Manuscript Click here to download Manuscript: Final Revised Manuscript CSR460 - Armstrong et al.doc 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. Armstrong1, Hanumant Singh2, Juan Torres1, Richard S. Nemeth3 Ali Can2, Chris Roman2, Ryan Eustice2, Lauren Riggs4, and Graciela Garcia-Moliner1,5 1 Bio-optical Oceanography Laboratory, University of Puerto Rico, Mayagüez, PR 00681-9013 2 Department of Applied Ocean Physics and Engineering, Woods Hole Oceanographic Institution, Woods Hole, MA 02543-1109 3 Center for Marine and Environmental Studies, University of the Virgin Islands, St. Thomas, USVI, 00802 4 Economics and Environmental Studies Program, Saint Mary’s College of Maryland, Saint Mary's City, MD 20686 5 Caribbean Fishery Management Council, 268 Muñoz Rivera Ave., Suite 1108, San Juan, PR 00918 * Corresponding author: Roy A. Armstrong, tel. (787) 899-6875, fax: (787) 899-5500, email: [email protected] 1 Abstract 2 3 The benthic communities of the deep insular shelf at the Hind Bank 4 Marine Conservation District (MCD), an important spawning grouper 5 aggregation site, were studied with the Seabed autonomous underwater vehicle 6 (AUV) at depths between 32 to 54 m. Four digital phototransects provided data 7 on benthic species composition and abundance of the insular shelf off St. 8 Thomas, U.S. Virgin Islands. Within the western side of the MCD, well- 9 developed coral reefs with 43% mean living coral cover were found. The 10 Montastrea annularis complex was dominant at all four sites between 33 to 47 11 m, the depth range where reefs were present. Maximum coral cover found was 12 70% at depths of 38 to 40 m. Quantitative determinations of sessile-benthic 13 populations, as well as the presence of motile-megabenthic invertebrates and 14 algae were obtained. The Seabed AUV provided new quantitative and 15 descriptive information of a unique coral reef habitat found within this deeper 16 insular shelf area. 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 Keywords: Atlantic; U.S. Virgin Islands; Coral reef; Deep hermatypic corals; AUV 40 1 1. Introduction 2 3 Shallow coral reefs in the tropical West Atlantic Ocean have 4 experienced living coral cover declines of up to 80% in the last 30 years 5 (Gardner et al., 2003). Pollution, sedimentation, hurricanes, and coral disease 6 are all contributors to the decline in living coral cover. While numerous 7 assessments of coral reef habitat have been conducted throughout the tropical 8 western Atlantic and elsewhere using SCUBA, there is limited information on 9 the deeper reefs zones, up to 100 m, which lie beyond the range of safe diving 10 operations. Mapping and monitoring deeper hermatypic (zooxanthellate) coral 11 reefs is necessary for selecting and protecting unique areas of high living coral 12 cover, diversity, and structural complexity that could aid in the recovery of the 13 shallower fish and benthic communities. 14 15 In the past, an evaluation of deep-water fish habitats and abundance 16 around Puerto Rico and the U.S. Virgin Islands, at depths ranging from about 17 100 - 450 m, was conducted using the Johnson-Sea-Link II submersible (Nelson 18 and Appeldoorn, 1985). To this day, the upper insular slope reefs of Puerto 19 Rico and the Virgin Islands, found at intermediate depths (30-100 m), remain 20 largely undescribed. 21 22 The Seabed autonomous underwater vehicle (AUV) is a new imaging 23 platform designed for high resolution optical and acoustic sensing (Singh et al., 24 2004). The Seabed was tested over the insular shelf slope off southwestern 25 Puerto Rico in March 2002 to a maximum depth of 125 m (Armstrong et al., 26 2002; Singh et al., 2004). This test dive established the potential of this new tool 27 for mapping and characterizing the upper insular slope coral reef habitat. 28 29 During June 2003, we used the Seabed for imaging the benthic habitats 30 of the shelf edge and upper insular slope of the Hind Bank Marine Conservation 31 District (MCD), south of St. Thomas, U.S. Virgin Islands. The MCD is a non- 32 fishing area that was established as per recommendation of fishers, scientists 33 and government officials in 1999, to conserve and manage representative 34 samples of marine habitats and ecosystems and to maintain marine biodiversity. 35 The area was originally managed as a seasonal closure for the protection of the 36 red hind (Epinephelus guttatus) spawning aggregation on the eastern side of the 37 MCD. The spawning aggregation primary site was described by Beets and 38 Friedlander (1997). The dominant coral present was flattened colonies of 39 Montastrea annularis. More recently, a limited number of diver surveys at 40 depths of 38-39 m were conducted by Nemeth et al. (2004). 41 42 Up to now, very little quantitative information has been available on the 43 structure and composition of the coral reefs at the MCD due to the depth of 44 these reefs. The goal of this study was to characterize the benthic community 45 structure and biodiversity of these deeper insular shelf corals reefs. The Seabed 46 high resolution color images were used to describe the major components of the 3 1 coral reef community of the MCD, which includes corals, gorgonians, sponges, 2 other macro-invertebrates, and algae, and to quantify living coral cover, species 3 richness, diversity, and evenness. 4 2. Study area 5 6 The Hind Bank MCD is located approximately 12 km south of St. 7 Thomas, U.S. Virgin Islands at the shelf edge encompassing an area of 8 approximately 41 km2 (Figure 1). According to the bathymetry map of this area, 9 depth ranges within the MCD from about 30 m on the western part to more than 10 430 m off the shelf edge. Within the insular shelf area of the MCD the 11 maximum depth is approximately 62 m with an average depth of 42 m on the 12 eastern half. Four, one-kilometer long transects were selected for the AUV 13 surveys based on the known distribution of coral reef habitats and fish spawning 14 grounds provided by previous SCUBA assessments. The two transects on the 15 western side of the MCD were parallel to shore over areas dominated by coral 16 reefs. The eastern transects were oriented perpendicular to shore at known fish 17 spawning sites. 18 3. Materials and methods 19 20 The Seabed AUV is composed of two torpedo-like body sections fixed 21 to each other with vertical structural members (Figure 2). The AUV was 22 designed to operate up to 2000 m depth to carry out photo transects, side-scan 23 sonar and bathymetric surveys. A Pixelfly 1024 x 1280 pixel resolution CCD 24 camera, with 12 bits of dynamic range, is the primary optical imaging sensor. 25 The camera was pointed straight down. A 150 watt-second strobe is used for 26 photographic illumination. The strobe is mounted 1.4 meters aft of the camera 27 to reduce the effects of lighting backscatter in the images (Figure 2a). The 28 frequency of photos is a function of strobe recharge time (2.5 seconds). 29 The size of the image was determined based on the altitude of the vehicle to the 30 bottom and the field of view of the camera. 31 32 Measurements of velocity over the bottom, heading, altitude, pitch, roll, 33 and integrated position are provided by a 300 kHz Acoustic Doppler Current 34 Profiler (ADCP), which projects four sonar beams into the water. These are 35 arranged in a Janus configuration pointing 30˚ fore, aft, to port and to starboard. 36 We utilize the forward pointing beam for obstacle avoidance. A Paroscientific 37 Model 8DP depth sensor provides depth information that, when combined with 38 a dedicated vertical thruster, delivers depth accuracies in the order of 3.5 cm 39 during our missions. More information on Seabed components, sensors, control 40 systems, and navigation can be found in Singh et al. (2003). 41 42 The AUV, which runs at speeds between 0.3 m/s to 1 m/s, was 43 programmed to run at minimum speed and to maintain a fixed distance from the 44 bottom to avoid collisions in case sudden changes in bottom relief were 4 1 encountered. The distribution of the four thrusters, coupled with the passive 2 stability of a two-hulled vehicle with a large meta-centric height (Figure 2b), 3 allows the Seabed to survey close to the sea floor, even in very rugged terrain 4 and to follow a greater than 45˚ slope. As opposed to most other AUV’s, the 5 Seabed was designed to be hover capable, that is, to be able to independently 6 drive in the X,Y and Z axes. 7 8 Each transect was approximately 1 km in length and produced over 500 9 images, each about 3.1 m wide by 2.5 m long, at the predetermined 4 m distance 10 from the bottom. The exact depth of each image was determined by combining 11 the depth of the AUV (measured with the Paroscientific sensor) with its altitude 12 (measured by the ADCP). The initial position of the vehicle at launch is 13 determined by shipboard GPS with an accuracy of tens of meters. 14 15 Underwater imagery is typically characterized by low contrast and low 16 color fidelity. The nonlinear attenuation of the visible spectrum in seawater 17 causes parts of the visible spectrum to be preferentially attenuated. Thus most 18 underwater images tend to be saturated in the blue-green region. We 19 reprocessed the raw imagery to obtain three channel, 1280x1024x12bit RGB 20 images. The 12 bits of dynamic range provided by the camera that is part of 21 Seabed allow us to compensate and color balance the imagery as shown in 22 Figure 3. 23 24 The enhanced benthic imagery was analyzed for percent cover of living 25 corals and other benthic components using random points generated by the 26 Point Count for Coral Reefs software (P. Dustan, personal communication). The 27 clearest, best quality images (50-70 per transect) were selected and individually 28 analyzed using 50 random points and the percent cover of corals, sponges, 29 algae, and bare substrate was recorded. From images acquired within coral reef 30 areas the diversity index for corals was calculated using the Shannon-Weaver 31 Diversity index: 32 33 H’ = - ∑ P ln P (1) i i 34 35 where P = n / n and n = total number of coral colonies. Evenness was i i 36 calculated using Pielou’s evenness index: 37 38 J = H’ / ln H (2) max 39 40 where H = total number coral species. max 41 42 4. Results 43 44 At the western side of the MCD, the reef at PR-03 had a depth range 45 from 33 to 42 m (Figure 4a) while at PR-04, the reef extended from 36 to 41 m 46 in depth (Figure 4b). Higher depths were found at the eastern side of the MCD 5 1 where the transect lines were perpendicular to the coast and extended beyond 2 the shelf edge. At PR-07 the reef was restricted to a narrow zone 45 to 47 m 3 deep in a transect that exceeded 60 m in depth (Figure 4c). In a similar way, at 4 PR-08 coral reefs were only present in the middle of the transect at depths from 5 41 to 46 m (Figure 4d). In all cases the AUV followed the bottom contour 6 maintaining a distance of approximately 4 m from the bottom. In addition to 7 depth, bottom composition varied considerably between the east and the west 8 transects of the MCD. Transects PR-03 and PR-04 were mainly composed of 9 coral reef areas (97 and 100%, respectively), whereas transects PR-07 and PR- 10 08 consisted of small coral reef patches separated by hardground areas along 11 with algal mats, sand, and sparse seagrasses (Figure 5). Hardgrounds are low- 12 relief areas where sediments have been lithified into rock. In these consolidated 13 substrates sparse sponges, gorgonians and algae may be present. Algal mats are 14 a dense layer of algae that blankets the bottom while “sand” refers to substrates 15 of unconsolidated calcium carbonate grains. Seagrasses were only present 16 within the algal mats and sand bottom categories. Within the reef areas, coral 17 cover at PR-04 was significantly higher (p<0.05) than at the other three sites 18 with coral cover ranging from 29 to 43% (Figure 6). Bare substrate was the 19 second most important category, in terms of percent cover, within coral reefs 20 areas, except for PR-03, where macroalgae was the dominant component. 21 Macroalgae was characterized by green, brown, and red algae, while the bare 22 substrate category includes dead coral colonies, uncolonized hard-bottom and 23 sand. 24 25 Within the scleractinian cover composition, the image data analysis 26 revealed dominance by Montastrea annularis complex at all transects, but 27 particularly at PR-03 (Figure 7) where the average cover of M. annularis was 28 about 90% and this dominance was nearly constant with depth. At PR-04 the 29 reef was also highly dominated by the M. annularis complex but with 30 increasing cover of the genus Agaricia with increasing depth (Figure 8). The M. 31 annularis complex was less dominant at PR-07 (Figure 9) at around 50-63% of 32 the total coral cover. The least dominance of this species was found at 46 m in 33 PR-08 where its coverage was reduced to less than 40% and the genus Agaricia 34 covered up to 30% of the substrate (Figure 10). Overall, Montastrea annularis 35 complex was the dominant coral species representing 92% of the coral cover. 36 One-Way ANOVA showed statistical differences in coral cover among depths 37 at PR-03 and PR-04 (p=0.0006 and p=0.005, respectively), but no significant 38 differences were found at PR-07 and PR-08 (p=0.138 and p=0.194, 39 respectively). 40 41 The relatively small Index of Diversity (range 1.48 to 1.85, Table 1) 42 further validates the dominance of the M. annularis complex over the rest of the 43 species. Species richness was highest on the western side of the MCD with 14 44 identifiable coral species at PR-04 followed by 11 at PR-03. On the eastern side 45 of the MCD, the coral reef at PR-07 and PR-08 consisted of only seven and nine 46 coral species, respectively (Table 1). 6 1 2 Within coral reef and hard ground areas, the benthic cover of other 3 macro invertebrates was dominated by sponges and gorgonians at all sites 4 (Table 2). The most dominant sponges were the giant barrel (Xestospongia 5 muta) and rope sponges of the genus Aplysina. A total of 17 species of sponges, 6 two hydrozoan corals (Millepora sp.), 10 gorgonians, one black coral 7 (Cirrhipathes sp.), and four lobster and crab species (crustaceans) could also be 8 identified in association with coral reef and hardground substrates. Most 9 gorgonian species were found in the western part of the MCD at PR-03 and PR- 10 04. The only invertebrates associated with sand bottom and algae substrate that 11 could be identified from the images were two species of conch (genus 12 Strombus), three species of echinoderms, and the upside-down jellyfish 13 Cassiopea. The echinoderms consisted of the sea urchin Meoma ventricosa and 14 two species of sea cucumbers (holothurians). Seven algae and one seagrass 15 species were identified in the phototransects (Table 3). Only functional groups 16 could be reported for most algae except for Udotea cyathiformis and Halimeda 17 sp. that could be distinguished due to their characteristic shape and were only 18 found in the eastern side of the MCD. Turf algae are a multispecific assemblage 19 of filamentous algae. The seagrass Halophila baillonis was found in sand 20 bottoms and algal substrates at both PR-07 and PR-08. 21 22 5. Discussion 23 24 The geomorphology of the eastern side of the MCD, as depicted by the 25 AUV transects, follows the description of Olsen and LaPlace (1978) who 26 described the red hind spawning site with the bottom topography as a series of 27 coral ridges, parallel to the 100 fathom (182 m) curve. These ridges were 28 usually 100 m across and separated by calcareous sand which ranged from 50 to 29 300 m in width, with the dominant coral, Montastrea annularis, measuring less 30 than 1 m in diameter. From a series of SCUBA dives at the red hind spawning 31 site along the shelf edge reef, at depths of 38-48 m, Clavijo and Tobias (1985) 32 described the substrate as dominated by dense scleractinian cover, primarily 33 plates of Montastrea annularis. Nemeth (2003) reported Montastrea franksi as 34 the dominant species in the MCD and within the spawning aggregation site. 35 36 The Montastrea annularis species complex is composed of the sibling 37 species M. annularis, M. faveolata, and M. franksi (Weil and Knowlton, 1994), 38 which are the dominant reef building corals in the tropical Western Atlantic 39 (Goreau, 1959). While the M. annularis complex represented 20% to 60% of 40 the corals found in near shore and mid-shelf reef areas in depths of 5 to 25 m 41 (Herzlieb et al., in press), it dominated in 30 to 40 m depths with cover ranging 42 from 60 to nearly 100%. In areas deeper than 40m it still represented 50 to 60% 43 of the coral cover. Also common in deep reef areas of the Caribbean are several 44 species of the genus Agaricia (Lang, 1974; Reed, 1985; Liddell and Ohlhorst, 45 1988). Within the MCD, Agaricia sp. represented up to 10% of the coral cover 46 at depths less than 41 m and over 20% of the cover in depths greater than 45 m. 7 1 The M. annularis complex and Agaricia sp. both grew as plate-like formations. 2 The plate-like morphology is an adaptation to reduced sunlight at deeper depths 3 (Grauss and Macintyre, 1982). Porites astreoides and Mycetophyllia sp. were 4 found at all depths between 33 m and 47 m. Agaricia sp. was only found in 5 small colonies that were easily identified by their spiraling ridge pattern and 6 brownish color. The average living coral cover seen in the photo transects was 7 43% with maximum cover of 70% found in some areas at 40 m. Also found in 8 abundance was an encrusting sponge, which is most likely Cliona sp. This 9 sponge accounted for 96% of all the sponges found, while most of the other 10 sponges present were columnar. Cliona is a brown-black sponge that grows 11 over corals as it dissolves them, which may account for the high percentage of 12 bare substrate. 13 14 In Bermuda, most deep-water hermatypic corals grow on reef substrate 15 down to 60 m (Fricke and Meischner, 1985). The coral growth in the deeper 16 reef areas in Bermuda appeared to be controlled by macroalgae growth. In the 17 Bahamas, coral assemblages less than 30m depth represented 3 to 25% of 18 benthic cover and were dominated by Agaricia agaricites, M. cavernosa, M. 19 annularis, P. astreoides, and P. porites (Liddell et al., 1997). The deeper (30-50 20 m) coral assemblages represented less than 3% of the benthic cover and were 21 dominated by M. cavernosa and Agaricia spp. (Liddell et al., 1997). Low water 22 temperatures favor macroalgae growth leading to a reduction of coral growth 23 (Johannes et al., 1983). In the fore-reef slope of Jamaica, the flattened growth 24 form of M. annularis is the dominant species of coral above 45 m (Goreau and 25 Land, 1974). 26 27 At a study site close to the PR-07 transect in this study, Nemeth et al. 28 (2004) reported mean values of 28% living corals, 35% macroalgae, 20% dead 29 coral with turf algae, 7% sand/sediment, 4% sponges, and 0% gorgonians. This 30 compares favorably with the AUV estimates of living coral (29%) and sponges 31 (3.8%) at PR-07. The bare substrate category was 28.5%, coralline algae 32 (16.6%), macroalgae (21.2%), and gorgonians (1%). In a separate study, 33 Herzlieb et al. (in press) found coral cover on the western end of the MCD to be 34 44.8% with coral cover of the M. annularis complex near 80%. Nemeth et al. 35 (2004) also reported dominance by the M. annularis complex at 76% of total 36 living coral cover, a 1.4 diversity index, coral bleaching at 6.5% and incidence 37 of diseased corals at 1.4%. The present study could not detect any evidence of 38 coral disease in any of the photo transects and at PR-07, the diversity index was 39 1.8 with dominance of M. annularis complex at 50-60%. The reefs at PR-07 are 40 substantially deeper (45-47 m) than the reef area described by Nemeth et al. 41 (2004) at 38-39 m. This could account for the observed differences in some of 42 these coral reef parameters. 43 44 Although the coral reefs of the MCD are largely unknown, they appear 45 to be relatively undisturbed by anthropogenic factors due to the considerable 46 distance to land sources of runoff and pollution (Herzlieb et al., in press). Their 8 1 greater depth also protects them from wave damage from storm surge. 2 Intermediate depth coral reefs, such as those of the MCD, could serve as refugia 3 for shallow water species in times of environmental stress (sensu Riegl and 4 Piller, 2003). 5 6 The high coral cover throughout the MCD suggests that these reefs are 7 largely unaffected by natural and human disturbances. Since the late 1970’s, 8 hurricanes (Woodley et al., 1981; Edmunds and Witman, 1991) and disease 9 (Gladfelter, 1982) have devastated shallow water reef communities throughout 10 the Caribbean. Edmunds and Witman (1991) reported that Hurricane Hugo 11 destroyed over 33% of the live cover of M. annularis on shallow (10 m) reefs 12 off St. John, US Virgin Islands. Although the turbulence from large hurricane 13 waves can cause fragmentation and abrasion damage in many shelf edge reefs, 14 these effects are limited to depths of 30 m (Blanchon and Jones, 1997). The 15 relatively high depths present at the MCD protects these coral reefs from the 16 high turbulence generated by hurricane waves, since the minimum depth of the 17 surveyed reefs was 33 m and most reefs were present at depths of 40 m or more. 18 Recently, the mortality rate of shallow water corals in the Caribbean has 19 increased dramatically due to a variety of diseases (Richardson et al., 1998a,b; 20 Harvell et al., 1999). Although no disease was evident in the Seabed AUV 21 digital imagery, Nemeth et al. (2004) reported that low levels of disease 22 occurred within the MCD in spring 2003, but higher levels of disease (over 23 10%) occurred on similar deep reefs outside the MCD. They speculated that the 24 unprotected deep reefs may have higher incidence of disease due to commercial 25 fish traps, which not only injure corals but may act as vectors of disease when 26 fishermen transfer traps from shallow to deep reefs. 27 28 There are numerous reports in the literature describing intermediate 29 depth coral reefs (30-50 m) present in the deep fore-reef and upper insular 30 slopes of many Caribbean Islands. However, the coral reefs of the MCD 31 described here are present within the deeper zones of the insular shelf, a unique 32 habitat that has not been described elsewhere. The distribution and 33 characterization of similar habitats at other locations needs to be established 34 before meaningful comparisons with similar coral communities can be made. 35 36 In summary, the quantitative characterization of benthic communities 37 associated with the deep coral reef habitats of the MCD, as provided by the four 38 AUV phototransects, compares favorably with the available information from 39 previous diving surveys in this area. In most cases, the resolution of the Seabed 40 digital imagery was adequate to identify the organisms to the species level. In 41 other cases, only genera could be identified, and for most of the algae, only 42 major groups could be identified. This study demonstrates that the high quality 43 digital imagery provided by the Seabed AUV can be used to characterize the 44 deeper zooxanthellate coral reef habitat present at depths greater than 30 m, 45 which for most areas of the world, remain largely unknown. Quantitative 9 1 determinations of sessile-benthic populations, as well as the presence of motile- 2 megabenthic invertebrates and algae were obtained. 3 4 Our initial results show that the deeper reefs of the MCD are largely 5 unaffected by hurricane disturbances, human impacts, and disease. Therefore, 6 they could serve as a potential refuge areas and a source of larvae for the 7 recovery of shallower coral reef communities present downstream and may play 8 a critical role in the reestablishment of fish populations in adjacent insular shelf 9 areas. 10 11 12 Acknowledgments: 13 14 We thank Captain Harry Montalvo and the crew of the R/V Chapman 15 and the field and technical assistance of members of the Seabed AUV 16 operations team. We appreciate the recommendations for transect locations 17 provided by Elizabeth Whiteman, the field assistance by Stacy Albritton, 18 Elizabeth Kadison, Yasmín Detrés, Fernando Gilbes, Liane Guild, and Jeannette 19 Arce, and the helpful comments and corrections of two anonymous reviewers. 20 Funding was provided in part by the CenSSIS ERC of the National Science 21 Foundation under grant EEC-9986821 and by the Caribbean Fishery 22 Management Council. University of the Virgin Islands staff time was supported 23 by a grant from Sea Grant (R-101-1-02) to R. Nemeth. 24 25 References: 26 27 Armstrong, R.A., Singh, H., Torres, J., 2002. Benthic survey of insular 28 slope coral reefs using the Seabed AUV. Backscatter 13 (3), 22-25. 29 Beets, J., Friedlander, A., 1997. Evaluation of the spawning aggregation 30 closure for red hind (Epinephelus guttatus), St. Thomas, US Virgin 31 Islands. Report to the Caribbean Fishery Management Council, San 32 Juan, Puerto Rico, 17 pp. 33 Blanchon, P., Jones, B., 1997. Hurricane control on shelf-edge reef 34 architecture around Grand Cayman. Sedimentology 44, 479-506. 35 Clavijo, I.E., Tobias, W.J., 1985. Virgin Islands commercial fisheries 36 research and development project (PL 88-309) Project No. 2-411-R- 37 1. Annual Report, April 1, 1984 to March 31, 1985, National Marine 38 Fisheries Service, St. Petersburg, Florida, 22 pp. 39 Edmunds, P.J., Witman, J.D., 1991. Effects of Hurricane Hugo on the 40 primary framework of a reef along the south shore of St. John, US 41 Virgin Islands. Marine Ecology Progress Series 78, 201-204. 42 10