EPA-260-R-06-004
December 2006
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in
the US Virgin Islands (St. Croix)
Final Report
Leska S. Fore
Statistical Design
136NW40thSt.
Seattle, WA 98107
William S. Fisher
U.S. Environmental Protection Agency
Office of Research and Development
National Health and Environmental Effects Research Laboratory
Gulf Ecology Division
1 Sabine Island Drive
Gulf Breeze, FL 32561
and
Wayne S. Davis
U.S. Environmental Protection Agency
Office of Environmental Information
Environmental Analysis Division
Environmental Science Center
701 Mapes Road
Ft. Meade, MD 20755-5350
Recocted /Recyclable
Printed wi'h Vegetable Oil on 100%
Chlorine
-------
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix)
SUMMARY
The goal of this study was to field test data collection and analysis protocols for stony
coral assemblages. From this study the most biologically meaningful and statistically
precise methods will be selected for inclusion in a long-term reef monitoring program for
the US Virgin Islands (US VI). Coral reef condition was measured at 61 reef stations in
St. Croix, USVI during 2006. Three observations for stony corals were recorded: species,
size, and percent live tissue. Stony corals were selected because they are primary
producers of the reef environment, they provide structure and habitat for other reef
organisms, and they support tourism and fisheries. Dive teams from the US
Environmental Protection Agency and the USVI Division of Environmental Protection
(DEP) collected physical measurements and recorded the condition of coral colonies
found within a radial belt transect. Different dive teams sampling the same reef station
reported very similar values indicating that the field protocol had good precision and low
measurement error associated with coral measurements. Indicators of coral condition
were tested against a gradient of human disturbance at three locations. Candidate metrics
for assessing coral condition were derived from four categories: species abundance and
composition, physical stature, biological condition, and coral community structure.
Human disturbance gradients were based on visual observations and narrative
descriptions of land use on shore. No quantitative or chemical measures of water quality
were collected. For the most intensely disturbed area, four metrics were highly correlated
with distance from an industrial point source: total surface area of coral, total live surface
area, taxa richness, and average colony size. For the other two gradients, changes in
indicator values were not associated with human influence, possibly because disturbance
in these areas was minimal or because indicators tested here were not be capable of
detecting subtle differences in reef condition. Many metrics were highly correlated with
depth, even when the range was only 20-40 ft. A statistical power analysis determined
that a survey area of 50 m2 was no more precise than a survey area half that size. Coral
metrics derived from this protocol had adequate precision to detect a reasonable level of
change in coral reef condition that would be protective of the resource.
-------
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix)
ACKNOWLEDGEMENTS
Divers were Jed Campbell and Bob Quarles (dive team leaders), Bill Fisher, and Rebecca
Hemmer (EPA ORD); Charles LoBue and Daniel Rodriquez (EPA Region 2); Alan
Humphrey (EPA ERT); Daniel Cooke (Lockheed-Martin REAC); Richard Henry (US
Fish and Wildlife); Kent Bernier, Diana Castro, Aaron Hutchins, Wesley Toller, and
Viol eta Villaneuva-Mayor (US VIDPNR-DEP); and Marcia Taylor (University of the
Virgin Islands). Conrad Knowles (DPNR-CZM) provided boat and navigational support.
Reef reconnaissance team was Valerie Chan (EPA ORD), Bill Fisher, Leska Fore,
Neelam Patel (EPA OW), and Heidi Bell (EPA OW). Data entry and metric calculation
were performed by Valerie Chan. This project was made possible by the US EPA Ocean
Research Vessel BOLD. Administrative support for the BOLD was provided by Ken
Potts (EPA OW) and Doug Pabst (EPA Region 2). Special thanks to Jarel Chamberlain,
Captain, and the crew of the BOLD who provided technical and logistic support for the
entire survey. Program and administrative support were provided by the EPA Coral Reef
Biocriteria Working Group chaired by Heidi Bell (EPA OW) and Lesa Meng (EPA
ORD). Discussions with Aaron Hutchins (USIDPNR), Patricia Bradley and Valerie
Chan (EPA ORD), and Heidi Bell were particularly constructive. Contract support was
provided by Leska S. Fore, Statistical Design, 136 NW 40th St., Seattle, WA 98107
under contract with Perot Systems Government Services, Inc., 8270 Willow Oaks
Corporate Drive, Suite 300, Fairfax, Virginia 22031.
The report should be cited as:
Fore, L. S., W.S. Fisher, and W. S Davis. 2006. Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix). EPA-260-R-
06-004. USEPA Office of Environmental Information. Washington. December 2006.
in
-------
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix)
CONTENTS
Summary ii
Acknowledgements iii
Table of Contents iv
List of Tables v
List of Figures v
Introduction 1
Project goals 3
Methods 4
Study area 4
Data collection 9
Candidate metrics for stony corals 10
Data analysis 14
Results 18
Sources of variance for the field survey protocol 18
Metric response to human disturbance 19
Natural variability associated with habitat differences 25
Biological comparison of CMZs 26
Power analysis 31
Population structure 32
Discussion 37
Efficacy of field protocols 37
Coral response to human disturbance 39
Influence of habitat on stony coral assemblages 41
Conclusions 42
References 43
iv
-------
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix)
TABLES
Table 1. Number of reef stations sampled in each CMZ 6
Table 2. Description of candidate coral metrics 11
Table 3. Natural history information for coral species observed in 2006 13
Table 4. Correlation of candidate coral metrics with disturbance and depth 20
Table 5. MOD values for candidate coral metrics 31
Table 6. Detectable change as a percent for candidate coral metrics 32
FIGURES
Figure 1. Land use/land cover classes and coastal management zones for St. Croix 5
Figure 2. Reef station locations in the West CMZ 7
Figure 3. Reef station locations in the North, East, and Buck Island CMZs 8
Figure 4. Reef stations near the industrial complex along the south side of the island in
the East, South, and Southwest CMZs 8
Figure 5. Schematic diagram of radial belt transect used to sample reef stations 9
Figure 6. Variance components for candidate coral metrics 19
Figure 7. Number of colonies and number of taxa plotted against distance from the
commercial dock on the south side of St. Croix 22
Figure 8. Percent live tissue (averaged over all colonies) and total SA plotted against
distance from the commercial dock on the south side of St. Croix 23
Figure 9. Live surface area and percent live surface area plotted against distance from the
commercial dock on the south side of St. Croix 24
Figure 10. Average surface area (averaged for all colonies) plotted against distance from
the commercial dock on the south side of St. Croix 25
-------
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix)
Figure 11. Depth and number of colonies for each CMZ 27
Figure 12. Number of taxa and percent live tissue (averaged for all colonies) for each
CMZ 28
Figure 13. Total surface area and live surface area for each CMZ 29
Figure 14. Percent live surface area and average surface area (for all colonies) for each
CMZ 30
Figure 15. Number of colonies by size class for Diploria strigosa andMontastraea
annularis 33
Figure 16. Number of colonies by size class for Montastraea cavernosa andM
faveolata 34
Figure 17. Number of colonies by size class for Porites astreoides and P. porites 35
Figure 18. Number of colonies by size class for Sidemstrea siderea 36
VI
-------
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix)
INTRODUCTION
Coral reef communities surrounding the US Virgin Islands (US VI) represent a valuable
economic and aesthetic resource for visitors and residents (USVIDEP & DPNR, 2004).
The government of USVI recognizes their value and supports a variety of coral reef
monitoring efforts (Nemeth et al., 2004). Along with rivers, streams, lakes and estuaries,
the Federal Clean Water Act (CWA, 1972) provides a regulatory framework for the
assessment, management and protection of near-shore water resources, including coral
reefs. Both the CWA and the US Virgin Islands Territorial Water Pollution Control Act
(1972) outline regulations for protection of surface waters and the biological assemblages
they support. Two programs specifically rely on biological monitoring data in coastal
marine areas: the 301(h) waiver program and the 403(c) ocean discharger program. The
waiver program allows marine dischargers to defer secondary treatment if they can show
the discharge does not affect biological communities. The ocean discharger program
requires all dischargers to marine waters to provide an assessment of the biological
community in the area of the discharge (Jameson et al., 1998).
The CWA authorizes the US EPA to determine appropriate minimum levels of protection
and provide national oversight to State, Territorial and Tribal programs; however,
considerable flexibility and discretion are left to States and Territories to design their own
programs and establish levels of protection beyond any national minimums (EPA, 2005).
The regulatory framework of the CWA requires States and Territories to adopt water
quality standards (WQS) to protect their waters. WQS are part of State law and define the
water quality goals for a water body by designating the use(s) and setting criteria
necessary to protect the use(s). WQS include three parts: 1) designated uses, 2) numeric
and narrative criteria that protect the uses, and 3) antidegradation policies to prevent
deterioration of high-quality waters (EPA, 2006). Examples of designated uses include
drinking water, navigation, and support of aquatic life. Once designated uses are
described, criteria for the protection of each use must be defined. Criteria may be tied to
threshold values of physical, chemical or biological measurements of aquatic condition.
When criteria fail to support the designated uses, a water body is listed as impaired.
States and Territories are required to assess and report whether their surface waters are
supporting or failing to support designated uses. At all levels, water quality standards are
much better defined for freshwater and estuarine environments than they are for coral
reefs where standards and guidelines are just beginning to emerge (Fisher, in press).
For the USVI, designated uses for surface waters are described as follows (USVI DEP &
DPNR, 2004; Hutchins, 2004):
Class A - Waters are for the preservation of natural phenomena requiring special
conditions with existing natural conditions that shall not be changed. Class A water
standards are the most stringent of the three classes because of the pristine or near
pristine state of waters in this classification.
Class B - Waters are for the propagation of desirable species of marine life and for
primary contact recreation.
1
-------
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix)
Class C - This classification is similar to Class B, except that it has slightly less
stringent water quality standards for a limited number of parameters.
Most States, Tribes, and Territories have similar types of narrative criteria that specify
the protection of aquatic life as a designated use. Phrases used above such as
"preservation of natural phenomena" and "propagation of desirable species" are an
example of this type of narrative criteria for the support of aquatic life (a type of
designated use). EPA's Office of Water has developed guidance to help States,
Territories and Tribes better characterize and more specifically define aquatic life uses as
part of their water quality standards (EPA, 2005). EPA recognizes that direct
measurements of the resource that is of greatest concern, e.g., coral reef communities, are
more protective than surrogate measures, e.g., water chemistry.
In 2001, the National Research Council (NRC) published a report on Assessing the
TMDL Approach to Water Quality Management in which the authors recommended
tiering designated uses for improving the decision-making related to setting water quality
standards (NRC, 2001). The NRC found the CWA's goals to be too broad to provide the
operational definition of designated uses needed to support aquatic life and recommended
greater specificity in defining aquatic life uses. For example, rather than stating that a
water body needs to be "fishable," the designated use should specifically describe the
expected fish assemblage (e.g., cold water fishery, warm water fishery, or salmon, trout,
bass, etc.). Tiered aquatic life uses (TALUs) are bioassessment-based statements of
expected biological condition in specific water bodies that allow more precise and
measurable definitions of designated aquatic life uses (EPA, 2005).
Designated uses are written in qualitative, narrative terms; therefore, the challenge is to
relate a water quality criterion to the designated use. Establishing this relationship is more
straightforward when the water quality measure, or criterion, is closely and meaningfully
related to the designated use. For this reason, the NRC recommended the use of
biological information to define more appropriate aquatic life uses. Specifically,
biological criteria, or biocriteria, define a desired biological condition for a water body
and can be used to evaluate the biological integrity of a water body (Karr and Chu, 1999).
The TALU approach provides an interpretative framework for developing a technical
program that will tighten the linkage between narrative use statements and numeric
biological criteria (EPA, 2005).
Individual WQS for States, Tribes, and Territories provide the foundation for the
management of surface waters and pollution control programs. WQS provide the basis
for determining whether a water body is impaired. Impairment triggers a process to
evaluate the total maximum daily load (TMDL) of pollutants at the site and management
actions are required to bring the site back into compliance with its designated use (Karr
and Yoder, 2004). Historically, States, Tribes, and Territories have taken different
approaches to defining their WQS. Different approaches are acceptable to EPA as long as
a minimum standard is protected. In other words, States, Tribes and Territories are
encouraged to be more protective than the national minimum. Once WQS are in place,
States, Tribes, and Territories are authorized to implement monitoring programs that
-------
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix)
allow them to report on the attainment of those standards and to identify and prioritize
waters not attaining standards for future management and abatement programs (EPA,
2005).
Project goals
The primary purpose of this project is to assist USVI in developing assessment tools, i.e.,
scientifically defensible protocols and a long-term monitoring program, for coral reefs.
This report describes the field testing of data collection and analysis protocols derived
from a coral survey conducted around St. Croix (USVI) during 2006. A companion report
described the long-term monitoring approach and survey design for USVI (Fore et al.,
2006b).
This report focuses on identification of biological indicators that can be used to define
biological criteria for the protection of coastal resources and for managing the local
human activities that threaten them. Although coral reefs are also sensitive to global
disturbance (e.g., elevated seawater temperatures), USVI DEP cannot manage human
disturbance at the global scale; therefore, our focus for this study was on developing tools
to assist local managers.
Data collected for this study were used to 1) determine the optimum size for a field
transect, 2) compare data collected by different dive teams, 3) evaluate the response of
coral indicators to a gradient of human disturbance, 4) characterize coral condition in
management zones expected to have different designated uses, and 5) measure the
potential ability of the coral metrics to detect change in reef condition.
-------
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix)
METHODS
Study area
Stony corals were surveyed at 61 reef stations around St. Croix, USVI, during February
2006. Reef stations were selected to satisfy two objectives: 1) test for association between
coral condition and human disturbance gradients in the watershed and near-shore
environment and 2) summarize coral condition in each of seven geographic areas
surrounding St. Croix. The type of coral reef observed in a near-shore environment
depends on geographic orientation (leeward or windward), patterns of water movement,
and depth profile. Seven coastal management zones (CMZ) were defined by resource
managers and scientists at the US VI Division of Environmental Protection (DEP) and
Department of Planning and Natural Resources according to the type of coral habitat
observed and the type of human land use within the water, along the shore, and inland
(Figure 1). Six of the CMZs are under the managerial jurisdiction of USVI DEP, but the
Buck Island Reef National Monument off the northeast coast of St. Croix is managed by
the National Park Service.
Starting on the west side and moving in a clockwise direction: the West CMZ includes
the city of Frederiksted and has the only large public dock on the island used by cruise
ships (Figure 2). Although cruise ships visit infrequently, traffic is expected to increase
during coming years. Recently, the area around the pier has a history of small boat and
yacht use which may be associated with anchor damage to reefs and nutrient enrichment
from wastewater. An earlier and more extensive history of anchorage north of the pier by
much larger ships has also been documented by Toller (2005) who identified > 21
hectares north of the pier that have been impacted by large ship anchors. Also located
near the pier is a small sewer overflow. The city of Frederiksted itself may be expected to
contribute to general disturbance. Further north and south from the pier, human influence
decreases. The Northwest CMZ has tourism associated with diving and fishing, but
otherwise only minimal human influence. Recent studies indicate that the impact of
recreational diving may be greater than previously expected (Barker and Roberts, 2004).
In the North CMZ is the city of Christiansted (Figure 3). The harbor at Christiansted has
boat traffic and 50-75 boats moored in the harbor. Potential non-point and point sources
of disturbance exist at Christiansted with urban development and a large wastewater
(sewage) treatment plant that discharges to the reef area.
Moving further east, Buck Island CMZ has no human development and visitors are
limited to daylight hours. Nonetheless, this area still supports recreational diving and
boating. The relatively low intensity of human disturbance means that Buck Island may
provide reference sites for other areas on the main island. The East CMZ excludes the
reef area surrounding Buck Island. Most of the East CMZ is included in the East End
Marine Park, designated in 2003; the park includes "no take" zones for fishing,
recreational areas, and a turtle preserve. Sources of disturbance from shore include run-
off and sediment from several unpaved and steep roads in this zone. Studies in St. John
and St. Thomas have shown that erosion from unpaved roads can be very high and that
-------
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix)
IN
A
024
8 12
: ParksJRecreation/Gpen Space
Undeveloped
: Agriculture
Lightly Impacted
I Metal/Resort
\ Highly Impacted
I Public Facilities
St. Croix US Virgin
Management - coral reef habitat
Legend
j [ f'.1arta(^iTieF'it zoites
r'""l Cnk.ni2f.fi BS'Jiocf.
: i Colonized Pavenwnt
: Colonized Pav*m*nt -.vitti Sand C.hantwls
r-J] Liswar Reef
"""I Patch Reef {AigiBiiated,
[ j Patch Reef .Individual)
;=3 S:attซed CoiaURook in UncorsolKiated Sediment
: : Spuf and Gicove Reef
Figure 1. Land use/land cover classes (upper panel) and coastal management zones and
coral reef habitat types for St. Croix (lower panel; Hutchins, 2004).
-------
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix)
sedimentation can reduce coral cover (Nemeth and Nowlis, 2001; Ramos-Scharron and
MacDonald, 2005). The South CMZ is an agricultural area (Figure 4). Adjacent to this is
the Southwest CMZ in which are located a large petroleum refinery, dredged channels for
commercial docks, the airport, the land fill, and a rum distillery that has discharged
effluent for >50 years. The South CMZ is nominally upstream of these disturbances
because prevailing wind and current are ENE. Most of the disturbance appeared to be
confined to the Southwest CMZ.
The number of reef stations surveyed varied within each CMZ (Table 1). In general,
stations were selected to test specific hypotheses and to evaluate the merit of the various
field sampling protocols. Our intention was not to identify locations for a long-term
monitoring design and no randomization was used in station selection. Most stations were
selected to provide data for the three independent tests of metric responsiveness to a
human disturbance gradient. Other stations were included to characterize stony coral
populations in different management zones. Three stations had duplicate samples
collected, that is, different dive teams surveyed the same radial belt transect to evaluate
the measurement error associated with the field sampling protocol and to determine how
easily the method could be transferred to new dive teams. Within a reef area, sampling
locations (stations) were selected to represent the best available habitat, that is, areas with
a variety of coral colonies. Areas with sand, seagrass, or only minimal coral cover were
avoided.
Table 1. Number of reef stations sampled in each CMZ.
CMZ name, number of reef stations sampled, number of stations included within a gradient of
human disturbance, and whether duplicate sampling occurred in the CMZ.
CMZ
Total
Number of stations
In Gradient
Duplicate sampling
Buck Island
East
North
Northwest
South
Southwest
West
Total
10
9
11
4
5
10
12
61
BIOS: 6 dive teams; full transect
4 (South) + 5 (North gradient)
11 (North gradient)
0
5 (South gradient)
10 (South gradient)
12 (West gradient)
WE07: 4 dive teams; 1/4 transect
WE14: 6 dive teams; % transect
-------
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix)
USVi 3D Coral Sampling Stations - Feb 2006
Western Region
-,:/
W :!ll
EEEB ', WE tf),
S ; WE If
-
Frederiksted
t
1 o
usgs-nwrc-5-200B-0119
Figure 2. Reef station locations in the West CMZ (WE). The public pier is just south of station WE13.
-------
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix)
USVI 3D Coral Sampling Stations - feto 2006
Eastern Region
Christiansted
Figure 3. Reef stations located in the North (NO), East (EE), and Buck Island (Bl) CMZs.
Industrial Area
V. s&RfTsyr 08s
^H i-- -
$&<<'
Figure 4. Reef stations along the south side of the island in the East (EE), South (SO), and Southwest
(SW) CMZs.
-------
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix)
Data collection
Each survey station was established by placing a tripod on the substrate which held an
upright pole in place. A 6-m line was attached to the top of the pole. During sampling,
one diver, the line tender, extended the line to its full extent and marked the spot with a
start/stop flag. Markers constructed of alligator clips and beads were suspended from the
line at 3- and 5-m distances from the pole so that surveyors could distinguish the 2-m
band of the radial transect (Figure 5). The line tender traveled around the pole as the
survey diver completed the coral measurements. At each 90ฐ increment of the belt
transect, the quadrant was marked as 1 through 4. Due to the high density of coral
colonies at many stations, only 1A of the radial belt was surveyed at 29 stations and the
full transect was surveyed at 32 stations. At one additional station (WE14) only 1A of the
radial belt was surveyed for training purposes only. For a full radial belt, the surveyed
area equaled 50.2 m2; for a /^-transect the area equaled 25.1 m2. For each colony found
within the transect, the following data were collected: species name, maximum diameter
parallel to the substrate (length), diameter perpendicular to first measurement at the
center of the colony (width), maximum height from the substrate, and % of the surface
area covered with living polyps. The % live tissue was recorded for each colony as 0, 0-
25, 25-50, 50-75, 75-100, or 100%. For metric calculation, the middle of the range was
used, e.g., for 0-25% live tissue, we used 13% for calculating sums and averages. Only
corals >10 cm in their longest dimension were recorded.
Figure 5. Schematic diagram of radial belt transect used to sample reef stations. Shaded
area represents area where corals are measured. Quadrant 1 through 4 are indicated and
"X" marks the location of the tripod.
-------
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix)
Candidate metrics for stony corals
Indicators of stony coral condition related to abundance and composition, physical
stature, biological condition and community structure were selected for testing (Table 2;
Jameson et al., 2001; Fisher et al., 2007). Indicators that show a consistent response to
different types of human disturbance in different habitat types and geographic areas are
considered metrics and used to define biocriteria (Karr and Chu, 1999; Fisher, in press).
The indicators described here are best qualified as "candidate" metrics because they have
not been extensively tested. This study represents one such test of these candidate metrics
for stony corals. Candidate metrics were calculated for each Vz- transect (25 m2 area).
Abundance and composition. The total number of colonies and the total number of
unique taxa that they represent are expected to decline as human disturbance increases.
For stony coral assemblages, tolerant and intolerant taxa have yet to be consistently
identified although some authors recommend Porites astreoides, P. porites, Siderastrea
siderea, andAgaricia agaricites as tolerant species (Tomascik and Sander, 1987).
Although rare or uncommon taxa are not necessarily sensitive or intolerant, some may be.
In the absence of information about which taxa are intolerant, rare taxa were simply
defined as those taxa for which <20 colonies were found (out of a total of 3720 colonies
for all taxa). Although the presence of dead coral may be indicative of poor reef
condition, we did not quantify the abundance of dead coral heads unless they could be
identified to genus. In general, coral rubble and pieces not connected to the reef could not
be identified to genus; often they could not even be identified as coral.
Physical stature. An earlier study in the Florida Keys used standard-sized cubes to
estimate colony size and calculated surface area by summing the areas of five sides of the
cube (Fisher et al, 2007). For this study, direct measurements of each colony provided
information to calculate a more exact surface area for each colony. Two equations were
used to calculate the surface area depending on the ratio of the colony height to radius
(measured as /^ of the maximum diameter). When the ratio is close to 1, a hemisphere is
a logical choice for the geometric model of colony size (SAhemi= 27ir2). When the height
is greater than the radius, a cylinder may be more appropriate (SAcyi = Tir2 + 27irh). For a
heightradius ratio >1.3 and <5.1, a cylinder was used to calculate surface area; for other
heightradius ratios, a hemisphere was used to calculate surface area. About half the
colonies measured for this study were h:r <1.3; only 8 colonies were h:r >5.1.
10
-------
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix)
Table 2. Description of candidate coral metrics.
Name of candidate metric, predicted response to an increase in human disturbance, and a
description of how each metric was calculated. Candidate metrics were calculated for each
1/4- transect (25 m2 area).
Candidate metric
Predicted Description
response
Abundance & Composition
Number of colonies
Taxa richness
% "Rare" colonies
% SA of "rare" taxa
decrease Number of stony coral colonies >10 cm in their
longest axis
decrease Number of unique taxa
decrease Percent of colonies at the station that were defined
as "rare" (<20 colonies found in taxon)
decrease Percent of total surface area from "rare" taxa
Physical stature
Total SA
Average radius of all
colonies
Average colony SA
Biological condition
% Live tissue
Live SA
Dead SA
% Live SA (vitality index)
% Hermaph. colonies
% SA of hermaph. taxa
% Gonochoristic colonies
% SA of gonochoristictaxa
% Brooder colonies
% SA of brooder taxa
% Spawner colonies
% SA of spawner taxa
Community structure
% SA Diploria
% SA Montastraea
% SA Porites
% SA Siderastrea
Percent dominance
decrease Total 3D surface area of all corals found (m )
decrease Average of the three measures of radius for each
colony, then average of all colonies
decrease Average of the total 3D surface area for each colony
decrease Percent live coral tissue on each colony averaged
for all colonies
decrease Sum of live colony surface areas for all colonies
increase Sum of dead (denuded) colony surface areas for all
colonies
decrease Live SA divided by Total SA
increase Percent of colonies that belong to hermaphroditic
taxa
increase Percent of total surface area from hermaphroditic
taxa
decrease Percent of colonies that belong to gonochoristic taxa
decrease Percent of total surface area from gonochoristic taxa
decrease Percent of colonies that belong to brooder taxa
decrease Percent of total surface area from brooder taxa
increase Percent of colonies that belong to spawner taxa
increase Percent of total surface area from spawner taxa
unknown Percent of total surface area from Diploria spp.
unknown Percent of total surface area from Montastraea spp.
unknown Percent of total surface area from Porites spp.
unknown Percent of total surface area from Siderastrea
siderea.
increase SA of the taxon with the greatest SA divided by total
SA
11
-------
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix)
Total surface area (SA), which is a measure of both live and dead portions of the coral
colony, is expected to decline as colonies die and are not replaced. Coral skeletons
lacking live tissue are vulnerable to erosion by both biological and physical processes.
Colony size measured as either average radius or as SA should also decrease with human
disturbance as larger colonies are eliminated by disturbance events over time. Coral
recruitment was not specifically measured for this study and only colonies >10 cm were
recorded. Nonetheless, we expect a trend toward smaller colonies to indicate a change in
the stony coral assemblage.
Biological condition. Coral are colonial animals that propagate somewhat like plants,
thus there are several ways to quantify their abundance and relative abundance.
Taxonomic groups of interest can be measured in terms of numbers of colonies or as
surface area of tissue. "Percent live tissue" was calculated as the average of the amount of
live coral tissue observed on each colony. Thus, a small or large colony will contribute
the same amount of information to the final value. In contrast, "percent live SA"
calculated the area of each coral colony that was alive, summed that area for all colonies,
and divided the value by the sum of total SA for all colonies. This calculation has also
been called the "vitality index" (Fisher, 2007). For this metric large colonies contributed
proportionately more information to the final station value. "Live SA" is the total area of
live tissue summed over all colonies. "Dead SA" was the total dead surface area summed
over all colonies.
When conditions are stable, we expect taxa with separate male and female colonies
(gonochoristic) to be more common than hermaphroditic taxa which may be more typical
in uncertain conditions. Similarly, in more stressful conditions we expect coral colonies
with a reproductive strategy designed to take advantage of changing conditions to be
more common. Thus, we might expect to find more brooders than spawners in less
disturbed locations and more spawners in locations with higher disturbance. Although not
tested specifically for stony corals, these ideas have been applied to numerous other
taxonomic groups to interpret patterns in demography and reproductive strategy across
species (Reznick et al., 2002). Richmond and Hunter (1990) list reproductive
characteristics for a subset of the taxa found in St. Croix (Table 3).
Community structure. Four genera contributed the greatest overall percentage of
surface area to St. Croix coral reefs. Diploria, Montastraea, andPorites included
multiple species; Siderastrea siderea was the only species in its genus. Expectations for
these particular taxa in response to human disturbance are unknown. Dominance was
measured as the taxon with the largest SA divided by the total SA for the station * 100%.
Dominance quantified the relative importance of the dominant species at a station,
regardless of which species it is. We expect dominance to increase with disturbance as
tolerant taxa dominate the assemblage.
12
-------
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix)
Table 3. Natural history information for coral species observed in 2006.
Stony coral name, the total number of colonies found, total surface area for all stations,
reproductive type (hermaphroditic or gonochoristic), reproductive mode (brooder or spawner),
and whether the taxon was designated as rare for this study.
Taxon name # colonies
Acropora cervicornis
Acropora palmata
Agaricia agaricites
Agaricia fragilis
Agaricia humilis
Agaricia larmarckii
Agaricia spp
Agaricia ten ui folia
Colpophyllia natans
Dendrogyra cylindrus
Dichocoenia stokesii
Diploria clivosa
Diploria labyrinthyformis
Diploria strigosa
Eusmilia fastigiata
Isophyllia rigida
Isophyllia sinuosa
Isophyllia spp
Madracis decactis
Madracis mirabilis
Madracis spp
Meandrina meandrites
Millepora complanata
Montastraea annularis
Montastraea cavernosa
Montastraea faveolata
Montastraea franksii
Montastraea spp
Mycetophellia larmarckiana
Mycetophyllia spp
Oculina varicosa
Porites astreoides
Porites porites
Siderastrea siderea
Solenastrea bournoni
Stephanocoenia intersepts
2
6
26
6
1
2
33
1
34
5
19
91
66
643
17
2
1
1
41
17
5
87
27
337
541
227
53
1
1
5
1
820
226
316
5
54
SA (cm2) Repro. type
4,919
76,261
6,461
1,135
147
339
14,901
579
129,184
35,374
8,796
99,672
75,006
668,383
10,144
510
179
101
18,679
12,995
3,083
53,917
47,395
2,392,281
1,139,876
1,270,845
181,716
179
472
1,153
693
411,344
431,571
491,697
18,626
13,237
H
H
H
H
G
H
G
G
G
H
G
H
G
G
Repro. mode Rare?
S
S
B
B
B
B
S
S
B
B
B
S
S
B
B
S
Rare
Rare
Rare
Rare
Rare
Rare
Rare
Rare
Rare
Rare
Rare
Rare
Rare
Rare
Rare
13
-------
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix)
Data analysis
Different sets of stations were used to answer different questions. For metric testing, only
stations located along the proposed gradients of human disturbance were used. Metrics
were tested independently across all three of the gradients. For the southern gradient of
human disturbance, metric response was tested using stations in three CMZs (East, South,
and Southwest). For the northern gradient all stations were in the North CMZ; for the
western gradient all the stations were in the West CMZ. The field protocol was evaluated
using specific approaches. For example, information was needed to examine sources of
variability associated with different station locations, different transect halves
(microhabitat differences), and different dive teams (measurement error). Only BIOS had
duplicate samples collected by different dive teams for both halves of the transect. When
comparing different sources of variance, e.g., measurement error and microhabitat
differences, the preferred method estimates variance components from the same set of
locations rather than using different sets of locations to estimate different variance
components because the results will be more reliable. The ideal design for our study to
estimate variance components would have included duplicate samples by different dive
teams and full transect sampling at every station; this was too time consuming and only a
few stations had duplicate sampling. For this analysis, only stations in Buck Island CMZ
were included; there were two reasons for this. First, inclusion of all 61 stations would
have made the design even more unbalanced than it was with only a few stations having
duplicate dive team samples. Second, we were interested in the relative contribution of
station differences within a management zone rather than across all zones around the
island.
For statistical power analysis, a third set of 16 stations was used to evaluate differences
associated with the two halves of the belt transect (microhabitat differences) and
differences associated with station locations in the larger reef. For this comparison we
were not interested in comparing measurement error associated with dive teams. Stations
in the Buck Island, East, and West CMZs were used for this analysis. Stations in the
North and Northwest CMZs were not used because most stations had only ^-transects
sampled. Stations in the South and Southwest CMZs were not included because coral
density and surface area were much lower than for the other CMZs. Power analysis tested
for the amount of change in metric values that could be detected within a CMZ.
Metric testing. Candidate metrics were tested for their correlation (Spearman's r) with
distance from the approximate center of human disturbance in three locations: from the
public dock on the west side, from the marina in Christiansted harbor, and from the
commercial dock on the south side. Not all of the 61 stations were located within a
gradient of human disturbance. For metric testing, 19 stations along the south side in the
East, South, and Southwest CMZs, 12 stations in the West CMZ, and 16 stations in the
North CMZ were used to test for metric response across a gradient of human disturbance
(see Table 1). No quantitative information related to water chemistry or other measures of
site condition were collected as part of this study to test metrics. All metrics were
calculated for /^-transects. For stations with full transects surveyed, metric values from
14
-------
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix)
the two /^-transects were averaged so that each station had only a single value for each
metric.
Evaluation of the field survey protocol. Coral reef communities represent a continuous
resource that could be divided into discrete sampling units in a variety of ways. For this
study radial belt transects were used to define the survey area at each station. Data
collected were used to determine whether the entire belt must be surveyed or if a smaller
area would suffice. For any monitoring study, the smallest area that will provide a
reliable estimate of site condition is preferred in order to minimize the time spent at each
station.
Components of variance analysis was used to evaluate the relative contribution to the
overall variance of coral metrics due to differences associated with stations, transects, and
measurement error. Components of variance uses an ANOVA model, one for each coral
metric, to partition the different sources of variance and compare their relative
contribution. Measurement error was defined as the variance associated with duplicate
surveys at the same station by different dive teams. This source of variance should be
small relative to the variance due to differences in stations or zones targeted in a
monitoring program. If dive teams obtain similar values for the same station, we can
assume that the protocol is robust and transferable to new situations. Variance associated
with the two sides of the radial belt represents the natural variability in coral condition at
the microhabitat level. Variance of metric values due to different station locations
represents the differences associated with depth, current, substrate, or other natural
features. The objective of any monitoring effort is to detect effects of human disturbances
beyond the natural variability associated with different station locations.
Ten stations in the Buck Island CMZ were used to compare the different sources of
variance for seven candidate coral metrics. Although duplicate samples were collected on
two other stations in the West CMZ, only BIOS had full transect surveys that allowed a
simultaneous evaluation of variance due to stations, transects within stations, and dive
teams. The design for this analysis was unbalanced (n = 23). Of the 10 stations used,
three had full transects surveyed while the others had only ^-transects, and all the
duplicate surveys were collected at a single station. Six dive teams sampled both l/2-
transects on BIOS.
Power analysis to detect changes in reef condition Statistical power is the probability
of detecting a change should a change truly occur. Statistical power is a function of
precision: the more variable a measure is, the more difficult it will be to detect a
difference between samples. Thus, a greater difference must be observed to detect change
in a more variable measure of coral condition. Variance is used in statistical power
equations to calculate the probability of detecting a difference for a selected statistical
model and known variance.
For this study, statistical power analysis was used to answer two questions. First, what is
the amount of change that we could expect to reliably detect using these candidate coral
metrics? This describes how sensitive the metrics are to change. Second, are full transect
15
-------
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix)
samples better than /^-transect samples for detecting change? If so, it must be determined
whether it is sufficiently better to offset the costs of surveying twice the area at the same
location. It is possible, for example, that sampling more stations will provide greater
power and a more efficient monitoring strategy.
For simplicity, a two-sample t test was used to compare and evaluate the relative
precision of seven candidate coral metrics. To estimate the amount of change in a
candidate metric that we can potentially detect, the minimum detectable difference
(MDD) for a two-sample t test can be calculated (Zar, 1984). A two-sample t test
evaluates the difference between two sets of samples. The samples could be from reef
stations in areas with different types of human disturbance or they could be from the
same reef areas sampled at different times. If the two samples are from the exact same
reef stations through time, a paired t test should be used instead. The MDD represents the
smallest difference between the mean metric values for the two sets of reef stations that
would indicate a statistically significant change.
Once the variance is known (or estimated), alternative sampling designs can be compared
for their relative sensitivity to detect change, either through time or between different reef
areas. For example, the relative sensitivity of a design with samples from 5, 10 or 15 reef
stations can be compared, even if 15 stations were not sampled in the original survey.
Data from the 2006 survey in St. Croix were used to estimate the variance. Because
values for the seven candidate coral metrics differed by CMZ, a one-way ANOVA design
was used to estimate the variance within each CMZ. The assumption for this approach to
statistical power analysis is that the two samples used in the t test would be from within
the same CMZ. Thus, the mean squared error from the ANOVA represents the variability
associated with different stations within a CMZ. This approach controls for the effect of
CMZ location.
Sixteen stations in three CMZ's with full transect surveys and minimal human
disturbance were used to calculate MDD: three stations in the Buck Island CMZ, six in
the East CMZ, and seven in the West CMZ. Statistical power was calculated separately
for seven candidate metrics using two data sets. The first data set used only data from a
single /^-transect at each of the 16 stations. The second data set used the average value
from the two /^-transects for each station. It was hypothesized that information provided
by the full transect (represented as an average of the two ^-transects) would be more
precise than a 1A transect, yielding a smaller MDD and a greater ability to detect change
in reef condition.
MDD was calculated as:
MDD>
Where s2 = the mean squared error from ANOVA for each metric,
n = the number of stations sampled on each occasion in a CMZ,
16
-------
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix)
ta(i), v = the t value for alpha of 0.1 for a 1-sided test,
^p(i), v = the t value for beta of 0.1 for a 1-sided test, and
v = 2n-2.
Patterns in population structure. To evaluate patterns in population structure at the
species level, the seven most common species were plotted according to the number of
colonies found in each size class. Size classes were defined based on the current data set
and used surface area to define six bins with a fairly even distribution of colonies across
bins. Bins doubled in size for each category for six species; for Porites astreoides the bin
size increased evenly. Patterns associated with both species abundance and distribution of
colonies across size classes were used to investigate possible differences in coral
composition, size and health in the different CMZs.
17
-------
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix)
RESULTS
A total of 4647 colonies were measured for this study; 3720 were observed on the 61
stations and an additional 927 colonies represented repeat measures by different dive
teams for training and protocol testing. A total of 31 taxa were recorded. The dominant
species as measured by both number of colonies and total S A were Diploria strigosa,
Montastraea annularis, M. faveolata, M. cavernosa, Porites porites, P. astreoides, and
Siderastrea siderea (see Table 3).
Sources of variance for the field survey protocol
This analysis compared the sources of variance for seven candidate coral metrics. When a
dive team surveys a field station and measures coral condition, there are multiple sources
of variability for the values observed. Some differences represent nuisance variance,
other differences are those we wish to detect. For example, when different dive teams
observe slightly different numbers of coral colonies in a transect, this represents
measurement error, a nuisance variance that should be minimized. Differences associated
with transect placement were captured by the different halves of the radial belt and
represent another source of nuisance variance. In contrast, station differences should be
relatively large compared to nuisance sources because these represent the types of
differences that may have been created by anthropogenic stressors.
The metrics least affected by nuisance sources of variance were number of colonies, total
SA, and average colony SA (Figure 6). For all seven coral metrics the percentage of the
total variance due to differences associated with duplicate sampling by the six different
dive teams was small compared to station differences. The greatest differences in dive
teams were associated with % live tissue and live SA. Both these metrics rely on
approximate measures of live tissue cover made by the surveyor. Differences associated
with transects within stations were larger than diver differences for four of the seven
metrics: # colonies, # taxa, % live tissue, and % live SA. Different dive teams obtained
very similar values for the seven metrics. Microhabitat differences at the transect level
were large enough for some metrics to be of concern and were evaluated separately (see
"Power analysis" below).
18
-------
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix)
CD
o
c
CO
c
CO
i
100
80
60
40
20
0
w
CD
'c
2
o
O
a
CO
0)
w
w
CO
Is
o
CO
CO
0)
O)
0)
Error (team)
Transect (Station)
Station
Figure 6. Variance components for candidate coral metrics. Total variance for stations in the Buck
Island CMZ was partitioned according to differences associated with 10 stations, two transects
within stations, and measurement error associated with six different dive teams.
Metric response to human disturbance
Changes in the coral assemblage were most obvious around the commercial docks on the
south side of St. Croix. The number of taxa, total SA, and average colony size all
declined at stations closest to the docks (Table 4). Live SA and dead SA both declined for
stations closest to the docks (Figures 7-10). Although we expect more disturbed stations
to have more dead coral, results from the south side indicate that stations further from the
disturbed area had more coral overall, both live and dead. Thus, less disturbed sites had
more dead coral (see parenthetic value in Table 4 for Dead SA).
19
-------
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix)
Table 4. Correlation of candidate coral metrics with disturbance and depth.
Name of candidate coral metric, its correlation with distance from the center of human
disturbance and correlation with depth below the surface for three gradients on the south, west,
and north sides of the island (Spearman's r, only correlation values >0.4 or <-0.4 are shown).
Correlation was calculated separately for distance and depth in each of the three areas. Metrics
used in subsequent analyses are noted by an "*". Correlation for dead SA for the south gradient is
noted parenthetically because correlation was in the opposite direction predicted.
Candidate metric
N =
Abundance & Composition
Number of colonies *
Total number of taxa *
% "Rare" colonies
% SA of "rare" taxa
Physical stature
Total SA *
Average radius of all colonies
Average SA of all colonies *
Biological condition
Average % live tissue *
Live SA *
Dead SA
% Live SA *
% Hermaph. colonies
% SA of hermaph. taxa
% Gonochoristic colonies
% SA of gonochoristictaxa
% Brooder colonies
% SA of brooder taxa
% Spawner colonies
% SA of spawnertaxa
Community structure
% SA Diploria
% SA Montastraea
% SA Porites
% SA Siderastrea
Percent dominance
Distance
(South)
19
0.53
0.54
0.52
0.79
0.67
0.66
0.66
(0.78)
0.45
Depth Distance Depth Distance
(ft) (West) (ft) (North)
19 12 12 16
0.66
0.69
0.50
0.77
0.75
-0.47
0.41
0.58
-0.46 -0.59
-0.54 0.53
-0.69 0.63 -0.46
-0.59
0.54 -0.73
0.72
0.47
-0.71
0.45 -0.84 -0.43
0.70
-0.54 0.42
-0.79
0.51
Depth
(ft)
16
0.51
-0.44
-0.54
-0.59
0.53
20
-------
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix)
In the West CMZ, metrics were much more strongly associated with depth than with
distance from the public dock even though station depth was relatively narrow (20-41 ft.).
The deeper stations were nearest the dock while more distant stations were in more
shallow water (r = -0.59). Many candidate coral metrics related to physical stature,
biological condition and community structure were strongly associated with depth. The
few metrics that were associated with distance from the dock were also highly correlated
with depth.
For the gradient in the North CMZ, distance from Christiansted was only associated with
two coral metrics, one of which (SA ofDiploria) was also correlated with depth. A few
additional metrics were also associated with depth.
Several candidate metrics were highly correlated with both distance from human
disturbance and depth. For example, percent rare colonies and percent SA of rare
colonies might be correlated with human disturbance if the influence of depth were
controlled. Other candidate metrics such as % SA ofDiploria were less promising given
its opposite response to disturbance for gradients on the north and west sides.
Seven coral metrics were selected for additional analysis. Four metrics were selected
because they were correlated with distance from the docks on the south side but not with
depth: # of taxa, total SA, average SA, and live SA. Number of colonies was lower near
the center of disturbance on the south side, although not significantly correlated with
distance, and was also selected. Average radius of all colonies was not selected because it
was highly correlated with average SA. Two other candidate metrics, although not
correlated with a gradient of human disturbance, were selected for additional analysis.
These two metrics were % live tissue (colony average) and percent live SA. We were
interested in these two candidate metrics for their potential as indicators of change within
a location compared to itself over time. These two metrics were included in the power
analysis to determine if they were simply too variable to reliably detect change in coral
condition.
21
-------
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix)
c
0
<3
=1*
t J
40
35
30
25
20
15
10
5
o
f
o >
<
ฐ 0
5
O a
w 99
O
0
OOOOOOOOOOO
2oooo ooooo
2oooo ooooo
ฐ CO CO -
-------
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix)
0)
95
90
85
80
75
70
65
60
55
c
c
c
c
T-
m
O >
0
O
o
0
0
= 20 ft
20 ft
,
>OOOOOOOOOOOOC
'oooo oooooooc
> o o o o oooooooc
> OO CO -
-------
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix)
s
CO
k_
CO
0)
o
CO
t
I
-3
o
3
2
2
1
1
0
c
0
5
0
5
0
5
n
ฐc
c
c
c
T-
0 :
:
;
O
o * :
o :
o :
.; s
'ooooooooooooc
Joooo oooooooc
Joooo oooooooc
'OOCD-^J-CN CN'^J-CDOOOCN-^J-'C
1111 1- 1- 1- 1-
Distance from dock(m)
r=0.66
CO
ฃ
CO
fl)
cc
w
0)
'3
\ n
1 . U
0.9
0.8
0.7
0.6
0.5
0.4
0.3
n o
0.2c
c
c
c
ฐ >
f
*
0^
.
*
O
Joooooooooo
Joooo ooooo
Joooo ooooo
'oOCD-'tCN (N-
-------
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix)
x-~ ซs
^
CD
CD
0
O
CD
t
W
0
D)
CD
0
/ uuu
6000
5000
4000
3000
2000
1000
0
In n n
u u u
c
c
c
c
T-
o
O
9
0 O
* *
^OOQOOOOOO
^OOQO 0000
^0000 0000
oOCD-^-CN CN-^TCDOO
9
ซ
O
0 >= 20 ft
< 20 ft
0 0 0 C
0 0 0 C
o o o c
O CN -^ CC
Distance from dock(m)
r=0.66
Figure 10. Average surface area (averaged for all colonies) plotted against distance from the
commercial dock on the south side of St. Croix. Negative distances were locations west of the
dock, positive distances were east of the dock. Correlation coefficients were calculated for the
absolute value of the distance from the dock (n = 19 stations; Spearman's /"-values from Table 4).
Correlation coefficients < 0.4 were not considered biologically significant.
Natural variability associated with habitat differences
Although the coral assemblages differed visibly from one geographic area to another (see
"Biological comparison of CMZs" below), distinct habitat types within a CMZ were
sometimes difficult to identify in the field. In the South and Southwest CMZs, back reef,
shallow fore reef, and deep fore reef habitat were observed. Fore reef along the most
seaward side was too deep (>60 ft.) to sample in these CMZs. In the North, Northwest
and West CMZs, most of the coral was found along a reef habitat that sloped away from
shore. Back reef habitat was evident in the West, Northwest, and North CMZs, but had
very little coral present. The East CMZ may include multiple habitat types and additional
sampling is needed to characterize this CMZ. The Buck Island CMZ had both back reef
and shallow fore reef habitat.
Differences in coral assemblages associated with depth were more obvious. For example,
Diploria spp. declined in relative abundance with depth while Montastraea spp.
increased. The number of colonies, the number of taxa, and total SA all increased with
depth across the 61 stations sampled. Percent live tissue (averaged across colonies) and
percent live SA both decreased with depth.
25
-------
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix)
Biological comparison of CMZs
Stations in the Buck Island and East CMZs varied in depth from -5-40 ft. The deepest
stations were located on the north and west sides of the island. The shallowest stations
were along the south side. Differences associated with depth may be driving some of the
differences observed in the candidate coral metrics across zones (see Table 4 above;
Figure 11). The number of colonies recorded was also a very strong predictor of the
number of taxa that were found (Figures 11, lower panel and Figure 12, upper panel). For
two of the least disturbed zones, Buck Island CMZ had a low number of colonies per l/2-
transect and fewer taxa; in contrast, the Northwest CMZ had more colonies and more
taxa compared to other zones. Total SA and live SA showed similar patterns across the
zones with lowest cover in the South and Southwest CMZs and higher cover in the Buck
Island, North, Northwest and West CMZs (Figure 13). The largest colonies as measured
by the average SA of individual colonies were found in the Buck Island and East CMZs
(Figure 14, lower panel). The smallest colonies were found in the Northwest and South
CMZs. The pattern for percent live SA was the opposite, with more live surface area
observed in zones with smaller colonies (Figure 14, upper panel).
Percent live tissue (colony average; see Figure 12, lower panel) and percent live SA (see
Figure 14, upper panel) also showed similar patterns across zones. Highest values for
both metrics were observed in the South and Southwest CMZs where disturbance was
highest. High values for live tissue in the most disturbed locations were the opposite of
original predictions. Corals in these areas were small and sparse, but exhibited high live
tissue coverage. Lower values for live tissue were observed in the Buck Island and West
CMZs.
26
-------
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix)
&
45
40
35
30
25
20
15
10
5
0
140
120
100
co 80
I 60
O
* 40
20
0
-20
T
T3 T3
JO LJJ
CO -*'
CO
J^ CO
O LU
3
DQ
CO
o
CO
0
CO
* Median
I I 25%-75%
| Non-Outlier Range
o Outliers
^ Extremes
Zone
T3 T3
C C
_CO LLJ
C/) *->
co
m co
O LLJ
m
CO
cu
CO
cu
o
CO
T3
m
to
0)
Median
I I 25%-75%
T" Non-Outlier Range
o Outliers
* Extremes
Zone
Figure 11. Depth and number of colonies for each CMZ. See Table 1 for
number of stations in each CMZ.
27
-------
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix)
CD
X
tt
ฃU
18
16
14
12
10
8
6
4
2
0
^.T
I
iฅ
T3 T3
j| LJJ
"w
.*: CD
O IJJ
T
T.
^P ^^^^n
I*J
I go
Z jl ฐ jl *-
0 | |
DO
Median
I | 25%-75%
I Non-Outlier Range
ฐ Outliers
%r Extremes
Zone
1 UU
90
^ 80
w
-ง-
m 70
d) ' ซ
i
^ 60
50
4n
-r ฐI
X
""ff""
pi
T"
4U T3 T3
C C
JD LJJ
"w
^ CD
0 LU
TH
i-i i ^
i
L-jJ X *
-
O
-c to J=
t CD 1ฑ
-5^0
^ ฃ CO
o
z
ป -y .
r-^-i
IN
w "2
CD c
S LU
ฃ "w
3 0)
ฐ 5
rn -^
DQ
Median
I | 25%-75%
I Non-Outlier Range
o Outliers
^ Extremes
Zone
Figure 12. Number of taxa and percent live tissue (averaged for all colonies) for
each CMZ. See Table 1 for number of stations in each CMZ.
28
-------
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix)
JU
45
40
35
30
< 25
o 15
10
5
0
-5
1
1
o
T :
o X,
ฎy *5y
"^ "o .c "GO .c "w "^
J5 LU ? s> ^ s> LU
"w Z 5 W 5 ^
DO
Zone
Zone
ป Median
I | 25%-75%
| Non-Outlier Range
o Outliers
^ Extremes
1 U
16
14
_ 12
^E 10
< g
0) c;
1
4
T1
2
0
o
o
X .
T -r \\
.
~T~ * -J-
f ^^^ nmJLiiiu
"*" 5X is i,
ป
~n -7-ป ^ -' ~n
^ T3 ^ w ฃ W C
_TO LJJ o ง ^ ง LU
(/) ^ ^ O , .
(/) i CO ฃ w
1s S 11
00
Median
25%-75%
Non-Outlier Range
Outliers
Extremes
Figure 13. Total surface area and live surface area for each CMZ. See Table 1 for
number of stations in each CMZ.
29
-------
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix)
1 n
I . U
0.9
0.8
< ฐ7
W 0.6
-------
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix)
Power analysis
Although transect location within a station contributed a considerable amount to the
overall variance of several coral metrics, the difference in precision associated with
metrics calculated from a single ^-transect vs. metrics calculated from the average of two
/^-transects (full transects) was minimal. For the 16 stations in 3 CMZs with full transect
surveys, several of the seven coral metrics differed in average values by zone. The
ANOVA model used to estimate variance compensated for these zonal differences and
provided an estimate of the variance associated with different stations within a zone.
For a single /^-transect sampled at each station, if five stations in a CMZ are surveyed,
the number of colonies would have to decline by 17 or more to represent a statistically
significant change (p = 0.1 for a one-sided ^test; Table 5). For 10 stations, a smaller
decline in the mean number of colonies (12 colonies) would be significant. Similarly for
all metrics, an increase in the number of stations surveyed corresponded to an increase in
sensitivity to detect a change in coral condition (and a smaller MDD). When more
information was used to calculate metrics from each of the 16 stations by averaging
metric values for both /^-transects, the MDD changed little. In a few cases the MDD
increased and for a few cases it declined (see parenthetic values in Table 5). For cases in
which power increased for smaller sample sizes, the differences were small and can be
attributed to noise in the data and should not be interpreted as meaningful differences. All
the values for MDD were close for both the analyses (using one or two ^-transects).
Lack of difference in MDD values for the different metrics supports the idea that no gain
in precision was associated with full vs. /^-transects.
Table 5. MDD values for candidate coral metrics.
Name of candidate metrics, mean value for 16 stations (n = 3 Buck Island, n = 6 East CMZ, n = 7
West CMZ), mean squared error from a 1-way ANOVA (with CMZ as the 1 factor), and minimum
detectable differences (MDD) for 5, 10 and 15 stations. The first values shown are the MDD
values based on one 1/2-transect at each station; in parentheses are the MDD values for the
average of two 1/2-transects for each station.
Candidate metric
# Colonies
#Taxa
% Live tissue (colony avg.)
Total SA (m2)
Live SA (m2)
% Live SA
Average SA
(cm2; colonies)
Mean
38.0
7.5
69.4
8.0
4.4
0.55
2,766
MSE
96.99
2.08
51.25
5.02
5.11
0.02
3.7 x106
MDD_5
17(20)
3(3)
13(12)
4(4)
4(3)
0.2 (0.2)
3,429 (2,907)
MDD_10
12(13)
2(2)
9(8)
3(3)
3(2)
0.2 (0.2)
2,303(1,953)
MDD_15
9(11)
1(2)
7(6)
2(2)
2(2)
0.1 (0.1)
1,852(1,570)
The relative sensitivity of each coral metric was compared by converting the MDD to a
percentage of the mean. For this comparison, average SA was the least sensitive and
percent live tissue could detect the smallest change relative to its mean (Table 6). For
31
-------
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix)
total SA, the metric most highly correlated with the disturbance gradient on the south
side, a 27% decline would be detectable if 15 stations were sampled each time.
Table 6. Detectable change as a percent for candidate coral metrics.
Candidate metric name, mean value for 16 stations (n = 3 Buck Island, n = 6 East CMZ, n = 7
West CMZ), and percent change that would be significantly different for 5, 10, and 15 stations.
Percentage calculated as MDD/Mean * 100%.
Candidate metric Mean(N=16)
# Colonies
#Taxa
% Live tissue (colony avg.)
Total SA (m2)
Live SA (m2)
% Live SA
Average SA (cm2;colonies)
38.0
7.5
69.4
8.0
4.4
0.55
2,766
MDD_5 MDD_10 MDD_15
46%
34%
18%
49%
92%
45%
124%
31%
23%
12%
33%
62%
30%
83%
25%
18%
10%
27%
49%
24%
67%
Population structure
The seven most abundant species differed in terms of their distribution around St. Croix.
More colonies ofDiploria strigosa were found in the North CMZ around Christiansted
(Figure 15). For all three species of Montastraea, the greatest number of colonies was
found in the West CMZ (Figure 16). Porites astreoides was much more common that P.
porites in all zones and most abundant in the West CMZ while P. porites was more
commonly found in the Buck Island and East CMZs (Figure 17). Siderastrea siderea was
least common in the Northwest and Buck Island CMZs, the two zones with the least
amount of human influence (Figure 18). For most species at most sites, a proportionately
larger number of colonies were found in the smaller size classes, indicating a larger
number of younger colonies. Montastraea faveolata and Siderastrea siderea showed
some tendency to have a more even spread of colonies across the size classes. For four
species, the most even distribution of colonies across class sizes was found in the Buck
Island CMZ. For/), strigosa, M. annularis, P. astreoides., and P. porites the greatest
proportion of large colonies was found in the Buck Island CMZ.
32
-------
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix)
80
60
40
20
0
80
60
40
20
0
Surface area of Diploria strigosa (m2)
Sin o in o
p <-
-------
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix)
80
60
40
20
0
Surface area of Montastraea cavernosa (m2)
8 S ?
00000
Zone: West End
-------
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix)
120
80
40
0
Surface area of Porites astreoides (m2)
Zone: Northwest
CO CO O
q q q q t-
00000
Zone: Buck Island
CN ^ co co o
q q q q <-
o o o o o
Zone: North
8 ง 8 8 ฐ
o o o o o
Zone: South
8 S S S ฐ
o o o o o
Zone: East End
Surface area of Porites porites (m2)
20
16
12
8
4
0
(/)
ง20
O 16
O 12
o 8
O ^
!_ 0
0
.Q
E
20
16
12
8
4
0
S ง 8 S S
o o o o o
Zone: V\test End
CO CO CN
q q q t- co
o o o o o
Zone: Northwest
S S 8
o o o o o
Zone: Buck Island
S-^- co CD CN
q q i- co
o o o o o
Zone: Southwest
CO CD CN
q q q t- co
o o o o o
Zone: North
S ง 8 S S
o o o o o
Zone: South
CO CD CN
q q q t- co
o o o o o
Zone: East End
Figure 17. Number of colonies by size class for Porites astreoides and P. porites. Shown are
number of colonies by size for each CMZ. Note that size classes differ by species.
35
-------
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix)
0
'c
"o
o
M
o
i_
0
E
30
20
10
0
30
20
10
0
Surface area of Siderastrea siderea (m2)
8 8 ? 8
o o o o o
Zone: West End
8 8 ? 8
o o o o o
Zone: Northwest
8 8 ? 8
o o o o o
Zone: Buck Island
8 8 ? 8
o o o o o
Zone: Southwest
8 8 ? 8
o o o o o
Zone: South
8 8 ? 8
o o o o o
Zone: North
8 8 ? 8
o o o o o
Zone: East End
Figure 18. Number of colonies by size class for Siderastrea siderea. Shown are number of
colonies by size for each CMZ. Note that size classes differ by species.
36
-------
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix)
DISCUSSION
The focus of this study was to identify meaningful biological indicators of reef condition
and to determine efficient data collection procedures to measure them. Once field tested,
these indicators can be used in a probabilistic survey design to estimate and monitor reef
condition throughout USVI (Fore et al., 2006b). A good bioindicator must be strongly
and consistently correlated with independent measures of human disturbance in a variety
of contexts, have a plausible biological connection to human-induced changes in the
environment, and have adequate precision to detect a change in resource condition should
a change occur (Yoder and Rankin, 1998; Fore, 2003). A good indicator should also be
relatively immune to differences associated with natural conditions such as depth,
seasonality, or annual variability. Biological measures that satisfy these criteria are
referred to as "metrics" in the biomonitoring literature (Karr and Chu, 1999). For stony
corals, metrics tested here were referred to as "candidate metrics" because they have yet
to satisfy all the above criteria. Candidate metrics in four categories were tested against a
gradient of human disturbance in three different areas surrounding St. Croix. A subset of
those metrics was then evaluated for their statistical precision and ability to detect change
over time.
Bioassessment is a relatively new endeavor for coral reef communities compared to
rivers, streams, lakes, wetlands, and estuaries (Barbour et al., 1999). Thus, a short list of
the best indicators, or metrics, has not been developed. Candidate metrics for stony corals
could be loosely divided into three categories. The first category would include coral
metrics associated with human disturbance in other locations that were also found to be
associated with disturbance for this study. An example of this would be coral cover
which can be measured in a variety of ways and typically declines with disturbance
(Jameson et al., 2001; Sealey, 2004). The second category includes metrics derived from
ecological theory and tested for this study only. Metrics may or may not have been
associated with disturbance in St. Croix, but may merit testing again in new locations.
Examples include metrics related to reproductive strategy such gonochoristic vs.
hermaphroditic development or brooder vs. spawner. The third category includes metrics
that could be derived from exploratory analysis of the current data set. For example, the
percentage of colonies observed in the largest size classes could be developed as an
indicator of stable environmental conditions that are adequate to support coral colonies
over a long time period. Smith et al. (2005) found thatAcropora assemblages had more
adults and juveniles in reef areas with less sediment influence. Similarly, presence and
absence of taxa at more disturbed and less disturbed sites could be compared to develop
sensitive and tolerant taxa lists for stony corals. Because the ideas behind these types of
metrics were derived from the current data set, hypothesis testing was not appropriate for
the current study, but awaits data from a new location.
Efficacy of field protocols
Field assessment of stony corals is intensive because survey divers must operate
underwater using SCUBA which requires a minimum of two divers, a boat, and a boat
37
-------
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix)
captain. Thus, if the same amount of information on biological condition is provided by a
25-m2 area as from a 50-m2 area, the smaller area would be preferred because it translates
into less field time. The smaller the area surveyed at each station, the quicker the dive
team can get to the next station, and the more stations will be visited during a field
season. From the statistical perspective, measurements from more locations are preferred
because as the number of survey locations (N) increases, so does the precision of the test
(Larsen, 1997). Thus, smaller survey areas are better from both logistical and statistical
viewpoints. The goal is to minimize the amount of effort required to get a reliable, and
repeatable, estimate of reef condition at each station.
This study evaluated nuisance variance at two different levels. Duplicate surveys by
different dive teams at the same reef station provided data to test the reliability of the
field protocol used to quantify coral condition. The two halves of the radial transect
provided data to assess the variability associated with microhabitat differences at a reef
and test whether a half or full transect provides a more precise measure of coral
condition. The analysis was complicated by the fact that at this point we do not know
which coral metrics will be the best bioindicators for tracking coral condition through
time. The seven metrics evaluated for their precision included four that were correlated
with human disturbance on the south side of St. Croix and three others related to density
and the percentage of live tissue observed.
The amount of variance associated with different dive teams was quite small for all seven
candidate metrics compared to variance due to transect location, station, and zone. Teams
of EPA divers with extensive experience using the field protocol in Florida recorded very
similar values as teams of US VI divers new to the EPA protocol for the same reef
stations. Although new to the EPA data collection protocol, USVI divers were very
experienced in terms of local conditions, coral identification, and underwater data
collection. The EPA field protocol used here was easily implemented by divers with
scientific knowledge of stony corals.
The density of coral colonies was much higher in St. Croix compared to reef stations
sampled in Florida (Fore et al., 2006a). High coral density necessitated a smaller survey
area at many reef stations in order to complete data collection during a reasonable time
period (i.e., one tank of air). Data from full transects were divided to provide replicate
samples from the same reef area in the form of two ^-transects. Variance component
analysis indicated a relatively high percentage of the overall metric variability was
associated with the two different halves of the radial belt transect. Differences were likely
due to microhabitat differences, such as placement of the transect in an area with a patch
of sand, a spur which creates more surface area, or a large coral colony.
Statistical power analysis compared the ability of the seven candidate metrics to detect
changes in a CMZ though time using a simple statistical test (the two-sample t test).
Metric precision was compared for two sampling scenarios, the first used a single 1/2-
transect from each station and the second used the average of two ^-transects for each
station. Although the second scenario provided twice as much information about coral
condition at a station, the increase in statistical power was minimal and inconsistent. In
38
-------
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix)
other words, the full transect did not increase our ability to detect change in these metrics
over the ^-transect. The percentage of metric variability associated with transect
differences was somewhat high, but station differences were greater. Thus, increasing the
precision at each station did not improve our ability to detect changes in stations because
station differences were greater than transect differences.
For a monitoring design that surveyed 15 stations within a CMZ either in two different
areas or on two different occasions, > 27% in total SA, > 18% decline in taxa richness
(-1.5 coral species), or > 67% decline in average colony SA would represent a
statistically significant change at the 90% confidence level. More reef stations in the
sampling design would yield greater precision to detect smaller changes. This level of
sampling effort (15 stations) provides a reasonable level of sensitivity for coral reef
protection. Note that if the exact same locations were sampled, the correct statistical test
would be a paired test rather than a two-sample test and a much smaller change in
condition could likely be detected because site differences would be eliminated by
comparing each site to itself. If only a few stations can be monitored each year (e.g.,
< 10), a paired test would be a better design for detecting change through time because a
much larger change would have to occur to be statistically significant when sample sizes
are small. For a regional comparison of stations in different areas, a paired design would
not be possible and the two-sample t test would be one example of an appropriate test.For
resource monitoring, the power to detect a change must be explicitly considered in order
to be protective of natural resources. In the past, monitoring designs have too often
ignored the ability of a sampling design to detect a change should it occur and instead
have focused on protecting against making a Type I error, that is, concluding that a
change has occurred when it has not (Dayton, 1998). In the context of resource
protection, many authors agree that the probability of making either a Type I (false
positive) or a Type II (false negative) error should be equal (Peterman, 1990; Steidl et al.,
1997; Yoccoz et al., 2001). The two types of error are mathematically related and should
be considered together. Although a 5% Type I error rate is more typically associated with
hypothesis testing, for power analysis a value of 5% for both types of error can be too
restrictive. For this study, we followed recommended guidelines and balanced both types
of error at 10% for power analysis calculations.
Coral response to human disturbance
Reef stations were selected to follow potential gradients of human disturbance.
Differences along the south side associated with commercial and industrial land use
centered at the docks was reflected in dramatic differences in coral condition. Human
disturbance was intense and associated with multiple land uses. Many of the disturbances
have also occurred over a long time period, e.g., the rum distillery has discharged effluent
to the near-shore environment for -80 years. The coral metrics quantified the changes in
the amount of coral surface cover, average size of coral colonies, and the number of coral
colonies. The number of coral species also increased for stations located further from the
dock; however, this could either be due to loss of taxa due to human influence or a
spurious correlation with the number of colonies found. The total number of taxa was
highly correlated with the number of colonies measured across all stations. A comparison
39
-------
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix)
of reefs near developed and undeveloped areas in the Bahamas found that taxa richness of
coral increased for developed areas, the opposite of what was found for St. Croix. For the
same study coral surface area declined with disturbance, which agreed with our results
(Sealey, 2004). A much larger scale study conducted across 135 reefs in the Great Barrier
Reef found a decline in species richness associated with agricultural run-off (DeVantier
et al., 2006). In contrast, a smaller scale study conducted in the area of agricultural run-
off failed to detect a difference in species richness for stony coral, although a decline in
both hard coral cover and richness of soft corals was observed for sites with greater
agricultural run-off (Fabricius and De'ath, 2004). Species richness of hard corals was also
shown to decline in Barbados across a eutrophication gradient as measured by water
quality samples (Tomascik and Sander, 1987).
For the other two gradients, distance from the Christiansted harbor and from the public
dock on the west side failed to correlate with the same metrics. Several candidate metrics
were correlated with distance in the West CMZ, but were also correlated, and typically
more strongly correlated, with depth. The deepest sites in the West CMZ were nearest the
dock and moving away from the dock the sampling stations were in more shallow water.
At the time of the survey, the 20 ft. difference in depth across these stations was not
considered as a potentially confounding factor because it represented such a narrow
range. Metrics calculated from samples collected at the same depth might show a more
direct correlation with disturbance. Based on the current analysis, the sensitivity of many
of the coral metrics to depth may represent an important consideration when designing
monitoring programs. Nonetheless, lack of consistent correlation between candidate coral
metrics and disturbance gradients in the North and West CMZs was not surprising
because the coral communities at all locations appeared healthy.
Several reasons could explain the lack of correlation between many candidate coral
metrics and the gradients of human disturbance. First, the human disturbances in the
North CMZ (Christiansted) and West CMZ (Frederiksted) could have been too small to
cause differences in the coral assemblage; there may have been no gradient to detect. In
contrast, the high intensity of human disturbance along the south side may have
overwhelmed more subtle measures of coral condition, such as those related to
reproductive strategy. It is also possible that we have yet to discover the coral metrics that
will quantify subtle changes in coral condition. An alternative explanation is that the
scale of disturbance may be larger than our survey design can detect. Given the
potentially complex mixing patterns associated with ocean currents, we do not know the
spatial scale at which corals respond to disturbance. For example, Edmunds (2002)
compared coral cover at two long-term monitoring stations (12 years). Both sites were
located in a marine park in St. John, USVI with minimal human disturbance, but one site
showed a dramatic increase in coral cover while the other declined just as dramatically.
Random sampling around these long-term stations suggested that coral change may be
occurring at larger spatial scales and over longer time scales than can be captured by a
typical survey.
For our study, rather than comparing stations along the northern gradient according to
their distance away from Christiansted harbor, stations should perhaps be compared with
40
-------
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix)
more distant pristine areas. The original sampling design was developed with this
approach in mind. Stations located around Buck Island were selected as potential
reference sites for the larger island of St. Croix. Unfortunately, the differences between
Buck Island stations and other stations around St. Croix were large enough in terms of
both summary metric values and species composition, that natural differences, such as
depth, and differences associated with human influence, such as proximity to a developed
area, could not be distinguished.
Another somewhat surprising result was the high positive correlation between the amount
of surface area that was live and the amount that was dead. High values of both live and
dead surface area were observed together. The original prediction was that live coral
would be replaced by dead coral as human disturbance increased. A study in the Great
Barrier Reef found a higher percentage of dead coral in more disturbed areas (Fabricius
and De'ath, 2004). Oddly, the most disturbed areas on the south side of St. Croix had the
highest values for percent live surface area and among the highest for percent live tissue
(averaged across all colonies). This result supports the value of calculating surface area.
At the disturbed areas along the south shore there were a few, small colonies with high
tissue survival; thus, high tissue survival alone would have led to an erroneous conclusion
regarding the health of the coral community. This result further suggests that the presence
of dead coral depends on live tissue and that once a colony dies, erosion may happen very
quickly, eliminating dead coral from the reef. Dead coral was only recorded if it could be
identified to at least the genus level. Physical structure that could not be identified to
genus was also difficult to identify as coral rather than rock substrate. In addition, coral
rubble was also difficult to distinguish from rock rubble and, therefore, was not recorded
as dead coral. The inability to account for all dead coral must be considered in any
interpretation of these results.
Influence of habitat on stony coral assemblages
The original sampling design for this study was to survey coral stations across a gradient
of human disturbance in the different habitat types found in that area. Although NOAA
maps indicated different types of benthic habitat within survey areas, in many cases the
differences were not visibly obvious (Kendall, 2001; NOAA, 2001). For example, spur
and groove, linear reef and colonized hard-bottom were difficult to distinguish in the
North CMZ. Larger differences between fore and back reef habitat types were easier to
recognize, but in many cases, e.g., the fore reef in the South, Southwest, and West CMZs,
the reef was too deep (>50 ft.) for efficient sampling. In the North CMZ, there were too
few corals in back reef areas to survey.
The high correlation between depth and many of the candidate coral metrics was
somewhat surprising, particularly in the West CMZ where the difference was only 20 ft.
across all stations. Given the additional differences associated with amount of coral
cover, average colony size, and species composition observed for different CMZs around
St. Croix, comparisons within similar geographic areas may be the best approach to
monitoring. Yet, even within a particular CMZ, the influence of depth should be carefully
considered.
41
-------
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix)
CONCLUSIONS
The EPA protocol for stony corals records three observations for each coral colony
within a transect: species, size, and tissue condition. Four coral metrics derived from
these measures were highly correlated with proximity to disturbance along the south side
of St. Croix. Candidate metrics related to taxa richness, total coral surface area, live
surface area, and average colony size all declined for stations closest to the disturbed
area. A similar response was not seen for coral metrics and the proposed gradients of
human disturbance in the West and North CMZs. The lack of correlation in these areas
was not too surprising because the corals at these stations all appeared healthy and
diverse. More subtle changes in coral condition may perhaps be present, but the current
candidate metrics did not detect them. Ben-Tzvi et al. (2004) failed to find differences in
taxa richness when comparing reefs with different levels of human disturbance but they
did document more subtle differences associated with coral recruitment and mortality.
The influence of depth on many of the candidate coral metrics was much greater than
anticipated. Although differences according to habitat type were expected, such strong
associations with metric values over only a 20 ft. range of depth were noteworthy. Depth
may covary with ocean currents or water chemistries that also influence stony corals.
Although the best coral metrics to use as biological indicators of reef condition are not
yet certain, the candidate metrics that correlated with human disturbance along the south
side of the island had adequate statistical precision to detect a level of change that would
provide reasonable protection of coral reef resources and could be used to monitor for
change in coral condition. Data from this type of field survey could also be used to test
additional candidate coral metrics not considered here but that may be more sensitive to
smaller changes in coral condition.
Four results from this study support the use of the EPA field protocol for the assessment
of stony corals. First, four candidate metrics were highly correlated with distance from a
known area of disturbance. Second, divers new to the method obtained nearly identical
results as experienced divers for the same survey area. Third, the metrics derived from
the field protocol were sensitive enough to document coral loss. Fourth, the method is
efficient because an area of only 25 m2 was needed to characterize a reef station.
Although additional testing in other coral reef areas both in the US VI and in other
geographic areas remains, the results for St. Croix provide a solid start toward the
development of biocriteria for the protection of US VI' s coral resources.
42
-------
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix)
REFERENCES
Barbour M.T., J. Gerritsen, B. D. Snyder, and J. B. Stribling. 1999. Rapid bioassessment
protocols for use in streams and wadeable rivers: Periphyton, benthic
macroinvertebrates, and fish. 2nd edition. EPA 841-B-99-002. US Environmental
Protection Agency, Office of Water, Washington, D.C.
Barker, N. H. L., and C. M. Roberts. 2004. Scuba diver behavior and the management of
diving impacts on coral reefs. Biological Conservation 120: 481-489.
Ben-Tzvi, O., Loya, Y., and Abelson, A. 2004. Deterioration Index (DI): a suggested
criterion for assessing the health of coral communities. Marine Pollution Bulletin
48: 954-960
Dayton, P. K. 1998. Reversal of the burden of proof in fisheries management. Science
279:821-822.
DeVantier, L. M., G. De'ath, E. Turak, T. J. Done, and K. E. Fabricius. 2006. Species
richness and community structure of reef-building corals on the nearshore Great
Barrier Reef. Coral Reefs 25: 329-340.
Edmunds, P. J. 2002. Long-term dynamics of coral reefs in St. John, US Virgin Islands.
Coral Reefs 21: 357-367.
Fabricius, K. E. and G. De'ath. 2004. Identifying ecological change and its causes: a case
study on coral reefs. Ecological Applications 14:1448-1465.
Fisher, W. S., W. P. Davis, R. L. Quarles, J. Patrick J. G. Campbell, P. S. Harris, B. L.
Hemmer M. Parsons. 2007. Characterizing coral condition using estimates of three-
dimensional colony surface area. Environmental Monitoring and Assessment 125:
347-360.
Fisher, W. S. In press. Stony coral rapid bioassessment protocol. Office of Research and
Development, NHEERL, Gulf Ecology Division, US EPA, Gulf Breeze, FL.
Fore, L. S. 2003. Developing biological indicators: Lessons learned from mid- Atlantic
streams. EPA 903/R-003/003. Ft. Meade, Maryland: Office of Environmental
Information and Mid- Atlantic Integrated Assessment Program, Region 3.
Available:
Fore, L. S., W. S. Fisher, and W. S. Davis. 2006a. Bioassessment tools for stony corals:
statistical evaluation of candidate metrics in the Florida Keys. EPA-260-R-06-002.
USEPA Office of Environmental Information, Washington. August 2006.
Fore, L. S., W. S. Fisher, and W. S. Davis. 2006b. Bioassessment tools for stony corals:
monitoring approaches and proposed sampling plan for the U.S. Virgin Islands.
EPA-260-R-06-003. USEPA Office of Environmental Information, Washington.
August 2006.
Hutchins, A. 2004. An Implementation Plan for Bio-Indicator Monitoring in Support of
Integrated Coastal Management in the U.S. Virgin Islands. M.S. thesis, Duke
University.
Jameson, S. C., M. V. Erdmann, G. R. Gibson Jr., K. W. Potts. 1998. Development of
biological criteria for coral reef ecosystem assessment. Atoll Research Bulletin,
September 1998, No. 450, Smithsonian Institution, Washington, DC, 102 pp.
Jameson, S. C., M. V. Erdmann, J. R. Karr, K. W. Potts. 2001. Charting a course toward
diagnostic monitoring: A continuing review of coral reef attributes and a research
strategy for creating coral reef indexes of biotic integrity. Bulletin of Marine
Science 69(2):701-744.
43
-------
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix)
Karr, J. R., and E. W. Chu. 1999. Restoring Life in Running Waters: Better Biological
Monitoring. Washington, DC: Island Press.
Karr, J. R., and C. O. Yoder. 2004. Biological assessment and criteria improve total
maximum daily load decision making. Journal of Environmental Engineering
130:594-604.
Kendall, M.S., C.R. Kruer, K.R. Buja, J.D. Christensen, M.Finkbeiner, and M.E.
Monaco. 2001. Methods Used to Map the Benthic Habitats of Puerto Rico and the
U.S. Virgin Islands. Silver Spring, MD: National Oceanic and Atmospheric
Administration, NCCOS Biogeography Program.
http://biogeo.nos.noaa.gov/projects/mapping/caribbean/startiip.htm
Larsen, D.P. 1997. Sample survey design issues for bioassessment of inland aquatic
ecosystems. Human and Ecological Risk Assessment 3: 979-991.
National Research Council (NRC). 2001. Assessing the TMDL approach to water quality
management. Water Science and Technology Board, Division on Earth and Life
Studies. National Academy Press, Washington, DC.
Nemeth, R.S., S. Herzlieb, E.S. Kadison, M. Taylor, P. Rothenberger, S. Herold, and W.
Toller. 2004. Coral reef monitoring in St. Croix and St. Thomas, United States Virgin
Islands. Year three final report submitted to Department of Planning and Natural
Resources. Center for Marine and Environmental Studies, University of the Virgin
Islands and Division of Fish and Wildlife, DPNR.
Nemeth, R. S., and J. S. Nowlis. 2001. Monitoring the effects of land development on the
near-shore reef environment of St. Thomas, US VI. Bulletin of Marine Science 69:
759-775.
Peterman, R. M. 1990. Statistical power analysis can improve fisheries research and
management. Canadian Journal of Fisheries and Aquatic Sciences 47: 2-15.
Ramos-Scharron, C. E. and L. H. MacDonald. 2005. Measurement and prediction of
sediment production from unpaved roads, St. John, US Virgin Islands. Earth
Surface Processes and Landforms. 30:1283-1304.
Reznick, D., M. J. Bryant, and F. Bashe. 2002. r- and K-selection revisited: the role of
population regulation in life-history evolution. Ecology 83: 1509-1520.
Richmond, R. H. and C. L. Hunter. 1990. Reproduction and recruitment of corals:
comparisons among the Caribbean, the Tropical Pacific, and the Red Sea. Marine
Ecology Series 60: 185-203.
Sealey K. S. 2004. Large-scale ecological impacts of development on tropical islands
systems: comparison of developed and undeveloped islands in the central Bahamas.
Bulletin of Marine Science 75(2): 295-320.
Smith, L. D., M. Devlin, D. Haynes, J. P. Gilmour. 2005. A demographic approach to
monitoring the health of coral reefs. Marine Pollution Bulletin 51: 399-407.
Steidl, R. J., J. P. Hayes, and E. Schauber. 1997. Statistical power analysis in wildlife
research. Journal of Wildlife Management 61:270-279.
Toller, W. 2005. F7 Interim Completion Report - October 1, 2003 to September 30 2005,
Part 2. Preliminary results from a study of anchor damage to the Frederiksted Reef
System: impacts to scleractinian corals and reef fish communities. Division of Fish
and Wildlife, Department of Planning and Natural Resources, Government of the
U.S. Virgin Islands.
44
-------
Bioassessment Tools for Stony Corals:
Field Testing of Monitoring Protocols in the US Virgin Islands (St. Croix)
Tomascki, R. and F. Sander. 1987. Effects of eutrophication on reef-building corals II.
Structure of scleractinian coral communities on fringing reefs, Barbados, West
Indies. Marine Biology 94: 53-75.
U.S. Environmental Protection Agency (EPA). 2005. Use of Biological Information to
Better Define Designated Aquatic Life Uses in State and Tribal Water Quality
Standards: Tiered Aquatic Life Uses. Office of Water, Washington, DC. EPA-822-
R-05-001.
2006 Draft. Consolidated Assessment and Listing Methodology - Toward a
compendium of best practices. Washington, D. C.: Office of Wetlands, Oceans,
and Watersheds. Available: http://www.epa.gov/owow/monitoring/calm.html.
U.S. National Oceanic and Atmospheric Administration (NOAA). 2001. National Ocean
Service, National Centers for Coastal Ocean Science Biogeography Program. (CD-
ROM). Benthic Habitats of Puerto Rico and the U.S. Virgin Islands. Silver Spring,
MD: National Oceanic and Atmospheric Administration.
US VI Division of Environmental Protection and Department of Planning and Natural
Resources (DEP and DPNR). 2004. State of the Environment: United States Virgin
Islands, http://www.dpnr.gov.vi/dep/pubs/index.htm.
Yoccoz, N. G., Nichols, J. D. and Boulinier, T. 2001. Monitoring of biological diversity
in space and time. Trends In Ecology & Evolution 16: 446-453.
Yoder, C. O., and E. T. Rankin. 1998. The role of biological indicators in a state water
quality management process. Environmental Monitoring and Assessment 51(1-2):
61-88.
Zar, J. H. 1984. Biostatistical analysis, 2nd ed. Englewood Cliffs, New Jersey: Prentice-
Hall, Inc.
45
------- |