Reef Fishes of Broward County Natural Substrate and Vessel-reefs in Meter Water Depth: Final Report. FWC Grant FWC-05017

Transcription

1 Reef Fishes of Broward County Natural Substrate and Vessel-reefs in Meter Water Depth: Final Report FWC Grant FWC submitted by: Richard Spieler, Ph.D., David Bryan, and Kirk Kilfoyle Oceanographic Center Nova Southeastern University

2 Reef Fish of Broward County Natural Substrate and Vesselreefs in Meter Water Depth: Report 2007 Executive Summary In Broward County previous research has described the reef fish community from nearshore hardbottom habitat to the outer reef in 30m depth. However at depths greater than 50m, little is known of the benthic habitat or fish community present on natural substrate as well as among 20+ artificial reefs that have been created at these depths. This is surprising as there is an important recreational and small commercial, fishery in southeastern Florida and Broward County for several deep-reef species. The deep-reef habitat can be expected to experience ever-increasing fishing pressure due to a paucity of desirable, shallow-water legal size reef fish in Broward County and requires effective management. The effective management of deep-reef fisheries resources, in turn, requires a baseline survey to determine change in those resources due to management techniques, or to any other anthropogenic, or natural, activity. We have examined the deep-reef fish assemblage on natural substrate and artificial reefs (vessel-reefs) in Broward County by use of remotely operated vehicles (ROVs). A previous report covered the first 18 months of the study. This document serves as the contract completion report for the last 21 months but, for continuity, it includes data from the first report as well. The work was plagued by a host of contractual, weather and technical problems. Nonetheless a total of 50,376 fishes of 65 species for 1331 min of swim time were recorded (or about 2271 fish/h). Small, Anthiinae basses numerically dominated the vessel-reefs with an estimate of 44,209 observed during 24 surveys. In the companion study of natural substrate, a total of 1009 fish of 45 species in 1002 min of swim time (or about 60 fish/h) were noted on the natural substrate during the 2005/06 drift transects. It is clear that the natural substrate at 50 to 120m water depth, especially in the areas around vessel-reefs, has a paucity of fishes. The natural substrate, at the depths examined, has the appearance of low-relief, soft sediment with widely scattered patches of low-relief hard bottom. In contrast, vessel-reefs provide an island of hard substrate with substantial amounts of variably sized refuge. It is premature to draw definitive conclusions about the ecological role and potential resource management value of vessel-reefs. Likewise, it would be unwise to assume the substrate photographed in 1002 min of video footage is a representation of the entire substrate, or even a substantial portion of the substrate, in 50 to 120 m of water off Broward County. However, our results clearly highlight the need for additional data to understand the fisheries resources in this depth range, as well as the management potential of artificial reefs. If the substrate at this depth is uniformly of low relief with few fishes, an artificial reef program targeting this area might dramatically enhance fishery resources. 1

3 Table of Contents Executive Summary... 1 TABLE OF CONTENTS...2 INTRODUCTION...4 METHODOLOGY...7 Project Descriptions... 7 Natural reef... 7 Vessel Reefs RESULTS AND DISCUSSION...16 CONCLUSION...28 LITERATURE CITED...29 APPENDIX...32 Appendix A. Species lists for 2005/06 drift transects Appendix B. Vessel Reef Surveys Appendix C. Field Notes /2004 Field Notes Field Notes Field Notes Field Notes

4 List of Figures Figure 1. LADS and side scan sonar of a section Broward County offshore habitat... 5 Figure 3. Location of southern transects (numbers correspond to drift number) Figure 5. Average species richness per hour for vessel-reefs ( ) and natural substrate ( ). Bars indicate standard error Figure 6. Average abundance per hour for vessel-reefs ( ) and natural substrate ( ). Bars indicate standard error Figure 7. Mean species richness per hour by depth category Figure 8. Mean abundance per hour by depth category List of Tables Table 1. Location of 2005 and 2006 drift transects...9 Table 2. List of species observed during 2005/06 natural substrate drift transects and abundances per depth category...17 Table 3. Vessel-reef fish abundances ( )

5 Reef Fish of Broward County Natural Substrate and Vessel-reefs in Meter Water Depth: Final Report Introduction In general, fish assemblages below the depth limits of recreational SCUBA are poorly described. Admittedly, there are a number of research reports on the deep-water communities using deep-dive submersibles. However, coral reef fish assemblages in the shallower 70 to 150m range have, for the most part, received little attention (Colin 1974; 1976; Thresher and Colin 1986; Itzkowitz et al. 1991; Pyle 2000) and there is only one report on this assemblage from the Broward County reef tracts (Shinn and Wicklund 1989). This is surprising as there is an important recreational, and small commercial, fishery in Broward for local members of the deep snapper/grouper complex (e.g. tilefish, blackfin, cubera, silk and vermilion snappers; snowy grouper, scamp, and Warsaw grouper) and many other species are taken as by-catch (eg. conger eels, spot-tail morays, and large scorpaenids). Annual commercial landings summaries from the Florida Fish and Wildlife Conservation Commission s Marine Fisheries Information System suggest that there is a small commercial fishing effort in Broward County for deep-water fishes Over the last ten years ( ) 7755 kg of gag grouper have been landed in Broward County, 3657 kg of scamp, 10,768 kg of snowy grouper, 2060 kg of vermilion snapper, 8045 kg of golden tilefish, and 4809 kg of gray tilefish ( These landings are extremely low, when compared to the rest of Florida. Currently, there is a very little commercial effort in these fisheries in Broward County (74 trips total for 2004). Amberjack was also a small commercial fishery in Broward County in the late 1990 s; however, during the last few years less than 1000 lbs a year have been landed. There does, however, appear to be a popular recreational fishery for amberjack from sport fishing boats throughout Ft Lauderdale. Charter boats and private vessels frequent the deep-water vessel-reefs and appear to be successful in catching grouper and amberjack. We suspect that this fishing pressure impacts the abundance of these species on the deep vessel-reefs. (Pers. obs. authors). The deep-water habitat can be expected to experience ever increasing fishing pressure due to a paucity of desirable, shallow-water bottom fish of legal size in Broward County (Ettinger et al. 2001; Ferro et al. 2005) and requires effective management. The effective management of the deep-water fisheries resources, in turn, requires a baseline survey to determine change in those resources due to management techniques, or to any other anthropogenic, or natural, activity. The narrow continental shelf offshore of Broward County, Florida consists of three relic reef tracts that run parallel to the coastline and which are separated by areas of sand flats. These reef tracts extend roughly 2.5km east of the shoreline to 40m depth. Laser airborne depth soundings (LADS) (Figure 1) has given scientists a comprehensive description of bottom topography in Broward County. East of the 40m depth contour, the continental slope or platform margin slope is relatively steep (Avent et al. 1977) and quickly slopes down to 200m into the Miami Terrace (Malloy and Hurley 1970; Mullins and Neumann 1979). 4

6 At depths greater than 40m, LADS are rarely effective and bottom topography is characterized with sonar. Avent et al. (1977) described many ridges near the shelf edge and on the slope between 70 and 80m that were possible tropical carbonate facies. Walker (NCRI, unpublished data) examined this region between the depths of 67m and 89m with side scan sonar and found an increase in slope (roughly six percent) compared to surrounding depths. Walker (pers. communication) suggests that this may be due to a lithified beach ridge created during the Holocene transgression. High relief substrate has not been reported out to the upper Miami Terrace and general habitat features have not been described from 50 to 120m depth. Several derelict vessels have been deployed in this area for recreational fishing and others are planned. By all reports these artificial reefs have been quite successful as fishing hot spots. The close access to the Gulf Stream, with its abundant larval supply, combined with the apparent low relief and low refuge of the natural substrate, may indicate that this area is ideal for using artificial reefs to increase reef-fish resources in Broward County. The apparently abundant and diverse fish assemblages on the vessel-reefs support such a hypothesis. However, without a full characterization of both the natural and vessel-reef assemblages it is difficult to distinguish between production and simple aggregation. 50m depth 120m depth INSHORE REEFS MIDDLE REEF TRACT OUTER REEF TRACT AREA OF INCREASED SLOPE 70-80m depth BILL BOYD 82m depth, 65m length Figure 1. LADS and side scan sonar of a section Broward County offshore habitat. We examined fish assemblages of deep-reef natural substrate and artificial reefs (vesselreefs) of Broward County by use of a remotely operated vehicle (ROV). The use of ROVs to 5

7 study fish communities is not unprecedented (Thompson et al. 1982) and there has been a steady increase in their scientific use since the 1990 s (Auster 1997). More recently, several researchers have used ROVs to investigate shelf edge prominences through the Gulf of Mexico and Florida (Koenig et al. 2000; Weaver et al. 2001; Gledhill and David 2004). Our preliminary work in m and a scuba/rov comparison study in shallow water also demonstrate the effectiveness of an ROV (Spieler 2004; Spieler and Bryan 2005, Bryan 2006). Two previous reports (Spieler, 2004; Spieler and Bryan 2005) covered the first 2 years of study. This document serves as the contract completion report of the third year of study; but, for continuity, it includes data from the first two years as well. Two separate funding agencies (Florida Fish and Wildlife Conservation Commission, FWC and National Marine Fisheries Service, NMFS) funded separate yet linked studies (FWC: deep vessel-reefs; NMFS: deep hardbottom fish assemblages). However, both agencies are interested in the results of the entire study. Therefore, this report contains data of both sub-studies and has been submitted to both agencies. The FWC project and the NOAA/NMFS project are stand-alone studies. However, because they were done in the same time frame, using the same visual census methodology (combination of roving swim and stationary counts), we were provided the opportunity to compare the fish assemblages on natural deep reef to the assemblages found on vessel-reefs. We have completed similar interfacing studies for NOAA and FWC with productive results (Arena et al., 2002; 2007; Jordan et al., 2002); the comparisons between studies provides added value to both projects. For example, if artificial reefs aid in the production of fish, or merely their aggregation allowing more efficient harvest (often referred to as the attraction/production question ), remains a central problem in the effective use of artificial reefs, including vessel-reefs, as resource management tools. Simply put, the answer to this question determines, from a fisheries perspective, the use to which an artificial reef can be effectively deployed. Data from our lab, and others, makes it clear that there is no single answer to the attraction/production question, but rather the answer depends upon the species, and even the life stage of a species, being discussed. Comparing the vessel-reefs to neighboring natural structure will provide insight into the diversity, abundance, and life-stage (based on total length) of species that associate with deep-water vessel-reefs as compared to natural habitat. This comparison will, in turn, offer insight into the species-specific functionality of these reefs. For example, a high abundance of a species on the vessel-reef and the complete absence of this species on natural reef would imply some aspect of habitat limitation on the natural reef and production of this species on the vessel-reef. This situation has been reported on shallow-water vessel-reefs (Spieler, 2001; Arena et al., 2002; 2007). 6

8 Methodology Project Descriptions Natural reef In 2004 fish censuses were taken at four sites in 70 to 100m of water offshore Broward County. All the sites were neighboring vessel-reefs that were censused in the accompanying study (see below). A non-destructive, visual census was utilized; and census location relative to the vessel depended on the prevailing current. To survey the adjacent substrate of each vessel-reef, timed 20 min roving dives were attempted on either side of the vessel. However, current and technical difficulties precluded us from completing all surveys. We planned to examine any hard bottom exhibiting substantial relief encountered on initial ROV inspection of a site, within 90m of a vessel-reef. However, no substantial natural relief was noted on any dive. Transects were video taped (digital) and the fish identified on return to shore. When feasible we made total length estimates of the fish by comparison with dual laser beams with 11cm separation. In 2005, in an attempt to better characterize fish communities on natural substrate at depths between 50 and 120m the methodology for surveys was modified. Instead of just sampling the area within 90 m of each vessel-reef, longer drift transects were conducted at latitudes near the vessel-reefs (Figure 2,3). Ten transects were accomplished ranging from 11 to 57 min each (355 min total)(table 1). Transect direction depended on prevailing current and wind patterns and varied throughout the study. When possible, the ROV was stopped for stationary counts, of about 5 min, at sites with large numbers of fish. Total transect time includes stationary counts. In 2006, the drift transect methodology was repeated to include a wider survey area and encompass some regions in greater detail (Figure 2,3). Seventeen transects were completed ranging from 15 to 117 minutes in length (Table 1). Three different remotely operated vehicles (ROVs) were used for this project. Two of the ROVs, named BOVs (Benthic Observation Vehicle), were manufactured and piloted by a local engineer, Bill Baxely. The first BOV was used for two deep-water surveys and then the vehicle was upgraded to a second unit to account for some difficulties encountered at depth. The vehicles consisted of four DC thrusters, initially with two 50-watt lights then 100 watts, a digital video camera, compass, depth sensor, two lasers (5cm apart), and a tubular aluminum structure. The ROV s dimensions were 1m length by 55cm width by 42cm height and weighed roughly 31 kg plus ballast. They were depth rated to 300m and had 3 knots of forward speed and 2 knots of reverse speed. Three-hundred m of umbilical, with 3,800lb breaking strength, supplied the 120 AC volts required for operation. The power system initially supplied 400 watts and was upgraded to 500 watts. The video from the ROV was sent back to the recording device through RCA cables and was then converted into digital format. The ROV was controlled from the topside with a viewing station and a PC based joystick/keyboard. Its engineer, Bill Baxley, operated the ROV with input for sampling procedures from D. Bryan Because of continued equipment and scheduling problems. We discontinued use of the BOV in August 2004 (Field Notes). 7

9 The third ROV used was a VideoRay Pro III. This commercial ROV consisted of two high thrust propellers for forward and reverse movement, two 20 watt high efficiency halogen lights, a 570 line high resolution 0.3 lux color camera with a wide angle lens (90 degree field of view, 90 degree tilt) along with a 430 line 0.1 lux rear facing black and white camera, compass, depth sensor, and two lasers spaced 11cm apart. The front camera s trim/tilt and focus could be adjusted from the control box. The vehicle was depth rated for 152m and could handle up to 200m of tether. The vehicle power consumption was less than 300 watts of VAC and maximum voltage in the tether was 48 VDC. The vehicle was also equipped with a forward manipulator. Steve Van Meter was the pilot for the VideoRay. If wind speeds were predicted to be greater than 10 knots, resulting in seas greater than 1m, trips were canceled. If the current offshore was known to be greater than 2 knots, the trip was also canceled. At the artificial reef sites, the RV Panacea was stationed above the vessel reef with a 10 kg danforth anchor and 7m of chain attached to 400m of 3/8 inch triple braided polypro line (black lobster line). Prior to deploying the anchor several drifts were conducted over the vessel reef to determine the current direction and speed, along with any wind influences. Typically the anchor was dropped in 100m depths about 350m up current of the vessel-reef (scope 3-4 to 1). Scope was increased with increased current speed. Generally the anchor system held; however, during the course of a day, wind and current shifts would move the RV Panacea away from the vessel reef and re-anchoring was required. Occasionally, on the first anchor attempt the influence of the wind or surface current was miscalculated and therefore the RV Panacea did not come to rest directly over the vessel reef and had to be re-anchored. With currents in excess of 3 knots, a larger anchor, chain, and line was needed for station keeping; however, ROV surveys were avoided in such conditions. On return to the laboratory the Mini DV video was played back on a 70cm color monitor. Fish identification was primarily done by a lead investigator (D. Bryan); however, for species difficult to ID, multiple staff viewed the video. Fish richness, abundance, and transect time was recorded from the video tapes. Useable video tape or swim time, defined as when the ROV was within 3 m of substrate, was used for analyses. In 2006, the drift transect methodology was repeated to include a wider survey area and encompass some regions in greater detail (Figure 2, 3). Seventeen transects were completed ranging from 15 to 117 minutes in length (Table 1). When reviewing digital video from natural surveys the depth at which each fish was observed was recorded, along with the amount of time spent within 7 depth categories (50-60m, 61-70m, 71-80m, m). Average species richness and abundance per hour for each depth category was calculated using surveys with over 10 minutes spent at a particular depth category. 8

10 Table 1. Location of 2005 and 2006 drift transects. Drift # Survey Date Latitude Longitude Length (m) Survey time (minutes) 1 March March April May May June June June Not Recorded Not Recorded 11 9 June July January January April May July July July July July August August

11 22 August August August August August August

12 Figure 2. Location of northern transects (numbers correspond to drift number). 11

13 Figure 3. Location of southern transects (numbers correspond to drift number). 12

14 Vessel Reefs Fish censuses were taken, or attempted (the Bud Krohn has not been successfully surveyed despite numerous attempts, see Field Notes), on four similar vessels deployed as artificial reefs between 1985 and 1989 in 70m to 140m of water offshore Broward County (Figure 4): Bud Krohn 56m freighter deployed m deep N ' W ' Bill Boyd 64m freighter m N ' W ' Caicos Express 56m freighter m N ' W ' Papa s Reef 52m freighter m N ' W ' These vessels lie on soft substrate east of the third reef tract. It was our intention to sequentially survey these vessel reefs quarterly, however a variety of technical, personnel and weather problems have prevented this (see Field Notes). The census methodology was the same as that used on the transects (a rover-swim with stationary counts in areas of high fish density). Because structure and attendant fishes can vary widely on a sunken vessel, typically these surveys were swam along the long axis of the vessels, on both sides from bow or stern. However, depending on where the fish were, and prevailing currents, these sites were altered to insure that maximum and minimum count areas were included. An attempt was made to spend an equal amount of time on the sand, along the gunwales, on the deck, and around the superstructure; however current patterns often dictated where we could and could not survey. In general it was easiest to survey along the sand, as surveys on the deck and around the superstructure dramatically increased the risk of entanglement. On a few occasions we were able to penetrate the vessels with the ROV. Transects were usually conducted from stern to bow. The same transect was not maintained on repeated visits to the same ship. The ROV was deployed from a stationary vessel using a clump weight system (Bryan 2006). Digital video was recorded as soon as the ROV was placed into in the water. However, for analysis only the video during which fish were actively being viewed was used (fish viewing time). Time during which the vessel was being located, time spent dealing with the clump weight and tether, time during which video was silted out, and any other time in which the vessel-reef could not be viewed was removed (less than 50% of captured video was used as fish viewing video for analysis). Until two hours prior to sunset, ambient light at depth was sufficient to conduct surveys on the Bill Boyd, Caicos Express, and Papa s Reef. At the Bud Krohn, in 135m depth, ambient light became more of a problem. In general, lights were used when looking inside of the vessels through scuttling holes, under the bow and stern, through ports and doors, and inside cargo holds. Lights were also used to aid with fish identification. In 2004 four vessel-reefs surveys were conducted on the Bill Boyd, Caicos Express, and Papa s Reef, with a total fish viewing time of 237 minutes. In 2005 eight surveys were conducted on vessel-reefs for a total of 466 minutes. In 2006 seven surveys were conducted for a total of 343 minutes and in 2007 five surveys were conducted with 257 minutes of fish viewing time. Over the 39 month period, 24 vessel reef surveys were conducted for totaling 13

15 1331 minutes of fish viewing time for an average of 55:30 minutes per survey. The Bill Boyd and Caicos Express were each surveyed nine times and Papa s Reef was surveyed six times. To compare species richness and abundances among vessel-reefs and with natural substrate surveys, average species and abundances per hour were calculated. Both T-tests and ANOVAs were used to determine significant differences ( p<0.05). Species richness values were calculated with fish that were identified to species, or, if identified only to genus, no other fish in the same genus were observed during the survey. Abundance per hour values were log10(x) transformed to homogenize variances. Individual species differences among vessel-reefs were compared by using survey averages. Surveys were dived into two seasons for analysis. Seven surveys were conducted during the dry season (October-March) and seventeen surveys were conducted during the wet season (April-September). Abundance values were log10(x) transformed and a T-test was used to compare seasonal abundance values. We have included some summary statistics (mean and standard error of the mean) and basic statistical analyses (T-test and ANOVA). However, there is a host of confounding variables (e.g., season, current, visibility, timing) that would call into question any statistical analyses at this point. A much larger database is required before we would be comfortable with either parametric or non-parametric analyses. However, it is unlikely that the dramatic difference in numbers between vessel-reefs and hardbottom (see below) are artifacts of methodology. 14

16 Figure 4. Flags mark locations of vessel reefs 15

17 Results and Discussion The 2004 data clearly demonstrated that the natural substrate at 70 to 100m water depth offshore Broward County, at least in the areas around vessel-reefs, has a paucity of fishes. A total of 5 fish of 4 species in 102 min of swim time (or 3 fish/h) were noted on the natural substrate. Of these five, three likely followed the ROV onto the sand from adjacent vesselreefs. In sharp contrast, the vessel-reefs had a total of fishes of 36 species for 343 min of swim time (or 210 fish/h). No significant vertical relief greater than 1m was found surrounding the vessel-reefs and this lack of habitat, rather than attraction of nearby fish to vessel-reefs, is likely the explanation for the lower species richness and fish abundance. In 2005/06 the drift transects revealed a dramatic increase in species richness and abundance when compared to the 2004 sampling near vessel-reefs; 45 species were recorded over the natural substrate from 27 different families (Table 2). Nonetheless, at 45 species, the Broward County deep-reef community is extremely depauperate, especially when compared to the outer reef tract, which is within a kilometer and harbors a diverse reef fish community of at least 173 species (Ferro et al. 2005). The total number of fish recorded during 2005/6 was 1009 for 1002 min of swim-time (594 were recorded to species) or about 60 fish/h of swim-time. However, the natural substrate still yielded significantly lower species richness and abundance values than vessel-reefs (Figure 5, 6). The increase in species richness over the 2004 dataset can be accounted for, in part, by the different habitats found at changing depth. Although we made no effort to quantify habitat type, subjective observations indicate a change in substrate with depth (i.e., algae/rubble at 50-55m, increased rubble/black corals/sparse rope sponge at 55-60m, rubble/octocorals/sparse algae 60-70m, sand and rubble, with an increase of rubble on sloping areas, 70-80m, sand and small scattered rubble at m.). 16

18 Table 2. List of species observed during 2005/06 natural substrate drift transects and abundances per depth category. Abundance (bold # indicate association with artificial structure) Depth categories 51-60m 61-70m 71-80m 81-90m m m m Common Name Scientific Name Eagle Rays Myliobatidae Spotted eagle ray Aetobatus narinari 1 Morays Muraenidae Morray spp Gymnothorax sp. 1 Batfish Ogcocephalidae Longnose batfish Ogcocephalus corniger 2 Batfish species Ogcocephalus sp. 1 Squirrelfish Holocentridae Squirrelfish Holocentrus adscensionis 1 Squirrelfish species Holocentrus sp. 1 Cornetfish Fistularidae Bluespotteds cornetfish Fistularia tabacaria 2 Sea Basses Serranidae Snowy grouper Epinephelus niveatus Bank seabass Centropristis ocyurus Tattler bass Serranus phoebe 2, , 1 1, 12 Roughtongue bass Pronotogrammus martinicensis 16 Anthiine fishes Anthiinids Grouper family Serranidae 1 Bigeyes Priacanthidae Bigeye Priacanthus arenatus 1 Short bigeye Pristigenys alta 4, Cardinalfish Apogonidae Twospot cardinalfish Apogon pseudomaculatus 1 Tilefish Malacanthidae Blueline tilefish Caulolatilus microps 2 Sand tilefish Malacanthus plumieri 3 Cobias Rachycentridae Cobia Rachycentron canadum 13 Jacks Carangidae Bar jack Carangoides ruber 1 Blue runner Caranx crysos 7 Amberjack Seriola dumerili Almaco jack Seriola rivoliana 4, Jack species Seriola spp Snappers Lutjanidae Mahogony snapper Lutjanus mahogoni 1 Vermilion snapper Rhomboplites aurorubens Snapper species Lutjanidae 1 17

19 50-60m 61-70m 71-80m 81-90m m m m Porgies Sparidae Sheepshead porgy Calamus penna 5 Calamus species Calamus spp Butterflyfish Chaetodontidae Bank butterflyfish Prognathodes aya Reef butterflyfish Chaetodon sedentarius 3, Angelfish Pomacanthidae French angelfish Pomacanthus paru 5 Angelfish species Holacanthus spp. 3 1 Angelfish species Pomacanthus spp. 2 Damselfish Pomacentridae Yellowtail reeffish Chromis enchrysurus Chromis species Chromis spp. 1 1 Damselfish family Pomacentridae 6, 15 3 Wrasses Labridae Spotfin hogfish Bodianus pulchellus 1 Hogfish Lachnolaimus maximus 2, Wrasse species Halichoeres sp. 2 2 Razorfish species Hemipteronotus sp. 1 Barracuda Sphyraenidae Great barracuda Sphyraena barracuda 1 Hovering Gobies Ptereleotridae Blue goby Ptereleotris calliura 2 2 Hovering gobies Ptereloetris spp Surgeonfish Acanthuridae Surgeonfish species Acanthurus spp. 2 Flying Gurnards Dactylopteridae Flying gurnard Dactylopterus volitans 1 Lefteye Flounders Bothidae Flounder species Bothidae spp. 1 1 Filefish Monacanthidae Unicorn filefish Aluterus monoceros 6 4 Filefish species Aluterus sp. 2 Boxfish Ostraciidae Scrawled cowfish Acanthostracion quadricornis 3 1 Trunkfish species Lactophyrs spp. 2 1 Leatherjackets Balistidae Gray triggerfish Balistes capriscus 55, 1 10, Puffers Tetraodontidae Bandtail puffer Sphoeroides splengeri 4 1 Spiny Puffers Diodontidae Burrfish species Chilomycterus sp

20 35 30 Average species per hour Bill Boyd Caicos Express Papa's Reef Natural Habitat Figure 5. Average species richness per hour for vessel-reefs ( ) and natural substrate ( ). Bars indicate standard error Average abundance per hour Bill Boyd Caicos Express Papa's Reef Natural Habitat Figure 6. Average abundance per hour for vessel-reefs ( ) and natural substrate ( ). Bars indicate standard error. 19

21 Although there was no significant difference in species richness or abundance among depth categories (ANOVA, P>.05), there does appear to be a trend of decreasing species richness with depth (Figures 7, 8). This change is likely due to the loss of suitable habitat past 80m. Twenty two percent of the fishes observed during drift transects were associated with small artificial structures such as tires. Typically, at least one fish was associated with each small structure. This association further indicates the lack of natural habitat within the 50 to 120m depth range. Mean species richness per hour Mean abundance per hour Depth Ranges Figure 7. Mean species richness per hour by depth category Depth Ranges Figure 8. Mean abundance per hour by depth category. 20

22 Thirty-five of the forty-five species found between 50 and 120m depth were also observed at depths less than 30m. Of these fish, 18 have foraging distances potentially greater than a kilometer, suggesting that over a third of the fish in the natural substrate deep-reef community may be migrants from the nearby outer reef tract in 30m depth. The low species richness and abundance of fishes on natural substrate is comparable to other studies at similar depths in the South Atlantic Bight. Offshore of North Carolina, species richness is lower in areas of low relief when compared to areas of higher relief at a similar depth (Lindquest and Clavijo, 1993; Parker and Ross 1986). Lindquest and Clavijo (1993) recorded a species richness of 16 at depths between 96 and 109m, while Parker and Ross (1986) found 34 species at the same depth but in a region of higher relief. Additionally, in a shallower depth zone (52m to 98m) with significant hardbottom coverage and low relief, Parker and Ross (1986) recorded 83 species. Parker and Ross (1986) found 61 recreationally and commercially important fish per hectare over sandy areas versus 774 over reefs. In Broward County, other than large mixed schools of amberjack, Seriola dumerili and cobia, Rachycentron canadum, few recreationally or commercially important adult fishes were observed on natural substrate. The low numbers were probably due to the lack of relief that usually supports a diverse and abundant fish assemblage. However, other possibilities include local over-fishing on a narrow continental shelf. Unlike deep-reefs/hardbottom in the South Atlantic Bight, the 50m depth contour in Broward County is directly adjacent to a diverse tropical reef tract. With this in mind, it is surprising to find such a depauperate fish community, especially between 50 and 60m depth where algae, sponges and octocorals still survive. An examination of trophic groups reveals the community was composed of a few benthic carnivores, piscivores, and planktivores. Herbivores and omnivores were rare exceptions. Scarids and non-planktivorous pomacentrids were completely absent; and acanthurids were represented only once. The serranids, lutjanids, and haemulids were also relatively rare. These families are reef associated and without habitat structure they do not appear to be present. Eighteen species observed are found on adjacent shallow reefs. Twelve of these species have foraging ranges of at least a kilometer and likely traveled off the outer reef to forage. Seven typically deep reef species were observed; snowy grouper, Epinephelus niveatus, bank seabass, Centropristis ocyurus, tattler bass, Serranus phoebe, short bigeye, Pristigenys alta, Blueline tilefish, Caulolatilus microps, bank butterflyfish, Prognathodes aya, and an unidentified Halichoeres sp. These fish were only observed with structure except for C. microps. Although there was a general lack of natural substrate throughout the 50 to 120m depth zone, there were numerous artificial structures less than 1m 2. Some were unidentifiable objects, although the majority were tires. These tires presumably came from a project initiated in 1967 to create an artificial tire reef in 20m depth with approximately two million tires. The tires were initially bound together, but over the years they have become separated and scattered throughout the Broward County offshore environment (Sherman and Spieler 2006). Generally, at least one fish, typically a serranid, was observed on each tire. Both E. niveatus, all C. ocyurus, and most S. phoebe surveyed were found on tires. C. ocyurus is a smaller serranid that is mostly found in temperate waters from Florida to Massachusetts and throughout the Gulf of Mexico (Robins and Ray 1986). Although there 21

23 have been reports of C. ocyurus in shallower water (Human and DeLoach 2003), they are typically found at depths of 55m or greater in association with hardbottom (Robins and Ray 1986; Bullock and Smith 1991). Since this habitat type is rarely found in deeper waters (>50m) in Broward County, they appear to utilize the artificial habitat created by tires. Similar to S. phoebe, another small solitary serranid, no C. ocyurus have been recorded on the vessel-reefs. It is possible that these smaller serranids are unable to compete with larger predators found on vessel-reefs. C. ocyurus is also a benthic invertebrate carnivore (Bullock and Smith 1991), and the proximity of food items to a tire reef with less competition may be more advantageous than a neighboring vessel-reef with a larger assemblage of carnivorous competitors. Off North Carolina, Parker and Ross (1986) noted a distinct fish assemblage at depths between 34 to 98m around flat rocks less than 1m 2 in diameter, which were encrusted with crinoids, sponges, corals, and often associated with burrows or caves. Some fish typically observed on these rocks include; squirrelfish, Holocentrus ascensionis, E. niveatus, S. phoebe, sand perch, Dipletrum formosum, wrasse bass, Liopropoma eukrines, red barbier, Hemanthias vivanus, P. alt, and yellowtail reeffish Chromis enchrysurus. The small artificial structures in Broward County had a similar assemblage; (i.e., E. niveatus, S. phoebe, anthiine fishes, P. alta, and C. enchrysurus). Small artificial reefs have been deployed on the Oculina Banks to replace and help restore Oculina thicket damage from years of trawling. Observations by Koenig et al. (2005) suggest that reestablished fish populations appear to have reached historical levels on these artificial reefs. Offshore of Broward County between 50 and 120m, where habitat is limited, artificial structures may be providing habitat for fish that, although common in other regions, would be rare or possibly absent in Broward County. Both the derelict vessels and the steel totaling machines, encountered by chance during natural surveys, indicate that artificial structures provide habitat for fishes that were otherwise uncommon on natural substrate. The derelict vessel supported a community of fishes typically found on the outer reef, which is likely due to depth and close proximity of the artificial structure to the outer reef. The steel totaling machines had a community more closely resembling that of larger vessel-reefs than the natural substrate and tire reefs. Similar to studies in shallower water, artificial reef size and location appear to influence associated fish communities at depths between 50 and 120m. The fish assemblage associated with the vessel-reef was different than that of the natural substrate. Serranids, although not common on natural substrate between 50 and 120m, were the most specious family on vessel-reefs. As previously mentioned, the two serranids commonly observed during natural transects, S. phoebe and C. ocyurus, were not recorded on vessel-reefs. Instead, larger Epinephelus spp. and Mycteroperca spp. serranids were observed. The major difference between vessel-reefs and natural substrate was anthiinid abundance. Typically thousands were observed on vessel-reefs, and with the exception of one group of tires, none were recorded on natural substrate transects. Large carangids, Seriola spp., were common on vessel-reefs yet only occasionally observed on natural substrate. Fishes in the family Lutjanidae are common in waters less than 30m, but only five species were observed past 50m (mutton snapper, Lutjanus analis, blackfin snapper L utjanus buccanella, gray snapper, Lutjanus griseus, and cubera snapper, Lutjanus cyanopterus were 22

24 limited to vessel-reefs). Similarly, species in the family Haemulidae were sparse past 50m and those recorded were on vessel-reefs. Interestingly, no Pomacanthids were observed on natural substrate while they were common on vessel-reefs. A significant difference in species per hour was found between natural substrate surveys (10.4 +/ SE) and combined vessel reefs (24.3 +/ SE) (T-test p< ). There was also a significant difference in abundance per hour between natural substrate (62.0 +/ SE) and vessel-reef surveys ( / SE (with anthiines fish removed)) (T-test, p< on log10(x) transformed abundance values). During 39 months of study (between 2004 and 2007) 50,376 fishes from 22 different families were counted (Table 3). The majority of these fishes were anthiinids with a total of 44,209 individuals counted. The 6,167 other fishes represented 65 different species. There were no significant differences among vessel reefs in species observed per hour (ANOVA, p= 0.38) and in general the fish assemblages were similar among vessel-reefs. On the Bill Boyd, 40 different species were observed during nine surveys. Forty-nine species were observed on the Caicos Express during nine surveys and 41 species were observed on Papa s Reef during six surveys. Among vessel-reefs there was no significant difference in abundance of fish observed per hour of survey (ANOVA, p=0.18, log10(x) transformed data). There was no significant difference in abundance between seasons among all vessel-reefs (T-test, p=0.88, log10(x) transformed data). Varying effort between vessel-reef surveys confounded rigorous statistical analysis of individual species differences between vessel reefs. However several qualitative differences were observed and may be attributed to differences in vessel-reef depth and structure. The planktivorous anthiines fishes numerically dominated all vessel-reefs, however there appeared to be a larger school associated with the two deeper vessels, Bill Boyd and Papa s Reef, in comparison to Caicos Express. The Caicos Express in return generally had a higher number of Chromis species as well as tomtates, Haemulon aurolineatum, a facultative planktivore. Previous research by Arena (2007) described the importance of planktivores on vessel-reefs in 20m depth and this importance seems to continue to vessel-reefs located below 50m depth. The Caicos Express, located at 74m depth, appears to represent a transition from shallower-water planktivores, such as Chromis species and H. aurolineatum, to the predominately deeper-water anthiines fishes. Large serranids are uncommon in Broward County on natural reefs and vessel-reefs at depths less than 30m, as well as on natural substrate between 50 and 120m (Ferro et al. 2005, Spieler and Bryan 2006, Arena 2007). It was hypothesized by the authors that the deeper vessel-reefs may provide a refuge from fishing pressure and hence a higher occurrence of larger serranids. One hundred and twenty-two scamp, Mycteroperca phenax, were observed during a combined 22.2 hours of survey. With an average of 5.5 observed per hour, M. phenax was the most common serranid observed on the vessel-reefs other than the smaller anthiines fishes. M. phenax are not common in shallower water and were absent from the deeper water natural substrate. The vessel reefs therefore appear to be providing suitable habitat for M. phenax in Broward County. 23

25 The following serranids were also observed on vessel-reefs but in lower numbers. One goliath grouper, Epinephelus itajara, was observed on the Caicos Express. Three adult warsaw grouper, Epinephelus nigritus, were recorded on two vessel-reefs. Several speckled hind, Epinephelus drummondhayi, were observed on the Bill Boyd during repeated surveys and two were observed on the Caicos Express. Typically found in deeper water, this was the first time this species has been observed in Broward County. Black, Mycteroperca bonaci, and Gag, Mycteroperca microlepis, grouper were infrequently observed during vessel-reef surveys. Six juvenile/small adult E. niveatus, were recorded on Papa s Reef and one was observed on Caicos Express. Juvenile E. niveatus were also observed in association with tires during natural substrate surveys. These fish were typically larger than the E. niveatus observed in association with vessel-reefs and other artificial debris in shallower water in Broward County (Arena et al and Bryan pers. obs.). E. niveatus adults are found at depths between 100 and 400m throughout the Atlantic and Gulf of Mexico continental slopes of United States, as well as Central and South America (Hemestra et al. 2002). No adult fish were recorded on the vessel-reefs in this study; however, in the lower Florida Keys, adult E. niveatus are commercially caught between 120 and 260m depth (Moore and Labisky 1984; Bielsa and Labisky 1987). They appear to only be found near bottom structure or within a close foraging distance (Dodrill et al. 1993) and have a positive trend of increasing size with depth (Wyanski et al. 2000). It is possible that vessel-reefs as well as tires serve as an important transitional habitat for E. niveatus as they migrate from nearshore artificial structures to deeper water as adults. L. buccanella and L. griseus, were two economically important lutjanids commonly observed on vessel-reefs. With an average of 27 observed per survey, the Bill Boyd had the greatest abundance of L. griseus. Vermilion snapper, Rhomboplites aurorubens, also an economically important species, were commonly observed on the Bill Boyd with an average school of 91 fish per survey. These fish were typically observed several meters above the vessel-reef and often in association with the extended superstructure. An average of 40 R. aurorubens were observed on Papa s Reef and none were recorded on the Caicos Express. During the last two surveys on the Caicos Express in 2007 a small school (34 fish) of lane snapper, Lutjanus synagris, were observed for the first time. L. synagris recruit to the nearshore habitat in Broward County and these fish may represent a pulse offshore from a strong 2006 year class (Spieler unpub. data). Both the deep water P. aya, and shallower reef butterflyfish, Chaetodon sedentarius, were common on each vessel-reef; however there appeared to be a higher abundance of C. sedentarius and lower number of P. aya on the shallower Caicos Express. The french butterflyfish, Prognathodes guyanensis, was recorded seven times on Papa s reef and once on Bill Boyd. This Caribbean species has only recently been described in continental United States waters and this is the first record of the fish in Broward County (Quattrini 2004). Numerous personal observations of charter boats fishing near vessel-reefs suggest that Carangids are the most frequently caught. Two species observed on almost every survey were S. dumerili, and almaco jack, Seriola rivoliana. Despite their occurrence on each 24

26 vessel-reef, they both appeared to be more common on the Bill Boyd. The Bill Boyd also had significantly greater superstructure when compared to the other vessel-reefs and both Seriola sp. were typically observed above the vessel-reef and in possible association with the superstructure. The dramatic difference in abundance between the natural substrate and the vessel-reefs in 50 to 120m depth is primarily due to thousands of anthiine fishes that were found on each vessel-reef. These smaller planktivores seem to thrive on higher relief and more complex habitat provided by vessel-reefs. Individual species abundances could not be determined, however it appeared that most schools were a mixture primarily of roughtongue bass, Pronotogrammus martinicensis, and a more slender anthiinid, possibly threadnose bass, Anthias tenuis. In the Northern Gulf of Mexico, North Carolina, and Oculina Banks, two anthiine fishes, the H. vivanus, and P. martinicensis, numerically dominate the fish community over deep high profile reefs (Linquest and Clavijo 1993; Koenig et al. 2000; Weaver et al. 2001). Determining the actual abundances of anthiine fishes was difficult due to their small size and tendency to dart in and out of the skin of the ship, which provided refuge. Although abundance among vessel-reefs was variable through time, there appeared to be more anthiines on both Bill Boyd and Papa s Reef than Caicos Express. Anthiine abundance is also in striking contrast to community structure on shallow vessel-reefs. Although planktivores are the most abundant trophic group on shallow water vessel-reefs, their numbers do not approach those of deep vessel-reefs (Arena et al. 2007). Planktivores are considered an important trophic link in tropical reef ecosystems (Hobson, 1991). It appears that the same relationship also exists at depths greater than 50m. Weaver et al. (2001) considered planktivorous anthiine fishes to be a keystone species on the Pinnacles reef system in Northern Gulf of Mexico, and they are also an important link in the food web of the Charleston Bump on the southeastern continental slope of the United States (Weaver and Sedberry 2001). Weaver et al. (2001) suggested that the large abundance of fish found at these intermediate depths are supported through the transfer of food from plankton to smaller fish. The high percentage of piscivores on vessel-reefs is likely due to the abundance of anthiine fishes. During one survey S. rivoliana were videoed feeding on anthiines fishes on the deck of Bill Boyd. 25

27 Table 3. Vessel-reef fish abundances ( ). Common Name Scientific Name Bill Boyd Caicos Express Papa's Reef Squirrelfish Holocentridae Squirrelfish Holocentrus adscensionis 2 18 Cardinal soldierfish Plectrypops retrospinis 3 2 Squirrelfish species Holocentrus sp. 2 8 Squirrelfish family Holocentridae Sea Basses Serranidae Graysby Cephalophis cruentata 1 Speckled hind Epinephelus drummondhayi 12 2 Goliath grouper Epinephelus itajara 1 Warsaw grouper Epinephelus nigritus 1 2 Snowy grouper Epinephelus niveatus 1 6 Black grouper Mycteroperca bonaci Gag grouper Mycteroperca microlepis 2 1 Yellowmouth grouper Mycteroperca interstitialis 1 Scamp Mycteroperca phenax Creole-fish Paranthias furcifer 37 Wrasse bass Liopropoma eukrines Spanish flag Gonioplectrus hispanus Roughtongue bass Pronotogrammus martinicensis Anthiine fishes Anthinids Grouper species Mycteroperca spp Grouper family Serranidae 1 Soapfish Grammistidae Greater soapfish Rypticus saponaceus Soapfish species Rypticus spp. 1 1 Cardinalfish Apogonidae Bigtooth cardinalfish Apogon affinis 4 Twospot cardinalfish Apogon pseudomaculatus 2 Cobias Rachycentridae Cobia Rachycentron canadum 4 Remoras Echeneididae Sharksucker Echeneis naucrates 1 Jacks Carangidae Bar jack Carangoides ruber Blue runner Caranx crysos 1 Amberjack Seriola dumerili Almaco jack Seriola rivoliana Permit Trachinotus falcatus 1 Jack species Seriola spp Jack species Carangid spp. 1 6 Snappers Lutjanidae Black snapper Apsilus dentatus 1 Mutton snapper Lutjanus analis Blackfin snapper Lutjanus buccanella Lane snapper Lutjanus synagris 67 Gray snapper Lutjanus griseus