USVI Public Records

Document text

Continental Shelf Research 26 (2006) 194–205 Characterizing the deep insular shelf coral reef habitat of the Hind Bank marine conservation district (US Virgin Islands) using the Seabed autonomous underwater vehicle Roy A. Armstronga,
, Hanumant Singhb, Juan Torresa, Richard S. Nemethc, Ali Canb, Chris Romanb, Ryan Eusticeb, Lauren Riggsd, Graciela Garcia-Molinera,e aBio-optical Oceanography Laboratory, University of Puerto Rico, Mayagu¨ez, PR 00681-9013, USA bDepartment of Applied Ocean Physics and Engineering, Woods Hole Oceanographic Institution, Woods Hole, MA 02543-1109, USA cCenter for Marine and Environmental Studies, University of the Virgin Islands, St. Thomas, USVI 00802, USA dEconomics and Environmental Studies Program, Saint Mary’s College of Maryland, Saint Mary’s City, MD 20686, USA eCaribbean Fishery Management Council, 268 Mun˜oz Rivera Avenue, Suite 1108, San Juan, PR 00918, USA Received 14 January 2005; received in revised form 24 October 2005; accepted 26 October 2005 Available online 4 January 2006 Abstract The benthic communities of the deep insular shelf at the Hind Bank marine conservation district (MCD), an important spawning aggregation site for groupers, were studied with the Seabed autonomous underwater vehicle (AUV) at depths between 32 and 54 m. Four digital phototransects provided data on benthic species composition and abundance of the insular shelf off St. Thomas, US Virgin Islands. Within the western side of the MCD, well-developed coral reefs with 43% mean living coral cover were found. The Montastrea annularis complex was dominant at all four sites between 33 and 47 m, the depth range where reefs were present. Maximum coral cover found was 70% at depths of 38–40 m. Quantitative determinations of sessile benthic populations, as well as the presence of motile megabenthic invertebrates and algae were obtained. The Seabed AUV provided new quantitative and descriptive information of a unique coral reef habitat found within this deeper insular shelf area. r 2005 Elsevier Ltd. All rights reserved. Keywords: Atlantic; US Virgin Islands; Coral reef; Deep hermatypic corals; AUV 1. Introduction Shallow coral reefs in the tropical West Atlantic Ocean have experienced living coral cover declines of up to 80% in the last 30 yr (Gardner et al., 2003). Pollution, sedimentation, hurricanes and coral disease are all contributors to the decline in living coral cover. While numerous assessments of coral reef habitat have been conducted throughout the tropical western Atlantic and elsewhere using SCUBA, there is limited information on the deeper reefs zones, up to 100 m, which lie beyond the range of safe diving operations. Mapping and monitoring deeper hermatypic (zooxanthellate) coral reefs is necessary for selecting and protecting unique areas ARTICLE IN PRESS www.elsevier.com/locate/csr 0278-4343/$ - see front matter r 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.csr.2005.10.004
Corresponding author. Tel.: +1 787 899 6875; fax: +1 787 899 5500. E-mail addresses: roy@cacique.uprm.edu, rarmstrong@whoi.edu (R.A. Armstrong). of high-living coral cover, diversity and structural complexity that could aid in the recovery of the shallower fish and benthic communities. In the past, an evaluation of deep-water fish habitats and abundance around Puerto Rico and the US Virgin Islands, at depths ranging from about 100–450 m, was conducted using the Johnson-Sea- Link II submersible (Nelson and Appeldoorn, 1985). To this day, the upper insular slope reefs of Puerto Rico and the Virgin Islands, found at intermediate depths (30–100 m) remain largely undescribed. The Seabed autonomous underwater vehicle (AUV) is a new imaging platform designed for high-resolution optical and acoustic sensing (Singh et al., 2004). The Seabed was tested over the insular shelf slope off southwestern Puerto Rico in March 2002 to a maximum depth of 125 m (Armstrong et al., 2002; Singh et al., 2004). This test dive established the potential of this new tool for mapping and characterizing the upper insular slope coral reef habitat. During June 2003, we used the Seabed for imaging the benthic habitats of the shelf edge and upper insular slope of the Hind Bank marine conservation district (MCD), South of St. Thomas, US Virgin Islands. The MCD is a non-fishing area that was established as per recommendation of fishers, scientists and government officials in 1999, to conserve and manage representative samples of marine habitats and ecosystems and to maintain marine biodiversity. The area was originally man- aged as a seasonal closure for the protection of the red hind (Epinephelus guttatus) spawning aggrega- tion on the eastern side of the MCD. The spawning aggregation primary site was described by Beets and Friedlander (1997). The dominant coral present was flattened colonies of Montastrea annularis. More recently, a limited number of diver surveys at depths of 38–39 m were conducted by Nemeth et al. (2004). Up to now, very little quantitative information has been available on the structure and composition of the coral reefs at the MCD due to the depth of these reefs. The goal of this study was to character- ize the benthic community structure and biodiver- sity of these deeper insular shelf corals reefs. The Seabed high-resolution color images were used to describe the major components of the coral reef community of the MCD, which includes corals, gorgonians, sponges, other macro-invertebrates and algae, and to quantify living coral cover, species richness, diversity and evenness. 2. Study area The Hind Bank MCD is located approximately 12 km South of St. Thomas, US Virgin Islands, at the shelf edge encompassing an area of approxi- mately 41 km2 (Fig. 1). According to the bathymetry map of this area, depth ranges within the MCD from about 30 m on the western part to more than 430 m off the shelf edge. Within the insular shelf area of the MCD the maximum depth is approxi- mately 62 m, with an average depth of 42 m on the eastern half. Four, 1-km-long transects were selected for the AUV surveys based on the known distribution of coral reef habitats and fish spawning grounds provided by previous SCUBA assessments. The two transects on the western side of the MCD were parallel to the shore over areas dominated by coral reefs. The eastern transects were oriented perpendicular to shore at known fish spawning sites. 3. Materials and methods The Seabed AUV is composed of two torpedo- like body sections fixed to each other with vertical structural members (Fig. 2). The AUV was designed to operate up to 2000 m depth, to carry out photo transects, side-scan sonar and bathymetric surveys. A Pixelfly 1024
1280 pixel resolution CCD cam- era, with 12 bits of dynamic range, is the primary optical imaging sensor. The camera was pointed straight down. A 150 W s strobe is used for photographic illumination. The strobe is mounted 1.4 m aft of the camera to reduce the effects of lighting backscatter in the images (Fig. 2a). The frequency of photos is a function of strobe recharge time (2.5 s). The size of the image was determined based on the altitude of the vehicle to the bottom and the field of view of the camera. Measurements of velocity over the bottom, heading, altitude, pitch, roll and integrated position are provided by a 300 kHz acoustic Doppler current profiler (ADCP), which projects four sonar beams into the water. These are arranged in a Janus configuration pointing 301 fore, aft, to port and to starboard. We utilize the forward pointing beam for obstacle avoidance. A Paroscientific Model 8DP depth sensor provides depth information that, when combined with a dedicated vertical thruster, delivers depth accuracies in the order of 3.5 cm during our missions. More information on Seabed components, sensors, control systems and navigation can be found in Singh et al. (2004). ARTICLE IN PRESS R.A. Armstrong et al. / Continental Shelf Research 26 (2006) 194–205 195 The AUV, which runs at speeds between 0.3 m/s to 1 m/s, was programmed to run at minimum speed and to maintain a fixed distance from the bottom, to avoid collisions in case sudden changes in bottom relief were encountered. The distribution of the four thrusters, coupled with the passive stability of a two-hulled vehicle with a large meta-centric height (Fig. 2b), allows the Seabed to survey close to the sea floor, even in very rugged terrain, and to follow a greater than 451 slope. As opposed to most other AUV’s, the Seabed was designed to be hover capable, i.e., to be able to independently drive in the X, Y and Z axes. Each transect was approximately 1 km in length and produced over 500 images, each about 3.1 m wide by 2.5 m long, at the predetermined 4 m distance from the bottom. The exact depth of each image was determined by combining the depth of the AUV (measured with the Paroscientific sensor) with its altitude (measured by the ADCP). The initial position of the vehicle at launch is determined by shipboard GPS with an accuracy of tens of meters. Underwater imagery is typically characterized by low contrast and low color fidelity. The nonlinear attenuation of the visible spectrum in seawater causes parts of the visible spectrum to be preferen- tially attenuated. Thus most underwater images tend to be saturated in the blue–green region. We ARTICLE IN PRESS Fig. 2. Seabed AUV: (A) schematic diagram of the major components and (B) external view. Fig. 1. Map of the Hind Bank MCD. The location of the MCD in relation to Puerto Rico and the US Virgin Islands is indicated by an arrow. Corner coordinates are: (A) 18113.20N, 651060W; (B) 18113.20N, 641590W; (C) 18110.70N, 651060W; (D) 18111.80N, 641590W (Federal Register 64 (213): 60132-60133; 4 November 1999) (50 CFR Part 622 y622.33 (b) (1)). R.A. Armstrong et al. / Continental Shelf Research 26 (2006) 194–205 196 reprocessed the raw imagery to obtain three channel, 1280
1024
12 bit RGB images. The 12 bits of dynamic range provided by the camera that is part of Seabed allow us to compensate and color balance the imagery, as shown in Fig. 3. The enhanced benthic imagery was analyzed for percent cover of living corals and other benthic components using random points generated by the Point Count for Coral Reefs software (P. Dustan, personal communication). The clearest, best quality images (50–70 per transect) were selected and individually analyzed using 50 random points, and the percent cover of corals, sponges, algae and bare substrate was recorded. From images acquired within coral reef areas the diversity index for corals was calculated using the Shannon–Weaver diversity index H0 ¼  X Pi ln Pi, (1) where Pi ¼ ni=n and n ¼ total number of coral colonies. Evenness was calculated using Pielou’s evenness index J ¼ H0= ln Hmax, (2) where Hmax ¼ total number coral species. 4. Results At the western side of the MCD, the reef at PR-03 had a depth range from 33 to 42 m (Fig. 4a) while, at PR-04, the reef extended from 36 to 41 m in depth (Fig. 4b). Higher depths were found at the eastern side of the MCD where the transect lines were perpendicular to the coast and extended beyond the shelf edge. At PR-07 the reef was restricted to a narrow zone, 45 to 47 m deep, in a transect that exceeded 60 m in depth (Fig. 4c). In a similar way, at PR-08, coral reefs were only present in the middle of the transect at depths from 41 to 46 m (Fig. 4d). In all cases, the AUV followed the bottom contour maintaining a distance of approximately 4 m from the bottom. In addition to depth, bottom composi- tion varied considerably between the East and the West transects of the MCD. Transects PR-03 and PR-04 were mainly composed of coral reef areas (97% and 100%, respectively), whereas transects PR-07 and PR-08 consisted of small coral reef patches separated by hardground areas along with algal mats, sand and sparse seagrasses (Fig. 5). Hardgrounds are low-relief areas where sediments have been lithified into rock. In these consolidated substrates sparse sponges, gorgonians and algae may be present. Algal mats are a dense layer of algae that blanket the bottom, while ‘‘sand’’ refers to substrates of unconsolidated calcium carbonate grains. Seagrasses were only present within the algal mats and sand bottom categories. Within the reef areas, coral cover at PR-04 was significantly higher (po0:05) than at the other three sites, with coral cover ranging from 29% to 43% (Fig. 6). Bare substrate was the second most important category, in terms of percent cover within coral reefs areas, except for PR-03 where macro-algae was the dominant component. Macro-algae was character- ized by green, brown and red algae, while the bare substrate category includes dead coral colonies, uncolonized hard-bottom and sand. Within the scleractinian cover composition, the image data analysis revealed dominance by M. annularis complex at all transects, but particularly at PR-03 (Fig. 7) where the average cover of M. annularis was about 90% and this dominance was nearly constant with depth. At PR-04, the reef was ARTICLE IN PRESS Fig. 3. Reef at a depth of 40 m at PR-04 before (A) and after (B) color enhancement. Scale bar ¼ 1 m. R.A. Armstrong et al. / Continental Shelf Research 26 (2006) 194–205 197 also highly dominated by the M. annularis complex, but with increasing cover of the genus Agaricia with increasing depth (Fig. 8). The M. annularis complex was less dominant at PR-07 (Fig. 9) at around 50–63% of the total coral cover. The least dominance of this species was found at 46 m in PR-08 where its coverage was reduced to less than 40% and the genus Agaricia covered up to 30% of the substrate (Fig. 10). Overall, M. annularis complex was the dominant coral species represent- ing 92% of the coral cover. One-way ANOVA showed statistical differences in coral cover among depths at PR-03 and PR-04 (p ¼ 0:0006 and p ¼ 0:005, respectively), but no significant differences were found at PR-07 and PR- 08 (p ¼ 0:138 and p ¼ 0:194, respectively). The relatively small index of diversity (range 1.48–1.85, Table 1) further validates the dominance of the M. annularis complex over the rest of the species. Species richness was highest on the western side of the MCD, with 14 identifiable coral species at PR-04 followed by 11 at PR-03. On the eastern side of the MCD, the coral reef at PR-07 and PR-08 consisted of only seven and nine coral species, respectively (Table 1). Within coral reef and hard ground areas, the benthic cover of other macro-invertebrates was dominated by sponges and gorgonians at all sites (Table 2). The most dominant sponges were the giant barrel (Xestospongia muta) and rope sponges of the genus Aplysina. A total of 17 species of sponges, two hydrozoan corals (Millepora sp.), ARTICLE IN PRESS Fig. 4. (a)–(d). Depth profiles of the AUV transects: (a) PR-03, (b) PR-04, (c) PR-07, and (d) PR-08. The box in (c) and (d) indicate where the coral reef was located within the transect. The reef was continuous over the entire transect in (a) and (b). R.A. Armstrong et al. / Continental Shelf Research 26 (2006) 194–205 198 10 gorgonians, one black coral (Cirrhipathes sp.) and four lobster and crab species (crustaceans) could also be identified in association with coral reef and hardground substrates. Most gorgonian species were found in the western part of the MCD at PR- 03 and PR-04. The only invertebrates associated with sand bottom and algae substrate that could be identified from the images were two species of conch (genus Strombus), three species of echinoderms and the upside-down jellyfish Cassiopea. The echino- derms consisted of the sea urchin Meoma ventricosa and two species of sea cucumbers (holothurians). Seven algae and one seagrass species were identified in the phototransects (Table 3). Only functional groups could be reported for most algae, except for Udotea cyathiformis and Halimeda sp. that could be distinguished due to their characteristic shape and were only found in the eastern side of the MCD. Turf algae are a multispecific assemblage of filamentous algae. The seagrass Halophila baillonis was found in sand bottoms and algal substrates at both PR-07 and PR-08. 5. Discussion The geomorphology of the eastern side of the MCD, as depicted by the AUV transects, follows the description of Olsen and LaPlace (1978) who described the red hind spawning site with the bottom topography as a series of coral ridges, parallel to the 100 fathom (182 m) curve. These ridges were usually 100 m across and separated by calcareous sand which ranged from 50 to 300 m in width, with the dominant coral, M. annularis, measuring less than 1 m in diameter. From a series of SCUBA dives at the red hind spawning site along the shelf edge reef, at depths of 38–48 m, Clavijo and Tobias (1985) described the substrate as dominated by dense scleractinian cover, primarily plates of M. annularis. Nemeth (2003) reported M. franksi as the dominant species in the MCD and within the spawning aggregation site. The M. annularis species complex is composed of the sibling species M. annularis, M. faveolata and M. franksi (Weil and Knowlton, 1994), which are the dominant reef-building corals in the tropical western Atlantic (Goreau, 1959). While the M. annularis complex represented 20–60% of the corals found in near-shore and mid-shelf reef areas in depths of 5–25 m (Herzlieb et al., in press), it dominated in 30–40 m depths with cover ranging from 60% to nearly 100%. In areas deeper than 40 m it still represented 50–60% of the coral cover. Also common in deep reef areas of the Caribbean are several species of the genus Agaricia (Lang, 1974; Reed, 1985; Liddell and Ohlhorst, 1988). Within the MCD, Agaricia sp. represented up to ARTICLE IN PRESS Coral Reef Hardground Algal Mat Sand + Algae PR 03 PR 04 PR 07 PR 08 0 20 40 60 80 100 Percentages Fig. 5. Percent distribution of various bottom types along the four transects. Scleractinians Bare Substrate Macroalgae Octocorallia Porifera Coralline Algae PR 03 PR 04 PR 07 PR 08 0 10 20 30 40 50 Percentage Fig. 6. Bottom composition within the coral reef areas of each transect. Bare substrate category includes dead coral colonies, uncolonized hard-bottom and sand. R.A. Armstrong et al. / Continental Shelf Research 26 (2006) 194–205 199 ARTICLE IN PRESS 33 34 35 36 37 38 39 40 41 42 L. cucullata C. natans S. bournoni P. porites P. astreoides S. siderea D. strigosa M. cavernosa Mycetophyllia sp. Agaricia sp. M. annularis complex 0 10 20 30 40 50 60 70 80 90 100 Percentage Depths (m) Fig. 7. Coral cover by species and depth at PR-03. 36 37 38 39 40 41 I. rigida Madracis sp. Scolymia sp. L. cucullata C. natans S. bournoni P. porites P. astreoides S. siderea D. strigosa M. cavernosa Mycetophyllia sp. Agaricia sp. M. annularis complex 0 10 20 30 40 50 60 70 80 90 100 Percetange Depths (m) Fig. 8. Coral cover by species and depth at PR-04. R.A. Armstrong et al. / Continental Shelf Research 26 (2006) 194–205 200 ARTICLE IN PRESS 45 46 47 I. rigida P. astreoides S. siderea M. cavernosa Mycetophyllia sp. Agaricia sp. M. annularis complex 0 10 20 30 40 50 60 70 80 90 100 Percentage Depths (m) Fig. 9. Coral cover by species and depth at PR-07. 41 42 43 44 45 46 Millepora sp. Madracis sp. P. astreoides S. siderea D. strigosa M. cavernosa Mycetophyllia sp. Agaricia sp. M. annularis complex 0 10 20 30 40 50 60 70 80 90 100 Percentage Depths (m) Fig. 10. Coral cover by species and depth at PR-08. R.A. Armstrong et al. / Continental Shelf Research 26 (2006) 194–205 201 10% of the coral cover at depths less than 41 m, and over 20% of the cover in depths greater than 45 m. The M. annularis complex and Agaricia sp. both grew as plate-like formations. The plate-like mor- phology is an adaptation to reduced sunlight at deeper depths (Grauss and Macintyre, 1982). Porites astreoides and Mycetophyllia sp. were found at all depths between 33 and 47 m. Agaricia sp. was only found in small colonies that were easily identified by their spiraling ridge pattern and brownish color. The average living coral cover seen in the photo transects was 43%, with maximum cover of 70% found in some areas at 40 m. Also found in abundance was an encrusting sponge, which is most likely a species of Cliona. This sponge accounted for 96% of all the sponges found, while most of the other sponges present were columnar. Cliona is a brown–black sponge that grows over corals as it dissolves them, which may account for the high percentage of bare substrate. In Bermuda, most deep-water hermatypic corals grow on reef substrate down to 60 m (Fricke and Meischner, 1985). The coral growth in the deeper reef areas in Bermuda appeared to be controlled by macro-algae growth. In the Bahamas, coral assem- blages less than 30 m depth represented 3–25% of benthic cover, and were dominated by Agaricia agaricites, M. cavernosa, M. annularis, P. astreoides and P. porites (Liddell et al., 1997). The deeper (30–50 m) coral assemblages represented less than 3% of the benthic cover and were dominated by M. cavernosa and Agaricia spp. (Liddell et al., 1997). Low water temperatures favor macro-algae growth leading to a reduction of coral growth (Johannes et al., 1983). In the fore-reef slope of Jamaica, the flattened growth form of M. annularis is the dominant species of coral above 45 m (Goreau and Land, 1974). At a study site close to the PR-07 transect in this study, Nemeth et al. (2004) reported mean values of 28% living corals, 35% macro-algae, 20% dead coral with turf algae, 7% sand/sediment, 4% sponges and 0% gorgonians. This compares favor- ably with the AUV estimates of living coral (29%) and sponges (3.8%) at PR-07. The bare substrate category was 28.5%, coralline algae (16.6%), macro-algae (21.2%) and gorgonians (1%). In a separate study, Herzlieb et al. (in press) found coral cover on the western end of the MCD to be 44.8%, with coral cover of the M. annularis complex near 80%. Nemeth et al. (2004) also reported dominance by the M. annularis complex at 76% of total living coral cover, a 1.4 diversity index, coral bleaching at 6.5% and incidence of diseased corals at 1.4%. The present study could not detect any evidence of coral disease in any of the phototransects and, at PR-07, the diversity index was 1.8 with dominance of M. annularis complex at 50–60%. The reefs at PR-07 are substantially deeper (45–47 m) than the reef area described by Nemeth et al. (2004) at 38–39 m. This could account for the observed differences in some of these coral reef parameters. Although the coral reefs of the MCD are largely unknown, they appear to be relatively undisturbed by anthropogenic factors due to the considerable distance to land sources of runoff and pollution (Herzlieb et al., in press). Their greater depth also protects them from wave damage from storm surge. Intermediate depth coral reefs, such as those of the MCD, could serve as refugia for shallow water species in times of environmental stress (sensu Riegl and Piller, 2003). The high coral cover throughout the MCD suggests that these reefs are largely unaffected by natural and human disturbances. Since the late 1970s, hurricanes (Woodley et al., 1981; Edmunds and Witman, 1991) and disease (Gladfelter, 1982) have devastated shallow water reef communities throughout the Caribbean. Edmunds and Witman (1991) reported that hurricane Hugo destroyed over 33% of the live cover of M. annularis on shallow (10 m) reefs off St. John, US Virgin Islands. Although the turbulence from large hurricane waves can cause fragmentation and abrasion damage in many shelf edge reefs, these effects are limited to depths of 30 m (Blanchon and Jones, 1997). The relatively high depths present at the MCD protects these coral reefs from the high turbulence generated by hurricane waves, since the minimum depth of the surveyed reefs was 33 m, and most reefs were present at depths of 40 m or more. Recently, the mortality rate of shallow water corals in the Caribbean has increased dramatically due to a variety of diseases (Richardson et al., 1998a, b; ARTICLE IN PRESS Table 1 Species richness, diversity and evenness at the four transects Transect Species richness Diversity index (H0) Evenness (J) PR-03 11 1.48 0.615 PR-04 14 1.73 0.655 PR-07 7 1.85 0.949 PR-08 9 1.50 0.684 R.A. Armstrong et al. / Continental Shelf Research 26 (2006) 194–205 202 ARTICLE IN PRESS Table 2 Macro-invertebrates (other than scleractinians) identified in the phototransects (CR—coral reef, HG—hardground, AM—algal mat, SA— sand bottom+algae) Group/scientific name PR 03 PR 04 PR 07 PR 08 CR HG CR CR HG AM SA CR HG AM SA PORIFERA Cliona sp. + + + + Aplysina cauliformis + + Aplysina archeri + Aplysina sp. + + + + + + + Chondrilla nucula + Diplastrella sp. + + + + + Agelas clathrodes + + + + + + + Agelas conifera + + + + + + Geodia neptuni + + + + + + Amphimedon compressa + + + + + Pseudoceratina crassa + + + Mycale laevis + + Xestospongia muta + + + + + Neofibularia nolitangere + Cinachyra sp. + Cribochalina vasculum + + Strongylacidon sp. + CNIDARIA Hydrozoans Millepora squarrosa + Millepora alcicornis + + + Scyphozoa Cassiopea sp. + Gorgonacea Plexaura sp. + + Ellisella barbadensis + + + Pseudopterogorgia sp. + + + Plexaurella nutans + + + Plexaurella sp. + + + Gorgonia sp. + + + Eunicea sp. + + + Leptogorgia hebes + + Pseudoplexaura sp. + + Muriceopsis sp. + Antipatharia Cirrhipathes sp. + + + ECHINODERMATA Eostichopus sp. + + Holothuria sp. + Meoma ventricosa + ARTHROPODA Scyllarides aequinoctialis + Panulirus argus + + Palinurellus guandlachi + Mithrax sp. + + MOLLUSCA Strombus gigas + + Strombus costatus + R.A. Armstrong et al. / Continental Shelf Research 26 (2006) 194–205 203 Harvell et al., 1999). Although no disease was evident in the Seabed AUV digital imagery, Nemeth et al. (2004) reported that low levels of disease occurred within the MCD in spring 2003, but higher levels of disease (over 10%) occurred on similar deep reefs outside the MCD. They speculated that the unprotected deep reefs may have higher incidence of disease due to commercial fish traps, which not only injure corals but may act as vectors of disease when fishermen transfer traps from shallow to deep reefs. There are numerous reports in the literature describing intermediate depth coral reefs (30–50 m) present in the deep fore-reef and upper insular slopes of many Caribbean Islands. However, the coral reefs of the MCD described here are present within the deeper zones of the insular shelf, a unique habitat that has not been described elsewhere. The distribution and characterization of similar habitats at other locations needs to be established before meaningful comparisons with similar coral commu- nities can be made. In summary, the quantitative characterization of benthic communities associated with the deep coral reef habitats of the MCD, as provided by the four AUV phototransects, compares favorably with the available information from previous diving surveys in this area. In most cases, the resolution of the Seabed digital imagery was adequate to identify the organisms to the species level. In other cases, only genera could be identified, and for most of the algae, only major groups could be identified. This study demonstrates that the high quality digital imagery provided by the Seabed AUV can be used to characterize the deeper zooxanthellate coral reef habitat present at depths greater than 30 m, which for most areas of the world remain largely unknown. Quantitative determinations of sessile benthic populations, as well as the presence of motile megabenthic invertebrates and algae were obtained. Our initial results show that the deeper reefs of the MCD are largely unaffected by hurricane disturbances, human impacts and disease. There- fore, they could serve as a potential refuge areas and a source of larvae for the recovery of shallower coral reef communities present downstream, and may play a critical role in the re-establishment of fish populations in adjacent insular shelf areas. Acknowledgments We thank Captain Harry Montalvo and the crew of the R/V Chapman, and the field and technical assistance of members of the Seabed AUV opera- tions team. We appreciate the recommendations for transect locations provided by Elizabeth Whiteman, the field assistance by Stacy Albritton, Elizabeth Kadison, Yasmı´n Detre´ s, Fernando Gilbes, Liane Guild and Jeannette Arce, and the helpful com- ments and corrections of two anonymous reviewers. Funding was provided in part by the CenSSIS ERC of the National Science Foundation under Grant EEC-9986821 and by the Caribbean Fishery Man- agement Council. University of the Virgin Islands staff time was supported by a grant from Sea Grant (R-101-1-02) to R. Nemeth. ARTICLE IN PRESS Table 3 Plants identified in the phototransects (CR—coral reef, HG—hardground, AM—algal mat, SA—sand bottom+algae) Group/scientific name PR 03 PR 4 PR 07 PR 08 CR HG CR CR HG AM SA CR HG AM SA CYANOBACTERIA ALGAE + + Brown algae + + + + + + + + + Red algae + + + Red coralline algae + + + + + + + Green algae + + + Udotea cyathiformis + + + + + Halimeda sp. + + + + + Turf algae + + + + + + + ANGIOSPERMAE Halophila baillonis + + + + R.A. Armstrong et al. / Continental Shelf Research 26 (2006) 194–205 204 References Armstrong, R.A., Singh, H., Torres, J., 2002. Benthic survey of insular slope coral reefs using the Seabed AUV. Backscatter 13 (3), 22–25. Beets, J., Friedlander, A., 1997. Evaluation of the spawning aggregation closure for red hind (Epinephelus guttatus), St. Thomas, US Virgin Islands. Report to the Caribbean Fishery Management Council, San Juan, Puerto Rico, 17pp. Blanchon, P., Jones, B., 1997. Hurricane control on shelf-edge reef architecture around Grand Cayman. Sedimentology 44, 479–506. Clavijo, I.E., Tobias, W.J., 1985. Virgin Islands commercial fisheries research and development project (PL 88-309) Project No. 2-411-R-1. Annual Report, 1 April 1984–31 March 1985, National Marine Fisheries Service, St. Peters- burg, Florida, 22pp. Edmunds, P.J., Witman, J.D., 1991. Effects of Hurricane Hugo on the primary framework of a reef along the south shore of St. John, US Virgin Islands. Marine Ecology Progress Series 78, 201–204. Fricke, H., Meischner, D., 1985. Depth limits of Bermudan scleractinian corals: a submersible survey. Marine Biology 88 (2), 175–187. Gardner, T.A., Cote, I.M., Gill, J.A., Grant, A., Watkinson, A.R., 2003. Long term region-wide declines in Caribbean corals. Science 301, 958–960. Gladfelter, W.G., 1982. White-band disease in Acropora palmata: implications for the structure and growth of shallow reefs. Bulletin of Marine Science 32, 639–643. Goreau, T.F., 1959. The ecology of Jamaican reefs. I: species composition and zonation. Ecology 40, 67–90. Goreau, T.F., Land, L.S., 1974. Fore-reef morphology and depositional processes, North Jamaica. In: Laporte, L.F. (Ed.), Reefs in Time and Space. Society of Economic Paleontologists and Mineralogists Special Publication No. 8. Tulsa, Oklahoma, pp. 77–89. Grauss, R.R., Macintyre, I.E., 1982. Variation in growth forms of the reef coral Montastrea annularis: a quantitative evaluation of growth response to light distribution using computer simulation. In: Rutzler, K., Macintyre, I.G. (Eds.), The Atlantic Barrier Reef Ecosystem at Carrie Bow Cay, Belize I. Structure and Communities. Smithsonian Contribu- tions in Marine Science 12. Smithsonian Institution Press, Washington, DC, pp. 441–464. Harvell, C.D., Kim, K., Burkholder, J.M., Colwell, R.R., Epstein, P.R., Grimes, D.J., Hofmann, E.E., Lipp, E.K., Osterhaus, A.D.M.E., Overstreet, R.M., Porter, J.W., Smith, G.W., Vasta, G.R., 1999. Emerging marine diseases–climate links and anthropogenic factors. Science 285, 1505–1510. Herzlieb, S., Kadison, E., Blondeau, J., Nemeth, R.S. Compara- tive assessment of coral reef systems located along the insular platform of St. Thomas, US Virgin Islands and the relative effects of natural and human impacts. Proceedings of tenth International Coral Reef Symposium, Okinawa, Japan, in press. Johannes, R.E., Wiebe, W.J., Crossland, C.J., Rimmer, D.W., Smith, S.W., 1983. On the latitudinal limits of coral reef growth. Marine Ecology Progress Series 11, 105–111. Lang, J.C., 1974. Biological zonation at the base of a reef. American Scientist 62, 272–281. Liddell, W.D., Ohlhorst, S.L., 1988. Hard substrata community patterns, 1–120 m., North Jamaica. Palaios 3, 413–423. Liddell, W.D., Avery, W.E., Ohlhorst, S.L., 1997. Patterns of benthic community structure, 10–250 m, The Bahamas. Proceedings of eighth International Coral Reef Symposium, Panama, vol. 1, pp. 437–442. Nelson, W.R., Appeldoorn, R.S., 1985. A submersible survey of the continental slope of Puerto Rico and the US Virgin Islands, 1–23 October 1985. Cruise report, R/V Seward Johnson, National Marine Fisheries Service, Pascagoula, Mississippi Laboratories, 76pp. Nemeth, R.S., 2003. Investigating movement patterns and spawning habitat of red hind grouper in a newly established marine fishery reserve in the US Virgin Islands. Final Project Report for MARFIN (Award NA97FF0348), National Marine Fisheries Service, 21pp. Nemeth, R.S., Herzlieb, S., Kadison, E.S., Taylor, M., Rothen- berger, P., Harold, S., Toller, W., 2004. Coral reef monitoring in St. Croix and St. Thomas, United States Virgin Islands. Final Report submitted to Department of Planning and Natural Resources, US Virgin Islands, 79pp. Olsen, D.A., LaPlace, J.A., 1978. A study of the Virgin Island grouper fishery based on a breeding aggregation. Proceedings of Gulf and Caribbean Fisheries Institute 31, 130–144. Reed, J.K., 1985. Deepest distribution of Atlantic hermatypic corals discovered in the Bahamas. Proceedings of fifth International Coral Reef Congress, Tahiti, vol. 6, pp. 249–254. Richardson, L.L., Goldberg, W.M., Carlton, R.G., Halas, J.C., 1998a. Coral disease outbreak in the Florida Keys: plague type II. Revista de Biologı´a Tropical 46 (Supl. 5), 187–198. Richardson, L.L., Goldberg, W.M., Kuta, K.G., Aronson, R.B., Smith, G.W., Ritchie, K.B., Halas, J.C., Feingold, J.S., Miller, S., 1998b. Florida’s mystery coral killer identified. Nature 392, 557–558. Riegl, B., Piller, W.E., 2003. Possible refugia for reefs in times of environmental stress. International Journal of Earth Science 92, 520–531. Singh, H., Armstrong, R.A., Gilbes, F., Eustice, R., Roman, C., Pizarro, O., Torres, J., 2004. Imaging coral I: imaging coral habitats with the Seabed AUV. Subsurface Sensing Technol- ogies and Applications 5 (1), 25–42. Weil, E., Knowlton, N., 1994. A multi-character analysis of the Caribbean coral Montastrea annularis (Ellis and Solander, 1786) and its two sibling species, M. faveolata (Ellis and Solander, 1786) and M. franksi (Gregory, 1895). Bulletin of Marine Science 55, 151–175. Woodley, J.D., Chornesky, E.A., Clifford, P.A., Jackson, J.B.C., Kaufman, L.S., Knowlton, N., Lang, J.C., Pearson, M.P., Porter, J.W., Rooney, M.C., Rylaarsdam, K.W., Tunnicliffe, V.J., Wahle, C.M., Wulff, J.L., Curtis, A.S.G., Dallmeyer, M.D., Jupp, B.P., Koehl, M.A.R., Neigel, J., Sides, E.M., 1981. Hurricane Allen’s impact on Jamaican coral reefs. Science 214, 749–755. ARTICLE IN PRESS R.A. Armstrong et al. / Continental Shelf Research 26 (2006) 194–205 205