Back Reef II - Perry Institute for Marine Science

Transcription

Back reef systems: important but overlooked components of tropical marine ecosystems
Craig Dahlgren and John Marr ....................................................................................................145
Processes influencing recruitment inferred from distributions of coral reef fishes
Aaron J. Adams and John P. Ebersole ...........................................................................................153
Fish density, diversity, and size-structure within multiple back reef habitats of Key West National
Wildlife Refuge
David B. Eggleston, Craig P. Dahlgren, and Eric G. Johnson ..................................................175
Habitat associations of adult queen conch (Strombus gigas L.) in an unfished Florida Keys back reef:
applications to essential fish habitat
Robert A. Glazer and James A. Kidney ........................................................................................205
BULLETIN OF MARINE SCIENCE
Dedication ................................................................................................................................................143
The seascape approach to coral ecosystem mapping: an integral component of understanding the habitat
utilization patterns of reef fish
Matthew S. Kendall, Ken R. Buja, John D. Christensen,
Curtis R. Kruer, and Mark E. Monaco .........................................................................................225
Coral Reef Watch 2002
Alan E. Strong, Gang Liu, Jill Meyer,
James C. Hendee, and Desiree Sasko .............................................................................................259
MARINE SCIENCE
(ISSN 0007-4977)
Volume 75
SEPTEMBER 2004
Number 2
Volume 75
The impact of Hurricane Georges on soft-bottom, back reef communities: site- and species-specific
effects in south Florida seagrass beds
James W. Fourqurean and Leanne M. Rutten ............................................................................239
Bulletin of
Transport processes linking shelf and back reef ecosystems in the Exuma Cays, Bahamas
Ned P. Smith .......................................................................................................................................269
•
Wind-mediated diel variation in flow speed in a Jamaican back reef environment: effects on ecological
processes
Salvatore J. Genovese and Jon D. Witman ..................................................................................281
Number 2
Back reef habitats: ecological analysis and characterization
•
John Marr
Guest Editor
Large-scale ecological impacts of development on tropical islands systems: comparison of developed
and undeveloped islands in the central Bahamas
Kathleen Sullivan Sealey.............................................................................................................295
Pages 143–334
Linking habitat protection and marine protected area programs to conserve coral reefs and associated
back reef habitats
Michelle A. Duval, Douglas N. Rader, and Kenyon C. Lindeman .............................................321
Lee Stocking Island, Bahamas
December 2001
Sponsors:
U.S. Environmental Protection Agency
National Oceanic and Atmospheric Administration
Environmental Defense
Perry Institute for Marine Science
(Continued on outside back cover)
SEPTEMBER 2004
ROSENSTIEL SCHOOL
OF MARINE AND ATMOSPHERIC SCIENCE
UNIVERSITY OF MIAMI
www.rsmas.miami.edu/bms
PROCESS BLUE BLACK
Bulletin of
MARINE SCIENCE
(ISSN 0007-4977)
The Bulletin of Marine Science is dedicated to the dissemination of high quality research dealing with the tropical
and subtropical waters of the world’s oceans. All aspects of marine science are treated by the Bulletin of Marine Science, including papers in marine biology, biological oceanography, fisheries, marine affairs, applied marine physics,
marine geology and geophysics, marine and atmospheric chemistry, and meteorology and physical oceanography.
The Bulletin of Marine Science is now in its 54th year of publication. Commencing in 1951 the first volumes were
published as the Bulletin of Marine Science of the Gulf and Caribbean. The name was changed in 1965 (volume
15) to the present one. It is published by the Rosenstiel School of Marine and Atmospheric Science, University of
Miami, 4600 Rickenbacker Causeway, Miami, Florida 33149-1098 U.S.A.
EDITORIAL BOARD
Su Sponaugle, Editor
Division of Marine Biology and Fisheries, Rosenstiel School of Marine and Atmospheric Science, University
of Miami, 4600 Rickenbacker Causeway, Miami, Florida 33149-1098 U.S.A.
Rafael J. Araújo, Assistant Editor
Ivan Goodbody, Associate Editor, Zoology
Department of Zoology, University of the West Indies, Kingston 7, Jamaica
Peter B. Ortner, Associate Editor, Biological Oceanography
Atlantic Oceanographic and Meteorological Laboratory, National Oceanic and Atmospheric
Administration, Miami, Florida 33149 U.S.A.
Daniel O. Suman, Associate Editor, Marine Affairs
Division of Marine Affairs and Policy, Rosenstiel School of Marine and Atmospheric Science,
University of Miami, 4600 Rickenbacker Causeway, Miami, Florida 33149-1098 U.S.A.
Nancy A. Voss, Associate Editor, Marine Biology
EDITORIAL OFFICE
Manuscripts should be sent to the Editor for consideration for publication. The submission of a manuscript will
be taken to imply that the manuscript is original and has not been submitted or published elsewhere in its present or
modified form. Instructions for manuscript preparation are located on the inside back cover. Other matters should
be sent to the Editorial Office, Bulletin of Marine Science, Rosenstiel School of Marine and Atmospheric Science,
University of Miami, 4600 Rickenbacker Causeway, Miami, Florida 33149-1098 U.S.A. or e-mailed to bms@rsmas.
miami.edu.
The Bulletin of Marine Science (ISSN 0007-4977) is published six times per year in two volumes for the Rosenstiel
School of Marine and Atmospheric Science of the University of Miami by Allen Press, Inc. 1041 New Hampshire
St., Lawrence, Kansas 66044 U.S.A. Periodicals postage paid at Lawrence, Kansas. POSTMASTER: Send address
changes to Bulletin of Marine Science, Rosenstiel School of Marine and Atmospheric Science, University of Miami,
P.O. Box 971, Key Biscayne, Florida 33149-0971 U.S.A.
SUBSCRIPTIONS AND BACK ISSUES
Bulletin of Marine Science publishes six issues per year in two volumes of three issues each commencing with
volume 34. U.S. Subscription rates for 2004 are: $260 for institutions ($280 other countries), $110 for individuals
($130 other countries). Individual subscriptions must be placed directly with the subscription office and paid for by
personal check, money order, or credit card. New institutional subscribers will receive series of back issues while
supplies last. Subscriptions and changes of address should be sent to Bulletin of Marine Science, P.O. Box 971,
Key Biscayne, Florida 33149-0971 U.S.A. Back issues are available for volumes 1–17 from Swets and Zeitlinger,
B.V., Backsets Department, Heereweg 347, Lisse, The Netherlands, and for volumes 18 to present from the Bulletin
of Marine Science, P.O. Box 971, Key Biscayne, Florida 33149-0971 U.S.A. The Bulletin is also availbale on CD
(PDF format).
Copyright © 2004 by the Rosenstiel School of Marine and Atmospheric Science of the University of Miami.
This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper).
INSTRUCTIONS TO AUTHORS
Potential contributors to the Bulletin of Marine Science should adhere strictly to the following instructions. Authors must submit an original and three true copies of the manuscript complete with title page, figures, and tables.
Manuscripts that are poorly written, not formatted according to the requirements of the Bulletin, or otherwise flawed,
will be returned to the author without review. Authors are also requested to submit the names and complete postal
addresses of up to five potential reviewers. Authors are required to retain a complete copy of their manuscript and
illustrations identical in every respect to the original submitted manuscript.
The Bulletin accepts monographs (16–30 pages), reviews and articles (10–15 pages), reports (up to 9 pages), notes
(up to 5 pages with no abstract), discussions and replies (up to 3 pages), and informative stand-alone black and white
photographs with legends (1 page). Papers over 30 pages may be considered for separate publication in the Bulletin’s
“Studies in Tropical Oceanography” series. Prior to the submission of photographs, the Editor should be consulted
for guidance. Color photographs will incur a page charge of up to $1200.
Papers are first reviewed for suitability by the Editor. Suitable papers are then critically reviewed by up to three
experts in the relevant discipline. Based on their advice, the Editor accepts the paper, accepts the paper contingent
on revision, or rejects the paper. Papers revised at the invitation of the Editor are sent to a fourth reviewer. Revised
papers that are not resubmitted within six months will be considered rejected. Final acceptance is made when the
paper has been satisfactorily revised and the author has submitted the final manuscript in the required electronic
format. Authors will be notified immediately of the acceptance of their manuscript for publication.
Language: Only papers in English will be accepted. Spelling and word division should follow either the Concise
Oxford Dictionary of Current English or The Merriam-Webster Third New International Dictionary.
Content of Manuscript: Authors are strongly advised to consult a recent issue of the Bulletin and follow the
style and general layout of articles.
The title page should contain (a) the full title of the paper in 20 words or less and reference the geographical
location of the research if relevant; (b) a running head of not more than 52 characters and spaces; (c) the full name
and address of each author and the name of the institution in which the research was conducted; (d) a list of up to 10
keywords; and (e) the name, current postal address, and E-mail address of the corresponding author.
The abstract should be a single paragraph of not more than 200 words and written in the passive voice. It should
briefly state the reason for the research, summarize the significant findings, and note the implication of those findings.
Figures and tables, with their legends and headings, should be self-explanatory and not require reference to the
text. Line drawings should be scaled for 50% reduction. Oversize figures and tables will not be accepted unless the
author is willing to assume the cost of size reduction.
Figures should be clear and legible on all copies, with 1-inch margins. Figures will be reduced to 12.5 × 20.0 cm
printed page size. The dimensions and the thickness of the individual lines, spaces, and letters will also be reduced
proportionally. Authors are urged to refer to the Council on Biological Editors Style Manual, 6th edition, for guidance in the preparation of their figures.
Do not send original drawings to the Editor; if needed they will be requested.
Each table should start on a separate page. Consistency in headings and format is desirable. Vertical rules should
be avoided.
Acknowledgments, if included, are placed at the end of the text.
Literature is cited in the text as : Smith and Jones (1968) or (Smith and Jones, 1968). All papers referred to in
the text should be listed alphabetically by the senior author’s surname under the heading “Literature Cited.” Only
the authors’ surnames and initials are required in the Literature Cited. The accuracy of the Literature Cited is the
responsibility of the author. Abbreviations of names of periodicals should conform to the Council of Biological Editors Style Manual, 6th edition, or the World List of Scientific Periodicals.
Common abbreviations and symbols, such as %, mm, m, g, ml, mg, oC (for Celsius), µm, and so forth, should be
used. Abbreviate units of measure only when used with numerals. Periods are only rarely used with abbreviations.
We prefer that measurements be given in metric units; other equivalent units may be given in parentheses.
Footnotes are not accepted. Such ancillary information should appear in the text set off in [brackets].
Form of the Manuscript: The submitted manuscript should be typed, double-spaced, on white paper. The
sequence of the material should be: TITLE PAGE, ABSTRACT, TEXT, LITERATURE CITED, AUTHOR ADDRESSES, APPENDIX, TABLES (each table should be numbered with an arabic numeral and heading provided),
LIST OF FIGURES (Entire figure legends), and FIGURES (each figure should be numbered with an arabic numeral).
When the figures form a collage the figure numbers should be permanently affixed to each figure or identified as
upper left, lower right, etc.
Final Manuscript Submission: Following acceptance of the revised manuscript, the author will be required to
submit an electronically-prepared, correctly-formatted manuscript for publication. Computer formatting instructions
will be sent to the author when the manuscript has been accepted for publication.
The formatted manuscript may be sent as an e-mail attachment or enclosure to bms@rsmas.miami.edu, or, a
disc or CD may be submitted in which case the author is required to identify the word processing program and version number (e.g., Microsoft Word® 6.0, WordPerfect® 6.1) and the computer platform (Apple Macintosh® or IBM
compatible PC). Original illustrations must be submitted separately by regular mail even if they are included in the
electronic submission.
Following submittal of the final manuscript, as per the submittal instructions, the manuscript will be set up for
publication and a galley proof returned to the author for final proof reading. The author will be responsible for any
errors that appear in the final printed version.
Page Charges: Page charges will be assessed at the rate of $70 per printed page. Generally, 2.5 manuscript pages
equal one printed page.
Dedication
W
e dedicate this volume to Mr. John Heisler
of the U.S. EPA for his leadership role in
ensuring that this workshop took place and
for focusing efforts on the importance of back
reef ecosystems. This publication is the result
of many scientists and managers convening at
a workshop held at the research station on Lee
Stocking Island, Bahamas, in December of 2001.
BULLETIN OF MARINE SCIENCE, 75(2): 145–152, 2004
BACK REEF SYSTEMS: IMPORTANT BUT OVERLOOKED
COMPONENTS OF TROPICAL MARINE ECOSYSTEMS
Craig Dahlgren and John Marr
Introduction to Back Reef Systems
Tropical marine ecosystems are composed of a complex mosaic of habitats and interconnected ecological communities that extend from the shoreline to the open ocean.
Because of their inherent beauty, high biodiversity, economic importance, and threatened status, coral reefs are the most widely recognized and best-studied tropical marine
ecosystem. While we still have much to learn about coral reefs, we have made significant
strides towards furthering our understanding of these ecological communities over the
past few decades. More recently, national and international attention has been focused
on coral reefs, as evident with 1997 being declared the International Year of the Reef,
the creation of government bodies such as the multi-agency Coral Reef Task Force in the
U.S., and the establishment of international organizations focused on research, monitoring, and management of coral reefs.
The more we learn about coral reefs, the more we appreciate the complexity of ecological interactions within this ecosystem and the interdependence between coral reefs
and the near shore and offshore marine systems that surround them. The complexity and
critical nature of these surrounding ecosystems, however, have received comparatively
little attention. The contribution of these back reef ecosystems to the overall biodiversity of tropical marine systems, to coral reef ecosystems, and value to humans has only
recently been addressed. Of particular importance are shallow water ecosystems that
separate the reef from the shoreline. Back reef systems refer to the ecosystems between
the reef crest and the shoreline’s upper intertidal zone. Back reef systems include a mosaic of interconnected habitats ranging from the coastal margin, lagoonal, patch reef
habitats, mangroves, tidal creeks, seagrass beds, intertidal and subtidal sand and mud
flats, and macroalgal beds. While this system represents a diverse mix of habitats and
ecological communities, the components of the system serve many similar essential
ecological functions, face many common threats, and are interconnected by biophysical
processes.
Overview of the Ecological Function of Back Reef Systems
Thriving coral reefs must be bathed by warm oligotrophic seawater and buffered from
terrestrial inputs by surrounding ecosystems. Thus, the health of coral reef ecosystems
is inextricably linked with surrounding offshore, inshore and coastal ecosystems (Ogden, 1988). Back reef systems serve as a critical buffer between coral reefs and terrestrial sources of stress. Mangroves and seagrass beds stabilize sediments and prevent the
smothering and scouring of coral reefs. Sediment loads also reduce light penetration to
the reefs and block the settlement of coral larvae (Ogden and Gladfelter, 1983; Rogers,
1990; Hemminga and Duarte, 2000). Mangroves and seagrasses may serve as a nutrient
sink to prevent corals from experiencing eutrophic conditions (Ogden and Gladfelter,
1983; Hemminga and Duarte, 2000). Although seagrasses and mangroves may also rely
on terrestrial sources of nutrients to support their high rates of primary productivity
Bulletin of Marine Science
© 2004 Rosenstiel School of Marine and Atmospheric Science
of the University of Miami
145
146
BULLETIN OF MARINE SCIENCE, VOL. 75, NO. 2, 2004
(Zieman, 1982), they incorporate a significant amount of dissolved nutrients in their biomass. For example, net primary productivity of seagrass beds across the Bahama Banks,
−
can reach levels of 1062 g C m−2
y−1, and net primary productivity of seagrasses for the
entire Bahama Banks may represent 0.2% of global ocean productivity (Dierssen et al.,
2003). The high rates of primary productivity within mangroves and seagrasses support diverse benthic and fish communities in these areas. Certain fish and invertebrates
commonly associated with reef structures make foraging forays into these habitats on a
regular basis (e.g., Ogden and Zieman, 1977; Nagelkerken et al., 2000c). Export of this
productivity to coral reefs includes the movements of fish and invertebrates from back
reef systems to coral reefs. Thus, back reef systems are critical for supporting fisheries,
and maintaining fish and invertebrate diversity in tropical marine ecosystems.
Many adult fish and invertebrate species common to coral reefs use back reef systems
as juveniles. The spatial proximity of certain back reef habitats to coral reefs can shape
the species structure, especially for the reef fishes that are dependent on the back reef
nursery habitats (Nagelkerken et al., 2000c, 2002). Understanding the functional role
of nursery habitats within back reef systems and their connectivity to adult habitats is
essential for protecting tropical marine biodiversity and managing many important fisheries (Gillanders et al., 2003).
Nursery habitats in back reef systems can also influence fisheries. Several important
fishery species in the Caribbean, such as the Caribbean spiny lobster, Panulirus argus
(Latreille, 1804), for example, settle from the plankton into back reef nursery systems
and reside there until they approach adult sizes (reviewed by Lipcius and Eggleston,
2000). Seagrass and hardbottom habitats within back reef systems are among the most
important juvenile and adult habitats for the queen conch, Strombus gigas (Linnaeus,
1758), one of the most important commercial and artisanal fishery species in the Caribbean (Stoner, 1997). In areas where these nurseries are rare, there may be lower adult
densities available to fishers, than in areas where nurseries are common in back reef systems (e.g., Lipcius et al., 1997; Acosta, 1999). Economically important finfish species,
including snapper and grunt species utilize seagrass, mangrove, macroalgal, and shallow patch reef habitats within the back reef system as nurseries (e.g., Nagelkerken et al.,
2000c,b, 2002). Other fishery species that are considered threatened or endangered, like
the Nassau grouper, Epinephelus striatus (Bloch, 1792), live in various habitats within
back reef systems as juveniles before moving onto coral reefs as adults (Eggleston, 1995;
Dahlgren and Eggleston, 2001).
Threats to Back Reef Systems
Like coral reef ecosystems, back reef systems face a wide range of anthropogenic
and environmental threats. Because they are located in close proximity to land and are
generally in shallow waters, humans have great access to back reef systems. As a result,
they are subject to a variety of uses and impacts resulting from coastal development,
point and non-point source pollution, shipping, species introductions, commercial fishing, and recreational activities (e.g., boating, fishing, snorkeling, diving, and shell collecting). While these activities and uses are of a tremendous economic value to local
communities, they reduce the ecological value of back reef systems when not properly
managed. Boating through the shallow waters of back reef systems can lead to groundings or anchoring that scar seagrass beds (Hudson and Goodwin, 2001) and dislodge or
DAHLGREN AND MARR: INTRODUCTION TO BACK REEFS
147
break coral heads on patch reefs (Rogers, 1985). Even such localized disturbances can
fragment or destroy habitats and recovery may take years to decades (e.g., Reigl, 2001).
Fishing may have a wide range of impacts to back reef systems. Overfishing of key
species may disrupt back reef community structure and species diversity and alter ecosystem functions. Many of the commercially important fish and invertebrates that inhabit back reef systems do so primarily as juveniles, therefore removal of these species
from back reef systems, either in a targeted fishery or as bycatch, can lead to the collapse
of stocks. Even when adults are targeted, intensive fishing in back reef systems can cause
population collapse (e.g., Berg et al., 1992). Destructive fishing practices that alter habitats, such as trawling or the use of poisons or explosives, also degrade these systems and
reduce their ability to support fisheries and potential for ecosystem recovery (e.g., Reigl
and Luke, 1998; Reigl, 2001).
Major disturbances to back reef systems resulting from land uses and coastal developments can increase runoff that leads to sediment loading, nutrient enrichment, or chemical contamination in back reef systems and ultimately the reef proper (e.g., Lapointe
et al., 1990; Lapointe and Matzie, 1996; Nemeth and Sladek Nowlis, 2001). These terrestrial inputs degrade water quality, alter habitats, and may cause key species in the
ecosystem to die off (Sebens, 1994). Even greater impacts to back reef systems can occur
through human alterations to coastlines and the sea floor, such as dredging and filling.
Dredging not only removes areas of the sea floor, often replacing habitats of high productivity and structural complexity with unvegetated areas of low structural complexity, but also increases sediment loads in the water column, smothering adjacent benthic
habitats, and reducing light penetration into the water. Similarly, dredge spoils used
as fill along the coastline often remove productive and structurally complex mangrove
shorelines and adjacent seagrass beds. This can affect fish and invertebrate communities
that rely on these habitats (Lindeman and Snyder, 1999). Removal of mangroves and
other coastal vegetation, whether directly (e.g., dredging) or indirectly (e.g., dredge spoil
placement), eliminates a natural buffer from terrestrial inputs to back reef systems and
results in even greater sediment loads and further reduced water quality with adverse
effects to the ecosystem. The role as a natural buffer to terrestrial inputs that mangroves
and emergent plant communities serve is critical for reducing impacts from erosion and
flooding, particularly for areas that are prone to hurricanes.
The shallow nature of back reef systems contributes to their high productivity, and
makes them extremely vulnerable to fluctuations in environmental conditions. Because
conditions within back reef systems are not buffered by depth or the regular influx of
oceanic water, as is the case on deeper and more exposed coral reefs, environmental
fluctuations are often greater in back reef systems than on adjacent coral reefs. For instance, extended periods of elevated temperatures during doldrums conditions can cause
waters within back reef systems to heat rapidly, often to extremely high temperatures.
Such prolonged periods of elevated water temperatures can cause corals on patch reefs
to bleach, and if conditions are extreme or persist for an extended period, such bleaching may lead to mass mortality of corals (e.g., Ogden and Wicklund, 1988; Glynn, 1996;
Wilkinson et al., 1999). Mass mortality of many coral species was observed on patch
reefs in lagoonal areas around the world during the bleaching event of 1998 (e.g., Aronson et al., 2000; Goreau et al., 2000; Aronson et al., 2002).
Despite the apparent economic and ecological importance of back reef systems, and
the extent of serious threats that these systems face, there is a limited understanding
about their ecosystem function. Although we understand the basic linkages between
148
back reef systems and surrounding ecosystems, we must now collect more quantitative
data to measure interactions and connectivity in such a way that will allow us to assess
the ecological importance and determine why these interactions vary at different spatial
and temporal scales. Similarly, it will be important to understand the interactions of
human impacts and natural stresses so to evaluate impacts to the structure and function
of back reef systems. To effectively manage back reef systems, we must improve our
understanding of these ecosystems, and how humans affect them.
Back Reef Systems in the Spotlight: Research, Monitoring, and Management
In 2001, the Caribbean Marine Research Center hosted a workshop to discuss science
and management of back reef systems. At this workshop, researchers specializing in
a wide range of disciplines presented current findings on various aspects of back reef
systems, including ecological processes affecting populations and communities within
back reef systems, the effects of human impacts and natural disturbances to back reef
systems, and various advances in tools and techniques used to study back reef systems.
In addition to these presentations, researchers and marine resource managers discussed
the current status of our understanding of back reef systems and identified future priorities for research, monitoring, and management of back reef systems (Table 1). This
workshop was the first of its kind to address back reef systems using a multidisciplinary
approach. Resulting from the workshop was an exchange of information and ideas that
had not been previously realized by the researchers and resource managers who normally tend to present their developments and findings within their own disciplines.
Several papers presented at this workshop are included in this volume. They represent research on various habitats, species, communities, ecological processes, human
impacts, and restoration of back reef systems. One focal topic of the workshop was an
examination of the importance of back reef systems to fish and invertebrate populations
and communities. Adams and Ebersole, for example, discuss the importance of lagoonal
patch reefs as a nursery area for various fish species in the Virgin Islands. Similarly,
Eggleston et al. discuss characteristics of fish communities in various habitats, including
mangroves, seagrasses, hardbottom areas, lagoonal patch reefs and channels, within the
back reef systems of the lower Florida Keys. Both of these papers indicate the importance of back reef systems for supporting reef fish populations, particularly as nursery
areas for several species.
Several studies examined how the spatial configuration of habitats within a seascape
affects fish and invertebrate populations. The contribution by Kendall et al. integrates in
situ censuses of fishes with habitat maps derived from satellite and aerial images, to examine the interaction between habitat configuration and populations of reef fishes within
a seascape. The paper by Glazer and Kidney focuses on habitat use and spatial population dynamics of an important invertebrate fishery species, the queen conch, S. gigas.
They examine the movement of queen conch within and between somewhat isolated
populations with respect to habitat features of the area. Both of these papers illustrate the
importance of seascape features to population dynamics and apply recently developed
technology such as remote sensing and biotelemetry to address these issues.
Other papers focus on ecological processes that may affect populations and communities within back reef systems. On a regional to global scale, Strong et al. described the
use of remote sensing and in situ monitoring stations to measure environmental parameters that are linked to physiological responses in corals and other species inhabiting back
149
Table 1. Summary of priority areas identified by workshop participants for (A) management
and (B) research and monitoring in back reef systems.
(A) Priority management goals for back reef systems
1.
Minimize disturbances over time
2.
Ensure sustainable use of resources
3.
Preserve ecological function of back reef systems
4.
Restore back reef systems that are deemed restorable
(B) Research and monitoring priorities for effective management of back reef systems
1.
Inventory and baseline data on the distribution and abundance of species and habitats to
prioritize areas for management
2.
Focused research and monitoring to assess effectiveness of management strategies,
including marine protected areas
3.
Monitoring human activities within back reef systems and the resulting effects or
impacts of these uses
4.
Studies that investigate the population dynamics of key species in the back reef system,
including habitat associations, ontogenetic habitat shifts, and connectivity between local
populations
5.
Studies that identify the spatial and temporal scales at which important ecological
processes and human impacts operate within back reef systems, and the resolution
necessary to detect them
6.
Hypothesis driven studies that examine the connectivity or linkages within back reef
systems and between back reef and surrounding ecosystems
7.
Comparative studies of areas subject to different levels of habitat quality, environmental
conditions, exploitation or other natural and anthropogenic stresses
8.
Monitoring of key populations, communities and environmental parameters that
influence to detect long-term trends and responses to natural or anthropogenic events
within the back reef system
9.
Studies of the socio-economic implications of different management strategies for back
reef systems
10.
Better coordination among research programs to integrate different disciplines
11.
Greater participation of resource users in research and monitoring activities
12.
Information on the importance or value of back reef systems is needed in a format that
is more accessible to marine resource managers and the public
reef habitats. Information gained from these complementary approaches to monitoring
can enhance our understanding of the factors that cause coral bleaching and will enhance our ability to predict where and when these events occur. At a much smaller spatial scale, on the order of kilometers to tens of kilometers, the paper by Smith examines
water transport within a back reef system and between the back reef system and adjacent
open ocean and coral reef ecosystems. Such physical linkages influence the transport
of larvae from the open ocean to nursery habitats within back reef systems. On even
smaller spatial scales, Genovese and Witman examine how daily variability in winds affects water flow and subsequently the foraging behavior of a species within the back reef
system of Jamaica. This collection of papers highlights the way that oceanographic and
atmospheric factors affect the dynamics of communities and populations within back
reef systems on a variety of spatial and temporal scales. They also provide examples of
traditional and new advances in tools or techniques for addressing these issues.
A number of papers examine both natural and anthropogenic disturbances to back
reef systems, and mechanisms for protecting back reef systems. Long-term monitoring
of seagrass beds within the Florida Keys by Fourqurean and Rutten, for example, illustrates both the short- and long-term impacts that hurricanes have on seagrass ecosystems. Sullivan-Sealey examines long-term temporal changes to back reef systems of the
150
Bahamas and how human population growth and development over several decades have
contributed to these changes. Finally, Duvall et al. examines the framework for managing back reef systems to protect both habitats and fishery species. Understanding the
existing mechanisms for management and how various mechanisms may be integrated
is essential to ensuring that the diverse habitats that comprise the back reef system and
the species that reside there are protected against critical threats.
Although these papers do not present a comprehensive overview of back reef systems
by any means and are primarily focused on tropical and subtropical back reef systems of
the western Atlantic, they provide representative examples of current studies of back reef
systems and methods, and the results and conclusions described in these papers have implications for the study and management of back reef systems around the world. Clearly
we are making progress towards understanding these systems, their importance, and
how humans impact them. With more effective management now possible, it is essential
to better protect the critical ecological roles of back reef systems. Further research and
monitoring will greatly improve resource management, which will in turn effectively
maintain the integrity and ensure sustainable use of back reef systems.
Acknowledgements
This paper is funded in part by a grant from the Caribbean Marine Research Center (CMRC
Project #CMRC-00-IXNR-03-01A) National Oceanic and Atmospheric Administration (NOAA)
Undersea Research Program, U.S. Environmental Protection Agency, and Environmental Defense. Views expressed herein are those of the authors and do not necessarily reflect the views of
CMRC, NOAA, or any of the supporting agencies. We are thankful for the use of the research
station at Lee Stocking Island, Bahamas, to host this workshop and to the staff at CMRC for
their logistical support. In addition, special thanks to J. O’Neill for working with the authors and
reviewers in finalizing the manuscripts and to the many reviewers for improving the content of
this series.
Literature Cited
Acosta, C. A. 1999. Benthic dispersal of Caribbean spiny lobsters among insular habitats: implications for the conservation of exploited marine species. Cons. Biol. 13: 603–612.
Aronson, R. B., W. F. Precht, I. G. Macintyre, and T. J. T. Murdoch. 2000. Coral bleach-out in
Belize. Nature 405: 36 p.
____________, __________, M. A. Toscano, and K. H. Koltes. 2002. The 1998 bleaching event
and its aftermath on a coral reef in Belize. Mar. Biol. 141: 435–447.
Berg, C. J. Jr., R. Glazer, J. Carr, J. Krieger, and S. Acton. 1992. Status of the queen conch, Strombus gigas, in Florida waters. Proc. Gulf. Carib. Fish. Inst. 41: 439–443.
Dahlgren, C. P. and D. B. Eggleston. 2001. Spatio-temporal variability in abundance, size and
microhabitat associations of early juvenile Nassau grouper Epinephelus striatus in an off-reef
nursery system. Mar. Ecol. Prog. Ser. 217: 145–156.
Dierssen, H. and R. Zimmerman, R. Leathers, T. Downes, and C. Davis. 2003. Remote sensing of
seagrass and bathymetry in the Bahamas banks using high resolution airborne imagery. Limnol. Oceangr. 48: 444–455.
Eggleston, D. B. 1995. Recruitment in Nassau grouper Epinephelus striatus: post-settlement abundance, microhabitat features and ontogenetic habitat shifts. Mar. Ecol. Prog. Ser. 124: 9–22.
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.
151
Glynn, P. W. 1996. Coral reef bleaching: facts, hypotheses and implications. Global Change Biol.
2: 495–509.
Goreau, T., T. McClanahan, R. Hayes, and A. Strong. 2000. Conservation of coral reefs after the
1998 global bleaching event. Cons. Biol. 14: 5–15.
Hemminga, M. and C. Duarte. 2000. Seagrass ecology. Cambridge University Press. Cambridge.
310 p.
Hudson, J. H. and W. B. Goodwin. 2001. Assessment of vessel grounding injury to coral reef and
seagrass habitats in the Florida Keys National Marine Sanctuary, Florida: protocol and methods. Bull. Mar. Sci. 69: 509–516.
Lapointe, B. E., J. D. O’Connel, and G. S. Garret. 1990. Nutrient couplings between on-site sewage disposal systems, groundwaters, and nearshore surface waters of the Florida Keys. Biogeochemistry 10: 289–307.
____________ and W. R. Matzie. 1996. Effects of stormwater nutrient discharges on eutrophication processes in nearshore waters of the Florida Keys. Estuaries 19: 422–435.
Lindeman, K. C. and D. B. Snyder. 1999. Nearshore hardbottom fishes of southeast Florida and
effects of habitat burial caused by dredging. Fish. Bull. 97: 508–525.
Lipcius, R. N. and D. B. Eggleston. 2000. Ecology and fishery biology of spiny lobsters. Pages
1–41 in B. F. Phillips and J. Kittaka, eds. Spiny lobsters: fisheries and culture. Fishing News
Books, Oxford.
__________, W. T. Stockhausen, D. B. Eggleston, L. S. Marshall Jr., and B. Hickey. 1997. Hydrodynamic decoupling of recruitment, habitat quality and adult abundance in the Caribbean spiny
lobster: source-sink dynamics? Mar. and Fresh. Res. 48: 807–815.
Nagelkerken, I. G. van der Velde, M. W. Gorissen, G. J. Meijer, and T. van’t Hof. 2000a. Importance of mangroves, seagrass beds and the shallow coral reef as a nursery for important coral
reef fishes, using a visual census technique. Est. Coas. Shelf Sci. 51: 31–44.
______________, M. Dorenbosch, W. C. E. P. Verberk, E. Cocheret de la Moriniere, and G. van
der Velde. 2000b. Importance 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.
_______________, ____________, ________________, ______________________, and _____
________. 2000c. 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.
________________, C. M. Roberts, G. van der Velde, M. Dorenbosch, M. C. van Reil, E. Cocheret
de la Moriniere, 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.
Nemeth, R. and J. Sladek Nowlis. 2001. Monitoring the effects of land development on the nearshore reef environment of St. Thomas, USVI. Bull. Mar. Sci. 69: 759–775.
Ogden, J. C. 1988. The influence of adjacent systems on the structure and function of coral reefs.
Proc. 6th Intl. Coral Reef Symp. 1: 123–129.
__________ and E. H. Gladfelter. 1983. Coral reefs, seagrass beds and mangroves: their interaction in the coastal zones of the Caribbean. UNESCO reports in marine science 23. 34 p.
__________ and R. Wicklund, eds. 1988. Mass bleaching of coral reefs in the Caribbean: a research strategy. National Undersea Research Program Research report 88-2. 51 p.
__________ and J. C. Zieman. 1977. Ecological aspects of coral reef-seagrass bed contacts in the
Caribbean. Proc. 3rd Intl. Coral Reef Symp. 377–382 p.
Reigl, B. 2001. Degredation of reef structure, coral and fish communities in the red sea by ship
groundings and dynamite fisheries. Bull. Mar. Sci. 69: 595–611.
______ and K. E. Luke. 1998. Ecological characteristics of dynamited reefs in the northern Red
Sea and their relevance to reef rehabilitation. Mar. Poll. Bull. 37: 488–498.
Rogers, C. S. 1985. Degredation of Caribbean and western Atlantic coral reefs and decline of associated fisheries. Proc. 5th Intl. Coral Reef Symp. 6:491–496.
152
__________. 1990. Responses of coral reefs and reef organisms to sedimentation. Mar. Ecol. Prog.
Ser. 62: 185–202.
Sebens, K. P. 1994. Biodiversity on coral reefs: what we are losing and why? Amer. Zool. 34:
115–133.
Stoner, A. W. 1997. The status of queen conch, Strombus gigas, research in the Caribbean. Mar.
Fish. Rev. 59: 14–22.
Wilkinson, C., O. Linden, H. Cesar, G. Hodgson, J. Rubens, and A. E. Strong. 1999. Ecological
and socioeconomic impacts of 1998 coral mortality in the Indian Ocean: an ENSO impact and
a warning of future change? Ambio 28: 188–196.
Zieman, J. C. 1982. The ecology of the seagrasses of south Florida: a community profile. U.S.
Fish and Wildlife Services, Office of Biological Services, Washington, D.C. FWS/OBS-82/25,
158 p.
Address: Perry Institute for Marine Science, 100 North U.S. Highway 1, Suite 202, Jupiter,
Florida 33477
33477.
PROCESSES INFLUENCING RECRUITMENT INFERRED
FROM DISTRIBUTIONS OF CORAL REEF FISHES
Aaron J. Adams and John P. Ebersole
ABSTRACT
To determine the relative impact of oceanographic vs. benthic processes on recruitment of coral reef fishes to back reef lagoon nursery habitats, we examined the distributions of small fishes (recruits < 3 cm), medium (juvenile 3–5 cm), and large (juvenile/
adults > 5 cm) of five fish taxa at six, 16-ha lagoons of St. Croix, USVI. Since all habitats of a location are influenced by the same oceanographic events, it was hypothesized
that rankings of recruit densities on the same habitat types across different lagoons
should be similar if oceanographic processes have a dominant influence on recruitment.
Concordance analysis of recruit densities produced no evidence of consistent amongsite differences. It was hypothesized that consistent rankings of habitats within lagoons
(e.g., density of post-settlers on rubble habitat ranked higher than seagrass habitat
within all lagoons), indicated post-settlement benthic processes were more influential.
Differential use of habitat by recruits was consistent among lagoons and over 2 yrs of
study. Patterns of habitat use by juveniles were different from the patterns of recruits.
Acanthurus spp. and Haemulon spp. moved from the lagoon (nursery habitats) to the
reef (adult habitat), and densities of large fishes of these species on back reefs were
strongly related to the availability of nursery habitat in adjacent lagoons. These ontogenetic changes in habitat use indicate continuing influence of benthic processes.
This publication is part in a series of papers resulting from a scientific workshop held at
the Caribbean Marine Research Center (December 2001) to evaluate the importance of
back reef systems for supporting biodiversity and productivity of marine ecosystems.
The life history of most coral reef fishes is a two-phase cycle that decouples local reproduction from recruitment of new individuals into the local population. Juveniles and
adults are demersal while larvae are planktonic. Furthermore, larvae of many prominent
fish taxa on coral reefs settle into habitats that are distinct from those of adults (Robertson et al., 1979; Shulman and Ogden, 1987), creating another stage within the demersal
phase of life history. Thus, for many coral reef fishes, the life history includes three
stages; larval (planktonic), juvenile (demersal), and adult (demersal) – further complicating population dynamics.
Oceanographic processes may determine settlement patterns of coral reef fishes by
transporting and influencing the survival of larvae (Choat et al., 1988). For example,
for some species in Barbados, the occurrence of late stage larvae in traps corresponded
with the first appearance of juveniles on reefs, suggesting that larval supply was a good
indicator of settlement (Sponaugle and Cowen, 1996). Moreover, larvae of some species
[e.g., Stegastes partitus (Poey, 1868) and Acanthurus bahianus (Castelnau, 1855)] were
consistently associated spatially and temporally, suggesting these species were influenced similarly by oceanographic processes such as prevailing currents, tidal currents,
wind-induced water flow, and large-scale externally forced events.
However, mixed results from recent studies of recruitment at multiple spatial scales
demonstrate the complexity of processes influencing larval supply, settlement, and postsettlement densities. Significant spatial trends in settlement for Thalassoma bifasciatum
(Bloch, 1791) were attributed to oceanographic influences by Caselle and Warner (1996).
153
154
However, these spatial trends were much weaker for all species combined, and others
have reported consistent spatial trends for some species, but not others (Fowler et al.,
1992; Green, 1998; Tolimieri et al., 1998; Vigliola et al., 1998). Furthermore, Caselle and
Warner (1996) concluded that although patterns of settler densities were determined by
oceanographic processes, spatial variation in recruitment resulted from a combination of
larval supply, settlement habitat patterns, and post-settlement movement and mortality.
In general, recruitment (entry into the benthic population after the high mortality of the
early post-settlement period; Booth and Brosnan, 1995) is probably affected by varying combinations of larval supply and benthic processes (Choat et al., 1988; Fowler et
al., 1992; Booth and Beretta, 1994; Gibson, 1994; Booth and Brosnan, 1995; Tolimieri,
1998a; Tolimieri et al., 1998), with the relative importance of oceanographic vs. benthic
processes varying among species and locations.
Our knowledge of the juvenile stage of the life cycle and the transition from juvenile
nursery habitat to adult habitat is even less developed than our knowledge of settlement,
and our grasp of population dynamics of coral reef fishes is hampered by this ignorance
(Jones, 1991). For example, during a post-settlement transition period, larvae may actively search out particular habitats (Sancho et al., 1997), or enter and exit habitats multiple times before permanently entering the benthic population (Kaufman et al., 1992),
but the reasons for these movements are unclear. That settling fishes search for suitable
habitat indicates that finding a good nursery habitat is worth the energy and predation
risk associated with exploratory movement. However, even though we know that recruitment patterns are modified after settlement (Eggleston, 1995; Caselle and Warner,
1996; Tolimieri, 1998b), and we know the habitat requirements of recruits and adults of
many coral reef fishes (Sale et al., 1984; Robertson, 1988; Tolimieri, 1998a,b), we do
not understand which processes are most important in determining spatial patterns of
abundance. Furthermore, much of the research into the importance of recruitment to the
abundance of adult fishes has been conducted on species that spend the entire demersal
portion of the life cycle in a single habitat (e.g., Green, 1996), or even a single site (e.g.,
Doherty, 1983), and these findings may not be applicable to coral reef fishes in general.
In this study, we focus on species that are not site-attached, and undergo ontogenetic
habitat shifts, such that they are more likely to be influenced by benthic processes (Ault
and Johnson, 1998). Determining the relative importance of oceanographic and benthic
processes to coral reef fishes is not only important for understanding the ecology of coral
reefs, but also has management implications. The design of marine reserves, for example, must consider not only the source and destination of larvae, but the extent, quality,
and dispersion of habitats. This is especially true for species that undergo ontogenetic
shifts that incorporate multiple habitats during their benthic life stages. By examining
the utilization patterns of back reef and lagoon habitats by the early life stages of coral
reef fishes, we can begin to determine the extent to which abundances of these species
on coral reefs are a function of surrounding shallow-water nursery habitats (Parrish,
1989), and incorporate this information into the design of marine reserves.
We used spatial distributions of post-settlement, juvenile, and sub-adult and adult
fishes to test three hypotheses concerning the relative importance of oceanographic and
benthic processes on populations of coral reef fishes that utilize lagoon habitats as nurseries: (H1) Recruitment is almost completely determined by oceanographic processes;
(H2) Recruitment is primarily determined by benthic processes; and (H3) Adult populations on adjacent reefs are influenced by the habitat composition of nearby lagoons.
ADAMS AND EBERSOLE: PROCESSES INFLUENCING FISH RECRUITMENT
155
We reasoned that in an extreme case where oceanographic processes completely determined recruitment by creating differences in recruit densities among lagoons/back
reefs (scale of kilometers), then recruitment to all habitats within a lagoon (scale of
hundreds of meters) should be influenced in the same manner. Thus, if oceanographic
processes are the overwhelmingly dominant influence on recruitment (H1), a study lagoon with high recruitment in one habitat should have high recruitment in all habitats.
By contrast, benthic influences would be based on local factors (scale of meters), such as
predation intensity and availability of food and shelter, so that the pattern of variation in
habitat quality among lagoons would be different for each habitat type. Thus, if benthic
processes are dominant (H2), high recruitment in one habitat of a given lagoon will not
be associated with high recruitment in other habitats of the lagoon. Instead, patterns of
habitat use by post-settlement and juvenile fishes should be similar for all lagoons, irrespective of habitat availability. It then follows that reefs adjacent to lagoons with more
of the most heavily used habitats should have more adults. This leads to the hypothesis
that if quality of lagoon habitats is influential in determining densities of post-settlement
and juvenile fishes for species that use lagoon habitats as nurseries, then lagoons with
the greatest amount of highly used nursery habitat would have the greatest densities of
sub-adults and adults on adjacent back reefs (H3).
Methods
Study Location and Habitat Types.—We used transects to visually census fishes in back
reef and lagoon habitats at six study lagoons spanning 14 km of coast on the eastern end of St.
Croix (Fig. 1). Each 16-ha study lagoon comprised a section of back reef and its associated lagoon. The study lagoons are similar in terms of bank-barrier reef orientation and size, lagoon
area, depth, and habitat types. Seagrass beds, in which coral rubble, patch-reef, algal plain, and
sand bottom habitats are patchily distributed, are the dominant habitats of these lagoons, which
are bounded on their seaward sides by the back reef of a bank-barrier reef system. Wide, continuous, shallow (< 1 m) reef crests prevent work on the reef platform and restrict access to the shallow fore reef in all but the calmest weather. The back reefs are shallow areas composed mostly of
highly inter-mixed calcareous pavement, patch reef (coral heads) and rubble, with smaller patches
of sand, algal plain, or seagrass mixed in. In contrast to the lagoon, potential nursery habitats of
the back reef are contiguous with the rest of the bank-barrier reef. Since considerable movement
of fishes may occur along a continuous reef tract but not among isolated sections of reef (Ault
and Johnson, 1998), isolated sections of bank-barrier reef and associated lagoons were chosen for
study lagoons to reduce movement of fishes that might occur between contiguous sections of reef,
or between lagoonal nurseries and non-adjacent sections of a contiguous reef tract.
We divided the lagoon habitats into five primary types: (1) Patch reef, isolated, high-relief,
calcareous structure with a vertical profile that often, but not always, contains live coral cover; (2)
Rubble, low-relief, calcareous structure composed primarily of conch shells or dead/dying coral
fragments that are not attached to the substrate; (3) Seagrass, consisting of Thalassia testudinum
Banks & Soland. ex Koenig with varying densities of Syringodium filiforme Kuetz. mixed in; (4)
Algal plain, sand bottom dominated by Halimeda spp., Penicillus spp., and Udotea spp., which
may include sparse stands of S. filiforme and T. testudinum; and (5) Sand, sand bottom with no or
very little (< 10% cover) plants or coralline material represented.
Field Methods.—To estimate patterns of use of back reef and lagoon habitats by recruit,
juvenile, and adult fishes, we censused the fishes at each lagoon in two seasons over 2 yrs: at the
beginning and end of the settlement season (June and October, respectively, of 1999 and 2000).
In each season of each year we completed 2 d-long lagoon/back reef fish censuses at each study
lagoon, with 14, 50 × 2 m transects covering 1400 m2 of back reef and 20, 50 × 2 m transects cov-
156
Figure 1. Locations of study lagoons on the east end of St. Croix, U.S. Virgin Islands. TH =
Turner Hole, RB = Rod Bay, TAG = Tague Bay, YC = Yellowcliff Bay, POW = Pow Point, SOL
= Solitude Bay.
ering 2000 m2 of lagoon each census day. The basic unit of analysis was the census data collected
from a single study lagoon in a single day.
To ensure consistency of counts, the same two individuals censused all transects. In each 50 ×
2 m transect, all fishes (except cryptic species such as gobiids, apogonids, bleniids) were counted,
and recorded in three size categories (< 3, 3–5, > 5 cm). Size categories were used to reduce
potential differences in the estimation of fish size by the two observers over the course of the
study, and to divide fishes into three general sizes that best reflected the following categories at
the community level: small (recruits < 3 cm); medium (juveniles 3–5 cm); and large (late juvenile
through adults > 5 cm). The clipboards that held the census data sheets were marked with size
increments for in situ size reference.
Within lagoons, the location and direction of transects were randomized to ensure adequate
and representative coverage of lagoon habitats. Fishes were recorded in association with a given
habitat, and the length of transect tape crossing each habitat type was noted. Lengths of tape overlaying different habitats were recorded to provide estimates of percent cover by habitat type, and
to allow calculation of fish densities. Back reef transects were parallel to the longitudinal axis of
the back reef, with no differentiation of habitat types due to the fine-grained, highly inter-mixed
nature of the back reef substrata.
Focal Fishes.—Our analyses focused on two abundant species, Halichoeres bivittatus (Bloch,
1791) and Stegastes leucostictus (Müller and Troschel in Schomburgk, 1848), and two abundant
genera, Acanthurus spp. and Haemulon spp. These taxa were chosen due to: (1) economic, and
thus conservation and management, importance ((Acanthurus spp. and Haemulon spp. are important in the commercial fishery; A.J.A, pers. obs.; W. Tobias, U.S. Virgin Islands Division of Fish
and Wildlife, pers. comm.); (2) trophic importance ((Acanthurus spp. are important grazers on
reefs); (3) habitat use (H. bivittatus is a habitat generalist, and was recorded in all lagoon habitats and back reef; S. leucostictus is a lagoon-attached species with relatively restrictive habitat
requirements); and (4) ontogenetic characteristics ((Acanthurus spp. and Haemulon spp. undergo
ontogenetic habitat migrations, while S. leucostictus settles into adult habitat). Since newly settled
individuals of Acanthurus ((Acanthurus chirurgus (Bloch, 1787) and A. bahianus) and Haemulon
(Haemulon flavolineatum (Desmarest, 1823) >90% of Haemulon spp.), Haemulon chrysargy-
157
reum Guenther, 1859, Haemulon plumieri (Lacepède, 1801), Haemulon sciurus (Shaw, 1803),
Haemulon carbonarium Poey, 1860, and Haemulon aurolineatum Cuvier, 1830) are difficult to
identify at the species level, we pooled these species within each genus to form two nominal species for data analysis.
Caribbean Acanthurus spp. have planktonic larval durations of 44–69 d (Sponaugle and Cowen 1996), and settle at a relatively large size ((A. bahianus mean size at settlement in Panama =
26.9 mm (Robertson, 1992), on St. Croix = 32.5 mm; Risk, 1998). Competent acanthurid larvae
can swim long distances (Sancho et al., 1997; Stobutzki and Bellwood, 1997) while delaying
metamorphosis (McCormick 1999), so they can explore potential settlement lagoons. Post-settlement A. bahianus and A. chirurgus may use a variety of habitat types (Robertson, 1988; Mahon
and Mahon, 1994; Risk, 1997; Lawson et al., 1999; Adams and Ebersole, 2002), so they can take
advantage of the post-settlement transition period (Kaufman et al., 1992) to find the most suitable
habitat for final settlement. Among reef habitats, juvenile A. bahianus persist longer on the back
reef than on the fore reef and reef crest (Risk, 1997), and are more abundant on the edges of large
patch reefs than patch reef interior (Robertson, 1988). Sub-adult and adult Acanthurus spp. are
primarily associated with reef habitats (Robertson, 1988; Lawson et al., 1999), although adults of
A. bahianus also utilize seagrass habitats adjacent to reefs (Robertson, 1988).
The most abundant Haemulon species in this study, H. flavolineatum, has a planktonic duration of ~14 d (McFarland et al., 1985), is 5–15 mm total length (TL) at settlement, and shows plasticity in settlement and post-settlement habitat use (Shulman and Ogden, 1987). Large Haemulon
spp. utilize back reef and fore reef habitats (Shulman and Ogden, 1987; Adams and Ebersole,
2002), which indicates a clear ontogenetic shift from plant-dominated habitats to calcium carbonate habitats.
Halichoeres bivittatus is the most common labrid in seagrass and shallow reefs of St. Croix.
After a larval stage of ~25 d, H. bivittatus settles at ~10 mm TL around the new moon, with juveniles most commonly found in small groups over sand-rubble substrates (Sponaugle and Cowen,
1997).
After a planktonic stage of ~28 d (Thresher and Brothers, 1989), S. leucostictus settles onto
benthic habitats at ~7–10 mm TL (McGehee, 1995). Settlement is highest in summer through
early fall (Booth and Beretta, 1994; McGehee, 1995). Among reef habitats, adult S. leucostictus
are found mostly in rubble habitat of the back reef (Itzkowitz, 1985; Wellington, 1992), where
there may be regular changing of territories among adults (Itzkowitz et al., 1995).
Analyses
Patterns of Recruit Densities.—To determine whether early post-settlement fishes exhibited differences in habitat utilization, densities of recruits (recruit and juvenile classes combined
for Acanthurus spp.) for each focal species and for all species combined were examined with a
two-way ANOVA, with Habitat (five lagoon habitats and back reef for each lagoon) and Year
(June and October pooled within each year) as factors. Data were log (x + 1)-transformed to meet
requirements of normality and homogeneity of variances. Lagoon was not used as a factor in
this analysis because we use the obvious and well-documented differences in total numbers of
fishes among lagoons (Adams and Ebersole, 2002) as a starting point, and limit our examination
to determining to what extent oceanographic and benthic processes influence these differences.
Habitats were excluded from analysis if the species was absent from that habitat at four or more
lagoons in a given year. Lagoonal sand habitat was excluded from all analyses. Tukey-Kramer
post-hoc comparisons were used to examine significant factor effects.
Densities of recruits of Haemulon spp., H. bivittatus, S. leucostictus, and all species (focal
species + other species) combined, and densities of fishes in recruit and juvenile size classes
combined for Acanthurus spp., were examined for the relative importance of oceanographic and
benthic processes on recruitment. Haemulon spp., H. bivittatus, and S. leucostictus all settle into
benthic habitat at sizes < 3 cm (see above); but because Acanthurus spp. settle at 25–32 mm TL,
we pooled recruit and juvenile size classes for this genus.
158
Within each year, fishes were censused at the beginning and end of the settlement season, June
and October, respectively, and the two counts within each year were pooled for analyses of habitat
use patterns of each settlement season. Growth rates of post-settlement fishes for all focal species
are such that recruits (< 3 cm) censused in June had grown into the juvenile (3–5 cm) size class by
the October census (~120 d between census periods), ensuring that fishes were not recorded twice
(June and October) in the same size class. Using published growth rates for the focal species,
approximate mean sizes of June settlers after 100 d are: Haemulon spp., 42 mm (Shulman and
Ogden, 1987); H. bivittatus, 38 mm (Victor 1991; Sponaugle and Cowen, 1997); S. leucostictus,
40 mm (McGehee, 1995); and Acanthurus spp., > 40 mm. Haemulon spp. were excluded from
calculations of densities of recruits for all species (focal species + other species) combined since
Haemulon spp. comprised 54.2 and 93.9% of recruits in lagoons in 1999 and 2000, respectively,
and 27% (1999) and 60% (2000) of all recruits in all habitats combined.
Densities of recruits (recruit and juvenile size classes for Acanthurus spp.) in each habitat
(five lagoon habitats and back reef) for the four focal species and for all species combined were
examined to determine the relative influence of oceanographic processes on recruitment (H1).
For this analysis, densities were used to rank lagoons within each habitat type, and analyzed
with Kendall’s coefficient of concordance (Sokal and Rohlf, 1995) to determine to what extent
rankings of lagoons within each habitat were consistent. Probabilities were Bonferroni-adjusted
to limit the experiment-wise error rate to 0.05 (Sokal and Rholf, 1995) for testing this hypothesis.
The expectation was that if oceanographic processes were more influential, significant and high
concordance would result. In other words, when lagoons are ranked according to recruit densities
in one habitat, then lagoon rankings will be similar for all other habitats. If benthic processes are
more influential the coefficient of concordance will be low and non-significant because lagoons
are not ranked similarly for the different habitats.
Densities of recruits (recruit juvenile size classes for Acanthurus spp.) in each habitat (five
lagoon habitats and back reef) for the four focal species and for all species combined were examined to determine the relative influence of oceanographic processes on recruitment (H2). For
this analysis, densities were used to rank habitats within each lagoon, and analyzed with Kendall’s coefficient of concordance (Sokal and Rohlf, 1995) to determine to what extent rankings of
habitats within lagoons were consistent (i.e., were habitat use patterns the same at all lagoons?).
Probabilities were Bonferroni-adjusted to limit the experiment-wise error rate to 0.05 (Sokal and
Rholf, 1995) for testing this hypothesis. The expectation was that if benthic processes were more
influential in determining small fish densities then high and significant concordance should result
(i.e., habitats within each lagoon are ranked similarly).
Patterns of Juvenile Densities.—To determine patterns of habitat use for juveniles, log
(x+1)-transformed densities for each focal species and for all species combined were examined
with a two-way ANOVA, with Habitat (five lagoon habitats and back reef for each lagoon) and
Year (June and October pooled within each year) as factors. Lagoon was not used as a factor in
this analysis because we used the obvious and well-documented differences in total numbers of
fishes among lagoons (Adams and Ebersole, 2002) as a starting point. Habitats were excluded
from analysis if the species was absent from that habitat at four or more lagoons in a given year.
Lagoonal sand habitat was excluded from all analyses. Tukey-Kramer post-hoc comparisons were
used to examine significant factor effects. Acanthurus spp. was excluded from these analyses
since the recruit and juvenile size classes were combined for the previous examination of recruits.
As with recruits, Kendall’s coefficient of concordance was used to examine the relative importance of oceanographic influences on densities of the juveniles. Densities of juveniles in each
habitat (five lagoon habitats and back reef) for the four focal species and for all species combined
were examined to determine the relative influence of oceanographic processes on the distribution
of juveniles (H1). Acanthurus spp. was not included in this analysis since the recruit and juvenile
size classes were combined for the analysis of post-settlement densities. For this analysis, densities were used to rank lagoons within each habitat type, and analyzed with Kendall’s coefficient of
concordance (Sokal and Rohlf, 1995) to determine to what extent rankings of lagoons within each
159
habitat were similar. Probabilities were Bonferroni-adjusted to limit the experiment-wise error
rate to 0.05 (Sokal and Rholf, 1995) for testing this hypothesis. The expectation was if oceanographic processes were extremely influential, then patterns resulting from settlement would remain evident in the juvenile size class, and significant and high concordance should result. Thus,
if benthic processes were more influential, there would be no measurable signal of settlement
patterns, and the coefficient of concordance would be low and non-significant.
Densities of juveniles on each habitat (five lagoon habitats and back reef) for the four focal species and for all species combined were examined to determine the relative influence of oceanographic processes on recruitment (H2). As before, Acanthurus spp. were not included in this
analysis since recruit and juvenile size classes were combined for the analysis of post-settlement
densities. For this analysis, densities were used to rank habitats within each lagoon, and analyzed
with Kendall’s coefficient of concordance (Sokal and Rohlf, 1995) to determine to what extent
rankings of habitats within lagoons were consistent. Probabilities were Bonferroni-adjusted to
limit the experiment-wise error rate to 0.05 (Sokal and Rholf, 1995) for testing this hypothesis.
The expectation was if benthic processes were most influential then high and significant concordance would result, because juvenile densities would reflect that patterns of habitat use were
similar in all lagoons.
Nursery Habitat Quality.—
Quality.—We examined densities of recruits and juveniles in lagoons and
adults on back reefs to test the hypothesis that adult populations on adjacent reefs are influenced
by the habitat composition of nearby lagoons (H3). Least squares regression (Sokal and Rohlf,
1995) was used to determine to what extent densities of juvenile/adult (> 5 cm) fishes on backreefs are dependent upon lagoon quality (as determined by availability and use of lagoon nursery
habitats) for the two focal species ((Acanthurus spp. and Haemulon spp.) that utilize lagoon habitats as nurseries. We derived a species-specific lagoon quality index (LQI) for each lagoon for
recruits, juveniles of Haemulon spp., and recruit and juveniles combined for Acanthurus spp. by
the following formula:
5
LQII ij = Aˆ aix ⋅ Pjx
LQ
x =1
where x = a given lagoon habitat (patch reef, rubble algal plain, seagrass, sand); aix = mean density
of species i on habitat xj (computed from values pooled from all six lagoons, to eliminate potential
lagoon bias, and over both years, to eliminate annual variation in recruitment and habitat use); Pjx
= relative cover (0–1) of habitat x in lagoon j (j
( = Turner Hole, Rod Bay, Tague Bay, Yellowcliff
Bay, Pow Point, Solitude Bay).
The LQI provides a measure of the value of a lagoon as a nursery for populations of large
fishes on adjacent reefs as a function of the standardized, habitat-specific use by each species and
the amount (availability) of each habitat in each lagoon. Since the availability of habitat in each
lagoon is a factor, LQI is a useful tool for quantifying the extent to which benthic differences
among lagoons are important to young fishes. Moreover, LQI characterizes the overall value of
the lagoon nursery over time as a source of ontogenetic migrants since recruitment data were
averaged over 2 yrs rather than relying on a single cohort.
The dependent variable —density of juvenile/adult (> 5 cm) fishes— was calculated separately
for each lagoon. Data were pooled across years because the hypothesis pertains to the quality
of lagoons as nursery habitats per se, not as a function of recruitment in a particular year. Since
the large size class includes fishes that were recruited over a number of different years, pooling
density data from 1999 and 2000 merely enhances the integration of several years of recruitment
that is already present with each population. Significant regression coefficients were expected if
lagoon nursery habitat availability was a highly influential factor in determining densities of large
(> 5 cm) fishes on adjacent back-reefs.
160
Results
A total of 103 species was recorded over the course of the study: 99 species on back
reef, 89 species in lagoon habitats, and 71 species in both back reef and lagoon. Total recruitment (June and October pooled for each year) of Haemulon spp. was lower in 1999
(9884 Haemulon spp. recruits < 3 cm) than in 2000 (20,943 Haemulon spp.). In contrast,
total recruitment for all species combined (not including Haemulon spp.) was higher in
1999 (27,467 post-settlement fishes) than in 2000 (13,949 small fishes). In total, 57,238
juvenile fishes were recorded in 1999 and 34,355 medium fishes in 2000. Densities of
recruits and juveniles differed among lagoons and years (Table 1). The percent cover of
lagoon habitats varied among lagoons, with algal plain the most abundant habitat overall
at three of six study lagoons, and rubble the least common habitat (Fig. 2).
Patterns of Recruit Densities.—Concordance results indicate that densities of
recruits did not reflect oceanographic influences since no Kendall coefficients of concordance were significant (probabilities were Bonferroni adjusted for an experimentwise
error rate of 0.05, so each individual test level of significance, P = 0.01: for all coefficients, P > 0.01).
When densities of recruits were examined, each of the four focal taxa and all species
combined showed significant habitat use patterns that were consistent across years (Table 2). Recruit and juvenile (< 3 cm and 3–5 cm size classes combined) Acanthurus spp.
use lagoonal patch reef and rubble habitats as nurseries more than back reef and other
lagoon habitats – seagrass, algae, and sand – which cover considerably more area (Fig.
3A). Although we did not record them as a separate size class, we observed that newly
settled (< 1 cm) Haemulon spp. were most frequently seen in seagrass and algal plain,
and only infrequently in back reef, patch reef, and rubble. Surviving Haemulon spp. settlers quickly moved from seagrass and algal plain to habitats that provided more shelter
(i.e., rubble, patch reef, and back reef). Haemulon spp. recruits (< 3 cm) were recorded
Figure 2. Percent cover of habitats in lagoons for each study lagoon. Values were pooled from all
lagoon transects for each lagoon.
0.105 (0.051)
0.121 (0.056)
0.13 (0.059)
0.128 (0.061)
< 3 cm
3–5 cm
< 3 cm
3–5 cm
< 3 cm
3–5 cm
Haemulon spp.
0.448 (0.24)
0.535 (0.318)
< 3 cm
3–5 cm
0.918 0.594)
1.532 (0.949)
0.479 (0.478)
0.364 (0.362)
1.154 (1.152)
0.005 (0.004)
0.351 (0.345)
2.922 (2.265)
0.838 (0.525)
1999
0.398 (0.249)
0.267 (0.135)
0.067 (0.032)
0.044 (0.029)
0.112 (0.052)
0.098 (0.054)
0.0614 (0.055)
0.143 (0.098)
2000
0.261 (0.153)
0.116 (0.078)
0.004 (0.004)
0.018 (0.016)
0.132 (0.125)
0.21 (0.207)
0.047 (0.045)
0.108 (0.079)
0.087 (0.056)
Rod Bay
0.106 (0.075)
* Densities for recruits (<3cm) of All Species Combined do not include Haemulon spp.
All Species Combined*
S. leucostictus
H. bivittatus
0.043 (0.03)
< 5 cm
Acanthurus spp.
0.273 (0.165)
0.162 (0.083)
Size
Class
Focal Taxa
Turner Hole
1999
2000
0.55 (0.311)
0.783 (0.279)
0.338 (0.164)
0.459 (0.224)
0.164 (0.092)
0.116 (0.085)
0.027 (0.027)
0.517 (0.303)
2000
0.319 (0.161)
0.172 (0.088)
0.935 (0.867)
1.566 (1.55)
0.012 (0.008)
0.276 (0.264)
0.046 (0.029)
0.76 (0.55)
0.002 (0.001)
Tague Bay
0.001 (0.001)
1999
0.486 (0.311)
0.235 (0.142)
0.131 (0.092)
0.108 (0.093)
0.197 (0.157)
0.038 (0.022)
0.013 (0.002)
0.131 (0.077)
2000
0.202 (0.112)
0.152 (0.087)
0.002 (0.002)
0.026 (0.02)
0.029 (0.02)
0.11 (0.076)
0.075 (0.065)
0.393 (0.168)
0.038 (0.029)
Pow Point
0.057 (0.051)
1999
0.626 (0.396)
0.45 (0.254)
0.117 (0.046)
0.163 (0.077)
0.174 (0.094)
0.2 (0.116)
0.222 (0.197)
0.133 (0.11)
0.082 (0.049)
2.409 (2.354)
0.479 (0.478)
0.055 (0.051)
0.003 (0.003)
0.008 (0.007)
0.005 (0.005)
0.004 (0.004)
0.692 (0.626)
1.278 (1.273)
Yellowcliff Bay
1999
2000
0.855 (0.499)
0.3 (0.165)
0.082 (0.03)
0.044 (0.026)
0.182 (0.146)
0.048 (0.023)
0.325 (0.167)
1.613 (1.526)
2000
0.388 (0.302)
0.056 (0.036)
0.113 (0.11)
0.011 (0.008)
0.01 (0.006)
0.004 (0.003)
0
0.333 (0.153)
0.192 (0.189)
Solitude Bay
0.232 (0.184)
1999
Table 1. Mean (standard error in parentheses) densities (number of fish m−2) of recruits (< 3 cm) and post-settlement/juveniles (3–5 cm) of focal taxa by lagoon and year. For Acanthurus spp. the recruit and post-settlement/juvenile size classes were
combined.
161
162
Table 2. Results of Two-way ANOVA on densities (log (x+1) transformed) of early postsettlement fishes for the four focal species ((Acanthurus spp., Haemulon spp., Halichoeres
bivittatus, Stegastes leucostictus) and for all species combined. Factors were Habitat (backreef, patch reef, rubble, seagrass, algal plain, sand) and Year (1999, 2000). For each Habitat,
densities were calculated for each lagoon within each Year (June and October census data were
pooled). Densities of small (< 3 cm) fishes were examined for Haemulon spp., H. bivittatus, S.
leucostictus, and all species combined; small (< 3 cm) and medium (3–5 cm) size classes were
combined for Acanthurus spp.
Species
Acanthurus spp.
Haemulon spp.
H. bivittatus
S. leucostictus
All species
Source
Habitata
Year
Habitat × Year
Error
Habitata
Year
Habitat × Year
Error
Habitata
Year
Habitat × Year
Error
Habitata
Year
Habitat × Year
Error
Habitata
Year
Habitat × Year
Error
df
4
1
4
50
4
1
4
50
4
1
4
50
2
1
2
30
4
1
4
50
MS
0.5768
0.1147
0.1317
0.0882
0.5511
0.2938
0.0345
0.1737
0.2704
0.0045
0.0583
0.0228
0.659
0.0055
0.1008
0.1272
0.9573
0.1338
0.0763
0.0480
F
6.5362***
1.2999NS
1.4919NS
3.1722*
1.6913NS
0.1988NS
11.8367***
0.1973NS
2.5519NS
5.1804*
0.0430NS
0.7923NS
19.9523***
2.7883NS
1.5901NS
a = Habitats were excluded from analysis if four or more lagoons had densities of zero for that habitat in
a given year. Sand was excluded from all analyses, and seagrass and algal plain were also excluded for S.
leucostictus. * = P < 0.05; ** = P < 0.01; *** = P < 0.001; NS = P > 0.05.
in all habitats, with highest densities on lagoon rubble and patch reef and lowest densities in back reef (Fig. 3B). Halichoeres bivittatus recruits were present in significantly
higher densities in lagoon rubble habitat, followed by patch reef, and in lowest densities
in seagrass and algal plain (Fig. 3C). Stegastes leucostictus recruits were present only
in rubble, patch reef, and back reef habitats, and were in greatest density in rubble (Fig.
3D). Rubble, patch reef, and back reef were the most heavily used habitats for recruits of
all species combined (Fig. 3E). Although there were some annual differences in densities of recruits for some species in some habitats (Fig. 3), there were no significant Year
or Year × Habitat effects (Table 2).
Patterns of habitat use by recruits were consistent across lagoons in both years in most
cases (Table 3), indicating that, overall, habitat selection by settlement and post-settlement fishes and/or benthic processes affecting post-settlement fishes were consistent at
all lagoons. In 2000, habitat use patterns were not significantly consistent for Haemulon
spp. (Table 3).
Patterns of Juvenile Densities.—Concordance analysis used to examine whether
there was a measurable signal of oceanographic processes on densities of juvenile (3–5
cm) fishes in lagoons produced non-significant results for all taxa examined (probabilities were Bonferroni adjusted for an experimentwise error rate of 0.05, so each individual test level of significance, P = 0.0125: for all coefficients, P > 0.0125).
163
Figure 3. Habitat use patterns by early post-settlement fishes. June and October were pooled
within each year for each lagoon (N = 6). Values represent means (+ 1 SE). Values are densities
of small (< 3 cm) fishes for all variables except Acanthurus spp., for which small (< 3 cm) and
medium (3–5 cm) fishes were combined due to large and variable size at settlement. (A) Acanthurus spp., (B) Haemulon spp., (C) Halichoeres bivittatus, (D) Stegastes leucostictus, (E) All
species combined.
Examination of habitat use patterns by juveniles of all taxa also indicated substantial
effects of benthic processes on distributions of juvenile fishes. For example, Haemulon
spp. recruits were present in all lagoon habitats, but juvenile Haemulon were absent
from seagrass, algal plain, and sand, and were in greatest densities in lagoonal patch reef
164
Table 3. Results of concordance analyses of densities of small (< 3 cm) fishes on five lagoon
habitats and back-reef at each of the six study lagoons for the four focal species ((Acanthurus
spp., Haemulon spp., Halichoeres bivittatus, Stegastes leucostictus) and for all species
combined to determine whether patterns of habitat use varied among lagoons.
Kendall’s Coefficient
Species
Acanthurus spp.a
Haemulon spp.
Halichoeres bivittatus
Stegastes leucostictus
All species combinedb
1999
0.622*
0.556*
0.911**
0.583NS
0.889**
2000
0.501*
0.304NS
0.406*
0.132NS
0.866**
a = Small (< 3 cm) and medium (3–5 cm) size classes were combined for Acanthurus spp. since size at
settlement incorporates both size classes.
b = Haemulon spp. were excluded from calculations of densities for all species combined for small fishes
since Haemulon spp. comprised 54.2% of all small fishes in 1999, and 93.9% of all small fishes in 2000.
Probabilities were Bonferroni adjusted for an experimentwise error rate of 0.05, so each individual test
level of significance, P = 0.01. * = P < 0.01, ** = P < 0.001, NS = P > 0.01.
and rubble and on back reef, indicating an ontogenetic niche shift from post-settlement
to juvenile habitat. Among rubble, patch reef, and back reef, there was no significant
difference in densities of juvenile Haemulon spp., probably due to high variability in estimates of density on rubble and patch reef resulting from the aggregative nature of this
species group (Fig. 4A; Table 4). In addition, the significant difference in habitat use by
S. leucostictus recruits was no longer evident in juvenile S. leucostictus (Fig. 4C; Table
4), again indicating post-settlement processes substantially modified distributions of S.
leucostictus between the post-settlement and juvenile life stages.
Figure 4. Habitat use patterns by juvenile fishes. June and October were pooled within each
year for each lagoon (N = 6). Values represent means (+ 1 SE). Values are densities of medium
(3–5 cm) fishes. Acanthurus spp. were not included because small and medium size classes were
combined for examination of early post-settlement fishes (see Fig. 3). (A) Haemulon spp., (B)
Halichoeres bivittatus, (C) Stegastes leucostictus, (D) All species combined.
165
Table 4. Results of Two-way ANOVA on densities (log (x+1) transformed) of medium (3–5 cm)
fishes for three focal species (Haemulon spp., Halichoeres bivittatus, Stegastes leucostictus)
and for all species combined. Acanthurus spp. was not included in these analyses because the
small and medium size classes were combined for this focal species for examination of spatial
patterns of settlement (results in Table 2). Factors were Habitat (back-reef, patch reef, rubble,
seagrass, algal plain, sand) and Year (1999, 2000). For each Habitat, densities were calculated
for each lagoon within each Year (June and October census data were pooled).
Species
Haemulon spp.
H. bivittatus
S. leucostictus
All species
Source
Habitata
Year
Habitat × Year
Error
Habitata
Year
Habitat × Year
Error
Habitata
Year
Habitat × Year
Error
Habitata
Year
Habitat × Year
Error
df
2
1
2
30
4
1
4
50
2
1
2
30
4
1
4
50
MS
0.1350
0.0843
0.0655
0.1562
0.1093
0.1640
0.0480
0.0086
0.1404
0.0035
0.0271
0.0628
2.1407
0.0001
0.0931
0.1072
F
0.8646NS
0.5397NS
0.4194NS
12.7755***
19.1736***
5.6113***
2.2345NS
0.0550NS
0.4309NS
19.9758***
0.0003NS
0.8691NS
a = Habitats were excluded from analysis if four or more lagoons had densities of zero for that habitat in a
given year. Sand was excluded from all analyses for all species; algal plain and seagrass were also excluded
from analyses for Haemulon spp. and S. leucostictus.
* = P < 0.05; ** = P < 0.01; *** = P < 0.001; NS = P > 0.05.
Unlike Haemulon spp. and S. leucostictus, which exhibited no significant differences,
there were significant Habitat, Year, and interactive effects (Fig. 4B; Table 4) for H.
bivittatus, and significant Habitat effects for all species combined (Table 4).
Habitat use patterns for juvenile fishes were consistent among lagoons for three of four
taxa examined in 1999, but only for all species combined in 2000 (Table 5). However,
examination of the data in graphical form (Figs. 4A–D) indicates overall consistent patterns of habitat use by each of these groups but with greater variance in 2000.
Nursery Habitat Quality.—Densities of juvenile/adult (> 5 cm) fishes on back
reefs were significantly and strongly related to lagoon quality for both focal species
that utilize lagoons as nurseries and undergo ontogenetic migrations to reef habitats.
Table 5. Results of concordance analyses on densities of medium (3–5 cm) size fishes in
five lagoon habitats and back-reef at each of the six study lagoons for the four focal species
(
(Acanthurus
spp., Haemulon spp., Halichoeres bivittatus, Stegastes leucostictus) and for
all species combined to determine whether patterns of habitat use varied among lagoons.
Acanthurus spp. was not included in these analyses because the small and medium size classes
were combined for this focal species for examination of spatial patterns of settlement.
Kendall’s Coefficient
Species
Haemulon spp.
All species combined
1999
0.021NS
0.906**
0.861*
0.911**
2000
0.049NS
0.225NS
0.006NS
0.920**
Probabilities were Bonferroni adjusted for an experimentwise error rate of 0.05, so each individual test
level of significance, P = 0.0125. * = P < 0.0125,** = P < 0.001, NS = P > 0.01.
166
Figure 5. Relationship between density of large (> 5 cm) Acanthurus spp. on back reefs and the
Lagoon Quality Index (LQI) for small (< 3 cm) and medium (3–5 cm) Acanthurus spp. combined.
See text for calculation of LQI.
Densities of juvenile/adult Acanthurus spp. on back reefs were significantly related to
the Lagoon Quality Index (LQI) computed from habitat composition of the lagoons and
data on habitat use of recruit and juvenile (< 5 cm) Acanthurus spp. (Fig. 5), although
Pow Point had large leverage (leverage > 0.5), with lagoon quality accounting for nearly
three quarters of the variation in densities of large fishes on back reefs. Similarly, densities of juvenile/adult Haemulon spp. on back reefs were significantly related to lagoon
quality for recruits (< 3 cm) of Haemulon spp. (Fig. 6). In contrast, as might be expected,
densities of large Haemulon spp. on back reefs were not significantly related to the LQI
computed for juvenile (3–5 cm) Haemulon spp. (Fig. 7). By the time Haemulon spp.
reach the juvenile size class they have already undergone, or are in the process of under-
Figure 6. Relationship between density of large (> 5 cm) Haemulon spp. on back reefs and the
Lagoon Quality Index (LQI) for small (< 3 cm) Haemulon spp. See text for calculation of LQI.
167
Figure 7. Relationship between density of large (> 5 cm) Haemulon spp. on back reefs and the Lagoon Quality Index (LQI) for medium (3–5 cm) Haemulon spp. See text for calculation of LQI.
going, ontogenetic shifts from nursery to sub-adult and adult habitats, so most juvenile
fish are not included in the computation of the LQI. There was no significant relationship
between densities of large fishes on back reefs and lagoon quality for either small (R2 =
0.00, F = 0.0018, P = 0.9697, df = 1,4) or medium (R2 = 0.00, F = 0.0079, P = 0.9333, df
= 1,4) fishes of all species combined.
Discussion
Relative Effect of Oceanographic Processes.—At the scale addressed in this
study, the relative importance of benthic processes effecting recruitment were more influential than oceanographic processes in determining densities of recruit and juvenile
fishes in lagoon and back reef habitats of St. Croix. Oceanographic processes transport and influence survival of larvae, which determines larval distribution (Choat et
al., 1988). In Barbados, several fish species shared the same temporal patterns, with
larvae appearing in traps when settlers were appearing on reefs, indicating that the same
transient factors acting on larval supply produced pulses of settlement in several species
(Sponaugle and Cowen, 1996). Variability in densities of settlers among years (Fowler
et al., 1992; Caselle and Warner, 1996; Vigliola et al., 1998) and variability in timing
of settlement among lagoons (Williams et al., 1994; Caselle and Warner, 1996; Green,
1998) has been considered evidence for the importance of variation in oceanographic
processes to larval distribution and settlement.
Consistency of spatial patterns of settlement over time has also been used to infer the
importance of oceanographic processes to recruitment: similar rankings in densities of
settlers in multiple years, even though the absolute densities may vary among years, may
indicate that oceanographic currents are distributing larvae of these species consistently
at large spatial scales (Fowler et al., 1992; Caselle and Warner, 1996; Tolimieri et al.,
1998; Vigliola et al., 1998). That spatial variation is generally greater than temporal
variation may imply that oceanographic processes provide locations with consistently
168
different proportions of larvae (Choat et al., 1988; Planes et al., 1993; Caselle and Warner, 1996).
Oceanographic processes may be most influential at large spatial scales (Fowler et al.,
1992), whereas benthic processes, such as habitat selection at settlement, may be most
important at small scales (Fowler et al., 1992; Wellington, 1992; Green, 1998). At the
spatial scale of this study (kilometers), any oceanographic influences on larval supply
were so substantially modified by benthic processes that no patterns in densities of recruits at the among-lagoon scale were evident.
Differences between our results and the findings of Caselle and Warner (1996), who
attributed the distinct West-East gradient they found in recruitment densities to oceanographic processes, may be partly explained by differences in geographic scales. The
West-East gradient in recruitment on the north shore of St. Croix reported by Caselle and
Warner (1996) was due largely to differences between their western-most (Butler Bay)
and eastern-most (Forereef) study locations, while differences between their two eastern
study locations were not significant (fig. 5 in Caselle and Warner, 1996). Similarly, the
East-West gradient on the south shore was driven largely by differences between their
extreme east and west study lagoons. All of our study lagoons were located between
their two most eastern lagoons on each shoreline, indicating oceanographic processes
may be operating at a larger scale (4 km of coast on the north shore and 2 km of coast
on the south shore) than was addressed in this study. Sponaugle and Cowen (1997) also
reported mostly non-significant differences among study locations for recruitment of
labrids to Barbados on a spatial scale similar to this study. Moreover, the back reef and
lagoonal habitats we censused may receive lower levels of recruitment than the outer
reef slope censused by Caselle and Warner (e.g., Ault and Johnson, 1998), leading to
different perceptions of larval supply. Finally, all of our study lagoons were part of the
bank-barrier reef system on the eastern portion of St. Croix, whereas Caselle and Warner
censused fishes on two different reef environments —the bank-barrier reef system on the
east end, and fringing reefs adjacent to steep reef walls on the west end of St. Croix.
Relative Effect of Benthic Influences.—Choat et al. (1988) suggested that
habitat-associated variables filter the larval supply, slightly modifying the settlement
patterns determined by oceanographic processes. Our results indicate that this filter effect is extremely selective and so severe that no among-lagoon spatial pattern that might
reflect differences in larval supply is evident. Rankings of lagoons by density of recruits
—which would have been similar for all habitats if oceanographic processes were overwhelmingly important— proved to be discordant, while habitat use patterns of juveniles
were consistent among lagoons and across years.
We agree with Robertson (1988), that seagrass and algal plain provide large target
areas for settlement of A. bahianus, which soon move to suitable recruit and juvenile
habitats within seagrass and algal plain such as rubble and patch reef, or to back reef.
We posit that this strategy may also be followed by A. chirurgus, Haemulon spp., and
other species that utilize lagoon habitats (Parrish, 1989). For example, Haemulon spp.
settle to algal plain and sparse seagrass, and then use rubble, patch reef, and back reef as
they grow (Shulman and Ogden 1987; the present study). Stegastes leucostictus heavily
utilize lagoon patch reef and rubble (this study), but show low site fidelity as juveniles
(McGehee, 1995). We have often noted that a conch shell or similar small structure
placed in a lagoon may quickly attract juvenile S. leucostictus that are well past settlement size (i.e., have recruited to the benthic population).
169
The settlement and movement strategy proposed by Robertson (1988) and Parrish
(1989) would allow post-settlement fishes to respond to benthic processes such as priority effects, competition, and predation. While we did not directly measure these factors,
patterns we discerned from our data are suitably explained by the findings of other authors. Priority effects act during the immediate post-settlement period (Leis and CarsonEwart, 1999): settlement and post-settlement persistence of Acanthurus spp. are reduced
by the presence of S. leucostictus (Shulman et al., 1983; Risk, 1998), and settlement
of Haemulon spp. is low where juveniles of predator species (e.g., lutjanids) have already settled (Shulman et al., 1983; Tupper and Juanes, 1999). Intraspecific competition
for limited shelter space for juveniles of a temperate labrid varies with recruit density
(Tupper and Boutilier, 1995), and this variable importance of competition with the level
of recruitment appears to be a general attribute of post-settlement dynamics (Booth
and Brosnan, 1995). Moreover, Beets (1997) found that the size of shelters influences
densities of post-settlement fishes, and Risk (1998) found that A. bahianus juveniles are
influenced by both intra- and inter-specific aggression, implying a limitation in suitable
shelter. Furthermore, Shulman and Ogden (1987) calculated that early post-settlement
predation was so high for Haemulon species on St. Croix that an increase in survivorship would have a much greater effect than an increase in larval supply, and predation
of post-settlement fishes is higher on reefs with shelter holes large enough for predators
(Hixon and Beets, 1993; Beets, 1997). We found that recruits and juveniles of many
coral reef fishes inhabit rubble, perhaps because predation is low in this habitat that lacks
the large holes that might harbor piscivores, but has small crevices for shelter of small
fishes (Nemeth, 1998).
Benthic processes may influence distributions of recruits through direct effects (e.g.,
loss of settlers to predators), or through selection of habitat by settling post-larvae or mobile recent settlers so as to avoid these direct effects. Direct and indirect effects of benthic processes on recruitment are difficult to discern when they act during the settlement
and early post-settlement phases of the life cycle, and still more difficult when larval
supply is low. In year 2000 of this study, when recruit densities were low, Haemulon spp.
and S. leucostictus did not show consistent habitat use patterns at all lagoons because
no individuals of these species were recorded in some habitats at some lagoons. In 1999,
when recruitment densities were higher, habitat utilization by post-settlement fishes was
consistent across all lagoons. Benthic processes may produce significant differences in
density among habitats in years of high settlement, but differences among habitats may
be undetectable during years of low settlement because densities are low in all habitats
(thus precluding significant among habitat differences). So although oceanographic processes influence overall larval supply, fishes may continue to exhibit habitat selection.
Demographic and Conservation Implications.—Both ontogenetic niche shifters
examined in this study ((Acanthurus spp. and Haemulon spp.) had strong relationships
between lagoon quality and densities of large fishes on back reefs. The LQI standardized
lagoon nursery habitat use to remove any potential lagoon bias by averaging use over
all lagoons. The LQI also averaged 2 yrs of recruitment data to address potential bias of
any single cohort. Given this conservative approach, the demonstration of a significant
effect on densities of large fishes (also pooled across years and composed of a multitude
of cohorts) on adjacent back reefs underscores the importance of lagoon nursery habitats
to these species. Both small and medium Acanthurus spp. must be included in the “nursery” class because they are so large at settlement. Nursery class Acanthurus spp. utilize
lagoonal patch reef and rubble, and move from lagoon nursery habitats to nearby reefs as
170
they grow into the juvenile/adult (> 5 cm) size class. The large leverage imposed by the
Pow Point study site does not negate the importance of the findings of this analysis, but
does suggest additional sites should be censused to further test this hypothesis.
Haemulon spp. undergo ontogenetic habitat shifts while still in the recruit (< 3 cm)
and juvenile (3–5 cm) size classes. These early life history habitat changes allow us to
examine two pre-adult size classes to determine for what life stages lagoon habitats are
most important as nurseries. The general premise that densities of juvenile/adult (> 5
cm) Haemulon spp. on the reef depend on the availability of nursery habitats in adjacent
lagoons yields different expectations for the two size classes, and our findings agree
with these expectations. Haemulon spp. settle in algal plain and seagrass at small sizes
(5–15 mm), and then migrate to nearby rubble, patch reef, and back reef habitats as they
grow (McFarland et al., 1985; Shulman, 1985; the present study). The use of lagoon
habitats by post-settlement and early juvenile stage fishes results in a strong relationship
between Lagoon Quality (LQI) for recruits (< 3 cm) and densities of juvenile/adult (>
5 cm) fishes on back reefs. By the time Haemulon spp. reach the juvenile (3–5 cm) size
class, most have already moved from lagoon habitats to back reef, or are in the process
of doing so, as indicated by the lack of a clear difference in habitat utilization (Table 5).
That the positive and significant relationship between density of juvenile/adult fishes on
back reefs and the LQI for recruits indicates a real demographic dependence of population densities of adults on availability of lagoon nursery habitat is reinforced by the
dissolution of the relationship between density of large fishes on back reefs and habitat
quality when the LQI is computed for juveniles. The findings for Acanthurus spp. and
Haemulon spp. underscore that the LQI, and the relationship between LQI and densities
of large fishes on reefs, is specific to taxon and to size class rather than a general index
between “good lagoons” and “bad lagoons.”
We know that distributions of post-settlement and juvenile fishes are influenced by
benthic processes and that distributions of settlers are the result of some combination of
oceanographic and benthic processes. However, it may be that settlement and early postsettlement patterns have a stronger benthic component than previously thought, and that
this is manifest through differential habitat-associated predation, competition, and migrations that begin as part of settlement (e.g., post-settlement transition and priority effects). That we are finding relationships between benthic nursery habitat utilization and
populations of large fishes on reefs despite inherent variation suggests benthic processes
are key components in the early life histories of coral reef fishes. This should be especially true for species such as Acanthurus spp. and Haemulon spp., that have plasticity
in their settlement and post-settlement requirements, so are able to actively pursue better
quality habitats during and after settlement. However, recent studies of lagoon-attached
species suggest that similar processes influence settlement and post-settlement densities
for these species as well (Schmitt and Holbrook, 1999).
Since oceanographic processes are important in determining distributions of larvae,
understanding oceanographic processes is important in the determination of where to
place marine reserves within a region (e.g., which islands or reef systems are larval
“sources” rather than “sinks”). This study indicates that when considering what section
of island shoreline or reef system should be protected (the scale at which such decisions
are made), benthic features must be paramount. Marine protected areas (MPAs) may be
especially important for artisanal fisheries that depend on fishes that are sedentary as
adults. Conventional theory holds that since sedentary fishes in an MPA are protected
from fishing pressure, they may realize the high reproductive rates achieved by large
171
adults, and so contribute offspring (eggs and larvae are typically planktonic) that can
maintain the population inside and outside the MPA. However, the sedentary adults
targeted by artisanal fishermen on coral reefs utilize nearshore habitats during early
stages of their life histories. Therefore, MPAs can help maintain fisheries only if they
protect all the benthic habitats required for the survival of target species. MPAs must be
designed to include essential nursery habitat for juveniles as well as habitats for adults,
or they will not succeed as population sources. In general, the observation that juveniles
of many reef fishes utilize non-reef habitats before moving to adult reef habitats is further demonstration of the need for a comprehensive approach to management of coral
reef fishes.
Understanding patterns of use of lagoon habitats by fishes that undergo ontogenetic
habitat shifts is critical to understanding the ecology of coral reef fishes. Habitat associated features of lagoons have a powerful influence on recruitment, with effects on
movement and survival through the post-settlement transition and juvenile phase of life
history. The demographic consequences of these benthic processes project forward to
influence densities of adult populations on reefs, and reflect backward to influence settlement (through habitat selection). Additional implications for our understanding of
community structure, and for proper management of resources in the tropical marine
environment (e.g., limitation of fishing in essential nursery areas, protection of essential
nursery habitats), will be revealed by continued research in this area.
Acknowledgments
This research was done in partial fulfillment of the requirements for a Ph.D. by A.J.A. and was
funded by a grant from NOAA/NMFS/Saltonstall-Kennedy program (NA97FD0070) to J.P.E.,
Sigma Xi Grants In Aid of Research, a University of Massachusetts Boston Dissertation Support
Grant, and a NSF-GRT award to A.J.A, and generous support from the Biology Department of
the University of Massachusetts Boston. This paper is funded in part by a grant from the Caribbean Marine Research Center (CMRC Project # CMRC-00-IXNR-03-01A) National Oceanic
and Atmospheric Administration (NOAA) National Undersea Research Program, U.S. Environmental Protection Agency, and Environmental Defense. Views expressed herein are those of the
authors and do not necessarily reflect the views of CMRC, or any supporting agencies. We thank
K. Gloger for excellent work in the field, G. Skomal for lodging on St. Croix, D. Ward of Seaward
Research for field support, NOAA/NOS for aerial photographs of benthic habitat, the St. Croix
Yacht Club, and the staff at Anchor Dive Center for excellent care and service. K. Duchon, M.
Hixon, and two anonymous reviewers provided constructive comments.
Literature Cited
Adams, A. J. and J. P. Ebersole. 2002. Use of lagoon habitats and back-reef by coral reef fishes.
Mar. Ecol. Prog. Ser. 228: 213–226.
Ault, T. R. and C. R. Johnson. 1998. Spatially and temporally predictable fish communities on
coral reefs. Ecol. Monogr. 68: 25–50.
Beets, J. 1997. Effects of a predatory fish on the recruitment and abundance of Caribbean coral reef
fishes. Mar. Ecol. Prog. Ser. 148: 11–21.
Booth, D. and G. A Beretta. 1994. Seasonal recruitment, habitat associations and survival of pomacentrid reef fish in the U.S. Virgin Islands. Coral Reefs 13: 81–89.
_______ and D. M. Brosnan. 1995. The role of recruitment dynamics in rocky shore and coral reef
fish communities. Adv. Ecol. Res. 26: 309–385.
Caselle, J. E. and R. R.Warner. 1996. Variability in recruitment of coral reef fishes: the importance
of habitat at two spatial scales. Ecology 77: 2488–2504.
172
Choat, J. H., A. M. Ayling, and D. R. Schiel. 1988. Temporal and spatial variation in an island fish
fauna. J. Exp. Mar. Biol. Ecol. 121: 91–111.
Doherty, P. J. 1983. Tropical territorial damselfishes: is density limited by aggression or recruitment? Ecology 64: 176–190.
Eggleston, D. B. 1995. Recruitment in Nassau grouper, Epinephelus striatus: post-settlement abundance, microhabitat features, and ontogenetic habitat shifts. Mar. Ecol. Prog. Ser. 124: 9–22.
Fowler, A. J., P. J. Doherty, and D. McB. Williams. 1992. Multi-scale analysis of recruitment of a
coral reef fish on the Great Barrier Reef. Mar. Ecol. Prog. Ser. 82: 131–141.
Gibson, R. N. 1994. Impact of habitat quality and quantity on the recruitment of juvenile flatfishes.
Neth. J. Sea Res. 32: 191–206.
Green, A. L. 1996. Spatial, temporal and ontogenetic patterns of habitat use by coral reef fishes
(Family Labridae). Mar. Ecol. Prog. Ser. 133: 1–11.
__________. 1998. Spatio-temporal patterns of recruitment of labroid fishes (Pisces: Labridae and
Scaridae) to damselfish territories. Env. Biol. Fish. 51: 235–244.
Hixon, M. A. and J. P. Beets. 1993. Predation, prey refuges, and the structure of coral-reef fish assemblages. Ecol. Monogr. 63: 77–101.
Itzkowitz, M. 1985. Aspects of the population dynamics and reproductive success in the permanently territorial beaugregory damselfish. Mar. Beh. Phys. 12: 57–69.
___________, D. E. Itzkowitz, and D. Shelly. 1995. Territory use and disuse in the beaugregory
damselfish. Bull. Mar. Sci. 57: 653–662.
Jones, G. P. 1991. Postrecruitment processes in the ecology of coral reef fish populations: a multifactorial perspective. Pages 294–328 in P.F. Sale, ed. The ecology of fishes on coral reefs.
Academic Press, New York.
Kaufman, L., J. Ebersole, J. Beets, and C. C. McIvor. 1992. A key phase in the recruitment dynamics of coral reef fishes: post-settlement transition. Env. Biol. Fish. 34: 109–118.
Lawson, G. L., D. L. Kramer, and W. Hunte. 1999. Size-related habitat use and schooling behavior
in two species of surgeonfish ((Acanthurus bahianus and A. coeruleus) on a fringing reef in
Barbados, West Indies. Env. Biol. Fish. 54: 19–33.
Leis, J. M. and B. M. Carson-Ewart. 1999. In situ swimming and settlement behaviour of larvae
of an Indo-Pacific coral-reef fish, the coral trout Plectropomus leopardus (Pisces: Serranidae).
Mar. Bio. 134: 51–64.
Mahon, R. and S. D. Mahon. 1994. Structure and resilience of a tidepool fish assemblage in Barbados. Env. Biol. Fish. 41: 171–190.
McCormick, M. I. 1999. Delayed metamorphosis of a tropical reef fish ((Acanthurus triostegus): a
field experiment. Mar. Ecol. Prog. Ser. 176: 25–38.
McFarland, W. N., E. B. Brothers, J. C. Ogden, M. J. Shulman, E. L. Bermingham, and N. M. Kotchian-Prentiss. 1985. Recruitment patterns in young french grunts, Haemulon flavolineatum
(family Haemulidae), at St. Croix, Virgin Islands. Fish. Bull. 83: 151–161.
McGehee, M. A. 1995. Juvenile settlement, survivorship and in situ growth rates of four species of
Caribbean damselfishes in the genus Stegastes. Env. Biol. Fish. 44: 393–401.
Nemeth, R. S. 1998. The effect of natural variation in substrate architecture on the survival of
juvenile bicolor damselfish. Env. Biol. Fish. 53: 129–141.
Parrish, J. D. 1989. Fish communities of interacting shallow-water habitats in tropical oceanic
regions. Mar. Ecol. Prog. Ser. 58: 143–160.
Planes, S., A. Levefre, P. Legendre, and R. Galzin. 1993. Spatio-temporal variability in fish recruitment to a coral reef (Moorea, French Polynesia). Coral Reefs 12: 105–113.
Risk, A. 1997. Effects of habitat on the settlement and post-settlement success of the ocean surgeonfish Acanthurus bahianus. Mar. Ecol. Prog. Ser. 161: 51–59.
________. 1998. The effects of interactions with reef residents on the settlement and subsequent
persistence of ocean surgeonfish, Acanthurus bahianus. Env. Biol. Fish. 51: 377–389.
Robertson, D. R. 1988. Abundances of surgeonfishes on patch-reefs in Caribbean Panama: due to
settlement, or post-settlement events? Mar. Biol. 97: 495–501.
173
_____________. 1992. Patterns of lunar settlement and early recruitment in Caribbean reef fishes
at Panama. Mar. Biol. 114: 527–537.
_____________, N. V. C. Polunin, and K. Leighton. 1979. The behavioral ecology of three Indian
Ocean surgeonfishes ((Acanthurus lineatus, A. leucosternum, and Zebrasoma scopas): Their
feeding strategies, and social and mating systems. Env. Biol. Fish. 4: 125–170.
Sale, P. F., P. J. Doherty, G. J. Eckert, W. A. Douglas, and K. J. Ferrell. 1984. Large scale spatial and
temporal variation in recruitment to fish populations on coral reefs. Oecologia 64: 191–198.
Sancho, G., D. Ma, and P. S. Lobel. 1997. Behavioral observations of an upcurrent reef colonization event by larval surgeonfish Ctenochaetus strigosus (Acanthuridae). Mar. Ecol. Prog. Ser.
153: 311–315.
Schmitt, R. J. and S. J. Holbrook. 1999. Settlement and recruitment of three damselfish species:
larval delivery and competition for shelter space. Oecologia 118: 76–86.
Shulman, M. J. 1985. Recruitment of coral reef fishes: effects of distribution of predators and
shelter. Ecology 66: 1056–1066.
_____________ and J. C. Ogden. 1987. What controls tropical reef fish populations: recruitment
or benthic mortality? An example in the Caribbean reef fish Haemulon flavolineatum. Mar.
Ecol. Prog. Ser. 39: 233–242.
____________, ______________, J. P. Ebersole, W. N. McFarland, S. L. Miller, and N. G. Wolf.
1983. Priority effects in the recruitment of juvenile coral reef fishes. Ecology 64: 1508–1513.
Sokal, R. R. and F. J. Rohlf. 1995. Biometry. W.H. Freeman and Company, New York. 887 p.
Sponaugle, S. and R. K. Cowen. 1996. Larval supply and patterns of recruitment for two Caribbean
reef fishes, Stegastes partitus and Acanthurus bahianus. Mar. Fresh. Res. 47: 433–447.
___________ and ____________. 1997. Early life history traits and recruitment patterns of Caribbean wrasses (Labridae). Ecol. Monogr. 67: 177–202.
Stobutzki, I. C. and D. R. Bellwood. 1997. Sustained swimming abilities of the late pelagic stages
of coral reef fishes. Mar. Ecol. Prog. Ser. 149: 35–41.
Thresher, R. E. and E. B. Brothers. 1989. Evidence of intra- and inter-oceanic regional differences
in the early life history of reef-associated fishes. Mar. Ecol. Prog. Ser. 57: 187–205.
Tolimieri, N. 1998a. The relationship among microhabitat characteristics, recruitment and adult
abundance in the stoplight parrotfish, Sparisoma viride, at three spatial scales. Bull. Mar. Sci.
62: 253–268.
__________. 1998b. Effects of substrata, resident conspecifics and damselfish on the settlement
and recruitment of the stoplight parrotfish, Sparisoma viride. Environ. Biol. Fish. 53: 393–
404.
___________, P. F. Sale, R. S. Nemeth, and K. B. Gestring. 1998. Replenishment of populations
of Caribbean reef fishes: are spatial patterns of recruitment consistent through time? J. Exp.
Mar. Biol. Ecol. 230: 55–71.
Tupper, M. and R. G. Boutilier. 1995. Effects of habitat on settlement, growth, predation risk and
survival of a temperate reef fish. Mar. Ecol. Prog. Ser. 151: 225–236.
_________ and F. Juanes. 1999. Effects of a marine reserve on recruitment of grunts (Pisces: Haemulidae) at Barbados, West Indies. Env. Biol. Fish. 55: 53–63.
Victor, B. C. 1991. Settlement strategies and biogeography of reef fishes. Pages 231–260 in P.F.
Sale, ed. The ecology of fishes on coral reefs. Academic Press, San Diego.
Vigliola, L., M. L. Harmelin-Vivien, F. Biagi, R. Galzin, A. Garcia-Rubies, J. G. Harmelin, J.
Y. Jouvenel, L. Direach-Boursier, E. Macpherson, and L. Tunesi. 1998. Spatial and temporal
patterns of settlement among sparid fishes of the genus Diplodus in the northwestern Mediterranean. Mar. Ecol. Prog. Ser. 168: 45–56.
Wellington, G. M. 1992. Habitat selection and juvenile persistence control the distribution of two
closely related Caribbean damselfishes. Oecologia 90: 500–508.
Williams, D. McB., S. English, and M. J. Milicich. 1994. Annual recruitment surveys of coral reef
fishes are good indicators of patterns of settlement. Bull. Mar. Sci. 54: 314–331.
174
Address: Biology Department, University of Massachusetts Boston, 100 Morrissey Blvd., Boston, Massachusetts 02125. Corresponding Author Present Address: (A.A.) Mote Marine
Laboratory, Charlotte Harbor Field Station, P.O. Box 2197, Pineland, Florida 33945. Phone:
(239) 283-1622, E-mail: <aadams@mote.org>.
FISH DENSITY, DIVERSITY, AND SIZE-STRUCTURE
WITHIN MULTIPLE BACK REEF HABITATS OF
KEY WEST NATIONAL WILDLIFE REFUGE
David B. Eggleston, Craig P. Dahlgren, and Eric G. Johnson
ABSTRACT
Tropical and subtropical back reef habitats such as seagrass meadows, mangrove
prop-roots, and channels bisecting mangrove islands presumably serve as important
nursery areas for numerous fishes. This study provides an initial step towards identification of the nursery role of specific habitats within multiple back reef habitats by
quantifying fish density, diversity, and size-structure, and was part of a larger study that
used aerial photographs, ground-truthing, and GIS software to map putative nursery
habitats in the Key West National Wildlife Refuge (KWNWR). Visual surveys assessed
fish density, diversity, and size-structure in the Lakes and Marquesas regions of the
KWNWR over a 3-mo period and across the marine habitats of concern (seagrass,
channels, mangroves, hardbottoms, patch reefs, offshore reefs). A combination of band
transects and 10-min surveys provided a more complete overall species assessment than
either method in isolation. Mangrove prop-root habitats contained the highest relative
mean density and diversity of fish, with abundant forage fish such as silverside minnows
(Atherinidae) and herrings (Clupeidae), as well as a high number of piscivores such as
gray snapper Lutjanus griseus (Linnaeus, 1758) and barracuda Sphyraena barracuda
(Walbaum in Artedi, 1792). Channel habitats contained the greatest diversity of microhabitats, and contained a relatively high diversity of fish compared to seagrass. Channel
habitats typically harbored juvenile snappers (Lutjanidae), grunts (Haemulidae), and
forage fish (Atherinidae). Qualitatively, we observed greater numbers of relatively large
gamefish, as well as rare and threatened species in channel and mangrove habitats than
any other habitat. Conversely, seagrass contained higher fish densities than channels.
Increases in the size-frequency of certain species, such as S. barracuda, Pomacanthus
arcuatus (Linnaeus, 1758), and Gerres cinereus (Walbaum in Artedi, 1792), from backreef habitats such as seagrass and mangroves, to channels and eventually patch and offshore reefs were suggestive of ontogenetic patterns of habitat use. In contrast, the smallest stages of L. griseus were found exclusively in seagrass, but remaining size classes,
including adults, were found at all of the habitats surveyed. In contrast, the smallest size
classes of Halichoeres bivitattus (Bloch, 1791), Lutjanis synagris (Linnaeus, 1758) and
Haemulon sciurus (Shaw, 1803) were found in nearly all of the habitats examined. We
found no relationship between fish density and diversity, or seagrass shoot density and
blade height. Inclusion of seagrass, mangrove, and channel habitats in future studies of
reef fish growth, survival, and emigration should produce a more complete picture of
their nursery role in tropical back reef environments.
This publication is part in a series of papers resulting from a scientific workshop held
at the Caribbean Marine Research Center (December 2001) to evaluate the importance
of back reef systems for supporting biodiversity and productivity of marine ecosystems.
Tropical and subtropical back reef habitats such as seagrass meadows and mangrove
prop-roots presumably serve as important nursery areas for numerous reef fishes (Weinstein and Heck, 1979; Stoner, 1983; Sogard et al., 1987; Morton, 1990; Eggleston, 1995;
Ley et al., 1999; Nagelkerken et al., 2000; Dahlgren and Eggleston, 2001, Laegdsgaard
and Johnson, 2001). These habitats are thought to intercept large numbers of larvae and
provide abundant food resources and protection from predators (Parrish, 1989; DahlBulletin of Marine Science
175
176
gren and Eggleston 2000; Laegsgaard and Johnson, 2001). Fish eventually migrate from
these nursery habitats to nearshore patch reefs and offshore reefs as they mature. The
term “back reef nursery” implies that juvenile fish density and ecological processes such
as growth, survivorship, and emigration success should be enhanced compared to adjoining juvenile habitat types (Beck et al., 2001). A habitat is a nursery if its contribution
per unit area to the production of individuals that recruit to adult populations is greater,
on average, than production from other habitats in which juveniles occur (Beck et al.,
2001). The ecological processes operating in nursery habitats, as compared with other
habitats, must support greater contributions to adult recruitment from any combination
of four factors: (1) density, (2) growth, (3) survival of juveniles, and (4) movement to
adult habitats (Beck et al., 2001). There are very few data that compare animal density
and ecological processes across multiple, structurally complex habitats that characterize
tropical and subtropical back reef environments (Nagelkerken et al., 2000; Beck et al.,
2001). This study provides an initial step towards identification of the nursery role (sensu
Beck et al., 2001) of specific habitats within multiple back reef habitats by quantifying
fish density, diversity and size-structure, and was part of a larger study that mapped
putative nursery habitats in the KWNWR and quantified distribution and abundance of
Caribbean spiny lobster in these habitats (Eggleston and Dahlgren, 2001).
The Florida Keys marine ecosystem in the U.S. supports important commercial and
recreational fisheries for both fish and invertebrates [e.g., snapper, Lutjanidae; grouper, Serranidae; Caribbean spiny lobster, Panulirus argus (Latreille, 1804); stone crab,
Menippe mercenaria (Say, 1818)], as well as a marine-based tourism industry. Despite
the ecological and economic significance of the Florida Keys coral reef ecosystem, it is
faced with a growing number of threats including water quality degradation (Lapoint
and Clark, 1992), habitat loss (Robblee et al., 1991; Durako, 1994; Herrnkind et al.,
1997), and overfishing (e.g., Ault et al., 1998). These multiple insults have led to the
Florida Keys coral reef ecosystem being classified as an “ecosystem-at-risk” (NMFS,
1996). To conserve this threatened ecosystem, a network of protected areas is being
established to safeguard its living resources.
The first protected area in the Florida Keys was the Key West National Wildlife Refuge (KWNWR), established in 1908. Marine habitats within the KWNWR include mangroves, seagrass meadows, hardbottom, macroalgal beds, sand flats, and coral reefs. Both
recreational and commercial fishing are allowed within a majority of the KWNWR.
Although most of the refuge consists of shallow bank and seagrass habitats interspersed
with mangrove islands, it also contains numerous patch reefs within Hawk Channel to
the south and north towards the Gulf of Mexico. Within the KWNWR are two smaller
areas referred to as the “Lakes” and “Marquesas” (122 km2; Fig. 1). These smaller areas contain a complex mosaic of habitat types including seagrass, channels, macroalgal
meadows, hardbottoms, mangroves, and patch reefs. In this study, we 1) mapped the
distribution and aerial cover of habitats for reef fish in the “Lakes” and “Marquesas,”
2) quantified fish density, diversity, and size-structure across the mosaic of five habitats
described above, as an initial step towards identification of the nursery role (sensu Beck
et al., 2001) of these habitats, and 3) quantified the relationship between fish abundance
and specific habitat features.
EGGLESTON ET AL.: FISH DENSITY AND DIVERSITY IN BACK REEF HABITATS
177
Methods and Materials
Habitat Mapping and Sampling Sites
The KWNWR is a rectangular-shaped (82o 10ʹ W × 24o 40ʹ N, 81o 49ʹ W × 24o 27ʹ N) area
measuring 766.9 km2. Initial site and habitat reconnaissance was conducted during July, 1999
by ground-truthing aerial photographs (1:48,000; obtained from the National Ocean Service,
NOAA) of the Lakes (24o 35ʹ N, 82o 55ʹ W) and Marquesas (24o 36ʹ N, 82o 8ʹ W) regions with a
small (7 m) boat, and by snorkeling and SCUBA diving. We identified six habitats within which
to quantify the density, diversity, and size-structure of fish: 1) submerged mangrove (Rhizophora
mangle L.) prop-roots; 2) channels bisecting mangrove islands or seagrass shoals; 3) subtidal
seagrass beds; 4) inshore hardbottoms; 5) inshore patch reefs; and 6) offshore reefs (Fig. 1). We
digitized the aerial photographs and estimated areal cover of seagrass beds (both subtidal and
intertidal) and channels, as well as the perimeter of mangroves, using GIS ArcView software. We
were unable to delineate hardbottom and reef habitats because these substrates were not clearly
visible.
Figure 1. Schematic of habitat types and locations of sampling stations within the Lakes and Marquesas regions within the Key West National Wildlife Refuge (KWNWR), Florida. Habitat maps
were generated with ground-truthed, geo-referenced, and digitized aerial photographs (1:48,000
scale obtained from the National Ocean Services, NOAA). C = channel habitats, S = seagrass
habitats, M = mangrove habitats, H = hardbottom habitats, and PR = patch reefs. Alphanumerics
(e.g., M1, S1, etc.) are sample sites. See text for details regarding the sampling approach.
178
Fish Density and Microhabitat Features
We used visual survey techniques to rapidly assess the distribution and abundance of reef
fishes in the Lakes and Marquesas regions over a 3-mo period and across the marine habitats of
concern (seagrass, channels, mangroves, hardbottoms, patch reefs, offshore reefs). All surveys
were conducted during the day when water visibility exceeded 10 m, and during a 7-d window
bracketing the new moon each month so as not to confound fish counts with possible variation in
abundance due to diel and lunar variation in migration behavior (Helfman et al., 1982; Rooker
and Dennis, 1991; Kasai et al., 2000 and references therein).
We used a stratified random survey procedure for fishes in subtidal seagrass beds (Fig. 1), and
randomly chose channels, mangroves, hardbottoms, patch reefs, and offshore reefs to survey
fish. The areal cover of seagrass and channels was ~3–6 times greater in the Lakes than in the
Marquesas (see below), thus our sample size for seagrass and channels was ~4 times higher in
the Lakes than in the Marquesas. Fish counts were conducted with two general approaches, each
using SCUBA divers: 1) visual band transects, which provided density estimates, and 2) 10-min
surveys with visual estimates of area searched, which provided an additional, although cruder,
measure of density. During September 2002, we assessed the accuracy of our visual estimates of
distance traveled during 10-min swims in seagrass and mangroves that were located in the nearby
Great White Heron National Wildlife Refuge.
Habitat-Specific Surveys
Seagrass.—To sample fishes in seagrass, a grid system containing cells measuring ~200 × 200
m was superimposed over a navigational chart of each region. We then randomly chose 19 cells
and six cells at the Lakes and Marquesas, respectively, and used SCUBA divers to quantify fish
density, diversity and size-frequency, as well as habitat characteristics, in each cell. If a randomly
chosen cell corresponded to an intertidal seagrass bed, we randomly chose another cell until a
subtidal seagrass site was selected. Two parallel band transect lines (60 × 2 m) were then placed
as close as possible to the center latitude and longitude coordinates for each grid cell using a GPS.
The transects were located parallel to each other ~100 m apart, and were identified with floats at
the ends. Divers initiated their surveys ~20 min after the transect lines were deployed, and began
at the downstream edge of each cell such that divers began their survey by swimming against
the current. All counts within a single band were made by two divers; the first diver would count
fishes and the second diver quantified habitat characteristics (see below). Fish total length (TL)
was estimated to the nearest 1 cm by comparing a fish to a ruler attached perpendicular to the far
end of a 70 cm rod held out from a diver (Eggleston, 1995; Eggleston et al., 1997). This device
helped avoid underwater magnification problems in estimating fish sizes. Divers counting fishes
slowly swam along each transect and used a 2-m long PVC-pipe to delineate the 2-m band width.
Although estimates of fish size were made at the resolution of 1 cm, fish sizes were compared
among habitats by categorizing size distributions into 5-cm size classes. The 2-m PVC-pipe was
also slowly pushed through seagrass or macroalgae to “herd” small, cryptic fishes for periodic
enumeration. The response variable produced from the band transects was the density (no. 120
−
m−2
) of fishes.
There is often a positive relationship between macrophyte structural complexity (e.g., seagrass
shoot density and biomass) and fish density in macroalgal and seagrass systems (Carr, 1994;
Eggleston, 1995; Levin and Hay, 1996). Thus, we measured seagrass habitat characteristics (mean
shoot density and blade height) adjacent to each band transect within a grid cell (N = 2 cell−1).
Seagrass shoot density was quantified by blindly tossing a 0.07 m2 quadrat near the starting point
of each band transect. Individual seagrass shoots and mean blade height (mean of 10 haphazardly
chosen shoots measured with a ruler) within a quadrat were counted by SCUBA divers. After seagrass characteristics were measured within a quadrat, percent cover of six benthic habitat categories (each covering > 1% of the total area) were estimated along each 2 × 60 m band to the nearest 5%: (1) seagrass (Thalassia testudinum Banks & Soland. ex Koenig, Syringodium filiforme
Kuetz., Halodule sp.); (2) Laurencia sp. (including Laurencia-covered coral clumps); (3) other
179
macroalgae (including Penicillus sp., Udotea sp., and Halimeda sp.); (4) relatively large sponges
[primarily Speciospongia vesparia (Lamarck, 1814)]; (5) coral (predominantly Porites porites
(Pallas, 1766) including both live and dead coral rubble); and (6) sand (usually sand or a thin sand
veneer over rock). To increase the accuracy of estimates, habitat data were recorded every 20
m and then combined for each band. In our previous studies using these techniques (Eggleston,
1995; Dahlgren and Eggleston, 2001), two divers independently estimated percent cover along a
given transect, which never varied much between divers (< 10% in all cases; Dahlgren and Eggleston, 2001). Therefore, we used one diver per transect to quantify habitat characteristics in this
study. Diver estimates, however, were averaged across transects to reduce individual diver bias.
After a band transect was completed, divers continued to swim up-current in a straight line,
and initiated a timed (10 min) visual survey for fishes. The timed transect started ~10 m from
the end of the band transect to reduce the chance of counting the same fishes. At the end of the
10-min survey the diver surfaced, visually estimated the distance back to the band transect float,
and subtracted 10 m to estimate the total distance traveled. The distance traveled averaged 120 m.
During September 2002, we used a differential GPS (accurate within 3 m) on a research boat to
assess the accuracy of diver estimates of distance traveled. We compared diver estimates of distance traveled with the known distance from latitude/longitude points taken at the start and end
of the 10-min swim. Although divers tended to over-estimate the distance traveled (mean = + 4.5
m, SE = 4.9 m, N = 12), there was no significant difference between diver estimates of distance
traveled and distances measured with the differential GPS (paired t-test, t = 0.90, P = 0.39). Thus,
diver estimates of distance traveled in seagrass were relatively accurate. The width of the 10-min
transect was determined by water visibility, which averaged 10 m. Thus, the 10-min survey in seagrass covered an estimated average area of 1200 m2. Although this method of estimating distance
traveled was somewhat crude, it provided an estimate of fish densities and variance that could be
used to 1) estimate required sample sizes for future studies, 2) make relative comparisons of density across structurally complex habitat types, and 3) qualitatively compare fish density between
band transect and timed swim survey methods. The mean values between the two surveys within
a seagrass cell (i.e., two band transects and two, 10-min swims) served as a single replicate (N =
19 and six at the Lakes and Marquesas, respectively).
Channel Habitats.—Channel habitats measuring 2–4 m deep that bisect mangrove islands and
intertidal seagrass (Fig. 1) probably serve as important conduits for ontogenetic migrations of
some species from nursery habitats within the Lakes and Marquesas to offshore reefs. In total, 14
and four channels were randomly chosen from available channels at the Lakes and Marquesas, respectively (Fig. 1). Two separate band transects and two separate 10-min surveys were conducted
within each channel as described above for seagrass. We quantified fishes and habitat characteristics as described above for seagrass habitats, with the additional quantification of sponge habitat characteristics. The mean values for fish density and habitat characteristics between the two
surveys (two bands or two, 10-min surveys) within a channel served as a single replicate (N = 14
and four at the Lakes and Marquesas, respectively). The average estimated area searched during
10-min surveys in channel habitats was 1000 m2.
Large sponges were a relatively common feature of channel bottoms probably due to high tidal
current speeds (1–1.5 m s−1), which scoured the bottom providing a hard substrate for sponge attachment, and delivered a high concentration of suspended food for these suspension-feeders. The
total number of sponges and sponge volume per transect (120 m 2) was estimated by divers. We
estimated sponge volume by measuring the radius (r) and height (cm) of each sponge with a ruler,
and then treating each sponge as a cylinder and multiplying height by π r2.
Hardbottom Habitats.—Within seagrass beds in the Lakes region, we observed hardbottom areas that were devoid of seagrass but contained solution holes, sponges, and coral rubble, and were
typically 1–4 ha in area. These hardbottom areas were absent for the most part in the Marquesas,
and relatively uncommon in the Lakes (H1 and H2 located east and west of “Archer Key”; Fig.
1). Hardbottom habitats provided some of the only crevices available for crevice-dwelling fishes
(e.g., Serranidae) within large seagrass beds, and so were included in our Lakes surveys during
September (N = 2 hardbottom sites). Fishes and habitat features were quantified using both band
180
transect and timed survey methods, as described above for seagrass habitats. The average estimated area searched during 10-min surveys in hardbottom areas was 400 m2.
Mangroves.—Snorkelers conducted 10-min surveys for fishes in mangrove prop-root habitats.
Snorkel was used instead of SCUBA to avoid becoming entangled on the prop-root canopy while
searching the shallow interstices of the prop-root canopy. Each mangrove survey was conducted
by 2–4 snorkelers surveying non-overlapping areas, with the mean values between divers used
in statistical analyses (N = 5 and seven in the Lakes and Marquesas, respectively). We estimated
the area covered during a 10-min search by recording the distance that we could reliably count
fish within the prop-root canopy (~2–4 m), and by placing floats at the beginning and end of
a survey. After a particular survey was completed, we visually estimated the linear distance
surveyed between the floats marking the beginning and end of a transect, and accounted for indentations along the mangrove fringe, which would add distance to the distance traveled. During
September 2002, we used a tape measure to assess the accuracy of diver estimates of distance
traveled. We compared diver estimates of distance traveled with the known distance from laying
out a tape measure along the mangrove fringe that was surveyed. In this case, divers tended to
underestimate the distance traveled (mean = −1.75 m, SE = 2.35 m, N = 20); however, there was no
significant difference between diver estimates of distance traveled and distances measured with
the underwater tape measure (paired t-test, t = -0.75, P = 0.47). Thus, diver estimates of distance
traveled along the mangrove fringe were relatively accurate. Our estimates of linear distance traveled during a 10-min search ranged from 10–80 m, and averaged 38 m. The average estimated
area covered during 10-min surveys in mangrove prop-root habitats was 152 m2. Although we
randomly chose seven out of all available mangrove habitats to sample at the Marquesas, we were
restricted to five mangrove areas in the Lakes (Fig. 1) because all of the other mangrove areas
were inaccessible by boat due to extremely shallow water.
Nearshore Patch and Offshore Reefs.—We counted reef fishes at all patch reef sites that we
could locate near the Lakes (N = 5) and Marquesas (N = 2; Fig. 1) with the 10-min survey
method. Patch reefs consisted primarily of clusters of patch coral heads surrounded by seagrass
south of the Marquesas, and a series of ledges, hardbottoms, and patch heads north of Cottrell
Key in the Lakes (Fig. 1). We also used this survey method to quantify the abundance of fishes at
the following randomly chosen offshore reefs within the KWNWR: “Sand Key,” “Western Dry
Rocks,” “Coalbin Rocks,” and the eastern portion of “Cosgrove Shoals.” These offshore reefs are
not shown on Figure 1, but were located 12–15 km south of the Lakes and Marquesas along the
southern boundary of the KWNWR. Fish counts and sizes were estimated as described above for
seagrass habitats, but with four divers. The divers surveyed areas that were 90º in the opposite
direction of each other. We used the mean counts from a total of four divers per reef in statistical
analyses. The areas searched per diver during 10-min surveys of patch reefs ranged from 300–
700 m2 and averaged 600 m2. The areas searched per diver during 10-min surveys of offshore
reefs ranged from 1000–1700 m2, and averaged 1500 m2. In summary, the band-transect and 10min survey methods were used in seagrass, channel, and hardbottom habitats, whereas only the
10-min survey method was used in mangrove prop-roots, patch reefs, and offshore reefs.
Statistical Analyses
We examined the effects of region (Lakes vs Marquesas) on the mean density of reef fishes
in seagrass, mangrove, channel, and patch reef habitats with separate t-tests. We used separate ttests, rather than an ANOVA approach that would include habitat type as a factor, because we did
not know how the accuracy of our visual survey techniques compared across habitat types, and
because of widely different search areas across habitat types (e.g., 152 m2 for mangroves and 1200
m2 for seagrass). For example, animal diversity often increases with area searched (Rosenzweig,
1995). The response variables from the band transect and timed surveys were the mean densities
of: 1) fishes (including atherinids, which dominated the counts in many surveys); 2) fishes without
atherinids; 3) fish families; and 4) fish species. The data were log of (x + 1) transformed when
necessary to meet the assumptions of normality (tested with a Kolmogorov-Smirnov test) and
homogeneity of variances (tested with Levene’s test). We calculated the mean density of fishes in
181
hardbottom and offshore reef habitats, but did not statistically contrast these data across regions
(Lakes vs Marquesas) because hardbottoms were only sampled in the Lakes, and offshore reefs
were located well outside of the Lakes and Marquesas.
A forward, stepwise multiple regression model was used to examine the relationship between
habitat characteristics and reef fish density measured during band transects in seagrass and channel habitats. Separate models were used for seagrass and channels. For seagrass, the regression
model included as independent variables: 1) seagrass shoot density, 2) mean seagrass blade height,
and 3) the percent cover of Thalassia, Syringodium, Halodule, Laurencia, other macroalgae,
sponges, coral, and sand. For channel habitats, the independent variables were similar to those of
seagrass, with the addition of 1) sponge density and 2) mean sponge volume. Alpha to enter and
remove factors from the model was 0.10.
Results
Habitat Areal Cover and Microhabitat Characteristics
The most visible components of the landscape in aerial photographs of the Lakes and
Marquesas were mangrove-fringed islands, dense seagrass beds, shallow intertidal seagrass, and channels (Fig. 1). Intertidal seagrass was the most extensive habitat in terms
of areal cover at both regions (Table 1), with an area of 58.27 km2 and 14.52 km2 at
the Lakes and Marquesas, respectively. The second most extensive habitat was subtidal
seagrass beds (Table 1). The areal cover of subtidal seagrass beds was ~6 times greater
at the Lakes (37.89 km2) than at the Marquesas (5.7 km2). Diver surveys covered an
estimated total subtidal seagrass area of ~0.05 km2 and ~0.02 km2 at the Lakes and Marquesas, respectively (Table 1). The areal cover of channels was ~3 times greater at the
Lakes (5.6 km2) than at the Marquesas (1.65 km2). We surveyed ~0.007% and 0.006%
of channel habitats at the Lakes and Marquesas, respectively (Table 1). The perimeter
of mangrove fringe surrounding small islands (keys) within the KWNWR was slightly
higher in the Marquesas (33.22 km) than in the Lakes (20.55 km).
Within subtidal seagrass beds, seagrass percent cover (primarily Thalassia) was always extremely high at the Marquesas, ranging from 94–100% (Appendix 1). We did not
observe any appreciable areas that contained a mixture of seagrass and macroalgal beds
at the Marquesas, whereas we sometimes observed a mosaic of both dense and moderate
density seagrass interspersed with macroalgal meadows at the Lakes. Seagrass percent
cover in transects at the Lakes ranged from a low of 5–14% in areas that contained a
mixture of seagrass and macroalgal beds (e.g., stations S6, S14, Appendix 1), to a high
of 100% (Appendix 1). The primary species of seagrass counted in quadrats at the Lakes
and Marquesas was T. testudinum (Appendix 1), although we observed expansive beds
of Syringodium filiforme and Halodule sp. at both the Lakes and Marquesas, particularly
in channel habitats at the Marquesas. Although mean seagrass shoot count and blade
height was higher in the Marquesas than the Lakes (Table 2), the mean values did not
vary significantly by region (t-test; P > 0.05). Mean water depths in subtidal seagrass
meadows were ~1.5 m in both the Lakes and Marquesas.
In general, channel habitats contained more diverse microhabitats than seagrass beds,
with macroalgal clumps (Laurencia sp. + “other macroalgae”) providing the greatest
percent cover (mean = 27%), followed by sand (mean = 25%) and seagrass (mean = 24%)
(Appendix 2). Sponges were absent in one of four channels in the Marquesas and two
of 14 channels in the Lakes (Appendix 2). The mean density and volume of sponges in
182
Table 1. Estimates of areal cover of fish habitat in the Lakes and Marquesas regions of the
KWNWR estimated from ground-truthed, geo-referenced, and digitized aerial photographs
using ArcView software. Also provided are estimates of percent areal cover by a particular
habitat within a region, and estimates of areal cover of visual surveys (band transects vs 10min surveys) by divers. Estimates of areal cover for band transects was based on the number of
seagrass or channel stations sampled in a region × 2 replicate transects/location (e.g., S1, C1,
etc.) × 120 m2. Estimates of areal cover for 10-min diver surveys were based on the number
of seagrass or channel stations sampled in a region × 2 replicate transects/location × 1200
m2. See text for details concerning areal cover of diver surveys. N/A = habitat not available to
sample. BT = band transects and 10-min = 10-min surveys.
Habitat
Areal cover in km2 and (% cover) within a region
Lakes
Marquesas
Subtidal seagrass
Intertidal seagrass
Seagrass/macroalgal beds
Channels
37.89 (34%)
58.27 (52%)
10.81 (10%)
5.60 (4%)
Mangroves
20.55
5.70 (26%)
14.52 (66%)
N/A
1.65 (8%)
Perimeter (km)
33.22
Area surveyed (km2)
Subtidal seagrass* (BT)
Subtidal seagrass* (10-min)
Subtidal seagrass* (total)
Channels (BT)
Channel (10-min)
Channels (total)
0.0046
0.0456
0.0502
* Includes seagrass/macroalgal beds for the Lakes region
0.0014
0.0144
0.0158
0.0010
0.0096
0.0106
channel habitats did not differ between the Lakes and Marquesas (Table 3; t-test, P >
0.05).
Fish Density, Diversity, and Size-Structure.
We conducted a total of 248 diver surveys (124 band transects and 124 timed surveys
combined) during a 21 d period in August–October, 1999 in the KWNWR, and recorded density, diversity, and size-structure of 114 species of fish representing 42 families (Appendix 3). We also observed commercially important stone crabs, queen conch
(Strombus gigas Linnaeus, 1758) and Caribbean spiny lobster, as well as mating pairs
of horseshoe crabs [Limulus polyphemus (Linnaeus, 1758)] within the KWNWR, but
did not record their numbers or estimate size-frequencies. We also observed juvenile
goliath grouper [Epinephelus itajara (Lichtenstein, 1822), ~30 cm TL] and Nassau grouper [Epinephelus striatus (Bloch, 1792); ~20 cm TL] residing in mangrove and crevice
habitats outside of our surveys. Nassau and goliath grouper are federally protected species in the U.S.
Comparisons of Band Transect vs 10-min Surveys.—Similarity in the number of fish
families, genera, and species between band transects and 10-min surveys was ~0.50
(Table 4). Divers swimming band transects often observed small, cryptic species such
as newly settled L. griseus, L. synagris, and G. cinereus that were missed during 10min swims, whereas divers conducting 10-min swims observed more large, transient
fish species. Higher numbers of fish families, genera, and species were observed during
10-min surveys than in band transects (Table 4). There were 57 unique species observed
in 10-min surveys, and eight unique species observed in band transects. Examples of
183
Table 2. Mean, range, standard deviation, and sample size for shoot count and blade height in
subtidal seagrass beds in the Lakes and Marquesas regions of the KWNWR.
Mean
Range
Standard deviation
n
Mean
Range
Standard deviation
n
Lakes
Shoot count (no. 0.07 m−2)
26.23
12.00–43.00
10.7
18
Marquesas
31.3
16.50–43.00
11.4
6
Blade height (cm)
29.74
15.50–47.35
9.52
18
36.50
27.30–45.10
6.29
6
unique species observed during 10-min surveys included: yellow jack Caranx bartholomaei Cuvier in Cuvier and Valenciennes, 1833, tarpon Megalops atlanticus Valenciennes in Cuvier and Valenciennes, 1847, red drum Scianops ocellatus (Linnaeus, 1766),
and lemon shark Negaprion brevirostris (Poey, 1868). Examples of unique species observed during band transects included: foureye butterflyfish Chaetodon capistratus Linnaeus, 1758, Hamlets Hypoplectrus sp., and Gobies Ioglossus sp. Thus, the band transect
and 10-min survey methods appear to be complementary in terms of characterizing fish
diversity in backreef habitats.
Habitat-Specific Rank Order of Fish Abundance
Seagrass Habitats.—The fish fauna inhabiting seagrass beds often consisted of patchily distributed and tightly packed schools of silversides and herrings (Atherinidae and
Clupeidae, respectively), which were typically located near the surface of the water column. Schools of demersal snapper (family Lutjanidae, primarily gray snapper, Lutjanus
griseus), mojarras (family Gerreidae, Gerres spp.), sea bream (family Sparidae, Archosargus rhomboidalis (Linnaeus, 1758)), and juvenile grunts (family Haemulidae, primarily Haemulon sciurus and H. plumieri) were observed amongst the seagrass blades and
bottom. We sometimes observed tarpon ((M. atlanticus) and bonnethead sharks Sphyrna
tiburo (Linnaeus, 1758) outside our band transects or after the 10-min surveys were
complete. The most numerically abundant fish family inhabiting seagrass during band
Table 3. Mean, range, standard deviation and sample size for sponge count and sponge volume in
channel habitats in the Lakes and Marquesas regions of the KWNWR.
Lakes
Mean
Range
Standard deviation
n
Mean
Range
Standard deviation
n
Sponge count (no. 120 m−2)
22.27
0.00–99.00
35.2
14
23.50
0.00–33.50
17.20
3
Sponge volume (cm3)
15,679.2
105.20–46352.70
17,679
7
Marquesas
16,283.90
717.10–33283.00
16,283.10
3
184
Table 4. Total number of fish families, genera, species, and unique species observed during
band transects and 10-min surveys at the Lakes and Marquesas regions of the KWNMR during
July–October, 1999. The Jaccard Index indicates the similarity of fish species, where 1 equals
complete similarity.
No. of families
No. of genera
No. of species
No. of unique species
Band transects
4
42
73
8
10-min surveys
48
66
122
57
Jaccard index
0.577
0.514
0.500
N/A
Figure 2. Rank order of abundance of fish from band transects in (A) seagrass, (B) channel, and
(C) hardbottom habitats pooled across the Lakes and Marquesas regions of the KWNWR.
185
Figure 3. Rank order of abundance of fishes from 10-min surveys in (A) seagrass; (B) channel;
(C) mangrove; (D) hardbottom; (E) patch reef; and (F) offshore reef habitats pooled across the
Lakes and Marquesas regions of the KWNWR. Note different y-axis values for (C) and (E).
transects was Atherinidae, followed by Gerreidae, Haemulidae, and Lutjanidae (Fig.
2A). Conversely, the family Sparidae (primarily Archosargus rhomboidalis) had the
highest density in seagrass using 10-min surveys, followed by Atherinidae, Gerreidae,
Haemulidae, and Lutjanidae (Fig. 3A). Estimates of fish density from band transects
were generally higher than 10-min surveys (Figs. 2A,3A).
Channel Habitats.—The most striking feature of fish assemblages inhabiting channel
habitats during band transects was the relatively high number of fish species and families compared to seagrass (Figs. 4C,D). Upon entering the water at channels and before
we began our transect surveys, we often saw sharks (primarily nurse sharks, Ginglymo-
186
Figure 4. The effects of region (Lakes vs Marquesas) and habitat (seagrass, channels) on the mean
(+ 1 SE) (A) density of fish; (B) fish without atherinids; (C) number of fish families; and (D) number of fish species measured with band transects. See text for details of statistical analyses.
stoma cirratum (Bonnaterre, 1788), lemon sharks, Negaprion brevirostris, and bonnethead sharks, Sphyrna tiburo), turtles [primarily loggerhead, Caretta caretta (Linnaeus,
1758) and green turtles, Chelonia mydas (Linnaeus, 1758)], and tarpon ((M. atlanticus).
These species typically swam out of the survey area by the time we initiated our 10-min
surveys. The family Lutjanidae (snappers) had the highest density in channel habitats
in band transects, followed by Labridae (wrasses), Haemulidae (grunts), and Scaridae
(parrotfishes) (Fig. 2B). The most common species of lutjanid was L. griseus, followed
by Lutjanis synagris. During the 10-min surveys within channel habitats, the family
with the highest density was Atherinidae, followed by Labridae (Halichoeres bivittatas),
Lutjanidae, and Scaridae (Fig. 3B). The top five families present in channel habitats
were similar when the band transect and 10-min survey rank order of abundance indices
were compared (Figs. 2B,3B). The exception was schooling atherinids, which were not
commonly observed during band transects when divers were searching methodically for
more cryptic species.
Hardbottom Habitats.—Within seagrass beds near “Archer Key” in the Lakes, there
were several low relief hardbottom areas that were devoid of seagrass but contained
solution holes, small sponges and patch corals (Fig. 1). These hardbottom areas represented some of the only crevice-type structure available within relatively large, monotypic stands of seagrass. The fish family with the highest density in hardbottom habitats
was Gerriedae, followed by Lutjanidae, Sparidae, Haemulidae, and Scaridae (Fig. 2C).
187
Figure 5. The effects of region (Lakes vs Marquesas) and habitat (seagrass, channels, mangrove,
patch reefs) on the mean (+ 1 SE) (A) density of fishes; (B) fishes without atherinids; (C) number
of fish families; and (D) number of fish species measured in 10-min surveys. The numbers above
each histogram in (A) denote the number of sampling sites. See text for details of statistical
analyses.
During the 10-min surveys, Lutjanidae had the highest density, followed by Gerreidae,
Sparidae, and Pomacentridae (Fig. 3D). Lutjanidae was the top family in both band
transect and 10-min surveys in hardbottoms; however, estimated densities were 5-times
higher in band transects than 10-min surveys (Figs. 2C,3D).
Mangrove Habitats.—The most striking feature of fish assemblages inhabiting mangrove prop-root habitats was the extremely high density and diversity of fish (Fig. 5),
and the clear domination by the family Atherinidae (silverside minnows; Fig. 3C), which
often hovered along the mangrove fringe in tightly packed schools containing 1000s of
individuals. The family Atherinidae had the highest density of any fish family across all
habitats surveyed. The second most abundant fish family in mangroves was the Clupeidae (herrings; Fig. 3C), which also formed pelagic schools along the mangrove fringe,
followed by the families Gerreidae, Lutjanidae (primarily L. griseus), and Haemulidae (primarily Haemulon aurolineatum Cuvier, 1830 and H. sciurus), which swam in
schools among the mangrove prop-roots. Several relatively large (> 40 cm TL) gamefish
species were observed residing within the mangrove prop-root canopy including tarpon
M. atlanticus, snook Centropomus undecimalis (Bloch, 1792), and red drum Sciaenops
ocellatus. We also observed Cubera snapper Lutjanus cyanopterus (Cuvier in Cuvier
and Valenciennes, 1828). Although not included in our 10-min surveys, we also observed the following species in mangrove habitats: lemon shark N. brevirostris, silver
188
porgy Diplodus argenteus (Valenciennes, 1830) and the mangrove terrapin Malaclemys
terrapin rhizophorarum (Fowler, 1906).
Patch Reefs.—Patch reef habitats had the second highest densities of fish after mangrove habitats, and were dominated by Atherinidae (skewed by three patch reefs), followed by Haemulidae (primarily Haemulon plumieri (Lacépede, 1801) and H. sciurus),
Labridae (primarily H. bivittatas), Lutjanidae (primarily L. griseus and L. apodus), and
Carangidae [primarily Caranx ruber (Bloch, 1793), Caranx latus Agassiz in Spix and
Agassiz, 1831, Caranx crysos (Mitchill, 1815)] (Fig. 3E).
Offshore Reefs.—The
.—The most abundant fish families residing in offshore reefs were Hae.—
mulidae (primarily Haemulon flavolineatum (Desmarest, 1823) and Haemulon plumieri),
followed by Labridae (primarily H. bivittatus and Halichoeres garnoti), Pomacentridae
(primarily Chromis multilineata (Guichenot, 1853) and Abudefduf saxatilis (Linnaeus,
1758)), Carangidae (primarily Caranx ruber, Caranx latus, Caranx crysas), and Acanthuridae [primarily Holocentrus marianus Cuvier in Cuvier and Valenciennes, 1829 and
Holocentrus adscensionis (Osbeck, 1765)] (Fig. 3F).
Effects of Region and Habitat Type on Fish Density
Band Transects.—In seagrass, there were significantly higher numbers of fish at the
Marquesas than at the Lakes, regardless of whether or not atherinids were included (Fig.
4A,B; t-test, P < 0.04). Conversely, in channels, there was no significant difference in
total numbers of fishes between the Lakes and Marquesas (Fig. 4A,B; t-test; P > 0.14).
There was also no difference in the number of fish species and families between the
Lakes and Marquesas, irrespective of habitat type (Fig. 4C,D; t-test; P > 0.26). Qualitatively, there was a trend towards higher diversity (fish species and families) in channels
than seagrass, despite relatively higher total numbers of fishes in seagrass (Fig. 4A,B).
Higher diversity of fishes in channels may have been due to the relatively high diversity
of microhabitats in channels (e.g., sponges, macroalgae, seagrass, corals; Appendix 2).
10-Min Surveys.—There were significantly higher total numbers of fishes observed at
the Marquesas than the Lakes, which was similar to the pattern observed during band
transects (Fig. 5A,B; t-test; all P < 0.04). There was no significant difference in the total
numbers (with and without Atherinids) and diversity (species and families) of fishes
between the Lakes and Marquesas in mangrove and patch reef habitats, and no difference in fish diversity between the Lakes and Marquesas in seagrass (Fig. 5, t-test; all P
> 0.18). Qualitatively, the mean density and diversity of fishes was 3–5 times higher in
mangroves than seagrass, channel, or patch reef habitats (Fig. 5). This pattern of highest
fish abundance and diversity in mangroves was consistent across both regions (Fig. 5).
Relationship Between Fish Density and Microhabitat Features
Although there was a positive trend between the density of lutjanids and seagrass
shoot density, there was no significant relationship between fish density and any of the
microhabitat features measured (multiple regression, all P > 0.09). The lack of a relationship between fish abundance and habitat characteristics (e.g., seagrass shoot density,
blade height, % macroalgal cover) may have been due to the relatively low number of
samples taken to characterize seagrass (i.e., one 0.07 m2 quadrat per transect line), or the
generally high amount of habitat available. For example, seagrass shoot density in the
Marquesas was very high, and percent cover averaged 95%. In the Lakes in areas where
seagrass percent cover was somewhat lower, these areas contained a mixture of alterna-
189
Figure 6. The size frequency distribution of Sphyraena barracuda (barracuda) observed during
10-min surveys in (A) seagrass; (B) channels; (C) mangroves; (D) patch reefs; and (E) offshore
reefs for the Lakes and Marquesas regions pooled. Note different y-axis values in (C).
tive microhabitats such as coral rubble and clumps of macroalgae (primarily Laurencia
spp.).
Ontogenetic Habitat Shifts
We examined body size-specific habitat use in several species as a possible indicator
of ontogenetic habitat shifts. One of the clearest examples of size-specific habitat use
was barracuda Sphyraena barracuda (Walbaum, 1792). The size-frequency of S. barracuda shifted from a mixture of size classes in seagrass and mangroves, to only the
largest stages in channels, patch reefs, and offshore reefs (Fig. 6). Gray angelfish Pomacanthus arcuatus (Linnaeus, 1758) also demonstrated size-specific habitat use, with the
smallest sizes in seagrass, hardbottom and channel habitats, and largest sizes in offshore
190
Figure 7. The size frequency distribution of Pomacanthus arcuatus (gray angelfish) observed
during 10-min surveys in (A) seagrass; (B) channels; (C) patch reefs; and (D) offshore reefs for
the Lakes and Marquesas regions pooled. Note lower density on y-axes of (C) and (D).
reefs (Fig. 7). Similarly, the smallest Gerres cinereus (Walbaum, 1792) (yellowfin mojarra) were found in seagrass, channels and mangrove habitats, and only the largest size
class occurred on patch reefs (Fig. 8). Although the smallest size classes of L. griseus
(gray snapper) occurred in seagrass, the largest size classes were observed in all habitats
(Fig. 9). Certain fish species appeared to use both back-reef and offshore reef habitats as
early juvenile habitat; these species included Lutjanus synagris (lane snapper; Fig. 10),
H. bivitatttus (slippery dick; Fig. 11), and H. sciurus (bluestripped grunt; Fig. 12).
Discussion
In this study we provide an initial assessment of the nursery role (based solely on
daytime fish densities) of multiple, structurally complex tropical backreef habitats for
fish, with comparisons of fish species composition and size-structure to reefs located
offshore. The combination of band transects and 10-min surveys provided a more complete overall species assessment than either method in isolation. Based on visual surveys,
mangrove habitats in the KWNWR contained the highest relative mean density and
diversity of fishes, with abundant forage fish such as silverside minnows (Atherinidae)
and herrings (Clupeidae), as well as a high number of piscivores such as L. griseus (gray
snapper) and S. barracuda (barracuda). Thus, based solely on the density criterion for
identifying nursery habitats (Beck et al., 2001), mangroves appear to be the most important backreef nursery habitat in the KWNWR. An important caveat to this conclusion,
191
Figure 8. The size frequency distribution of Gerres cinereus (yellowfin mojarra) observed during
10-min surveys in (A) seagrass; (B) channels; (C) mangroves; and (D) patch reefs for the Lakes
and Marquesas regions pooled. Note different y-axis values.
however, is the nocturnal use of nearby seagrass meadows by mangrove fishes (pers.
obs.). Mangrove prop-roots provide small juvenile fish with an architecturally complex
substrate that provides maximum food availability and minimizes the risk of predation
(Laegdsgaard and Johnson, 2001). Mangrove fishes probably also rely on food in adjacent seagrass meadows during nighttime foraging, such that the nursery role of mangroves is invariably linked to nearby seagrass habitat. Channel habitats contained the
greatest diversity of microhabitats, and contained a relatively high diversity of fish compared to seagrass. Channel habitats typically harbored juvenile snappers (Lutjanidae),
grunts (Haemulidae), and forage fish (Atherinidae). Qualitatively, we observed greater
numbers of relatively large gamefish, as well as rare and threatened species in channel
and mangrove habitats than any other habitat. Conversely, seagrass contained higher
fish densities than channels. Increases in the size-frequency of certain species, such as
S. barracuda, P. arcuatus, and G. cinereus, from backreef habitats such as seagrass and
mangroves, to channels and eventually patch and offshore reefs were suggestive of ontogenetic patterns of habitat use. In contrast, the smallest stages of L. griseus were found
exclusively in seagrass, but remaining size classes, including adults, were found at all of
the habitats surveyed. In contrast, the smallest size classes of H. bivitattus, L. synagris,
and H. sciurus were found in nearly all of the habitats examined.
Potential Sampling Biases.—
Biases.—We used visual survey techniques to rapidly assess
fish distribution and abundance patterns in the Lakes and Marquesas regions over a 3mo period and across a broad range of marine habitats (seagrass, channels, mangroves,
192
Figure 9. The size frequency distribution of Lutjanus griseus (gray snapper) observed during 10min surveys in (A) seagrass; (B) channels; (C) mangroves; (D) hardbottoms; and (E) patch reefs
for the Lakes and Marquesas regions pooled. Note different y-axis values.
hardbottoms, patch reefs, offshore reefs). We recognize that some of the differences observed between fish abundance and diversity and the six habitats surveyed may be due to
differences in the efficiency of our visual survey methods across habitats, as well as area
searched. Poor water visibility and the cryptic nature of certain fish species in complex
benthic habitats can reduce the accuracy and precision of visual survey techniques. For
example, very small stages (0–5 cm TL) of serranids and sparids were absent from our
10 min surveys, but present in our band transects, in which divers spent more time slowly
searching through seagrass and crevice habitats than in 10-min surveys. Moreover, some
fish fled divers before they could be identified or their size estimated. We tried to reduce
habitat- and observer-specific biases in our visual survey techniques through the use of
1) divers with experience in identifying coral reef fishes; 2) replicate surveys within a
given sampling cell or site; 3) slow and methodical searches in complex benthic habitats;
193
Figure 10. The size frequency distribution of Lutjanus synagris (lane snapper) observed during
10-min surveys in (A) seagrass; (B) channels; (C) hardbottoms; and (D) patch reefs for the Lakes
and Marquesas regions pooled. Note different y-axis values.
4) a complementary combination of band transect and timed surveys in many habitat
types; 5) waiting for fish that initially fled an area to return; and 6) conducting visual
surveys when water visibility was high (> 10 m). This use of a complementary combination of band transect and timed surveys likely increased the accuracy with which we
assessed both cryptic species, such as recently settled lutjanids and gerreids, as well as
more transient species such as large snappers, jacks, and barracuda. For example, the
density of relatively small G. cinereus was up to 50 times greater in band transects than
10-min surveys, which was likely due to divers using the 2-m PVC-pipe that delineated
the 2-m band width to methodically “herd” small fish from within the interstices of the
seagrass canopy. Conversely, transient S. barracuda were rarely observed during band
transects in seagrass, but were commonly observed during 10-min surveys. In a related
study, Schmitt et al. (2002) employed a combination of roving diver surveys and band
transects to assess coral reef fish assemblages off southeastern Hispanola. Roving diver
surveys involved a diver swimming around a reef site for ~45–60 min, recording all fish
species observed (Schmitt et al., 2002). Although roving diver surveys did not provide
estimates of fish density, they did provide a rapid assessment of fish species presence and
absence, and when combined with transect surveys, provided a more complete assessment of fish assemblages than either survey technique alone (Schmitt et al., 2002).
Accurate measures of fish density for our 10-min surveys depended on accurate estimates of distance surveyed. We assessed the accuracy of our visual estimates of distance
traveled in seagrass and mangrove habitats, and although we slightly overestimated and
194
Figure 11. The size frequency distribution of Halichoeres bivitattus (slippery dick wrasse) observed during 10-minute surveys in (A) seagrass; (B) channels; (C) mangroves; (D) hardbottoms;
(E) patch reefs; and (F) offshore reefs for the Lakes and Marquesas regions pooled. Note different
y-axis values.
underestimated distance traveled in seagrass and mangroves, respectively, our estimates
of distance traveled did not differ significantly from the actual distances measured using a differential GPS system on a boat (seagrass) or a tape measure (mangroves). Thus,
our visual estimates of distance traveled during 10-min surveys were relatively accurate.
The efficiency of our visual survey methods was likely similar to or higher than trawling
in seagrass (e.g., 16–69% efficient; Kjelson and Colby, 1977), or block-net and rotenone
sampling methods in mangroves (~75%; Thayer et al., 1987), but may have been lower
than the use of drop-nets in seagrass (Nagelkerken et al., 2000).
There are fundamental sources of bias when comparing animal diversity across different habitat types when search area is different, or different areas are searched within a
given habitat. We tried to reduce bias associated species/area relationships (Rosenzweig,
195
Figure 12. The size frequency distribution of Haemulon sciurus (bluestripped grunt) observed
during 10-min surveys in (A) seagrass; (B) channels; (C) mangroves; (D) hardbottoms; (E) patch
reefs; and (F) offshore reefs for the Lakes and Marquesas regions pooled. Note different y-axis
values.
1995) by not making statistical comparisons of fish diversity across habitat types, and by
standardizing area searched for a given habitat type according to its relative availability
when making comparisons between regions. For example, we surveyed ~3 times more
channels in the Lakes than the Marquesas because there was ~3 times greater areal cover
of channels in the Lakes than the Marquesas. Qualitatively, it does not appear that area
searched was a critical determinant of fish diversity in this study given that the habitat
with one of the lowest areas searched, mangroves, had much higher diversity than habitats with much greater areas searched (e.g., seagrass and channels).
Habitat Characteristics.—
Characteristics.—The Lakes and Marquesas regions within the KWNWR contain a relatively large and diverse mosaic of habitats for fishes (this study) and
Caribbean spiny lobster (Eggleston and Dahlgren, 2001). The use of aerial photographs
was critical in (1) designing our field surveys that targeted the most conspicuous habitat
196
types in the KWNWR, (2) efficiently navigating this shallow region, and (3) determining
samples sizes within a given habitat and region based on the areal cover of the habitat.
The major limitation to relying on the aerial photographs was the inability to see patch
reefs and hardbottom habitats. We overcame this limitation by hiring local guides at the
initiation of the study. These local guides were especially helpful in terms of identifying
hardbottom “lobster holes” not indicated on navigational charts.
We found significantly higher fish densities at the Marquesas than the Lakes regions.
The reason for the higher fish densities in the Marqueasas than the Lakes is unclear, but
may have been due to a combination of factors including, but not limited to: (1) higher
−
seagrass density at the Marquasas (31.3 shoots 0.07 m−2
) than the Lakes (22 shoots 0.07
−
−2
m ) (although the trend was not significant and there was 81% greater seagrass areal
cover at the Lakes), (2) 40% more mangrove perimeter at the Marquesas than the Lakes,
and (3) relatively close proximity of the Lakes to Key West, which may have resulted in
poorer water quality and greater fishing pressure than in the Marquesas.
We did not observe any obvious areas of seagrass or sponge die-off, as indicated by
denuded mud patches, as has been seen in nearby Florida Bay (Robblee et al., 1991; Durako, 1994; Butler et al., 1995; Herrnkind et al., 1997). Seagrass percent cover was high
(75–100%) in areas outside of macroalgal beds, with relatively high average shoot densi−
ties (467–500 shoots m−2
) and blade heights (30–37 cm), compared to similar measurements in Florida Bay (Durako, 1994), Mexico (Gallegos et al., 1993), and the Bahamas
(Eggleston, 1995).
Density and Diversity of Fish in Back-Reef Habitats.—
Habitats.—The density of fish in
seagrass was similar to that in mangroves when atherinids were removed; however, the
diversity of fish in mangrove fringe and channel habitats was greater than in adjacent
seagrass meadows. Overall, the fish species present in the Lakes and Marquesas appeared somewhat transitional between the more subtropical estuarine environment of
Florida Bay basins, and the tropical oceanic environment of the Bahamas and Caribbean. For example, in seagrass, we tended to find greater numbers of reef fish species
(e.g., haemulids and lutjanids) than did Sogard et al. (1987) in Florida Bay, whereas they
found greater abundances of more estuarine fish from the families Cyprinodontidae
(killifishes) and Batrachoididae (toadfishes). Similarly, Thayer et al. (1989) working in
seagrass, found greater numbers of more estuarine fish such as Cyprinodontidae, Batrachoididae, as well as spotted seatrout Cynoscion nebulosus (Cuvier in Cuvier and Valenciennes, 1830) and pinfish Lagodon rhomboids (Linnaeus, 1766), and fewer species
of reef fish than our study. The fish assemblages observed by Ley et al. (1999) working
along an estuarine salinity gradient in Florida Bay mangrove prop-roots were dominated
by the families Engraulidae and Atherinidae, which was similar to this study; however,
they found higher numbers of more estuarine fishes such as Poeciliidae (mosquitofish)
and Cyprinodontidae than we did. Conversely, we identified greater numbers of more
estuarine species in mangrove habitats than did Rooker and Dennis (1991), who worked
in mangroves in Puerto Rico that were in close proximity to coral reefs.
Our visual estimates of fish density (converted to fish ha−1 for comparative purposes)
in seagrass in the Lakes and Marquesas using band transects ranged from a low of ~3 ×
103 fish ha−1 in the Lakes to a high of ~2.5 × 104 fish ha−1 in the Marquesas. Our fish density estimates are within the range of those reported from other tropical and subtropical
seagrass systems. For example, using otter trawls in Florida Bay seagrass beds and channels, Thayer and Chester (1989) reported an average density of 2 × 103 fish ha−1. Using
throw-traps in Florida Bay, Sogard et al. (1987) observed highest mean densities of 11
197
× 104 fish ha−1 in seagrass bank habitats in Florida Bay. Adams (1976) used throw traps
in eelgrass (Zostera marina L.) beds in North Carolina and found an average density
of 1.8 × 104 fish ha−1, and Weinstein and Brooks (1983) found densities <103 fish ha−1 in
Chesapeake Bay seagrass beds using an otter trawl.
Ontogenetic Habitat Shifts.—Ontogenetic habitat shifts from back-reef to offshore reefs were observed for some species. One of the clearest examples of an ontogenetic habitat shift was S. barracuda, in which the smallest size class was found exclusively
in mangroves, and the largest size class in all habitats. A similar ontogenetic habitat
shift was reported for S. barracuda in Bonaire, Netherlands Antilles (Nagelkerken et
al., 2000). Other species that showed relatively strong evidence for ontogenetic habitat
shifts, particularly from seagrass and channels to reefs, included G. cinereus and P. arcuatus. Conversely, the smallest size class of L. griseus occurred exclusively in seagrass,
with adult stages occupying all habitats. A similar pattern was observed for L. griseus in
Bonaire (Nagelkerken et al., 2000).
The fish families and species (Appendix 1) identified during patch reef and offshore
reef surveys in this study were similar to those reported from elsewhere in the Florida
Keys (Bohnsack and Bannerot, 1986; Ault et al., 1998) and Dry Tortugas (Rydene and
Kimmel, 1995). For example, Bohnsack and Bannerot (1986) reported a total of 117 species of fish observed in 160 visual surveys from the forereef at Looe Key, Florida. Their
species assemblages were dominated by haemulids, labrids, pomacentrids, and acanthurids (Bohnsack and Bannerot, 1986). The jacks (Carangidae) comprised ~10% of the
reef fish in this study, but less than 5% off of Looe Key (Bohnsack and Bannerot 1986).
Although bluehead wrasse Thalassoma bifasciatum (Bloch, 1791) was the predominant
labrid at Looe Key (Bohnsack and Bannerot, 1986), the slippery dick (H. bivittatus) was
the most common labrid in this study.
In conclusion, mangroves (based on 10-min surveys) and seagrass (based on band
transects) appear to be key nursery habitats for fishes, particularly given that seagrass
consistently harbored the smallest size classes of many species, and mangroves contained the highest density and diversity of fishes. Although channels containing large
sponges represented only 0.06% of the total area of the Lakes and Marquesas, they contained the highest diversity of microhabitats, and a relatively high diversity and density
of fishes. Moreover, we qualitatively observed more sharks, gamefish (e.g., tarpon), and
turtles in channels than any other habitat. Channels also provide a likely corridor for
fishes migrating from back reef habitats such as seagrass and mangroves to patch reef
and offshore reefs (Parrish, 1989). Thus, inclusion of seagrass, mangrove, and channel
habitats in future studies of reef fish growth, survival, and emigration should produce
a more complete picture of their nursery role (sensu Beck et al., 2001) in tropical back
reef environments.
Acknowledgements
We thank the following people for their expert assistance in the field: L. Etherington, T. Kellison, N. Reyns, S. Searcy, A. Drew, D. Nadeau, M. Darcy, and D. Blackmon. We are especially
grateful to L. Etherington for assembling the sampling equipment, S. White for assistance with
initial field reconnaissance and Key West logistics, B. Lockwood for logistics in the Great White
Heron National Wildlife Refuge, the U.S Fish and Wildlife Service for use of their boats, T. McKellar for help in producing the figures of the study sites, and J. Sobel for facilitating the research
funding. Funding for this project was provided by a Challenge Cost-Share Agreement between
the Center for Marine Conservation and U.S. Fish and Wildlife Service for Contract 1448-40181-
198
99-6, the National Science Foundation (OCE 99-86567), North Carolina State University, the Caribbean Marine Research Center (CMRC Project # CMRC-00-IXNR-03-01A), National Oceanic
and Atmospheric Administration (NOAA) National Undersea Research Program, U.S. Environmental Protection Agency, and Environmental Defense. Views expressed herein are those of the
authors and do not necessarily reflect the views of CMRC, or any of the supporting agencies.
Literature Cited
Adams, S. M. 1976. The ecology of eelgrass, Zoostera marina (L.), fish communities: II. Functional analysis. J. Exp. Mar. Biol. Ecol. 22: 293–311.
Ault, J. S., J. A. Bohnsack, and G. A. Meester. 1998. A retrospective (1979-1996) multi-species
assessment of coral reef fish stocks in the Florida Keys. U.S. Fish. Bull. 96: 395–414.
Beck, M. W., K. L. Heck, Jr., K. Able, D. Childers, D. Eggleston, B. Gillanders, B. Halpern, C.,
Hays, K Hoshino, T. Minello, R. Orth, P. Sheridan, and M. Weinstein. 2001. Towards better
identification, conservation and management of estuarine and marine nurseries. BioScience
51: 633–641.
Bohnsack, J. A. and S. P. Bannerot. 1986. A stationary visual censusing technique for quantitatively assessing community structure of coral reef fishes. NOAA Tech. Report, NMFS 41, 15 p.
Butler, M. J., IV, 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.
Carr, M. H. 1994. Effects of macroalgal dynamics on recruitment of a temperate reef fish. Ecology
75: 1320–1333.
Dahlgren, C. P. and D. B. Eggleston. 2000. Ecological processes underlying ontogenetic habitat
shifts in a coral reef fish. Ecology 81: 2227–2240.
____________ and _____________. 2001. Spatio-temporal variability in abundance, size and microhabitat associations of early juvenile Nassau grouper Epinephelus striatus in an off-reef
nursery system. Mar. Ecol. Prog. Ser. 217: 145–156.
Durako, M. J. 1994. Seagrass die-off in Florida Bay (USA): changes in shoot demographic characteristics and population dynamics in Thalassia testudinum. Mar. Ecol. Prog. Ser. 110: 59–66.
Eggleston, D. B. 1995. Recruitment in Nassau grouper Epinephelus striatus: post-settlement abundance, microhabitat features, and ontogenetic habitat shifts. Mar. Ecol. Prog. Ser. 124: 9–22.
______________, R. N. Lipcius, and J. Grover. 1997. Predator and shelter effects on coral reef fish
and spiny lobster prey. Mar. Ecol. Prog. Ser. 149: 43–59.
______________ and C. P. Dahlgren. 2001. Distribution and abundance of Caribbean spiny lobsters in the Key West National Wildlife Refuge: relationship to habitat features and impact on
an intensive recreational fishery. Mar. Freshwater Res. 52: 1567–1576.
Gallegos, M. E., M. Merino, N. Marba, and C. M. Duarte. 1993. Biomass and dynamics of Thalassia testudinum in the Mexican Caribbean: elucidating rhizome growth. Mar. Ecol. Prog. Ser.
95: 185–192.
Helfman, G. S., J. L. Meyer, and W. N. McFarland. 1982. The ontogeny of twilight migration patterns in grunts (Pisces: Haemulidae). Anim. Behav. 30: 317–326.
Herrnkind, W. F., M. J. Butler, IV, J. H. Hunt, and M. Childress. 1997. Role of physical refugia:
implications from a mass sponge die-off in a lobster nursery in Florida. Mar. Freshwater Res.
48: 759–769.
Kasai, A., W. Sakamoto, Y. Mitsunaga, and S. Yamamoto. 2000. Behavior of immature yellowtails (Seriola quinqueradiata) observed by electronic data-recording tags. Fish. Oceanogr. 9:
259–270.
Kjelson, M. A. and G. N. Johnson. 1978. Catch efficiencies of a 6.1-meter otter trawl for estuarine
fish populations. Trans. Am. Fish. Soc. 107: 246–254.
Laegdsgaard, P. and C. Johnson. 2001. Why do juvenile fish utilize mangrove habitats? J. Exp.
Mar. Biol. Ecol. 257: 229–253.
199
Lapoint, B. E. and M. W. Clark. 1992. Nutrient inputs from the watershed and coastal eutrophication in the Florida Keys. Estuaries 15: 465–476.
Levin, P. S. and M. E. Hay. 1996. Responses of temperate ref fishes to alterations in algal structure
and species composition. Mar. Ecol. Prog. Ser. 134: 37–47.
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. Est., Coast. Shelf
Sci. 48: 701–723.
Morton, R. M. 1990. Community structure, density and standing crop of fishes in a subtropical
Australian mangrove area. Mar. Biol. 105: 385–394.
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 nursery for important coral reef fishes, using visual census technique. Est. Coast. Shelf Sci. 51: 31–44.
NMFS. 1996. Our living oceans: report of the status of U.S. living marine resources, 1995. U.S.
Department of Commerce, NOAA Tech. Memo. NMFS-F/SPO-19, 160 p.
Robblee, M. B., T. R. Barber, P. R. Carlson, M. J. Durako, J. W. Fourqurean, L. K. Muehlstein,
D. Portoer, L. A. Yarbro, R. T. Zeiman, and J. C. Zieman. 1991. Mass mortality of the tropical
seagrass Thalassia testudinum in Florida Bay (USA). Mar. Ecol. Prog. Ser. 71: 297–299.
Rooker, J. R. and G. D. Dennis. 19991. Diel, lunar and seasonal changes in a mangrove fish assemblage off southwestern Puerto Rico. Bull. Mar. Sci. 49: 684–698.
Rosenzweig, M. L. 1995. Species diversity in space and time. Cambridge University Press, Cambridge. 436 p.
Rydene D. A. and J. J. Kimmel. 1997. A five-year assessment (1990-1994) of coral reef fish assemblages within Dry Tortugas National Park, Florida, using visual censusing techniques. FL
Mar. Res. Inst. Report. 23 p. and append.
Schmitt, E. F., R. D. Sluka, and K. M. Sullivan-Sealey. 2002. Evaluating the use of roving diver
and transect surveys to assess the coral reef fish assemblage off southeastern Hispaniola. Coral
Reefs 21: 216–223.
Sogard, S. M., G. V. N. Powell, and J. G. Holmquist. 1987. Epibenthic fish communities on Florida
Bay banks: relations with physical parameters and seagrass cover. Mar. Ecol. Prog. Ser. 40:
25–39.
Stoner, A. W. 1983. Distribution of fishes in seagrass meadows: role of macrophyte biomass and
species composition. U. S. Fish. Bull. 81: 837–846.
Thayer, G. W. and A. J. Chester. 1989. Dsitribution and abundance of fishes among basin and channel habitats in Florida Bay. Bull. Mar. Sci. 44: 200–219.
___________, 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.
Weinstein, M. P. and H. A. Brooks. 1983. Comparative ecology of nekton residing in a tidal creek
and adjacent seagrass meadow: community composition and structure. Mar. Ecol. Prog. Ser.
12: 15–27.
_____________ and K. L. Heck, Jr. 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.
Addresses: (D.B.E., E.G.J.) North Carolina State University, Department of Marine, Earth and
Atmospheric Sciences, Raleigh, North Carolina 27695-8208. (C.P.D.) The Oceans Conservancy,
1725 DeSales St., NW, Suite 600, Washington, D.C. 20036. Present Address: Perry Institute
for Marine Science/Caribbean Marine Research Center, 100 North Highway 1, Jupiter, Florida
33477. Corresponding Author: (D.B.E.) E-mail: <eggleston@ncsu.edu>.
S1
S2
S3
S4
S5
S6
S7
S8
S9
S10
S11
S12
S13
S14
S15
S16
S17
S18
S19
S1
S2
S3
S4
S5
S6
Lakes
Marquesas
Station
Region
31.0
41.5
22.0
43.0
18.0
N/A
37.0
12.0
43.0
20.5
21.0
21.0
12.5
N/A
36.5
19.0
27.5
15.5
18.5
40.5
40.5
43.0
24.0
16.5
23.0
Shoot Ct
25.8
42.4
31.4
47.7
27.0
N/A
31.7
23.6
36.1
18.1
19.3
39.6
16.9
N/A
41.5
15.5
26.9
25.4
29.5
27.3
45.1
38.5
33.2
34.4
40.8
Blade Ht
99.3
99.7
54.0
99.5
17.5
14.2
68.3
75.8
85.8
21.7
15.0
100.0
75.0
5.0
99.2
50.0
97.0
99.2
93.8
99.0
96.2
94.0
94.7
95.5
100.0
Thalassia
0.0
0.0
0.0
0.5
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
Syringodium
0.0
0.0
0.0
0.0
0.0
0.0
0.0
8.3
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
Halodule
S1
S2
S3
S4
S5
S6
S7
S8
S9
S10
S11
S12
S13
S14
S15
S16
S17
S18
S19
S1
S2
S3
S4
S5
S6
Station
31.0
41.5
22.0
43.0
18.0
N/A
37.0
12.0
43.0
20.5
21.0
21.0
12.5
N/A
36.5
19.0
27.5
15.5
18.5
40.5
40.5
43.0
24.0
16.5
23.0
Shoot Ct
25.8
42.4
31.4
47.7
27.0
N/A
31.7
23.6
36.1
18.1
19.3
39.6
16.9
N/A
41.5
15.5
26.9
25.4
29.5
27.3
45.1
38.5
33.2
34.4
40.8
% Cover
Blade Ht
0.7
0.0
0.0
0.0
17.5
26.7
2.5
13.0
0.0
6.7
22.5
0.0
0.0
5.0
1.0
0.0
0.0
0.0
0.0
0.0
0.0
1.0
0.0
0.0
0.0
Laurencia
Other
Macroalgae
0.0
0.3
25.0
0.0
20.5
7.4
22.5
2.9
10.7
29.1
40.0
0.0
11.7
40.0
0.2
15.1
0.5
0.3
6.2
0.2
0.0
4.8
3.3
0.0
0.0
0.0
0.0
0.0
0.0
3.7
0.0
0.0
0.0
0.0
3.3
5.0
0.0
0.0
0.0
0.0
1.7
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
Sponge
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
2.5
2.5
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
Coral rubble
0.0
0.0
21.0
0.0
40.8
51.7
6.7
0.0
3.5
36.7
15.0
0.0
13.3
50.0
0.0
33.2
2.5
0.5
0.0
0.8
3.8
0.2
2.0
4.5
0.0
Sand
Appendix 1. Habitat characteristics from seagrass habitats within the Lakes region of the KWNWR. Shoot counts and blade heights are 0.07 m−2. Percent cover is 120 m−2. Counts represent the average of two band
transects within a sampling station (grid cell). N/A indicates that measurements were not taken.
200
3.0
4.5
3.5
N/A
25.5
N/A
29.0
0.0
N/A
0.5
2.5
2.5
83.0
99.0
33.0
33.5
8.0
0.0
N/A
3,308.4
N/A
N/A
3,348.9
N/A
10,474.5
N/A
N/A
N/A
105.2
24,875.2
46,352.7
32,803.1
717.1
33,283.0
17,144.1
0.0
20.8
0.0
5.0
0.0
15.0
3.3
21.7
44.2
66.7
61.7
26.3
15.0
8.3
3.3
5.8
15.0
13.3
16.7
39.2
0.0
0.0
0.0
0.0
0.0
0.0
10.0
0.0
0.8
0.0
0.0
0.0
0.0
16.7
40.0
26.7
46.6
0.0
0.0
0.0
0.0
0.0
0.0
0.0
1.7
0.0
0.0
2.5
0.0
0.0
0.0
70.5
20.8
0.0
0.0
0.0
1.8
0.0
37.5
14.2
43.8
18.2
0.0
0.0
0.0
6.3
29.2
38.9
0.0
0.0
12.2
0.0
0.0
Station Sponge count Sponge volume Thalassia Syringodium Halodule Laurencia
C1
C2
C3
C4
C5
C6
C7
C8
C9
C10
C11
C12
C13
C14
Marquesas C1
C2
C3
C4
Lakes
Region
C1
C2
C3
C4
C5
C6
C7
C8
C9
C10
C11
C12
C13
C14
C1
C2
C3
C4
Station
3.0
4.5
3.5
N/A
25.5
N/A
29.0
0.0
N/A
0.5
2.5
2.5
83.0
99.0
33.0
33.5
8.0
0.0
% Cover
Sponge
count
N/A
3,308.4
N/A
N/A
3,348.9
N/A
10,474.5
N/A
N/A
N/A
105.2
24,875.2
46,352.7
32,803.1
717.1
33,283.0
17,144.1
0.0
Sponge
volume
19.2
64.0
75.0
50.0
18.5
51.9
28.4
23.3
3.3
7.5
13.6
15.0
14.8
56.7
2.3
2.7
45.0
0.0
Other
Macroalgae
0.0
0.0
0.0
0.0
0.5
0.0
0.0
0.0
0.0
0.0
0.0
0.0
24.4
18.3
0.0
0.8
5.0
0.0
Sponge
0.0
7.5
0.0
12.5
21.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
Coral
rubble
20.8
26.7
20.0
0.0
30.8
1.0
31.7
20.8
30.0
30.0
51.3
40.8
13.6
21.7
4.7
8.5
10.0
36.7
Sand
Appendix 2. Habitat characteristics (120 m−2) from channel habitats within the Lakes region of the KWNWR. Counts represent the average of two band transects within a sampling
station. N/A indicates that measurements were not taken.
201
202
Appendix 3. List of fish families and species observed during band transects and timed surveys at the Lakes and Marquesas regions of the KWNMR during July–October, 1999. Habitat
codes are: C (channels), H (hardbottoms), M (mangroves), PR (patch reefs), R (offshore reefs)
and S (seagrass).
Family
Acanthuridae
Atherinidae
Balisitidae
Batrachoididae
Belonidae
Blennidae
Carangidae
Carcharhinidae
Centropomidae
Chaetodontidae
Clinidae
Clupeidae
Cyprinidae
Dasyatidae
Echeneidae
Elopidae
Ephippidae
Exocoetidae
Fundulidae
Gerreidae
Gobiidae
Haemulidae
Species
Acanthurus bahianus
Acanthurus chirurgus
Acanthurus coeruleus
Acanthurus sp.
sp.
Balistes capriscus
Monacanthus tuckeri
Opsanus sp.
sp.
sp.
Caranx bartholomaei
Caranx crysos
Caranx latus
Caranx ruber
Caranx sp.
Negaprion brevirostris
Centropomus undecimalis
Chaetodon capistratus
Chaetodon ocellatus
Chaetodon striatus
Malococtenus sp.
sp.
sp.
Dasyatis americana
Ecenies sp.
Megalops atlanticus
Chaetodipterus faber
Hemirhamphus brasiliensis
Fundulus sp.
Eucinostomus melanopterus
sp.
Coryphopterus glaucofraenum
Gobiosoma oceanops
Ioglossus sp.
Anisostremus virginicus
Haemulon aurolineatum
Haemulon flavolineatum
Haemulon macrostomum
Haemulon parra
Haemulon plumieri
Haemulon sciurus
Haemulon sp.
Habitats of occurrence
C, PR, R
C, H, PR, R
C, PR, R
M, S
C
R, S
C
M, S
C, PR, R
H, S
M, PR, R
PR
C, H, PR, R, S
PR
C
M
C, PR, R, S
C, PR, R
C, PR, R
C, R
M, PR, S
M, S
C, PR
R
M, PR, C
PR
S
H, S
S
S
C, H, PR, S
PR, R
C, S
C, H, M, PR, R, S
C, M, PR, R, S
C, H, M, PR, R, S
M, PR, S
C, H, M, PR, R
C, H, M, PR, R, S
C, H, M, PR, R, S
C, M, PR, S
Appendix 3. Continued.
Family
Holocentridae
Kyphosidae
Labridae
Lutjanidae
Mullidae
Ostraciidae
Pomacantidae
Pomacentridae
Rhincodontidae
Scaridae
Species
Holocentrus adscensionis
Holocentrus marianus
Kyphosis sectatrix
Bodianus rufus
Halichoeres garnoti
Halichoeres maculipinna
Halichoeres poeyi
Halichoeres radiatus
Lachnolaimus maximus
Thalassoma bifasciatum
Lutjanus griseus
Lutjanus analis
Lutjanus apodus
Lutjanus cyanopterus
Lutjanus jocu
Lutjanus mahogoni
Lutjanus synagris
Ocyurus chrysurus
Mulliodichthys martinicus
Opisthognathus aurifrons
Lactophrys quadricornis
Holacanthus bermudensis
Holacanthus ciliaris
Pomacanthus arcuatus
Pomacanthus paru
Pomacantus sp.
Abudefduf saxatilis
Chromis cyanea
Chromis multilineata
Microspathadon chrysurus
Stegastes fuscus
Stegastes partitus
Steagastes planifrons
Stegastes sp.
Stegastes variabilis
Ginglymostoma cirratum
Scarus sp.
Scarus coeruleus
Scarus croicensis
Scarus guacamaia
Scarus taeniopterus
Scarus vetula
Sparisoma aurofrenatum
Sparisoma chrysopterum
Sparisoma radians
Sparisoma rubripinne
Sparisoma viride
R
R
C, M, PR, R
R
C, H, M, PR, R, S
R
R
R
C, PR, R
C, H, PR, R, S
C, R
C, H, M, PR, S
C, H, M, PR, S
C, H, M, PR, R, S
M
M
R
C, H, PR, S
C, PR, R, S
R
R
PR
C, PR
C, PR, S
C, PR, R, S
C, H, PR, S
C
C, H, M, PR, R
R
R
PR, R
C, H, PR, R
C
C, H PR, R, S
R
PR, R
C, PR, R
C
C, R, S
H, R
C, PR, R, S
C, M, PR
C, H, PR, R, S
M, R
C, H, PR, R, S
C, H, M, PR, R, S
C, H, PR, S
C, M, PR, R
C, H, M, PR, R, S
203
204
Appendix 3. Continued.
Family
Sciaenidae
Serranidae
Serranidae
Sparidae
Sphyraenidae
Syngnathidae
Synodontidae
Tetraodontidae
Urolophidae
Species
Equetus acuminatus
Equetus umbrosus
Odontoscion dentex
Sciaenops ocellatus
Diplectrum formosum
Epinephelus adscenionis
Epinephelus cruentatus
Epinephelus itajara
Epinephelus morio
Epinephelus striatus
Hypoplectrus puella
Hypoplectrus sp.
Mycteroperca bonaci
Mycteroperca phenax
Rypticus maculatus
Serranus tabacarius
Serranus tigrinus
Archosargus probatocephalus
Archosargus rhomboidales
Calamus bajonado
Calamus penna
Lagodon rhomboidales
Sphyraena barracuda
sea horse sp.
Synodus sp.
Canthigaster rostrata
Diodon hystrix
Sphoeroides spengleri
Sphoeroides testudinus
Urolophus jamaicensis
C, PR, S
PR, R
PR, R
M
C, H, S
PR
PR
M
C, PR
C, H
C, PR
C, PR, R, S
C, M, PR, R
PR
C, M, PR, R
R
R
PR
C, H, M, PR, S
C, PR, R
H
C, M , S
R
S
S
R
PR
C
C
C, PR, R, S
HABITAT ASSOCIATIONS OF ADULT QUEEN CONCH
(STROMBUS
STROMBUS GIGAS L.) IN AN UNFISHED FLORIDA KEYS
BACK REEF: APPLICATIONS TO ESSENTIAL FISH HABITAT
Robert A. Glazer and James A. Kidney
ABSTRACT
The identification of essential fish habitat (EFH) is critical for the effective management of an aquatic species. In Florida, the fishery for queen conch, Strombus gigas L.,
has been closed since 1986 due to declines attributed in part to habitat degradation and
overfishing. To examine the habitat requirements of adult conch, 39 conch were tagged
with acoustic transmitters and tracked from June 1997 through June 1998 within two
back reef sites in the Florida Keys. Sampling occurred biweekly and consisted of recording each conch’s habitat, geospatial position, and reproductive activity. The area of
the available habitat and the frequencies of observations in each habitat were used to estimate habitat utilization, selection, and preference. At one site, conch showed a strong
preference for the coarse-sand and rubble/coarse-sand habitats and avoided rubble. At
the other site, no habitat was preferred or avoided. Actively reproducing conch avoided
seagrass at one site and rubble at the other. In general, conch selected the coarse-sand
habitat for reproduction. It is likely that the harvest moratorium influenced the spatial
distribution of conch. We suggest that EFH for Florida’s conch be considered at a scale
that incorporates a mosaic of habitats supporting a range of biological functions in the
back reef.
This publication is part of a series of papers resulting from a scientific workshop held
The analysis of habitat utilization and preference is commonly employed in resource
management to conserve exploited populations of fish and wildlife. In 1996, the U.S.
Congress amended the Magnuson-Stevens Fishery Conservation and Management Act
recognizing that “one of the greatest long-term threats to the viability of commercial
and recreational fisheries is the continuing loss of marine, estuarine, and other aquatic
habitats. Habitat considerations should receive increased attention for the conservation
and management of fishery resources of the United States” (16 U.S.C. 1801 (A)(9)).
Within this context, the U.S. Congress defined Essential Fish Habitat (EFH) as “those
waters and substrate necessary to fish for spawning, breeding, feeding, or growth to
maturity.” The legislation mandates that all federal fishery management plans identify
habitat needs of, and address human impacts upon, regulated species so that they can
be managed effectively. By necessity, the identification of EFH for a species requires
a comprehensive understanding of the biological and ecological requirements of that
organism (Hartwell, 1998).
The waters of the Florida Keys support coral, seagrass, and other communities associated with the greater Caribbean region. These habitats are increasingly impacted
by nutrification, physical disturbances including ship groundings, commercial interests,
and extensive recreational activities (Department of Commerce, 1996). Recognizing the
fragile nature of this environment, the Florida Keys National Marine Sanctuary was
designated by the U.S. Congress in 1990 to protect, conserve, and sustain the resources
found within its boundaries.
205
206
The queen conch is a large, commercially valuable gastropod found throughout the
Caribbean bioprovince. Although queen conch still support an important commercial
fishery, overfishing and habitat loss have been implicated in a region-wide decline in
their abundance (Berg and Olsen, 1989; Appeldoorn, 1994) resulting in their listing in
Appendix II with the Convention on the International Trade in Threatened and Endangered Species of Flora and Fauna (CITES).
In Florida, queen conch are found almost exclusively within the boundaries of the
Florida Keys National Marine Sanctuary. A decline in conch abundance during the
1980s was attributed both to overfishing and to habitat loss resulting from coastal development (Berg and Glazer, 1995; Glazer and Quintero, 1998), and since 1986, the collection of live conch has been prohibited (Florida Administrative Code Chapter 68B-16).
The biology of conch has been well described. During the summer months, adults
aggregate in shallow water to spawn (Randall, 1964; D’Asaro, 1965; Brownell, 1977),
making them vulnerable to exploitation. Female conch lay egg masses (Davis et al.,
1984) and the eggs hatch into veligers. After ~3 wk (Davis et al., 1993) they settle on
suitable substrate (Stoner et al., 1996b) where they are part of the infaunal community
(Iverson et al., 1986). Approximately 3.5–4 yrs later, they become sexually mature (Appeldoorn, 1988).
The ecological requirements of adult queen conch have been documented throughout
their range, at least qualitatively. In general, they occupy shallow seagrass beds (Randall, 1964; Stoner, 1994; Stoner and Schwarte, 1994; de Jesús et al., 1999; Delgado,
1999) although they are also found occupying rubble habitats (Alcolado, 1976; Stoner
and Schwarte, 1994). Queen conch have been observed in water as deep as 25 m (Rathier, 1993; Stoner and Schwarte, 1994; Mateo et al., 1998). They are herbivores that
feed on algae (Robertson, 1961). Reproduction has been documented in sand habitats
(Robertson, 1959; Randall, 1964; Stoner and Schwarte, 1994) although there are also
reports of conch reproducing in seagrass (Weil and Laughlin, 1984).
Seasonal migrations of adult conch have been reported in several locations. In the
Bahamas, conch were observed migrating from the food rich rubble community to sand
habitats for reproduction (Stoner and Sandt, 1992). In the Turks and Caicos, adult conch
moved from a seagrass dominated community to a sand-algal community associated
with the onset of winter (Hesse, 1979).
In Florida, conch have been observed more commonly in shallow-water, hard-bottom habitats adjacent to land, and in back reef habitats associated with offshore reefs
and characterized by rubble and coarse sediment (Glazer and Berg, 1994). Conch inhabiting the nearshore locations have never observed reproducing; reproduction occurs
exclusively in the back reef zone adjacent to the reef crest and in deeper waters further
offshore (McCarthy et al., 2000).
Management strategies for conch stocks have recognized the need to identify and
conserve habitat in Florida (Berg and Glazer, 1995) and the wider Caribbean (Mahon,
1990). The queen conch fishery management plan for the U.S. Virgin Islands and Puerto
Rico addressed critical habitats as required by law (Caribbean Fisheries Management
Council, 1996).
We used acoustic telemetry to quantitatively examine the habitat associations of adult
queen conch, Strombus gigas L., in a Florida Keys back reef community over a 1-yr
period. This study focused only on the reproductive conch population found offshore. It
represents the first quantitative evaluation of habitat utilization and preference for adult
queen conch.
GLAZER AND KIDNEY: HABITAT ASSOCIATIONS — QUEEN CONCH
207
Site Description
This study was conducted at Conch Reef in the upper Florida Keys (Fig. 1). Conch Reef is a
bank reef with well-defined fore-reef and back reef areas and is representative of the intermittent,
shoal-water bank reefs along the Florida Keys reef tract. For the purposes of this study, we are
defining the back reef zone as an area that runs the length of the Florida Keys, ~1500 m wide,
situated immediately to the lee of the reef crest encompassing rubble, seagrass, and sediment
habitats. Within the back reef, the depth is typically 1–10 m (Fig. 1).
Conch were tagged and released at two sites where reproduction has been commonly observed
(Florida Fish and Wildlife Conservation Commission, unpubl. data). Because conch harvest has
been prohibited, these aggregations may be considered unfished, although poaching still occurs
Figure 1. Sampling sites (C1 and C2) for queen conch tag-recapture study using acoustic telemetry. Sites are at Conch Reef in the Florida Keys. Water depth (m) is presented. UK represents an
area of unknown depth. Stippling on the Florida map represents the reef crest and is presented
for reference.
208
occasionally (Glazer, pers. obs.). Site C1 was in ~1–4 m of water and was situated immediately
to the northeast of the shallowest zone of emergent reef (Fig. 1). This site contained a mixture
of juvenile and adult conch. In May 1997, we estimated that the density of adult conch was 111
conch ha−1 and the abundance was ~80 individuals (Florida Fish and Wildlife Conservation Commission, unpubl. data). Site C2 was situated ~600 m to the southwest of C1 in deeper-water (2–10
m) to the southeast of the emergent reef zone (Fig. 1). In August 1997, we estimated that the adult
conch density was 13.4 conch ha−1 and the abundance was ~270 individuals (Florida Fish and
Wildlife Conservation Commission, unpubl. data).
Materials and Methods
Spatial Analyses.—In order to determine how conch utilized the habitats that were available, we first constructed a map within a Geographic Information System (GIS: ArcView v.3.1,
ESRI, Inc., Redlands, California, U.S.) that quantitatively defined the available habitats. The
data consisted of a habitat layer that was digitized in our laboratory from 10× enlargements of
1:48,000 scale, high-altitude, conventional color photographs of the Florida Keys reef tract (National Ocean Service, 1991) scanned at 300 dpi with a resolution of 4 m. From this, the available
area of each habitat type at each site was determined in a two-step process.
First, we defined a Minimum Convex Polygon (MCP) at each site. The MCP encompassed the
perimeter defined by the outermost locations of all observations of conch tagged with acoustic
tags and subsequently recaptured at each site (see tagging section below) over the entire time
period of the study (Fig. 2). We used the MCP method because it reduces the subjectivity associated with defining available habitat (Porter and Church, 1987) and, when testing for preference,
reduces the probability of encountering Type I errors (McClean et al., 1998). We estimated the
area of the MCP by using the ArcView extension Animal Movement Analysis (Hooge et al.,
2001). The second step consisted of intersecting the MCP with the underlying habitat layer. The
area (m2) of each habitat type located within each MCP was then calculated. Habitat codes were
consistent with those used in field surveys.
The map was ground-truthed by comparing the habitat occupied by each conch with the habitat
shown on the map. There was a high degree of concordance between the habitat polygons derived
from the aerial photographs and the observed habitats. In some cases, the shape of the seagrass
and sand habitats needed slight adjustments. No additional patches were added to the habitat layer; changes were only made to the shape of polygons constructed from the aerial photographs.
Tagging/Telemetry.—All conch for this study were selected from the representative habitats
at each site. We tagged each conch with an acoustic transmitter that emitted a unique combination of pulse and frequency (68–76 kHz). The tag was bound to the shell’s spire with monel wire
(Fig. 3). A second tag, consisting of a monel plate stamped with a unique identifying number, was
attached to the spire using the same method. We also attached brightly colored flagging tape to
the spire to facilitate locating the conch during recapture surveys. The sex was determined for
each conch at the time of tagging using well-described methods (Randall, 1964). We tagged new
individuals when tagged conch died and their shells were recovered, when tags were lost (i.e., a
tag was found on the bottom without a conch), or when tags failed (i.e., a tag was found attached
to a conch but was emitting no signal).
Recapture surveys were conducted biweekly or as weather conditions permitted from June
1997 through June 1998. We chose this interval to minimize the effect of serial autocorrelation
associated with multiple observations from relatively few animals, a common problem in telemetry studies (Aebischer et al., 1993). To locate each individual, we used a hydrophone and receiver
(Sonotronics, Tucson, Arizona) tuned to a selected frequency. Signals were more difficult to locate in highly rugose habitats thereby necessitating careful and methodical surveys. The signals
were detectable up to 1 km from the tag. Searches were conducted up to 10 km from the location
of tagging to ensure that conch that had traversed significant distances would be located. When
the signal was equally strong in all directions, a snorkeler searched the area until the tagged conch
was located. The position (i.e., latitude and longitude) of the tagged conch was determined using a
209
Global Positioning System receiver and was recorded. We also noted the habitat within which the
conch was observed. Habitats were characterized based on both the substrate and benthos (Table
1). Only data from conch that were directly observed by the snorkeler were used in the analyses.
Statistical Analyses.—For statistical purposes, the sites were treated independently and
not as replicates because of the differences in population structure, available habitat types, and
depth profiles between the two sites. For the purposes of this study, we define habitat selection as
a descriptive term that does not necessarily indicate statistical significance. Habitat preference
and habitat avoidance are used to indicate that there was a statistically significant preference for
or avoidance of the habitat.
Habitat availability was calculated using the simple proportion:
Πi = Ai AMCP
MC ,
(1)
where Πi = the proportion of habitat i, Ai = area (m2) of habitat i, and A MCP = area (m2) of the
MCP.
Habitat utilization was calculated using the Neu index (Neu et al., 1974) as refined by Manly
et al. (1993):
oi = ui ut
(2)
where oi = the used sample proportion of habitat i, ui = the sample count of conch in each habitat i, ut = the total number of observations of conch in all habitats.
To examine habitat utilization, we calculated a variety of parameters using
Π i ut = Π i ⋅ ut
(3)
where Πiut = the expected number of observations of conch in each habitat i.
Whereas the results derived from equations (2) and (3) do not necessarily indicate statistical
significance for preference or avoidance, the parameters are useful for explaining how conch
utilize each habitat relative to the availability of that habitat.
We also examined how conch utilized and selected for each habitat by calculating several selection ratios that are useful for illustrating how the population utilizes the resources relative to
their availability:
Wi = ui Πi ut
(4)
Bi = wi wt , wher
eree wt = ∑ wi
(5)
The wi is useful for illustrating how the population utilizes the resources relative to their availability and the Bi was calculated to facilitate comparisons between habitats and has the utility of
summing to one. Thus, a habitat with a B = 0.50 is selected for with a probability twice that of a
habitat with B = 0.25.
To test for statistically significant differences in habitat preference at each site, we employed a
log-linear Chi Squared test (Neu et al., 1974; Manly et al., 1993). If the test indicated that there
was a statistically significant preference or avoidance at a site, we examined which habitats were
preferred or avoided by comparing the habitat proportion for each habitat, Πi, with the bounds
on the 95% confidence limits of the used sample proportions, oi, for that habitat. The confidence
limits were calculated using a Bonferroni correction to the z statistic to adjust for nonindependent
multiple tests (Manly et al., 1993). There was no statistical significance indicated if Πi fell within
210
the range defined by the confidence limits. Conversely, if Πi was below the lower bound of the
confidence limit, preference was indicated; a Πi value above the upper limit indicated avoidance.
The tests were conducted at ∝ = 0.05. We used the G test to compare differences between the
habitat utilization of males and females.
We examined how conch used the available habitats during reproduction in two ways. In both
cases, we used the MCP calculated from all conch observations for the entire year because this
area represented habitat that was used by conch at some time during the study, and thus we reasoned that it must be suitable for their general requirements. In the first case, we examined the
utilization and preference of the habitats occupied by conch that were actively reproducing (i.e.,
mating or spawning) relative to how much of that habitat was available within the MCP. Utilization and preference were examined using the methods described above. Second, we examined
the utilization and preference of available habitats for all tagged conch during the season of
reproduction (R: April–September) and during the season when conch do not reproduce (NR:
October–March). To compare these seasons, we used the G test to examine the frequencies of
observations in the habitats.
Results
All conch were tagged and released from 27 June 1997 through 2 December 1997.
After 31 July, tagging was only conducted to replace lost or failed tags.
At site C1, 19 conch were tagged. Of these, 11 were males and eight were females. The
mean number of observations per conch was 7.4, with a maximum of 11 observations
and a minimum of two. The mean number of days that a conch was tracked was 150,
and the maximum was 330. A total of 140 observations was made at C1. Males were
observed 68 times; females were observed 72 times.
Twenty conch were tagged at site C2. Nine conch were males and ten were females;
the sex of one was unknown. The mean number of observations per conch was 7.6, with
a maximum of 15 observations and a minimum of one. The mean number of days each
animal was tracked was 162, and the maximum was 334. A total of 149 observations
was made at C2. Males were observed 65 times, females were observed 81 times, and
the conch of unknown sex was observed three times.
Habitat Availability
For site C1, we estimated that the MCP defined by the conch positions was ~98,000 m 2
and encompassed three habitats (Table 2). The dominant habitat in C1 was Rb (84.1%;
see Table 1 for habitat abbreviations and definitions), followed by SgsSd (11.0%); an additional 5% of the area was Cs. Within the Rb, there were small and isolated patches (i.e.,
microhabitats) of Cs whose areas could not be quantified from the aerial photographs.
The observations made of two conch that migrated from C1 to C2 (Fig. 2) were excluded
from the determination of the MCP and from the analyses because their positions were
beyond the limits of the aggregation and no other tagged or untagged conch were found
nearby.
Site C2 encompassed ~132,000 m2, and five habitats were present. Again, Rb was the
most common habitat (64.5%), followed by SgsCs (20.0%), RbCs (10.9%), Cs (4.6%),
and SgdCs (0.1%).
Habitat Utilization and Preference
Site C1.—At site C1, 11 conch were found exclusively in Rb; they were not found in
other habitats at any time during the study. Seven of the remaining eight conch were
observed in SgsSd at least once.
211
Table 1. Classifications of habitats delineated in maps used in acoustic telemetry study of queen
conch in the Florida Keys. Asterisks indicate habitats that were occupied by conch during the
study.
Code
Rf
Sd
Cs*
Rb*
RbCs*
SgsSd*
SgsCs
SgdSd*
SgdCs*
Habitat
Reef – continuous barren carbonate substrate
Sand – particles pass through 2-mm sieve but are retained on 0.5-mm sieve
Coarse Sand – particles pass through 12-mm sieve but are retained on 2-mm sieve
Rubble – particles are retained on 12-mm sieve
Rubble/Coarse Sand – homogenous mix of Rb and Cs
Mixed seagrass community (Thalassia testudinum and Syringodium filiforme),
sparse, with Sd substrate; seagrass blade density < 1200 m-2 and canopy height <
15 cm
Mixed seagrass community (T. testudinum and S. filiforme), sparse, with Cs substrate;
seagrass blade density < 1200 m−2 and canopy height < 15 cm
Mixed seagrass community (T. testudinum and S. filiforme), dense, with Sd substrate;
seagrass blade density > 1200 m−2 and canopy height > 15 cm
Mixed seagrass community (T. testudinum and S. filiforme), dense, with Cs substrate;
seagrass blade density > 1200 m−2 and canopy height > 15 cm
The highest proportion of observations of conch occurred in Rb (Table 2). Despite the
fact that Cs was selected for at approximately twice the probability of Rb or SgsSd (BCs
= 0.526, BRb = 0.252, BSgsSd = 0.222), there was no statistically significant preference for,
or avoidance of, any habitat type at C1 (χ
χ2 = 5.963, df = 2, P > 0.05). There also was no
significant difference between habitats occupied by males and females (G = 3.218, df =
2, P = 0.200).
Site C2.—At C2, 18 conch used more than one habitat during the study. Of these, 14
utilized one of the seagrass habitats at least once. One conch was only observed in RbCs
(9×).
Conch at site C2 were observed most often in the RbCs; approximately half of all
observations occurred in this habitat resulting in a selection for this habitat (Table 2).
Cs and SgdCs were also selected. The tagged conch used the SgdCs with almost a six
times greater probability than the next highest selected habitat RbCs (Table 3). Rb was
selected against.
There was a significant difference in the use of habitats relative to their availability (χ
χ2
= 307.567, df = 4, P < 0.001). Conch showed a significant preference for the Cs and RbCs
habitats and a significant avoidance of the Rb (Fig. 4). There was no significant preference for or avoidance of either SgsCs or SgdCs. There was no significant difference between males and females for the habitats occupied at C2 (G = 4.075, df = 4, P = 0.396.)
Reproductive Behavior
Site C1.—At C1, we observed 14 instances of conch reproducing (Table 3). Of these,
nine observations were of spawning (i.e., egg-laying); however, one individual accounted
for four of these nine observations. We also observed five instances of mating; one individual was observed mating on two different occasions.
One observation of reproduction occurred in Cs and 13 in Rb. However, nine of these
13 conch were observed occupying RbCs microhabitats isolated within the large expanse of Rb. Because this microhabitat was not discernable in the aerial photographs, we
could not directly quantify its contribution to overall habitat availability of RbCs at this
site. For these conch, we recorded their habitat occupied as Rb because, in the context
C2
C1
Site
Area (m2)
4,921
82,595
10,757
98,273
6,013
85,226
14,445
26,445
168
132,129
Habitat
Cs
Rb
SgsSd
Total
Cs
Rb
RbCs
SgsCs
SgdCs
Total
0.050
0.841
0.110
1.000
0.046
0.645
0.109
0.200
0.001
1.000
Habitat
proportion
(Πi)
14
113
13
140
22
6
78
39
4
149
Sample count
(ui)
0.100
0.807
0.093
1.000
0.148
0.040
0.523
0.262
0.027
1.000
Used sample
proportion
(oi)
7.0
117.7
15.3
140.0
6.9
96.1
16.2
29.8
0.1
149.1
Expected
count
(Πiut)
2.000
0.960
0.844
3.804
3.210
0.062
4.803
1.309
26.846
36.229
Selection
ratio
(wi)
0.526
0.252
0.222
1.000
0.089
0.002
0.133
0.036
0.741
1.000
Standardized
selection ratio
(Bi)
Table 2. Habitat utilization and selection by queen conch at sites C1 and C2 in the Florida Keys. Data are from observations of acoustically tagged conch.
Habitats were as follows: Cs = coarse sand; Rb = rubble; SgsSd = sparse, mixed seagrass community over sand; RbCs = homogeneous mix of rubble and
coarse sand; SgsCs = sparse, mixed seagrass community over coarse sand; SgdCs = dense, mixed seagrass community over coarse sand.
212
213
Figure 2. Habitat layer used to determine habitat availability and utilization for queen conch.
The data were digitized from 1:48,000 high-altitude conventional color photography (National
Ocean Service, 1991). The black polygons represent the Minimum Convex Polygons associated
with site C1 and site C2 and were used to estimate habitat availability. The symbols within the
MCPs represent the observations of individual conch at recapture. The habitats were identified
using high-altitude photography and digitized into ArcView v. 3.1. Habitat types were as follows:
Rf = reef; Sd = sand; Cs = coarse sand; Rb = rubble; SgsSd = sparse, mixed seagrass community
over sand; SgsCs = sparse, mixed seagrass community over coarse sand; SgdSd = dense, mixed
seagrass community over sand; SgdCs = dense, mixed seagrass community over coarse sand.
of habitat availability, they were within a Rb field. For reproducing conch, there was no
selection for any habitat; the habitats were used in proportion to their availability (χ
χ2 =
3.2295, df = 2, P > 0.05).
Site C2.—At C2, we observed reproduction occurring 12 times; one conch was observed simultaneously spawning and mating, but for analysis, this was treated as one
observation of reproduction. Of the 12 observations, six were of spawning conch. One
individual was observed laying eggs at two different times. Additionally, we observed
tagged conch mating six times. One conch was observed mating at two separate times.
Conch selected SgdCs and Cs for reproduction. The Cs was selected almost six times
more than RbCs or SgsCs.
There was a significant difference in how reproducing conch utilized the habitats relative to what was available (χ
χ2 = 32.670, df = 4, P < 0.001). Conch avoided Rb (Fig. 5).
No habitat was preferred.
214
Table 3. Habitat utilization and preference of queen conch observed reproducing (i.e., mating
and/or spawning) at sites C1 and C2 in the Florida Keys. Data are from acoustic telemetry
observations. At C2, one individual was observed mating and spawning simultaneously and
thus is presented here as one reproductive event. Habitat types were as follows: Cs = coarse
sand; Rb = rubble; SgsSd = sparse, mixed seagrass community over sand; RbCs = homogeneous
mix of rubble and coarse sand; SgsCs = sparse, mixed seagrass community over coarse sand;
SgdCs = dense, mixed seagrass community over coarse sand. “-” indicates where calculations
could not be made because the sample count was 0.
Site
C1
C2
Habitat
Sample
count (ui)
Used sample
proportion
(oi)
Expected
count
(Πiut)
Selection
ratio (wi)
Standardized
selection ratio
(Bi)
Cs
Rb
SgsSd
Total
Cs
Rb
RbCs
SgsCs
SgdCs
Total
1
13
0
14
4
0
2
4
1
11
0.071
0.929
0.000
1.000
0.364
0.000
0.182
0.364
0.091
1.000
0.7
11.8
1.5
14.0
0.5
7.1
1.2
2.2
0.0
11.0
1.429
1.104
2.533
7.905
0.000
1.668
1.818
90.909
102.300
0.564
0.436
0.000
1.000
0.077
0.000
0.016
0.018
0.889
1.000
Seasonal Habitat Use
Site C1.—The tagged conch at C1 occupied the different habitats at significantly different frequencies during R and NR (G = 7.988, df = 2, P = 0.018). There was a seasonal
shift within the Cs: two conch were observed there during NR and 12 conch during R
(Fig. 6). The number of individuals observed in Rb and SgsSd were approximately equal
during the two seasons. During R, there was a significantly disproportionate use of the
available habitat (χ
χ2 = 12.753, df = 2, P < 0.01). The conch showed a preference for Cs;
no other habitat was preferred or avoided (Fig. 7). During NR, there was no significant
difference in the utilization of habitat relative to habitat availability.
Site C2.—The frequency with which conch at C2 were observed in each habitat during
NR and R was not significantly different (G = 8.373, df = 2, P = 0.079). During both NR
and R, there was a significantly disproportionate use of the available habitat at C2 (NR:
χ2 = 180.202, df = 4, P < 0.01; R: χ2 = 135.949, df = 4, P < 0.01). RbCs was preferred and
Rb was avoided during both seasons (Fig. 8).
Discussion
In this study, we found that most conch occupied the rubble, coarse sand, and rubblecoarse sand habitats typical of the shallow-water, back reef zone of the Florida Keys;
adjacent seagrass meadows were generally avoided. We also identified coarse sand and
rubble-coarse sand habitats as those selected or preferred by adult conch.
The preference for coarse sand and rubble-coarse sand habitats was observed during
both the reproductive and nonreproductive seasons. These results were not unexpected;
reproducing conch often utilize sand habitats (Randall, 1964; Weil and Laughlin G.
1984; Stoner and Sandt, 1992). However, what we did not expect was that, in general
conch avoided the rubble habitat relative to its availability. Even at C1, where 84% of the
conch observations occurred in rubble, the conch did not use the habitat in proportion
to its availability.
215
Habitat
Figure 3. Adult queen conch with a sonic transmitter attached to the shell’s spire. The transmitter
was affixed with monel stainless steel wire strapped tightly around the spire. A unique frequency
and pulse combination was used to identify each animal in the field.
Figure 4. Test of significance of habitat preference by queen conch at site C2 in the Florida Keys.
The confidence intervals are on the used sample proportion (oi) and were calculated using a
Bonferonni correction (∝ = 0.05). The diamond indicates the habitat proportion. Significant
avoidance is indicated by the habitat proportion greater than the upper limit on the used sample
proportion; preference is indicated by the habitat proportion less than the lower limit on the used
sample proportion. Habitat types were Cs = coarse sand; Rb = rubble; RbCs = homogeneous mix
of rubble and coarse sand; SgsCs = sparse, mixed seagrass community over coarse sand; SgdCs
= dense mixed seagrass community over coarse sand.
216
Habitat
Selection index (w)
Figure 5. Test of significance of habitat preference or avoidance by conch observed reproducing
(i.e., spawning or mating) at site C2 in the Florida Keys. The confidence intervals are on the used
sample proportion (oi) and were calculated using a Bonferonni correction (∝ = 0.05). The diamond indicates the habitat proportion. Significant avoidance is indicated by the habitat proportion
greater than the upper limit on the used sample proportion; preference is indicated by the habitat
proportion less than the lower limit on the used sample proportion. Habitat types were characterized as follows: Cs = coarse sand; Rb = rubble; SgsSd = sparse, mixed seagrass community over
sand; RbCs = homogeneous mix of rubble and coarse sand; SgsCs = sparse, mixed seagrass community over coarse sand; SgdCs = dense, mixed seagrass community over coarse sand.
Habitat
Figure 6. Selection index (w) for habitats occupied by queen conch at site C1 in the Florida Keys
during the season when conch do not reproduce (NR: October–March) and the season when
reproduction occurs (R: April–September). Cs = coarse sand, Rb = rubble, and SgsSd = sparse
seagrass with sand substrate.
217
Habitat
Figure 7. Test of significance of habitat preference or avoidance by queen conch at site C1 in the
Florida Keys during the reproductive season (April–September). The confidence intervals are on
the used sample proportion (oi) and were calculated using a Bonferonni correction (∝ = 0.05).
The diamond indicates the habitat proportion. Significant avoidance is indicated by the habitat
proportion greater than the upper limit on the used sample proportion; preference is indicated
by the habitat proportion less than the lower limit on the used sample proportion. Habitat types
were as follows: Cs = coarse sand; Rb = rubble; SgsSd = sparse, mixed seagrass community over
sand.
The use of rubble may be driven, in part, by density-dependant mechanisms. The
Fretwell-Lucas theory (sensu MacCall, 1990) states that as densities increase, animals
will occupy less desirable habitats. At C1, densities were almost 10× greater than at C2
(Florida Fish and Wildlife Conservation Commission, unpubl. data). Coincidentally, the
conch at C1 were observed in relatively higher abundance in the rubble zone than were
conch in the rubble zone at C2. Thus, at C1, the high densities coupled with the limited
availability of the preferred habitats (only about 5% of the available habitat was represented by the coarse sand habitat) may have driven conch into the less desirable rubble
habitat.
In general, conch showed no preference for any of the seagrass habitats. This is in
contrast to the findings of studies in the Caribbean and Bahamas where adult conch
were reported to most often occupy the seagrass habitats (e.g., Randall, 1964; Weil and
Laughlin, 1984; de Jesús et al., 1999; Delgado, 1999). We suggest that these disparities
are most likely a result of fishing pressure.
Florida’s conch population is essentially unfished except for a small amount of poaching (Glazer, pers. obs.). After the harvest ban was instituted, recolonization was relatively rapid; from 1992–1997, we estimated that the population of adult conch in the
shoal-water, back reef zone throughout the Florida Keys increased from ~6000–21,000
individuals (Florida Fish and Wildlife Conservation Commission, unpubl. data). Conch
that aggregate in this zone may be more vulnerable to harvest because these areas are
easier to fish due to the shallower water; however, in a closed fishery, the conch are able
to colonize these areas from other, more difficult to harvest habitats (e.g., deeper seagrass or deep-water sand plains). Thus, in a biological sense, there may be no selection
for seagrass; however, the effect of fishing may indicate a preference for it. Conversely,
avoidance of the critically important shallow-water, back reef habitats may be observed,
218
Habitat
Figure 8. Test of significance of habitat preference by queen conch at site C2 in the Florida Keys
during the nonreproductive (NR; October–March) and reproductive (R; April–September) seasons. The confidence limits are on the used sample proportion (oi) and were calculated using
a Bonferonni correction (∝ = 0.05). The diamond indicates the habitat proportion. Significant
avoidance is indicated by the habitat proportion greater than the upper limit on the used sample
proportion; preference is indicated by the habitat proportion less than the lower limit on the used
sample proportion. Habitat types were as follows: Cs = coarse sand; Rb = rubble; RbCs = homogeneous mix of rubble and coarse sand; SgsCs = sparse, mixed seagrass community over coarse
sand; SgdCs = dense, mixed seagrass community over coarse sand.
but these observations may result from the intensive harvest in those easily exploited
aggregations.
These conclusions are supported by studies in other locations. In a harvest refuge
in the Bahamas, adult conch were found in greater abundance and density in shallow
waters than in a fished population (Stoner and Ray, 1996). In the fished population,
adult conch were most commonly found on deep-water sand plains, where the prohibition against using SCUBA or hookah essentially created a deep-water refuge. The same
pattern occurred in the U.S. Virgin Islands where conch were found in shallower waters
during years when the conch fishery was closed than in years when fishing was permitted (Friedlander et al., 1994). In this case, conch were most common in rubble when
the fishery was closed and in seagrass when the fishery was open. In Venezuela, conch
in an unfished population were observed in shallower locations than conch in a fished
219
population (Weil and Laughlin, 1984). The same pattern was observed in Martinique
(Rathier, 1993).
There are other explanations besides fishing pressure that may account for this discrepancy. Outside Florida, conch may occur in relative proportion to the expansive seagrass
beds. Therefore, conch may be more frequently observed in seagrass habitats because
their abundance is higher; however, there may be no selection for seagrass relative to its
availability. At site C1 in this study, the habitat most occupied by conch was rubble, yet
there was no preference indicated. To the casual observer, it may appear as if conch are
preferring rubble because of the much greater abundance there than in other habitats.
There also may be differences between the substrata underlying seagrass habitats in
Florida and the seagrass habitats in both the Bahamas and the Turks and Caicos Islands
(Glazer, pers. obs.). In Florida, most of the seagrass habitat appears to be on a substrate
of relatively fine sediment (Glazer, pers. obs.), which is poor habitat for conch (Glazer
and Berg, 1994). Substrate differences in seagrass habitats may account for differences
in conch distributions and confound comparisons between areas.
The spatial distribution of conch may be affected by more than habitat quality. For juvenile conch, distribution has been influenced by factors such as responses to predators
(Stoner and Waite, 1990; Marshall, 1992), forage (Stoner and Waite, 1990), hydrology
(Jones, 1996; Stoner et al., 1996a), sediment organics (Stoner et al., 1995) and larval
supply (Stoner and Waite, 1990; Jones, 1996; Stoner et al., 1996c). It is likely that at
least some of these variables may also affect adult conch distribution. It is clear that
habitat alone does not fully explain the distribution of conch and that what constitutes
preferred and avoided habitats may be related to the synergistic effects of habitat with
other variables.
Conch used different habitats during the reproductive and nonreproductive seasons.
During the nonreproductive time of the year, the majority of observations occurred in
the RbCs habitat, whereas during the reproductive season, the majority of observations
were made in the homogeneous coarse-sand plains. In the Bahamas, a similar shift in
habitat utilization was observed (Stoner and Sandt, 1992).
Several problems arose in this study. The first related to habitat availability. All preference studies suffer from the subjective determination of what constitutes available
habitat (McLean et al., 1998). In an effort to inject less subjectivity into that determination, we chose a method that estimates the available habitats in a fairly objective manner
(Porter and Church, 1987). However, it may be likely that using a different definition of
available habitat would have resulted in different results relative to habitat preference.
We encountered two problems with the tags. First, the factory neglected to glue the
tags to the base and, thus, some tags failed from saltwater intrusion. It was solely by
chance that we recovered a tagged conch with a failed tag. Second, the monel wire we
first received was defective and in some instances, corroded, thus necessitating retagging with replacement wire partway through the experiment. For these reasons, it was
difficult to assess whether a lost signal was due to a conch emigrating from the site or
from the failed tag. However, we believe most signals that were lost were due to a failed
tag because of our extensive searches well beyond the boundaries of the sites.
Additionally, the utilization of microhabitats within large expanses of homogeneous
habitat types made the determination of what constitutes preferred or selected habitats
difficult. A conch may choose a very small expanse of microhabitat to spawn within
and, in some cases, a conch may even physically move a piece of rubble to get at the
220
underlying coarse sand habitat (Glazer, pers. obs.). In these cases, it is very difficult to
quantitatively define the contribution this microhabitat makes to habitat availability.
Sample size is typically a problem in resource selection studies (Manly et al., 1993). In
general, the Neu method of analyzing resource selection is robust to small sample sizes
and small numbers of available habitats (Aldredge and Ratti, 1986). However, the small
sample number of tagged conch and/or the relatively small proportion of some available
habitats may confound interpretation of the results. For example at C2, conch selected
the scarce, dense seagrass/coarse sand (SgdCs) habitat. However, only four conch were
observed in the SgdCs. Because this habitat accounted for only 0.1% of the available
habitat, the small number of observations coupled with the limited availability of SgdCs
within the MCP make it difficult to draw conclusions about the importance of this habitat. It is certain that there is more SgdCs throughout the region that was not considered as
available habitat in this study. Had it been considered available habitat, we suspect that
there would have been either no selection for SgdCs or avoidance of the SgdCs.
The Magnuson-Stevens Fisheries Conservation Act defines EFH as those habitats necessary for the species of concern to grow to maturity, spawn, breed, or feed. The Act
stipulates that those habitats identified as EFH should be conserved or rehabilitated to
ensure the productivity of that species.
Based on this study, EFH for conch may be assessed at different spatial scales. This
study identified preferred and avoided habitats at a scale localized within a back reef environment. Examining the habitats at this scale, the coarse sand and rubble-coarse sand
habitats were clearly preferred and the rubble avoided. Yet, each of these habitats may
provide different ecological functionality. Additionally, despite the avoidance of rubble
relative to its availability, most conch were found occupying this habitat. Therefore, not
recognizing rubble as essential may have the overall effect of negatively impacting most
of the individuals that one is trying to protect or manage. Thus, it probably would not
be either biologically prudent or practical to manage these habitats differently at this
scale.
It may be more useful to view EFH for conch as a mosaic of habitats which, when
taken in the aggregate, support a full range of biological functionality. Ultimately, the
availability of suitable habitats will determine the distribution of the population of organisms. These habitats must provide areas for reproduction, cover, foraging, and dispersal
(Morrison et al., 1998). The course sand habitat is, without doubt, critical for spawning
(Randall, 1964; Stoner and Sandt, 1992; this study). It is likely that the adjacent seagrass habitats provide forage. The rubble habitats probably are suitable for supporting a
full range of biological functions; at C1, the conch occupying rubble often remained in
the rubble. Additionally, the use of coarse-sand microhabitats for spawning within the
rubble field adds complexity to this habitat.
It may be most useful to define EFH for conch in Florida at an even greater scale to
include the continuous back reef zone along the Florida Keys reef tract. This strategy
provides a variety of benefits. First, it will protect the contiguous habitats comprising
those areas that sustain conch populations. Furthermore, this approach will conserve
those areas where conch do not currently exist yet where colonization may yet occur.
This approach is particularly useful from a management perspective as it avoids defining smaller management units that may be difficult to manage individually. Additionally, it will protect nursery areas for juvenile conch because they settle in locations
closely associated with the adult aggregations (Florida Fish and Wildlife Conservation
Commission, unpubl. data). These aggregations were beyond the scope of this study but
221
determination of EFH for them is critical for sustaining the population. Finally, habitats
will be conserved that were not identified in this study but may provide important biological resources to conch in other south Florida locations.
A new paradigm that embraces the establishment of fishery reserves has emerged for
managing marine fisheries. A goal of this strategy is to conserve the spawning stock
to ensure a consistent supply of larvae to fished populations, and to enhance fishery
yields (Russ and Alcala, 1996). The identification of potential fishery reserves requires
a broad understanding of the utilization of resources by the species of concern and what
constitutes available and optimal habitat (Recksiek and Appeldoorn, 1998). Future studies should incorporate a multidisciplinary approach to examine the variables that drive
conch distribution at both the organismal and ecological levels (sensu Jones, 1996) and
to determine how these variables interact.
Acknowledgments
This study was supported by a grant from the Florida Fish and Wildlife Commission’s NonGame grants program (NG97-005). The paper was funded by a grant from the Caribbean Marine Research Center (CMRC Project #CMRC-00-IXNR-03-01A), National Oceanic and Atmospheric Administration (NOAA) National Underseas Research Program, U.S. Environmental
Protection Agency, and Environmental Defense. Views expressed herein are those of the authors
and do not necessarily reflect the views of CMRC, or any of the supporting agencies. C. Bartels,
R. Bertelsen, J. Quinn, J. Hunt, and three anonymous reviewers provided valuable comments on
the manuscript. J. Leiby and L. French assisted in the editorial process. The Nature Conservancy
provided volunteer support for assistance in the field sampling. The authors wish to thank E. Webber and L. Lawrence for their assistance in the field.
Literature Cited
Aebischer, N. J., P. A. Robertson, and R. E. Kenward. 1993. Compositional analysis of habitat use
from animal radio-tracking data. Ecology 74: 1313–1325.
Alcolado, P. M. 1976. Crecimiento, variaciones morfologicas de la concha y algunas datos biologicos del cobo Strombus gigas L. (Mollusca, Mesogastropoda). Acad. Cienc. Cuba, Ser.
Oceanol. 34: 1–36.
Alldredge, J. R. and J. T. Ratti. 1986. Comparison of some statistical techniques for analysis of
resource selection. J. Wildl. Manage. 50: 157–165.
Appeldoorn, R. S. 1988. Age determination, growth, mortality and age of first reproduction in
adult queen conch, Strombus gigas L., off Puerto Rico. Fish. Res. 6: 363–378.
_______________ 1994. Queen conch management and research: status, needs and priorities.
Pages 301–319 in R. S. Appeldoorn, ed. Proceedings of the 1st Latin American Malacological
Conference. Special workshop on the management and culture of queen conch. Fundacíon
Científica Los Roques, Caracas.
Berg, C. J., Jr. and R. A. Glazer. 1995. Stock assessment of a large marine gastropod (Strombus
gigas) using randomized and stratified towed-diver censusing. Rapp. P.−v. Réun. Cons. int.
Explor. Mer. 199: 247–258.
____________ and D. A. Olsen. 1989. Conservation and management of the queen conch (Strombus gigas) fisheries in the Caribbean. Pages 421–442 in J. F. Caddy, ed. Marine invertebrate
fisheries: their assessment and management. Wiley and Sons, New York.
Brownell, W. N. 1977. Reproduction, laboratory culture, and growth of Strombus gigas, S. costatus, and S. pugilus in Los Roques, Venezuela. Bull. Mar. Sci. 27: 668–680.
Caribbean Fishery Management Council. 1996. Fishery management plan, regulatory impact review, and final environmental impact statement for the queen conch resources of Puerto Rico
and the United States Virgin Islands. San Juan, Puerto Rico. 41 p.
222
D’Asaro, C. N. 1965. Organogenesis, development and metamorphosis in the queen conch, Strombus gigas, with notes on breeding habits. Bull. Mar. Sci. 15: 359–416.
Davis, M., C. A. Bolton, and A. W. Stoner. 1993. A comparison of larval development, growth, and
shell morphology in three Caribbean Strombus species. Veliger 36: 236–244.
________, B. A. Mitchell, and J. L. Brown. 1984. Breeding behavior of the queen conch Strombus
gigas Linné held in a natural enclosed habitat. J. Shellfish Res. 4: 17–21.
De Jesús, A. E. González, J. Oliva. A. Pelayo, and G. Medina. 1999. Advances over some ecological aspects of queen conch, Strombus gigas L. in southern Quintana Roo, México. Proc. Gulf
Carib. Fish. Inst. 45: 932–943.
Delgado, G. A. 1999. Influence of habitat on queen conch abundance and size distribution in softsediment marine communities in Parque del Este, Dominican Republic. M.S. Thesis, University of Miami, Coral Gables. 74 p.
Department of Commerce. 1996. Florida Keys national marine sanctuary final management plan/
environmental impact statement (FMP/EIS), Volume I. National Oceanic and Atmospheric Administration, Sanctuaries and Reserves Division. 319 p.
Friedlander, A., R. S. Appeldoorn, and J. Beets. 1994. Spatial and temporal variations in stock
abundance of queen conch, Strombus gigas, in the U.S. Virgin Islands. Pages 51–60 in R. S.
Appeldoorn, ed. Proceedings of the 1st Latin American malacological conference. Special
workshop on the management and culture of queen conch. Fundacíon Científica Los Roques,
Caracas.
Glazer, R. A. and C. J. Berg, Jr. 1994. Queen conch research in Florida: an overview. Pages 79–95
in R. S. Appeldoorn, ed. Proceedings of the 1st Latin American malacological conference.
Special workshop on the management and culture of queen conch. Fundacíon Científica Los
Roques, Caracas.
__________ and I. Quintero. 1998. Observations on the sensitivity of queen conch to water quality: implications for coastal development. Proc. Gulf Caribb. Fish. Inst. 50: 78–93.
Hartwell, S. I. 1998. Biological habitat quality indicators for essential fish habitat. NOAA Tech.
Memo. NMFS-F/SPO-32. 124 p.
Hesse, K. O. 1979. Movement and migration of the queen conch, Strombus gigas, in the Turks and
Caicos Islands. Bull. Mar. Sci. 29: 303–311.
Hooge, P. N., W. M. Eichenlaub, and E. K. Solomon. 2001. Using GIS to analyze animal movements in the marine environment. Pages 37–51 in G.H. Kruse, N. Bez, A. Booth, M. W. Dorn,
S. Hills, R. N. Lipcius, D. Pelletier, C. Roy, S. J. Smith, and D. Witherell, eds. Spatial processes
and management of marine populations. University of Alaska Sea Grant, AK-SG-01-02. Fairbanks.
Iversen, E. S., D. E. Jory, and S. P. Bannerot. 1986. Predation on queen conch, Strombus gigas, in
the Bahamas. Bull. Mar. Sci. 39: 61–75
Jones, R. L. 1996. Spatial analysis of biological and physical features associated with the distribution of queen conch, Strombus gigas, nursery habitats. 1996. M.S. Thesis, Florida Inst. Tech.,
Melbourne. 98 p.
MacCall, A. D. 1990. Dynamic geography of marine fish populations. Washington Sea Grant,
Univ. Washington, Seattle. 153 p.
Mahon, R. 1990. Fishery management options for the Lesser Antilles counties. Food and Agriculture Organization of the United Nations, Technical Paper 313. 126 p.
Manly, B. F. J., L. L. McDonald, and D. L. Thomas. 1993. Resource selection by animals: statistical design and analysis for field studies. Chapman and Hall, London. 177 p.
Marshall, L. S., Jr. 1992. Survival of juvenile queen conch, Strombus gigas, in natural habitats: impact of prey, predator and habitat features. Ph.D. Diss., College of William and Mary, Gloucester Point.144 p.
Mateo, I., R. Appeldoorn, and W. Rolke. 1998. Spatial variations in stock abundance of queen
conch, Strombus gigas, (Gastropoda: Strombidae) in the west and east coast of Puerto Rico.
Proc. Gulf Caribb. Fish. Inst. 50: 32–48.
223
McCarthy, K. J., C. T. Bartels, M. C. Darcy, G. A. Delgado, J. R. Styer, and R. A. Glazer. 2000.
Habitat induced reproductive failure of queen conch, Strombus gigas, in the Florida Keys: a
final report. U.S.F.W.S. Project P-3. Partnerships for Wildlife. 30 p.
McClean, S. A., M. A. Rumble, R. M. King, and W. L. Baker. 1998. Evaluation of resource selection methods with different definitions of availability. J. Wildl. Manage. 62: 793–801.
Morrison, M. L., B. G. Marcot, and R. W. Mannan. 1998. Wildlife-habitat relationships, 2nd Ed.
University of Wisconsin Press, Madison. 435 p.
National Ocean Service. 1991. High altitude conventional color photographs of the Florida Keys
reef tract. 1:48,000. National Ocean Service, NOAA. Frame 2407.
Neu, C. W., C. R. Byers, J. M. Peek, and V. Boy. 1974. A technique for analysis of utilization availability data. J. Wildl. Manage. 38: 541–545.
Porter, W. F. and K. E. Church. 1987. Effects of environmental pattern on habitat analysis preference. J. Wildl. Manage. 51: 681–685.
Randall, J. E. 1964. Contributions to the biology of the queen conch, Strombus gigas. Bull. Mar.
Sci. 14: 246–295.
Rathier, I. 1993. Le stock de lambi (Strombus gigas l.) en Martinique: analyse de la situation
1986-1987, modelisation de l’exploitation, option d’amenagement. Ph.D. Diss., Université de
Bretagne Occidentale, Brest. 273 p.
Recksiek, C. W and R. S. Appeldoorn. 1998. In pursuit of design criteria for marine fishery reserves. Proc. Gulf Caribb. Fish. Inst. 50: 372–384.
Robertson, R. 1959. Observations on the spawn and veligers of conchs (Strombus) in the Bahamas.
Proc. Malacological Soc. 33: 164–170.
__________. 1961. The feeding of Strombus and related herbivorous marine gastropods. Notulae
Naturae. 343: 1–9.
Russ, G. R. and A. C. Alcala. 1996. Do marine reserves export adult fish biomass? Evidence from
Apo Island, central Philippines. Mar. Ecol. Prog. Ser. 132: 1–9.
Stoner, A. W. 1994. Significance of habitat and stock pre-testing for enhancement of natural fisheries: experimental analyses with queen conch. J. World Aquacult. Soc. 25: 155–165.
___________ and M. Ray. 1996. Queen conch, Strombus gigas, in fished and unfished locations
of the Bahamas: effects of a marine fishery reserve on adults, juveniles, and larval production.
Fish. Bull., U.S. 94: 551–565.
___________, J. Lin, and M. D. Hanisak. 1995. Relationships between seagrass characteristics
and juvenile queen conch (Strombus gigas Linné) abundance in the Bahamas. J. Shellfish Res.
14: 315–323.
____________, P. A. Pitts, and R. A. Armstrong. 1996a. Interaction of physical and biological
factors in the large-scale distribution of juvenile queen conch in a seagrass meadow. Bull. Mar.
Sci. 58: 217–233.
___________, M. Ray, R. A. Glazer, and K. J. McCarthy. 1996b. Metamorphic responses to natural substrata in a gastropod larva: decisions related to postlarva growth and habitat preference.
J. Exp. Mar. Bio. Ecol. 205: 229–243.
___________, ______, _________, and P. J. Barile. 1996c. Larval supply to queen conch nurserie: relationships with recruitment process and population size in Florida and the Bahamas. J.
Shellfish Res. 15: 407–420.
___________ and V. J. Sandt. 1992. Population structure, seasonal movements and feeding of
queen conch, Strombus gigas, in deep-water habitats of the Bahamas. Bull. Mar. Sci. 51: 287–
300.
____________ and K. C. Schwarte. 1994. Queen conch, Strombus gigas, reproductive stocks in
the central Bahamas: distribution and probable sources. Fish. Bull., U.S. 92: 171–178.
____________ and J. M. Waite. 1990. Distribution and behavior of queen conch Strombus gigas
relative to seagrass standing crop. Fish. Bull. 88: 573–585.
Weil, E. and R. Laughlin. 1984. Biology, population dynamics, and reproduction of the queen
conch Strombus gigas Linné in the Archipelago de los Roques National Park. J. Shellfish Res.
4: 45–62.
224
Address: Florida Fish and Wildlife Conservation Commission, Fish & Wildlife Research Institute, 2796 Overseas Highway, Suite 119, Marathon, Florida 33050. Corresponding Author:
(R.A.G.) Tel: 305-289-2330, Fax: 305-289-2334, E-mail: <bob.glazer@fwc.state.fl .us>.
THE SEASCAPE APPROACH TO CORAL ECOSYSTEM
MAPPING: AN INTEGRAL COMPONENT OF UNDERSTANDING
THE HABITAT UTILIZATION PATTERNS OF REEF FISH
Matthew S. Kendall, Ken R. Buja, John D. Christensen,
Curtis R. Kruer, and Mark E. Monaco
ABSTRACT
Benthic maps provide the fundamental analytical framework to design reef-monitoring programs, organize spatial data, and to conduct spatially-explicit assessments of
various components of the reef ecosystem. To meet the need for benthic maps, novel
software that emphasizes simplicity but maintains capability in the mapping process
was developed to delineate benthic features directly into a Geographic Information
System. The software expedites mapping relative to more commonly used mapping
techniques while maintaining excellent thematic and spatial accuracies. A hierarchical
classification scheme was used in which the major bottom types were unconsolidated
sediment, coral reef and hard bottom, and submerged vegetation. Mapping in the U.S.
Virgin Islands covered 23.9 km2 of unconsolidated sediment, 160.5 km2 of submerged
vegetation, and 298.7 km2 of coral reef and hard bottom. Mapped features were also
given a location attribute such as back reef or fore reef according to their position relative to the shoreline and lagoon-forming reefs. There were large differences in spatial
extent among zones such as lagoon and back reef (22.9 km 2, 4.7% of the total area) and
bank/shelf (433.9 km2, 88% of the total area). Thematic accuracy of maps produced
using the new approach was measured by comparing ground survey data to map attributes and was similar to the accuracy of maps produced with an analytical stereo
plotter. Recent literature indicates that analysis of fish census data in concert with such
benthic maps can be used to more clearly evaluate fish distributions relative to analysis
of census data without a seascape framework.
This paper emphasizes the need for comprehensive maps of reef ecosystems to define the
spatial distribution of benthic habitats and develop a more complete understanding of the
coupling between organisms and their ecological setting.
Simple studies of fish communities within single habitat types without reference to the
influence of the adjacent seascape elements often do not completely explain patterns in
community structure (Brandt and Kirsch, 1993). This is because the fish community at a
given site will not only be influenced by the habitat at that site, but also by direct or indirect interaction with other habitats occurring some distance away (Parrish, 1989). Interactions across two or more habitat types occur at a range of spatial scales. For example,
there is ample evidence of relatively short foraging migrations of reef organisms into
seagrass areas in the form of grazed halos of bare sand several meters wide between reef
and vegetated habitats (Ogden, 1976; Ogden and Zieman, 1977; Tribble, 1981). Broader
scale interactions between reef and adjacent sand and seagrass habitat occur in the Haemulidae (grunts) and Mullidae (goatfishes) which utilize the reef as a structural refuge
and migrate tens to hundreds of meters into sand and seagrass areas to forage (Randall,
1967; Helfman et al., 1982; Burke, 1995; Randall, 1996). Carangids (jacks) that cruise
225
226
higher in the water column are even more wide ranging, with daily travels encompassing
many different habitat types. While the distribution of these organisms is directly affected by seascape heterogeneity, there are also indirect influences on the distribution of
other organisms through a web of ecological pathways. For example, haemulids obtain
most of their food energy from sand and seagrass foraging areas each night. Much of
this energy is transferred to various trophic pathways of the reef via fecal matter from
the grunts or through direct predation on the grunts themselves by piscivorous reef fish.
Excretion from grunts has been demonstrated to double the amount of some essential
nutrients available to coral reefs (Meyer and Shultz, 1985) and result in higher growth
rates for some corals (Meyer et al., 1983). In addition, predatory reef fish such as the
sand diver (Synodus intermedius Agassiz) and peacock flounder (Bothus lunulatus Linnaeus) have been observed to position themselves along grunt migration corridors at
reef/seagrass margins (Helfman et al., 1982). Differences in intensity of piscivory have
also been documented based on location on a reef (Connell, 1996; Letourneur, 1996) or
distance from reefs (Shulman, 1985). Only by using benthic maps to incorporate knowledge of the spatial configuration of all seascape elements into analyses of fish distribution and community structure can a more complete understanding of the ecosystem be
obtained.
To address this need for the tropical marine environment we developed a novel approach to coral ecosystem mapping that allows rapid delineation of discrete habitat types
within a geomorphological context. We define coral ecosystems as those integrated biological and physical habitats that include not only structures created from scleractinian
coral such as patch reefs and barrier reefs but also other commonly associated habitats
such as seagrass, mangrove, and soft sediment areas. In addition to delineations of bottom type, the spatial arrangement of these seascape elements relative to each other can
often be divided into discrete geomorphological zones such as lagoons, back reefs, and
fore reefs that are recognizable in most coral reef ecosystems throughout the world.
Categorization of mapped habitats into such zones provides a useful spatial framework
within which to interpret fish distribution since these zones typically encompass multiple habitats subjected to similar exposure regimes of currents, waves, sedimentation,
and other physiographic forces (Alevison et al., 1985; McGehee, 1994). For example,
habitats occurring in the lagoon zone such as patch reefs and seagrass beds are often exposed to elevated sedimentation rates and low intensities of wave surge whereas
the inverse often occurs for habitats located in zones seaward of the reef crest. These
physiographic and oceanographic forces are often just as important as substrate type in
determining fish community structure. Structurally identical bottom types can be functionally very different when they occur in different locations on the shelf (Lindeman et
al., 1998). By mapping these zones, which are representative of specific physiographic
regimes, it is possible to identify fish communities associated with particular areas of the
coral reef ecosystem such as the lagoon, back reef, fore reef, and shelf edge.
To be useful in understanding patterns of reef fish distribution and community structure, habitat maps must be created at an appropriate scale and contain sufficient spatial
and thematic detail to track the interactions of organisms among habitats. In this paper
we describe a conceptual framework and method to develop maps of the benthic seascape that can be used to evaluate the distribution of fish communities. The objectives
of this paper are to: 1) introduce novel software for efficient mapping of benthic features
from georeferenced aerial imagery, 2) discuss the evolution of studies of fish/habitat associations in coral ecosystems through selected examples from the scientific literature,
KENDALL ET AL.: SEASCAPE APPROACH TO CORAL MAPPING
227
and 3) emphasize that a thorough understanding of habitat utilization patterns of fish
must incorporate knowledge of the spatial distribution of seascape elements comprising
the coral ecosystem.
Methods
We created maps of the coral reefs, seagrass beds, mangrove forests, and other common marine habitats in the U.S. Virgin Islands. Benthic features were mapped directly into a geographic
information system (GIS) using software specially developed to facilitate visual interpretation of
orthorectified aerial imagery. While the image data used in this project was derived from aerial
photographs it is important to note that any georeferenced imagery could be used with our approach including satellite or sonar based data.
Classification Scheme.—Among the first steps in creating the benthic maps was devising
a classification scheme that would accomplish the management and scientific objectives of the
maps within the limitations imposed by the source data (1:48,000 scale aerial photos in this case).
To meet this need, a hierarchical classification scheme was used to describe and delineate bottom
features. A draft scheme was created based on preliminary evaluation of the spectral and spatial
resolution of the aerial photography, the desired ecological applications of the map products, and
published schemes used for similar mapping projects on coral reef ecosystems (Kruer, 1995; Lindeman et al., 1998; Reid and Kruer, 1998; Beets et al., 1986; Boulon, 1986; Holthus and Maragos,
1995; Shepard et al., 1995; Vierros, 1997; Chauvaud et al., 1998; Florida Fish and Wildlife Conservation Commission, Florida Marine Research Institute and NOAA, 1998; Mumby et al.,1998;
NOAA et al., 1998; NOAA et al., 2000). This draft scheme was then reviewed and revised by a
panel of local experts in the U.S. Caribbean. The final scheme defines benthic communities on
the basis of two attributes: large geographic “zones”, which are composed of smaller “habitats”.
Zone refers only to benthic community location and habitat refers to substrate and/or cover type.
Therefore, every polygon on the benthic community map was assigned a habitat within a zone
(e.g., sand in the lagoon, or sand on the bank) and not only defined discrete bottom types but also
provided a spatial framework in which they occur.
By examining aerial photos of the U.S. Virgin Islands we identified nine, mutually exclusive
zones from land to open water corresponding to typical insular shelf and coral reef geomorphology (Fig. 1). These zones included: land, shoreline/intertidal, lagoon, back reef, reef crest, fore
reef, bank/shelf, bank/shelf escarpment, and dredged (since this condition eliminates natural
geomorphology). In addition, 26 distinct and non-overlapping habitat types were identified that
could be mapped by visual photointerpretation. Habitats were defined in a collapsible hierarchy
ranging from four broad classes (submerged vegetation, unconsolidated sediment, coral reef/hard
bottom, and other), to more detailed categories (e.g., mangrove, seagrass, algae, individual patch
reefs, bedrock, etc.), to patchiness of some specific features (e.g., percent cover of seagrass and
macroalgal categories, as suggested by the Seagrass Research Workshop, NOAA et al., 1998).
Figure 1. Cross section of geomorphic zones from across the shelf and reef areas to deep, ocean
water that is seaward of the shelf escarpment.
228
The complete description of each zone and habitat including example images is too lengthy to
include here but is available in Kendall et al. (2001).
List of Habitats.—
Unconsolidated Sediments (0 to <10% submerged vegetation)
Sand
Mud
Submerged Vegetation
Seagrass
Continuous Seagrass (90–100% cover)
Patchy (Discontinuous) Seagrass (70 to <90% cover)
Macroalgae
Continuous Macroalgae (90–100% cover)
Patchy (Discontinuous) Macroalgae (50 to <90% cover)
Patchy (Discontinuous) Macroalgae (10 to <50% cover)
Coral Reef and Hardbottom
Coral Reef and Colonized Hardbottom
Linear Reef
Spur and Groove
Individual Patch Reef
Aggregated Patch Reefs
Scattered Coral/Rock in Unconsolidated Sediment
Colonized Pavement
Colonized Bedrock
Colonized Pavement with Sand Channels
Uncolonized Hardbottom
Reef Rubble
Uncolonized Pavement
Uncolonized Bedrock
Uncolonized Pavement with Sand Channels
Other Delineations
Land
Mangrove
Artificial
Unknown
Creating Georeferenced Mosaics.—In 1999, ~300 color, 1:48,000 scale, aerial photos
were taken of the coastal waters of the U.S. Virgin Islands. Prints and color transparencies were
created from the original negatives. Transparencies were then scanned at a resolution of 500 pixels per inch using a metric scanner, yielding 2.4 × 2.4 m pixels. All scans were saved in tagged
image file (TIF) format for the purposes of orthorectification. Photos were orthorectified using
Socet Set Version 4.2.1 software and ground control points collected with differential global
positioning system (GPS) equipment. Average spatial accuracy of the completed orthorectified
mosaics was measured using independent ground control points and generally fell within 0–3
pixel widths from the true location (Table 1).
Table 1. Mosaic specifications for each island. Accuracy is in m ± SD.
Location
St. John
St. Thomas
St. Croix
# of photos
14
20
27
Avg. spatial accuracy X
4.31 ± 5.2
1.48 ± 1.3
1.21 ± 3.0
Avg. spatial accuracy Y
2.14 ± 8.4
1.05 ± 3.4
0.69 ± 3.4
229
Only the best segments of each photo were used to create the final mosaic from which benthic
features were to be interpreted. Segments of each photo were selected to minimize sun glint,
cloud interference, turbidity, and other factors that reduce interpretability. As a result, segments
from 61 out of the ~300 original aerial photos were selected to create the final mosaic. Final mosaics for the U.S. Virgin Islands were created in geoTIF format (georeferenced image file) with
the following projection parameters: North American Datum 83, Universal Transverse Mercator
Zone 20.
Digitizing Benthic Habitats.—We created a new software extension for ArcView 3.2 that
allows polygons to be rapidly digitized and attributed using a simple graphical user interface.
Simple pull-down menus and dialog boxes allow users to specify a minimum polygon size, set
the scale at which digitizing will be conducted (the scale at which images will appear on the
computer monitor), create custom hierarchical classification schemes, and efficiently delineate
and attribute polygons around features visible in georeferenced imagery. By working directly
with experts who routinely use more technical mapping approaches and equipment, we designed
the software to minimize the complexities of the digitizing process for users with a minimum of
ArcView and GIS experience while maintaining the range of options and attributes needed to
create robust map products (Kendall et al., 2001).
For this project, a minimum polygon size of 1 ac was used and all digitizing was conducted
at a scale of 1:6000 on the computer monitor. Minimum polygon size was selected to balance
several considerations including the technical limitations of the source data (e.g., scale of aerial
photos, scan resolution), the size of the area to be mapped, the needs of user communities, and the
resolution required for the maps to have ecological relevance (Alevison et al., 1985; Parrish, 1989;
McGehee, 1994; Lindeman et al., 1998; Sale, 1998; Kendall et al., 2003). Once the orthorectified
mosaic was loaded into ArcView, polygons were delineated around contiguous areas with similar
spectral signatures (e.g., areas with specific color and texture patterns) appearing on the computer
screen, which corresponded to habitat types described in the classification scheme. The same
individual conducted all digitizing.
Following careful evaluation of the aerial photography, and in some cases creation of a first
draft habitat map through the process outlined above, selected sites were visited in the field for
typological validation. This validation included: (1) areas in the aerial photography and mosaic
with confusing or difficult to interpret signatures; (2) transects across representative habitat types
occurring in different depths and water conditions; (3) a survey of the zones; and (4) confirmation
of preliminary habitat delineations if a first draft of the habitat maps was produced. Individual
sites were visually evaluated by snorkeling and free diving or directly from a boat in shallow,
clear water. Once field data were processed, polygon boundaries and habitat classifications were
created or revised where necessary, and zone attributes were assigned to each polygon. Additionally, historical field data and habitat maps were used when available to assist with draft interpretation (Kumpf and Randall, 1961; Adey, 1979; Beets et al., 1986; Boulon, 1986; Bacle 1995). Draft
habitat maps were then reviewed and revised with the guidance of local experts at peer review
sessions in the U.S. Virgin Islands and over the Internet. Thematic accuracy was assessed for
these final maps.
Assessment of Classification Accuracy.—Thematic accuracy of the habitat maps was
evaluated for the three most general habitat categories: unconsolidated sediment, submerged vegetation, and coral reef/hard bottom. Logistical constraints, such as the prohibitively large amount
of field time required for in situ data collection of robust sample sizes, prevented more detailed
classification categories from being evaluated. Accuracy was estimated for the maps of Buck
Island National Monument, St. Croix, and the surrounding ecosystem since this area includes the
full complement of habitat types (except mud and mangroves), depth ranges, and water conditions found around the U.S. Virgin Islands. The accuracy of maps measured at this location was
assumed to be similar to map accuracy elsewhere in the project area. This approach enabled a
statistically robust evaluation of thematic accuracy to be conducted at least for the three main
habitat types without the logistical difficulty of collecting data for accuracy assessment over the
entire project area.
230
Figure 2. Mapped habitats (% of area excluding land and unknown).
In addition, since new mapping software was used to enhance production time it was necessary
to ensure that maps produced directly in a GIS environment had comparable accuracy to maps
produced using more routinely used techniques. To accomplish this goal, the thematic accuracy
of ArcView derived maps was compared to the accuracy of maps produced using published and
well-known photogrammetric techniques. Maps of the Buck Island area were created using two
techniques; the ArcView Extension and on-screen digitizing process described above, and a standard photogrammetric technique using an analytical stereoplotter to visually interpret benthic
features directly from hard copy photos. Maps derived using the stereoplotter were created by
the NOAA Coastal Services Center using Coastal Change Analysis Program (C-CAP) protocols,
which include widely accepted and commonly used photogrammetric techniques and instruments
(Dobson et al., 1995).
While map production was underway, habitat type at 109 sites was evaluated in the Buck Island
area to compare with habitat delineations derived from each mapping technique. A stratified sampling protocol was used where sample sites were pre-selected so that overall thematic accuracy
of the three major habitat types across the range of depths and water conditions found in the field
could be evaluated. Data recorded at each site included habitat type, depth, and other descriptive information such as dominant species of seagrass or coral where applicable. Data recorded
at each of the ground truth sites was overlaid onto the habitat maps and compared against the
classification assigned by the photointerpreters. After comparing the map classification to each
ground truth site, an error matrix was produced displaying overall accuracy (total number of correct classifications/total number of ground truth points evaluated), errors of inclusion (number of
correct classifications within a given category/number of ground truth locations classified as that
category) and omission (number of correct classifications within a given category/actual number
of ground truth points of that category), and Kappa Statistic ([overall classification accuracyexpected accuracy if classifications were randomly assigned]/[1- expected accuracy if classifications were randomly assigned]) (Congalton, 1991).
Results
Coral Reef Ecosystem Maps.—A two dimensional area of ~486 km2 was mapped
excluding areas labeled as “unknown” and “land”. Of this area 23.9 km2 was unconsolidated sediment (5.0% of area mapped), 160.5 km2 was submerged vegetation (33.3% of
Figure 3. Mapped geomorphic zones (% of area excluding land and unknown).
231
Table 2. Matrix of area and percent of total area mapped for each of the unique zone/habitat combinations excluding polygons delineated as land and unknown. The upper value in each cell denotes the area mapped in km2 for each zone/habitat combination. The
lower, italicized value in each cell is the percentage of the total area that was mapped which is covered by each zone/habitat combination. Values in bold denote row and column totals. Blank cells indicate that no features were mapped that met those zone and habitat
criteria. All values are rounded to two decimal places and therefore may result in slight apparent discrepancies in presentation of row
and column totals.
Artificial
Colonized bedrock
0.08
0.02
Colonized pavement
Colonized pavement with sand channels
Linear reef
Macroalgae
Mangrove
Mud
2.11
0.43
0.53
0.11
Patch reef (aggregated)
Patch reef (individual)
Reef rubble
Sand
0.02
0.00
0.04
0.01
Scattered coral/rock in unconsolidated sed.
Seagrass
0.49
0.01
0.79
0.16
1.04
0.21
0.30
0.06
0.21
0.04
0.32
0.07
0.00
0.00
0.63
0.13
0.18
0.04
0.03
0.01
0.05
0.01
0.16
0.03
0.06
0.01
1.47
0.30
1.51
0.31
0.17
0.03
1.11
0.23
0.12
0.02
12.51
2.57
0.16
0.03
0.01
0.00
0.00
0.00
0.00
0.00
0.08
0.02
2.12
0.44
0.38
0.07
14.82
3.05
0.00
0.00
0.11
0.02
Spur and groove reef
Uncolonized bedrock
0.29
0.06
0.00
0.00
Uncolonized pavement
Uncolonized pavement with sand channels
Column totals
3.56
0.73
20.84
4.28
1.87
0.39
1.72
0.35
0.24
0.05
0.33
0.07
0.05
0.01
0.38
0.08
0.10
0.02
14.49
2.98
0.09
0.02
12.21
2.51
100.91
20.76
120.46
24.78
3.53
0.73
76.37
15.71
0.02
0.00
0.29
0.06
2.17
0.45
4.72
0.97
7.21
1.48
1.30
0.27
0.19
0.04
18.59
3.82
23.67
4.81
67.55
13.90
0.26
0.05
0.095
0.02
432.88
89.00
7.40
1.52
0.08
0.02
0.05
0.01
Row totals
Dredged
Bank/shelf
escarpment
Bank/shelf
Forereef
Reef crest
Backreef
Lagoon
Shoreline/
intertidal
Zones
Habitats
0.09
0.02
13.36
2.75
103.52
21.30
123.06
25.32
23.94
4.92
0.13
76.82
0.03
15.81
2.13
0.44
1.59
2.74
0.33
0.56
7.41
1.52
1.34
0.27
0.45
0.09
0.29
21.11
0.06
4.34
24.10
4.89
0.42
83.64
0.09
17.21
0.69
0.14
0.34
0.07
0.38
0.08
0.10
0.02
2.43 486.00
0.50 100.00
area mapped), 2.1 km2 was mangroves (0.4% of area mapped), and 298.7 km2 was coral
reef and hardbottom (61.7% of area mapped; Fig. 2). The zone with the largest area was
bank/shelf (433.9 km2) followed distantly by lagoon (21.0 km2), fore reef (14.5 km2),
bank/shelf escarpment (7.4 km2), dredged (2.4 km2), shoreline/intertidal (3.6 km2), back
reef (1.9 km2), and reef crest (1.7 km2; Fig. 3). Table 2 summarizes the area mapped in
km2 and as the percentage of the total area covered by each zone/habitat combination.
For example, 14.8 km2 of “seagrass” was mapped in the lagoon zone, which covered 3%
of the total area mapped. The largest areas mapped occurred on the bank/shelf zone and
included “colonized pavement with sand channels” (24.8% of the total area mapped),
“colonized pavement” (20.8% of mapped area), seagrass (13.9% of mapped area), and
232
Table 3. Error matrix for habitat classification of the Buck Island area using on-screen digitizing.
Numbers in the matrix indicate class coincidence, (I) indicates inclusion errors, and (O) indicates
omission errors based on analysis of 109 ground truth points.
Mapped habitat type
Actual habitat type
Coral reef/hardbottom
Submerged vegetation
Unconsolidated sediment
Coral reef/
hardbottom
35
97.2% (I)
85.4% (O)
0
6
Submerged
vegetation
0
30
100% (I)
100% (O)
0
Unconsolidated
sediment
1
0
37
86.1% (I)
97.4% (O)
macroalgae (15.7% of mapped area). Together, these categories covered nearly 75% of
the total area mapped. In contrast, areas identified as occurring in the lagoon zone covered only 4.3% of the total area mapped excluding polygons delineated as land and
unknown.
Thematic Accuracy of On-screen vs Stereoplotter Digitizing.—Comparison
with the ground truth data revealed very similar levels of thematic accuracy between the
two maps (Congalton, 1991). Overall accuracy was 93.6% (Kappa 0.90) for on-screen
digitizing (Table 3) and 87.8% (Kappa 0.82) for maps digitized directly from stereo
pairs (Table 4). These findings indicate that on-screen digitizing results in similar or
even superior levels of thematic accuracy for maps produced at this scale with this type
of classification scheme.
Discussion
We have demonstrated the successful application of new software designed to facilitate benthic mapping which results in thematic accuracy that is comparable to other,
commonly used but more technically demanding mapping approaches. Indeed, using the
process and simplifying software outlined above, benthic maps can be quickly created
Table 4. Error matrix for habitat classification of the Buck Island area using an analytical
stereoplotter. Numbers in the matrix indicate class coincidence, (I) indicates inclusion errors, and
(O) indicates omission errors based on analysis of 98 ground truth points. Slightly fewer points
were used in this analysis since the extent of this map was smaller than the distribution of ground
truth points.
Mapped habitat type
Actual habitat type
Coral reef/hardbottom
Unconsolidated sediment
Coral reef/
hardbottom
35
0
92.1% (I)
89.7% (O)
3
25
75.8% (I)
100% (O)
1
0
Unconsolidated
sediment
3
5
26
96.3% (I)
76.5% (O)
233
using readily available software packages and a normal desktop computer (Kendall et
al., 2001); there is no need for the expensive and highly specialized equipment (e.g., analytical stereo plotter) associated with producing this type of map in the past. The simple
software and approach described here builds a mapping capability for individuals and
agencies with access to limited financial and computing resources and expertise.
Evaluation of the benthic maps of the U.S. Virgin Islands suggests that the dominant
benthic types in the area are coral reef and hard bottom followed by submerged vegetation and unconsolidated sediment. This type of information is critical for natural
resource managers responsible for protecting marine habitat. For example, the National
Action Plan to Conserve Coral Reefs produced by the U.S. Coral Reef Task Force identified five conservation objectives (USCRTF, 2000), the most substantial of which calls
for a phased in protection of 20% of U.S. coral reefs and associated habitats via no-take
ecological reserves by the year 2010. Clearly, accurate maps of the benthic seascape
are required to meet this objective and identify appropriate areas for protection that are
representative of regional coral reef ecosystems.
Just as important as the area covered by each benthic type (e.g., patch reef, submerged
vegetation) is knowledge of where those types occur in a spatial context. For example,
the vast majority of the mapped area in the U.S. Virgin Islands was classified as bank/
shelf with only small areas designated as back reef or lagoon; however, the low energy
environment of the lagoon, back reef, and bay habitats are commonly considered to be
important nursery areas for many reef organisms (Nagelkerken et al., 2000a,b). In addition, even though the lagoon and back reef represent a small percentage of the total
shelf their proximity to terrestrial processes means they often bear the brunt of deleterious human activities known to reduce the health of coral ecosystems such as sediment
laden runoff resulting from coastal construction (Rogers, 1990). These considerations
suggest that lagoon and back reef areas warrant special consideration for implementing
protective management even though they encompass a relatively small proportion of the
benthic resources of a region. The zone attribute of our benthic maps provides a useful
framework for evaluating reef structures and fishes within this spatial context.
Area measurements for each zone/habitat combination such as those provided here
for the U.S. Virgin Islands can provide useful guidelines for natural resource managers;
however, these numbers do not speak to the relative ecological importance of each of
the seascape elements. Only directed studies of fish assemblages coupled with analysis
of seascape maps can do so. Many excellent studies of fish communities associated with
a single habitat type exist (e.g., coral reef or seagrass respectively); however, relatively
few have conducted assessments of fish assemblages across the entire seascape of the
coral ecosystem (Parrish, 1989). One of the earliest works to couple such analyses for
coral reef ecosystems at a broad scale was performed in the Virgin Islands National
Park, Biosphere Reserve. In this area, Beets et al. (1986) produced maps of the benthic
seascape from aerial photography. Boulon (1986) then published a companion report
describing the fisheries resources associated with each of the habitat types identified and
mapped by Beets et al. More recently, Lindeman et al. (1998) provided another example
of an advance in reef ecosystem analysis by utilizing a more complex spatial framework
to describe the distribution of nursery areas for some reef fish of Biscayne Bay, Florida.
Both individual benthic types and cross-shelf zonation were used to create a matrix
similar to Table 2 but which was populated with occurrence patterns of young grunts
and snappers. Values in the matrix were then applied to maps of the benthic seascape to
identify the distribution of nursery habitat. Lindeman et al. (1998) was among the first to
234
clearly assert that structurally similar benthic features (e.g., patch reefs) are often home
to different assemblages of reef fish depending on their cross-shelf position.
A missing factor in all of these studies is the explicit consideration of the relative spatial configuration of habitat types in the seascape. An excellent review of the research
on interactions among seagrass, algal, mangrove, and reef communities was provided
by Parrish (1989). This paper culminated in a call for quantitative studies emphasizing
the trophic dynamics among and distances over which these neighboring habitats are
linked. Over a decade has since passed and few empirical studies on interactions among
habitats have been conducted. This must be partially attributed to the lack of benthic
maps for most regions. Fortunately, as we note in this study, the technical demands of
benthic mapping are becoming more manageable.
A recently completed analysis that coupled habitat maps and fish distributions demonstrates an important step in the evolution of this type of work by emphasizing the need
to consider the spatial configuration of the elements in the marine seascape. Kendall et
al. (2003) recently examined the distribution of juvenile French grunts, Haemulon flavolineatum (Desmarest, 1823), within the spatial framework defined by our maps of the
marine seascape at Buck Island. Haemulids are known to undergo foraging migrations
from daytime resting sites on hard bottom to soft bottom areas such as sand and seagrass
each night (Ogden and Zieman, 1977; Helfman et al., 1982). Therefore, the occurrence
of juvenile H. flavolineatum on hard bottom sites during the day was evaluated relative
to distance to nearest soft bottom and total area of soft bottom within nightly migration
distance. Logistic regression revealed that the probability of occurrence of juvenile H.
flavolineatum on hard bottom sites is positively correlated with the area of adjacent soft
bottom and negatively correlated with distance to adjacent soft bottom. The results of
this type of analysis further emphasize the need to consider the relative spatial position
of elements in the mosaic of benthic types in marine seascapes rather than relying on
simple correlation between organisms and single bottom types. Indeed, the distribution of coral reef organisms at any given site is often a function of not only the bottom
features at that specific site but also of bottom features present some distance away, a
concept that has received much discussion in literature of coral ecosystems (Parrish,
1989) but for which little empirical evidence has been demonstrated due to the lack of
adequate habitat maps to use for spatial analyses.
The management of coral reef ecosystems is in part a spatial problem. The ability to
make informed decisions on defining essential fish habitats as discussed above, on the
placement of marine protected areas, and on optimal location for coastal development
is severely limited without highly resolved and accurate benthic maps created using an
appropriate classification framework. In fact, the National Action Plan for U.S. coral reef
protection endorsed by USCRTF identified mapping of all U.S. coral reefs as one of the
highest priorities for protecting these important ecosystems (USCRTF, 2000). When
these types of maps are combined with in situ monitoring programs, a robust assessment
capability is formulated to support management needs (Monaco et al., 2001). As the use
of marine refuges continues to increase, it is important that they be based on biologically
relevant boundaries. The analysis of fish census data in the context of the spatial framework provided by maps of the benthic seascape provides a significant improvement on
efforts to define fish distributions.
235
Acknowledgements
Project #CMRC-00-IXNR-03-01A) National Oceanic and Atmospheric Administration (NOAA)
National Undersea Research Program, U.S. Environmental Protection Agency, and Environmental Defense. Views expressed herein are those of the authors and do not necessarily reflect the
views of CMRC, or any of the supporting agencies. NOS coral mapping activities can only be
conducted through a suite of federal, state, local, territory, commonwealth, academic, and private
sector partners. We thank all of these entities and especially our colleagues in the NOS/NCCOS Biogeography Program and the NOS Coastal Services Center in development of technical
mapping approaches. We especially thank our colleagues in the National Geodetic Survey who
conducted the aerial photography missions in the Caribbean. In addition, special thanks to Z.
Hillis-Starr, B. L. Phillips, J. Tutien, and other staff at the Buck Island Reef National Monument
for devoting their time, facilities, and advice.
Literature Cited
Adey, W. H. 1979. Maps of U.S. Virgin Islands reef habitats for the sediment reduction program.
Department of Conservation and Cultural Affairs, U.S. Virgin Islands Government. Prepared
by CH2Mhill, Consulting Engineers, Gainesville.
Alevison, W., R. Richardson, P. Pitts, and G. Serviss. 1985. Coral zonation and patterns of community structure in Bahamian reef fish. Bull. Mar. Sci. 36: 304–318.
Bacle, J. P. 1995. Mapping coastal habitat features Great Pond Bay, St.Croix, U.S. Virgin Islands.
Photo-interpretation. 4: 264–268.
Beets, J., L. Leewand, and E. S. Zullo. 1986. Marine community descriptions and maps of bays
within the Virgin Islands National Park/Biosphere Reserve. Biosphere Reserve Research Report Number 2. National Park Service, Department of Interior, Washington DC. 118 p.
Boulon, R. H. 1986. Distribution of fisheries habitats within the Virgin Islands Biosphere Reserve.
Biosphere Reserve Research Report Number 8, National Park Service, Department of Interior,
Washington DC. 56 p.
Brandt, S. B. and J. Kirsch. 1993. Spatially explicit models of striped bass growth potential in
Chesapeake Bay. Trans. Amer. Fish. Soc. 122: 845–869.
Burke, N. C. 1995. Nocturnal foraging habitats of French and bluestriped grunts, Haemulon flavolineatum and H.sciurus, at Tobacco Caye, Belize. Env. Biol. Fish. 42: 365–374.
Chauvaud, S., C. Bouchon, and R. Maniere. 1998. Remote sensing techniques adapted to high
resolution mapping of tropical coastal marine ecosystems (coral reefs, seagrass beds, and mangrove). Intl. J. Remote Sens. 19: 3525–3639.
Congalton, R. G. 1991. A review of assessing the accuracy of classifications of remotely sensed
data. Remote. Sens. Environ. 37: 35–46.
Connel, S. D. 1996. Variations in mortality of a coral-reef fish: links with predator abundance. Mar.
Biol. 126: 347–352.
Dobson, J. E., E. A. Bright, R. L. Ferguson, D. W. Field, L. L. Wood, K. H. Haddad, H. Iredale III,
J. R. Jensen, V. V. Klemas, R. J. Orth, and J. P. Thomas. 1995. NOAA Coastal Change Analysis
Program (C-CAP): Guidance for regional implementation. NOAA Technical Report, National
Marine Fisheries Service 123. Department of Commerce.
Florida Fish and Wildlife Conservation Commission, Florida Marine Research Institute (FMRI),
and National Oceanic and Atmospheric Administration. 1998. Benthic Habitats of the Florida
Keys. FMRI Technical Report No. TR-4. 52 p.
Helfman, G. S., J. L. Meyer, and W. N. McFarland. 1982. The ontogeny of twilight migration patterns in grunts (pisces: Haemulidae). Anim. Behav. 30: 317–326.
Holthus, P. F. and J. E. Maragos. 1995. Marine ecosystem classification for the tropical island Pacific. Pages 239–278 in J.E. Maragos, M.N.A. Peterson, L.G. Eldredge, J.E. Bardach, and H.F.
236
Takeuchi, eds. Marine and coastal biodiversity in the Tropical Island Pacific Region. East-West
Center, Honolulu.
Kendall, M. S., C. R. Kruer, K. R. Buja, J. D. Christensen, M. Finkbeiner, R. Warner, and M. E.
Monaco. 2001. Methods used to map the benthic habitats of Puerto Rico and the U.S. Virgin
Islands. NOAA NOS NCCOS CCMA Tech. Rpt. 152. National Oceanic and Atmospheric Administration, Silver Spring.
Kendall, M. S., J. D. Christensen, and Z. Hillis-Starr. 2003. Multi-scale data used to analyze the
spatial distribution of French grunts, Haemulon flavolineatum, relative to hard and soft bottom
in a benthic seascape. Env. Biol. Fish. 66: 19–26.
Kumpf, H. E. and H. A. Randall. 1961. Charting the marine environments of St. John, U.S.Virgin
Islands. Bull. Mar. Sci. 11: 543–551.
Kruer, C. 1995. Mapping and characterizing seagrass areas important to manatees in Puerto RicoBenthic communities Mapping and Assessment. Report prepared for Department of Interior,
Natl. Biol. Surv., Sirenia Project, Order No. 83023-5-0161. 14 p.
Letourneur, Y. 1996. Dynamics of fish communities on Reunion fringing reefs, Indian Ocean. I.
Patterns of spatial distribution. J. Exp. Mar. Biol. Ecol. 195: 1–30.
Lindeman, K. C., G. A. Diaz, J. E. Serafy, and J. S. Ault. 1998. A spatial framework for assessing
cross-shelf habitat use among newly settled grunts and snappers. Proc. Gulf Carib. Fish. Inst.
50: 385–416.
McGehee, M. A. 1994. Correspondence between assemblages of coral reef fishes and gradients of
water motion, depth, and substrate size off Puerto Rico. Mar. Ecol. Prog. Ser. 105: 243–255.
Meyer, J. L. and E. T. Shultz. 1985. Migrating haemulid fishes as a source of nutrients and organic
matter on coral reefs. Limnol. Oceanogr. 30: 146–156.
_________, ______________, and G. S. Helfman. 1983. Fish schools: an asset to corals. Science.
220: 1047–1049
Monaco, M. E., J. D. Christensen, and S. O. Rohmann. 2001. Mapping and monitoring of US coral
reef ecosystems. Earth Syst. Monitor. 12: 1–7.
Mumby, P. J., A. R. Harborne, and P. S. Raines. 1998. Draft classification scheme for marine habitats of Belize. UNDP/GEF Belize Coastal Zone Management Project. 44 p.
NOAA, US Geological Survey, Florida Fish, and Wildlife Conservation Commission and Florida
Marine Research Institute. 1998. Seagrass and aquatic habitat assessment workshop summary,
and accompanying participant survey data. July 28-29, 1998 Technical Workshop at University
of South Florida, St. Petersburg. 22 p.
NOAA, U.S. Environmental Protection Agency, U.S. Coast Guard, Puerto Rico Departmento de
Recursos Naturales y Ambientales, and Department of Interior. 2000. Sensitivity of coastal
and inland resources to spilled oil - Puerto Rico atlas. Hazardous Materials Response Division
(NOAA), Seattle. 68 p.
Nagelkerken, I., M.Dorenbosch, W.Verberk, E.Cocheret de la Moriniere, and G. van der Velde.
2000a. Importance 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.
2000b. Importance of mangroves, seagrass beds and the shallow coral reefs as a nursery for important coral reef fishes, using a visual census technique. Est. Coast. and Shelf Sci. 51: 31–44
Ogden, J. C. 1976. Some aspects of herbivore-plant relationships on Caribbean reefs and seagrass
beds. Aquat. Bot. 2: 103–116.
_________ and J. C. Zieman. 1977. Ecological aspects of coral reef-seagrass bed contacts in the
Caribbean. Proc. Third Annual Coral Reef Symposium. 377–382 p.
Randall, J. E. 1967. Food habits of reef fishes of the West Indies. Proc. Intern. Conf. on Trop.
Oceanogr. 5: 665–847.
__________. 1996. Caribbean reef fishes, 3rd ed. TFH Publications, Neptune City. 368 p.
237
Reid, J. P. and C. R. Kruer. 1998. Mapping and characterization of nearshore benthic habitats
around Vieques Island, Puerto Rico. Report to U.S. Navy. U.S. Geological Survey/BRD, Sirenia Project, Gainesville. 11p.
Rogers, C. S. 1990. Responses of coral reefs and reef organisms to sedimentation. Review. Mar.
Ecol. Prog. Ser. 62: 185–202.
Sale, P. F. 1998. Appropriate spatial scales for studies of reef-fish ecology. Aust. J. Ecol. 23: 202–
208.
Sheppard, C. R., K. Matheson, J. C. Bythel, P. Murphy, C. B.Myers, and B. Blake. 1998. Habitat
mapping in the Caribbean for management and conservation: use and assessment of aerial
photography. Aquat. Cons. 5: 277–298.
Shulman, M. J. 1985. Recruitment of a coral reef fish: effects of the spatial distribution of predators
and shelter. Ecol. 66: 1056–1066.
Tribble, G. W. 1981. Reef-based herbivores and the distributin of two-seagrasses (Syringodium
filiforme and Thalassia testudinum) in the San Blas Islands (Western Caribbean). Mar. Biol.
65: 277–281.
USCRTF. 2000. The National Acton Plan to Conserve Coral Reefs. Washington, D.C. 34 p + appendices. (<http://coralreef.gov/>).
Vierros, M. K. 1997. Integrating multisource imagery and GIS analysis for mapping Bermuda’s
benthic habitats. Presented at the 4th International Conference on Remote Sensing for Marine
and Coastal Environments. Orlando. March 997, I: 649–656.
Addresses: (M.S.K., K.R.B., J.D.C., M.E.M.) NOAA/NOS, National Centers for Coastal Ocean
Science, Biogeography Program N/SCI-1, 1305 East West Highway, Silver Spring, Maryland
20910. (C.R.K.) P.O. Box 753, Sheridan, Montana 59749. Corresponding Author: (M.S.K.)
E-mail: <matt.kendall@noaa.gov>.
THE IMPACT OF HURRICANE GEORGES ON SOFT-BOTTOM,
BACK REEF COMMUNITIES: SITE- AND SPECIES-SPECIFIC
EFFECTS IN SOUTH FLORIDA SEAGRASS BEDS
James W. Fourqurean and Leanne M. Rutten
ABSTRACT
Seagrass beds are the dominant benthic marine communities in the back reef environment of the Florida Keys. At a network of 30 permanent monitoring stations in this
back reef environment, the seagrass Thalassia testudinum Banks & Soland. ex Koenig
was the most common marine macrophyte, but the seagrasses Syringodium filiforme
Kuetz., and Halodule wrightii Aschers., as well as many taxa of macroalgae, were also
commonly encountered. The calcareous green macroalgae, especially Halimeda spp.
and Penicillus spp., were the most common macroalgae. The passage of Hurricane
Georges on September 25, 1998 caused an immediate loss of 3% of the density of T.
testudinum, compared to 19% of the S. filiforme and 24% of the calcareous green algae.
The seagrass beds at three of the stations were completely obliterated by the storm. Stations that had little to moderate sediment deposition recovered from the storm within 1
yr, while the station buried by 50 cm of sediment and the two stations that experienced
substantial erosion had recovered very little during the 3 yrs after the storm. Early colonizers to these severely disturbed sites were calcareous green algae. Hurricanes may increase benthic macrophyte diversity by creating disturbed patches with the landscape,
but moderate storm disturbance may actually reduce macrophyte diversity by removing
the early successional species from mixed-species seagrass beds.
Hurricanes and cyclones are important structuring forces in nearshore marine communities. High winds, large waves, torrential rains, and storm surge impact submarine
communities like seagrass beds and coral reefs, emergent plant communities like salt
marshes and mangroves, and upland ecosystems. Tropical cyclones are an often-cited
cause of damage to seagrass beds. For example, 70% of the seagrass cover in the Gulf
of Carpinteria, Australia, was uprooted by the 12 m waves generated by the 220 km h−1
winds of Cyclone Sandy (Poiner et al., 1989), and over 1000 km2 of seagrass beds were
lost due to wave action, flooding and turbidity caused by a cyclone in Hervey Bay, Australia (Preen et al., 1995).
In the Gulf of Mexico/Caribbean region, smaller-scale hurricane impacts to seagrass
beds are commonly reported (e.g., Glynn et al., 1964; Wanless et al., 1988; Rodríguez et
al., 1994; van Tussenbroek, 1994), but it is interesting to note that some of the most severe storms in south Florida during the past half-century have had relatively little impact
on seagrass beds. Hurricane Donna, a category 4 hurricane which passed directly over
the Florida Keys in 1960, had negligible impact on seagrass biomass and mud bank topography of Florida Bay (Ball et al., 1967), and despite depositing 2 m deep wracklines
of seagrass blades along the shore, had little impact on seagrass distribution in Biscayne
Bay (Thomas et al., 1961). Hurricane Andrew, another category 4 storm in 1992, had a
devastating impact on the mangrove forests and terrestrial ecosystems of south Florida
(Smith et al., 1994; Armentano et al., 1995; Doyle et al., 1995; McCoy et al., 1996), but
caused little damage to seagrass beds immediately seaward of the mangroves (Tilmant et
al., 1994). It is apparent that factors other than maximum wind speed of a hurricane are
important determinants of the degree to which a hurricane will affect seagrass beds.
239
240
Some have proposed a positive impact of mechanical storm damage on seagrass beds
over the long term. Moderate disturbance may help maintain species diversity in ecosystems (Connell and Slatyer, 1977; Paine and Levin, 1981). Ecological succession in seagrass beds of the tropical Atlantic tends towards a seagrass bed dominated by Thalassia
testudinum Banks & Soland. ex Koenig (Den Hartog, 1971; Zieman, 1982). Storm-generated bare patches in dense T. testudinum meadows are colonized by early successional
macrophytes, like calcareous green algae and the seagrasses Halodule wrightii Kuetz.
and Syringodium filiforme Aschers., thereby enhancing the diversity of these disturbed
beds (Patriquin, 1975). In south Florida, mixed-species seagrass beds support higher
densities of fish and invertebrates than T. testudinum -dominated seagrass beds (Thayer
and Chester, 1989). In addition to this role in maintaining diversity, it has even been
proposed that periodic removal of sediment, organic matter, and biomass by hurricanes
can be necessary for the long-term homeostasis of some seagrass beds (Zieman et al.,
1989).
The time required for seagrass beds to recover from a disturbance is a function of
the size of the disturbance. Recolonization of even a small gap (<1 m2) in a seagrass
bed can take up to a decade, since most seagrass colonization occurs through vegetative growth into the gap instead of through seedling recruitment (Zieman, 1976; Rollon
et al., 1998). Dynamic erosional features <10 m wide in T. testudinum beds caused by
storm events are colonized by calcareous algae, H. wrightii, and S. filiforme years before
T. testudinum can become reestablished (Patriquin, 1975). Larger disturbances, whether
caused by tropical cyclones, mass die-off (Robblee et al., 1991; Seddon et al., 2000), catastrophic grazing (Maciá and Lirman, 1999; Rose et al., 1999; Peterson et al., 2002), or
any other factor, may persist for many decades. Studying the dynamics of recovery from
such a large-scale disturbance is made difficult by this time scale, but understanding the
rate and mechanisms of recovery of seagrass beds from such disturbances is important
not just for the scientific aspects of the exercise, but also for aiding in guiding restoration
and management of damaged seagrass ecosystems.
Studying the effects of hurricanes is also problematic because of the unpredictability
of their occurrence. Since 1920, 24 hurricanes have passed over the Florida Keys National Marine Sanctuary, but those hurricanes have not been equally distributed over
those 81 yrs (data from the U.S National Oceanographic and Atmospheric Administration’s National Hurricane Center, <www.nhc.noaa.gov>). The 1940s and 1960s were
particularly active times for hurricanes in the region, with six hurricanes in each decade.
In contrast, there was only one hurricane (Floyd, a small, category 1 storm in 1987)
during the period 1970–1991, and since 1992 there have been three hurricanes. One reliable way to ensure that the effects of hurricanes on the living resources of an area can
be measured is to establish a spatially comprehensive, long-term monitoring program.
Such a program will increase the chances of a storm influencing at least some areas of
known character during the course of the program. Further, a long-lived program is necessary to assess the long-term impacts of the hurricane disturbance. Such a monitoring
program was established in 1995 as part of the Water Quality Protection Plan for the
Florida Keys National Marine Sanctuary (FKNMS). In this paper, we examine the data
from 6 yrs of this monitoring program to describe the relative abundance of seagrasses
and macroalgae in south Florida and to assess the immediate impacts and first 3 yrs of
recovery of benthic soft-bottom marine communities from the passage of a hurricane.
This publication is part of a series of papers resulting from a scientific workshop held at
FOURQUREAN AND RUTTEN: HURRICANE IMPACTS IN SEAGRASS BEDS
241
Methods
Site Description.—The FKNMS encompasses about 9000 km2 of near-shore, subtropical
marine communities (Fig. 1). Over most of this area, the water is very shallow (<20 m); the
Florida Keys barrier coral reef on the Atlantic Ocean side of the Florida Keys parallels the islands
7–10 km offshore. Between the barrier reef and the islands, and on the Gulf of Mexico side of the
islands, lie extensive seagrass beds. There are some relatively small areas where active patch reefs
and the lack of sediment preclude the presence of seagrass beds, but seagrass communities dominate over 80% of the bottom between the barrier reef and the islands (Fourqurean et al., 2001,
2002). Although these communities are called seagrass beds, it is important to note that they
contain a wide variety of benthic plants and animals in addition to the seagrasses themselves;
in fact, macroalgal primary production may rival seagrass primary production over much of the
backreef environment of the Florida Keys (Davis and Fourqurean, 2001).
Hurricanes in the Study Area, 1995–2001.—Fortunately for the residents of south Florida,
few hurricanes have passed over the study area since 1995, and these (Georges in 1998, Irene in
1999) were relatively mild over the Florida Keys. Hurricane Georges was the most intense storm
during this period; it was a category 2 when it passed over Key West on September 25, 1998 (see
storm track on Fig. 1, inset). The hurricane force winds were restricted to a relatively narrow
band in the lower and middle Florida Keys. Maximum sustained winds of 90 knots (145 km h−1)
were recorded in Key West, 81 knots (130 km h−1) offshore of Marathon, but only 46 knots (75
km h−1) offshore in the upper Keys. Irene was a small category 1 hurricane when it passed over
the upper Keys in 1999, maximum sustained winds were not recorded over 57 knots (92 km h−1).
These relatively small hurricanes did have very substantial, but localized, effects on the natural
resources and economic interests Florida Keys.
Station Selection.—In order to assess the dynamics of the soft-bottom benthic communities
in the back-reef environment in south Florida, data were gathered from the network of 30 permanent seagrass monitoring stations in the FKNMS (Fig. 1, see Fourqurean et al., 2001; Fourqurean
and Rutten, 2003 for a description of the monitoring program). These stations were originally
located using a stratified-random approach, with two random locations being chosen within three
strata (inshore, offshore, and intermediate) in each of five defined segments of the Sanctuary
(Klein and Orlando, 1994). At each seagrass monitoring station, a permanent 50 m long transect
was established at the beginning of the study period by driving steel rods into the substratum at
both ends of the transect. Data on the species composition and abundance of the benthic macrophyte community was collected from these stations four times per year, beginning in December
1995. By serendipity, a survey was completed the week prior to the passage of Hurricane Georges
on September 25, 1998. In response to this storm, all 30 stations were resurveyed the week after
the storm passed. Thereafter, surveys continued four times per year until the end of 2001. During
these resurveys, recent sediment deposition and/or erosion were noted. For cases of minor deposition, the depth of recent sediments above the previous sediment horizon and covering the seagrass
blades was measured. Bare areas in the seagrass beds created by the storm were excavated, and
the depth to the rhizome layer of the previously extant seagrass community was recorded if there
had been deposition. In the event of erosion, the amount of erosion was estimated by examining
the remaining seagrass short shoots for signs of previous sediment depth (often, this is indicated
by the location of chlorophyll in outer leaves in a short shoot); if the previous seagrass community
was completely missing and not buried, a rough estimate of the extent of erosion was made by
examining the depth of the rhizomes of nearby but unaffected seagrass beds.
Braun Blanquet Monitoring.—We assessed the species composition and relative abundance of benthic macrophytes at the permanent monitoring stations using Braun-Blanquet coverabundance surveys (Fourqurean et al., 2001, 2002). In order to codify the data collection, a list of
taxa was established to define the monitoring targets (Table 1). In some instances, the taxa consist
242
Figure 1. Map of study area, showing geographic location, position of monitoring stations, track
of Hurricane Georges relative to study area, and maximum sustained winds experienced during the passage of Hurricane Georges on September 25, 1998 in Key West, Marathon, and Key
Largo.
of single species (e.g., the seagrasses T. testudinum, S. filiforme, etc.), while other taxa were more
broadly defined to simplify the in situ data collection. All conspicuous macrophytes fell into
one of the categories in Table 2. Ten quadrats (0.25 m2) were placed along each 50 m transect
at predetermined random distances from one of the marker rods. A new set of random sampling
positions were chosen before each visit to a station. Each quadrat was examined using SCUBA.
For each quadrat, the taxa occurring in the quadrat were listed, and a score based on the cover of
each taxon in that quadrat was assigned (Table 2).
From the observations of cover in each quadrat at a station, three statistics were computed for
each taxon: frequency, abundance, and density. Frequency was calculated as
Fi = N i n
Eq. 1
where Fi = the frequency of taxon i, n = the total number of quadrats sampled at a station, and Ni
is the number of quadrats at a station in which taxon i was present; such that 0 ≤ Fi ≤ 1.
Abundance was calculated as
n
Ai = ∑ Siijj N i
j =1
Eq. 2
where Ai = the abundance of taxon i, j = quadrat number from 1 to n, the total number of quadrats
sampled at a station; Sij = the Braun-Blanquet score for taxon i in quadrat j; and Ni is the number
of quadrats at a station in which taxon i was present. For any taxon, A can range between 0 and 5,
the maximum Braun-Blanquet score.
Density was calculated as
243
Table 1. Taxa list used to describe macrophyte species composition of soft-bottom back reef areas
in south Florida.
Code
Seagrasses:
T
S
H
Hd
He
Macroalgae:
CGH
CGU
CGP
CGR
CGA
CGO
CGT
CA
GO
DR
CR
RO
SAR
BO
O
Taxa
Example species
Thalassia testudinum
Syringodium filiforme
Halodule wrightii
Halophila decipiens
Halophila engelmanni
Calcareous green Halimeda spp.
Calcareous green Udotea spp.
Calcareous green Penicillus spp.
Calcareous green Rhipocephalus spp.
Calcareous green Acetabularia spp.
Calcareous green other
Calcareous green total, all species
Caulerpa spp.
Green other
Drift red
Coralline red
Red other
Sargassum spp.
Brown other
Other, unidentified
Halimeda incrassata
Udotea flabellum
Penicillus capitatus
Rhipocephalus phoenix
Acetabularia crenulata
Cymopolia barbata
Caulerpa prolifera
Anadyomene stellata
Laurencia intricata
Jania adhaerens
Heterosiphonia gibbesii
Sargassum natans
Dictyota cervivornis
n
Di = ∑ Sij n
j =1
Eq. 3
where Di = density of taxon i; j = quadrat number from 1 to n, the total number of quadrats
sampled at a station, and Sij = the Braun-Blanquet score for taxon i in quadrat j. For any species,
D can range between 0 and 5, the maximum Braun-Blanquet score (note that Di = Fi × Ai, and Di ≤
Ai). At a station, however, the sum of all taxa D values can be greater than 5, becausse of the relatively broad cover ranges for each Braun-Blanquet value and the fact that seagrass canopies are
three dimensional. It should also be noted that a taxon may be observed at a station by the sample
collector, but unless the taxon falls within one of the randomly-placed observation quadrats, the
taxon receives a D = 0.
Changes in benthic macrophyte abundance caused by the storm were quantified by calculatiing
the fractional change (∆) in the Braun-Blanquet A, F and D scores for a taxon at each site:
∆ i=
∆X
Xi aafte
fter − Xi before
Xi bbefore
Eq. 4
where X represents one of the three Braun Blanquet metrics ((A,
A, F, or D) for taxon i either immediately before or after the passage of Georges. To facilitate the visualization of the spatial pattern
in storm-induced change, contour plots of ΔD for the three most common taxa (T. testudinum, S.
filiforme, and Calcareous Green Algae) were generated using a kriging algorithm (point kriging
using a linear variogram and no nugget).
244
Table 2. Braun-Blanquet abundance scale used to assess seagrass density. Cover is defined as
the fraction of the bottom that is obscured by the species when viewed by a diver from directly
above.
Cover class
0
0.1
0.5
1
2
3
4
5
Description
Absent
Solitary individual or ramet, less than 5% cover
Few individuals or ramets, less than 5% cover
Many individuals or ramets, less than 5% cover
5–25% cover
25–50% cover
50–75% cover
75–100% cover
Results
Thalassia testudinum was the most common macrophyte encountered on the BraunBlanquet surveys, it was found on 96% of the 690 transects samples from the 30 stations
over the period December 1995–December 2001 (Table 3). On those surveys, it was
found with a mean F of 0.77, but a median F of 1.00, since it was found in every quadrat
of more than half of all transects (Fig. 2). It was often found with high A, and it had the
highest mean A (2.02) of any of the monitored taxa. Of the other seagrass species, S. filiforme was the next most common taxa, occurring on 64% of all transects. Syringodium
filiforme was rarely encountered with high A; mean A was 1.33 while the median A was
only 1.00. Halodule wrightii was found on only 14% of transects, with a mean A of 0.14.
Halophila spp. were rarely encountered. Thalassia testudinum and S. filiforme tended
to occur as continuous, the most common F class for each was 90–100% cover (Fig. 2).
In contrast, H. wrightii was most often distributed in small patches; it was most often
observed with a frequency of 0–10% when it was present on transects.
Calcareous green algae were by far the most common macroalgae encountered on the
Braun-Blanquet surveys, with at least one taxon being recorded on 92% of the surveys
(Fig. 2). Halimeda spp. and Penicillus spp. were the most common calcareous green
algae; both occurred on ca. 80% of all transects. Both “Other Red Algae” and Caulerpa
spp. were relatively common, occurring on 48 and 24% of all transects, respectively.
Macroalgae rarely had A scores over 2, indicating they rarely covered more than 25%
of a quadrat, but occasionally Halimeda spp., Penicillus spp., or red algae were found to
cover the majority of a quadrat (Table 3).
The Braun-Blanquet surveys produced data of sufficient precision to identify intra-annual patterns in the relative abundance of the target taxa (Fig. 3). Station 223 was a dense
T. testudinum bed, with an understory of other macrophyte taxa. Thalassia testudinum
was always found in every quadrat at this station (F = 1). The most common taxa in the
understory at this station were S. filiforme and calcareous green algae (CGT), but the
frequency of encountering these understory plants followed a pronounced seasonal pattern, with summertime peaks in F for both taxa. Peaks for CGT were ca. 0.6, and peaks
for S. filiforme were around 0.2. Abundance of T. testudinum peaked around A = 4.2
each summer, while the common understory taxa had peaks in abundance ≤ 1. Note that
there was no measurable alteration of the seasonal patterns in frequency, abundance, or
density associated with the passage of Georges at this station.
In contrast to station 223 (Fig. 3), at many stations there was a marked decrease in
0.96
0.64
0.14
0.01
0.00
0.81
0.62
0.78
0.41
0.18
0.08
0.92
0.24
0.34
0.19
0.13
0.48
0.02
0.26
Thalassia testudinum
Halodule wrightii
Halophila engelmanni
Halophila decipiens
Calcareous green algae:
Halimeda spp.
Udotea spp.
Penicillus spp.
Ripocephalus spp.
Acetabularia spp.
Calcareous green other
Calcareous green total
Caulerpa spp.
Green other
Drift red
Coralline red
Red other
Sargassum spp.
Brown other
Occurrence
0.04
0.11
0.06
0.04
0.21
0.00
0.08
0.51
0.26
0.43
0.10
0.08
0.02
0.70
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.80
0.70
1.00
1.00
1.00
1.00
0.20
1.00
1.00
1.00
1.00
1.00
1.00
0.90
1.00
Frequency (F)
F
F)
Mean Median Max
0.77
1.00
1.00
0.48
0.80
1.00
0.03
0.00
0.80
0.00
0.00
0.20
0.00
0.00
0.10
0.16
0.18
0.20
0.12
0.41
0.01
0.17
0.65
0.25
0.46
0.13
0.08
0.07
0.97
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.66
Abundance ((A)
Mean
Median
2.02
1.88
1.33
1.00
0.14
0.00
0.00
0.00
0.00
0.00
2.00
2.44
4.00
3.00
3.50
1.00
3.71
5.00
1.78
3.00
2.00
1.65
4.50
4.50
Max
5.00
5.00
3.00
2.00
0.10
0.03
0.07
0.08
0.04
0.22
0.00
0.07
0.44
0.13
0.31
0.03
0.04
0.02
0.83
Mean
1.74
1.14
0.03
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.51
Density (D)
Median
1.80
0.75
0.00
0.00
0.00
1.20
2.20
3.50
1.75
3.50
0.15
2.60
2.35
1.60
2.70
0.67
1.65
2.05
3.70
Max
5.00
5.00
1.40
0.20
0.01
Table 3. Summary statistics of the Braun-Blanquet survey data for the surveyed taxa. The minimum score for frequency, abundance, and density for all taxa was 0. Occurrence is the proportion of all transects (n = 691) on which the taxa was present.
245
Figure 2. Histograms of observed Braun Blanquet frequency (F), abundance (A) and density (D) for the six most common macrophyte taxa. Data were collected
four times a year during the period December 1995–December 2001 from the 30 monitoring stations (Fig. 1). The total number of transects was 691. Top row:
Seagrass species. Bottom row: macroalgal taxa.
246
247
Figure 3. Example time series of Braun-Blanquet monitoring data from station 223, which was
not affected by the passage of Hurricane Georges on September 25, 1998 (dashed vertical line).
This station was located at 25º 03.613ʹ N, 80º 25.382ʹ W, and was 3.5 m deep.
248
Table 4. Mean change in Braun-Blanquet frequency (∆F),
F),
F
), abundance (∆
(∆A
A), and density (∆D)
of the four most common taxa of benthic macrophytes at the 30 monitoring stations as a result
of Hurricane Georges. Change at a station was computed as [(score after storm − score before
storm) /score before storm]. Values are means ± 1 SE for the 30 stations.
Taxon
Thalassia testudium
Halodule wrightii
Calcareous green algae
Frequency (∆F)
F
F)
0.00 ± 0.11
−0.11 ± 0.11
−0.10 ± 0.11
−0.17 ± 0.09
Abundance (∆
(∆A)
−0.09 ± 0.08
−0.19 ± 0.08
−0.08 ± 0.11
−0.24 ± 0.10
Density (∆D)
−0.03 ± 0.09
−0.19 ± 0.08
−0.11 ± 0.12
−0.24 ± 0.12
the frequency, abundance, and/or density of benthic macrophytes between the surveys
conducted immediately before and after the passage of Georges. Averaging across all 30
stations, the average immediate change in F for T. testudinum was near zero, because of
small-scale patchiness in its distribution and the fact that independent observations of F
were made before and after the storm. However, there was a decrease in T. testudinum A
by an average of 9 ± 8% (± 1 standard error) and a decrease in D by 3 ± 9% (Table 4). Because D = F × A, the apparent variability in F for T. testudinum obscured any net change
in D across all stations. The storm had a similar impact on H. wrightii, which decreased
by ca. 10% for F, A, and D. Syringodium filiforme was affected to a greater degree, with
a 11 ± 0.07% loss in F, 19 ± 8% loss in A, and 19 ± 8% loss in D. Calcareous green algae
as a group showed the greatest immediate net loss because of the storm: F decreased by
17 ± 9%, A by 24 ± 10%, and D decreased by 24 ± 12% on average. We did not compute
the average changes of other less common taxa, because the random chance of encountering these other taxa on our surveys made estimates of net change unreliable.
There was a spatial pattern in the degree of loss caused by the storm for most taxa;
in general losses were greater in the western parts of our study area, and in the east-
Figure 4. Spatial pattern in the short-term loss in density (∆D) of the seagrass Thalassia testudinum across the study area, based on surveys made the week before and the week after the passage
of Hurricane Georges on September 25, 1998. Loss was calculated as [∆D = (Dafter-Dbefore)/Dbefore].
Contours generated by interpolating between values from the 30 permanent monitoring stations,
indicated by crosses.
249
Figure 5. Spatial pattern in the short-term loss in density of the seagrass Syringodium filiforme
(∆D) across the study area, based on surveys made the week before and the week after the passage
of Hurricane Georges on September 25, 1998. Loss was calculated as [∆D = (Dafter-Dbefore)/Dbefore].
Figure 6. Spatial pattern in the short-term loss in density (∆D) of calcareous green algae (CGA)
across the study area, based on surveys made the week before and the week after the passage of
Hurricane Georges on September 25, 1998. Loss was calculated as [∆D = (Dafter-Dbefore)/Dbefore].
250
Figure 7. Time series of the frequency, abundance and density of Thalassia testudinum, Syringodium filiforme, Halodule wrightii, and calcareous green algae (CGA) at stations 216 and 243,
which experienced significant erosion during the passage of Hurricane Georges on September 25,
1998 (dashed vertical line). Station 216 was located at 25º 10.772ʹ N, 80º 15.322ʹ W, and was 10.6
m deep. Station 243 was located at 24º 44.588ʹ N, 80º 47.779ʹ W, and was 9.2 m deep.
ern parts of our study area only offshore stations experienced large losses. There was
complete loss of T. testudinum from three stations (stations 216 and 243 on the Atlantic
Ocean side of the Florida Keys, and station 309 on the Gulf of Mexico side of the Keys;
Fig. 4). These large losses in D occurred mostly on the periphery of our study area, and
there tended to be little change in the benthic communities close to shore. Note also that
there were other stations along the periphery that exhibited no change during the storm,
even in the westernmost parts of our study area that were exposed to the strongest winds.
The same general spatial pattern was evident in the loss of S. filiforme during the storm
(Fig. 5), but the magnitude of the losses tended to be higher than for T. testudinum.
Losses of CGT were both greater in magnitude and more widespread than the losses of
the seagrass species (Fig. 6).
There were three causes of loss in macrophyte density observed on our post-storm
surveys: thinning of the canopy by removal of standing leaves of seagrass or thalli of
macroalgae, burial of the entire benthic community, or erosion of many centimeters of
sediment and consequently the removal of the community. Thinning was responsible
for the moderate losses of seagrass and macroalgae density, but burial and erosion were
responsible for the severe losses. Erosion was the factor that removed the seagrass beds
at the two offshore stations (216 and 243). The effect of the storm at these two stations
was dramatic and long-lived (Fig. 7). Station 216 was a deep (10.6 m) seagrass bed with
low-density, continuous cover of S. filiforme with occasional shoots of T. testudinum;
calcareous green algae were seasonally common but not very abundant before the storm.
251
Figure 8. Time series of the frequency, abundance and density of Thalassia testudinum, Syringodium filiforme, Halodule wrightii, and calcareous green algae (CGA) at stations 291 and 309,
which experienced significant sediment deposition during the passage of Hurricane Georges on
September 25, 1998 (dashed vertical line). Station 291 was located at 24º 49.008ʹ N, 81º 08.828ʹ
W, and was 2.7 m deep. Station 309 was located at 24º 41.907ʹ N, 81º 40.967ʹ W, and is 4.5 m
deep.
Immediately after the storm, the station was an unvegetated sand plain, with no macrophytes of any kind. Since the storm, calcareous green algae have become reestablished,
but other taxa have not recolonized the area. Station 243 was also deep (9.2 m), with a
denser but patchy seagrass bed with an even mixture of the seagrasses T. testudinum
and S. filiforme. Calcareous green algae were seasonally common, but not abundant.
The storm eroded many of the patches of seagrass away, and reduced the size of many
others, so that the probability of randomly sampling a patch was very low immediately
following the storm. Since the storm, calcareous green algae have reestablished, and the
remnant seagrass patches have regained their previous abundance and the patches are
growing in size.
In contrast to the erosional losses on the Atlantic Ocean side of the study area, large
losses on the Gulf of Mexico side were caused by burial (Fig. 8). Sediment was deposited
by the storm at many of the monitoring stations, with accumulations immediately after
the storm ranging from <1–50 cm at station 309 (pers. obs.). The longer-term effects of
the storm were dependent on the amount of sediment deposited. At most stations, subsequent sediment redistribution following the storm deposition allowed for a complete
recovery of the seagrass beds within 1 yr of the storm (see time series from sample
station 291; Fig. 8). However, at station 309, the recovery from the storm resembled the
trajectory of the eroded station 243 (Figs. 7,8). Seasonal blooms of calcareous green
algae became reestablished in the year following the hurricane, remnant patches of T.
252
testudinum were slowly expanding and their abundance was increasing; but A was still
well below pre-hurricane levels 3 yrs after the storm.
Discussion
The passage of the relatively small category 2 Hurricane Georges over the western
end of the FKNMS on September 5, 1998, decreased the density of marine macrophytes
in the soft-bottom back reef environments, but the magnitude of the decrease varied by
species. On average, density of T. testudinum, the most common seagrass in the study
area, decreased by only 3%, compared to 11% losses for H. wrightii, 19% losses for S. filiforme, and 24% losses of calcareous green algae. Localized losses were much more severe; at three of 30 monitoring stations, seagrass beds were completely obliterated by the
hurricane. The hurricane caused losses of benthic macrophytes by three mechanisms:
mechanical thinning, sediment deposition, and erosion of the underlying sediments.
There were not large increases in turbid freshwater discharge from land associated with
this hurricane, as were responsible for much of the loss of seagrasses in Hervey Bay,
Australia (Preen et al., 1995). Losses caused by thinning were largely regained quickly
after the storm. Losses caused by sediment deposition varied in magnitude. Stations
with just a few cm of sediment deposition recovered very rapidly, but one station that experienced 50 cm of deposition was still largely devoid of seagrasses 3 yrs after the storm.
Recovery was slowest at stations where the seagrass beds were removed by erosion.
The apparent differences in the susceptibility of different macrophytes to storm damage is probably a result of the architecture of the macrophytes. Thalassia testudinum,
the species that was least affected by the storm, has a large investment of biomass in rhizomes and roots that are typically buried 10–40 cm deep, compared to the higher shoot:
root ratios and shallower rooting depths of S. filiforme and H. wrightii. Calcareous green
algae are anchored to the sediment by relatively shallow rhizoid bulbs. Among the seagrasses, there is apparently also a difference in the breaking strength of attached leaves,
with S. filiforme leaves being much more susceptible to breakage than T. testudinum.
After even moderate storms in Florida Bay, it is common to see large rafts of S. filiforme
leaves floating on the surface over mixed T. testudinum – S. filiforme beds (J.W.F., pers.
obs.). These differences in architecture that lead to a gradient in susceptibility to storm
damage are likely a result of the different life-history strategies of the different species
(Bazzaz, 1979). Thalassia testudinum, the late-succesional species, has a slower growth
rate, greater longevity, reproduces almost exclusively by vegetative propagation, and has
the high investment in belowground structures typical of a late-successional plant. These
life-history characteristics also appear to provide T. testudinum with greater resistance
to hurricane-induced thinning and erosion than the early-successional seagrasses or calcareous green algae. These conjectures about rooting strength and breaking strength
remain to be tested with biophysical measurements, however.
The relative resilience of seagrasses to burial was not surprising, since seagrasses often inhabit areas with mobile sediment. Seagrasses can respond to burial by increasing
the amount of internode elongation on the vertical shoots of buried plants to compensate
for the burial (Patriquin, 1973; Marbà and Duarte, 1994). The elongation rate of buried
shoots increases in response to reduced light reaching the meristems (Terrados, 1997),
but there is a limit to the ability of seagrasses to overcome burial, with survival of shoots
decreasing as burial depth increases (Marbà and Duarte, 1994). In fact, sediment redistribution during Hurricane Gilbert in the Mexican Caribbean left a clear signal in the
253
pattern of internode elongation rates of rhizomes of T. testudinum that was measurable
4 yrs after the storm (Marbà et al., 1994). Not all species have an equivalent ability to
deal with burial. In an experiment conducted in a mixed-species seagrass bed in the
Philippines, congeners of the species found in our study differed in their ability to survive burial, such that the early-successional Halodule uninervis Boiss. and Syringodium
isoetifolium (Aschers.) Dandy survived better than Thalassia hemprichii Aschers. (Duarte et al., 1997). It is possible then, that frequent burial by storms could alter species
composition of seagrass beds in our study area by favoring H. wrightii and S. filiforme
over T. testudinum.
At the three stations where the seagrass beds were eradicated by the storm (Stations
216, 243, and 309), the first macrophytes to become reestablished after the storm were
the calcareous green algae. Compared to the seagrasses, these plants have a very high
investment in sexual reproduction and higher dispersal ability (Clifton, 1997); hence
they exhibit the life-history characteristics expected of early colonizers in a successionary sequence (Bazzaz, 1979). Although our monitoring data have yet to detect invasion
of these disturbed stations by seagrasses, we expect that they will be colonized by the
comparatively rapidly colonizing species H. wrightii and S. filiforme before T. testudinum can reinvade the disturbed areas (Fonseca et al., 1987).
The densities of calcareous green algae reached only the pre-hurricane levels during
the 3 yrs subsequent to the hurricane disturbance. One would expect that the disturbed
stations would support larger densities of calcareous green algae, since seagrasses compete for resources with the algae (Davis and Fourqurean, 2001). But, as noted by Williams (1990), competitive exclusion of calcareous algae by seagrasses rarely occurs,
calling into question the importance of competition between seagrasses and macroalgae
as a mechanism underlying the often-observed successional pattern in Caribbean seagrass beds (Den Hartog, 1971; Zieman, 1982). It has been suggested that calcareous
green algae facilitate the colonization of disturbed areas by seagrasses (Zieman, 1982),
but this contention has not been experimentally addressed.
The mechanism often invoked to explain the increase in diversity at intermediate levels of disturbance is that the disturbance removes the late-successional species that are
superior competitors, thereby allowing early-successional species with greater dispersal ability to occupy space once occupied by the former competitor (e.g., Connell and
Slatyer, 1977; Tilman, 1994). The greater susceptibility of early-successional species to
storm damage compared to late successional species may act counter to this mechanism,
since low-to-moderate levels of disturbance can actually decrease the density of the
early-successional species in a late-successional community. In this case in particular,
the removal of S. filiforme, H. wrightii, and calcareous green algae may be beneficial
to T. testudinum, since both the seagrass species (Williams 1987, 1990; Fourqurean et
al., 1995) and calcareous green algae (Davis and Fourqurean, 2001) compete with T.
testudinum for resources.
The spatial pattern in the degree of storm impact was likely due to five main factors: distance from the center of the hurricane, fetch, water depth, offshore bathymetry,
and the direction of the winds associated with the storm. In general, there was greater
loss in macrophyte density in the western parts of our study area, commensurate with
peak winds associated with the storm. Greater losses also tended to occur offshore compared to onshore. This is likely a result of offshore bathymetry and fetch. As the storm
approached the study area from the SE, the southeasterly winds associated with the
northern side of a cyclone in the northern hemisphere blew over hundreds of kilometers
254
of deep (<100 m) ocean; the resulting oceanic swells transmitted a lot of energy to the
bottom as soon as they reached shallow water at the outer reef face of the Florida Keys
barrier reef (roughly coincident with the SE boundary of the FKNMS). Much of this
reef face is characterized by a very shallow (ca. 5 m) coral reef; this reef line absorbed
the brunt of the ocean swells. The wave energy was largely dissipated by the 10 km of
shallow water between the coral reef and the Florida Keys. Our two stations 216 and
243, however, are situated in gaps in the shallow coral reefs, hence the greater degree of
macrophyte loss at these stations than nearby offshore stations. For example, station 216
sits between stations 215 and 225, which are ca. 300 m directly behind Carysfort and
Molasses Reefs, respectively. These two stations in the lee of the reefs exhibited much
less loss than the exposed station 216. The offshore stations in the western part of our
study (stations 273, 276, and 267) were similarly sheltered from the large oceanic swells.
Storm wave energy absorbed by the coral reef and the back reef environment has been
an important geomorphological force during the geologic history of the area. The back
reef sand shoals that today are covered with seagrass beds may have originated as storm
deposits resulting from past hurricanes (Ball et al., 1967).
It is likely that the spatial pattern and the magnitude of changes caused by the passage
of Hurricane Georges are peculiar to this storm. The relative exposure of locations in
the FKNMS to the force of a hurricane depends on many complex and interacting factors, including location, fetch, wind speed and duration, storm history and water depth.
As such, it is likely that other hurricanes will generate a different pattern of impact. The
existence of the seagrass-monitoring program for the FKNMS has provided a unique
opportunity to study the effects of small hurricanes in the seagrass beds of the Gulf of
Mexico and Caribbean. Detailed, pre-disturbance data on the species composition and
abundance of multiple stations affected by one storm are rare. If this program continues
over the long term, it is likely that it will provide data on the impacts of major hurricanes
on seagrass beds. The time series data also allow us to be sure of the causes of the changes in the seagrass beds at these stations. This is important because the changes wrought
by the storm at the most highly impacted stations (increase in the abundance of early
successional species compared to late successional species) are similar to the changes
predicted by models of eutrophication of seagrass beds (see Duarte, 1995; Fourqurean
and Rutten, 2003). In a region as populated as the Florida Keys, recognizing early signs
of eutrophication are of great importance. Hence, time-series monitoring data from fixed
stations are very important to help interpret the causes of change in marine macrophyte
communities.
Acknowledgments
This research was funded by grant X994620-94-5 from the Environmental Protection Agency
as part of the Florida Keys National Marine Sanctuary Water Quality Protection Program. Salary support to J.W.F. for data analysis and document preparation was provided by the Secretariá
de Estado de Educación y Universidades, Ministerio de Educación, Cultura y Deporte, Spain.
During this sabbatical, C. Duarte kindly provided space in his laboratory. Invaluable field and
laboratory work was done by B. Davis, A. Willsie, C. Rose, M. Ferdie, and S. Escorcia. Logistical support and transportation were provided by Captain D. Ward (R/V Magic) and Captain M.
O’Connor (R/V Expedition II). R. Price provided comments on earlier drafts of this paper. Publication of this paper was funded in part by a grant from the Caribbean Marine Research Center
(CMRC Project # CMRC-00-IXNR-03-01A), National Oceanic and Atmospheric Administration (NOAA) National Undersea Research Program, U.S. Environmental Protection Agency, and
255
Environmental Defense. Views expressed herein are those of the authors and do not necessarily
reflect the views of CMRC, or any of the supporting agencies. This is contribution no. x from the
Southeast Environmental Research Center at FIU.
Literature Cited
Armentano, T. V., R. F. Doren, W. J. Platt, and T. Mullins. 1995. Effects of Hurricane Andrew on
coastal and interior forests of Southern Florida: overview and synthesis. J. Coast. Res. SI 21:
111–144.
Ball, M. M., E. A. Shinn, and K. W. Stockman. 1967. The geologic effects of Hurricane Donna in
South Florida. J. Geol. 75: 583–597.
Bazzaz, F. A. 1979. The physiological ecology of plant succession. Ann. Rev. Ecol. System. 10:
351–371.
Clifton, K. E. 1997. Mass spawning by green algae on coral reefs. Science 275: 1116–1118.
Connell, J. H. and R. O. Slatyer. 1977. Mechanisms of succession in natural communities and their
role in community stability and organization. Am. Nat. 111: 1119–1144.
Davis, B. C. and J. W. Fourqurean. 2001. Competition between the tropical algae, Halimeda incrassata, and the seagrass, Thalassia testudinum. Aquat. Bot. 71: 217–232.
Den Hartog, C. 1971. The dynamic aspect in the ecology of sea-grass communities. Thalassia
Jugoslavica 7: 101–112.
Doyle, T. W., T. J. Smith, III, and M. B. Robblee. 1995. Wind damage effects of Hurricane Andrew on mangrove communities along the southwest coast of Florida, USA. J. Coast. Res. 18:
159–168.
Duarte, C. M. 1995. Submerged aquatic vegetation in relation to different nutrient regimes. Ophelia 41: 87–112.
___________, J. Terrados, N. S. R. Agawin, M. D. Fortes, S. Bach, and W. J. Kenworthy. 1997.
Response of a mixed Philippine seagrass meadow to experimental burial. Mar. Ecol. Prog. Ser.
147: 285–294.
Fonseca, M. S., G. W. Thayer, and W. J. Kenworthy. 1987. The use of ecological data in the
implementation and management of seagrass restorations. Pages 175–187 in M. J. Durako, R.
C. Phillips, and R. R. Lewis, III, eds. Proceedings of the symposium on subtropical-tropical
seagrasses of the southeastern United States. Florida Marine Res. Publications, No. 42. Florida
Department of Natural Resources, St. Petersburg.
Fourqurean, J. W. and L. M. Rutten. 2003. Competing goals of spatial and temporal resolution:
monitoring seagrass communities on the regional scale. Pages 257–288 in D.E. Busch and J.
C. Trexler, eds. Monitoring ecosystems: interdisciplinary approaches evaluating ecoregional
initiatives. Island Press, Washington DC.
_______________, G. V. N. Powell, W. J. Kenworthy, and J. C. Zieman. 1995. The effects of
long-term manipulation of nutrient supply on competition between the seagrasses Thalassia
testudinum and Halodule wrightii in Florida Bay. Oikos 72: 349–358.
_______________, A. W. Willsie, C. D. Rose, and L. M. Rutten. 2001. Spatial and temporal pattern in seagrass community composition and productivity in south Florida. Mar. Biol. 138:
341–354.
_______________, M. J. Durako, M. O. Hall, and L. N. Hefty. 2002. Seagrass distribution in south
Florida: a multi-agency coordinated monitoring program. Pages 497–522 in J. W. Porter and
K. G. Porter, eds. The Everglades, Florida Bay, and the coral reefs of the Florida Keys. CRC
Press, Boca Raton.
Glynn, P. W., L. R. Almodóvar, and J. G. Gonzales. 1964. Effects of hurricane Edith on marine life
in la Parguera, Puerto Rico. Caribb. J. Sci. 4: 335–345.
Klein, C. J. I. and S. P. J. Orlando. 1994. A spatial framework for water-quality management in the
Florida Keys National Marine Sanctuary. Bull. Mar. Sci. 54: 1036–1044.
256
Maciá, S. and D. Lirman. 1999. Destruction of Florida Bay seagrasses by a grazing front of sea
urchins. Bull. Mar. Sci. 65: 593–601.
Marbà, N. and C. M. Duarte. 1994. Growth response of the seagrass Cymodocea nodosa to experimental burial and erosion. Mar. Ecol. Prog. Ser. 107: 307–311.
________, M. E. Gallegos, M. Merino, and C. M. Duarte. 1994. Vertical growth of Thalassia testudinum: seasonal and interannual variability. Aquat. Bot. 47: 1–11.
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.
Paine, R. T. and S. A. Levin. 1981. Intertidal landscapes: disturbance and the dynamics of pattern.
Ecol. Monogr. 51: 145–178.
Patriquin, D. 1973. Estimation of growth rate, production and age of the marine angiosperm
Thalassia testudinum König. Caribb. J. Sci. 13: 111–123.
__________. 1975. “Migration” of blowouts in seagrass beds at Barbados and Carricou, West
Indies, and its ecological and geological implications. Aquat. Bot. 1: 163–189.
Peterson, B. J., C. D. Rose, L. M. Rutten, and J. W. Fourqurean. 2002. Disturbance and recovery
following catastrophic grazing: studies of a successional chronosequence in a seagrass bed.
Oikos 97: 361–370.
Poiner, I. R., D. I. Walker, and R. G. Coles. 1989. Regional studies: seagrasses of tropical Australia. Pages 279-303 in A. W. D. Larkum, J. A. McComb, and S. A. Shepherd eds. Biology
of seagrasses: a treatise on the biology of seagrasses with special reference to the Australian
region. Elsevier, Amsterdam.
Preen, A. R., W. J. Lee Long, and R. G. Coles. 1995. Flood and cyclone related loss, and partial
recovery, of more than 1000 km2 of seagrass in Hervey Bay, Queensland, Australia. Aquat.
Bot. 52: 3–17.
Robblee, M. B., T. R. Barber, P. R. Carlson, M. J. Durako, J. W. Fourqurean, L. K. Muehlstein,
D. Porter, L. A. Yarbro, R. T. Zieman, and J. C. Zieman. 1991. Mass mortality of the tropical
seagrass Thalassia testudinum in Florida Bay (USA). Mar. Ecol. Prog. Ser. 71: 297–299.
Rodríguez, R. W., R. M. T. Webb, and D. M. Bush. 1994. Another look at the impact of Hurricane
Hugo on the shelf and coastal resources of Puerto Rico, USA. J. Coast. Res. 10: 278–296.
Rollon, R. N., E. D. de Ruyter van Steveninck, W. van Vierssen, and M. D. Fortes. 1998. Contrasting
recolonization strategies in multi-species seagrass meadows. Mar. Poll. Bull. 37: 450–459.
Rose, C. D., W. C. Sharp, W. J. Kenworthy, J. H. Hunt, W. G. Lyons, E. J. Prager, J. F. Valentine,
M. O. Hall, P. E. Whitfield, and J. W. Fourqurean. 1999. Overgrazing of a large seagrass bed
by sea urchins in outer Florida Bay. Mar. Ecol. Prog. Ser. 190: 211–222.
Seddon, S., R. M. Connolly, and K. S. Edyvane. 2000. Large-scale seagrass dieback in northern
Spencer Gulf, South Australia. Aquat. Bot. 66: 297–310.
Smith, T. J., III, M. B. Robblee, H. R. Wanless, and T. W. Doyle. 1994. Mangroves, hurricanes and
lightning strikes. BioSci. 44: 256–262.
Terrados, J. 1997. Is light involved in the vertical growth response of seagrass when buried by
sand? Mar. Ecol. Prog. Ser. 152: 295–299.
Thayer, G. W. and A. J. Chester. 1989. Distribution and abundance of fishes among basin and channel habitats in Florida Bay. Bull. Mar. Sci. 44: 200–219.
Thomas, L. P., D. R. Moore, and R. C. Work. 1961. Effects of Hurricane Donna on the turtle grass
beds of Biscayne Bay, Florida. Bull. Mar. Sci. Gulf Caribb. 11: 191–197.
Tilman, D. 1994. Competition and biodiversity in spatially structured habitats. Ecology 75: 2–16.
Tilmant, J. T., R. W. Curry, R. D. Jones, A. Szmant, J. C. Zieman, M. Flora, M. B. Robblee, D.
Smith, R. W. Snow, and H. Wanless. 1994. Hurricane Andrew’s effect on marine resources.
BioSci. 44: 230–237.
van Tussenbroek, B. I. 1994. Aspects of the reproductive ecology of Thalassia testudinum in Puerto Morelos Reef Lagoon, Mexico. Bot. Mar. 37: 413–419.
Wanless, H. R., L. P. Tedesco, and K. M. Tyrrell. 1988. Production of Subtidal Tubular and Surficial Tempestites by Hurricane Kate, Caicos Platform, British West Indies. J. Sed. Petrol. 58:
739–750.
257
Williams, S. L. 1987. Competition between the seagrasses Thalassia testudinum and Syringodium
filiforme in a Caribbean lagoon. Mar. Ecol. Prog. Ser. 35: 91–98.
____________. 1990. Experimental studies of Caribbean seagrass bed development. Ecol. Monogr.
60: 449–469.
Zieman, J. C. 1976. The ecological effects of physical damage from motor boats on turtle grass
beds in southern Florida. Aquat. Bot. 2: 127–139.
__________. 1982. The ecology of the seagrasses of south Florida: a community profile. FWS/
OBS- 82/25, U.S. Fish and Wildlife Service, Washington, DC. 123 p.
___________, J. W. Fourqurean, and R. L. Iverson. 1989. Distribution, abundance and productivity of seagrasses and macroalgae in Florida Bay. Bull. Mar. Sci. 44: 292–311.
Address: Department of Biological Sciences and Southeast Environmental Research Center,
Florida International University, Miami, Florida 33199. Corresponding Author: (J.W.F)
Telephone: (305) 348-4084, Fax: (305) 348-4096, E-mail: <fourqure@fiu.edu>.
CORAL REEF WATCH 2002
Alan E. Strong, Gang Liu, Jill Meyer,
James C. Hendee, and Desiree Sasko
ABSTRACT
The National Oceanic and Atmospheric Administration’s new Coral Reef Watch
(CRW) program, led out of its National Environmental Satellite, Data, and Information
Service (NESDIS) and Oceanic and Atmospheric Research (OAR) offices will strive
to fully utilize NOAA coral resources to monitor and predict changes in coral reef
ecosystems worldwide. CRW inaugurated its first Coral Reef Early Warning System
(CREWS) station in 2001 at “Rainbow Gardens,” Lee Stocking Island, Great Exuma,
Bahamas, with the installation of its first of 20 new in situ monitoring stations slated for
many domestic reefs during this decade. A major objective is to discern the relationship
between the magnitude and persistence of anomalously high sea surface temperatures
in coral reef areas and coral reef bleaching and mortality. By coordinating both in situ
point observations with the overview provided through satellite imagery this program is
designed to actively support coral reef managers and researchers through near real-time
Web-access to coral reef environmental data and coral bleaching alerts.
This paper is part in a series of papers resulting from a scientific workshop held at
the Caribbean Marine Research Center (CMRC, December 2001) to evaluate the importance of back reef systems for supporting biodiversity and productivity of marine
ecosystems. Coral reefs are one of the most diverse ecosystems in the World, supporting
essential coastal fisheries, offering potential medicines, protecting coasts from erosion,
and supporting coastal tourism industries.
Over the past few years, anomalously warm sea surface temperatures have led to increased incidence of coral reef bleaching around the globe (Goreau et al., 1998; Goreau
et al., 2000; Wilkinson et al., 2000; Wellington et al., 2001b), such as that occurred in
early 2001 at Ningaloo, South Indian Ocean (Fig. 1). This stress compounds stresses
already incurred via natural factors such as hurricanes and our changing climate and
a myriad of factors associated with detrimental human activities, such as overfishing,
anchor damage, sediment and nutrient run-off, and unregulated tourism. Increased deterioration of coral ecosystems is of major concern worldwide as human impacts undoubtedly are playing an increasing role.
Recognizing the need to protect these fragile ecosystems, in 1998 the U.S. federal
government called for increased research and monitoring of coral reefs for improved
management by establishing the multi-agency U.S. Coral Reef Task Force. In 1999,
NESDIS sponsored a Workshop at the East West Center in Honolulu, Hawaii: “International Workshop on the Use of Remote Sensing Tools for Mapping and Monitoring
Coral Reef.” The fortuitous timing of these two events during and immediately following the major bleaching event brought on by the most recent El Niño (1997–98) assisted
in galvanizing much of the coral reef community in recognizing both the benefits of the
new Internet-dependent society and the often basin-wide orchestration of many of these
bleaching events.
As early as 1995, NESDIS began producing worldwide, web-accessible, satellite-derived, sea surface temperature products to monitor for potential coral reef bleaching
(Strong et al., 1997). Additionally, NESDIS has been providing technical support for
259
260
Figure 1. Bleaching at Ningaloo Reef, South Indian Ocean, February 2001.
coral reef mapping efforts, developing a robust and comprehensive international coral
reef data management system, using paleo-climate records to describe the coral reef
environment in the past (>100 yrs), and building interagency and international collaborations in coral monitoring and research. Simultaneously, OAR’s Atlantic Oceanographic and Meteorological Laboratory (AOML) had been developing the Coral Reef
Early Warning System (CREWS), an integration of meteorological and oceanographic
instrumented arrays (buoys and dynamic pylons) with artificial intelligence software.
These CREWS stations are being deployed as coral reef environmental monitoring stations to monitor for conditions theoretically conducive to coral reef bleaching (e.g., high
sea temperature alone, high temperature plus high irradiance; see Hendee et al., 2001),
and provide long-term data sets for other coral reef ecosystem modeling and for Marine Park Area (MPA) decision support. The CREWS concept grew out of protyping
and experimentation under the Florida Institute of Oceanography and NOAA’s similarly
instrumented-array SEAKEYS program, developed in the early 1990s for the Florida
Keys National Marine Sanctuary (Ogden et al., 1994).
Coral Reef Watch (CRW)
In an effort to expand NOAA’s coral reef monitoring and bleaching alert capabilities
NESDIS and OAR joined their complimentary coral activities under the Coral Reef
Watch initiative in 2000. CREWS temporally intensive sea temperature (and other) data
serve to validate NESDIS satellite-derived spatially intensive sea surface temperature
monitoring products, while NESDIS satellite products extend coral reef bleaching monitoring to larger spatial scales and remote locations. Within NESDIS and within OAR,
CRW maximizes coral reef resources by joining the existing coral reef strengths under
a coordinated program.
CRW seeks to fully utilize space-based sea surface temperature (SST) observations
combined with CREWS in-water derived data to continually monitor for early indications of thermally induced coral bleaching worldwide. As part of CRW, the NESDIS
satellite coral bleaching monitoring program has been using 50-km, twice-weekly, nighttime-only satellite advanced very high resolution radiometer (AVHRR) SST to derive its
core coral bleaching “early warning” products, bleaching HotSpot anomaly charts, and
bleaching Degree Heating Weeks (DHW) charts as indices of coral reef bleaching related thermal stress (Strong et al., 1999, Toscano et al., 1999). The “HotSpot” technique
STRONG ET AL.: CORAL REEF MONITORING: SATELLITE AND IN SITU
261
is proving to be highly successful in providing early warnings of thermally induced
coral reef bleaching to the coral reef community (Goreau et al., 2000, Wellington et al.,
2001a). The coral bleaching HotSpot is a type of SST anomaly showing positive anomaly (potential thermal stress) compared to a “static” bleaching threshold SST climatology.
A satellite maximum monthly mean SST climatology derived from the satellite AVHRR
SSTs over a period of 1985–1993 has been used as the threshold climatology, which
is static in time but varies in space (Strong et al., 1997). Only the positive anomalies
are calculated and highlighted in the HotSpot charts as the indices of coral bleaching
inducing thermal stress. The DHW represents the accumulation of HotSpots during a
previous 12 wk time period and the HotSpot anomalies have to be at least +1°C to be
accumulated. One DHW is equivalent to 1 wk of HotSpot levels staying at 1°C or 0.5 wk
of HotSpot levels at 2°C, etc. To assess the accuracy of the HotSpot technology, comparing the satellite SST with hourly measured in situ water temperature observations at the
developing CREWS network should be a necessary first assessment step.
In the Caribbean alone, considerable bleaching variability has been seen over the past
decade from episodes of high SSTs. At NESDIS we are actively bringing on-line newlyimproved Pathfinder (AVHRR) SST data with higher resolution SSTs from what has
been available operationally (mapping at 9 versus 50 km resolution) and more consistently derived observations, being compared methodically against point-source information from drifting buoy SSTs (Kilpatrick et al., 2001).
HotSpot Charts, highlighting regions of possible thermally induced coral reef bleaching, are shown for the Caribbean in Figure 2. HotSpots are shown as an anomaly above
the expected maximum climatological SST for the entire year (Toscano et al., 2001).
Yellow/orange indicates bleaching potential and white indicates no bleaching potential
– blue/purple indicates levels just below critical. Numbers correspond to representative reef locations shown in Table 1. Over the years of derived Pathfinder satellite SST
observations, 1985–1998, 1985 (Fig. 2, top) was a relatively cool year, while 1998 (Fig.
2, bottom) was relatively warm. The bar graph in Figure 3 highlights the variability in
the number of the 12 representative reef sites (above) that appear to have experienced
sea surface temperatures at sufficient levels (at least 1°C above climatological maximum
SST for the year) to cause coral reef bleaching during each year from 1985–1998.
At the Lee Stocking Island CREWS station, R/V Kristina (buoy), shown on station
in Figure 4, a critical part of the effort is the local maintenance and calibration of the
sea temperature sensor to ensure quality data; these data can then be automatically compared with satellite monitored temperatures and thus provide near real-time feedback on
the accuracy of the satellite-monitored temperatures. The CREWS buoy R/V Kristina
routinely measures and transmits directly via satellite the following parameters (<http://
coral.aoml.noaa.gov/crw/crw_data_bahamas_72.html>): Barometric pressure; air temperature; water temperature; tide; wind (speed, gust, direction); conductivity; salinity;
photo-synthetically active radiation – PAR (Surface, 1 m); and UVB (Surface, 1 m).
The staff at Lee Stocking Island maintaining the station also give critical feedback on
the presence and progress of coral bleaching and thus validate coral bleaching predictions made by both satellite HotSpot anomalies and in situ CREWS information products. Buoy R/V Kristina is the first of ~20 CREWS stations to be deployed throughout
U.S. domestic reefs. Its location is near Rainbow Gardens Reef where the CMRC has
maintained in situ data loggers. These data supply another source of high temporal resolution sea temperature data to further compare with the more spatially comprehensive
50-km satellite SSTs that are derived from nearby pixels at the relatively shallow Great
262
Figure 2. 1985 and 1998 Caribbean Pathfinder SST Annual Composite HotSpot Charts – at sites
1,2,6,7,9 HotSpot levels (white = warmest) were reached in 1998.
Bahama Bank (west) and much deeper Exuma Sound (east). During summer, 2001,
buoy R/V Kristina successfully transmitted via satellite its in situ temperature measurements, which showed good agreement with our satellite SSTs (Fig. 5) but averaged
nearly a constant 1°C cooler than the SST loggers at Rainbow Gardens Reef. This was
presumed to be due to the increased flow and mixing at the site of R/V Kristina compared with the shallower site of Rainbow Gardens Reef nearly 1 mi to the north. It will
be interesting to see if this difference reverses during the winter months. Logger data
helped to validate and interpret the satellite SST and CREWS station readings.
263
Table 1. Reef locations located in Fig. 2.
Representative reef locations
Belize (1)
Bermuda (2)
Bonaire (3)
Dry Tortugas, Florida (4)
Flower Garden, Texas (5)
Grand Cayman (6)
Jamaica (7)
Lee Stocking Island (8)
Panama – Atlantic (9)
Puerto Rico (10)
St. Croix, U.S.V.I. (11)
St. Lucia (12)
A Case Study — Rainbow Gardens, Lee Stocking Island
This particular study attempted to compare the new CREWS buoy (R/V Kristina)
with existing in situ CMRC data from Rainbow Gardens Reef to make this new data
source, nearly 1 mi from the CMRC site, comparable (if possible) to previous historical
data. The study also attempted to relate the new data stream from satellite information
somewhat farther from shore to both the CREWS buoy and the CMRC data logger.
Satellite SST time series extracted from NOAA’s global 50-km nighttime SST datasets
at the two closest pixel locations (centered at 23°30Ń, 76°30´W, noted as SW pixel and
24°N, 76°W, as NE pixel) at Lee Stocking Island (LSI), Bahamas were used to compare
with the CMRC in situ logger water temperature time-series at Rainbow Gardens Reef
(23°47.78Ń,
°°47.78Ń, 76°08.78´W). The two SST pixels are the pixels closest to the Rainbow
Gardens logger, one over the shallow Great Bahama Bank to the west and one over a
considerably deeper Exuma Sound to the east. The Rainbow Gardens logger is located
in shallow water (seawater depth of 4 m) close to a small island (Iguana Cay) and adjacent to an active tidal channel between the Bank and Sound. The temperature sensor is
0.3 m above the bottom. The logger is approximately in the middle of the two non-land
contaminated satellite data pixel centers.
The polar-orbiting satellite nighttime passes over the Rainbow Gardens Reef were
usually between midnight and before sunrise local time. The logger water temperature
observations are available at hourly intervals. For the purpose of comparison, the same
value of a composite twice-weekly nighttime SST analysis at a pixel was applied as the
SST daily value during the twice-weekly period. This “daily” SST was then compared
with the daily mean (0:00–23:00 LT) and also nighttime mean (18:00–05:00 LT) of the
Rainbow Gardens logger temperature over the period: September 2000 through July
2001. The results are shown in Figure 6.
The mean differences between the SST at the SW pixel and the CMRC in situ logger temperature were −0.10°C and −0.02°C (STD: 0.63°C and 0.72°C) for in situ daily
mean and nighttime mean, respectively, and the NE pixel, −0.26°C and −0.13°C (STD:
0.70°C and 0.76°C). The comparison between the mean SST of the two pixels and the
264
Figure 3. Bar chart frequency of reef locations in Table 1 coinciding with HotSpot data.
logger temperature showed a mean difference of −0.18°C and −0.06°C (STD: 0.60°C and
0.68°C) for in situ daily mean and nighttime mean, respectively.
The visual examination of the difference between the SST at these two adjacent pixels and CMRC logger temperature found that at each pixel the difference periodically
increases and decreases and the difference is usually out-of-phase at the two pixels.
This suggests that the phase difference may be caused by the change in direction of the
current in the tidal channel passing by the Rainbow Gardens Reef logger. Physical characteristics of the shallow bank and the deep sound are significantly different. With this
assumption, two new SST time series were constructed by choosing the SST value, for
each day, from one of the two pixels with the smaller temperature difference to the logger temperature. One time series was to best fit the daily in situ mean temperature and
another daily nighttime in situ mean temperature. The resulting comparisons showed a
decrease in the mean difference and standard deviation: −0.11°C with STD of 0.50°C
and −0.02°C with STD of 0.57°C for daily mean and nighttime mean in situ values, respectively. From our inspection at this LSI site, the scatter plots showed that the best-fit
SST time series have the least scatter. But all scatter plots showed that satellite-derived
SSTs are ‘lower’ than in situ temperature at the high temperature end and “higher” at the
low temperature end. We will be examining the hypothesis that this might be attributed
to land (island)-sea temperature differences as we continue to compare these values over
the next few months.
The comparison between the monthly means derived from the SST and in situ Rainbow Gardens Reef logger showed that the mean difference was −0.10°C with STD of
0.28°C, 0.03°C with STD of 0.36°C for daily mean and nighttime mean in situ values,
respectively, for the SW pixel and −0.23°C with STD of 0.4°C and −0.11°C with STD of
0.39°C for NE pixels; −0.17°C with STD of 0.26°C and −0.04°C with STD of 0.29°C for
mean of the two pixels; and −0.1°C with STD of 0.2°C and −0.005°C with 0.24°C for the
best-fit SST values.
The comparison between the weekly means derived from the SST and CMRC logger
temperature showed that the mean difference was −0.10°C with STD of 0.42°C, 0.02°C
with STD of 0.49°C for daily mean and nighttime mean, respectively, for the SW pixel;
−0.25°C with STD of 0.54°C and −0.13°C with STD of 0.56°C for the NE pixel; −0.18°C
with STD of 0.40°C and −0.06°C with STD of 0.46°C for mean of the two pixels; and
−0.11°C with STD of 0.29°C and −0.015°C with 0.35°C for the best-fit pixel.
265
Figure 4. Lee Stocking Island CREWS buoy: R/V Kristina
Figure 5. 12 d of SSTs from Rainbow Garden logger and R/V Kristina: 27 June–8 July 2001.
266
Figure 6. Daily comparisons of SST from: R/V Kristina, Rainbow Gardens logger and satellite
pixels to the SW (Bank) and NE (Sound) of Kristina location: 29 May–8 July 2001.
The results from this initial CREWS-site data logger comparison lead us to conclude
that the 50-km satellite nighttime SST analysis is providing a very good match with the
in situ temperature observation even though it is at a depth of nearly 4 m. Occasionally we noted the absence of an SST observation due to cloud cover. Sub-resolution
cloud elements may decrease the satellite’s SST accuracy. Our findings also suggest that
the diurnal vertical mixing due primarily to high tidal velocities at Rainbow Gardens
Reef were active over the comparison period. The SST algorithm aims to calibrate its
measured temperature to be representative to a 1-m bulk water temperature. From our
brief study it would appear that the 50-km nighttime satellite SST is providing water
temperatures for the shallow water areas, typical for LSI reefs, at the required accuracy
level. The extreme 2–3º temperature swings observed during the period provide useful water temperature information representative alternately of both the Bank and the
Sound. However, the SSTs at both pixels are needed to give more accurate monitoring
of water temperature at the Rainbow Gardens Reef. More observation and research are
needed to determine which one at what time is suitable for monitoring the water temperature at the location.
Benefits
For coral reef managers, CRW near real-time Web-based monitoring products permit immediacy in response to changing ecosystem character, which has allowed for
improved regulation of access to the reefs in question. By maintaining a more constant
vigilance and carefully coordinating with reef (MPA) managers having oversight of the
threatened jurisdictions, we stand an increased likelihood of being able to reduce stress
resulting from fishing and recreational activities during periods of high stress, e.g., high
267
water temperatures. If we are successful, although coral mortality cannot be completely
stopped, stresses can be minimized and recovery maximized through improved awareness and management practices. CRW data and products have begun to deliver the tools
necessary to alert managers and researchers, for the first time, to be responsive as soon
as adverse environmental conditions begin to develop. Their on-site feedback not only
acquires initial bleaching information first hand but contributes greatly to our understanding of coral bleaching phenomena. Moreover, the accumulation of both satellite
and in situ CRW long-term data sets will aid in our understanding of our coral reefs’
response to climate change as well as coral reef ecosystem function.
Plans
In 2002, NESDIS and OAR will seek to improve spatial coverage, reliability, quality,
and accessibility of CRW data and products by:
1. Expanding the network of coral reef environmental monitoring stations to the U.S.
Virgin Islands and American Samoa
2. Adding pollutant indicator sensors to existing environmental monitoring stations to
provide a more complete set of environmental parameters for monitoring and modeling
coral reef ecosystems
3. Improving national and international collaboration and information exchange in order to validate monitoring data and bleaching alert products as well as better understand
the coral bleaching phenomenon
4. Securing technical support for satellite near real-time coral reef bleaching and
monitoring products to ensure their availability during critical seasons
5. Increasing the spatial resolution of satellite monitoring and bleaching alert products, thus improving applicability and relevance to smaller scale ecosystems
6. Performing temporal assessments of coral reef bleaching using high-resolution satellite data
7. Providing automated bleaching event maps in user-friendly formats (e.g., Geographic Information System)
8. Extending SST records using the coral paleo-climate proxy record, thereby promoting an understanding of coral’s response to environmental conditions in the past
9. Continuing development of the NOAA international Coral Reef Information System
(CoRIS) that enhances access to NOAA, national, and international coral reef data and
information worldwide.
Summary
The Coral Reef Watch 2002 Project embodies a coordinated NESDIS and OAR coral
monitoring and bleaching research program that responds to a need for improved understanding of coral reef ecosystems and fulfills NOAA’s mission to Sustain Healthy
Coasts. The planned 2002 activities fully exploit NESDIS and OAR expertise in data
management, satellite mapping and monitoring, while hopefully encouraging diverse
partnerships and communication.
268
Acknowledgments
This paper is funded, in part, by a grant from the Caribbean Marine Research Center (CMRC
Project # CMRC-00-IXNR-03-01A) National Oceanic and Atmospheric Administration (NOAA)
National Undersea Research Program, U.S. Environmental Protection Agency, and Environmental Defense. Views expressed herein are those of the authors and do not necessarily reflect the
views of CMRC, or any of the supporting agencies.
Literature Cited
Goreau, T. J., R. Hayes, A. E. Strong, E. Williams, G. Smith, J. Cervino, and M. Goreau. 1998.
Coral reefs and global change: impacts of temperature, bleaching, and emerging diseases. Sea
Wind 12: 2–6.
___________, T. McClanahan, R. Hayes, and A. E. Strong. 2000. Conservation of coral reefs after
the 1998 global bleaching event. Conserv. Biol. 14: 5–15.
Hendee, J. C., E. Mueller, C. Humphrey, and T. Moore. 2001. A data-driven expert system for
producing coral bleaching alerts at Sombrero Reef in the Florida Keys. Bull. Mar. Sci. 69:
673–684
Kilpatrick, K. A., G. P. Podestá, and R. Evans. 2001. Overview of the NOAA/NASA advanced
very high resolution radiometer Pathfinder algorithm for sea surface temperature and associated matchup database. J. Geoph. Res. 106: 9179–9197.
Ogden, J., J. Porter, N. Smith, A. Szmant, W. Jaap, and D. Forcucci. 1994. A long-term interdisciplinary study of the Florida Keys seascape. Bull. Mar. Sci. 54: 1059–1071.
Strong, A. E., C. S. Barrientos, C. Duda, and J. Sapper. 1997. Improved satellite techniques for
monitoring coral reef bleaching. Pages 1495–1498 in Proc 8th International Coral Reef Symposium, Panama City, Panama.
___________, T. J. Goreau, and R. Hayes. 1999. Ocean HotSpots and coral reef bleaching: January–July 1998. Reef Encounters 24: 20–22.
Toscano, M. A., A. E. Strong, and I. C. Guch. 1999. New analyses for ocean HotSpots and coral
reef bleaching. Reef Encounters 26: 31.
_____________, G. Liu, I. C. Guch, K. S. Casey, A. E. Strong, and J. E. Meyer. 2001. Improved
prediction of coral bleaching using high-resolution HotSpot anomaly mapping. Pages 1143–
1147 in Proc. 9th Intl. Coral Reef Symp., Bali, Indonesia.
Wellington, G. M., P. W. Glynn, A. E. Strong, S. Navarrete, and E. Wieters. 2001a. Crisis on coral
reefs linked to climate change, EOS 82: 1, 5.
_______________, A. E. Strong, and G. Merlen. 2001b. Sea surface temperature variation in the
Galápagos Archipelago: a comparison between AVHRR nighttime satellite data and in-situ
instrumentation (1982-1988). Bull. Mar. Res. 69: 24–42.
Wilkinson, C., O. Linden, H. Cesar, G. Hodgson, J. Rubens and A. E. Strong. 1999. Ecological and
socioeconomic impacts of 1998 coral mortality in the Indian Ocean: An ENSO impact and a
warning of future change? Ambio 28: 188–196.
ADDRESSES: (A.E.S., G.L., J.M.) National Oceanic and Atmospheric Administration, National
Environmental Satellite, Data and Information Service, Oceanic Research and Applications
Division (NOAA/NESDIS/ORAD), 5200 Auth Road Camp Springs, Maryland 20746. (J.C.H.)
National Oceanographic and Atmospheric Administration, Oceanic and Atmospheric Research,
Atlantic Oceanographic and Meteorological Laboratory, 4301 Rickenbacker Causeway, Miami,
Florida 33149-1026. (D.S.) Perry Institute of Marine Science—Caribbean Marine Research
Center, 100 North U.S. Highway 1, Jupiter, Florida 33477. CORRESPONDING AUTHOR: (A.E.S.)
Telephone: 301-763-8102 x170, Fax: (301) 763-8572, E-mail: <Alan.E.Strong@noaa.gov>.
Website: <coralreefwatch.noaa.gov>.
TRANSPORT PROCESSES LINKING SHELF AND BACK
REEF ECOSYSTEMS IN THE EXUMA CAYS, BAHAMAS
Ned P. Smith
ABSTRACT
Shallow-water transport processes are reviewed to describe the physical linking of
inner shelf and back reef environments. Data from Exuma Sound and Great Bahama
Bank near Lee Stocking Island, Exuma Cays, Bahamas provide examples of tidal and
wind-driven transport. Along-shelf currents on the seaward side of Lee Stocking Island
are predominantly northwestward, and speeds are usually <25 cm s−1. Amplitudes of
tidal currents are 2 cm s−1 or less. Across-shelf currents are generally between ± 5 cm
s−1, and amplitudes of across-shelf tidal currents are less than 1 cm s−1. Near-surface
flow is deflected landward, and a weak downwelling pattern holds surface-layer water
close to the coast. Water passing Adderley Cut on the flood tide is carried onto Great
Bahama Bank. Maximum flood current speeds in the channel reach 60–70 cm s−1. In
back reef areas, maximum tidal current speeds are 20–40 cm s−1, but the tide-induced
residual speed is <1 cm s−1. Wind-drift currents at three back reef sites are 3–4 cm s−1.
Plots of salinity recorded in tidal channels show “spikes” that represent high-salinity
bank water leaving during the ebb. Hyperpycnal ebb tide plumes increase the efficiency
of tidal exchanges and enhance the linkage of shelf and back reef ecosystems.
This publication is part in a series of papers resulting from a scientific workshop held at
The exchange of water between shelf and back reef environments is usually some
combination of two-way tidal and wind-driven transport processes and the one-way,
seaward movement of run-off from the adjacent continental margin. Transport pathways
and the relative importance of transport processes vary greatly from one location to the
next due to significantly different physical settings. In nearshore waters seaward of the
reef, transport is commonly a response to wind forcing, and current speeds may be only
a few tens of cm s−1. Accompanying the wind-driven along-shelf flow is a more subtle
across-shelf component, producing an upwelling or downwelling pattern that holds material close to the coast, or carries it seaward. Along-shelf tidal currents can be significant, but across-shelf tidal currents are often minimal due to the coastal boundary.
Tidal currents in the shallow water landward of the reef and especially in tidal channels are often much stronger than they are in middle and outer shelf waters. And it is
not uncommon for tidal currents to be an order of magnitude stronger than wind-driven
currents. Flood and ebb tide excursions provide the actual linking of shelf and back reef
regions. Tidal excursions vary with spring and neap tide conditions, diurnal inequalities
and wind forcing, which can aid or inhibit purely tidal exchanges significantly.
Essential to the exchange of shelf and back reef water is the mixing that occurs
throughout the tidal cycle (Officer, 1976). As a result of turbulent mixing, some fraction
of the incoming shelf water remains in the back reef, and water entrained into the plume
leaves on the ebb. Water left in the back reef area is then transported farther in two ways.
A tide-induced residual transport (van de Kreeke, 1978) arises in shallow water from
the interaction of the rise and fall in water level with the ebb and flood of the current.
269
270
In addition, the wind-drift current can be significant, especially during short-lived and
infrequent events, such as the passage of cold fronts (Smith, 1990).
A series of field studies conducted in the early 1990s examined transport processes in
the vicinity of Lee Stocking Island, one of the Exuma Cays in the Bahamas. While the
specific objective was to assemble data needed to characterize the transport of queen
conch larvae from shelf spawning sites to back reef nursery areas, the database that
accumulated over a 4-yr period was used for a broader range of investigations. Results
of individual studies focused on larval transport in tidal channels (Smith and Stoner,
1993); salt, heat, and mass transport in tidal channels (Smith, 1995a); the effect of tidal
exchanges on back reef hydrography (Stoner, et al., 1996) and across-shelf transport on
the Exuma Sound side of Lee Stocking Island (Stoner and Smith, 1997).
The purpose of this paper is to describe physical transport processes that link shelf
and back reef ecosystems, giving examples from studies conducted in the vicinity of Lee
Stocking Island. Exchanges may be tide-driven, influenced by wind forcing that is common for a given season, or in the form of distinct but transient wind events such as cold
fronts. Patterns that emerge are specific to the study area; nevertheless, they demonstrate
a close coupling of shelf and back reef environments.
Study Area
The Exuma Cays lie along the western side of Exuma Sound and along the eastern margin of
Great Bahama Bank (Fig. 1). The cays extend 130 km from Norman Cay (24o 39.3ʹN) southeast
to Rat Cay (23o 44.2ʹN) at the northern end of Great Exuma Island.
Water depths over Great Bahama Bank are commonly 2–3 m. On the Exuma Sound side of
the cays, the shelf is only 1–2 km wide, and the shelf break is defined by the top of a “wall” at a
depth of 25–30 m. The base of the wall is at a depth of about 200 m, and from there the sea floor
descends with a slope of about 60º. Water depths of 1000 m can occur just a few kilometers from
the coast.
Approximately 30 tidal channels exchange significant amounts of water between Exuma Sound
and Great Bahama Bank. Thus the Exuma Cays constitute a leaky barrier between two distinctly
different hydrographic regimes. In the deep waters of Exuma Sound, the response to local heating
and cooling is slow, and variations in temperature and salinity are much less than those occurring
in bank waters. The annual temperature cycle in shelf waters includes a late summer maximum of
29–30oC and a midwinter minimum of about 24oC (Pitts and Smith, 1993; Wicklund et al., 1993).
In the shallow waters of Great Bahama Bank, responses to local heat fluxes and to precipitation-evaporation differences are more distinct (Pitts and Smith, 1993). Midsummer temperatures
rarely exceed 31oC, though the period of maximum temperatures may be of longer duration than
in shelf waters. Midwinter temperatures recorded on the bank can be as low as 20oC with the passage of strong cold fronts. Salinity in Exuma Sound remains between about 36.5 and 37.0 (Hickey
et al., 2000), while unpublished data from Great Bahama Bank show salinities as low as 35–36
and, more commonly, as high as 39–40.
Weather records (Pitts et al., 1993) define a subtle progression of the seasons. A long summer
season extends from July–October, and mean daily temperatures are generally within a degree of
29oC. Daily mean temperatures are lowest in January, February, and March when they may reach
16–17oC. Interannual variability can be large, however, depending on the timing, frequency, and
intensity of cold fronts. Similarly, monthly average air temperatures in January, February, and
March are as low as 24oC in some years or as high as 26oC in others. Wind directions are most
commonly out of the easterly octant (090 ± 22.5o). During a 1-yr period from July 1990 through
June 1991, 31% of the hourly wind directions were easterly, and 74% of the observations were in
the northeasterly, easterly, or southeasterly octants. Wind speeds are generally in the 3–6 m s−1
SMITH: TRANSPORT PROCESSES
271
Figure 1. Map showing study sites (1) in shelf waters of Exuma Sound off Lee Stocking Island,
(2) in Adderley Cut, and (3)–(5) in shallow waters of Great Bahama Bank between Lee Stocking
Island and the Brigantine Cays.
range. During the same 1990–91 time period, 21% of the readings were 4–5 m s−1, and 58% were
between 3 and 6 m s−1.
Climatological data needed to define wet and dry seasons are sparse, but rainfall data from
1990–92 (Pitts et al., 1993) suggest that the May–October period receives about 70% of the total
annual rainfall, while the November–April period receives only 30%. The annual total rainfall
is ~30 cm. The average net excess of evaporation over precipitation in this area is about 150
cm (Schmitt et al., 1989). Unpublished evaporation calculations, combined with precipitation
measurements (Pitts et al., 1993) indicate that during a 1990–91 study evaporative water loss
exceeded precipitation by 140 cm. As a result of higher salinities on Great Bahama Bank, there is
a net salt export from the bank to the sound. During the 1990–91 study, the average tide-induced
net salt loss through Adderley Cut was 2.2 × 106 kg per semidiurnal tidal cycle (Smith, 1995a).
Higher ebb-tide salinity serves as a natural tracer for monitoring the exchange of bank and sound
water, and it results in density currents that cascade down the shelf and enhance the efficiency of
tidal exchanges (Smith, 1996).
Data
Weather data needed to characterize wind forcing were recorded with a Campbell weather
station located along the southwest side of a landing strip on Lee Stocking Island, ~350 m from
Exuma Sound at its closest point. When winds were out of the northerly to southeasterly quadrants, speed and direction measurements were minimally influenced by topography or vegetation.
Small hills lying southwest of the weather station affected wind measurements when winds were
out of the southwesterly and northwesterly quadrants, but the blocking effect of topography was
more to reduce wind speeds than to alter wind directions. Wind speed and direction were recorded with accuracies of ± 0.11 m s−1 and ± 0.5o, respectively. Air temperature was recorded with
an accuracy of ± 0.5oC, and relative humidity was recorded with an accuracy of 5%. The relative
humidity sensor was calibrated approximately bimonthly using a sling psychrometer.
Current speeds and directions were recorded 13 m above the bottom in 22 m of water at a
mid-shelf location on the Exuma Sound side of Lee Stocking Island (Station 1 in Fig. 1). Current
and water level records were obtained at a study site at the south end of Adderley Cut (Station
2) and at three locations on Great Bahama Bank (Stations 3–5) to characterize shallow-water
272
tidal and wind-driven transport, including the perturbing effect of transient but energetic wind
events. Currents in shelf waters of Exuma Sound and in Adderley Cut, measured with a General
Oceanics Mark II current meter, had speed and direction accuracies of ± 1.0 cm s−1 and ± 2o,
respectively. The Adderley Cut current meter also recorded conductivity and temperature with
accuracies of ± 2.5 mS cm−1 and ± 0.25oC. This permits salinity to be calculated with an accuracy
of approximately ± 2 (Lewis, 1980). The 0.1 mS cm−1 and 0.02oC resolution of the conductivity
and temperature values, however, permit calculations of hour-by-hour changes in salinity with a
resolution of 0.1.
At the bank study sites, time-averaged, half-hourly current speeds and directions were obtained from an Endeco Type 105 current meter, with speed and direction accuracies of ± 5.9 cm
s−1 and ± 3.6o, respectively. Water level at the bank stations and in Adderley Cut was determined
from a Sea Data TDR-3 pressure recorder with an accuracy of ± 1.2 cm.
Methodology
Currents recorded at Station 1 were decomposed into along-shelf (310–130o) and across-shelf
components and filtered to remove tidal fluctuations, as well as any response to diurnal variations in wind stress (Bloomfield, 1976). The low-pass filter removes 50% of the input variance
at a periodicity of 37 hrs. At periods of 30 and 48 hrs, 10 and 90% of the input variance passes
through the filter.
The flow at Station 2 was decomposed into along-channel and across-channel components
by determining the headings at which the two orthogonal components were uncorrelated. The
336–156o component was used to describe along-channel flow. Inflow, toward 156o, was defined
to be positive. Conductivity and temperature at Station 2 were used to calculate salinity (Lewis,
1980). Biofouling was minimal, but salinity values calculated at the end of the deployment were
compared with data obtained with a Sea-Bird Electronics Seacat SBE 19 profiler.
Surface-to-bottom transport in the shallow waters of Great Bahama Bank (Stations 3–5) was
calculated by extrapolating mid-depth current measurements to the bottom and to the surface.
Assuming a logarithmic current profile, the vertically-integrated transport rate, T2D, is given by
T2 D =

Z
u*
 Z ln   − ( Z − zo )  ,
k 
 zo 

Eq. 1
(Smith, 1994) where u* is the friction velocity, k is the Karman constant, Z is the total water
depth, and zo is the roughness length. Water depth was obtained from bottom pressure measurements. Units of m2 s−1 for transport rates, or m2 for cumulative net transport result from verticallyintegrated current speeds. Volume transport, in m3 s−1, can be obtained if lateral variations in
current speed are known, but this was not practical at the three bank study sites. The flow was not
locally bounded, as in a channel, and horizontal gradients in current speed were unknown.
Equation (1) was used to calculate both tidal and total transport at the three bank stations. Current vectors were decomposed into north-south and east-west components, and component transports were accumulated individually. Cumulative net transport was obtained by reconstructing
transport vectors from the components. Harmonic analysis of the current components (Dennis
and Long, 1971) identified the principal tidal constituents. Harmonic constants of the M2, S2, N2,
K1, O1 and P1 constituents were used to predict the tidal rise and fall in water level and the ebb and
flood of the current (Schureman, 1958). The tide-induced residual transport was calculated from
predicted water levels and currents.
273
Amplitudes of the principal tidal constituents in shelf and bank waters were used to estimate
the distance traveled during the flood or ebb half of a tidal cycle. For a tidal constituent with amplitude A, expressed in km hr−1, the tidal excursion, E, is given by
E=
AT
,
π
Eq. 2
where T is the period of the constituent in hours.
Wind stress was calculated from wind speed and direction, air temperature and relative humidity recorded at the Lee Stocking Island weather station using the drag coefficient suggested by
Wu (1980). The anemometer was positioned approximately 5 m above the land surface, but no
correction was applied to ~10-m level winds over water.
Results
Figure 2 shows across-shelf (a) and along-shelf (b) flow recorded at Station 1.
The two plots provide an example of the first phase of the linking of shelf and back reef
ecosystems. The mean across-shelf flow is landward at −1 cm s−1. This is at the accuracy
of the current meter, but it suggests that flow just above mid depth is being held close to
the coast. In view of the predominantly landward winds during the study, it is likely that
surface layer currents are also landward, and thus the recorded shoreward flow is part of
a weak downwelling pattern. Along-shelf flow is generally northwestward with an aver-
Figure 2. (A) Across-shelf and (B) along-shelf current components recorded at Station 1, October 22, 1991–March 20, 1992. Positive across-shelf flow is seaward; positive along-shelf flow is
toward 310o.
274
Figure 3. (A) Wind stress, (B) low-pass filtered along-channel current speed, and (C) salinity
recorded at Station 2, March 21–May 18, 1992. Wind stress components are positive toward 208o;
positive current components represent an inflow from Exuma Sound to Great Bahama Bank.
age speed of about 7 cm s−1. The amplitude of the along-shelf M2 tidal current is 2.4 cm
s−1, and the tidal excursion is 0.35 km. Reversals are infrequent and relatively short-lived,
but 22% of the along-shelf current components are in a southeastward direction. The
remainder of the time, the flow is in the direction of the tidal channels northwest of Lee
Stocking Island. The first of these channels is Adderley Cut. Water reaching the northern
275
end of Lee Stocking Island during the flood part of the tidal cycle will be drawn into the
channel, thereby starting the second phase of the physical linking of shelf and back reef
ecosystems. Transport at the mouth of Adderley Cut was described by Smith (1995a).
The amplitude of the M2 tidal current is 69 cm s−1. During a 1-yr study in 1990–91, the
mean current speed at mid depth in the center of the channel was an inflow of 9.5 cm s−1,
and the mean volume transport was 220 m3 s−1.
Figure 3 combines (a) wind forcing with (b) low-frequency fluctuations in along-channel flow and (c) salinity recorded at Station 2. Wind stress is represented by the 028–208o
component that was most highly correlated with along-channel flow. Even though wind
stress was not low-pass filtered, the correlation with low-pass filtered along-channel flow
is +0.868. Salinity spikes occur in bursts, especially during the first three and a half
weeks of the study. Comparison of low-pass filtered, nontidal flow with salinity shows
that salinity spikes appear during and following seaward (negative) nontidal flow.
Spikes disappear during and following nontidal inflow, which in turn is a response to
low-frequency southward wind stress. The strong inflow event that begins in mid April,
for example, floods the basin between Lee Stocking Island and the Brigantine Cays with
Exuma Sound water. For the next 10 d, spikes do not appear during the ebb, and salinity
remains relatively stable at ~37. Results suggest that wind events alternately increase the
residence time of Exuma Sound water on the eastern fringe of Great Bahama Bank, then
encourage the export of bank water through tidal channels and onto the shelf.
Figure 4 includes the cumulative net transport diagrams of water moving past (a)
Station 3, (b) Station 4, and (c) Station 5 on Great Bahama Bank. These three study
sites exemplify the final phase of the linking of shelf and back reef ecosystems. Flow
recorded at Station 3 from June 5–August 15, 1991, includes the ebb and flood of the
tide along a major axis that is oriented northeast–southwest, and a net flow that is generally westward. The long-term net transport is nearly perpendicular to the ebb and flood
of the tide at this location. No significant weather-related events occurred during this
midsummer study.
Flow recorded at Station 4 from December 8, 1990–January 10, 1991, includes a period of south–southeastward transport in response to a cold front that moved through
the area on December 9 (Pitts et al., 1993) and a longer period of westward transport in
response to westward wind stress. Two additional cold fronts arrived on December 21
and 29, but the response does not stand out distinctly. Flood tide and ebb tide headings
are southwestward and northeastward, respectively.
Flow past Station 5 is fundamentally different in the sense that the long-term net
transport is east–northeastward. In view of the location of the study site off the northern
end of Great Exuma Island, it is hypothesized that this generally upwind transport is part
of a counterclockwise gyre that occupies the southeastern part of the basin between Lee
Stocking Island and the Brigantine Cays. At this location, the ebb and flood of the tide
are nearly parallel to the direction of the long-term net transport.
Tidal transport in the back reef can be expressed in two ways. The tide-induced residual current, obtained using equation (1), is a resultant current speed over time scales
much longer than a tidal period. The tidal excursion, estimated by equation (2), is the
distance water moves over individual flood and ebb cycles. Tide-induced residual currents were calculated for Stations 3–5 using predicted tidal currents and water levels.
All magnitudes are 0.3 cm s−1 or less, and thus the tide-driven transport is of negligible
importance in the semi-enclosed basin between Lee Stocking Island and the Brigantine
Cays. Tidal waves in this region exhibit a standing wave pattern, and the small magni-
276
Figure 4. Cumulative net transport past (A) Station 3, June 5–August 15, 1991; (B) Station 4,
December 8, 1990–January 10, 1991, and (C) Station 5, October 28–December 15, 1991. Positive
east-west and north-south transport is eastward and northward, respectively.
tude of the tide-induced residual is a result of the near coincidence of high or low tide
with slack water. Tidal excursions calculated from M2 constituent amplitudes at Stations
3, 4, and 5 are 4.5, 2.7, and 4.9 km, respectively.
Wind-driven transport was estimated by subtracting the tidal residual transport from
the total transport during each of the three studies. Results indicate that the wind-drift
currents at Stations 3, 4, and 5 averaged 3.6, 4.2, and 3.3 cm s−1, respectively. At these
speeds, dissolved and suspended material can be transported 1–2 km during a semidiurnal tidal cycle.
Discussion and Conclusions
Data assembled here provide evidence of the importance of both tidal and wind-forced
transport mechanisms in linking shelf and back reef ecosystems. Wind-driven upwelling
and downwelling patterns embedded in the along-shelf flow are important for holding
material close to the coast where it can be drawn into the back reef area on the flood, or
277
carrying it seaward where it will no longer be available to either shelf or back reef habitats. This is true whether the dissolved or suspended material is concentrated in nearsurface or near-bottom layers. Only when material is uniformly distributed throughout
the water column will upwelling and downwelling patterns be of negligible importance.
Even under vertically-mixed conditions, however, ebb tide plumes pushing well across
the shelf can interrupt the along-shelf flow and force shelf water seaward where it will
not be drawn into the tidal channels.
Tidal and wind-driven transports are both important in channels and in the back reef,
but they play very different roles. Stoner et al. (1996) used drogue tracking to investigate
the coupling of shelf and back reef areas. Direct coupling was restricted to a relatively
narrow but variable band along the eastern margin of Great Bahama Bank. Flood tide
excursions starting at the mouth of Adderley Cut were generally between 5 and 10 km.
Wind forcing was of secondary importance, but wind drift could extend or shorten flood
and ebb tide excursions by several kilometers. In the back reef, normal wind drift during a semidiurnal tidal cycle moves water 1–2 km. Thus, in the short term, wind-driven
transport is smaller than the tidal excursion. But over longer time scales, wind drift is
much larger than the tide-induced residual transport.
Wind conditions most effective in controlling bank-shelf exchanges vary only slightly
from tidal channels to adjacent regions of Great Bahama Bank. For example, the 218–
038o component of the wind stress vector is most highly correlated with nontidal flow
at the mouth of Adderley Cut (Smith, 1995a). In the present study, the 208–028o wind
stress component was most highly correlated with nontidal along-channel flow at the
southern end of Adderley Cut. Further analysis of the drogue trajectories presented by
Stoner et al. (1996) suggests that the length of the flood-tide drift is most sensitive to the
209–029o component of the wind stress vector. Meteorological data from Lee Stocking
Island (Pitts et al., 1993) indicate that winds out of the north-northeast are most common during winter months. Thus, the importance of winds in linking shelf and back reef
environments probably varies seasonally, with the strongest influence occurring during
winter months.
The exchange of bank and sound water that results from horizontal mixing during the
ebb and flood of the tide constitutes another mechanism for linking shelf and back reef
ecosystems. While it is difficult to quantify the amount of sound water that remains on
the bank, or the amount of bank water that is exported to Exuma Sound over a semidiurnal tidal cycle, evidence of a net exchange is obtained by using salinity as a natural
tracer. At times of low rainfall, a mixture rich in bank water appears as a ‘spike’ in the
salinity plot (Fig. 3C), and this water generally leaves the bank as a hyperpycnal, nearbottom plume. Smith (1996) showed that almost none of this bank water returns on the
following flood, thereby increasing the amount of sound water carried onto the bank.
This increases the efficiency of the exchange and enhances the transport of dissolved
and suspended material from shelf to back reef environments on the flood and from back
reef to shelf environments on the ebb.
The regional-scale wind-driven or tide-induced movement of water across Great Bahama Bank, toward or away from Exuma Sound, can also be important in linking shelf
and back reef areas. Regional flow patterns can aid one part of the tidal cycle and oppose the other. The role played by large-scale flow patterns over Great Bahama Bank
cannot be determined in the vicinity of the Exuma Cays, however, because results are
available from only three locations. Smith (1995b) reported a net northward transport
over the bank during a 398-d study at a site ~40 km west-northwest of Lee Stocking
278
Island (23o 52.9ʹN, 76o 26.1ʹW). This would carry bank water toward the Exuma Cays
and thus oppose Exuma Sound water flooding onto the bank. Unpublished data from two
more recent studies conducted closer to Lee Stocking Island, however, are contradictory.
A 132-d time series collected 24 km west of the northern end of Lee Stocking Island
(23o 46.6ʹN, 76o 20.5ʹW) showed a net southwestward transport, directly away from the
Exuma Cays. A 267-d study conducted only 11 km to the southeast at West Barracouta
Rock (23o 42.9ʹN, 76o 15.4ʹW) showed a net northwestward transport that paralleled the
Exuma Cays and aided neither the flood nor the ebb. In an open, flow-through system,
regional circulation patterns may be important in linking shelf and back reef ecosystems, but the local effect in the Exuma Cays will be determined only when additional
studies are conducted to reveal a coherent large-scale pattern.
Acknowledgments
I would like to thank P. Pitts, who participated in the fieldwork on Lee Stocking Island and did
the downloading and initial editing of the current meter and pressure recorder data files. Thanks
also to the staff of the Caribbean Marine Research Center on Lee Stocking Island for the support
they provided during the data collection phase of the study. This paper is funded by a grant from
the Caribbean Marine Research Center (CMRC Project # CMRC-00-IXNR-03-01A), National
Oceanic and Atmospheric Administration (NOAA) National Undersea Research Program, U.S.
Environmental Protection Agency, and Environmental Defense. Views expressed herein are those
of the author and do not necessarily reflect the views of CMRC, or any of the supporting agencies.
Harbor Branch Oceanographic Institution Contribution 1530.
Literature Cited
Bloomfield, P. Fourier analysis of time series: an introduction. John Wiley & Sons, New York.
258 p.
Dennis, R. E. and E. E. Long. 1971. A user’s guide to a computer program for harmonic analysis
of data at tidal frequencies. NOAA Tech. Rept. NOS 41, U.S. Dept. of Comm., National Ocean
Survey, Rockville.
Hickey, B. M., P. MacCready, E. Elliott, and N. B. Kachel. Dense saline plumes in Exuma Sound,
Bahamas. J. Geophys. Res. 105: 11, 471–11, 488.
Lewis, E. L. 1980. The practical salinity scale 1978 and its antecedents. IEEE Ocean. Eng. OE-1:
3–8.
Officer, C. B. 1976. Physical oceanography of estuaries (and associated coastal waters). John
Wiley & Sons, New York. 465 p.
Pitts, P. A. and N. P. Smith. 1993. Annotated summary of temperature and salinity data from the
vicinity of Lee Stocking Island, Exuma Cays, Bahamas. Caribbean Marine Research Center,
Tech. Rept. Series 93-3, 43 p.
________, _____________, and K. W. Mueller. Summary of meteorological data from Lee Stocking Island, Exuma Cays, Bahamas: 1990-92. Caribbean Marine Research Center, Tech. Rept.
Series 93-2, 48 p.
Schmitt, R. W., P. S. Bogden, and C. E. Dorman. 1989. Evaporation minus precipitation and density fluxes for the North Atlantic. J. Phys. Ocean. 19: 1208–1221.
Schureman, P. 1958. Manual of harmonic analysis and prediction of tides. Special Publ. 98 (Rev.
Ed.), U.S. Dept of Comm., U.S. Govt. Printing Office, Washington, D.C.
Smith, N. P. 1990. Wind domination of residual tidal transport in a coastal lagoon. Pages 123–133
in R.T. Cheng, ed. Residual currents and long-term transport. Coastal and estuarine studies,
vol. 38. Springer-Verlag, New York.
279
_________. 1994. Long-term Gulf-to-Atlantic transport through tidal channels in the Florida
Keys. Bull. Mar. Sci. 54: 602–609.
_________. 1995a. Observations of steady and seasonal salt, heat and mass transport through a
tidal channel. J. Geophys. Res. 100: 13, 713–13, 718.
_________. 1995b. On long-term net flow over Great Bahama Bank. J. Phys. Ocean. 25: 679–
684.
_________. 1996. The effect of hyperpycnal water on tidal exchange. Pages 107–116 in D. E.
Aubrey and C. T. Friedrichs, eds. Buoyancy effects on coastal and estuarine dynamics, coastal
and estuarine studies, vol. 53. American Geophysical Union, Washington, D.C.
_________ and A. W. Stoner. 1993. Computer simulation of larval transport through tidal channels: role of vertical migration. Est. Coas. Shelf Sci. 37: 43–58.
Stoner, A. W., P. A. Pitts, and R. A. Armstrong. 1996. Interaction of physical and biological factors
in the large-scale distribution of juvenile queen conch in seagrass meadows. Bull. Mar. Sci.
58: 217–233.
Stoner, A. W. and N. P. Smith. 1997. Across-shelf transport of gastropod larvae in the central Bahamas: rapid response to local wind conditions. J. Plankton Res. 20: 1–16.
Van de Kreeke, J. 1978. Mass transport in a coastal channel, Marco River, Florida. Est. Coas. Mar.
Sci. 7: 203–214.
Wicklund, R. I., G. D. Dennis, and K. W. Mueller. 1993. Summary of data from the water Temperature monitoring network: Lee Stocking Island, Bahamas, 1998-91. Caribbean Marine Research Center Technical Report Series, Nr. 93-1, 142 p.
Wu, J. 1980. Wind-stress coefficients over sea surface near neutral conditions—a revisit. J. Phys.
Ocean. 10: 727–740.
Address: Harbor Branch Oceanographic Institution, 5600 U. S. Highway 1, North Fort Pierce,
Florida 34946. E-mail: <nsmith@hboi.edu>.
WIND-MEDIATED DIEL VARIATION IN FLOW SPEED
IN A JAMAICAN BACK REEF ENVIRONMENT:
EFFECTS ON ECOLOGICAL PROCESSES
Salvatore J. Genovese and Jon D. Witman
ABSTRACT
The movement of water plays an important role in a number of physiological (e.g.,
metabolic rate, nutrient uptake) and ecological (e.g., foraging, fertilization) processes
for coral reef organisms. In the back reef of Discovery Bay, Jamaica, daytime mean
flow speeds were on average, 61% greater than at night during a given 24 hr period.
Wind speed was a significant predictor of flow speed in these shallow water environments, with the variation in wind speed able to explain 30% of the variation in flow
speed. Porter’s (1985) yearlong wind speed record in Discovery Bay indicated that the
time of maximum daily wind speed occurred during daylight hours for 93% of the year.
Activity of the fireworm, Hermodice carunculata (Pallas, 1766), represented by total
abundance in six, 1 × 30 m transects was negatively correlated with flow speed. Atmospheric and oceanographic conditions enhancing wind-dependent water flow in back
reef environments include prevalent tradewinds and negligible tidal currents, which
suggests that the diel variation in flow speed documented for Discovery Bay may be a
common phenomenon in similar environments. Such predictable environmental variability may be an important selective agent shaping the evolution of diel rhythms of reef
invertebrates and algae. Therefore, recent atmospheric and climatological shifts (e.g.,
frequency of El Niño events, global climate change) may exert additional selective pressure on the organisms found in these environments.
Although Darwin (1842) noted the influence of water motion on the structure and distribution of coral reefs one-and-a-half centuries ago (Dennison and Barnes, 1988), its
effect on the physiology and ecology of corals and coral reefs has only received significant attention over the past three decades. Early experimental work by Jokiel (1978),
demonstrating the influence of water motion on the growth, reproduction, and mortality
of three coral species has been followed by numerous studies examining flow-mediated
effects on various aspects of coral physiology, including: calcification and photosynthesis (Dennison and Barnes, 1988), respiration (Patterson et al., 1991), and the uptake of
nutrients such as phosphate (Atkinson and Bilger, 1992) and ammonium (Atkinson et
al., 1994; Thomas and Atkinson, 1997), or both (Atkinson et al., 2001). Similarly, the
effects of water motion on marine macroalgae have been well documented (see Hurd,
2000), with examples demonstrating flow-dependent primary productivity (Carpenter et
al., 1991), nitrogen fixation (Williams and Carpenter, 1998) and nutrient uptake (Larned
and Atkinson, 1997) for coral reef algae.
Several ecological studies have attributed the distribution of coral species on reefs
to gradients in flow regime (Brown and Dunne, 1980; Done, 1983; Sebens and Done,
1993). While wave force may be the most obvious flow-related effect on the distribution
of reef corals (Chamberlain and Graus, 1975; Geister, 1977; Tunnicliffe, 1982; Graus
and Macintyre, 1989), ecological processes such as interspecific competition may also
be influenced by gradients in flow speed (Genin et al., 1994). Work by Sebens et al.
281
282
(1996, 1997, 1998) examining the feeding ecology of scleractinian corals demonstrate
the importance of flow speed, as do similar studies with alcyonacean corals (Fabricius
et al., 1995) and gorgonians (Lasker, 1981).
When mapping the global distribution of coral reefs, the 20°C average minimum water
temperatures necessary to support most reef-building corals limits their range to within
~30° north and south of the equator (Barnes, 1987). This also corresponds to the region
of the earth’s surface influenced by the trade winds (Pinet, 1998), which are a product of
the uneven distribution of solar heat and the rotating earth. By using a specific example
from a Caribbean coral reef, we suggest a possible effect of this global pattern in wind
circulation on physical, and corresponding biological processes in specific shallow coral
reef ecosystems. Surface waves are predominately wind-generated, and dependent upon
the wind’s velocity, duration, and fetch. Shallow portions of the reef, such as the reef
crest, reef flat, and back reef may receive greater impact from wave-induced flow, rather
than the current-induced flow that usually dominates deeper reef habitats (Roberts and
Suhayda, 1983; Pickard, 1986). Several studies have empirically quantified and modeled
the contribution of wind-driven flow on circulation patterns in shallow reef systems
(Prager, 1991; Symonds et al., 1995; Kraines et al., 1998; Yamano et al., 1998).
Here we present results demonstrating significant diel variation in flow speed in back
reef habitats of Discovery Bay, Jamaica. We then provide evidence that these differences
are modulated by daily variation in the intensity of the trade winds. Next, an example
of the ecological consequences of these differences in flow speed is illustrated by examining foraging activity of the fireworm, Hermodice carunculata (Pallas, 1766). This
polychaete is an important omnivore on Caribbean coral reefs that preys on a variety
of scleractinian corals, milleporid hydrocorals, anemones, and gorgonians (reviewed in
Witman, 1988). Finally, we discuss the generality of these results, their broader implications with respect to the evolution of diel rhythms of reef invertebrates and algae, and
possible consequences of recent atmospheric and climatological shifts.
Study Site.—This study was conducted in the West Back Reef (WBR) of Discovery Bay
(18°28ʹ00ʺN, 77°24ʹ30ʺW), on the north coast of Jamaica (Fig. 1). The following summary of
climatological conditions experienced at this location is provided by Gayle and Woodley (1998).
In short, the northeast trade winds, along with the presence of an alternating sea- and land breeze
system, are the dominant weather features in this area. Daily trade winds act in concordance with
the local sea breeze system to produce a steady northeast wind that builds during daylight hours
to a maximum in the mid-afternoon. A nightly land breeze counteracts the diminished trade
winds, resulting in a period of minimal wind between 1900 and 0700 hrs. These patterns can be
obscured by the arrival of cold fronts generated from North America that disrupt local weather
for 3–4 d during December–March (Gayle and Woodley, 1998). Over the entire year, average
wind speeds are highest in May and lowest in September (Porter, 1985). In addition, Porter’s data
(1985) indicate that winds were nearly always from the northeast, and that wind speed, but not
direction, was subject to change. The mean diurnal tidal range is only 24.5 cm (NOAA/NOS,
2001), and tidal currents are negligible.
The WBR is dominated by a seagrass bed composed primarily of Thallasia testudinum (Banks
and Solandt ex. Koenig, 1805) with patches of sand, coral rubble, and hard pavement, and water depths between 1–3 m. Isolated coral colonies are scattered throughout the WBR, and are
dominated by members of Porites spp. and Siderastrea spp. There are a variety of mobile benthic
fauna, including echinoderms (Sides and Woodley, 1983; Aronson, 1993) and the fireworm, H.
carunculata.
GENOVESE AND WITMAN: DIEL VARIATION IN BACK REEF FLOW SPEEDS
283
Figure 1. Map of Jamaica showing location of Discovery Bay, and Discovery Bay Marine Laboratory (*) on the north shore of the island. The West Back Reef study site was located ~200 m
northeast of the laboratory, and 100 m south of the reef crest.
Flow Speed Measurement.—
Measurement.—The flow regime was quantified in the West Back Reef using
an electromagnetic current meter (InterOcean Model S4) fixed 0.5 m above the substratum, in 2
m of water. Flow speed was burst sampled at 2 Hz for 5 min every hour, yielding a sample of 600
data points. From this we calculated the mean flow speed (independent of flow direction) and the
standard deviation in flow. The duration for each deployment varied between 24–37 hrs, and was
limited by the internal memory of the instrument. Six instrument deployments were made during
January–March of 1991, with an additional nine deployments in January–March of 1992.
Flow records were analyzed by comparing mean hourly flow speeds between daylight (0600–
1800) and night (1800–0600) hrs. Although continuous data records could span several days in
instances where we were able to quickly download and re-deploy the current meter, comparisons
remained within a given day, due to the daily variation in atmospheric conditions which result in
differences in the magnitude of flow speed. Thus, in most instances, the day/night comparison
was a balanced design with a sample size of 12. In cases where record length exceeded 24 hrs
by up to 6 hrs, with no subsequent deployment, a single dataset was analyzed. Alternatively, in
multiple day deployments there were instances where the daily dataset at the start or end of the
deployment was < 24 hrs.
For a given sampling date, mean hourly flow speeds were compared between day and night
with a Mann-Whitney U test. Paired t-tests were employed to examine differences between paired
values of mean day and night flow speeds, as well as standard deviation in flow speeds, over the
entire course of this study. If an Fmax test indicated that sample variances were unequal between
day and night, the critical probability level was adjusted downward to the appropriate value.
Normality of the data was confirmed by comparing against an ideal normal distribution with a
Kolmogorov-Smirnov test of normality.
Wind Speed Measurement.—Hourly average wind speeds were measured using an anemometer (Casella London, Model W12041) mounted 12 m above sea level, atop the water tower
at Discovery Bay Marine Lab. The location of the water tower was ~200 m from the back reef
deployment location of the current meter. An hourly average wind speed measurement required
a manual reading at the beginning and end of the hour. The second of each paired anemometer
readings was scheduled to coincide with the end of an hourly S4 current meter burst sample, consistent with our hypothesis of wind-induced water flow. Wind speed measurements were opportunistically recorded whenever the current meter was deployed. The resulting paired wind speed
and flow speed data were used to construct a linear regression of flow speed as a function of the
square of wind speed, which is a better measure of the magnitude of wind forcing.
Annual Weather Record.—The dataset provided in Porter’s (1985) continuous weather record from Discovery Bay Marine Lab during the period from August 1983–July 1984 was used
284
to determine: 1) the time of maximum wind speed for each day and 2) the yearly average of the
daily ratio of maximum:mean wind speed. While Porter primarily used the PAR irradiance data
presented in his study to construct an annual carbon budget for the coral, Montastrea annularis
(Ellis and Solander, 1786), the wind data proved valuable for our present study.
Fireworm Foraging Activity.—To quantify the foraging activity of the fireworm, H. carunculata, under varying flow speeds, six permanent 1 × 20 m transects were established in the
WBR of Discovery Bay. Habitat type included patches of sand and coral rubble, within a seagrass
bed composed primarily of T. tesudinum. Fireworms often retreat within reef cavities when not
actively foraging, thus the number visible serves as a good indicator of predatory activity. The total density of fireworms observed in all six transects was recorded within an hour, and compared
with the coinciding flow measurements. All surveys were opportunistically conducted during
daylight hours in February 1991, as scheduling permitted. Fireworm activity (density +1) was
plotted as function of flow speed, with an exponential regression fitted to test if flow speed is a
significant predictor of fireworm density. Back reef flow speeds corresponding to the time each
set of transects were conducted, were obtained from S4 current meter records described above.
Results
Flow Speed Measurement.—Representative records of daily flow speeds in the
West Back Reef (WBR) of Discovery Bay for 1991 (Fig. 2A) and 1992 (Fig. 2B) clearly
illustrate the differences in flow speed between day and night. In addition to greater
mean flow speeds during the day, there is also an increased variability in flow speed, due
to the oscillatory nature of this wind-driven flow. This is evidenced by a concomitant
increase in the standard deviations associated with greater mean flow speeds (Fig. 2).
Plots comparing the mean day and night flow speeds for each of the sampling dates in
1991 (Fig. 3A) and 1992 (Fig. 3B) show that in all cases, mean flow speeds were greater
during daylight hours. Of the 21 sampling dates, there were significant differences in
seventeen (81%) cases. However it is interesting to note the variation in the magnitude
of mean flow speeds among sampling dates (Fig. 3). During the given 24 hr sampling
periods, mean daytime flow speeds were an average of 61% greater than at night. Despite unequal variances between mean day and night flow speeds (Fmax= 2.95, P = 0.02)
daytime flow speeds were still greater when compared with an adjusted critical P-value
(paired t-test: t = 5.497, 20 df, P < 0.0001). A similar comparison of the average standard deviation in flow speeds indicated greater variation for daytime flow measurements
(paired t-test: t = 2.24, 20 df, P = 0.037), with daytime values an average of 231% greater
than nighttime variation in flow speeds.
Wind Speed Measurement and Annual Weather Record.—A total of 177 recorded hourly wind speed measurements was paired with simultaneous flow speed measurements (Fig. 4). Wind speed was a significant predictor of flow speed (F = 71.01, P <
0.0001, df = 1, 176), with the variation in the square of wind speed able to explain 29% of
the variation in flow speed. A frequency histogram plotting the time of maximum wind
speed from Porter’s (1985) yearlong climatological record for Discovery Bay shows that
in 1984, this occurs during daylight hours during 93% of the year (Fig. 5). In fact, the
shape of the frequency distribution is a good proxy for the daily pattern of wind speed
intensity in Discovery Bay; wind speeds increase during daylight hours to a maximum
in the mid-afternoon, and then decrease into the evening. In addition, for the entire annual weather record, the daily maximum wind speed was on average 2.95 times greater
than the daily mean wind speed (± 0.70 SD).
Fireworm Foraging Activity.—Foraging activity of the fireworm, H. carunculata,
represented by total abundance in six, 1 × 30 m transects was negatively correlated with
285
Figure 2. Representative records of mean hourly flow speeds in the West Back Reef of Discovery
Bay for A) 1991 and B) 1992. Flow speeds were sampled at 2 Hz during a 5 min hourly block (n
= 600); error bars represent standard deviations. White and shaded backgrounds correspond to
day and night hours, respectively.
flow speed (Fig. 6). A total of 19 surveys of the field transects indicated a strong decline
in fireworm abundance with an increase in flow speed. This relationship was best fitted to an exponential regression model given both the high rate of decline, and the fact
that we could expect asymptotes on both ends of the fitted line, as there is no reason to
believe both flow speed and fireworm density could not continue to increase beyond the
limits of our data on their respective ends. From the exponential transformation, Y= 31.8
* e −0.18x, the linear form of the model, ln Y= ln 36.37 − 0.19 X, can be used to assess
the adequacy of the independent variable (flow speed) in the model with a t-test. In this
model flow speed is a strong predictor of fireworm density (t = −6.572, P < 0.0001, df
= 18).
Discussion
Taken together, the results presented above make a compelling case for wind-mediated diel variation in flow speed in the back reef environment of Discovery Bay, Jamaica.
Wind speed is a significant predictor of flow speed (Fig. 4), which in turn is consistently
286
Figure 3. Plots comparing the mean day (open circles) and night (solid squares) flow speeds for
each of the sampling dates in A) 1991 and B) 1992. Each data point represents a mean flow speed
value, with n = 12, unless otherwise noted by the actual sample size in parentheses above or below
the error bars (= standard deviation) for day and night values, respectively. Arrows in A and B
point to mean day and night flow speeds for records presented in Figure 2A,B, respectively. Significance levels as follows: *, P < 0.05; **, P < 0.01; ***, P < 0.001; ns, not significant.
greater during the daytime (Figs. 2,3). Data extracted from Porter’s (1985) yearlong
climatological record at Discovery Bay indicate that maximum daily wind speeds occur
during daytime hours for 93% of the year (Fig. 5), and that the daily maximum wind
speed is on average, nearly three times greater than the mean wind speed. These data are
consistent with observations by Gayle and Woodley (1998), and support the conclusion
that throughout the year, flow speeds are consistently greater during the day, in the back
reef of Discovery Bay. The increased variability (by 231%) in flow speed noted during
the day (Figs. 2,3) is indicative of the increase in wave-induced oscillatory (bi-directional) flow. The increase in both mean flow speed and variance may have significant
effects on several physiological and ecological parameters discussed below.
287
Figure 4. Regression of mean flow speed in the West Back Reef of Discovery Bay Jamaica as a
function of wind speed squared. The square of wind speed was a significant predictor of flow
speed (F = 71.01, P < 0.0001, df = 1, 176).
However, we point out that these observations were conducted at a single site on a
single reef, on a single island in the Caribbean Sea. While there is sufficient temporal
replication of our data, with sampling conducted during 3 mo in each of 2 yrs, there is
no spatial replication. The obvious question becomes, “How typical are these results at
other locations?” Addressing this question requires an understanding of the conditions
under which we would expect to observe this phenomenon. Both the circumglobal trade
winds as well as any local sea breeze system can provide the daily variation in wind
stress on the sea surface, which in turn generates wind waves that can directly affect
local flow speeds in shallow reef habitats. Water flow induced by this wind stress can be
very significant in shallow reef habitats of 1–3 m (Andrews and Pickard, 1990). For these
patterns in flow speed to exhibit the diel variation presented herein additionally requires
a lack of strong local oceanographic currents, and a weak tidal effect or range. Thus,
while tide may be the most dominant factor influencing water flow on the Great Barrier
Reef (Parnell, 1988), in the Caribbean, Glynn (1973) found that total volume flow across
a reef in Puerto Rico was strongly influenced by prevailing winds.
In many respects Discovery Bay’s geographic location could be considered a best-case
scenario for the proposed mechanism of wind-mediated diel variation in flow speeds;
located on the north coast of Jamaica, it is directly impacted by the northeasterly trade
winds, while enjoying shelter from oceanic swells by Cuba, located 150 km to the north
(Gayle and Woodley, 1998). The daily sea breeze acts in conjunction with the trade
winds to produce steady inshore winds, while the nightly land breeze counteracts the
diminished trade winds. However similar observations of daily variation in wind speed
and sea state on the western (Negril) and southern (Kingston) coasts of Jamaica (S.
Genovese, pers. obs.), suggest that overall, a similar wind regime exists around the
island. Additional data for wind, tidal, and current conditions at 20 other island and
coastal locations throughout the Caribbean (UNESCO, 1998) suggest that the condi-
288
Figure 5. Frequency histogram plotting the time of maximum wind speed on a daily basis in 1984,
from Porter’s (1985) yearlong climatological record for Discovery Bay.
tions reported for Discovery Bay, Jamaica may be widespread throughout the region.
However, it is clear that empirical measurements throughout the Caribbean are the only
means to definitively confirm this hypothesis.
It is worth noting that as biologists, we have provided an Eulerian view of water flow
in the back reef. That is, we are interested in flow past a fixed point on the reef, corresponding to an organism’s perspective. Oceanographic studies examining reef flow
(Prager, 1991; Kraines et al., 1998; Yamano et al., 1998) have generally adopted a Lagrangian view that is concerned with overall circulation on the reef. These studies are
primarily interested in parameters such as inflow and outflow, residence time, and turnover of water masses on the reef.
There was a strong, negative effect of increased flow speed on the daytime foraging
activity of the fireworm, H. carunculata (Fig. 6). To avoid the addition of light regime
as a confounding factor, no surveys were conducted at night. However, given the results
presented in this study, it is possible to conclude that fireworm foraging activity will be
greater during the night, providing one example of how such predictable environmental
variability (reduced nightly flow speed) could possibly serve as an important selective
agent shaping the evolution of diel rhythms of reef organisms. The mechanism most
likely to explain this pattern is that the ambulatory ability of fireworms decreases with
increasing flow speeds. This effect may be more pronounced in an oscillatory flow regime, noted by the increased variance with increasing mean flow speeds (Figs. 2,3).
However, high flow has also been demonstrated to provide hydrodynamic refuge from
mobile predators utilizing olfactory-mediated search and guidance mechanisms (Weissburg and Zimmer-Faust, 1993), and could be important for similar predator-prey interactions in shallow reef ecosystems (Levitan and Genovese, 1988). Finally, the activity
patterns of reef fishes inhabiting shallow reef habitats (Adams and Ebersole, 2002) may
also be influenced by the local flow regime.
Other situations where diel variation in flow speed could be significant are for mass
transfer rates and suspension feeding abilities of cnidarians. The interactive effects of
289
Figure 6. Foraging activity of the fireworm, Hermodice carunculata, represented by total abundance in six, 1 × 30 m transects, was negatively correlated with flow speed in the West Back Reef
of Discovery Bay, Jamaica. Mean flow speeds corresponding to the time each set of transects
were conducted, were obtained from S4 current meter records.
water flow and morphology on both mass transfer of nutrients and gases across the diffusive boundary layer adjacent to the surface of cnidarians (Patterson et al., 1991; Patterson, 1992; Lesser et al., 1994; Thomas and Atkinson, 1997; Gardella and Edmunds,
2001) and suspension feeding by cnidarians (Lasker, 1981; Fabricius et al., 1995; Sebens
et al., 1996, 1997, 1998) have been well documented. Given that mass transfer is critical
for autotrophic productivity during the day, and that suspension feeding occurs primarily at night, the diel variation in flow speeds documented in this study suggests selection
pressure on morphological characteristics may need to balance these potentially conflicting performance demands on cnidarians.
Algal turf canopies provide another example related to mass transfer issues, and how
diel variation in flow speed may enhance existing gradients in oxygen concentrations. In
flume experiments, daytime algal photosynthesis enriched the near bottom water with
oxygen, while at night algal respiration caused a substantial uptake of oxygen from the
seawater, with the water within the bottom boundary layer commonly becoming hypoxic (Breitbarth, 2000). Similar trends were reported in the same study for field measurements in the back reef of Discovery Bay, Jamaica. Because oxygen concentrations within
the bottom boundary layer were significantly affected by water velocity, a decrease in
nightly flow speed would exacerbate any hypoxic conditions present within the canopy,
especially since algal turf canopies reduce flow speed directly above the substratum
(Carpenter and Williams, 1993). Breitbarth (2000) further postulated that a hypoxic bottom boundary layer caused by algal respiration could provide a competitive advantage
for turf algae via inhibition (sensu Connell and Slayter, 1977).
On a global scale, the trade winds driving the patterns of flow speed reported here are
intertwined with El Niño events. During an El Niño event, trade winds in the South Pacific are reduced, specifically causing increased sea surface temperatures in the eastern
290
Pacific due to reduced upwelling in the region (Pinet, 1998). However, coincidental circumglobal reports of coral bleaching (references in Bruno et al., 2001), in likely response
to the 1–4°C increase in maximum summer temperatures associated with the 1997–98
El Niño event, indicate that effects on shallow reef systems could be widespread. Calm
weather (i.e., a lack of wind) has been implicated in coral bleaching (Jaap, 1979) due
to an increase in seawater temperature (Glynn, 1973; Jaap, 1979) and irradiance levels
(Lesser et al., 1990). In either case the lack of wind results in decreased flow speeds in
shallow reef habitats where coral bleaching is common. Decreased daytime flow speeds
in turn will reduce both the circulation of shallow water masses (which increases local
water temperature), and the scattering of visible and UV light in the water column. Low
wind speeds have also been implicated in coral bleaching due to decreased evaporative
cooling (Hendee et al., 2001), but Smith (2001) suggests that, in addition, wind-induced
currents may influence advective heat transport.
The uncertainty of whether global warming will affect the frequency and intensity
of El Niño events (Fedorov and Philander, 2000) has potential implications for the biological and ecological processes impacted by the diel variation in flow speed discussed
above. While reef-building corals will most certainly be impacted by global climate
change, the topographic structure and complexity they provide to the reef ecosystem
will have cascading effects upon the associated community. A better understanding of
the prevalence of wind-mediated diel variation in flow speed in shallow reef environments will be useful in predicting these impacts.
Acknowledgements
The East/West Marine Biology Program provided logistical and financial support during our
visits to Jamaica, and the students of E/W VII assisted with fireworm transects. Fieldwork at
Discovery Bay Marine Lab was facilitated by J. Woodley and P. Gayle. P. Edmunds kindly provided the graphics used in Figure 1. S.J.G would like to extend special thanks to C. Dahlgren
and J. Marr for an invitation to the Back Reef Habitat Workshop. J. Leichter, K. Sebens, and
an anonymous reviewer provided comments that significantly improved this manuscript. This
paper is funded in part by a grant from the Caribbean Marine Research Center (CMRC Project
# CMRC-00-IXNR-03-01A) National Undersea Research Program, National Oceanic and Atmospheric Administration (NOAA), U.S. Environmental Protection Agency, and Environmental
Defense. Views expressed herein are those of the authors and do not necessarily reflect the views
of the CMRC, or any of the supporting agencies. This is Contribution Number 247 of the Marine
Science Center of Northeastern University, and Contribution Number 662 of the Discovery Bay
Marine Laboratory.
Literature Cited
Adams, A. J. and J. P. Ebersole. 2002. Use of back-reef and lagoon habitats by coral reef fishes.
Andrews, J. C. and G. L. Pickard. 1990. The physical oceanography of coral-reef systems. Pages
11–48 in Z. Dubinsky, ed. Ecosystems of the world, Vol. 25. Elsevier, Amsterdam.
Aronson, R. B. 1993. Hurricane effects on backreef echinoderms of the Caribbean. Coral Reefs
12: 139–142.
Atkinson, M. J. and R. W. Bilger. 1992. Effects of water velocity on phosphate uptake in coral
reef-flat communities. Limnol. Oceanogr. 37: 261–272.
___________, E. Kotler, and P. Newton. 1994. Effects of water velocity on respiration, calcification and ammonium uptake of a Porites compressa community. Pac. Sci. 48: 296–303.
291
___________, J. L. Falter, and C. J. Hearn. 2001. Nutrient dynamics in the Biosphere 2 coral reef
mesocosm:water velocity controls NH4 and PO4 uptake. Coral Reefs. 20: 341–346.
Barnes, D. J. 1987. Invertebrate zoology, 5th Edition. Saunders, New York. 893 p.
Breitbarth, E. 2000. Oxygen dynamics in turf algae canopies on coral reefs. M.S. Thesis. Northeastern University, Boston. 92 p.
Brown, B. E. and R. P. Dunne. 1980. Environmental controls of patch reef growth and development, Anegada, British Virgin Islands. Mar. Biol. 56: 85–96.
Bruno, J. F., C. E. Siddon, J. D. Witman, P. L. Colin, and M. A. Toscano. 2001. El Niño related
coral bleaching in Palau, Western Caroline Islands. Coral Reefs. 20: 127–136.
Carpenter, R. C., J. M. Hackney, and W. H. Adey. 1991. Measurements of primary productivity and
nitrogenase activity of coral reef algae in a chamber incorporating oscillatory flow. Limnol.
Oceanogr. 36: 40–49.
_____________ and S. L. Williams. 1993. Effects of algal turf canopy height and microscale
topography on profiles of flow speed in a coral forereef environment. Limnol. Oceanogr. 38:
687–694.
Chamberlain, J. A. and R. R. Graus. 1975. Water flow and hydromechanical adaptations of
branched reef corals. Bull. Mar. Sci. 25: 112–125.
Connell, J. H. and R. O. Slayter. 1977. Mechanisms of succession in natural communities and their
role in community stability and organization. Am. Nat. 111: 1119–1144.
Darwin, C. 1842. The structure and distribution of coral reefs. Smith, Elder and Co., London.
214 p.
Dennison, W. C. and D. J. Barnes. 1988. Effect of water motion on coral photosynthesis and calcification. J. Exp. Mar. Biol. Ecol. 115: 67–77.
Done, T. J. 1983. Coral zonation:its nature and significance. Pages 107–147 in D.J. Barnes, ed.
Perspectives on coral reefs. Australian Institute of Marine Science, Manuka.
Fabricius, K., A. Genin, and Y. Benayahu. 1995. Flow-dependent herbivory and growth in zooxanthellae-free soft corals. Limnol. Oceanogr. 40: 1290–1301.
Fedorov, A. V. and S. G. Philander. 2000. Is El Niño changing? Science 288: 1997–2002.
Gardella, D. J. and P. J. Edmunds. 2001. The effect of flow and morphology on boundary layers
in the scleractinians Dichocoenia stokesii (Milne-Edwards and Haime) and Stephanocoenia
michilini (Milne-Edwards and Haime). J. Exp. Mar. Biol. Ecol. 256: 279–289.
Gayle, P. M. H. and J. D. Woodley. 1998. Discovery Bay, Jamaica. Pages 17–33 in B. Kjerve, ed.
Caricomp: Caribbean coral reef, seagrass and mangrove sites. UNESCO, Paris.
Geister, J. 1977. The influence of wave exposure on the ecological zonation of Caribbean coral
reefs. Proc. 3rd Intl. Coral Reef Symp., Miami. 1: 23–30.
Genin, A., L. Karp, and A. Miroz. 1994. Effects of flow on competitive superiority in scleractinian
corals. Limnol. Oceanogr. 39: 913–924.
Glynn, P. W. 1973. Ecology of a Caribbean coral reef: the Porites reef-flat biotope: Part 1. Meteorology and hydrography. Mar. Biol. 20: 297–318.
Graus, R. R. and I. G. Macintyre. 1989. The zonation patterns of Caribbean coral reefs as controlled by wave and light energy input, bathymetric setting, and reef morphology: computer
simulation experiments. Coral Reefs 8: 9–18.
Hendee, J. C., E. Mueller, C. Humphrey, and T. Moore. 2001. A data-driven expert system for
producing coral bleaching alerts at Sombrero reef in the Florida Keys, USA. Bull. Mar. Sci.
69: 673–684.
Hurd, C. L. 2000. Water motion, marine macroalgal physiology, and production. J. Phycol. 36:
453–472.
Jaap, W. C. 1979. Observations on zooxanthellae expulsion at Middle Sambo reef, Florida Keys.
Bull. Mar. Sci. 29: 414–422.
Jokiel, P. L. 1978. Effects of water motion on reef corals. J. Exp. Mar. Biol. Ecol. 35: 87–97.
Kraines, S. B., T. Yanagi, M. Isobe, and H. Komiyama. 1998. Wind-wave driven circulation on the
coral reef at Bora Bay, Miyako Island. Coral Reefs 17: 133–143.
292
Larned, S. T. and M. J. Atkinson. 1997. Effects of water velocity on NH4 and PO4 uptake and nutrient-limited growth in the macroalga Dictyosphaeria cavernosa. Mar. Ecol. Prog. Ser. 157:
295–302.
Lasker, H. R. 1981. A comparison of the particulate feeding abilities of three species of gorgonian
soft coral. Mar. Ecol. Prog. Ser. 5: 61–67.
Lesser, M. P., W. R. Stochaj, D. W. Tapley, and J. M. Schick. 1990. Bleaching in coral reef anthozoans: effects of irradiance, ultraviolet radiation, and temperature on the activities of protective
enzymes against active oxygen. Coral Reefs 8: 225–232.
____________, V. M. Weis, M. R. Patterson, and P. L. Jokiel. 1994. Effects of morphology and
water motion on carbon delivery and productivity in the reef coral, Pocillopora domicornis
(Linaeus): diffusion barriers, inorganic carbon limitation, and biochemical plasticity. J. Exp.
Mar. Biol. Ecol. 178: 153–179.
Levitan, D. R. and S. J. Genovese 1989. Substratum-dependent predator-prey dynamics: patch
reefs as refuges from gastropod predation. J. Exp. Mar. Biol. Ecol. 130: 111–118.
NOAA/National Ocean Service. 2001. Tide tables 2002: East coast of North and South America.
McGraw-Hill, Camden. 396 p.
Parnell, K. E. 1988. The hydrodynamics of fringing reef bays in the Great Barrier Reef Marine
Park. Proc. 6th Intl. Coral Reef Symp., Townsville. 2: 503–508.
Patterson, M. R. 1992. A chemical engineering view of cnidarian symbioses. Am. Zool. 32: 566–
582.
______________, K. P. Sebens, and R. R. Olson. 1991. In situ measurements of flow effects on
primary production and dark respiration in reef corals. Limnol. Oceanogr. 36: 936–948.
Pickard, G. L. 1986. Effects of wind and tide on upper-layer currents at Davies Reef, Great Barrier
Reef, during MECOR (July-August 1984). Aust. J. Mar. Freshwater Res. 37: 545–565.
Pinet, P. R. 1998. Invitation to oceanography. Jones and Bartlett, Boston. 508 p.
Porter, J. W. 1985. The maritime weather of Jamaica: its effects on annual carbon budgets of the
massive reef-building coral Montastrea annularis. Proc. 5th Intl. Coral Reef Symp., Tahiti. 6:
363–379.
Prager, E. J. 1991. Numerical simulation of circulation in a Caribbean-type backreef lagoon. Coral
Reefs 10: 177–182.
Roberts, H. H. and J. N. Suhayda. 1983. Wave-current interactions on a shallow reef. Coral Reefs
1: 209–214.
Sebens, K. P. and T. J. Done. 1993. Water flow, growth form and distribution of scleractinian corals: Davies reef (GBR), Australia. Proc. 7th Intl. Coral Reef Symp., Guam. 1: 557–568.
____________, K. S Vandersall, L. A. Savina, and K. R. Graham. 1996. Zooplankton capture by
two scleractinian corals, Madracis mirabilis and Montastrea cavernosa, in a field enclosure.
Mar. Biol. 127: 303–318.
Sebens, K. P., J. Witting, and B. Helmuth. 1997. Effects of water flow and branch spacing on particle capture by the reef coral Madracis mirabilis. J. Exp. Mar. Biol. Ecol. 209: 233–259.
___________, S. P. Grace, B. Helmuth, E. J. Maney Jr., and J. S. Miles. 1998. Water flow and prey
capture by three scleractinian corals, Madracis mirabilis, Montastrea cavernosa, and Porites
porites, in a field enclosure. Mar. Biol. 131: 347–360.
Sides, E. M. and J. D. Woodley. 1985. Niche separation in three species of Ophiocoma (Echinodermata:Ophiuroidea) in Jamaica, West Indies. Bull. Mar. Sci. 36: 701–715.
Smith, N. P. 2001. Weather and hydrographic conditions associated with coral bleaching: Lee
Stocking Island, Bahamas. Coral Reefs 20: 415–422.
Symonds, G., K. P. Black, and I. R. Young. 1995. Wave-driven flow over shallow reefs. J. Geophys.
Res. 100: 2639–2648.
Thomas, F. I. M. and M. J. Atkinson. 1997. Ammonium uptake by coral reefs: effects of water
velocity and surface roughness on mass transfer. Limnol. Oceanogr. 42: 81–88.
Tunnicliffe, V. 1982. The effects of wave-induced flow on a reef coral. J. Exp. Mar. Biol. Ecol.
64: 1–10.
293
UNESCO. 1998. CARICOMP – Caribbean coral reef, seagrass and mangrove sites. Coastal region
and small island papers 3. UNESCO, Paris. 347 p.
Weissburg, M. J. and R. K. Zimmer-Faust. 1993. Life and death in moving fluids: hydrodynamic
effects on chemosensory-mediated predation. Ecol. Monogr. 74: 1428–1443.
Williams, S. L. and R. C. Carpenter. 1998. Effects of unidirectional and oscillatory water flow on
nitrogen fixation (acetylene reduction) in coral reef algal turfs, Kaneohe Bay, Hawaii. J. Exp.
Mar. Biol. Ecol. 226: 293–316.
Witman, J. D. 1988. Effects of predation by the fireworm Hermodice carunculata on milleporid
hydrocorals. Bull. Mar. Sci. 42: 446–458.
Yamano, H., H. Kayanne, N. Yonekura, and H. Nakamura. 1998. Water circulation in a fringing
reef located in a monsoon area: Kabira Reef, Ishigaki Island, Southwest Japan. Coral Reefs
17: 89–99.
Address: (J.D.W.) Department of Ecology and Evolutionary Biology, Brown University, Box
G-W, Providence, Rhode Island 02912. E-mail: <Jon_Witman@brown.edu>. Corresponding
Author: (S.J.G.) East/West Marine Biology Program, Marine Science Center, Northeastern
University, East Point, Nahant, Massachussetts 01908. E-mail: <s.genovese@neu.edu>.
LARGE-SCALE ECOLOGICAL IMPACTS OF DEVELOPMENT ON
TROPICAL ISLANDS SYSTEMS: COMPARISON OF DEVELOPED
AND UNDEVELOPED ISLANDS IN THE CENTRAL BAHAMAS
Kathleen Sullivan Sealey
ABSTRACT
The relationship between density of development and the health of nearshore marine
habitats is explored through spatial and temporal comparisons of patch reef environments in the central Bahamas. Nearshore patch reefs are important fish habitats, and
tend to have high, but variable, coral cover and benthic diversity in the Bahamian archipelago. Twelve patch reef stations were established off developed and undeveloped
islands in the central Bahamas. Environmental parameters were measured over an
18-mo period to examine seasonal, tidal, and diurnal variability. Water quality measurements were not significantly different between developed and undeveloped sites
for temperature, salinity, dissolved oxygen, chlorophyll-a, total nitrogen and total
phosphorus. Only turbidity measurements were significantly different among sites, attributed to storm events. Ecological surveys recorded macroalgae species, stony coral
species, coral cover, and coral vitality. Significant differences in species composition
between developed and undeveloped stations were seen, with a higher coral diversity,
lower coral cover, and higher incidence of coral lesions on developed patch reefs. A 53yr comparison of nearshore environments from aerial imagery showed significant loss
of patch reefs and seagrass areas with increasing development density. Results stress
the importance of comparison reefs in marine protected areas for evaluating impacts of
coastal development on nearshore marine habitats.
This paper is one in a series resulting from a workshop held at the Caribbean Marine
Research Center (December 2001) to evaluate the importance of back reef systems for
supporting biodiversity and productivity of marine ecosystems. This paper specifically
addresses the potential impacts of coastal development on nearshore patch reef environments by spatial and temporal comparisons of heavily developed (“developed”) and
less-developed (“undeveloped”) islands in the central Bahamas. The Bahamian archipelago is made up of shallow-water carbonate limestone environments with few surface
water resources; a critical feature is the continuous solution of the limestone rock, and
consequent permeation of seawater beneath all the islands (Sealey, 1994). The marine
and terrestrial habitats adjacent to the shoreline interact ecologically in terms of nutrient
flux through rainfall run-off, seepage of ground water, and detrital material moving on
and off the islands. Most people in the Bahamian archipelago live within 2 km of the
sea, and both countries that occupy the islands (The Bahamas and the Turks and Caicos
Islands) share a common culture as well as economic interest in the coastal and marine
resources.
The typical pattern of development on these carbonate islands requires completely
clearing the site of all vegetation, and leveling the rocky terrain, with fill if necessary.
Threats to nearshore marine environments, particularly patch reefs, occur on several levels. The most severe threat is the physical elimination of reefs for the alteration and restructuring of the shoreline to accommodate marina, harbor, and residential development.
A second threat to these reefs comes from the acute and chronic sedimentation generated
when coastal vegetation is removed, and the soil/sand exposed to erosion by rain during
construction. Often exotic plants are used in landscaping, and native coastal plants are
295
296
removed permanently, leaving little or no vegetation buffer between buildings and the
sea. Lastly, a threat exists, though poorly documented in the Bahamas specifically, from
nutrient seepage from cesspit systems very close to the sea, or in contact with ground water. With the exception of downtown Nassau and some large resorts, the common means
of wastewater treatment and disposal is the use of “soak-aways” or cesspit systems constructed onsite (Cant, 1997). All onsite disposal systems installed on limestone islands
are shown to enrich nearshore water with inorganic nutrients and some amount of organic
matter (U.S. EPA, 1983, 1991).
Human activities on land inevitably increase nutrient inputs to coastal waters from
deforestation, wastewater, fertilizer, and other sources (Bell, 1992). Human population
growth in the Bahamian archipelago is one of the highest in the insular Caribbean, with
a 10-yr intercensal growth rate of 19% for the Bahamas (Department of Statistics, 2002),
and 41.7% for the Turks and Caicos Islands (UNEP, 2002). The population of the Bahamas is projected to increase from 303,000 in 2000 to 357,000 in 2010. Seventy percent of
the population in the Bahamas live on New Providence Island. Although the population
is small in comparison to the islands of the Greater Antilles, population size is similar to
eastern Caribbean island nations. The small island nature of the archipelago places all
development in close proximity to nearshore reef environments. The Bahamas is not unlike the Florida Keys (Hoffmeister, 1974), where the process of eutrophication has been
studied intensively. Major pathways of nutrient input to waters of the Florida Keys include
onsite sewage disposal systems (Lapointe et al., 1992), and submarine groundwater discharge (Lapointe and Matzie, 1992). A major difference between the Florida Keys and the
central Bahamas is circulation. The Bahamian islands are not adjacent to the peninsula
of south Florida, and have not been targeted for any significant fill between islands for
causeway construction (See Lott et al., 1996 for historical overview of island fill in the
Florida Keys).
Considerable concern exists over the loss of live coral and the decline of coral reefs
caused by macroalgae proliferation in the tropics (e.g., Dustan and Halas, 1987; Hallock
et al., 1993; overview by McCook et al., 2001). Coral reef ecosystems are very sensitive to environmental perturbations for several reasons: 1) key reef-building organisms
such as corals have very narrow physiological tolerances, 2) interactions of key species (plant-herbivore, algae-coral) are easily disrupted by human-induced perturbations,
and 3) the toxic effects of introduced materials are enhanced because of tropical water
temperatures and higher metabolic rates (Pastorak and Bilyard, 1985). The stress of eutrophication or nutrient-enrichment on coral reefs likely contributes to this macroalgae
proliferation, and appears to be related to the proximity of coastal development as a
source of pollution, particularly sewage (Bilyard, 1985; Marszalek, 1987; Pastorak and
Grigg and Dollar, 1990). Throughout developed areas of the tropics, nutrient input from
human sources constitutes one of the greatest threats to coral reefs, often resulting in a
chronic eutrophication process (Mee, 1988). Indeed, this threat may be even greater in
the Bahamian archipelago because, despite having a large area of shallow-water banks,
fisheries production is in fact tied to habitats very close to islands along the platform
margin (see Stoner et al., 1994; Colin, 1995; Sluka et al., 1996; Lipcius et al., 1997). How
can nearshore habitats be monitored and evaluated to detect impacts from development?
Clark and Green (1988) summarized several criteria to be used in pollution assessment,
including the importance of selecting one or more sites as controls (relatively non-impacted), and the documentation of nuisance physical and biological variables from seasonal,
tidal, or diurnal cycles.
SULLIVAN SEALEY: DEVELOPMENT IMPACTS ON BAHAMIAN NEARSHORE REEFS
297
The present study evaluates criteria for the selection of control sites for nearshore patch
reefs based on their proximity to coastal development. With coastal development, nearshore patch reefs can be physically destroyed (loss of habitat) or degraded by sediment
and/or nutrient enrichment (habitat degradation or phase-shifts). Habitat degradation
should be detected by changes in both environmental parameters (physical and chemical
properties of the water column and sediments) and ecological parameters (benthic species
composition and coverage). No islands in the archipelago can be considered “undeveloped” in the purest sense (Jackson et al., 2001) as historical changes in land cover and
vegetation, as well as extinctions and extirpations, have accounted for unknown ecological
changes over the past two millennia. For this paper, “undeveloped” is less-developed, and
at the lowest level of contemporary human occupation and use in the central Bahamas.
Both the undeveloped and developed patch reef stations are in areas of restricted fishing;
undeveloped patch reefs are within the Exuma Cays Land and Sea Park, a marine fisheries
reserve, and the developed patch reef stations are within the no-fishing limits adjacent to
settlements.
This study examined three questions: 1) Are spatial comparisons made between patch
reefs adjacent to developed and undeveloped islands valid, and are there appropriate “reference sites” within a larger ecological system (e.g., banks)? 2) What environmental and
ecological parameters are most useful in characterizing the condition of reefs based on
proximity to coastal development in the central Bahamas? and 3) What is the nature of
temporal changes in reef environments as development density increases (e.g., Montagu
Bay off New Providence ) given a detailed history of aerial photography and historical
records? The answers are critical to developing both a better ecological understanding of
nutrient flux in the coastal zone as well as management guidelines for development practices within the archipelago.
Study Area
The central Bahamas includes the eastern half of the Great Bahama Bank, and is comprised of
about ten main islands and many smaller cays and rocks extending along the platform margin of
the bank (Fig. 1). The bank is bounded by the western Atlantic to the east, and is penetrated by a
submarine trough, probably of tectonic origin (Exuma Sound). Most of the bank area is very shallow (mean depth of 5–10 m). The selection of patch reef stations was based on previous studies of
the reefs as grouper habitat, and physical similarities in size, proximity of islands, and exposure to
prevailing winds. The twelve patch reefs selected were within 5 km of the islands of New Providence
(developed) and Warderick Wells, Halls Pond, and Bell Island (undeveloped; Table 1; Fig. 2). All
patch reefs were located on the leeward sides of islands, somewhat protected from direct wave and
wind action. The islands themselves are located along the Great Bahama Bank platform margin, the
Exuma Cays facing the western edge of Exuma Sound, and New Providence facing the southern
edge of the Northeastern Providence Channel. Several previous studies have reported on the local
oceanography, and most notably the tidal-generated currents adjacent to islands and tidal inlets;
tidal currents can be quite strong, as high as 150 cm s−1 (Lang et al., 1988; Stoner et al., 1994; Colin,
1995; Lipcius et al., 1997). Initial observations of the sites showed well developed, dome-type patch
reefs.
Dome-type patch reefs have a diversity of dominant biota and exhibit great variability in topographic complexity (Jones, 1977). Mature patch reefs in the Bahamas are often dominated by large
colonies of the species Montastraea annularis (Ellis and Solander, 1786), Montastraea cavernosa
Linnaeus, 1767, Siderastrea sidereal (Ellis and Solander, 1786), Colpophyllia natans (Houttuyn,
1772), and Diploria clivosa (Ellis and Solander, 1786) (Sluka et al., 1996). The patch reefs selected
were previously investigated as a habitat for juvenile groupers (Sluka et al., 1994, 1999) and the
298
Figure 1. Map of the central Bahamas with the location of New Providence Island and the northern Exuma Cays. Bathymetric contours represent 200 m and define the platform margin of the
Great Bahama Bank.
patch reefs’ benthic species composition had been characterized in terms of coral and sponge cover
(Sullivan and Chiappone, 1992). Small patch reefs in sheltered, nearshore environments of the
central Bahamas were shown to be important fisheries habitats, and were highly variable in coral
species present as well as overall coral cover. All patch reefs tended to have higher coral cover than
other reef and non-reef hard bottom habitats (Sluka et al., 1997). The twelve patch reefs were small,
ranging from 98–660 m2 (size measured by two perpendicular survey tapes stretched along the
sides of the patch, area reported in square meters), with an average size of 350 m2 (Table 1).
The methods used can be divided into three sections: measurement of environmental parameters
(water quality and sediment); measurement of ecological parameters (species composition and cover, coral lesions); and temporal image analysis of 50-yr changes to the Montagu Bay environment
as development density increased fourfold.
Environmental Parameters.—Measurement of water quality and sediment parameters adjacent to the patch reef stations was aimed at first characterizing the natural variability of these
parameters for nearshore patch reefs, and second, evaluating the usefulness of the Exuma Cays
as an undeveloped comparison site to New Providence (heavily developed). Variability in water
quality on the patch reef stations was likely attributed to tidal cycles, diurnal cycles, season,
and distance from shore. Water quality assessment involved surveying the water column directly
above the patch reefs (Fig. 2). Water quality sampling was carried out over a 2-wk period in each
of February 1998 and October 1998 (1 wk at developed island stations, and 1 wk at undeveloped
island stations). Water quality samples were taken via Niskin bottles deployed at the reef sites at
1-m depth. Salinity, temperature, and dissolved oxygen were measured in situ with a YSI probe
(Yellow Springs Instruments, Model 85, oxygen electrode calibrated daily). Sample bottles were
placed in a cooler and returned to shore for turbidity measurements via a LaMotte turbidity me-
299
Table 1. Summary of patch reef stations with location, area, depth, and brief description. Twelve
patch reefs were selected for sampling, six adjacent to a heavily developed island, New Providence,
and six adjacent to undeveloped islands in the Exuma Cays Land and Sea Park, central Bahamas.
Site name
Station
number
Patch reef
area (m2)
Latitude
Longitude
Depth range Distance
at low tide from shore
(m)
(m)
Montagu Bay off New Providence Island (developed)
Lighthouse
DEV1
98
N 25º 02.65ʹ
Triplets
DEV2
100
N 25º 03.02ʹ
Midchannel
DEV3
425
N 25º 03.35ʹ
Wreck
DEV4
490
N 25º 04.52ʹ
Porgy Rocks 2
DEV5
575
N 25º 03.54ʹ
Porgy Rocks 1
DEV6
660
N 25º 03.79ʹ
Exuma Cay Land and Sea Park (undeveloped)
W 77º 15.61ʹ
W 77º 16.71ʹ
W 77º 16.02ʹ
W 77º 16.55ʹ
W 77º 14.51ʹ
W 77º 14.57ʹ
2–4 m
1–3 m
1–3 m
1–4 m
3–4 m
4–6 m
50
230
650
2,600
4,700
5,200
Bell South
Big Eye
Hall’s Pond
Emerald
Malabar
Channel Rocks
W 76º 33.44ʹ
W 76º 38.15ʹ
W 76º 35.45ʹ
W 76º 37.76ʹ
W 76º 35.42ʹ
W 76º 33.36ʹ
1–3 m
1–4 m
2–5 m
1–3 m
1–3 m
3–5 m
10
40
150
450
900
1,200
UNDEV1
UNDEV2
UNDEV3
UNDEV4
UNDEV5
UNDEV6
180
297
250
99
440
550
N 24º 18.02ʹ
N 24º 23.97ʹ
N 24º 22.02ʹ
N 24º 23.04ʹ
N 24º 21.98ʹ
N 24º 17.38ʹ
ter (Model 2008, calibrated daily with 0.5 NTU calibration standard). Five hundred-ml of each
sample was then filtered through a GF/F filter, and the filter was subsequently labeled and frozen
for chlorophyll-a determination. At the University of Miami, filters were processed for chlorophyll-a content as per Szmant and Forester (1996). Filters were placed into vials with 10-ml of
20% tetrahydrofuran and 80% methanol (D’Elia et al., 1983). Samples were mixed, refrigerated
for 4 hrs, and then centrifuged. A flourometer was used to determine chlorophyll-a concentration
of the supernatant, and values were used to calculate micrograms per liter of seawater (Lorenzen,
1996). Unfiltered seawater (250 ml) was collected and frozen for laboratory analysis of water
column nutrients. Total nitrogen (TN) and total phosphorus (TP) were determined by the South
Florida Environmental Research Center at Florida International University. TN was measured by
a Nitrogen Anlyzer (Jones and Boyer, 2002) and TP was determined using a dry ashing hydrolysis technique. Data are reported in micromoles. Sediment samples were taken from the “halo”
adjacent to the patch reefs by using a PVC core. Sediment from ~25-cm depth from five random
locations around the patch reef was collected and frozen. Thawed sediment samples were mixed.
Subsamples were sieved for grain size distribution (Stoddart, 1978). Sediment (100 g) was dried,
and then ground with an agate mortar and pestle. Samples were analyzed for TN with a Carlo
Erba Model 1106 Elemental Analyzer (Szmant and Forrester, 1996).
Water quality values tend to be skewed to low concentrations, particularly in oligotrophic tropical waters. Values do not follow a normal distribution, thus a more appropriate means to evaluate
natural variability of parameters is through examining median values, as measures of central
tendency. Outliers (<5th and >95th percentiles) were excluded; thus the 5th and 95th percentiles
are the minimum and maximum values given (Christian et al., 1991). Differences in values were
tested between developed and undeveloped patch reef sites, using the Wilcoxon Ranked Sign test
(P = 0.05).
Ecological Parameters.—Species presence-absence surveys were used to inventory the
dominant and conspicuous benthos on the patch reef stations. Surveys consisted of 1-hr searches each for benthic macroalgae and stony corals based on standardized checklists from previous
surveys in the Bahamas (Sullivan and Chiappone, 1992, 1993). Macroalgae identification was reviewed using Littler and Littler (2000). The algae and coral data were analyzed using similarity
coefficients and clustering strategies to evaluate similarities in species composition among several
300
Figure 2. Map of the developed (A) and undeveloped (B) patch reef stations in the central Bahamas. Developed patch reef stations were located in Montagu Bay, New Providence. Undeveloped
patch reef stations were located in the Exuma Cays Land and Sea Park.
survey locations. The Jaccard coefficient was used in pair-wise comparisons between patch reefs
(Pielou, 1977; Hubalek, 1982). This matrix was used to construct a dendogram based on cluster
analysis of similarity values. The dendograms were used to identify patterns of species that could
be indicative of ecological change.
One-m2 quadrats were used to inventory the coral cover and incidence of lesions for reef-building
corals. Preliminary surveys were carried out to determine effective sample sizes for the coral vitality surveys on patch reefs. Pilot study results indicated that surveys of 20, 1-m2 quadrats captured
78–85% of the total colonies larger than 7 cm in maximum diameter on the small patches. Within
the quadrats, corals were identified to species and measured for diameter or length and width to
the nearest 0.5 cm. Colony size measurements were used to calculate the planar size (cm2) of each
colony, these measurements were used to calculate coral cover (with a maximum of 10,000 cm2 per
quadrat). The recording of coral lesions was a modification of Dustan (1993). Data were analyzed
by presenting vitality as a percentage of all coral colonies surveyed with lesions, defined as all
conditions except “unblemished” and “almost unblemished”.
Image Analysis of Temporal Trends.—The purpose of the temporal trends analysis was to
understand landscape scale changes in a bay adjacent to a developed island. Montagu Bay was
historically dominated by three types of hard-bottom habitats: A) patch reefs dominated by large
boulder corals, B) hard bar or hard bottom co-dominated by corals, octocorals and sponges, and
C) nearshore hard bottom or rocky platform habitats dominated by corals and sponges. Montagu
Bay and the adjacent northeastern shore of New Providence were selected as the image analysis
site because of the historical record of nearshore coral reefs in the area, the uniformity in the type
of development (all single family homes with onsite sewage disposal), and the available historical
data.
The Montagu Bay study site was 443.8 ha of seafloor and coastal zone (land). The bay was
bounded by New Providence Island to the south, Paradise (Hogg) Island to the northwest, Athol
Island to the north, and Porgy Rocks to the northeast (Fig. 2). This bay has a strong tidal circulation through Nassau Harbour to the west, and through the channel between Paradise and
Athol Island. The bay contains many reefs that have been used historically and today for tourism
(snorkeling, SCUBA diving, and glass-bottom boats). The reefs are well documented both in
aerial photographs, and early underwater photographs. The classification of the benthic habitats
was based on previous benthic surveys and mapping the central Bahamas (Sluka et al., 1999;
Chiappone et al., 2000). The classification was based on the hierarchical habitat classification in
Cowardin et al. (1979).
Two sets of aerial photography were selected: 1) 1943 photography taken by a Canadian survey
company in high-resolution black/white film at a scale of 1:30,000, and 2) 1995 photography
301
taken by a U.S. survey company also in black/white film at a scale of 1:10,000. Both photo sets
provided an excellent view of reefs and shallow marine habitats. Images were scanned, orthocorrected, and assembled in mosaics in a GIS database (ESRI ArcInfo). Photo-mosaics were
examined for both the density of development (number of houses within 500 m of the shoreline)
and mapping benthic habitats. Stereo pairs of the photographs were used to aid in photo-interpretation. Benthic habitats were identified as one of 13 classes, these classes were grouped for change
detection into three broader categories based on substrate and lifeform coverage: 1) Bare or algalcovered sand/mud bottom, 2) Seagrass habitats on sand or mud, and 3) Reefal and hard bottom
habitats. Bare or sparsely covered sand or mud bottom included nearshore mud bottom adjacent to
mangrove wetlands, mud bottom with bioturbation (shrimp) mounds, algae-dominated mud bottom, and dredged “pits”. Seagrass habitats included sparse, patchy, or dense seagrass beds. Hard
bottom communities included nearshore rocky platform, hard bar (non-reefal hard bottom), and
patch reefs. The 1995 image was ground-truthed, and the benthos classified based on verification
in field surveys. Change detection was accomplished by comparison of the two maps, then identification of the areas that changed habitat classes. Analysis included comparing the total area of
coverage of a habitat as well as the number of polygons (habitat units) that changed between the
respective years for the physical loss of reef habitats.
Results
Environmental Parameters.—Water quality data are summarized in Table 2;
temperature, salinity, dissolved oxygen, turbidity, TN, and TP are shown in pair-wise
comparisons, grouped by season and tidal cycle. Median values are presented for all
developed and all undeveloped reefs, for season and tide. These water quality parameters show no significant difference between developed and undeveloped reefs with one
exception. The only significant differences between the developed and undeveloped sites
occurred in turbidity (P = 0.05). Turbidity measurements were significantly higher off
undeveloped reefs in February. Seasonal and tidal variability of parameters was much
greater than diurnal changes. Patterns and extent of seasonal and tidal variation were
similar between developed and undeveloped patch reef stations.
Temperature on nearshore patch reefs varied as much as a 4ºC over the tidal or diurnal
cycle. Temperatures were very high in October of 1998, with long-term temperature loggers recording a maximum summer (June 1988–October 1998) temperature of 34.4ºC
on developed reefs and 34.8ºC on undeveloped reefs. Salinity was likewise high at all
the patch reefs, ranging between 36.2 and 40.9.
Dissolved oxygen levels measured at all stations were high and reflected surface mixing and tidal circulation in shallow water at both developed and undeveloped stations.
Dissolved oxygen levels were higher during the winter, when the water temperature was
also lower. Levels of dissolved oxygen were not different between sunrise and sunset.
Overall, chlorophyll-a concentrations were lower with less variation during February
(0.00–0.83 mg L−1) compared to October (0.04–4.36 mg L−1). Total nitrogen at patch
reefs ranged from 0.00–11.49 μM during February and from 1.96–16.59 μM during October. TN was also slightly higher during low tide compared to high tide. TP concentrations were very low, ranging from 0.00–0.18 μM in October, and 0.00–0.46 μM in
February. Variation in TP was greatest at low tide during February.
Table 3 presents the water quality parameters grouped for reefs near developed and
undeveloped sites, based on distance from shore during the February sampling. The
distance from shore represents a linear distance from island shoreline, but also distance
out onto the banks (away from the platform margin), thus farther “offshore” does not
302
Table 2. Water quality parameters for 12 patch reefs adjacent to developed and undeveloped coasts. Water quality sampling
was carried out over 2 wks in February 1998 and 2 wks in October 1998. Median, maximum, and minimum values
(exclusive of outliers) are given for temperature, salinity, dissolved oxygen, turbidity, chlorophyll a, total nitrogen, and
total phosphorus. (**) indicates significant differences between developed and undeveloped reefs.
Temperature Salinity Dissolved
oxygen
(oC)
(ppt)
(mg L−1)
Turbidity
Chl a
Total N
Total P
(NTU)
(µg L−1)
(µM)
(µM)
SEASONAL VARIABILITY-FEBRUARY
All developed
Median
21.6
38.3
6.4
0.16 **
0.06
4.89
0.05
High and low tide
Max
23.8
39.2
7.1
0.3
0.18
9.61
0.37
Min
20.9
36.2
5.9
0.0
0.00
0.25
n
130
130
130
70
52
60
0.00
60
All undeveloped
Median
21.15
39.1
6.4
0.8
0.22
3.50
0.06
High and low tide
Max
24
39.6
7.1
4.6
0.83
11.49
0.46
Min
19
38.5
5.9
0.0
0.01
0.00
n
120
120
120
89
57
60
0.00
60
SEASONAL VARIABILITY-OCTOBER
All developed
Median
30.25
40.5
4.9
0.1
0.01
6.93
High and low tide
Max
31.4
40.9
6.4
0.4
4.56
13.83
Min
28.2
37.8
4.1
0.0
0.04
1.96
n
100
100
100
70
58
58
All undeveloped
Median
29.9
39.6
4.7
0
0.09
6.86
High and low tide
Max
31.2
40.1
5.0
0.5
0.12
16.59
Min
29.4
39.1
3.9
0
0.04
3.24
n
90
90
90
90
60
53
0
0.17
0.00
58
0
0.17
0
53
TIDAL VARIABILITY (WITHIN SEASON)
All developed
Median
21.4
38.2
6.06
0.2
0.08
5.04
0.04
February
Max
21.5
38.6
6.14
0.3
0.17
9.19
0.37
Low
Min
21.2
38
5.88
0.1
0.01
0.78
n
30
30
All developed
Median
22.2
38.5
6.45
0.1
0.04
4.75
0.06
February
Max
23.8
39
6.57
0.3
0.18
9.61
0.33
High
Min
21.5
37.8
6.26
0.0
0
0.25
n
30
30
30
30
30
30
30
28
30
30
0
30
0
30
infer proximity to oceanic (deeper) water. There were no significant differences in water
quality parameters based on distance from shore.
Thirty sediment samples were processed from both developed and undeveloped patch
reef stations. Sediment composition did not vary significantly between developed and
undeveloped reefs; for both areas the sediment adjacent to reefs was made up of mostly
sand (0.5 mm fraction), ranging from 54–73% of sample. The smallest fraction from
both areas was rubble (4.75 mm), ranging from 8–20%. Rubble was entirely made up of
shell fragments. Nitrogen content of the sediments was similar between sites, ranging
between 20–57% dry weight in the 60 samples.
Ecological Parameters.—More macroalgae species were found on developed patch
reef stations (mean = 16.2 ± 5.6) compared to undeveloped patch reef stations (mean =
9.3 ± 1.7), but these differences were not significant (t[10,.05] = 0.088). Significantly more
stony corals species were found on developed patch reef stations (mean = 21.5 ± 1.2)
compared to undeveloped patch reef stations (mean = 14.0 ± 2.9; t[10,.05] = 9.59). There
303
Table 2. Continued.
Temperature Salinity Dissolved
oxygen
(oC)
(ppt)
(mg L−1)
Turbidity
Chl a
Total N
Total P
(NTU)
(µg L−1)
(µM)
(µM)
All undeveloped
Median
21.2
39.1
6.4
1.4
0.30
3.6
0.06
February
Max
24
39.6
7.1
3.6
0.83
9.78
0.26
Low
Min
19
38.7
6.3
0
0.11
0.88
n
40
40
All undeveloped
Median
21.7
39.2
6.3
0.4
0.12
3.31
0.04
February
Max
23.5
39.4
6.85
4.6
0.79
11.49
0.46
High
Min
19
38.9
5.91
0
0.01
n
40
40
All developed
Median
30.4
40.5
5.22
0.0
0.09
6.45
October
Max
31.4
40.8
6.36
0.1
4.56
12.32
Low
Min
30.0
37.8
4.11
0.0
0.04
1.96
n
35
35
All developed
Median
29.2
40.5
4.6
0.1
0.11
7.58
October
Max
29.8
40.9
5.3
0.4
0.13
13.83
High
Min
28.2
39.5
4.2
0.0
0.08
4.62
n
35
35
All undeveloped
Median
29.8
39.6
4.6
0.0
0.09
6.57
October
Max
30.3
40.1
5.0
0.5
0.12
10.89
Low
Min
29.5
39.1
3.9
0.0
0.04
4.65
n
45
45
All undeveloped
Median
30
39.5
4.7
0.0
0.08
7.14
October
Max
31.2
39.7
4.9
0.2
0.11
16.59
High
Min
29.4
39.3
4.6
0.0
0.04
3.24
n
45
45
40
40
35
35
45
45
40
40
35
35
45
45
29
29
29
29
30
30
30
0
30
0
0
30
30
28
30
23
30
0
0.03
0
28
0
0.17
0
30
0
0.16
0
23
0
0.17
0
30
was no significant difference in the number of sponge species recorded on undeveloped
and developed patch reefs (t[10,.05] = 0.818; Table 4).
Figure 3 illustrates the cluster analysis of presence-absence similarity values for benthic algae species, with a segregation of the developed vs. undeveloped patch reefs. Only
14 algae species were found on reefs from both sites. Macroalgae composition on the
patch reef stations is variable from patch to patch, with the highest similarity between
DEV3 and DEV4, with 52% of algae species shared.
More species of stony corals were found on the patch reefs in Montagu Bay, but the
cluster analysis did not segregate the patch reefs by developed and undeveloped stations
(Fig. 4). Patch reef stations that were clustered with macroalgal species composition
were not clustered based on stony coral species present. Percent coral cover was significantly different between developed patch reef stations (mean = 17.8% ± 4.7) compared
to undeveloped patch reef stations (mean = 24.2% ± 12.1; based on cm2 m−2 coral cover
t[10,.05] = 0.701). The percent of all coral colonies showing lesions was significantly higher
for developed patch reef stations (mean = 30.3% ± 9.8) compared to undeveloped patch
reef stations (mean = 16.2% ± 5.4; based on number of lesion colonies t[10,.05] = 0.481;
Table 4).
The species composition of macroalgae was quite different between the developed
and undeveloped reefs. Only 14 species out of 46 were found on both developed and
304
Table 3. Water quality parameters for 12 patch reefs at varying distance from islands in the central Bahamas. Stations are grouped by distance from shore. UNDEV1, UNDEV2, DEV1, and UNDEV4 were <500
m from shore, DEV2, UNDEV4, DEV3, and UNDEV5 were 650–900 m from shore. UNDEV6, DEV4,
DEV5, and DEV6 were over 1000 m from shore. Water quality sampling was carried out over 2 wks in
February 1998. Median, maximum, and minimum values (exclusive of outliers) are given for temperature,
salinity, dissolved oxygen, turbidity, chlorophyll a, total nitrogen, and total phosphorus.
Temp
Salinity
Dissolved
oxygen
Turbidity
Chl a
Total N
Total
P
(oC)
(ppt)
(mg L−1)
(NTU)
(mg L−1)
(μM)
(μM)
Distance from shore
All reefs within 500 m
Median
21.6
39.1
6.36
0.29
0.11
3.2
0.05
Max
24
39.6
6.85
3.42
0.83
10.85
0.39
Min
19
37.6
n
77
77
Median
21.4
38.9
6.38
0.23
0.06
5.19
0.06
Max
22.6
39.4
7.14
4.62
0.37
9.78
0.46
Min
19.6
36.2
81
81
Median
22.1
39
6.39
0.19
0.08
3.65
0.05
Max
23.8
39.6
6.85
3.77
0.79
11.49
0.33
Min
19
38.1
5.92
0.01
0.02
0.25
n
75
75
5.88
77
0.01
54
0.01
39
0
0
40
40
Reefs 650–900 m
n
6
0
0
81
55
36
0.88
40
0
40
Reefs over 1,000 m
75
40
34
40
0
40
undeveloped reef stations. No species were found on all 12 patch reefs, though Halimeda
tuna (Ellis and Solander) and Amphiroa fragilissima (Linnaeus) were found on all developed patch reefs (Table 5). Some alga species only found on developed reefs included
Schizothrix spp., Rhipocephalus phoenix (Ellis and Solander), Caulerpa cuppressoides
Agardh, Dictyosphaeria cavernosa (Forskaål), and Galaxaura oblongata (Ellis and Solander). Species found only on undeveloped patch reefs included Microdictyon marinum
(Bory), Turbinaria turbinata Barton, and Caulerpa ventricillata Agardh. Stony coral
species were much more widely distributed between developed and undeveloped reefs
(Table 6). Some coral species were found on all 12 patch reefs: Agaricia agaricites agaricites (Linnaeus), Porites porites porites (Pallas), Siderastrea radians (Pallas), and M.
annularis. Many species were only missing from one or two of the 12 reefs, and common
in both areas: Millepora alcicornis Linnaeus, Siderastrea siderea (Ellis and Solander),
Porites asteroides Lamarck, Diploria labyrinthiformis Linnaeus, Favia fragum (Esper),
Meandrina meandrites meandrites Linnaeus, and Montastraea cavernosa Linnaeus.
The coral vitality surveys revealed the differences between patch reefs of developed
and undeveloped islands. For M. annularis, most colonies in Montagu Bay had more
lesions per colony (Fig. 5A) and covered a larger area of the colony. Most of the lesions
were identified as algal overgrowth in Montagu Bay. In the Exumas, most of the lesions
on M. annularis were classified as damage to tissue only (predation). The most common coral to both sites was A. agaricites, a lettuce coral. This coral species showed the
greatest difference in number of lesions per colony between the two sites (Fig. 5B). The
305
Table 4. Summary of number of species, percent coral cover, and percent of all coral colonies with lesions from
12 patch reef sites in the central Bahamas. Species present were recorded from entire patch reef (ranging in size
from 98–660 m2). Coral cover and percent of all coral with lesions was evaluated from the measurement and
inspection of coral colonies from 20, 1-m2 permanent quadrats.
Site name
Station
number
# of macroalgae species
# of stony
coral species
# of
sponge
species
% coral
cover
% of all coral colonies with lesions
(n = sample size)
Reefs adjacent to coastal development
Lighthouse
DEV1
16
20
12
15.1
18.2
(n = 303)
Triplets
DEV2
23
21
14
23.4
38.0
(n = 332)
Midchannel
DEV3
22
23
18
23.1
25.8
(n = 299)
Wreck
DEV4
13
23
22
11.9
22.8
(n = 307)
Porgy Rocks 1
DEV5
15
21
12
18.6
44.4
(n = 178)
Porgy Rocks 2
DEV6
8
21
12
14.6
32.4
(n = 222)
Reefs adjacent to undeveloped islands
Bell South
UNDEV1
10
12
20
24.9
10.5
(n = 429)
Big Eye
UNDEV2
6
13
14
26.3
17.3
(n = 510)
Hall’s Pond
UNDEV3
9
16
14
22.3
12.8
(n = 450)
Emerald
UNDEV4
10
11
8
15.9
11.5
(n = 477)
Malabar
UNDEV5
11
13
14
10.1
23.7
(n = 312)
Channel Rocks
UNDEV6
10
19
24
45.5
21.1
(n = 360)
majority of the lesions recorded were bleaching of Agaricia coral on both developed and
undeveloped patch reef stations. Lesions consisting of algal overgrowth, bleaching, fresh
tissue damage, or predation and sediment damage accounted for up to 70% of the colony
area for developed patch reef corals.
Image Analysis of Temporal Trends.—Over 52 yrs the housing density along the
10.5 km shoreline of Montagu Bay increased from 1.5 houses ha−1 in 1943 to 5.7 houses
ha−1. Aerial imagery showed 327 houses in 1943 compared to 1270 houses in the same
Figure 3. Cluster analysis (Q-mode) of presence-absence similarity values for benthic algae on
six patch reefs in Montagu Bay and six patch reefs in the northern Exuma Cays, Bahamas. Values
were calculated using the Jaccard coefficient and clustered using a group-average sorting strategy.
306
Figure 4. Cluster analysis (Q-mode) of presence-absence similarity values for stony corals on six
patch reefs in Montagu Bay and six patch reefs in the northern Exuma Cays, Bahamas. Values
were calculated using the Jaccard coefficient and clustered using a group-average sorting strategy.
area in 1995. All houses have onsite wastewater disposal with the exception of one package plant and injection well at a condominium complex. One hundred percent of the
linear extent of the shoreline changed from 1943–1995; shoreline alterations came from
dredging canals, filling of coastal wetlands, and seawall construction. Between 1943–
1995, 2.9 ha of land were added on New Providence with the filling in of a mangrove
wetland, and building the shoreline out with seawalls. Spatial change analysis showed
47% of the marine benthic habitats changed from 1943–1995 (Figs. 6,7). Changes were
analyzed for major shifts in three habitat categories: 1) bare or algal-covered sand/mud
bottom, 2) seagrass habitats, and 3) reefs and hard-bottom habitats. The area of mud or
algal-covered bottom increased, while the area of seagrass and hard-bottom habitats
decreased (Table 7). The number of patch reefs in the bay decreased from 214 to 133.
The total number of seafloor polygons decreased overall within the 50-yr time span from
280 to 216 habitat units. Disturbed seagrass areas (with visible propeller scars, dumping
and anchor damage) increased from 0.39 to 15.76 ha. In 1943, propeller scars were seen
very close to docks and anchorages, but in 1995, scars impacted much of the shallow
seagrass in the western areas of the bay. Dredged pits and spoil areas were not visible in
1943, but covered 0.33 ha in 1995. The tidal channel reef visible in 1943 aerial images,
and documented with underwater photographs did not exist in 1995; much was cleared
and physically removed for a navigation channel.
Seagrass areas showed the most extensive changes, with a 24.1% loss in area of dense
seagrass habitats (Table 7). The entire bay lost coral habitat, both in the loss of patch
reefs, and in the loss of coral and sponge dominated hard-bottom habitats (locally
called “hard bar”). Along the southeastern extent of Montagu Bay, hard-bottom habitats
changed from sponge-coral-octocoral co-dominated to algal-dominated.
An analysis of habitat change over 52 yrs illustrated some important trends in the
Montagu Bay environment (Fig. 8). For each of the three categories, each polygon or
habitat unit was examined to evaluate the nature of change over time. Polygons identified
DEV2
1
1
Udotea flabellum
Valonia macrophysa
Ventricaria ventricosa
1
1
1
1
1
Rhaphocephalus phoenix
Penicillus dumetosus
Penicillus capitatus
Microdictyon marinum
Halimeda tuna
Halimeda opuntia
Halimeda goreauii
Dictyosphaeria cavernosa
Dictyosphaeria cavernosa
Derbesia sp.
Caulerpa verticillata
Caulerpa taxifolia
Caulerpa racemosa
Caulerpa cupressoides
Batophora oerstedii
Avrainvillea longicaulis
DEV3
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
Midchannel
1
1
1
1
1
1
Anadyomene stellata
1
Acetabularia crenulata
Triplets
Acetabularia calyculus
1
Species
CHLOROPHYTA
DEV1
Lighthouse
Algae
1
1
1
1
1
1
1
Wreck
DEV4
1
1
1
1
1
Porgy 1
DEV5
1
Porgy 2
DEV6
1
1
1
Bell South
UNDEV1
1
Big Eye
UNDEV2
1
1
Hall’s Pond
UNDEV3
1
1
1
Emerald
UNDEV4
1
1
1
Malabar
UNDEV5
1
1
Channel Rock
UNDEV6
Table 5. Macro-algae species checklist with species present on 12 patch reef stations in the central Bahamas. Six patch reefs were adjacent to developed coastlines, and six were adjacent to
undeveloped coasts. Surveys were carried out in 1 hr.
307
DEV3
1
1
Galaxaura oblongata
1
1
1
Coelothrix irregularis
1
1
1
Cladophoropsis macromeres
Amphiroa rigida
Amphiroa fragilissima
RHODOPHYTA
Turbinaria turbinata
Turbinaria tricostata
Sargassum pteropleuron
Sargassum polyceratium
Sargassum hystrix
1
1
1
Padina sanctae-crucis
Rosenvingea intricata
1
1
1
Lobophora variegata
1
1
Midchannel
Dictyota divaricata
1
DEV2
Triplets
Dictyota dichotoma
Dictyota cervicornis
Dictyota humifusa
1
Species
PHAEOPHYTA
DEV1
Lighthouse
Algae
Table 5. Continued.
DEV4
1
1
1
1
Wreck
DEV5
DEV6
1
1
1
1
1
1
1
1
Porgy 2
1
1
1
1
1
Porgy 1
UNDEV1
1
1
1
1
1
1
1
Bell South
1
1
1
1
Big Eye
UNDEV2
UNDEV3
1
1
1
Hall’s Pond
1
1
1
1
1
1
1
Emerald
UNDEV4
1
1
1
1
1
Malabar
UNDEV5
UNDEV6
1
1
1
1
1
Channel Rock
308
DEV2
10
UNDEV3
1
Hall’s Pond
13
15
1
8
1
6
9
10
1
1
Malabar
UNDEV5
16
23
1
22
1
1
Emerald
UNDEV4
11
1
1
1
1
1
Big Eye
UNDEV2
1
1
UNDEV1
Bell South
1
1
DEV6
Porgy 2
1
1
DEV5
Porgy 1
1
1
DEV4
Wreck
TOTAL SPECIES
1
1
1
DEV3
Midchannel
Thalassia testudinium
SEAGRASSES
Szhizothrix spp.
BLUE-GREEN ALGAE
Titanoderma prototypum
Peysonnelia sp.
Neogoniolithon spectabile
Laurencia poitei
1
1
Laurencia intricata
1
Triplets
Kallymenia limminghii
Species
Hydrolithon boergesenii
DEV1
Lighthouse
Algae
Table 5. Continued.
UNDEV6
10
1
1
1
Channel Rock
309
310
Table 6. Stony coral species checklist with species present on 12 patch reef stations in the central Bahamas. Six patch reefs were
adjacent to developed coastlines, and six were adjacent to undeveloped coasts. Surveys were carried out in 1 hr.
Species list
DEV1
DEV2
DEV3
DEV4
DEV5
DEV6
Lighthouse
Triplets
Midchannel
Wreck
Porgy 1
Porgy 2
1
1
1
1
1
1
1
1
1
STONY CORALS
1
Acropora cervicornis
Agaricia agaricites
1
Agaricia fragilis
Bartholomea annulata
1
Colpophyllia natans
1
1
1
1
Condylactis gigantea
1
1
1
Dichocoenia stokesi
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
Discosoma sanctithomae
1
1
1
1
1
Eusmilia fastigata
1
1
1
1
1
1
Favia fragum
1
1
1
1
1
1
1
1
1
1
1
Diploria clivosa
Diploria labyrinthiformis
Diploria strigosa
1
Isophyllastrea rigida
Isophyllia sinuosa
Lebrunia danae
Leptoseris cucullata
1
1
1
1
1
Madracis formosa
1
1
1
1
Madracis mirabilis
1
1
1
1
Millepora alcicornis
1
1
1
1
Millepora complanata
1
Montastraea annularis
1
1
1
1
Montastraea cavernosa
1
1
1
1
1
1
1
1
1
1
1
Manicina areolata
Meandrina meandrites
1
1
Mussa angulosa
Mycetophyllia danaana
1
Mycetophyllia ferox
Mycetophyllia lamarckiana
Porites astreoides
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
21
21
0
0
Porites branneri
Porites porites
Scolymia lacera
Siderastrea radians
1
Siderastrea siderea
Stephanocoenia intersepta
1
1
Total species
20
21
1
23
23
1
CORALLIMORPHS / AHERMATYPES / ANTIPATHARIANS
1
Palythoa caribaeorum
Ricordea florida
Total species
0
1
1
1
1
2
1
311
Table 6. Contiued.
Species list
UNDEV1
UNDEV2
UNDEV3
UNDEV4
UNDEV5
UNDEV6
Bell South
Big Eye
Hall’s pond
Emerald
Malabar
Channel Rock
1
1
1
1
1
1
STONY CORALS
Acropora cervicornis
Agaricia agaricites
Agaricia fragilis
Bartholomea annulata
1
Colpophyllia natans
Condylactis gigantea
1
Dichocoenia stokesi
1
1
1
1
1
1
1
1
1
1
Eusmilia fastigata
1
1
1
Favia fragum
1
1
1
Diploria clivosa
Diploria labyrinthiformis
Diploria strigosa
Discosoma sanctithomae
1
1
1
1
1
Isophyllastrea rigida
Isophyllia sinuosa
Lebrunia danae
Leptoseris cucullata
1
Madracis formosa
1
Madracis mirabilis
Manicina areolata
1
1
1
1
Meandrina meandrites
Millepora alcicornis
1
1
1
1
1
1
1
1
1
Millepora complanata
Montastraea annularis
1
1
1
1
1
1
Montastraea cavernosa
1
1
1
1
1
1
1
Mussa angulosa
Mycetophyllia danaana
1
Mycetophyllia ferox
Mycetophyllia lamarckiana
1
1
1
1
1
1
1
1
1
1
1
Siderastrea radians
1
1
1
1
1
1
Siderastrea siderea
1
1
1
1
1
1
Porites astreoides
Porites branneri
1
Porites porites
Scolymia lacera
Stephanocoenia intersepta
Total species
12
13
1
1
16
11
13
19
0
0
0
0
CORALLIMORPHS / AHERMATYPES / ANTIPATHARIANS
Palythoa caribaeorum
Ricordea florida
Total species
0
0
312
Figure 5. (A) Mean number of lesions per colony on Montastraea annularis on the six developed
patch reef stations and six undeveloped patch reefs in the central Bahamas (B) Mean number of
lesions per colony for Agaria agaricites on same six developed patch reef stations and six undeveloped patch reef stations in the central Bahamas.
in aerial imagery are shown in Figures 6 and 7, and these habitat units vary in size. In
addition to looking at area percent change, the change analysis of habitat units indicates
how benthic habitats have changed over time. For the bare sand/mud or algal-bottom
habitats, 57% of the polygons changed to land or dredge spoils, 9% remained within
this category, and 33% changed to seagrass habitats. For all seagrass habitats, 12% of
the polygons changed to land or dredge spoils, 12% changed to bare sand/mud or algal bottom habitats, 38% remained seagrass habitats, and 38% changed to hard-bottom
habitats. For reefal and hard-bottom habitats, 10% changed to bare mud/sand or algal
bottom habitats, 63% changed to seagrass habitats, and only 25% remained hard-bottom habitats. For patch reefs that changed to seagrass communities, these areas were
observed to be rubble-strewn seagrass beds with scattered coral heads; patch reefs were
either buried in sand, or physically broken up as navigation hazards. Most changes occurred in the western half (inner half) of the bay, with high conservation of reefs and
benthic habitats in the eastern (outer half) of the bay.
Paradise Island
Athol Island
313
Nearshore mud bottom
Mud bottom with shrimp mounds
Sparse seagrass
Dense seagrass meadow
Patchy or broken seagrass meadow
Disturbes seagrass areas
Algae mud bottoms
Dredge spoils
Nearshore rocky platform
Nearshore algae platform
Hard bar
Patch reef
Tidal channel reef
Land
New Providence
Figure 6. Map of the distribution of marine benthic habitats in Montagu Bay based on 1943 aerial
photography used in the image analysis of temporal trends.
Paradise Island
Athol Island
Nearshore mud bottom
Mud bottom with shrimp mounds
Sparse seagrass
Dense seagrass meadow
Patchy or broken seagrass meadow
Disturbes seagrass areas
Algae mud bottoms
Dredge spoils
Nearshore rocky platform
Nearshore algae platform
Hard bar
Patch reef
Tidal channel reef
Land
New Providence
Figure 7. Map of the distribution of marine benthic habitats in Montagu Bay based on 1995 aerial
photography and ground-truthing used in the image analysis of temporal trends.
314
Figure 8. Change in benthic habitats of a tropical marine bay adjacent to a heavily developed
island (Montagu Bay). 280 benthic habitat polygons from image analysis of 1943 were categorized based on changes mapped from 1995 aerial imagery. Three broad habitat categories were
used: bare and algal-covered sand/mud; seagrass habitats; or reef and hard-bottom habitats. Lines
indicate the direction and magnitude of change over the 52 yrs. Most sand/mud habitats were converted to land; most seagrass habitats were converted to hard bottom; and most reef habitats were
converted to seagrass. Dashed lines indicate changes within the same habitat category.
Discussion
The results initially looked somewhat surprising. How could patch reefs adjacent to a
densely populated island be so similar in water quality conditions and ecological character to patch reefs off undeveloped islands? Paul et al. (1995) documented the ease at
which fecal bacteria can move from cesspits in the Florida Keys to nearshore waters and
canals. Onsite sewage disposal on carbonate islands contributes nutrients, organic material and pathogens to nearshore waters (U.S. EPA 1983, 1991). Nutrient loading would
appear to be significant to developed patch reef stations based on coastal housing density. Where do nutrients go? Ecological phase shifts that are catalyzed by human activities
are complex and are manifested in abrupt changes triggered by catastrophic events such
as the impacts of Hurricane Gilbert on Jamaican reefs as described by Hughes (1984).
Although the phase shifts on reefs represent profound ecological change, it is often difficult to identify the early indicators of this shift or the contribution of causal mechanisms
(e.g., trophic shifts, algal overgrowth).
The results suggest two things. First, spatial comparisons made between patch reefs
adjacent to developed and undeveloped islands should be valid and appropriate as reference sites. The patterns of variability in water quality parameters between sites were not
significantly different, and showed strong seasonal changes. Without undeveloped shorelines, it will be difficult to assess the changes from chronic eutrophication, or to begin to
develop ecological criteria for coastal restoration and mitigation programs. The patch reefs,
though variable in coral cover and diversity, appear to be good sampling units, and provide
valuable comparisons. Marine parks have been reviewed for their importance as marine
fisheries reserves (PDT, 1990; Bahamas park, Chiappone and Sullivan Sealey, 2000), but
an even more critical role may be for marine parks as comparison sites to understand
changes in coastal ecology. Simple water quality sampling programs can help determine
the control sites that have a similar range and variability on environmental parameters.
315
In the central Bahamas, only the Exuma Cays Land and Sea Park has large islands and is
sufficiently isolated to utilize its reefs for comparisons to Nassau.
Secondly, there may be no single parameter or single taxa group that can characterize
the shifting mosaic of a coral reef, but stony corals remain critical to reef identity both
in the status of individual colonies, and the status of coral habitats (Brown and Howard,
1985; Wittenburg and Hunte, 1992). The value of historical changes to reefs and coral habitats is essential in the evaluation of current condition and trends. Montagu Bay appeared
in 1943 as an incredible coral environment, with over 200 patch reefs, and large areas of
hard-bottom coral habitats. In light of the historical loss of coral habitat, the lower coral
coverage today on developed patch reef stations becomes more significant. There are
fewer coral colonies, less coral habitat, and smaller populations of coral now in Montagu
Bay than in 1943. These changes may not be linked to nutrient enrichment from cesspits
alone, but also to a combination of sediment movement and coastal zone alterations associated with construction and development of the island. The increasing density of houses
on or near the coastline introduces two ubiquitous stresses on patch reef systems: 1)
chronic eutrophication from “soak-away” onsite sewage disposal systems, and 2) the
loss of native coastal vegetation and wetland buffers to limit sediments and nutrients
from run-off. The indirect impact of chronic nutrient enrichment on a variety of coral reef
organisms has been documented (Pastorak and Bilyard, 1985; Birkeland, 1988), but the
ecology of eutrophication is not clear. Coral mortality is likely due to a variety of causes,
not only the stimulation of algal growth (McCook et al., 2001).
The central Bahamas is unique in its shallow-water banks system, and the oceanographic
dynamics at the bank margins. Patch reefs often develop in channels adjacent to the platform margin, and experience strong tidal circulation (Sealey, 1994). Circulation around
the islands has been a key focus in fisheries recruitment studies (Ray and Stoner, 1995)
and may be a key variable in maintaining good water quality for nearshore reefs, despite
development. The absolute nutrient load combined with the coastal circulation determines the severity and extent of nutrification. Restricted embayments are more sensitive
to nutrient loading or chronic eutrophication (U.S. EPA, 1991). The effects of coastal
eutrophication have been well documented in the tropics, such as the Great Barrier Reef
(Bell, 1992; Bell and Elmetri, 1995); Hawaii (Smith et al., 1981; Laws and Redalje,
1982; Maragos et al., 1985); Barbados (Tomascik and Sander, 1985, 1987a,b; Hunte and
Wittenburg, 1992); and the Florida Keys (Tomasko and Lapointe, 1991; Lapointe et al.,
1994). Strong circulation around islands at the platform margin of the Bahamas suggests
a higher tolerance to eutrophication and less obvious indications of change.
Taken as a single assessment, the differences between developed and undeveloped
patch reef stations are not striking. Macroalgae species present were different between
reefs, but not linked to proximity to development (Fig. 3). The most similar patch reefs
(DEV1, DEV2, DEV3, and DEV4) share calcareous green algal species, and differ in
species of frondose red and brown algae (e.g., Laurencia spp. and Dictyota spp.). Undeveloped patch reef stations had fewer algal species, and especially fewer calcareous
green algae (e.g., Halimeda spp., Udotea spp., Penicillus spp.). On the developed patch
reefs more pockets of sand and sediment accumulated, appropriate for the growth of
these green calcareous species (more typically found in seagrass beds, or sparsely vegetated sand bottom communities of the central Bahamas). The three least similar patch
reefs in macroalgae species present (UNDEV3, DEV5, and DEV6) were reefs closest to
the platform margin, and were the least protected of all 12 patch reefs stations. These
same three patch reefs formed a separate cluster for stony coral species present (Fig. 4).
316
Stony coral species presence did not segregate developed from undeveloped patch reef
stations, nor did the reefs segregate by size or distance from shore (with the exception of
the three previously mentioned).
The visual differences between developed and undeveloped patch reefs are attributed
to algae coverage. Reefs adjacent to developed shorelines had a distinctly “fuzzy” appearance. Developed patch reef stations not only supported more calcareous green algae
but also species such as Schizothrix spp., C. cupressoides, D. cavernosa, B. oerstedii,
and K. limminghii (found only on developed patch reef stations). More brown algae species were found on undeveloped patch reefs, especially rooted Sargasum species. These
algal species differences are not an indication of eutrophication. Eutrophication represents the combination of nutrient enrichment, increased sedimentation, and introduction
of toxins due to human activity (Pastorak and Bilyard, 1985; Tomascik and Sander, 1985;
Rogers, 1990). A higher percentage of lesions for all coral colonies is interpreted as a
higher likelihood for loss of coral cover (Hallock and Schlager, 1986; Porter and Meier,
1992). Coral cover remains one of the most commonly evaluated ecological parameters
on reefs, yet is at best only an indirect indicator of algal overgrowth and eutrophication
(Grigg 1995; McCook et. al., 2001).
Water quality parameters were not different between developed and undeveloped
patch reef stations, suggesting that these two areas are very similar environments. Seasonal variability of patch reefs is perhaps the most troublesome. Measurements were
made in October, typically a hot wet month, and February, typically a cool dry month.
Most consistent in the seasonal trends were temperature records: the median temperatures for October and February 1998 differed by 9ºC. Rainfall, in contrast to temperature, was highly variable across the central Bahamas from location to location, and from
year to year. Rainfall for the central Bahamas overall was below average for 1998, but
there were heavy rains in February, 1998 (13.2 cm total rainfall for the February 1998
compared to 2.4 cm in 1997; Sealey, 1999). Seasonal changes in temperature and rainfall
impact surface water salinity. Salinity can be high (40.5 median salinity for October) on
the shallow banks of the Bahamas when evaporation is high. The greater variation and
slightly lower salinity values during February were the result of low temperatures (low
evaporation) and high rainfall.
Large storms such as cold fronts or hurricanes may impact the undeveloped and developed patch reef stations in a similar manner, but with a 128 km distance between
these stations, rainfall variability is likely the greatest environmental difference. In any
monitoring effort between developed and undeveloped patch reef stations, meteorological records of air temperature and rainfall would be very important components for
comparing stations.
Nutrient levels measured on nearshore patch reefs in the central Bahamas are lower than TN and TP recorded for nearshore middle and lower Florida Keys, stations in
the water quality monitoring program of the Florida Keys National Marine Sanctuary
(Jones and Boyer, 2002). Nearshore Florida Keys surface water stations have a median
TN of 13.35 μM, and a maximum value of 85.88 μM (n = 478). This is twice as high
as recorded in this study and represents quarterly sampling events. TP was also higher,
with a median value of 0.020 μM, and a maximum of 0.62 μM (n = 478). In the central
Bahamas, TN was higher in October than in February, but TP was higher in February.
These results suggest that the Bahamas should be characterized as extremely oligotrophic (LaPointe et al., 1994) and water column nutrients are therefore poor indicators of
chronic eutrophication.
317
Sediment plays an important role in the reef community by trapping organic materials
(such as detritus) and releasing nutrients. Smaller sediment particles will be able to trap
more nutrients, and possibly become anoxic. Where there had been dredging, very small
sediment particles (0.125 mm) dominated the sediment samples. This was not the case
for in situ sediment adjacent to the developed patch reef stations. Sediment movement
buried small patch reefs in the inner reaches of Montagu Bay, but existing patch reefs
did not show elevated sediment nutrients or reduced grain size.
Most of the change over 52 yrs was the loss of small patch reefs from sedimentation;
and the physical removal of material during coastal development projects. The physical restructuring of the shoreline and sea floor had a large impact on the Montagu Bay
environment. Major changes in sedimentation occurred as a result of jetty construction
and beach erosion. Analysis of temporal changes to nearshore patch reefs indicates a
major impact of sedimentation and movement of sediments associated with construction
projects (Fig. 8). Historical aerial photographs show that the major erosion and sedimentation events occurred with major construction projects in and near Montagu Bay from
1961 to the present. The physical restructuring of the shoreline was massive – 100% of
the shoreline along Montagu Bay from 1943–1995 was altered. The physical loss and
change of habitats, particularly seagrass and reef habitats, can have a profound impact
on coral populations, as well as the ecological function and fisheries production of the
bay as a whole (Eggleston, 1995; Sluka et al., 1996). Image analysis for temporal trends
in nearshore patch reefs is a critical procedure for in understanding long-term function
of individual patch reef monitoring stations.
Acknowledgments
The following people are gratefully acknowledged for their assistance: M. Chiappone, E.
Schmitt, J. Kelly, A. Lowe, K. Pronzati, and T. Hollis (field work); N. Sealey (boat support and
editing); R. Wright, and T. Benham (GIS support); A. Szmant (chlorophyll and sediment sample
analyses); R. Lightbourne, Capt. W. Munro, and P. Balfe (historical photograph and interviews),
and Department of Archives, Nassau, Bahamas. Funding for the fieldwork was provided by the
U.S. Environmental Protection Agency (X825122-01-0), Jeniam Foundation, the Nature Conservancy, and University of Miami. Permission to conduct research was granted by the Bahamas
National Trust and Department of Fisheries, Government of The Bahamas. This paper is funded
in part by a grant from the Caribbean Marine Research Center (CMRC Project #CMRC-001XNR-03-01A) National Oceanic and Atmospheric Administration (NOAA) Undersea Research
Program, US Environmental Protection Agency, and Environmental Defense. Views expressed
herein are those of the author and do not necessarily reflect the views of CMRC, or any supporting agencies.
Literature Cited
Bell, P. R. F. 1992. Eutrophication and coral reefs – some examples in the Great Barrier Reef lagoon.
Water Res. 26: 553–568.
_________ and I. Elmetri. 1995. Ecological indicators of large-scale eutrophication in the Great
Barrier Reef lagoon. Ambio 24: 208–215.
Birkeland, C. 1988. Second-order ecological effects of nutrient input into coral communities. Galaxea 7: 91–100.
Brown, B. E. and L. S. Howard. 1985. Assessing the effects of “stress” on reef corals. Adv. Mar.
Biol. 22: 1–63.
318
Cant, R.V. 1997. Developing water supply strategies for the Bahamas. Bahamas J. Sci. 4: 5–7.
Chiappone, M. and K. M. Sullivan Sealey. 2000. Marine reserve design criteria and measures of
success: lessons learned from the Exuma Cays Land and Sea Park, Bahamas. Bull. Mar. Sci.
66: 691–705.
____________, R. Sluka, and K. M. Sullivan Sealey. 2001. Groupers (Pices: Serranidae) in fished
and protected areas of the Florida Keys, Bahamas and northern Caribbean. Mar. Ecol. Prog.
Ser. 198: 261–272
Christian, R., J. Boyer, D. Stanley, and M. Rizzo. 1991. Multi-year distribution of patterns of nutrients in the Neuse River Estuary, North Carolina. Mar. Ecol. Prog. Ser. 71: 259–274,
Clarke, K. R. and R. H. Green. 1988. Statistical design and analysis for a ‘biological effects’ study.
Colin, P. L. 1995. Surface currents in Exuma Sound, Bahamas and adjacent areas with reference to
potential larval transport. Bull. Mar. Sci. 56: 48–57.
Cowardin. L., V. Carter, F. Golet, and E. LaRoe. 1979. Classification of wetlands and deepwater
habitats of the United States. U.S. Fish and Wildlife Service. FWS/OBS-79/31: 67 p.
D’Elia, C. F., S. L. Pomotor, and K. L. Webb. 1983. Nutrient uptake kinetics in freshly isolated
zooxanthellae. Mar. Biol. 75: 157–167.
Department of Statistics. 2002. Commonwealth of The Bahamas Report of the 2000 census of
population and Housing. Ministry of Economic Development, Government of the Bahamas,
Nassau. 112 p.
Dustan, P. 1993. Developing methods for assessing coral reef vitality: A tale of two scales. Pages
M8–M14 in R.N. Ginsburg, compiler. Global aspects of coral reefs — Health, hazards, and
history. University of Miami, Miami.
Dustan, P. and J. C. Halas. 1987. Changes in the reef-coral community of Carysfort Reef, Key
Largo, Florida: 1974 to 1982. Coral Reefs 6: 91–106.
Eggleston, D. B. 1995. Recruitment in Nassau grouper Epinephelus striatus: post-settlement abundance, microhabitat features, and ontogenetic habitat shifts. Mar. Ecol. Prog. Ser. 124: 9–22.
Grigg, R. W. 1995. Coral reefs in an urban embayment in Hawaii: a complex case history controlled by natural and anthropogenic stress. Coral Reefs 14: 253–266.
__________ and S. J. Dollar. 1990. Natural and anthropogenic disturbance on coral reefs. Pages
439–452 in Y. Achituv and Z. Dubinsky, eds. Ecosystems of the world, Volume 25: Coral reefs.
Elsevier, New York.
Hallock, P., F. Muller-Karger, and J. C. Halas. 1993. Coral reef decline. Nat. Geogr. Res. Expl. 9:
358–378.
_________ and W. Schlager. 1986. Nutrient excess and the demise of coral reefs and carbonate
platforms. Palaios 1: 389–398.
Hoffmeister, J. E. 1974. Land from the sea. Univeristy of Miami Press, Miami. 140 p.
Hubalek, Z. 1982. Coefficients of association and similarity based on binary (presence/absence)
data: an evaluation. Biol. Rev. 57: 669–689.
Hughes, T. P. 1984. Population dynamics based on individual size rather than age: A general model
with a reef coral example. Amer. Nat. 123: 778–795.
Hunte, W. and M. Wittenburg. 1992. Effects of eutrophication and sedimentation on juvenile corals. II. Settlement. Mar. Biol. 114: 625–631.
Jackson, J., M. Kirby, W, Berger, K. Borndal, L. Botsford, B. Norque, R. Bradbury, R. Cooke, J.
Erlander, J. Ester, T. Hughes, S. Kidwell, C. Lange, H. Lenihan, J. Pandolfi, C. Peterson, R.
Steneck, M. Tegner, and R. Warner. 2001. Historical overfishing and the recent collapse of
coastal ecosystems. Science 293: 629–637.
Jones, J. A. 1977. Morphology and development of southeastern Florida patch reefs. Proc. Third
Int. Coral Reef Symp., Miami, 2: 231–235.
Jones, R. and J. N. Boyer. 2002. FY2001 Annual Report of the water quality monitoring project for
the Water Quality Protection Program in the Florida Keys National Marine Sanctuary. Southeast Environmental Research Center, Florida International University, Miami. Tech. Rept.
T181: 121 p.
319
Lang, J. C., R. I. Wicklund, and R. F. Dill. 1988. Depth- and habitat- related bleaching of zooxanthellate reef organisms near Lee Stocking Island, Exuma Cays, Bahamas. Proc. Sixth Int. Coral Reef
Symp., Townsville, 2: 269–274.
____________ and W. R. Matzie. 1992. The impacts of stormwater discharges on eutrophication in
the Florida Keys. International Marine Research, Inc., Big Pine Key. 47 p.
____________, M. M. Littler, and D. S. Littler. 1992. Modification of benthic community structure
by natural eutrophication: the Belize Barrier Reef. Proc. Seventh Intl. Coral Reef Symp., Guam,
1: 323–334.
_____________, D. A. Tomasko, and W. R. Matzie. 1994. Eutrophication and trophic state classification of seagrass communities in the Florida Keys. Bull. Mar. Sci. 54: 696–717.
Laws, E. A. and D. G. Redalje. 1982. Sewage diversion effects on the water column of a subtropical
estuary. Mar. Envir. Res. 6: 265–279.
Lipcius, R. N., W. Stockhousen, D. B. Eggleston, L. S. Marshall, Jr., and B. Hickey 1997. Hydrodynamic decoupling of recruitment, habitat quality and adult abundance in the Caribbean spiny
lobster: source-sink dynamics? Aust. Mar. Freshw. Res. 48: 807–815.
Littler, D. S. and M. M. Littler. 2000. Caribbean marine plants. Offshore Graphics, Inc, Washington, D.C. 526 p.
Lorenzen, C. J. 1996. A method for continuous measurement of in vivo chlorophyll concentration.
Deep-Sea Res. 13: 223–227.
Lott, C., R. Dye, and K. M. Sullivan. 1996. Historical overview of the development and natural history of the Florida Keys. Volume 3: Site characterization for the Florida Keys National Marine
Sanctuary and Environs. The Preserver, Farley Court of Publishers, Zenda. 49 p.
Maragos, J. E., C. Evans, and P. Holthus. 1985. Reef corals in Kaneohe Bay six years before and
after termination of sewage discharges (Oahu, Hawaiian archipelago). Proc. Fifth Int. Coral
Reef Congr., Tahiti, 4: 189–194.
Marszalek, D. 1987. Sewage and eutrophication. Pages 77–90 in B. Salvat, ed. Human impacts on
coral reefs: Facts and recommendations. Antenne Museum, Moorea, French Polynesia.
McCook, L. J., J. Jompa, and G. Diaz-Pulida. 2001. Competition between corals and algae on coral
reefs: a review of mechanisms and evidence. Coral Reefs: 400–417.
Mee, L. D. 1988. A definition of “critical eutrophication” in the marine environment. Rev. Biol.
Trop. 36: 159–161.
Pastorak, R. A. and G. R. Bilyard. 1985. Effects of sewage pollution on coral-reef communities.
Paul, J. H., J. B. Rose, S. Jiang, C. Kellogg, and E. A. Shinn. 1995. Occurrence of fecal indicator bacteria in surface waters and the subsurface aquifer in Key Largo, Florida. Appl. Envir.
Microbiol. 61: 2235–2241.
Pielou, E. C. 1977. Mathematical ecology. Wiley & Sons, New York. 385 p.
Plan Development Team (PDT). 1990. The potential of marine fisheries reserves for reef fish management in the US Southern Atlantic. NOAA Tech. Mem. NMFS-SEFC-261. 40 p.
Porter, J. W. and O. W. Meier. 1992. Quantification of loss and change in Floridian reef coral populations. Amer. Zool. 32: 625–640.
Ray, M. and A. W. Stoner. 1995. Growth, survivorship, and habitat choice in a newly settled seagrass gastropod, Strombus gigas. Mar. Ecol. Prog, Ser. 123: 83–94.
Rogers, C. S. 1990. Responses of coral reefs and reef organisms to sedimentation. Mar. Ecol. Prog.
Ser. 62: 185–202.
Sealey, N. E. 1994. Bahamian landscapes: An introduction to the geography of the Bahamas (2nd
Edition). Media Publishing, Nassau. 126 p.
__________. 1999. Rainfall distribution on New Providence for 1997 and 1998. Bahamas J. Sci.
6: 55–58.
Sluka, R., K. M. Sullivan, and M. Chiappone. 1994. Comparison of juvenile grouper populations
in southern Florida and the central Bahamas. Bull. Mar. Sci. 54: 871–880.
320
________, M. Chiappone, K. M. Sullivan, and R. Wright. 1996. Habitat and life in the Exuma
Cays, The Bahamas: The Status of Groupers and Coral Reefs in the Northern Cays. Media
Publishing Ltd., Nassau. 83 p.
________, ____________, ____________, E. Schmitt, T. Potts, and J. Levy. 1997. Density, species and size distribution of groupers (Serranidae) in three habitats at Elbow Reef, Florida
Keys. Bull. Mar. Sci. 62: 219–228.
________, M. Chiappone, K. M. Sullivan, and M. de Garine. 1999. Habitat characterization and
space utilization by epinepheline groupers in the central Bahamas. Proc. Gulf Caribbean Fish.
Inst. 45: 23–36
Smith, S. V., W. J. Kimmerer, E. A. Laws, R. E. Brock, and T. W. Walsh. 1981. Kaneohe Bay sewage diversion experiment: perspectives on ecosystem responses to nutrient perturbation. Pac.
Sci. 35: 239–396.
Stoddart, D. 1978, Mechanical analysis of reef sediments. Pages 55–63 in D.R. Stoddart and R.E.
Johannes, eds. Coral reefs: research methods. UNESCO, Page Brothers. 147 p.
Stoner, A. W., P. A. Pitts, and R. A. Armstrong. 1994. Interaction of physical and biological factors
in the large-scale distribution of juvenile queen conch in seagrass meadows. Bull. Mar. Sci.
58: 14–22.
Sullivan, K. M. and M. Chiappone. 1992. A comparison of belt quadrat and species presence/absence sampling of stony coral (Scleractinia and Milleporina) and sponges for evaluating species patterning on patch reefs in the central Bahamas. Bull. Mar. Sci. 50: 464–488.
_________ and ____________. 1993. Hierarchical methodologies and sampling design for conservation monitoring of tropical marine hard-bottom communities. Aquat. Conserv.: Mar.
Freshw. Ecos. 3: 169–187.
Szmant, A. M. and A. Forrester. 1996. Water column and sediment nitrogen and phosphorus distribution patterns in the Florida Keys, USA. Coral Reefs 15: 21–41.
Tomascik, T. and F. Sander. 1985. Effects of eutrophication on reef building corals. I. Growth rate of
the reef-building coral Montastrea annularis. Mar. Biol. 87: 143–155.
__________ and ________. 1987a. Effects of eutrophication on reef-building corals III. Reproduction of the reef building coral Porites porites. Mar. Biol. 94: 77–94.
__________ and ________. 1987b. Effects of eutrophication on reef-building corals II. Structure of
scleractinian coral communities on fringing reefs, Barbados West Indies. Mar. Biol. 94: 53–75.
Tomasko, D. A. and B. E. Lapointe. 1991. Productivity and biomass of Thalassia testudinum as
related to water column nutrient availability and epiphyte levels: field observations and experimental studies. Mar. Ecol. Prog. Ser. 75: 9–17.
United National Environmental Programmes (UNEP). 2002. Global environmental outlook -3.
UNEP. 435 p.
U.S. Environmental Protection Agency (USEPA). 1983. Ecological impacts of sewage discharges
on coral reef communities. Contract Number 68-01-5906, Office of Water Program Operations,
Washington, D.C. 86 p.
________________________________________. 1991. Water quality protection program for the
Florida Keys National Marine Sanctuary. Phase I Report. Office of Wetlands, Oceans, and Watersheds, Contract No. 68-C8-0105, Work Assignment 3-225. Continental Shelf Associates, Jupiter.
Wittenburg, M. and W. Hunte. 1992. Effects of eutrophication and sedimentation on juvenile corals. I. Abundance, mortality and community structure. Mar. Biol. 112: 131–138.
Address: Department of Biology, University of Miami, PO Box 249118, Coral Gables, Florida
33124.
LINKING HABITAT PROTECTION AND MARINE
PROTECTED AREA PROGRAMS TO CONSERVE CORAL
REEFS AND ASSOCIATED BACK REEF HABITATS
Michelle A. Duval, Douglas N. Rader, and Kenyon C. Lindeman
ABSTRACT
A variety of mechanisms exist at the state, territorial, and federal level in the U.S.
to protect habitats critical in the development of reef fish species. The most recently
formulated are the Essential Fish Habitat designations in federal waters, in addition to
various National Environmental Policy Act-associated and state-level permitting processes, as well as provisions of the Coastal Zone Management Act and National Marine
Sanctuaries Act. Similarly, several mechanisms are available to implement marine protected areas, including existing federal National Marine Sanctuary processes and varied
Fishery Management Council initiatives under development. Linking habitat management and MPA implementation is critical to developing whole-ecosystem protection to
threatened habitats and populations. Spatially explicit science remains key to coordinating such efforts as diverse stressors occur across the shelf from land- and water-based
sources. However, fragmentation of jurisdictional authority significantly impacts the
ability to institute effective protection. Solutions to this must include: (1) Filling critical
personnel shortages at the field staff level, (2) development of integrated regulations for
agencies with jurisdiction over marine fisheries, water quality and coastal development,
and (3) dedicated money for monitoring and enforcement efforts as a prerequisite to
implementation of management regimes.
This publication is part of a series of papers resulting from a scientific workshop held
Coral reefs can be considered “charismatic macrofauna” of the marine realm, along with
marine mammals and sea turtles often known as charismatic megafauna. There is a high
level of public awareness regarding coral reefs and the environmental impacts to which
they are susceptible. However, most of the popular media has focused public attention
on forereef-associated habitats. Many back reef habitat types – seagrasses, sandy bottoms, hard bottoms, patch reefs, mangroves, channels – do not receive the same level of
consideration, although the threats to these areas are also pressing. Back reef habitats
provide many important ecosystem functions: serving as sources of primary production,
forage areas, critical juvenile nursery habitats, as well as larval settlement areas (Boesch
and Turner, 1984; Thayer et al., 1987; Keener et al., 1988; Nelsen et al., 1991; Able and
Fahay, 1998). These habitats and their functions are threatened by a suite of stressors that
are cross-shelf in nature and include: direct loss from coastal development, non-point
source pollution (e.g., increased stormwater runoff from impervious surfaces), alteration
of habitat and hydrology from channel dredging, prop scarring in shallow waters, damage from bottom-tending fishing gears (Fonseca et al., 1984; Van Dolah et al., 1987), as
well as increased turbidity from all of the above activities. Moreover, the effects of many
of these stressors are cumulative in nature and difficult to measure on a case-by-case
basis. Back reef systems themselves are also cross-shelf in nature, and are a fundamental component of a continuum between inshore and outer shelf regions. The utilization
patterns of many reef fish species reflect the use of multiple habitats that are distributed
adjacent to each other (Parrish, 1989; Parker and Mays, 1998).
321
322
We examine the potential for linking administrative mechanisms and available scientific information to protect back reef habitats with recent efforts to establish marine protected areas. Both habitat management and protected area design can be aided by similar
data, e.g., identification of common developmental and habitat utilization patterns for
key species groups (Lindeman et al., 2000). In addition, we attempt to address obstacles
to integrated management actions, such as fragmentation of jurisdictional authority, by
providing a state-level example of a coordinated, ecosystem approach to protection of
marine resources.
HABITAT PROTECTION
A variety of administrative tools exists at federal, regional, state, and local levels
which can be employed to provide real protection for reef-associated habitats as well as
raise awareness of the need for such protection. Traditionally, most avenues for administrative oversight of marine habitat impacts have been provided through the National
Environmental Policy Act (NEPA), the Clean Water Act (CWA, particularly through
Section 404), and the Coastal Zone Management Act (CZMA). Through these and additional laws (e.g., the Fish and Wildlife Coordination Act), the National Marine Fisheries
Service (NMFS) and U.S. Fish and Wildlife Service (FWS) have had limited comment
authority on federal projects that affect coastal habitats. Authority is limited because ultimately most permitting decisions rest with the U.S. Army Corps of Engineers (ACOE),
or with federally approved state CZMA implementing agencies. As ACOE permit denials are extremely rare, the quantity and quality of a number of coastal habitats is increasingly degraded (Odum, 1982; Lindeman, 1997).
ESSENTIAL FISH HABITAT (EFH).—The most recent opportunity for targeted habitat
protection is the Essential Fish Habitat (EFH) provision of the federal 1996 Sustainable Fisheries Act, which reauthorized and amended the Magnuson Act (NOAA, 1996).
The mandate to identify and designate EFH recognized and formalized the need to link
coastal land management with fishery management. Although councils were given regulatory authority to stop or minimize impacts to EFH from fishing activities, they were
not given authority to minimize impacts to EFH from land-based activities. However,
consultation opportunities are now encouraged.
In addition, a subset of high-value EFH can be designated as Habitat Areas of Particular Concern (HAPC), designed to protect rare, ecologically important areas that are
sensitive to environmental degradation and potentially in danger from the effects of expanded development (NOAA, 1996; SAFMC 1998a,b). Although the HAPC designation
is supposed to provide an additional measure of protection, regulatory mechanisms beyond those of EFH do not exist for HAPCs. At a minimum, certain large-scale activities
(e.g., offshore oil/gas drilling, sand mining) should be prohibited within all designated
HAPCs, either by statute or regulation.
Despite completed and ongoing research to determine the impact of specific gear types
on habitat (NEFSC, 2002), the problem of habitat alteration as a result of cumulative and
long-term gear impacts (i.e., a shifting baseline condition) is difficult to assess (SAFMC,
1998a) and can be less of an impact in most back reef areas than on deeper fore reefs. No
requirement exists for comprehensive protection, or for the development of management
plans to achieve that protection. Long term studies on common anthropogenic habitat
stressors involving both fishing gear impacts and land-based impacts are badly needed.
DUVAL ET AL.: LINKING MANAGEMENT OF HABITATS AND MPAS
323
When possible these studies need to control for confounding variables and track effects
of both direct and indirect perturbations.
The expanded consultation requirements for federal agencies either conducting or
permitting projects that may negatively impact EFH theoretically provide another layer
of defense against habitat destruction. Although the fishery management councils can
comment on non-fishing based federal projects and permits, they have only indirect authority with which to pursue action or mitigation. However, as of June 1999, the NMFS
Habitat Conservation Division Southeast Region commented on 3916 projects that were
believed to have the potential to adversely impact EFH. Of those, 262 projects resulted in
an expanded consultation process between fishery managers and other agency staff. Of
the completed expanded consultations, projects have either been denied or modified to
protect EFH in ~220 of these cases (A. Mager, NMFS, pers. comm.). Therefore, 5.6% of
the 3916 projects showed measurable responses to the EFH issue. It is important to note
that the impacts of EFH may be higher in less measurable up-front project modifications,
which are designed to avoid expanded EFH consultations.
NATIONAL MARINE SANCTUARIES ACT.—Another tool for habitat protection is the Marine
Protection, Research and Sanctuaries Act (i.e., the National Marine Sanctuaries Act) of
1972 (16 U.S.C § 1431 et. seq.). The Act provides the authority to set aside protection
for marine areas, which possess certain ecological, historical, aesthetic, research, and/or
educational qualities of national or international significance and for which current state
or federal laws do not provide sufficient protection. Although the Act does not contain
specific habitat mandates, it does encompass the concepts of habitat protection and restoration within its purposes and policies (16 U.S.C. § 1431(b)). In addition, theoretically
prohibited activities within any sanctuary include those resulting in the destruction, loss,
or injury of sanctuary resources and are subject to federal enforcement and fines. Similar to the EFH consultation under the Sustainable Fisheries Act, federal agencies are
required to consult with the Secretary of Commerce if their actions are likely to affect
sanctuary resources. Finally, each sanctuary develops regulations for allowed activities
in coordination with other federal, state, and local authorities; the regulations are specific to each sanctuary and can vary greatly in terms of the level of protection afforded.
OTHER REGULATORY TOOLS.—Additional tools available to enhance habitat protection
include the federal Coastal Zone Management Act of 1972 (CZMA, 16 U.S.C.§ 1451
et. seq.), interstate compacts (e.g., the Atlantic States Marine Fisheries Commission),
and state waters special use designations. The CZMA does not establish federal regulations or even a management framework for coastal areas, but rather encourages states to
develop their own programs through the use of federal grant monies. However, in order
to receive federal approval a stateʼs program must meet several criteria, one of which is
inclusion of measures to protect natural resources in the coastal zone, e.g., fish and wildlife and their habitats. A very important component of CZMA is the requirement that
all federal activities, which directly affect lands, waters, and natural resources within a
stateʼs coastal zone be “consistent” with its management program. This is analogous to
consultation under the Sustainable Fisheries Act, and a state may object to the consistency certification or analysis provided by a federal agency. In many cases, the state may
choose to suggest measures that would modify the project in a manner that is consistent
with the stateʼs coastal management program.
Interstate compacts such as the Atlantic States Marine Fisheries Commission
(ASMFC), which is responsible for the development of interstate fishery management
plans, have also developed habitat protection policies. While the policies do not require
324
member states to adopt regulations, they are public documents and are developed and
endorsed by the member states. For example, the ASMFCʼs Submerged Aquatic Vegetation (SAV) policy (ASMFC, 1997) recognizes the importance of SAV as habitat, and has
generated further study on fishing gear impacts to SAV. Again, although the ASMFC
cannot require the adoption of regulations regarding habitat protection, the awareness
that such policies generate among both managers and decision-makers is extremely beneficial to future protection efforts.
Regulations and policies adopted by individual states are also tools for habitat protection. Significant reef-associated habitat resources in the southeastern U.S. are located in
state as well as federal waters. Special designations (e.g., primary or secondary nursery
areas, outstanding resource waters, etc.) often restrict the types of activities or development, which can occur in these regions. Arguably, impacts to these habitats still occur
despite such restrictions.
MARINE PROTECTED AREAS
The efforts to protect marine habitats utilizing the tools outlined above have been
paralleled by a growing movement toward the use of marine protected areas (MPAs)
to aid in the restoration of depleted fish populations. The term “MPA” is defined by the
World Conservation Union (IUCN) as “any area of the intertidal or subtidal terrain,
together with its overlying water and associated flora, fauna, historical and cultural features, which has been reserved by law or other effective means to protect part of all of
the enclosed environment” (IUCN, 1988). There are clearly many levels of protection
encompassed by this definition, from “no-take” areas, which allow no consumptive or
potentially harmful uses, to “zoned areas” which regulate many different activities at
various levels.
MPAS AND FISHERY MANAGEMENT.—The use of MPAs as a fishery management tool has
generated much controversy, particularly with regard to the use of no-take MPAs. The
type of MPA used for fishery management purposes depends on the management goal,
as well as site characteristics; however, there are benefits to the use of no-take MPAs,
which should be carefully considered. First, enforcement personnel have indicated that
no-take MPAs are easier to enforce than MPAs with restrictions on the use of certain
types of gear or harvest of particular species. As a corollary, this scenario is likely to be
easier for the fishing community as well, rather than having to keep abreast of multiple
regulations for a given area under protection. Second, use of no-take MPAs can reduce
user conflicts regarding access to the resource (i.e., nobody is allowed access). Third,
no-take MPAs have the benefit of protecting all the habitat characteristics of a given
area, and afford some level of protection to not just one or two species, but all that may
use it. Finally, use of no-take MPAs in conjunction with traditional fishery management
measures could prove to be less costly both economically and ecologically, rather than
use of traditional measures alone. Size, bag, and trip limits often have direct economic
impacts on the fishing community that could be eased by incorporation of MPAs into a
management program.
Interest in utilizing MPAs as a fishery management tool is particularly strong in the
southeastern U.S. and Caribbean, where many commercially and recreationally important species use reef or back reef type habitats during development (Sedberry and Van
Dolah, 1984; Parker and Mays, 1998; Ley et al., 1999). In particular, the South Atlantic
Fishery Management Council has considered the use of protected areas for some time
325
(Table 1) and the initial scoping document (Plan Development Team, 1990) remains a
fundamental guide to MPA rationale. Most of the Councilʼs managed species are slow
growing, late to mature, and have a reproductive capacity that increases with age. Several are sequential hermaphrodites, spending the first portion of their lives as females before becoming males in response to social cues. Many aggregate at known sites to spawn
and are extremely vulnerable to fishing pressure during that time. Reef fish also exist
in mixed associations within several habitat types (Rooker and Dennis, 1991). These
characteristics can make successful implementation of traditional fishery management
measures for a single species difficult (Plan Development Team, 1990).
All three of the federal fishery management councils with reef fish resources (South
Atlantic, Gulf of Mexico, and Caribbean) manage them as multispecies complexes under
single management plans: the South Atlantic Snapper Grouper plan (SAFMC 1983a,c),
and separate Reef Fish plans for the Gulf (GMFMC, 1984) and Caribbean (CMFC,
1985). Designing management measures, which will have the desired effect upon a single species, could in some instances result in unintended consequences on other species
within the complex. Both scientists and managers have suggested that the use of MPAs,
in conjunction with traditional management measures, may be necessary for the restoration and future management of several reef fish species (Plan Development Team, 1990;
Allison et al., 1998; Johnson et al., 1999).
Establishment of MPAs through the authority of the federal fishery management council is a mechanism that is still under development. The authority is contained within
the Magnuson-Stevens Act as a discretionary provision under the development of fishery management plans. The legal justification for such action is provided by the Actʼs
national standards as well as the requirement to minimize fishing-related impacts to
EFH. Admittedly, there is a limit to the actions that the councils can take, as they only
have the direct authority to manage fishing. The South Atlantic Fishery Management
Council first formally explored the opportunity to utilize MPAs as a management tool in
1990 (Table 1), at the recommendation of the Snapper Grouper Plan Development Team
(PDT). The Councilʼs effort generated resistance within the fishing community and was
sidelined for several years. However, the Council had previously established the Oculina
Bank restricted use zone in 1983 through the Coral Fishery Management Plan; this area
can be considered the first MPA established under the Councilʼs jurisdiction (Table 1).
In 1998, the SAMFC decided to reexamine the issue of MPAs as a management tool
for its reef fish (Snapper Grouper) complex and created an MPA Committee and Advisory Panel. The Advisory Panel consists of representatives from both commercial and
recreational fishing interests, conservation organizations, the diving community, and education and outreach specialists, and was tasked with (1) providing an opinion statement
to the Council regarding the utility of MPAs in the South Atlantic, and (2) developing a
set of criteria with which to evaluate any potential sites. The Council conducted several
rounds of public scoping with regard to the two issues described above (Table 1). As a
result of the scoping meetings, the Council decided to move forward with identifying a
list of potential sites which would protect members of the deep water grouper complex
(warsaw grouper, snowy grouper, speckled hind, yellowedge grouper, misty grouper, all
tilefish species) and is currently pursuing an MPA design that would allow some pelagic
trolling while prohibiting bottom fishing.
EXECUTIVE ORDER ON MPAS.—Other processes exist which have brought attention to
the utility of marine protected areas, namely the Executive Order on Marine Protected
Areas issued by President Clinton in May 2000, and the National Marine Sanctuaries
326
Table 1. Historical overview of consideration of MPA use by South Atlantic Fishery Management
Council (SAFMC). FMP = Fishery Management Plan; HAPC = Habitat Area of Particular
Concern; PDT = Plan Development Team.
Year
1983
1990
1992
1994
1995
1997
1998
1999
2000
2001
2002
Council action
Oculina Banks protected area under Coral FMP; further restrictions implemented as
listed in remainder of table.
Snapper Grouper PDT recommends MPAs as management tool.
SAFMC conducts public scoping regarding MPAs; concept receives little support from
fishing public.
Oculina Banks HAPC becomes Experimental Oculina Research Reserve; closed to
bottom fishing and anchoring for 10 yrs.
Scientific Review Panel concludes properly designed MPAs can be effective tool in
conjunction w/other management measures.
SAFMC faces possibility of severe reductions in allowable catch after passing of SFA;
concept of MPAs resurfaces.
SAFMC engages in scientific fact-finding process using MPA Advisory Panel.
MPA Advisory Panel unanimously recommends use of MPAs as management tool for
snapper grouper stocks.
SAFMC solicits stakeholder input through informal presentations and formal scoping;
Council votes to move forward with use of MPAs.
Council conducts scoping to solicit site recommendations from public regarding
potential MPAs.
Council conducts another round of public scoping regarding list of potential sites;
conducts outreach workshops to communicate science to fishing community.
Act. The former is really an informational umbrella; it does not authorize any federal
agency to create MPAs. Rather, it directs agencies within the Interior and Commerce
departments to develop a national “framework” for MPAs with the input of an appointed
national advisory committee. It establishes a national MPA research center, as well as
an administrative headquarters that is tasked with completing a national inventory of
MPAs and developing a website for dissemination of information to the public. The most
important component of the Executive Order is the definition it establishes for “marine
protected area”, which is “any area of the marine environment that has been reserved by
Federal, State, territorial, tribal, or local laws or regulations to provide lasting protection for part of all of the natural and cultural resources therein.” This broad definition is
very similar to that of the IUCN and includes national wildlife refuges, national marine
sanctuaries, as well as other areas where consumptive uses are allowed. Although the
Executive Order does not provide a mechanism to establish marine protected areas, it
does raise the awareness of the public, managers, and decision-makers of the need to
protect certain areas of the marine environment and to establish a coordinated plan for
doing so.
NATIONAL MARINE SANCTUARIES ACT.—The National Marine Sanctuaries Act provides
a mechanism to designate geographic areas to acknowledge their national significance,
provided they meet several biological and/or cultural criteria. As stated previously, the
regulations pursuant to the act do not restrict activities within sanctuaries as a matter of
course. Any restrictions must be developed in cooperation with federal, state, and local
agencies that may hold jurisdiction in the area. The successful designation of no-take
MPAs within the Dry Tortugas followed just such a multi-stakeholder process; although
lengthy, all input was carefully considered and the outcome endorsed by all participants.
The key is that sanctuary designation can be the first step toward a closer examination
327
of marine resources, threats to those resources, and the best means to protect and sustain them for the future. Management plan revisions under the National Marine Sanctuaries Act are currently underway at the Channel Islands, Grayʼs Reef, and Stellwagen
Banks National Marine Sanctuaries; several of the options under consideration involve
measures, which would restrict activities in order to enhance protection of sanctuary
resources.
FORGING LINKS: THE NEED AND THE PROBLEM
THE PROBLEMS.—Forming linkages between habitat protection and MPA initiatives to
protect the biology and ecology of reef species and habitats is necessary to the success of
both efforts. However, it is made complicated by fragmentation of jurisdictional authority. At both the national and state level, different agencies carry different yet overlapping mandates, which affect the management of marine resources (Fig. 1). The chemical
characteristics of water – the medium in which organisms live and breathe – are rarely
managed within the same agency as living marine resources. This is also true of coastal
development – which impacts both water quality and bottom habitat. At the federal level,
NMFS is responsible for the review and implementation of the fishery management plans
that the councils are charged with developing. Yet, both the South Atlantic and Gulf of
Mexico Fishery Management Councils, and to a lesser extent the Caribbean Fishery
Management Council, have jurisdiction over the same species—which can sometimes
result in conflicting management schemes. The Environmental Protection Agency (EPA)
administers the Clean Water Act, much of which is delegated to the states for execution.
Implementation of Section 404 of the Clean Water Act, which regulates the disposal of
dredged material in waters of the U.S., is administered by the Army Corps of Engineers,
which permits a wide array of projects affecting reef systems, often in very problematic
manners (Lindeman, 1997; Peterson et al., 2000). Finally, coastal development has no
associated federal regulations under the CZMA and is instead regulated at the state and
local level; the result is varying sets of rules, despite the guidelines that CZMA provides
in order for states to receive federal approval, and therefore federal money, for their
coastal management programs.
Figure 1. Differing agencies with responsibility for the management of fishery resources, water
quality, and coastal development often have separate administrative mandates, but effective ecosystem management requires active involvement of all three administrative categories.
328
Not surprisingly, this separation of jurisdiction often continues at the state government level. In North Carolina, for example, the Division of Marine Fisheries, Division of
Coastal Management, and Division of Water Quality are all housed within the Department of Environment and Natural Resources, yet operate under different state mandates.
This does not imply that no communication takes place between the divisions; interagency permit review does occur. However, projects are often not elevated before a sister
agency prior to permits being issued; a more coordinated approach to development and
habitat protection within the stateʼs coastal region is necessary.
THE NEEDS.—Given the practical limitations on both staff and budgetary resources at
all levels of government, it is difficult to consider the indirect or unintended impacts of
management decisions on resources not directly within an agencyʼs purview. Proactive,
interagency coordination would streamline the permit review process while in all likelihood providing increased protection to marine habitats. In order for this to succeed,
a comprehensive and upfront, i.e., “pre-permitting” evaluation of cumulative impacts
– before they are allowed to occur – is necessary. The current approach to environmental impact analysis is one in which impacts are assessed ad hoc on a project-by-project
basis, and only within the geographic boundaries of the project. Rarely do they consider
the cumulative effects of repeated habitat modification (Odum, 1982; Vestal and Rieser,
1995), as in large beach dredging projects that routinely bury EFH-HAPC (Lindeman
and Snyder, 1999). Unfortunately, this results in the permitting of many small impacts,
which appear benign on an individual basis, but collectively constitute serious and permanent impact. Furthermore, this type of analysis allows for the piecemealing of large
projects that might otherwise not receive permit approval without significant mitigation.
Advance analysis of an allowable level of impact to marine resources would greatly
inform the permitting process and subsequent mitigation efforts, as well as improve
interagency communication. Unfortunately, habitat databases are often limited in scope,
and subject to repeated review for permitting considerations. Both this and the “caseby-case” method of impact review promote an institutional policy whereby inadequate
information regarding cumulative impacts is equated with no impact.
IMPROVING THE SYSTEM
As Congress considers reauthorization of the Magnuson-Stevens Fishery Conservation and Management Act (MSFCMA) and the U.S. Coral Reef Task Force continues
its work, there is both a need and an opportunity for improving the current matrix of
marine habitat protection mechanisms. At the federal level, this includes promoting an
ecosystem level of management, rather than traditional species-by-species management.
Although difficult to define, this approach recognizes the importance of species interactions, the value of a species to the ecosystem as a whole, and the need to conserve
biodiversity. Only recently have the federal fishery management councils taken small,
yet painful, steps toward the planning and implementation of such an approach. The
extensive research undertaken by many of the councils with regard to designation of
EFH, while incomplete, provides an ideal foundation for pursuing an ecosystem-based
approach.
However, for the above approach to be successful, several changes in the current legislative and regulatory structure are necessary: (1) Modify National Standard One of the
MSFCMA to reflect that resource conservation and achievement of optimum yield may
be somewhat conflicting mandates that can only be mediated appropriately through the
329
use of ecosystem-based approaches; (2) Enhance the use and development of multispecies
models; although several councils have multispecies management plans, the impact of
conservation and management measures are still typically assessed using single-species
models; (3) Expand and standardize the national assessment and monitoring program
to be comprehensive in its coverage, including both exploited and unexploited species.
Components (habitat mapping, life history characterizations, ecosystem response) of
such a program exist in every region of the country, yet are limited in geographic scope.
There is ample opportunity to partner with research universities, as well as state-level
natural resource agencies. An example of this is the Atlantic Coastal Cooperative Statistics Program (ACCSP), which is a joint effort among the ASMFC, the three fishery
management councils with jurisdiction on the Atlantic coast, NMFS and various state
agencies. Also, the comprehensive reef-mapping program undertaken by the U.S. Coral
Reef Task Force is an example of a habitat mapping initiative that can be expanded and
linked to other mapping efforts. Dedicated funding for such monitoring programs needs
to be a prerequisite of any management effort; and (4) Finally, a method of addressing personnel shortages at the permit review and field staff level is absolutely critical;
although advances in technology can aid management efforts, meaningful changes in
management cannot occur without adequate on-the-ground staff to implement and oversee them.
CASE STUDY: COASTAL HABITAT PROTECTION PLANS.—One example of interagency integration at the state level is occurring in North Carolina. In 1997, the state of North
Carolina passed a Fisheries Reform Act which included a mandate to the North Carolina Marine Fisheries Commission (which oversees the regulation and harvest of living
marine resources) to develop an integrated Coastal Habitat Protection Plan (CHPP) in
cooperation with the Environmental Management Commission (which oversees water
quality permitting and regulations) and Coastal Resources Commission (which oversees
development within the stateʼs coastal zone). The goal of the plan is the “long term enhancement” of the value of coastal habitats to coastal fisheries. All three of the above
rulemaking commissions must approve the plans and ensure that their actions and rules
are consistent with the plan recommendations. The law also requires that the plans be
reviewed and updated every 5 yrs. Although the Division of Marine Fisheries is the lead
agency, the Plan Development Team consists of staff from the state divisions of Water
Quality, Coastal Management, Marine Fisheries, the Wildlife Resources Commission
(regulates inland fisheries and game), as well as the U.S. Fish and Wildlife Service and
the National Marine Fisheries Service. This level of coordination is unique for a statebased planning effort.
The plan is also being subdivided into detailed geographic management units that
reflect the boundaries of the stateʼs major coastal river basins, with the exception of separate management units for the sounds, the estuaries in the southern part of the state, and
the “coastal ocean” —which includes the stateʼs barrier island system and ocean-facing
beaches. The first iteration of the plan includes a Background Document, which contains general information common to all management units, and the main CHPP, which
includes specific management recommendations and issues of concern for each of the
geographic subunits. The plan development process includes several phases of internal
review (Fig. 2), including review by a six-member committee (Intercommission Review
Committee) that is composed of two members of each of the three rulemaking commissions before being approved for public comment. After public input, the plans are revised
before being sent to each of the full commissions for approval.
330
Figure 2. Coastal Habitat Protection Plan (CHPP) process under development in North Carolina.
CHPPs are subject to multiple reviews from the parent agency (North Carolina Department of
Environment and Natural Resources = DENR). Two rounds of public scoping will be conducted
to ensure public participation. EMC = Environmental Management Commission; CRC = Coastal
Resources Commission; MFC = Marine Fisheries Commission; IRC = Intercommission Review
Committee.
We are unaware of any other state planning effort as comprehensive and integrative
in scope as North Carolinaʼs CHPP initiative. Habitat protection will be achieved using all of the tools available to coastal management, fishery management, and water
quality management; success lies in the mandated adoption of similar regulations by
all three rulemaking commissions. Although it would prove logistically challenging to
extrapolate this model to a federal level, the merits are obvious and it could certainly be
employed on a regional basis. This plan, when completed, will be coordinated with the
SAFMC Habitat Plan and the MPA process under development.
331
CONCLUSIONS
Both the biological and administrative needs to link habitat protection initiatives with
marine protected area processes are evident. The habitat utilization patterns of the many
species in the South Atlantic and Caribbean are cross-shelf in nature and can involve sequential use of both vegetation and reef structures with development (Claro and GarciaArteaga, 1993; Ley et al., 1999); likewise, anthropogenic impacts can be spread across
wide arrays of habitats. Indirect and direct responses to stressors, therefore, can occur
during several ontogenetic stages within many interacting species. Although population
level effects of these cascades can be difficult to measure, they can undermine feeding, growth, and reproductive processes in important manners. Therefore, administrative processes that integrate fisheries, MPAs, and other habitat protection efforts are
needed.
As described previously, both habitat protection and MPA initiatives have the common
goal of sustainable management of coastal resources – integration of these processes
can effectively combine tools used for coastal land management, water quality management, and fishery management. There are obvious places where procedural links can be
forged. First, the massive data collection and synthesis efforts of the Councils with regard to EFH provide a substantial framework within which scientifically sound MPA design can be pursued. Second, as the gaps in EFH data are filled, Councils are better able
to determine HAPC locations – another logical platform from which to construct MPAs,
particularly when attributes like spawning aggregations are emphasized (SAFMC, 1998
<a,b?>; Lindeman et al., 2000). Similarly, several National Marine Sanctuary initiatives are primary models of no-take MPAs. An example is the successful multiyear development of the Tortugas Ecological Reserve no-take marine reserve within the broader
Florida Keys National Marine Sanctuary. Building upon information and mechanisms
already available will reduce the time and effort (and therefore money) needed to ensure
that the most scientifically valid and effective means of protection are being employed
(Coleman et al., 1999). Finally, it must be reiterated that success in any of the above areas
is critically dependent upon the filling of significant staff shortages, dedicated sources
of funding for monitoring and enforcement, and development of integrated regulations
among resource agencies at all levels of government.
ACKNOWLEDGEMENTS
Project #CMRC-00-IXNR-03-01A), National Oceanic and Atmospheric Administration (NOAA)
Undersea Research Program, U.S. Environmental Protection Agency, and Environmental Defense. We would also like to thank the Pew Charitable Trusts, Rockefeller Brothers Foundation
and the Curtis and Edit Munson Foundation for their generous support of our work in the south
Atlantic region. Finally, we thank the many scientists, resource managers, and fishermen who
have helped us in the development of these ideas. Views expressed herein are those of the authors,
and do not necessarily reflect the views of CMRC, or any of the supporting agencies.
LITERATURE CITED
Able, K. W. and M. P. Fahay. 1998. The first year in the life of estuarine fishes in the Middle Atlantic Bight. Rutgers University Press, New Brunswick. 342 p.
332
Allison, G. W., J. Lubchenco, and M. H. Carr. 1998. Marine reserves are necessary but not sufficient for marine conservation. Ecol. Appl. 8: S79–S92.
ASMFC (Atlantic States Marine Fisheries Commission). 1997. Submerged aquatic vegetation
Policy. ASMFC Habitat Management Series 3. ASMFC, Washington, DC. 9 p.
Ault, J. S., J. A. Bohnsack, and G. Meester. 1998. A retrospective (1979-1995) multispecies assessment of coral reef fish stocks in the Florida Keys. Fish. Bull. 96: 395–414.
_________, ____________, ________, J. Luo, S. G. Smith, J. E. Serafy, J. D. Wang, R. Humston,
and G. Diaz. 1999. A spatial dynamic multistock production model. Can. J. Fish. Aquat. Sci.
56: 1–22.
Boesch, D. F. and R. E. Turner. 1984. Dependence of fishery species on salt marshes: the role of
food and refuge. Estuaries 7: 460–468.
CFMC (Caribbean Fishery Management Council). 1985. Fishery management plan, final environmental impact statement, and draft regulatory impact review, for the shallow-water reeffish
fishery of Puerto Rico and the U.S. Virgin Islands. CFMC, San Juan. 69 p.
Christensen, J. D., T. A. Battista, M. E. Monaco, and C. J. Klein. 1997. Habitat suitability index
modeling and GIS technology to support habitat management: Pensacola Bay, Florida, Case
study. Report to NOAA Strategic Environmental Assessment Division, Washington, D.C. 90 p.
Claro, R. and J. P. García-Arteaga. 1993. Estructura de las comunidades de peces asociados a los
manglares del grupo insular Sabana-Camaguey, Cuba. Revista Oceanol. Ecol. Trop. Avicennia,
0: 60–82.
Coleman, F. C., C. C. Koenig, A. Eklund, and C. B. Grimes. 1999. Management and conservation
of temperate reef fishes in the grouper-snapper complex of the southeastern United States. Am.
Fish. Soc. Symp. 23: 233–242.
Fonseca, M. S., G. W. Thayer, A. J. Chester, and C. Foltz. 1984. Impact of scallop harvesting on
eelgrass (Zostera marina) meadows: implications for management. North Am. J. Fish. Manag.
4: 286–293.
GMFMC (Gulf of Mexico Fishery Management Council). 1989. Final environmental impact statement, fishery management plan, and regulatory analysis for the reef fish resource of the Gulf of
Mexico. GMFMC, Tampa. 356 p.
IUCN (World Conservation Union). 1988. Proceedings of the 17th Session of the General Assembly of IUCN and 17th Technical Meeting. San Jose, Costa Rica, 1–10 February, 1988. Gland,
Switzerland. np.
Johannes, R. E. 1981. Working with fishermen to improve coastal tropical fisheries and resource
management. Bull. Mar. Sci. 31: 673–680.
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. N. Am. J. Fish. Manage. 19:
436–453.
Keener, P., G. D. Johnson, B. W. Stender, E. B. Brothers, and H. R. Beatty. 1988. Ingress of postlarval gag, Mycteroperca microlepis (Pisces: Serranidae), through a South Carolina barrier
island inlet. Bull. Mar. Sci. 42: 376–396.
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. Est. Coast. Shelf
Sci. 48: 701–723.
Lindeman, K. C. 1997. Comparative management of beach systems of Florida and the Antilles: applications using ecological assessment and decision support procedures. Pages 134–164 in G.
Cambers, ed. Managing beach resources in the smaller Caribbean islands. UNESCO Coastal
Region and Small Island Papers 1. UNESCO, Paris.
_____________ and D. B. Snyder. 1999. Nearshore hardbottom fishes of southeast Florida and
effects of habitat burial caused by dredging. Fish. Bull. 97: 508–525.
_____________, R. Pugliese, G. T. Waugh, and J. S. Ault. 2000. Developmental patterns within
a multispecies reef fishery: management application for essential fish habitats and protected
areas. Bull. Mar. Sci. 66: 929–956.
333
Nelson, D. M., E. A. Irlandi, L. R. Settle, M. E. Monaco, and L. Coston-Clements. 1991. Distribution and abundance of fishes and invertebrates in southeast estuaries. ELMR Report No. 9,
NOAA/NOS Strategic Environmental Assessments Division, Silver Spring. 167 p.
NEFSC (Northeast Fisheries Science Center). 2002. Workshop on the effects of fishing gear on
habitats off the northeastern U.S. October 23-25, 2001. NEFSC Reference Document 02-01.
86 p.
NOAA (National Ocean and Atmospheric Administration). 1996. Magnuson-Stevens Fishery Conservation and Management Act, as amended through October 11, 1996. NOAA Tech. Memo.
NMFS-F/SPO-23. 121 p.
Odum, W. E. 1982. Environmental degradation and the tyranny of small decisions. BioScience
32: 728–729.
Parker, R. O., Jr. and R. W. Mays. 1998. Southeastern United States deepwater reef fish assemblages, habitat characteristics, catches, and life history summaries. NOAA Tech. Rpt. 138. 41 p.
Parrish, J. D. 1989. Fish communities of interacting shallow-water habitats in tropical ocean regions. Mar. Ecol. Prog. Ser. 58: 143–160.
Peterson, C. H., D. H. M. Hickerson, and G. G. Johnson. 2000. Short-term consequences of nourishment and bulldozing on the dominant large invertebrates of a sandy beach. J. Coastal Res.
16: 368–378.
Plan Development Team. 1990. The potential of marine fishery reserves for reef fish management
in the U.S. southern Atlantic. Snapper-Grouper Plan Development Team report for the South
Atlantic Fishery Management Council. NOAA Tech. Memo. NMFS-SEFC-261. National Marine Fisheries Service, Southeast Fisheries Center, Miami. 45 p.
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.
SAFMC (South Atlantic Fishery Management Council). 1982. Fishery management plan for coral
and coral reefs of the Gulf of Mexico and south Atlantic. Charleston, South Carolina. Prepared
in collaboration with Gulf of Mexico FMC, Tampa. 316 p.
____________________________________________. 1983a. Fishery management plan, regulatory impact review and final environmental impact statement for the snapper grouper fishery
of the south Atlantic region. SAFMC, Charleston. 237 p.
____________________________________________. 1983b. Source document for the snapper
grouper fishery of the south Atlantic region. SAFMC, Charleston. 173 p.
____________________________________________. 1998a. Final comprehensive amendment
addressing essential fish habitat in fishery management plans of the south Atlantic region. Including a final environmental impact statement/supplemental environmental impact statement,
initial regulatory flexibility analysis, regulatory impact review, and social impact assessment/
fishery impact statement. SAFMC, Charleston. 136 p.
____________________________________________. 1998b. Final habitat plan for the south
Atlantic region: essential fish habitat requirements for fishery management plans of the South
Atlantic Fishery Management Council. SAFMC, Charleston. 639 p.
Sedberry, G. R. and R. F. Van Dolah. 1984. Demersal fish assemblages associated with hard-bottom habitat in the South Atlantic Bight of the U.S.A. Environ. Biol. Fishes 11: 241–258.
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: 25–28.
Van Dolah, R. F., P. H. Wendt, and N. Nicholson. 1987. Effects of a research trawl on a hard bottom
assemblage of sponges and corals. Fish. Res. 5: 39–54.
Vestal, B. and A. Reiser. 1995. Part I—Syntheses, with annotated bibliography. Pages 1–135 in
Methodologies and mechanisms for management of cumulative coastal environmental impacts. Decision Analysis Series no. 6, NOAA Coastal Ocean Office, Silver Spring.
334
ADDRESSES: (D.N.R.) Environmental Defense, 2500 Blue Ridge Rd., Suite 330, Raleigh, North
Carolina 27607. (K.C.L.) Environmental Defense, Miami Field Office, 14630 SW 144 Terrace,
Miami, Florida 33186. CORRESPONDING AUTHOR: (M.A.D.) Environmental Defense, 2500 Blue
Ridge Rd., Suite 330, Raleigh, North Carolina 27607.

Back Reef II - Perry Institute for Marine Science

Transcription

Similar documents

The truth about… The truth about… habitats

Terumbu Karang dan Perubahan Iklim Coral Reefs

48x96 poster template

Hawai`i Wildlife Watching Guide

Brochure - Habitat for Zoomanity A Life Science Webquest

Reef Check Malaysia

Presentation Link

Oyster Restoration and Fishing Assessment

Seychelles Coral Reef Fish

2011 Impact Statement - Future Foundation Atlanta