 1846) 2 3 1 1 E 195 SIGN Siganidae 196 SCANA Siganus canaliculatus (Park, 1797) 2 3 0 1 E 197 SPHYR Sphyraenidae 198 SPUTN Sphyraena putnamae Jordan and Seale, 1905 2 4 0 1 E 199 TRICH Trichiuridae 200 TLEPT Trichiurus lepturus Linnaeus, 1758 1 4 0 1 C 201 SCOMB Scombridae 202 SMACU Scomberomorus maculatus (Mitchill, 1815) 1 4 0 1 C 203 PARAL Paralichthyidae 204 PELEV Pseudorhombus elevatus Ogilby, 1912 2 5 0 1 E 205 ACHIR Achiridae 206 ADUME Apionichthys dumerili Kaup, 1858 3 5 1 1 C 207 ALINEA Achirus lineatus (Linnaeus, 1758) 3 5 1 1 C 208 TRIAC Triacanthidae 209 TWEBER Trixiphichthys weberi (Chaudhuri, 1910) 2 6 0 1 E 210 MONAC Monacanthidae 211 ATOMEN Acreichthys tomentosus (Linnaeus, 1758) 2 6 0 1 E 212 MCHINE Monacanthus chinensis (Osbeck, 1765) 2 3 0 1 E 213 PJAPO Paramonacanthus japonicus (Tilesius, 1809) 1 6 0 1 E 214 TETRA Tetraodontidae 215 AIMMA Arothron immaculatus (Bloch and Schneider, 1801) 2 5 1 1 E 216 CPATO Chelonodon patoca (Hamilton, 1822) 2 5 0 1 E 217 TERYTH Tetraodon erythrotaenia Bleeker, 1853 3 5 1 1 E 218 CPSIT Colomesus psittacus (Bloch and Schneider, 1801) 3 5 1 1 C * Originally cited in Appendix as Arius mastersi (non-valid). ** Originally cited in Appendix as Hexanematichthys parkeri (non-valid). *** Originally cited in Appendix as Hexanematichthys proops (non-valid). BULLETIN OF MARINE SCIENCE, 80(3): 681–720, 2007

Diel variation in mangrove fish abundances and trophic guilds of northeastern Australian estuaries with a proposed trophodynamic model

Janet A. Ley and Ian A. Halliday

ABSTRACT Diel activity patterns of tropical fish assemblages in turbid, mangrove-dominated estuaries remain largely undocumented, leading to uncertainty about ecological processes in these systems. To capture active fishes by day and night, gill nets were set perpendicular to mangrove shorelines, in six northeastern Australian estuaries during 13 bimonthly trips. Fish were sampled with eight large mesh (102–151 mm) nets, set for 6 hrs (1500–2100), and checked hourly (1146 day, 635 dusk, 872 night checks). Four smaller mesh (19–51 mm) nets were also set for 1 hr before and after sunset (77 day, 78 night checks). Of 157 total species, 22 were netted exclusively before sunset and 47 exclusively after sunset. All of the top 26 species were present both day and night, but of these, 46% were primarily nocturnal (diel index > 0.65). An average of 77.2 fish hr–1 were netted by day vs 171.4 by night. Within the 400 km coastal region, assemblages differed between two northern wave-dominated (WD) estuaries and four southern tide-dominated (TD) estuaries. In all six estuaries Lates calcarifer (Bloch, 1790) dominated night assemblages. In TD estuaries, night assemblages were also dominated by Thryssa hamiltoniG ray, 1835 and Eleutheronema tetradactylum (Shaw, 1804); while in WD estuaries Herklotsichthys castelnaui (Ogilby, 1897), Leiognathus equulus (Forsskål, 1775), and Megalops cyprinoids (Broussonet, 1782) were dominant at night. Nocturnal species included planktivores and carnivores, while daytime assemblages were dominated by detritivores (Mugillidae). Higher night catch rates are attributed to increased activity by mobile fishes moving from mangrove to adjacent habitats to forage, especially immediately post-sunset. Although day–night diets and forage resources have yet to be compared in mangrove systems, previously unrecognized trophic relationships involving variation in diel activity among important fishery species (Centropomidae,P olynemidae, Ca- rangidae) and their prey may be key ecological processes in these tropical mangrove estuaries. A proposed hypothesis explaining diel variation in mangrove fish assemblages of tropical estuaries is presented through a conceptual model.

The 24-hr (diel) periodicity of the earth’s rotation creates a predictable pattern of light and dark, and in synchrony with this cycle, the majority of animals have evolved to be active (e.g., gathering food) either by day, by night, or during crepuscular periods (dawn, dusk) (Helfman et al., 1997, Kronfeld-Schor and Dayan, 2003). Such partitioning of the daily cycle among fish taxa may promote co-existence in ecological communities among competitors, and among predators and their prey (Schoener, 1974). Thus, habitat partitioning in time should lead to variation in species composition and use of resources, ecological properties which can only be measured through surveys encompassing multiple temporal dimensions. Ecological monitoring programs commonly cover a range of spatial and seasonal scales, but despite the fact that many aquatic ecosystems are known to have strong diel variations in species activity (tropical lakes and streams, Lowe-McConnell, 1987; coral reefs, Hobson, 1991; nearshore mesopelagic zones, Benoit-Bird and Au, 2006), investigations are usually limited to daytime observations. Faunal surveys that include only daytime sampling have been

Bulletin of Marine Science 681 © 2007 Rosenstiel School of Marine and Atmospheric Science of the University of Miami 682 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 shown to underestimate abundance, diversity, and biomass (Stoner, 1991; Rountree and Able, 1993; Griffiths, 2001), and may fail to identify important ecological interactions. The current study focuses on comparing the composition and ecological characteristics of fish assemblages in several tropical estuaries at a hierarchy of spatial and temporal scales, emphasizing diel variation in mangrove fringe habitats. Diurnally active fishes mainly rely on vision to find food, while nocturnally active animals are more likely to obtain their food by tactile probing, smell, and hearing (Kronfeld-Schor and Dayan, 2003). Specialized adaptations lead to decreased effectiveness outside normal activity periods, temporally limiting individuals to actively seeking food during certain times of the day or night (Hobson and Chess, 1977). Thus, even though relatively fine-scale environmental factors may influence diel- ac tivity within a fish assemblage, the capacity of individuals to vary basic patterns is limited by physiological, morphological, and behavioral constraints (Kronfeld-Schor and Dayan, 2003). Consequently, intrinsic adaptations to the day–night cycle may lead to predictable community patterns in diel habitat partitioning in fish assemblages in similar habitats and settings. Diel variation in fish assemblages has been studied most comprehensively on coral reefs, revealing the following patterns (see Livingston, 2003). Diel Variation: Coral Reef Fishes.—Predator-prey interactions on coral reefs are largely structured around the diel cycle. By day, small copepods in the water column provide a forage base for schooling planktivorous fishes that are themselves well protected from diurnal predators such as Sphyraenidae and Carangidae by “many eyes” (Hobson, 1991). After feeding throughout daylight hours, some species of planktivores (e.g., Pomacentridae, Caesionidae) descend at dusk to the reef structure. However, also at dusk, zooplankton densities increase over the reef as demersal fauna, including decapod crustaceans and polychaetes, emerge from the reef substrate and rise into the water column (Hobson, 1991; Yahel et al., 2005). Nocturnally active planktivores (e.g., Apogonidae, Pemphridae) that school within or near the reef substrate by day, ascend to feed. Once darkness falls completely, these small fishes are protected by reduced predator effectiveness at low light levelsH ( obson, 1972, 1974). However, both diurnal and nocturnal planktivores are most vulnerable to predation during twilight (dusk, dawn) when they descend to, or ascend from, the reef structure. Crepuscular predators such as Serranidae assume positions low on the reef and lunge upwards to capture exposed planktivores (Hobson, 1991). Adap- tations of crepuscular predators include retina composition suited to reduced light levels, morphology and tail shape for upward bursts of speed, and camouflage coloration. Do diel adaptations and species interactions identified for coral reef fish communities provide a model that can be applied to other aquatic ecosystems? Other habitats are far less well-studied, but diel processes frequently involve both vertical and horizontal movements by mobile fishes between habitats (Benoit-Bird and Au, 2006), as discussed below. Diel Variation: Habitat Shifts.—Some species of Haemulidae and Lutjanidae migrate from coral reefs at dusk to forage on emergent invertebrates in nearby habitats such as seagrass beds and lagoons (Robblee and Zieman, 1984; Helfman, 1993; Yahel et al., 2005). Recently, similar diel habitat shifts have also been documented for temperate planktivores (i.e., Pemphridae) that ascend at dusk from their day shel- ters in caves to forage on emergent fauna such as amphipods, ostracods, and mysids in the water column above rocky reefs (Annese and Kingsford, 2005). Most studies LEY AND HALLIDAY: DIEL VARIATION IN MANGROVE FISH ASSEMBLAGES 683 detailing diel dynamics have been conducted in clear water, allowing direct observation of transitional movements between habitats by individuals and groups of fishes. Generally, in a variety of clear water settings, diel habitat shifts can be summarized as processes in which certain fishes: (1) rest in a darkened or structurally complex location, relatively safe from predators by day; (2) move to a feeding habitat that may provide less protective structure at dusk; and (3) return to daytime resting habitats at dawn. Critically, surviving such activity requires a fish to risk crossing a transition zone between safe and more open habitats. Sheaves (2005) suggested that, for fishes, such transition zones are bottlenecks in the spatial complex of habitats within an aquatic ecosystem due to the high risk of predation. Intersecting boundaries of habitats would thus be critical passages for mobile species where predator avoidance must be accomplished through defensive behavioral or morphological adaptations. Progress toward identifying general trends in diel variation in a variety of ecosystems, including mangroves, might be achieved by focusing survey efforts on areas of transition at a time when movement is most likely to occur, i.e., twilight. Diel Variation: Mangroves.—Well-developed mangrove shorelines are nursery habitats for late-stage juveniles and sub-adults of transient species of fisheries importance (Rooker and Dennis, 1991; Ley et al., 1999). However, numerous scientific questions remain regarding how mangroves function as nursery habitats, and as habitats for fishes in general (e.g., Sheridan and Hays, 2003). Key among these questions concerns connectivity of mangroves to other habitats through movement of fishes. Given processes discussed so far, diel movement of fishes between more structured resting locations (i.e., mangrove roots) and less structured feeding areas may be critical in mangrove food webs supporting transient juvenile species and maintaining biodiversity in fish assemblages. Mangroves occupy a variety of environments, including limited areas along clear water shorelines, primarily in offshore island settings. Focusing on fringing mangroves adjacent to a coral reef in Puerto Rico, Rooker and Dennis (1991) conducted a diver-based survey that included twilight and day–night observations. During the day, planktivores (e.g., Clupeidae), intermediate carnivore/predators (e.g., Haemu- lidae), and herbivores (e.g., Acanthuridae), were among the abundant families observed within fringing mangrove habitats. At night, almost all of these fishes left the mangrove habitat, with no replacement species observed entering the mangroves. Twilight movements out of the mangroves began about 1 hr before sunset and ended about 30 min after sunset. While Haemulidae gathered at one launching point and streamed away from the mangroves, other species exited haphazardly. Similarly, in a small tropical embayment without live coral reefs, Haemulidae, Lutjanidae, and Di- odontidae left the safety of mangroves and other non-coral structures (i.e., boulders, rocky crevices) at dusk and swam to seagrass and macroalgal beds to feed on emergent crustaceans and mollusks (Nagelkerken et al., 2000). These studies suggest that, not unlike coral reef habitats, diel variation in fish assemblages within clear-water mangrove habitats associated with coral reefs includes predictable movement across the fringe by planktivorous and intermediate carnivorous species into other habitats, primarily focused around dusk. Most tropical mangrove habitats are located in sediment-laden waters along tropical coasts (Woodroffe, 1992) making visual censuses an impractical sampling method. Diel variation hypotheses are yet to be tested in these more commonly occurring mangrove habitats. The aim of this study is to investigate ecological implications of 684 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 diel variation in fish assemblages along the mangrove fringe in a series of relatively turbid, tropical estuaries, seeking to: (1) Characterize diel variation in fish assemblages, considering overall day–night patterns and components of the assemblages responsible for the observed patterns; (2) Identify consistency in diel variation patterns for key diurnally- and nocturnally-active species, considering trends among the estuaries and months; (3) Identify activity patterns during day, dusk, and at night for key predator species; and (4) Examine implications of the findings for trophodynamic processes in tropical mangrove estuaries.

Methods

Study Sites.—The study area extends 400 km along the northeastern Australian coast adjacent to the Great Barrier Reef (Fig. 1). Six large estuaries (catchment > 150 km2, combined area of water and mangrove forests > 4 km2) were selected. The estuaries have very little clearing along continuously mangrove-covered shorelines [Avicennia marina (Forssk.) Vierh., Rhi- zophora stylosa Griff., Ceriops spp.], no wastewater input, but moderate levels of agricultural runoff (AED, 2005). Regulated commercial gill-net fishing is permitted in three of the estuaries (Nobbies, Barrattas, and Hull), while limited recreational line-fishing is permitted in all six estuaries. Although regional semi-diurnal tidal levels range from < 2 m on neap to > 4 m on spring tides, all sampling was conducted on six consecutive days over the period of neap tides (0.5–1.8 m range). This tropical coast has a rainy season (Nov/Dec–Feb/Mar), but average annual rainfall is greater in the north (Table 1). The sampling design (partially nested, full factorial) differentiated the estuaries for comparison using the following framework: (i) Region (North, Mid, South): recognizing the gradient in climate and landscape conditions; (ii) Fishing (Open, Closed) nested within region: for each region, one estuary in which commercial fishing has been banned (closed), one nearby (within 80 km) estuary open to commercial net fishing (open).

Figure 1. Study sites in northeastern Queensland, Australia. Estuaries are either closed to commercial fishing (cl); or open (o). North, Mid, and South refer to regions. LEY AND HALLIDAY: DIEL VARIATION IN MANGROVE FISH ASSEMBLAGES 685

Thus, for example, Haughton (mid closed) is located in the mid region and is closed to commercial fishing (Table 1). The present study is part of a larger study testing the effects of commercial netfishing on estuarine ecosystems. Further details can be found in Halliday et al. (2001), Ley et al. (2002), Ley and Halliday (2003), and Ley (2005). Sampling Procedures.—Bimonthly sampling was conducted from Mar 1998 through Mar 2000 (n = 13 trips). Two crews with identical sets of gill nets fished upstream sites (as far as navigable, 2–7 km from the mouth) and downstream sites (within 1 km of the mouth) simultaneously. Each set of six nets consisted of two replicate 102 mm and 152 mm mesh nets, a single 51 mm mesh net, and a multi-panel net comprised of 19, 25, and 32 mm mesh segments (see Halliday et al., 2001). With one end tied to a mangrove trunk on top of the bank, nets were set perpendicularly to the bank (and hence, tidal flow), passing through 1–3 m of the fringe of overhanging mangrove roots and branches. The outer end of the net was fixed in position midstream by attaching an anchor. The replicate 102 mm and 152 mm mesh nets were set at approximately 1500, checked for the presence of meshed fish every 1–1.5 hrs after setting, and retrieved as close to 2100 as possible. In order to avoid excessively large catches of small schooling fishes, the smaller mesh nets (51 mm and multi-panel) were only fished for 1 hr either side of sunset. For the daytime soak, both small mesh nets were set at approximately 1630 and retrieved at 1730; they were re-set for the night-time soak at 1900 with final retrieval at 2000. Throughout the 6 hr net-soak period, salinity and temperature were measured at 5 min intervals by one or two Hydrolab Datasonds near net-deployment sites. Water level was also measured at each time interval and resulting values were analyzed to identify tidal influence (i.e., rising, falling water levels). Data Analysis.—Raw data for each net set and check were converted to catch per hour (CPUE) to account for variations in duration of periods between checks. These values were aggregated for two types of analyses: (1) day–night and (2) day/dusk/night. For the day–night analysis, catch hr–1 check–1 net–1 was summed for all nets, and then totaled for all post-sunset checks (night, n = 78 samples) and all pre-sunset checks (day, n = 77 samples). This set of 155 day–night samples was subjected to multivariate and ANOVA analyses (see below). For the day/dusk/night analysis (151 and 102 mm mesh nets only), each observation of catch hr–1 net–1 check–1 was assigned to one of five check periods based on the mid-point of the time interval of each check: 1 = 1430–1600 (day), 2 = 1601–1730 (day), 3 = 1731–1900 (dusk, sunset), 4 = 1901–2030 (night), and 5 = 2031–2300 (night). The resulting set of day/dusk/night samples contained 2653 observations (1446 day, 635 crepuscular, 872 night) distributed among the estuaries as shown in Table 1 and analyzed using ANOVA (see below). Multivariate analyses of the 155 day–night samples were facilitated by Primer V6 (Primer- E Ltd, Plymouth, 2004). Following 4th-root transformation, the Bray-Curtis index was used to derive a matrix of similarity values between pairs of samples (Clarke, 1993). Relationships among groups were ascertained using non-metric multidimensional scaling (MDS). Relative within-group variability was analyzed by the index of multivariate dispersion (IMD, MVDISP routine) (Warwick and Clarke, 1993). Significance of each grouping factor (region, month, fishing, diel) was tested using one-way analysis of similarities (ANOSIM). Species responsible for distinguishing groups were identified by a discrimination index, defined as the ratio of the average contribution to similarity between groups (numerator), to the standard deviation of similarity between groups (denominator) (SIMPER). The higher the value of the discrimination index, the more informative the species was for discriminating between groups (Clarke and Warwick, 2001). A second index indicated the probability that a family would occur in a group (typifying index) based its average similarity within a group (numerator) divided by the standard deviation (denominator) (Clarke and Warwick, 2001). BIO-ENV revealed correlations between environmental variables (diel period, salinity, temperature, and tide) and the similarity matrix based on CPUE by species. For the first set of ANOVAs, focusing on day–night catch rates for selected species as dependent variables (n = 155 samples), independent variables were region, fishing, diel period 686 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

77 78 23 23 27 26 551 635 218 595 654 Total (all estuaries) Total

6 0.2 43 13 13 13 17 26 25 102 109 109 106 North Closed Russell 3,016 Wave, True estuary True Wave,

0.1 96 12 98 35 13 13 17 95 17 26 26 Hull Open North 116 2,855 Wave, True estuary True Wave,

3 91 32 47 13 13 26 27 27 27 Mid 888 128 106 108 Closed Haughton Tide, Delta Tide,

3 91 18 93 34 13 13 23 95 22 28 27 111 Mid 888 Open Barrattas Tide, Delta Tide, 0.8 84 19 44 12 13 27 25 26 26 100 106 101 South Closed 1,045 Yellow Gin Yellow Tide, Creek Tide,

2.5 87 14 15 13 13 89 31 31 27 26 Open 107 107 South Nobbies 1,045 Tide, Creek Tide, ) 2 6 (all nets) 1 2 (102 and 5 4 3 1601–1730 1731–1900 1901–2030 2031–2300 1430–1600 st nd rd th th Fishing River Mouth width (km) (mm) rainfall ann. Av. No. samples 151 mm mesh) 1 Table 1. Overview of design features and general catch data. Table Design feature Region Estuary type Mangrove area (km Day Day 2 4 Salinity mean Night No. checks Night 3 5 Temperature mean (°C) Temperature Day Night LEY AND HALLIDAY: DIEL VARIATION IN MANGROVE FISH ASSEMBLAGES 687 i in 110 135 157 7,616 17,292 24,908 56 83 94 1,200 4,407 5,607 48 69 78 1,127 2,349 3,476 64 72 85 1,304 2,356 3,660 56 69 77 1,193 2,667 3,860 63 61 75 1,808 3,157 4,965 63 75 90 984 2,356 3,340 Check = “hourly” removal of fish caught during each net soak period from initial set time (approx. 1500) to pickup at (approx. 2100) (variation due to logistical difficulties Sample = sum of 12 nets set during each trip in river (see text) Day = pre-sunset Night = post-sunset near net sites during soak periods Salinity and temperature averaged from Hydrolab data recorded at 5 min intervals Wave = wave-dominated, Tide= tide-dominated; creek, delta, true estuary. Details on estuary type, mangrove area, mouth width, rainfall in Ley (2005) tide-dominated; creek, delta, true estuary. Tide= = wave-dominated, Wave removing sporadic large catches from the nets) removing sporadic large Table 1. Continued. Table Catch overview No. species (all nets) Day 1 2 3 4 5 6 Night Total abund. (all nets) Tot. Day Night Total 688 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

(day, night), and month (Jan, Mar, May, Jul, Sep, Nov with n = 2, 3, 2, 2, 2, and 2 replicates respectively). For the second set of ANOVAs, focusing on day/dusk/night samples (n = 2653), dependent variables were catch rates for selected species; independent variables were identical to the day/night dataset except that the five diel periods (check times 1 through 5) replaced day/night as a factor. Prior to analyses, catch rates were natural log (x + 1) transformed. All interactions up to 3-way were estimated. For each significant main effect and interaction, post hoc multiple comparison tests were conducted (Fishers LSD test). Most species were caught both day and night, and in order to discern nocturnal or diurnal tendencies, a diel ratio was calculated as the average catch rate collected at night (n = 78) divided by the sum of the average night and day catch rates (n = 155) [adapted from Saigusa (2001) and Saigusa et al. (2003)]. In this system, diel variation patterns were classified into four nocturnal and four diurnal categories (each) on the basis of the diel ratio value, ranging from no clear diel activity (N0 and D0, “continually active”) to increasingly nocturnal (or diurnal) (N1–N3, D1–D3) (Table 2). Diets and residency information for key species were obtained from FishBase (Froese and Pauly, 2005) and other published sources. Fishes were classified into five trophic categoriesH ( alliday and Young, 1996; Ronnback et al., 1998): detritivores (D); herbivores (H); plantivore/microinvertebrate carnivores (Pl) (feeding on plankton, small invertebrates, and small amounts of detritus); intermediate carnivores (I) (feeding on large invertebrates with some vegetation and small fish); predators (P) (feeding on fish, with some large invertebrates). Residency status was categorized by considering which of four life history stages (larval, juvenile, adult, spawning) occurs in estuaries, resulting in three categories: residents (all stages in the estuary), transients (juvenile stages in the estuary), occasional visitors (adults entering the estuary mainly to feed) (Day et al., 1989).

Results

Environmental Conditions Salinity ranged from 0.1 (Mar, 1999, north open estuary) to 43 (May, 1999, south open estuary), averaging 23 both day and night (Table 1). Water temperature ranged from 20 °C (July, 1999) to 32 °C (Mar, 1999), averaging 27 °C by day and 26 °C at night. A predominant seasonal pattern was evident in all estuaries, i.e., higher water temperatures associated with lower salinity periods during Nov–Mar. However, lower and more variable salinity conditions were recorded in the two estuaries located in the north region.

Overall Patterns in Catch Rates In total, 157 species from 55 fish families were netted (Table 1). Twenty-two species were netted only during the day, compared with 47 species netted only at night (Appendix I). All of these 69 species were rare (< 10 individuals each). Of the total catch of 24,908 fish, only 30.6% were caught during the day. Fifty percent of the total catch was comprised of three species: Herklotsichthys castelnaui, Thryssa hamiltoni, and Lates calcarifer, and each of these species was more active at night. An average of 77.2 fish were netted hr–1 by day and 171.4 by night in the 155 day–night samples (Table 2). Of the 26 top species that comprised 90% of the cumulative catch, none was caught exclusively in the day or night (Table 2). Based on the diel ratio, 46.2% of the top 26 species were nocturnal (N1, N2 or N3); these 12 species comprised 77.5% of the total catch (Table 2, Figs. 2A,B). For all species, 57.1% were nocturnal, comprising 74.9% of total catch (Appendix 1, Figs. 2C,D). Only three of the top 26 species were diurnally active (D1, D2 or D3), and of these, three trophic categories were represented by one species each, i.e., a detritivore, an herbivore, and a plankti- LEY AND HALLIDAY: DIEL VARIATION IN MANGROVE FISH ASSEMBLAGES 689 3 Res Res Res Res Res Res Res Res Res Tran Tran Tran Tran Tran Tran Tran Tran Tran Tran Tran Residency 2 I I I I I I I I P P P D D D D Pl Pl Pl Pl Pl group Trophic Trophic C C C C C C C C N N N N N N N N N D N N Sum. 1 N1 N2 N2 N0 N0 N2 N2 N2 N1 N1 N0 N1 D0 D1 D0 N1 N0 N2 N0 D0 Group Diel categories 0.73 0.89 0.82 0.51 0.53 0.84 0.86 0.83 0.69 0.71 0.60 0.76 0.43 0.30 0.39 0.71 0.56 0.82 0.63 0.44 Ratio 4.72 2.58 1.85 3.00 2.25 8.04 1.91 5.86 4.45 0.92 2.96 5.06 1.66 2.38 1.71 1.16 –1 Avg. night 57.33 20.58 12.45 10.29 (n = 78) Catch hr 0.61 4.67 2.50 1.67 2.46 0.48 2.08 1.00 3.32 1.27 1.82 5.78 2.14 4.70 2.05 1.30 0.51 0.99 1.49 day Avg. 20.85 (n = 77) 31 72 42 75 76 48 77 53 79 57 80 64 61 81 67 70 82 83 85 86 Cum. % % 66 43 60 46 56 77 43 63 21 60 69 52 63 29 63 51 35 18 50 50 occur. Freq. of mm 32–67 58–84 49–150 42–240 46–990 21–235 48–630 80–900 54–380 54–380 52–390 44–315 65–520 220–540 170–475 180–568 151–734 220–450 204–650 Size range 170–1,020 Herklotsichthys castelnaui Ambassis vachelli Species Thryssa hamiltoni Arius nella Scomberoides commersonianus Scomberoides Lates calcarifer Nibea soldado Leiognathus equulus Arius spp. Arius proximus Pomadasys kaakan Eleutheronema tetradactylum Eleutheronema Valamugil cunnesius Valamugil Valamugil buchanani Valamugil Liza subviridis Nematalosa come Strongylura strongylura Strongylura Escualosa thoracata Pomadasys argenteus Liza vaigiensis Table 2. Overview of characteristics of the top cumulative 90% of fish caught by summed catch rate (n = 26 species). See Appendix I for full list. 2. Overview of characteristics the top cumulative 90% fish caught by summed catch rate (n = 26 species). See Table Clupeidae (herrings) Ambassidae (glassfish) Family (common name) Engraulidae (anchovies) Ariidae Carangidae (trevally) Centropomidae (barramundi) Sciaenidae (drums) Leiognathidae (ponyfish) Ariidae Ariidae (sea catfish) Haemulidae (grunts) Polynemidae (threadfin) Mugilidae (mullet) Mugilidae Mugilidae Clupeidae Belonidae (needlefish) Clupeidae Haemulidae Mugilidae 690 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 3 Res Res Res Tran Tran Tran Tran Residency 2 P P P D H Pl Pl group Trophic Trophic I C C C D D N Sum. 1 D0 D0 D1 D2 N2 N0 N1 N0 Group Diel categories 0.44 0.36 0.32 0.18 0.83 0.61 0.69 0.63 Ratio 1.16 0.78 0.67 0.35 1.47 1.01 1.00 –1 Avg. night 171.4 (n = 78) Catch hr 1.49 1.39 1.41 1.59 0.30 0.64 0.57 day Avg. 77.2 (n = 77) 86 87 87 88 89 89 90 Cum. % 7 % 50 22 23 32 52 30 occur. Freq. of mm 50–95 42–130 58–230 204–650 175–525 340–860 25–1,040 20–1,520 Size range Liza vaigiensis Anodontostoma chacunda Species Arrhamphus sclerolepis Arrhamphus Sardinella brama Sardinella Megalops cyprinoides Caranx ignobilis Polydactylus macrochir Diel groups: Diurnal D = D3 (r < 0.05); D2 (0.05 ≤ r < 0.20); and D1 (0.20 ≤ r < 0.35); Continually active C = D0 (0.35 ≤ r < 0.50); N0 (0.50 ≤ r < 0.65); Nocturnal N = N1 (0.65 = predators (see text) groups: D = detritivore; H herbivore; Pl planktivore/microinvertebrate-carnivore; I intermediate carnivore; P Trophic (see text) Transients Tran: Residency: Res: Estuarine residents; X ≤ r < 0.80; N2 (0.80 0.95); and N3 (0.95 r). (Adapted from Saigusa, 2001) Table 2. Continued. Table Mugilidae 1 2 3 Clupeidae Family (common name) Hemiramphidae (halfbeak) Clupeidae Megalopidae (tarpon) Carangidae Polynemidae Total (all 157 sp) Total LEY AND HALLIDAY: DIEL VARIATION IN MANGROVE FISH ASSEMBLAGES 691

B Percent of top Percent 26 of species

Diel groups Figure 2. Summary fish netted divided into three general diel groups determined from the Sai- gusa diel index (see text and Table 2). Percent of (A) the top 26 species which comprised 90% of the cumulative catch by diel group; (B) the catch per hour per sample (n = 155) for the top 26 species by diel group; (C) (see next page) all 157 species by diel group; and (D) the catch per hour per sample for all species by diel group.

vore (Table 2). The group that was active both day and night (continually active, D0 or N0) included 11 species represented by all trophic groups except herbivores. The nocturnal group was comprised of 12 species including planktivores, intermediate carnivores, and predators. Twelve of the 26 top species were life-time residents of estuaries and of these, 7 were nocturnal (Table 2). The other 19 of the top species were estuarine transients and of these, five were nocturnal. None of the top 26 species was considered to be an occasional visitor.

Diel Patterns in Fish Assemblages The 155 day–night samples were ordinated by rank similarity, generating an initial 2-dimensional MDS diagram (not illustrated). Night samples clustered at the upper central portion of this initial MDS diagram. ANOSIM revealed that fish as- 692 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

D Percent of top Percent 26 of species

Diel groups Figure 2. Continued. semblages differed significantly between night and day, but Global R was moderate (Table 3). The index of multivariate dispersion was 0.667 for the night samples and 1.342 for the day samples, indicating that night assemblages tended to be comprised of a distinctive and coherent group of species, while the day assemblages were more variable. Month was also a significant source of variation but Global R was only 6.15%. In- spection of the paired comparisons revealed that May, July, and Sept (dry season) tended to separate from Nov and Jan (wet season). Region was a relatively important source of variation with a Global R of 22.6%. Paired comparison results indicated little separation between mid and south regions in the original three region analysis; furthermore, inspection of the MDS diagram revealed that mid and south samples were well intermixed. When the mid and south regions were combined into one meta-region (i.e., mid+south), ANOSIM revealed a substantial separation from fish assemblages in the north region (Global R = 32.3%). Finally, the two strongest factors LEY AND HALLIDAY: DIEL VARIATION IN MANGROVE FISH ASSEMBLAGES 693

Table 3. Summary of the results of analyses of similarity (ANOSIM) based on 155 day-night samples (all species included in the analyses). In all cases, P < 0.001.

Grouping factor Number of Global Pairwise comparisons groups relationship R Groups R (%) (%) 1. Diel 2 17.1 Day vs night 2. Month 6 6.15 Jan, Mar, May, Jul, Sep, Nov 3. Fishing 2 4.3 Open vs closed 4. Region 3 3 22.6 Mid, North 35.9 Mid, South 8.0 North, South 24.1 5. Region 2 2 32.3 North vs Mid+south 6. DN Region 2 4 31.2 dMid+south, nMid+south 14.6 dMid+south, dNorth 40.2 dMid+south, nNorth 25.6 nMid+south, dNorth 69.1 nMid+south, nNorth 35.0 dNorth, nNorth 27.6

[DN (day, night) and Region 2 (north, mid+south)] were combined and each of the 155 samples was assigned to one of the four new groups, defined as: (1) dNorth: day, north region (Hull, Russell); (2) nNorth: night, north region; (3) dMid+south: day, mid+south meta-region (Barrattas, Haughton, Yellow Gin, Nob- bies); and (4) nMid+south: night, mid+south meta-region. The significant Global R of 31.2% indicated that this four group configuration (DN Region 2) provided the most comprehensive grouping factor, accounting for both regional effects and diel variation. The modified MDS diagram, with symbols representing the four groups, clarified group differences (Fig. 3). Thus, composition of the fish assemblages was markedly influenced by both diel period and regional characteristics

Figure 3. Multidimensional scaling diagram illustrating the four group configuration based on diel group (Day–night, d, n) and meta-region (north, mid+south). Symbols: solid = night; hollow = day; circles = north region; triangles = mid+south meta-region 694 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 -root vs 1.4 1.5 1.3 1.4 1.0 1.5 th south nMid+ nNorth vs 1.2 1.2 1.7 1.6 1.5 0.8 south nMid+ dNorth vs 1.3 1.2 1.4 0.6 1.5 1.4 dNorth nNorth vs 1.2 1.1 1.4 1.2 1.2 0.7 dMid +south nMid+south C. Discrimination index vs 1.9 1.8 1.3 1.2 1.2 1.6 dMid +south nNorth vs 1.2 1.2 0.8 1.1 1.1 0.6 dMid +south dNorth 3.5 3.8 1.1 0.1 2.6 1.4 nNorth 1.0 1.4 0.1 0.1 0.4 0.3 dNorth + 1.0 0.8 1.7 1.3 2.4 0.1 nMid south B. Typifying index Typifying B. + 0.3 0.3 0.5 0.7 0.8 0.1 dMid south 0.3 9.8 3.5 90.9 21.9 10.4 nNorth 4.0 0.5 0.1 0.1 0.4 41.6 dNorth cluster + 4.6 8.6 0.5 40.6 25.7 13.8 nMid south (untransformed data) + 1.1 6.8 2.7 3.2 0.2 10.3 A. Average catch per hour by Average A. dMid south Planktivore/microinvertebrate carnivores (Pl) Herklotsichthys castelnaui Leiognathus equulus Intermediate carnivores (I) and predators (P) Thryssa hamiltoni tetradactylum Eleutheronema Lates calcarifer Megalops cyprinoides transformed for the analysis. The species listed in this summary table had typifying and/or discrimination indices > 1.5 for at least one comparison. transformed for the analysis. Table 4. Comparison of four MDS sub-groups by top discriminating species based on SIMPER analysis of catch per hour data for 155 samples. Data were 4 were Data samples. 155 for data hour per catch of analysis SIMPER on based species discriminating top by sub-groups MDS four of Comparison 4. Table Groups species LEY AND HALLIDAY: DIEL VARIATION IN MANGROVE FISH ASSEMBLAGES 695 0.28** Fishing

Region 2.19*** 5.08*** Mean squares Mean squares Interactions of Diel with: Interactions of Diel with:

(significance ranges below) 4.35* 0.85** 1.98* 1.75** Month 12.87***

8.44** 6.87**** 1.67* 7.10**** Fishing(R) 10.63**** Region 5.22**** 4.71**** 13.48*** 35.63**** 10.45**** 42.44**** 18.56**** Main effects Main effects Mean squares Mean squares Month

1.32*** 2.33*** 3.48**** 3.16*** 1.31* (significance ranges below) Diel 6.64**** 5.58*** 35.58**** 70.72**** 79.13**** 26.58**** 84.19**** 1.47 1.85 1.01 4.72 5.86 Night 10.29 20.58 12.45 57.33 2.08 4.67 2.46 0.30 1.67 0.64 0.61 1.82 Day hour per sample Average catch per Average 20.85 208 366 182 536 691 1,383 2,263 1,656 7,494 Total catch Total Table 5. Summary of results for analysis of variance (ANOVA) based on 155 day–night samples. The natural log-transformed (ln + 1) data were analyzed for nine for analyzed were data 1) + (ln log-transformed natural The samples. day–night 155 on based (ANOVA) variance of analysis for results of Summary 5. Table key species. None of the 3- or 4-way interactions were significant. Leiognathus equulus Intermediate carnivores (I) and Predators (P) Thryssa hamiltoni Lates calcarifer Megalops cyprinoides commersonianus Scomberoides P < 0.0001 P < 0.001; **** P < 0.01; *** P < 0.05; ** Significance ranges: * Caranx ignobilis Tropic category / species Tropic Planktivore/microinvertebrate carnivores (Pl) Ambassis vachelli Eleutheronema tetradactylum Eleutheronema Herklotsichthys castelnaui 696 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 Figure 4. Average catch all perfor hour Figurenets day (pre-sunset, per= by 155) Average 4. sample (n lighter line diamond withand hollow night symbols) (post-sunset, darker line with Error solid square bars symbols). represent confidence upper 95% limitsfor each data point. LEY AND HALLIDAY: DIEL VARIATION IN MANGROVE FISH ASSEMBLAGES 697

leading to distinct groups of samples. However, although each paired group comparison was significant P( < 0.001), R values ranged from 14.6 (quite similar) for dMid+south vs nMid+south to 69.1 (very different) for nMid+south vs dNorth (Table 3). SIMPER analysis of the four MDS groups revealed species responsible for distinctions in fish assemblages (Table 4). A criterion was selected to identify key species in each group, such that either the typifying or discriminating indices calculated for a species exceeded 1.5 in at least one comparison. This strict criterion indicates that a key species was both abundantly and frequently netted in a particular MDS group. Six key species were thus identified as most important in distinguishing the four groups. Two of the key species were planktivore/microinvertebrate feeders (Pl) and four were intermediate carnivores and top predators (I and P). Lates calcarifer typified night samples in both regions (typifying index > 2.4 nMid+south and nNorth). Night mid+south samples were also typified by greater catch rates ofEleutheronema tetradactylum and T. hamiltoni, while night north samples were most typified by H. castelnaui, Leiognathus equulus, and Megalops cyprinoides. Day samples in the mid+south estuaries (dMid+south) were most typified by T. hamiltoni, L. calcarifer, and E. tetradactylum but index values were relatively low (< 0.8). Day samples in the north estuaries (dNorth) were most typified byL. equulus (typifying index = 1.4), but the nNorth index value was much greater for this species (3.8).

Analysis of Fish Assemblages and Environmental Variables The BIO-ENV analysis of rank correlations between fish and environmental matrices revealed that diel period was the only variable of importance. Tide, water temperature, and salinity each had correlations with the ranks of the fish assemblage matrix of < 10%. Adding any or all of these variables to the abiotic variable matrix did not improve the correlations. Thus, tested sources of temporal variation other than diel period had little influence on the fish assemblages.

Analysis of Individual Species Using the original design based on three regions, ANOVAs were conducted for the six key species identified in the SIMPER analysis as strongly influencing the nocturnal assemblages. In addition, three other ecologically important species were analyzed: Ambassis vachelli, Caranx ignobilis, and Scomberoides commersonianus. Thus, spatial and temporal consistencies in diel patterns were analyzed for three planktivorous species and six intermediate carnivore/predator species. Planktivore-microinvertebrate Carnivores (PI).—Ambassis vachelli catch rates averaged 7.7 times greater at night, and they were consistently nocturnal among the estuaries (Table 5). However, although night catch rates were only slightly greater than day in most months, in May 1998 and July 1999, catch rates more than doubled, but only at night (Fig. 4A). Herklotsichthys castelnaui catch rates averaged 2.7 times greater at night, and they were also consistently nocturnal among the estuaries. However, while during the wet season (Nov–Jan), night and day catch differed little (averaging < 50 fish sample–1), during the dry season (May–Sep) average night catch rates were much greater (up to 94 fish sample-1) than day (Fig. 4B). Leiognathus equulus catch rates averaged 5.0 times greater at night than day, a trend that was consistent in all estuaries. Nocturnal catch rates were higher in most months, especially during May 1998 (Fig. 4C). 698 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 3 way 0.07*** 0.04* Fishing 0.12** 0.57**** 0.13**** 0.08* Region 0.11* 0.11**** 0.13** Interaction of check with: Month 0.13**** 0.04* Fishing 1.51**** 3.90**** 0.88**** Region 1.41**** 3.30**** 1.01**** 0.37*** Main effects Month 0.69**** 1.32**** 0.08** 0.25**** Check 0.29**** 8.62**** 0.20**** 0.03 0.09 0.53 0.04 0.05 Check 5 0.04 0.16 0.74 0.08 0.08 Check 4 0.07 0.14 0.50 0.08 0.13 Check 3 0.04 0.06 0.12 0.01 0.10 Check 2 Average catch per hour net check Average 0.04 0.13 0.16 0.03 0.06 Check 1 tetradactylum cyprinoides commersonianus Species Caranx ignobilis Table 6. Table Summary of results for analysis of based variance on (ANOVA) 2,653 samples (catch per hour for each net set of the 102 and The natural log-transformed (ln + 1) data were analysed for five key species. check). 151 mm mesh nets by Eleutheronema Eleutheronema Time periods for each check (hrs): 1 = 1430–1600; 2 1601–1730; 3 1731–1900; 4 1901–2030; 5 2031–2300 Time P < 0.0001 P < 0.001; **** P < 0.01; *** P < 0.05; ** Significance ranges: * Lates calcarifer Megalops Scomberoides Scomberoides LEY AND HALLIDAY: DIEL VARIATION IN MANGROVE FISH ASSEMBLAGES 699

Intermediate Carnivores (I) and Predators(P).—Thryssa hamiltoni catch averaged over 4.0 times greater at night (Table 5). This trend was consistent among the estuaries, with one exception, the south open estuary, in which average day catch rates were greater. Two estuaries, south closed and mid open, had particularly high catch rates, especially at night. Night catch exceeded day for all months except Nov 1998 and Mar 2000 (Fig. 4D). Caranx ignobilis catch rates averaged 1.6 times greater at night than during the day (Fig. 4E). No significant sources of variation were detected for diel period or any other spatial or temporal factors tested. For the second ANOVA designed to detect movements at dusk based on data for the 102 and 151 mm mesh nets, no consistent patterns were observed in catch rates based on time of check (Table 6). Eleutheronema tetradactylum catch rates averaged 3.2 times greater at night, a trend which was consistent among the three estuaries in which this species was abundant. In Sept–Mar, catch rates were lower overall, tending to be equal night and day (Fig. 4F). The high average catch rates observed at night in May and July 1998, and July 1999, were due to numerous E. tetradactylum netted in the south closed estuary (i.e., up to 54 fish hr–1). For the crepuscular analysis, peak average catch rates occurred during checks 3 and 4. Lates calcarifer catch rates averaged 6.0 times greater at night, consistently in all estuaries. Catch rates were highest in the closed mid estuary, especially at night. Overall catch rates were consistently higher at night in all months (Fig. 4G), increasing at check 3 with a peak by check 4. The relatively high value of the mean squares term (i.e., check) indicated that this trend in catch rate was very strong and remarkably consistent among the estuaries. Catch rates for M. cyprinoides averaged 5.0 times greater at night, a trend that was consistent among the estuaries. However, this species was only netted abundantly in the north region. Night catch rates were higher in the dry season (Fig. 4H), increasing at check 3 with a peak by check 4. Catch rates for S. commersonianus averaged only 1.1 times greater at night and diel period was not a significant factor in the model. Day and night catch rates did not vary consistently among the months (Fig. 4I), with no consistent patterns among the check times.

Discussion

Day/Night Assemblages.—Due to similarities in species composition, night samples comprised a distinct subset of day–night gill net observations (n = 155). Similarly, day and night fish assemblages have differed significantly in a wide variety of aquatic habitats (e.g., coral reefs, Hobson, 1972, 1974; temperate estuarine seagrass beds, Gray et al., 1998; temperate surf zone, Layman, 2000; tropical freshwater river sandbanks, Arrington and Winemiller, 2003). As found by Young et al., (1997) for temperate estuarine assemblages, the index of multivariate dispersion for night samples was about one-half the value for day samples, suggesting that nocturnal assemblages comprise a spatially and temporally consistent group of species within a given habitat type and region relative to more variable day assemblages. Of the 26 species which cumulatively comprised 90% of the fishes netted in the six estuaries, 53.8% were either primarily diurnal or continually active (i.e., equally active day and night). This percentage falls within the 50%–60% expected to be diurnal 700 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 Night only Sillaginidae Mugillidae Tetraodontidae Sparidae Gerreidae Top families Top (night) Mugillidae Teraponidae Cichlidae Gobiidae Scatophagidae Channel Arripidae Mugillidae Carangidae Galaxidae Anguillidae Forest Mugilidae Arripidae Galaxidae Anguillidae Sparidae Top families Top (day) Mugillidae Leiognathidae Cichlidae Eleotridae Gobiidae Channel Arripidae Mugillidae Tetraodontidae Galaxidae Sillaginidae Forest Mugillidae Galaxidae Tetraodontidae Sillaginidae Arripidae na Multi-variate groups ANOSIM < 1% Diel P ANOSIM < 1% Diel, P 42 No. sp. (night) 21 Channel: 11 Forest: 14 No. sp. (day) 18 Channel: 13 Forest: 8 Night only 3,320 Relative number (night) 623 Channel: 203 Forest: 190 Relative number (day) 531 Pooled both gears Channel: 246 Forest: 54 (1)

Method (mesh) Fyke nets Creek mouth Fyke nets and Gill nets set within forest 25–76 mm mesh (both gears set in both habitats) Enclosure net Mangrove species Kandelia candel Avicennia Avicennia marina (100%) Rhizophora stylosa Site Subtropical Taiwan Temperate Temperate Australia Subtropical Queensland Table 7. Published analyses of diel variation in mangroves and related habitats. In several cases, explicit lists of species or families by relative abundance were not provided in the published material published the in provided not were abundance relative by families or species of lists explicit cases, several In habitats. related and mangroves in variation diel of analyses Published 7. Table so that the five families listed in last 2 columns were deri ved by interpretation of secondary analyses. Author Lin and Shao, 1999 Smith and Hindell, 2005 Halliday and 1996 Young, LEY AND HALLIDAY: DIEL VARIATION IN MANGROVE FISH ASSEMBLAGES 701 Top families Top (night) Night only Ambassidae Atherinidae Tetraodontidae Gobiidae Mugillidae Haemulidae Lutjanidae Gerreidae Belonidae Sphyraenidae Mangrove Haemulidae Lutjanidae Diodontidae Chaetodontidae Gerreidae Seagrass and algal beds Haemulidae Diodontidae Lutjanidae Sparidae Acanthuridae Top families Top (day) Haemulidae Lutjanidae Sparidae Pomacentridae Gerreidae Mangrove Haemulidae Lutjanidae Scaridae Chaetodontidae Pomacentridae Seagrass and algal beds Haemulidae Scaridae Lutjanidae Gerreidae Pomacentridae Multi-variate groups na na Cluster analysis Diel difference in each habitat No. sp. (night) 37 12 Seagrass: 4 Mangrove: 3 Algae: 3.5 Channel: 3 Limestone: 2 Boulder: 2 No. sp. (day) 25 Seagrass: 9 Mangrove: 9 Algae: 2 Channel: 13 Limestone: 4 Boulder: 7 Relative number (night) Night only 10,209 2,200 hrs 151 Avg Density Avg Seagrass: 3.8 Mangrove: 13.1 Algae: 2.3 Channel: 4.6 Limestone: 12.5 Boulder: 145.5 Relative number (day) 1,200 hrs 2,838 Top 90% Top Density Avg Seagrass: 68.7 Mangrove: 242 Algae: 2.4 Channel: 132.4 Limestone: 86.8 Boulder: 1556.4 Method (mesh) Enclosure net (2) Visual census with Visual dive light at night Visual census with Visual dive light by night Mangrove species Avicennia Avicennia marina Rhizophora apiculatata Rhizophora mangle Rhizophora mangle Site Tropical Philippines Tropical Tropical Caribbean Puerto Rico Tropical Tropical Caribbean Curacao Table 7. Continued. Table Author Ronnback et al., 1999 Rooker and Dennis,1991 Nagelkerken et al., 2000 702 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 ( 6 ) (66.3) Clupeidae Engraulidae (21.9) Mugillidae (11.9) Ariidae (12.9) Centropomidae (12.5) Leiognathidae (12.3) Polynemidae (6.9) Ambassidae (5.7) Carangidae (3.9) Haemulidae (3.8) Top families Top (night) Carcharhinidae (3) Engraulidae Polynemidae Megalopidae Chirocentridae Polynemidae Mugillidae Chirocentridae Lethrinidae Platycephalidae ( 5 ) (26.6) Clupeidae Engraulidae (5.4) Mugillidae (15.2) Ariidae (6.8) Centropomidae (2.5) Leiognathidae (2.7) Polynemidae (2.4) Ambassidae (0.9) Carangidae (2.9) Haemulidae (2.3) Top families Top (day) Sparidae (4) Bothidae Gerreidae Sillaginidae Carangidae ANOSIM < 0.1% Diel, P Multi-variate groups na Correspondence analysis: gradient of species along day-night axis 135 61–83 in any one estuary No. sp. (night) na na 110 48–63 in any one estuary No. sp. (day) na na –1 Abundance 17,292 Avg 171.4 fish hr Relative number (night) na na –1 Abundance 7,616 Relative number (day) na na Avg 77.2 fish hr

–1 for the top 10 families. –1 Method (mesh) Gill nets set perpendicular to mangrove fringe 50–150 mm mesh Gill nets set perpendicular to mangrove fringe 50–80 mm mesh Gill nets set perpendicular to the mangrove fringe 19–151 mm mesh sample –1 sample –1 Mangrove species na Rhizophora mucronata Avicennia marina Rhizophora stylosa Avicennia marina Site Tropical Tropical Australia Tropical Tropical Madagascar Tropical Tropical Queensland Table 7. Continued. Table Author na = not available avg = average (1) Flume type net set along paths cleared of branches running perpendicular to mangrove shoreline back dry land; at hig h tide and retrieved low tide. (2) Square section net set along paths within mangrove forest at high tide and retrieved low tide. relative abundances day vs night. (3) Families with significantly different relative abundances day vs night. axis and having significantly different (4) Species as ordinated along day-night CCA (5) In parentheses, average day catch hr (6) In parentheses, average night catch hr Blaber et al., 1995 Laroche et al., 1997 Ley, present Ley, study LEY AND HALLIDAY: DIEL VARIATION IN MANGROVE FISH ASSEMBLAGES 703 or continually active in a typical fish assemblage H( elfman et al., 1997). However, considering the entire assemblage, only 44% of all species were diurnal or continually active. Furthermore, the catch rate for all species combined was 2.2 times greater at night. In fact, if sampling had only occurred during the day, 30% of the species would not have been recorded at all due to their presence only in night samples; in contrast, by netting only at night, 14% of the species would not have been sampled. Thus, based on catch rates and total species composition, overall assemblages tended to be more strongly nocturnal than expected. Mangrove estuaries were not included in the comprehensive review of diel variation in fish assemblages presented in Helfman et al. (1997), but extending their analysis to include recent relevant studies has proven difficult due to the variety of sampling and reporting methods, as summarized in Table 7 and discussed below. To date, eight previous studies have investigated nocturnal assemblages in mangrove habitats (Table 7). Including the current study, most have been conducted within tropical latitudes (n = 6) and in the Indo-Pacific region (n = 7). For all five studies in which multivariate analyses were conducted, night assemblages differed significantly from day. However, in comparing the current results with other studies, careful interpretation of the sampling regime employed by the investigators is required. Sampling methods have included: (1) fyke nets set in creeks and forests—two studies; (2) gill nets set perpendicular to the edge—three studies; (3) gill nets set within the forest—1 study; (4) enclosure nets—two studies; and (5) visual census—two studies. For both enclosure net studies, the data represent night assemblages only (Halliday and Young, 1996; Ronnback et al., 1999), while fyke nets were deployed day and night (Smith and Hindell, 2005; Lin and Shao, 1999). These enclosure and fyke net sampling procedures captured fishes moving off flooded intertidal mangrove forests as the tide drained away, precluding separation of the influence of diel variation and tidal effects. Only two previous studies have compared day–night fish assemblages by sampling with set gill nets along subtidal mangrove fringe habitats (as in the current study). Unfortunately, neither reported the overall number of species or relative abundances netted, and only one reported data for taxonomic composition day and night (Laroche et al., 1997). Relative to the current investigation, the most informative and relevant studies were conducted by Rooker and Dennis (1991) and Nagelkerken et al. (2000) using visual census methods to survey fishes along mangrove fringe habitats in clear-water settings near coral reefs in the Caribbean. Both investigators reported dramatically reduced assemblages (in terms of fish abundance and numbers of species) in mangrove fringe habitats at night. Furthermore, they also directly observed movement of fishes away from mangrove fringe habitats at dusk. In Curacao, destinations of fishes moving from the mangrove fringe were foraging habitats in dispersed seagrass and algal beds (Nagelkerken et al., 2000). These studies support the interpretation that the higher night catch rates in the current study were due to increased activity along the mangrove fringe, probably as mobile fishes moved from mangroves to adjacent habitats to forage. Thus, the distinct group of fishes netted at night appears to include more mobile components of fish assemblages that may be adapted for foraging in low light conditions. As seagrass and macroalgal habitats were sparse or absent in the estuaries under study, night destinations of fishes moving from mangrove fringes in the current study must have been other types of feeding habitats, as discussed further below. 704 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Regional Patterns in Diel Trends.—Regional distinctions existed in composition of day and night assemblages, possibly reflecting differences in habitat conditions among the estuaries at broad landscape scales. In a previous analysis of 11 geographically dispersed (1100 linear km) estuaries (including the six analyzed here), the two estuaries in the north region (north closed, north open) were identified as wave-dominated (WD) systems, situated in high rainfall, mountainous catchments, dominated by lush rain forests; the four mid and south estuaries (mid closed, mid open, south closed, south open) were tide-dominated (TD) systems, situated in lower rainfall catchments draining broad, sparsely vegetated floodplains (Ley, 2005). These large-scale features coincided with differing estuarine habitat features, as WD estuaries had less turbid, lower salinity conditions, with narrow entrances, relatively smaller areas of mud banks and mangroves; and TD systems had very turbid, higher salinity conditions, with wider entrances, extensive mud banks and networks of mangrove-lined channels. In the current study, night assemblages of fishes in WD (north) systems were distinguished from TD (mid+south) systems by greater catch rates of H. castelnaui, L. equulus, and M. cyprinoides; in contrast, night TD systems had greater catch rates of T. hamiltoni and E. tetradactylum. Day assemblages in WD systems were distinguished from TD systems by greater catch rates of H. castelnaui and L. equulus; day TD assemblages had greater catch rates of T. hamiltoni, E. tetradactylum, and L. calcarifer than WD systems. Within a given region, however, day catch rates were consistently lower than night for all six key species. Clearly, adaptations by individual species led to regional differences in fish assemblages based on both habitat conditions and diel activity. Diel variation: Key Species.—As diel activity is primarily related to food gathering by fishes H( elfman et al., 1997), analysis of dietary and related physiological, morphological, and behavioral adaptations may explain the patterns observed in how the nine key species used the 24-hr daily cycle. For example, the three key species of planktivores (A. vachelli, H. castelnaui, L. equulus) have fine gill rakers, protrusible jaws, and mouths capable of a large gape, morphological traits which equip them for feeding on invertebrates in the water column at night (Martin and Blaber, 1984; Blaber, 1997 and 2000). Each of these species may also be protected from nocturnal predators by camouflage coloration, having lighter to silvery sides and bellies, with dark green, bronze or grey upper bodies, helping these fishes to “disappear in the water column,” thereby reducing detection from above and below (Helfman et al., 1997). Individual species have different food habits, leading to diel foraging patterns, as discussed below. The planktivore H. castelnaui was moderately nocturnal (N1) and by far the most abundant of all species netted, comprising 30% of the total catch. Salini et al. (1990) reported diets comprised mainly of copepods, penaeid shrimp, amphipods and non- chitinous invertebrates. Clupeids are known to primarily forage at night and sound production may facilitate foraging-related communication within nocturnal shoals (Wilson et al., 2004). Although no specifically nocturnal dietary data is available for H. castelnaui, Herklotsichthys quadrimaculatus diets included fish and pelagic amphipods obtained during nocturnal foraging migrations from coral reefs to open waters of a nearby adjacent lagoon (Milton et al., 1994). If H. castelnaui be- haves similarly, during seasons when resources are limited, they may migrate from daytime mangrove habitats in order to forage in adjacent open waters at night. This foraging-related behavior may explain the seasonal trends in diel activity noted for LEY AND HALLIDAY: DIEL VARIATION IN MANGROVE FISH ASSEMBLAGES 705

this species. In daytime sampling, previous investigators have shown that sesarmid crab larvae become very dense in flooded mangrove forests in the wet season, corresponding with peak densities of planktivorous fishes which feed opportunistically on these seasonally abundant zooplankton (Robertson et al., 1988; Robertson and Duke, 1990). In the dry season, crab larvae are absent and densities of zooplankton are lower everywhere in the mangrove habitat (Robertson et al., 1988; McKinnon and Klumpp, 1998). Although no night zooplankton data are available for estuaries in the region, it is known that juvenile prawns (primarily Penaeus merguiensis) recruit to deeper mangrove creeks during the dry season (Robertson, 1988), providing an alternative food resource for H. castelnaui. This resource would primarily be available at night, as larger juvenile P. merguiensis remain buried in the sediments by day, but emerge at dusk to forage throughout the night (Wassenberg and Hill, 1994). Amphipods and other invertebrates may also emerge from the sediments in deeper mangrove creeks at dusk and dark (see below). In our gill nets, H. castelnaui catch rates were lower both day and night in the wet season; during the dry season, catch rates were very high, but only at night. Thus, during the wet season when crab larvae are abundant, H. castelnaui apparently obtain most of their food requirements in the broad, flooded mangrove forest habitat. During the dry season, when plankton densities are lower, H. castelnaui may exploit nocturnally available resources such as juvenile prawns and other emergent invertebrates in the open waters of deeper creeks adjacent to the forest. To access these resources, H. castelnaui must migrate from the mangrove fringe during dusk and at night in the dry season months, leading to the markedly increased catch rates in our gill nets. The trophic implications of this movement are discussed in the next section. Another abundant planktivore, A. vachelli, was strongly nocturnal (N2) overall, but night catch rates peaked sharply in two dry season months (May 1998, July 1999). Ambassids are extremely abundant in local mangrove forests by day (Robertson and Duke, 1990). Although food habits of A. vachelli have not been studied, other estuarine ambassids consume copepods and a variety of crustaceans (Coates, 1987, 1990; Hayward et al., 1998; Hajisamae et al., 2004; Baker and Sheaves, 2005), with feeding peaks in the early evening and early morning (Martin and Blaber, 1983). Thus, as discussed earlier for H. castelnaui, A. vachelli may have been netted in greater numbers along the mangrove fringe at night during certain dry season months because they moved to adjacent open waters to forage when densities of crab larvae and other zooplankton within the mangrove forest declined. Leiognathus equulus was strongly nocturnal (r = 0.83) but unlike the other two major planktivores (H. castelnaui, A. vachelli), nocturnal catch rates did not vary systematically among the months. Diatoms, copepods, and plant material are dominant items consumed by this species (Tiews et al., 1972). Like all leiognathids, nocturnal adaptations include a highly producible mouth and bioluminescent organs in the throat (Myers, 1999), possibly advantageous in attracting plankton while feeding at night in the water column. Intermediate carnivores and predators also displayed a variety of traits related to foraging during lower light conditions. The small intermediate carnivore, T. hamiltoni, forms schools by day and was strongly nocturnal (r = 0.82) with no systematic seasonal trends detected. Diets consist of crustaceans, pelagic fishes, and penaeids (Brewer et al., 1995; Salini et al., 1998; Baker and Sheaves, 2005). Thryssa hamiltonii were often netted at night by encountering the net with an open mouth and becom- 706 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

ing entangled by their protruded jaw parts; pelagic fish and prawns, items frequently occurring in the diets of T. hamiltoni, could be captured at night through this foraging behavior. The three key species of non-schooling intermediate carnivore/predators, E. tetradactylum, L. calcarifer, and M. cyprinoides, all very nocturnal, have seldom been listed among the dominant species in previous sampling surveys in the region, probably because these surveys were conducted by day (e.g., Robertson and Duke, 1990; Sheaves, 2005). Lates calcarifer, the third most abundant species netted (6.6% of the total catch), was strongly nocturnal (r = 0.83) displaying a remarkably consistent increase in activity at sunset. Catch rates rose through the second nighttime check, and began to decrease during the final check (after 2030). Surprisingly little information is available in the scientific literature about the nocturnal behavior and adaptations of this major commercial species. Juvenile Lates spp. in Lake Tanganyika were most active at night, feeding on nocturnally emergent shrimps (Kondo and Abe, 1995). In other Centropomidae species (Psammoperca waigiensis, Centropomus undecimalis) a ta- petum lucidum, enhances the capture of light (Linke et al., 2001); L. calcarifer may have similar eye morphology as evidenced by red eye shine at night. Thus,L. calcarifer has an advantage in crepuscular foraging on large penaeids which emerge in the evening; in fact, juveniles are major predators of Penaeus merguiensis in deeper mangrove creeks (Robertson, 1988). In laboratory studies, L. calcarifer were observed locating buried penaeids by manipulating the sand with their snout and pelvic fins, indicating that other senses may be involved in locating prey at low light levels (Pri- mavera, 1997). Acetes spp. and fishes are also major components of the L. calcarifer diet (Davis, 1985; Salini et al., 1990, 1998; Brewer et al., 1995; Baker and Sheaves, 2005; Ley, 2005). Bronze coloration may make them nearly invisible to prey fishes in estuaries at dusk, and with a wedge-shaped tail and streamlined morphology, L. calcarifer use bursts of speed to capture prey fishes in the water column. Thus, when planktivores (e.g., H. castelnaui) move from the mangrove edge as darkness falls, they are vulnerable to predation by L. calcarifer. Moderately nocturnal (r = 0.77), few E. tetradactylum were caught in the two WD (north) estuaries, while catch rates were higher in the four TD estuaries (mid and south) where turbidity tends to be greater. The diet of E. tetradactylum consists mainly of fish, penaeids, and other crustaceans (Salini et al., 1990, 1998; Baker and Sheaves, 2005). In turbid and low light conditions, the four free pectoral rays of E. tetradactylum, may be used to locate foods such as burrowing stomatopods, mollusks, and prawns (Kailola et al., 1993). At night in the dry season, catch rates peaked sharply, possibly due to an upward trophic cascade initiated by juvenile prawn recruitment in deeper mangrove creeks (as previously discussed for H. castelnaui). At dusk, E. tetradactylum are also known to rise from the bottom and prey on small fishes in the water column G( rant, 1995), accounting for increased activity during the sunset net check. Strongly nocturnal (r = 0.83), M. cyprinoides catch rates were greater in the WD estuaries, possibly due to lower average salinities and other habitat conditions (Ley, 2005). Although M. cyprinoides is a marine transient, many sub-adults apparently live in freshwater, surviving low oxygen episodes by breathing air (Coates, 1987; Pusey et al., 2004). Diets consist of pelagic fishes, shrimps, and insects (Coates, 1987, 1990; Rao and Padam, 1999; Pusey et al., 2004; Ley, 2005). For all months, night catch LEY AND HALLIDAY: DIEL VARIATION IN MANGROVE FISH ASSEMBLAGES 707

Figure 5. Conceptual diagram illustrating daytime activity along the shoreline of a well-developed mangrove habitat in a turbid Indo-Pacific estuary. Fish families characteristically (A) resting and (B) active are indicated. Invertebrates (C) are shown as buried in the sediments of permanently submersed channels.

Figure 6. Conceptual diagram illustrating the night time counterpart to the diagram in Figure 7. The upward arrow indicates (A) emergence of invertebrates from the sediments into the channel water column. The downward arrow indicates (B) movement of planktivorous and intermediate carnivorous fishes from the mangrove prop root habitat. (C) Predators active at night in the channels adjacent to the mangrove fringe forage on emergent invertebrates and small active fishes, especially at dusk. 708 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

rates were greater than day except during the wet season when numbers were lower overall, possibly due to availability of freshwater flooded habitats. Thus, in the wet season, M. cyprinoides may have spent less time foraging along the mangrove fringe where our research gill nets were deployed; during the dry season, catch rates were consistently high in the gill nets at night, indicating that M. cyprinoides were actively feeding along and adjacent to the mangrove fringe, probably on fishes and prawns. On average, activity began to increase during the sunset net check, rose through the second night check, and decreased during the final check (after 2030). Morphologi- cally adapted for high speed, and having a wide producible jaw, with an eye structured to capture light at low light levels, M. cyprinoides are equipped for crepuscular foraging on water column fishes and emergent invertebrates. The two abundant carangid species, C. ignobilis (r = 0.61) and S. commersonianus (r = 0.53), were active both night and day. Caranx ignobilis catch rates were approximately equal during both day and night in all estuaries and months, and no systematic trends during crepuscular periods were detected. The relatively well-studied C. ignobilis has a broad diet dominated by penaeids, pelagic fishes, crabs, amphipods, and mollusks (Blaber and Cyrus, 1983; Sudekum et al., 1991; Brewer et al., 1995; Blaber, 1997, 2000; Smith and Parrish, 2002; Baker and Sheaves, 2005; Ley, 2005). Like other carangids, pack-foraging tactics and sound production through stridula- tion may enhance effectiveness of night (and day) predation on small fishes (Tavolga, 1971; Sudekum et al., 1991; Helfman et al., 1997). However, C. ignobilis tracked with acoustic telemetry in a coral reef reserve foraged along the reef edge at night with long quiescent periods during the day (Wetherbee et al., 2004). In the two estuaries where S. commersonianus were particularly abundant (mid closed, south closed), catch rates peaked before and during sunset. Pelagic fish and penaeids are the main dietary components of S. commersonianus (Salini et al., 1990, 1998; Brewer et al., 1995; Baker and Sheaves, 2005; Ley 2005). Implications for Trophodynamics.—The foregoing analyses of diets and foraging adaptations suggest that during and just after dusk, gill nets set along the mangrove fringe captured fishes as they moved from fringing mangroves to other habitats in the estuary to forage. Two conceptual diagrams illustrate a hypothetical model of resultant trophodynamics (Figs. 5,6). This model generally parallels diel variation patterns described for clear-water Caribbean mangrove shorelines near coral and seagrass habitats (Rooker and Dennis, 1991; Nagelkerken et al., 2000). Thus, we propose that similar processes exist along mangrove shorelines in turbid, Indo-Pacific estuaries, even if they are virtually lacking in seagrass beds and coral reefs. Exist- ing evidence (and gaps) relating to this hypothesis are summarized in the following paragraphs. During the day, relatively few small planktivores (Pl) and intermediate carnivores (I) in the families Clupeidae, Engraulidae, Leiognathidae, and Ambassidae were caught, indicating they were less active along the mangrove fringe (Fig. 5A). Also by day, predators (P) such as Carangidae were active along the mangrove fringe (Fig. 5B) and if small pelagic fishes strayed from the mangroves, they would be vulnerable to predation. We propose that larger invertebrates, food items for Pl and I, remain buried in the sediments of creeks adjacent to the mangrove fringe during the day (Fig. 5C). Under this model, invertebrate fauna would begin to emerge from creek sediments at dusk (Fig. 6A). Densities of emergent penaeids, larger copepods, amphipods, Ace- LEY AND HALLIDAY: DIEL VARIATION IN MANGROVE FISH ASSEMBLAGES 709

Figure 7. Diagram illustrating the trophic relationships arising from diel shifts from mangrove shoreline habitats to adjacent creeks as night falls. Arrow 1 indicates foraging on plankton by emergent invertebrates. Arrow 2 indicates foraging by planktivorous and intermediate carnivorous fishes following a rapid shift from mangrove prop roots to open waters at sunset. Arrow 3 indicates active foraging by fishes from commercially important families primarily during crepuscular periods. tes spp., and other invertebrates have been shown to increase throughout the water column at night in estuarine waters (Williams and Bynum, 1972; Day et al., 1989; Clark et al., 2003). Following crepuscular emergence of invertebrates, Pl and small I apparently moved out of the mangrove fringe, leading to increased catch rates in the smaller mesh gill nets (Fig. 6B). Feeding requirements may motivate these fishes to leave the shelter of the forest to consume the emergent invertebrates in adjacent waters. Although comprehensive studies of day–night food habits have not been published for tropical estuarine fishes, stomach contents samples collected by day include remnants of a variety of invertebrates including penaeids, copepods, and other crustaceans. Furthermore, emergent invertebrates may also be important items consumed by larger I and P species (e.g., Carangidae, Polynemidae, Centropomidae, Megalopidae), for which small pelagic fishes are equally important food items (Fig. 6C). Crepuscular movements detected during the current study showed that these I and P species began activity at dusk, peaking just after dark with reduced movement later in the night. 710 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Reported stable isotope analyses are consistent with trophic relationships identified in the model discussed above. Mangrove leaves are the primary production base for a food web involving large juvenile Serranidae and Lutjanidae which remain near subtidal snags in creeks at low tide, but forage in the mangrove forest on leaf-eating crabs at high tide (Sheaves and Maloney, 2000). These fish families were not well sampled in the current study apparently because they do not undertake regular crepuscular movements along the mangrove fringe. However, Sheaves and Maloney (2000) also found that carangids (S. commersonianus, C. ignobilis) had isotopic signatures indicative of a food web based on non-mangrove sources of primary production, as yet unidentified. Stable isotopic analysis of nocturnal species including L. calcarifer, E. tetradactylum, M. cyprinoides, and T. hamiltoni have not been reported. Given the nighttime results of the current study and other daytime studies, diel variation patterns may allow at least two food webs to operate at different times during the 24- hr cycle, overlapping within the same mangrove habitats. One web links mangrove leaves to Serranidae and Lutjanidae, includes mangrove forest crabs as the intermediate prey, and primarily functions by day on tidally driven cycles (Sheaves, 2005). The other food web suggested in the current study, links plankton and nocturnally emergent benthic invertebrates which become available in the water column adjacent to the mangrove fringe at dusk, with species of Centropomidae, Polynemidae, Megalopidae, and many other taxa; cues involving this food web primarily activate at dusk and continue at least through the early part of the night (Fig. 7). Like the well-studied movements of coral reef planktivores and intermediate carnivores to seagrass or pelagic foraging habitats, the food web suggested by the current findings involves movement from highly structured shelter (mangrove fringe habitats) to less structured areas for foraging (open water) at dusk. Subjects for future research include comprehensive surveys quantifying diel variation in food habits of fishes with concurrent sampling of invertebrate water column assemblages in estuarine mangrove creeks and basins. Limitations of the Study.—Variation in fish assemblages was not influenced by tidal conditions which prevailed concurrent with sampling. Gill nets are less effective at capturing fish if they are taut (von Brandt, 1984), as would occur in high-flow spring tides. Thus, the design of the sampling program was limited to 3 d before and 3 d after the weakest tide in the lunar cycle. Tidal variation may have been a more significant factor if the study design had included spring tides (Krumme et al., 2004; Sheaves, 2005). Sampling with gill nets has selectivity biases which differ from visual census and other methods. As a passive gear, gill nets only sample that portion of the fish assemblages that move rapidly enough and have the morphological features to become gilled or entangled in the nets (Sparre and Venema, 1992). Thus, differing sampling practices and efficiencies may explain some of the discrepancies between the current study findings and other diel mangrove studies. To determine whether gill netting, or any other method, gives a true indication of diel variation, tests comparing a range of sampling methods by night and day are needed. A further limitation is the reliance on some literature sources for information on diet composition. Although the diets of some of the species have been well studied by day, the diel period in which foods were obtained remains unknown. LEY AND HALLIDAY: DIEL VARIATION IN MANGROVE FISH ASSEMBLAGES 711

Conclusions

Fish assemblages occupying these tropical mangrove estuaries were characterized by distinct diel variation in composition that appeared to be driven by trophic relationships. Similar to assemblages observed in other aquatic ecosystems, at least half of the species of fishes netted in these tropical mangrove estuaries were primarily active at night. These results suggest that movement patterns observed relate to foraging behavior motivated by changes in the availability of forage base organisms at sunset, and to predator avoidance strategies. Emergent benthic fauna may provide a forage base for small planktivores with upwardly cascading effects on nocturnal intermediate carnivores and predators (Centropomidae, Polynemidae, Megalopidae, and scientists with nets). By day, predators apparently occupied deeper subtidal habitats, while planktivores sheltered in the mangrove fringe. Food webs based on crepuscular habitat shifts by mobile fishes may spatially overlap other mangrove-based food webs identified by previous investigators thereby enhancing the diversity and fisheries productivity in tropical estuarine mangrove systems.

Literature Cited

Annesea, D. M. and M. J. Kingsford. 2005. Distribution, movements and diet of nocturnal fishes on temperate reefs. Environ. Biol. Fish. 72: 161–174. Arrington, D. A. and K. O. Winemiller. 2003. Diel changeover in sandbank fish assemblages in a neotropical floodplain river. J. Fish Biol. 63: 442–459. Australian Estuarine Database (AED). 2000 Available from: http://www.ozestuaries.org/. Ac- cessed March, 2005. Baker, R. and M. Sheaves. 2005. Redefining the piscivore assemblage of shallow estuarine nursery habitats. Mar. Ecol. Prog. Ser. 291: 197–213. Benoit-Bird, K. J. and W. W. L. Au. 2006. Extreme diel horizontal migrations by a tropical nearshore resident micronekton community. Mar. Ecol. Prog. Ser. 319: 1–14. Blaber, S. J. M. 1997. Fish and fisheries of tropical estuaries. Chapman & Hall, London, 367 p. ______. 2000. Tropical estuarine fishes: ecology, exploitation and conservation. Black- well Sciences Ltd., London, 372 p. ______and D. P. Cyrus. 1983. The biology of Carangidae (Teleostei) in Natal estuaries. J. Fish Biol. 22: 173–188. ______, D. T. Brewer, and J. P. Salini. 1995. Fish communities and the nursery role of the shallow inshore waters of a tropical bay in the Gulf of Carpentaria, Australia. Estuar. Coast. Shelf Sci. 40: 177–193. Brewer, D. T., S. J. M. Blaber, J. P. Salini, and M. J. Farmer. 1995. Feeding ecology of predatory fishes from Groote Eylandt in the Gulf of Carpentaria, Australia, with special reference to predation on penaeid prawns. Estuar. Coast. Shelf Sci. 40: 577–600. Clarke, K. R. 1993. Non-parametric multivariate analyses of changes in community structure. Aust. J. Ecol. 18: 117–143. ______and R. M. Warwick. 2001. Change in Marine Communities: An Approach to Sta- tistical Analysis and Interpretation (2nd edition). Primer-E Ltd, Plymouth Marine Labora- tory, UK. ______, G. M. Ruiz, and A. H. Hines. 2003. Diel variation in predator abundance, predation risk and prey distribution in shallow-water estuarine habitats. J. Exp. Mar. Biol. Ecol. 287: 37–55. Coates, D. 1987. Observations on the biology of Tarpon, Megalops cyprinoides (Broussonet) (Pisces: Megalopidae) in the Sepik River, Northern Papua New Guinea. Aust. J. Mar. Freshw. Res. 38: 529–535. 712 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

______. 1990. Aspects of the Biology of the Perchlet Ambassis interrupta Bleeker (Pisces : Am- bassidae) in the Sepik River, Papua New Guinea. Aust. J. Mar. Freshw. Res. 41: 267–274. Davis, T. L. O. 1985. The food of barramundi,Lates calcarifer (Bloch), in coastal and inland waters of Van Dieman Gulf and the Gulf of Carpentaria, Australia. J. Fish Biol. 26: 669–682. Day, J. W., C. A. S. Hall, W. M. Kemp, and A. Yanez-Arancibia. 1989. Estuarine Ecology John Wiley and Sons, New York. 558 p. Froese, R. and D. Pauly, eds. 2005. FishBase. World Wide Web electronic publication. Available from: www.fishbase.org. Grant, E. M. 1995. Fishes of Australia. E. M. Grant Pty. Limited, Brisbane. 457 p. Gray, C. A., R. C. Chick, and D. J. McElligott. 1998. Diel changes in assemblages of fishes associated with shallow seagrass and bare sand. Estuar. Coast. Shelf Sci. 46: 849–859. Griffiths, S. P. 2001. Diel variation in the seagrass ichthyofaunas of three intermittently open estuaries in south-eastern Australia: implications for improving fish diversity assessments. Fisheries Manage. Ecol. 8: 123–140. Hajisamae, S. L., M. Chou, and S. Ibrahim. 2004. Feeding habits and trophic relationships of fishes utilizing an impacted coastal habitat, Singapore. Hydrobiologia 520: 61–71. Halliday I. A. and W. R. Young. 1996. Density, biomass and species composition of fish in a subtropical Rhizophora stylosa mangrove forest. Mar. Freshw. Res. 47: 609–615. ______, J. A. Ley, A. J. Tobin, R. N. Garrett, N. A. Gribble, and D. A. Mayer. 2001. The effects of net fishing: addressing biodiversity and bycatch issues in Queensland inshore waters. Fisheries Research and Development Corporation 97/206 (Canberra, ACT), Aus- tralian Institute of Marine Science and Queensland Department of Primary Industries, De- ception Bay, Queensland, Australia. Hayward, M. D. E., D. S. Heales, R. A. Kenyon, N. R. Loneragan, and D. J. Vance. 1998. Pre- dation of juvenile tiger prawns in a tropical Australian estuary. Mar. Ecol. Prog. Ser. 162: 201–214. Helfman, G. S. 1993. Fish behaviour by day, night and twilight. In Pitcher, T. J., ed. Behaviour of teleost fishes. nd2 ed. Chapman & Hall, London. 715 p. ______, B. B. Collette, and D. E. Facey. 1997. The diversity of fishes. Blackwell Science, Inc. Malden. 529 p. Hobson, E. S. 1972. Activity of Hawaiian reef fishes during the evening and morning transition between daylight and darkness. Fish. Bull. 70: 710–740. ______. 1974. Feeding relationships of teleostean fishes on coral reefs in Kona, Hawaii. Fish. Bull. 72: 915–1031. ______. 1991. Trophic relationships of fishes specialized to feed on zooplankters above coral reefs. In P. F. Sale, ed. The ecology of fishes on coral reefs. Academic Press Inc., San Diego. 754 p. ______and J. R. Chess. 1977. Trophic relationships among fishes and plankton in the lagoon at Enewetak Atoll, Marshall Islands. Fish. Bull. 76: 133–153. Kailola, P. J., M. J. Williams, P. C. Stewart, R. E. Reichelt, A. McNee, and C. Grieve. 1993. Aus- tralian Fisheries Resources. Fisheries Research and Development Corporation, Common- wealth of Australia, Canberra. 442 p. Kondo, T. and N. Abe. 1995. Habitat preference, food habits and growth of juveniles of Lates angustifrons and Lates mariae (Pisces: Centropomidae) in Lake Tanganyika. Ecol. Res. 10: 275–280. Kronfeld-Schor, N. and T. Dayan. 2003. Partitioning of time as an ecological resource. Annu. Rev. Ecol. Evol. Syst. 34: 153–181. Krumme U., U. Saint-Paul, and H. Rosenthal. 2004. Tidal and diel changes in the structure of a nekton assemblage in small intertidal mangrove creeks in northern Brazil. Aquat. Living Resour. 17: 215–229. Laroche, J., E. Baran, and B. Rasoanandrasana. 1997. Temporal patterns in a fish assemblage of a semiarid mangrove zone in Madagascar. J. Fish Biol. 51: 3–20. LEY AND HALLIDAY: DIEL VARIATION IN MANGROVE FISH ASSEMBLAGES 713

Layman, C. A. 2000. Fish assemblage structure of the shallow ocean surf-zone on the eastern shore of Virginia barrier islands. Estuar. Coast. Shelf Sci. 51: 201–213. Ley, J. A. 2005. Linking fish assemblages and attributes of mangrove estuaries in tropical Aus- tralia: criteria for regional marine reserves. Mar. Ecol. Prog. Ser. 305: 41–57. ______and I. A. Halliday. 2003. A key role for marine protected areas in sustaining a regional fishery for barramundi Lates calcarifer in mangrove-dominated estuaries? Evidence from northern Australia. Am. Fish. Soc. Symp. 42: 225–236. ______, C. C. McIvor, and C. L. Montague. 1999. Fishes in mangrove prop-root habitats of northeastern Florida Bay: distinct assemblages across and estuarine gradient. Estuar. Coast. Shelf Sci. 48: 701–723. ______, I. A. Halliday, A. J. Tobin, R. N. Garrett, and N. A. Gribble. 2002. Ecosystem effects of fishing closures in mangrove estuaries of tropical Australia. Mar. Ecol. Prog. Ser. 245: 223–238. Lin, H.-J. and K.-T. Shao. 1999. Seasonal and diel changes in a subtropical mangrove fish assemblage. Bull. Mar. Sci. 65: 775–794. Linke, T. E., M. E. Platell, and I. C. Potter. 2001. Factors influencing the partitioning of food resources among six fish species in a large embayment with juxtaposing bare sand and seagrass habitats. J. Exp. Mar. Biol. Ecol. 266: 193–217. Livingston, R. J. 2003. Trophic organization in coastal systems. CRC Press, Boca Raton. 388 p. Lowe-McConnell, R. H. 1987. Ecological studies in tropical fish communities. Cambridge Uni- versity Press, Cambridge. 382 p. Martin, T. J. and S. J. M. Blaber. 1984. Morphology and histology of the alimentary tracts of Ambassidae (Cuvier) (Teleostei) in relation to feeding. J. Morphol. 182: 295–305. McKinnon, A. D. and D. W. Klumpp. 1998. Mangrove zooplankton of north Queensland, Aus- tralia. Hydrobiologia 362: 127–143. Milton, D. A., S. J. M. Blaber, and N. J. F. Rawlinson. 1994. Diet, prey selection and their energetic relationship to reproduction in the tropical herring Herklotsichthys quadrimaculatus in Kiribati, Central Pacific. Mar. Ecol. Prog. Ser. 103: 239–250. Myers, R. F. 1999. Micronesian reef fishes: a field guide for divers and aquarists. CoralG raphics Barrigada, Guam. 216 p. Nagelkerken I., M. Dorenbosch, W. C. Verberk, E. Cocheret de la Morinière, and G. van der Velde. 2000. Day–night shifts of fishes between shallow-water biotopes of a Caribbean bay, with emphasis on the nocturnal feeding of Haemulidae and Lutjanidae. Mar. Ecol. Prog. Ser. 194: 55–64. Primavera, J. H. 1997. Fish predation on mangrove-associated penaeids: the role of structures and substrate. J. Exp. Mar. Bio. Ecol. 215: 205–216. Pusey, B., M. Kennard, and A. Arthington. 2004. Freshwater fishes of north-eastern Australia. CSIRO Publishing, Collingwood, Victoria. 694 p. Rao, L. M. and G. Padam. 1999. Variations in food and feeding habits of Megalops cyprinoides from coastal waters of Visakhapatnam. Indian J. Fish. 46: 407–410. Robblee, M. B. and J. C. Zieman. 1984. Diel variation in the fish fauna of a tropical seagrass feeding ground. Bull. Mar. Sci. 34: 335–345. Robertson, A. I. 1988. Abundance, diet and predators of juvenile banana prawns, Penaeus merguiensis, in a tropical mangrove estuary. Aust. J. Mar. Freshw. Res. 39: 467–478. ______and N. C. Duke. 1990. Recruitment, growth and residence time of fishes in a tropical Australian mangrove system. Estuar. Coast. Shelf Sci. 31: 723–743. ______, P. Dixon, and P. A. Daniel. 1988. Zooplankton dynamics in mangrove and other nearshore habitats in tropical Australia. Mar. Ecol. Prog. Ser. 43: 139–150. Ronnback P., M. Troell, N. Kautskya, and J. H. Primavera. 1999. Distribution pattern of shrimps and fish amongAvicennia and Rhizophora microhabitats in the Pagbilao mangroves, Philip- pines. Estuar. Coast. Shelf Sci. 48: 223–234. Rooker, J. R. and G. D. Dennis. 1991. Diel, lunar, and seasonal changes in a mangrove fish assemblages off southwestern Puerto Rico. Bull. Mar. Sci. 49: 684–698. 714 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Rountree, R. A. and K. W. Able. 1993. Diel variation in decapod crustacean and fish assemblages in New Jersey polyhaline marsh creeks. Estuar. Coast. Shelf Sci. 37: 181–201. Saigusa, M. 2001. Daily rhythms of emergence of small invertebrates inhabiting shallow subtidal zones: a comparative investigation at four locations in Japan. Ecol. Res. 16: 1–28. ______, T. Okochi, and S. Ikei. 2003. Nocturnal occurrence, and synchrony with tidal and lunar cycles, in the invertebrate assemblage of a subtropical estuary. Acta Oecologica 24: S191–S204. Salini, J. P., S. J. M. Blaber, and D. T. Brewer. 1990. Diets of piscivorous fishes in a tropical Australian estuary, with special reference to predation on penaeid prawns. Mar. Biol. 105: 363–374. ______, D. T. Brewer, and S. J. M. Blaber. 1998. Dietary Studies on the predatory fishes of the Norman River estuary, with particular reference to penaeid prawns. Estuar. Coast. Shelf Sci. 46: 837–847. Sheridan, P. and C. Hays. 2003. Are mangroves nursery habitat for transient fishes and decapods? Wetlands 23: 449–458. Schoener, T. W. 1974. Resource partitioning in ecological communities. Science 185: 27–39. Sheaves, M. 2005. Nature and consequences of biological connectivity in mangrove systems. Mar. Ecol. Prog. Ser. 302: 293–305. ______and B. Malony. 2000. Short-circuit in the mangrove food chain. Mar. Ecol. Prog. Ser. 199: 97–109. Smith, G. C. and J. D. Parrish. 2002. Estuaries as Nurseries for the Jacks Caranx ignobilis and Caranx melampygus (Carangidae) in Hawaii. Estuar. Coast. Shelf Sci. 55: 347–359. Smith, T. M. and J. S. Hindell. 2005. Assessing effects of diel period, gear selectivity and predation on patterns of microhabitat use by fish in a mangrove dominated system in SE Austra- lia. Mar. Ecol. Prog. Ser. 294: 257–270. Sparre, P. and S. C. Venema. 1992. Introduction to tropical fish stock assessment. Part 1. FAO Fisheries Technical Paper No. 306.1 Rev. 1, Rome. 376 p. Stoner, A. W. 1991. Diel variation in the catch of fishes and penaeid shrimps in a tropical estuary. Estuar. Coast. Shelf Sci. 33: 57–69. Sudekum, A. E., J. D. Parrish, R. L. Radtke, and S. Ralston. 1991. Life history and ecology of large jack in undisturbed, shallow, oceanic communities. Fish. Bull. 89: 493–513. Tavolga, W. N. 1971. Sound production and detection. In W. S. Hoar and D. J. Randall, eds. Fish physiology, vol. V: sensory systems and electric organs. Academic Press, New York. 600 p. Tiews, K. P., I. A. Divtno, A. Ronquilfo, and J. Marque. 1972. On the food and feeding habits of eight species of leiognathids found in Manilla Bay and San Miguel Bay. Proc. Indo-Pacific Fish. Comm. 13 (III): 85–92. von Brandt, V. 1984. Fish catching methods of the world. (3rd ed.). Fishing News Books, Ltd. Farnham Surrey, England. 418 p. Warwick, R. M. and K. R. Clarke. 1993. Increased variability as a symptom of stress in marine communities. J. Exp. Mar. Biol. Ecol. 172: 225–244. Wassenberg, T. J. and B. J. Hill. 1994. Laboratory study of the effect of light on the emergence behaviour of eight species of commercially important adult penaeid prawns. Aust. J. Mar. Freshw. Res. 45: 43–50. Wetherbee, B. M., K. N. Holland, C. G. Meyer, and C. G. Lowe. 2004. Use of a marine reserve in Kaneohe Bay, Hawaii by the giant trevally, Caranx ignobilis. Fish. Res. 67: 253–263. Williams, A. B. and K. H. Bynum. 1972. A ten-year study of meroplankton in North Carolina estuaries: amphipods. Chesap. Sci. 13: 175–192. Wilson, B., R. S. Batty, and L. M. Dill. 2004. Pacific and Atlantic herring produce burst pulse sounds. Proc. R. Soc. Lond., B (Biology Letters Suppl. 3) 271: S95–S97. Woodroffe, C. 1992. Mangrove sediments and geomorphology. In A. I. Robertson and D. M. Alongi, eds. Tropical mangrove ecosystems. American Geophysical Union, Washington DC. LEY AND HALLIDAY: DIEL VARIATION IN MANGROVE FISH ASSEMBLAGES 715

Yahel, R., G. Yahel, and A. Genin. 2005. Near- bottom depletion of zooplankton over coral reefs: I: diurnal dynamics and size distribution. Coral Reefs 24: 75–85. Young, G. C, I. C. Potter, G. A. Hyndes, and S. de Lestang. 1997. The ichthyofauna of an intermittently open estuary: implications of bar breaching and low salinities on faunal composition. Estuar. Coast. Shelf Sci. 45: 53–68.

Addresses: (J.A.L.) Australian Maritime College, PO Box 21, Beauty Point, Tasmania 7270 Australia. (I.H.) Queensland Department of Primary Industries, Southern Fisheries Centre, P.O. Box 76, Deception Bay, Queensland 4508, Australia. Corresponding Author: (J.A.L.) E-mail: . 716 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Appendix 1.

Catch hr–1 sample–1 Diel Family / species Total Day Night pref. Ambassidae Ambassis gymnocephalus (Lacépède, 1802) 151 0.36 0.98 N1 Ambassis nalua (Hamilton, 1822) 1 0.01 D3 Ambassis spp. 4 0.02 N2 Ambassis vachelli Richardson, 1846 536 0.61 4.72 N2 Apogonidae Apogon hyalosoma Bleeker, 1852 2 0.01 N3 Ariidae Arius nella (Valenciennes, 1840) 556 2.50 2.58 N0 Arius proximus Ogilby, 1898 1,216 3.32 8.04 N1 Arius spp. 372 1.00 2.25 N1 Atherinidae 10 0.13 N3 Batrachoididae Halophryne diemensis (Lesueur, 1824) 4 0.03 D3 Belonidae Strongylura incisa (Valenciennes, 1846) 1 0.01 N3 Strongylura leiura (Bleeker, 1850) 1 0.01 N3 Strongylura sp. 3 0.03 N3 Strongylura strongylura (van Hasselt, 1823) 328 1.30 1.66 N0 Tylosurus crocodilus (Péron and Lesueur, 1821) 3 0.02 N3 Tylosurus punctatus (Günther, 1872) 1 0.01 D3 Tylosurus sp. 1 0.01 N3 Bothidae Pseudorhombus sp. 1 0.01 D3 Carangidae Alectis indicus (Rüppell, 1830) 8 0.08 N3 Carangoides hedlandensis (Whitley, 1934) 1 0.01 N3 Caranx bucculentus Alleyne and Macleay, 1877 1 0.01 D3 Caranx ignobilis (Forsskål, 1775) 185 0.64 1.01 N0 Caranx sexfasciatus Quoy and Gaimard, 1825 9 0.02 0.06 N1 Gnathanodon speciosus (Forsskål, 1775) 3 0.03 N3 Scomberoides commersonianus Lacépède, 1801 366 1.67 1.85 N0 Scomberoides lysan (Forsskål, 1775) 21 0.12 0.10 D0 Scomberoides sp. 2 0.02 N3 Scomberoides tala (Cuvier, 1832) 83 0.31 0.59 N1 Scomberoides tol (Cuvier, 1832) 5 0.05 D3 Trachinotus bailloni (Lacépède, 1801) 6 0.06 D3 Trachinotus blochi (Lacépède, 1801) 27 0.06 0.19 N1 Carcharhinidae Carcharhinus ambionensis (Müller and Henle, 1839) 2 0.03 D3 Carcharhinus leucas (Müller and Henle, 1839) 153 0.56 0.83 N0 Carcharhinus sp. 13 0.15 D3 Rhizoprionodon acutus (Rüppell, 1837) 2 0.02 N3 LEY AND HALLIDAY: DIEL VARIATION IN MANGROVE FISH ASSEMBLAGES 717

Appendix 1. Continued.

Catch hr–1 sample–1 Diel Family / species Total Day Night pref. Centropomidae Lates calcarifer (Bloch, 1790) 1656 2.46 12.45 N2 Psammoperca waigiensis (Cuvier, 1828) 1 0.01 N3 Chanidae Chanos chanos (Forsskål, 1775) 38 0.25 0.09 D1 Chirocentridae Chirocentrus dorab (Forsskål, 1775) 15 0.08 0.07 D0 Clupeidae Anodontostoma chacunda (Hamilton, 1822) 204 1.39 0.78 D0 Escualosa thoracata (Valenciennes, 1847) 248 0.51 2.38 N2 Herklotsichthys castelnaui (Ogilby, 1897) 7494 20.85 57.33 N1 Herklotsichthys koningsbergi (Weber and de Beaufort, 1912) 12 0.12 N3 Nematalosa come (Richardson, 1846) 761 2.05 5.06 N1 Nematalosa erebi (Günther, 1868) 30 0.14 0.17 N0 Sardinella albella (Valenciennes, 1847) 20 0.11 0.07 D0 Sardinella brachysoma Bleeker, 1852 179 1.59 0.35 D2 Cynoglossidae Paraplagusia sp. 1 0.01 N3 Dasyatidae Himantura granulata (Macleay, 1883) 1 0.01 D3 Himantura sp. 1 0.01 N3 Himantura toshi Whitley, 1939 1 0.01 D3 Himantura uarnak (Forsskål, 1775) 2 0.01 0.01 N0 Drepanidae Drepane punctata (Linnaeus, 1758) 128 0.70 0.56 D0 Echeneidae Echeneis naucrates Linnaeus, 1758 2 0.02 N3 Elopidae Elops hawaiensis Regan, 1909 63 0.26 0.37 N0 Engraulidae Stolephorus commersoni Lacépède, 1803 72 0.11 0.50 N2 Stolephorus nelsoni Wongratana, 1987 18 0.23 D3 Stolephorus sp. 89 0.42 0.79 N1 Thryssa baelama (Forsskål, 1775) 4 0.04 N3 Thryssa hamiltoni (Gray, 1835) 2263 4.67 20.58 N2 Ephippidae Platax batavianus Cuvier, 1831 1 0.01 N3 Platax orbicularis (Forsskål, 1775) 1 0.01 N3 Platax teira (Forsskål, 1775) 1 0.01 N3 Zabidius novemaculeatus (McCulloch, 1916) 1 0.01 N3 Gerreidae Gerres erythrourus (Bloch, 1791) 17 0.02 0.15 N2 Gerres longirostris (Lacépède, 1801) 3 0.04 D3 Gerres filamentosusCuvier, 1829 41 0.23 0.17 D0 Gerres oblongus Bleeker, 1854 4 0.02 0.02 N1 Gerres oyena (Forsskål, 1775) 14 0.01 0.11 N2 718 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Appendix 1. Continued.

Catch hr–1 sample–1 Diel Family / species Total Day Night pref. Gobiidae Psammogobius biocellatus (Valenciennes, 1837) 1 0.01 D3 Periophthalmus sp. 1 0.01 D3 Haemulidae Plectorhinchus gibbosus (Lacépède, 1802) 27 0.04 0.19 N2 Pomadasys argenteus (Forsskål, 1775) 284 0.99 1.71 N0 Pomadasys kaakan (Cuvier, 1830) 353 1.27 1.91 N0 Pomdasys maculatus (Bloch, 1793) 3 0.03 N3 Hemiramphidae Arrhamphus sclerolepis Günther, 1866 213 1.41 0.67 D1 Hemiramphidae 5 0.03 N3 Hemiramphus robustus Günther, 1866 7 0.03 0.03 D0 Hyporhamphus quoyi (Valenciennes, 1847) 5 0.01 0.04 N1 Hyporhamphus regularis (Whitley, 1931) 2 0.01 0.01 N0 Zenarchopterus buffonis (Valenciennes, 1847) 48 0.25 0.18 D0 Zenarchopterus gilli Smith, 1945 1 0.01 N3 Zenarchopterus sp. 1 0.01 D3 Lactariidae Lactarius lactarius (Bloch and Schneider, 1801) 14 0.01 0.11 N2 Leiognathidae Gazza achlamys Jordan and Starks, 1917 3 0.02 N3 Gazza minuta (Bloch, 1795) 17 0.01 0.15 N2 Gazza sp. 5 0.03 N3 Leiognathus bindus (Valenciennes, 1835) 9 0.03 0.05 N0 Leiognathus blochi (Valenciennes, 1835) 4 0.03 0.01 D1 Leiognathus brevirostrus (Valenciennes, 1835) 143 0.28 1.15 N2 Leiognathus daura (Cuvier, 1829) 5 0.05 N3 Leiognathus equulus (Forsskål, 1775) 1,383 2.08 10.29 N2 Leiognathus sp. 20 0.04 0.14 N1 Leiognathus splendens (Cuvier, 1829) 42 0.14 0.30 N1 Secutor insidiator (Bloch, 1787) 3 0.02 N3 Secutor ruconius (Hamilton, 1822) 10 0.06 0.04 D0 Leptobramidae Leptobrama mulleri Steindachner, 1878 63 0.62 0.11 D2 Lutjanidae Lutjanus argentimaculatus (Forsskål, 1775) 40 0.07 0.26 N1 Lutjanus johnii (Bloch, 1792) 28 0.07 0.16 N1 Lutjanus russelli (Bleeker, 1849) 6 0.06 N3 Lutjanus sp. 1 0.01 N3 Megalopidae Megalops cyprinoides (Broussonet, 1782) 208 0.30 1.47 N2 Monodactylidae Monodactylus argenteus (Linnaeus, 1758) 18 0.03 0.14 N2 LEY AND HALLIDAY: DIEL VARIATION IN MANGROVE FISH ASSEMBLAGES 719

Appendix 1. Continued.

Catch hr–1 sample–1 Diel Family / species Total Day Night pref. Mugilidae Liza subviridis (Valenciennes, 1836) 750 4.70 2.96 D0 Liza vaigiensis (Quoy and Gaimard, 1825) 281 1.49 1.16 D0 Mugil cephalus Linnaeus, 1758 121 0.14 N1 Mugilidae 16 0.25 0.93 N3 Valamugil buchanani (Bleeker, 1854) 308 2.14 0.92 D1 Valamugil cunnesius (Valenciennes, 1836) 997 5.78 4.45 D0 Valamugil seheli (Forsskål, 1775) 166 0.60 0.92 N0 Valamugil sp. 71 0.21 0.42 N1 Muraenesocidae Muraenesox cinereus (Forsskål, 1775) 2 0.02 N3 Myliobatidae Aetobatus narinari (Euphrasen, 1790) 6 0.04 0.02 D1 Platycephalidae Platycephalus fuscus Cuvier, 1829 19 0.08 0.12 N0 Platycephalus indicus (Linnaeus, 1758) 19 0.05 0.16 N1 Plotosidae Paraplotosus albilabris (Valenciennes, 1840) 11 0.01 0.10 N2 Polynemidae Eleutheronema tetradactylum (Shaw, 1804) 691 1.82 5.86 N1 Polydactylus macrochir (Günther, 1867) 172 0.57 1.00 N0 Filimanus heptadactylus (Cuvier, 1829) 6 0.01 0.07 N2 Rhinobatidae Aptychotrema rostrata (Shaw, 1794) 2 0.02 D3 Scatophagidae Scatophagus argus (Linnaeus, 1766) 30 0.16 0.12 D0 Selenotoca multifasciatus (Richardson, 1846) 28 0.11 0.12 N0 Sciaenidae Nibea soldado (Lacépède, 1802) 384 0.48 3.00 N2 Otolithes ruber (Bloch and Schneider, 1801) 1 0.01 N3 Sciaenidae 3 0.03 N3 Scombridae Scomberomorus semifasciatus (Macleay, 1883) 58 0.48 0.12 D1 Scombridae 6 0.06 N3 Scorpaenidae Notesthes robusta (Günther, 1860) 1 0.01 N3 Serranidae Epinephelus coioides (Hamilton, 1822) 2 0.01 0.01 D1 Epinephelus lanceolatus (Bloch, 1790) 2 0.01 N3 Epinephelus malabaricus (Bloch and Schneider, 1801) 2 0.02 D3 Siganidae Siganus guttatus (Bloch, 1787) 1 0.01 N3 Siganus lineatus (Valenciennes, 1835) 2 0.02 N3 720 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Appendix 1. Continued.

Catch hr–1 sample–1 Diel Family / species Total Day Night pref. Sillaginidae Sillago analis Whitley, 1943 5 0.02 0.03 N1 Sillago ciliata Cuvier, 1829 5 0.04 D3 Sillago sihama (Forsskål, 1775) 109 0.52 0.54 N0 Soleidae Synaptura nigra Macleay, 1880 4 0.03 N3 Sparidae Acanthopagrus australis (Günther, 1859) 5 0.04 N3 Acanthopagrus berda (Forsskål, 1775) 15 0.02 0.11 N2 Sphyraenidae Sphyraena barracuda (Edwards, 1771) 17 0.05 0.10 N1 Sphyraena jello Cuvier, 1829 31 0.08 0.19 N1 Sphyrnidae Sphyrna lewini (Griffith and Smith, 1834) 3 0.02 0.01 D1 Stromateidae Parastromateus niger (Bloch, 1795) 7 0.05 0.02 D1 Teraponidae Mesopristes argenteus (Cuvier, 1829) 4 0.01 0.03 N1 Terapon jarbua (Forsskål, 1775) 5 0.05 N3 Tetraodontidae Arothron hispidus (Linnaeus, 1758 17 0.10 0.08 D0 Arothron manilensis (Marion de Procé, 1822) 3 0.04 N3 Arothron reticularis (Bloch and Schneider, 1801) 9 0.02 0.06 N1 Chelonodon patoca (Hamilton, 1822) 4 0.03 0.02 D0 Marilyna pleurosticta (Günther, 1872) 3 0.04 D3 Tetraodontidae 2 0.02 N3 Toxotidae Toxotes chatareus (Hamilton, 1822) 59 0.27 0.26 D0 Toxotes jaculatrix (Pallas, 1767) 8 0.04 0.02 D1 Trichiuridae Trichiurus lepturus Linnaeus, 1758 1 0.01 N3 Total abundance 24,908 7,916 17,292 Total species 157. 110. 135. Total families 55. 42. 48. BULLETIN OF MARINE SCIENCE, 80(3): 721–737, 2007

An assessment of ichthyofaunal assemblages within the mangal of the Belize offshore cays

D. Scott Taylor, Eric A. Reyier, Carole C. McIvor, and William P. Davis

abstract We assessed ichthyofaunal diversity within offshore mangrove cays in Belize during three, 2-wk surveys (2003, 2004, 2005). Nine sampling gears were deployed in pre-defined micro-habitats: fringe, transition, dwarf red mangrove, internal creeks, ponds, and sinkholes. Water quality data (temperature, salinity, DO) were taken during most collections. A total of 2586 gear sets was completed and 8131 individuals collected, comprising 75 taxa. Minnow trap data from the various micro-habitats tested indicates some overlap in assemblages. Significant differences in water quality were also noted, with the fringe representing the most benign and the sinkhole the most harsh microhabitats, respectively. We also conducted extensive visual surveys around the fringe at a number of cays, tallying an additional 67 taxa. The fringe is the most diverse (128 taxa) and sinkhole least (12 species). An overall total of 142 taxa from 55 families has therefore been documented from the cays, and all but eight were found on Twin Cays alone. This figure is among the highest reported for oceanic mangroves in this biogeographic realm. Our comprehensive approach with a variety of gear-types in a wide range of micro-habitats, combined with visual observation, lends credence to the conclusion that most ichthyological species inventories for the mangal are commonly underestimates.

The biological connectivity between mangroves and adjoining seagrasses and coral reef systems of the worldwide tropics is now clearly established. However, there is an increasing need to better understand the complex association between the mangal and adjacent ecosystems, as removal or degradation of the world’s mangroves is of great concern, with 30%–50% already removed (World Resources Institute, 1996; Alongi, 2002), and the rate of destruction apparently accelerating (Valiela et al., 2001). Both mangroves and seagrasses have established values as feeding and nursery grounds for a number of coral reef–dependent fish species (extensively reviewed in Sheridan and Hays, 2003; Manson et al., 2005; Verweij et al., 2006). In addition to the physical shelter, predation reduction and feeding opportunities offered by mangroves to juvenile coral reef fishes (Laegdsgaard and Johnson, 2001; Verweij et al., 2006), mangroves offer an important transitional habitat, allowing some juveniles, notably lutjanids and haemulids, to migrate to coral reefs with ontogenetic shifts in diet (Cocheret de la Morinère et al., 2003). However, significant differences may exist between the mangal of the Indo-Pacific and western Atlantic in terms of nursery value (Parrish, 1989; Sheaves, 2005). In Netherland Antilles’ embayments with both seagrass and mangroves, Doren- bosch et al. (2004) studied 17 species of fishes and 15 were also observed on adjacent reefs. Four “nursery” species were found at their highest densities on reefs associated with seagrass/mangroves, and another three species were similarly associated, but with abundance varying more over different reef types. Within one of those same Netherland Antilles embayments, Nagelkerken and van der Velde (2002) concluded that non-estuarine mangroves have greater densities of juvenile fishes than adjacent seagrasses, mud flats or channels. Densities within mangroves were more similar to those of the coral reef. They attributed the attractiveness

of mangroves to juvenile fishes to their structural complexity. These authors also raised an important point in their discussion of “estuarine dependence”. Namely, that not all mangroves are found in estuarine “conditions”, with some having purely an oceanic influence. A more basic issue is knowing which fish species utilize mangroves and at what life stages or seasonality. Part of this uncertainty is due to the inherent difficulty of sampling in the complex 3-dimensional environment of the mangal, but Clynick and Chapman (2002) have noted that most studies are inconsistent in sampling method or design and some studies examine fishes only in mangroves, while others include adjacent habitats (seagrasses/mudflats). A more recent review (Faunce and Serafy, 2006) pointed out that one out of five published studies of mangrove fishes did not actually sample in mangroves. The number of fish species reported from Caribbean mangroves varies widely. Baelde (1990) studied fish assemblages in seagrass adjacent to non-estuarine mangroves and coral reefs in Guadeloupe. A fixed net capéchade( ) collected 61 species adjacent to the mangroves, with more estuarine-associated species found here. The same gear-type was used by Louis et al. (1995) in Martinique, where the net was deployed offshore from estuarine mangroves in mud or seagrass bottom. Seagrass sets caught more species, and a total of 87 species was taken. Rooker and Dennis (1991) visually surveyed mangroves in Puerto Rico and documented 41 species of fish. Acosta (1997) used gill and trammel nets and visual census to record fishes within mangroves and coral reefs on the southwest coast of Puerto Rico. Ninety- six species were documented from the two habitats. Stoner (1986), also studying a Puerto Rico mangrove estuary, used trawls to collect 41 species of fish. Two studies from Cuba have documented 87 species by visual census (Claro and Garcia Arteaga, 1993) and 149 species from “soft bottoms” and mangroves (Claro et al., 2001). In the latter study (utilizing commercial trawls, gill nets, visual census, and rotenone) no distinction was made on the number of species from the respective habitats. A comprehensive summary of sampling methodology is provided by Faunce and Serafy (2006). However, they point out that more than one-third of studies utilized towed gears, a technique not effective in sampling fishes actually found within mangroves. Thayer et al. (1987) were among the first to undertake the difficult task of -quan tifying fishes within mangroves of south Florida (Everglades National Park) estuaries. Working with blocknets and rotenone, they found 36 species exclusively in mangroves, 24 species in adjacent seagrass and 27 species in both habitats, giving a total of 63 species for mangroves. Gilmore (1995) exhaustively sampled estuarine mangroves in the Indian River Lagoon (east central Florida) over a period of more than 20 yrs, and recorded 88 species of fish. Ley et al. (1999), also collecting in south Florida estuaries with block nets with rotenone and visual surveys, recorded 76 species within prop roots. Layman and Silliman (2002) visually surveyed five habitats (including mangrove) within an estuarine creek in Andros Island, Bahamas. They documented a total of 57 species with 42 found in mangroves. Also in Andros, Layman et al. (2004) examined how fragmentation affects fish assemblages in 30 different estuaries. An extensive visual survey of four different habitats (sandflat, seagrass, mangrove, rock) recorded 86 species. In an earlier study from Exuma, Bahamas, Wilcox et al. (1975) collected with rotenone in creek channels of island mangroves, and 56 species were taken. TAYLOR ET AL.: MANGROVE FISH ASSEMBLAGES IN THE BELIZE CAYS 723

The Belize barrier reef system has been previously documented as an area rich in fish species, with 497 species reported from the seagrasses, patch mangroves and reefs of the Pelican Cays in the southern area of the reef tract (Smith et al., 2003). However, mangrove habitats were specifically excluded in this study.H ere, we report the results of a study conducted in the cays of the Belize barrier reef which appears to be among the most comprehensive of surveys from the Caribbean region.

Methods

Study Location.—We assessed mangrove fish assemblages on the mangrove cays of the Belize barrier reef, with a focus on Twin Cays, a 92 ha complex of mangrove islands with two main cays (East/West) located 1.5 km west of the barrier reef (16°50´N, 88°06´W). We also completed supplemental visual surveys on various cays north to Tobacco Range and south to the Pelican Cays (Fig. 1). Our survey encompassed two week visits during Oct.–Nov. 2003, Sept. 2004, and April–May 2005. In addition to visual surveys, we utilized the following nine primary gear-types in our collections: fyke net (5 hoops, 4.7 m wings/1.7 m high, 6 mm mesh bag and wings), 1 m2 throw-trap, center bag seine (10 m long/6.7 mm mesh), Gee™wire minnow traps, “cup” traps (Taylor, 1990), Breder traps, oval mesh traps, 1 m2 drop-trap, gill net (38 mm stretch mesh). In addition, we also made opportunistic collections with four other gears: cast net, hook and line, hand spear, and hand nets. We identified the following micro-habitat types within Twin Cays for deployment of these gears, with some of the descriptions following those of Woodroffe (1995) and Feller et al. (2002): fringe (shoreline, Rhizophora mangle Linnaeus 1753 dominant), transition (landward of fringe, Avicennia germinans (Linnaeus) Stearn 1958/Laguncularia racemosa (Linnaeus) Gaertn.f./R. mangle mix), dwarf red (internal, stunted R. mangle), pond (internal, open shallow water bodies), internal creek (interior tidal channels), and sink hole (small, deep, infrequently tidally flushed, internal ponds). Table 1 illustrates the deployment of the various gears within these micro-habitats. The fyke and gill nets, minnow, cup and oval traps were deployed for 24 hrs. Minnow and oval traps were baited with cut fish bait. Cup traps were covered with debris/leaf litter, with the intent of providing a refuge for small fishes in very shallow water. At most collections, water quality data were obtained with a YSI 6920 sonde, including temperature, dissolved oxygen (DO), and salinity. All collected fishes were identified, counted, and released, except for needed specimens for vouchers or stable isotope studies, with nomenclature following Froese and Pauly (2006). Although we have previously documented the abundance and habitat preferences of Kryp- tolebias (formerly Rivulus) marmoratus on the Belize cays (Taylor et al., 2004), we have not attempted to quantify its abundance. Accordingly, we established fiberglass screen enclosures within the transition habitat on Twin Cays at low tide, at which time K. marmoratus is “fos- sorial” (underground in crab burrows or inside logs or leaf litter). The lower edge of the enclosures was buried into the peat substrate and the upper edge was supported by stakes, keeping it above water at high tide. Four areas varying from 0.75 to 2.2 m2 were enclosed and the areas trapped with cup traps over several days, until all K. marmoratus were removed. The visual surveys were conducted by snorkelers swimming around the outside of the fringe or within internal tidal creeks, ponds, and dwarf red mangrove. Observers compiled species lists with the strict criteria of a specimen being observed within the prop root zone, under the overhanging canopy, or within internal creeks, ponds, or dwarf stands. Species within adjacent seagrass beds were not included. These observations were anecdotal, the areal extent surveyed was not quantified and observations were qualitative only. However, surveys were completed at several cays in addition to Twin Cays (Fig. 1). Data Analysis.—An overall species list was compiled from gear collections and visual surveys using the semi-quantitative abundance categories modified from Starck (1968): Abundant, noted in large numbers during most collection efforts and visual surveys; Com- 724 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Figure 1. The cays of the central Belize barrier reef (from Taylor et al., 2004). Twin Cays indicated and other sites where visual surveys conducted are marked with an asterisk (k). mon, observed regularly, sometimes in large numbers; Occasional, encountered infrequently or in small numbers; and Rare, seldom encountered. We examined minnow trap data (the one gear consistently used across all microhabitats) in greater detail using non-metric multidimensional scaling (MDS). Minnow trap catch rates were relatively low, with many “zero” catches, so catches were pooled by habitat type across years with effort standardized as fish per trap day in order to generate a single “sample” for each. These values were then square- root transformed to downweight the otherwise dominant influence of abundant taxa (allowing less common species to contribute to habitat discrimination) and a two-dimensional MDS ordination was generated using PRIMER 6.0 software. Shannon diversity (H’) indices were calculated for each microhabitat for these data. Water quality data were analyzed with a two-way ANOVA (with microhabitat and year as main effects) and Tukey HSD post-hoc tests to identify differences in water quality between microhabitats, using SPSS 12.0 software. TAYLOR ET AL.: MANGROVE FISH ASSEMBLAGES IN THE BELIZE CAYS 725

Table 1. Summary of gears used in various mangrove micro-habitats on Twin Cays, Belize.

Micro-habitat Gears deployed Fringe Minnow, cup, oval trap, hand net, hand spear Transition Minnow, cup, Breder trap, 1 m2 drop trap Dwarf red Minnow, oval, cup trap, cast net Pond Gill net, seine, minnow trap, 1 m2 throw trap, cast net Internal creek Fyke net, minnow, oval trap, hand spear Sinkhole Minnow, cup trap, hook and line

Results

We completed 2586 gear-sets of the nine primary gear types over the three collection cycles. Within the fringe, there were 1277 sets, 321 in the transition (low water frequently prevented deployment of traps here, particularly in 2005 when this zone did not flood at all), 614 in dwarf red, 203 in sinkhole, 66 in internal creek, and 101 in pond. A total of 8131 individual fishes was taken with all 13 gears, comprising 75 taxa. Many gear-sets (38.2%), particularly with smaller traps, resulted in no catch. The most common species taken were Poecilia orri (35% of total catch), followed by Gerres cinereus (16%), and Gambusia yucatana (11%). Fyke net sets within internal creeks produced the largest number of individuals (n = 2672), consisting mostly of gerreids, followed by gears set within dwarf and sinkhole zones, with high overall catches consisting mostly of two species, P. orri and G. yucatana. Catches from the fringe had 98 taxa unique to that micro-habitat, far exceeding other micro-habitats (Table 2). The MDS plot shows modest discrimination of species assemblages for minnow trap collections across micro-habitats (Fig. 2). Visual Surveys.—The visual surveys described account for 104 person-hrs over the three collection periods, including some nighttime surveys. These surveys documented an additional 67 species. A heavy advection of Sargassum sp. algae into the windward R. mangle fringe on the East Cay in 2005 probably accounted for the collection of a few species that would not commonly be encountered in mangroves, including Microphis brachyurus, Syngnathus pelagicus, Balistes capriscus, Histrio histrio, and Cantherhines pullus. Table 2 shows the breakdown of collections and visual survey within microhabitats, with the number of taxa collected and observed and number of unique taxa. Between the gear collections and the visual surveys, the total number of taxa documented was 142 (Table 3). Water Quality.—In conjunction with fish collections, 115 water quality measurements were recorded. Significant differences (P < 0.05) in water temperature, salinity and dissolved oxygen were observed across micro-habitats (Table 4). The lowest water temperature was found in the fringe and highest in the transition. Sa- linity was highest in sinkhole and lowest in the transition. Dissolved oxygen was highest in pond and lowest in sinkhole.

Discussion

Our documentation of 142 taxa from 55 different families from the Belize offshore mangal represents perhaps the highest number of taxa reported from this biogeographic realm, including both estuarine and oceanic/island mangrove systems. Remarkably, 726 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Figure 2. A comparison of the ichthyofaunal species assemblages collected with minnow traps between mangrove micro-habitats on Twin Cays, Belize, as demonstrated by non-metric multidimensional scaling. Interpoint distances are proportional to overall faunal similarity. 2-D stress = 0.02. Values in parentheses indicate Shannon diversity (H'). all but eight taxa were recorded from Twin Cays alone. The high species diversity of this site can be accounted for by its large size, the diversity of micro-habitats found there and its close proximity to the barrier reef, allowing a number of unexpected species to be found there (i.e., the Sargassum fauna described above). Our 2-wk collections over three consecutive years and at two different seasons also allowed a wider range of opportunities to document new species. However, the gradually increasing species count over the 3 yrs illustrates the difficulty of collection/observation in the mangal. Our initial collection in 2003 documented 86 species, followed by 22 new species in 2004, and 34 in 2005. Should this effort have continued, specifically with the addition of ichthyocides, it is likely that other species, notably the cryptic (e.g., anguilliform, gobioid, and blennioid fishes) and “infaunal” species (e.g., the one synbranchid known from Belize) that are still missing could be included. This is analogous to the situation in Caribbean coral reefs, where species counts by standard collection or visual census techniques were significantly augmented by use of rotenone (Collette et al., 2003; Smith-Vaniz et al., 2006). In examining species distributions across micro-habitats by all techniques, we documented 128 species from the fringe, 30 each in creeks and ponds, 20 from the dwarf red, 18 from the transition, and 12 from the sinkhole. The fringe dominated both in number of families (n = 48) and number of species within the dominant families. Within the fringe, serranids, pomacentrids, gobiids, haemulids, lutjanids, and scarids combined accounted for 51 species, 36% of the total species documented. Outside of the fringe, gerreids, cyprinodontids, and poeciliids were the most speci- ose, and taxa from these families generally dominated the entire gear collections in all microhabitats. The lower diversity from the sinkhole is no doubt indicative of the poorer water quality found here. However, three species were unique to this microhabitat: the two eleotrids Guavina guavina and Dormitator maculatus, and Mega- lops atlanticus, all species known to be tolerant of low DO (Graham, 1997). This site had the lowest daytime DO, and we suspect that at night DO falls precipitously, as fish from this site were frequently found dead in the traps. Only during very high tides, would water quality conditions improve enough to allow less hardy species to TAYLOR ET AL.: MANGROVE FISH ASSEMBLAGES IN THE BELIZE CAYS 727 (10%) (6%) (31%) (53%) (68%) (36%) (65%) spp. (10%) Dominant taxa (Percent of total gear catch) Poecilia orri Lutjanus apodus (12%) Bathygobius Poecilia orri Floridichthys polyommus Kryptolebias marmoratus (6%) Poecilia orri Gambusia spp. (20%) (16%) cinereus Gerres Poecilia orri Gambusia spp. (23%) Dormitator maculatus (6%) (37%) cinereus Gerres Eucinostomus spp. (35%) Floridichthys polyommus Eucinostomus spp. (33%) Floridichthys polyommus (16%) cinereus Gerres Dominant families (No. of taxa) Serranidae (10) Pomacentridae (9) Gobiidae (8) Haemulidae (8) Lutjanidae (8) Scaridae (8) Gerreidae (5) Poeciliidae (3) Cyprinodontidae (2) Gerreidae (5) Cyprinodontidae (3) Poeciliidae (3) Poeciliidae (3) Cyprinodontidae (2) Eleotridae (2) Lutjanidae (2) Gerreidae (5) Apogonidae (3) Carangidae (2) Cyprinodontidae (2) Haemulidae (2) Poeciliidae (2) Sciaenidae (2) Gerreidae (5) Lutjanidae (5) Cyprinodontidae (3) 408 189 1,234 1,758 1,870 2,672 No. fish collected 7 48 10 12 17 18 observed collected/ No. families 0 0 3 3 1 98 (total) Taxa unique Taxa to microhabitat 18 20 12 30 30 128 Taxa (total) 0 1 0 2 7 56 Taxa observed 72 18 19 12 28 23 Taxa collected Table 2. Table Summary of collection (13 gear types) and visual observations within six microhabitats Cays, at Twin Belize. Shown are taxa collected, taxa visually observed, total taxa, unique taxa, number of families Also(collected and shownobserved), areand dominantnumber familiesof withfish number collected. of taxa in parentheses and dominant with percent of total catch parentheses. Microhabitat Fringe Transition Dwarf Sinkhole Internal Creek Pond 728 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 R R R C C R R R R R C A C A R O O O O O O O O R* Relative abundance V V V V O O V H O O O M M M M M M F H F O F O F O F O type H M Gear C H M F F F F F F F S F F F F F F F F F F S T S C P C F type C F P C F P Habitat C D F P nurse shark Common name requiem shark sargassumfish southern stingray shortnose batfish Chupare stingray clingfish yellow stingray hardhead silverside tarpon mangrove rivulus bonefish chain moray green moray spotted moray speckled worm eel manytooth conger false herring redear herring dwarf round herring sand diver key anchovy viviparous brotula viviparous brotula (Gosse, 1851) (Cuvier, 1829) (Cuvier, (Cuvier, 1816) (Cuvier, Valenciennes, 1847 Valenciennes, (Cuvier, 1829) (Cuvier, (Bloch, 1795) (Linnaeus, 1758) Ginglymostoma cirratum (Bonnaterre, 1788) Species unidentified carcharhinid Histrio histrio (Linnaeus, 1758) Hildebrand and Schroeder, 1928 Dasyatis americana Hildebrand and Schroeder, (Cuvier, 1829) Ogcocephalus nasutus (Cuvier, (Werner, 1904) (Werner, Himantura schmardae Gobiesocid sp. A Gobiesocid sp. Urobatis jamaicensis Urobatis (Müller and Troschel, 1848) Troschel, Atherinomorus stipes (Müller and Megalops atlanticus (Poey, 1880) Kryptolebias marmoratus (Poey, Albula vulpes Echidna catentata Gymnothorax funebris Ranzani, 1840 (Cuvier, 1829) Gymnothorax moringa (Cuvier, Myrophis punctatus Lütken, 1852 Myrophis Conger triporiceps Kanazawa, 1958 Harengula humeralis Harengula Jenkinsia lamprotaenia Agassiz, 1829) Synodus intermedius (Spix and Harengula clupeola Harengula (Fowler, 1906) Anchoa cayorum (Fowler, A Bythitid sp. Bythitid sp. B Ginglymostomatidae Table 3. Fishes collected or observed within mangrove micro-habitats of Twin Cays, Belize 2003–2005. Micro-habitat Types: (C) Internal Creek, (D) Dwarf Red, Dwarf (D) Creek, Internal (C) Types: Micro-habitat 2003–2005. Belize Cays, Twin of micro-habitats mangrove within observed or collected Fishes 3. Table (F) Fringe, (P) Pond, (S) Sinkhole, (C) (T) Gear Transition. Cup Type: (F) Trap, Fyke Net, (H) Spear or Hand Net, (M) Minnow Observation (V) Trap, Visual (O) Only, Other techniques (hook and line, oval trap, gill net, cast net, throw trap, drop trap, bag seine, Asterisk indicates taxa collected/observed from nearby cays only. Breder Common, (O) Occasional, (R) Rare. trap). Relative Abundance: (A) Abundant, (C) Family Carcharhinidae Antennariidae Dasyatidae Ogcocephalidae Gobiesocidae Urolophidae Atherinidae Megalopidae Rivulidae Albulidae Muraenidae Ophichthidae Congridae Clupeidae Synodontidae Engraulidae Bythitidae TAYLOR ET AL.: MANGROVE FISH ASSEMBLAGES IN THE BELIZE CAYS 729 C C A A A A R R R R R R R O O O O O O O O O R* R* Relative abundance V V V V V V V V V V V H H H O O M type M O M O Gear F H O C F H M O C F H M O C F H M O C F H M O F F F F F F F F F F F F F F F F F D P type D S T D S Habitat D F P T D F P C D P S T S C D P T S C D P C D F P S T S C D F P C D F P S T S C D F P pike killifish flat needlefish Common name s. Yucatán mosquitofish Yucatán s. mangrove molly Yucatán pupfish Yucatán ocellated killifish flagfish Yucatán redfin needlefish Timucu butter hamlet longspine squirrelfish Mayan hamlet longsnout seahorse black grouper opossum pipefish yellowmouth grouper sargassum pipefish sargassum flying gurnard graysby red hind goliath grouper Nassau grouper black hamlet barred hamlet (Poey, 1860) (Poey, Hubbs, 1936 (Lacépède, 1802) (Poey, 1852) (Poey, (Linnaeus, 1758) (Walbaum, 1792) (Walbaum, Linnaeus, 1758 (Bloch, 1792) (Linnaeus, 1758) Kner, 1860 Kner, (Cuvier, 1828) (Cuvier, (Lichtenstein, 1822) Ginsburg, 1933 Ginsburg, (Valenciennes, 1846) (Valenciennes, Fowler, 1943 Fowler, Belonesox belizanus Species Gambusia yucatana Regan, 1914 Poecilia orri Cyprinodon artifrons Hubbs, 1936 Cyprinodon artifrons Floridichthys polyommus Garmanella pulchra Hubbs, 1936 Ablennes hians (Poey, 1860) notata (Poey, Strongylura 1792) timucu (Waulbaum, Strongylura Hypoplectrus unicolor (Walbaum, 1792) Holocentrus rufus (Walbaum, Hypoplectrus sp. Hippocampus reidi Hippocampus reidi (Poey, 1860) bonaci (Poey, Mycteroperca Microphis brachyurus lineatus (Kaup, 1856) Microphis Mycteroperca interstitialis Mycteroperca Syngnathus pelagicus Dactylopterus volitans Cephalopholis cruentata Epinephelus guttatus Epinephelus itajara Epinephelus striatus Hypoplectrus nigricans Hypoplectrus puella Poeciliidae Table 3. Continued. Table Family Cyprinodontidae Belonidae Holocentridae Syngnathidae Dactylopteridae Serranidae 730 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 R R R A A A A C R R A C C R R A O O O O O O R* Relative abundance F F V V V V V V V H H H M F O F O F O type Gear F M O F M O F M O F M O F M O F H M O F H M O F F F F F F F F C F P F P C F C F C F type C F P Habitat C F P S C F P C D F P C D F P C D F P T C D F P C D F P T C D F P T C D F P C D F P T C D F P T C D F P C D F P S T S C D F P barred cardinalfish flamefish dusky cardinalfish freckled cardinalfish remora slender mojarra flagfin mojarra yellowfin mojarra Common name yellow jack bar jack horse-eye jack leatherjack mutton snapper schoolmaster cubera snapper gray snapper dog snapper mahogany snapper lane snapper yellowtail snapper tripletail jenny mojarra tidewater mojarra (Poey, 1860) (Poey, (Günther, 1879) (Günther, (Sylvester, 1915) (Sylvester, Poey, 1867) Poey, spp. spp. Apogon binotatus ( 1860) Apogon maculatus (Poey, Phaeoptyx conklini Echeneis Eucinostomus jonesii (Bleeker, 1863) Eucinostomus melanopterus (Bleeker, 1792) (Walbaum, cinereus Gerres Species Phaeoptyx pigmentaria (Cuvier, 1833) Carangoides bartholomaei (Cuvier, Carangoides ruber (Bloch, 1793) Caranx latus Agassiz, 1831 Oligoplites (Cuvier, 1828) Lutjanus analis (Cuvier, (Walbaum, 1792) Lutjanus apodus (Walbaum, (Cuvier, 1828) Lutjanus cyanopterus (Cuvier, Lutjanus griseus (Linnaeus, 1758) 1801) Lutjanus jocu (Bloch and Schneider, (Cuvier, 1828) Lutjanus mahogoni (Cuvier, Lutjanus synagris (Linnaeus, 1758) Ocyurus chrysurus (Bloch, 1791) Lobotes surinamensis (Bloch, 1790) Eucinostomus gula (Quoy and Gaimard, 1824) Eucinostomus harengulus Goode and Bean, 1879 Eucinostomus harengulus Apogonidae Echeneidae Table 3. Continued. Table Family Carangidae Lutjanidae Lobotidae Gerreidae TAYLOR ET AL.: MANGROVE FISH ASSEMBLAGES IN THE BELIZE CAYS 731 C C C R C A C C A R R R R C R R R O O O O O O O O O Relative abundance F F F F V V V V V V V V V V V V V V V M M M M type Gear F M O F M O F M O F F F F F F F F F F F F F F F F F C C F P F P F P C F type Habitat C D F P C D P S T S C D P C D F P T C D F P tomtate porkfish sergeant major sergeant Common name black grunt smallmouth grunt night sergeant french grunt dusky damselfish beaugregory threespot damselfish sailor’s choice sailor’s white grunt cocoa damselfish bluestriped grunt sheepshead sea bream spotted goatfish ground croaker reef croaker sea chub foureye butterflyfish queen angelfish gray angelfish French angelfish mullet Mayan cichlid (Bloch, 1793) (Desmarest, 1823) Cuvier, 1830 Cuvier, Linnaeus, 1758 (Müller and Troschel, 1848) Troschel, (Müller and (Linnaeus, 1758) (Lacépède, 1801) (Cuvier, 1830) (Cuvier, (Linnaeus, 1758) (Castelnau, 1855) (Müller and Troschel, 1848) Troschel, (Müller and Anisotremus virginicus (Linnaeus, 1758) virginicus Anisotremus Haemulon aurolineatum Species 1830 Haemulon bonariense Cuvier, 1859 Günther, Haemulon chrysargyreum Abudefduf saxatilis Abudefduf taurus Haemulon flavolineatum (Troschel, 1865) Stegastes adustus (Troschel, (Cuvier, 1830) (Cuvier, Stegastes planifrons Haemulon parra (Desmarest, 1823) Stegastes leucostictus Stegastes variablis Haemulon plumierii (Shaw, 1803) Haemulon sciurus (Shaw, 1792) (Walbaum, probatocephalus Archosargus Archosargus rhomboidalis (Linnaeus, 1758) rhomboidalis Archosargus (Cuvier, 1830) (Cuvier, ronchus Bairdiella Odontoscion dentex Pseudupeneus maculatus Kyphosus spp. Chaetodon capistratus Holocanthus ciliaris (Linnaeus, 1758) Pomacanthus arcuatus Pomacanthus paru (Bloch, 1787) unidentified mugilid 1862) (Günther, Cichlasoma urophthalmus Haemulidae Table 3. Continued. Table Family Pomacentridae Sparidae Sciaenidae Mullidae Kyphosidae Chaetodontidae Pomacanthidae Mugilidae Cichlidae 732 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 C C C R C C C R C C C R A O O O O O O O O O O O* O* O* O* Relative abundance V V V V V V V V V V H H H H H H H H H M M M type C M C M C M H M Gear C M O F F F F F F F F F F F F F F F F F F F F F F F S F S F T F type Habitat Spanish hogfish pallid goby Common name slippery dick bridled goby yellowhead wrasse masked goby puddingwife darter goby schooling wrasse neon goby leopard goby bluehead wrasse crested goby rainbow parrotfish striped parrotfish princess parrotfish queen parrotfish redband parrotfish redtail parrotfish bucktooth parrotfish stoplight parrotfish hairy blenny rosy blenny fat sleeper emerald sleeper guavina frillfin goby Böhlke and Robins, 1960 (Bloch, 1791) Jordan, 1904 (Robins, 1960) (Valenciennes, 1837) (Valenciennes, Bloch and Schneider, 1801 Bloch and Schneider, spp. Bodianus rufus (Linnaeus, 1758) Species Halichoeres bivittatus Halichoeres Coryphopterus glaucofraenum Gill, 1863 Coryphopterus eidolon (Valenciennes, 1839) garnoti (Valenciennes, Halichoeres (Jordon and Thompson, 1905) Coryphopterus personatus (Jordon and Halichoeres radiatus (Linnaeus, 1758) Halichoeres Ctenogobius boleosoma (Jordan and Gilbert, 1882) Halichoeres socialis Randall and Lobell, 2003 Halichoeres Elacatinus oceanops Thalassoma bifasciatum (Bloch, 1791) Lophogobius cyprinoides (Pallas, 1770) Tigrigobius saucrus Tigrigobius Cuvier, 1829 Scarus guacamaia Cuvier, Scarus iserti (Bloch, 1789) Scarus taeniopterus Desmarest, 1831 Scarus vetula 1840) (Valenciennes, Sparisoma aurofrenatum Bloch and Schneider, 1801) Sparisoma chrysopterum ( Bloch and Schneider, (Valenciennes, 1840) Sparisoma radians (Valenciennes, Sparisoma viride (Bonnaterre, 1788) Labrisomus nuchipinnis (Quoy and Gaimard, 1824) (Poey, 1868) (Poey, Malacoctenus macropus Dormitator maculatus (Bloch, 1792) (Valenciennes, 1837) smaragdus (Valenciennes, Erotelis Guavina guavina Bathygobius Labridae Table 3. Continued. Table Family Scaridae Labrisomidae Eleotridae Gobiidae TAYLOR ET AL.: MANGROVE FISH ASSEMBLAGES IN THE BELIZE CAYS 733 C R R R R R R C R O O O O O O O O O Relative abundance V V V V V V V V O H H H H H H M type Gear F M O F H M O P F F F F F F F F F F F F F F F type Habitat C D F P T C D F P C D F P T C D F P Atlantic spadefish Common name ocean surgeon doctorfish blue tang great barracuda gray triggerfish sole scrawled filefish orangespotted filefish fringed filefish pygmy filefish honeycomb cowfish spotted trunkfish buffalo trunkfish buffalo sharpnose puffer bandtail puffer checkered puffer porcupinefish Poey, 1876 Poey, (Linnaeus, 1758) Bloch and Schneider, 1801 Bloch and Schneider, (Mitchill, 1818) (Broussonet, 1782) (Bennett, 1831) Gmelin, 1789 Chaetodipterus faber Species Acanthurus bahianus Castelnau, 1855 Acanthurus chirurgus (Bloch, 1787) Acanthurus chirurgus Acanthurus coeruleus (Walbaum, 1792) Sphyraena barracuda (Walbaum, Achirus spp. Balistes capriscus Aluterus scriptus (Osbeck, 1765) Cantherhines pullus (Ranzani, 1842) Cantherhines Monacanthus ciliatus Stephanolepis setifer Acanthostracion polygonius Lactophrys bicaudalis Lactophrys trigonus (Linnaeus, 1758) Canthigaster rostrata (Bloch, 1786) Canthigaster rostrata Sphoeroides spengleri (Bloch, 1785) Sphoeroides Sphoeroides testudineus (Linneaus, 1758) Sphoeroides Diodon hystrix Linnaeus, 1758 Total Species: 142 Total Ephippidae Table 3. Continued. Table Family Acanthuridae Sphyraenidae Achiridae Balistidae Monacanthidae Ostraciidae Tetraodontidae Diodontidae Total Families: 55 Total 734 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Table 4. Means of three water quality parameters at six micro-habitats at Twin Cays, Belize, and pairwise comparison of water quality data. Group similarity indicated by similar letter designation, as determined by Tukey HSD Test.

Micro-habitat Temperature (°C) Salinity DO (mg l–1) Fringe 30.2 ± 1.1a 36.7 ± 1.4ab 4.8 ± 1.5b Transition 35.2 ± 4.4b 35.9 ± 1.3a 3.8 ± 1.4b Dwarf 34.7 ± 4.4b 40.6 ± 3.9c 5.6 ± 1.4bc Creek 32.6 ± 2.4ab 38.6 ± 2.6bc 5.6 ± 1.2bc Pond 32.6 ± 2.5ab 37.9 ± 2.3ab 7.0 ± 0.3c Sink hole 30.8 ± 1.2a 43.9 ± 7.9d 1.7 ± 1.4a

survive, as was the case in 2003 when we collected a very few Lutjanus griseus, Lut- janus apodus, and G. cinereus in the sinkhole (Table 3). Only four species, L. apodus, G. cinereus, G. yucatana, and P. orri were found across all micro-habitats. The four other gerreids taken were also notable for their presence in all but the sinkhole habitat. Another rather surprising “generalist” was Cichla- soma uropthalmus, a species normally associated with oligohaline or fresh water, found in all five micro-habitats. The high salinities encountered in sinkhole (mean = 43.9; max. = 55.0) would seem to present challenges to this species. Another species not typically recorded from mangroves or other polyhaline habitats is the pike killifish, Belonesox belizanus, which we collected in dwarf, sinkhole, and transition. Among fish species of recreational or commercial importance, we recorded 10 species of serranids, all but three by visual observation. All species were recorded as sub-adults from the fringe only, and were relatively rare. Eight species of lutjanids were documented, all mostly sub-adult and from the fringe, with only a few species in creeks/ponds. We commonly observed juveniles of Lutjanus cyanopterus, ranging in size from 20–50 cm TL, with most < 30 cm TL, within the fringe, dwarf red, ponds, and creeks. Information on the juvenile ecology of this commercially important species is scarce (Lindeman and DeMaria, 2005), but our observations reinforce the importance of mangroves as juvenile habitat. Haemulids were represented by eight species, the most common of which was Haemulon sciurus, observed/collected everywhere but the sinkhole habitat. With few exceptions, other species of grunt occurred only in the fringe. Other reef-associated species (e.g., Pomacentridae, Lab- ridae, Scaridae, etc.) were observed/collected strictly in the fringe, and most records were visual. With few exceptions, individuals of these families were sub-adult. The MDS plot (Fig. 2) of the minnow trap catches shows that micro-habitats are somewhat similar in their ichthyofauna, with the exception of creek and pond. Sink- hole, transition, fringe, and dwarf are most similar mainly due to the high abundances of cyprinodontiform fishes found in these micro-habitats. The separation of the other two micro-habitats, which also had many cyprinodontiforms, was due to the presence of unique species in the collections not found elsewhere. The lack of great distinction among Shannon diversity indices (H') probably indicates the non- quantitative nature of trapping. The lower value for pond reflects a relatively lower trapping effort here. We collected a number of K. marmoratus in cup traps, primarily in transition and sinkhole sites. However, our screen enclosures and trapping until depletion indicated the number of fish per m2 from this area at between 7 and 26, certainly making this species one of the more abundant in the transition. TAYLOR ET AL.: MANGROVE FISH ASSEMBLAGES IN THE BELIZE CAYS 735

Our data from an oceanic/island mangrove system reinforce the importance of the association between the mangal and adjacent ecosystems, and the abundance and diversity of fishes that we have documented includes juveniles of taxa commonly found on coral reefs (e.g., haemulids, lutjanids, and some serranids). Our comprehensive approach, over three different years and two seasons, with a variety of gear-types applied across a wide range of micro-habitats and combined with extensive visual observation, lends credence to the notion that most ichthyological species accounts for the mangal, mostly focused on the fringe, are undoubtedly underestimates.

Acknowledgments

We are grateful for the financial support of the Caribbean Coral Reef Ecology Program (CCRE), Smithsonian Institution, for travel and logistic support in the field. This paper is CCRE Contribution #779. Financial support, database support, and equipment for CCM was provided by USGS. In addition, further travel support was from a National Science Founda- tion grant (DEB-9981535) to I. Feller. Finally, we thank K. Kuss for data support and the Belize Fisheries Department for permission to collect in the mangal.

Literature Cited

Acosta, A. 1997. Use of multi-mesh gillnets and trammel nets to estimate fish species composition in coral reef and mangroves in the southwestern coast of Puerto Rico. Carib. J. Sci. 33: 45–57. Alongi, D. M. 2002. Present state and future of the world’s mangrove forests. Environ. Conserv. 29: 331–349. Baelde, P. 1990. Differences in the structures of fish assemblages in Thalassia testudinum beds in Guadeloupe, French West Indies, and their ecological significance. Mar. Biol. 105: 163– 173. Claro, R. and J. P. Garcia Arteaga. 1993. Fish community structure on mangroves from the In- sular Group Sabana-Camaguey, Cuba. Rev. Ecol. Oceanol. Biodivers. Trop. No. 0:60-83. ______, J. P. Garcia Arteaga, and F. Pina Amargos. 2001. Ichthyofauna of the soft bottoms of the Sabana-Camaguey Archipelago, Cuba. Rev. Invest. Mar. 22: 117–128. Collette, B. B., J. T. Williams, C. E. Thacker, and M. L. Smith. 2003. Shore fishes of Navassa Island, West Indies: a case study on the need for rotenone sampling in reef fish biodiversity studies. Aqua 6: 89–131. Clynick, B. and M. G. Chapman. 2002. Assemblages of small fish in patchy mangrove forests in Sydney Harbour. Mar. Freshw. Res. 53: 669–677. Cocheret de la Morinière, E., B. J. A. Pollux, I. Nagelkerken, and G. van der Velde. 2003. Diet shifts of Caribbean grunts (Haemulidae) and snappers (Lutjanidae) and the relation with nursery-to-coral reef migrations. Estuar. Coast. Shelf Sci. 57: 1079–1089. Dorenbosch, M., M. C. van Riel, I. Nagelkerken, and G. van der Velde. 2004. The relationship of reef fish densities to the proximity of mangrove and seagrass nurseries. Estuar. Coast. Shelf Sci. 60: 37–48. Faunce, C. H. and J. E. Serafy. 2006. Mangroves as fish habitat: 50 years of field studies. Mar. Ecol. Prog. Ser. 318: 1–18. Feller, I. C., K. L. McKee, D. F. Whigham, and J. P. O’Neill. 2002. Nitrogen vs. phosphorus limitation across an ecotonal gradient in a mangrove forest. Biogeochem. 62: 145–175. Froese, R. and D. Pauly. Editors. 2006. FishBase. World Wide Web electronic publication. June 2006. Available from: www.fishbase.org Gilmore, R. G. Jr. 1995. Environmental and biogeographic factors influencing ichthyofaunal diversity in the Indian River Lagoon. Bull. Mar. Sci. 57: 153–170. 736 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Graham, J. B. 1997. Air-breathing fishes: evolution, diversity, and adaptation. Academic Press, New York. 299 p. Laegdsgaard, P. and C. Johnson. 2001. Why do juvenile fish utilize mangrove habitats? J. Exp. Mar. Biol. Ecol. 257: 229–253. Layman, C. A. and B. R. Silliman. 2002. Preliminary survey and diet analysis of juvenile fishes of an estuarine creek on Andros Island, Bahamas. Bull. Mar. Sci. 70: 199–210. ______, D. A. Arrington, R. B. Langerhans, and B. R. Silliman. 2004. Degree of fragmentation affects fish assemblage structure in Andros Island (Bahamas) estuaries. Carib. J. Sci. 40: 232–244. Ley, J. A., C. C. McIvor, and C. L. Montague. 1999. Fishes in mangrove prop-root habitats of northeastern Florida Bay: distinct assemblages across an estuarine gradient. Estuar. Coast. Shelf Sci. 48: 701–723. Lindeman, K. C. and D. DeMaria. 2005. Juveniles of the Caribbean’s largest coral reef snapper do not use reefs. Coral Reefs 24: 359. Louis, M., C. Bouchon, and Y. Bouchon-Navaro. 1995. Spatial and temporal variations of mangrove fish assemblages in Martinique (French West Indies). Hydrobiologia 295: 275–284. Manson, F. J., N. R. Loneragan, G. A. Skilleter, and S. R. Phinn. 2005. An evaluation of the evidence for linkages between mangroves and fisheries: a synthesis of the literature and identification of research directions. Ocean. Mar. Biol. Ann. Rev. 43: 485–515 Nagelkerken, I. and G. van der Velde. 2002. Do non-estuarine mangroves harbour higher densities of juvenile fish than adjancent shallow-water and coral reef habitats in Curaçao (Neth- erland Antilles)? Mar. Ecol. Prog. Ser. 245: 191–204. Parrish, J. D. 1989. Fish communities of interacting shallow-water habitats in tropical oceanic regions. Mar. Ecol. Prog. Ser. 58: 143–160. Rooker, J. R. and G. D. Dennis. 1991. Diel, lunar and seasonal changes in a mangrove fish assemblage off southwestern Puerto Rico. Bull. Mar. Sci. 49: 684–698. Sheaves, M. 2005. Nature and consequences of biological connectivity in mangrove systems. Mar. Ecol. Prog. Ser. 302: 293–305. Sheridan, P. and C. Hays. 2003. Are mangroves nursery habitat for transient fishes and decapods? Wetlands 23: 449–458. Smith, C. L., J. C. Tyler, W. P. Davis, R. S. Jones, D. G. Smith, and C. C. Baldwin. 2003. Fishes of the Pelican Cays, Belize. Atoll Res. Bull. No. 497. 90 p. Smith-Vaniz, W. F., H. L. Jelk, and L. A. Rocha. 2006. Relevance of cryptic fishes in biodiversity assessments: a case study at Buck Island Reef National Monument, St. Coix. Bull. Mar. Sci. 79: 17–48. Starck, W. A. II. 1968. A list of fishes of Alligator Reef, Florida, with comments on the nature of the Florida reef fish fauna. Undersea Biol. 1: 4–40. Stoner, A. W. 1986. Community structure of the demersal fish species of Laguna Joyuda, Puerto Rico. Estuaries 9: 142–152. Taylor, D. S. 1990. Adaptive specializations of the cyprinodont fish Rivulus marmoratus. Fla. Sci. 53: 239–248. ______, W. P. Davis, and B. J. Turner. 2004. Groveling in the mangroves: 16 years in pursuit of the cyprinodont fish Rivulus marmoratus. Atoll Res. Bull. No. 525. 14 p. Thayer,G . W., D. R. Colby, and W. F. Hettler, Jr. 1987. Utilization of the red mangrove prop root habitat by fishes in south Florida. Mar. Ecol. Prog. Ser. 35: 25–38. Valiela, I., J. L. Bowen, and J. K. York. 2001. Mangrove forests: one of the world’s threatened major tropical environments. BioScience 51: 807–815. Verweij, M. C., I. Nagelkerken, D. de Graaff, M. Peeters, E. J. Bakker, andG . van der Velde. 2006. Structure, food and shade attract juvenile coral reef fish to mangrove and seagrass habitats: a field experiment. Mar. Ecol. Prog. Ser. 306: 257–268. Wilcox, L. V., T. G. Yocom, R. C. Goodrich, and A. M. Forbes. 1975. Ecology of mangroves in the Jewfish Chain, Exuma, Bahamas. Pages 305–343in G. Walsh, S. Snedaker, and H. Teas, eds. Proc. Int. Symp. on Biology and Management of Mangroves. IFAS, Gainesville. TAYLOR ET AL.: MANGROVE FISH ASSEMBLAGES IN THE BELIZE CAYS 737

Woodroffe, C. D. 1995. Mangrove vegetation of Tobacco Range and nearby mangrove ranges, central Belize barrier reef. Atoll Res. Bull. No. 427. 35 p. World Resources Institute. 1996. World Resources 1996-97: the urban environment. The World Resources Institute, NEP, NEDP, Work Bank. Oxford Univ. Press, New York. 400 p.

Addresses: (D.S.T.) Brevard County Environmentally Endangered Lands Program, 91 East Dr., Melbourne, Florida 32904. (E.A.R.) Dynamac Corporation, Mail Code: Dyn-2, Kennedy Space Center, Florida 32899. (C.C.M.) U.S. Geological Survey, Florida Integrated Science Cen- ter, Center for Coastal and Watershed Studies, St. Petersburg, Florida 33701. (W.P.D.) PO Box 607, Quincy, Florida 32353. Corresponding Author: (D.S.T.) E-mail: . BULLETIN OF MARINE SCIENCE, 80(3): 739–754, 2007

NOTE

Spatial differences and seasonal cyclicity in the intertidal fish fauna from four mangrove creeks in a salinity zone of the Curuçá estuary, North Brazil

Tommaso Giarrizzo and Uwe Krumme

Intertidal creeks are vital component of the mangrove forest. They serve as hydrological connectors between the subtidal area and the mangrove interior and provide important temporary foraging grounds for juvenile fishes. In the large estuarine mangrove areas of the tropics, intertidal creeks can be found in different zones of salinity ranging from low salinity environments at the head, to high salinity environments at the mouth of the estuary. Mangrove creeks in different estuarine salinity zones may vary in their importance as nurseries for juvenile fishes. To test for such a longitudinal relationship, however, variability in fish habitat use in intertidal creeks located in the same salinity zone must be better understood to avoid attributing creek variations to salinity dynamics. Typically, most fish fauna studies from tropical intertidal mangrove creeks compare this habitat with other estuarine habitats (e.g., Blaber et al., 1989; Robertson and Duke, 1990; Leh and Sasekumar, 1991; Sasekumar et al., 1992; Ikejima et al., 2003), or only analyze temporal patterns (e.g., Wright, 1986; Batista and Rêgo, 1996; Laroche et al., 1997). Studies on spatial variation of tropical intertidal fish assemblages are less common and have been usually either carried out on a small scale [< 2 km between sites, e.g., Barletta et al. (2003) and Krumme et al. (2004)], or on a large scale [e.g., about 20 km between sites, Vidy et al. (2004)]. Information is lacking on the inter-creek variability of fish habitat use on a medium spatial scale (2–20 km), where the estuarine salinity zones are usually located. To better understand medium scale differences in juvenile fish habitat use, we studied the fish assemblages of four intertidal mangrove creeks separated by just 4 km and located in the same salinity zone. The study was carried out in the macrotidal Curuçá estuary, a conservation unit located close (50 km) to the southern channel of the Amazon mouth (Fig. 1). The specific objectives of our study were to (1) describe the composition and seasonal changes of the intertidal fish assemblages, and (2) compare density and biomass of juvenile fishes (as a measure of habitat quality) among the four creeks located in the same salinity zone.

Material and Methods

Study Area.—The study area is located in the estuary of the Curuçá River near the city of Curuçá (0°10´S, 47°50´W) and town of Abade, in Pará, north Brazil (Fig. 1), and is part of the second largest continuous mangrove area of the world. The Curuçá estuary is a marine- dominated system formed by the confluence of the Curuçá River and the Muriá channel, a natural tidal connection with the São Caetano de Odivelas estuary to the west (Fig. 1). The perimeter, length, and area of the Curuçá estuary is 73 km, 21 km, and 200 km2, respectively.

Figure 1. (A) Coast of Pará and Maranhão states covering the second largest mangrove area in the world, north Brazil, and location of the Curuçá estuary at the eastern tip of the mouth of the southern channel of the Amazon delta (Pará River); (B) Location of the four first-order intertidal mangrove creeks (D–F) in the Curuçá estuary.

It is covered by approximately 116 km2 of mangrove forest dominated by Rhizophora mangle Linnaeus mixed with Avicennia germinans (Linnaeus) Stearn. Mean air temperature is 27 °C and mean annual rainfall is 2526 mm (ANA, 2005; n = 16 yrs, range: 1085–3647 mm), with the wet season from January to June and the dry season from July to December. The mean depth of the estuary at slack low water is 3 m, with depres- sions of up to 8 m near Abade. A complex network of branching intertidal mangrove creeks is flooded twice daily by the semidiurnal tide. The tidal range is 3–4 m at neap tides and 4–5 m at spring tides. Sampling and Catch Analysis.—The study was conducted in four intertidal first-order mangrove creeks in the upper and middle reaches of the Curuçá estuary: (1) the Muriá channel near Abade, (2) the western side of the middle estuary, (3) the eastern side of the middle estuary where the Curuçá River meets the bay, and (4) the uppermost creek in the study, close to Curuçá (Fig. 1). We define the stream order as follows: no order is assigned to the subtidal section, intertidal creeks draining into the subtidal section are first-order creeks. The four creeks were all dead-ending and had no connection to other creek systems at neap tide. These creeks were sampled every second month between September 2003 and July 2004 during the waxing of the moon (neap tide), with a fyke net. Over a sampling period of four consecutive days each second month, each creek was sampled once. This yielded a total of 24 samples (6 samples from each creek). The fyke net was set at the mouth of the creeks at daytime slack high water H( W). It consisted of two wings (20 × 6 m, 20 mm stretched mesh size) and a fyke net body (seven circular stainless steel hoops; total length: 7.5 m; 13 mm stretched meshed size until the 3rd hoop, 5 NOTES 741

Table 1. Summary of landscape features of the four first-order intertidal mangrove creeks in the Curuçá estuary, Pará, north Brazil, at neap tides. The presence of different creek features is described subjectively using: blank = absence, + = low, ++ = medium, +++ = high.

Creek Physical features 1 2 3 4 Inundated area (m2) 9,224 4,685 2,598 20,000 Perimeter of inundated area (m) 978 795 460 777 Creek length (m) 263 336 157 277 Width at the mouth at slack HW (m) 17 17 18 20 Subtidal low tide depth near the creek (m) 8 3 2.5 1.3 Outcropping bedrock ++ ++ Soft, muddy substrate ++ +++ ++ + Silt bank in front of the creeks’ mouth ++ + Branches of fallen trees in the creek ++ ++ +++ + Scarps + +++ Habitat complexity + + ++ ++ mm from the 3rd to the innermost hoop) connected by an inlet funnel. The bottom line of the wings and the inlet funnel was pushed into the mud and the headline was fixed about 1 m above HW level using poles of mangrove. The fish were collected from the fyke net body during ebb tide until total drainage of the creek. Fish were kept on ice and transported to the laboratory. Specimens were identified to the lowest possible taxonomic level. Fish taxa were assigned to five ecological guilds following McHugh (1967) and to six feeding categories according to published literature and our own stomach content inspections. Gonad inspection was used to distinguish juveniles from adults. The topography of the creeks was surveyed GP( S, compass, tape measure) and a GIS map for each creek was generated (Fig. 1C–F). Several physical characteristics of the four creeks are given in Table 1. The inundated area (m2) for a medium neap tide event was quantified for each creek using ArcGIS. Thus, catch weights and abundance were standardized to density (fish –2m ) and biomass (g m–2), respectively. For each sampling event, water surface salinity, temperature, pH, Secchi depth, seston, and dissolved oxygen concentrations were recorded. Data Analysis.—Differences in environmental parameters (see above) and fish assemblage variables [species richness, species/family (S/F) ratio, Shannon-Weaver diversity H´

(log10), evenness (J´), density, and biomass] were compared using 2-factor analyses of variance (ANOVA). On the species level, only fishes with 100% frequency of occurrence or n ≥ 850 were analyzed. The experimentwise error was set to 0.05 using the Bonferroni approach. Season (dry season: September and November 2003 and July 2004; wet season: January, March and May 2004) and site (creek 1, 2, 3, 4) were treated as fixed factors. Each treatment combination had three pseudoreplicates (e.g., dry-season creek 1: September and November 2003 and July 2004). If necessary, log(x+1) or 4th-root transformations were performed to fulfill the ANOVA assumptions of normal distribution (Kolmogorov-Smirnov test) and homoscedasticity (Co- chran test). When a significant difference for main effects was detectedP ( < 0.05), Tukey’s multiple comparison test was used to identify individual group differences.V ariables that did not meet the assumptions were tested using the non-parametric Kruskal-Wallis test (K-W). The similarities in fish species composition among the samples were assessed using cluster analysis (complete linkage) and non-metric multidimensional scaling (MDS) (PRIMER package; Clarke and Warwick, 1994). The original data matrix containing the total catch weight per sample and of all species was square root transformed to generate the Bray-Curtis similarity matrix. The density of the dominant fish families was superimposed on the MDS. Two-way crossed analysis of similarities (ANOSIM) (Clarke, 1993) was used to test whether the species compositions differed significantly between season (dry and wet season) and site (creeks 1–4). Multivariate dispersion (MVDISP) was used to describe the relative multivariate 742 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Figure 2. Bimonthly variation in environmental parameters in four intertidal first-order mangrove creeks of the Curuçá estuary, north Brazil, in 2003/2004. Creek location: refer to Figure 1.

variability on the MDS plot between dry and wet season and between the four creeks (Clarke and Warwick, 1994). The RELATE routine was used to test whether bimonthly changes in the ichthyofauna composition were related to a seasonal cycle, i.e., if the assemblage returned to the initial structure.

Results

Environmental Parameters.—Site had no significant effect on any of the environmental parameters (Fig. 2). Seasonal effects were detected for salinity K( -W: P < 0.001), water temperature (K-W: P < 0.01), and pH (ANOVA: P < 0.001) (Fig. 2A–C). The sampling area was hypersaline in November 2003 (40.8; dry season). Through- out the wet season (January–May), salinities were < 15 (minimum: 6.1). Fish Fauna Composition.—In total, 56,090 individuals (745.4 kg) belonging to 65 species of 29 families were caught (Table 2). The dominant families in terms of total catch weight were Tetraodontidae (52%), Ariidae (19%), and Engraulidae (13%). Six species accounted for 84% of the total catch weight: Colomesus psittacus (51%), Sciades herzbergii (14%), Anchovia clupeoides (7%), Cathorops sp. (5%), Cetengraulis edentulus (4%), and Mugil curema (3%). Mean fish density and biomass of all species (± SD) were 0.3 fish –2m ± 0.3 (range: 0.1–1.0) and 6.0 g m–2 ± 7.1 (range: 0.8–25.0), respectively. Ninety percent of all fish caught were juveniles. Thirty-one species (48% of all species) were represented only by juveniles, 30 species (46%) by both juveniles and adults, and four species (6%) by adults only (Astianax sp., Cathorops spixii, My- rophis punctatus, and Tomeurus gracilis). Mean fish length was 7.4 cm ± 4.5 SD. Mean fish weight was 13.3 g ± 37.7 DS . In terms of catch weight, the ecological guilds NOTES 743

were dominated by estuarine species (82%) (Table 2). Marine species, estuarine-marine species, freshwater fishes and occasional marine visitors accounted for 9, 8, < 1, and < 0.1%, respectively. The feeding categories were dominated by benthophage fishes (77%). Herbivorous, benthoichthyophage, zooplanktivorous, lepidophage, and ichthyophage species accounted for 18, 3, 1, 0.7, and 0.3%, respectively. Spatio-Temporal Patterns in Fish Fauna Composition.—Univariate analysis of spatio-temporal changes in the fish fauna indicate that the number of fish species (mean: 26 ± 4 SD) and the S/F ratio (mean: 1.8 ± 0.2 SD) were not influenced by either season or site (Table 3). H′ and J′ were significantly higher in the wet season but did not differ among sites. The most important result was that fish density and biomass did not differ significantly between seasons, but did among sites, with 2.0 and 3.7 times higher values in creeks 2 and 3 than in creeks 1 and 4, respectively (Fig. 3A,B). Creeks 1 and 4 were several times larger in size than creeks 2 and 3 (Table 1) but had lower density and biomass. Consequently, fish density and biomass were not proportional to creek size. On the species level, significantly higher values in the wet than in the dry season were recorded for density of Cathorops sp. and S. herzbergii (Table 3). In contrast, higher dry season values were found for density of Mugil sp. Differences among sites were statistically significant for density ofC. psittacus, Genyatremus luteus, and Oli- goplites saurus with higher values in the outer creeks 2 and 3. Cluster analysis separated the samples into four groups according to an interaction of the factors season and site (A–D in Fig. 4A). Group A contained only two dry season samples in the outer creek 3. Group B consisted of dry season samples, primarily from the creeks 1 and 2. Group C was comprised of wet season samples of the outer creeks (creeks 2 and 3). The largest group D contained mainly wet season samples, especially from the innermost creek 4. The combination of season and site had a significant effect on changes in the structure of the fish fauna compositions in the intertidal mangrove creeks (two-way cross-ANOSIM; season: global R = 0.30, P < 0.05; creek: global R = 0.22, P < 0.05) as displayed in the MDS ordination (Fig. 4B). There was a diagonal gradient in Figure 4B regarding the catch weights, with the highest values on the lower left (group C: wet season, outer creek) and lowest values on the upper right of the MDS plot (group A/B: dry season). Wet season samples had a lower dispersion value than dry season samples (MVDISP: 0.756 vs 1.244). Wet season samples were more tightly clustered, indicating more uniform fish fauna assemblages whereas dry season samples formed a widely dispersed group in the upper region of the plot, pointing to more distinct fish fauna compositions among the creeks, especially in September (Fig. 4B). Superimposing the main fish family densities on Fig. 4B showed that the MDS group C (outer creeks, wet season, highest catch weights) was characterized by highest densities of Tetraodontidae, Ariidae, Sciaenidae (Fig. 4C–E), and Haemulidae (not shown). The MDS group A/B (dry season, lowest catch weights) was typified by highest densities of Engraulidae that occurred mainly in September (Fig. 4E). However, while fish densities of the September samples in the outer creeks 2 and 3 were dominated by juveniles of A. clupeoides (Sep2 and Sep3), Anchoa hepsetus dominated in the innermost creek 4 (Sep4). The high density in May and July in creek 2 was due to adult C. edentulus and juvenile A. hepsetus, respectively. Furthermore, Engraulidae occurred in the inner creek 1 only with lowest densities. The group D was intermediate because none of the dominant fish families had their density or biomass peak 744 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 8.5 0.4 3.2 5.3 D 17.4 23.5 79.0 55.9 70.1 47.7 61.9 34.9 25.1 23.8 112.1 137.3 768.8 103.9 590.4 360.4 224.1 1,769.7 1,320 1.8 463.9 606.8 834.8 257.3 180.4 245.0 593.4 582.2 313.7 BM 6,111.1 1,258.3 2,559.4 6,445.8 1,266.1 4,707.6 3,194.5 1,146.9 8,347.9 3,127.7 1,158.7 1,012.7 Creek 2 72,071.4 8 3 9 5 6 2 7 1 4 Rk 11 19 10 18 44 15 14 20 25 22 12 16 17 13 4.9 0.7 2.9 5.8 0.7 5.1 D 11.9 11.7 10.1 14.5 12.5 17.9 74.5 24.8 16.1 44.3 22.0 29.3 14.6 165.5 214.8 181.6 144.0 43.9 47.1 70.3 65.6 54.6 263.7 296.3 312.0 366.0 470.1 973.5 801.8 590.4 554.1 123.0 255.4 255.3 226.5 138.2 BM 1,065.5 1,521.7 4,454.2 Creek 1 10,055.8 9 6 7 8 3 4 5 1 2 11 13 12 10 23 22 19 20 21 15 16 17 18 14 Rk ) of all fish species collected with a fyke net from four intertidal first- –1 L Z Z B B B B B B B B H H H H H H H B/I B/I B/I B/I B/I TC F M M M M M TE TE TE TE TE TE TE TE TE EG EM EM EM EM EM EM EM EM ), mean density (D, fish ha –1 (Linnaeus, 1766) (Fowler, 1911) (Fowler, (Agassiz, 1829) (Cuvier, 1829) (Cuvier, (Swainson, 1839) (Linnaeus, 1758) (Lacépède, 1801) (Steindachner, 1864) (Steindachner, (Jordan, 1889) Hancock, 1830 Cynoscion acoupa (Bloch and Schneider, 1801) Lutjanus jocu (Bloch and Schneider, Genyatremus luteus (Bloch, 1790) Genyatremus (Bloch and Schneider, 1801) Oligoplites saurus (Bloch and Schneider, Anableps anableps Lycengraulis grossidens Lycengraulis Cetengraulis edentulus Anchoviella lepidentostole sp. Cathorops (Steindachner, 1879) amazonica (Steindachner, Rhinosardinia Mugil sp. Stellifer microps Centropomus undecimalis (Bloch, 1792) Centropomus Anchovia clupeoides Valenciennes, 1836 Valenciennes, Mugil curema Stellifer naso Pterengraulis atherinoides Pterengraulis Mugil incilis testudineus (Linnaeus, 1758) Sphoeroides Diapterus auratus Ranzani, 1842 (Bloch and Schneider, 1801) Colomesus psittacus (Bloch and Schneider, (Bloch, 1794) Sciades herzbergii Poey, 1860 pectinatus Poey, Centropomus Species order mangrove creeks at diurnal neap tides between September 2003 and July 2004 in the Pará, Curuçá north estuary, Brazil. Ecological guild (EG) according to McHugh (1967): truly estuarine resident species (TE), marine occasional visitors (MO), marine species (M), estuarine-marine species (EM), and freshwater I: herbivorous, H: benthophage, B: inspections: stomach own and (2005) Pauly and Froese (2004), al. et Krumme to according (TC) category Trophic (F). fishes ichthyophage, B/I: benthophage-ichthyophage, Z: zooplanktivorous, L: lepidophage. Mean total length (TL) (cm) and the size range is given for each species. Table 2. Table Rank by biomass (Rk), mean biomass (BM, g ha NOTES 745

TL 4.5 (2–15.6) 7.7 (2.7–26) 15.6 (3–33.1) 8.4 (2.1–23.3) 6.2 (1.1–19.6) 7.4 (2.2–13.8) 8.2 (2.3–23.5) 5.9 (2.4–18.2) 8.7 (2.7–17.5) 9.8 (2.2–18.3) 11.4 (3.0–24.6) 11.4 10.0 (2.3–20.4) 10.6 (2.6–27.5) 14.8 (2.9–31.5) 10.5 (3.0–18.8) 15.8 (1.7–26.5) 12.0 (4.3–17.6) 18.8 (4.1–36.5) 6.0 (1.58–17.3) 13.3 (2.7–35.1) 24.8 (4.8–38.5) 10.7 (0.4–32.6) 12.6 (3.4–21.6) 3.1 3.0 6.3 2.2 3.9 9.2 7.5 0.6 0.3 D 17.0 10.3 21.5 26.8 18.3 32.6 21.9 21.8 16.8 76.5 145.1 188.7 862.3 130.8 6.2 39.2 27.3 60.4 63.3 21.9 83.5 66.6 43.7 16.6 BM 127.8 158.0 123.2 274.5 378.3 146.2 277.5 220.2 361.4 Creek 4 1,253.7 1,517.3 3,058.0 3,891.3 9 6 4 8 3 7 2 1 Rk 11 21 24 18 13 14 17 25 15 16 20 12 26 10 28

6.4 7.7 8.3 0.6 1.3 25.0 12.2 99.4 75.7 45.6 38.2 28.9 13.5 26.9 39.1 16.0 D 183.0 521.5 426.0 394.6 359.9 1,097.1 1,086.9 37.1 10.61 211.7 127.0 452.1 273.9 708.6 388.7 498.0 216.0 287.4 941.0 433.7 775.4 589.1 271.6 BM 1,415.8 1,625.1 2,564.6 2,973.0 3,227.6 16,969.7 47,789.9 Creek 3 7 6 5 8 4 9 3 2 1 11 25 14 21 10 18 13 24 23 20 32 15 35 22 Rk (Linnaeus, 1766) (Fowler, 1911) (Fowler, (Agassiz, 1829) (Cuvier, 1829) (Cuvier, (Swainson, 1839) (Linnaeus, 1758) (Lacépède, 1801) (Steindachner, 1864) (Steindachner, (Jordan, 1889) Hancock, 1830 Pterengraulis atherinoides Pterengraulis Poey, 1860 pectinatus Poey, Centropomus Cynoscion acoupa (Bloch and Schneider, 1801) Lutjanus jocu (Bloch and Schneider, Genyatremus luteus (Bloch, 1790) Genyatremus Anableps anableps (Bloch and Schneider, 1801) Oligoplites saurus (Bloch and Schneider, Stellifer microps Mugil sp. Anchoviella lepidentostole (Steindachner, 1879) amazonica (Steindachner, Rhinosardinia Cetengraulis edentulus sp. Cathorops Stellifer naso Lycengraulis grossidens Lycengraulis Valenciennes, 1836 Valenciennes, Mugil curema Centropomus undecimalis (Bloch, 1792) Centropomus Diapterus auratus Ranzani, 1842 Anchovia clupeoides Sphoeroides testudineus (Linnaeus, 1758) Sphoeroides (Bloch, 1794) Sciades herzbergii Mugil incilis Table 2. Continued. Table 1801) Colomesus psittacus (Bloch and Schneider, Species 746 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 0.4 0.4 5.7 1.8 0.4 3.6 0.4 4.6 6.8 1.1 1.1 1.4 0.4 1.4 6.8 0.4 0.4 3.6 5.3 3.9 1.8 D 477.1 6.6 6.3 0.1 4.9 3.5 13.2 29.7 12.0 27.5 98.4 49.3 34.3 18.6 17.9 12.9 43.8 10.2 247.0 238.6 205.4 179.5 157.1 BM Creek 2 Rk 39 40 36 32 48 38 33 21 28 23 24 26 27 29 31 34 35 37 41 42 30 45 0.2 0.5 0.4 0.2 3.8 0.2 0.4 2.3 0.2 0.2 2.0 0.5 0.2 0.2 0.2 0.2 D 36.5 2.9 1.6 7.1 2.0 9.7 8.5 7.4 0.6 0.5 0.1 0.6 15.5 16.2 31.9 18.7 16.7 36.2 BM Creek 1 36 34 35 30 31 32 33 28 29 26 27 25 24 39 40 37 38 Rk I I Z Z Z Z Z Z Z Z B B B B B B B B B B B B/l B/I B/I B/I B/I TC F F F F M TE TE TE TE TE TE TE TE TE TE EG EM EM EM EM EM EM EM EM EM EM MO (Linnaeus, 1766) (Linnaeus, 1766) Günther, 1862 Günther, (Quoy and Gaimard, 1825) (Vaillant and Bocourt, 1883) (Vaillant (Valenciennes, 1847) (Valenciennes, (Broussonet, 1782) (Günther, 1868) (Günther, (Lichtenstein, 1822) Linnaeus, 1758 Bloch and Schneider, 1801 Bloch and Schneider, (Linnaeus, 1758) (Linnaeus, 1758) (Linnaeus, 1758) sp. sp. Lycengraulis batesii Lycengraulis Echeneis naucrates Chaetodipterus faber Cynoscion jamaicensis Stellifer Pseudauchenipterus nodosus (Bloch, 1794) Hyporhamphus roberti Hyporhamphus Astianax Citharichthys spilopterus (Bleeker, 1863) Eucinostomus melanopterus (Bleeker, (Bloch and Schneider, 1801) Batrachoides surinamensis (Bloch and Schneider, (Cuvier, 1830) (Cuvier, Cynoscion microlepidotus (Walbaum, 1792) timucu (Walbaum, Strongylura Achirus lineatus (Ranzani, 1840) Ophichthus cylindroideus spixii (Agassiz, 1829) Cathorops Atherinella brasiliensis Lagocephalus laevigatus Achirus achirus (Linnaeus, 1758) 1837) Ctenogobius smaragdus (Valenciennes, Poecilia vivípara Anchoa hepsetus Selene vomer Chloroscombrus chrysurus Chloroscombrus Epinephelus itajara Species Table 2. Continued. Table Myrophis punctatus Lütken, 1852 Myrophis NOTES 747

TL 10.9 18.9 13.2 12.8 4.4 (3.7–5) 4.1 (2.0–15) 8.8 (3–15.1) 2.9 (1.2–5.2) 7.2 (6.2–7.8) 6.4 (3.2–9.3) 7.0 (1.9–15.7) 6.7 (2.7–15.2) 9.6 (6.2–16.2) 7.9 (3.0–13.1) 6.5 (3.0–10.2) 5.4 (2.2–14.5) 12.7 (11.2–15) 18.9 (4.5–32.1) 14.7 (8.2–24.1) 10.1 (3.2–15.5) 30.0 (3.3–49.2) 21.7 (9.5–37.8) 27.1 (3.1–51.0) 10.9 (10.5–11.2) 15.2 (12.5–17.3) 19.9 (16.6–22.1) 1.7 1.1 0.2 0.3 0.3 1.1 0.1 5.1 0.1 0.3 0.1 0.1 0.1 D 599.5 4.2 4.9 0.8 2.8 3.4 0.8 0.5 0.2 0.2 10.8 51.0 33.3 <0.1 BM 426.2 Creek 4 5 27 31 30 33 32 37 38 19 22 39 40 41 45 Rk

4.5 0.6 1.3 1.9 9.6 0.6 0.6 0.6 6.4 0.6 0.6 1.3 5.8 16.0 54.1 D 121.3 4.5 0.8 6.0 8.9 0.5 11.3 40.0 52.0 62.8 57.6 90.3 28.5 117.4 402.0 296.6 513.3 BM Creek 3 31 40 44 17 30 28 36 38 19 37 29 27 12 33 45 26 Rk (Linnaeus, 1766) (Linnaeus, 1766) Günther, 1862 Günther, (Quoy and Gaimard, 1825) (Vaillant and Bocourt, 1883) (Vaillant (Valenciennes, 1847) (Valenciennes, (Broussonet, 1782) (Günther, 1868) (Günther, (Lichtenstein, 1822) Linnaeus, 1758 Bloch and Schneider, 1801 Bloch and Schneider, (Linnaeus, 1758) (Linnaeus, 1758) (Linnaeus, 1758) sp. sp. Echeneis naucrates punctatus Lütken, 1852 Myrophis Selene vomer chrysurus Chloroscombrus Epinephelus itajara Achirus lineatus (Ranzani, 1840) Ophichthus cylindroideus Cynoscion jamaicensis Pseudauchenipterus nodosus (Bloch, 1794) roberti Hyporhamphus Astianax batesii Lycengraulis Chaetodipterus faber (Walbaum, 1792) timucu (Walbaum, Strongylura 1863) Eucinostomus melanopterus (Bleeker, 1801) Batrachoides surinamensis (Bloch and Schneider, 1830) (Cuvier, Cynoscion microlepidotus Citharichthys spilopterus Stellifer Achirus achirus (Linnaeus, 1758) 1837) Ctenogobius smaragdus (Valenciennes, Poecilia vivípara Lagocephalus laevigatus Cathorops spixii (Agassiz, 1829) Cathorops Species Table 2. Continued. Table Anchoa hepsetus Atherinella brasiliensis 748 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

0.4 0.4 0.4 0.4 0.4 D 19.9 0.1 0.2 0.1 2.0 0.4 51 25 < 0.1 331.3 BM 5,854.4 Creek 2 16,455 117,898.9

Rk 50 51 47 49 43 46

D 40 21 128.3 BM 5,635 1,018.2 Creek 1 23,180.8

Rk I

L Z B B B B B B B B H H B/I B/I B/I TC

F F F M TE TE TE TE TE TE TE EM EM EM EM EM EG (Valenciennes, 1840) (Valenciennes, (Eigenmann, 1912) (Pallas, 1770) (Cuvier, 1832) (Cuvier, ) ) –1 –1 (Jordan, 1889) (Bloch, 1790) Gobiosoma hemigymnum (Eigenmann and Eigenmann, 1888) Hypostomus plecostomus (Linnaeus, 1758) 1830) (Cuvier, Cynoscion leiarchus Stellifer rastrifer Lacépède, 1800 Odontognathus mucronatus Oligoplites palometa Gobionellus oceanicus Ctenogobius sp. Agassiz, 1831 Caranx latus 1801) Symphurus plagusia (Bloch and Schneider, lepturus Linnaeus, 1758 Trichiurus Number of species Number of families catch weight (kg) Total Total number of individuals Total Overall BM (g ha Aspredinichthys tibicen Aspredinichthys Anchoviella guianensis Steindachner, 1876 Steindachner, Thalassophryne nattereri Stellifer stellifer Tomeurus gracilis Eigenmann, 1909 Tomeurus Overall D (fish ha Species Table 2. Continued. Table NOTES 749

TL 2.4 3.3 5.0 7.6 5.8

9.3 (7–12) 4.2 (3.2–5.8) 4.9 (4.4–5.9) 7.8 (7.0–8.5) 3.6 (3.4–3.7) 4.3 (3.0–5.5) 2.2 (1.6–2.6) 9.6 (2.7–16.5) 9.7 (4.5–14.2) 23.3 (19.0–34.5) 49.6 (38.5–68.0)

3.9 0.3 0.1 0.1 0.1 0.3 0.1 0.3 D 1.9 2.5 0.2 2.1 4.9 0.1 45 22 29.9 < 0.1 BM 153.6 Creek 4 2,241.2 26,895 12,796.6

34 42 44 35 29 43 23 36 Rk

3.2 2.6 5.8 0.6 0.6 0.6 0.6 1.3 3.8 2.6 D 4.6 0.9 0.2 0.1 0.1 1.7 1.8 21.4 49 23 < 0.1 409.3 132.4 BM 7,150 4,541.2 84,927.0 Creek 3

16 34 39 43 46 47 48 42 41 49 Rk (Valenciennes, 1840) (Valenciennes, (Eigenmann, 1912) (Pallas, 1770) (Cuvier, 1832) (Cuvier, ) ) –1 –1 (Jordan, 1889) (Bloch, 1790) Gobiosoma hemigymnum (Eigenmann and Eigenmann, 1888) Hypostomus plecostomus (Linnaeus, 1758) 1830) (Cuvier, Cynoscion leiarchus Stellifer rastrifer Lacépède, 1800 Odontognathus mucronatus Oligoplites palometa Gobionellus oceanicus Ctenogobius sp. Agassiz, 1831 Caranx latus 1801) Symphurus plagusia (Bloch and Schneider, lepturus Linnaeus, 1758 Trichiurus Number of species Number of families catch weight (kg) Total number of individuals Total Overall BM (g ha Species Table 2. Continued. Table 1876 Steindachner, Thalassophryne nattereri Stellifer stellifer gracilis Eigenmann, 1909 Tomeurus tibicen Aspredinichthys Anchoviella guianensis Overall D (fish ha 750 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Table 3. Summary of results of two-factor analysis of variance (site and season) and Kruskal-Wallis tests for six fish assemblage variables and 10 fish species collected in four first-order intertidal mangrove creeks in the Curuçá estuary, Pará, north Brazil, using a fyke net. The experimentwise error was set to 0.05 using the Bonferroni approach. Significant critical values (F-values in case of ANOVA; H in case of Kruskal-Wallis test) are indicated by an asterisk. W: wet season; D: dry season. df: degrees of freedom. In multiple post-hoc comparisons (Tukey’s honest significant difference (HSD) method) significantly different pairs are indicated by the same font style.

Main effect Interaction Tukey HSD test Creek (1) Season (2) (1) × (2) (1) (2) (1) × (2) Variable (df 3, 16) (df 1, 16) (df 3, 16) Species richnessa 0.85 1.94 1.38 Species/Familya 2.16 5.14 2.98 a Diversity Hʹ (log10) 3.84 25.08* 4.10 W > D Evenness Jʹa 5.47 19.17* 4.57 W > D Density (fish –2m )b 8.29* 3.78 1.23 2 > 3 > 4 > 1 Biomass (g m–2)c 13.13* 3.20 Density Anchoa hepsetusc 3.33 3.13 Anchovia clupeoidesc 5.91 3.41 Cathorops sp.b 0.59 33.44* 0.27 W > D Cetengraulis edentulusc 2.12 1.60 Colomesus psittacusb 13.41* 0.29 0.39 2 > (3) > 1 > (4) Genyatremus luteusb 8.65* 3.6 0.34 2 > (3) > (4) >1 Sciades herzbergiib 0.75 10.63* 2.29 W > D Mugil sp..b 3.55 13.18* 0.18 2 > 3 > 1 > 4 D > W Oligoplites saurusc 13.08* 0.36 Rhinosardinia amazonicac 2.05 3.43 a Two-way ANOVA, log(x+1) transformation b Two-way ANOVA, 4th-root transformation c Kruskal-Wallis test in these samples. Superimposing salinity (not shown) confirmed the dry-wet season alternation but did not help to explain the spatial patterns. The results of the RELATE routine showed that bimonthly changes in the fish fauna composition over all samples followed a cyclical (annual) pattern (sample sta- tistic: 0.1, P < 0.01). To track the site-specific changes in the fish fauna throughout the year, the samples were ordinated separately for each creek (Fig. 5). Starting with September 2003, the samples followed a cyclic change, anticlockwise in creeks 1 and 3 (Fig. 5A,C) and clockwise in creek 2 (Fig. 5B). In the innermost creek 4 with the lowest dispersion value, all samples except the September sample clustered together, however, the January–July samples still followed a cyclic pattern (Fig. 5D).

Discussion

Temporal Patterns.—Similar seasonal changes in fish assemblages occurred in all creeks. Cyclic changes in the intertidal fish fauna compositions did not affect overall density and biomass throughout the year. Similar numbers and standing stock were apparently maintained year-round. This suggests that increases in abundance or weight in some species were compensated by reductions in others. For NOTES 751

Figure 3. Mean density (A) and biomass (B) (± SD) of the intertidal mangrove fish fauna captured at diurnal neap tides between September 2003 and July 2004 from four intertidal mangrove creeks at slack high water, Curuçá estuary, Pará. Bimonthly sampling yielded n = 6 samples per creek. instance, Mugil sp. dominated both in density and biomass in the dry season while Ariidae dominated in the wet season. Spatial Patterns.—The outer creeks 2 and 3 had higher fish biomass and density year-round without a site × season interaction. This pattern was driven by four dominant fish families. Consequently, the two outer creeks per se were associated with higher abundances of juvenile fishes per unit of area. The creeks were located in a zone of homogenous salinity and none of the environmental parameters recorded could explain spatial differences between the creeks’ fish assemblages. Creek size is another possible factor as suggested by results of Barletta et al. (2003), Barletta-Ber- gan et al. (2002), and Paterson and Whitfield (2003). However, differences between our creeks were not related to creek size because the largest creek 4 (2–8 times larger than the other creeks; Table 1) did not have the highest densities. Thus, it seems other factors might have a role in the significant differences between the fish fauna compositions of intertidal mangrove creeks. Paterson and Whitfield (2003) have suggested that topographical characteristics and not physico-chemical variables (temperature, salinity, turbidity) explain the distribution patterns of fish in three intertidal salt marsh creeks in a South African estuary. We propose that landscape factors may be important in structuring intertidal fish assemblages of mangrove creeks. In our system, landscape features that likely play an important role are the position relative to the ocean or to the mainland, irrespective of salinity (e.g., location at the edge or in the center of a mangrove area); proximity to an extensive subtidal resting area where the intertidal fishes may spend the low tide period; topography at the mouth of a creek that allows the unrestricted inflow of the flood tide and the immigration of fish species; and structural complexity and topography (e.g., canopy cover, ramifications, inundation period). For example, creeks 1 and 4 were characterized by natural obstacles at the mouth (extensive sand bank and rocks, respectively), and closer proximity to the mainland. Moreover, the subtidal area in front of creek 1 was characterized by a deep depression (8 m at low tide) and bedrock as bottom substrate. Fishermen reported that adult predatory fishes can be caught in the Muriá channel close to the entrance of creek 1. Where mangrove, seagrass, and coral reef biotopes co-occur, strong connectedness of their fish fauna often exists. Many studies compare samples from different biotopes under the assumption that they are representative for the particular shallow-water biotope (e.g., Nagelkerken et al., 2000). However, prior to this type 752 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Figure 4. Ordination of total catch weight data for all fish species collected bimonthly between September 2003 and July 2004 from four intertidal mangrove creeks. (A) Cluster dendrogram, (B) MDS plot. Open symbols: dry season. Black symbols: wet season. The number inside the symbol in the MDS plot refers to the month sampled. (C–F) The same MDS as in (B), but with superimposed circles, sized according to changes in density of four dominant fish families.

Figure 5. MDS plot of samples for each creek (A–D). Dispersion values (MVDISP) are shown. NOTES 753 of comparison, a proper understanding of the habitat variability within a biotope is required. Our results from one mangrove biotope emphasize that intra-biotope variability can be significant. Our findings from the Curuçá estuary suggest that spatial differences in the composition of the intertidal fish fauna within a zone of homogenous salinity are dependent on landscape features. However, there is still not enough information to precisely determine general spatial relationships between mangrove creek features and fish assemblage structure. Due to the uniqueness of each creek, it is difficult to define a minimum number of intertidal creeks that must be sampled in order to provide a representative picture of the spatial variation within a salinity zone. Given the spatial heterogeneity encountered in fish assemblages among creeks of the same salinity zone, however, a preliminary classification of mangrove habitat landscape features seems necessary (sensu Valesini et al., 2003) to establish, for example, how many similar types of intertidal mangrove creek exist in an area. Only then can different types of creeks be reasonably related to spatial differences in fish habitat use.

Acknowledgments

The authors thank A. Jesus, E. Lameira, B. Almeida, D. Monteiro, F. Arnour, and J. Kiku- chi for assistance in field collections and laboratory analyses. We thank the editors and all referees for their valuable comments on the manuscript. T. Giarrizzo acknowledges financial support by the CNPq grant 303958/2003-0. U. Krumme acknowledges financial support by the German Ministry for Education, Science, Research and Technology (BMBF). This work was funded by the Millennium Initiative Project “Coastal Resources” of the Brazilian Minis- try of Science and Technology and the Brazilian-German project Mangrove Dynamics and Management—MADAM (project number: 03F0253A5).

Literature Cited

ANA (Agência Nacional das Águas). 2005. Accessed April 30, 2005. Available from: http:// www.ana.gov.br. Barletta, M., A. Barletta-Bergan, U. Saint-Paul, and G. Hubold. 2003. Seasonal changes in density, biomass, and diversity of estuarine fishes in tidal mangrove creeks of the lower Caeté Estuary (northern Brazilian coast, east Amazon). Mar. Ecol. Prog. Ser. 256: 217–228. Barletta-Bergan, A., M. Barletta, and U. Saint-Paul. 2002. Community structure and temporal variability of ichthyoplankton in North Brazilian mangrove creeks. J. Fish Biol. 61: 33–51. Batista, V. S. and F. N. Rêgo. 1996. Análise de associações de peixes, em igarapés do estuário do Rio Tibiri, Maranhão. Rev. Bras. Biol. 56: 163–176. Blaber, S. J. M., D. T. Brewer, and J. P. Salini. 1989. Species composition and biomasses of fishes in different habitats of a tropical northern Australian estuary: Their occurrence in the adjoining sea and estuarine dependence. Estuar. Coast. Shelf Sci. 26: 509–531. Clarke, K. R. 1993. Non-parametric multivariate analyses of changes in community structure. Austr. J. Ecol. 18: 17–143. ______and R. M. Warwick. 1994. Change in marine communities: an approach to statistical analysis and interpretation. Natural Environment Research Council, Plymouth. 144 p. Froese, R. and D. Pauly (eds.). 2005. FishBase. Accessed January 30, 2006. Available from: http://www.fishbase.org. Ikejima, K., P. Tongnunui, T. Medej, and T. Taniuchi. 2003. Juvenile and small fishes in a mangrove estuary in Trang province, Thailand: seasonal and habitat differences. Estuar. Coast. Shelf Sci. 56: 447–457. NOTES 754

Krumme, U., U. Saint-Paul, and H. Rosenthal. 2004. Tidal and diel changes in the structure of a nekton assemblage in small intertidal creeks in northern Brazil. Aquat. Living Resour. 17: 215–229. Laroche, J., E. Baran, and N. B. Rasoanandrasana. 1997. Temporal patterns in a fish assemblage of a semiarid mangrove zone in Madagascar. J. Fish. Biol. 51: 3–20. Leh, M. U. C. and A. Sasekumar. 1991. Ingression of fish into mangrove creeks in Selangor, Ma- laysia. Pages 495–501 in Alcala, ed. Proc. Reg. Symp. Living Resour. Coast. Areas. Marine Science Institute, University of the Philippines, Quezon. Nagelkerken, I., G. van der Velde, M. W. Gorissen, G. J. Meijer, T. Van’t Hof, and C. den Hartog. 2000. Importance of mangroves, seagrass beds and the shallow coral reef as a nursery for important coral reef fishes, using a visual census technique. Estuar. Coast. Shelf Sci. 51: 31–44. McHugh, J. L. 1967. Estuarine nekton. Pages 581–620 in G. H. Lauff, ed. Estuaries. American Association for the Advancement of Science, Washington, D.C. Paterson, A. W. and A. K. Whitfield. 2003. The fishes associated with three intertidal salt marsh creeks in a temperate southern African estuary. Wetlands Ecol. Manage. 11: 305–315. Robertson, A. I. and S. J. M. Duke. 1990. Mangroves as nursery sites: comparisons of the abundance and species composition of fish and crustaceans in mangroves and other nearshore habitats in tropical Australia. Mar. Biol. 96: 193–205. Sasekumar, A., V. C. Chong, M. U. Leh, and R. D’Cruz. 1992. Mangroves as a habitat for fish and prawns. Hydrobiologia 247: 195–207. Valesini, F. J., K. R. Clarke, I. Eliot, and I. C. Potter. 2003. A user-friendly quantitative approach to classifying nearshore marine habitats along a heterogeneous coast. Estuar. Coast. Shelf Sci. 57: 163–177. Vidy, G., F. S. Darboe, and E. M. Mbye. 2004. Juvenile fish assemblages in the creeks of the Gambia Estuary. Aquat. Living Resour. 17: 56–64. Wright, J. M. 1986. The ecology of fish occurring in shallow water creeks of a Nigerian mangrove swamp. J. Fish Biol. 29: 431–441.

Addresses: (T.G.) Laboratório de Biologia Pesqueira - Manejo dos Recursos Aquáticos, Uni- versidade Federal do Pará (UFPA), Av. Perimetral 2651, Terra Firme, 66040170 Belém, Pará,- Brazil. (U.K., T.G.), Center for Tropical Marine Ecology (ZMT), Fahrenheitstr. 6,�� 283��5�9� ��Bre- men, Germany. Corresponding Author: (T.G.) E-mail: . BULLETIN OF MARINE SCIENCE, 80(3): 755–772, 2007

MANGROVES-FISHERIES LINKAGES— THE MALAYSIAN PERSPECTIVE

V. C. Chong

ABSTRACT Mangroves contribute substantially to coastal fisheries in terms of providing trophic and refuge support, and larval retention. Malaysia’s existing 566,856 ha of mangrove forests are believed to sustain more than 50% of its annual offshore fishery landings of 1.28 million t. Here I review research that relates to the contribution of Malaysian mangroves as nursery areas, and to coastal food chains and fisheries. I examine salient features of mangrove productivity, food resources, refugia, and fish nursery-ground value of selected mangrove areas. The net fisheries contribution from 1 ha of mangrove forest amounted to US$846 yr–1. Various anthropogenic activities such as unsustainable forestry practices, aquaculture, coastal development, and pollution influence the carrying capacity of mangroves. The nursery function of Malaysian mangroves is variable depending on several factors, including hydrogeomorphology and the extent of human disturbance, both of which affect carrying capacity.

Mangroves of South and Southeast Asia account for 41.4% of the world’s mangroves of over 18 million ha (Spalding et al., 1997). Malaysia’s total mangrove forests cover 566,856 ha. These forests are located mainly in the eastern region in the States of Sabah (340,689 ha) and Sarawak (126,400 ha), while the most populated, western region or peninsular Malaysia has 97,780 ha (17%) (Shaharuddin et al., 2004). The total annual fishery landings of Malaysia amount to 1.28 million t with a total value of US$1.06 billion, of which 72% is derived from peninsular Malaysia (Annual Fish- eries Statistics, 2003). However, this is more indicative of the unequal development of the fishing industry in Malaysia, which is developed or even overdeveloped in the western or peninsular region but underdeveloped in the eastern region. Peninsular Malaysia has 90,041 ha or 92% of its total mangroves on its sheltered western margin facing the Malacca Straits, while its eastern margin facing the open South China Sea has only 7739 ha of mangroves. Finfish catches derived from its western waters total 491,624 t (57%) annually while 370,925 t are from its eastern waters (Annual Fisheries Statistics, 2003). Shrimp catches are even more starkly different: 45,950 t (90%) from the west in contrast to 4934 t from the east. Further evidence are the various quantified relationships obtained between mangrove area and fisheries production, although the most convincing examples are from shrimp catches, as in Malaysia (Sasekumar and Chong, 1987), Indonesia (Martosubroto and Naamin, 1977), Philippines (Paw and Chua, 1989), and Australia (Staples et al., 1985). However, such empirical relationships have been criticized by many as not necessarily showing cause and effect (e.g., Robertson and Blaber, 1992; Baran and Hambrey, 1998), and by being confounded by various physical and biological factors (Chong and Ooi, 2001). Among these are differences in species with varying degrees of mangrove-dependence, type of mangrove (e.g., riverine or basin), presence of other coastal biotopes and the influence of covariables such as depth, rainfall, and salinity. Nevertheless, Chong and Ooi (2001) were able to associate and rank the abundance of specific species of shrimp to mangrove area, mudflats, and other coastal biotopes

using a multivariate or canonical ordination technique. Thus, the argument for a mangrove-fisheries connection remains compelling. Various hypotheses—which are not necessarily mutually exclusive—have been advanced to explain why mangroves are important to fish and shrimps and thus support coastal fisheries (see e.g., Robertson, 1993; Chong, 1996). Of these, the most popular are the food supply and refuge hypotheses, both of which have been subjected to extensive testing and research. This paper examines the mangrove-fisheries linkages as are known in Malaysia, based mainly on our continuing work in the last 20 yrs, in an effort to understand and elucidate connectivity based on these hypotheses.

Food Supply Hypothesis

Mangrove Productivity.—The “food supply hypothesis” is based on the prem- ise that mangroves are estuarine areas of high productivity providing a ready supply of organic matter which is utilized directly or indirectly by marine fishes, shrimps, crabs, and others. Ong et. al. (1984) estimated a mean net primary productivity of 23.3 t ha–1 yr–1 for a 15-yr-old stand of Rhizophora apiculata Blume, 1827 in the Matang mangrove forest, but productivity values varied from 16 to 50 t ha–1 yr–1 depending on stand age, with the highest for 15 to 20-yr-old stands. The same-age stand produced 10 t ha–1 yr –1 of mangrove litter, measured over a 12-mo period (Gong et al., 1984). On the other hand, Sasekumar and Loi (1983) obtained mangrove litter production values of 15.4, 14.0, and 15.8 t ha–1 yr–1 for Avicennia, Sonneratia, and Rhizophora stands respectively, in Kuala Selangor over a 22-mo period. The total standing biomass of the Matang mangrove (with an area of 40,151 ha) has been estimated at 8.26 million t of dry matter, with an estimated biomass released of about 1 million t annually (Gong and Ong, 1990). Leaf litter biomass and nutrients export alone would account for 3.9 and 0.1 t ha–1 yr–1 respectively. Such exports are expected to have significant impact on the surrounding environment, whether on the sediment or pelagic processes. Energy from Mangrove Detritus.—Fragmentation of mangrove leaves (for instance, by herbivorous crabs) and their subsequent decomposition, including other woody material, form and sustain a detritus-based food chain in mangrove waterways and immediate nearshore waters (Odum and Heald, 1975; Robertson et al., 1992). Apparently, during the leaching process of dissolved organic matter (DOM) there is no microbial intervention (Camilleri and Ribi, 1986), but subsequent min- eralization of leaf detritus depends greatly on bacterial and fungal communities (Benner al., 1988). As much as 40% of the mangrove leaf organic matter consists of leachable materials which are rapidly converted into microbial biomass (Benner et al., 1986). Flocculation into flakes stimulates rapid colonization by protozoa and meiofauna (Camilleri and Ribi, 1986), forming a nutritive food source for a variety of small invertebrates in the mangroves. Refractory lignocellulosic components are left behind or very slowly mineralized by microbes, and appear to be of little nutritional value since few animals can directly assimilate these refractory materials. Because of this, the food value of outwelled coastal mangrove detritus has been questioned. Using stable carbon isotope ratios of mangrove leaves, leaf litter, plankton, and a variety of fish and invertebrates in the Klang mangroves, Rodelli et al. (1984) demonstrated that while the latter showed a wide range of δ13C values (−28.6 to −15.5‰) indicating consumption of both man- CHONG: MANGROVE-FISHERIES LINKAGES IN MALAYSIA 757

grove and algal carbon, those residing in offshore waters (2–8 km) did not consume mangrove carbon (−19.1 to −13.1‰). Since mangrove carbon was clearly distinguish- able in suspended particulates and sediments they suggested that the outwelled or exported mangrove materials are highly refractory to metabolism and hence play no part in coastal food chains. A review by Robertson et al. (1992) similarly supported this view. However, inside the mangrove creek, Rodelli et al. (1984) showed that juvenile banana prawn Penaeus merguiensis de Man, 1888 assimilated on average 65% mangrove carbon, with remainder from other sources. Their result is consistent with that derived through stomach content analysis by Chong and Sasekumar (1981) who found that juvenile banana prawns from a tidal creek in Klang mangrove had organic detritus that varied seasonally between 10%–80% (average 45%). Newell et al. (1995) also working in the Klang mangroves further substantiated the contribution of mangrove detritus to the nutrition of juvenile banana prawns living in tidal mangrove creeks by using triple stable (C, N, S) isotope ratio analysis. In vivo 14C feeding experiments by the same authors, showed no difference in the assimilation efficiency of chemically-defined, lignocellulosic crude fibre mangrove carbon by adult (10.94%) and juvenile (10.02%) rainbow prawn Parapenaeopsis sculptilis (Heller, 1862) living in offshore (13 km) waters. Interestingly, while juvenile banana prawns living inside mangrove creeks showed closely similar assimilation efficiency (13.42%), adult banana prawns living offshore had only 2.07% (± 0.84%; n = 17) assimilation efficiency. Although the discrepancy between assimilation efficiency of adults of both species is not readily explained, it suggests an ontogenetic loss of cellulase enzymes in adult banana prawns that may be related to the shortfall of mangrove detritus in open offshore waters. Thus, their study shows that despite the intrinsic ability of both prawn species to digest even refractory mangrove detritus, mangrove material widely dispersed in offshore waters is not sufficient to support prawn nutrition. In contrast, Chong et al. (2001) show the substantial contribution of mangrove detritus to the nutrition of five shrimp species, which amounted to as much as 84% for juvenile shrimps residing 14 km upstream in Matang rivers. However, this amount rapidly declined offshore as algal carbon became more important. In comparison to results by Rodelli et al. (1984) and Newell et al. (1995) for the Klang mangroves, these results suggest that different mangrove ecosystems provide variable support of coastal production. It is conceivable that the larger area of Matang mangrove forests contributes to the larger quantity of outwelled mangrove detritus which is also less dispersed due to weaker tidal currents when compared to the Klang region. Stable C isotopes of seven pelagic fishes and a squid in Matang ranged from −21.0 to −25.0‰, suggesting mangrove carbon assimilation of between 40%–80%; the highest was observed in the anchovy Stolephorus commersonnii Lacépède, 1803, while the squid Loligo duvaucelli Orbigny, 1848 incorporated as much as 40% mangrove carbon in its tissue (Hayase et al., 1999). These results indicate that suspended mangrove detrital particles, fed upon directly or indirectly via intermediaries, may be more important to fish nutrition than previously thought. The main detrital food chain is supported by the large store of decomposing mangrove litter at the bottom of mangrove forests, which is consumed by the various micro- and macrofauna. These animals form the potential food sources for fishes and shrimps. Thus, mangrove detritus represents a direct or indirect food source. δ15N values indicate that shrimps are at the second or third trophic levels (2‰–3‰ 758 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

enrichment; Chong et al., 2001), whilst the pelagic fishes were at the second to fourth trophic levels (3‰–7‰ enrichment) (Hayase et al., 1999). Macrobenthos Abundance in Mangroves.—The study of the abundance and distribution of macrobenthos should contribute to our understanding of the issue of food supply of mangroves. However, quite unexpectedly, the main waterways of Matang mangrove are generally poor in macrobenthos, and those present largely occur at river mouths. In a 14-mo study at Matang, Muhammad Ali (2004) recorded 80 taxa of macrobenthos, comprising polychaetes (38), small crustaceans (22), gastropods (8), bivalves (6), and others (6). Total macrobenthos density (numbers m–2) values ranged from 22.23 (±46.45) from upstream (ca. 8 km) to 47.91 (±49.56) at the river mouth of Sg. Sangga Besar, and from 12.84 (±17.71) at 14 km upstream to 38.55 (±74.73) at the river mouth of Sg. Selinsing. In a cross-river study of macrobenthos abundance at Sg. Selinsing, Muhammad Ali et al. (1999) found that the mangrove forest had an influence on animal abundance, in that sampling sites closer to the forest fringe yielded more macrobenthos as compared to those located at the midsection of the river, with the highest observed at the forest floor itself. Also, more polychaetes and crustaceans were found at the forest fringe, whereas midsection sites contained nearly all bivalves. As will be discussed later, this “edge” distribution and abundance of macrobenthos has an important impact on the abundance of their predators. Fish Diet Based on Stomach Content Analyses.—Stomach content analysis has yielded evidence that macrobenthos are being consumed as food by fish. Yap et al. (1994) reported that the major food of sciaenid fishes in Matang were penaeid shrimps, and other crustaceans such as amphipods, Acetes shrimps, hermit crabs, mysids, polychaetes, gastropods, and mangrove detritus were also in their diet. Large juvenile (> 10 cm TL) mangrove snappers, Lutjanus johnii (Bloch 1792) feed largely on mysids, prawns, crabs, anomurans, and isopods, but small juveniles prefer smaller organisms including copepods and mysids (Kiso and Mahyam, 2003). Six ariid catfishes in Matang were reported to feed on a variety of mangrove macrobenthos that included polychaetes, penaeids, crabs, gastropods, as well as fish and sergestids (Singh, 2003). He suggested that the high diet overlap among the ariid species is due to food abundance in the Matang mangrove, allowing for a small, but varying degree of diet specialization. In the Klang mangrove, several fish diet studies have shown the direct consumption of mangrove detritus by iliophagous fishes such as gray mullets (Chong, 1977), gizzard shads (Fong, 1977), and Acetes shrimps (Tan, 1976), with volumetric compositions that reached 50%, 25%, and 38%, respectively. Chong and Sasekumar (1981) showed an ontogenetic shift in the food items of the banana prawn as they migrate from nursery (mangrove) to spawning grounds in offshore waters. In mangrove creeks, young postlarvae consumed largely copepods while the larger juveniles consumed 40%–70% mangrove detritus, with small amounts of copepods, and brachyuran and gastropod larvae. In offshore waters, the prawns’ diet shifted to that comprising of Acetes, mysids, polychaetes, bivalves, and small quantities of detritus (10%–30%). In shallow coastal grounds (up to 3 km offshore) adjacent to the Klang mangrove, Leh and Sasekumar (1984) reported that the diet of 14 penaeid shrimp species comprised between 12%–36% mangrove detritus, the rest being animal material from copepods, ostracods, and crab larvae. From nearshore waters, large numbers of at least CHONG: MANGROVE-FISHERIES LINKAGES IN MALAYSIA 759

11 species of fishes, including catfishes, eel catfish, gray mullets, silver grunters, and treadfins, make feeding forays into the coastal mangroves via the coastal mudflats, feeding on bottom items such as mangrove detritus, grapsid crabs, shrimps, sergestids, amphipods, stomatopods, molluscs, sipunculids, and on the fauna of mangrove tree trunks such as Sphaeroma and Brachiodontes (Sasekumar et al., 1984). How Important are Plankton?—The contribution of mangrove phytoplankton and microphytobenthos could be substantial in some cases. For instance, the phytoplankton contribution in less turbid, large channels and open lagoons may exceed that from the fringing mangrove plants (Robertson and Blaber, 1992). Although it is generally accepted that mangrove detritus forms the basis of the food chain in mangrove waters, this contention should be reviewed and modified in the wake of stable isotope evidence that demonstrates the increasing importance of phytoplankton downstream within the mangroves (e.g., Hayase et al., 1999; Chong et al., 2001). Banana prawns located at the Sg. Sangga Besar river mouth have as much as 70% algal carbon in their tissues (Chong et al., 2001). The downstream reaches of such rivers, particularly at their mouths, are characteristically wide and not shaded by mangrove trees. Incident light is strong and water is typically green. Chlorophyll a concentrations as high as 50 µg L–1 have been recorded at Matang estuaries (Alongi et al., 2003). Thus, the food chain is likely supported by abundant phytoplankton in the open water downstream, which may be vital to planktonic organisms. The zooplankton commonly consumed by young fishes in Matang is represented by copepods and planktonic shrimps such as mysids and Acetes. Young juveniles of Arius maculatus Thunberg, 1792 consume 70%–100% copepods, while the adults consume mainly fish and penaeids (Singh, 2003). Based on the examination of stomach contents of 2123 postlarval, juvenile, and small fishes that were sampled over a year, Chew et al. (2007) concluded that most of the 26 major fish species in Matang mangroves depended on zooplankton for food. The fish diet comprised 47 taxa of prey food such as copepods (53%), Acetes shrimps (16%), mysids, ostracods, polychaetes, larval mollusks, and cirripedes (3%–8%). Detritus (40%) was present in variable, but small amounts. Frequent consumers of copepods (50%–85%) included ambassids, leiognathids, engraulids, gerreids, and juvenile ariids and sciaenids. Planktonic shrimps such as Acetes and mysids were consumed by many common or economically-important fish species such as carangids, threadfins, snappers, grunts, anchovies, sciaenids, mullids, and gobiids, with occurrences that ranged from 9% to 61% and mean volumetric composition of 7%–60%. Stable isotope studies by Chew et al. (2007) show that fish larvae and juvenile fishes in the mangrove consumed mainly zooplankton that derived most of their energy from phytoplankton and/or benthic diatoms (Fig. 1). Thus, young fish in the mangrove forest shift their diet preference ontogenetically from one that is phytoplankton-based to one that is detritus-based, although some species such as anchovies and at least one sciaenid, Johnius vogleri (Bleeker, 1853), continue to derive their nutrition from phytoplankton.

The Refuge Function Hypothesis

Role of Refugial Space.—The “refuge function hypothesis” is based on the prem- ise that mangrove forests provide habitat complexity or heterogeneity in the form of above-ground root structures and plant litter, and soft muddy substrates associated with turbid and shallow waters. Root structures such as prop roots and pneumato- 760 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 15N δ

Figure 1. δ13C and δ15N values of primary producers and consumers in the estuaries of Matang Mangrove Forest Reserve, Malaysia.

phores, fallen litter, soft mud, and shallow water provide the refugial space for prey and young fishes. High turbidity further reduces perception distance of visual predators. In a study of shrimp predation by sciaenid fishes in Matang mangroves, Yap et al. (1994) found that the mid-water species Panna microdon (Bleeker, 1849), Joh- nius vogleri (Bleeker, 1853), and Pennahia macropthalmus (Bleeker, 1850) consumed more penaeid shrimps (65%–95% volumetric composition) than did demersal species Johnius weberi Hardenberg, 1936, Johnius carouna (Cuvier, 1830), and Dendrophysa russelii (Cuvier, 1829); (23%–43%). Shrimps become more vulnerable to their piscine predators when they leave the litter-strewn benthos, despite very turbid water conditions during spring tide. A major demersal predator in the Klang mangroves is the gray eel catfish, Plotosus canius Hamilton, 1822 which feeds on a wide variety of macrofauna including crabs, shrimps, bivalves, amphipods, polychaetes, sipunculids and other fish (Leh and Sasekumar, 1993). Shrimps are however not the major food item (only 2%–10%) of the fish despite their abundance in the mangroves. Apparently the juvenile shrimps are able to avoid these voracious barbeled predators that are well adapted to feed in turbid mangrove waters. Various studies have shown how mangrove root structures and substrata may provide refugia for shrimps from their predators. Through a series of laboratory experiments using two species of penaeid shrimps (Penaeus merguiensis De Man, 1888 and Penaeus monodon Fabricius, 1798) and predaceous fish Lates[ calcarifer (Bloch, 1790 and Lutjanus argentimaculatus (Forsskål, 1775)], Primavera (1997) found that pneumatophores of Sonneratia provide more safety to shrimps (30% mortality) if compared to leaf bracts (44%) and bare sand (49%), but concluded that prey mortality varied greatly depending not only on shelter structure and density, but also on prey and predator behavior. In an indirect study of predator avoidance by shrimps CHONG: MANGROVE-FISHERIES LINKAGES IN MALAYSIA 761

in their natural habitat, Affendy and Chong (2006) found that catch per unit effort (CPUE) of shrimps by shrimp pots varied greatly according to inundation height (or water depth), whether sites were open or among roots, and mangrove (Rhizophora spp.) stand age. Their study revealed shrimp incursions of up to 60 m into the mangrove forest (Matang) during high tide, and increasing shrimp densities with decreasing water depth (i.e., higher elevation stations) and mangrove stand age. Also, root-shelter sites were more attractive than open-space sites. The lowest CPUE was obtained from the clear-felled forest site whilst the highest was from the 24-yr-old forest which had the highest root occupied area. However, tree density (highest in the 5-yr-old forest), appeared to be not as important as forest maturity in attracting shrimps. Chong et al. (1994) recorded higher penaeid shrimp catches at the banks of the Sg. Sangga Besar and Sg. Selinsing (Matang) than at their mid-sections. For instance, at Sg. Sangga Besar, mean monthly density (biomass) varied between 250–500 shrimps ha–1 (0.55–1.30 kg ha–1) and 430–1800 shrimps ha–1 (0.82–3.46 kg ha–1), for mid-section and bank, respectively. The study also found proportionally higher numbers of P. merguiensis at the banks of the river than at the middle, whereas at the latter, Metapenaeus brevicornis (H. Milne Edwards, 1837) and Metapenaeus affinis (H. Milne Edwards, 1837) were relatively more abundant. Such spatial differences in abundance are thought to be related to diel variation in feeding activity and/or to predator avoidance in response to turbidity changes associated with the spring and neap tide regimes (Chong et al., 1994). Thus, P. merguiensis, a non-burrower, was distributed more evenly across the river during spring tides (high turbidity), but aggregated at the banks during neap tides (lower turbidity) to avoid predation. Low and Chong (1999) similarly observed higher shrimp abundance at the banks than at the middle of Sg. Selinsing, but found no significant difference between shrimp abundance during night and day, nor between spring and neap tide. Only M. brevicornis was observed to be significantly more abundant during night than day, which is sug- gestive of nocturnal feeding activity. More studies in the Matang mangroves as well as the Dinding mangroves clearly show the preference of shrimps for the mangrove forest fringe at the river bank (Table 1), which supports the idea that shrimp aggregation in shallow water is likely a predator-avoidance strategy. In an attempt to assess the protective roles of different mangrove plant species to banana prawns of Australia, Vance et al. (1996, 2002) chose forest types dominated by the structurally-complex Rhizophora stylosa Griffith, 1854 and the more open forests of Ceriops tagal C. B. Robinson, 1908 and Avicennia marina (Forsskål) Vierhapper, 1907. They found no clear-cut differences in shrimp abundance between mangrove forest types, and noted decreased catch with increased distance (decreased water depth) into the forest at a river site, while at a creek site, the highest catches were recorded both at the creek mangrove fringe and further inland. They concluded that shrimp abundance depends largely on the local topography and water current pattern. Ronnback et al. (1999) who sampled shrimps and fishes from four microhabitats that differed in dominant mangrove species (A. marina, Avicennia officinalis Lin- naeus, 1753 and R. apiculata) in the Pagbilao mangroves, Philippines, observed extensive use of intertidal mangrove forests by the vagile fauna, presumably in response to predation. They observed that the fish community preferred pneumatophore over prop root habitats, while the highest shrimp density was obtained in a replanted Rhizophora habitat which had the highest root structural complexity. Their findings 762 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Table 1. Comparison of the abundance of shrimps in the Matang Mangrove Forest Reserve (Sungai Selinsing, Sungai Sangga Besar) and the Dinding Mangrove Forest (1996–97) (data from Low et al., 1999).

Mean abundance (no. of shrimps ha–1) Selinsing Sangga Besar Dinding Site Mean SD Mean SD Mean SD Middle region 476 254 462 341 432 881 River bank 3,715 4,446 751 702 70 42 Pooled abundance Selinsing Sangga Besar Dinding River Mean SD Mean SD Mean SD No. of shrimps ha–1 1,758 1,292 593 391 379 744 kg ha–1 2.11 1.30 1.19 0.57 1.75 3.57 on successful recolonization of the replanted Rhizophora habitat by shrimps and fish imply that mangrove rehabilitation can be a potential management tool for enhancing fisheries stocks. Fish Movements and Diversity.—Some of the most intriguing questions concerning fish movements between mangrove and offshore waters concern the type of movement, the purpose of the movement, and the time and developmental stage that performs this movement into the mangrove. Obviously, fish movements are very much related to the requirements of maximum survival, growth, and body condition, and these in turn are related to the conditions prevailing in the nursery and spawning areas. Blaber (2000) classified four major categories of tropical and subtropical estuarine species, based on the type of movement and purpose of use of the habitats. These are: (1) estuarine species that can complete their life cycle inside the estuary (e.g., mudskippers, sciaenids, mullets), (2) marine migrants from the sea that spawn in more saline offshore waters, but their young or adults migrate into estuaries to nurture or feed for varying periods (e.g., threadfins, carangids, stingrays, penaeid prawns), (3) anadromous species that migrate upstream to spawn (e.g., shads), and (4) freshwater species that usually spawn in freshwater, but migrate into estuaries, (e.g., eleotrids). Matang mangrove fishes (138 identified species) include 40% estuarine residents and 60% marine migrants, with 105 species or 76% that are commercially exploited (Chong, 2005). Except for five species of caridean shrimps, all 15 penaeid shrimp species are marine migrants, and all are commercially exploited. Sasekumar et al. (1994) reported that 87% of the fishes in Matang mangrove waterways and 83% in adjacent mudflats were juveniles, suggesting a larger role of the mangrove as nursery ground. Major marine migrant species included the economically-important threadfins, trev- allies, queenfish, snappers, and groupers. Economically-important estuarine species included most of the sciaenids, several species of clupeids (e.g., gizzard shads), catfish eels, and gray mullets. Yap et al. (1994) and Singh (2003) who respectively studied the feeding and reproductive ecology of Sciaenidae and Ariidae, concluded that species belonging to these families breed, and their young are nurtured, inside the mangrove forest. Matang waterways and adjacent mudflats are the major nursery areas for shrimps which comprise 70%–89% and 40%–90% of all juveniles, respectively (Chong et al., 1994). For banana prawns in Matang, Ahmad Adnan et al. (2002) reported a broad period of postlarval and juvenile recruitment from December to May and subse- CHONG: MANGROVE-FISHERIES LINKAGES IN MALAYSIA 763 quent juvenile offshore migration from February to July. Low et al. (1999) reported that high numbers of M. brevicornis (5–23 mm CL) occurred in the Matang estuaries in May and October, but recruitment was continuous. Parapenaeopsis sculptilis in Matang is the only species of the genus to penetrate significantly upstream. Con- tinuous recruitment and the presence of multiple modal populations or cohorts with sizes that ranged between 6–34 mm CL, indicate a longer period of residence in the mangroves. This species is known to spawn just off the subtidal edge of coastal mudflats. The mangrove caridean shrimp, Exopalaemon styliferus (H. Milne Edwards, 1840) belongs in the estuarine category, and comprises a seasonal nearshore fishery. In Klang mangrove, 114 species of fishes and 12 species of shrimps have been identified from coastal and island creeks or inlets. Half of the fish species present were common to the 132 species recorded from the adjacent coastal mudflats, as well as to the 113 species found in the inshore waters (Chong et al., 2005a). However, the mudflat fish community is transient, comprising periodic wanderers from both mangrove creeks and immediate offshore habitats during high tide. Chong et al. (1990) suggested that the Klang mangrove system, with its small but abundant coastal creeks and inlets functions more as a feeding area for marine fishes. In the Johor mangroves, located in the south of the peninsula, 130 fish species were recorded from the Sungai Johor estuary, of which 91 species (70%) were commercially exploited, while 105 fish species were recorded from the Sungai Pulai estuary, of which 67 species (63.8%) were commercially exploited (Chong and Sasekumar, 2002b). Both estuarine systems in Johor function as feeding, nursery and/or breeding area for marine fish and shrimps. To the north of the peninsula, Malaysia’s only type of mangrove growing on limestone platforms, or Type VI mangrove setting (sensu Thom, 1984), are found on the northeastern coast of Langkawi Island. A total of 91 species of fish belonging to 42 families has been recorded, but the fish fauna include not only common mangrove species but also coral reef species such as sparids, labrids, haemulids, and monacan- thids (Chong et al., 2005b). Apparently these species also utilize the high salinity (21–33.5) waters of the mangrove-fringed estuaries as their living habitat and feeding area. The Sungai Johor coastal system also reflects that of conjoint habitats, where fish movements between mangroves, seagrass beds, and coralline sponge beds were evident (Chong and Sasekumar, 2002b). These two unique scenarios warrant further research on coupling of coastal habitats for their rational management. An often-asked question is “do fish generally enter their nursery areas (mangroves) as larvae?” Certainly the zooplankton resources in mangroves are sufficient to support large numbers of larvae, as they do for postlarval shrimps (see below). However, for fish some interesting results have emerged. From an 11-mo study in which more than 70,000 fish larvae were sampled, Ooi et al. (2007) reported that the larval fish densities inside the Matang mangrove estuary seasonally ranged from 92 to 1248 larvae 100 m–3. The larvae were numerically dominated by the families Gobiidae (53%) and Engraulidae (42%), and larvae of 14 other families including the Scatophagidae, Ambassidae, Blennidae, Sciaenidae, Cynoglossidae, Carangidae, Scorpaenidae, Syn- gnathidae, Teraponidae, Haemulidae, Ogcocephalidae, and Bregmacerotidae constituted the remaining 5%. However, 49 fish families have been recorded from the Matang mangrove by Chong (2005), suggesting that other species may move into mangrove waters post-settlement the early juvenile stage. It has been reported that young mangrove snappers, L. johnii, of mostly 10–15 cm SL, continuously recruit 764 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

into their mangrove nursery area, staying in for approximately a year, before they start to migrate back to the open sea at ca. 25 cm SL (Ogawa, 2003; Amy Then et al., 2006). Larvae of marine, mangrove-dependent fishes may not be as tolerant of brackish waters as juveniles. Mangroves as Retention Areas for Larvae.—Mangroves may function as retention areas for the planktonic stages of fish and shrimps which would otherwise be swept away by water currents and tides. The process of larval accumulation in nearshore waters is critical to their successful recruitment into estuaries or mangrove areas, and appears prerequisite to mangroves functioning as nursery areas. Two-dimensional, numerical modelling of the shallow water of the Klang Strait has demonstrated that the coastal mangrove forests have the ability to retain passing shrimp larvae as a result of lateral trapping of a coastal boundary layer, retaining as many as 65 × 109 penaeid shrimp postlarvae annually (Chong et al., 1996). Based on the model, it appears that if the coastal mangroves are destroyed for harbor, aquaculture, and other development, not only will the Klang Strait silt up, but up to 50% less shrimp larvae from offshore spawning grounds may be advected near to their nursery area, and the loss rate of these larvae thereafter may double. Nursery-Ground Value of Mangroves.—Mangrove waterways in Matang support the highest density of fish, 8517 fish –1ha (biomass of 40 kg ha–1), followed by mudflats with 6699 fish –1ha (30.5 kg ha–1) and nearshore waters with 385 fish ha–1 (2.5 kg ha–1), i.e., fish density and biomass decreases in the offshore direction (Sase- kumar et al., 1994). Shrimp densities are also highest inside mangrove waterways, ranging from 292–34,000 shrimps ha–1 (mean = 4092 ± 6616 shrimps ha–1), with biomass of 0.63–38.28 kg ha–1 (mean = 6.86 kg ha–1 ± 7.55), as compared to the mudflat with 255–7788 shrimps ha–1 (mean = 2294 ± 2165 shrimps ha–1), and biomass of 1.64–31 kg ha–1 (mean = 6.78 ± 6.93 kg ha–1) (Chong et al., 1994). In the Klang area, the same trend of fish and shrimp densities decreasing from mangrove to offshore waters occurs: from 12,860 to 492 fish ha–1 and 17.7 to 3.74 kg fish ha–1, respectively, while shrimp densities and biomass were from 1238 to 328 shrimps ha–1 and 3.18 to 1.14 kg shrimp ha–1, respectively (Chong et al., 1990). However, water depth appears to influence shrimp abundance, with higher density shallow water. The nursery-ground value of mangrove habitats is likely variable, depending on several factors, but one such factor as gleaned from our data on shrimp distribution and abundance, is the presence of river banks closest to the mangrove forest (see Table 1) or the forest-water channel interface. Such an area would increase proportionately with the number of creeks or waterways present. Thus, the many waterways that traverse the Matang mangrove not only increase accessibility into the mangrove forest during flood tide, but also increase the edge area for shrimps to rest, feed, and shelter during ebb tide (Chong et al., 1994). The Matang mangrove thus has a much higher nursery-ground value for shrimps when compared to Dinding mangroves. Nevertheless, comparison of the density and biomass of fish and shrimp among several mangrove areas indicates that mangrove nursery value may have decreased due to greater human disturbance (Table 2). Further analysis of the implication of this interface or edge area to shrimp abundance shows that it may be related to the mangrove:water channel (surface area) ratio calculated to be 4.70 for Matang mangrove (Sasekumar et al., 1994) and 1.09 for Dinding mangrove (Chong et al., 1999). But while a large mangrove:channel ratio does not necessarily, and also geometrically, generate the largest interface area, an CHONG: MANGROVE-FISHERIES LINKAGES IN MALAYSIA 765 beam trawl, cast net beam trawl, cast net beam trawl otter trawl, pushnet otter trawl, pushnet beach seine; otter trawl# Sampling gears ) −1 (3.18) #(1.75) Shrimp 0.03–5.60 0.41–15.93 1.03–38.28 (7.32) 1.57–14.33 (4.63) EIA (1997) for fish, Low et al. (1999) shrimp. EIA 6 Biomass (kg ha Fish (9.3) (17.7) (23.3) 0.84–43.51 10.00–30.30 40.00–49.10 ) −1 #(379) (1,238) Shrimp 32–1,417 236–1,714 495–20,536 (4,315) 292–34,082 (4,677) Density (no. ha Fish (274) (3,378) (12,860) 62–4,847 646–4,736 4,331–23,810 Sasekumar et al. (1994) for fish, Chong shrimp; 4, 5 Mangrove status partially degraded partially degraded partially degraded pristine, silvicultured pristine, silvicultured highly degraded Chong et al. (1990); 3 State Johor Johor Selangor Perak Perak Perak Chong (1999); 2 Sasekumar (1999); Pulai River Johor River Klang Sementa Kecil River, Matang Selinsing River, Matang Sangga Besar River, Lumut Dinding River, Site Table 2. A comparison A 2. Table of the abundance of fish and shrimp in a number of mangrove estuaries in Malaysia. Range and/or mean (parenthesis) given (modified 2002). from Chong and Sasekumar, 1 2 3 4 5 6 1 766 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 optimum ratio depends on the number of waterways and a minimum requirement of adjacent mangrove area. Thus, there are valid grounds in Marshall’s (1994) reason- ing that the total mangrove area in some regions may be far in excess of that needed to stock the coastal fisheries, based on the large discrepancy between the biomass of shrimps in nurseries and coastal catches. He further pointed out that in quantifying nursery function, creek area rather than mangrove area per se should be used to quantify nursery function.

Fisheries Contribution to Total Economic Value of Mangrove

An economic valuation of peninsular Malaysia’s west coast mangrove reserves published in 1999 shows the net benefit of direct use due to fish catches from the mangrove forests was US$2.7 million per annum, while the net benefit of indirect use due to their nursery role for marine fishes/shrimps (i.e., from coastal fisheries), was US$67.7 million per annum (MPP-EAS, 1999). The net benefit of total fisheries contribution due to mangroves accounted for US$70.4 million or 5% of the total economic value of mangroves of US$1.38 billion per annum. On a unit hectare basis, the net fisheries contribution of peninsular Malaysia’s mangroves amounted to US$846 per annum.

Human Impacts and Mangrove Conservation

Some of the conservation issues that need to be addressed concern the reasons for the loss of mangrove forests, which are in some cases by natural forces (e.g., typhoons and tsunamis) but mostly by human impacts due to development. These include unsustainable forestry practices, illegal harvests, agriculture, aquaculture, construction of airports and harbors, industrialization and urbanization, land reclamation for coastal development, and waste disposal and pollution, which had taken a combined toll of some 111,046 ha or 16% of Malaysia’s mangrove forest reserves from 1973 to 2000 (Chong and Sasekumar, 2002a), with specific losses in five states ranging from 20%–70% (Chan et al., 1993). The combined effects of various anthropogenic impacts have invariably eroded the carrying capacity of these mangrove habitats to support aquatic fauna. Even the Matang mangrove despite its largely intact area is not im- mune to some of these impacts. Waste disposal and increasing upstream pollution over the years are believed to have reduced its carrying capacity as a nursery ground for shrimps (Table 3). Offshore shrimp catches in the peninsula waters appear to be affected as mangrove forests shrunk by 20% (1980–2000) in area, and are becoming increasingly polluted. Shrimp catches in the peninsula reached 79,591 t in 1990, but catches dropped to 71,505 t in 1994, 62,635 t in 1998, and to 50,885 t in 2003, despite the reduction in fishing effort (Annual Fisheries Statistics, 2003).

Summary

Mangroves are unique irreplaceable ecosystems that sustain coastal productivity. Their ecological functioning not only maintains shoreline stability, but also supports rich biodiversity and coastal fisheries. The linkages between mangroves and fishes or fisheries are real and strong in Malaysia, as attested by the important contribu- CHONG: MANGROVE-FISHERIES LINKAGES IN MALAYSIA 767

Table 3. Comparison of shrimp densities (no. ha–1) between two mangrove estuaries in Matang Mangrove Forest Reserve, Malaysia, after a lapse of 6 yrs.

Chong et al. (1994) Low et al. (1999) Year 1990–93 1996–97 Site Selinsing River Sangga Besar River Selinsing River Sangga Besar River Mean 4,315 4,677 1,758 593 SD 7,202 8,663 1,292 391 Max 20,536 34,082 4,474 1,372 Min 495 292 271 62 tions of mangroves to trophic energy and refugial space that are prerequisites to their functioning as vital feeding and nursery areas. Although Malaysian mangroves offer attractive nursery grounds for a rich diversity and abundance of fishes, shrimps, and other marine organisms, nursery support is variable for different mangrove areas, depending on several factors, including the hydrogeomorphology and extent of human disturbance, which affect carrying capacity. On a broader geographical scale, there are also apparent differences as well as similarities among tropical Malaysian, Philippine, Australian, and New World mangroves in terms of usage for food (e.g., detritus vs plankton) and refugia (mangrove root structures) by vagile fauna. For instance, mangrove detritus appeared important to the shrimp food-web in small creeks of large mangrove forests in Malaysia, whereas it is not apparent in Puerto Rico (Stoner and Zimmerman, 1988), and the Philippines (Primavera, 1996), and only to a some extent in tropical Australia (Lon- eragan et al., 1997). This difference appears to be a function of location where the food resources available to organisms depend largely on the geomorphological and physical characteristics of the environment, and thus parallel the salt marsh situation, where the relative contribution of Spartina detritus and algae to estuarine food webs has been a long standing issue (Haines, 1977; Peterson et al., 1985). Hydrogeomorphology, as a factor affecting carrying capacity, may be human-en- gineered or manipulated to increase the mangrove:channel interface area and possibly increase the coastal fisheries production. Addressing the second factor, which reflects multiple-use conflict, warrants an integrated coastal zone management plan to replace the existing management plans based on sectoral interests. The present analysis highlights the need for further research on (1) the role or effects of hydrogeomorphology, habitat complexity, and pollution on mangrove carrying capacity and fish production, and (2) developing standard procedures and protocols for quantifying mangrove carrying capacity, in a range of geographical locations. Although Malaysia is not free from temptation of the lucrative shrimp farming industry, or the need to develop its valuable coastal lands for settlement, agriculture, and industry, the recent mangrove loss for such purposes is relatively small, due to a previous moratorium on mangrove deforestation, and more recently, to affirmative government effort to replant mangroves after the 2004 tsunami disaster. It is hoped that more mangrove forests in Malaysia can assume a status as exemplified by the 40,151-ha Matang Mangrove Forest Reserve, which has been sustainably-managed for more than a century without loss of a single hectare of forest. 768 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Acknowledgments

I would like to thank JIRCAS (Japan) and MOSTI (Malaysia) for financial support in mangrove research in the last 10 yrs, and two anonymous reviewers for their constructive criti- cisms of the manuscript.

Literature Cited

Affendy, N. and V. C. Chong. 2006. Shrimp ingress into mangrove forests of different age stands, Matang Mangrove Forest Reserve, Malaysia. Bull. Mar. Sci. 80: 915. Ahmad Adnan, N., N. R. Loneragan, and R. M. Connolly. 2002. Variability of, and the influence of environmental factors on, the recruitment of postlarval and juvenile Penaeus merguiensis in the Matang mangroves of Malaysia. Mar. Biol. 141: 241–251. Alongi, D. M., V. C. Chong, P. Dixon, A. Sasekumar, and F. Tirendi. 2003. The influence of fish cage aquaculture on pelagic carbon flow and water chemistry in tidally-dominated mangrove estuaries of peninsular Malaysia. Mar. Environ. Res. 55: 313–333. Amy Then, Y. H., V. C. Chong, H. H. Moh, and Y. Hanamura. 2006. Size frequency, abundance and feeding habits of young snappers (Lutjanus spp.) and groupers (Epinephelus spp.) in the Matang mangrove estuary, Malaysia. Pages 1–5 in K. Nakamura, ed. Sustainable Produc- tion Systems of Aquatic Animals in Brackish Mangrove Areas, JIRCAS Working Report No. 44, Japanese International Center for Agricultural Sciences, Tsukuba, Japan. Annual Fisheries Statistics. 2003. Annual Fisheries Statistics. Volume 1. Department of Fisher- ies, Malaysia. Kuala Lumpur, Malaysia. Baran, E. and J. Hambrey. 1998. Mangrove conservation and coastal management in Southeast Asia: What impact on fisheries resources? Mar. Pollut. Bull. 37: 431–440. Benner, R., R. E. Hodson, and D. Kirchman. 1988. Bacterial abundance and production on mangrove leaves during initial stages of leaching and biodegradation. Archiv fuer Hydro- biologie 31: 19–26. ______, R. Peele, and R. E. Hodson. 1986. Microbial utilization of dissolved organic matter from leaves of the red mangrove, Rhizophora mangle, in the Fresh Creek estuary, Bahamas. Estuar. Coast. Shelf Sci. 23: 607–620 Blaber, S. J. M. 2000. Tropical estuarine fishes: Ecology, exploitation and conservation. Black- well Science, Britain. 372 p. Camilleri, J. C. and G. Ribi. 1986. Leaching of dissolved organic carbon (DOC) from dead leaves, formation of flakes of DOC, and feeding on flakes by crustaceans in mangroves. Mar. Biol. 91: 337–344. Chan, H. T., J. E. Ong, W. K. Gong, and A. Sasekumar. 1993. Socio-economic, ecological and environmental values of mangrove ecosystems in Malaysia and their present state of conservation. Pages 41–82 in B. F. Clough, ed. The economic and environmental values of mangrove forests and their present state of conservation in the South-East/Pacific region, ISME/ITTO/JIAM, Okinawa, Japan. Chew, L. L., V. C. Chong, and Y. Hanamura. 2007. How important are zooplankton to juvenile fish nutrition in mangrove ecosystems? JIRCAS Working Report 56. Chong, V. C. 1977. Studies on the small grey mullet Liza melinoptera (Valenciennes). J. Fish Biol. 11: 293–308 ______. 1996. The prawn – mangrove connection – fact or fallacy? Pages 3–20 in M. Su- zuki, S. Hayase, and S. Kawahara, eds. Sustainable utilization of coastal ecosystems, JIRCAS Working Report No. 4, Tsukuba, Japan. ______. 2005. Fifteen years of mangrove fisheries research in Matang – what have we learnt? Pages 1–20 in M. I. Shaharuddin, M. Azahar, U. Razani, A. B. Kamaruzzaman, K. L. Lim, R. Suhaili, M. S. Jalil and A. Latiff, eds. Sustainable management of Matang mangroves: 100 years and beyond. Forest Biodiversity Series 4. Forestry Department Peninsular Malaysia. CHONG: MANGROVE-FISHERIES LINKAGES IN MALAYSIA 769

______and A. L. Ooi. 2001. Prawn abundance and mangroves: quantified relationships and new perspectives. International Workshop on Mangrove Systems of South East Asia, 6–8 Nov 2001, ICLARM, Penang, Malaysia. 12 p. ______and A. Sasekumar. 1981. Food and feeding habits of the white prawn Penaeus merguiensis. Mar. Ecol. Prog. Ser. 5: 185–191. ______and ______. 2002a. Coastal habitats (mangroves, coral reefs and seagrass beds) of the ASEAN region: Status, utilization and management issues. Fish. Sci. 68: 566–571. ______and ______. 2002b. Fish communities and fisheries of Sungai Johor and Sungai Pulai Estuaries (Johor, Malaysia). Malay. Nat. J. 56: 279–302. ______, C. B. Low, and T. Ichikawa. 2001. Contribution of mangrove detritus to juvenile prawn nutrition: a dual stable isotope study in a Malaysian mangrove forest. Mar. Biol. 138: 77–86. ______, A. Sasekumar, and K. H. Lim. 1994. Distribution and abundance of prawns in a Ma- laysian mangrove system. Pages 437–445 in S. Sudara, C. R. Wilkinson, and L. M. Chou, eds. Proc. Third-ASEAN-Australia Symp. on Living Coastal Resources, Vol. 2: Research Papers, Chulalongkorn University, Bangkok, Thailand. ______, ______, and E. Wolanski. 1996. The role of mangroves in retaining penaeid prawn larvae in Klang Strait, Malaysia. Mangrove and Salt Marshes 1: 11–22. ______, ______, and S. W. Zgozi. 2005a. Ecology of fish and shrimp communities. Pages 174–200 in A. Sasekumar and V. C. Chong, eds. Ecology of Klang Strait, University of Ma- laya Press, Kuala Lumpur. ______, ______, M. U. C. Leh, and R. D’Cruz. 1990. The fish and prawn communities of a Malaysian coastal mangrove system, with comparisons to adjacent mud flats and inshore waters. Estuar. Coast. Shelf Sci. 31: 703–722. ______, ______, C. B. Low, and S. H. Muhammad Ali. 1999. Physico-chemical environment of the Matang and Dinding Mangroves (Malaysia). Pages 115–121 in K. Kiso and P. S. Choo, eds. Proc. Fourth JIRCAS Seminar on Productivity and Sustainable Utilization of Brackish Water Mangrove Ecosystems, 8–9 December 1998, Penang, Malaysia, Japan International Research Center for Agricultural Sciences, Tsukuba, Japan. ______, Y. P. Ng, B. J. Hairi, A. L. Ooi, L. L. Chew, M. Amirah, and B. N. Affendy. 2005b. Update of the fishes of mangrove and coastal waters of northeastern Langkawi. Malay. J. Sci. 24: 167–184. EIA. 1997. Environmental impact assessment for Lekir coastal development: Fisheries resources of the Sungai Dinding and Selat Dinding, Perak, Malaysia. Pages 8-39 to 8-53 in Perund- ing Utama Private Limited, Kuala Lumpur, Malaysia. Fong, C. H. 1977. Studies on Dorosoma chacunda (Bleeker) and Haplochilus melastigma (Mc Clelland) in a Malayan mangrove estuary. B.Sc. Thesis, Department of Zoology, University of Malaya, Kuala Lumpur, Malaysia. 60 p. Gong, W. K. and J. E. Ong. 1990. Plant biomass and nutrient flux in a managed mangrove forest in Malaysia. Estuar. Coast. Shelf Sci. 31: 519–530. ______, ______, C. H. Wong, and G. Dhanarajan. 1984. Productivity of mangrove trees and its significance in a managed mangrove ecosystem in Malaysia. Pages 216–225 in E. Soepadmo, A. N. Rao, and D. J. Macintosh, eds. Proc. UNESCO Asian Symp. on Mangrove Environment - Research and Management: University of Malaya, Kuala Lumpur, Malaysia. Ardyas Publication, Kuala Lumpur. Haines, E. B. 1977. The origin of detritus in Georgia salt marsh estuaries. Oikos 29: 254–260. Hayase, S., T. Ichikawa, and K. Tanaka. 1999. Preliminary report on stable isotope ratio analysis for samples from Matang mangrove brackish water ecosystem. Japan Agriculture Research Quarterly 33: 215–221. Kiso, K. I. and M. I. Mahyam. 2003. Distribution and feeding habits of juvenile and young John’s snapper Lutjanus johnii in the Matang mangrove estuary, west coast of Peninsular Malay- sia. Fish. Sci. 69: 563–568. 770 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Leh, M. U. C. and A. Sasekumar. 1984. Feeding ecology of prawns in shallow waters adjoining mangrove shores. Pages 331–353 in E. Soepadmo, A. N. Rao, and D. J. Macintosh, eds. Proc. Asian Symp. on Mangrove Environment: Research and Management, University of Malaya and UNESCO, Ardyas Publication, Kuala Lumpur. ______and ______. 1993. Ingression of fish into mangrove creeks in Selangor, Ma- laysia. Pages 495–502 in A. C. Alcala, ed. Proc. Regional Symp. on Living Resources in Coastal Areas, Marine Science Institute, University of the Philippines, Quezon City, Philip- pines. Loneragan, N. R., S. E. Bunn, and D. M. Kellaway. 1997. Are mangroves and seagrasses sources of organic carbon for penaeid prawns in a tropical Australian estuary? A multiple stable- isotope study. Mar. Biol. 130: 289–300. Low, C. B. and V. C. Chong. 1999. Changes in a prawn population due to tidal and day/night effects in a Matang mangrove river. Pages 92–102in V. C. Chong and P. S. Choo, eds. Proc. Third JIRCAS Seminar on Productivity and Sustainable Utilization of Brackishwater Man- grove Ecosystems, 8–9 Dec. 1997, University of Malaya, Kuala Lumpur. Japan International Center for Agricultural Sciences, Tsukuba, Japan. ______, ______, L. H. S. Lim, and S. Hayase. 1999. Prawn production of Matang and Dind- ing River Mangroves: Species distribution and seasonal recruitment. Pages 89–101 in K. Kiso and P. S. Choo, eds. Proc. Fourth JIRCAS Seminar on Productivity and Sustainable Utilization of Brackishwater Mangrove Ecosystems, 8–9 December 1998, Penang, Malay- sia, Japan International Center for Agricultural Sciences, Tsukuba, Japan. Marshall, N. 1994. Mangrove conservation in relation to overall environmental considerations. Pages 303–310 in A. Sasekumar, N. Marshall, and D. J. Macintosh, eds. Ecology and conservation of southeast asian marine and freshwater environment including wetlands. Hydro- biologia 285. Kluwer Academic Publishers, The Netherlands. Martosubroto P. and N. Naamin. 1977. Relationship between tidal forests (mangroves) and commercial shirmp production in Indonesia. Mar. Res. Indones. 18: 81–86. MPP-EAS. 1999. Total Economic Valuation: Coastal and Marine Resources in the Straits of Malacca. MPP-EAS Technical Report No. 24, GEF/UNDP/IMO, Philippines. Muhammad Ali, B. S. H. 2004. Diversity and abundance of macrobenthos in the mangrove estuaries of Matang and Dinding, Peninsular Malaysia. Unpublished M.Sc Thesis, Faculty of Science, University of Malaya, Kuala Lumpur, Malaysia. 129 p. ______, V. C. Chong, and A. Sasekumar. 1999. Benthic macrofauna distribution in the Sungai Selinsing, Matang Mangrove Forest Reserve, Malaysia. Pages 36–48 in K. Kiso and P. S. Choo, eds. Productivity and Sustainable Utilization of Brackish Water Mangrove Ecosystem, Proc. 4th Seminar on Results for 1997/98 Research Projects, Japan International Research Center for Agricultural Sciences, Tsukuba, Japan. Newell, R. I. E., N. Marshall, A. Sasekumar, and V. C. Chong. 1995. Relative importance of benthic microalgae, phytoplankton, and mangroves as sources of nutrition for penaeid prawns and other coastal invertebrates from Malaysia. Mar. Biol. 123: 595–606. Odum, W. E. and E. J. Heald. 1975. The detritus-based food web of an estuarine mangrove community. Pages 265–286 in L. E. Cronin, ed. Estuarine Research, Academic Press Inc., New York. Ogawa, Y. 2003. Migration, growth and feeding habit of John’s snapper Lutjanus johnii and duskytail grouper Epinephelus bleekeri in Merbok mangrove brackish river. Pages 1–10 in Y. Ogawa, H. Y. Ogata, Y. Maeno, Y. Shimoda, T. Fujioka, and Y. Fukuda, eds. Sustainable production systems of aquatic animals in brackish mangrove areas. JIRCAS Working Re- port No. 35, Japan International Center for Agricultural Sciences, Tsukuba, Japan. Ong, J. E., W. K. Gong, C. H. Wong, and G. Dhanarajan. 1984. Contribution of aquatic productivity in a managed mangrove ecosystem in Malaysia. Pages 209–215 in E. Soepadmo, A. N. Rao, and D. J. Macintosh, eds. Proc. UNESCO Asian Symp. on Mangrove Environment — Research and Management: University of Malaya, Kuala Lumpur, Malaysia. Ardyas Pub- lication, Kuala Lumpur. CHONG: MANGROVE-FISHERIES LINKAGES IN MALAYSIA 771

Ooi, A. L., V. C. Chong, Y. Hanamura, and Y. Konishi. 2007. Occurrence and recruitment of fish larvae in Matang mangrove estuary, Malaysia. JIRCAS Working Report. Paw J. N. and T. E. Chua. 1989. An assessment of the ecological and economic impact of mangrove conversion in Southeast Asia. Mar. Pollut. Bull. 20: 335–343. Peterson, B. J., R. W. Howarth, and R. H.Garritt. 1985. Multiple stable isotopes used to trace the flow of organic matter in estuarine food webs. Science 227: 1361–1363. Primavera, J. H. 1996. Stable carbon and nitrogen isotope ratios of penaeid juveniles and primary producers in a riverine mangrove in Guimaras, Philippines. Bull. Mar. Sci. 58: 675–683. ______. 1997. Fish predation on mangrove-associated penaeids: The role of structures and substrate. J. Exp. Mar. Biol. Ecol. 215: 205–216. Robertson, A. I. 1993. Fish, prawn and mangroves: patterns and processes. Pages 114–130 in A. Sasekumar, ed. Mangroves fisheries and connections. ASEAN-Australia Marine Science Project on Living Coastal Resources (Malaysia), Ipoh, Malaysia. Ministry of Science, Tech- nology and the Environment, Malaysia/Australia International Development Assistance Bureau. ______and S. J. M. Blaber. 1992. Plankton, epibenthos and fish communities. Pages 173–224 in A. I. Robertson and D. M. Alongi, eds. Tropical mangrove ecosystems. Coastal and Estuarine Studies 41. American Geophysical Union, Washington. ______, D. M. Alongi, and K. G. Boto. 1992. Food chains and carbon fluxes. Pages 43–62 in A. I. Robertson and D. M. Alongi, eds. Tropical mangrove ecosystems. Coastal and Es- tuarine Studies 41. American Geophysical Union, Washington. Rodelli, M. R., J. N. Gearing, P. J. Gearing, N. Marshall, and A. Sasekumar. 1984. Stable isotope ratio as a tracer of mangrove carbon in Malaysian ecosystems. Oecologia 61: 326–333. Ronnback, P., M. Troell, N. Kautsky, and J. H. Primavera. 1999. Distribution pattern of shrimps and fish amongAvicennia and Rhizophora microhabitats in the Pagbilao mangroves, Philip- pines. Estuar. Coast. Shelf Sci. 48: 223–234. Sasekumar, A. and V. C. Chong. 1987. Mangroves and prawns: Further perspectives. Pages 10–22 in A. Sasekumar, S. M. Phang, and E. L. Chong, eds. Towards conserving Malaysia’s heritage. Proc. 10th Annual Seminar of the Malaysian Society of Marine Sciences, Kuala Lumpur, Malaysia. ______and J. J. Loi. 1983. Litter production in three mangrove forest zones in the Malay peninsula. Aquat. Bot. 17: 283–290. ______, T. L. Ong, and T. K. Thong. 1984. Predation of mangrove fauna by marine fishes. Pages 378–384 in E. Soepadmo, E., A. N. Rao, and D. J. Macintosh, eds. Proc. Asian Symp. on Mangrove Environment: Research and Management, University of Malaya and UNES- CO, Ardyas Publication, Kuala Lumpur. ______, V. C. Chong, K. H. Lim, and H. R. Singh, 1994. The fish community of Matang mangrove waters. Pages 457–464 in S. Sudara, C. R. Wilkinson, and L. M. Chou, eds. Proc. Third- ASEAN-Australia Symp. on Living Coastal Resources, Vol. 2: Research Papers, Chulalong- korn University, Bangkok, Thailand. Shaharuddin, B. M. I., B. M. Azahar, B. M. Abdul Latiff, B. N. K. Nazir Khan, and K. L. Lim. 2004. Matang Mangroves: A Century of Sustainable Development. Sasyaz Holdings Private Limited, Petaling Jaya, Malaysia. 112 p. Singh, R. H. 2003. The biology of the estuarine catfishes (Fam: Ariidae) of the Matang mangrove ecosystem (Perak, Malaysia). Unpublished Ph.D. Diss., Faculty of Science, University of Malaya, Kuala Lumpur. 332 p. Spalding, M., F. Blasco, and C. Field. 1997. World Mangrove Atlas. The International Society for Mangrove Ecosystems, Okinawa, Japan. Staples, D. J., D. J. Vance, and D. S. Heales. 1985. Habitat requirements of juvenile penaeid prawns and their relationship to offshore fisheries. Pages 44–54in P. C. Rothlisberg, B. J. Hill, and D. J. Staples, eds. Second Australian National Prawn Seminar. Cleveland, Australia. Stoner, A. W. and R. J. Zimmerman. 1988. Food pathways associated with penaeid shrimps in a mangrove-fringed estuary. Fish. Bull. 86: 543–551. 772 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Tan, K. F. 1976. Some aspects on the biology of the Acetes erythraeus in the Sungei Sementa Besar, Malaysia. B.Sc. thesis, Department of Zoology, University of Malaya, Kuala Lumpur, Malaysia. 48 p. Thom, B. G. 1984. Coastal landforms and geomorphic processes. Pages 3–17in S. C. Snedaker and J. G. Snedaker, eds. The mangrove ecosystem: research methods. UNESCO, Paris. Vance, D. J., M. D. E. Haywood, D. S. Heales, R. A. Kenyon, N. R. Loneragan, and R. C. Pendrey. 1996. How far do prawns and fish move into mangroves? Distribution of juvenile banana prawns Penaeus merguiensis and fish in a tropical mangrove forest in northern Australia. Mar. Ecol. Prog. Ser. 131: 115–124. ______, M. D. E. Haywood, D. S. Heales, R. A. Kenyon, N. R. Loneragan, and R. C. Pendrey. 2002. Distribution of juvenile penaeid prawns in mangrove forests in a tropical Austra- lian estuary, with particular reference to Penaeus merguiensis. Mar. Ecol. Prog. Ser. 228: 165–177. Yap, Y. N., A. Sasekumar, and V. C. Chong. 1994. Sciaenid fishes of the Matang mangrove waters. Pages 491–498 in S. Sudara, C. R. Wilkinson, and L. M. Chou, eds. Proc. Third- ASEAN-Australia Symp. on Living Coastal Resources, Vol. 2: Research Papers: Chulalong- korn University, Bangkok, Thailand.

Address: Institute of Biological Sciences, Faculty of Science, University of Malaya, 50603 Kuala Lumpur, Malaysia. E-mail: . BULLETIN OF MARINE SCIENCE, 80(3): 773–793, 2007

Relationships between estuarine habitats and coastal fisheries in Queensland, Australia

Jan-Olaf Meynecke, Shing Yip Lee, Norman C. Duke, and Jan Warnken

Abstract Worldwide, estuaries have been recognized as critical habitats for nearshore fish productivity through their capacity as nursery grounds and nutrient sources. The purpose of this study was to demonstrate the importance of the habitat characteristics of estuaries to commercial fish catch in Queensland, Australia, with particular focus on the role of mangrove, saltmarsh, and seagrass habitats, and their connectivity. Traditionally, such analyses have taken the single-habitat approach, i.e., assessing the value of individual habitat types. Combined occurrence of these habitats and their collective accessibility may better explain the importance of estuaries to nekton. A literature review identifies the role of estuaries as integrated systems for fisheries. ��Our study provides strong supportive evidence for habitat-based, not species-based, management of fisheries in Queensland. Outcomes from preliminary analyses in Queensland suggest that collective spatial characteristics of estuarine habitats such as size and structural connectivity significantly correlate with fish catch data with r2 values > 0.7 for 17 commercial species groups. The catch of one quarter of the investigated species was best explained by the presence of mud- and sandflats. An exploration of currently available data on habitat distribution and fisheries catch shows the need to scrutinise their spatial and temporal accuracy, and how best to use them to understand estuarine-fisheries links. We conclude that structural connectivity of estuarine habitats is fundamental to the size of fish stocks and to optimizing the sustainable yield for commercial and recreational fishers.

Estuaries play an important, often essential, role in the life histories of many aquatic organisms (Blaber, 2000), including fish species of importance to indigenous, commercial, and recreational fishers D( unning et al., 2001). In Australia, estuarine ecosystems comprising mangroves, seagrasses, and shallow-water areas provide critical habitats to about 70%–75% of fish and crustacean species in the fisheries of Queensland (Quinn, 1993) and New South Wales (Pollard, 1981). This value is lower for south-western Australia (20%) and Australia as a whole (32%), which are less dominated by estuary-dependent species (Lenanton and Potter, 1987). Australia’s total annual economic value of the fishery industry in 2002 was AUD$7.4 billion (Williams, 2002) including the gross value of production (GVP) of AUD$2.3 billion (ABARE, 2005). The fisheries production from some coastal estuaries and lagoons is estimated at up to 3.3 t km–2 yr–1 (Pollard, 1994). The importance of estuarine habitats rests not only with their extent, but also their combined occurrence and relation to each other. Attributes of the estuarine “seascape” may also be critical for sustaining productivity of coastal fisheries. We expect that estuary-dependent species catch would increase with the size of connected estuarine habitats. Previous studies have found positive correlations between mangrove extent and fisheries catch (Baran et al., 1999; Manson et al., 2005b) only a few have examined multiple habitats (Saintilan, 2004; Lee, 2005). Most estuary-dependent species use a wide range of habitats during their life history. Not all habitats contribute equally as nurseries (Beck et al., 2001), but comparison of their values has rarely been rigorously undertaken.

The purpose of this study is to demonstrate the importance of estuarine habitat characteristics to commercial fish catch, with particular focus on the role of mangroves, saltmarsh, seagrass, and mud- and sandflats as connected habitats driving observed fluctuations in fish catch. The connectivity of habitats can be a key factor in fish production by allowing exchange of energy and organisms (Merriam, 1984) among estuarine habitats. This also includes “edge effects”, which reflect changes in ecological factors at the boundary between habitats, thus supporting the “chain of habitats” concept as a useful management system for Australian and worldwide species (e.g., Nagelkerken et al., 2001). Here we: (a) briefly analyze current knowledge and literature in regard to the importance and role of estuaries for fisheries; (b) examine deficiencies of studies that address the relationship between estuarine habitats and fish catch; (c) present information available for assessing the linkage between estuaries and commercially important fish species; (d) provide a case study relating combined occurrence of Queensland estuarine habitats with fish catch data using multiple regression models and non-metric multidimensional scaling (nMDS); and (e) discuss ways of assessing the data and benefits for fisheries management. An Overview of Studies Linking Estuarine Characteristics and Fisher- ies Using Fish Catch Data.—The importance of biotic and abiotic characteristics of estuaries to fish assemblages has been demonstrated in a number of studies H( orn and Allen, 1976; Monaco et al., 1992; Pease, 1999; Blaber, 2000). More specifically, estuary size characteristics have been demonstrated to be correlated with fish assemblages, e.g., mouth width (Horn and Allen, 1976); mouth depth (Monaco et al., 1992), and water area (Pease, 1999). Pollard (1994) attempted to collate and contrast the historical data for the south coast estuaries of New South Wales, Australia. He found that only rank abundance could be derived from this information for comparative purposes. The outcomes were affirmed in analyses undertaken by Pease (1999) and Saintilan (2004). These studies used similar multivariate techniques to Stergiou and Pollard (1994) and were based on commercial fisheries data for New South Wales. Interactions between abiotic and biotic estuarine characteristics and fish assemblages have been widely investigated in temperate estuaries (Perkins, 1974; Barnes, 1980; Day et al., 1989; Pease, 1999). A number of studies comparing commercial fish catch with estuarine habitats, in particular, mangroves and prawns in tropical and subtropical zones, have been completed in the last few decades. Most of these studies showed positive relationships, assuming that the area of tidal wetland habitat translates to the secondary production and catch of commercial fisheries (Baran and Hambrey, 1998; Baran et al., 1999; Manson et al., 2005a). Rönnbäck (1999), for example, listed the proportion of mangrove-related species in fisheries around the world giving a figure of 67% for eastern Australia. In Malaysia, it was estimated that 32% of the biomass (total fish catch) of the 1981 fish harvest could be linked to mangroves; while in theP hilippines, about 72% of the catch between 1982 and 1986 was associated with mangroves (Paw and Chua, 1991). Analyses of estuaries and fish catch data have demonstrated the potential of data sets available for Australia (Pease, 1999; Duffy et al., 2003; Saintilan, 2004; Manson et al., 2005b). Several studies have found correlations between the extent of mangroves and the catch in nearby fisheries (Staples et al., 1985; Manson et al., 2005b; Table 1). MEYNECKE ET AL.: ESTUARIES AND COASTAL FISHERIES IN QUEENSLAND 775

Table 1. Overview of studies investigating relationships between estuarine habitat distribution and fisheries production for prawn (+) or fish (++) or both if not indicated in the last three decades (Baran, 1999; Manson et al., 2005a). Not reported is shown as “nr”. The r2 value indicates goodness of fit of a linear model between habitat and catch. r2 (n) Region Variable Reference Positive Malaysia Mangrove area Macnae, 1974 0.54 (27)+; 0.64 (14)+ World, tropical Intertidal area Turner, 1977 0.89 (nr)+ Indonesia Mangrove area Martosubroto and Naamin, 1977 0.58 (6)+ Gulf of Carpentaria, Linear extent of Staples et al., 1985 Australia mangroves 0.48 (10)++ Gulf of Mexico Coastal vegetation area Yáñez-Arancibia, 1985 0.53 (nr)++ World, tropical Mangrove area Pauly and Ingles, 1986 Positive Philippines Mangrove area Camacho and Bagarinao, 1987 0.89 (10)+ Peninsular Malaysia Mangrove area Sasekumar and Chong, 1987 0.32 (nr)+ U.S.A Salt marsh length Browder et al, 1989 interface 0.61 (18)+; 0.66 (18)+; Philippines Mangrove area Paw and Chua, 1991 0.34 (15)++; 0.88 (5)++ 0.53 (18); 0.40 (20); 0.66 (12); 0.40 (18); 0.4 (34); 0.45 (39); 0.95 (nr) 0.95 (nr); 0.88 (5)++ Vietnam Mangrove area de Graaf and Xuan, 1998 Positive++ Philippines Mangrove area Gilbert and Janssen, 1998 0.38 (37)++ World, tropical Coastline length, Lee, 2004 mangrove area, tidal amplitude 0.32–0.75 (49) NewSouthWales Total area of mangrove, Saintilan, 2004 Australia saltmarsh, seagrass 0.46–0.63 (8) Malaysia Mangrove area Loneragan et al, 2005 0.37–0.70 (36)+; East coast Queensland, Mangrove perimeter, Manson et al., 2005b 0.57–0.77 (36)++ Australia area shallow water, mangrove area/length coast line

Lee (2004) suggested that the amount of intertidal habitats and energy available for material exchange, as indicated by tidal amplitude, rather than just the area of mangroves functions as a major driver for prawn production. The correlation of mangrove, saltmarsh, seagrass distribution as well as channels and flats in estuaries supports the idea of a link between these habitats. A study by Saintilan (2004) found a strong correlation between mangroves and saltmarsh distribution (correlation coefficient 0.79) for estuaries in New South Wales. Although studies have documented greater abundances of juvenile species in mangroves compared to other estuarine and inshore habitats (e.g., in Australia; Laegdsgaard and Johnson, 2001), other estuarine habitats and the strong links between them have been neglected (Sheridan and Hays, 2003). However, the continuity and interdependence of riverine, estuarine, and marine environments is an ecological reality for coastal fish resources (Baran et al., 1999). To maintain fisheries production, the combination of estuarine habitats that are utilized by different species and their various life history stages must be taken into account to ensure the inclusion of all resources required by these species (Lee, 776 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

2004). It is essential to determine the dependency of fish resources on estuarine environments by addressing the following questions regarding the capacity of any given species in completing its life cycle: (1) is the estuarine zone essential?, (2) how strong is the dependency on habitat?, and (3) which habitat characteristics are critical?

Methods

Data Collection.—Data on catch, effort (number of days and boats), and gross value of production for estuary-dependent species or species groups were provided by the Depart- ment of Primary Industries and Fisheries, Queensland (DPI&F) Assessment and Monitoring Unit (Table 2). This data set is based on daily logbook records reported by commercial fishers providing details of their catch and effort, covering the years 1988–2004, and recorded in half-degree grids (30-nmi) for the entire coast of Queensland. Data from each of the four fisheries (trawl, line, net, and pot) can be distinguished. The trawl fishery for prawns has two components: a within-estuary (river) beam-trawl fishery, and an offshore (and coastal fore- shore) otter-trawl fishery.H owever, the Gulf of Carpentaria has only an offshore trawl fishery (Staunton-Smith et al., 2004). The fish catch data have an estimated error of 10% due to the type of recording, market fluctuations, policies, and management changes (L. Olyott,DP I and F, pers. comm.). Data on Queensland coastal wetland vegetation were obtained from DPI&F Assessment and Monitoring Unit. The 1:100,000 coastal wetland vegetation map includes information on mangrove communities, saltmarsh and open water based on Landsat TM images and ground truthing (1987–1999). Data on channels, intertidal flats, and sandflats for estuaries were taken from Geoscience Australia (Geoscience_Australia, 2004). This information was primarily sourced from the 1:100,000 scale National Topographic Map series produced by Geoscience Australia. Fish Catch Data.—The coast of Queensland was divided into 13 separate sections for the purpose of this study. These sections were selected because of their importance to the commercial fisheries (L. Olyott, DPI and F, pers. comm.), likely independence between the catches, and their value which represented two thirds of Queensland’s total fish catch [or 65% of the total fish catch in 2004 DP( I and F, 2005)]. Selection criteria for species were that they should be/have (1) relatively constant and high market values, (2) well known, (3) estuary-dependent, and (4) widespread throughout Queensland [based on Yearsley et al. (1999); L. Williams, DPI and F, pers. comm.]. Annual summaries of the fish catch data were calculated together with the following variables: latitudinal section, species or species group, fishery type, total catch (tonnes) and catch- per-unit-effort (CPUE, kg day–1). Daily catch data recorded from compulsory commercial fishing logbooks were standardized for fishing effort by dividing the catch of each species by the number of days fished in each section (Tanner and Liggins, 2000). We separated the four fisheries (trawl, net, pot, line) but also used total CPUE, as the majority of the techniques are considered passive and their effort is measured in days with one technique dominating the fish catch for a species group. We compared total fish catch with the habitat parameters and did the same for CPUE for all fisheries and for individual fisheries to see whether different combinations would give similar outcomes. Yearly catch values for a total of 31 species or species groups were used. The fish species were selected according to their estuary dependence (Table 2). These species are mainly “marine-estuarine species” [see alsoW hitfield (1999) for a detailed classification], which use inshore areas and estuaries for significant periods of time, often (but not limited to) during their juvenile phase. Several marine-estuarine species have juveniles that are only found within estuarine habitats (e.g., Penaeus merguiensis De Man, 1898) (Staples et al., 1985; Vance et al., 1996). Some catadromous species travelling between freshwater and marine habitats use estuarine habitats at certain life-stages, e.g., Lates calcarifer (Bloch, 1790; Russell and Garrett, 1983). The fish species groups have been divided into dif- MEYNECKE ET AL.: ESTUARIES AND COASTAL FISHERIES IN QUEENSLAND 777

Table 2. Selected species for the analyses with the total catch of 13 selected geographical areas along the coast of Queensland from 1988–2004 and suggested habitat dependence as described in Williams (2002). Channels and saltmarsh are not included. FL–Mud- and Sandflats, SG–Seagrass, MG–Mangrove, CH–Channels, SM–Saltmarsh.

Common name and fish Taxa Habitat Catch (t) catch class Barramundi Lates calcarifer (Bloch, 1790) MG 8,212 Bream Monodactylus argenteus (Linnaeus, 1758), MG, FL, SG 10,870 Pomadasys maculatum (Bloch ,1793), Acanthopagrus australis (Akazaki, 1984), Acanthopagrus berda (Forsskål, 1775), Plectorhinchus gibbosus (Lacépède, 1802), Sparidae spp. Bugs Thenus indicus (Leach, 1815), FL 2,921 Thenus orientalis (Lund, 1793) Blue Swimmer Crab Portunus pelagicus (Fox, 1924) FL 5,277 Mud Crab Scylla serrata (de Haan, 1833) MG 513 Dart Trachinotus anak (Ogilby, 1909), FL 1,025 Trachinotus blochii (Lacépède, 1801), Trachinotus botla (Shaw, 1803), Trachinotus spp. Flathead Platycephalus fuscus (Cuvier, 1829), SG, FL 6 Platycephalus spp. Flounder Pseudorhombus jenynsii (Bleeker, 1855), FL 465 Pseudorhombus arsius (Amaoka, 1969), Pseudorhombus spinosus (McCulloch, 1914) Grunter Hephaestus fuliginosus (Macleay, 1883), FL 14 Pomadasys spp. Milkfish Chanos chanos (Forsskål, 1775) MG 7,954 Mullet Liza vaigiensis (Quoy and Gaimard, 1825), FL 28,447 Liza subviridis (Valenciennes, 1836), Liza argentea (Quoy and Gaimard, 1825), Valamugil georgii (Bleeker, 1858), Valamugil seheli (Forsskål, 1775), Mugil cephalus (Linnaeus, 1758), Parupenaeus spp. Trachystoma petardi (Castelnau, 1875), Mugilidae spp. Prawns-bait Family Penaeidae – 152 Prawns-banana Fenneropenaeus indicus (Milne-Edwards, 1837), MG, FL 7,801 Fenneropenaeus merguiensis (de Man, 1888) Prawns-bay Metapenaeus macleayi (Haswell, 1879), – 5,906 Metapenaeus insolitus (Racek and Dall, 1965) Prawns-endeavour Metapenaeus endeavouri (Schmitt, 1926), SG 11,803 Metapenaeus ensis (De Haan, 1844) Prawns-greasy Metapenaeus bennettae (Racek and Dall, 1965) MG, FL 1,415 Prawns-king Penaeus monodon (Fabricius, 1798), SG 27,389 Penaeus semisulcatus (De Haan, 1844) Prawns-school Metapenaeus macleayi (Haswell, 1879) SG 654 Prawns-tiger Penaeus esculentus (Haswell, 1879) SG 23,438 Prawns-unspecified Penaeidae spp. – 69 778 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Table 2. Continued.

Common name and fish Taxa Habitat Catch (t) catch class Rays Gymnura australis (Ramsay and Ogilby, 1886), FL 142 Himantura toshi (Whitley, 1939), Myliobatis australis (Macleay, 1881), Urolophus paucimaculatus (Dixon, 1969), Rhynchobatus djiddensis (Forsskål, 1775), Dasyatis kuhlii (Müller and Henle, 1841), Aetobatus narinari (Euphrasen, 1790, Trygonorrhina sp. Dasyatidae spp. Sawfish-unspecified Pristis zijsron (Bleeker, 1851), FL 13 Pristidae spp. Sea Perch-mixed Lutjanidae spp. MG 27 Sea Perch-mangrove jack Lutjanus argentimaculatus (Forsskål, 1775) MG 1,046 Snapper Pagrus auratus (Foster, 1801), SG, FL 726 Etelis carbunculus (Cuvier, 1828), Caesionidae spp., Lutjanidae spp. Tailor Pomatomus saltatrix (Linneo, 1766) FL 2,599 Tarwhine Rhabdosargus sarba (Forsskål, 1775) – 10 Threadfin-blue Eleutheronema tetradactylum (Shaw, 1804) FL 33 Threadfin-king Polydactylus macrochir sheridani (Macleay, 1884) MG 2,045 Threadfin-unspecified Polynemidae spp. – 5,226 Whiting Sillago ciliata (Cuvier, 1829), MG, FL 17,739 Sillago analis (Whitley, 1943), Sillago maculata (Quoy and Gaimard, 1824), Sillago burrus (Richardson, 1842), Sillago ingenuua (McKay, 1985), Sillago sihama (Forsskål, 1775), Sillago robusta (Stead, 1908), Sillaginidae spp. ferent categories according to their known or suggested habitat requirements. These groups were then tested for their dependence using forward stepwise regression with the respective fish species as the dependent variable to identify any relationship between the geomorphic variables and fisheries catches. Geomorphic Data.—The GIS software ArcGis v. 9.0 was used to access and visualize the estuarine habitat data and to produce raster maps of the habitats with a 10 × 10 m grid size. TheDP I and F 1:100,000 vector data of the coastal wetland vegetation were combined with intertidal vector data from Geoscience Australia. The resulting vector data layer was converted to a 10 × 10 m grid for each for the 13 regions applying the assigned GDA 94 projection for each section using ArcGis v. 9.0. Each grid layer was analyzed with ��the spatial metrics analysis package Fragstats 3.3 (McGarigal et al., 2002) to calculate the total area (CA), perimeter (Peri), mean perimeter to area ratio (PARA), number of patches (NP) and a connectivity index (CONNECT), giving a value for the distances between patches (threshold 100 m) within the 13 separate sections. The connectivity index (CONNECT) is defined as:

R n V S / cijk W CONNECT = S jk= W # 100 S nnii- 1 W S ]g2 W T X MEYNECKE ET AL.: ESTUARIES AND COASTAL FISHERIES IN QUEENSLAND 779

The index is equal to the number of functional connections between all patches of a particular patch type (Σ cijk, where cijk = 0 if patches j and k are not within the specified distance of 100 m of each other and cijk = 1 if patch j and k are within the specified distance of 100 m), divided by the total number of possible connections between all patches of the corresponding patch type, and converted to a percentage (Clarke and Warwick, 2001). We used the lowest possible threshold to avoid the inclusion of habitat patches that were divided by a barrier (e.g., terrestrial vegetation). The perimeter to area ratio was measured for each habitat as a mean which equals the sum, across all patches of the corresponding patch type and patch metric values, divided by the number of patches (McGarigal et al., 2002). Over 140 estuaries from > 300 recognized Queensland estuaries were covered, ranging from the Gulf of Carpentaria (17°S, 141°E) to Coolangatta (28°S, 153°E). The following classes were used: channels, mangroves, saltmarsh, seagrass, combined mud- and sandflats, total number of estuaries, and latitude per section. Mangroves, saltmarsh, channel, mud- and sandflats were calculated separately for total wetland parameters to look at the overall effect of the wetland on fish catch data. Seagrass was not included due to its overlapping with the channel habitats on the digital map, which did not allow its separate calculation. The habitat parameter of each geographical area was assigned the corresponding grid code from the DPI and F fishery grid (Fig. 1). Data Analysis.—Statistical analyses were carried out using the PRIMER 5.0 (Clarke and Ainsworth, 1993) and SPSS 12.01 software packages. Correlation analyses and non-metric multidimensional scaling (nMDS) were used to explore dependency and similarity between perimeter, area, mean perimeter to area ratio, and connectivity of mangroves, saltmarsh, seagrass, mud- and sandflats, using Bray Curtis similarity and Euclidean distance. Data were square root transformed prior to analysis to normalize the variances. CPUE variables were not transformed whereas other variables were square root, fourth root or log10 transformed to reduce the right-skewness of the data. nMDS was used to represent the similarity of estuaries on the basis of habitats using Euclidean distance as a common method for modeling habitat distances (Conner et al., 2002). The extent to which habitat variables explained the variability in fish catch and CPUE was determined using the BIO-ENV procedure (Clarke and Ainsworth, 1993). This procedure generated various similarity matrices from subsets of the environmental variables and displayed the best of these various correlations (Mantel’s tests). In addition, multiple regression models were used to investigate which of the estuarine habitats accounted for the variation in fisheries production throughout the region. For this approach the CPUE for each fishery as well as an averaged CPUE for all fisheries was applied. In order to exclude auto-correlation within the data (Pyper and Peterman, 1998), results from correlation analysis were reviewed for consistency with the theoretical mechanisms proposed before the analysis, with the expectation that high correlation exists between habitats themselves and estuarine fish catch. Further considerations were given to “outliers”, where species groups had less than 100 t catch in more than half of the investigated sections. These groups were excluded from the later analysis.

Results

Fisheries and Environmental Data.—The total estuarine fish catch from all 13 geographical areas between 1988 and 2004 was 174,000 t compared to 255,000 t for the whole coast of Queensland (Fig. 2). There was little overlap in the species caught in the four fisheries, except for blue swimmer crabs Portunus pelagicus (Linnaeus, 1758), which were caught in both the pot (> 60% of the catch) and trawl fisheries, and whiting (Sillago spp.) which were caught in the trawl (> 50% of the catch) specifically directed towards stout whiting (Sillago robusta Stead, 1908) and net fisheries. There have been differences in catches throughout the regions with the highest catch in Moreton Bay and Fraser Island, which also had the highest effort. The highest CPUE for all fisheries (pot, line, net, and trawl) was in the Moreton Bay and Fraser region, 780 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Figure 1. Location of investigated area showing major cities and river systems in Queensland and selected geographical areas with fish catch grids for preliminary analyses of the study (modified from Geoscience Australia, 2004). where mullet (e.g., Mugil cephalus Linnaeus, 1758) contributed 28% and whiting (Sil- lago spp.) 12% towards the total catch. The CPUE in the line fishery was high in the central region (e.g., Hinchinbrook) and also in the north where effort was in general low and for some species (e.g., Platycephalus spp.), too low to provide a meaningful CPUE. The net fishery dominated in the Gulf regions where the total catch and effort are low. The prawn fishery is not included within inshore grids as the trawl fishery in the Gulf of Carpentaria is undertaken only offshore. Barramundi L.( calcarifer) and mud crab Scylla serrata (Forskål, 1775) catches accounted for more than half of the total catch in these regions. MEYNECKE ET AL.: ESTUARIES AND COASTAL FISHERIES IN QUEENSLAND 781

Figure 2. Total catch of 31 estuary-dependent species groups for 13 selected regions in Queensland, Australia, between 1988–2004 and their extent of estuarine habitats. A large range of different estuarine habitat characteristics was covered by the selected geographical areas (Fig. 2). One extreme occurred in the north with the Albert River section having a large area of saltmarsh/saltpan and high number of mangrove patches and the other extreme at the southern end of Queensland with Moreton Bay providing large areas of flats, channel perimeter, and seagrass perimeter. Large areas of mangrove occurred in the Hinchinbrook region whereas the Burdekin region had a high number of channel and patches of flats. The Fraser region dominated in the 782 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Table 3. Data on selected environmental variables for 13 locations for wetland components of: channels, flats (mud and sand), mangroves, saltmarsh, seagrass, and total wetlands. Variables include: area in km2, perimeter in km, number of patches and estuaries, mean perimeter to area ration (Mn_P:A) in km, and connectivity in %. Locations include: Al–Albert River, Ar–Archer Bay, Bu–Burdekin River, Ca–Cairns, Fi–Fitzroy River, Fr–Fraser, Hi–Hinchinbrook, Mi–Mitchell River, Mo–Moreton Bay, No–Normanby, Pa–Pascoe, We–Wenlock, Wh–Whitsundays.

Parameters Al Ar Bu Ca Fi Fr Latitude 18 14 20 16 24 25 Estuaries 7 5 11 13 12 6 Channel area 11 6 20 3 26 366 Perimeter 1,412 813 2,132 483 1,659 1,828 Patches 27 16 431 22 23 186 Mn_P:A 2,193 276 333 1,167 779 115 Connectivity 3.70 0.83 0.07 3.03 1.58 0.00 Flats area 15 7 23 2 8 51 Perimeter 400 888 2,409 231 629 2,737 Patches 33 104 486 58 117 138 Mn_P:A 321 262 355 719 503 443 Connectivity 0.57 0.15 0.07 0.36 0.22 0.30 Mangrove area 25 5 20 8 24 15 Perimeter 5,720 763 2,132 776 4,478 2,038 Patches 2,471 312 431 126 1,281 544 Mn_P:A 447 342 333 256 432 367 Connectivity 0.01 0.05 0.07 0.46 0.03 0.06 Saltmarsh area 280 8 23 1 35 6 Perimeter 11,760 1,221 2,409 132 4,081 1,421 Patches 1,016 458 486 75 1,032 726 Mn_P:A 463 331 355 334 404 411 Connectivity 0.07 0.08 0.07 0.14 0.03 0.03 Seagrass area 4 4 4 3 2 132 Perimeter 123 112 157 95 126 292 Patches 8 6 28 18 29 9 Mn_P:A 36 49 140 248 158 146 Connectivity 0 0 0 0 0 0 Wetlands area 358 27 66 13 92 434 Perimeter 5,622 2,175 3,320 954 3,841 2,623 Patches 514 368 328 70 210 113 Mn_P:A 461 344 382 323 443 433 Connectivity 0.09 0.12 0.08 0.54 0.29 0.51 channel category with Hervey Bay having large areas of seagrass. The Mitchell River section had the highest channel connectivity index and Normanby River, the highest flats connectivity index (Table 3). Furthermore, we detected significant positive correlation between estuarine habitat variables. Out of 304 estuaries and their adjacent habitats in Queensland, we found significant positive correlation between mangroves and saltmarsh area (r = 0.41, P < 0.01, n = 304) and between seagrass and mangrove area (r = 0.40, P < 0.01, n = 304) (Table 4), suggesting that these habitats mostly occur together and may be regarded as a network of adjacent ecosystems. MEYNECKE ET AL.: ESTUARIES AND COASTAL FISHERIES IN QUEENSLAND 783

Table 3. Continued.

Parameters Hi Mi Mo No Pa We Wh Latitude 19 16 27 15 12 12 21 Estuaries 21 4 21 6 6 8 22 Channel area 19 6 16 3 3 15 7 Perimeter 329 1,300 1,910 627 441 1,344 1,177 Patches 114 8 129 14 10 12 55 Mn_P:A 2,882 1,407 1,299 392 899 331 1,078 Connectivity 1.41 10.8 0.57 0.00 2.22 0.00 1.28 Flats area 5 2 73 3 2 11 12 Perimeter 70 314 3,290 175 154 645 1,109 Patches 133 64 344 20 42 82 240 Mn_P:A 527 555 680 250 540 337 445 Connectivity 0.21 0.25 0.15 1.05 0.46 0.39 0.24 Mangrove area 30 6 15 10 14 27 21 Perimeter 2,584 1,128 2,026 1,647 1,114 3,341 2,420 Patches 280 618 539 688 186 725 512 Mn_P:A 290 344 315 335 282 281 359 Connectivity 0.17 0.02 0.06 0.03 0.16 0.04 0.08 Saltmarsh area 3 37 3 32 1 7 5 Perimeter 520 4,842 632 3,315 350 1,526 1,040 Patches 251 1,157 447 833 196 838 601 Mn_P:A 360 310 422 336 339 347 425 Connectivity 0.05 0.03 0.04 0.04 0.14 0.03 0.03 Seagrass area 8 0 25 47 14 5 7 Perimeter 400 0 1,307 409 250 129 409 Patches 59 0 277 14 15 14 105 Mn_P:A 96 0 289 87 32 61 155 Connectivity 0 0 0.05 1.10 0 0 0.05 Wetlands area 56 53 208 51 20 61 44 Perimeter 2,660 5,810 2,664 3,432 1,194 4,369 2,575 Patches 108 914 73 755 125 626 242 Mn_P:A 355 337 504 343 293 342 385 Connectivity 0.33 0.04 0.76 0.04 0.19 0.07 0.11

Relationships Between Estuarine Habitats Total Fish Catch and Cpue Data.—We compared total catch with different variables of the data set and found the best fit between section data of total fish catch and the wetland connectivity index (r2 = 0.62) (Fig. 3). Furthermore, total catch correlated with the wetland mean perimeter to area ratio (r2 = 0.45) and averaged CPUE for all fisheries correlated with total wetland area (r2 = 0.55). An nMDS based on square root transformed total fish catch data using Bray Curtis similarity showed a clear grouping of three classes for combined CPUE: (1) South eastern Queensland; (2) Central and Northern Queensland; and (3) the Gulf of Car- pentaria (Fig. 4A). The nMDS using standardized habitat parameters and applying Euclidean distance as a similarity measure resulted in some overlap with the nMDS based on standardized catch with the following groups: (1) Fraser and Moreton Bay (2) Whitsundays, Hinchinbrook, Cairns, Pascoe, and Wenlock (3) Albert, Mitchell, and Normanby. The Fitzroy, Archer Bay, and Burdekin sections did not show clear 784 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Table 4. Pearson correlation of different estuarine habitat types in Queensland for 304 estuaries. P < 0.05*, P < 0.01**. Habitats Saltmarsh Seagrass Flats Channels Mangroves .411** .403** .576** .627** Saltmarsh – .120* .180* Seagrass .224** .897** Flats .499**

grouping. The distribution of wetland connectivity is an important factor separating the regions, with the Fraser and Moreton Bay region having the highest connectivity index (Fig. 4B). The highest correlation was found in BIO-ENV (0.729) between CPUE for all fisheries and the wetland mean perimeter to area ratio, followed in rank order by the wetlands and flats connectivity index, and the total number of estuaries. Stepwise Multiple Regression Analysis of Key Species.—The most important environmental parameter for nine out of 24 species or species groups for predicting CPUE with stepwise regression was the total area of mud- and sandflats as well as patches and their connectedness, followed by the wetlands parameters, which gave good CPUE prediction for another four species. The fitted model for flats, for- ex ample, accounted for 70%–90% of the variation in CPUE of mullet (e.g., M. cephalus), blue swimmer crab (P. pelagicus), and dart (Trachinotus spp.) (Table 5). Mangrove parameters were less important than expected for the mangrove related species such as barramundi (L. calcarifer) and mud crabs (S. serrata) [for comparison see Manson et al. (2005b)], for which total wetland and seagrass parameters gave the best fit, al- beit weak. This fit was, however, rather weak. Prawn CPUE data were in general best explained by the presence of flats, channels and total wetland parameter. For the seagrass related group it was wetlands or other parameters. Seagrass parameters did not emerge as significant.

Figure 3. Relation between wetland connectivity and total catch from 13 geographical areas along the coast of Queensland (catch has been fourth root transformed). MEYNECKE ET AL.: ESTUARIES AND COASTAL FISHERIES IN QUEENSLAND 785

Figure 4. nMDS for 13 selected geographical areas (A) (based on untransformed standardized habitat parameters and latitude using Euclidean distance) showing the three groups: Southern Queensland, Central, and Northern Queensland, and the Gulf of Carpentaria; and (B) (based on square root transformed catch data and Bray Curtis similarity) in Queensland, Australia. The value of the wetland connectivity index is indicated by the size of circles. The models identified mangrove parameters (connectivity) as being important in predicting CPUE for whiting (Sillago spp.) and greasy prawns (Metapenaeus bennettae Racek and Dall, 1965). Latitude was the only parameter fitted to the model for king prawns (Penaeus monodon Fabricius, 1798, Penaeus semisulcatus de Haan, 1844) that explained the variation in CPUE of this species group, accounting for 56% of variation (Table 5). The total number of estuaries per geographical area was only important as a predictor for CPUE for banana prawns, e.g., Fenneropenaeus indicus (H. Milne Edwards, 1837). Some species groups did not show any significant relationship with any of the variables [grunter, Pomadasys spp.; snapper, Pagrus auratus (Forster in Bloch and Schneider, 1801) or had no significant r2 value (flathead, Platy- cephalus spp.; flounder, Pseudorhombus spp.; rays, e.g., Gymnura australis (Ramsay and Ogilby, 1886); threadfin unspecified, Polynemidae; sawfish, Pristidae]. Their to- 786 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Table 5. Significant 2r values for the two most important variables in stepwise multiple regressions predicting CPUE for 24 species groups showing: A CPUE–all fisheries, N CPUE–net, T CPUE– trawl, C CPUE–pot, L CPUE–line. P < 0.05*, P < 0.01**.

Habitat Species Parameters Type Adjusted r2 Df MG Barramundi Seagrass NP, Wetlands NP A CPUE 0.684* 12 Seagrass Peri, Saltmarsh Peri L CPUE 0.648* 10 FL Blue Swimmer crab Flats CA, Channel CA A CPUE 0.896** 12 Channel CA, Seagrass NP C CPUE 0.935** 11 Wetlands CA, Saltmarsh PARA T CPUE 0.817* 8 MG, FL, SG Bream Flats CA, Channel NP A CPUE 0.861** 12 Flats CA, Channel connect N CPUE 0.900** 12 FL Bugs Flats NP A CPUE 0.535* 12 Flats NP, Seagrass Peri T CPUE 0.734* 8 FL Dart Flats CA, Wetlands CA A CPUE 0.944** 12 Flats CA, Wetlands CA N CPUE 0.913** 12 MG Milkfish Flats connect, Seagrass PARA, A CPUE 0.649* 12 MG Mud Crabs Flats connect, Mangrove CA, A CPUE 0.543* 12 Wetlands NP Wetlands NP, Mangrove CA, C CPUE 0.546* 12 Flats connect FL Mullet Flats CA A CPUE 0.743** 12 Flats CA N CPUE 0.789** 12 – Prawns, Bait Flats connect, Saltmarsh PERI A CPUE 0.553* 12 Flats connect T CPUE 0.658* 8 MG, FL Prawns, Banana Flats PARA, number of estuaries A CPUE 0.653* 11 Flats PARA T CPUE 0.522 8 – Prawns, Bay Flats NP A CPUE 0.634* 12 Flats NP T CPUE 0.915** 8 SG Prawns, Endeavour Wetlands Peri, Flats connect A CPUE 0.578* 12 Latitude T CPUE 0.591 8 MG, SG, FL Prawns, Greasy Mangrove connect, Saltmarsh connect A CPUE 0.701** 12 Channel Peri, Saltmarsh Peri T CPUE 0.958** 6 SG Prawns, King Latitude A CPUE 0.563* 12 Latitude T CPUE 0.492 8 SG Prawns, School Channel CA, Wetlands connect A CPUE 0.880** 12 Wetlands all CA, Saltmarsh connect T CPUE 0.956** 8 SG Prawns, Tiger Wetlands Peri A CPUE 0.293 12 Wetlands PARA, Saltmarsh CA T CPUE 0.938** 8 – Prawns unsp. Channel NP, Flats NP A CPUE 0.867** 12 Channel NP, Flats NP T CPUE 0.844* 8 MG Sea perch Wetlands NP A CPUE 0.750** 12 Channel CA, Channel PARA L CPUE 0.924** 12 MG Mangrove Jack Channel connect, Wetlands NP A CPUE 0.849** 12 Channel connect, Wetlands NP N CPUE 0.854** 12 FL Tailor Channel connect, Saltmarsh connect A CPUE 0.858** 12 Channel connect, Saltmarsh connect N CPUE 0.898** 12 – Tarwhine Flats CA A CPUE 0.622* 12 FL Threadfin Blue Seagrass NP A CPUE 0.885** 12 Channel Peri N CPUE 0.566* 12 MG Threadfin King Saltmarsh CA A CPUE 0.828** 12 Saltmarsh CA N CPUE 0.935** 12 MG, FL Whiting Wetlands all CA, Mangrove connect A CPUE 0.959** 10 Wetlands PARA, Seagrass CA N CPUE 0.773** 12 Channel CA, Wetlands connect T CPUE 0.995** 8 MEYNECKE ET AL.: ESTUARIES AND COASTAL FISHERIES IN QUEENSLAND 787 tal catch throughout the 13 geographical areas was generally low. When leaving out areas with no catch, we found fewer differences between CPUE for all fisheries and CPUE for the most important individual fisheries. The habitat to CPUE relation was relatively consistent throughout the different data sets.

Discussion and Conclusion

This study has shown an empirical link between estuarine habitat and fishery production for estuary-dependent species. Intertidal flats were one of the most important variables explaining fish catch variation in Queensland according to regression analyses for single species. These outcomes are similar to findings by Saintilan (2004) for New South Wales, Australia. The BIO-ENV results with CPUE for all fisheries further demonstrated that flats were one of the most important variables. This suggests that the significance of mud- and sandflats has been greatly underestimated in broad scale analyses to date. Mud- and sandflats are often in proximity to mangroves, salt marshes, and seagrass beds, suggesting connectivity. Larger fish of the same species use the flats and unvegetated areas in inshore and open waters of estuaries as feeding grounds (Chong et al., 2001; Laegdsgaard and Johnson, 2001). These ecosystems are an important habitat for larger fish species, have high microbial activity and large quantities of MPB (microphytobenthos). Mullet (Mugilidae spp.) for example feed on detritus, diatoms, algae, and small invertebrates that they filter from mud and sand (Williams, 2002). According to our results, their catch is best explained by the size of mud- and sandflats. udfludflatsM ats are common in tropical Australia with extensive oc-occurrence in northern Australia due to large tidal ranges. However, defining just one habitat as the major driver of CPUE from the data is misleading. Our results support the model that fish species depend on a number of habitats with the overall catch and CPUE being dependent on the whole estuarine wetland habitat suite rather than any single habitat. For example, mangrove jack Lutjanus argentimaculatus (Forsskål, 1775) is known to utilize all types of tidal wetlands as juveniles (Williams, 2002) and prefers sheltered areas in channels. Therefore, areas with high channel connectivity and relatively large wetland patch perimeter are likely to promote higher mangrove jack catches. Although mangroves were well represented in the 13 geographical areas they were not the only important habitat. One reason is that many fish and crustaceans only use mangrove forests for a part of the tidal cycle (Vance et al., 2002). Noting that mangroves are flooded < 50% of the time D( uke, 2006), the availability of adjacent habitats must be important as well. Some estuarine species move and migrate between habitat types, localities, and regions, (e.g., sea mullet M. cephalus) (Cappo et al., 1998). This implies that it is difficult and possibly misleading to separate the value of each habitat type from the broader estuarine values when looking at fish catch data. The connectivity index calculated with the program “Fragstats” appears to be a useful parameter that did not strongly correlate with other habitat parameters but combines their values in biologically meaningful ways. The index was positively related to total catch and combined CPUE of all fisheries. One explanation for this relationship may be the importance of easily accessible estuarine habitats to fish for the provision of food and shelter. The connectivity index reflects the proximity of habitat patches and therefore their potential accessibility. The differences between CPUE in 788 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

the north and the south may also be related to a reduction in habitat connectivity in the northern habitats. Studies of habitat connectivity are important; as connectivity influences other ecological processes such as species distribution, food availability, and population dynamics. Extraneous Influences.—Associated with the habitats are a number of abiotic factors such as rainfall, which may influence fish catch (Meynecke et al., 2006) but have not been considered in this study. Data on environmental factors could be useful in regression analyses to explain more of the variation in catch and it is essential to better understand how the interactions of such factors affect the distribution of fish within various estuary types and broad geographic areas (Blaber, 2000). Other factors such as stock-recruitment relationships, habitat changes, pollution impacts, competition and predation, management plans, fuel prices, and market forces also influence the fish catch data and may be considered in future analyses. The analyses are limited by the resolution of the data set, the type of fish selected and spatial classification of habitats. An a priori classification of the data is essential as unclassified aggregation can lead to misinterpretation. Assessments may be improved by: (1) refining parameters; (2) including species that are not estuary-dependent; and (3) conducting analyses over a range of spatial and temporal scales since the spatial scale of data collection can affect covariance and correlation statistics (Dungan et al., 2002). Analyses, for example, between the fish catch data and small areas of estuarine habitats are not meaningful, as the spatial resolution of fish catch data does not support such a spatial scale. Some data in Queensland are now collected at the 6 nmi scale, which can significantly improve the usefulness of the data. The spatial scale problem may also be overcome by newly available spatial analysis techniques (Campbell et al., 2006). Additional limitations in fish catch data include the recording of fish by common names, which can often result in confusion of specific identity. In order to refine the fish catch data, harvest by recreational anglers and charter operators should be included in analyses as well (Hancock, 1995). The correlations in our study were generally strong, with r2 values usually above 0.7, however, this cannot be used to assume causality because of possible non-linearity of linking mechanisms (Baumann, 1998) and the small size of the data sets. In general, we found that species or species groups with an overall low catch record had weak or non-significant results when nMDS and stepwise regression analyses were used. Significance test for most of the adjusted r2 values resulted in significance at 0.01 probability levels. A standard experiment wise error test showed a 0.26 probability that the significance tests resulted in a Type I error. The actual experiment wise error rate will range between the computed experiment wise error rate (0.26) and the test wise error rate (0.01). The computed experiment wise error rate can be reduced further by reducing the size of the allowable error (significance level) for each comparison but may fail to identify an unnecessarily high percentage of actual significant differences in the data (Olejnik et al., 1997). Another way of assessing such correlated parameters can be achieved with the application of multivariate decision trees (Breiman et al., 1984; De’ath and Fabricius, 2000). Unfortunately, there is a lack of essential life-history information for most of the major fishery species in Australia. There is a need for additional information on “critical” habitat requirements and processes such as recruitment, post-recruitment mortality and competition, spawning, and species interactions: information that is important in assessing the value of habitats for fishery species (Beck et al., 2001) and MEYNECKE ET AL.: ESTUARIES AND COASTAL FISHERIES IN QUEENSLAND 789 modelling. Beck et al. (2001) argued that the importance of a habitat as a nursery site depended not only on whether it supported more juvenile fish, but more importantly, its actual contribution to the adult (breeding) population. This juvenile-adult link is almost non-existent in analyses conducted to date. It is important therefore to include different life history stages in future analyses. In conclusion, healthy and functional estuarine wetlands are fundamental to the health of fish stocks and for optimizing sustainable yield for commercial and recreational fishers. The protection of one habitat type will only benefit a minority of commercially important fish species. To optimize habitat and fisheries management, a combination of estuarine habitats should be considered in fisheries management, supporting the move from conventional single-species or single-habitat management to ecosystem-based management (NRC, 1999). Fishery systems are complex, and the management systems needed to optimize the benefits accruing from fisheries require institutions and knowledge systems that are able to cope with the multi-disciplin- ary requirements of the fisheries management function. Fisheries managers need to broaden their knowledge base in order to make informed decisions. The controversy over estuarine outwelling has benefited from recent tracer techniques, such as stable isotope analyses, which enable the tracking of food sources for individual species (Lee, 2005). In many cases, the results question the extent of the outwelling hypothesis (Odum, 1968), suggesting the need for further studies and refinement of current knowledge. Deficiencies of studies that address the relationship between estuarine habitats and fish catch are mainly due to their single-habitat approach. The present data on fish catch and estuarine habitat distribution allowed for modelling but has its spatial limitations. The analyses showed a strong dependency by 24 estuarine species or species groups or different estuarine wetland habitats, with mud- and sandflats being most important for 25% of the species. Overall, the results show that a broad diversity of coastal habitats, other than those receiving the most attention, are essential to the completion of fish life cycles. Our investigation contributes to the broader knowledge of coastal habitats and how they influence fisheries catch and productivity. Future studies concerning fish catch and estuarine habitats could benefit from: (1) refined collection of fisheries-dependent data; (2) considerations of linked processes between estuarine and coastal habitat types; and (3) enhanced knowledge of life history stages of estuarine dependent fish species.

Acknowledgments

We would like to thank M. Dunning, L. Williams, L. Olyott, K. Danaher (DPI and F), and M. Arthur (Griffith University) for help and useful discussions during this study, and DPI&F for provision of data on fisheries and habitat.

Literature Cited

ABARE. 2005. Australian Fisheries Statistics 2004. Australian Bureau of Agricultural and Re- source Economics, Canberra. 65 p. Baran, E. and J. Hambrey. 1998. Mangrove conservation and coastal management in southeast Asia: what impact on fishery resources? Mar. Pollut. Bull. 37: 431–440. ______, J. J. Albaret, and P. S. Diouf. 1999. Les peuplements de poissons des mangroves des Rivières du Sud. Pages 98–117 in M. C. Cormier-Salem, ed. Rivières du Sud - sociétés et mangroves ou est-africaines IRD Éditions, Paris. 790 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Barnes, R. S. K. 1980. Coastal Lagoons. Cambridge University Press, Cambridge. 94 p. Baumann, M. 1998. The fallacy of the missing middle: physics -> ... -> fisheries. Fish. Oceanogr. 7: 63–65. Beck, M. W., K. L. Heck, K. W. Able, D. L. Childers, D. B. Eggleston, B. M. Gillanders, B. Halp- ern, C. G. Hays, K. Hoshino, T. J. Minello, R. J. Orth, P. F. Sheridan, and M. P. Weinstein. 2001. The identification, conservation and management of estuarine and marine nurseries for fish and invertebrates. BioScience 51: 633–641. Blaber, S. J. M. 2000. Tropical estuarine fishes. Ecology, exploitation and conservation. Black- well Science, Oxford. 372 p. Breiman, L., J. H. Friedman, R. A. Olshen, and C. J. Stone. 1984. Classification and Regression Trees. Wadsworth and Brooks/Cole, Monterey. 358 p. Browder, J. A., J. L. N. May, A. Rosenthal, J. G. Gosselink, and R. H. Baumann. 1989. Modeling future trends in wetland loss and brown shrimp production in Louisiana using thematic mapper imagery. Remote Sens. Environ. 28: 45–59. Camacho, A. S. and T. U. Bagarinao. 1987. Impact of fishpond development on the mangrove ecosystem in the Philippines. Pages 441–446 in Mangroves of Asia and the Pacific: status and management. Technical Report UNDP/UNESCO Research and Training Pilot Pro- gramme on Mangrove Ecosystems in Asia and the Pacific (RAS/79/002) Natural Resources Management Center, Manila. 504 p. Campbell, A. B., B. Pham, and Y. -C. Tian. 2006. A Framework for Spatio-temporal Data Analy- sis and Hypothesis Exploration. In A. Voinov, A. J. Jakeman, and A. E. Rizzoli, eds. Pro- ceedings of the iEMSs Third Biennial Meeting: “Summit on Environmental Modelling and Software”. International Environmental Modelling and Software Society, Burlington, USA, July 2006. CD ROM. Internet: http://www.iemss.org/iemss2006/sessions/all.html. Cappo, M., M. Alongi, D. M. Williams, and N. C. Duke. 1998. A Review and Synthesis of Aus- tralian Fisheries Habitat Research. Major Threats, Issues and Gaps in Knowledge of Coastal and Marine Fisheries Habitats - a Prospectus of Opportunities for the FRDC “Ecosystem Protection Program”. Australian Institute of Marine Science, Townsville. 431 p. Chong, V. C., C. B. Low, and T. Ichikawa. 2001. Contribution of mangrove detritus to juvenile prawn nutrition: a dual stable isotope study in a Malaysian mangrove forest. Mar. Biol. 138: 77–86. Clarke, K. R. and M. Ainsworth. 1993. A method of linking multivariate community structure to environmental variables. Mar. Ecol. Prog. Ser. 92: 205–219. ______and R. M. Warwick. 2001. Primer-E (5) computer program. Natural Environmental Research Council, Swindon, United Kingdom. Conner, L. M., M. D. Smith, and L. W. J. Burger. 2002. A comparison of distance-based and classification-based analyses of habitat use: reply. Ecology 86: 3125–3129. Day, J. W. J., C. A. S. Hall, W. M. Kemp, and A. Yáñez-Arancibia. 1989. Estuarine Ecology. John Wiley and Sons, New York. 258 p. De’ath, G. and K. Fabricius. 2000. Classification and Regression Trees: a powerful yet simple technique for the analysis of complex ecological data. Ecology 81: 3178–92. de Graaf, G. J. and T. T. Xuan. 1998. Extensive shrimp farming, mangrove clearance and marine fisheries in the southern provinces of Vietnam. Mangr. Salt Marsh. 2: 159–166. Duke, N. 2006. Australia’s Mangroves - The Authoritative Guide to Australia’s MangroveP lants. The University of Queensland, Brisbane. 200 p. Duffy, R., M.D unning, and J. Kirkwood. 2003. Draft: A preliminary classification of Queensland’s estuaries based on commercial and recreational fish catch information. CRC Coastal and DPI Queensland Fisheries Service, Brisbane. 36 p. Dungan, J. L., J. N. Perry, M. R. T. Dale, P. Legendre, S. Citron-Pousty, M. -J. Fortin, A. Jako- mulska, M. Miriti, and M. S. Rosenberg. 2002. A balanced view of scale in spatial statistical analysis. Ecography 25: 626–640. MEYNECKE ET AL.: ESTUARIES AND COASTAL FISHERIES IN QUEENSLAND 791

Dunning, M., M. Bavins, and C. Bullock. 2001. Australian Estuary Assessment, Queensland Fisheries Component, a Report for the National Land and Water Resources Audit, Theme 7. Queensland Fisheries Service, Brisbane. 15 p. Geoscience Australia. 2004. Geodata Coast 100K 2004. Geoscience Australia, Canberra. Gilbert, A. J. and R. Janssen. 1998. Use of environmental functions to communicate the values of a mangrove ecosystem under different management regimes. Ecol. Economics 25: 323–346. Hancock, D. A. 1995. Recreational Fishing: what’s the catch? Pages 177–182 in Proc. Australian Society for Fish Biology Workshop, 30–31 August 1994. AGPS, Canberra. 274 p. Horn, M. H. and L. G. Allen. 1976. Numbers of species and faunal resemblance of marine fishes in California bays and estuaries. Bull. So. Calif. Acad. Sci. 75: 159–170. Laegdsgaard, P. and C. Johnson. 2001. Why do juvenile fish utilise mangrove habitats? J. Exp. Mar. Biol. Ecol. 257: 229–253. ______. 2004. Relationship between mangrove abundance and tropical prawn production: a re-evaluation. Mar. Biol. 145: 943–949. ______. 2005. Exchange of organic matter and nutrients between mangroves and estuaries: myths, methodological issues and missing links. Int. J. Ecol. Environ. Sci. 31: 163–175. Lenanton, R. C. J. and I. C. Potter. 1987. Contribution of estuaries to commercial fisheries in temperate Western Australia and the concept of estuarine dependence. Estuaries 10: 28–35. Loneragan, N. R., N. A. Adnan, S. E. Bunn, and D. M. Kellaway. 2005. Prawn landings and their relationship with the extent of mangroves and shallow waters in western peninsular Malay- sia. Estuar. Coast. Shelf. Sci. 63: 187–200. Macnae, W. 1974. Mangrove Forests and Fisheries. FAO, Rome. 35 p. Manson, F. J., N. R. Loneragan, G. A. Skilleter, and S. R. Phinn. 2005a. An evaluation of the evidence for linkages between mangroves and fisheries: A synthesis of the literature and identification of research directions. Oceanogr. Mar. Biol. Ann. Rev. 43: 483–513. ______, ______, B. D. Harch, G. A. Skilleter, and L. E. Williams. 2005b. A broad- scale analysis of links between coastal fisheries production and mangrove extent: A case- study for northeastern Australia. Fish. Res. 74: 69–85. Martosubroto, P. and N. Naamin. 1977. Relationship between tidal forests and commercial shrimp production in Indonesia. Mar. Res. Indones. 18: 81–86. McGarigal, K., S. A. Cushman, M. C. Neel, and E. Ene. 2002. FRAGSTATS: Spatial Pattern Analysis Program for Categorical Maps. University of Massachusetts, Amherst. Available from http://www.umass.edu/landeco/research/fragstats/fragstats.html. Merriam, G. 1984. Connectivity: a fundamental ecological characteristic of landscape pattern. Pages 5–12 in Proc. First Int. Seminar on Methodology in Landscape Ecological Research and Planning. Roskilde Universitetsforlag GeuRuc, Roskilde, Denmark. 214 p. Meynecke, J. O., S. Y. Lee, N. C. Duke, and J. Warnken. 2006. Effect of rainfall as a component of climate change on estuarine fish production in Queensland, Australia. Estuar. Coast. Shelf Sci. 69: 491–504. Monaco, M. E., T. A. Lowery, and R. L. Emmett. 1992. Assemblages of US West coast estuaries based on the distribution of fishes. J. Biogeogr. 19: 251–267. Nagelkerken, I., S. Kleijnen, T. Klop, R. A. C. J. van den Brand, E. Cocheret de la Morinière, and G. van der Velde. 2001. Dependence of Caribbean reef fishes on mangroves and seagrass beds as nursery habitats: a comparison of fish faunas between bays with and without mangroves/seagrass beds. Mar. Ecol. Prog. Ser. 214: 225–235. NRC. 1999. Sharing the Fish: Towards a National Policy on Individual Fishing Quotas. National Research Council, Washington, D.C. 422 p. Odum, E. P. 1968. A research challenge: evaluating the productivity of coastal and estuarine water. Pages 63–64 in Proc. 2nd Sea Grant Conference. University of Rhode Island, Kingston. Olejnik, S., J. Li, S. Supattathum, and C. J. Huberty. 1997. Multiple testing and statistical power with modified Bonferroni procedures. J. Educational and Behavioral Statistics, 22: 389–406. 792 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Paw, J. N. and T. -E. Chua. 1991. An assessment of the ecological and economic impact of mangrove conversion in Southeast Asia. Pages 201–212 in L. M. Chou, ed. Towards an Integrated Management of Tropical Coastal Resources. ICLARM Conference Proceedings ICLARM, Manila. Pauly, D. and J. Ingles. 1986. The relationship between shrimp yields and intertidal vegetation (mangrove) areas: a reassessment. IOC/FAO Workshop on Recruitment in Tropical Coastal Demersal Communities. IOC, UNESCO, Paris. Pease, B. C. 1999. A spatially oriented analysis of estuaries and their associated commercial fisheries in New South Wales, Australia. Fish Res. 42: 67–86. Perkins, E. J. 1974. The Biology of Estuaries and Coastal Waters. Academic Press, London. 448 p. Pollard, D. A. 1981. Estuaries are valuable contributors to fisheries production. Aust. Fish. 40: 7–9. ______. 1994. A comparison of fish assemblages and fisheries in intermittently open and permanently open coastal lagoons on the south coast of New South Wales, south-eastern Australia. Estuaries 17: 631–646. Pyper, B. and R. Peterman. 1998. Comparisons of methods to account for autocorrelation in correlation analyses of fish data. Can. J. Fish. Aquat. Sci. 55: 2127–2140. DPI&F. 2005. Commercial fisheries in Queensland. Department of Primary Industries and Fisheries, Brisbane. Quinn, R. 1993. What is the importance of wetland and estuarine habitats? Queensl. Fisher- man 11: 27–30. Rönnbäck, P. 1999. The ecological basis for economic value of seafood production supported by mangrove ecosystems. Ecol. Econ. 29: 235–252. Russell, D. J. and R. N. Garrett. 1983. Use by juvenile barramundi, Lates calcarifer (Bloch), and other fishes of temporary supralittoral habitats in a tropical estuary in northern Australia. Aust. J. Mar. Freshw. Res. 34: 805–811. Saintilan, N. 2004. Relationships between estuarine geomorphology, wetland extent and fish landings in New South Wales estuaries. Estuar. Coast. Shelf Sci. 61: 591–601. Sasekumar, A. and V. C. Chong. 1987. Mangroves and prawns: further perspectives. Pages 10–22 in A. Sasekumar, S. M. Phang, and E. L. Chong, eds. Proc. 10th Annual Seminar of the Malaysian Society of Marine Sciences. Univ. of Malaya, Kuala Lumpur. Sheridan, P. and C. Hays. 2003. Are mangroves nursery habitat for transient fishes and decapods? Wetlands 23: 449–458. Staples, D. J., D. J. Vance, and D. S. Heales. 1985. Habitat requirements of juvenile penaeid prawns and their relationship to offshore fisheries. Pages 47–54 in Second Australian Na- tional Prawn Seminar. CSIRO, Kooralbyn, Queensland. Staunton-Smith, J., J. B. Robins, D. G. Mayer, M. J. Sellin, and I. A. Halliday. 2004. Does the quantity and timing of fresh water flowing into a dry tropical estuary affect year-class strength of barramundi (Lates calcarifer)? Mar. Freshw. Res. 55: 787–797. Stergiou, K. I. and D. A. Pollard. 1994. A spatial analysis of the commercial fisheries catches from the Greek Aegean Sea. Fish Res. 20: 109–135. Tanner, M. and G. W. Liggins. 2000. New South Wales Commercial Fisheries Statistics 1998/99. Cronulla Fisheries Centre, Cronulla, New South Wales. 22 p. Turner, R. E. 1977. Intertidal vegetation and commercial yields of penaeid shrimp. Trans. Am. Fish Soc. 106: 411–416. Vance, D. J., M. D. E. Haywood, D. S. Heales, and D. J. Staples. 1996. Seasonal and annual variation in abundance of postlarval and juvenile grooved tiger prawns, Penaeus semisulcatus (Decapoda: Penaeidae), and environmental variation in the Embley River, Australia: a six- year study. Mar. Ecol. Progr. Ser. 135: 43–55. ______, ______, ______, ______, N. R. Loneragan, and R. C. Pen- drey. 2002. Distribution of juvenile penaeid prawns in mangrove forests in a tropical Aus- tralian estuary, with particular reference to Penaeus merguiensis. Mar. Ecol. Prog. Ser. 228: 165–177. MEYNECKE ET AL.: ESTUARIES AND COASTAL FISHERIES IN QUEENSLAND 793

Whitfield, A. K. 1999. Ichthyofaunal assemblages in estuaries: a South African case study. Rev. Fish Biol. Fish. 9: 151–186. Williams, L. E., ed. 2002. Queensland’s fisheries resources: Current condition and recent trends 1988–2000. Queensland Department of Primary Industries and Fisheries, Brisbane. 180 p. Yearsley, G. K., P. R. Last, and R. D. Ward. 1999. Australian Seafood Handbook — An Identifi- cation Guide to Domestic Species. CSIRO Division of Marine Research, Canberra. 461 p.

Addresses: (J.-O.M., S.Y.L., J.W.) Griffith School of Environment and Australian Rivers Institute, Griffith University–Gold Coast campus, PMB 50 GCMC Bundall, Queensland, 9726, Austra- lia. (N.C.D.) Centre for Marine Studies, University of Queensland, St Lucia, Queensland, 4072, Australia. Corresponding Author: (J.-O.M.) Telephone: ++61 7 555 28983, Fax: ++61 7 555 28067, E-mail: . BULLETIN OF MARINE SCIENCE, 80(3): 795–804, 2007

MANGROVES AND SHRIMP POND CULTURE EFFLUENTS IN AKLAN, PANAY IS., CENTRAL PHILIPPINES

J. H. Primavera, J. P. Altamirano, M. J. H. L. Lebata, A. A. delos Reyes Jr., and C. L. Pitogo

ABSTRACT The capacity of a natural mangrove system in Ibajay, Aklan province, central Phil- ippines to process shrimp pond culture effluents was assessed through analysis of mangrove community structure and 24-hr monitoring of water quality parameters

(NH3-N, NO3-N, PO4-P, sulfide, and total suspended solids). Results from the latter showed decreased nutrient levels within 6 hrs after daytime draining of effluents into the mangrove stand, but only nitrate reduction was statistically significant. Based on nitrate loss, volume of water drained, mangrove area, and shrimp farming data (e.g., N loss from ponds, feed composition, feeding rate), calculations show that 1.8–5.4 ha of mangroves are required to remove nitrate wastes from 1 ha of shrimp pond. N uptake by the mangrove macroflora was supported by data showing longer nipa palm leaflets and faster mangrove seedling growth in the experimental mangrove receiving effluents compared to a control mangrove, but not from mangrove biomass measurements. These results have significant implications for the Philip- pine brackishwater pond culture industry to conserve or rehabilitate mangroves as potential pond biofilters, to implement legally mandated 20- and 50-m greenbelts, and to reverse the national 0.5 ha mangrove: 1.0 ha pond ratio.

Although mangrove systems contribute significantly to the well-being of coastal communities through a wide array of services, fishery and forestry products, society has viewed them as wastelands to be developed. Hence their worldwide decline from 18 million ha in the early to mid-1990s (Spalding et al., 1997) to 15 million ha as of 2000, of which a third is found in Southeast Asia (Wilkie and Fortuna, 2003). Valiela et al. (2001) estimated a yearly loss of 2.1% of existing area, and an average loss of 35% since the 1980s, with half of such loss due to shrimp and fish culture, where more than 50% of total present mangrove area was represented. Around half of the 279,000 ha of Philippine mangroves lost from 1951 to 1988 was developed into culture ponds (Primavera, 2000a). Conversely, 95% of brackishwater ponds in the same period were constructed in mangrove areas. In the Philippines and Southeast Asia, the shrimp farming industry has followed a boom-and-bust pattern, and many shrimp ponds have declined in productivity or have been abandoned altogether (Primavera, 1998). Elsewhere, shrimp culture is shifting from technologies associated with contamination of receiving waters to environment-friendly practices, including zero/reduced water exchange, polyculture with seaweeds/bivalves/fish, in-pond biofilters and use of adjacent, but separate wetlands such as mangroves (Rivera-Monroy et al., 1999; Gautier, 2002). Tidal mangrove estuaries are able to process nutrient inputs from culture ponds at least over short spatial and temporal scales (Trott and Alongi, 2000), and pioneering trials have used planted mangroves to treat shrimp pond wastes (Sansanayuth et al., 1996; Ahmad, 2000). In 1996, the Southeast Asian Fisheries Development Center (SEAFDEC) Coun- cil proactively mandated the development of environment-friendly shrimp culture

Bulletin of Marine Science 795 © 2007 Rosenstiel School of Marine and Atmospheric Science of the University of Miami 796 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 technology, both aquasilviculture that physically integrates the low-density culture of crabs with mangroves, and biofilters where separate mangrove stands are used to absorb intensive pond effluents (Primavera, 2000b). The present study was conducted by the Philippine-based SEAFDEC Aquaculture Department to assess the capacity of mangroves to process aquaculture effluents in terms of water quality and mangrove growth and structure.

Materials and Methods

The Bugtongbato-Naisud mangrove covering 70.2 ha in Ibajay, Aklan province in central Philippines (122°11.4´–122°12.7´E; 11°47.8´–11°48.5´N) is a multispecies community dominated by Avicennia rumphiana Hall. f. and Avicennia officinalisLinn. (Lebata, 2006). Located ~1 km from the sea (Fig. 1), the study site comprises a tidal creek as water source, reservoir (1480 m2 with a few trees on elevated mounds and estimated water area of 890 m2) for conditioning inflow water, shrimp pond (880 m2), and experimental mangrove (320 m2) bordered by earthen dikes to retain water for sampling (Fig. 2). Shrimp Culture.—To grow natural food and eliminate unwanted fish, agricultural lime, chicken manure, urea, and teaseed powder were applied to the shrimp pond at standard rates (Apud et al., 1985). MilkfishChanos chanos Forsskål of 20–300 g average body weight (ABW) were stocked in the shrimp pond (21–27 kg) and reservoir (30–40 kg) following the greenwa- ter technique. After one week, shrimp Penaeus monodon Fabricius postlarvae were stocked in the shrimp pond at 25 ind m–2. Commercial feeds (crude protein 40%) were given at 4%–8% BW. Shrimp were harvested after 5–6 mo, and production and survival were determined. Number and biomass of extraneous species were also noted. Two trials were conducted— Trial 1 (26 June–1 December 2001) and Trial 2 (25 June–16 November 2002). Water Quality.—Every spring tide, sea water from the creek was allowed to flow into the reservoir for conditioning for ~12 hrs, pumped to the shrimp pond and used for 1–5 d, drained to the experimental mangrove and retained for at least 12 hrs before release back to the creek (Fig. 2). Physicochemical parameters were monitored monthly in the creek, reservoir, shrimp pond, and experimental mangrove–salinity (Atago refractometer), temperature and pH (Hach SensION portable pH meter), dissolved oxygen (YSI 550-20 DO meter), NH3-N,

NO3-N, PO4-P, sulfide, total suspended solids, or TSS (Hach DREL 2010 portable water qual-

Figure 1. Location of study site in Bugtongbato-Naisud, Ibajay, Aklan province in central Philip- pines. PRIMAVERA ET AL.: MANGROVES AND SHRIMP-POND EFFLUENTS IN CENTRAL PHILIPPINES 797

Figure 2. Schematic experimental set-up of Reservoir-Shrimp Pond-Experimental Mangrove including area (m2) and duration (d) of water holding in Ibajay, Aklan, central Philippines. Not to scale. ity laboratory), and 5-d biochemical oxygen demand or BOD5 (WTW incubator TS 606). In Trial 2, intensive 24-hr monitoring of water quality was undertaken in the 3rd and 4th mo of culture when shrimp biomass and effluent levels were highest. Nutrient and TSS levels were measured in the experimental mangrove as effluents were partially drained from the shrimp pond (0 hr) and at regular intervals thereafter (6, 12, and 24 hrs). Water was drained at 0800 (for a day cycle) or at 2000 (night cycle). Mangrove Ecosystem.—At the start of the study, community structure of the mangrove stand was analyzed, including counts of seedlings (< 1 m ht), saplings (> 1 m ht, < 4 cm diameter at breast height or DBH), and trees (> 1 m ht, > 4 cm DBH) (English et al., 1994). For each shrimp production trial, the experimental mangrove was monitored at the same time as a control mangrove site of similar elevation, but upstream of the experimental site (therefore unaffected by pond effluents). Wild seedlings of the mangrove species A. officinalis, A. rumphiana, Bruguiera cylindrica (Linn.) Blume, Ceriops decandra (Griff.) Ding Hou, Ceriops tagal (Perr.) C. B. Rob., Rhizopho- ra mucronata Lamk., and Xylocarpus granatum Koen. were collected from the nearby forest and planted in both experimental and control areas. The seedlings were labeled with plastic tags for identification and measured regularly. For Trial 2, the average leaflet length of the mangrove palm Nypa fruticans (Thunb.) Wurm. (n = 3–5 leaflets per plant, 5–10 plants per site) was determined monthly in the experimental and control mangrove. Mean lengths were compared by t-test. The number of leaves per plant was not counted because nipa leaves are harvested by locals every 3–4 mo for making house shingles.

Results

Shrimp Production and Effluents.—Although shrimp survival was similar in both trials, total production in Trial 2 was only half that of Trial 1 because of a shorter grow-out period (by 2 wks) and smaller shrimp sizes (Table 1). Competition from many gobies and other unwanted fish species could account for smaller shrimp sizes. Monthly water quality measurements (Fig. 3) indicate increased loadings of 798 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Table 1. Summary of shrimp Penaeus monodon culture in 880 m2 pond in Bugtong Bato, Ibajay, Aklan Province, central Philippines.

Trial 1 Trial 2

Duration 26 Jun–1 Dec 2001 25 Jun–16 Nov 2002 P. monodon: Initial no. 24,000 23,000 Final no. 13,233 11,003 Survival rate (%) 55.51 47.84 Ave. body wt. (g) 19.1 12.8 Total shrimp production (kg) 252.7 141.4 Other species (kg): Milkfish 31.8 46.1 Other fish 2.0 27.8 Shrimps, crabs 9.6 0.0 Total biomass harvested (kg) 296.1 215.3 Temperature (°C) 28.4–34.5 27.6–31.7 Salinity 15–29 23–32 pH 7.18–10.18 6.93–8.89

suspended solids and some nutrients later in the grow-out period in the shrimp pond as well as the experimental mangrove. Intensive 24-hr monitoring during the months of high shrimp biomass showed that nutrient levels in pond effluents decreased within 6 hrs if drained into the mangroves during the day, but increased if draining was at night (Fig. 4). In contrast, TSS levels showed a decrease regardless of whether draining was done by day or night. Because the reduced nutrient levels were significant only for nitrate (Table 2), this value was used to estimate mangrove area requirement (for comparison with earlier calculations based on data from commercial shrimp farms and literature on mangrove pro- –1 –1 –1 ductivity). Multiplying the removal rate of 0.4830 mg NO3-N L d by water volume drained (62,000 L), then dividing by experimental mangrove area (320 m2) gave a ni- –2 –1 trate removal rate of 93.58 mg NO3-N m mangrove d . Assuming 35% loss of pond N through water exchange (Briggs and Funge-Smith, 1994), 6% N content of shrimp feed (I. Borlongan, Central Philippine University, Iloilo City, pers. comm.), feeding rate of 4% of shrimp biomass and harvest sizes of 20–30 g average body weight, calculations –2 –1 show removal of 267.4 mg NO3-N m mangrove d which is equivalent to 0.0044 kg feed m–2 mangrove d–1 or 1112 kg shrimp ha–1 mangrove. Put another way, the removal of nitrate wastes from 1 ha of shrimp pond will require treatment or processing by 1.8–2.7 ha of mangroves for semi-intensive culture (10 postlarvae m–2) and 3.6–5.4 ha of mangroves at intensive levels (20–30 postlarvae m–2). Mangroves.—In May 2001 prior to the experiment, the Bugtongbato-Naisud mangrove community had a high density of 349,379 plants ha–1 mainly due to numerous seedlings in May, immediately following the March–April peak of Avicennia flowering (Primavera et al., 2004). Plant numbers in both mangrove sites sharply decreased to 16%–27% of original 2001 levels in April 2002 (when blooming flowers had not yet turned to fruits and (fallen) seedlings, and slightly increased afterwards (Table 3). Initial stand basal area (BA, representing biomass) of 18.99 m2 ha–1 did not change much in April 2002, but increased by 48%–154% in November 2002. The latter was higher in the experimental than the control mangrove (34.7 vs 28.1 m2 PRIMAVERA ET AL.: MANGROVES AND SHRIMP-POND EFFLUENTS IN CENTRAL PHILIPPINES 799

Figure 3. Monthly levels of nutrients, total suspended solids (TSS), dissolved oxygen (DO) and

biochemical oxygen demand (BOD5) (mean ± SE) in Creek (C), Reservoir (R), Shrimp Pond (SP), and Experimental Mangrove (M) in Ibajay, Aklan, central Philippines, Trial 1 (2001).

ha–1), although this difference was not statistically significant (t-test, P > 0.05). Ini- tial Shannon Index of Diversity of 0.6703 showed slight fluctuations throughout the study, but values did not significantly differ between experimental and control mangroves. For the palm N. fruticans (Table 4), average leaflet length was significantly higher in the experimental mangrove (162.4–201.5 cm) compared to the control mangrove (151.5–157.7 cm) over a 4-mo period. Growth rates in terms of height increase of planted seedlings (Table 5) in Trial 1 were higher in the experimental mangrove (0.5%–24.2%) compared to the control mangrove (−9.5% to 6.8%). For Trial 2, C. decandra, C. tagal, and A. officinalis had similar growth rates but A. rumphiana, B. cylindrica, and Xylocarpus spp. showed negative growth in the control mangrove.

Figure 4. Levels of nutrients, sulfide and TSS in effluents in the Experimental Mangrove (drained from the Shrimp Pond). Start of draining was 0800 for the day cycle and 2000 for the night cycle. 800 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Table 2. Comparison of mean (±1 SE) levels of solids and nutrients (mg L–1) in pond effluents at 0 and 6 hrs after draining into experimental mangrove (start 0800, day cycle).

0 hrs 6 hrs P value (t test) Total suspended solids 102.833 ± 52.218 36.833 ± 5.192 0.261 Sulfide 0.0503 ± 0.0108 0.0330 ± 0.0056 0.179

PO4 0.02833 ± 0.0048 0.0767 ± 0.0274 0.108

NH3-N 0.6370 ± 0.1830 0.4783 ± 0.0634 0.345

NO3-N 2.5830 ± 0.2060 2.1000 ± 0.2860 0.0028* *Significantly different

Discussion

Despite lower shrimp production in Trial 2, the combined biomass (2500–3400 kg–1 ha crop–1) of the shrimp+milkfish harvest+extraneous species produced adequate pond effluents for draining to the experimental mangrove. Monthly monitoring revealed increased nutrient loading, suspended solids, and BOD5 in the shrimp pond during the later months of the cropping period. Wetlands may process nutrients in pond effluents either through physical dilution and tidal flushing, or biological transformation. In this study, however, dilution was eliminated as the cause of decreased nutrient levels because an actual increase in these levels was observed within the first 6 hrs of night-time draining. The presence of earthen dikes enclosing the experimental mangrove further precludes tidal flushing, and leaves biochemical processes—nutrient uptake by micro- and macroflora, denitrification, etc.—to explain nutrient removal. Denitrification and volatilization to the atmosphere accounted for only 3%–13% of total N loss from shrimp pond bud- gets (Briggs and Funge-Smith, 1994; Jackson et al., 2003; Trott et al., 2004). Seventy percent of effluent N input from a shrimp farm in Queensland, Australia reached the lower reaches of a mangrove creek and was removed by tidal flushing and biological

Table 3. Changes in mangrove plant numbers (stems ha–1) and biomass (stand basal area in m2 ha–1) of experimental mangrove (with pond effluents) and control mangrove in Ibajay, Aklan, central Philippines. Values are means ± SE (in parenthesis).

May 2001 April 2002 November 2002 Exp Control Exp Control A. Mangrove plants Seedlings 343,333 82,667 45,600 93,467 74,933 (135,572.35) (34,856.15) (27, 365.18) (43,180.45) (34,085.64) Saplings 5,733 11,600 10,533 19,933 9,067 (1,866.67) (5,281.41) (2,969.47) (4,015.53) (1.763.83) Trees 313 469 900 656 967 (0.00) (0.00) (100.00) (0.00) (88.19) Total 349,379 94,736 57,033 113,456 84,967 B. Biomass Seedlings 6.74 1.62 0.90 1.84 1.47 (2.66) (0.69) (0.54) (0.85) (0.67) Saplings 1.80 3.64 3.31 6.16 2.85 (0.59) (1.66) (0.93) (1.26) (0.56) Trees 10.45 13.58 17.50 26.74 23.77 (0.00) (0.00) (8.51) (0.01) (5.83) Total 18.99 18.84 21.71 34.74 28.09 C. Species diversity 0.6703 0.7108 0.7194 0.6661 0.7020 PRIMAVERA ET AL.: MANGROVES AND SHRIMP-POND EFFLUENTS IN CENTRAL PHILIPPINES 801

Table 4. Leaflet length (means ± SE in cm) of Nypa fruticans (n = 3–5 leaflets per plant, 5–10 plants per site) in experimental mangrove and control mangrove in Bugtong Bato, Ibajay, Aklan, central Philippines.

Date Experimental mangrove Control mangrove P value (t test) 25 Jul 2002 162.4 ± 0.76 151.1 ± 7.89 0.339 20 Aug 2002 197.3 ± 5.79 155.0 ± 11.05 0.003** 22 Sep 2002 195.9 ± 6.50 156.2 ± 11.00 0.006** 19 Oct 2002 196.8 ± 6.06 157.4 ± 11.52 0.007** 20 Nov 2002 201.5 ± 6.28 157.7 ± 11.58 0.004** **Highly significant

transformation through algal and bacterial production, only 30% remained in the upper reaches (Trott et al., 2004). However, the Queensland study did not consider nutrient uptake by mangrove trees. Monitoring of mangrove macroflora indicated that growth in length of nipa leaflets was significantly higher in the experimental mangrove than in the control site. Regu- lar harvest or removal of mangrove biomass (nipa leaflets in this study) is necessary to maximize the efficiency of plant uptake as a N sink (Gautier, 2002). The growth rate of planted mangrove seedlings for all species in Trial 1 and majority of species in Trial 2 was similarly higher in the experimental mangrove compared to the control site. Survival and/or growth of seedlings of various mangrove species have been reported in shrimp pond effluents (Rajendran and Kathiresan, 1996), intertidal dumpsites near shrimp farms (Macintosh, 1996), and an abandoned shrimp pond (Tantipuknont et al., 1994). Natural recruitment of mangrove seedlings in a biofilter receiving shrimp farm effluents was double that of a nearby control site (Gautier, 2002). Nevertheless, experimental planting of Rhizophora apiculata Bl. in a tidal flat resulted in high seedling mortality and decline in growth under high sediment accumulation of > 4 cm (Terrados et al., 1997), and burial of roots in > 10 cm of sediments caused death of various mangrove species of Avicennia, Sonneratia, and Rhizophora (Ellison, 1998). Therefore, shrimp farms with high effluent levels of solids must- in corporate settling ponds before release of wastes (to adjacent mangrove biofilters). Table 5. Growth of mangrove seedlings in the experimental mangrove (with shrimp pond effluents) and control mangrove in Bugtong Bato, Ibajay, Aklan, central Philippines.

Experimental mangrove Control mangrove Init ht Final ht % Init ht Final ht % (cm) (cm) change (cm) (cm) change Trial 1 (Jul–Nov 2001) Avicennia rumphiana 69.33 84.00 21.15 67.50 70.17 3.95 Ceriops decandra 31.00 38.50 24.19 19.33 17.50 −9.48 Rhizophora mucronata 69.67 70.00 0.48 61.00 60.33 −1.09 Xylocarpus spp. 100.50 114.00 13.43 83.33 89.00 6.80 Trial 2 (Jul–Nov 2002) Avicennia officinalis 30.20 39.20 29.80 36.13 48.40 33.96 Avicennia rumphiana 89.38 93.17 4.24 64.10 54.25 −15.37 Bruguiera cylindrica 40.75 41.83 2.66 27.50 24.70 −10.18 Ceriops decandra 58.30 66.80 14.58 17.67 20.60 16.58 Ceriops tagal 24.50 25.25 3.06 28.67 29.00 1.15 Rhizophora mucronata 69.00 67.60 −2.03 67.13 71.00 5.76 Xylocarpus granatum 109.38 121.13 10.74 91.67 81.00 −11.64 802 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

On the other hand, biomass (stand BA) did not significantly differ between the experimental and control mangrove. The similarity in plant numbers, biomass, and species diversity in both mangrove sites indicates the benign effects of passing pond effluents through a mangrove system. Our results generated estimates of 1.8–2.7 ha and 3.6–5.4 ha of mangrove area required to process N (nitrate) wastes from one ha of semi-intensive and intensive shrimp pond, respectively. Robertson and Phillips (1995) calculated that 2.4–7 ha and 3–22 ha of mangroves, respectively, could absorb N and P wastes from a 1 ha intensive shrimp pond. Kautsky et al. (1997) estimated that more than 6 ha of mangroves are required to absorb the N and P wastes from a 1-ha semi-intensive shrimp pond in Colombia. These theoretical calculations based on published data are complemented by experimental results with figures of 8.96 ha and 7.82 ha of mangroves needed for N and P removal, respectively (Boonsong and Eiumnoh, 1995). The substantial areas required for wetland processing of aquaculture wastes provide a compelling argument for conserving mangroves and carry positive implications for the rehabilitation of wide expanses of degraded and abandoned brackishwater ponds throughout the country. Interestingly, the earliest guideline on mangrove: pond ratio gives an approximate figure of not more than 20% for pond conversion (Saenger et al., in 1983), or a 4 mangrove: 1 pond ratio. Processing of N wastes requires higher mangrove-pond ratios of up to 6–9: 1, and adding other ecological functions (e.g., fish/invertebrate nurseries, coastal protection) suggest an even higher coverage of mangroves. Philippine mangroves have declined from 500,000 ha in 1918 to 120,000 in 1994, while brackishwater ponds have quadrupled from 61,000 ha in 1940 to 232,000 ha in 1994 (Primavera, 2000a). These figures give an actual ratio of only 0.5 ha of mangrove for every ha of pond, which is well below the 6–9: 1 and 4: 1 ratios recommended above. Ironically, many brackishwater ponds are underutilized and suffer from low productivity because they are poorly managed (Primavera, 2005). The typical Filipino pond operator resides in urban centers (to practice law, medicine, and other pro- fessions as the primary source of income)—and leaves daily pond operations to a caretaker. Therefore abandoned and/or low productivity ponds should be reverted to mangroves, in compliance with the 20/50/100-m wide greenbelts along waterways mandated by law.

Acknowledgments

This study was funded by the Government of Japan Trust Fund for the Mangrove-Friendly Shrimp Culture Project. Grateful thanks are extended to Bugtong Bato, Ibajay Barangay Head N. Soliva for facilitating field arrangements, A. Traje for assistance in the pond and mangrove work, N. Kautsky and M. Beveridge for helpful comments on the manuscript, and J. Binas.

Literature Cited

Ahmad, T. 2000. The use of mangrove stands for shrimp ponds waste-water treatment. Pages 147–153 in Proc. JIRCAS Int. Workshop on Brackishwater and Mangrove Ecosystems Pro- ductivity and Sustainable Utilization, Tsukuba, Japan, 29 Feb.–1 Mar. 2000. Japan Interna- tional Research Center for Agricultural Sciences, Tsukuba, Japan. PRIMAVERA ET AL.: MANGROVES AND SHRIMP-POND EFFLUENTS IN CENTRAL PHILIPPINES 803

Apud, F. D., J. H. Primavera, and P. L. Torres, Jr. 1985. Farming of prawns and shrimps. Aqua- culture Extension Manual No. 5, SEAFDEC Aquaculture Department, Iloilo, Philippines. 66 p. Boonsong, K. and A. Eiumnoh. 1995. Integrated management system for mangrove conservation and shrimp farming: a case of Kung Krabaen Bay, Chanthaburi, Thailand. Pages 48–65 in C. Khenmark, ed. Ecology and management for mangrove restoration and regeneration in East and Southeast Asia. UNESCO Regional Office for Science and Technology in South- east Asia, Thailand. Briggs, M. R. P. and S. J. Funge-Smith. 1994. A nutrient budget of some intensive marine shrimp ponds in Thailand. Aquacult. Fish. Mngmnt. 25: 789–811. Ellison, J. C. 1998. Impacts of sediment burial on mangroves. Mar. Pollut. Bull. 37: 420–426. English, S., C. Wilkinson, and V. Baker, eds. 1994. Survey Manual for Tropical Marine Resourc- es. Australian Institute for Marine Science, Townsville, Australia. 368 p. Gautier, D. 2002. The Integration of Mangroves and Shrimp Farming: A Case Study on the Caribbean Coast of Colombia. Report prepared under the World Bank, NACA, WWF and FAO Consortium Program on Shrimp Farming and the Environment. Work in Progress for Public Discussion, 26 p. Jackson, C., N. Preston, P. J. Thompson, and M. Burford. 2003. Nitrogen budget and effluent nitrogen components at an intensive shrimp farm. Aquaculture 218: 397–411. Kautsky, N., H. Berg, C. Folke, J. Larsson, and M. Troell. 1997. Ecological footprint for assessment of resource use and development limitations in shrimp and tilapia aquaculture. Aquacult. Res. 28: 753–766. Lebata-Ramos, M. J. H. 2006. Stock enhancement of the mud crab Scylla spp. in the mangroves of Naisud and Bugtong Bato, Ibajay, Aklan, Philippines. Ph.D. Diss. University of Wales, Bangor, U.K. 274 p. Macintosh, D. 1996. Mangroves and coastal aquaculture: doing something positive for the environment. Aquacult. Asia 1: 3–8. Primavera, J. H. 1998. Tropical shrimp farming and its sustainability. Pages 257–289 in S. De Silva, ed. Tropical mariculture. Academic Press, London. ______.2000a. Development and conservation of Philippine mangroves: institutional issues. Ecol. Econ. 35: 91–106. ______. 2000b. Integrated mangrove-aquaculture systems in Asia. Pages 121–130 in A. Pink, ed. Integrated coastal zone management. Autumn edition. ______. 2005. Mangroves, fishponds, and the quest for sustainability. Science 310: 57–59 Primavera, J. H., R. B. Sadaba, M. J. H. L. Lebata, and J. P. Altamirano. 2004. Handbook of Man- groves in the Philippines – Panay. SEAFDEC Aquaculture Department (Philippines) and UNESCO Man and the Biosphere ASPACO Project, 106 p. Rajendran, N. and K. Kathiresan. 1996. Effect of effluent from a shrimp pond on shoot biomass of mangrove seedlings. Aquacult. Res. 27: 747–747. Rivera-Monroy, V. H., L. A. Torres, N. Bahamon, F. Newmark, and R. R. Twilley. 1999. The potential use of mangrove forests as nitrogen sinks of shrimp aquaculture pond effluents: the role of denitrification. J. World Aquacult. Soc. 30: 12–25. Robertson, A. I. and M. J. Phillips. 1995. Mangroves as filters of shrimp effluent: predictions and biogeochemical research needs. Hydrobiologia 295: 311–321. Saenger, P., E. J. Hegerl, and J. D. S. Davie. 1983. Global status of mangrove ecosystems. IUCN Commission on Ecology Papers No. 3. Gland, Switzerland. 88 p. Sansanayuth, P., A. Phadungchep, S. Ngammontha, S. Ngdngam, P. Sukasem, H. Hoshino, and M. S. Tabucanon. 1996. Shrimp pond effluent: pollution problems and treatment by constructed wetlands. Water Sci. Technol. 34: 93–98. Spalding, M., F. Blasco, and C. Field, eds. 1997. World Mangrove Atlas. International Society for Mangrove Ecosystems, Okinawa, Japan. 178 p. 804 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Tantipuknont, S., N. Paphavasit, and S. Aksornkae. 1994. Growth and survival rate of three mangrove seedlings planted on the abandoned shrimp pond, Changwat Samutsongkram. Pages 373–375 in S. Sudara, C. Wilkinson, and L. M. Chou, eds. Proc. Third ASEAN-Aus- tralia Symp. on Living Coastal Resources: Research Papers 2. Australian Institute of Marine Science, Townsville, Australia. Terrados, J., U. Thampanya, N. Srichai, P. Kheowvongsri, O. Geertz-Hansen, S. Boromthana- rath, N. Panapitukkul, and C. M. Duarte. 1997. The effect of increased sediment accretion on the survival and growth of Rhizophora apiculata seedlings. Estuar. Coast. Shelf Sci. 45: 697–701. Trott, L. A. and D. M. Alongi. 2000. The impact of shrimp pond effluent on water quality and phytoplankton biomass in a tropical mangrove estuary. Mar. Pollut. Bull. 40: 947–951. ______, A. D. McKinnon, D. M. Alongi, A. Davidson, and M. A. Burford. 2004. Carbon and nitrogen processes in a mangrove creek receiving shrimp farm effluent. Estuar. Coast. Shelf Sci. 59: 197–207. Valiela, I., J. L. Bowen, and J. K. York. 2001. Mangrove forests: one of the world’s threatened major tropical environments. Bioscience 51: 807–815. Wilkie, M. L. and S. Fortuna. 2003. Status and trends in mangrove area extent worldwide. Working Paper FRA 63, Forest Resources Division, Forestry Department, UN Food and Agriculture Organization, Rome. 292 p.

Addresses: (J.H.P., M.J.H.L., C.L.P.) Aquaculture Department, Southeast Asian Fisheries Development Center, Tigbauan, Iloilo 5021, Philippines. (J.P.A.) Laboratory of Global Fisher- ies Science, Department of Global Agricultural Sciences, The University of Tokyo, 1-1-1 Yayoi, Bunkyo-ku, Tokyo 113-8657, Japan. (A.A.R.,Jr) Fish Farming Centre, North Obhur, Jeddah, P.O. Box 9612, Jeddah 21423, Saudi Arabia. Corresponding Author: E-mail: . BULLETIN OF MARINE SCIENCE, 80(3): 805–821, 2007

HABITAT ASSOCIATIONS OF LARGE-BODIED MANGROVE- SHORELINE FISHES IN A SOUTHWEST FLORIDA ESTUARY AND THE EFFECTS OF HURRICANE DAMAGE

Marin F. D. Greenwood, Charles F. Idelberger, and Philip W. Stevens

ABSTRACT We used 183-m seines to study the effects of canopy damage by Hurricane Char- ley (August 2004) on the use of red mangrove Rhizophora mangle Linnaeus shorelines by large-bodied (> ~100 mm SL) fishes in Charlotte Harbor (Florida, USA). No significant relationship was found between proportion of the shoreline that was damaged and fish-assemblage structure in the year following the hurricane; however, temperature, water depth, and salinity did influence fish-assemblage structure. Post-hurricane fish assemblages were not notably different from pre-hurricane (1996–2004) assemblages. Despite wide-ranging loss of canopy, defoliation, and mortality of red mangroves, prop-root structure remained largely intact, which may have minimized gross changes in the assemblages of large-bodied fish along the mangrove shorelines during the first post-hurricane year. The effects of habitat damage on the fish assemblages may yet be manifested due to continued deteriora- tion of existing prop-root structure and the apparent slow rate of recolonization of red mangroves in the area. Also, the effect of widespread mangrove mortality on the many other mangrove functions (e.g., primary productivity, food-web, and nutrient dynamics) may ultimately lead to system-level changes that affect the abundance and condition of large-bodied fish.

Mangrove shorelines are important habitats for fishes (Kathiresan and Bingham, 2001), mostly because they provide areas that are valuable for feeding and refuge from predation (Manson et al., 2005a). Evidence suggests that mangrove carbon does not directly contribute to fish diets, but instead that mangroves form substrates for various epibionts (both primary producers and consumers) that fish may eat (Kieckbusch et al., 2004). Refuge from predators may be provided by canopy shade and prop-root structural complexity (Laegdsgaard and Johnson, 2001, Cocheret de la Morinière et al., 2004, Ellis and Bell, 2004). There is debate over the value of mangroves as nursery habitat, but juveniles of some species appear to be mangrove-dependent (reviewed by Manson et al., 2005a). Very few studies have documented clear links between fishery catches of finfish and extent of mangroves in adjacent areas (Manson et al., 2005b). Correlations tend to be relatively weak and merely imply enhanced nursery function without directly addressing causative factors (Manson et al., 2005a). Two recent fishery-independent studies suggested relatively high fish abundance and species richness in mangrove areas with low anthropogenic degradation (Dankwa and Gordon, 2002, Singkran and Sudara, 2005), although both studies had confounding variables (e.g., distance to sea) that may compromise the overall conclusions. One of the most destructive natural threats to mangroves is the passage of hurricanes. Hurricane-associated wind effects include defoliation, snapping of trunks/ limbs, and uprooting, all of which can lead to immediate or delayed mortality of mangroves (Baldwin et al., 2001, Sherman et al., 2001). Susceptibility to damage, mortality rates, and potential for regeneration all vary depending upon mangrove species, location (e.g., fringe or inland), and tree size (Roth, 1992, Imbert et al., 1996, McCoy et al., 1996, Baldwin et al., 2001, Sherman et al., 2001). Regardless, hurri-

Bulletin of Marine Science 805 © 2007 Rosenstiel School of Marine and Atmospheric Science of the University of Miami 806 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 cane-associated wind damage could indirectly affect mangrove fish assemblages by altering key aspects of the habitat that benefit fish. Loss of canopy and defoliation decreases shade and increases light penetration (Baldwin et al., 2001, Sherman et al., 2001), which may influence shelter quality (Laegdsgaard and Johnson, 2001; Cocher- et de la Morinière et al., 2004; Ellis and Bell, 2004) and composition of the epibiont community (Farnsworth and Ellison, 1996). Mangrove mortality results in decreased tree density and basal area and ultimately a reduction in structural complexity (Roth, 1992, Baldwin et al., 2001). This could also negatively affect shelter and feeding. Oth- er hurricane-induced effects such as lowered dissolved oxygen and salinity may also be of consequence to fish assemblages (Burkholder et al., 2004; Stevens et al., 2006). To our knowledge, no published studies have closely examined hurricane-associated changes in fish assemblages in mangrove habitats. Silverman (2006) demonstrated significant differences in nekton-community structure of Big Sable Creek (South Florida) when comparing intact mangrove areas with areas of former mangrove that were converted to mudflats by major hurricanes in 1935 and 1960. In this study, we examined the effects of Hurricane Charley on assemblages of large-bodied (> ~100 mm standard length, SL) fishes associated with red mangrove shorelines of Charlotte Harbor, southwest Florida. We sought answers to two main questions: (1) Was post-hurricane fish-assemblage structure related to degree of wind-associated mangrove damage, and (2) Was the post-hurricane assemblage of mangrove-shoreline fishes notably different from long-term (1996–2004) pre-hurricane assemblages?

Methods

Study Area.—Charlotte Harbor is a shallow (generally < 4 m), drowned-river-valley estuary (Sheng, 1998) located in southwestern Florida (Fig. 1) and has a surface area of about 700 km2 (McPherson et al., 1996). A shallow shelf < 1.5 m deep extends 500–1200 m from the shoreline, beyond which deeper, unvegetated habitats (i.e., sand substrates) predominate. Tidal range is typically < 0.7 m (NOAA, 2006). Principal freshwater inputs are from the Ca- loosahatchee, Peace, and Myakka rivers. Charlotte Harbor contains considerably large areas of vegetated habitat, including seagrass (predominantly turtle grass, Thalassia testudinum Banks and Soland. ex Koenig, shoal grass, Halodule wrightii Achers., and manatee grass, Syringodium filiforme Kuetz. and fringing mangroves (principally red mangrove, Rhizopho- ra mangle Linnaeus). Fringing mangroves cover an area of approximately 143 km2 (L. Kish, Manatee County Geographic Information Systems Division, unpubl. data). A more detailed description of Charlotte Harbor is available from McPherson et al. (1996). Hurricane Charley.—Hurricane Charley made landfall near North Captiva Island, Flor- ida, on Friday, 13 August 2004 (Weisberg and Zhen, 2006). The storm was notable in that its inner core was unusually compact and intense, with a path 16–24 km wide and maximum wind gusts of nearly 280 km h–1. Physical damage to Charlotte Harbor was primarily wind- related because the storm’s landfall location, direction of travel (NNE), high speed, and small eye radius kept storm surge minimal (Weisberg and Zheng, 2006). Following the storm’s passage, dissolved oxygen in one of Charlotte Harbor’s major tributaries (the Peace River) fell below 1 mg L–1, and hypoxic conditions extended approximately 15 km into the estuary (To- masko et al., 2006). A long-term fishery-independent monitoring program (see Greenwood et al., 2006) resumed within 2 wks of the storm’s landfall and sampling was intensified in the most heavily impacted areas to document changes in fish assemblages as a result of the hurricane (Stevens et al., 2006) At the mouth of the Peace River and upper Charlotte Harbor, fish abundance decreased dramatically, and typical estuarine fish assemblages were replaced by GREENWOOD ET AL.: POST-HURRICANE MANGROVE-SHORELINE FISH ASSEMBLAGES 807

Figure 1. Map of the study area. Sampling sites from the post-hurricane study are indicated with circles, and color coding indicates degree of mangrove canopy damage based on visual assessment: white (no damage), light gray (10%–30% damaged), dark gray (40%–70% damaged), and black (80%–100% damaged). The path of Hurricane Charley’s eye wall is shown by dashed poly- gons (digitized by T. Walker, Southwest Florida Regional Planning Council, from local Doppler radar imaging). Dashed straight lines indicate the southern limits of the study area for the pre- and post-hurricane comparison analysis. Cross-hatching indicates mangrove habitat. 808 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Figure 2. Examples of red mangrove damage from winds associated with Hurricane Charley. Note that trunks and major branches are broken and pushed landward, but mangrove prop roots remain. (A) A 50%-damaged mangrove shoreline, with bar indicating undamaged, verdant areas; (B) a 100%-damaged area. those dominated by a few resilient estuarine and freshwater species including the mosquitofish,Gambusia holbrooki Girard, 1859; Florida gar Lepisosteus platyrhincus Dekay, 1842; and exotic armored catfishes, Hoplosternum littorale (Hancock, 1828) and Pterygoplichthys spp. Fish abundance and composition in other areas of the estuary did not noticeably differ from those expected during summer and fall, including those in the Myakka River, located only a few kilometers west of the storm track. The hypoxic event was short-lived as dissolved oxygen levels and estuarine fish assemblages recovered within a month (Stevens et al., 2006). The most conspicuous, perhaps long-lasting, effect of Hurricane Charley was the extensive mangrove damage caused by high winds (Fig. 2). Damage to mangroves included defoliation and snapping of limbs and trunks; damage was particularly substantial along the eastern edge of the hurricane’s path. Decreases in mangrove canopy density attributable to hurricane winds ranged from ~70% at < 4 km from the hurricane eye-wall to ~25% at 19 km from the eye-wall, with very little evidence of recovery in mangrove canopy density 8 mo after the hurricane’s passage (Milbrandt et al., 2006). The effects of the mangrove damage to fish assemblages in Charlotte Harbor, particularly large-bodied shoreline associates, have not been assessed until now. Sampling Methods.—Sampling sites were randomly selected on a monthly basis using established protocols of the State of Florida Fish and Wildlife Conservation Commission’s Fish and Wildlife Research Institute’s Fisheries-Independent Monitoring program (Poulakis et al., 2003). We sampled the assemblage of large-bodied fish (> ~100 mm SL) along red mangrove shorelines with 183- × 3-m center-bag haul seines (38-mm stretched mesh) set from a small research vessel in an approximate rectangle of 4120 m2 (103-m length along shoreline and 40-m width away from shore), with the shoreline forming one of the long edges of the rectangle. The net was pulled ashore by hand and the sample was concentrated in the bag. GREENWOOD ET AL.: POST-HURRICANE MANGROVE-SHORELINE FISH ASSEMBLAGES 809

Sampling effort was ~10 samples mo–1 from November 1996 to October 2004 and ~13 samples mo–1 thereafter. Each site was sampled once. Fish were identified to lowest practical taxon (generally species); nomenclature follows Nel- son et al. (2004). Fish were released after data collection, except in cases when laboratory identification was necessary. Temperature and salinity were recorded at surface, bottom, and 1-m intervals between surface and bottom with Hydrolab® or YSI Inc. probes at the location of the center bag. Water depth was also recorded to the nearest 10 cm at each site near the center bag. Visual estimates of habitat characteristics at each sampling site included proportion of shoreline covered by red mangroves, whether mangroves were inundated at the time of sampling, approximate width of inundated mangrove habitat from its seaward edge to the land-water interface, type of mangrove shoreline (i.e., fringing mangrove backed by a berm or intermittent overwash island), and proportion of sampling area covered by submerged aquatic vegetation (SAV, to the nearest 10%). Tidal stage at the time of sampling was noted (Table 1). More details of sampling procedures were described by Kupschus and Tremain (2001). Fol- lowing Hurricane Charley, the proportion of red mangroves damaged along the shoreline of each sampling site was visually estimated (to the nearest 10%; recorded from November 2004 onwards) (Table 1). Typical mangrove damage consisted of broken trunks and branches and leafless portions of trees or entire trees (no canopy), but prop roots remained largely intact (Fig. 2). In only a few cases were mangroves completely uprooted and toppled. Data Analysis.—We limited data analysis to shorelines with at least 80% red mangrove coverage during November 1996–October 2005. Mangrove shoreline samples extended 40 m out into adjacent nearshore habitat that often included large areas of seagrass beds. There- fore, only common species that could be considered shoreline associates were included in the analysis. The selection of these species was based upon a comparison of data collected with the 183-m haul seines and data collected away from the shoreline with 183-m purse seines (R. H. McMichael Jr., Fish and Wildlife Research Institute, unpubl. data; Wessel and Winner, 2003), as well as consideration of other relevant studies (Ley et al., 1999; Ley and McIvor, 2002; Faunce et al., 2004). Species included in analyses were striped mullet, Mugil cephalus Linnaeus, 1758; whirligig mullet, Mugil gyrans (Jordan and Gilbert, 1884); redfin needlefish, Strongylura notata (Poey, 1860); common snook, Centropomus undecimalis (Bloch, 1792); gray snapper, Lutjanus griseus (Linnaeus, 1758); striped mojarra, Eugerres plumieri (Cuvier, 1830); sheepshead, Archosargus probatocephalus (Walbaum, 1792); red drum, Sciaenops ocellatus (Linnaeus, 1766); and Atlantic spadefish, Chaetodipterus faber (Broussonet, 1782). Post-hurricane data (November 2004–October 2005) were used to address the study’s first question: Was fish assemblage structure related to degree of wind-associated mangrove damage? For each sample, abundance data were square-root-transformed to reduce the influence of unusually high catches (ter Braak, 1986). We conducted canonical correspondence analysis (CCA; ter Braak, 1986) to assess significant environmental variables correlating with assemblage structure (Table 1). The variables considered were percentage of shoreline with hurricane damage, average temperature, average salinity, water depth (square-root-transformed), submerged aquatic vegetation (SAV) coverage at the sampling site, inundation of mangrove, width of inundated mangrove habitat, type of mangrove shoreline, and tidal stage. We followed the procedure of Marshall and Elliott (1998) by initially conducting nine exploratory CCA analyses with each environmental variable included separately, in order to determine an approximate rank of explanatory power for each variable. Following this determination, variables were added to a CCA model in order of their rank, with statistical significance of each additional term to the model’s explanatory power being assessed with a permutation test incorporating 1000 random data configurations (Lepš and Šmilauer, 2003). Variables were accepted into the model if statistically significant at P < 0.05. In this manner, a final model was obtained that included only significant explanatory variables and we assessed the relationship of the fish species to the environment with a CCA biplot. A dummy species with unit abundance in all samples was added to the analysis to keep potentially important all-zero samples under consideration and allow the CCA model to calculate weighted averages for the 810 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 Factor level (% of total samples) or continuous variable seasonal means (ranges) °C) (19.1–34.1 °C ’05–Oct. Jul. °C); 29.8 ’05: (13.7–30.9 °C 24.6 ’05–Jun. Mar. ’05: °C); (14.2–25.1 °C ’04–Feb. 21.5 ’05: Nov. ’05–Jun. ’05: 99.0 cm (50–250 cm); Jul. ’05–Oct. 98.6 (40–270 cm) ’04–Feb. ’05: 93.6 cm (40–240 cm); Mar. Nov. ’05–Jun. ’05: 23.9 (4.3–34.6); Jul. ’05–Oct. ’05 : 17.9 (2.6–33.1) ’04–Feb. ’05: 26.6 (13.6–35.4); Mar. Nov. slack (1.3%); mid, rising (14.3%); low, falling (7.8%); low, High, falling (15.6%); high, rising (18.8%); slack (5.8%); low, falling (10.4%); mid, rising (26.0%) 0% (31.8%), 10–50% (18.8%), > 50% (49.4%) < 1 m (24.7%); 1–2 (32.5%); 3–9 (26.6%); ≥ 10 (16.2%) 0% (27.3%); 10%–30% (23.4%); 40%–70% (26.0%); 80%–100% (23.4%) Inundated (91.6%); not inundated (8.4%) Fringing (90.3%); overwash island (9.7%) Table 1. Summary of Table variables considered in stepwise Analysis Canonical of Correspondence Charlotte Harbor 183-m haul-seine data from November 2004 to October 2005. Levels of factors are noted with percentages of the total number of samples (n = 154) that they constituted. Means (and range of monthly means are ranked in order of inclusion models. Variables in parentheses) are shown for continuous variables. Variable Temperature depth Water Salinity stage Tidal coverage SAV Mangrove habitat width Mangrove shoreline damage Inundation Shoreline type GREENWOOD ET AL.: POST-HURRICANE MANGROVE-SHORELINE FISH ASSEMBLAGES 811 nine study species (CCA uses a weighted-average algorithm that otherwise cannot include all-zero samples). We compared the post-hurricane fish assemblage with pre-hurricane assemblages to address the study’s second question: Was the post-hurricane mangrove-shoreline fish assemblage notably different from the pre-hurricane assemblage? This analysis excluded data from the more southerly portions of the study area (Fig. 1), because these areas were not sampled prior to November 2003. Abundance data from 1996 to 2005 were square-root-transformed, grouped into three 4-mo periods for each year (November–February, March–June, and July– October) and averaged (number of fish per set), generating 27 “samples” for analysis. Data were transformed prior to averaging to reduce the influence of intersample variability (K. R. Clarke, PRIMER-E Ltd, pers. comm.). The three periods were defined in this manner not only to be consistent with the period for which mangrove damage data had been collected (i.e., November 2004 onwards), but also because they were appropriate demarcations of the main meteorological seasons (November–February: cool and dry; March–June: warm and dry; July–October: hot and wet). Note that the data for the July–October period in 2004 therefore encompassed both pre-hurricane and post-hurricane conditions. The data were then subjected to correspondence analysis (CA; Lepš and Šmilauer, 2003), the unconstrained (i.e., free of environmental variables) ordination analog of CCA (see above). From the resulting CA biplot, we assessed the extent to which post-hurricane fish assemblages differed from pre-hurricane assemblages. By conducting an unconstrained analysis, we concentrated on fish-assemblage structure without regard to environmental conditions. All analyses were conducted using the vegan Community Ecology Package v. 1.6-10 (Oksanen et al., 2005), implemented in R v. 2.2.1 (R Development Core Team, 2005).

Results

Overall Sample Composition.—In total, 14,230 fishes of the nine study species were collected (Table 2). This represented just over 11% of the total abundance of the 119 taxa collected during the study period. Nearly 90% of the remaining fish were one of five abundant species that could not be considered shoreline associates [pinfish, Lagodon rhomboides (Linnaeus, 1766); tidewater mojarra, Eucinostomus harengulus Goode and Bean, 1879; silver jenny, Eucinostomus gula (Quoy and Gaimard, 1824); silver perch, Bairdiella chrysoura (Lacépède, 1802); pigfish, Orthopristis chrysoptera (Linnaeus, 1766); and hardhead catfish, Ariopsis felis (Linnaeus, 1766)]. Centropomus undecimalis (Bloch, 1792) was both the most abundant and most frequently collected of the study species, contributing more than 27% of the total catch. Sciaenops ocellatus and E. plumieri were the two least abundant species, with each contributing just under 6% of the total catch. Chaetodipterus faber was the least frequently collected species and was present in just over 15% of samples. Mean sizes of fish ranged from 151 mm SL (L. griseus) to 449 mm (C. undecimalis) (Table 2). Post-hurricane Analysis: Was Fish-Assemblage Structure Related to Degree of Wind-associated Mangrove Damage?—There was no evidence that wind-associated damage affected the structure of the fish assemblages along red mangrove shorelines in Charlotte Harbor from November 2004 to October 2005. The final CCA model retained temperature, water depth, and salinity as variables that were significantly related to fish-assemblage structure. The overall model was statistically significant (P < 0.001), but explained only a modest proportion of the total inertia (data variability). The inclusion of the percentage-of-damaged-mangrove-shoreline variable did not significantly increase the explanatory power of the model including the above three covariables (P = 0.36; Table 3). CCA axis 1 explained 812 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 185.4 (42–331) 185.4 180.9 (55–333) 180.9 151.1 (45–280) 151.1 159.4 (32–300) 159.4 394.8 (57–727) 394.8 211.3 (24–440) 211.3 289.9 (83–466) 289.9 449.4 (101–960) 449.4 344.4 (160–507) SL (Mean,SL range) CPUE n = 360) n = 360) (Jul.–Oct., 0.72 ± 0.11 0.72 0.89 ± 0.14 0.89 1.13 ± 0.20 1.13 0.58 ± 0.12 0.58 0.50 ± 0.13 0.50 4.60 ± 0.70 4.60 2.49 ± 0.73 2.49 1.85 ± 0.22 1.85 1.86 ± 0.20 1.86 CPUE n = 357) 0.74 ± 0.10 0.74 1.37 ± 0.18 1.37 1.11 ± 0.22 1.11 ± 0.19 0.77 0.64 ± 0.12 1.55 ± 0.20 1.55 3.36 ± 0.39 3.36 0.67 ± 0.08 0.67 0.50 ± 0.09 0.50 (Mar.–Jun., CPUE n = 344) 1.22 ± 0.15 1.22 0.46 ± 0.18 0.46 0.29 ± 0.10 0.29 2.75 ± 0.50 2.75 0.94 ± 0.12 0.94 1.53 ± 0.30 1.53 3.34 ± 0.57 3.34 0.25 ± 0.05 0.25 (Nov.–Feb., (Nov.–Feb., 2.35 ± 0.037 2.35 15.11% 31.13% 29.41% 24.71% 49.86% 16.56% 16.92% 44.43% 36.29% (n = 1,105) (n occurrence Frequency of CPUE (n = 1,105) (n 1.58 ± 0.15 1.58 0.84 ± 0.11 0.84 0.75 ± 0.09 0.75 0.74 ± 0.06 0.74 3.50 ± 0.31 3.50 2.35 ± 0.21 2.35 1.30 ± 0.27 1.30 0.93 ± 0.09 0.93 0.88 ± 0.07 0.88 821 833 973 926 Total Total 1,435 1,744 1,030 3,873 2,595 number Chaetodipterus faber Strongylura notata Mugil gyrans Eugerres plumieri Mugil cephalus Lutjanus griseus Archosargus probatocephalus Sciaenops ocellatus Gray snapper, Sheepshead, Common snook, Centropomus undecimalis Striped mojarra, Red drum, Whirligig mullet, Redfin needlefish, Striped mullet, Atlantic spadefish, Table 2. Catch composition of principal fish species collected along red mangrove shorelines of Charlotte Harbor, November 1996–October 2005. Catch per unit per Catch 2005. 1996–October November Harbor, Charlotte of shorelines mangrove red along collected species fish principal of composition Catch 2. Table only calculated were means Seasonal present. was species a which in samples all of percentage is occurrence of Frequency SE). (± haul per fish is (CPUE) effort is standard length (mm). Sample sizes are noted in parentheses. for areas included in the pre- and posthurricane comparison analysis. SL GREENWOOD ET AL.: POST-HURRICANE MANGROVE-SHORELINE FISH ASSEMBLAGES 813

Table 3. Results of stepwise Canonical Correspondence Analysis model analysis of Charlotte Harbor 183-m haul-seine data from November 2004 to October 2005. Statistically significant additions are noted in bold, with conditional variables at each step in parentheses.

Constraining variables Significance of added term 1. (Temperature) + water depth 0.002 2. (Temperature + water depth) + salinity 0.017 3. (Temperature + water depth + salinity) + tidal stage 0.4 4. (Temperature + water depth + salinity) + SAV coverage 0.121 5. (Temperature + water depth + salinity) + habitat width 0.092 6. (Temperature + water depth + salinity) + mangrove damage 0.363 7. (Temperature + water depth + salinity) + inundation 0.287 8. (Temperature + water depth + salinity) + shoreline type 0.582

7.8% of total inertia and described fish-assemblage differences attributable to temperature (Fig. 3). Species most strongly associated with lower temperatures included S. notata and M. cephalus, whereas C. faber tended to be most abundant at higher temperatures. CCA axis 2 explained 2.8% of total inertia and was related to assemblage differences caused by differing salinities. This demonstrated that E. plumieri, for example, tended to be found at very low salinities. Water depth was quite strongly associated with both of the CCA axes. Pre- and Post-hurricane Comparison Analysis: Was the Post-hurricane Mangrove-shoreline Fish Assemblage Notably Different from Pre-hurricane Assemblages?—There was little evidence of post-hurricane fish assemblages associated with red mangrove shorelines greatly differing from pre-hurricane assemblages. Post-hurricane assemblages tended to be well mixed among those found in pre-hurricane years, as seen on the CA plot (Fig. 4). The same was true of the assemblage present from July to October 2004 (i.e., encompassing August 2004 when Hurricane Charley affected the area; Fig. 4). The post-hurricane assemblage from March to June 2005 was somewhat different from this season’s assemblages in most of the remaining years, primarily because of relatively high abundance of S. notata and M. gyrans, and also because of relatively low abundance of E. plumieri, A. probatocephalus, and C. undecimalis. The March to June assemblage from 2004 was more of an outlier, however (Fig. 4). CA axis 1 explained nearly 50% of the total inertia and was related to seasonal differences in abundance: species were generally either most abundant from November to February (M. gyrans, S. ocellatus, S. notata, M. cephalus, and A. probatocephalus) or most abundant from July to October (L. griseus, C. faber, and S. ocellatus). Only E. plumieri had highest mean abundance from March to June. The centroids ofC. undecimalis and A. probatocephalus were close to March to June in the biplot because their annual peaks in abundance—from July to October (C. undecimalis) and from November to February (A. probatocephalus)—were not substantially greater than during March–June. CA axis 2 explained just over 15% of total inertia and was related to interannual differences in abundance.

Discussion

Results of our study suggested that wind-associated damage to red mangroves along the Charlotte Harbor shoreline did not greatly affect the structure of assemblages of large-bodied fish. Detailed experimental studies of the importance of man- 814 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Figure 3. Canonical Correspondence Analysis species-environment biplot of the post-hurricane study to assess effects of mangrove damage on fish-assemblage structure. Only environmental variables significantly related to assemblage structure are shown. Each species label is adjacent to the point representing its centroid.

grove-habitat attributes such as shade and structure have tended to focus on the nursery function of these areas for small fishes (Primavera, 1997; Laegdsgaard and Johnson, 2001; Cocheret de la Morinière et al., 2004; Ellis and Bell, 2004). Field studies have provided more information on fishes comparable in size to those described in the present study. Ley and McIvor (2002), working in Florida Bay, observed that L. griseus were positively associated with a compound environmental axis representing increasing mangrove fringe width, increasing mangrove tree height, decreasing prop-root density, increasing water depth, and increasing SAV volume. Faunce et al. (2004) found positive correlations between L. griseus abundance and width of the prop-root habitat (and also creek depth; see below) in the southeastern saline Everglades, but found no relationship between abundance and prop-root density or diameter. Of other species that we studied in Charlotte Harbor, Faunce et al. (2004) could not establish any relationship between abundance of S. notata and any of the aforementioned features of the mangrove habitat. It is therefore apparent that while some species-habitat correlations involving mangroves exist, relationships are not consistent for different species or for the same species from different locations. In this light, it is perhaps not surprising that we did not observe any differences in fish assemblage with various mangrove habitat characteristics. The only study that we could locate addressing the relationship between changes to mangrove canopy and GREENWOOD ET AL.: POST-HURRICANE MANGROVE-SHORELINE FISH ASSEMBLAGES 815

Figure 4. Canonical Correspondence Analysis species-sample biplot of the pre- and post-hurricane comparison study to assess effects of mangrove damage on fish-assemblage structure. Sample labels indicate year and period, where NF = November–February, MJ = March–June, and JO = July–October. Note that samples labeled NF represent samples from November of one year to February of the following year. Full-box labels indicate post-hurricane samples, dashed-box label indicates samples collected from July to October 2004 (i.e., encompassing Hurricane Charley in August 2004). Each label represents the mean of all 183-m seine samples collected during that period. Each species or sample label is adjacent to the point representing its centroid. fish-assemblage structure was that of Ellis (2003). Using drop block nets, he found mangrove-canopy trimming to have no effect on overall fish assemblage structure or on abundance of large roving species (including species included in our study) in Rookery Bay, Florida. Ellis (2003) also noted that variability among samples within the same habitat is often so great that alteration to habitat may have to be substantial to produce discernible changes in fish abundance or assemblage structure. The same may have been true of our study. With regard to other features of mangrove structure that we investigated, it is important to note that larger fish will tend not to penetrate very far into inundated mangroves (Vance et al., 1996; Rönnbäck et al., 1999); this could explain why mangrove habitat width appeared unimportant in structuring the fish assemblage. To our knowledge, no attempt has been made to ascertain the differences between fish assemblages in fringing mangrove forests and those in overwash-island mangrove forests. Gilmore and Snedaker (1993) did not differentiate the two forests in terms of overall lists of fish species. Some habitat attributes of the two types of forest differ (e.g., average tree height and stand diameter; Gilmore and 816 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Snedaker, 1993). Our limited sampling of the overwash habitat precludes our making firm conclusions regarding any differences in the fish assemblages of these habitats. Our study demonstrated that the post-hurricane large-bodied mangrove-shoreline fish assemblage was not markedly different from pre-hurricane assemblages, suggesting a degree of resiliency to seemingly great change. Any direct mortality due to the effects of the hurricane was likely to have been minimal (cf. Tabb and Jones, 1962), but some fish kills and short-term changes to the large-bodied-fish assemblage structure were observed as a result of hypoxic conditions attributable to mangrove and riparian-vegetation defoliation (Stevens et al., 2006). These effects were temporary and apparently induced changes to fish-assemblage structure that were within the range of variability observed from 1996 to 2004. Small-bodied-fish assemblage structure also changed in the short term, but generally reverted back to a more typical composition within a period of weeks (Stevens et al., 2006). Greenwood et al. (2006) showed that the change in small-bodied shoreline fish assemblage following Hurricane Charley was outside the range of variability observed in the previous 8 yrs, with apparent declines in some vegetation-associated species. Additional years of data collection are required to examine the long-term effects of vegetation damage, both to mangroves and seagrasses, from Hurricane Charley. Regarding the lack of substantial change in the post-hurricane large-bodied mangrove-shoreline fish assemblage compared to the pre-hurricane period, it is important to note that the pre-hurricane period was rather variable in terms of environmental conditions: the winter of 1997–98 included a strong El Niño event, characterized by relatively high rainfall, followed by a period of more than 2 yrs of La Niña-influenced drought (Schmidt et al., 2004). Also, an important fishery-management measure (elimination of entangling gill nets) was enacted in July 1995 and caused a major decrease in fishing mortality of several of the species considered in this study, particularly M. cephalus. This produced increases in overall abundance over the first few years of our study period (Pierce et al., 1998). In some areas, the mortality of red mangroves following Hurricane Charley is gradually leading to the lifting of prop roots from the substrate as standing dead trees dry out and straighten; the prop roots are also gradually decaying (T.A. Tat- tar, University of Massachusetts, pers. comm.). In general there appears to be very poor recruitment of red mangrove saplings to damaged areas, suggesting that standing dead trees may decay quicker than fringing mangrove can recover, ultimately converting forest to mud flat. This phenomenon was observed in Big Sable Creek, Florida (Silverman, 2006). These longer-term aspects of change may eventually lead to greater effects on shoreline biotic communities than we observed in the year following the hurricane. Also, the effect of widespread mangrove mortality on the many other mangrove functions (e.g., primary productivity, food-web and nutrient dynamics) may ultimately lead to system-level changes that affect abundance of large-bodied fishes. Our continued study of Charlotte Harbor will allow us to re-evaluate the shoreline fish assemblages in the future. Our results showed that of the variables we assessed, physical attributes of the environment (temperature, water depth, and salinity) most influenced the structure of the fish assemblage. This agrees with the overall conclusions of Kupschus and Tre- main (2001), who sampled the ichthyofauna of the Indian River Lagoon (east Florida) with the same gear and procedures. Some very obvious similarities existed between the fish-assemblage associations in Charlotte Harbor and those in the Indian River GREENWOOD ET AL.: POST-HURRICANE MANGROVE-SHORELINE FISH ASSEMBLAGES 817

Lagoon. For example, S. ocellatus, M. cephalus, and S. notata tended to be found in cooler temperatures, whereas L. griseus and C. faber were more abundant during warmer periods and in deeper areas. The importance of temperature in structuring the fish assemblage is most likely the result of seasonal changes in abundance as opposed to avoidance of cooler areas, although the latter effect should not be dis- counted for some species (e.g., C. undecimalis; Pattillo et al., 1997). Ley at al. (1999) found positive associations between L. griseus abundance and increasing temperature, salinity, and water depth. Ley and McIvor (2002) and Faunce et al. (2004) also found a positive association between L. griseus and water depth. Thus it appears that greater water depth is an important factor governing abundance of L. griseus, which may be due to improved access to prey at greater depths. We did not find any association between fish-assemblage structure and SAV coverage, which was not surprising given our selection of species that are principally shoreline associates. Although Kupschus and Tremain (2001) found that SAV coverage was of importance in structuring the Indian River Lagoon fish assemblage, eight of the nine species that we considered in this study had no clear association with SAV in their study area (E. plumieri was the exception and was less common in vegetated areas). Furthermore, our study was limited to red mangrove shorelines and thus eliminated one of the major factors structuring fish assemblages in tropical and subtropical estuaries, because overhanging shorelines often have markedly different assemblages compared to beach or non-overhanging shorelines (Sheaves, 1992, 1996; Kupschus and Tremain, 2001). Regardless, the presence of red mangrove prop-root habitat may be of greatest importance in determining fish distribution; changes to the habitat may not be as important as its absence (Silverman, 2006). It is appropriate to acknowledge the study’s main limitations. Firstly, our sampling technique probably did not provide the same degree of semi-quantitative information on fishes associated with mangrove-prop-root habitats as the information obtained from studies using block nets or visual surveys (Sheridan and Hays, 2003). Although we consider the species chosen for analysis to be shoreline-oriented, we acknowledge that our sampling technique could have captured them away from the shoreline. Secondly, data from this study were collected during a fishery-independent monitoring program designed to monitor interannual trends in fish abundance—we had to subset a larger database of samples to find red mangrove shorelines meeting our criteria for inclusion, which inevitably led to a relatively low sample size and lack of control of other environmental variables. A dedicated study (perhaps with fixed sampling stations) aiming to control as many confounding environmental variables as possible may provide clearer indications of the effects of mangrove damage on fish assemblages. Thirdly, the correlations between fish assemblage structure and environment were relatively weak, although statistically significant (and fairly typical for this type of analysis). This may be because of large intersample variability in abundance (with many zero catches) or perhaps because important predictors of assemblage structure had been excluded from the analysis, e.g., food abundance or habitat structure of adjacent areas. The study should therefore be regarded as largely exploratory as opposed to predictive. Finally, the mangrove-damage metric that we employed was based on a simple visual assessment and did not attempt to quantify potentially important variables such as below-canopy light penetration or structural complexity. 818 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

We conclude that hurricane damage to red mangrove shorelines had little apparent effect on the structure of the shoreline-associated large-bodied fishes. We focused solely on abundance—and, by extension, assemblage structure—but other potentially important aspects of change may be worthy of investigation (e.g., mean size, biomass, condition, and disease prevalence). To more accurately gauge the influence of mangrove damage on shoreline fish assemblages would require a dedicated study in which sampling gear that is more effective at accurately assessing abundance within the prop-root habitat is used (e.g., visual census [video or direct observation] or block-netting) and in which detailed measurements are made of mangrove characteristics such as prop-root density, below-canopy light penetration, and epibiont coverage. Such a study would ideally be undertaken in a relatively small geographical area to minimize the influences of differing physico-chemical conditions (e.g., salinity, distance to inlets) on assemblage structure (see Ellis, 2003). Although not without limitations (see above), our sampling regime had the strength of addressing fishery issues at broad temporal and spatial scales. For example, it was clear that recreationally important fishes included in this study (e.g., C. undecimalis, S. ocellatus) had not abandoned damaged areas and were not found solely in areas where mangroves remained undamaged. Also, populations of larger fishes were not (as yet) grossly affected by Hurricane Charley and the associated wind damage to mangroves. These issues will remain important while more is learned about the rate of mangrove decay and recovery. A comprehensive oversight of the situation will ultimately come from broad, random sampling of the estuary, which would allow the state of fisheries to be viewed from the perspective of long-term trends in abundance and habitat use on an estuary-wide scale.

Acknowledgments

We thank the field crews of the Florida Fish and Wildlife Conservation Commission’s Fish- eries-Independent Monitoring program (Charlotte Harbor Field Laboratory) for collecting and processing the data. Support for this study was provided in part by funds from Florida Recreational Saltwater Fishing License sales and U.S. Department of the Interior, Fish and Wildlife Service, Federal Aid for Sportfish Restoration Project Number F-43. Thanks to J. Leiby, J. Quinn, R. McMichael, D. Blewett, and two anonymous reviewers for constructive comments on an earlier draft of the manuscript. L. French’s assistance with figure preparation is greatly appreciated.

Literature Cited

Baldwin, A., M. Egnotovich, M. Ford, and W. Platt. 2001. Regeneration in fringe mangrove forests damaged by Hurricane Andrew. Plant Ecol. 157: 151–164. Burkholder, J., D. Eggleston, H. Glasgow, C. Brownie, R. Reed, G. Janowitz, M. Posey, G. Melia, C. Kinder, R. Corbett, D. Toms, T. Alphin, N. Deamer, and J. Springer. 2004. Comparative impacts of two major hurricane seasons on the Neuse River and western Pamlico Sound ecosystems. Proc. Natl. Acad. Sci. U. S. A. 101: 9291–9296. Cocheret de la Morinière, E., I. Nagelkerken, H. van der Melj, and G. van derVelde. 2004. What attracts juvenile coral reef fish to mangroves: habitat complexity or shade? Mar. Biol. 144: 139–145. Dankwa, H. R. and C. Gordon. 2002. The fish and fisheries of the Lower Volta mangrove swamps in Ghana. African journal of science and technology (AJST). Science and Engi- neering Series 3: 25–32. GREENWOOD ET AL.: POST-HURRICANE MANGROVE-SHORELINE FISH ASSEMBLAGES 819

Ellis, W. L. 2003. Mangrove canopy damage: implications for litter processing and fish utilization of the intertidal zone. PhD Diss. Univ. South Florida, Tampa, Florida. 157 p. ______and S. S. Bell. 2004. Conditional use of mangrove habitats by fishes: depth as a cue to avoid predators. Estuaries 27: 966–976. Farnsworth, E. J. and A. M. Ellison. 1996. Scale-dependent spatial and temporal variability in biogeography of mangrove root epibiont communities. Ecol. Monogr. 66: 45–66. Faunce, C. H., J. E. Serafy, and J. J. Lorenz. 2004. Density–habitat relationships of mangrove creek fishes within the southeastern saline Everglades (USA), with reference to managed freshwater releases. Wetl. Ecol. Manag. 12: 377–394. Gilmore, R. G., Jr. and S. C. Snedaker. 1993. Mangrove forests. Pages 165–198 in W. H. Martin, S. G. Boyce, and A. C. Echternacht, eds. Biodiversity of the southeastern United States. John Wiley and Sons, Inc., New York. Greenwood, M. F. D., P. W. Stevens, and R. E. Matheson, Jr. 2006. Effects of the 2004 hurricanes on the fish assemblages in two proximate southwest Florida estuaries: change in the context of multi-annual variability. Estuaries and Coasts 29: 985–996. Imbert, D., P. Labbe, and A. Rousteau. 1996. Hurricane damage and forest structure in Guade- loupe, French West Indies. J. Trop. Ecol. 12: 663–680. Kathiresan, K. and B. L. Bingham. 2001. Biology of mangroves and mangrove ecosystems. Adv. Mar. Biol. 40: 81–251. Kieckbusch, D. K., M. S. Koch, J. E. Serafy, and W. T. Anderson. 2004. Trophic linkages among primary producers and consumers in fringing mangroves of subtropical lagoons. Bull. Mar. Sci. 74: 271–285. Kupschus, S. and D. Tremain. 2001. Associations between fish assemblages and environmental factors in nearshore habitats of a subtropical estuary. J. Fish Biol. 58: 1383–1403. Laegdsgaard, P. and C. Johnson. 2001. Why do juvenile fish utilise mangrove habitats? J. Exp. Mar. Biol. Ecol. 257: 229–253. Lepš, J. and P. Šmilauer. 2003. Multivariate analysis of ecological data using CANOCO. Cam- bridge University Press, Cambridge, UK. 282 p. Ley, J. A. and C. C. McIvor. 2002. Linkages between estuarine and reef fish assemblages: enhancement by the presence of well-developed mangrove shorelines. Pages 539–562 in J. W. Porter and K. G. Porter, eds. The Everglades, Florida Bay, and coral reefs of the Florida Keys. CRC Press, Boca Raton. ______, ______, and C. L. Montague. 1999. Fishes in mangrove prop-root habitats of northeastern Florida Bay: distinct assemblages across an estuarine gradient. Estuar. Coast. Shelf Sci. 48: 701–723. Manson, F. J., N. R. Loneragan, G. A. Skilleter, and S. R. Phinn. 2005a. An evaluation of the evidence for linkages between mangroves and fisheries: a synthesis of the literature and identification of research directions. Oceanogr. Mar. Biol. Ann. Rev. 43: 485–515. ______, ______, B. D. Harch, G. A. Skilleter, and L. Williams. 2005b. A broad-scale analysis of links between coastal fisheries production and mangrove extent: a case-study for northeastern Australia. Fish. Res. 74: 69–85. Marshall, S. and M. Elliott. 1998. Environmental influences on the fish assemblage of the Hum- ber Estuary, U.K. Estuar. Coast. Shelf Sci. 46: 175–184. McCoy, E. D., H. R. Mushinsky, D. Johnson, and W. E. Meshaka Jr. 1996. Mangrove damage caused by Hurricane Andrew on the southwestern coast of Florida. Bull. Mar. Sci. 59: 1–8. McPherson, B. F., R. L. Miller, and Y. E. Stoker. 1996. Physical, chemical, and biological characteristics of the Charlotte Harbor basin and estuarine system in southwestern Florida–a summary of the 1982–89 U.S. Geological Survey Charlotte Harbor assessment and other studies. U.S. Geological Survey water-supply paper 2486. United States Government Print- ing Office, Washington. 32 p. Milbrandt, E. C., J. M. Greenawalt, P. D. Sokoloff, and S. A. Bortone. 2006. Response of southwest Florida mangroves to the 2004 hurricane season. Estuaries and Coasts 29: 979–1003. 820 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

NOAA (National Oceanic and Atmospheric Administration). 2006. NOAA Tides and Cur- rents. Silver Spring, Maryland: National Oceanographic and Atmospheric Administration/ Center for Operational Oceanographic Products and Services. Available from: http://tide- sandcurrents.noaa.gov/ Accessed 17 October 2006. Nelson, J. S., E. J. Crossman, H. Espinosa-Pérez, L. T. Findley, C. R. Gilber, R. N. Lea, and J. D. Williams. 2004. Common and scientific names of fishes from the United States, Canada, and Mexico, sixth edition. American Fisheries Society, Bethesda, Maryland. 386 p. Oksanen, J., R. Kindt, and R. B. O’Hara. 2005. Vegan: Community Ecology Package version 1.6- 10. Available from: http://cc.oulu.fi/~jarioksa/. Accessed 3 March 2006. Pattillo, M. E., T. E. Czapla, D. M. Nelson, and M. E. Monaco. 1997. Distribution and abundance of fishes and invertebrates in Gulf of Mexico estuaries, Volume II: species life history summaries. ELMR Rep. No. 11, NOAA/NOS Strategic Environmental Assessments Divi- sion, Silver Spring, MD. 377 p. Pierce, D. J., J. E. Wallin, and B. Mahmoudi. 1998. Spatial and temporal variations in the species composition of bycatch collected during a striped mullet (Mugil cephalus) survey. Gulf Mex. Sci. 16: 15–27. Poulakis, G. R., D. A. Blewett, and M. E. Mitchell. 2003. The effects of season and proximity to fringing mangroves on seagrass-associated fish communities in Charlotte Harbor, Florida. Gulf Mex. Sci. 21: 171–184. Primavera, J. H. 1997. Fish predation on mangrove-associated penaeids. The role of structures and substrate. J. Exp. Mar. Biol. Ecol. 215: 205–216. R Development Core Team. 2005. R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing. Available from: http://www.R- project.org. Accessed 3 March 2006. Rönnbäck, P., M. Troell, N. Kautsky, and J. H. Primavera. 1999. Distribution pattern of shrimps and fish amongAvicennia and Rhizophora microhabitats in the Pagbilao mangroves, Philip- pines. Estuar. Coast. Shelf Sci. 48: 223–234 . Roth, L. C. 1992. Hurricanes and mangrove regeneration: effects of Hurricane Joan, October 1988, on the vegetation of Isla de Venado, Bluefields, Nicaragua. Biotropica. 24: 375–384. Schmidt, N., M. E. Luther, and R. Johns. 2004. Climate variability and estuarine water resources: a case study from Tampa Bay, Florida. Coast. Manage. 32: 101–116. Sheaves, M. J. 1992. Patterns of distribution and abundance of fishes in different habitats of a mangrove-lined tropical estuary, as determined by fish trapping. Aust. J. Mar. Freshw. Res. 43: 1461–1479. ______. 1996. Habitat-specific distributions of some fishes in a tropical estuary. Mar. Freshw. Res. 47: 827–830. Sheng, P. S. 1998. Circulation in the Charlotte Harbor estuarine system. Pages 91–110 in S. F. Treat, ed. Proc. Charlotte Harbor Public Conference and Technical Symp., March 1997. South Florida Water Management District and Charlotte Harbor National Estuary Pro- gram, West Palm Beach and North Fort Myers, FL. Sheridan, P. and C. Hays. 2003. Are mangroves nursery habitat for transient fishes and decapods? Wetlands 23: 449–458. Sherman, R. E., T. J. Fahey, and P. Martinez. 2001. Hurricane impacts on a mangrove forest in the Dominican Republic: damage patterns and early recovery. Biotropica 33: 393–408. Silverman, N. L. 2006. Assessing the consequences of hurricane-induced fragmentation of mangrove forest on habitat and nekton in Big Sable Creek, Florida. MS Thesis. Univ. South Florida, College of Marine Science, St. Petersburg, Florida. 40 p. Singkran, N. and S. Sudara. 2005. Effects of changing environments of mangrove creeks on fish communities at Trat Bay, Thailand. Environ. Manage. 35: 45–55. Stevens, P. W., D. A. Blewett, and J. P. Casey. 2006. Short-term effects of a low dissolved oxygen event on estuarine fish assemblages following the passage of Hurricane Charley. Estuaries and Coasts 29: 997–1003. GREENWOOD ET AL.: POST-HURRICANE MANGROVE-SHORELINE FISH ASSEMBLAGES 821

Tabb, D. C. and A. C. Jones. 1962. Effect of Hurricane Donna on the aquatic fauna of north Florida Bay. Trans. Am. Fish. Soc. 91: 375–378. Tomasko, D. A., C. Anastasiou, and C. Kovach. 2006. Dissolved oxygen dynamics in Charlotte Harbor and its contributing watershed, in response to Hurricanes Charley and Frances and Jeanne—Impacts and Recovery. Estuaries and Coasts 29: 932–938. ter Braak, C. J. F. 1986. Canonical correspondence analysis: a new eigenvector technique for multivariate direct gradient analysis. Ecology 67: 1167–1179. Vance, D. J., M. D. E. Haywood, D. S. Heales, R. A. Kenyon, N. R. Loneragan, and R. C. Pendrey. 1996. How far do prawns and fish move into mangroves? Distribution of juvenile banana prawns Penaeus merguiensis and fish in a tropical mangrove forest in northern Australia. Mar. Ecol. Prog. Ser. 131: 115–124. Weisberg, R. H. and L. Zheng. 2006. A simulation of the Hurricane Charley storm surge and its breach of North Captiva Island. Fla. Sci. 69: 152–165. Wessel, M. R. and B. L. Winner. 2003. Using a modified purse seine to collect and monitor estuarine fishes. Gulf Caribb. Res. 15: 61–71.

Addresses: (M.F.D.G.) Florida Fish and Wildlife Conservation Commission, Fish and Wild- life Research Institute, 100 8th Ave SE, St Petersburg, Florida 33705. (C.F.I., P.W.S.) Florida Fish and Wildlife Conservation Commission, Fish and Wildlife Research Institute, Charlotte Harbor Field Laboratory, 1481 Market Circle Unit 1, Port Charlotte, Florida 33953. Corre- sponding Author: (M.F.D.G.) E-mail: . BULLETIN OF MARINE SCIENCE, 80(3): 823–837, 2007

Important Considerations to Achieve Successful Mangrove Forest Restoration with Optimum Fish Habitat

Roy R. Lewis III and R. Grant Gilmore

Abstract Mangrove forest restoration projects commonly fail to achieve significant plant cover for two reasons: because there is a misunderstanding of mangrove forest hydrology, or, acceptance of the false assumption that simply planting mangroves is all that is required to establish a fully-functional mangrove ecosystem. Even restoration projects that meet a restoration goal within 3–5 yrs often fail to provide adequate habitat for fish and invertebrates. Here we discuss how fish and mangrove ecosystems are coupled in time and space, offer several restoration strategies that match these couplings, and provide simple sequential checklist of design tasks to use to prevent most failures. Tidal hydrology must be carefully designed to incor- porate fish habitat, including tidal creeks, to provide access and low tide refuge for mobile nekton because the mangrove forest floor is generally flooded by tidal waters less than 30 percent of the time. A fully successful restoration design must mimic tidal stream morphology and hydrology along an estuarine gradient across a heterogeneous mixture of mangrove ecosystem communities.

Mangrove ecosystems currently cover 146,530 km² of the tropical shorelines of the world (FAO, 2003). This represents a decline from 198,000 km² of mangroves in 1980, and 157,630 km² in 1990 (FAO, 2003). These losses represent about 2.0% per year between 1980–1990, and 0.7% per year between 1990–2000. These figures show the magnitude of mangrove loss, and hence the magnitude of mangrove restoration opportunity, presented by areas like mosquito control impoundments in Florida, USA (Brockmeyer et al., 1997), and abandoned shrimp aquaculture ponds in Thailand and the Philippines (Stevenson et al., 1999). There are a multitude of opportunities to restore mangrove forests, and reasons to do so. The reasons are often based on fish habitat value, but include others such as direct wildlife use and dependence on fisheries production as a source of food. Here we discuss the basis for the fish—mangrove coupling and restoration efforts meant to restore these habitats. We offer several restoration strategies that match these couplings, and provide a simple sequential checklist of design tasks to be used to prevent most failures.

Coupling of Fish and Mangrove Habitats

We have previously written about the value of mangroves as habitat for fish and invertebrates (Gilmore et al., 1983; Lewis et al., 1985) and have noted the pioneering work of Heald (1971), Odum (1971), and Odum and Heald (1975) in Florida. Wolanski and Boto (1990) looked at similar issues with regard to specific mangrove forests in Thailand, Brazil, Malaysia, Japan, West Africa, and northern Australia and noted that “mangrove forests are a highly productive and valuable marine ecosystem.” Dan- iel and Robertson (1990) state that “[C]omparisons of the densities of fish and Old World mangrove sites have generally supported the contention of a nursery ground

Bulletin of Marine Science 823 © 2007 Rosenstiel School of Marine and Atmospheric Science of the University of Miami 824 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 function for mangrove habitats…” but note that “…there is often high variance in the nursery ground value of mangrove habitats in any one region.” While the controversy about the role of mangrove forests as “nursery habitat” for fish and invertebrates continues (Sheridan andH ays, 2003; Manson et al., 2005), it is generally accepted that mangroves and adjacent nearshore habitats provide important coastal habitats for fish and invertebrates, as well as coastal wildlife species that depend on mangrove forests for both habitat and food resources, such as wading and seabirds, crocodilians, monkeys, and even large mammals such as tigers (Saenger, 2002). A good example of this is the Roseate spoonbill in Florida (Ajaia ajaja Lin- naeus, 1758) about which Lorenz et al. (2002) state “…small fishes of the orders Cy- prinodontidae and Poecilidae appear to be the primary dietary items.” As noted in Lewis (1992), declines in both the commercial and recreational harvests of fishes have been attributed to overharvesting, water pollution, and habitat loss (Stroud, 1983; Royce, 1987). In response to these declines, various fishery management tools have been proposed and implemented. They include legally limiting harvests (by size and number), closing seasons, limiting entry, stocking from hatch- eries, and enhancing fisheries, the latter most often defined as the creation of artificial reefs (Royce, 1987). Fish and invertebrate habitat protection is often discussed in relation to declining fishery harvests, but restoration of degraded or lost coastal wetland habitats has only recently been recognized as having great potential as a cost-effective fishery maintenance and restoration strategy (Lewis, 1992; Benaka, 1999). Mangrove communities worldwide often contain a suite of fish with widely varying life history strategies. A few “resident” species spend their entire lives reproduc- ing and feeding within mangrove habitats. These fish often small and numerous, numerically dominating the mangrove fish fauna (Blaber, 1997). Ephemeral visitors to mangrove communities, or “transients” typically utilize mangrove systems to forage, or to seek refuge from predation. Mangrove ecosystems are sought as nursery sites for many transient species, such as the elopids, megalopids, centropomids, and lutjanids, which are major tropical estuarine predators as adults. These species enter the mangrove system as small larvae or early juveniles, often grow rapidly, and leave the system before maturing as top predators in the adjacent estuary. The most common transient omnivorous primary consumers are represented by mugiliids, clupeids, sparids, and gerreids (Gilmore and Snedaker, 1993). The omnivorous/detritivorous transient fishes may use the mangrove system as a nursery ground, but they also enter these systems in large numbers as adults to feed on the abundant primary producers, algal/bacterial conglomerates, and detritus that is abundant in tropical mangrove habitats. Resident species that reproduce and feed primarily in mangrove forest ecosystems are typically micro/macro-consumers, herbivores or omnivores (Table 1). This group typically consists of small fishes with highly variable diets, often capable of consum- ing algae, cyanobacteria, and micro-invertebrates. In the tropical western Atlantic, resident mangrove species typically consist of cyprinodonts, fundulids, rivulins, poeciliids, eleotrids, and gobiids. To remain within mangrove ecosystems throughout the year, a resident species must have eco-physiological capabilities that allow it to survive the extreme physical conditions characteristic of mangrove habitats: periodic dewatering, low dissolved oxygen, high sulfide levels, extremes of high temperature, and salinity variation. Resident mangrove forest species can typically with- LEWIS AND GILMORE: MANGROVE RESTORATION AND FISH HABITAT 825

Table 1. Typical east Florida restored mangrove ecosystem fish community: common species of a total of 93 captured in 5,731 collections at 326 sites in the Indian River Lagoon, Florida. Species are divided into residential and trophic groups from data collected simultaneously in three seperate barrier island mangrove systems for 24 hrs every 2 wks over 2 yrs.

Biomass Wet weight (g) % No. Individuals % Omnivores grazing on detritus, algae, cyanobacteria, micro and macro invertebrates Cyprinodon. variegatus 253,700.16 57 405,944 41 Poecilia latipinna 146,412.94 32 234,495 24 Gambusia holbrooki 50,993.30 11 350,769 35 Total residents 451,106.40 49.3 991,208 95

Mangove transient species Top predators feeding on fish and macro invertebrates Centropomus undecimalis 5,1487.77 44 12,431 53.0 Elops saurus 25,442.41 22 10,279 44.0 Megalops atlanticus 22,649.31 20 408 2.0 Lutjanus griseus 16,544.94 14 193 0.8 Total transient predators 116,124.43 12.7 23,311 2.2

Omnivores grazing on detritus, algae, cyanobacteria as well as micro and macro invertebrates Mugil cephalus 283,754.38 81.0 21,184 74.1 Mugil curema 564,66.10 16.0 3,064 10.7 Brevoortia smithi 935.22 0.3 1,544 5.4 Diapterus auratus 1,596.04 0.5 1,612 5.6 Archosargus probatocephalus 4,934.83 1.4 169 0.6 Gerres cinereus 806.74 0.2 1,033 0.4 Total transient omnivores 348,493.31 38.1 28,606 2.7 Total fish 915,724.14 1,043,125

stand major variations in physical parameters: water temperature, salinity, pH and dissolved oxygen (Harrington and Harrington, 1961; Gilmore, et al., 1982). Transient species such as tarpon, Megalops atlanticus Valenciennes, 1847, and common snook, Centropomus undecimalis (Bloch, 1792), have adaptations that allow them to withstand low dissolved oxygen conditions and variable salinity regimes characteristic of mangrove ecosystems (Peterson and Gilmore, 1991). During the 1950s and 1960s the majority of wetland mangrove habitat on the central east coast of Florida between latitude 27°00´ and 29°00´N, approximately 225 coastal km, was impounded to control mosquito populations. Impounding destroyed thousands of hectares of mangrove forest habitat and caused significant perturbation to the mangrove ecosystem and its dependent biota. The first studies implemented to determine the influence of impounding- man grove ecosystems on associated fish communities were conducted in the 1950s by Harrington and Harrington (1961, 1982). As the first study to document and suggest positive restoration measures Gilmore et al. (1982) demonstrated that even though impoundments destroyed thousands of hectares of fish habitat, these habitats could be restored for fish use through simple tidal reconnection (hydrologic restoration; Brockmeyer et al., 1997). In fact, tidally connected ditches dug to produce impoundment dikes were found to offer new habitat that mimicked tidal creeks, thus allowing 826 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 water to remain in the wetland during low water periods, offering more habitat to transient species such as common snook and striped mullet, Mugil cephalus Lin- naeus, 1758. As an example, fish fauna data from pre- and post-impoundment of a restored mangrove habitat show that hydrologic restoration restored transient fish communities, previously precluded from the mangrove wetland by impoundment dikes (Table 2). Invertebrate and vegetative communities were also restored through the restoration of tidal access and natural plant succession. Tidal connections to adjacent estuary restored biological transport of biomass from the mangrove wetland to the adjacent estuary and allowed recruitment to the mangrove wetland of juvenile transient species that normally utilize this habitat as nursery habitat (centropomids, elopids, megalopids, lutjanids, and mugiliids; Gilmore 1987a,b). Extensive impoundment and subsequent restoration of mangrove forest wetlands throughout coastal east Florida allowed researchers to demonstrate the direct value of mangrove restoration by hydrologic restoration on fish communities, particularly

Table 2. Comparison of the east Florida mangrove fish fauna before and after restoration of a 26 ha mangrove wetland ecosystem, Round Island, Indian River County, Florida. These data were collected from 1955–1968 by Harrington and Harrington (1961, 1982) and from 1978–1986 by Gilmore et al. (1982) and Gilmore (1987a). Pre-impoundment collections were from 10 Septem- ber 1956 to 17 October 1956 and post-impoundment collections from September 1956 to October 1965 and from September 1984 to August 1986.

Pre-impound. Post-impound. Restored Harrington and Harrington and Gilmore Harrington, 1961 Harrington, 1982 et al., 1985 No. Individuals % No. Individuals % No. Individuals % Mangrove resident species Omnivores grazing on detritus, algae, cyanobacteria, micro and macro invertebrates Cyprinodon variegatus 1,986 40.0 235 25 258,597 41.0 Poecilia latipinna 1,215 25.0 400 43 160,317 26.0 Gambusia holbrooki 1,769 36.0 300 32 210,080 33.0 Total residents 4,970 76.0 935 100 628,994 94.0

Mangrove transient species Top predators feeding on fish and macro invertebrates Centropomus undecimalis 172 38.0 0 0.0 4,005 53.0 Elops saurus 33 7.0 0 0.0 8,849 44.0 Megalops atlanticus 254 55.0 0 0.0 316 2.0 Lutjanus griseus 0 0.0 0 0.0 34 0.8 Total transient predators 459 7.0 0 13,204 2.0

Omnivores grazing on detritus, algae, cyanobacteria as well as micro and macro invertebrates Mugil cephalus 1,124 100.0 0 0.0 17,856 74.1 Mugil curema 0 0.0 0 0.0 3,064 10.7 Brevoortia smithi 0 0.0 0 0.0 78 5.4 Diapterus auratus 0 0.0 0 0.0 102 5.6 Archosargus probatocephalus 0 0.0 0 0.0 15 0.6 Gerres cinereus 0 0.0 0 0.0 328 0.4 Total transient omnivores 1,124 17.0 21,443 3.0 Total fish 6,553 900 670,194 LEWIS AND GILMORE: MANGROVE RESTORATION AND FISH HABITAT 827

transient fish species which supported regional sport and commercial fisheries. For example, over 1500 juvenile common snook were captured during a single 3-hr cul- vert trap set in a restored, previously impounded, mangrove wetland, Jack Island State Park, on the barrier island in St. Lucie County, Florida. This is the single largest capture of this species as juveniles recorded in Florida waters and demonstrated the value of mangrove habitat for this species and its utilization of a restored tidal connection between the mangrove forest and the estuary. The trophic work of Lucz- kovich et al. (1995) included this species, verifying the major role mangrove residents play in the diet of predaceous transients in restored mangrove wetlands first demonstrated by Harrington and Harrington (1961, 1982) and Gilmore et al. (1983). One of the first definitive research efforts in the United States to target early juvenile fish habitat modification as a major limiting factor to adult fishery stock considered of commercial importance was that of Gilmore et al. (1983) documenting the life history of the common snook in the Indian River on the east coast of Florida. This species is a prized game fish that had been primarily managed by closure of its fishery to commercial harvest in 1957, and subsequent recreational harvest limits, neither of which has been shown to maintain or increase the adult stock. Gilmore et al. (1983) determined that while adult common snook spawn in higher salinity coastal passes, the early juvenile snook (mean 27.5 mm standard length, SL) are found in shallow freshwater tributary streams entering estuarine waters and in polyhaline to marine barrier island mangrove forests. Juvenile common snook typically leave peripheral mangrove and freshwater habitats as they reach 100–150 mm SL, moving into seagrass meadows as they mature, at a mean of 240 mm SL. However, larger juvenile and mature adults will periodically visit tidal fringe mangrove forest habitats, rivers and creeks where adequate prey and water depth allow. Important food items for juveniles included fish, shrimp, and microcrustaceans, with the eastern mosquitofish, Gambusia holbrooki Girard, 1859, being the dominant identifiable fish species in specimens < 100 mm SL (Gilmore et al., 1983; Luczkovitch et al., 1995). Along the coasts of central Florida, the role of mangroves and the mangrove-marsh mixture in the production of small forage fish species [eastern mosquitofish; rainwater killifish: Lucania parva Baird and Girard, 1855; sailfin molly: Poeicilia latipinna (Leseur, 1821); sheepshead minnow: Cyprinodon variegates Lacépède, 1803; Fundu- lus spp.; diamond killifish: Adinia xenica (Jordan and Gilbert, 1882)] essential as dietary items for larger and more commercially important fishes (as well as wading birds and seabirds) is underappreciated and rarely documented (Fore and Schmidt, 1973; Gilmore et al., 1983; Luczkovitch et al., 1995). The few studies focusing on this function of mangroves and mangrove-marsh communities include Harrington and Harrington (1961), Gilmore et al. (1982, 1983), Rey et al. (1990), Ley et al. (1994), and Whitman and Gilmore (1993). Robertson and Duke (1990) make a similar point about the valuable Australian fishery for barramundi (Lates calcarifer Alexander, 1845): “[O]nly a small number of fish species examined in this study are of direct economic importance…(H)owever, the valuable commercial gill net and recreational line fisheries for barramundi…are intimately linked to mangrove habitats…our data…show that more than 50% of the diet of subadult L. calcarifer is composed of the small fish which dominate the mangrove habitats…” In spite of this, the recent review of Manson et al. (2005), in discussing groups of mangrove dependent fish species states “(A) final grouping is the true estuarine species that complete their entire life cycle within estuaries. These species are clearly es- 828 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

tuarine-dependent but many are small and short lived…and few contribute directly to fisheries,. . . (and). . . will not be discussed further. Instead, the focus will be on the marine-estuarine category that includes a number of economically important species.” We believe ignoring the role of small mangrove-dependent fish in the families Gobiidae, Atherinidae, Cyprinodontidae, and Poecilidae, for example, as important mangrove-dependent species is neither an ecologically sound approach to mangrove forest management and restoration, nor estuarine fisheries management. Gilmore et al. (1983) concluded that, “loss of habitat and general degradation of water quality has undoubtedly had a more permanent and, therefore, far greater effect on reducing snook populations than the fishery. Removal of juvenile habitat curtails recruitment from the largest portion of the snook population and that portion which is most susceptible to natural mortality.” Similar life history implications are reported for other key commercial and recreationally important estuarine fish species, such as redfish Sciaenops( ocellatus Linnaeus, 1766), spotted seatrout [Cy- noscion nebulosus (Cuvier, 1830)], and tarpon (Lewis et al., 1985; Lewis, 1990). Similar restoration is reported by Barimo and Serafy (2003) where fish utilization of a 30 ha mangrove restoration site in Biscayne Bay, Florida, with 28 ponds connected to adjacent waters by a series of tidal creeks, supported 25 fish species 1–3 yrs after restoration. The most common of these species were the gold spotted killifish [Floridichthys carpio (Gunther, 1866)]; yellowfin mojarra Gerres[ cinerus (Walbaum, 1792)]; and rainwater killifish. Sampling of ponds of various ages revealed a general trend in increasing fish community diversity over time. Finally, fragmented or hydrologically restricted or isolated mangrove habitat in the Bahamas has shown reduced numbers of fish species (Layman et al., 2004) similar to the data of Harrington and Harrington (1961, 1982), Gilmore et al. (1982), and Gilm- ore (1987a). Community-based hydrologic restoration of these areas led to increases in fish and invertebrate utilization after restoration (Layman and Arringon, 2005). Layman et al. (2004) proposed that hydrologic restrictions influence fish and invertebrate abundance through: (1) reduced tidal exchange producing decreased habitat quality (e.g., greater salinity extremes), (2) reduced tidal exchange lowering the influx of planktonic larvae and juveniles of estuarine dependent species, and (3) non-per- meable landscape features (e.g., roads without culverts) halting upstream movement by transient marine species [e.g., bonefish, Albula vulpes (Linnaeus, 1758)].

Improving the Success of Restoration

Great potential exists to reverse the worldwide loss of mangrove forests, and thus enhance fish and invertebrate habitats, through application of basic principles of ecological restoration using ecological engineering approaches. Previously documented attempts to restore mangroves, where successful, have largely concentrated on creation of plantations consisting of a few mangrove species targeted for harvest as wood products or temporarily used to collect eroded soil and raise intertidal areas to elevations usable for terrestrial agriculture [see review by Lewis (2005)]. But many expensive efforts to restore mangroves have failed (deLeon and White, 1999; Erfte- meijer and Lewis, 2000), and most attempts to restore mangroves never even achieve significant cover by mangrove species, or meet their stated goals (Erftemeijer and Lewis, 2000; Lewis and Streever, 2000; Lewis, 2000, 2005). We suggest there should be three restoration strategies for fish habitat optimization: LEWIS AND GILMORE: MANGROVE RESTORATION AND FISH HABITAT 829

Strategy 1: Achieve plant cover similar to that in an adjacent relatively undisturbed and mature mangrove ecosystem.—Due to the length of time it may take to achieve full maturity in a restored mangrove forest (25–50 yrs minimum), the likely actual measurable criterion will be the establishment of a trajectory (based upon regular monitoring of multiple parameters, such as plant cover by species and density and height of mangroves) that, if continued, would establish such mature forest cover [see Lewis et al. (2005) for a typical trajectory curve and successful forest cover establishment at 36 mo post restoration]. Such vegetation development trends are rarely documented or reported, however, in part due to the high failure rate of mangrove restoration. Strategy 2: Establish a network of channels that mimics the shape and form of a natural tidal creek system.—Tidal creeks are important as pathways for the movement of mangrove detritus and for their function in maintaining water quality (particularly with regard to the dissolved oxygen levels required to support fish and invertebrates), as well as for their role as entry points on flood tides and exit points on ebb tides for mobile nekton. Although this seems obvious from a fisheries perspective, mangrove restoration projects are frequently completed without concern for reconnection or construction of tidal creeks. Even when a network of channels is incorporated into the design, it too often consists of just rough-cut ditches dug without consideration of the resulting tidal prism, the desired exchange of tidal waters with adjacent estuarine or oceanic waters, or whether velocities will be high enough to keep the new channels open over time (Turner and Streever, 2002). Instead, channels should be designed to serve natural functions and be as self-maintaining as possible. Strategy 3: Establish a heterogeneous landscape similar to that exhibited by the local mangrove ecosystem.—Mangrove ecosystems include the mangrove forest itself, its associated animal community, and adjacent and hydrologically linked wetlands. Important contiguous plant communities include terrestrial plant communities in the watershed, and shallow marine communities including tidal ponds, tidal flats, seagrass meadows, and coral reefs. The landward portion of the mangrove ecosystem includes salt flats or salinas, as well as freshwater streams and wetlands (both marshes and forested wetlands) that ultimately drain to the sea via both surficial and underground waterways through mangrove forests. The seaward edge of the mangrove ecosystem is the outer edge of the influence of water flows coming from drainage through the mangrove forests and may extend several kilometers offshore. Mangroves have also been ecologically linked with fish movements from adjacent coral reefs and seagrass meadows (Ley and McIvor, 2002; Mumby et al., 2004). Lewis (2005) reviews the most likely points of failure in successful mangrove forest restoration and offers a simple sequential checklist of design tasks, that, if followed rigorously, would prevent most of these failures. Callaway (2001) recommends seven similar steps be taken in designing for hydrologic and geomorphologic development of tidal marshes in California. Both Vivian-Smith (2001) and Sullivan (2001) suggest addressing these factors through the use of a reference tidal marsh for restoration planning and design, which applies equally well here for mangrove forest restoration. Five sequential design tasks have been suggested by Stevenson et al. (1999) as essential to implementing the strategies listed above, and successfully restoring mangrove forests. Following these design tasks in sequence will prevent most of the common errors in mangrove forest restoration design. 830 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

1. Understand the autecology (individual species ecology) of the mangrove species in the vicinity of the restoration site, paying particular attention to patterns of reproduction, zonation, propagule distribution, and seedling establishment. 2. Understand the normal hydrologic patterns controlling the distribution and successful establishment and growth of the targeted mangrove species. 3. Determine what modifications and stresses of the previous mangrove environment are currently preventing natural secondary succession. 4. Design the restoration program to first reestablish the appropriate hydrology at an appropriate restoration site, and then utilize natural volunteer mangrove propagule recruitment for plant establishment. 5. Only plant propagules or seedlings after determining through steps 1–4 that natural recruitment will not provide the quantity of successfully established seedlings, rate of stabilization, or rate of growth of saplings set as goals for the restoration project. The more common alternative approach to mangrove restoration is to first build a nursery to grow mangrove seedlings, then produce large numbers of potted mangroves for planting, and finally identify a location to install the nursery-grown plant materials. This is referred to by Lewis (2005) as the “gardening” approach. Often unvegetated mudflats that have never supported mangroves are chosen for such garden sites, and frequently most or all of the plants installed die in a short time, or remain stunted due to excessive flooding and stress due to extended soil saturation (Erftemeijer and Lewis, 2000). In reality, actual planting of mangroves is rarely needed to establish plant cover unless a site has been identified as “propagule limited” (Lewis, 2005). Previous research has documented the general principle that mangrove forests worldwide occur on raised and sloped platforms situated above mean sea level and inundated less than thirty percent of the time by tidal waters (Lewis, 2005). More frequent flooding can cause stress and death of these tree species (Turner and Lewis, 1997). Preventing such damage requires a detailed understanding of local mangrove hydrology. Again the inclusion of correctly designed tidal creeks in a mangrove restoration design provides access to this platform on high tides, and refuge within the creeks on low tides if properly designed. Because degraded mangrove forests may recover without active restoration efforts, restoration projects should begin with a search for hydrologic sources of stress (such as blocked tidal inundation), or roadways within the forest without adequate cross connections both upstream and downstream of the road, that might be preventing secondary succession, then focus on removing that stress (Hamilton and Snedak- er, 1984; Cintrón-Molero, 1992; Layman et al., 2004; Layman and Arringon, 2005). Once normal hydrology has been reestablished, the next step is to observe whether natural seedling recruitment occurs. Planting should be considered only if this does not happen. It is important to understand that mangrove forests occur in a wide variety of hydrologic and climatic conditions that result in a broad array of mangrove community types. In Florida, Lewis et al. (1985) identified at least four variations on the original classic mangrove zonation pattern described by Davis (1940), all of which include a tidal marsh component dominated by such species as smooth cordgrass (Spartina alternifloraLoisel.) or saltwort (Batis maritima Linnaeus). Lewis (1982a,b) describes the role that smooth cordgrass plays as a “nurse species” that initially colonizes bare soil and facilitates primary or secondary succession to a climax community domi- LEWIS AND GILMORE: MANGROVE RESTORATION AND FISH HABITAT 831

nated by mangroves, but with some remnants of the original tidal marsh vegetation typically remaining. Crewz and Lewis (1991) further discuss the tidal marsh component typical of Florida mangrove forests. The typical “mangrove restoration” project aimed at establishing a monospecific stand of planted Rhizophora spp. fails to restore the freshwater-marine salinity gradient and plant biodiversity characteristic of intact mangrove ecosystems. Establish- ing this salinity gradient and a related variety of habitat types must be the goal of any mangrove restoration project intended to provide fish habitat and enhance biodiversity. Besides providing food and habitat for a broad spectrum of fish, invertebrates, and associated wading birds, seabirds, and other wildlife, incorporating this estuarine gradient maximizes both above- and belowground productivity of mangroves and their ability to deal with rising sea levels (Snedaker, 1993).

Restoration Opportunities

There are both large and small opportunities for mangrove restoration. Both have a role in mangrove forest management, but it is important to understand that very small restoration projects, such as the typical shoreline plantings of a few hundred or thousand mangroves are generally neither as cost-effective on a per hectare basis nor as ecologically valuable as larger hydrologic restoration projects, such as restoring normal seasonal freshwater flows to mangroves in Florida Bay (Ley et al., 1994; Ley and McIvor, 2002) or Biscayne Bay (Braun et al., 2004). Large scale hydrologic restoration, such as that described by Brockmeyer et al. (1997), has produced real ecological benefits for US$250 ha–1. Small local citizen-based restorations still have educational value and can play a supplementary role, especially in regard to rees- tablishing species and community diversity along the urban interface (Layman and Arringon, 2005). But where money is limited and/or fish habitat optimization is the goal, we should listen to the words of Spurgeon (1998): “[I]f coastal habitat rehabilitation/creation is to be widely implemented, greater attempts should be made to: find ways of reducing the overall costs of such initiatives; devise means of increasing the rate at which environmental benefits accrue; and to identify mechanisms for appropriating the environmental benefits.” There are many opportunities to restore mangroves worldwide, including at least several hundred thousand hectares of abandoned shrimp aquaculture ponds, primarily in Southeast Asia, but also in Ecuador and Brazil (Stevenson et al., 1999).

Case Study

Aerial photographs of the same mangrove restoration site over a period of 4 yrs represent a portion of a 500 ha hydrologic restoration project at West Lake (WL) in Hollywood, Florida (Fig. 1; described in Lewis (1990)). Success resulted from using a reference site, and targeting final constructed grades to be the same as the adjacent undisturbed forest. This resulted in a final sloped grade from +27 cm to +42 cm mean sea level (MSL). Extensive constructed tidal creeks and shallow mudflats were also added to the original plans which had been designed and originally permitted without them. No planting of mangroves took place or was necessary. All three of the Florida species of mangroves (red mangrove Rhizophora mangle Linnaeus, black mangrove Avicennia germinans (Linnaeus) Linnaeus, and white mangrove La- 832 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Figure 1. Oblique aerial photographs of a portion of the hydrologic mangrove restoration project at West Lake Park, Hollywood, Florida over three time periods: (A) Preconstruction, July 1987, (B) completion of construction, August 1989, and (C) 2 yrs post-construction, May 1991. No planting of mangroves occurred. All vegetation derived from volunteer mangrove propagules. LEWIS AND GILMORE: MANGROVE RESTORATION AND FISH HABITAT 833

guncularia racemosa (Linnaeus) Gaertn.f.) volunteered on their own. High marshes dominated by B. maritima were also hydrologically restored. This type of major ex- cavation to restore mangrove habitat is similar to the successful efforts reported by Barimo and Serafy (2003) for Biscayne Bay. Another form of hydrologic restoration is to reconnect impounded mangroves to normal tidal influence (Brockmeyer et al., 1997; Turner and Lewis, 1997; Layman and Arringon, 2005). Comparative sampling of fish populations within both restored and natural reference areas at WL, and adjacent natural and restored sites at John U. Lloyd Park (JUL; Roberts, 1994), resulted in the collection of 9937 fishes representing 34 species in the JUL reference sites, 6835 fishes representing 28 species at the JUL mangrove restoration sites, 9141 fishes representing 36 species at the WL reference sites, and 11, 374 fishes representing 31 species from theW L restoration site. The predominant species were bay anchovy Anchoa mitchilli Valenciennes, 1848; juvenile mullet Mugil spp.; yellowfin mojarra Gerres cinereus (Walbaum, 1792); silver jenny Eucinostomus gula (Quoy and Gaimand, 1824); striped mojarra Eugerres plumie (Cuvier, 1830); eastern gambusia; sailfin molly; marsh killifishFundulus confluentusG oode and Bean; 1879, and goldspotted killifish. Statistical tests of the differences in fish assemblages, and total mean abundance in the natural and 3–5 yr old restored mangroves sites were not significant. Lewis (1992) noted in reviewing similar reports on fish use of tidal marsh restoration sites in Florida, North Carolina, and California that “…rapid (3–5 yrs) establishment of comparable fish communities in created and restored coastal wetlands, when compared with natural wetlands, is a generally documented observation…”

Conclusions

Ellison (2000) asks the question “mangrove restoration: do we know enough?” His answer is that “[R]estoration of mangal does not appear to be especially difficult…” and comments that in contrast to the complexity of restoring inland wetlands, “…it is more straightforward to restore tidal fluctuations and flushing to impounded coastal systems where mangroves could subsequently flourish…” Thus, ecological restoration of mangrove forests is feasible, has been done on a large scale in various parts of the world, and can be done cost-effectively. Fish utilization of these restored ecosystems has, however, not been a routine measure of successful restoration, and we recommend more use of this criterion in the future. Lewis (2005) recommends than an ecological engineering approach be applied to mangrove ecosystem restoration projects. The simple application of the five steps outlined by Stevenson et al. (1999) would at least insure an analytical thought process and minimize resources wasted on futile mangrove “gardening” efforts. To optimize fish use of restored mangrove ecosystems, we recommend planning restoration projects around the three key strategies outlined earlier: establishing plant cover, a network of tidal channels, and a landscape mosaic, all of which mimic natural undisturbed systems and reflect the zonation and diversity of the local mangrove ecosystem, including freshwater inputs to the landward edge of a normal estuarine gradient. 834 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Literature Cited

Barimo, J. F. and J. E. Serafy. 2003. Fishes of a restored mangrove habitat on Key Biscayne, Florida. Fla. Sci. 66: 12–22. Benaka, L. R. 1999. Fish habitat: Essential Fish Habitat and Rehabilitation. American Fisheries Society, Bethesda, Maryland. 459 p. Blaber, S. J. M. 1997. Fish and fisheries of tropical estuaries. Fish and Fisheries Series 22. Chap- man & Hall, New York. 367 p. Braun, G., R. Lewis, and J. Watson. 2004. Freshwater flow and ecological relationships in Bis- cayne Bay. Tasks 1-6. South Florida Water Management District Contract C-15967-Wo0- 06. West Palm Beach, Florida. Available from: http://www.sfwmd.gov/org/wsd/mfl/biscay- nebay/project_doc.htm. Accessed 15 January 2007. Brockmeyer, R. E. Jr., J. R. Rey, R. W. Virnstein, R. G. Gilmore, and L. Ernest. 1997. Rehabilita- tion of impounded estuarine wetlands by hydrologic reconnection to the Indian River La- goon, Florida (USA). Wetl. Ecol. Manage. 4: 93–109. Callaway, J. C. 2001. Chapter three. Hydrology and substrate. Pages 89–117 in J. B. Zedler, ed. Handbook for restoring tidal wetlands. CRC Press, Boca Raton. 439 p. Cintrón-Molero, G. 1992. Restoring mangrove systems. Pages 223–277 in G. W. Thayer, ed. Restoring the nation’s marine environment, Maryland Seagrant Program, College Park, Maryland. 716 p. Crewz, D. W. and R. R. Lewis. 1991. An evaluation of historical attempts to establish emergent vegetation in marine wetlands in Florida. Florida Sea Grant Technical Publication No. 60. Florida Sea Grant College, Gainesville, Florida. 76 p + appends. Daniel, P. A. and A. I. Robertson. 1990. Epibenthos of mangrove waterways and open embayments: community structure and the relationship between exported mangrove detritus and epifaunal standing stocks. Estuar. Coast. Shelf Sci. 31: 599–619. Davis, J. H. 1940. The ecology and geologic role of mangroves in Florida. Carnegie Inst. Wash. Pap. Tortugas Lab. No. 32. Publ. 517: 305–412. De Leon, T. O. D. and A. T. White. 1999. Mangrove rehabilitation in the Philippines. Pages 37–42 in W. Streever, ed. An international perspective on wetland rehabilitation. Kluwer Academic Publishers, Dordrecht. 338 p. Ellison, A. M. 2000. Mangrove restoration: do we know enough? Rest. Ecol. 8: 219–229. Erftemeijer, P. L. A. and R. R. Lewis. 2000. Planting mangroves on intertidal mudflats: habitat restoration or habitat conversion? Pages 156–165 in Proc. ECOTONE VIII Seminar En- hancing Coastal Ecosystems Restoration for the 21st Century, Ranong, Thailand, 23–28 May 1999, Royal Forest Department of Thailand, Bangkok, Thailand. 201 p. Food and Agricultural Organization (FAO). 2003. New global mangrove estimate. Avail- able from: http://www.fao.org/documents/show_cdr.asp?url_file=/docrep/007/j1533e/ j1533e52.htm. Accessed 1 October 2006. Fore, P. L. and T. W. Schmidt. 1973. Biology of juvenile and adult snook, Centropomus undecimalis, in the Ten Thousand Islands, Florida. Chapter 16 in M. R. Carter, ed. Ecosystems Analyses of the Big Cypress Swamp and Estuaries. U.S. Environmental Protection Agency, Surveillance and Analysis Division, Athens, Georgia. 356 p. Gilmore, R. G. 1987a. Fish, macrocrustacean and avian population dynamics and cohabitation in tidally influenced impounded subtropical wetlands. Pages 373–394 in W. R. Whitman and W. H. Meredith, eds. Proc. Symp. on Waterfowl and Wetlands Management in the Coastal Zone of the Atlantic Flyway. Delaware Depart. Nat. Res. and Environ. Control, Dover, Delaware. 567 p. ______. 1987b. Subtropical-tropical seagrass communities of the Southeastern United States: Fishes and fish communities. Pages 117–137 in M. J. Durako, R. C. Phillips, and R. R. Lewis, eds. Proc. Symp. on Subtropical-tropical seagrasses of the Southeastern United States. Marine Research Publication 42, Fla. Dept. Nat. Res. 209 p. LEWIS AND GILMORE: MANGROVE RESTORATION AND FISH HABITAT 835

______and S. C. Snedaker. 1993. Chapter 5: Mangrove forests. Pages 165–198 in W. H. Martin, S. G. Boyce, and A. C. Echternacht, eds. Biodiversity of the southeastern United States: lowland terrestrial communities. John Wiley & Sons, Inc., Publishers, New York. 502 p. ______, D. W. Cooke, and C. J. Donohoe. 1982. A comparison of the fish populations and habitats in open and closed salt marsh impoundments in east-central Florida. North- east Gulf Sci. 5: 25–27. ______, C. J. Donohoe, and D. W. Cooke. 1983. Observations on the distribution and biology of east-central Florida populations of the common snook, Centropomus undecimalis (Bloch). Fla. Sci. 46: 313–336. Hamilton, L. S. and S. C. Snedaker. 1984. Handbook for mangrove area management. East-west Center, Honolulu, Hawaii. 123 p. Harrington, R. W., Jr. and E. S. Harrington. 1961. Food selection among fishes invading a high subtropical salt marsh: from onset of flooding through the progress of a mosquito brood. Ecology 41: 646–666. ______and ______. 1982. Effects on fishes and their forage organisms of impounding a Florida salt marsh to prevent breeding by salt marsh mosquitoes. Bull. Mar. Sci. 32: 523–531. Heald, E. J. 1971. The production of organic detritus in a south Florida estuary. Seagrant Tech- nical Bulletin No. 6., University of Miami Sea Grant Program, Miami, Florida. 110 p. Layman, C. A. and D. A. Arringon. 2005. Community-based collaboration restores tidal flow to an island estuary (Bahamas). Ecol. Rest. 23: 58–59. ______, ______, R. B. Langerhans, and D. R. Silliman. 2004. Degree of fragmentation affects assemblage structure in Andros Island (Bahamas) estuaries. Caribb. J. Sci. 40: 232–244. Lewis, R. R. 1982a. Mangrove forests. Pages 153–172 in R. R. Lewis, ed. Creation and restoration of coastal plant communities. CRC Press, Boca Raton. 219 p. ______. 1982b. Low marshes, peninsular Florida. Pages 147–152 in R. R. Lewis, ed. Cre- ation and restoration of coastal plant communities. CRC Press, Boca Raton. 219 p. ______. 1990. Creation and restoration of coastal plain wetlands in Florida. Pages 73–101 in J. A. Kusler and M. E. Kentula, eds. Wetland creation and restoration: the status of the science. Island Press, Washington, D. C. 591 p. ______. 1992. Coastal habitat restoration as a fishery management tool. Pages 169–173 in R. H. Stroud, ed. Stemming the tide of coastal fish habitat loss, Proc. Symp. on Conser- vation of Coastal Fish Habitat, Baltimore, Md., March 7–9, 1991. National Coalition for Marine Conservation, Inc., Savannah, Georgia, 321 p. ______. 2000. Ecologically based goal setting in mangrove forest and tidal marsh restoration in Florida. Ecol. Eng. 15: 191–198. ______. 2005. Ecological engineering for successful management and restoration of mangrove forests. Ecol. Eng. 24: 403–418. ______and W. Streever. 2000. Restoration of mangrove habitat. Tech Note ERDC TN- WRP-VN-RS-3. U.S. Army, Corps of Engineers, Waterways Experiment Station, Vicksburg, Mississippi. 7 p. ______, A. B. Hodgson, and G. S. Mauseth. 2005. Project facilitates the natural reseeding of mangrove forests (Florida). Ecol. Rest. 23: 276–277. ______, R. G. Gilmore, Jr., D. W. Crewz, and W. E. Odum. 1985. Mangrove habitat and fishery resources of Florida. Pages 281–336 in W. Seaman, ed. Florida aquatic habitat and fishery resources, Florida Chapter, American Fisheries Society, Eustis, Florida. 543 p. Ley, J. A. and C. C. McIvor. 2002. Linkages between estuarine and reef fish assemblages: enhancement by the presence of well-developed mangrove shorelines. Pages 539–562 in J. W. Porter and K. G. Porter, eds. The Everglades, Florida Bay, and coral reefs of the Florida Keys – an ecosystem sourcebook. CRC Press, Boca Raton. 1000 p. 836 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

______, C. L. Montague, and C. C. McIvor. 1994. Food habits of mangrove fishes: a comparison along estuarine gradients in northeastern Florida Bay. Bull. Mar. Sci. 54: 881–899. Lorenz, J. J., J. C. Ogden, R. D. Bjork, and G. V. N. Powell. 2002. Nesting patterns of Roseate spoonbills in Florida Bay 1935–1999: implications of landscape scale anthropogenic impacts. Pages 563–606 in J. W Porter and K. G. Porter, eds. The Everglades, Florida Bay, and coral reefs of the Florida Keys: an ecosystem source book. CRC Press, Boca Raton. 1000 p. Luczkovich, J. J., S. F. Norton, and R. G. Gilmore. 1995. The influence of oral anatomy on prey selection during the ontogeny of two percoid fishes,Lagodon rhomboides and Centropomus undecimalis. Environ. Biol. Fish. 44: 79–95. Manson, F. J., N. R. Loneragan, G. A. Skilleter, and S. R. Phinn. 2005. An evaluation of the evidence for linkages between mangroves and fisheries: a synthesis of the literature and identification of research directions. Oceanogr. Mar. Biol., Annu. Rev. 43: 485–515. Mumby, P. J., A. J. Edwards, J. E. Arias-Gonzalez, K. C. Lindeman, P. G. Blackwell, A. Gall, M. I. Gorczynska, A. R. Harborne, C. L. Pescod, H. Renken, C. C. C. Wabnitz, and G. Llewellyn. 2004. Mangroves enhance the biomass of coral reef fish communities in the Caribbean. Nature 27: 533–536. Odum, W. E. 1971. Pathways of energy flow in a south Florida estuary. Seagrant Technical Bul- letin No. 7., University of Miami Sea Grant Program (Living Resources), Miami, Florida. 162 p. ______and E. J. Heald. 1975. The detritus based food web of an estuarine mangrove community. Page 265–286 in I. E. Cronin, ed. Estuarine research. Academic Press, Inc., New York. 468 p. Peterson, M. S. and R. Grant Gilmore, Jr. 1991. Eco-physiology of juvenile snook Centropomus undecimalis (Bloch): Life-history implications. Bull. Mar. Sci. 48: 46–57. Rey, J. R., J. Shaffer, D. Tremain, R. A. Crossman, and T. Kain. 1990. Effects of re-establishing tidal connections in two impounded subtropical marshes on fishes and physical conditions. Wetlands 10: 27–45. Roberts, K. 1994. The distribution of fish populations in the natural and mitigated mangrove forests in Southwest Florida. M.Sc. thesis, Nova Southeastern University, Ft. Lauderdale, Florida. 174 p. Robertson, A. I and N. C. Duke. 1990. Recruitment, growth and residence time of fishes in a tropical Australian mangrove system. Estuar. Coast. Shelf Sci. 31: 723–743. Royce, W. F. 1987. Fishery development. Academic Press, New York. 248 p. Saenger, P. 2002. Mangrove ecology, silvaculture and conservation. Kluwer Academic Publish- ers, Dordrecht. 360 p. Sheridan, P. and C. Hays. 2003. Are mangroves nursery habitat for transient fishes and decapods? Wetlands 23: 449–458. Snedaker, S. C. 1993. Impact on mangroves. Pages 282–305 in G. A. Maul, ed. Climatic change in the intra-American seas: implications of future climate change on the ecosystems and socio-economic structure of the marine and coastal regimes of the Caribbean Sea, Gulf of Mexico, Bahamas and N.E. Coast of S. America. United Nations Environment Program. Edward Arnold, London. 389 p. Stevenson, N. J., R. R. Lewis, and P. R. Burbridge. 1999. Disused shrimp ponds and mangrove rehabilitation. Pages 277–297 in W. J. Streever, ed. An international perspective on wetland rehabilitation. Kluwer Academic Publishers, Dordrecht. 338 p. Stroud, R. 1983. Marine recreational fisheries at the crossroads. Marine recreational fisheries 8: Proc. Eighth Annual Marine Recreational Fisheries Symp. Sport Fishing Institute, Wash- ington, D.C. 236 p. Spurgeon, J. 1998. The socio-economic costs and benefits of coastal habitat rehabilitation and creation. Mar. Pollut. Bull. 37: 373–382. Sullivan, G. 2001. Chapter four. Establishing vegetation in restored and created coastal wetlands. Pages 119–155 in J. B. Zedler, ed. Handbook for restoring tidal wetlands. CRC Press, Boca Raton. 439 p. LEWIS AND GILMORE: MANGROVE RESTORATION AND FISH HABITAT 837

Turner, R. E. and R.R. Lewis. 1997. Hydrologic restoration of coastal wetlands. Wetl. Ecol. Manage. 4: 65–72. ______and B. Streever. 2002. Approaches to coastal wetland restoration: northern Gulf of Mexico. SPB Academic Press bv. The Hague. 147 p. Vivian-Smith, G. 2001. Box 2.1. Reference data for use in restoring Tijuana Estuary. Pages 59–63 in J. B. Zedler, ed. Handbook for restoring tidal wetlands. CRC Press, Boca Raton. 439 p. Wolanski, E. and K. Boto. 1990. Mangrove swamp oceanography and links with coastal waters. Estuar. Coast. Shelf Sci. 31: 503–504. Whitman, R. L., Jr. and R. G. Gilmore, Jr. 1993. Comparative evaluation of fisheries community structure and habitat relationships in natural and created saltmarsh ecosystems. Southwest Florida Water Management District, Brooksville, FL. 46 p.

Addresses: (R.R.L.) Lewis Environmental Services, Inc., P.O. Box 5430, Salt Springs, Florida 32134. (R.G.G.) Estuarine, Coastal and Ocean Science, Inc., 5920 First Street SW, Vero Beach, Florida 32968. Corresponding Author: (R.R.L.) E-mail: . BULLETIN OF MARINE SCIENCE, 80(3): 839–861, 2007

Altered mangrove wetlands as habitat for estuarine nekton: Are dredged channels and tidal creeks equivalent?

Justin M. Krebs, Adam B. Brame, and Carole C. McIvor

ABSTRACT Hasty decisions are often made regarding the restoration of “altered” habitats, when in fact the ecological value of these habitats may be comparable to natural ones. To assess the “value” of altered mangrove-lined habitats for nekton, we sampled for 1 yr within three Tampa Bay wetlands. Species composition, abundance, and spatial distribution of nekton assemblages in permanent subtidal portions of natural tidal creeks and wetlands altered by construction of mosquito-control ditches and stormwater-drainage ditches were quantified through seasonal seine sampling. Results of repeated-measures analysis of variance and ordination of nekton community data suggested differences in species composition and abundance between natural and altered habitat, though not consistently among the three wetlands. In many cases, mosquito ditches were more similar in assemblage structure to tidal creeks than to stormwater ditches. In general, mosquito ditches and stormwater ditches were the most dissimilar in terms of nekton community structure. These dissimilarities were likely due to differences in design between the two types of ditches. Mosquito ditches tend to fill in over time and are thus more ephemeral features in the landscape. In contrast, stormwater ditches are a more permanent altered habitat that remain open due to periodic flushing from heavy runoff. Results indicate that environmental conditions (e.g., salinity, current velocity, vegetative structure) may provide a more useful indication of potential habitat “value” for nekton than whether the habitat has been altered. The type of ditching is therefore more important than ditching per se when judging the habitat quality of these altered channels for fishes, shrimps and crabs. Planning should entail careful consideration of environmental conditions rather than simply restoring for restoration’s sake.

Mangrove systems provide habitat for numerous species of estuarine fishes and invertebrates, many of which are ecologically or economically valuable (Thayer et al., 1978; Weinstein, 1979; Odum et al., 1982; Thayer et al., 1987;H ettler, 1989; Sheridan, 1992; Kneib, 1997; Ley et al., 1999). Due to human population growth in coastal regions, mangroves have experienced a substantial reduction in coverage since 1950 with worldwide losses approaching one third (Alongi, 2002). Anthropogenic alteration of mangroves is evident in the landscape along the Gulf Coast of Florida in the form of dredged channels, spoil mounds, and water-control structures. The ecological consequences of such habitat alteration, however, are not always as apparent. Un- derstanding the fundamental differences between natural vs altered wetlands, and how and why faunal assemblages vary between habitats, is necessary to evaluate the effects of alteration on the ecological functioning of the system. In an effort to repair the effects of wetland disturbance, wetland restoration and creation have become widespread management tools over the last few decades (Lewis, 1992; Weinstein et al., 2000; Alongi, 2002; Warren et al., 2002; Lewis, 2005). However, the creation of a functional wetland system often proves more of a challenge than simply creating wetland structure (i.e., tidal creeks, ponds, salt barrens). A thorough understanding of faunal ecology, including that of the nekton communities (i.e., fishes, shrimps,

Bulletin of Marine Science 839 © 2007 Rosenstiel School of Marine and Atmospheric Science of the University of Miami 840 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 crabs) that utilize wetland habitats, is necessary for the successful creation of a functional wetland. Many of the wetlands in the Tampa Bay estuary on the Gulf coast of Florida have been subjected to dredge-and-fill activities since the mid-1900s (Johnston, 1981; Lewis and Estevez, 1988; Clark, 1992), while the overall coverage of the bay’s wetlands has been reduced by at least 44% over the last century (Lewis et al., 1985). Current wetland-restoration plans in the bay seek to restore areas that have been modified for mosquito control, stormwater drainage, and agriculture. However, hydrological changes induced by human activities such as dredge-and-fill and even restoration efforts may affect biotic components such as fish communities (Freeman et al., 2001; Bunn and Arthington, 2002; Roy et al., 2005). Whereas little is known of the composition and structure of nekton communities in shallow-wetland habitats of Tampa Bay, even less is known about comparative nekton use of natural and altered wetlands. Thus, fish-community data provide a useful baseline to assist resource managers with pre-restoration planning and post-restoration assessment. To determine fish-community response to habitat alteration and to provide baseline community conditions from which to plan and assess wetland-restoration projects, we initiated a three-year sampling program in several Tampa Bay wetlands. Objectives were to: (1) compare habitat characteristics of naturally formed tidal creeks and man-made ditches; and (2) compare spatial use of natural and man-made mangrove habitat by nekton relative to habitat characteristics. Here we report results from the first year of that ongoing study.

Methods

Study Site Tampa Bay is a large (1031 km2) urbanized estuary located on the central west coast of Flor- ida (27°45´N, 82°33´W) near the northern biogeographic limit of well-developed mangrove forests (Odum et al., 1982). Its watershed drains approximately 5700 km2 from six counties, including Pinellas, the most densely populated county in the state (1173 km²) according to recent census data. Approximately 2 million people live within the bay’s watershed, many of these in the coastal regions of the estuary. Coastal wetlands can be defined as those vegetated lands and associated intertidal features (e.g., creeks) that are alternately flooded and drained by lunar tides. By this definition, a wetland extends from the upland interface with terrestrial systems to the low tide interface with permanent subtidal waters (e.g., creeks, channels, tidal rivers, embayments, lagoons). For the purpose of this study, we consider small, primarily subtidal linear features nested within the wetland landscape to be functional parts of the wetland and we refer to these features including creeks, mosquito-control ditches, stormwater-drainage ditches, as “habitat types” (sensu Minello et al., 2003). Tidal creeks are naturally formed features (i.e., existing prior to dredge-and-fill disturbance) characterized by a meandering, sinuous channel with deeper erosional and shallower depositional areas (McIvor and Rozas, 1996), and a current regime that fluctuates depending on freshwater inflow from upstream as well as tidal ebb and flow. In Tampa Bay, tidal creek shorelines are typically lined by red mangroves (Rhizophora mangle Linnaeus, 1753), occasionally interspersed with white mangroves [Laguncularia racemosa (Linnaeus) Gaert- ner, 1807]. Substrates are muddier in depositional areas and sandier in erosional areas with patchy expanses of live and dead oyster habitat [Crassostrea virginica (Gmelin, 1791)]. Mos- quito-control ditches are similar to tidal creeks in terms of the homogenous red mangroves that often line the banks of both habitat types, but often differ in terms of bank inundation. KREBS ET AL.: NEKTON IN TIDAL CREEKS AND DITCHES 841

The steep banks of the mosquito-control ditches constrain the timing and extent of tidal inundation relative to that observed along the more gently sloped banks of the tidal creeks. In contrast to tidal creeks, mosquito ditches are typically linear features, lacking the sinuosity and variable bottom topography of creeks. Most mosquito ditches are generally narrower and in most cases receive less tidal circulation than creeks. A second type of man-made feature, stormwater-drainage ditches differ from mosquito-control ditches in that they typically have higher current velocities, are wider than mosquito ditches, and have lower mean salinities and firmer, sandy, shell-hash substrates, all a result of their function in conveying large volumes of stormwater runoff from freshwater retention ponds to the estuary.

Experimental Design Three wetland regions in Tampa Bay were chosen for study (Fig. 1) within county or state preserves and were sampled seasonally from December 2003 through November 2004. Sam- ple sites were selected by randomly generating points along creeks and ditches within each wetland. Fixed sites were designated as close as possible to the random points while avoiding areas of heavy vegetative structure that would prevent efficient and standardized gear deployment and retrieval. These particular wetland preserves were selected for study because of proposed restoration plans in these areas, and thus the need for pre-restoration data. Tidal creeks were sampled at each of the three regions: Mobbly Bayou (upper bay), Weedon Island (mid-bay), and Terra Ceia (lower bay). Nine fixed sites within Mobbly Bayou in the upper bay, six sites within Grassy Creek in the mid-bay, and six sites within Frog Creek in the lower bay were chosen randomly along the tidal extent of each creek. For comparison with naturally-formed tidal creeks, nine and ten fixed sample sites were established in mosquito- control ditches at the upper- and mid-bay regions, respectively. Grid ditching typical of mosquito control throughout much of the bay was minimal in the lower bay and thus precluded the sampling of mosquito ditches in this region. In addition to creek and mosquito ditch habitat types, the upper-bay region provided a second type of man-made mangrove habitat type in the form of stormwater-drainage ditches. The availability of interconnected habitat types both naturally formed and man-made allowed an opportunity to examine nekton habitat use at greater resolution within this wetland. Nine fixed sites were sampled in stormwater ditches for comparison with creek and mosquito ditch sites in the upper bay.

Sampling Methodology Sample sites were characterized by documenting channel width, water depth, current velocity, and percent cover of shoreline vegetation on each bank. Channel width was measured once at both upstream and downstream ends of the site and averaged. Water depth was measured in the middle of the channel at both ends of the site at the start of the sample and averaged. Current velocity was recorded at the start of sampling by timing a floating object over a known distance and averaging three measurements. It should be noted that nekton samples were not taken when current velocity exceeded 0.4 m s–1, thus current data are representative of these sites when current allowed sampling. Water-quality measurements of temperature, salinity, and dissolved oxygen were also taken at each site during sampling with a YSI 556 MPS unit. Sites were sampled by isolating nekton in a 9-m section of the creek or ditch using two block nets (3-mm mesh) and then seining from one block net to the other using a center-bag haul seine (5 × 1.2 m, 10 × 1.2 m, or 15 × 1.2 m, with 3-mm mesh) stretched from bank to bank. Size of block nets (5, 6, 10, 12, or 15 m) and haul seines used depended on channel width. Three successive hauls were made at each site to allow more precise estimates of nekton abundance. For data analysis, abundances were pooled across all three hauls at a given site. Preliminary estimates of gear efficiency [based on the depletion method of Zippin (1958)] for the block net method was approximately 87% after three consecutive seine hauls (J. Krebs, unpubl. data), suggesting that nekton densities presented herein are conservative estimates of abso- 842 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Figure 1. Map of Tampa Bay depicting the location of wetland study sites. lute abundances in these habitat types. Each sample was processed in the field by identifying and enumerating fish and decapod crustacean species collected in replicate seine samples. Up to 20 individuals of each species were measured to the nearest 1-mm standard length, SL (fishes), carapace width, CW (blue crabs), or post-orbital head length, HL (pink shrimp) and released. Taxonomy follows Nelson et al. (2004).

Habitat Analysis Tidal creek, mosquito ditch, and stormwater ditch sites within each wetland region were compared in terms of the habitat conditions (i.e., temperature, salinity, dissolved oxygen, water depth, and current velocity) recorded at each site (Table 1). Because habitat parameters were not normally distributed, we used a nonparametric two-way Analysis of Variance (ANOVA) developed by Scheirer et al. (1976) as an extension of the Kruskal-Wallis ANOVA (both procedures described in Sokal and Rohlf, 1997). Habitat variables were also analyzed using principal components analysis (PCA) to examine the similarity among sites and the relation between habitat type (i.e., creek, mosquito-ditch, and stormwater-ditch) and habitat KREBS ET AL.: NEKTON IN TIDAL CREEKS AND DITCHES 843 UB (n = 9) 25 (7.8) 56 (9.6) 13 (6.6) 0.5 (0.03) 0.12 (0.02) 10.96 (0.15) Stormwater ditch 0 MB 53 (9.6) 40 (8.2) (n = 10) 0.4 (0.02) 4.59 (0.14) 0.07 (0.01) Mosquito ditch UB 7 (5.2) (n = 9) 41 (8.5) 60 (9.7) 0.4 (0.02) 5.24 (0.11) 0.06 (0.01) LB 4 (2.9) (n = 6) 32 (11.0) 73 (10.7) 0.5 (0.04) 9.19 (0.47) 0.08 (0.02) 0 MB (n = 6) Creeks 33 (9.4) 72 (10.6) 0.7 (0.04) 9.23 (0.51) 0.12 (0.02) 0 UB (n = 9) 47 (8.1) 62 (8.5) 0.4 (0.02) 7.37 (0.25) 0.09 (0.02) Habitat variable Table 1. Mean (SE) physical description of sample sites in natural and man-made wetland channels, Tampa Bay, Florida. Channel width and shoreline vegetation shoreline and width Channel Florida. Bay, Tampa channels, wetland man-made and natural in sites sample of description physical (SE) Mean 1. Table were measured once and water depth current velocity four times during the year at each of n-sample sites. UB = upper bay; MB mid-bay; LB lower bay Channel width (m) Mean water depth (m) Mean current velocity (m/s) Shoreline vegetation (% cover) Rhizophora mangle Laguncularia racemosa Schinus terebinthifolius 844 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 descriptors. Variables included in the analysis were salinity, temperature, dissolved oxygen, channel width, water depth, and percent shoreline cover of R. mangle, L. racemosa, and Bra- zilian pepper (Schinus terebinthifolius Raddi). Data were normalized prior to analysis to ensure that variables were scaled similarly. Variables were considered to be explanatory on an axis if the corresponding values were > 0.40.

Nekton Analysis Wetland-associated nekton include a variety of fishes and macroinvertebrates that can be grouped by their use of the wetland based on life-history strategy (sensu Deegan and Thomp- son, 1985). Those species that use the wetland for their entire life cycle are considered residents. Although many of the species we classify as residents may not be exclusively wetland residents (i.e., many of them are found throughout the greater estuary; Nordlie, 2003), the strong site fidelity of many of these species (Lotrich, 1975; Teo and Able, 2003) and their ability to feed, grow, and reproduce without leaving the wetland (Coorey et al., 1985; Smith, 1995; Nordlie, 2000) suggests that many of the organisms in our collections are resident to the wetland in which they were collected. In contrast, those species that use the wetland (or the estuary) during only a specific life stage are considered transients. Some transient species may be of economic importance to recreational and/or commercial fisheries (e.g., black drum, Pogonias cromis and blue crab, Callinectes sapidus) and were analyzed separately to examine the relative importance of each of the habitat types to the economic class. A third group composed of schooling taxa in the families Atherinidae, Clupeidae, and Engraulidae is a dominant component (in terms of total abundance) in estuarine systems along the Gulf and Atlantic coasts (Kilby, 1955; Weinstein, 1979; Kneib, 1997). These three classes represented the majority of nekton collected and were tested for differences in density between habitat types and bay regions. Taxa from the non-economic transients, freshwater, and exotic classes (Table 2) were not collected in sufficient numbers for meaningful analyses. Nekton Density Among Habitat Types.—To examine trends in nekton density between creeks and man-made ditches within Tampa Bay, we modeled the dependent variable of mean density for: (1) total nekton, (2) resident, (3) economic, and (4) schooling taxa in three habitat types (i.e., tidal creeks, mosquito-control and stormwater-drainage ditches). Specifically, we performed repeated measures ANOVA (Littell et al., 1998) using the Statistical Analysis Software (SAS) package (SAS Institute, 2000). Nekton density was calculated as number of individuals per 100 m2 to standardize catch data from sample sites of different areas (range: 24–120 m2). The first model compared creeks among bay regions using sample site nested within bay region as the subject effect. The second model compared creeks and mosquito ditches between the upper- and mid-bay regions with sample site nested within region and habitat type. The third model compared the creek, mosquito ditch and stormwater ditches within the upper-bay region with the subject effect of sample site nested within habitat type. Because of an unbalanced sample design (i.e., no lower-bay mosquito ditches, no mid-bay or lower-bay stormwater ditches), it was necessary to run three ANOVA models for each dependent variable (e.g., total nekton, resident, etc.). Each model was run with sample season as the repeated variable and using a first-order autoregressive covariance structure which assumes that spatial autocorrelation at a sample site decreases as time between sample collections increases. In each instance, we tested the null hypothesis of no difference in nekton density among treatments (habitat types). Season was included in all three models as a blocking factor because of the strong seasonality of the nekton community and the improved fit of the model when accounting for seasonal variation in abundance. However, because our primary research questions focused on habitat-related differences, season was not considered in the interpretation of the ANOVAs (Sokal and Rohlf, 1997). Pair-wise comparisons were run on least square means using the post-hoc Tukey test for all significant main effects and/or interactions and were considered significant if P < 0.05. Data were evaluated for normality (Shapiro-Wilk test) and equal variance (Levene’s test) prior to analyses and log-transformed when either of the assumptions were not met. KREBS ET AL.: NEKTON IN TIDAL CREEKS AND DITCHES 845

Species Diversity.—Shannon diversity index was calculated as: ) HPl =-/ iePi ,log ^ ] gh where Pi equals the proportional abundance of species i. We tested the null hypothesis of no difference in species diversity between regions and habitat types using a GLM followed by a test of the least squares means for significant main effects or interactions to identify the regions and habitat types that were statistically different. Diversity data met the assumption of normality using the Shapiro-Wilk test. Community Analysis.—Community analysis included ordination (non-metric multidimensional scaling MDS) and rank correlation (BVSTEP, BIO-ENV) to examine the relation between nekton assemblage structure (i.e., species densities averaged by site for 2004) and habitat type. Prior to analysis, Bray-Curtis sample similarity was calculated on square-root transformed densities for the 26 most abundant taxa (i.e., those representing ≥ 1% of all individuals collected). Sample sites were then plotted using MDS to visualize the relation between habitat type and nekton assemblage structure. Using MDS, samples with nekton assemblages having similar species composition and abundance were considered more similar and were grouped more closely in two-dimensional space than those samples that were less similar. Relegating a multidimensional dataset to two dimensional space imposes a “stress” which typically distorts the true multidimensional relations between samples (i.e., two samples appear to be similar based on spatial proximity when in fact, they are not). Stress values for each MDS plot were deemed acceptable at a level of ≤ 0.2 (Clarke and Gorley, 2006). Differences in community structure between habitat types were tested using the Analysis of Similarity (ANOSIM) routine, a non-parametric test analogous to ANOVA. The ANOSIM R-value typically ranges from 0 to 1 and indicates the degree of similarity between habitat types. An R-value of 0 would indicate no difference in assemblage structure between habitat types, whereas an R-value of 1 would indicate greater similarity within, than between, habitat types. To define the smallest subset of the nekton community that best explained the pattern observed in the MDS plot, we used the BVSTEP routine. To test the strength of the habitat- nekton relation and to determine how well the measured habitat conditions explained nekton structure at each sample site, we correlated habitat data used in the PCA with nekton data used in the MDS using the BIO-ENV routine. Community analyses described here, and the habitat PCA described above, were performed using the PRIMER v6 software package (Clarke and Gorley, 2006).

Results

Habitat Water temperature did not differ between habitat types (ANOVA P = 0.385) or bay regions (P = 0.054; Fig. 2). Lowest salinities were observed in the lower bay creek (P < 0.002) and in the stormwater ditches in the upper bay, whereas the highest salinities occurred in mosquito ditches in the mid-bay (P < 0.0001). Low dissolved oxygen (3.8–4.4 mg L–1) was observed in both creek and mosquito ditch sites in the mid-bay and was statistically less (P < 0.0001) than that observed at upper and lower bay creeks or upper bay mosquito ditches (5.7–7.2 mg L–1). Mean water depth varied positively with current velocity and was shallowest in the mosquito ditches (P < 0.0001) which experienced significantly less current than the deeper stormwater ditches in the upper bay and the mid-bay creek (P < 0.05). Tidal creeks at upper- and lower-bay regions did not differ from mosquito ditches, which had some of the lowest current velocities (P > 0.05). 846 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 0 0 0 1.4 (1.4) UB 0.79 (0.3) 1.51 (0.6) 0.9 (0.26) 0.19 (0.19) 0.03 (0.03) 0.15 (0.06) 0.06 (0.04) 0.08 (0.04) 0.03 (0.03) 0.03 (0.03) 4.28 (4.23) 0.22 (0.08) 0.24 (0.12) 0.39 (0.25) 1.93 (1.93) (n = 36) 57.68 (11.88) Stormwater ditches 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2.98 (1) MB (n = 39) 0.13 (0.09) 0.61 (0.41) 4.69 (2.36) 0.06 (0.06) 43.77 (13.68) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Mosquito ditches UB 4.99 (2.2) 5.74 (2.33) 0.25 (0.15) (n = 36) 47.97 (24.92) Habitat 0 0 0 0 0 0 0 0 LB (n = 24) 0.16 (0.11) 0.04 (0.04) 0.08 (0.05) 2.02 (0.86) 0.23 (0.16) 2.97 (2.97) 0.22 (0.22) 0.13 (0.13) 0.19 (0.19) 0.04 (0.04) 15.16 (3.24) 37.52 (17.41) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2.92 (1) MB Creeks (n = 24) 3.05 (1.16) 0.77 (0.47) 1.66 (0.78) 0.09 (0.07) 70.97 (33.4) 0 0 0 0 0 0 0 0 0 0 UB (n = 36) 0.15 (0.11) 0.19 (0.11) 2.32 (0.68) 3.87 (0.81) 0.58 (0.46) 0.06 (0.06) 0.77 (0.38) 0.03 (0.03) 1.69 (1.43) 83.84 (21.24) Strategy Resident Economic Resident Resident Economic Schooling Resident Economic Transient Exotic Transient Exotic Transient Schooling Transient Schooling Economic Schooling Economic Schooling in naturally formed tidal creeks and man-made mosquito-control and FL. The stormwater-drainage ditches Bay, number in Tampa of samples is indicated for 2 Species Poey, 1865 Poey, spp. Family Scientific name each habitat type and region. Families are arranged alphabetically regardless of whether they fish or crustaceans. Achiridae Achirus lineatus (Linnaeus, 1758) Table 2. Mean Table nekton density (SE) per 100 m Rafinesque, 1819 Lepomis macrochirus (Bloch and Schneider, 1801) maculatus (Bloch and Schneider, Trinectes Unidentified sole Micropterus salmoides (Lacépède, 1802) Micropterus Aplocheilidae 1880) Kryptolebias marmoratus (Poey, Centropomidae undecimalis (Bloch, 1792) Centropomus Batrachoididae Opsanus beta (Goode and Bean, 1880) Atherinidae Menidia Cichlidae Cichlid Belonidae 1860) notata (Poey, Strongylura (Steindachner, 1864) (Steindachner, aureus Oreochromis (Walbaum, 1792) timucu (Walbaum, Strongylura Clupeidae spp. Brevoortia Carangidae Oligoplites saurus (Bloch, 1793) Clupeidae Herring Centrarchidae Lepomis spp. Harengula jaguana Harengula (Valenciennes, 1831) Lepomis punctatus (Valenciennes, (Lesueur, 1818) Opisthonema oglinum (Lesueur, KREBS ET AL.: NEKTON IN TIDAL CREEKS AND DITCHES 847 0 0 0 0 UB 2.38 (0.9) 13.7 (6.65) 0.05 (0.05) 0.22 (0.22) 4.13 (1.33) 7.19 (2.57) 0.49 (0.22) 0.03 (0.03) 5.84 (2.54) 1.51 (0.53) 0.03 (0.03) 0.93 (0.35) 0.36 (0.15) 6.58 (1.71) (n = 36) 11.87 (4.39) 11.87 34.56 (7.94) 12.02 (2.86) 50.41 (28.58) 46.66 (21.04) Stormwater ditches 0 0 0 0 0 0 0 0 0 MB 1.6 (0.78) (n = 39) 4.11 (1.42) 4.11 2.07 (0.88) 2.48 (1.79) 0.12 (0.12) 4.52 (1.41) 0.84 (0.39) 0.45 (0.35) 17.93 (11.9) 27.12 (7.67) 12.63 (4.87) 58.6 (22.16) 92.35 (43.86) 197.89 (55.48) 0 0 0 0 0 0 0 Mosquito ditches UB 0.11 (0.08) 0.11 2.49 (0.73) 0.06 (0.06) 5.58 (2.69) 1.29 (0.66) 0.25 (0.15) 1.74 (0.63) 0.38 (0.27) 4.23 (2.92) 4.96 (2.15) (n = 36) 30.88 (7.25) 23.18 (5.33) 91.99 (40.1) 15.38 (7.73) 29.32 (7.13) 14.92 (12.05) Habitat 0 0 0 0 0 0 0 0 0 3.1 (2) LB 0.2 (0.11) (n = 24) 4.27 (3.31) 0.41 (0.21) 1.82 (1.27) 0.27 (0.16) 0.07 (0.07) 4.57 (2.39) 2.77 (0.87) 3.26 (1.22) 20.67 (4.18) 84.22 (27.7) 38.64 (36.25) 47.46 (19.97) 0 0 0 0 0 0 0 0 0 0.2 (0.2) MB Creeks 8.8 (4.58) (n = 24) 0.26 (0.17) 1.13 (0.39) 3.18 (0.97) 0.94 (0.68) 0.29 (0.18) 0.67 (0.48) 5.15 (1.33) 7.91 (4.16) 5.01 (2.27) 14.33 (4.75) 48.8 (16.89) 186.8 (75.99) 0 0 0 0 UB 4.6 (1.72) 0.8 (0.53) 0.13 (0.1) (n = 36) 7.72 (2.36) 3.16 (0.81) 1.51 (0.53) 0.03 (0.03) 0.04 (0.04) 5.32 (1.82) 0.88 (0.37) 0.82 (0.56) 3.29 (1.08) 48.01 (11.5) 15.29 (4.28) 14.02 (7.19) 12.41 (5.64) 13.34 (5.14) 31.74 (5.28) 54.24 (21.87) Strategy Resident Resident Freshwater Transient Resident Resident Resident Transient Resident Transient Freshwater Transient Resident Schooling Resident Transient Resident Resident Resident Resident Resident Resident Resident (Mitchill, 1814) Species Lacépède, 1803 Goode and Bean, 1879 (Broussonet, 1782) (Günter, 1866) (Günter, (Valenciennes, 1848) (Valenciennes, (Jordan and Gilbert, 1882) Family Scientific name Fundulus grandis Baird and Girard, 1853 Table 2. Continued. Table Cynoglossidae Symphurus plagiusa (Linnaeus, 1766) Fundulus similis (Baird and Girard, 1854) Floridichthys carpio Cyprinidae Notemigonus crysoleucas Cyprinodontidae Cyprinodon variegatus Fundulus spp. Dasyatidae 1824) Dasyatis sabina (Lesueur, Lucania parva (Baird and Girard, 1855) (Lesueur, 1817) Dasyatis say (Lesueur, Gerreidae 1830) plumieri (Cuvier, Eugerres Elopidae Elops saurus Linnaeus, 1766 Eucinostomus gula (Quoy and Gaimard, 1824) Engraulidae Anchoa mitchilli Goode and Bean, 1879 Eucinostomus harengulus Ephippidae Chaetodipterus faber Eucinostomus spp. Fundulidae Adinia xenica Gobiidae 1837) Ctenogobius smaragdus (Valenciennes, Fundulus confluentus Gobiosoma bosc (Lacépède, 1800) Gobiosoma spp. Microgobius gulosus (Girard, 1858) Microgobius 848 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 0 0 0 0 0 0 0 0 UB 0.17 (0.1) 0.82 (0.47) 3.76 (1.19) 1.88 (1.31) 0.06 (0.04) 0.94 (0.65) 4.88 (2.32) 2.64 (0.68) 1.75 (0.54) 0.54 (0.17) 8.52 (2.01) 0.24 (0.07) 3.26 (0.94) (n = 36) 28.66 (8.81) 327.60 (44.47) Stormwater ditches 0 0 0 0 0 0 0 0 0 0 (154.04) MB 1,025.07 0.3 (0.22) 1.4 (0.84) (n = 39) 0.31 (0.31) 1.39 (0.89) 4.66 (2.67) 4.54 (1.67) 3.53 (1.49) 0.15 (0.15) 23.55 (13.9) 10.63 (2.64) 309.11 (73.01) 309.11 188.99 (32.86) 0 0 0 0 0 0 0 0 0 Mosquito ditches UB 5.54 (4.8) 0.1 (0.07) 0.53 (0.31) 2.46 (0.89) 1.49 (0.94) 2.43 (1.19) 2.05 (1.52) 0.49 (0.33) 0.07 (0.07) (n = 36) 44.45 (10.9) 13.67 (11.86) 30.46 (10.14) 99.86 (25.71) 490.09 (81.48) Habitat 0 0 0 0 LB 3.6 (1.52) 19.0 (4.5) (n = 24) 0.15 (0.11) 0.35 (0.35) 0.32 (0.17) 0.24 (0.14) 1.86 (1.86) 0.18 (0.18) 0.07 (0.07) 1.36 (0.51) 0.04 (0.04) 0.66 (0.62) 4.76 (3.45) 4.42 (1.73) 0.04 (0.04) 34.37 (7.76) 32.64 (6.95) 26.02 (11.36) 401.84 (60.60) 0 0 0 0 0 0 0 0 MB Creeks 0.42 (0.2) (n = 24) 1.15 (0.61) 0.35 (0.15) 0.32 (0.19) 0.25 (0.17) 3.19 (1.01) 0.06 (0.06) 4.06 (1.79) 0.39 (0.28) 13.04 (5.57) 10.09 (2.26) 51.54 (22.71) 82.93 (31.19) 72.91 (27.02) 603.63 (123.02) 0 0 0 0 UB 1.39 (0.5) (n = 36) 0.06 (0.06) 2.12 (0.85) 9.93 (4.98) 0.06 (0.06) 0.05 (0.05) 1.13 (0.54) 8.96 (6.14) 6.86 (2.47) 5.39 (2.05) 8.54 (3.21) 0.15 (0.15) 0.12 (0.12) 10.63 (6.87) –0.14 (0.08) 14.29 (1.66) 16.53 (10.25) 93.89 (21.05) 492.23 (52.62) Strategy Economic Economic Economic Economic Economic Economic Economic Economic Economic Economic Transient Economic Economic Economic Resident Transient Freshwater Transient Transient Resident Economic Transient Species Gerard, 1959 (Linnaeus, 1766) (Evermann and Kendall, 1896) Rathbun, 1896 (Lesueur, 1821) (Lesueur, (Linnaeus, 1758) Linnaeus, 1758 Family Scientific name Table 2. Continued. Table Lutjanidae Lutjanus griseus (Linnaeus, 1758) (Cuvier, 1830) Cynoscion nebulosus (Cuvier, Mugilidae Mugil cephalus Leiostomus xanthurus Lacépède, 1802 americanus (Linnaeus, 1758) Menticirrhus Mugil gyrans (Jordan and Gilbert, 1884) (Linnaeus, 1766) Pogonias cromis Mugil spp. Sciaenops ocellatus Myliobatidae Rhinoptera bonasus (Mitchill, 1815) Unidentified sciaenid Penaeidae Farfantepenaeus duorarum (Burkenroad, 1939) Sparidae 1792) (Walbaum, probatocephalus Archosargus Poeciliidae Gambusia holbrooki (Linnaeus, 1766) Lagodon rhomboides Agassiz, 1855 Heterandria formosa Syngnathidae Syngnathus scovelli Poecilia latipinna Synodontidae Synodus foetens nekton density Total UB = upper bay; MB mid-bay; LB lower bay Portunidae Callinectes sapidus Sciaenidae chrysoura (Lacépède, 1803) Bairdiella Ginsberg, 1930 Ginsberg, Cynoscion arenarius KREBS ET AL.: NEKTON IN TIDAL CREEKS AND DITCHES 849

Figure 2. Water quality, depth and velocity in tidal creek (■), mosquito-control (■), and stormwater-drainage (®) ditches at each of the three bay regions. UB = upper bay; MB = mid-bay; LB = lower bay. Values within a plot that are not statistically different (P < 0.05) are identified by similar letters.

Shoreline vegetation along creeks at lower- and mid-bay regions was similar and consisted primarily of red mangrove, R. mangle, with some interspersed white mangrove, L. racemosa (Table 1). However, percent cover along creek shorelines in the upper bay was more evenly composed of R. mangle and L. racemosa, with five of the nine creek sites having at least 50% cover by both species. Many of the mosquito ditches in the upper bay (7 of 9 sites) and mid-bay (6 of 10 sites) were adjacent to higher elevation upland or dredge spoil and tended to have more L. racemosa (> 60% cover on at least one shoreline), while the remaining sites were along the main path of tidal circulation and had greater cover by R. mangle (> 80%). Shorelines along stormwater ditches consisted primarily of L. racemosa and had the greatest coverage of Brazilian pepper, S. terebinthifolius, of any habitat type (Table 1). Mosquito ditches and stormwater ditches were most dissimilar among the three habitat types (Fig. 3). These habitat types were separated on the first principal component axis (PC1) by channel width (factor loading = 0.461), salinity (−0.413) and percent composition of Brazilian pepper (0.467). Percent composition of red (0.481) and white (−0.459) mangroves in addition to water depth (0.420) separated deeper creek and mosquito ditch sites with more red mangroves from shallower sites with more white mangroves along the secondary component axis (PC2). These two principal components explained 48.6% of the variation in habitat factors between sites.

Nekton Sixty-five taxa (including 19 of economic value) were collected from 195 samples in tidal creeks, mosquito-control ditches, and stormwater-drainage ditches (Table 2). Fewer than half of the taxa (n = 29) constituted 99% of all nekton collected. The most common taxa (i.e., those occurring in > 50% of samples) included several species of residents: killifish and pupfish (Adinia xenica, Cyprinodon variegatus, Fundulus grandis, Lucania parva), two livebearers (Gambusia holbrooki and Poecilia latipin- 850 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Figure 3. Principal components analysis (PCA) of habitat conditions in tidal creeks (●), mosquito-control (●), and stormwater-drainage (°) ditches at three wetlands within Tampa Bay. Vec- tors represent the direction of increasing value for the habitat variable indicated at the end of the vector. Only factors with eigenvalues exceeding 0.40 were displayed as vectors. The first two PC axes explain 48.6% of the variability between habitat types.

na), and a goby (Microgobius gulosus). Schooling taxa Menidia spp. and two economically valuable species, C. sapidus and Leiostomus xanthurus (highly abundant but only collected in 37% of samples), were also among the dominant taxa. These ten taxa represented 81% of all nekton collected. Nekton Density Among Habitat Types.—Tidal Creeks.—Fifty-six taxa were collected in tidal-creek samples. Total nekton density did not differ between the three creeks (RMANOVA P = 0.535; Fig. 4), nor did mean density of residents (P = 0.216). There was a trend for greater density of schooling taxa in the upper bay creek, but it was not significant (P = 0.052; Fig. 4). The most dominant species in the creeks were residents or schooling taxa and included L. parva, P. latipinna, M. gulosus, Menidia spp., G. holbrooki, A. xenica, F. grandis, and Anchoa mitchilli though the abundances of any one species varied considerably between creeks (Table 2). Mean density of economic taxa was similar across all tidal creeks (P = 0.528). Economically important species were 18.3% of the overall assemblage in tidal creeks and consisted primarily of L. xanthurus, C. sapidus, Sciaenops ocellatus, Mugil spp. (including M. cephalus), and Farfantepenaeus duorarum (Table 2). Mosquito Ditches.—Thirty-eight taxa were collected in mosquito-ditch samples. However, the number of taxa collected in mosquito ditches at mid- (n = 32) and upper-bay (n =34) sites and in the mid-bay creek (n = 37) was considerably less than that collected from the upper-bay creek (n = 47). Total nekton densities did not differ between creeks and mosquito ditches (RMANOVA P = 0.198; Fig. 5) but, in general, were slightly greater at mid-bay sites, in general (P = 0.048). Significantly greater nekton densities in mosquito ditches in the mid-bay (P = 0.034) caused the difference be- KREBS ET AL.: NEKTON IN TIDAL CREEKS AND DITCHES 851 - SE) of (1) SE) total of (1) nekton, residents,(2) schooling,(3) and economic(4) taxa in tidal creeks and(crk) mosquito-control ditches (msq) SE) of (1) total schooling,economic residents,nekton, and SE) of (3) (1) taxa (2) (4) in three tidal creeks located mid-at upper-, and lower-bay ± ± Figure 4. Mean density ( regions in Tampa Bay, FL. Significancevalues Bay, regions in Tampa are reportedfor the maineffect. Note that the scale differs between the firsttwo and two last panels. Figure 5. Mean density ( located at upper- and mid-bay regions in Tampa Bay, FL. Significance values are reported for main effects of bay region (R) and habitat type (H) and the interac type habitat the (H) and and (R) region bay of effects mainFL.reported Significancevalues for are Bay, Tampa in regions mid-bay and upper- locatedat tion. Bars with the same letters not significantly were Note differentthat 0.05. at P < the scale differs between the firsttwo and two last panels. 852 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 SE) of (1) total economicschooling, residents,and nekton,taxa (3) (4) of SE) (1) (2) in a tidal mosquito-control creek (crk), ditchesand (msq), ± Figure 6. Mean density ( stormwater-drainage locatedditches (stm) at the FL.region upper-bay in Bay, SignificanceTampa values are reportedfor the maineffect. Bars with the same letters not significantly were Note differentthat 0.05. at P < the scale differs between the firsttwo and two last panels. KREBS ET AL.: NEKTON IN TIDAL CREEKS AND DITCHES 853 tween upper- and mid-bay regions. Specifically, high abundances of resident nekton in these mosquito ditches likely contributed to the observed difference, as residents were significantly more abundant in mid-bay mosquito ditches than elsewhere (P = 0.002; Fig. 5). Five resident taxa dominated (87%) the mosquito ditches in the mid- bay and included P. latipinna, L. parva, G. holbrooki, C. variegatus, and A. xenica, and the schooling taxa, Menidia spp. The same taxa comprised slightly less (78%) of the assemblage collected at the equally species-poor mid-bay creek sites. In contrast, the abundant residents identified above were considerably less dominant at upper- bay creek (54%) and mosquito ditch (60%) sites. Schooling taxa were absent from mid-bay creek and mosquito ditch sites with the exception of Menidia spp. (Table 2), which explained the significantly higher abundance of schooling taxa observed in the upper bay creek (P = 0.022). Economic taxa did not differ in their abundance between bay regions (P = 0.154) or between creeks and mosquito ditches (P = 0.063), but species-specific abundances varied widely by region and habitat type (Table 2). Specifically, juvenile blue crabs C.( sapidus) and black drum (Pogonias cromis) were considerably more abundant in upper-bay mosquito ditches than in any of the other habitat types sampled in this study. Stormwater Ditches (Habitat Comparisons Restricted to Upper Bay Site).—Fifty taxa were collected in stormwater ditches, including eight unique taxa that were primarily schooling or of freshwater affinity. In comparison, 47 taxa were collected in the creek and 34 taxa in mosquito ditches. Total nekton density did not differ between the three habitat types in the upper bay (RMANOVA P = 0.072; Fig. 6). Despite high species richness in stormwater ditches (Table 2), this habitat type had statistically lower densities of resident (P = 0.044) and economic (P = 0.032) taxa compared to the creek. In terms of individual species, F. grandis and P. latipinna were dominant in all three habitat types though densities were lower in stormwater ditches (Table 2). Similarly, L. parva and M. gulosus were much less abundant in stormwater ditches. In contrast, A. mitchilli and Eucinostomus spp. (those individuals < 40 mm SL that could not be reliably identified to species) were much more abundant relative to creek and mosquito ditch sites. Other schooling taxa, including Menidia spp. and several clupeids (Harengula jaguana, Opisthonema oglinum, Brevoortia spp.) were also characteristic of the stormwater ditches (Table 2). However, their densities differed statistically only from those in mosquito ditches (P < 0.0001). Species Diversity.—Mean Shannon diversity ± SE in Tampa Bay wetlands was 1.64 ± 0.03. Diversity was statistically greatest at upper-bay creek sites (1.88 ± 0.06; P < 0.05 for all pair-wise comparisons) and least at mosquito ditches in the mid-bay (1.40 ± 0.06; P < 0.05). Diversity was statistically similar in all other habitat types (range = 1.52–1.66; GLM P > 0.05), though the mid-bay creek was among the least diverse of these (1.52 ± 0.08) and did not differ from mosquito ditches in the same region (P = 0.25). Community Analysis.—Nekton assemblage structure, defined as taxonomic composition and relative abundance, differed significantly among creeks and ditches (ANOSIM P < 0.001; Global R = 0.31), as is evident in the separation of sample sites in multidimensional space (Fig. 7). Eight taxa explained 88% of the pattern observed in the ordination (BVSTEP) and typically differed between habitat types in terms of rank and/or differential abundance (Table 2). Mosquito ditches and stormwater ditches were the most dissimilar of the habitat types, occurring at opposite ends of the first MDS axis (R = 0.64; Fig. 7). Distinguishing species were A. mitchilli and 854 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Figure 7. Non-metric multidimensional scaling plot of nekton assemblages in tidal creek (●), mosquito-control (●), and stormwater-drainage (°) ditches. Points represent the species composition and abundance of each sample site averaged across four seasonal samples taken during 2004.

Eucinostomus spp. (including Eucinostomus gula and Eucinostomus harengulus) in the stormwater ditches vs greater relative abundances of L. parva, P. latipinna, G. holbrooki, and C. variegatus in the mosquito ditches (Table 2). Tidal creeks in the center of the MDS plot overlapped with both ditch types in assemblage structure. Although creeks and mosquito ditches were characterized by high absolute abundances of P. latipinna, L. parva, and Menidia spp., these taxa were relatively more abundant in mosquito ditches, a pattern that contributed to the dissimilarity between the two habitat types. Additionally, G. holbrooki reached its greatest abundance in mosquito ditches whereas M. gulosus reached its greatest abundance in creeks, thus further distinguishing between the two habitat types. Finally, stormwater ditches were best distinguished from tidal creeks by the high abundance of Eucinostomus spp. and A. mitchilli, the same taxa that discriminated stormwater and mosquito ditches. Patterns of nekton assemblage structure were correlated with patterns observed in habitat data from tidal creeks, mosquito-control, and stormwater-drainage ditches (BIO-ENV; r = 0.43). This correlation was best explained by channel width, salinity, and dissolved oxygen.

Discussion

Our finding that nekton community structure in mosquito-control ditches was similar to that observed in natural tidal creeks suggests that the nature of the alteration (i.e., for mosquito-control vs stormwater-drainage) may be more important in determining habitat use than the fact that the habitat has been altered. For example, mosquito ditches in this study were similar to tidal creeks in terms of substrate type (J. Krebs, unpubl. data), channel width, and shoreline vegetation in many instances, but stormwater ditches were wider, had sandier, less organically enriched substrates, KREBS ET AL.: NEKTON IN TIDAL CREEKS AND DITCHES 855 and swifter current velocities. In some cases, man-made wetlands may support greater nekton densities than natural wetlands. In a study of created wetlands in Tampa Bay, four man-made tidal creeks (3 yrs old) supported a greater abundance of economic taxa (e.g., C. undecimalis, S. ocellatus, and Mugil spp.) than a naturally formed tidal creek (Kurz et al., 1998). On a local scale, salinity, dissolved oxygen, and channel width were the habitat factors that best linked with patterns of nekton distribution among tidal creeks, mosquito-control ditches, and stormwater-drainage ditches (BIO-ENV). Higher salinity, shallower depths, and narrower channel widths were characteristic of mosquito ditches, just as lower salinity, greater depths, and wider channels were characteristic of stormwater ditches (Table 1). Given the differences in terms of the habitat factors, it is not surprising that the two man-made ditch types were used by different nekton assemblages. Salinity regime is known to influence distribution patterns among nekton in estuaries (e.g., Gunter, 1961; Daiber, 1977; Rowe and Dunson, 1995; Nemerson and Able, 2004). In our study, site-specific salinity differences were a function of both freshwater inflow (qualitatively defined) and location within the estuary. Salinities at creek sites in the lower bay and stormwater ditches in the upper bay were lowest overall. The volume of freshwater input in the lower-bay creek with a sizable catchment (data not shown) and in upper-bay stormwater ditches that drain residential areas was likely greater than that at other locations. Water depth has also been shown to be a key determinant in structuring fish assemblages in shallow water habitats (McIvor and Odum, 1988; Paterson and Whitfield, 2000; Ellis and Bell, 2004) by excluding larger piscivores and providing refuge for smaller-sized nekton. Mosquito-control ditches, because of their characteristically shallow water depths, may play a role as a nursery area for newly recruited blue crab, C. sapidus (Yeager et al., 2007) and black drum, P. cromis (the present study) as both species were observed in greater abundance in mosquito-ditched mangrove wetlands. As such, mosquito-control ditches, despite their stigma as stagnant and anoxic, can provide shallow water habitat and possible spawning sites (Nordlie, 2000) for resident and forage taxa. Further, mosquito ditches may provide a source of prey to piscivorous predators during falling tides when small nekton are forced to move into adjacent deeper water habitats. Channel width and openness may also affect habitat value for small nekton. In our study, average width of stormwater ditches (11.0 m) was considerably greater than that of mosquito ditches (5.1 m). In contrast to the short distances and shallow depths between dense red mangrove prop roots on opposite banks of mosquito ditches, stormwater ditches provided relatively sparse white mangroves separated by greater depths and wider distances between banks. In areas which have been altered for mosquito-control or stormwater-drainage, increased surface elevation adjacent to ditches may create more suitable conditions for white mangrove recruitment (Craighead and Gilbert, 1962; Ball, 1980). In contrast, elevation along tidal creeks is naturally lower with more frequently inundated banks, conditions suitable for red mangroves. For this reason, stormwater ditches may have provided less structurally complex habitat than what is typically found in narrower mosquito ditches. Further- more, the closed mangrove canopy created by narrow distances between shorelines in mosquito ditches provides shade, an additional factor controlling habitat use by smaller-sized forage taxa (Ellis and Bell, 2004). 856 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

On a larger scale, spatial location, both within the estuary and relative to prevailing tidal circulation throughout the estuary, may influence the degree of tidal exchange between wetland and bay, and thus access by migrating nekton to particular wetland regions in an estuary (Weinstein et al., 1988; Galperin et al., 1992). Spatial location may also influence the probability of settlement by passively transported planktonic larvae of nekton, which are spawned in the estuary or in offshore waters (Weinstein et al., 1988; Olmi, 1995; Wenner et al., 1998). The high abundance of resident nekton in mid-bay mosquito ditches in our study and concurrent low abundances of juvenile sportfish and other offshore-spawned species at this location (Table 2) would suggest a physical disconnect between these mosquito ditches and the larger estuary. Particularly notable in mid-bay mosquito ditches was the low abundance of young- of-the-year blue crabs (see also Yeager et al. 2007) and the absence here of hogchoker, Trinectes maculatus, an otherwise well-distributed species whose larvae are transported up-estuary by tidal currents (Dovel et al., 1969). Furthermore, similarities in species composition and diversity between the mid-bay mosquito ditches and tidal creek suggest that the spatial distribution of nekton in this region is driven by large- scale processes (i.e., tidal circulation) that may restrict dispersal of larvae of transient species and obscure small-scale selection of tidal creek or mosquito ditch habitat. Our finding of contrasting nekton assemblages from mosquito-control ditches at two mangrove wetlands in Tampa Bay, and similarity between tidal creek and man- made ditch habitat at the upper bay wetland further reinforces the argument that spatial location within the estuary may play a large role in structuring the nekton community. Our results also suggest that mosquito-control ditches, given the proper physical conditions, can serve as habitat for nekton assemblages typically associated with tidal creeks. Species composition and richness of the wetland-associated nekton collected in Tampa Bay (65 species) are similar to those observed in previous studies in the southeastern United States. Taxa from the families Poeciliidae, Cyprinodontidae, Fundu- lidae, Atherinidae, Mugilidae, Gobiidae, Gerreidae, and Sciaenidae have consistently been reported as dominant members of intertidal salt marsh communities along the Atlantic and Gulf coasts (Kilby, 1955; Subrahmanyam and Drake, 1975; Weinstein, 1979; Rountree and Able, 1992; Kneib, 1997; Kurz et al., 1998). Several studies on the Gulf coast of Florida have observed 50–65 taxa (Kilby, 1955; Subrahmanyam and Drake, 1975; Kurz et al., 1998). For mangrove prop root habitat, Thayer et al. (1987) recorded 64 fish taxa inW hitewater, Coot, and Florida bays, Ley et al. (1999) recorded 76 fish species in northeastern Florida Bay, Serafy et al. (2003) recorded 38 fish species in Biscayne Bay, and Sheridan (1992) documented 13 fish species from the intertidal forest floor in Rookery Bay. This relatively large range in species richness in South Florida mangroves likely reflects differences in microhabitats, gear biases, and sampling effort. Stormwater-drainage ditches were characterized by schooling species from the families Atherinidae, Clupeidae, and Engraulidae. Association of schooling taxa with open water habitat types similar to that observed in stormwater-drainage ditches has been previously documented. Sheridan (1992) found the greatest abundances of schooling taxa (i.e., Brevoortia smithi, H. jaguana, and Anchoa hepsetus) in open waters in Rookery Bay, Florida, while Silverman (2006) documented large schools of engraulids, Anchoa spp., on intertidal mudflats (but not in adjacent mangrove forests) in Big Sable Creek, Everglades National Park, FL. Hettler (1989) cites a complete KREBS ET AL.: NEKTON IN TIDAL CREEKS AND DITCHES 857

absence of another schooling taxon, Brevoortia spp. in collections from Spartina salt marsh surfaces in North Carolina. In the present study, several other non-schooling species were collected almost solely in stormwater ditches. These included the transients, timucu (Strongylura timucu), blackcheek tonguefish Symphurus ( plagiusa), and Atlantic spadefish Chaetodipterus( faber), which are common in the adjacent estuary. Several taxa common to adjacent freshwater habitat, bluegill (Lepomis macrochirus), largemouth bass (Micropterus salmoides), and golden shiner (Notemigo- nus crysoleucas) occurred solely in stormwater ditches. The presence of these unique species from adjacent habitats suggests that stormwater ditches may act as a corridor between bay and freshwater habitats. Nekton densities observed in subtidal creeks and man-made ditches in this study averaged from 400–600 nekton 100 m–2, with the exception of mosquito-control ditches at Weedon Island where mean nekton density exceeded 1000 nekton 100 m–2. The nekton densities from our study are similar to those reported for mangrove prop root habitat in south Florida. Densities in Coot, Whitewater and Florida bays typically averaged 575 fish 100 –2m and rarely exceeded 1500 fish 100 –2m (Thayer et al., 1987), while mean fish density in Rookery Bay mangroves was estimated to be 551 fish 100 –2m (Sheridan, 1992). A recent study of mangrove-lined tidal creeks in Charlotte Harbor, FL, compared creeks having natural and altered watersheds and observed nekton densities in both types of watershed ranging from 100–600 nekton 100 m–2 during winter and spring, but exceeding 1000 nekton 100 m–2 in natural creeks during the summer and fall (Adams, 2005). Mean fish density, not including blue crabs (C. sapidus) or pink shrimp (F. duorarum) along Spartina alterniflora edge in Texas salt marshes was 771 fish 100 –2m (Minello, 1999), well within the range of nekton densities observed in our Tampa Bay study. In terms of specific species of economic value in Tampa Bay, mean densities of juvenile Mugil spp. and S. ocellatus did not typically exceed 20 fish 100 –2m in tidal creeks or mosquito-control ditches, while juvenile C. sapidus ranged from 10–44 crabs 100 m–2 in these habitat types (Table 2). In contrast, Minello (1999) reported densities for these three species in emergent marsh and over submerged aquatic vegetation, oyster reef, and bare substrate that varied considerably from the present study: blue crabs were considerably less abundant, striped mullet had similar densities, and red drum were more abundant in Tampa Bay wetlands compared with Texas wetlands. Although habitat and nekton assemblage structure were relatively similar between tidal creeks and mosquito-control ditches in this study, the mosquito ditches that we sampled were better flushed by the tide than were other, unsampled mosquito- control ditches. We were unable to sample with haul seines in the excessively muddy, more overgrown, and largely infilled ditches that typically occur in the poorly flushed interior portions of ditched wetlands in Tampa Bay. As a result, our conclusions about mosquito-control ditches as nekton habitat apply only to the sampled subset of mosquito ditches. Although Brown (1987) used small plexiglass Breder traps (Breder, 1960) in both relatively open and in older, more infilled mosquito ditches near our Weedon Island sites, his finding of greater dominance of resident species in infilled ditches is difficult to interpret as Breder traps are biased against transient species (Fulling et al., 1999). The similarity in nekton assemblages observed between naturally formed tidal creeks and mosquito-control ditches, coupled with the dissimilarity between sampled mosquito-control and stormwater-drainage ditches highlights the potential 858 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 ecological differences between types of man-made mangrove habitat. Based on our data, we suggest that nekton community structure can not be predicted from habitat origin alone (i.e., naturally formed vs man-made). Instead it is necessary to consider the habitat conditions that result from alteration, as these conditions can be more similar between natural and man-made tidal channels than between man-made channels created for different purposes. For example, sampled mosquito-control ditches apparently provided comparable habitat to natural tidal creeks for forage fish and juvenile sportfish because habitat factors such as frequency of tidal flushing and presence of shoreline structure (e.g., Rhizophora prop roots) were similar. The caveat that must be considered, however, is that many of Tampa Bay’s mosquito ditches were constructed in areas of low tidal circulation and as a result have accumulated sediment that would have made them difficult to sample had they been included in our study. Such in-filling has resulted in habitat conditions that differ greatly from self-maintaining meandering channels typical of naturally formed tidal creeks. In summary, the results of this study demonstrate that existing physical conditions and patterns of habitat use by aquatic fauna need to be carefully assessed and integrated prior to implementing restoration projects in mosquito-ditched wetlands and in other systems affected by habitat alteration.

Acknowledgments

We thank, first and foremost, D. Nielsen, L. Yeager, No. Hansen, Na. Hansen, J. Sanford, G. Craig, K. Hart, G. Hill, K. Hill, R. Krebs, R. Poyner, N. Silverman, and K. Sommers for their tremendous effort in collecting and processing the samples.W e thank E. Sherwood and M. Greenwood for their statistical advice. We also acknowledge the continuing cooperation of county and state managers who have provided access to sample sites as well as historical information and assistance in the field. These include: K. Hill, P. Leasure and K. Thompson (Pinellas County), L. Rives (City of Oldsmar), and R. Runnels (Florida DEP). Special thanks to T. Burress for acquiring much of the literature cited in this study. We thank K. Hart, J. Kind- inger (U.S. Geological Survey) and R. Edwards (University of South Florida) as well as the editors and two anonymous reviewers whose comments have greatly improved this manuscript. Funding for this project was provided by the U.S. Geological Survey’s Tampa Bay Integrated Science Study.

Literature Cited

Adams, A. J. 2005. Evaluating the effects of restoration of subtropical oligohaline marshes on abundance and habitat use by juvenile snook, Centropomus undecimalis and associated fish communities. Final Report to Charlotte Harbor National Estuary Program. 54 p. Alongi, D. M. 2002. Present state and future of the world’s mangrove forests. Environ. Conser- vat. 29: 331–349. Ball, M. C. 1980. Patterns of secondary succession in a mangrove forest of southern Florida. Oecologia 44: 226–235. Breder, C. M. 1960. Design for a fry trap. Zool. N. Y. 45: 155–164. Brown, J. E. 1987. Distribution and diversity of fishes in a Tampa Bay mangrove swamp and the effects of rotary ditching. MS Thesis, University of South Florida. 61 p. Bunn, S. E. and A. H. Arthington. 2002. Basic principles and ecological consequences of altered flow regimes for aquatic biodiversity. Environ. Manage. 30: 492–507. Clark, P. A. 1992. Management directions and needs for Tampa Bay tidal tributaries. Pages 497–510 in S. F. Treat and P. A. Clark, eds. Proc. Tampa Bay Area Scientific and Informa- KREBS ET AL.: NEKTON IN TIDAL CREEKS AND DITCHES 859

tion Symp. 2, Feb. 27–Mar. 1, 1991. Tampa Bay Regional Planning Council, St. Petersburg, FL. Clarke, K. R. and R. N. Gorley. 2006. PRIMER v6: User Manual/Tutorial. PRIMER-E Ltd: Plym- outh, United Kingdom. 190 p. Coorey, D. N., K. W. Able, and J. K. Shisler. 1985. Life history and food habits of the inland silverside, Menidia beryllina, in a New Jersey salt marsh. Bull. New Jersey Acad. Sci. 30: 29–38. Craighead, F. C. and V. C. Gilbert. 1962. The effects of hurricane Donna on the vegetation of southern Florida. Quar. J. Fla. Acad. Sci. 25: 1–27. Daiber, F. C. 1977. Saltmarsh animals: distributions related to tidal flooding, salinity and vegetation. Pages 79–108 in V. Chapman, ed. Ecosystems of the World I: Wet Coastal Ecosys- tems. Elsevier Scientific Publishing Company, Amsterdam. Deegan, L. A. and B. A. Thompson. 1985. The ecology of fish communities in the Mississippi River deltaic plain. Pages 35–56 in A. Yáñez-Arancibia, ed. Fish community ecology in estuaries and coastal lagoons: Towards an Ecosystem Integration. UNAM Press, Mexico City. Dovel, W. L., J. A. Mihursky, and A. J. McErlean. 1969. Life history aspects of the hogchoker, Trinectes maculatus, in the Patuxent River estuary, Maryland. Chesap. Sci. 10: 104–119. Ellis, W. L. and S. S. Bell. 2004. Conditional use of mangrove habitats by fishes: Depth as a cue to avoid predators. Estuaries 27: 966–976. Freeman, M. C., Z. H. Bowen, K. D. Bovee, and E. R. Irwin. 2001. Flow and habitat effects on juvenile fish abundance in natural and altered flow regimes. Ecol. Appl. 11: 179–190. Fulling, G. L., M. S. Peterson, and G. J. Crego. 1999. Comparison of Breder traps and seines used to sample marsh nekton. Estuaries 22: 224–230. Galperin, B., A. Blumberg, and R. Weisberg. 1992. A time-dependent three dimensional model of circulation in Tampa Bay. Pages 77–97 in S. F. Treat and P. A. Clark, eds. Proc. Tampa Bay Area Scientific Information Symp. 2, February 27–March 1, 1991. Tampa Bay Regional Planning Council, St. Petersburg, FL. Gunter, G. 1961. Some relations of estuarine organisms to salinity. Limnol. Oceanogr. 6: 182– 190. Hettler, W. F. 1989. Nekton use of regularly-flooded saltmarsh cordgrass habitat in North Caro- lina, USA. Mar. Ecol. Prog. Ser. 56: 111–118. Johnston, Jr., S. A. 1981. Estuarine dredge and fill activities: A review of impacts. Environ. Man- age. 5: 427–440. Kilby, J. D. 1955. The fishes of two Gulf coastal marsh areas of Florida. Tul. Stu. Zool. 2: 175– 247. Kneib, R. T. 1997. The role of tidal marshes in the ecology of estuarine nekton. Oceanogr. Mar. Biol. Annu. Rev. 35: 163–220. Kurz, R. C., R. W. Fenwick, and K. A. Davis. 1998. A comparison of fish communities in restored and natural salt marshes in Tampa Bay, FL. Tech. Rept., Southwest Florida Water Management District, 49 p. Lewis, R. R. 1992. Coastal habitat restoration as a fishery management tool. Stemming the tide of coastal fish habitat loss. Pages 169–173 in R. H. Stroud, ed., Proc. Symp. on Conserva- tion of Coastal Fish Habitat, March 7–9, 1991. National Coalition for Marine Conservation, Inc., Savannah, GA. ______. 2005. Ecological engineering for successful management and restoration of mangrove forests. Ecol. Eng. 24: 403–418. ______and E. D. Estevez. 1988. The Ecology of Tampa Bay, Florida: An Estuarine Profile. U.S. Fish and Wildlife Service, Biol. Rept. 85(7.18), 132 p. ______, R. G. Gilmore Jr., D. W. Crewz, and W. E. Odum. 1985. Mangrove habitat and fishery resources of Florida. Pages 281–336in W. Seaman, ed. Florida Aquatic Habitat and Fishery Resources. Florida Chapter of the American Fisheries Society, Eustis, FL. Ley, J. A., C. C. McIvor, and C. L. Montague. 1999. Fishes in mangrove prop-root habitats of northeastern Florida Bay: distinct assemblages across an estuarine gradient. Estuar. Coast. Shelf Sci. 48: 701–723. 860 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Littell, R. C., P. R. Henry, and C. B. Ammerman. 1998. Statistical Analysis of Repeated Mea- sures Data Using SAS Procedures. J. Anim. Sci. 76: 1216–1231. Lotrich, V. A. 1975. Summer home range and movements of Fundulus heteroclitus (Pisces: Cyprinodontidae) in a tidal creek. Ecology 56: 191–198. McIvor, C. C. and W. E. Odum. 1988. Food, predation risk, and microhabitat selection in a marsh fish assemblage. Ecology 69: 1341–1351. McIvor, C. C. and L. P. Rozas. 1996. Direct nekton use of intertidal saltmarsh habitat and linkage with adjacent habitats: a review from the southeastern United States. Pages 311–333 in K. F. Nordstrom and C. T. Roman, eds. Estuarine Shores: Evolution, Environments and Human Alterations, John Wiley and Sons Ltd., New York. 486 p. Minello, T. J. 1999. Nekton densities in shallow estuarine habitats of Texas and Louisiana and the identification of essential fish habitat. Pages 43–75in L. Benaka, ed. Fish habitat: essential fish habitat and rehabilitation. American Fisheries Society, Symp. 22, Bethesda, Mary- land, 400 p. ______, K. W. Able, M. P. Weinstein, and C. G. Hays. 2003. Salt marshes as nurseries for nekton: Testing hypotheses on density, growth and survival through meta-analysis. Mar. Ecol. Prog. Ser. 246: 39–59. Nelson, J. S., E. J. Crossman, H. Espinosa-Pérez, L. T. Findley, C. R. Gilbert, R. N. Lea, and J. D. Williams. 2004. Common and scientific names of fishes from the United States, Canada and Mexico. American Fisheries Society, Special Publications 29, Bethesda. Nemerson, D. M. and K. W. Able. 2004. Spatial patterns in diet and distribution of juveniles of four fish species in Delaware Bay marsh creeks: Factors influencing fish abundance. Mar. Ecol. Prog. Ser. 276: 249–262. Nordlie, F. G. 2000. Patterns of reproduction and development of selected resident teleosts of Florida salt marshes. Hydrobiologia 434: 165–182. ______. 2003. Fish communities of estuarine salt marshes of eastern North America, and comparisons with temperate estuaries of other continents. Rev. Fish Biol. Fish. 13: 281–325. Odum, W. E., C. C. McIvor, and T. J. Smith III. 1982. The ecology of the mangroves of south Florida: a community profile.W ashington, D.C. Fish and Wildlife Service, Office of Biologi- cal Services FWS/OBS-81/24. 144 p. Olmi III, E. J. 1995. Ingress of blue crab megalopae in the York River, Virginia, 1987–1989. Bull. Mar. Sci. 57: 753–780. Paterson, A. W. and A. K. Whitfield. 2000. Do shallow-water habitats function as refugia for juvenile fishes? Estuar. Coast. Shelf. Sci. 51: 359–364. Rountree, R. A. and K. W. Able. 1992. Fauna of polyhaline subtidal marsh creeks in southern New Jersey: Composition, abundance and biomass. Estuaries 15: 171–185. Rowe, C. L. and W. A. Dunson. 1995. Individual and interactive effects of salinity and initial fish density on a salt marsh assemblage. Mar. Ecol. Prog. Ser. 128: 271–278. Roy, A. H., M. C. Freeman, B. J. Freeman, S. J. Wenger, W. E. Ensign, and J. L. Meyer. 2005. Investigating hydrological alteration as a mechanism of fish assemblage shifts in urbanizing streams. J. N. Am. Benthol. Soc. 24: 656–678. SAS Institute Inc. 2000. SAS OnlineDoc. Version 8, Cary, NC: SAS Institute Inc. Scheirer, C. J., W. S. Ray, and N. Hare. 1976. The analysis of ranked data derived from completely randomized factorial designs. Biometrics 32: 127–129. Serafy, J. E., C. E. Faunce, and J. J. Lorenz. 2003. Mangrove shoreline fishes of Biscayne Bay, Florida. Bull. Mar. Sci. 72: 161–180. Sheridan, P. F. 1992. Comparative habitat utilization by estuarine macrofauna within the mangrove ecosystem of Rookery Bay, Florida. Bull. Mar. Sci. 50: 21–39. Silverman, N. L. 2006. Hurricane-induced conversion of mangrove forest to mudflat: impacts on nekton, Big Sable Creek, Florida. MS thesis. University of South Florida, Saint Peters- burg, FL. 50 p. KREBS ET AL.: NEKTON IN TIDAL CREEKS AND DITCHES 861

Smith, K. J. 1995. Processes regulating habitat use by salt marsh nekton in a southern New Jersey estuary. Ph.D Diss. Rutgers University, New Brunswick, New Jersey. 166 p. Sokal, R. R. and F. J. Rohlf. 1997. Biometry, 3rd Ed. W.H. Freeman and Co., New York. 887 p. Subrahmanyam, C. B. and S. H. Drake. 1975. Studies on the animal communities in two north Florida salt marshes Part I. Fish communities. Bull. Mar. Sci. 25: 445–465. Teo, S. L. H. and K. W. Able. 2003. Habitat use and movement of the mummichog (Fundulus heteroclitus) in a restored salt marsh. Estuaries 26: 720–730. Thayer,G . W., D. R. Colby, and W. F. Hettler, Jr. 1987. Utilization of the red mangrove prop root habitat by fishes in south Florida. Mar. Ecol. Prog. Ser. 35: 25–38. ______, H. H. Stuart, W. J. Kenworthy, J. F. Ustach, and A. B. Hall. 1978. Habitat values of salt marshes, mangroves, and seagrasses for aquatic organisms. Pages 235–247 in P. E. Greeson, J. R. Clark, and J. E. Clark, eds. Wetland functions and values: the state of our understanding. American Water Resources Association, Minnesota. 674 p. Warren, R. S., P. E. Fell, R. Rozsa, A. H. Brawley, A. C. Orsted, E. T. Olson, V. Swamy, and W. A. Niering. 2002. Salt Marsh Restoration in Connecticut: 20 Years of Science and Manage- ment. Restor. Ecol. 10: 497–513. Weinstein, M. P. 1979. Shallow marsh habitats as primary nurseries for fishes and shellfish, Cape Fear River, North Carolina. Fish. Bull. 77: 339–357. ______. 1988. Larval fish and shellfish transport through inlets. Amer. Fish. Soc. Symp. 3. Bethesda, MD. 165 p. ______, J. M. Teal, J. H. Balletto, and K. A. Strait. 2000. Restoration principles emerging from one of the world’s largest tidal marsh restoration projects. Wetl. Ecol. Manage. 9: 387–407. Wenner E., D. Knott, J. Blanton, C. Barans, and J. Amft. 1998. Roles of tidal and wind-generated currents in transporting white shrimp (Penaeus setiferus) postlarvae through a South Carolina (USA) inlet. J. Plankton Res. 20: 2333–2356. Yeager, L. A., J. M. Krebs, A. B. Brame, and C. C. McIvor. 2007. Comparison of juvenile blue crab abundances in three mangrove habitat types in Tampa Bay, Florida USA. Bull. Mar. Sci. 80: 555–565. Zippin, C. 1958. The removal method of population estimation. J. Wildl. Manage. 22: 82–90.

Addresses: (J.M.K., A.B.B.) ETI Professionals, 4902 Eisenhower Boulevard, Suite 150, Tampa Florida 33634. (C.C.M.) United States Geological Survey, Florida Integrated Science Center, 600 4th Street South, Saint Petersburg, Florida 33701. Corresponding Author: (J.M.K.) Email: . BULLETIN OF MARINE SCIENCE, 80(3): 863–877, 2007

Habitat fragmentation decreases fish secondary production in Bahamian tidal creeks

L. Valentine-Rose, C. A. Layman, D. A. Arrington, and A. L. Rypel

Abstract We provide an example of how fish secondary production can be used to assess effects of anthropogenic stress on ecosystem function. Specifically, we demonstrate that fragmentation of hydrologic connectivity in small tidal creeks on Andros Is- land, Bahamas, decreases secondary production by as much as an order of magnitude for five economically important fish species: gray snapper (Lutjanus griseus Linnaeus, 1758), schoolmaster snapper (Lutjanus apodus (Walbaum, 1792)), cubera snapper (Lutjanus apodus cyanopterus (Cuvier, 1828)), blue striped grunt (Haemu- lon sciurus (Shaw, 1803)), and sailor’s choice (Haemulon parra (Desmarest, 1823)). We conducted more than 800 individual quadrat surveys, in eight creeks that varied in degree of habitat fragmentation, to estimate total biomass for the focal species. Biomass values were multiplied by in situ or published growth rates to calculate annual secondary production. Differences in estimates of secondary production among creeks were attributable to fewer species, fewer individuals, and smaller individuals in fragmented creeks. Secondary production estimates are a useful way to measure responses of ecosystem function to anthropogenic stress because they are a composite value that incorporates many important aspects of ecosystem function.

Habitat degradation and alteration is a fundamental cause of global declines in biodiversity and loss of critical ecosystem services (Naeem et al., 1994; Heywood and Watson, 1995; Chapin et al., 1997, 1998; Vitousek et al., 1997; Tilman, 1999). Habitat fragmentation, or the sub-division of natural ecosystems (Saunders et al., 1991; Fahrig, 2002, 2003), is one of the most common forms of habitat alteration. In aquatic systems, fragmentation of hydrologic connectivity, i.e., blocking water- mediated transfer of matter, energy, or organisms within or between elements of the hydrologic cycle (Pringle, 2001, 2003a,b), is one of the most common anthropogenic impacts (Holmquist et al., 1998, Benstead et al., 1999, March et al., 2003). In estuarine ecosystems, habitat fragmentation is wide-spread and particularly damaging to ecological integrity of the systems (Ray, 2005). Tidal flow is a primary physical driver in estuaries, and restricting tidal flow alters physicochemical parameters as well as floral and faunal community structure (Roman et al., 1984; Warren et al., 2002). While there are many studies on effects of fragmentation in rivers (For- man and Alexander, 1998; Trombulak and Frissell, 1999; Pringle, 2003b; Taylor et al., 2006) and in temperate estuaries (Burdick et al., 1997; Eertman et al., 2002; Roman et al., 2002; Tanner et al., 2002; Raposa and Roman, 2003; Ray, 2005), there has been little research in tropical systems (but see Holmquist et al., 1998; Benstead et al., 1999; March et al., 2003; Layman et al., 2004b). Layman et al. (2004b) reported that > 80% of mangrove-lined tidal creek ecosystems on the east coast of Andros Island (the largest island in the Bahamas archipelago) were fragmented to some degree, causing substantial shifts in fish assemblage structure. Whereas this and other studies conducted in tropical coastal systems have documented aspects of altered community structure following fragmentation, effects on composite measures of ecosystem function (i.e., ecosystem processes) remain poorly understood (but see Mallin and Lewitus, 2004).

Secondary production, the accumulation of animal biomass over time, is one composite measure of ecosystem function (Benke, 1993). Secondary production is an ideal measure in assessing the nursery function of coastal ecosystems, i.e., ability to support large densities of juvenile marine biota and export these biota to offshore habitats, as it is a comprehensive measure of biomass, growth, and survival of organisms—three of the four core criteria for a nursery area (Beck et al., 2001; Dahlgren et al., 2006). Secondary production estimates recently have been used to assess the functional value of putative nursery habitats, e.g., oyster reefs (Peterson et al., 2003), artificial patch reefs P( owers et al., 2003), seagrass (Edgar and Shaw, 1995c; McCay and Rowe, 2003), and salt marshes (McCay and Rowe, 2003). For example, Peterson et al. (2003) used secondary production to show that oyster reef restoration could increase fish biomass values by as much as 5 g –2m y–1 over 30 yrs. In similar fashion, secondary production studies may be an excellent way to quantify effects of estuarine habitat fragmentation, as fragmentation may result in decreased organism abundance and size (Valentine-Rose et al., in press), two core components of production. In the present study, we describe differences in secondary production between unfragmented and fragmented tidal creeks on Andros Island, Bahamas. We discuss underlying reasons why fragmentation affected secondary production of these systems, and how these data can be used to attribute monetary value to a specific ecosystem service provided by these creek systems. This is the first attempt to estimate secondary production for fishes in mangrove habitat (Faunce and Serafy, 2006).

Methods

Site Description.—Secondary production estimates were calculated for eight tidal creeks located on the eastern shore of Andros Island, Bahamas. Tidal amplitude in these creeks is ~1 m, with water depths at high tide flooding extensive, shallow intertidal flats. At low tide, water is restricted to the main creek channels providing permanent habitat for fishes (i.e., fishes remain in the main creek channels even on the lowest tides). As an example, inundated area of two surveyed tidal creeks (Somerset and Bowen) at mean high tide level were 270.32 ha and 82.77 ha, while main channel areas inundated at the mean low tide level were 1.60 ha and 0.24 ha, respectively. Main channel areas may contain multiple habitat types (e.g., mangrove, rock, seagrass, or sand), and vary substantially in microtopography among systems. Intertidal flats typically have a sandy substrate with interspersed dwarf red (Rhizophora mangle Linnaeus), black (Avicennia germinans Linnaeus), white (Laguncularia racemosa (Linnaeus) Gaertn. f.), and buttonwood (Conocarpus erectus Linnaeus) mangroves. In the Bahamas, estuary fragmentation frequently results from construction of roads lacking flow conveyance structures (e.g., bridge or culverts) across some portion of the system, which decreases habitat quality and alters floral and faunal assemblage structure (Layman et al., 2004b, Valentine-Rose et al. in press). Following Dynesius and Nilsson (1994), we have identified four categories of estuary fragmentation, ranging from unfragmented systems with uninterrupted tidal flow to totally fragmented systems in which no tidal flow occurs (Layman et al., 2004b). Degree of fragmentation of hydrologic connectivity in creeks was defined a priori (following Layman et al., 2004b) as unfragmented (U), minimally fragmented (MF), partially fragmented (PF), or completely fragmented (CF). For the purposes of this study, in MF, PF, and CF systems, “upstream” refers to areas with restricted tidal flow, while “downstream” refers to areas on the ocean side of the road or causeway. Creek main-channel area, maximum channel depth at low tide, habitat area at low tide, and degree of fragmentation for each surveyed system are provided in Table 1. General Survey Methodology.—Secondary production estimates were derived from quantitative measures of density, biomass, and in situ or published growth rates for five ec- VALENTINE-ROSE ET AL.: HABITAT FRAGMENTATION AND PRODUCTION 865

Table 1. Summary characteristics of surveyed tidal creeks. U = unfragmented, PF = partially fragmented, CF = completely fragmented (sensu Layman et al., 2004b). D = downstream areas of fragmented tidal creeks and Up = upstream areas of fragmented tidal creeks. Proportion of total creek habitat area Main Max. channel Degree of channel depth at low Creek fragmentation area (m2) tide (m) Rock Mangrove Sand/Silt Seagrass White Bight U 1,008 1 0.59 0.18 0.23 0.01 Somerset U 16,245 0.5 0.10 0.14 0.03 0.74 Buryin Peace U 1,810 0.5 0.60 0.25 0.06 0.10 Love Hill PF-D 8,350 3 0.38 0.20 0.33 0.09 PF-Up 831 0.5 0.45 0.18 0.36 0.01 Davis PF-D 10,020 0.5 0.04 0.10 0.84 0.02 PF-Up 384 0.5 0 0.54 0.46 0 Man o War PF-D 11,718 0.25 0.39 0.02 0.59 0 PF-Up 1,650 0.25 0.04 0.10 0.87 0 Conch Sound PF-D 1,415 0.05 0.68 0.11 0.21 0 PF-Up 150 0.25 0.33 0.67 0 0 Bowen CF-D 2,439 2 0.30 0.33 0.10 0.27 CF-Up n/a 0.1 n/a 0 n/a 0 onomically important fish species (gray snapper: Lutjanus griseus Linnaeus, 1758; schoolmaster snapper: Lutjanus apodus (Walbaum, 1792); cubera snapper: Lutjanus cyanopterus (Cuvier, 1828); blue striped grunt: Haemulon sciurus (Shaw, 1803); sailor’s choice: Haemu- lon parra (Desmarest, 1823)). Density and biomass (based on length-weight regressions) of these species were estimated in 1 m2 sample plots using underwater visual census (UVC) with mask and snorkel (Brock, 1954; Layman et al., 2004b). Surveys at high tide in creeks are extremely difficult because of the spatial extent of inundated intertidal areas and dispersal of fish throughout these shallow-water areas. Thus, surveys were taken within 2 hrs of low tide, which facilitated estimating fish biomass because focal fish species were constricted to the main channel. While this resulted in superficially high per unit area (m–2) estimates in all creeks, extrapolated values of whole-creek production (see below) are more robust than high tide estimates because surveys were conducted when fish biomass was easiest to quantify. Secondary production estimates were based upon UVC samples collected in May 2005. Because secondary production is calculated on an annual time scale, a single-month survey of different size-classes and abundance is often combined with size-specific growth rates to calculate annual production (Huryn, 1996, 1998; Peterson et al., 2003; Powers et al., 2003). We estimated fish density (N) using 1 m2 visualized quadrats (i.e., we visually estimated the 1 m × 1 m sample area without actually deploying a quadrat prior to observing fish). We used quadrats instead of belt transects (e.g., Schmitt and Sullivan, 1996; Gust et al., 2001, Thompson and Mapstone, 2002; Mumby et al., 2004) because of specific underlying aspects of the tidal creek habitat template. Creeks included in this study are relatively small and, in most cases, it was not possible to lay out transects > 5 m in length within a single habitat type. Further, both quadrat and transect methods may bias fish density estimates because of induced behavioral responses of the target species by overestimating curious individuals that approach a diver, or underestimating shy individuals that move into refugia (Kulbicki, 1998; Sneddon, 2003). However, since study creeks had relatively shallow low tide water depths (Ta- ble 1), swimming along transect lines likely would have induced more behavioral responses in fishes than use of the quadrat method (due to increased disturbance of benthic habitat during continual swimming). Both techniques result in an underestimation of all fishes that utilize creek habitats, as some transient individuals may not remain in the main channel at low tide and only move into creeks at high tide to feed. However, based on observation of hundreds of 866 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 individually tagged juvenile snapper (J. E. Allgeier, University of Georgia, unpubl. data) and hundreds of hours of UVC (Layman et al., 2004b, L. Valentine-Rose, University of Alabama, unpubl. data), it is apparent that the overwhelming majority of individuals of the focal species are tidal creek residents, i.e., do not move out of creek systems over a period of days to weeks (although they may undergo ontogenetic habitat shifts out of creeks as adults). Additionally, use of quadrats allowed for extensive within-site replication and a more precise estimate of mean fish densities within habitat types. Although any survey methodology will likely result in biased fish density estimates, we employed a standardized survey protocol among all sites, which allowed for robust among-system comparisons. Density Estimation.—Since fish densities vary significantly among habitat types (Blaber et al., 1989; Sheaves, 1992; Mateo and Tobias, 2001; Eggleston et al., 2004), we conducted surveys separately in each habitat type in each creek (see Table 1). Quadrat areas were visually estimated one minute prior to obtaining densities and lengths because use of actual quadrats resulted in fish behavioral avoidance of the quadrat area. To scale production values for an entire creek, we calculated the total extent of each of four main habitat types (rock, mangrove, sand/silt, and seagrass) in the creek channel. We then divided each habitat into 25 equal sections, and in each sub-section we randomly selected a 1 m2 area for the UVC. At survey time, one author (L.V.-R.) approached the pre-determined quadrat area within ~2 m, waited 1 min for fish to acclimate, and then noted the density (N) and length (L) of each of the focal species. For fragmented tidal creeks, this protocol was conducted in each habitat type both upstream and downstream of the road blockage. Age Class and Biomass Estimation.—Age class distributions for each species were estimated by running published or in situ derived Von Bertalanfy growth rate parameters (Table 2) in reverse from known lengths estimated in the field:

AgeL=-ln 33-+LLage kt0 ^h6@ where Lage = field estimated lengths,L ∞ = asymptotic maximum length, t0 = age at zero length, and k = Brody growth coefficient. When multiple growth rates were available for a species, growth rates from similar habitats and environmental conditions were averaged and used to calculate age classes. Error in production estimates may have resulted from using the same growth rates for both fragmented and unfragmented tidal creeks for three of the focal species (Table 2). However, since growth rates are likely slower in fragmented creeks due to stressful conditions (see Discussion), differences in secondary production estimates presented here are conservative (i.e., we likely underestimated production differences between fragmented and unfragmented creeks). To calculate overall fish biomass, length/weight (L/W) regressions for each species were derived based on individuals ranging over the entire size distribution in creeks for that species. Individuals were collected by seining or hook-and-line from each tidal creek in May 2004 (1 yr prior to survey period). Standard length was measured in the lab with calipers to the nearest 0.1 mm, and wet weight was measured with an electronic scale to the nearest 0.01 g. L/W regressions were derived from this data (Table 3) based on the following formula: logWa=+ logbL log with lengths in millimeters, weights in grams, and species-specific constants a and b. Age class biomass (B) was calculated by multiplying the N of each age class by the weights calculated from the L/W regression, and total B was found by summing age class B estimates. Biomass m–2 estimates were calculated for each habitat in each creek to (1) compare fish distributions across habitats within and among creeks, and (2) for secondary production estimates (see below). While L/W relationships may be affected by fragmentation (i.e., individuals of equal length may be different weights in different creeks), using the same regressions for VALENTINE-ROSE ET AL.: HABITAT FRAGMENTATION AND PRODUCTION 867

Table 2. Parameters and sources given for the Von Bertalanfy (VB) growth equation Lage+1 = L∞ −k(t−t0) (1 − e ) where Lage = field estimated lengths, L∞ = asymptotic maximum length, t0 = age at zero length, and k = Brody growth coefficient with all lengths L representing standard length (SL) in millimeters. Size range (min–max) given for fish collected in situ are all within size ranges used for published VB equations. Temperatures given for published VB equations used (°C), while for those experienced in situ during the sampling month were 33–35 °C in upstream areas of fragmented creeks and 25–30 °C in downstream areas of all creeks. Growth rates for schoolmaster in unfragmented tidal creeks fit a linear model, with the equation y = 24.711x + 50.826, where y is standard length in millimeters, and x is age in years (A. Rypel, University of Alabama, unpubl. data).

Species L∞ k t0 °C Source Gray snapper UF 947 .036 −2.5 in situ FD 238 .303 .14 in situ FU 191 .352 .1 in situ Schoolmaster FD-U 416 .061 −1.9 in situ Cubera snapper 877 .125 27.2 Valle et al., 1997 Blue striped grunt 305 .22 −.14 27.2 Valle et al., 1997 334 .3 0 27.2 Appeldoorn, 1992 Sailors choice 349 .24 −.27 27.2 Valle et al., 1997 each creek again likely resulted in conservative estimates of differences between fragmented and unfragmented tidal creeks. Secondary Production Calculation.—For each species, annual secondary production (P) was calculated in each habitat for each age class by subtracting field B estimates from B estimates predicted for each individual one year later by Von Bertalanfy growth parameters:

--kt t0 LLage 1 =-1 e ] g + 3 ] g.

–2 In other words, P = Bage+1 − Bage field. Mean P m was calculated for each habitat, and then each habitat was weighted by the proportion of that habitat type in the survey area (Table 1). Weighted habitat P m–2 values were then summed, giving total creek P m–2 (Appendix 1). Secondary production values for each species in unfragmented tidal creeks were statistically compared to those in fragmented tidal creeks using a paired t-test assuming unequal variance. It is important to note that these P values are intended to be used for comparisons among tidal creeks within this study. Care should be taken in extending these numbers for comparisons with other systems. Due to the limitations of our sampling method, they are not exact whole-ecosystem values, as they do not include in situ growth, survival, or emigration and immigration rates for all species. The use of published growth rates from similar environmental conditions to estimate secondary production is a common practice when in situ growth rates cannot be obtained (Waters, 1977; Benke, 1993; Peterson et al., 2003; Powers et al., 2003). The lack of survival, mortality, and migration estimates surely affects our production estimates, but these estimates are extremely difficult to obtain in coastal and estuarine ecosystems (Adams et al., 2006). Nonetheless, the goal of this study was to employ a standardized protocol across sites allowing for robust among-system comparisons of secondary production estimates. Therefore, these limitations do not compromise the value of the estimates in this study to indicate ecosystem functional response to fragmentation. 868 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Results

Extent of habitat type coverage varied among tidal creeks (Table 1). However, unfragmented tidal creeks tended to be characterized by a combination of mangrove, rock, and seagrass habitats, while fragmented creeks had greater proportion of silt/ sand flats. In areas inundated at low tide, rock was the most common habitat type in White Bight (58%), Buryin Peace (59%), Love Hill (38%), and Conch Sound (67%), mangrove was the most common in Bowen (67%), seagrass in Somerset (74%), and sand/silt flats inD avis (83%) and Man o War (59%) (Table 1). Both natural variability and human impacts affect the extent of each habitat type. Geology and local hydrology drive the extent of exposed rock and can affect areas most conducive for mangrove and seagrass growth. Human impacts can modify natural habitat coverage, e.g., by altering abiotic conditions such as light and salinity, or edaphic conditions through sedimentation, which renders creeks less than optimal for seagrass and/or mangrove growth and fills in important rock habitat. Highest fish biomass for each species at low tide was found in association with mangrove or rock habitat, with the exception of seagrass habitat in Bowen Sound for blue striped grunt (Fig. 1). For example, in Somerset, > 85% of gray snapper, schoolmaster snapper, blue striped grunt, and sailor’s choice was associated with mangrove habitat. In the three creeks with the most disrupted tidal regime (MOW- partially fragmented, CNC-partially fragmented, BOW-completely fragmented), the highest fish biomass was never associated with mangrove habitat (with the exception of schoolmaster snapper in Bowen). In Man o War, fish were never associated with mangrove habitat, likely because no large-mangroves were found along the main channel areas. Fish typically were not associated with sand and seagrass habitat (0%–2% across all creeks) at low tide, with the exception of blue striped grunt in Bowen seagrass habitat (100%), and for schoolmaster (35%) and cubera snapper (25%) in Buryin Peace seagrass. Secondary production of fish biomass was significantly higher in unfragmented tidal creeks than fragmented creeks (t-test P < 0.05 for each species), sometimes as much as an order of magnitude (Fig. 2). For example, for schoolmaster snapper, values ranged from 34 to 107 g m–2 yr–1 in unfragmented creeks, and from 0.2 to 6.5 g m–2 yr–1 in fragmented creeks. Cubera snapper, sailor’s choice, and blue striped grunt were not found in the three creeks with highly disrupted tidal flow (MOW, CNC, and BOW) with the exception of cubera snapper in MOW (4 g m–2 yr–1) and sailor’s choice in BOW (2 g m–2 yr–1). Gray and schoolmaster snapper consistently had higher production values than the other three species (maximum 150 g m–2 yr–1 compared

Table 3. Parameters and sources given for the length-weight (L/W) conversion equation log W = log a + b log L where a and b are species-specific constants, # = number of individuals used to develop in situ L/W regressions, with all lengths L representing standard length (SL) in millimeters and all weights W representing weight in grams. Size range (min–max) given for fish collected in situ.

Species a b min–max (mm) r2 # Gray snapper −4.2751 2.8754 40–290 0.9917 115 Schoolmaster −4.3009 2.9107 50–180 0.9509 100 Cubera snapper −4.0684 2.7574 74–310 0.997 37 Blue striped grunt −4.4 2.9437 63–174 0.989 35 Sailors choice −4.4476 2.9803 50–135 0.9839 31 VALENTINE-ROSE ET AL.: HABITAT FRAGMENTATION AND PRODUCTION 869

Figure 1. Relative contribution (out of 100%) of habitat to biomass (“B”) of (A) gray snapper, (B) schoolmaster, (C) cubera, (D) blue striped grunt, and (E) sailor’s choice. WB = White Bight, SOM = Somerset, BP = Buryin Peace, LH = Love Hill, DAV = Davis, MOW = Man o War, CNC = Conch Sound, BOW = Bowen. U = Unfragmented, PF = Partially fragmented, CF = Completely fragmented (sensu Layman et al., 2004b). to maximum values of 12, 80, and 55 g m–2 yr–1 for cubera, blue striped grunt, and sailor’s choice, respectively), largely a function of abundance. For example, there were a mean of 2.0 gray snapper m–2 and 1.5 schoolmaster snapper m–2 in White Bight, with only 0.08 blue striped grunts m–2 and 0.5 sailor’s choice m–2. Production values in Man o War were more similar to unfragmented creeks than to fragmented creeks for the three snapper species, perhaps a result of restoration efforts 1 yr prior to sampling (see Layman et al., 2004a). Downstream areas of fragmented tidal creeks always had higher production values than upstream areas (Fig. 3). For example, production values for gray snapper in Love Hill downstream areas were 150 g m–2 yr–1 compared to 11 g m–2 yr–1 upstream. Up- stream areas in Love Hill, Davis, and Bowen consistently had production values close or equal to zero (due to low abundance and small size classes) in upstream areas. Within fragmented creeks, blue striped grunt were never found in upstream areas, sailor’s choice were only found upstream in Love Hill, and cubera snapper were only found upstream in Man o War. Man o War had similar upstream and downstream values of schoolmaster snapper production (33 and 25 g m–2 yr–1, respectively), suggesting that recent restoration of hydrologic connectivity (Layman et al., 2004a) may have mitigated effects of previous fragmentation. Lower production and species numbers in upstream areas may be an artifact of shallow waters associated with low- tide surveys. However, these results also occur in downstream areas, suggesting that this pattern is to some extent a consequence of fragmentation. 870 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Figure 2. Estimated annual production values (g m–2 yr–1) for (A) gray snapper, (B) schoolmaster, (C) cubera, (D) blue striped grunt, and (E) sailor’s choice. Note different y-axis values for each species. WB = White Bight, SOM = Somerset, BP = Buryin Peace, LH = Love Hill, DAV = Davis, MOW = Man o War, CNC = Conch Sound, BOW = Bowen. U = Unfragmented, PF = Partially fragmented, CF = Completely fragmented (sensu Layman et al., 2004b).

Discussion

Our study suggests that fragmentation of tidal creeks significantly reduces secondary production of five economically important fish species in the Bahamas. Lower production values in fragmented systems resulted from a combination of fewer species, lower densities of those species, and smaller size classes of fishes. Secondary production is directly related to food quality and quantity (Benke et al., 1988; Benke, 1993; Edgar and Shaw, 2004a,b,c) which is likely reduced in fragmented creeks (Val- entine-Rose et al. in press). Higher salinity and temperature in upstream areas of fragmented tidal creeks (Valentine-Rose et al. in press) may decrease productivity by increasing energy costs in juvenile fishes (Wuenschel et al., 2004, 2005), as well as by influencing their survival and viability G( reen and Fisher, 2004). Recruitment of post-settlement fishes may be lower in fragmented creeks due to reduction in supply rates and connectivity to larval sources (Gaines and Roughgarden, 1985; Cowen et al., 2006) or reduced quantity and quality of preferred habitat (Carr, 1994; Jordan et al., 1998; Adams and Ebersole, 2004; Mumby et al., 2004; Dorenbosch et al., 2005). In our study, highest fish production typically was associated with creeks with the most available mangrove and rock habitat, suggesting habitat complexity likely is a major factor affecting the levels of secondary production. The correlation between habitat complexity and fish production also has been observed in freshwater stream studies (Benke et al., 1988). VALENTINE-ROSE ET AL.: HABITAT FRAGMENTATION AND PRODUCTION 871

Figure 3. Estimated annual production values (g m–2 yr–1) for (A) gray snapper, (B) schoolmaster, (C) cubera, (D) blue striped grunt, and (E) sailor’s choice in downstream (filled bars) and upstream (unfilled bars) areas of fragmented tidal creeks. LH = Love Hill, DAV = Davis, MOW = Man o War, CNC = Conch Sound, BOW = Bowen. Production values in fragmented tidal creeks were typically lower, often substantially so, than in unfragmented tidal creeks. Nonetheless, differences between fragmented and unfragmented creeks may have been more pronounced had we not used the same growth rates in all creeks for three of the five species. It is likely that many fish species’ growth rates are slower in fragmented creeks due to loss of preferential food sources (Cocheret de la Morinière, 2004b; Kieckbusch et al., 2004; Nemerson and Able, 2004), inadequate refuge habitat (Cocheret de la Morinière et al., 2004a), or adverse environmental conditions (Green and Fisher, 2004; Wuenschel et al., 2004, 2005). Furthermore, fish in fragmented creeks were generally smaller and may have lower survival rates (Fuiman and Werner, 2002). In sum, our methodology did not employ in situ growth rates for all species or mortality estimates for any species, and we believe these measures would further the discrepancy between secondary production estimates for fragmented and unfragmented tidal creeks. Because secondary production encompasses three of the four criteria defining a nursery, i.e., density, growth, and survival (Benke, 1993; Beck et al., 2001; Dahlgren et al. 2006), secondary production is a first step in assessing nursery function of estuaries. Complemented with direct evidence of export of juveniles from fragmented estuaries to off-shore adult habitats (i.e., neighboring back reef and coral reef ecosystems D( ahl- gren and Marr, 2004; Mumby et al., 2004; Dorenbosch et al., 2005), nursery function of these systems could be quantitatively estimated. The species in this study are believed to experience ontogenetic habitat shifts from juvenile to adult habitats (Cocheret de la Morinière et al., 2003a,b, 2004; Serafy et al., 2003), and therefore decreases in productivity in important juvenile nursery habitat through fragmentation most likely results 872 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 in decreased contribution of individuals to offshore adult habitats (Mumby et al., 2004; Dorenbosch et al., 2005). Secondary production values may be helpful in putting an economic value on ecosystem services, which is often crucial for prioritizing management and conservation efforts (Costanza et al., 1997). Based on economic value of commercially important species, secondary production estimates can be directly translated into a dollar value of one service (i.e., production of consumable animal protein) provided by these creek systems. For example, an extensive survey of local fisherman and restaurants on An- dros Island revealed the average price for the five focal species sold by fishermen is $18/kg for all species 15 cm or longer, while Androsian restaurants charge an average of $7/fish 15 cm or longer. Applying these numbers to fish secondary production estimates from our study creeks, hydrologic fragmentation (even of our small creeks) can result in a loss of up to $9000 of value (for the five focal species). Differences between estimates of secondary production and loss of economic worth of fish biomass between unfragmented and fragmented creeks may be even more pronounced in larger creek systems. In our study, we examined differences in secondary production between small creeks (< 2 ha). However, some tidal creek systems in the Bahamas are thousands of hectares in size, such as Fresh Creek, Andros (~4100 ha). Larger creek systems have a higher abundance of fish, a higher number of species (L. Val- entine, University of Alabama, unpubl. data), and generally have larger individuals (pers. obs.), all of which would serve to increase secondary production estimates and economic value. Our study systems appear to be ideal “natural” systems in which to work, because they are small enough to thoroughly sample in a quantitative manner and they can be physically manipulated (e.g., restored) at the ecosystem-scale (e.g., Gribsholt et al., 2005). Therefore, they can serve as natural experimental units with which to make comparisons, as well as to scale-up responses to much larger wetland systems throughout the Bahamas and Caribbean. In conclusion, we provide a survey methodology to assess the effects of anthropogenic stress on an ecosystem function: secondary production. This is the first attempt to estimate secondary production for mangrove fishes (Faunce and Serafy, 2006). We suggest secondary production may be an appropriate measure of ecosystem functional responses to other anthropogenic stresses, as it is a single, composite value that incorporates many factors within an ecosystem. Other studies can build upon this framework to provide more accurate production estimates, e.g., by determining in situ growth for all species. Although in almost all situations in situ survival and migration will be difficult to estimate, secondary production estimates without these measures still provide for a powerful tool to make among-system comparisons. Comparative studies of secondary production may provide insight into patterns of ecosystem functional response following environmental change, and may also be useful in applying monetary values to ecosystem responses to anthropogenic stress or habitat restoration.

Acknowledgments

We owe sincere thanks to M. Blackwell, Bahamas Environmental Research Center, and the College of the Bahamas for providing a research facility and logistical support. We would like to thank J. Grabowski, R. Appledoorn, and J. Bohnsack for advice in various steps of this work. We would also like to thank the many Androsians that assisted with field work, particularly C. “Tutu” Bains and M. “Chili” Gibson. This research was funded by the National Science VALENTINE-ROSE ET AL.: HABITAT FRAGMENTATION AND PRODUCTION 873

Foundation (DGE 0319143), an RJKOSE grant from The Nature Conservancy, and The Acorn Alcinda Foundation. Work was conducted under permits granted by the Bahamian Depart- ment of Fisheries.

Literature Cited

Adams, A. J. and J. P. Ebersole. 2004. Processes influencing recruitment inferred from distributions of coral reef fishes. Bull. Mar. Sci. 75: 153–174. ______, C. P. Dahlgren, G. T. Kellison, M. S. Kendall, C. A. Layman, J. A. Ley, I. Nagelkerken, and J. E. Serafy. 2006. Nursery function of tropical back-reef systems. Mar. Ecol. Prog. Ser. 318: 287–301. Appeldoorn, R. S. 1992. Interspecific relationships between growth parameters, with application to Haemulid fishes. Proc. 7th Int. Coral Reef Symp., Guam 2: 899–904. Beck, M. W., K. L. Heck, Jr., K. W. Able, D. L. Childers, D. B. Eggleston, B. W. Gillanders, B. Halpern, C. G. Hays, K. Hoshino, T. J. Minello, R. J. Orth, P. F. Sheridan, and M. P. Wein- stein. 2001. The identification, conservation, and management of estuarine and marine nurseries for fish and invertebrates. Bioscience 51: 633–641. Benke, A. C. 1993. Edgardo Baldi Memorial Lecture: Concepts and patterns of invertebrate production in running waters. Verh. International Verein. Limnol. 25: 15–38. ______, C. A. S. Hall, C. P. Hawkins, R. H. Lowe-McConnell, J. A. Stanford, K. Suberkropp, and J. V. Ward. 1988. Bioenergetic considerations in the analysis of stream ecosystems. J. N. Am. Benthol. Soc. 7: 480–502. Benstead, J. P., J. G. March, C. M. Pringle, and F. N. Scatena. 1999. Effects of a low-head dam and water abstraction on migratory tropical stream biota. Ecol. Appl. 9: 656–668. Blaber, S. J. M., D. T. Brewer, and J. P. Salini. 1989. Species composition and biomasses of fishes in different habitats of a tropical northern Australian estuary: their occurrence in the adjoining sea and estuarine dependence. Estuar. Coast. Shelf Sci. 29: 509–531. Brock, R. E. 1954. A preliminary report on a method of estimating reef fish populations. J. Wildl. Manage. 18: 297–308. Burdick, D. M., M. Dionne, R. M. J. Boumans, and F. T. Short. 1997. Ecological responses to tidal restorations of two northern New England salt marshes. Wetl. Ecol. Manage. 4: 129–144. Carr, M. H. 1994. Effects of macroalgal dynamics on recruitment of a temperate reef fish. Ecol- ogy 75: 1320–1333. Chapin III, F. S., B. H. Walker, R. J. Hobbs, D. U. Hooper, J. H. Lawton, O. E. Sala, and D. Tilman. 1997. Biotic control over the functioning of ecosystems. Science 277: 500–504. ______, O. E. Sala, I. C. Burke, J. P. Grime, D. U. Hooper, W. K. Lauenroth, A. Lombard, H. A. Mooney, A. R. Mosier, S. Naeem, S. W. Pacala, J. Roy, W. L. Steffen, and D. Tilman. 1998. Ecosystem consequences of changing biodiversity. Bioscience 48: 45–52. Cocheret de la Morinière, E., B. J. A. Pollux, I. Nagelkerken, and G. van der Velde. 2003a. Diet shifts of Caribbean grunts (Haemulidae) and snappers (Lutjanidae) and the relation with nursery-to-coral reef migrations. Estuar. Coast. Shelf Sci. 57: 1–11. ______, ______, ______, and ______. 2003b. Post-settlement life cycle migration patterns and habitat preference of coral reef fish that use seagrass and mangrove habitats as nurseries. Estuar. Coast. Shelf Sci 55: 309–321. ______, B. J. A. Pollux, I. Nagelkerken, M. A. Hemminga, A. H. L. Huiskes, and G. van der Velde. 2004. Ontogenetic dietary changes of coral reef fishes in the mangrove-seagrass-reef continuum: stable isotopes and gut-content analysis? Mar. Ecol. Prog. Ser. 246: 279–289. Costanza, R. R. d’Arge, R. de Groot, S. Farber, M. Grasso, B. Hannon, K. Limburg, S. Naeem, R. V. ONeill, J. Paruelo et al. 1997. The value of the world’s ecosystem services and natural capital. Nature 387: 253–260. Cowen, R. K., C. B. Paris, and A. Srinivasan. 2006. Scaling of connectivity in marine populations. Science 311: 522–527. 874 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Dahlgren, C. P. and J. Marr. 2004. Back reef systems: important but overlooked components of tropical marine ecosystems. Bull. Mar. Sci. 75: 145–152. ______, G. T. Kellison, A. J. Adams, B. M. Gillanders, M. S. Kendall, C. A. Layman, J. A. Ley, I. Nagelkerken, and J. E. Serafy. 2006. Marine nurseries and effective juvenile habitats: concepts and applications. Mar. Ecol. Prog. Ser. 312: 291–295. Dorensbosch, M., M. G. G. Grol, M. J. A. Christianen, I. Nagelkerken, and G. van der Velde. 2005. Indo-Pacific seagrass beds and mangroves contribute to fish density and diversity on adjacent coral reefs. Mar. Ecol. Prog. Ser. 302: 63–76. Dynesius, M. and C. Nilsson. 1994. Fragmentation and flow regulation of river systems in the northern third of the world. Science 266: 753–762. Edgar, G. J. 1995a. The production and trophic ecology of shallow-water fish assemblages in southern Australia II. Diets of fishes and trophic relationships between fishes and benthos at Western Port, Virginia. J. Exp. Mar. Biol. Ecol. 194: 83–106. ______. 1995b. The production and trophic ecology of shallow-water fish assemblages in southern Australia III. General relationships between sediments, seagrasses, invertebrates, and fishes. J. Exp. Mar. Biol. Ecol. 194: 53–81. ______and C. Shaw. 1995c. The production and trophic ecology of shallow-water fish assemblages in southern Australia I. Species richness, size-structure, and production of fishes in Western Port, Virginia. J. Exp. Mar. Biol. Ecol. 194: 53–81. Eertman, R. H. M., B. A. Kornman, E. Stikvoort, and H. Verbeek. 2002. Restoration of the Sie- perda Tidal Marsh in the Sheldt Estuary, the Netherlands. Restor. Ecol. 10: 438–449. Eggleston, D. B., C. P. Dahlgren, and E. G. Johnson. 2004. Fish density, diversity, and size-structure within multiple back reef habitats of Key West National Wildlife Refuge. Bull. Mar. Sci. 75: 175–204. Fahrig, L. 2002. Effect of habitat fragmentation on the extinction threshold: a synthesis. Ecol. Appl. 12: 346–353. ______. 2003. Effects of habitat fragmentation on biodiversity. Annu. Rev. Ecol. Syst. 34: 487–515. Faunce, C. H. and J. E. Serafy. 2006. Mangroves as fish habitat: 50 years of field studies. Mar. Ecol. Prog. Ser. 318: 1–18. Forman, R. T. T. and L. E. Alexander. 1998. Roads and their major ecological effects. Annu. Rev. Ecol. Syst. 29: 207–231. Fuiman, L. A. and R. G. Werner. 2002. Fishery science: the unique contributions of early life stages. Blackwell Science, Malden. Gaines, S. and J. Roughgarden. 1985. Larval settlement rate: a leading determinant of structure in an ecological community of the marine intertidal zone. Proc. Nat. Acad. Sci. 82: 3707–3711. Green, B. S. and R. Fisher. 2004. Temperature influences swimming speed, growth, and larval duration in coral reef fish larvae. J. Exp. Mar. Biol. Ecol. 299: 115–132. Gribsholt, B., H. T. S. Boschker, E. Struyf, M. Andersson, A. Tramper, L. D. Brabandere, S. V. Damme, N. Brion, P. Meire, F. Dehairs, J. J. Middelburg, and C. H. R. Heip. 2005. Nitro- gen processing in a tidal freshwater marsh: A whole ecosystem 15N labeling study. Limnol. Oceanogr. 50: 1945–1959. Gust, N., J. H. Choat, and M. I. McCormick. 2001. Spatial variability in reef fish distribution, abundance, size, and biomass: a multi-scale analysis. Mar. Ecol. Prog. Ser. 214: 237–251. Heywood, V. H. and R. T. Watson. 1995. Global Biodiversity Assessment. UNEP, Cambridge University Press, New York. 1140 p. Holmquist, J. G., J. M. Schmidt-Gengenback, and B. B. Yoshioka. 1998. High dams and marine- freshwater linkages: effects on native and introduced fauna in the Caribbean. Conserv. Biol. 12: 621–630. Huryn, A. D. 1996. An appraisal of the Allen paradox in a New Zealand trout stream. Limnol. Oceanogr. 41: 243–252. VALENTINE-ROSE ET AL.: HABITAT FRAGMENTATION AND PRODUCTION 875

______. 1998. Ecosystem-level evidence for top-down and bottom-up control of production in a grassland stream system. Oecologia 115: 173–183. Jordan, F., K. J. Babbitt, and C. C. McIvor. 1998. Seasonal variation in habitat use by marsh fishes. Ecol. Freshw. Fish. 7: 159–166. Kieckbusch, D. K., M. S. Koch, J. E. Serafy, and W. T. Anderson. 2004. Trophic linkages among primary producers and consumers in fringing mangroves of subtropical lagoons. Bull. Mar. Sci. 74: 271–285. Kulbikci, M. 1998. How the acquired behavior of commercial reef fishes may influence the results obtained from visual censuses. J. Exp. Mar. Biol. Ecol. 222: 11–30. Layman, C. A., D. A. Arrington, and M. A. Blackwell. 2004a. Community-based restoration of an Andros Island (Bahamas) estuary. Ecol. Restor. 23: 58–59. ______, ______, R. B. Langerhans, and B. R. Silliman. 2004b. Degree of fragmentation affects fish assemblage structure in Andros Island (Bahamas) estuaries. Caribb. J. Sci. 40: 234–244. Mallin, M. A. and A. J. Lewitus. 2004. The importance of tidal creek ecosystems. J. Exp. Mar. Biol. Ecol. 298: 145–149. March, J. G., J. P. Benstead, C. M. Pringle, and F. N. Scatena. 2003. Damming tropical island streams: problems, solutions, and alternatives. BioScience 53: 1069–1078. Mateo, I. and W. J. Tobias. 2001. Distribution of shallow water coral reef fishes on the northeast coast of St. Croix, USVI. Carib. J. Sci. 37: 210–226. McCay, D. P. F. and J. J. Rowe. 2003. Habitat restoration as mitigation for lost production at multiple trophic levels. Mar. Ecol. Prog. Ser. 264: 233–247. Mumby, P. J., A. J. Edwards, J. E. Aria-Gonzalez, K. C. Lindeman, P. G. Blackwell, A. Gall, M. I. Gorczynska, A. R. Harborne, C. L. Pescod, H. Renken, C. C. C. Wabnitz, and G. Llewellyn. 2004. Mangroves enhance the biomass of coral reef fish communities in the Caribbean. Nature 427: 533–536. Naeem, S., L. J. Thompson, S.P . Lawler, J. H. Lawton, and R. M. Woodfin. 1994.D eclining biodiversity can alter the performance of ecosystems. Nature 368: 734–736. Nemerson, D. M. and K. W. Able. 2004. Spatial patterns in diet and distribution of juveniles of four fish species in Delaware Bay marsh creeks: factors influencing fish abundance. Mar. Ecol. Prog. Ser. 276: 249–262. Peterson, C. H., J. H. Grabowski, and S. P. Powers. 2003. Estimated enhancement of fish production resulting from restoring oyster reef habitat: quantitative valuation. Mar. Ecol. Prog. Ser. 264: 249–264. Powers, S. P., J. H. Grabowski, C. H. Peterson, and W. J. Lindberg. 2003. Estimating enhancement of fish production by offshore artificial reefs: uncertainty exhibited by divergent scenarios. Mar. Ecol. Prog. Ser. 264: 265–277. Pringle, C. 2001. Hydrologic connectivity and the management of biological reserves: a global perspective. Ecol. Appl. 11: 981–998. ______. 2003a. The need for a more predictive understanding of hydrologic connectivity. Aquat. Conserv. Mar. Freshw. Ecos. 13: 467–471. ______. 2003b. What is hydrologic connectivity and why is it ecologically important. Hydrol. Proc. 17: 2685–2689. Raposa, K. B. and C. T. Roman. 2003. Using gradients in tidal restriction to evaluate nekton community responses to salt marsh restoration. Estuaries 26: 98–105. Ray, G. C. 2005. Connectivities of estuarine fishes to the coastal realm. Estuar. Coast. Shelf Sci. 64: 18–32. Roman, C. T., W. A. Niering, and R. S. Warren. 1984. Salt marsh vegetation change in response to tidal restriction. Environ. Manage. 8: 141–150. ______, K. B. Raposa, S. C. Adamowicz, and J. G. Catena. 2002. Quantifying vegetation and nekton response to tidal restoration of a New England salt marsh. Restor. Ecol. 10: 450–460. 876 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Saunders, d. A., R. J. Hobbs, and C. R. Margules. 1991. Biological consequences of ecosystem fragmentation: a review. Conserv. Biol. 5: 18–32. Schmitt, E. F. and k. M. Sullivan. 1996. Analysis of a volunteer method for collecting fi sh presence and abundance data in the Florida keys. Bull. Mar. Sci. 59: 404–416. Serafy, J. E., C. H. Faunce, and J. J. Lorenz. 2003. Mangrove shoreline fi shes of Biscayne Bay, Florida. Bull. Mar. Sci. 72: 161–180. Sheaves, M. J. 1992. patterns of distribution and abundance of fi shes in diff erent habitats of a mangrove-lined tropical estuary, as determined by fi sh trapping. Aust. J. Mar. Freshw. Res. 43: 1461–1479. Sneddon, L. U. 2003. Th e bold and the shy: individual diff erences in rainbow trout. J. Fish Biol. 62: 971–976. Tanner, C. d., J. R. Cordell, J. Rubey, and L. M. Tear. 2002. Restoration of freshwater intertidal habitat functions at Spencer Island, Everett, Washington Restor. Ecol. 10: 564–576. Taylor, C. M., T. L. Holder, R. A. Fiorillo, L. R. Williams, R. B. Th omas, and M. L. Warren, Jr. 2006. distribution, abundance, and diversity of stream fi shes under variable environmental conditions. Can. J. Fish. Aquat. Sci. 63: 43–54. Th ompson, A. A. and B. d. Mapstone. 2002. Intra-versus inter-annual variation in counts of reef fi shes and interpretations of long-term monitoring studies. Mar. Ecol. prog. Ser. 232: 247–257. Tilman, d. 1999. Th e ecological consequences of changes in biodiversity: a search for general principles. Ecology 80: 1455–1474. Trombulak, S. C. and C. A. Frissell. 1999. Review of ecological eff ects of roads on terrestrial and aquatic communities. Conserv. Biol. 14: 18–30. Valentine-Rose, L., d. A. Arrington, J. A. Cherry, J. Culp, C. Layman, k. perez, and J. pollock. In press. Floral, faunal, and physiochemical diff erences between fragmented and unfragmented Bahamian tidal creeks. Wetlands. Valle, S. V., J. p. garcía-Arteaga, and R. Claro. 1997. growth parameters of marine fi shes in Cuban waters. Naga ICLARM Q. 20: 34–37. Vitousek, p. M., H. A. Mooney, J. Lubchenco, and J. M. Melillo. 1997. Human domination of earth’s ecosystems. Science 277: 494–499. Warren, R. S., p. E. Fell, R. Rozsa, A. H. Brawley, A. C. Orsted, E. T. Olson, V. Swamy, and W. A. Niering. 2002. Salt Marsh Restoration in Connecticut: 20 years of science and management. Restor. Ecol. 10: 497–513. Waters, T. F. 1977. Secondary production in inland waters. Adv. Ecol. Res. 10: 91–164. Wuenschel, M. J., A. R. Jugovich, and J. A. Hare. 2004. Eff ect of temperature and salinity on the energetics of juvenile gray snapper: implications for nursery habitat value. J. Exp. Mar. Biol. Ecol. 312: 333–347. ______, ______, and ______. 2005. Metabolic response of juvenile gray snapper to temperature and salinity: physiological cost of diff erent environments. J. Exp. Mar. Biol. Ecol. 321: 145–154.

Addresses: (L.V.-R., A.L.R.) Department of Biological Sciences, University of Alabama, Box 870206, Tuscaloosa, Alabama 35487. (C.A.L.) Marine Biology Program, Department of Bio- logical Sciences, Florida International University, 3000 NE 151st Street, North Miami, Flor- ida 33181. (d.A.A.) Director of Water Resources, Loxahatchee River District, 2500 Jupiter Park Drive, Jupiter, Florida 33458-8964. Corresponding Author: (L.V.-R.) E-mail: . VALENTINE-ROSE ET AL.: HABITAT FRAGMENTATION AND PRODUCTION 877

Appendix 1. Example of calculation of annual production: schoolmaster snapper in White Bight tidal creek. N = density m–2, B = biomass m–2, P = secondary production, and P (weighted) = production multiplied by percentage of habitat contribution for each habitat. Equations can be found in the “Methods” section. Habitat area contributions can be found in Table 1. All necessary parameters for the calculations (length-weight, length-age) appear in Table 2.

P P P (weighted) –2 –2 –1 –2 –1 White Bight Rock Age N Bage field B age + 1 g 25m yr g m yr g m yr 0 24 400.58 0 1 18 1,563.75 3,078.7 2 7 1,745.04 7,593.75 3 1 811.44 0 4 0 0 930.69 Total 4,520.81 11,603.35 7,082.33 283.29 166.57

Mangrove Age N Bage field B age + 1 0 41 266.68 0 1 8 1,320.72 5,313.18 2 22 5,484.42 11,926.4 3 20 12,887.3 6,585.5 4 0 0 9,306.9 Total 19,959.12 33,131.98 13,172.9 526.91 96.42

Seagrass Age N Bage field B age + 1 0 3 12.10 0 1 0 0 244.76 Total 12.10 244.76 232.66 58.17 0.23

Sand Age N Bage field B age + 1 0 1 4.41 0 1 0 0 85.19 Total 4.41 85.19 80.78 3.23 0.77 Total 293.99 BULLETIN OF MARINE SCIENCE, 80(3): 879–890, 2007

Mangrove removal in the Belize cays: effects on mangrove-associated fish assemblages in the intertidal and subtidal

D. Scott Taylor, Eric A. Reyier, William P. Davis, and Carole C. McIvor

abstract We investigated the effects of mangrove cutting on fish assemblages in Twin Cays, Belize, in two habitat types. We conducted visual censuses at two sites in adjoining undisturbed/disturbed (30%–70% of shoreline fringe removed) sub-tidal fringing Rhizophora mangle Linnaeus, 1753. Observers recorded significantly more species and individuals in undisturbed sites, especially among smaller, schooling species (e.g., atherinids, clupeids), where densities were up to 200 times greater in undisturbed habitat. Multivariate analyses showed distinct species assemblages between habitats at both sites. In addition, extensive trapping with wire minnow traps within the intertidal zone in both undisturbed and disturbed fringing and transition (landward) mangrove forests was conducted. Catch rates were low: 638 individuals from 24 species over 563 trap-nights. Trap data, however, indicated that mangrove disturbance had minimal effect on species composition in either forest type (fringe/ transition). Different results from the two methods (and habitat types) may be explained by two factors: (1) a larger and more detectable species pool in the subtidal habitat, with visual “access” to all species, and (2) the selective nature of trapping. Our data indicate that even partial clearing of shoreline and more landward mangroves can have a significant impact on local fish assemblages.

Among subtropical and tropical coastal marine habitats, mangroves are regarded as unusually biologically diverse and quite productive. In addition, the importance of mangroves in shoreline stabilization and protection of the terrestrial interface, and the linkages to adjacent marine systems (e.g., seagrasses/coral reefs) was recognized long ago (Lugo and Snedaker, 1974). Within the perceived resource values of the mangal is that of a nursery area for juvenile coral reef fishes. The literature is replete with discussions of this nursery concept, and the debate is ongoing. Beck et al. (2001) defined general guidelines for describing marine nursery areas, which included (1) greater densities of juveniles, (2) faster growth rates, (3) enhanced juvenile survival, and (4) successful movement to adult habitats. Nagelkerken et al. (2000) provided a summary of the various aspects of the nursery hypotheses specific to the mangal, the primary bases of which are predator avoidance, food abundance, and larval settlement. Contributing to these aspects are (1) the high structural complexity of the mangal, which both protects against predators and intercepts entrained fish larvae, (2) relatively remote locations from reef or offshore waters, reducing proximity to predators, (3) enhanced turbidity, reducing predation efficiency, (4) high autochthonous food production, and (5) areal extent that could intercept planktonic fish larvae at greater frequency than the reef itself. Many early assumptions of nursery value for both mangroves and seagrasses are now being challenged, since they were based on qualitative observations, with no separation of information on fish size or juvenile/adult ratios (Nagelkerken et al., 2000). Faunce and Serafy (2006) indicated that there is evidence that many fish species use the mangal as a diurnal refuge, and thus some of the fisheries production assumed to have originated there may have other sources.

The nursery role of mangroves may be strongest at broad scales. For example, the areal extent of mangroves is assumed to correlate with secondary production in the commercial fish catch (Manson et al., 2005). Mumby et al. (2004) estimated fish biomass at three Caribbean atolls with no or very limited mangrove cover and compared them with “mangrove-rich” systems (defined as having ≥ 70 km2 of fringing Rhizophora mangle Linnaeus, 1753 within an area of 200 km2). Fish biomass was much greater (up to 25 times for some species) on the reef systems with abundant, adjacent mangroves. The study concluded that mangroves clearly offer an intermediate nursery area that provides a transition between seagrass beds and patch reefs. The data of Mumby et al. (2004) (comparing intact vs naturally “barren” areas) provide, at best, an indication of the consequences on fisheries of human-caused mangrove destruction. The rate of worldwide mangrove deforestation (now at 35% globally) is highest in the Americas, where it is now estimated at 2251 km2 yr–1 (Va- liela et al., 2001). With destruction of the mangal so common, it is not surprising that studies assessing the effects of human changes are abundant, but most of this work has focused on the Pacific, and much of the emphasis has been on invertebrates and the physical changes following mangrove alteration (water chemistry/sediment characteristics; Manson et al., 2005). Surprisingly, studies examining the effects of alteration on fishes directly associated with the mangal are sparse. We note only one study (Ellis, 2003) that has examined mangrove canopy trimming (Rookery Bay, FL, USA) and the effect on fish assemblages. No effect on overall assemblage structure was found with moderate trimming and the author concluded that alterations to the habitat may have to be significant before changes occur, as variability between samples was very great in the sites studied. In addition, there are few data on how fishes respond to natural environmental variability within the mangal (Faunce and Serafy, 2006). We believe that to slow the rate of mangrove deforestation, empirical data are needed on the effects of mangrove shoreline alteration on associated fish assemblages. Here, we examine the effects of mangrove removal on fishes inhabiting the intertidal and subtidal mangroves on one of the Belize cays.

Methods

Study Location.—The study site was located on Twin Cays on the Belize barrier reef in Central America. Twin Cays is a complex of mangrove islands 1.5 km west of the central barrier reef, about 12 km offshore fromD angriga, Belize (16°50´N, 88°06´W). Twin Cays consists of two main cays (East and West) bisected by a central channel (Fig. 1). The vegetation of the Cays has been previously described (summarized in Feller et al., 2002), but the shorelines are dominated by what is termed the “fringe” mangrove forest, consisting of R. mangle in a stand 5–20 m wide and 4–6 m tall. Tree height decreases landward into the “transition” zone, an area characterized by the increasing presence of Avicennia germinans (Linnaeus) Stearn, 1958 and Laguncularia racemosa (Linnaeus) Gaertn.f., although R. mangle is also present. Trees are shorter (2–4 m) in the transition zone which varies from 5–30 m wide, and elevations are the highest along a gradient across the island. On Twin Cays, as well as other Beliz- ean cays (Woodroffe, 1995), we have observed clearing or partial cutting of mangroves over a two-decade period, generally to allow construction of small buildings on the islands. The typical pattern of clearing is to cut a significant portion of the shorelineR. mangle fringe and then clear-cut the transition zone, except for a few larger A. germinans (D.S.T., pers. obs.) We chose two sites, one each on East and West Twin Cays. At both sites, a portion of the shoreline fringe about 300 m in length had been partially cleared by a “thinning” of the shore- TAYLOR ET AL.: EFFECTS OF MANGROVE REMOVAL ON FISHES IN BELIZE 881

Figure 1. Study locations on Twin Cays, Belize (Adapted from M. K. Ryan). line fringe (estimated at 30%–70%), as well as almost total clearing of the landward transition zone (as above). Each cleared site (hereafter termed “disturbed”) directly adjoins an uncleared (“undisturbed”) area of otherwise identical habitat. This clearing activity results in much greater light penetration, less structural complexity, and increase in wave action, although we did not quantify these parameters. Water depth outside the R. mangle fringe at all sites varied from 1–2 m. We assessed mangrove fish assemblages at both sites with the following methods: (1) Visual surveys: Along adjacent portions of disturbed and undisturbed fringe shoreline on both East and West Cays, we conducted visual censuses of fishes. Four observers (snorkelers) were positioned about 75 m apart along the shoreline of a disturbed or undisturbed portion of each site. Each observer marked a beginning point in the fringe and slowly swam for 5 min along the shoreline, identifying and enumerating all fishes observed within the prop-roots and beneath the canopy extending beyond the prop-root zone. No fishes outside the canopy or prop-root zone were included. Most fishes were identified to species. Smaller schooling taxa (e.g., engraulids, clupeids), which were often quite abundant, were estimated in number and classified only to family, or on occasion, as unidentified. After 5 min, a marker was placed and another 5 min swim begun, for a total of four 5 min observation periods. Only very rarely did these observations overlap those of an adjacent observer. After all four observations were complete, each observer would immediately move to the adjacent undisturbed/ 882 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 disturbed site and repeat the process. Individual placement of observers along the shoreline was randomized and direction of travel was reversed in subsequent census events. When both disturbed and undisturbed site surveys were complete, swimmers measured the distance surveyed in each 5 min swim with a tape measure, following the contours of the shoreline. For each observer, all four sequential swims were considered a transect. We completed five total sets of visual surveys over two week periods within two different years (Oct.–Nov. 2003 n = 2 surveys; Sept. 2004 n = 3 surveys), for a total of 80 individual transects with a mean length (± 1 SD) of 85.6 ± 2.6 m. (2) Fish trapping: We placed Gee™ wire minnow traps (6 mm mesh) within both disturbed and undisturbed intertidal fringe and transition areas on the East Cay over 5 d periods during three different years (2003, 2004, 2005). Since the transition is completely within the intertidal zone during seasonal periods of higher tides, depending upon tidal cycles, some traps placed in the transition on occasion did not flood. For example, in 2005, the transition did not flood at all, and no traps could be deployed here. Traps were baited with cut-fish bait and deployed for 24 hrs. All collected fishes were identified to species and released, except for specimens needed for stable isotope analysis and voucher specimens. Overall, we deployed 563 traps, with the following breakdown by year: 2003 (Oct–Nov: 75 disturbed, 104 undisturbed). 2004 (Sept: 96 disturbed, 96 undisturbed), and 2005 (April–May: 192 disturbed, 0 undisturbed). In conjunction with trap deployments, during 2003 and 2004 we established two 120 m linear transects through adjacent disturbed/undisturbed transition areas on the East Cay to ascertain possible water quality differences resulting from mangrove clearing. Transects ran parallel to the shoreline and to each other. The first of the two transects was located at the beginning of the transition, about 25 m from the open water of the shoreline, and the second was 30 m landward of this. Both were located one-half within each of the disturbed/ undisturbed areas. Five sampling stations were located (~12 m apart) along each of the 60 m segments in respective zones, for total of 10 fixed, equally spaced stations. Temperature, dissolved oxygen (DO), and salinity were measured at each station with a YSI 6920 sonde. Data were collected on eight different days at differing hours. Data Analysis.—To assess spatial and temporal differences in fish density (individuals 10 m–1 of shoreline), visual census transect data were analyzed with a 3-way ANOVA with site (East/West Cay), habitat (undisturbed/disturbed), and year (2003–2004) as main effects. Prior to analysis, density means were log (1+x) transformed to meet parametric assumptions of normality and equal variance. Separate ANOVAs were run for total fish density, schooling fish density, and non-schooling fish density. Abundance data from minnow traps (expressed as catch-per-unit-effort: CPUE) were not compared statistically because of generally low catch rates that necessitated data pooling by year for each habitat type. To examine differences in the species community composition, non-metric multi-dimensional scaling (MDS) was employed to compare species assemblages between undisturbed and disturbed mangrove habitats and between years for each cay. Visual census and minnow trap data were analyzed separately. Density estimates were square-root transformed to downweight the influence of numerically dominant species (allowing less abundant species to contribute more fully to discrimination between samples) and a sample similarity matrix was created using the Bray-Curtis similarity coefficient (Bray and Curtis, 1957). Sample ordinations were then constructed with MDS, where interpoint distances are proportional to faunal similarity. These ordinations were followed by a two-way (visual data) or one-way (trap data) crossed analysis of similarities (ANOSIM) tests to more formally test the hypothesis that community differences existed between habitat types and years. Finally, a similarity percentages (SIMPER) routine was performed on the visual data to determine which species contributed most to observed differences (Clarke, 1993). All multivariate procedures were run with Primer 6.0 software. In addition, water-quality data were analyzed with a 3-way ANOVA with year, habitat, and disturbance as main effects. TAYLOR ET AL.: EFFECTS OF MANGROVE REMOVAL ON FISHES IN BELIZE 883

Results

Visual Census/Minnow Traps.—In the visual surveys of the mangrove fringe, we documented a total of 62 taxa. At both locations (East Cay, West Cay), more species were tabulated at the undisturbed site compared to the disturbed site. Forty-seven taxa were recorded at the East Cay undisturbed site, with 39 at the corresponding disturbed site. At the West Cay site, the ichthyofauna at the disturbed site was more depauperate, with 49 and 34 species, respectively. However, for many taxa, density differences between undisturbed and disturbed sites were pronounced, especially for smaller schooling species (Table 1). Significant differences in total fish density and schooling fish density were observed between habitat type (undisturbed/disturbed) and year (P < 0.001), but not across sites: Non-schooling fish density and taxa 10 m–1 shoreline differed by habitat and site (East/West Cay) (P < 0.001), but was similar across years. On a species basis, the differences in two grunt and two snapper species are particularly striking (Table 2). Despite different abundances between locations (East Cay, West Cay), Haemulon sciurus, Haemulon flavolineatum, Lutjanus apodus, and Lutjanus griseus were consistently found in greater abundance in undisturbed sites. MDS analysis showed differences in the fish assemblage at the East Cay site over both years (Fig. 2). The ANOSIM demonstrated significant differences in both habitat (undisturbed/disturbed: R = 0.40; P < 0.001) and year (2003/2004: R = 0.515; P < 0.001). The SIMPER analysis showed the five taxa contributing most to the observed difference by habitat and year. While schooling species contributed greatly to yearly differences, this could have been an artifact of the random movement of these species, as there was no “seasonal” difference between the 2 yrs.H owever, non-schooling species also were very clearly a component of observed habitat differences, in that two species of grunt (H. flavolineatum and H. sciurus) and one snapper (L. apodus) also contributed significantly to habitat distinction. West Cay is a more striking example of the effects of mangrove cutting. The ANOSIM was significant for both habitat (R = 0.259; P = 0.003) and year (R = 0.693; P < 0.001) (Fig. 3). The SIMPER showed that schooling species contributed most to both yearly and habitat differences, with only one non-schooling species (H. sciurus) appearing among the top five species contributing most to both habitat and year distinction. MDS of the minnow trap data for the five most abundant species (Table 3) was somewhat inconclusive (Fig. 4). Trap catch-rates were low (638 individuals, 24 species in 563 trap-nights). Of the five most abundant species captured in minnow traps Table 1. Density of fishes (mean individuals 10 –1m shoreline ± 1 SE) and number of taxa recorded during visual censuses within undisturbed and disturbed Rhizophora mangle fringe at East and West Twin Cays, Belize. Values averaged across years (2003–2004). Schooling taxa include only clupeids, engraulids, and atherinids and unknown smaller schooling taxa.

East Cay West Cay Taxon Undisturbed Disturbed Undisturbed Disturbed Total Density (all taxa) 261.7 ± 73.4 54.0 ± 13.1 809.8 ± 239.3 12.0 ± 2.0 284.4 ± 142.2 Density 66.3 ± 7.3 27.7 ± 4.0 31.3 ± 2.6 7.4 ± 0.6 33.2 ± 3.2 (non-schooling taxa) Density 195.4 ± 72.2 26.3 ± 11.8 778.6 ± 239.3 4.6 ± 2.0 251.2 ± 70.7 (schooling taxa) No. taxa 47 39 49 34 62 No. visual transects 20 20 20 20 80 884 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 Total 6.1 ± 0.6 8.6 ± 5.4 9.7 ± 1.7 0.4 ± 0.1 0.6 ± 0.1 0.6 ± 0.1 0.7 ± 0.3 1.1 ± 0.2 1.4 ± 0.2 1.5 ± 0.3 2.0 ± 0.2 5.8 ± 1.0 0.2 ± 0.0 0.2 ± 0.1 0.2 ± 0.1 0.2 ± 0.1 0.2 ± 0.1 0.2 ± 0.3 ± 0.1 0.3 ± 0.0 0.4 ± 0.1 0.4 ± 0.1 0.1 ± 0.0 0.1 ± 0.0 0.1 ± 0.2 ± 0.1 0.2 ± 0.0 13.9 ± 6.3 118.9 ± 53.5 118.9 109.8 ± 48.5 Rhizophora mangle 0.1 ± 0 2.1 ± 0.1 0.6 ± 0.5 0.5 ± 0.1 0.0 ± 1.1 ± 0.4 ± 0.1 0.2 ± 0.1 0.0 ± 0.2 ± 0.1 1.0 ± 0.1 0.0 ± 0.9 ± 0.1 0.1 ± 2.8 ± 1.8 0.5 ± 0.1 0.0 ± 0.1 ± 0.0 0.0 ± 0.0 ± 0.0 ± 0.0 ± 0.1 ± 0.0 0.4 ± 0.1 0.3 ± 0.2 0.0 ± 0.0 ± 0.0 ± 0.1 ± 0.1 ± 0.0 Disturbed West Cay West 6.8 ± 0.6 6.5 ± 0.8 0.7 ± 0.3 0.5 ± 0.2 0.5 ± 0.1 0.3 ± 0.1 1.4 ± 0.5 1.5 ± 0.2 2.7 ± 0.6 1.9 ± 0.3 3.9 ± 1.0 0.1 ± 0.0 ± 0.5 ± 0.3 0.3 ± 0.2 0.0 ± 0.0 ± 0.5 ± 0.3 0.4 ± 0.1 0.6 ± 0.3 1.0 ± 0.4 0.0 ± 0.1 ± 0.0 0.4 ± 0.2 0.1 ± 0.1 ± 33.3 ± 20.9 34.6 ± 22.8 Undisturbed 402.2 ± 180.7 308.5 ± 197.2 5.6 ± 0.9 0.3 ± 9.1 ± 2.6 7.3 ± 4.9 0.4 ± 0.1 0.5 ± 0.1 1.0 ± 0.2 0.0 ± 1.0 ± 0.4 0.5 ± 0.1 0.9 ± 0.3 2.3 ± 0.4 3.0 ± 1.1 8.1 ± 6.8 0.0 ± 0.5 ± 0.2 0.0 ± 0.3 ± 0.1 0.2 ± 0.1 0.9 ± 0.1 ± 0.3 ± 0.1 0.2 ± 0.1 0.1 ± 0.0 0.1 ± 0.0 ± 0.1 ± 0.0 0.0 ± 0.1 ± Disturbed 10.6 ± 9.6 East Cay 9.8 ± 1.7 0.0 ± 0.2 ± 0.1 1.1 ± 0.2 0.9 ± 0.1 2.4 ± 1.0 1.9 ± 0.3 2.6 ± 0.5 2.3 ± 0.6 2.8 ± 0.3 0.0 ± 0.1 ± 0.0 0.1 ± 0.1 ± 0.0 0.6 ± 0.4 0.0 ± 0.4 ± 0.1 0.3 ± 0.1 0.3 ± 0.1 0.2 ± 0.1 0.1 ± 0.2 ± 0.1 0.1 ± 0.3 ± 0.2 0.3 ± 0.1 22.7 ± 5.1 13.8 ± 9.5 16.0 ± 2.4 25.3 ± 18.9 Undisturbed 156.4 ± 73.2 shoreline ± 1 SE) recorded during visual censuses within fringing –1 (Desmarest, 1823) Linnaeus, 1758 (Bloch, 1791) (Müller and Troschel in Schomburgk, 1848) in Schomburgk, Troschel (Müller and (Lacépède, 1801) (Linnaeus, 1758) (Castelnau, 1855) Haemulon flavolineatum (Walbaum, 1792) Lutjanus apodus (Walbaum, Atherinidae (Shaw, 1803) Haemulon sciurus (Shaw, Clupeidae Engraulidae Halichoeres bivittatus Halichoeres (Valenciennes in Cuvier and Valenciennes, 1840) Valenciennes, in Cuvier and Sparisoma radians (Valenciennes Thalassoma bifasciatum (Bloch, 1791) Chaetodon capistratus Lutjanus synagris (Linnaeus, 1758) Scarus iseri (Bloch, 1789) Stegastes leucostictus Lutjanus griseus (Linnaeus, 1758) Abudefduf saxatilis Taxon Schoolers (unidentified) at Twin Cays, Belize. Values averaged across years (2003–2004). Values Cays, Belize. Twin at Stegastes variables Eucinostomus spp. Acanthurus bahianus Castelnau, 1855 Haemulon spp. Sparisoma viride (Bonnaterre, 1788) (Valenciennes in Cuvier and Valenciennes, 1840) Valenciennes, in Cuvier and (Valenciennes Sparisoma aurofrenatum Haemulon plumierii (Valenciennes in Cuvier and Valenciennes, 1839) Valenciennes, in Cuvier and garnoti (Valenciennes Halichoeres Table 2. Density of the 30 most abundant fishes (mean individuals 10 m Table (Walbum, 1792) Sphyraena barracuda (Walbum, (Walbaum, 1792) (Walbaum, cinereus Gerres (Troschel in Müller, 1865) in Müller, Stegastes adustus (Troschel (Linnaeus, 1758) virginicus Anisotremus (Poey, 1860) notata (Poey, Strongylura (Bloch and Schneider, 1801) Sparisoma chrysopterum (Bloch and Schneider, TAYLOR ET AL.: EFFECTS OF MANGROVE REMOVAL ON FISHES IN BELIZE 885

SIMPER by habitat SIMPER by year Species % Contribution Species % Contribution Haemulon flavolineatum 12.05 Schoolers (unk) 20.13 Schoolers (unk) 10.91 Engraulidae 10.43 Haemulon sciurus 9.51 Clupeidae 7.91 Clupeidae 5.49 Haemulon sciurus 7.17 Lutjanus apodus 5.46 Haemulon flavolineatum 6.49 Figure 2. A comparison of the ichthyofaunal species assemblage between undisturbed and disturbed Rhizophora mangle prop-root habitats at East Cay, Twin Cays, Belize, as demonstrated by non-metric multi-dimensional scaling. Each point represents mean density of all fishes (fish linear meter–1) averaged by observer for each survey date. Interpoint distances are proportional to overall faunal similarity. 2-D stress = 0.18. The table shows the SIMPER analysis by habitat and year.

(Table 3), four (Poecilia, Bathygobius, Floridichthys, and Gambusia) reach small adult size (< 70 mm SL), and are associated with mangrove or other shallow wetlands for their entire life cycle. In contrast, L. apodus is most likely to be found on reefs as adults (Greenfield and Thomerson, 1997). The primary intent of the minnow trapping effort was to establish if any differences existed in the transition, where visual observations are not possible, but this technique did not capture any differences in community structure, if they exist. The ANOSIM was not significant for year (R = 0.240; P = 0.114), so habitats (fringe vs transition) were pooled across years. ANOSIM (two-way crossed) still revealed no significant differences (habitat: R = −0.343; P = 1.0 and undisturbed/disturbed: R = −0.042; P = 0.47). Water Quality.—Data from the two water-quality transects through the disturbed/undisturbed transition areas showed consistently higher temperatures in the disturbed habitat, clearly a result of increased insolation with total canopy removal (mean disturbed: 33.4 °C, undisturbed: 30.0 °C; ANOVA P < 0.001). There was no significant difference in DO and salinity, although significantly higher salinity was found in the disturbed transition (ANOVA P < 0.001) in the more interior of the two transects (mean salinity at the waterward line was 33.7, and 39.5 in interior). 886 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

SIMPER by habitat SIMPER by year Species % Contribution Species % Contribution Engraulidae 20.80 Schoolers (unk) 19.89 Schoolers (unk) 17.87 Engraulidae 17.68 Atherinidae 6.52 Atherinidae 6.00 Clupeidae 6.36 Clupeidae 5.26 Haemulon sciurus 5.53 Eucinostomus spp. 3.57 Figure 3. A comparison of the ichthyofaunal species assemblage between undisturbed and disturbed Rhizophora mangle prop-root habitats at West Cay, Twin Cays, Belize, as demonstrated by non-metric multi-dimensional scaling. Each point represents mean density of all fishes (fish linear meter–1) averaged by observer for each survey date. Interpoint distances are proportional to overall faunal similarity. 2-D stress = 0.13. The table shows the SIMPER analysis by habitat and year.

Discussion

While it may seem obvious that removal or physical alteration of mangroves will affect the fish assemblages found there, we found few studies that document this. One study, Layman et al. (2004), investigated the effects of fragmentation in Baha- mas mangrove estuaries. Hydrologic fragmentation of surface water connection to the ocean by roads resulted in lower fish species diversity and shifts to assemblages more tolerant of extremes in salinity and temperature. Reef-associated species (e.g., pomacentrids, scarids) and transient marine species (e.g., carangids) were typically found only in unfragmented estuaries. Most of the literature addressing the effects of mangrove destruction extrapolates directly to the effects on coral reef fishes on the reefs themselves. Our data indicate that even partial removal of fringing R. mangle and landward mangrove stands affects fish assemblages found there. Although species diversity was reduced by disturbance, the effect on density was even more striking, especially in schooling species. At the West Cay site, schooling densities were nearly 200 times less at the disturbed site. Though both yearly and site differences were factors in observed variability, habitat (undisturbed/disturbed) was a significant variable in all ANOVA analyses. Table 2 shows the numerical dominance of schooling species at all sites. With few exceptions, among the 30 most abun- TAYLOR ET AL.: EFFECTS OF MANGROVE REMOVAL ON FISHES IN BELIZE 887

Table 3. CPUE of the five most abundant fishes (fish trap-night–1 ± 1 SE) recorded from wire minnow traps within fringe and transition mangrove habitats at East and West Twin Cays, Belize. Values averaged across years (2003–2005).

Fringe Transition Taxon Undisturbed Disturbed Undisturbed Disturbed Total Poecilia orri Fowler, 1943 0.37 ± 0.36 0.60 ± 0.60 1.30 ± 0.36 1.52 ± 1.09 0.85 ± 0.29 Bathygobius spp. 0.16 ± 0.03 0.27 ± 0.11 0.09 ± 0.05 0.10 ± 0.04 0.17 ± 0.04 Floridichthys polyommus 0.01 ± 0.01 0.03 ± 0.03 0.04 ± 0.04 0.44 ± 0.04 0.11 ± 0.06 Hubbs, 1936 Lutjanus apodus 0.10 ± 0.05 0.11 ± 0.08 0.02 ± 0.02 0.17 ± 0.02 0.10 ± 0.03 Gambusia yucatana 0.00 ± 0.00 0.00 ± 0.00 0.08 ± 0.08 0.18 ± 0.14 0.05 ± 0.03 Regan, 1914 CPUE (All Taxa) 0.68 ± 0.37 1.10 ± 0.49 1.54 ± 0.55 2.42 ± 0.99 1.33 ± 0.31 No. Trap Sets 196 184 100 83 563 No. Taxa 14 11 20 17 24

dant species listed, all were more abundant in undisturbed than in corresponding disturbed sites. Two species of grunt (H. sciurus, H. flavolineatum) and one species of snapper (L. apodus) accounted for 26% of the variability between the undisturbed/disturbed sites of the East Cay, with schooling species accounting for 10%. The role of schooling species was much more pronounced on the West Cay, where they accounted for 52% of the variability. Because counts of schooling species are subject to some inaccuracy, the MDS was run without inclusion of this group for both East and West Cays (data not shown), with little effect on separation of the undisturbed/disturbed sites. As expected, both lutjanids and haemulids were the species primarily responsible for discrimination of site differences in these analyses. The minnow trap data failed to indicate any effect of mangrove cutting on assemblages within the intertidal fringe and transition. However, the separation of 2005 from 2003 and 2004 is probably due to seasonal effects and the very low tides encountered in 2005 as opposed to the other 2 yrs. As expected, the MDS indicated a slight difference in assemblages between fringe and transition habitat-types. These differing results from the two methods (visual vs minnow traps) may be explained by the presence of a larger species pool in the sub-tidal fringe and the more comprehensive nature of visual census, as opposed to the selective nature of trapping. In addition, species typically found within the transition (e.g., poeciliids, cyprinodontids) are likely more tolerant of the extreme environmental fluctuations found here even in the undisturbed state, so disturbance in the transition may have less of an effect. The effects of mangrove cutting on fringe assemblages may extend beyond the loss of canopy/structure provided by R. mangle themselves. At both sites, the transition area landward of the fringe had been nearly completely cleared. While our water- quality data from the two transects indicated higher temperature in the disturbed transition, during ebb tides this heated, highly tannin-stained water is transported back through the fringe, frequently resulting in very hot “runout” channels of highly turbid water (D.S.T, pers. obs.). During ebb tides, we measured water temperature 10 m offshore from the fringe and typically found temperature 2–4 °C higher off the disturbed shoreline than off the undisturbed one. Our limited water-quality analysis may not have documented other significant differences, notably inD O, that occur at night. We suspect that DO is severely limiting to fishes in the disturbed transition at 888 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Figure 4. A comparison of the ichthyofaunal species assemblage between undisturbed and disturbed fringe and transition mangrove habitats at East Cay, Twin Cays, Belize, as demonstrated by non-metric multi-dimensional scaling. Each point represents mean catch rate of all fishes (fish trap–1 day–1) for 2003, 2004, and 2005. The transition was not sampled in 2005 due to low water. Interpoint distances are proportional to overall faunal similarity. 2-D stress = 0.04. night, since peat appears to be decomposing here. Many fishes captured in our minnow traps in the disturbed transition were found dead in the morning, even species notably very tolerant of low DO, such as Kryptolebias (formerly Rivulus) marmoratus (Poey, 1880). Peat erosion is clearly associated with disturbance at both East and West Cays. The East Cay disturbed shoreline has receded landward up to several meters over the last decade (D.S.T., pers. obs.) and within the transition, significant surface erosion appears to be occurring, as at flood tide, water depths here are greater than in adjacent undisturbed areas, sediments are notably less consolidated, and there is an absence of the characteristic, large burrowing mangrove land crab Ucides cordatus (Linnaeus, 1763). The West Cay shoreline is also eroding landward, but since this site is exposed to higher wave energy, erosion is also occurring at the bottom, where the peat is being undercut, exposing coral rubble and woody debris. In contrast, along the adjacent undisturbed shoreline of the West Cay, the substrate was stable and seagrass occurred right to the edge of the prop-root zone (D.S.T. pers. obs.). Our data show the value of intact mangroves to fishes within oceanic mangrove cays on the Belize barrier reef. The dramatic reduction in both species diversity and, especially, density between adjacent undisturbed/disturbed sites at two different locations implies that these effects could cascade to the adjacent barrier reef (< 2 km distant). In addition, our observations of higher water temperatures in cut-over transition habitat and in the fringe during ebbing tides as well as incidental observations of general habitat degradation (erosion, turbidity) suggest that the effects of mangrove disturbance may extend beyond the immediate area of impact. Woodroffe (1995) indicated that many of the mangrove ranges on the Belize barrier reef are undergoing recession, possibly due to the effects of sea-level rise. If this is the case, any additional losses of mangrove habitat due to human activity may aggravate a compounding problem for fisheries management. TAYLOR ET AL.: EFFECTS OF MANGROVE REMOVAL ON FISHES IN BELIZE 889

While many coral reef fishes are well-documented as using the mangal as nursery habitat, many of these species may also use alternative habitats, such as seagrass. However, the relative importance of these two habitats as nursery areas is yet to be determined. In addition, given the presumed nursery value of the mangal, the “distance effects” of the mangal on fish inhabiting adjacent reefs of varying proximity needs further attention (Halpern, 2004). While these and other questions remain, our data indicate that even incomplete clearing of fringe and transition mangrove significantly affects fish assemblages found in the fringe. It is noteworthy that of the 24 taxa found in fringing mangroves in this study that could be identified, 16 species are found primarily on coral reefs as adults. Our data indicate fewer individuals of coral reef affinity in mangroves that have been disturbed by cutting. We recommend further study at other locales to confirm our observations on the effects of mangrove clearing on local fish assemblages.

Acknowledgments

We are grateful for the financial support of the Caribbean Coral Reef Ecology Program (CCRE, Smithsonian Institution) for travel and logistic support in the field. This paper is CCRE Contribution #778. Financial support, database support, and equipment for C. C. M. was provided by U.S.G.S. In addition, further travel support was from a National Science Foundation grant (DEB-9981535) to I. Feller. Finally, we thank K. Kuss for data support, and the Belize Fisheries Department for permission to collect in the mangal.

Literature Cited

Beck, M. W., K. L. Heck, K. W. Able, D. L. Childers, D. B. Eggleston, B. M. Gillanders, B. Halp- ern, C. G. Hays, K. Hoshino, T. J. Minello, R. J. Orth, P. F. Sheridan, and M. P. Weinstein. 2001. The identification, conservation, and management of estuarine and marine nurseries for fish and invertebrates. BioScience 51: 633–641. Bray, J. R. and J. T. Curtis. 1957. An ordination of the upland forest communities of southern Wisconsin. Ecol. Monogr. 27: 325–349. Clarke, K. R. 1993. Non-parametric multivariate analyses of changes in community structure. Aust. J. Ecol. 18: 117–143. Ellis, W. L. 2003. Mangrove canopy damage: implications for litter processing and fish utilization of the intertidal zone. PhD Diss. University of South Florida, Tampa, FL. 157 p. Faunce, C. H. and J. E. Serafy. 2006. Mangroves as fish habitat: 50 years of field studies. Mar. Ecol. Prog. Ser. 318: 1–18. Feller, I. C., K. L. McKee, D. F. Whigham, and J. P. O’Neill. 2002. Nitrogen vs. phosphorus limitation across an ecotonal gradient in a mangrove forest. Biogeochem. 62: 145–175. Greenfield, D. W. and J. E. Thomerson. 1997. Fishes of the Continental Waters of Belize. Uni- versity of Florida Press, Gainesville. 311 p. Halpern, B. S. 2004. Are mangroves a limiting resource for two coral reef fishes? Mar. Ecol. Prog. Ser. 272: 93–98. Layman, C. A., D. A. Arrington, R. B. Langerhans, and B. R. Silliman. 2004. Degree of fragmentation affects fish assemblage structure in Andros Island (Bahamas) estuaries. Caribb. J. Science 40: 232–244. Lugo, A. E. and S. C. Snedaker. 1974. The ecology of mangroves. Ann. Rev. Ecol. Syst. 5: 39–64. Manson, F. J., N. R. Loneragan, G. A. Skilleter, and S. R. Phinn. 2005. An evaluation of the evidence for linkages between mangroves and fisheries: a synthesis of the literature and identification of research directions. Ocean. Mar. Biol. Ann. Rev. 43: 485–515. 890 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Mumby, P. J., A. J. Edwards, J. E. Arias-González, K. C. Lindeman, P. G. Blackwell, A. Gall, M. I. Gorczynska, A. R. Harborne, C. L. Pescod, H. Renken, C. C. C. Wabnitz, and G. Llewellyn. 2004. Mangroves enhance the biomass of coral reef fish communities in the Caribbean. Nature 427: 533–536. Nagelkerken, I., G. van der Velde, M. W. Gorissen, G. J. Meijer, T. van’t Hof, and C. den Hartog. 2000. Importance of mangroves, seagrass beds and the shallow coral reef as a nursery for important coral reef fishes, using a visual census technique. Estuar. Coast. Shelf Sci. 51: 31–44. Valiela, I., J. L. Bowen, and J. K. York. 2001. Mangrove forests: one of the world’s threatened major tropical environments. BioScience 51: 807–815. Woodroffe, C. D. 1995. Mangrove vegetation of Tobacco Range and nearby mangrove ranges, central Belize barrier reef. Atoll. Res. Bull. No. 427. 35 p.

Addresses: (D.S.T.) Brevard County Environmentally Endangered Lands Program, 91 East Dr., Melbourne, Florida 32904. (E.A.R.) Dynamac Corporation, Mail Code: Dyn-2, Kennedy Space Center, Florida 32899. (W.P.D.) PO Box 607, Quincy, Florida 32353. (C.C.M.) U.S. Geo- logical Survey, Florida Integrated Science Center, Center for Coastal and Watershed Studies, St. Petersburg, Florida 33701. Corresponding Author: (D.S.T.) E-mail: . BULLETIN OF MARINE SCIENCE, 80(3): 891–904, 2007

BASELINE ASSESSMENT OF FISHERIES FOR THREE SPECIES OF MUD CRABS (SCYLLA SPP.) IN THE mangroves of Ibajay, Aklan, PhilippineS

Ma. Junemie Hazel L. Lebata, Lewis Le Vay, Jurgenne H. Primavera, Mark E. Walton, and Joseph B. Biñas

ABSTRACT Stock enhancement through habitat restoration and habitat release have both been considered as approaches to the management of declining Scylla spp. Prior to stock enhancement trials, the present study was conducted to monitor recruitment and yields of three Scylla spp. in ~70 ha of natural mangroves in Aklan, Panay, Phil- ippines. Results showed that Scylla olivacea (Herbst, 1796) was the most abundant mud crab species, comprising 95% of the catches over the 4 yr sampling period. Size distribution for this species indicated year-round recruitment with peaks in the numbers of smaller, immature crabs during the summer months. The decreasing mean size at capture, yield and CPUE in terms of weight throughout the 4-yr sampling period is an indication that the area has been subjected to heavy fishing pressure. The constant CPUE in terms of numbers of crabs suggests that recruitment is constant, though this is likely to be lower than in other mangrove areas due to the topography of the site with limited access to the open sea, resulting in relatively low crab abundance and yields. Combined with the fidelity of S. olivacea to the mangrove habitat, this indicates a suitable population for investigation of the effectiveness of a hatchery-release program.

The mud crabs of the genus Scylla de Haan, 1833 are large, edible crustaceans associated with mangroves throughout the Indo-West Pacific region (Macnae, 1968; Ng, 1998) and form an important component of small-scale fisheries. They are highly prized as food and have a high value in domestic and export markets. Thus they are subject to heavy fishing pressure by subsistence and commercial fisheries which are an important source of income for coastal fishing communities (Overton and Ma- cintosh, 2002). In recent years, over-exploitation has resulted from increasing human populations and rising demand for live seafood products (Bell and Gervis, 1999) in affluent and developing countries in Asia (Overton et al., 1997). In the absence of -ef fective regulations, as is the case in many countries where crab fisheries exist, fishers target all size classes of crabs (Le Vay et al., 2001) with juveniles being collected for pond culture, and adult and sub-adults for short-term fattening, consumption, or for sale in seafood markets. To date, most aquaculture of mud crabs, such as fattening and grow-out in ponds and pens, depends on wild-caught juveniles (Chandrasekaran and Natarajan, 1994; Le Vay, 2001). In addition to these more conventional culture methods, the recent increasing interest in soft-shelled crab farming has increased pressure on Scylla populations in Southeast Asia (Overton et al., 1997; Dat, 1999). Aside from overexploitation, loss of habitat has also caused wild fisheries to decline. For example, in the Philippines, uncontrolled industry development and human demographic changes have contributed to the degradation of the coastal environment and its aquatic resources (Kuhlmann, 2002). Between 1918 and 1994, the country lost more than 70% of its mangroves, mostly to aquaculture ponds (Primavera, 2000). Loss of mangroves can be associated with loss of coastal productivity, biodiversity,

and socio-economic value of fisheries (Ronnback, 1999). The loss of mangrove habitat with which mud crabs are closely associated (Le Vay, 2001; Walton et al., 2006b) and overfishing have resulted in reduced landings and smaller mean size at capture of mud crabs (Overton et al., 1997; Kosuge, 2001). Despite such widespread economic importance, prior to the re-classification by Keenan et al. (1998), the four Scylla species were not differentiated in most ecological and biological studies. Early work by Estampador (1949a,b) recognized four distinct groupings of mud crabs based on color patterns, relative size, cheliped spination, chromosome form, and the process of gamete development, but most researchers followed Stephenson and Campbell (1960) in identifying all morphs as Scylla serrata (Korskål, 1775). Keenan et al. (1998) made a clear case for recognizing four species of Scylla: Scylla olivacea (Herbst, 1796), Scylla paramamosain Estampador, 1949, S. serrata, and Scylla tranquebarica (Fabricius, 1798). This re-classification is further supported by the recent findings of Imai et al. (2004) identifying all four species of Scylla using nuclear and mitochondrial DNA. In several studies in Southeast Asian mangroves, sympatric occurrence of two of the Scylla species has been confirmed (Overton et al., 1997; Le Vay et al., 2001; Overton and Macintosh, 2002; Walton et al., 2006a,b). However, the factors controlling the distributions of the four species have only been inferred from sampling over a wide geographic scale (Keenan et al., 1998) and little is known about their relative abundance over regional scales and within habitats. There are several possible approaches to address the problem of declining mud crab populations. These may include regulation of fishing effort, rehabilitation of mangrove habitats, mangrove-friendly aquaculture, and hatchery-based enhancement of wild crab stocks. Reforestation of previously-cleared mangrove areas has proven to be a highly successful approach to restoring depleted Scylla populations (Walton et al., 2006b). Stock enhancement by release of hatchery-reared crabs also offers potential for replenishing depleted stocks and managing fishery yields due to the post-recruitment fidelity of mud crab species to mangroves (Le Vay, 2001) but requires preliminary long-term studies to determine the presence and abundance of each species, their population dynamics, and fisheries yields in the proposed release area. This information may support the decision to implement release trials, selection of species for release, identification of an optimal strategy for seasons and habitats for releases, projection of potential yields, as well as providing a baseline against which success of releases can be monitored (Blankenship and Leber, 1995; Bell, 2004). In advance of stock enhancement trials, the present study was conducted to provide a baseline assessment of the population and fisheries of wild Scylla spp. and to monitor recruitment and yields in the natural mangroves of Naisud and Bug- tong Bato, Ibajay, Aklan, Panay Island, Philippines.

Materials and Methods

The study was conducted in ~70 ha of natural mangroves adjacent to the villages of Naisud and Bugtong Bato, municipality of Ibajay, in Aklan province on the north coast of Panay Is- land, central Philippines (Fig. 1). The mangrove area is situated between 122°11.4´–122°12.7´E and 11°47.8´–11°48.5´N and an area of high species diversity, with areas of both primary and secondary forest containing 27 of the 35 mangrove floral species that are found in the Philip- pines (Primavera et al., 2004). The area was surveyed with the aid of a hand-heldG PS (Garmin LEBATA ET AL.: POPULATION AND FISHERIES OF SCYLLA SPP. 893

Figure 1. The location of mangroves in the villages of Naisud and Bugtong Bato, Ibajay, Aklan, Philippines. 12 XL) and mapped using GPS mapping software (OziExplorer version 3.90.3a), refined in CorelDraw version 11.633. Six fishers were contracted to undertake standardized fishing every spring tide for the entire sampling period. Throughout the study period they were the only fishermen regularly catching crabs in the study area and thus represented the total fishing effort on the exploited populations. All catches were recorded for a period of 43 mo (April 2002–November 2005, except January 2004). From May 2005, sampling was only on the first spring tide of the month. Each fisher was trained to record data for type and number of gear used, percentage of traps containing crabs, deployed time and duration, crab species, sex, sexual maturation of females, carapace width, and body weight (noting missing limbs). Sampling every spring tide ranged from 4 to 7 d and the data were retrieved from each collector at the end of each sampling period. Three types of gear, all of which are used locally, were employed in the study. Baited lift nets are made from a 30-cm square panels of 1.5 cm mesh nylon net attached on four cor- ners to two pliable crossed bamboo slats with the bait hanging at the junction of the crossed slats. They are operated for 4–6 hrs commencing 2–3 hrs before the peak of the highest tide. Cylindrical traps (length 90 cm, diameter 25–30 cm) are made from bamboo slats with funnel-shaped entrances at each end, with the bait centrally positioned. The traps are set during low tide prior to the highest tide and retrieved on the next low tide. Baited lines attached to a bamboo pole are operated with a scoop net made of the same nylon mesh used in lift nets. The scoop net is used to catch the crabs that are attracted to the bait. These baited lines are deployed at the same time as the lift net. For all three types of gear, bait consisted mainly of 894 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

frogs or fish depending on availability. All three fishing methods were utilized until Decem- ber 2003. From February 2004, samples were collected using cylindrical bamboo traps only. Crabs were identified to species level following Ng (1998) and Keenan et al. (1998). Crabs were identified by collectors as native S.( olivacea), king crab (S. serrata), and “manginlawod” (S. tranquebarica), with these identifications being confirmed by the first author throughout the study. Length was measured as internal carapace width (mm CW) with a ruler or a dial caliper. Body weight (g BW) was measured using a balance. Sex and ovarian maturation of each crab were noted and females categorized as immature, mature, gravid, berried, or spent using the ovary and abdomen morphology. The indicators of maturity were abdominal morphology, the presence of setae on the fringe of the abdominal flap and abdominal disengage- ment from the thorax (Walton et al., 2006a). Temperature and salinity were measured using a hand-held YSI temperature-conductivity meter. From the data collected, monthly species composition, sex ratio, size distribution, total monthly crab catches, monthly yield, catch per unit effort (CPUE) in terms of number and biomass were analyzed. Size at sexual maturity was estimated following Jennings et al. (2001). Monthly mean CW (mm), BW (g), CPUE (number and biomass), and yield (kg) were each correlated with time, temperature, and salinity. All analyses was performed using the Minitab software package (version 14).

Results

Temperature ranged from 26.1 to 32.2 °C (mean ± SE = 28.7 ± 0.2) and salinity from 3.8 to 35.3 (mean ± SE = 22.6 ± 1.4). The highest temperature and salinity were both observed in April 2003, the lowest temperature in January 2005 and the lowest salinity in September 2005. Of the total crabs recorded (n = 8025), 95.0% were S. olivacea, 3.2% S. serrata, and 1.8% S. tranquebarica. Two-way ANOVA showed no significant difference in the relative abundance of each species in catches between the three gears but overall differences between catches of the three species were highly significant (ANOVA: P < 0.001, F = 2235.1, df = 2). There were no significant differences in the male to female ratio of crabs among species and among gears, all being close to unity. Immature females dominated the total female crab landings throughout the 43-mo sampling period (Fig. 2) regardless of the species and the gear used. Overall, the occurrence of immature females was significantly higher than mature and gravid animals, which in turn were significantly more abundant than spent females (ANOVA: P < 0.001, F = 187.9, df = 3). Of the 3976 female crabs caught, 75.2% were immature, 12.6% mature, 11.3% gravid, and 1.0% spent. No berried crabs carrying eggs were caught inside the mangroves. By species, immature crabs comprised 74.8% of female S. olivacea (n = 3767), 98.5% of S. serrata (n = 134), and 49.3% of S. tranquebarica (n = 75). Of the gears operated, immature females made up 82.6% and 71.2% of the female crabs caught in bamboo trap (n = 3102) and lift net (n = 198), respectively. In contrast to the baited line and scoop net (n = 671) which had an almost equal proportion of mature (43.1%) and immature females (42.0%). The size of crabs recorded ranged from 20.8 to 195.0 mm CW (20.8–140.0 mm for S. olivacea, 24.4–172.0 mm for S. serrata, and 40.0–195.0 mm for S. tranquebarica). Monthly length-frequency histograms for pooled male and female S. olivacea showed no clear modal progression. However, a seasonal cycle in both monthly mean CW (Fig. 3A) and BW (Fig. 3B) was observed throughout the 4 yr sampling period. Size increased between February and April, peaking between August and October with a subsequent decline. These lows and highs in size correspond to the dry (summer) LEBATA ET AL.: POPULATION AND FISHERIES OF SCYLLA SPP. 895

Figure 2. Monthly distribution (%) of different stages of sexual maturation in female (F) Scylla spp. caught from April 2002–November 2005. (Black) spent ovary; (light gray) mature, gravid ovary; (dark grey) mature, ovary not well-developed; (white) immature

and wet (rainy) seasons in the Philippines, respectively. In addition an overall general trend of decreasing monthly means for both CW and BW was observed. These decreasing trends for both length and weight are also evident in the annual histograms of sizes from 2002 to 2005 (Fig. 4). Monthly mean CW and monthly mean BW were not correlated with salinity and temperature but were both significantly negatively correlated with time (Pearson correlation = −0.649 and −0.510, respectively; P < 0.001 for both). The size at which 50% of female S. olivacea attained maturity was estimated as 87.8 mm CW by fitting a logistic curve to the relationship between the proportion of mature females and CW (Jennings et al., 2001). This analysis was not applied to S. serrata or S. tranquebarica due to the small number of samples. The smallest mature female observed was 55 mm in S. olivacea, 121 mm in S. serrata, and 60 mm in S. tranquebarica. The largest immature female recorded was 120 mm CW in S. olivacea, 117 mm in S. serrata, and 90 mm in S. tranquebarica. Scylla olivacea and S. tranquebarica were observed to start developing gravid ovaries at 65 and 70 mm, respectively. Bamboo traps deployed throughout the sampling period caught 6503 crabs (81.0%). Line and scoop and lift nets, operated April 2002–December 2003, caught 1185 and 337 crabs (14.8% and 4.2% of landings during this period), respectively. A Kruskal- Wallis test suggested the line and scoop net caught significantly larger crabs (mean ± SE = 79.6 ± 0.4) compared to the lift net (mean ± SE = 75.7 ± 0.9) and bamboo trap (mean ± SE = 68.3 ± 0.2) (H = 568.29; P < 0.001). The mean monthly yield from all traps deployed was 17.0 kg, varying between 8–48 kg or 0.34 kg ha–1, giving an annual fishing yield of 4.1 kg ha–1 (Fig. 5). Catch per unit effort (CPUE) was analyzed monthly using only bamboo trap catches to allow comparison of the entire survey period with consistent fishing effort (Fig. 6). Mean monthly CPUE for number ranged from 0.32 ± 0.02 (April 2003) to 1.02 ± 0.16 crabs gear–1 d–1 (April 2002). In terms of biomass, CPUE ranged from 32.0 ± 3.0 to 94.5 ± 17.7 g gear–1 d–1. A decreasing trend was observed for CPUE in both number and 896 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Figure 3. Monthly mean (A) carapace width (mm CW) and (B) body weight (g BW) of pooled males and females of Scylla olivacea from April 2002–November 2005. Error bars represent 95% confidence interval from the mean. biomass from the start of the sampling in April 2002–December 2003. CPUE in terms of both number and biomass more than doubled in February 2004, increased until June 2004, then started to drop again. Significant variation in CPUE characterized by high median values in April–July 2002 and February–July 2004 was observed both for number (H = 218.55; P < 0.001) and biomass (H = 215.14; P < 0.001). Both yield and CPUE (number) were not correlated with temperature, salinity, or time. However, CPUE (biomass) was significantly negatively correlated with time (Pearson correlation = −0.372, P < 0.05).

Discussion

Crabs of the genus Scylla are widely distributed throughout the Indo-Pacific (Keenan et al., 1998; Ng, 1998; Keenan, 1999; Imai et al., 2004; Imai and Takeda, 2005). According to Keenan et al. (1998), of the four Scylla species, S. serrata is the LEBATA ET AL.: POPULATION AND FISHERIES OF SCYLLA SPP. 897

Figure 4. Annual (A) length-frequency and (B) weight-frequency distributions (%) of Scylla olivacea caught (A) April–December 2002, (B) January–December 2003, (C) February–December 2004, and (D) January–November 2005. Mean, standard deviation, and number of samples per year (N) are shown. most widespread, occurring naturally throughout the Indo-Pacific. Although it can tolerate reduced salinity, it is commonly associated with mangrove forests inundated for most of the year with full salinity oceanic water. It is dominant in oceans where salinity is > 34. Scylla tranquebarica, on the other hand, is commonly found in the South China Sea and has been recorded in some specific locations in the Indo-Pacif- ic. It is commonly found in mangrove forests and coastlines inundated with reduced salinity for part of the year and often associated with S. olivacea. Scylla paramamosain is abundant in the continental coast of the South China Sea and the south of the Java Sea. Although associated with mangrove forests, it also may inhabit coral reef rubble, shallow subtidal flats, and estuarine ponds. In Vietnam, it is associated with estuarine habitats that experience seasonal periods of extremely low salinity (Le Vay et al., 2001; Walton et al., 2006a). Scylla olivacea, though commonly found in the South China Sea, also occurs in specific locations in the Indo-West Pacific. Often associated with S. tranquebarica, it is very common in mangrove forests and coastlines inundated with reduced salinity seawater. Of Keenan’s (1999) collection, 898 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Figure 5. Total monthly catch (kg) of Scylla spp. from the mangroves of Naisud and Bugtong Bato, Ibajay, Aklan, Philippines, April 2002–November 2005. Black bars represent months sampled in two spring tides while gray bars in only one spring tide.

S. olivacea has a strong representation in samples from the Philippines and Malaysia. Recent literature reports the existence of S. olivacea in the mangroves in Australia (Le Vay, 2001); Thailand (Moser et al., 2002; Overton and Macintosh, 2002; Moser et al., 2005); Vietnam (Macintosh et al., 2002; Walton et al., 2006a); Malaysia (Ko- suge, 2001); Indonesia, Taiwan and Japan (Imai et al., 2004; Imai and Takeda, 2005). Moreover, Ng (1998) reported that S. olivacea is the most common species found in the markets in Southeast Asia, reflecting the high percentage of S. olivacea sourced from wild fisheries. With recent developments in the species classification within the genus Scylla (Keenan et al., 1998; Imai et al., 2004), species identification in reports prior to these revisions is uncertain. In the Philippines, most published works refer to all mud crabs as S. serrata (Arriola, 1940; Escritor, 1972; Baliao et al., 1981; Alcala and Alcazar, 1984; Agbayani et al., 1990). In the present study, the sympatric occurrence of three out of four of the Scylla species was observed in the mangroves in Aklan, Philippines. The predominance ofS. olivacea in this mangrove system is consistent with a recent study of crab populations in nearby replanted mangroves at Kalibo, Aklan (Walton et al., 2006b). Moser et al. (2005) have suggested a strong, though not exclusive, preference of S. olivacea for the intertidal zone and this close linkage to mangrove habitats is thought to make it a useful indicator species for assessing the condition of mangrove habitats (Walton et al., 2006b). Although it was first described by Estampador (1949a), and reported to exist in considerable numbers in certain regions in the Philippines where extensive mangrove forests are found, S. paramamosain was not found at the study site. This is consistent with Keenan (1999), who also reported only the other three species of Scylla from Panay. The absence of S. paramamosain is difficult to explain, given the presence of suitable habitat and the fact that it has been reported to coexist with S. olivacea over a wide range including estuarine mangroves of the Mekong Delta in Vietnam (Le Vay et al., 2001; Walton et al., 2006a), Ban Don Bay in Thailand (Over- ton and Macintosh, 2002), and Shizouka, Japan (Imai et al., 2004). LEBATA ET AL.: POPULATION AND FISHERIES OF SCYLLA SPP. 899

Figure 6. Monthly mean catch per unit effort (CPUE) of Scylla spp. caught from April 2002–No- vember 2005. CPUE expressed as (A) number (crabs gear–1 d–1) and (B) biomass (g gear–1 d–1) were all from bamboo trap data. Error bars represent 95% confidence interval from the mean.

Moser et al. (2005) reported a male-dominated population of S. olivacea both in commercial and experimental fisheries during full moon spring tides in the mangroves in Ranong province, Thailand. Dominance of male crabs has also been observed in Scylla sp. crab landings in Kakinada Region in India (Devi, 1985) and in Eastern Cape estuaries in South Africa (Robertson, 1996). However, the present study and that of Hill et al. (1982) for S. serrata in Australia showed an almost 1:1 male to female ratio. The absence of berried mud crabs from catches inside mangroves, using traps, has also been reported by Hyland et al. (1984), Robertson and Kruger (1994), and Moser et al. (2005). However, Moser et al. (2005) reported that berried females were caught occasionally in gill nets on the mangrove fringe and drainage channels. In the present study area, a fisherman reported catching berried S. olivacea on two occasions (August 2002 and October 2003) in gill nets set at the mouth of Naisud River, the only channel connecting the mangroves to the open sea. The absence of berried crabs inside the mangroves and their occasional presence near the mouth of the river is consistent with migration from the mangrove estuarine habitat to oce- 900 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

anic or near oceanic waters for spawning, as reported by Arriola (1940), Hill (1975), Hyland et al. (1984), and Robertson and Kruger (1994). However, berried females may be unrepresented in sampling with baited traps as their feeding rate may be depressed (Heasman, 1980). The consistent prevalence in high numbers of immature female crabs throughout the sampling period and the lack of modal progression in size-frequency distributions may indicate year-round recruitment. Smaller male crabs were also observed, except that the stages of sexual maturation were unknown. However, peaks of occurrence of immature females were observed in February–March 2003, February–April 2004, and December 2004–April 2005 coinciding with the dry season. These coincided with the drop in mean size of crabs sampled, suggesting episodes of increased recruitment of smaller juveniles. The smallest crab caught by bamboo traps was 20.0 mm CW, indicating first recruitment into the trap fishery at this size. Seasonality in recruitment indicates an increase in migration, offshore spawning, and recruitment several months prior to these peaks, during the rainy season which lasts from June to December. Arriola (1940) and Estampador (1949b) also reported peaks of spawning activity in the Philippines from May to October and Perrine (1979) reported that offshore migration in tropical mud crab populations is stimulated by a decrease in salinity. Similarly, seasonal changes in salinity have been reported by Walton et al. (2006a) as an important factor in relation to recruitment of S. paramamosain. A common fishery management practice is to set a minimum legal size based on the size at first sexual maturity. Two common indices of sexual maturity are the smallest size of sexually mature female and the largest immature female (Overton and Macintosh, 2002). In the present study, classification of females into different stages of sexual maturity was based on the morphology of the abdominal flap and externally-observed stage of ovarian development but maturity in male crabs is not easily determined from external characteristics (Robertson and Kruger, 1994), hence stages of sexual maturity in males were not classified. For females, the estimated size at first sexual maturity (87.8 mm CW) is close to the value reported by Overton and Macintosh (2002) in Thailand where the smallest mature female recorded for S. olivacea was 83 mm CW and the largest immature female 118 mm. The largest S. olivacea recorded in the present study was 140 mm CW, comparable to the largest specimen recorded by Walton et al. (2006b) in Kalibo, Aklan (135 mm CW). In contrast, the maximum sizes of S. serrata and S. tranquebarica were larger (172 and 195 mm CW, respectively). Although relatively low in abundance, S. olivacea co-exists with S. serrata in northern Queensland, Australia, and females rarely reach the minimum legal landing size for mud crabs of 150 mm CW (Le Vay, 2001). In Kenya, female S. serrata have been observed to mature at 77 mm and become berried at 139 mm (Onyango, 2002). This was not found in the present study: females up to 117 mm did not attain sexual maturity. This corresponds with results of Robertson (1996), who reported all S. serrata sampled in Eastern Cape estuaries, South Africa, < 130 mm CW were immature. However, Robertson and Kruger (1994) reported mature female S. serrata from the same region as small as 104 mm. In Australia, S. serrata juveniles (20–99 mm CW) are resident in the mangrove zone, with subadults (100–149 mm CW) living in subtidal waters at low tide and foraging in the intertidal zone at high tide, and adults (> 150 mm CW) being mainly subtidal (Hill et al., 1982). This suggests possible under-representation of S. serrata, particularly mature animals, during the 4-yr sampling period of the present study. LEBATA ET AL.: POPULATION AND FISHERIES OF SCYLLA SPP. 901

In Vietnam, Walton et al. (2006a) also observed a pattern of differences in habitat preferences among different sizes of S. paramamosain with successively larger size classes in the mangrove fringe, intertidal burrows, and outside the mangroves, respectively. In the present work, S. olivacea did not exhibit such a clear distribution with life stage, but an overall high degree of fidelity to the intertidal mangrove habitat. Except for the absence of berried crabs, all sizes and stages were caught inside the mangroves. Indeed, they may go out offshore to spawn, but the presence of spent females in the samples show that they return to the mangroves after spawning. The use of traps is one of the easiest and more convenient ways of collecting mud crabs in mangrove habitats that are difficult to survey using other methods, and have been widely used for research purposes ��(Heasman et al., 1985; Robertson, 1996; Mo- ser et al., 2005; Walton et al., 2006b). According to Robertson (1989), it is the most practical and widely used gear for catching crabs for fisheries and research purposes. Of the three gears used, the least selective for species, sex or size was the bamboo trap. The trap design in this particular locality where bamboo slats are closely arranged resulted in very small size of crabs (20 mm CW) at first recruitment into the fishery. Through a repeated cycle of highs and lows, both CPUE and size of crabs were observed to decline each year of the present study. Overton et al. (1997) noted that as a result of high fishing pressure, the average size of mud crabs caught in Southeast Asian countries appears to be decreasing. This result is further supported by anecdotal claims of crab collectors in the present study area that a decade or two ago more and larger crabs could be caught over a short period of time (filling a 20-l bucket with crabs in 2 or 3 hrs of hand picking). The mean monthly yield of mud crabs of 4.1 kg ha–1 yr–1 in the present study is substantially lower than that in the reforested mangrove, in Kalibo, Aklan (65.4 kg ha–1 yr–1) (Walton et al., 2006b). In terms of CPUE using similar traps, the value obtained in the present study (0.61 crab trap–1 d–1) was half that of Walton et al. (2006b) (1.26 crabs trap–1 d–1). Higher estimates of mud crab fisheries production from natural mangroves have been reported elsewhere, 13 kg ha–1 yr–1 in Chanthaburi, Thailand and 64 kg ha–1 yr–1 in the mangroves in Kosrae, Micronesia (Ronnback, 1999). Both the Kalibo, Aklan mangroves and the mangroves in the present study are open access areas with no fishing regulation enforced. The present study area has been subject to heavy fishing pressure as evidenced by the relatively low yield and the decreasing crab size caught throughout the sampling period. The decreasing trend in CPUE mirrors that reported by Walton et al. (2006b) at Kalibo, Aklan over the same time period. In traps that are non size-selective and that are not saturated due to competitive interactions among crabs (Robertson, 1989), CPUE may be a good index of relative abundance (Le Vay et al., 2001). Despite the heavy fishing pressures on both sites, the lower abundance and yield in the present study may be explained by lower recruitment in the basin mangrove with a single channel connecting to the open coast and the lack of suitable crab habitat along the adjacent coastline of exposed sand beaches. The lack of any significant change in CPUE (crab gear–1 d–1) over time indicates that recruitment of crabs was relatively stable over the 4-yr study period. However, the significant negative correlation of CPUE (g gear–1 d–1) and decreasing mean CW and BW through time indicates growth-overfishing. The mangroves of Naisud and Bugtong Bato represent an interesting opportunity for experimental stock enhancement trials, given the observed high level of fidelity of S. olivacea to the intertidal mangrove habitat and the relatively low abundance of this 902 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 species compared to other sites, apparently due to limited recruitment. Recent trials have shown that stocking of juvenile crabs can increase yields without any control of fishing effort (Lebata, 2006). However, the apparent growth-overfishing indicated in the present study, which is likely to be representative of unregulated open access fisheries in many mangrove areas in southeast Asia, suggests that any hatchery-release program should be considered only as a component of broader initiatives to develop effective co-management.

Acknowledgments

This study was supported by EU-INCO contractIC-CT-2001-10022.�� We thank�� the crab collectors in the villages of Naisud and Bugtong Bato, Ibajay, Aklan for collecting crabs during sampling regardless of time and weather, J. Altamirano, J. Inguillo, R. Ramos, W. Sayson, and A. Traje for the assistance with mapping the study site.

Literature Cited

Agbayani, R. F., D. D. Baliao, G. P. B. Samonte, R. E. Tumaliuan, and R. D. Caturao. 1990. Eco- nomic feasibility analysis of the monoculture of mudcrab (Scylla serrata) Korskål. Aqua- culture 91: 223–231. Alcala, A. and S. Alcazar. 1984. Edible molluscs, crustaceans, and holothurians from North and South Bais Bays, Negros Occidental, Philippines. Siliman J. 31: 25–52. Arriola, F. J. 1940. A preliminary study of the life history of Scylla serrata (Korskål). Philippine J. Sci. 73: 437–458. Baliao, D. D., E. M. Rodriguez, and D. D. Gerochi. 1981. Culture of the mud crab, Scylla serrata (Korskål) at different stocking densities in brackish water ponds. SEAFDEC Q. Res. Rep. 5: 10–14. Bell, J. 2004. Management of restocking and stock enhancement programs: the need for different approaches. Pages 213–114 in K. M. Leber, S. Kitada, H. L. Blankenship, and T. Svasand, eds. Stock enhancement and sea ranching: developments, pitfalls and opportunities. Black- well Publishing Ltd., Oxford. _____ and M. Gervis. 1999. New species for coastal aquaculture in the tropical Pacific -- constraints, prospects and considerations. Aquacult. Int. 7: 207–223. Blankenship, H. L. and K. M. Leber. 1995. A responsible approach to marine stock enhancement. Am. Fish. Soc. Symp. 15: 167–175. Chandrasekaran, V. S. and R. Natarajan. 1994. Seasonal abundance and distribution of seeds of mud crab Scylla serrata in Pichavaram mangrove, Southeast India. J. Aquacult. Trop. 9: 343–350. Dat, H. D. 1999. Description of mud crab (Scylla spp.) culture methods in Vietnam. Pages 67–71 in C. P. Keenan and A. Blackshaw, eds. Mud crab aquaculture and biology. Darwin. Devi, S. L. 1985. The fishery and biology of crabs of Kakinada region. Indian J. Fish. 32: 18–32. Escritor, G. L. 1972. Observations on the culture of the mud crab, Scylla serrata. Pages 355–361 in T. V. R. Pillay, ed. Coastal Aquaculture in the Indo-Pacific Region. Estampador, E. P. 1949a. Studies on Scylla (Crustacea: Portunidae). I. A revision of the genus. Philippine J. Sci. 78: 95–108. ______. 1949b. Studies on Scylla (Crustacea; Portunidae). II. Comparative studies on spermatogenesis and oogenesis. Philippine J. Sci. 78: 301–353. Heasman, M. P. 1980. Aspects of the general biology and fishery of the mud crabScylla serrata (Korskål) in Moreton Bay, Queensland. Ph.D. Diss., University of Queensland, St. Lucia. 506 p. LEBATA ET AL.: POPULATION AND FISHERIES OF SCYLLA SPP. 903

______, D. R. Fielder, and R. K. Shepherd. 1985. Mating and spawning in the mudcrab, Scylla serrata (Korskål) (Decapoda, Portunidae), in Morton Bay, Queensland. Aust. J. Mar. Freshw. Res. 36: 773–783. Hill, B. J. 1975. Abundance, breeding and growth of the crab Scylla serrata in two South Afri- can estuaries. Mar. Biol. 32: 119–126. ______, M. J. Williams, and P. Dutton. 1982. Distribution of juvenile, subadult and adult Scylla serrata (Crustacea: Portunidae) on tidal flats in Australia. Mar. Biol. 69: 117–120. Hyland, S. J., B. J. Hill, and C. P. Lee. 1984. Movement within and between different habitats by the portunid crab Scylla serrata. Mar. Biol. 80: 57–61. Imai, H. and M. Takeda. 2005. A natural hybrid mud crab (Decapoda, Portunidae) from Japan. J. Crustac. Biol. 25: 620–624. ______, C. Jin-Hua, K. Hamasaki, and K. I. Numachi. 2004. Identification of four mud crab species (genus Scylla) using ITS-1 and 16S rDNA markers. Aquat. Living Resour. 17: 31–34. Jennings, S., M. J. Kaiser, and J. D. Reynolds. 2001. Marine Fisheries Ecology. Blackwell Science Ltd., Oxford. 417 p. Keenan, C. P. 1999. The fourth species ofScylla . Pages 48–58 in C. P. Keenan and A. Blackshaw, eds. Mud crab aquaculture and biology. Darwin. ______, P. J. F. Davie, and D. L. Mann. 1998. A revision of the genus Scylla de Haan, 1833 (Crustacea: Decapoda: Brachyura: Portunidae). Raffles Bull. Zool. 46: 217–245. Kosuge, T. 2001. Brief assessment of stock of mud crabs Scylla spp. in Matang Mangrove Forest, Malaysia and proposal for resources management. JARQ-Japan Agr. Res. Q. 35: 145–148. Kuhlmann, K. J. 2002. Evaluations of marine reserves as basis to develop alternative livelihoods in coastal areas of the Philippines. Aquacult. Int. 10: 527–549. Lebata, M. J. H. L. 2006. Stock enhancement of the mud crabs Scylla spp. in the mangroves of Naisud and Bugtong Bato, Ibajay, Aklan, Philippines. Ph.D. Diss., University of Wales, Bangor, Anglesey, United Kingdom. 274 p. Le Vay, L. 2001. Ecology and management of mud crabs, Scylla spp. Asian Fish. Sci. 14: 101– 111. ______, V. N. Ut, and D. A. Jones. 2001. Seasonal abundance and recruitment in an estuarine population of mud crabs, Scylla paramamosain, in the Mekong Delta, Vietnam. Hydrobiologia 449: 231–239. Macintosh, D. J., J. L. Overton, and H. V. T. Thu. 2002. Confirmation of two common mud crab species (genus Scylla) in the mangrove ecosystem of the Mekong Delta, Vietnam. J. Shellfish Res. 21: 259–265. Macnae, W. 1968. A general account of the fauna and flora of mangrove swamps and forests in the Indo-West-Pacific region. Pages 73–270 in S. F. S. Russel and S. M. Yonge, eds. Ad- vances in marine biology. Academic Press Inc., London. Moser, S., D. Macintosh, S. Prinpanapong, and N. Tongdee. 2002. Estimated growth of the mud crab Scylla olivacea in the Ranong mangrove ecosystem, Thailand, based on a tagging and recapture study. Mar. Freshw. Res. 53: 1083–1089. ______, ______, S. Laoprasert, and N. Tongdee. 2005. Population ecology of the mud crab Scylla olivacea: a study in the Ranong mangrove ecosystem, Thailand, with emphasis on juvenile recruitment and mortality. Fish. Res. 71: 27–41. Ng, P. K. L. 1998. Crabs. Pages 1045–1155 in K. E. Carpenter and V. N. Niem, eds. Fao species identification guide for fishery purposes, the living marine resources of the Western Cen- tral Pacific, vol. 2 Cephalopods, crustaceans, holothurians and sharks. Food and Agricul- ture Organization of the United Nations, Rome. Onyango, S. D. 2002. The breeding cycle ofScylla serrata (Korskål, 1755) at Ramisi River estuary, Kenya. Wet. Ecol. Manage. 10: 257–263. Overton, J. L. and D. J. Macintosh. 2002. Estimated size at sexual maturity for female mud crabs (Genus Scylla) from two sympatric species within the Ban Don Bay, Thailand. J. Crustac. Biol. 22: 790–797. 904 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

______, ______and R. S. Thorpe. 1997. Multivariate analysis of the mud crab Scylla serrata (Brachyura: Portunidae) from four locations in Southeast Asia. Mar. Biol. 128: 55–62. Perrine, D. 1979. The mangrove crab (Scylla serrata) on Ponape. Marine Resources Division, Ponape, East Caroline Islands. 66 p. Primavera, J. H. 2000. Development and conservation of Philippine mangroves: institutional issues. Ecol. Econ. 35: 91–106. ______, R. B. Sadaba, M. J. H. L. Lebata, and J. P. Altamirano. 2004. Handbook of Man- groves in the Philippines - Panay. SEAFDEC Aquaculture Department, Iloilo. 106 p. Robertson, W. D. 1989. Factors affecting catches of the crab Scylla serrata (Korskål) (Decapo- da, Portunidae) in baited traps - soak time, time of day and accessibility of the bait. Estuar. Coast. Shelf Sci. 29: 161–170. ______. 1996. Abundance, population structure and size at maturity of Scylla serrata (Korskål) (Decapoda: Portunidae) in Eastern Cape estuaries, South Africa. S. Afr. J. Zool. 31: 177–185. ______and A. Kruger. 1994. Size at maturity, mating and spawning in the portunid crab Scylla serrata (Forsskål) in Natal, South Africa. Estuar. Coast. Shelf Sci. 39: 185–200. Ronnback, P. 1999. The ecological basis for economic value of seafood production supported by mangrove ecosystems. Ecol. Econ. 29: 235–252. Stephenson, W. and B. Campbell. 1960. The Australian portunids (Crustacea: Portunidae) IV: Remaining genera. Aust. J. Mar. Fresh. Res. 11: 73–122. Walton, M. E. M., L. Le Vay, L. M. Truong, and V. N. Ut. 2006a. Significance of mangrove- mudflat boundaries as nursery grounds for the mud crab Scylla paramamosain. Mar. Biol. 149: 1199–1207 ______, ______, J. H. Lebata, J. Binas, and J. H. Primavera. 2006b. Seasonal abundance, distribution and recruitment of mud crabs (Scylla spp.) in replanted mangroves. Estuar. Coast. Shelf Sci. 66: 493–500.

Addresses: (M.J.H.L.L., L.L.V, M.E.W) School of Ocean Sciences, University of Wales, Bangor, Menai Bridge, Anglesey LL59 5AB, United Kingdom. (J.H.P., J.B.B.) Aquaculture Department, Southeast Asian Fisheries Development Center, Tigbauan, Iloilo 5021, Philippines. Corre- sponding Author: (L.L.V.) E-mail: . BULLETIN OF MARINE SCIENCE, 80(3): 905–914, 2007

NOTE

The impacts of red mangrove (Rhizophora mangle) deforestation on zooplankton communities in Bocas del Toro, Panama

Elise F. Granek and Kaitlin Frasier

Deforestation impact studies have generally focused on tropical rainforests or temperate coniferous woodlands. However, extensive clear-cutting is currently occurring in a wide-ranging, but far less recognized habitat: the world’s mangrove forests. These coastal forests thrive in areas of low wave action and high sediment availability, where mangrove trees develop dense and productive ecosystems (Alongi, 2002). Extensive aerial and subtidal prop root networks, a dense canopy, and varying water conditions allow these forests to support unique assemblages of flora and fauna. Even as mangrove deforestation continues to alter coastlines, these forests are increasingly recognized as important nursery habitats and feeding grounds for many larval, juvenile and adult fish and invertebrate species (e.g., Nagelkerken et al., 2001, 2002; Mumby et al., 2004; Nagelkerken and van der Velde, 2004). Larval populations of a wide variety of marine species recruit to these sheltered, structurally complex, shaded and nutrient-rich ecosystems (Krishnamurthy, 1982; Dennis, 1992). Whether zooplankton communities differ between intact and cleared mangrove areas is unknown. Structural complexity should affect flow and hence food availability and larval retention rates. However, results from different systems are inconsistent. In kelp forests, structural complexity inhibits deposition of suspended particles, possibly reducing food availability for benthic organisms and retarding zooplankton dispersal (Eckman et al., 1989). As with inhibited flow in kelp forests, Toffart (1983) suggests that there is a rapid decrease in species diversity from the seaward edge of a mangrove forest towards the shore because of reduced flow. This would particularly affect less active swimmers. However, few direct measurements of zooplankton inside and outside mangroves exist (e.g., Ambler et al., 1991). Mangroves may be preferred settlement sites for some highly mobile species that can actively select this habitat. Capacity for swift directional travel among zooplankton (Luckenbach and Orth, 1992; Ferrari et al., 2003) may allow these species to actively select mangrove habitat over less complex, more open environments (i.e., cleared mangrove areas). Low flow in mangroves may increase (or decrease) zooplankton abundance and create retention zones where zooplankton can develop near suitable adult habitat (Paula et al., 2004). Zooplankton survival may be high in mangroves due to favorable substrate and increased niches (due to root complexity), greater food availability, and reduced predation (Laegdsgaard and Johnson, 1995; Cocheret de la Morinière et al., 2004). The dense mangrove root networks retain nutrients and sediments carried in runoff from adjacent land or produced in situ, having the dual effect of fueling productivity (Bouillon et al., 2000) and creating murky water conditions, reducing visibility for predators. The structural complexity of mangrove roots may provide settling larvae with shelter from predators and open water currents. On the other hand, differences in structural complexity within man-

Bulletin of Marine Science 905 © 2007 Rosenstiel School of Marine and Atmospheric Science of the University of Miami 906 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 grove forests may also lead to variable larval supply and diversity within different areas of intact mangroves (Krumme and Liang, 2004; Osore et al., 2004). To date few studies have examined zooplankton communities in mangrove habitats (but see Ambler et al., 1991; Paula et al., 2001; Barletta-Bergan et al., 2002; Osore et al., 2004). In this study, we examined the effects of mangrove habitat loss on zooplankton communities by comparing diversity, abundance, and community composition between areas cleared of mangroves and areas with intact mangroves in the Bocas del Toro region of Panama. We tested the hypothesis that different meroplankton and holoplankton communities inhabit intact and cleared mangrove environments. Because higher complexity habitats are sometimes characterized by higher biological diversity (i.e., Kohn and Leviten, 1976; Taniguchi and Tokeshi, 2004; Gratwicke and Speight, 2005a,b; Kostylev et al., 2005; Lassau et al., 2005; Le Hir and Hily, 2005), we expected to find higher diversity in the intact mangrove habitat.

Methods

Study Area.—This study was conducted adjacent to Isla Colón in the Bocas del ToroP rov- ince off the Caribbean coast of Panama. The coastline at the study area was characterized by red mangrove, Rhizophora mangle Linneaus, 1773 trees, except where stands had been removed for agriculture, construction, or viewsheds. Cleared mangrove areas ranged from 100 to 300 m in length along the shore and were bordered on either end by intact mangrove habitat. Mangrove removal occurred approximately 8 yrs prior to our study. Intact mangrove areas were characterized by submerged prop roots colonized by oysters, sponges, sporadic coral heads, and, infrequently, epibiotic algae, with occasional unforested channels between trees. Cleared areas were similar to forested areas in depth and distance from the shoreline, but they lacked the complex 3-dimensional underwater root structure (generally limited to a few remaining snags), overhead cover under the mangrove canopies, and high sediment levels and nutrients resulting from organic production of a healthy mangrove community (Krish- namurthy, 1982; Granek, 2006). Mangrove-removed areas were characterized by submerged decaying prop roots on the substrate with significant macroalgal growth inshore and seagrass (Thalassia testudinum) growth further from shore. Tidal exchange in the region is small: ~50 cm and not variable between forested and nonforested areas. Six sites on Isla Colón were selected for this study because they met the following criteria: (1) at least 100 m long stretch of cleared R. mangle adjacent to stretches of at least 100 m of intact red mangroves; (2) fringing or patch coral reefs within 100 m of the seaward edge; and (3) > 2 km from major human development or construction, to exclude potential immediate anthropogenic sources of nutrients. Nearby development in the study region was limited, primarily consisting of subsistence farming and mangrove clear-cutting. Commercial and industrial development was > 10 km from all sampling sites. Below, we refer to areas of intact mangrove as +mangrove areas and to areas cleared of mangroves as –mangrove areas. Zooplankton Sampling.—Previous assessments demonstrate that community composition differs between light trap and plankton tow sampling (Hickford and Schiel, 1999; Porter et al., 2002). Because zooplankton display a range of swimming abilities and photosensitiv- ity, we simultaneously used light traps and plankton tows to assess zooplankton communities in the +mangrove and -mangrove areas. Positively phototactic swimming zooplankton are drawn into light traps, while non-phototactic and slow-moving or negatively phototactic zooplankton are more effectively sampled by plankton tows (Doherty, 1987). In June 2004, sampling was conducted for six nights around the new moon, the period of the lunar cycle when fish and invertebrate spawning is most common, and therefore, when the larval community is likely at its peak density (McFarland et al., 1985). We began collections 2 d prior to and continued 3 d after the new moon. Two sites were sampled per night (two intact and their NOTES 907 two associated cleared areas), and each site was sampled twice in the course of the study (n = 12 site × night combinations) using both sampling methods simultaneously. Zooplankton Light Traps.—Zooplankton light traps have the potential to trap positively phototactic, mobile organisms (Watson et al., 2001; Porter et al., 2002). The light trap design was based on that used by Roegner et al. (2003). The traps were constructed using 7.6-l (2 gal) clear plastic water jugs, inverted, with an attached, 220-µm, mesh-lined cod-end made of perforated PVC tubing (Fig. 1). A yellow glow stick, suspended inside the bottle from the top of each inverted trap, was used as a light source. Three funnel-shaped entry points were available to zooplankton in the bottle’s sides, each leading inward to a hole measuring approximately 1 cm in diameter. The small size of entry points and the funnel shapes were designed to limit, as much as possible, the ability of zooplankton to leave the traps after entering. Zoo- plankton were flushed into the cod-end of the trap when it was lifted from the water. Traps were deployed for 1 hr after sunset, between 1900 and 2030. In intact mangrove areas, light traps were anchored by suspended dive weights within the root structure approximately 1–1.5 m above the substrate. In the cleared areas, traps were deployed within the area previously occupied by mangroves. Plankton Tows.—Diver-pulled plankton tows were conducted in the vicinity of the traps for 1 min (approximately 20 m) during the time period in which the light trap was deployed. The 200-µm mesh plankton net had an opening diameter of 30 cm. In intact mangrove areas, the net was pulled through partially open waters found behind the most seaward trees and through small channels within the mangrove forest. Tows were conducted as close to the light traps as feasible within the root structure, given the size of the net. In the cleared mangrove areas, the tows were pulled along a straight line, parallel to shore and adjacent to light traps. All tows were pulled at a similar speed, during calm nights, so that water flow was consistent between + and –mangrove areas to control for water volume sampled. Sample Processing.—The contents of the cod-ends of the traps and tows were preserved in 2%–4% formalin solution. A light microscope was used for sample identification to count all individuals in each sample and identify them to phylogenetic order when possible. Deca-

Figure 1. Diagram of larval trap design (modified from Roegner et al., 2003; not to scale). Our design utilized a smaller bottle, glow stick instead of fluorescent bulb, and modified cod end. 908 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 pods were further categorized by developmental (i.e., zoea, megalopae, or postlarval) and reproductive stage. Statistical Analysis.—Prior to analysis, data were log transformed for analysis of variance (ANOVA) tests and square-root transformed for non-parametric multidimensional scaling analysis to meet model assumptions. Three-factor ANOVA was used to determine how much of the variability in taxonomic abundance was accounted for by physical location (site), mangrove presence (+mangrove vs –mangrove), and sampling night. A Shannon-Wein- er diversity index was used to determine differences in taxon diversity between +mangrove and –mangrove areas. A non-parametric multidimensional scaling analysis (nMDS) was run to examine differences between communities at each site, and whether +mangrove sites were more similar to each other than to –mangrove sites. A paired t-test compared overall zooplankton abundance in the samples. All analyses were run separately for light trap and plankton tow data. Communities were then separated into meroplankton and holoplankton, and a nMDS was run for each sub-community (meroplankton; holoplankton). A single factor ANOVA was used to test how high temperature events differed between -mangrove and +mangrove habitat.

Results

Light Traps.—In light trap samples, abundance of meroplankton taxa including amphipods, isopods, ostracods, crab and mysid postlarvae, and Daphnia were greater in +mangrove areas (ANOVA: P < 0.05; Fig. 2A; see online Appendix 1); Daphnia, crab megalopae and isopod abundance also varied by site and/or night. Abundance of three meroplankton taxa (crab zoeae, shrimp zoeae and megalopae, and polychaetes) and two holoplankton taxa (copepods and cumaceans) were greater in –mangrove areas (Fig. 2A, online Appendix 1); cumacean abundance also varied by site. For all other taxa sampled in light traps, there was no difference in taxon abundance between + and –mangrove areas (Fig. 2A). Total zooplankton abundance was not significantly different between +mangrove (mean = 2479 ± 596) and –mangrove (mean = 9857 ± 3801) areas (Paired t-test: t = −1.8141, df = 10, P = 0.10). Overall zooplankton diversity in light traps was more than 50% higher in intact mangrove areas relative to cleared areas (Shannon Weiner diversity index: intact = 1.4, removed = 0.92). Structural community differences between +mangrove and - mangrove areas were revealed by nMDS analysis (r2 = 0.46, Stress in randomized runs: P < 0.01; Fig. 3A). Community differences in light traps were driven primarily by higher abundance or presence in +mangrove areas of meroplankton including amphipods, reproductive mysids, and porcellanid megalopae as well as Daphnia and jellyfish. nMDS analysis examining only the meroplankton community demonstrated significant differences between +mangrove and -mangrove areas (r2 = 0.50, Stress in randomized runs: P < 0.01; Fig. 3B) as did an analysis of the holoplankton community (r2 = 0.42, Stress in randomized runs: P < 0.01; Fig. 3C). Zooplankton Tows.—In plankton tows, several meroplankton taxa (including amphipods, euphausids, mysids, and ostracods) were more abundant in +mangrove than -mangrove areas (Fig. 2B; online Appendix 1) with some taxa being 10–100 times greater in +mangrove areas. For all other taxa sampled in plankton tows, there was no significant difference in taxon abundance between + and -mangrove areas (Fig. 2B). There was also no difference in the total zooplankton abundance between +mangrove (mean = 351 ± 86) and –mangrove (477 ± 109) areas (Paired t-test: t = 1.3487, df = 10, P = 0.20). NOTES 909

Figure 2. Difference in mean taxon abundance between +mangrove areas and –mangrove areas. * indicates significant difference (P < 0.05). ° indicate taxa that are not known as important food items for reef fish; all others are considered key food items for reef fish (Randall, 1967). Overall zooplankton diversity in tows was more than 50% higher in +mangrove areas compared to -mangrove areas (Shannon Weiner diversity index: intact = 2.13, removed = 1.34). nMDS analysis revealed differences in zooplankton communities between +mangrove and -mangrove areas (r2 = 0.51, Stress in randomized runs: P < 0.01; Fig. 3D). Community differences in plankton tows were primarily driven by higher abundances in +mangrove areas of holoplankton including Daphnia, jellyfish, and rotifer larvae; and meroplankton including comatulids, euphausids, hydrozoan 910 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Figure 3. Nonparametric multidimensional scaling ordinations of differences in communities between intact and cleared mangrove areas from light traps (A–C) and plankton tows (D–F). Closed circles are +mangrove areas and open circles are -mangrove areas. larvae, mantis shrimp, and snails. nMDS analysis examining only meroplankton revealed differences in communities between +mangrove and –mangrove areas (r2 = 0.45, Stress in randomized runs: P < 0.01; Fig. 3E) although there was no difference in holoplankton communities between +mangrove and -mangrove areas (r2 = 0.21, Stress in randomized runs: P < 0.03; Fig. 3F).

Discussion

Taxonomic diversity of zooplankton communities, and, in particular, meroplankton, was different between adjacent areas with and without intact mangroves. We found that amphipods, euphausids, mysids, and ostracods were more abundant in undisturbed mangrove areas, regardless of sampling method. Crab megalopae and isopods were more abundant in light trap samples from undisturbed mangroves, whereas copepods, cumaceans, polychaetes, and crab and shrimp zoeae were more abundant in cleared areas. In addition to the dominant taxa sampled, less common taxa (e.g., jellyfish, comatulids, euphausids) also contributed to the patterns of community differences observed between intact and cleared areas, possibly due to their near or complete absence from cleared areas coupled with the high variability across samples for some more abundant species. Although light traps in +mangrove areas were likely sampling a smaller effective area than light traps in –mangrove areas due to decreased light penetration from roots and higher turbidity in +mangrove areas, the general patterns for these taxa were similar in both tows and traps, suggesting that the results are informative and a true representation of community differences. Furthermore, our use of higher order taxonomic groupings in the diversity analyses as well as the limited temporal scale of this study (seasonal, diel) may underestimate diversity differences at the generic or species level or across seasons between intact and cleared mangrove areas. Three major processes could be responsible, separately or in conjunction, for the difference in community structure between +mangrove and –mangrove areas: (1) differential proximity to source populations, (2) differential mortality, and/or (3) differential habitat preference among taxa. For taxa where adults inhabit and spawn NOTES 911 in mangrove habitats, proximity to such areas may well influence the relative abundance of early life stages. The community composition sampled in the two habitats could be affected by differential mortality through increased three-dimensional structure (Primavera, 1997; Sheridan and Hays, 2003) and turbidity in mangroves, both of which increase the ability of prey to escape or hide from predators, potentially increasing survival rates. Mangroves also provide shelter and protection for juvenile reef fish, including several zooplankton feeders (Randall, 1967), and are an important nursery and feeding area (Nagelkerken et al., 2001, 2002; Mumby et al., 2004; Nagelkerken and van der Velde, 2004). If the abundance of these juvenile planktivorous reef fish is greater in +mangrove areas (see Mumby et al., 2004), then predation pressure on zooplankton in mangrove areas should be high; however, it is unknown how lower predator abundance combined with reduced shelter against predators interact to affect the abundance of preferred prey items such as shrimp larvae, cumaceans, copepods, and polychaete worms in cleared areas where reef fish populations are purported to be lower (e.g., Mumby et al., 2004). Teasing apart the contribution of these conflicting mechanisms to differences in community structure between +mangrove and –mangrove areas necessitates further experimental investigation. Habitat preference may also structure the zooplankton community. Mangrove habitat may be favorable to meroplankton because of lower thermal stress (Granek, 2006), increased structure for predator avoidance (Mohan et al., 1997; Kingsford et al., 2002), and greater food availability (Schwamborn et al., 2002). Zooplankton are capable of responding to temperature cues (Yurista, 2000; Metaxas, 2001; Ouimet, 2001; Bell and Weithoff, 2003), and thermally stressful events are significantly more frequent in –mangrove areas (Granek, 2006). Therefore, zooplankton may be attracted to the less thermally stressful environment of mangrove areas relative to cleared areas. Variation in swimming abilities among meroplankton (Holzman et al., 2005) may contribute to the taxonomic variability in intact mangrove areas, as only strong swimmers may be able to actively select mangrove root structure. Because meroplankton ultimately need to settle and holoplankton do not, holoplankton may be more patchily distributed across inshore habitats with varying benthic complexity (Ambler et al., 1991; Stewart, 1996) and swarming behavior in certain taxa may contribute to this patchiness (Ambler, 2002). Patchy distribution in the water column and micro-site differences may explain some of the observed between-site and -night variability. Our finding that meroplankton are more abundant in +mangrove areas whereas holoplankton are similar (in tows) or only slightly more abundant (in traps) in –mangrove areas supports this tenet. The observed differences in zooplankton diversity and community composition between +mangrove and -mangrove areas suggest a potential impact of mangrove removal on coastal marine communities. Previous studies in diverse habitats demonstrate that habitat loss or transformation can lead to changes in community composition and species diversity (Boulinier et al., 2001; Silliman and Bertness, 2004; Stoner and Joern, 2004; Watson et al., 2004). Mangrove deforestation may change zooplankton communities due to a decrease in physical features, structure, changes in food availability, and water flow. Most zooplankton taxa sampled in this study are common or preferred food items for juvenile and adult reef fish (Randall, 1967; Fig. 2). Further research is needed to determine whether changes in zooplankton community composition observed in this study may lead to reduced food availability for 912 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 juvenile and adult reef fish feeding in mangrove habitat. The source of larval invertebrates for adult populations on adjacent reefs may decline as the meroplankton population shifts in nearby cleared mangrove areas. Further research is needed to determine whether changes in zooplankton abundance and diversity cascades into changes in fish communities on adjacent reefs following shifts in preferred prey items subsequent to mangrove removal.

Acknowledgments

Thanks to A. Shanks and C. Roegner for sharing their light trap design; C. Christenson of Searings Electric, Corvallis, OR for design construction ideas; N. Ehlers for construction assistance; G. Boehlert for suggestions and ideas; and E. Brown for late-night boat tending. This manuscript was significantly improved through discussions and suggestions from G. Rilov, S. Dudas, J. Tyburczy, G. L. Eckert, B. Menge, J. Lubchenco, M. Kavanaugh, and E. Seabloom. Special thanks to G. Rilov and S. Dudas for multiple readings of this manuscript. Funding for this study was provided in part by a Fulbright Fellowship, an NSF-GRF, a Smithsonian Tropi- cal Research Institution—Supplemental Research Award, an Anchor Environmental Scholar- ship, and funding from TheD avid and Lucile Packard Foundation (to EFG), and OSU Howard Hughes Medical Institute, OSU URISC and IURP Fellowships (to KF).

Literature Cited

Alongi, D. M. 2002. Present state and future of the world’s mangrove forests. Environ. Cons. 29: 331–349. Ambler, J. W., F. D. Ferrari, and J. A. Fornshell. 1991. Population-structure and swarm formation of the cyclopoid copepod Dioithona oculata near mangrove cays. J. Plank. Res. 13: 1257–1272. Ambler, J. W. 2002. “Zooplankton swarms: characteristics, proximal cues and proposed advantages.” Hydrobiologia 480: 155–164. Barletta-Bergan, A., M. Barletta, and U. Saint-Paul. 2002. Community structure and temporal variability of ichthyoplankton in North Brazilian mangrove creeks. J. Fish Biol. 61: 33–51. Bell, E. M. and G. Weithoff. 2003. Benthic recruitment of zooplankton in an acidic lake.J. Exp. Mar. Biol. Ecol. 285: 205–219. Bouillon, S., P. C. Mohan, N. Sreenivas, and F. Dehairs. 2000. Sources of suspended organic matter and selective feeding by zooplankton in an estuarine mangrove ecosystem as traced by stable isotopes. Mar. Ecol. Prog. Ser. 208: 79–92. Boulinier, T., J. D. Nichols, J. E. Hines, J. R. Sauer, C. H. Flather, and K. H. Pollock. 2001. For- est fragmentation and bird community dynamics: Inference at regional scales. Ecology 82: 1159–1169. Cocheret de la Morinière, E. C., I. Nagelkerken, H. van der Meij, and G. van der Velde. 2004. What attracts juvenile coral reef fish to mangroves: habitat complexity or shade? Mar. Biol. 144: 139–145. Dennis, G. D. 1992. Island mangrove habitats as spawning and nursery areas for commercially important fishes in the Caribbean. Proc. 41st Ann. Gulf Carib. Fish. Inst. 41: 205–225. Doherty, P. J. 1987. Selective but useful devices for quantifying the distributions and abundances of larval fishes. Bull. Mar. Sci. 41: 423–431. Eckman, J., D. Duggins, and A. Sewell. 1989. Ecology of understory kelp environments. 1. Ef- fects of kelps on flow and particle transport near the bottom. J. Exp. Mar. Biol. Ecol. 129: 173–187. Ferrari, F. D. J. A. Fornshell, L. Ong, and J. W. Ambler. 2003. Diel distribution of copepods across a channel of an overwash mangrove island. Hydrobiologia 499: 147–159. NOTES 913

Granek, E. F. 2006. Linkages between mangrove forests and coral reefs: quantifying disturbance effects and energy flow between systems. Ph.D. Diss., Oregon State University. Gratwicke, B. and M. R. Speight. 2005a. Effects of habitat complexity on Caribbean marine fish assemblages. Mar. Ecol. Prog. Ser. 292: 301–310. ______and ______. 2005b. The relationship between fish species richness, abundance and habitat complexity in a range of shallow tropical marine habitats. J. Fish Biol. 66: 650–667. Hickford, M. J. H. and D. R. Schiel. 1999. Evaluation of the performance of light traps for sampling fish larvae in inshore temperate waters. Mar. Ecol. Prog. Ser. 186: 293–302. Holzman, R., M. A. Reidenbach, S. G. Monismith, J. R. Koseff, and A. Genin. 2005. Near-bottom depletion of zooplankton over a coral reef - II: relationships with zooplankton swimming ability. Coral Reefs 24: 87–94. Kingsford, M. J., J. M. Leis, A. Shanks, K. C. Lindeman, S. G. Morgan, and J. Pineda. 2002. Sen- sory environments, larval abilities and local self-recruitment. Bull. Mar. Sci. 70: 309–340. Kostylev, V. E., J. Erlandsson, M. Y. Ming, and G. A. Williams. 2005. The relative importance of habitat complexity and surface area in assessing biodiversity: Fractal application on rocky shores. Ecol. Complex. 2: 272–286. Kohn, A. J. and P. J. Leviten. 1976. Effect of habitat complexity on population density and species richness in tropical intertidal predatory gastropod assemblages. Oecologia 25: 199–210. Krishnamurthy, K. 1982. Environmental management and utilization of tropical coastal ecosystems. Atlantica 5: 67–68. Krumme, U. and T. H. Liang. 2004. Tidal-Induced Changes in a Copepod-Dominated Zoo- plankton Community in a Macrotidal Mangrove Channel in Northern Brazil. Zool. Stud. 43: 404–414. Lassau, S. A., D. F. Hochuli, G. Cassis, and C. A. M. Reid. 2005. Effects of habitat complexity on forest beetle diversity: do functional groups respond consistently? Div. Distrib. 11: 73–82. Le Hir, M. and C. Hily. 2005. Macrofaunal diversity and habitat structure in intertidal boulder fields. Biol. Conserv. 14: 233-250. Laegdsgaard, P. and C. R. Johnson. 1995. Mangrove habitats as nurseries: Unique assemblages of juvenile fish in subtropical mangroves in eastern Australia. Mar. Ecol. Prog. Ser. 126: 67–81. Luckenbach, M. W. and R. J. Orth. 1992. Swimming velocities and behavior of blue crab (Cal- linectes sapidus Rathbun) megalopae in still and glowing water. Estuaries 15: 186–192. McFarland, W. N., E. B. Brothers, J. C. Ogden, M. J. Shulman, E. L. Bermingham, and N. M. Lotchian-Prentiss. 1985. Recruitment patterns in young French grunts, Haemulon flavolineatum (Family Haemulidae), at St Croix, Virgin Islands. Fish. Bull. 83: 413–426. Metaxas, A. 2001. Behaviour in flow: perspectives on the distribution and dispersion of mero- planktonic larvae in the water column. Can. J. Fish. Aquat. Sci. 58: 86–98. Mohan, P. C., R. G. Rao, and F. Dehairs. 1997. Role of Godavari mangroves (India) in the production and survival of prawn larvae. Hydrobiologia 358: 317–320. Mumby, P. J., A.J. Edwards, J. E. Arias-Gonzalez, K. C. Lindeman, P. G. Blackwell, A. Gall, M. I. Gorczynska, A. R. Harborne, C. L. Pescod, H. Renken, C. C. C. Wabnitz, and G. Llewellyn. 2004. Mangroves enhance the biomass of coral reef fish communities in the Caribbean. Nature 427: 533–536. Nagelkerken, I. and G. van der Velde. 2004. Are Caribbean mangroves important feeding grounds for juvenile reef fish from adjacent seagrass beds? Mar. Ecol. Prog. Ser. 274: 143– 151. ______, S. Kleijnen, T. Klop, R. van den Brand, E. Cocheret de la Morinière, and G. van der Velde, 2001. Dependence of Caribbean reef fishes on mangroves and seagrass beds as nursery habitats: a comparison of fish faunas between bays with and without mangroves/ seagrass beds. Mar. Ecol. Prog. Ser. 214: 225–235. ______, C. M. Roberts, G. van der Velde, M. Dorenbosch, M. C. van Riel, E. Cocheret de la Morinere, and P. H. Nienhuis. 2002. How important are mangroves and seagrass beds for 914 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

coral-reef fish? The nursery hypothesis tested on an island scale. Mar. Ecol. Prog. Ser. 244: 299–305. Osore, M. K. W., J. M. Mwaluma, F. Fiers, and M. H. Daro. 2004. Zooplankton composition and abundance in Mida Creek, Kenya. Zool. St. 43: 415–424. Ouimet, C. 2001. Implications of Chaoborus pupation and ecdysis in cold water. Freshw. Biol. 46: 1169–1177. Paula, J., C. Bartilotti, T. Dray, A. Macia, and H. Queiroga. 2004. Patterns of temporal occurrence of brachyuran crab larvae at Saco mangrove creek, Inhaca Island (South Mozam- bique): implications for flux and recruitment. J. Plankton Res. 26: 1163–1174. Paula, J., T. Dray, and H. Queiroga, 2001. Interaction of offoffshore shore and inshore processes control-controlling settlement of brachyuran megalopae in Saco mangrove creek, Inhaca Island (South Mozambique). Mar. Ecol. Prog. Ser. 215: 251–260. Porter, S. S., G. Eckert, C. Byron, and J. Fisher. 2002. A comparison of light traps and plankton tows. Integr. Compar. Bio. 42: 1296–1296. Primavera, J. H. 1997. Fish predation on mangrove-associated penaeids: The role of structures and substrate. J. Exp. Mar. Biol. Ecol. 215: 205–216. Randall, J. E. 1967. Food habits of reef fishes of the West Indies. Stud. Trop. Oceanogr. Miami 5: 665–757. Roegner, G. C., D. A. Armstrong, B. M. Hickey, and A. L. Shanks. 2003. Ocean distribution of Dungeness crab megalopae and recruitment patterns to estuaries in southern Washington State. Estuaries 26: 1058–1070. Schwamborn, R., W. Ekau, M. Voss, and U. Saint-Paul 2002. How important are mangroves as a carbon source for decapod crustacean larvae in a tropical estuary? Mar. Ecol. Prog. Ser. 229: 195–205. Sheridan, P. and C. Hays. 2003. Are mangroves nursery habitat for transient fishes and decapods? Wetlands 23: 449–458. Silliman, B. R. and M. D. Bertness. 2004. Shoreline development drives invasion of Phragmites australis and the loss of plant diversity on New England salt marshes. Biol. Conserv. 18: 1424–1434. Stewart, S. E. 1996. Field behavior of Tripedalia cystophora (class Cubozoa). Mar. Freshw. Be- hav. Physiol. 27: 175–188. Stoner, K. J. L. and A. Joern. 2004. Landscape vs. local habitat scale influences to insect communities from tallgrass prairie remnants. Ecol. Appl. 14: 1306–1320. Taniguchi, H. and M. Tokeshi. 2004. Effects of habitat complexity on benthic assemblages in a variable environment. Freshw. Biol. 49: 1164–1178. Toffart, J. L. 1983. Les peuplements des racines de paletuviers en guadeloupe (Antilles Fran- caises): I - Analyse floristique et faunistique; Methodologie et premiers resultats. Bulletin Ecologique 14: 227–239. Watson, J. E. M., R. J. Whittaker, and T. P. Dawson. 2004. Avifaunal responses to habitat fragmentation in the threatened littoral forests of south-eastern Madagascar. J. Biogeogr. 31: 1791–1807. Watson, M., R. Power, and J. L. Munro. 2001. Use of light-attracted zooplankton for rearing post-settlement coral reef fish. Proc. Gulf Carib. Fish. Inst. 52: 340–351. Yurista, P. M. 2000. Cyclomorphosis in Daphnia lumholtzi induced by temperature. Freshw. Biol. 43: 207–213.

Addresses: (E.F.G.) P.O. Box 751; Environmental Sciences and Resources, Portland State University, Portland, Oregon 97207. (K.F.) Biology Department, Cordley Hall, Oregon State University, Corvallis, Oregon 97331. Corresponding Author: (E.F.G.) Telephone: 503-725- 4241, Fax: 503-725-9040, E-mail: . BULLETIN OF MARINE SCIENCE, 80(3): 915–935, 2007

ABSTRACTS

DISTINGUISHING THE EFFECTS OF ANTHROPOGENIC AND NATURAL (HUR- RICANE) DISTURBANCES ON MANGROVE CREEK FISHES by Aaron J. Adams and R. Kirby Wolfe.—Disturbances are considered important factors influencing biological organization. Hurricanes are disturbances of particular importance in warm latitude coastal ecosystems. However, disturbance effects are often modified by interactions with anthropogenic disturbances, which tend to have different ecological effects than, and often alter the ecological response to, natural disturbances. Four creeks (two “natural”, two “degraded”) had been sampled for 22 mo prior to, and 14 mo after hurricane Charley. Prior to the hurricane, fish assemblages were different between creek types, with more species and higher densities in natural creeks. Species richness and density in natural creeks declined precipitously post-hurricane, and species generation times and isolation from likely source populations has resulted in a delayed recovery in abundance, but species richness is similar to pre-hurricane levels. Fish assemblages in degraded creeks appeared less impacted by the hurricane, likely because they had already experienced a phase shift.— Mote Marine Lab., Charlotte Harbor Field Station, Pineland, Florida, U.S.A.

PIT-TAGS AND REMOTE ANTENNAE: EFFECTIVE TOOLS FOR STUDYING JUVE- NILE FISH USE OF MANGROVE CREEKS by Aaron J. Adams and R. Kirby Wolfe.—Man- groves are purported nurseries for many marine and estuarine fishes, but information on density, growth, and survival is necessary to define nursery habitats. Recent technological advances have made large-scale tagging efforts a viable approach to studying juvenile fishes by increasing recapture rates and enabling the use of individual-identification tags. Passive Integrated Transponder (PIT) tags and autonomous antennae, used extensively in freshwater environments, have been successfully adapted to estuarine mangrove creeks to study juvenile snook, Centropomus undecimalis. Tag retention rate in juveniles > 120 mm standard length was 100%, with no mortality. The antenna detected approximately 67% of tagged fish that swam through, and the overall “recapture” rate by the antenna was > 40%. The high recapture rate allowed population model estimates of survival and population size, and physical recaptures allowed direct measures of growth. This is a viable approach to examining the nursery function of difficult-to-sample mangrove creek habitats.—Mote Marine Lab., Charlotte Har- bor Field Station, Pineland, Florida, U.S.A.

SHRIMP INGRESS INTO MANGROVE FORESTS OF DIFFERENT AGE STANDS, MATANG MANGROVE FOREST RESERVE, MALAYSIA by N. Affendy and V. C. Chong .—Although the importance of mangrove forests as nursery areas for shrimps has been well documented, why shrimps are associated with mangroves is not fully understood. This study investigates the relationship between shrimp abundance and mangroves of different age stands. Shrimp abundance in a cleared area and four different age stands of mangroves (5, 17, 24 and > 30 yrs) were compared. These forest stands are silvicultured, varying in tree densities as well as structural complexity of the root system. Shrimps were sampled using baited shrimp pots. A total of 3609 penaeid, caridean, and alpheid shrimps belonging to 10 species were sampled over a 12-mo period. The majority of prawns moved at least 56.4 m into the mangrove forest at high tide. The highest mean catch of prawns recorded in the 24-yr-old age stand was 51.08 prawns per hour and the lowest was 2.52 prawns per hour in the clear- felled mangrove site. Prawn catch increased with increasing distance from the mangrove river fringe into the forest. More prawns were caught among prop roots than in the open spaces between trees. The role of mangroves as nursery areas and implications for forestry management are discussed.— Institute of Biological Sciences, University of Malaya, Kuala Lumpur, Malaysia.

VARIATIONS IN JUVENILE FISH DENSITY ALONG THE MANGROVE–SEAGRASS– CORAL REEF CONTINUUM IN SOUTHWESTERN PUERTO RICO.—by Alfonso Aguilar- Perera and Richard S. Appeldoorn.—Despite extensive study of the fish community off southwestern Puerto Rico, little information is available on the relative importance of mangroves, seagrass beds, and shallow-water coral reefs as nursery areas. We investigated the extent to which 20 selected, reef-associated fish species use the mangrove and seagrass as nurseries in contrast to the use of shallow-water coral reefs. Stratified sampling was applied to quantify the variability of juvenile fish densities along the mangrove–seagrass–coral reef continuum following an inshore-offshore gradient. We recorded 28,758 individuals (in 7 families), with juveniles accounting for 80%. Significant variations in juvenile densities were evident. The importance of mangroves and seagrass for harboring juveniles was found to be relative and species-specific. In the majority of cases, shallow-water coral reefs showed higher densities than mangroves and seagrass. Ontogenetic migration of juveniles through the continuum was evident. Results highlight the importance of including this continuum within coastal management through marine reserves.—University of Puerto Rico, Mayagüez, Puerto Rico.

THE INTERNATIONAL MARINE SHRIMP ENVIRONMENTAL GENOMICS INITIA- TIVE (IMSEGI): MONITORING ECOSYSTEM, ANIMAL, AND PUBLIC HEALTH — CUR- RENT STATUS AND PERSPECTIVES FOR THE FUTURE by Acacia Alcivar-Warren 1,2,3,5, John Keating 1,6, Louise Maranda 1,3,4,5, Martha Delaney 1,2,5, Dawn Meehan-Meola 1,2,5, Wil- liam Moomaw 1,7, Caleb McClennen 1,7, Jorge Echevarria 1,8, Adan Alvarado 1,9, Cecilia Serrano 1,10, Cesar Valarezo 1,10, Luis Mejia 1,11, Marco Saavedra 1,11, Luis A. Alcivar 1,11, Miriam Alcivar 1,11, Margarita Palmieri 1,12, Suwanna Panutrakul 1,13, Wansuk Senanan 1,13, Praparsiri Bar- nette 1,13, Nongnud Tangrock-Olan 1,13, Jim Enright 14, Benjamin Brown 15, and Jianghai Xiang 1,16.—Marine shrimp populations in their natural habitat are threatened by a variety of pressures including habitat destruction, pollution (i.e., pathogens, heavy metals, PCBs, antibiot- ics) and gene pool depletion. To conserve penaeid shrimp species and develop a sustainable shrimp aquaculture industry, the IMSEGI was initiated with the purpose of yearly monitoring (1) the structure of the meta-population of wild penaeid shrimp species, (2) the levels of genetic differentiation of selected species, and (3) pollutant load (pathogens, heavy metals, PCBs, etc.) in penaeid shrimp populations along their natural range. We invite academic and industry groups as well as governmental and non-govermental organizations to join IMSE- GI.—1 IMSEGI. 2 Environmental & Comparative Genomics Section. 3 Center for Conservation Medicine. 4 International Program. 5 Department of Environmental and Population Health. 6 Pathology Laboratory, Department of Biomedical Sciences, Cummings School of Veterinary Medicine at Tufts University, N. Grafton, Massachusetts, U.S.A. 7 The Fletcher School of Law and Diplomacy, Tufts University, Medford, Massachusetts, U.S.A. 8 Departamento de Biología y Bioquímica, Facultad de Ciencias de la Salud, Universidad Nacional de Tumbes, Peru. 9 Departamento de Acuacultura, Facultad de Ingeniería Pesquera, Universidad Nacional de Tumbes, Peru. 10 Escuela de Acuacultura, Universidad Técnica de Machala, Ecuador. 11 Cli- nica de Especialidades Médicas, Santo Domingo, Ecuador. 12 Universidad del Valle de Guate- mala, Guatemala. 13 Department of Aquatic Science, Faculty of Sciences, Burapha University, Chonburi. 14 Mangrove Action Project, Amphur Muang, Trang, Thailand. 15 Mangrove Action Project, Yayasan Akar Rumput Laut, Yogyakarta, Indonesia. 16 Institute of Oceanology, Chi- nese Academy of Sciences, Qingdao, China.

DECLINING MANGROVES AND FISHERIES IN THE BATAN ESTUARY, PANAY IS- LAND, CENTRAL PHILIPPINES by Jon P. Altamirano.—In the Philippines, being an archipelago of 7100 islands with a vast 184,000 km2 continental shelf area, fishing is a major means of livelihood. With recent trends, coastal fishing grounds that were once abundant with fishery resources and vast coral reefs, seagrasses and mangrove belts, have been becoming increasingly depleted. Mangrove cover has dramatically declined and these important ecosystems were converted to other uses. About 279,000 ha were lost from 1951 to 1988 in FIRST INT. SYMP. ON MANGROVES AS FISH HABITAT: ABSTRACTS 917 the Philippines, 50% of which were attributed to aquaculture. Population increase that led to a more intense fishing pressure and alongside the loss of these important mangrove ecosystems is the noticeable decline in municipal capture fisheries in coastal fishing grounds. One important fishing ground in central Philippines is the Batan Estuary in Panay Island where about 10,000 households are dependent upon its fisheries, especially on shrimps. Shrimp catch in the Batan Estuary showed that especially the high-priced tiger shrimp Penaeus monodon composition declined from 61.9% of the total shrimp landings in 1976–80 to only 6.22% in 1990–91. It is important to note that mangrove cover along the estuary decreased from 4800 ha in 1953 to only less than 300 ha of scattered patches in the 1990s. There are also extensive (4596.81 ha) culture ponds in the area. In recent interview and sampling data, fishers who used to catch 10–25 kg d−1 of shrimps and fish in the 1970s only catch an average of 1.5 kg −1d at present. These trends and status of mangroves and fisheries of the Batan Estuary in central Philippines are presented in this paper.—Global Fisheries Development Science Laboratory, University of Tokyo, Tokyo, Japan.

LARVAL, JUVENILE, AND ADULT FISHES IN MANGROVE TIDAL CREEK HABI- TATS: SPATIOTEMPORAL CHANGES IN A BRAZILIAN TROPICAL ESTUARY by Mário Barletta 1, Audrey Barletta-Bergan 2, and Ulrich Saint-Paul 2.—Mangroves, open water channels, and adjacent marine areas are considered important spawning grounds and nursery areas for riverine, estuarine, and marine fishes. This study describes the seasonal changes of fish species composition in relation to biomass, density, and biodiversity in the mangrove tidal creeks and in different reaches of the main channel of the estuary. The fish fauna of each habitat was different in density, biomass, and species composition. Seasonal salinity fluctuation is the main factor which structures the fish assemblages in the different habitats of the estuary. The results are also discussed in relation to the different levels of estuarine dependence from the most important species captured in Caeté Estuary. At least 85% of the species captured by the artisanal and subsistence fisheries in the coastal areas of this region require estuarine conditions to complete their life cycle.—1 Laboratório de Ecologia de Ecossistemas Aquáticos, Departamento de Oceanografía, Universidade Federal do Pernambuco, Recife, Pernambuco, Brazil. 2 Zentrun für Marine Tropenökologie, Bremen, Germany.

GRACKLES FORAGING MAY ENHANCE FEEDING SUCCESS OF MANGROVE FISH- ES by Laura E. Bavaro, Lizeth M. Szelistowski, and William A. Szelistowski.—Positive interactions (mutualism, commensalism) are widespread in tropical environments such as coral reefs, but are not well known in mangroves. This study presents evidence that in the Gulf of Nicoya, Costa Rica, mangrove fishes use disturbances made by foraging grackles to locate and consume tree crabs. Historically, artisanal fishermen in the gulf were known to mimic grackle disturbances to attract snappers and catfish. We found that: (1) four common snappers, Lut- janus colorado, L. aratus, L. jordani, and L. argentiventris, and the catfish Hexanematichthys seemanni feed on the grapsid crabs Aratus pisonii and Goniopsis pulchra; (2) fishes frequently prey on tree crabs which enter the water to escape grackles; and (3) snapper and catfish at- tack rates on submerged crabs are increased following experimentally-produced disturbances resembling those of grackles. The difficulty of making visual observations in mangroves may contribute to the paucity of positive interactions known to occur there.—Eckerd College, St. Petersburg, Florida, U.S.A.

SPOIL DREDGE MANGROVE FORESTATION AS A POTENTIAL SUPPORT FOR FISH- ERIES IN A MEXICAN COASTAL LAGOON by D. Benitez-Pardo 1, F. J. Flores-Verdugo 2, and M. Casas-Valdéz 3.—The main problem in most coastal lagoons in Mexico is the silting of tidal channels. In the Bay of Navachiste (northwest coast of Mexico) a dredging program was developed by the Federal Fisheries Ministry with the purpose to restore the natural tidal hydrodynamics. The spoils were used for the construction of islands and forested with mangroves from a nursery with 25,000 seedlings of Rhizophora mangle and Avicennia germinans. 918 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Mangrove plantation was carried out considering flooding period for each species, according to their tidal ranges and soil salinities. After a year the survival rate was from 68% to 59% for R. mangle and from 80% to 85% for A. germinans. The growing rate for R. mangle was 1.4 cm mo−1 compared to 0.8 cm mo−1 for seedlings in the natural forest. For A. germinans growing rate was 3.2 cm mo−1 vs 1.5 cm mo−1 in the natural forest.—1 Facultad de Ciencias del Mar de la Universidad Autónoma de Sinaloa, Mazatlán, Sinaloa, Mexico. 2 Unidad Académica Mazatlán del Instituto de Ciencias del Mar y Limnología de la UNAM, Mazatlán, Sinaloa, Mexico. 3 Centro de Investigaciones de Ciencias Marinas del IPN, La Paz, Baja California Sur, Mexico.

HURRICANE IMPACTS ON FISH COMMUNITIES ASSOCIATED WITH MAN- GROVES by Stephen A. Bortone , A. J. Martignette, Brad Klement, Jon Guinn, and Eric Mil- brandt.—Hurricane Charley (Category 4) hit the southwest Florida coast in August 2004 and had a significant impact on mangrove structure. Ongoing surveys on fishes and mangroves presented the opportunity to investigate changes in the mangrove-associated fish community relative to the impact of the hurricane. Presumably, waters and substrates proximate to the mangrove shoreline were altered relative to increased exposure to sunlight with concomitant increases in temperature and decreases in oxygen. Consequently, a scenario was provided to examine the effects that these physical features of the mangrove community had on the associated fish community. Three replicate seine hauls, using a 21.3-m × 1.8-m (3.2-mm mesh) net, were deployed to sample the mangrove-adjacent fish community at numerous locations before and after the hurricane. Analyses of community structure (Bray-Curtis dissimilarity index) in simultaneous comparisons with associated environmental variables using CCA (Canonical Correspondence Analysis) indicated changes in the basic community structure were detectable but short lived, and were overwhelmed by other “catastrophic” events, such as red tide. Mangrove re-vegetation continues with concomitant evaluation of the fish community.— Marine Laboratory, Sanibel-Captiva Conservation Foundation, Sanibel, Florida, U.S.A.

MANGROVE MYTHS AND REALITY: THEIR CONNECTIVITY TO FISHERIES PRO- DUCTION by Dawn Couchman , John Beumer, and John Kirkwood.—Mangrove protection in Queensland, Australia, has a long history (ongoing for 100 yrs) and is unique in the developed world. Considerable resources have been invested in recent years to map and identify the extent, diversity, structure, and function of mangrove communities and their productivity, based on leaf-fall and nutrient export to coastal food webs. However, the linkages between fish and fisheries production and mangroves remains poorly understood. This presentation explores the current evidence for and against the reliance of fisheries productivity on mangrove communities and highlights the challenges resource managers face in an era of high development pressure in Queensland’s coastal cities. New legislation supporting development threatens to erode the protection afforded to Queensland’s mangroves. The pressure is on to provide hard evidence to support the continuing conservation of these marine plants.— Queensland Depertment of Primary Industries and Fisheries, Brisbane, QLD, Australia.

VARIABILITY IN CARIBBEAN MANGROVE CREEK FISH POPULATIONS AND COM- MUNITIES: IMPACTS OF HUMAN THREATS AND ENVIRONMENTAL FACTORS by C. P. Dahlgren 1, P. Kramer 2, Craig A. Layman 3, D. Albrey Arrington 4, L. Burke, and Lori M. Valentine 4.—Mangrove fringes of tidal creeks serve as important habitats for juvenile and adults of numerous fish species in the Caribbean. Fish using these habitats include those of importance to marine resource managers due to their economic value, ecological function or their threatened status. Thus, understanding factors that influence mangrove fish communities and populations of key species residing in mangroves is essential for effective management of mangrove systems and species that depend on them. Here we combine GIS analyses of human threats to mangrove systems, in situ measurements of environmental parameters, FIRST INT. SYMP. ON MANGROVES AS FISH HABITAT: ABSTRACTS 919 and surveys of fish from mangrove creek systems from the Bahamas and Virgin Islands. Using these datasets, we examine the relationship that human threats and environmental factors have with mangrove creek fish population and community structure. Preliminary analyses show that environmental variables such as water depth have the greatest influence on fish community structure, but several human threats also contribute to fish community structure and populations of fishery species.—1 Perry Institute for Marine Science, Jupiter, Florida, U.S.A. 2 National Science Foundation, Washington, DC, U.S.A. 3 Department of Ecology and Evolutionary Biology, Yale University, New Haven, Connecticut, U.S.A. 4 Loxahatchee River District, Jupiter, Florida, U.S.A. 3 University of Alabama, Tuscaloosa, Alabama, U.S.A.

CATCH-AND-RELEASE ANGLING IN MANGROVE CREEKS: SURVIVAL AND BE- HAVIOR OF BONEFISH (ALBULA SPP.) IN ELEUTHERA, BAHAMAS by Sascha A. Danyl- chuk 1,3,4, Andy J. Danylchuk 1,3, Steve J. Cooke 1,2,3, Tony L. Goldberg 1,4, Jeff Koppelman 1,5, and David P. Philipp 1,3,4.—Bonefish (Albula spp.) inhabit shallow tropical and subtropical mangrove environments worldwide and are economically important due to their popularity among recreational anglers. Despite their importance, little is known about the biology and ecology of bonefish. The purpose of this study was to examine how different angling, -han dling, and release techniques affect the short-term post-release survival of bonefish inhabiting mangrove creeks in Eleuthera, Bahamas. A total of 87 fish were angled, released, and visually tracked using small surface floats. A total of 15 (17%) bonefish were preyed upon. Fish released near protective mangrove cover did not experience decreased predation risk. However, all but one of the predation events occurred over 10 m distance from mangrove cover. These results indicate that handling practices and release strategies can significantly affect the short-term survival of bonefish in mangrove creeks. Angler education and management plans that en- courage conservative handling can help to sustain this important fishery.—1 Cape Eleuthera Institute, Eleuthera, The Bahamas. 2 Institute of Environmental Science and Department of Biology, Carleton University, Ottawa, Canada. 3 Illinois Natural History Survey, Center for Aquatic Ecology and Conservation, Champaign, Illinois, U.S.A. 4 Department of Natural Re- sources and Environmental Sciences, University of Illinois, Urbana, Illinois, U.S.A. 5 Missouri Department of Conservation, Columbia, Missouri, U.S.A.

DIFFERENT FISH COMPOSITION IN SEAGRASS BEDS ADJACENT TO EXTENSIVE MANGROVE AREAS AS OPPOSED TO CORAL REEFS by M. Dorenbosch, M. G. G. Grol , I. Nagelkerken, B. R. Lugendo, and G. van der Velde.—Little is known about fish assemblages on seagrass beds located adjacent to different habitats. Visual census surveys were used to study the fish composition of two types of seagrass habitats in Zanzibar (Tanzania): seagrass beds adjacent to extensive mangrove areas in an embayment (bay seagrasses) and seagrass beds situated on the continental shelf adjacent to coral reefs (reef seagrasses). At species level, 39 fish species were common in the seagrass habitats, of which nine showed significantly higher densities in bay seagrasses, and four species were exclusively observed in bay seagrasses. Seine net data supported these data and showed that five species occurring only or in higher densities in bay seagrasses were species typically associated with the mangroves as juveniles (Leth- rinus harak, Lutjanus argentimaculatus, L. ehrenbergii, L. fulviflamma and Scarus ghobban). This suggests that for some species, the occurrence of fishes on seagrass beds is related to the connectivity with mangroves.—Radboud University, Department of Animal Ecology and Ecophysiology, Nijmegen, The Netherlands.

LOCAL JUVENILE FISH DENSITIES IN FLORIDA KEYS MANGROVES CORRELATE WITH REGIONAL LANDSCAPE CHARACTERISTICS by C. Ashton Drew and David B. Eggleston .—Juvenile fish density and diversity vary greatly among mangrove prop root habitat in the Great White Heron National Wildlife Refuge, Florida Keys, USA. We tested relationships between juvenile fish density and diversity and patch- (100s m) and landscape-scale (1 km) habitat characteristics by using backwards elimination, multiple regression models, 920 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 and Akaike’s Information Criterion and adjusted R2 values to evaluate model fit. We observed that: (1) variability in juvenile fish density was better explained by landscape- than patch- scale habitat characteristics; (2) each species’ density responded uniquely to patch and landscape characteristics; and (3) juvenile fish diversity was not strongly related to either patch- or landscape-scale habitat characteristics. Contour maps of predicted relative juvenile species density and diversity in mangrove habitat (based on landscape-scale regression models) provided a good fit to field data. Our conclusions urge caution where experimental design or conservation strategies generalize the importance of mangrove habitat to juvenile fish.— 1 North Carolina State University, Department of Marine, Earth, and Atmospheric Sciences, Raleigh, North Caroliona, U.S.A.

INTEGRATED APPROACHES FOR MANGROVE CONSERVATION AND FISHERIES RESOURCES DEVELOPMENT by P. Eganathan , H. M. S. R Subramanian, and M. S. Swami- nathan.—The past few years have been marked by the exploitation of the mangrove ecosystem for coastal development. Indiscriminate felling of mangrove trees has been going on for timber, charcoal, etc. Aquatic faunas are depleted for commercial gains without any awareness toward ecological harm that it might lead to. Of late there is an increasing awareness at various levels of governments as well as the society in general with regards to the crucial ecological functions that the mangrove ecosystem plays. One good example is the strategy of Joint Mangrove Management (JMM) that involves coastal community along with government bodies in the conservation of mangrove ecosystem. This method has received some success in Tamilnadu wherein the JMM involved people who live along the coasts near the mangroves and linked them to the Forest Department. The presentation will explain in detail the methods of implementing JMM and thereby achieving sustainable management of mangrove and fishery resources.— M. S. Swaminathan Research Foundation, Chennai, India.

SPINY LOBSTER AND FISH IN MULTIPLE BACKREEF HABITATS IN THE LOWER FLORIDA KEYS, USA by David B. Eggleston 1, Geoff W. Bell, Eric G. Johnson, and G. Todd Kel- lison 2.—Aerial photographs and ground-truthing were used to map putative backreef nursery habitats in the lower Florida Keys, followed by diver-surveys of spiny lobster and juvenile fish. Back reef habitats appear to serve as an integrated mosaic of nursery habitats for reef fish and spiny lobster. Channels connecting the Atlantic Ocean and Gulf of Mexico provide conduits for larval ingress into back reef nursery habitats, as well as for ontogenetic migrations of animals from back reef to offshore reef habitats. Expansive seagrass meadows contained the smallest stages of fish and spiny lobster, whereas mangroves contained the highest densities of fish. Isolated patch coral heads contained the highest diversity of fish and density of spiny lobster. The relatively high densities of fish observed in the lower Keys compared to other Florida and Caribbean backreef systems highlight the important nursery role of this tropical system, and the need for strict conservation.—1 North Carolina State University, Raleigh, North Carolina, U.S.A. 2 National Oceanic and Atmospheric Administration, Southeast Fish- eries Science Center, Miami, Florida, U.S.A.

A META-ANALYSIS OF SNAPPER AND GRUNT DATA FROM MANGROVE SHORE- LINE HABITATS IN THE GREATER CARIBBEAN by Craig H. Faunce.—Worldwide field studies (109) were examined for size (mean, range, and maturation) and utilization (standardized abundance, density, or biomass) of mangrove habitats by snappers and grunts. The majority of comparable data are derived from visual surveys conducted in the Netherland Antilles (NA) and Florida (FL). The average size of each species examined was roughly half its respective size-at-maturity, though adult-sized fish were recorded for most species. Aver- age sizes of six species were comparatively larger in FL, though differences were significant in only two cases. The mean density of snappers and grunts were significantly greater than either seagrass or reef habitats in three of four comparisons. Agglomerative cluster-analysis and Indicator Species Analysis of data from 21 studies identified three distinct groupings; FIRST INT. SYMP. ON MANGROVES AS FISH HABITAT: ABSTRACTS 921

Lutjanus apodus and Haemulon flavolineatum indicated most island countries, Lutjanus griseus indicated continental western margins, and Lutjanus jocu, Haemulon aurolineatum, and Haemulon bonairi indicated Guadeloupe.—Florida Fish and Wildlife Research Institute, Tequesta Field laboratory, Tequesta, Florida, U.S.A.

UTILIZATION OF MANGROVES BY JUVENILE LEMON SHARKS (NEGAPRION BRE- VIROSTRIS) IN THEIR PRIMARY NURSERY AREAS by Bryan R. Franks 1,3 and Samuel H. Gruber 2,3.—Lemon sharks and other coastal shark species utilize shallow, protected waters during their early stages of life. Many of these highly productive, shallow nurseries in tropical and subtropical areas tend to be mangrove-fringed coastlines and estuaries. Sharks are thought to use these areas to reduce predation risk and maximize prey availability. An ongoing, multi-year telemetry study of juvenile lemon sharks in two primary nursery areas demonstrated that these sharks are disproportionately found near the mangrove roots. During the first two years of this study, we found that 55% of all tracking locations from 26 individuals in two separate nurseries were within 50 m of the mangroves and over 88% within 200 m of the mangroves. Prey sampling showed that prey density significantly decreases beyond 100 m from the mangroves suggesting that more prey are available close to shore. Additionally, predator presence increases proportionately with increasing distance from the mangroves suggesting that it is safer close to shore. These results stress the importance of mangroves for a properly functioning coastal shark nursery.—1 Drexel University, Philadelphia, Penn- sylvania, U.S.A. 2 University of Miami, Rosenstiel School of Marine and Atmospheric Science, Miami, Florida, U.S.A. 3 Bimini Biological Field Station, Bimini, Bahamas.

SPECIAL MANAGEMENT NEEDS OF THE GOLIATH GROUPER, EPINEPHELUS ITA- JARA, IN THE WESTERN ATLANTIC OCEAN by Sarah Frias-Torres.—Groupers are extremely vulnerable to commercial extinction, due to their late maturity, longevity, site fidelity, and formation of spawning aggregations. The Goliath grouper, Epinephelus itajara, is the largest reef fish in the Atlantic Ocean. It is found in tropical and subtropical waters. The species has been protected in U.S. waters since 1992. Epinephelus itajara is mangrove-dependent, with juveniles restricted to mangrove habitats and adults found in coral reefs and artificial structures. The status of E. itajara populations in the western Atlantic Ocean is evaluated. Lack of no-take mangrove reserves in the region might affect juvenile survival if fishing bans are lifted. Since existing marine protected areas (MPAs) rarely link nursery and adult habitats, I propose that, for E. itajara conservation, future MPA design should aim to link coral reefs with adjacent nursery habitats. Such strategy will also be useful to protect other coral reef fish species.—University of Miami, Rosenstiel School of Marine and Atmospheric Science, Miami, Florida, U.S.A.

BIOLOGICAL, FISHING, AND SOCIOECONOMIC STUDIES WITHIN THE CAYAPAS- MATAJE MANGROVE ECOLOGICAL RESERVE IN THE PROVINCE OF ESMERALDAS, ECUADOR by Nikita Gaibor, Leonor de Cajas, Mónica Prado, Dialhy Coello, Jacqueline Cajas, Rosa García, and Juan Moreno.—The Cayapas-Mataje Mangrove Ecological Reserve (REMACAM) is one of the most important ecological areas of the Ecuadorian coastal zone. It is located on the northwestern part of Ecuador. The reserve has extensive mangroves forests, which are very important for its high productivity, and where hundreds of aquatic species inhabit. The REMACAM has ecological and economic value, which serve as food for the local communities. The main objective was to obtain biological, fishing, and social information, such as the structure of phytoplankton and zooplankton within the reserve, the current biodiversity, and to have a baseline of biological and socioeconomic information, to help the knowledge and sustainable management of the reserve. Fourteen classes, with a wide variety of species, represent the zooplankton in which Acartia lilljeborg and A. tonsa predominated during the rainy and dry season, respectively. The northern area of the reserve is very productive and it showed the highest plankton biomass, with a diversity that varied between 3 922 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 and 5 bits, which means clean waters and low presence of human activities. In the central area of the reserve, the zooplankton increased and the phytoplankton decreased in relation to the northern area. It means that in this central area there are more human activities. The diversity oscillated between 3 and 4.3 bits, which is still an indication of clean water. In the southern area, the diversity decreases (1.4 and 2.8 bits). This means that these waters might be lightly polluted due to the physical and chemical conditions, especially due to the increase organic matter that is good for those species from inner waters, such as Pseudodiaptomus cf. marshi and Pseudodiaptomus sp., and also due to the anthropogenic activities within the REMACAM. There was more abundance of zoea and megalopa of brachyuran (70%), especially along the central zone. In addition, we have identified Macrobrachium larvae and larvae of Litopenaeus shrimp, which are of commercial importance, and local consumption (18%), but also species of ecological value such as mysidaceae, caridea, anomura, fish larvae, chaeto- gnata, among others (12%). Concerning cockles’ landings, the highest were registered in the Tambillo community with an annual average of 392,109 cockles in 2000 and 497,591 cockles in 2001. The capture per unit of effort was 151 cockles ofA. tuberculosa and 45 of A. similis in 2000, while the CPUE in the 2001 was 169 cockles of A. tuberculosa and 51 of A. similis. The average income was of US$1757 for the community of Santa Rosa, US$7787.5 for Tambillo, and US$3127 for the community of El Viento.—Instituto Nacional de Pesca – Fondo Ecuato- riano Populorum Progressio, Guayaquil, Ecuador.

ADULT SURVIVAL, PROBABILITY OF CAPTURE, AND ABUNDANCE ESTIMATES FOR MANGROVE DIAMONDBACK TERRAPINS (MALACLEMYS TERRAPIN) IN EVER- GLADES NATIONAL PARK, FLORIDA by Kristen M. Hart 1, Catherine A. Langtimm 2, and Carole C. McIvor 1.—Diamondback terrapins are distributed along the US east coast from Massachusetts to Texas in brackish water, but little is known about terrapins living in mangrove habitats. To estimate adult survival rate, capture probability, and abundance for mangrove terrapins, we conducted a capture-recapture study in the Big Sable Creek (BSC) complex of the Florida Everglades (November 2001–December 2003), and analyzed individual turtle encounter histories over five sampling occasions. We determined the first adult survival rate (φ = 0.79) and population estimate (mean n = 1545 individuals) for mangrove terrapins. We also determined that terrapin distribution within BSC lies largely in small headwater creeks. Moreover, terrapins in BSC showed a higher than expected injury rate (16%) compared to other populations surveyed in more northern salt marsh habitats. This high rate of injury may be due to a high proportion of unfished predators in the protected waters of BSC.—1 U.S. Geological Survey, Center for Coastal and Watershed Studies, St. Petersburg, Florida, U.S.A. 2 U.S. Geological Survey, Florida Integrated Science Center, Gainesville, Florida, U.S.A.

FISH ASSEMBLAGES IN THE CAUSEWAY MARGINS by Joán Irán Hernández-Alber- nas.—In 2003, 80 visual censuses were performed along the Caibarién-Santa María causeway to know the environmental impact on fish communities. The families better represented were Haemulidae, Lutjanidae, and Scaridae. The abundance and biomass showed significant differences (P < 0.05) for the interaction of factors strata and margins. The sizes most seen varied between 10–20 cm fork length. The dominant trophic guild was the benthivore. The composition per species showed high similarities to the ones registered at the mangroves of the region, even though the community structure based on the density, differed to those reported in the nearest biotopes. The biomass estimates in some areas were similar to those obtained in the productive reef patches at the Sabana-Camagüey archipelago. The results suggest that the spatial heterogeneity (increase of refuges) contributed by the rocky margins of the causeway has created favorable conditions for the development of fish assemblages exported from the adjacent mangroves areas.— Centro de Estudios y Servicios Ambientales de Villa Clara, Santa Clara, Villa Clara, Cuba. FIRST INT. SYMP. ON MANGROVES AS FISH HABITAT: ABSTRACTS 923

ECOLOGICAL EFFECTS OF THE EXTENSIVE BIMINI BAY RESORT DEVELOPMENT ON THE JUVENILE LEMON SHARK (NEGAPRION BREVIROSTRIS) POPULATIONS OF BIMINI, BAHAMAS: A BACI ANALYSIS by D. E. Jennings, S. H. Gruber 1, S. T. Kessel, B. R. Franks 2,3, and A. L. Robertson.—The shallow waters around Bimini (25°43.70´N, 79°18.00´W) provide an ideal nurseries for lemon sharks (Negaprion brevirostris), but this habitat is under threat from an extensive resort development. Using BACI analysis, effects of the development were investigated by studying three aspects; comparing growth rates of juvenile lemon sharks in the North Sound, Sharkland, and South Bimini nurseries; comparing first-year survival rates of neonate lemon sharks in the North Sound and Sharkland between 1995–2005; and a comparing of habitat structures in the North Sound and South Bimini between 2003 and 2005. BACI analysis did not identify statistically significant differences between growth rates of sharks before and after March 2001. However there was a highly significant statistical difference between first-year survival rates of both North Sound and Sharkland neonates before and after March 2001. Significant changes were also identified between 2003 and 2005 in the North Sound nursery.—1 University of Miami, Rosenstiel School of Marine and Atmospheric Science, Miami, Florida, U.S.A. 2 Drexel University, Philadelphia, Pennsylvania, U.S.A. 3 Bi- mini Biological Field Station, Bimini, Bahamas.

VARIATION OF OTOLITH ELEMENTAL SIGNATURES AMONG THREE SPECIES OF JUVENILE SNAPPERS INHABITING NURSERY REGIONS WITHIN SOUTH FLORIDA by David L. Jones 1, Monica R. Lara 1, and John T. Lamkin 2.—The snappers inhabiting south Flor- ida’s marine ecosystems are a commercially, recreationally, and ecologically important group of fishes that use mangrove and seagrass nursery habitats before migrating to the reef tract as young adults. Trace elements permanently incorporated into the otoliths of a fish during growth will vary in composition and proportion depending on the environmental conditions to which the fish was previously exposed. ICP-MS was used to determine the microchemical constituents of fish otoliths defining a distinct “elemental signature”. These signatures can differ among fishes exposed to different water masses and environmental conditions allowing them to serve as natural tags for tracking fishes. The trace elemental composition of otoliths extracted from 227 individuals of four species of snapper (Lutjanus apodus, L. chrysurus, L. griseus, and L. synagris) collected from 14 sites within and around Florida Bay were examined in order to assess the extent of spatial, temporal, and taxonomic variability.—1 University of Miami, Rosenstiel School of Marine and Atmospheric Science, Cooperative Institute for Ma- rine and Atmospheric Studies (CIMAS), Miami, Florida, U.S.A. 2 NOAA Fisheries Service, Miami, Florida, U.S.A.

AN EVALUATION AND COMPARISON OF DUAL-FREQUENCY SONAR AND STE- REOSCOPIC VIDEO FOR ESTIMATING FISH ABUNDANCE AND SIZE STRUCTURE IN MANGROVE HABITATS by G. Todd Kellison 1, Jiangang Luo 2, Jack Javech 1, Peter S. Rand 3, Peter Johnson 4, and Joseph E. Serafy 1.—Underwater visual fish surveys have become the most commonly used method for estimating fish abundance and diversity in coral reef environments, and more recently in adjacent environments such as mangroves, which provide habitat for a wide range of ecologically and economically important species. Limitations associated with visual surveys (e.g., restricted to daylight hours and relatively clear-water conditions) have resulted in a potentially incomplete assessment and understanding of fish community composition, structure and dynamics. In the present study, we examine the utility of three techniques for underwater fish community assessment. In mangrove and coral reef habitats, under varying conditions of light and water clarity, we compare and contrast: (1) a newly-developed, dual-frequency sonar system (DIDSON); (2) a stereo-video system; and (3) a standard visual survey. We discuss the benefits and limitations of each method for assessing fish community structure, and recommend combined survey approaches to maximize our knowledge of fish utilization of reef and mangrove habitats.— 1 National Oceanic and Atmo- spheric Administration, Southeast Fisheries Science Center, Miami, Florida, U.S.A. 2 Univer- 924 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 sity of Miami, Rosenstiel School of Marine and Atmospheric Science, Miami, Florida, U.S.A. 3 Wild Salmon Center, Jean Vollum Natural Capital Center, Portland, Oregon, U.S.A. 4 LGL Northwest, North Bonneville, Washington, U.S.A.

SPATIAL VARIATIONS IN MANGROVE NURSERY BOUND JUVENILE LEMON SHARK (NEGAPRION BREVIROSTRIS) PREY ITEMS, BIMINI, BAHAMAS by S. T. Kes- sel, S. H. Gruber 1, and S. P. Newman.—The North Sound and South Bimini study sites, Bi- mini Bahamas, are important nurseries for lemon sharks Negaprion brevirostris and a variety of finfish and invertebrates. Both sites are fringed by Rhizophora mangle, but Laguncularia racemosa and Avicennia germinans also occur. Following ground truthing of a false-color Landsat image, 17 habitats were recognized and overlain in GIS software with teleost catches comprising 574 daytime seine and block nets hauls. Positive relationships were identified between prey communities and specific habitats. The mangrove prop-roots showed highest prey densities, which decreased steadily with increasing distance from shore. These relationships suggest mangroves and near-shore habitats are essential for the survival of juvenile lemon sharks at Bimini. Uncontrolled development has already resulted in extirpation of approximately 30% of the North Sound fringing mangroves and sedimentation of much of the nearshore fringing habitat. Site plans which call for complete destruction of the fringing habitat have been approved. This would likely result in complete sterilization of this irreplaceable nursery ground.—1 University of Miami, Rosenstiel School of Marine and Atmospheric Sci- ence, Miami, Florida, U.S.A.

PATTERNS IN TIDAL MIGRATION OF MANGROVE FISH: A HYDROACOUSTIC AP- PROACH by Uwe Krumme and Ulrich Saint-Paul.—A 200-kHz split-beam echosounder with a 6° circular-beam transducer was applied in a macrotidal mangrove channel in north Brazil to study the migratory patterns of intertidal fish. Each tidal cycle, the first flood rise caused a clear response in the upstream movements of the entire fish population. Riding the first flood rise likely enables fish to early access the productive intertidal zone and to avoid significant expenditure of energy for movement, thereby gaining capacity for faster growth. Together with a net upstream longitudinal current in the study area, it is likely sufficient for the fish to achieve retention in the tidal tributaries from one tidal cycle to the next, emphasizing the importance of these habitats in providing a significant nursery function. When sampling the spatiotemporal distribution of fish in intertidal environments, the 3D spatial heterogeneity versus time should be considered. Hydroacoustics provides a valuable non-intrusive, high- resolution method, even in a turbid mangrove environment.—Center for Tropical Marine Ecology (ZMT), Bremen, Germany.

ECOLOGICAL CHARACTERIZATION OF MANGROVE ECOSYSTEM AS A FISH HABITAT IN TERMINOS LAGOON, IN THE SOUTHERN GULF OF MEXICO by Ana Laura Lara Dominguez 1, David Zarate Lomeli 2, Guillermo Villalobos Zapata 3, and Alejan- dro Yañez-Arancibia 1.—During seasonal sampling in 1991 and 1992, 909 individuals of 25 different fish species were collected with a trap net along the fringe red mangrove in Estero Pargo. Juvenile and pre-adults stages predominanted in the community. With more than 70% of the catch species having commercial importance; we conlcude that the mangrove ecosystem is an important critical habitat.—1 Instituto de Ecología, Xalapa, Veracruz, Mexico. 2 Consultores en Gestión Política y Planificación Ambiental, Cancún, Quintana Roo, Mexico. 3 Centro EPOMEX, Campeche, Campeche, Mexico.

THE CHRONIC ABSENCE OF CUMULATIVE IMPACT ANALYSES IN COASTAL CONSTRUCTION PERMITTING by Ken Lindeman.—Sustainable fishery management is contradicted when juvenile growth and mortality rates are compromised by long-term degradation of key habitats. However, construction projects that modify habitats continue at high rates in many areas. Primary permitting agencies have been consistently reluctant to FIRST INT. SYMP. ON MANGROVES AS FISH HABITAT: ABSTRACTS 925 say no or to seek significant reductions in project impacts. Many problems, well known in permitting, yet absent from most workshops, require attention. Cumulative impacts and associated habitat and trophic cascades are still rarely considered, even in massive documents regarding areas that have already undergone many anthropogenic disturbance events. Too often, environmental impact assessments only represent large project justification exercises. Expensive monitoring projects often do not meet minimum standards of peer-review (e.g., no replication, an absence of BACI designs). Euphemistic assumptions about project impacts gain administrative momentum and are continuously repeated in subsequent assessments. Tractable, but underutilized, analytic and administrative alternatives exist and are summarized using examples from coastal areas of the northern Caribbean.— Environmental De- fense, Miami, Florida, U.S.A.

PUBLICATION OF A 2ND EDITION OF THE WORLD MANGROVE ATLAS by M. Loyche Wilkie 1, S. Baba 2, M. Kainuma 2, S. Johnson 3, M. Clusener-Godt 4, E. Corcoran 5, and Z. Adeel 6.—Mangrove ecosystems are unique and highly productive. They support the livelihoods of millions living along tropical and subtropical coasts, providing diverse services including supporting fisheries. Drastic losses of mangrove forest are being experienced. Accurate information on of the status of mangroves is essential to achieve conservation and sustainable use of these services. The publication of a 2nd edition of the World Mangrove Atlas, first produced in 1997, is intended to inform not only scientists but also managers and conservation experts, and provide a reliable and consistent baseline. TheWorld Mangrove Atlas will include national and local-level case studies and thematic case studies that cut across national boundaries. Also country data will include profile, map, threat data where available and updated information on the current extent and changes in mangrove areas. This publication is being prepared as a joint initiative of FAO, ISME, ITTO, UNESCO-MAB, the UNEP-WCMC and UNU-IN- WEH.—1 Food and Agriculture Organization of the United Nations (FAO), Rome, Italy. 2 In- ternational Society for Mangrove Ecosystems (ISME), Okinawa, Japan. 3 International Tropi- cal Timber Organization (ITTO), Yokohama, Japan. 4 United Nations Educational, Scientific and Cultural Organization, Man and the Biosphere Programme (UNESCO-MAB). 5 United Nations Environment Programme World Conservation Monitoring Center (UNEP-WCMC), Cambridge, U.K. 6 United Nations University, International Network on Water, Environment and Health (UNU-INWEH), Hamilton, ON, Canada.

SPATIAL AND TEMPORAL VARIATION IN FISH COMMUNITY STRUCTURE OF A MARINE EMBAYMENT IN ZANZIBAR, TANZANIA by Blandina R. Lugendo 1,2, Arjan de Groene 1, Ilse Cornelissen 1, Annelies Pronker 1, Ivan Nagelkerken 1, Gerard van der Velde 1,3, and Yunus D. Mgaya 2.—Spatial-temporal variations in the fish community structure were studied from a tropical non-estuarine embayment in Chwaka Bay, Zanzibar. Fish samples were collected bimonthly for 1 yr from mangroves, mud/sand flats and seagrass beds. Environmental variables were examined to determine their relationship with fish the community structure. The fish community structure together with the environmental variables in mangroves and mud/sand flats remained constant for most part of the year; however, a marked decline in fish densities, biomass, and species richness and in environmental variables was observed during the rainy period. A significant relationship was found between density and species richness of fish, and temperature, salinity, and water clarity. Salinity was the most conspicu- ously changing environmental variable with seasons; we therefore propose that, salinity alone or in combination with the other environmental variables, was probably the most important environmental factor structuring the fish assemblage in the mangroves and mud/sand flats habitats.—1 Department of Animal Ecology and Ecophysiology, Institute for Water and Wet- land Research, Faculty of Science, Radboud University, Nijmegen, The Netherlands. 2 Faculty of Aquatic Sciences and Technology, University of Dar es Salaam, Dar es Salaam, Tanzania. 3 Naturalis National Natural History Museum, Leiden, The Netherlands. 926 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

CONNECTIONS BETWEEN ROOT EPIBIONTS AND FISH COMMUNITIES IN MAN- GROVE HABITATS by James A. MacDonald and Judith S. Weis.—Many fish and epibiont species utilize red mangrove (Rhizophora mangle) prop roots as habitat. There is often great variation between mangrove areas; fish communities and populations may vary significantly between nearby, similar mangrove areas. This study examined the relationship between fish and prop-root epibiont communities. Fish communities at 15 sites in Bocas del Toro, Pana- ma, were examined by visual census and trapping. Artificial mangroves consisting of stakes driven into seagrass, adorned with artificial epifauna (combinations of blocks and dowels) were also established. Prop-root organisms were censused using a random sample of roots per site. There was significant correlation between epibiont diversity and fish diversity, and a correlation between percent coverage of epibionts and fish diversity. The artificial mangroves with the most artificial epifauna attracted the most diverse fish community. Data will also be presented on community composition between sites, and on results of epibiont manipulation experiments.—Rutgers University, Biological Sciences, Newark, New Jersey, U.S.A.

THE IMPACTS OF NATURAL DISTURBANCE ON THE NEKTON COMMUNITY STRUCTURE IN MICRONESIAN MANGROVE FORESTS by Richard A. MacKenzie and Nicole Cormier.—Six and 18 mo after a category 3–4 typhoon devastated the island of Yap, we surveyed nekton communities from mangroves hypothesized to be more impacted and less impacted from the typhoon using lift and float nets. Fish densities were not significantly different among any sites in either year. Densities were generally higher in disturbed sites compared to undisturbed sites in 2004, and were dominated by Gobiidae. In 2005, fish densities were higher in undisturbed sites compared to disturbed sites, and were dominated by Apogonidae. Shrimp densities were higher in disturbed sites compared to undisturbed sites in 2004, but were similar between all sites in 2005. Results suggest that demersal fish may be better adapted at recovering from typhoons compared to other species. Elevated shrimp densities may have resulted from increased deposition of mangrove leaves after the typhoon, which may provide an important mechanism to help fish species recover after typhoons.— U.S. Department of Agriculture Forest Service, Institute of Pacific Islands Forestry, Hilo, Ha- waii, U.S.A.

ECONOMIC VALUE AND SOCIAL DIMENSIONS OF MANGROVE FISHERIES IN VIL- LAGES AFFECTED BY THE 2004 TSUNAMI by Rusyan Jill E. Mamiit.—On December 26, 2004, mangrove fisheries in coastal villages in Sri Lanka were subject to severe damage caused by the tsunami. The impacts of the disaster on the mangrove ecosystem magnified the significance of the resource as important source of fish, which supports the livelihood of the village residents. A socioeconomic assessment of the value of mangrove fisheries in Kapuhenwala and Waduruppa, villages characterized by functioning and degraded mangrove ecosystems, respectively, was carried out following the tsunami. Results indicate that intact mangroves generate an average annual fishing economic value of approximately SLR641,148 (US$6411) per fishing household; whereas, a degraded mangrove ecosystem has a value of SLR243,662 (US$2437). The economic values suggest that areas with degraded mangroves generate lower fishing benefits to the community. The findings further provide a sound basis for the inclusion of mangrove rehabilitation efforts in the post-tsunami reconstruction and rebuilding programs.—Joint Institute for Marine and Atmospheric Research, University of Hawai‘i at Manoa, Manoa, Hawaii, U.S.A.

EVALUATING TROPHIC LINKAGES IN MANGROVE-BASED FOOD WEBS USING STABLE ISOTOPES OF CARBON AND NITROGEN by Debashish Mazumder 1, Ron Szymc- zak 1, Neil Saintilan 2, and Robert J. Williams 3.—An understanding of the energy flow pathways and trophic linkages in estuarine food webs is essential for managing estuaries and their ecosystems sustainably. These pathways are complex, given the dynamics in physico-chemical processes, variety and area of habitats. Carbon and nitrogen stable isotope ratios (δ13C and FIRST INT. SYMP. ON MANGROVES AS FISH HABITAT: ABSTRACTS 927

δ15N) were measured for a variety of fish, invertebrate, and crustacean species collected from saltmarsh and mangrove habitats in Botany Bay and Homebush Bay, NSW, Australia. The work is on-going, however, initial observations indicate specific prey-predator linkages evident within a complex trophic structure. Results also advocate the role of certain non-commercial estuarine species as important conduits of energy and nutrition to higher trophic- order commercially valuable species, linking these with specific estuarine habitats. This work seeks to model the source of energy and nutrition in mangrove and saltmarsh-based food webs and to determine the chemical linkages between high trophic order species and different habitat resources.—1 Australian Nuclear Science and Technology Organisation, Menai, NSW, Australia. 2 River and Wetlands Unit, Department of Environment and Conservation, Sydney, NSW, Australia. 3 Department of Primary Industry, Cronulla, NSW, Australia.

VARIABILITY OF STABLE ISOTOPE RATIOS OF MANGROVE GLASSFISH (AMBASS- IS JACKSONIENSIS) FROM SOUTHEAST AUSTRALIA AND THE IMPLICATIONS FOR ECOSYSTEM STUDIES by Debashish Mazumder 1, Robert J. Williams 2, Ron Szymczak 1, Dennis Reid 2, and Neil Saintilan 3.—Scientists concerned with organic matter flow and food web structures in aquatic ecosystems are increasingly realizing the potential of stable isotope ratios as natural tracers. Stable isotopes offer an accurate and cost effective way to understanding critical pathways of energy and pollutant transfer. Further, many aquatic habitats have been degraded and isotope ratios offer insights into appropriate conservation and rehabilitation techniques to manage these valuable resources. So far, the literature shows little attention has been paid to spatial and temporal variations in isotope signatures of samples taken from saltmarsh and mangrove environments. This study reports on investigations into the differences in isotopic signatures within a single species, Ambassis jacksoniensis, surveyed from two locations at two different times. The results suggest significant variation in δ13C between different season and location for glassfish, but not for15 δ N. The results also suggest that care is needed in interpreting previously published results.—1 Australian Nuclear Science and Technology Organisation, Menai, NSW, Australia. 2 Department of Primary In- dustry, Cronulla, NSW, Australia. 3 River and Wetlands Unit, Department of Environment and Conservation, Sydney, NSW, Australia.

EPSCOR PHASE II: FUTURE DIRECTIONS — LAND-SEA INTERACTIONS by Amber McCammon, Richard Nemeth, Meri Whitaker, and Nasseer Idrisi.—The Experimental Pro- gram to Stimulate Competitive Research (EPSCoR) is a program of NSF to strengthen research, development and competitiveness in participating states and territories. The Virgin Islands EPSCoR is a partnership between NSF and the higher education, government, and business communities of the Virgin Islands to promote the development of the territory’s science and technology resources through targeted research, education, and outreach activities. The current VI-EPSCoR research thrust is the study of the Biocomplexity of Caribbean Coral Reefs (BCCR). BCCR research covers areas of oceanography, biodiversity, and ecology of Ca- ribbean coral reef ecosystems. In accordance with the environmental mission of VI-EPSCoR, phase II will cover the land-sea interface, including mangroves, an important component to the island ecosystem. Mangroves act as nursery grounds for many coral reef fishes. Man- groves also act as a buffer to filter nutrient and sediment run-off that would potentially upset the fragile balance of coral reef ecosystems.—University of the Virgin Islands, U.S.V.I.

USE OF CARBON AND NITROGEN STABLE ISOTOPIC TO INFER FOOD SOURCES IN BELIZE OFFSHORE MANGROVES by C.C. McIvor 1, M.L. Fogel 2, D. S. Taylor 3, W. Davis 4, and E. Reyier 5.—As part of an ichthyological survey of Twin Cays, Belize, we sampled fish and crustaceans for carbon and nitrogen stable isotopes from multiple habitats (fringe forests, dwarf Rhizophora mangle forests, ponds, internal channels, sinkholes). Of 12 species with at least three individuals in a given habitat (individuals analyzed separately), eight species differed significantly in their C isotopic values and five species differed in N isotopic values 928 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 in different habitats. Species using multiple habitats, therefore, are likely to derive both carbon and nitrogen from different within-habitat sources. Animals from hydrologically isolated sinkholes had carbon isotopic values significantly more negative than in other habitats, indicating either greater dependence on a carbon source from mangroves or from highly-respired, microbial carbon. Comparison of frequency histograms of carbon isotopic values from Twin Cays’ consumers with those from other mangrove studies indicates a greater incorporation of a variety of carbon sources than previously reported.— 1 U.S. Geological Service, Florida Inte- grated Science Center, Center for Coastal & Watershed Studies, St. Petersburg, Florida, U.S.A. 2 Carnegie Institution of Washington, Geophysical Laboratory, Washington, D.C., U.S.A. 3 Bre- vard County Environmentally Endangered Lands Program, Melbourne, Florida, U.S.A. 4 PO Box 607, Quincy, Florida, U.S.A. 5 Dynamac Corporation, Kennedy Space Center, Florida, U.S.A.

CONNECTIVITY CONTROLS MANGROVE-FISH ASSEMBLAGE STRUCTURE: EVI- DENCE FROM 12 MANGROVE-LINED WETLAND PONDS IN TAMPA BAY, FLORIDA by C. C. McIvor 1, J. M. Krebs 1,2, and A. B. Brame 1,2.—We collected seasonal nekton samples (fish and decapod crustaceans) from 12, mangrove-lined tidal ponds as part of a larger study to characterize the coastal wetlands of Tampa Bay. We captured 43,493 individuals of 63 species using a center-bag seine. Pond assemblages differed between regions of the bay and were distinct from tidal-creek assemblages from which they are likely derived (PRIMER, ANOSIM, Global r = 0.619, P = 0.004). The differences between pond and creek assemblages indicate predictable macrohabitat associations for some species. When averaged over the 2004 sampling year, mean nekton density in ponds ranged from 79 to 1160 per 100 m2. Species composition of ponds can be explained by degree of connectivity with nearby estuarine waters (PRIMER, P = 0.001). Three ponds with only intermittent connection to bay waters were depauperate, whereas ponds with moderate or good connectivity supported estuarine transient as well as estuarine resident species.— 1 U.S. Geological Service, Florida Integrated Science Center, Center for Coastal & Watershed Studies, St. Petersburg, Florida, U.S.A. 2 ETI Professionals, Tampa, Florida, U.S.A.

THE INSULAR ECOLOGY OF THE DIAMONDBACK TERRAPIN, MALACLEMYS TERRAPIN, ON THE MANGROVE KEYS OF SOUTH FLORIDA by Brian K. Mealey 1,3, Greta B. Mealey 2, John D. Baldwin 3, Gregory D. Bossart 4, and Michael R. J. Forstner 5.— The insular ecology and population dynamics of Malaclemys terrapin macrospilota and M. t. rhizophorarum on mangrove keys permeating Florida Bay and the Florida Keys are under investigation by collaborative investigators from many management, research, and educational institutions in south Florida. Results indicate extreme site fidelity over this time period for terrapins. Survival of individual terrapins over a 20-yr span has been documented using data from previous researchers in the region. Many are currently residing in the same highly localized black mangrove system, Avicennia germinans, in which they were first captured in the early 1980s. Malaclemys have a varied diet consisting of fish, crustaceans, and molluscs. The analyses of mitochondrial DNA (mtDNA) and nuclear DNA (nDNA) microsatellite markers support the physical evidence of limited dispersal. Florida terrapin populations represent localized assemblies with strong tendencies for long-term site fidelity. This has significant implications for management of these populations.—1 Institute of Wildlife Sciences, Palmetto Bay, Florida, U.S.A. 2 Miami Museum of Science, Miami, Florida, U.S.A. 3 Department of Biology, Florida Atlantic University, Davie, Florida, U.S.A. 4 Harbor Branch Oceanographic Institute, Ft. Pierce, Florida, U.S.A. 5 Department of Biology, Texas State University, San Mar- cos, Texas, U.S.A.

RESTORING ESSENTIAL FISH HABITAT IN SOUTHEAST FLORIDA: MANGROVE AND SEAGRASS HABITAT DESIGN COMPONENTS AND SUCCESS MONITORING by Gary R. Milano 1, Neil Hammerschlag 2, John Barimo 2, and Joseph E. Serafy 3.—Rapid urban- FIRST INT. SYMP. ON MANGROVES AS FISH HABITAT: ABSTRACTS 929 ization of the south Florida area has resulted in a loss of vital fisheries habitat for many commercially and recreationally important fish and invertebrate species. Large-scale wetlands restoration efforts are being designed and implemented regionally to maximize habitat heterogeneity, and provide critical fish habitat. To provide diverse seagrass and mangrove habitats for larval and juvenile fish development, the designs include a network of tidal flushing channels inter-connecting low energy tidal pools, and shallow open-water areas with specific hydrological criteria. To document the efficacy of the restored habitat, fish assemblages are being monitored at a 30-ha mangrove wetlands restoration site on Key Biscayne, Florida. To date, a total of 29 fish taxa have been identified in the restored tidal pools, and the diversity of fish species has increased from five to ten species per tidal pool over the five year monitoring effort. Fish species richness has increased by four species since the baseline was established in 2000/2001, and the fisheries inventory has documented the restored areas functioning to support important fisheries species.—1 Miami-Dade County, Department of Environmental Resource Management, Miami, Florida, U.S.A. 2 University of Miami, Rosenstiel School of Marine and Atmospheric Science, Miami, Florida, U.S.A. 3 NOAA Fisheries, Miami, Florida, U.S.A.

SEASONAL VARIATION IN FISH ABUNDANCE IN MANGROVE ECOSYSTEMS: COMPARING FORESTED AND UN-FORESTED HABITATS by H.O.D. Mirera 1, J. G. Kairo 2, N. E. Kimani 2, and K. F. Waweru 1.—The research investigated fish abundance in forested and unforested sites at Ungwana Bay, Kenya. Four forested sites having paired unforested sites were studied for comparison. Samples were collected with nets for 8 mo over two years (2003 and 2004). Mean fish abundance ranged from 6.11 fish 36−2 m to 80.08 fish 36 −2m in forested sites and 3.08 fish 36 −2m to 125.89 fish 36 −2m in unforested sites while biomass varied from 37.87 to 326.75 in forested sites and 7.88 to 303.92 in unforested ones. The results indicate a high abundance of fish in forested sites compared to unforested ones a part from site 4 where the unforested area had more fish abundance due to one big sample of Pel- lona ditchella that accounted for 73.1% of the fish in the site. A total of 35 fish species were sampled from both forested and unforested sites with 11 being exclusively forested and five unforested. There were significant differences in fish abundance and biomass with respect to substratum type indicating that the fish community preferred muddy bottom forested sites to sandy bottom forested sites. Fish abundance was significantly higher in all sites (forested and unforested) during northeastern monsoon compared to southeastern monsoon, however, the gap between the seasons was more pronounced in muddy substratum sites compared to sandy ones. The unforested sites showed significantly lower density of meiofauna in the sediments compared to forested sites, while muddy substratum sites also had significantly higher meiofauna density. The results support the hypothesis that fish visit mangrove habitats to feed and to avoid predators. They also raise the idea of substrate type as an influencing factor in fish habitat preference.—1 Egerton University, Department of Natural Resources, Njoro, Nakuru, Kenya. 2 Kenya Marine and Fisheries Research Institute, Mombasa, Kenya.

NUTRIENT DYNAMICS IN A CLOSED SYSTEM WITH MANGROVE SEEDLINGS AND POECILID FISHES by Leonardo Moroyoqui-Rojo 1, Francisco J. Flores-Verdugo 2, Di- ana Escobedo-Urías 1, and Maria Nancy Herrera-Moreno 1.—Six closed recirculation systems (1000 L each) with poecilid fishes and mangrove seedlings were designed to estimate the nutrient uptake by mangroves and the survival and growth rate of poecilid fishes. Each system has a biological filter of gravel and sand, 34 mangrove seedlings Rhizophora( mangle and Laguncularia racemosa) and 200 poecilid fishes. The results show that mangroves removed

71%–94% of the dissolved inorganic nitrogen (DIN) and 36%–47% of the orto-PO4. The systems without mangroves removed 35%–52% of DIN and 21%–25% during a cycle of 10 d. The poecilid fish survival was 100% with a good growth rate in all the systems (1 cm mo−1). Even though growth rate and survival of the poecillid fishes were similar in all the systems, we consider that in a long term, the mangroves play an important role in keeping good water 930 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 quality conditions for fish communities.—1 Centro Interdisciplinario de Investigación para el Desarrollo Integral Regional, Guasave, Sinaloa, Mexico. 2 Unidad Mazatlán del Instituto de Ciencias del Mar y Tecnología (UNAM), Mazatlán, Sinaloa, Mexico.

ADOPTION RATES OF SILVOFISHERIES INNOVATIONS AS AN OPTION FOR MAN- GROVE CONSERVATION; CHALLENGES AND RESEARCH OPPORTUNITIES ALONG THE EAST AFRICA COASTLINE by D. O. Obura 1, H. O. D. Mirera 2, and E. M. Arriesgado 2.— Mangrove wetlands offer refuge and nursery ground for juvenile fish, crabs, shrimps, and molluscs. Apart from being nutrient rich ecosystems they also act as buffers protecting the coastline from erosion, storm damage, wave action inclusive the recent traumatic tsunamis. This vital ecosystem has been jeopardised along the East African coast due to poor realisation of its diversified importance. Though efforts have been made by mangrove experts and conservationists, the big problem has been convincing the people on the immediate use of this resource to attain their daily food requirement while maintaining sustainability. Getting a new idea adopted, even when it has obvious advantages is difficult; hence a common problem for many individuals and organisations is how to speed up the rate of diffusion of an innovation. In trying to empower the local people to own and conserve their resource, intense awareness campaigns on silvofisheries and mangrove eco-tourism are prioritized along the East African coastline. Presently, ten communities along the Kenyan coast and five along Tanzania’s Tanga regions are covered in silvofisheries on farm trials to assess technology adoption rate. Social changes and social problems facing the East African coastal communities affect the diffusion of the innovations. The main challenge lies in the extreme poverty of the people and lack of dedication in culturing crabs and fish to market size coupled with mangrove conservation. The diffusion of innovations and consequent adoption is essentially important in silvofisheries technology in East Africa so as to ascertain meaningful mangrove conservation. Despite the success of a few communities, the researchers are challenged on how their day to day findings can reach the custodians and beneficiaries of our vital mangrove wetland. In due respect, the question still remains of whether there is enough silvofisheries research along East Africa coastline.1 Kenya Marine and Fisheries Research Institute, Mombasa, Kenya. 2 Egerton University, Department of Natural Resources, Njoro, Nakuru, Kenya.

FOOD HABITS OF FISH ASSOCIATED TO COASTAL HABITATS IN SAN ANDRÉS ISLAND, COLOMBIAN CARIBBEAN by Vivian Ochoa , Arturo Acero, Adriana Santos, and Jaime Polanía.—Feeding habits of fish captured at Hooker-Honda and Sea Horse Bays in San Andrés Island (Colombian Caribbean) were characterized. Four hundred and twenty-one individuals distributed in 18 families and 37 species were caught and purchased from local fishermen in the area. Three species of some commercial importance, formed 60% of the abundance; Opisthonema oglinum (32%), Gerres cinereus (15%), and Harengula humeralis (13%). Other species, for instance Oligoplites palometa and Caranx hippos demonstrated very low numerical abundances, with only one individual caught each. In stomach content analysis, 91 nutritional categories of different taxa, including seaweeds, seagrass, mollusks, annelids, crustaceans, fish, among others, were identified. In addition, each species was trophically categorized yielding the following proportion: 76% carnivores, 19% omnivores, 5% herbivores; planktivores were absent. This proportion has been widely observed in a variety of tropical systems. The dominant group was the carnivores, which exercise a strong predatory activity over crustaceans, mollusks, and fishes.—Universidad Nacional de Colombia Sede Medellín, Medellín, Colombia.

EXPLAINING PATTERNS OF ABUNDANCE FOR FISH USING MANGROVES: A MULTI-SCALE SEASCAPE APPROACH by S. J. Pittman, C. Caldow, S. Davidson Hile , and M. E. Monaco.—Mangroves are often considered an important resource for many species of fish, supporting high densities of juveniles including many commercially important species. Typically, these species use mangroves as part of a chain of interlinked resources through FIRST INT. SYMP. ON MANGROVES AS FISH HABITAT: ABSTRACTS 931 daily home ranges and to support developmental shifts (“ontogenetic stepping stones”). Con- siderable spatial variability is found in the patterns of abundance from one area to another. This study applies a landscape ecology approach to explore the influence of seascape composition on the abundance of fish using mangroves in southwestern Puerto Rico. We quantified within-patch structure (1 m2 quadrat) and seascape structure (50, 100, 300, 500 m) using landscape metrics applied to NOAA’s benthic habitat map. Results indicate that the amount of seagrass surrounding mangroves explains more of the variability in fish abundance than fine-scale mangrove structure. Fish community composition is significantly different in -man groves with high adjacent seagrass cover than mangroves with little or no seagrass cover. This has important implications for resource protection, restoration efforts, and water quality management.—National Oceanic and Atmospheric Administration, NOS, CCMA Biogeogra- phy Team, Silver Spring, Maryland, U.S.A.

INTEGRATION OF AQUACULTURE AND MANGROVES by J. H. Primavera.—South- east Asia has the highest concentration of mangroves and brackish water aquaculture ponds. This paper describes studies that integrate mangroves as biofilters, and as pen culture sites for mud crab farming. In one study, passing shrimp pond effluents through a natural mangrove stand reduced levels of TSS, sulfide, NH3-N and NO3-N by 18.7%–64.2%. Estimates show that 1.4–6.5 ha of mangroves are needed to assimilate nitrogen wastes from one hectare of shrimp pond. Mangrove biomass increase was 2.5 times greater with effluents compared to a control mangrove, although plant numbers remained similar. Present mud crab Scylla spp. farming still depends on raw (“trash”) fish and wild seed. To lessen such dependence, another study compared the stocking of hatchery vs wild juveniles, and feeding of pellet + raw fish (“trash fish”) vs fish alone. Preliminary results show that low-cost pellets can reduce raw fish requirement, and that hatchery crab juveniles need immediate feeding whereas wild crabs can subsist on natural mangrove productivity for one month. Mud crab pen culture is commercially viable but technological refinements and land tenure issues remain.—Aquaculture Department, Southeast Asian Fisheries Development Center, Tigbauan, Iloilo, Philippines.

THE MANGROVE ACTION PROJECT: ITS GENESIS, MISSION AND CHALLENGES by Alfredo Quarto.—Founded in 1992, the Mangrove Action Project (MAP) is an environmental, non-governmental organization dedicated to reversing the degradation and destruction of mangrove forest ecosystems worldwide. Early work focusing on mangrove conservation and restoration in Latin America and Asia has expanded to include science, education and outreach projects around the globe. Its mission is to promote the rights of local coastal peoples, including fishers and farmers, in the sustainable management of their coastal environment. MAP provides five essential functions for grassroots associations and other propo- nents of mangrove conservation: (1) serving as an information clearinghouse on the status of, and future threats to, mangrove systems around the world; (2) coordinating an international network of over 450 NGOs on issues relevant to mangrove protection; (3) promoting public education and awareness of mangrove forest issues; (4) developing of technical and financial support for relevant NGO projects; and (5) communicating, both within and outside im- poverished coastal fishing and farming communities, how consumer demand affects coastal livelihoods and environments. MAP is addressing the challenges of conserving diverse and productive mangrove systems in the face of poverty, shrimp aquaculture and development for tourism through participatory resource management, promoting responsible consumer choices and implementing sound environmental and socio-economic impact studies.— Man- grove Action Project,Port Angeles, Washington,. U.S.A.

MANGROVE COVER, FISHERIES, AND ENVIRONMENTAL PERTURBATIONS IN THE CIÉNAGA GRANDE DE SANTA MARTA (CGSM), COLOMBIAN CARIBBEAN by Jorge Restrepo, Jacobo Blanco, Carlos Villamil , Efraín Viloria, Juan Carlos Narváez, and Mario Rueda.—Hydrological manipulations in the Ciénaga Grande de Santa Marta (CGSM) 932 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 since 1956, mostly due to road and channel building, as well as historically-low precipitation in the 1990s, caused high rise in salinity in soil and water. In 1995, 56% of the mangrove cover disappeared, and fish catch declined by 49%. Channel dredging supplying freshwater from the Magdalena River in 1996 and 1998, and high rainfall in 1999 (2509 mm) caused the reduction of salinity in soil and water, which fostered a 10% increment in mangrove cover and 40% in fish catches. Nile tilapia Oreochromis( niloticus), an exotic freshwater species favored by environmental changes, represented 62% of the catch in 1999–2000. After 2000, draught and channel sedimentation affected mangrove cover, fish catches, and species composition. Climatic variation associated with ENSO seemed to account for these changes, involving hydrological and community perturbations, which in turn affected vegetation and fish resource availability in the CGSM system.— Instituto de Investigaciones Marinas y Costeras José Benito Vives de Andréis (INVEMAR), Santa Marta, Colombia.

ICHTHYOLOGIC DYNAMICS IN THE MANGROVE OF TÉRRABA, PACIFIC COSTA RICA: TOOLS FOR MANGROVE MANAGEMENT by José Rodrigo Rojas Morales.—Abun- dance and diversity of the estuarine fish of the mangroves in the south of the Pacific Coast of Costa Rica were determined between June 2001 and June 2004. Eighty-five species distributed in 34 families and 48 genera were found. The total number of individuals captured was 4200 and the total biomass was 547.6 kg. We found 31 families of osteichthyes and three species of chondrichthyes (C. limbatus, Dasyatis longus and Urotrygon nana). The families with more species were Carangidae (10), Haemulidae (7), Centropomidae (6), Lutjanidae (6), Scianenidae (6), Ariidae (5), Gerridae (5). Seventeen families were represented only with one species. The main community component in the site ware occasional visitors, representing 43% of species. Anchoa spinifer (876), Centropomus armatus (610) and Bairdiella ensifera (479) accounted 46.7% of the total number of individuals. Based on these results, we proposed the establishment of an integral management plan for the area.—Instituto Costarricense de Electricidad, San José, Costa Rica.

PREDATION RATES VARY WITH WATER DEPTH AND OTHER BIOLOGICAL FAC- TORS IN MANGROVE-LINED TIDAL CREEKS by Andrew L. Rypel 1, Craig A. Layman 2, D. Albrey Arrington 3, and Lori M. Valentine 1.—Changing water depths are thought to drive many ecological processes in estuarine intertidal zones. We examined how predation rates varied as a function of water depth in mangrove-lined tidal creeks of the Bahamas by tether- ing an abundant prey fish, mojarra (Eucinostomus spp.). Results revealed a significant expo- nential relationship between predation rate and water depth. Tethered mojarra in shallow water mangroves had significantly longer survival times than did fish tethered in deeper water, and we were able to identify a “critical depth” of predation, i.e., a water depth threshold at which predation rates increased abruptly. Subsequently, we explored three additional factors (local predator density, prey fish size, and creek flooding regime) which contributed to variation in time until predation along the water depth gradient. Predation rates are influenced by all of the above mentioned variables, and we suggest that water depth especially should be incorporated into studies linking mangroves and fisheries ecology.—1 University of Alabama, Tuscaloosa, Alabama, U.S.A. 2 Department of Ecology and Evolutionary Biology, Yale Univer- sity, New Haven, Connecticut, U.S.A. 3 Loxahatchee River District, Jupiter, Florida, U.S.A.

INTERRELATIONS BETWEEN MANGROVES, LOCAL ECONOMY, AND SOCIAL SUSTAINABILITY: AN EXAMPLE FROM A CASE STUDY IN NORTH BRAZIL by Ulrich Saint-Paul.—The littoral region of coastal Pará in NE Brazil is part of the world’s second- largest continuous mangrove region. The Bragança peninsula is the specific study area of the interdisciplinary still ongoing joint German Brazilian project on “Mangrove Dynamics and Management” (MADAM), which started in 1995. Human use in this mangrove ecosystem is characterized by about 15 products, which have either subsistence value or generate monetary income for the local rural population. The importance of these functions for the rural FIRST INT. SYMP. ON MANGROVES AS FISH HABITAT: ABSTRACTS 933 households increases with the distance from the urban center. In the primary production sector, agriculture and artisanal fisheries are the main source of income in the wider Bra- gantinian region. Both industries are characterized by many small operators. The industrial sector is very underrepresented throughout the region. Presently the control of the allocation of resources within this region rests predominantly in the hands of local individuals. This paper examines the conditions for the successful co-management of diverse species, resource use patterns and household income portfolios in a mangrove environment. Therefore stake- holders have been incorporated directly e.g., by participation in workshops. This is part of the support of the formation of the local RESEX (reserves extrativistas) movement, a Brazilian model of natural resources co-management.— Center for Tropical Marine Ecology, Bremen, Germany.

CHANGES TO FISH ASSEMBLAGES VISITING ESTUARINE WETLANDS FOLLOW- ING THE CLOSURE OF COMMERCIAL FISHING IN BOTANY BAY, AUSTRALIA by Neil Saintilan 1, Debashish Mazumder 2, and Karen Cranney 1.—Data on commercial landings of fish and crustaceans are available for 52 estuaries in New South Wales. These same estuaries have been mapped with regards to the distribution of fish habitat, including mangrove, saltmarsh, and seagrass, along with a suite of geomorphic units. A multiple regression analysis demonstrated strong relationships between the area of mangrove and the catch of a number of commercially important species, including long-finned eel, and the mud crab Scylla serrata.— 1 Centre for Environmental Restoration and Stewardship, ACU National, North Sydney, NSW, Australia. 2 Australian Nuclear Science and Technology Organisation, Menai, NSW, Australia.

HURRICANE-INDUCED CONVERSION OF MANGROVE FOREST TO MUDFLAT: IMPACTS ON NEKTON, BIG SABLE CREEK, FLORIDA by Noah L. Silverman 1,2, Carole C. McIvor 1, Justin M. Krebs 1,2 and Victor A. Levesque 1.—The passage of two Category 4-5 hurricanes across SW Florida (Labor Day Hurricane of 1935, Hurricane Donna 1960) resulted in patchy conversion of mangrove forest to mudflat habitat within Big Sable Creek, Everglades National Park, Florida. Our goal was to determine the consequence of this habitat conversion on nekton (i.e., fish and decapod crustaceans) inhabiting the intertidal zone. We used block nets across intertidal rivulets to compare nekton leaving replicate forest and mudflat sites. Overall nekton density (individuals m−3) was significantly greater for forested habitats than non-vegetated mudflats (RM ANOVA, P < 0.001). Species composition also differed between habitat types (PRIMER, ANOSIM). Benthic forage species dominated forested sites, whereas schooling species dominated mudflats. Results are consistent with previous studies regarding the influence of vegetation structure on density of fish. These results can assist managers in assessing the long-term impact that strong hurricanes can have on mangrove-associated nekton.— 1 U.S. Geological Survey, Florida Integrated Science Center, Center for Coastal & Watershed, St. Petersburg, Florida, U.S.A. 2 ETI Professionals, Tampa, Florida, U.S.A.

THE IMPORTANCE OF MANGROVES AS NURSERY HABITAT FOR SMALLTOOTH SAWFISH (PRISTIS PECTINATA) IN SOUTH FLORIDA by Colin A. Simpfendorfer.—The use of mangrove habitats by smalltooth sawfish (Pristis pectinata) was investigated using surveys, acoustic telemetry, a public encounter database and habitat suitability models. Neo- nates (70–99 cm) occur on shallow mud banks, normally associated with mangrove coves or shorelines. Movements of these animals are strongly affected by the tide, with individuals electing to use very shallow parts of the bank (< 30 cm), presumably to avoid predation by sharks that occur in adjacent deeper water. When depths on banks become too deep to avoid predators, sawfish refuge within mangrove prop root habitats, either on the edges of banks or in drainage channels running out of mangrove stands. Acoustic tracks of juveniles 100–199 cm in length show that they inhabit slightly deeper water (< 100 cm) and their movements are strongly associated with mangrove shorelines. Analysis of sawfish encounter data indicated 934 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 that juvenile sawfish < 200 cm are positively associated with mangroves, and that there is a positive relationship between size and distance from mangroves. To identify priority conservation areas for the endangered smalltooth sawfish habitat suitability models including mangrove distribution were constructed.— Mote Marine Laboratory, Sarasota, Florida, U.S.A.

SPATIAL-TEMPORAL PATTERNS OF ISLAND MANGROVE CREEK HABITATS: A CHARACTERIZATION AS RELATED TO SIZE, ENERGETICS, AND FEEDING MODE OF FISHES by Kathleen Sullivan Sealey , Vanessa L. Nero, and Sherry Constantine.—We examine the benthic and oceanographic characteristics of small mangrove creek systems with adjacent soft-bottom embayments in the northern and central Bahamas to understand differences in the diversity and biomass of fishes. Fish assemblages were evaluated by standard beach seines, visual surveys and traps. Key species such as bonefish (Albula vulpes), snappers (Lutjanidae) and mojarras (Gerridae) vary in their abundance, size-frequency distribution and abundance based on season, tidal state and habitat (site). Benthic characteristics focused on floral (e.g., algae and seagrass) species composition, density and canopy height. Habitat maps of three different islands are used to first characterize mangrove systems by size, variability in temperature and salinity, geomorphology as well as proximity to platform margin reefs. Associations of fishes with particular habitat and oceanographic features can be crucial to monitoring population dynamics of commercially important species. The use of more detailed mangrove creek habitat-complex characterization can improve Marine Protected Area (MPA) planning.— 1 University of Miami, Department of Biology, Miami, Florida, U.S.A.

A HYBRID MAXIMUM-LIKELIHOOD METHOD FOR DERIVING LARGE-SCALE MANGROVE DISTRIBUTIONS IN FLORIDA FROM THEMATIC MAPPER DATA b y J e ff Ueland.—The mangrove habitats of Florida have experienced significant change in recent years due to both anthropogenic and natural forces. Understanding regional-scale changes of mangroves, so as to be utilized in conjunction with large-scale datasets on marine fisheries, requires a systematically derived and comparable spatial inventory of these important habitats. This paper offers a method for this purpose by examining the spatial extent and change of mangrove habitat in a 14 county area of south Florida between 1987 and 2000. To this end, remotely sensed, Thematic Mapper satellite data were classified by coupling posterior prob- abilities from a maximum-likelihood, supervised classification process with spatially explicit ancillary data sources. The results showed a decline in mangrove habitat over the 13 yr period, but the changes were uneven and varied across space.— Ohio Univeristy, Athens, Ohio, U.S.A.

MANGROVES AND SEAGRASS BEDS AS DIURNAL FEEDING HABITATS FOR JUVE- NILE HAEMULON FLAVOLINEATUM by Marieke C. Verweij 1, Ivan Nagelkerken 1, Susanne L. J. Wartenbergh 1, Ido R. Pen 2, and Gerard van der Velde 1.—Caribbean seagrass beds suppos- edly are important feeding habitats for so-called nocturnally active zoobenthivorous fish, but the extent to which these fishes use mangroves and seagrass beds as feeding habitats during daytime remains unclear. Therefore, we studied daytime behavior of large juvenile (5–10 cm) and sub-adult (10–15 cm) Haemulon flavolineatum in mangroves and seagrass beds in Cu- raçao. Sub-adults occurred in mangroves only, spent most time on resting, and showed rare opportunistic feeding events, regardless of their social mode (solitary or schooling). They probably feed predominantly during the night in seagrass beds. Large juveniles were present in both habitat types and solitary fishes mainly foraged, while schooling fishes mainly rested. Large juveniles showed more feeding activity in seagrass beds than in mangroves. The study shows that both mangroves and seagrass beds provide daytime feeding habitats for some life- stages of H. flavolineatum, which is generally considered a nocturnal feeder.— 1 Department of Animal Ecology and Ecophysiology, Institute for Water and Wetland Research, Radboud University, Nijmegen, The Netherlands.2 Theoretical Biology Group, Centre for Ecological and Evolutionary Studies, University of Groningen, Haren, The Netherlands. FIRST INT. SYMP. ON MANGROVES AS FISH HABITAT: ABSTRACTS 935

DEVELOPING SHALLOW-WATER ACOUSTIC TELEMETRY METHODS FOR JU- VENILE SNAPPER HABITAT STUDIES IN THE FLORIDA KEYS NATIONAL MARINE SANCTUARY by Samantha R. Whitcraft 1, Bill Richards 2, John Lamkin 2, Trika Gerard 2, Tom Carlson 3, Geoff McMichael 4, Jessica Vucelick 4, Greg Williams 5, and Lisa Pytka 6.—The recent availability of customized coded micro-transmitters for juvenile fish tagging may provide useful tools in determining the early life history habitat requirements of commercially valuable snapper species within and outside marine protected areas. This pilot study utilizes micro-acoustic tags (dry wt = 0.65 g; excess mass = 0.39 g; 417 kHz) and standard acoustic telemetry methodologies developed by Battelle and NOAA Fisheries to investigate questions of habitat-use patterns of juvenile snappers in Florida Keys National Marine Sanctuary’s reserve areas. We apply these methodologies to examine specific physical, environmental, and biological factors with the goal of optimizing telemetry instrumentation and field techniques for this application. Optimization requires trade-offs between limitations imposed by biological constraints and factors influencing the propagation/detection of signals in estuarine/marine environments. Important physical factors are operating frequency, encoded pulse length, and vegetation density in mangrove and seagrass habitats in 1.5–6 m depths. Environmental factors include determining baseline configuration within the geometry of distinct habitat patches and effects of dense mangrove prop roots on acoustic signal propagation and reception. Biological factors include tag-effects on survivorship of juvenile snappers. We present initial data from shallow-water testing of acoustic reception in obstruction-rich environments, compare juvenile salmon and snapper survivorship, and range-testing strategies.— 1 Cooperative Institute for Marine and Atmospheric Studies (CIMAS), Rosenstiel School of Ma- rine and Atmospheric Science, University of Miami, Miami, Florida, U.S.A. 2 Early Life History Lab, NOAA Southeast Fisheries Science Center, Miami, Florida, U.S.A. 3 Pacific Northwest National Lab/Battelle, Portland, Oreon, U.S.A. 4 Ecology Group, Pacific Northwest National Lab/Battelle, Richland, Washington, U.S.A. 5 Sequim Marine Sciences Lab, Pacific Northwest National Lab/Batelle, Sequim, Washington, U.S.A. 6 New College of Florida, Sarasota, Flori- da, U.S.A.