METHODS

Cruise Activities

The initial reconnaissance and sampling cruise USGS 97-01 was conducted on the R/V SUNCOASTER 4-13 August 1997 (92 stations). The remaining six USGS-funded cruises (234 stations) were completed on the R/V TOMMY MUNRO (Table 1.2). The combined sampling effort for all cruises comprised 326 stations, apportioned into 112 angling, 63 trap, 22 bottom trawl, 58 remotely operated vehicle (ROV), 15 dredge/core/grab, and 37 plankton stations, with a small number of stations representing bottom surveys, CTD casts, light meter deployments, and exploratory sampling gear (Table 1.2). All sampling combined returned over 6,000 specimens for food habits analyses, taxonomic verification and documentation, and subsequent life history analyses.  Shipboard photographs documenting 113 species were obtained. ROV tapes resulted in about 85 hours of observation of reef and reef-associated biotopes and the resident fish fauna for analysis.

Sampling was conducted during all seven research cruises spanning the time interval of 1997-2000 (Table 1.2). Positions for all sampling sites were determined by Differential Global Positioning System (DGPS) and charts were provided by CSA-TAMU collaborators. Sampling conducted during the 2000-01 cruise was further georeferenced using digital maps of HRMBS topography and vessel position tracked in ArcView software to display station location (Fig. 1.14). Operations were conducted continuously 24 hours per day.

Hook and Line Angling

During all cruises, demersal and pelagic fishes from hard-bottom reefs were obtained using hook and line angling with bait and large hooks (for macro-carnivores) and angling with unbaited, multi-hook "Sabiki" jigs (for planktivores and micro-carnivores) (Fig. 1.15). Sabiki jigs are constructed of a variety of small hook sizes (5 to 20 mm in length) with added strips of iridescent film to imitate shrimps or other small planktonic organisms, and each pre-packaged unit contains from 4 to 7 hooks.  Sabiki jigs were deployed using 142 to 340 g lead sinkers, depending upon water depth and currents, and were successfully used to collect fishes to 110 m depth. Fishes were primarily collected during daylight hours, as the majority of individuals captured using this technique were visual planktivores.

Traps

A variety of sizes and mesh traps including large (1.8 m X 1.5 m X 0.6 m) chevron traps with 25 mm X 12 mm mesh to box traps (1.2 m X 1.2 m X 0.6 m) of 25 mm hexagonal wire were used to obtain fish specimens to confirm visual identification and sample cryptic taxa not easily observed by the ROV.   Some traps were equipped with 3 or 6 mm Vexar mesh liner to retain small specimens.  Traps were deployed in trap lines made up of two to four traps. Also attached to these lines were smaller collection devices such as minnow traps, PVC bundles, and bottles in an attempt to attract and capture small shelter seeking fishes.

Trawls

A 4.9 m semi-balloon trawl with a 3.8 cm mesh body and a 0.6 cm mesh liner was used to collect bottom fishes near reefs at 13 stations and a 7.6 m semi-balloon trawl with a 3.8 cm mesh body and a 0.6 cm mesh liner at five stations. Both trawls were equipped with a "Texas" roller rig made up of 7.6 cm rubber disks on a chain in front of the head rope to reduce hangs on reef structures.

Specimen Disposition

In-situ collections provided voucher specimens to verify and document species identifications, and material for food habits studies.  On board the research vessel, individual voucher fish specimens were carefully preserved, pinned out to display coloration and diagnostic features, and photographed using digital and 35 mm cameras on a water bath light table (Randall 1961).  Photographs were taken by D. Weaver, W. Smith-Vaniz, and J. Caruso. Specimens were preserved in 10% buffered formalin or frozen in the field, and later transferred to 70% ethanol in the laboratory. Selected voucher specimens were catalogued in the National Museum of Natural History (NMNH) and Florida Museum of Natural History (FMNH) ichthyology collections. The remaining voucher material, metadata, and voucher photographs are maintained at the Florida Integrated Science Center (FISC).

Other gears

Additional shipboard activities included plankton sampling to document pelagic prey availability to the deep reef fish community, and collection of physicochemical parameters of the water column.  These data will be reported elsewhere.

ROV Methodology

Collection of specimens was complemented by use of ROVs to conduct habitat reconnaissance and document faunal composition via color video camera.  All dives were recorded to videotape for subsequent laboratory analyses.  A Phantom DS4 ROV, equipped with a color video camera, was the primary system used for in-situ observations.  Three different Phantom ROVs were employed during the seven Pinnacles research cruises.  United States Navy (97-01) and NMFS (99-03) ROVs were operated by NMFS Mississippi Laboratories during the first three years of the program, and the NOAA/NURP Phantom DS4 was operated by NURC-UNCW on the remaining cruises.  Due to technical difficulties and image quality, only ROV data from cruises 97-01, 99-03, and 2000-01 were used for quantitative purposes (Appendix A). ROV observations were used to define and compare fish faunal composition at each main target study reef by depth, reef height (high, medium and low profile), and biotope.

Table 1.2. Summary of research activities conducted from 1997-2000. Number of stations listed by study site and sampling gear.

Table 1.2. (continued).

Table 1.2. (continued).

Figure 1.14. Distribution of sampling effort at Roughtongue Reef (NEGOM-CMEP Site 1) for cruises 97-01 through 2000-01. Each symbol represents the position of the research vessel during sampling gear deployment. Gray circles-angling, black triangles-traps, open squares-ROV dives.
 

Figure 1.15. Sabiki jigs used to collect small reef fishes during this study, including Pronotogrammus martinicensis (top and middle) and Hemanthias vivanus (bottom).
 

Transecting methodology were drawn from those previously employed in ROV/ submersible/camera sled studies that have shown that videotape analysis of large mobile demersal fishes requires a moving image and a low oblique perspective to capture diagnostic features of morphology and coloration (Grassle et al. 1975, Cohen and Pawson 1977, Uzmann et al. 1977, Rice et al. 1979, Parker and Ross 1986, Hecker 1987, Butler et al. 1991, Sulak and Ross 1993, 1996, Adams et al. 1995, Felley and Vecchione 1995). ROV transects were conducted at a constant speed of 0.100.15 m/sec across the bottom, with altitude of the ROV held to a maximum of 1.0 m (distance between sea floor and video camera lens, varying somewhat according to terrain). Horizontal swath width was resolved as the lower border of the video frame with the video camera zoom set to full wideangle, and pan and tilt angles set to about 25 and 20, respectively.  Video was recorded on board the ship on S-VHS or Hi-8 videotape with audio annotation.

In situ collections of small reef fishes were accomplished using a suction sampler fitted to the ROV (Fig. 1.16) that dispensed a suspended rotenone solution (co-designed by D. Weaver and L. Horn, UNC-Wilmington).  Powdered rotenone (Prentiss, Inc.) was suspended in 3 L of concentrated Ivory Liquid dishwashing detergent, and mixed with 8 L of seawater. Approximately 3 kg of standard table salt and 10 kg of granular sugar was added to increase the specific gravity of the solution to help maintain contact with the bottom. The suspended rotenone was passed through a standard kitchen strainer to remove large clumps, and loaded into the dual reservoir chambers of the sampler. A standard marine bilge pump, encased in an aluminum housing and powered from the surface, was used to dispense the rotenone solution and collect fishes at select reef sites during ROV dives.  Fishes were retained in a 5 mm mesh liner fitted inside the first PVC chamber (Fig. 1.16).

Videotape Analysis

Videotapes were transferred to S-VHS format and analyzed on a commercial quality S-VHS VCR with variable speed editor/jogger control (slow-motion, frame by frame advance) to facilitate species identification at slow speed advance and enumerate individuals.  Videotape analysis was used to quantify fish abundance by taxa and biotope.  Incidental observations of pelagic fishes also were recorded and individuals identified.  In addition to moving videotape transects, the zoom function of the NURC-UNCW ROV video camera was periodically used to capture close-up images of individuals to facilitate identification during cruises 2000-01 and 2000-02.  These still video segments were used to verify species identifications made during transect surveys.   All fish species were identified to the lowest possible taxonomic unit, with confirmation from close-up video images and by reference to voucher specimens collected by angling, trapping, or ROV suction sampler.

Habitat Parameters

In addition to defining fish community structure, videotape screening was used to make qualitative notes on the physical habitat and megafaunal invertebrates associated with particular biotopes and fish assemblages. Where appropriate (e.g., high profile, flat topped features) fish observed were assigned to one of six biotopes: reef top, reef crest, reef base, reef talus (circum-reef sediment apron), and soft bottom.
 

Figure 1.16. A. Suction sampler device schematic.  B. NURC-UNCW Phantom DS4 ROV with suction sampler.  C. A typical collection of fishes made by the suction sampler.
 

Reef Top is referred to as flat reef tops or reef flats by Gittings et al. (1992a) and describes the biotope that develops on large, flat-topped mounds such as Roughtongue Reef, Yellowtail Reef, and CMEP-Site 5. This biotope is identified by extensive pockets of accumulated sediments and rocky debris, interspersed with low profile rock outcrops, ridges, and crevices. Invertebrate assemblages on the reef flat top community exhibit high density and species richness, and are dominated by erect sponges, octocorals (particularly sea fans, including Nicella sp.), comatulid crinoids, antipatharians, gorgonocephalid basket stars, bryozoans, and coralline algae (Gittings et al. 1992a, Hardin et al. 2001).  Our reef top category corresponds to top interior locations of Hardin et al. (2001) and continuous hard bottom category of Snyder (2001).

Reef Crest biotopes were defined as the interface between the near vertical, rocky reef face and the sediment covered flat top characterized by dense assemblages of invertebrates. Reef crest biotopes typically were characterized by extensive rocky outcrops with small areas of sediment cover and low invertebrate densities.  We distinguished the reef crest ecotone from the adjacent flat reef top and vertical reef face biotopes to identify the possible influence of position in the leading edge of water currents on the reef fish community. Our reef crest category corresponds to the top edge location listed by Hardin et al. (2001).

Reef Face as referred to by Gittings et al. (1992a), are rugged, often near-vertical rocky surfaces that characterize the steep-sided edges of large flat-topped features and the steep walls of narrow, spire shaped deep water pinnacles. Reef face communities are characterized by a lower overall density of epifauna than reef tops, with an abundance of ahermatypic corals, including R. manuelensis, Madrepora sp., and Madracis/Oculina sp., comatulid crinoids, a variety of octocoral fans, antipatharians including the spiral whip Stichopathes lutkeni Brooks, 1889, coralline algae (to about 75 m), and sea urchins (Gittings et al. 1992a, Hardin et al. 2001). This biotope is described as reef side by Hardin et al. (2001) and Snyder (2001).

Reef Base was defined as the ecotone between the steep reef face and the talus zone, where the rugged rocky face of the reef was often undercut and formed small caves and overhangs. This region contained both rocky vertical faces with abundant solitary corals and the coarse sediments and rocky debris/talus resulting from bioerosion occurring on the reef above. This biotope is also identified by Hardin et al. (2001) and Snyder (2001).

Reef Talus (circum-reef sediment apron) is listed as debris fields or rubble flats by Gittings et al. (1991, 1992a). These are flat areas of reef debris and coarse carbonate sediments occurring on the flanks of large, high-relief mounds (Gittings et al. 1992a, Sager and Schroeder 2001). Coarse sediments and debris are produced by shell material resulting from bioerosion, and rock fragments from the main reef.  Small rocky outcrops in this biotope are often encrusted with solitary corals, small octocoral fans, and antipatharians (Gittings et al. 1992a, Hardin et al. 2001).  The reef talus zone is identifiable on HRMBS maps as the area of highest acoustic backscatter (red) to the south and west of hard bottom features (Gardner et al. 2000) (Fig. 1.5B).

Soft Bottom/Sand Plain refers to the relatively flat substrate surrounding reef features, characterized by fine to coarse quartz sand without visible portions of shell hash or rock rubble.  These areas are often flat and featureless, but occasionally are contoured by ripples, gentle waves, or numerous excavated burrows, pits, and mounds that add to the topography.  Sessile invertebrates in this habitat are limited to occasional small octocorals or antipatharians that attach to buried rock surfaces. Non-reef associated soft bottom is distinguishable in acoustic backscatter maps as areas of lowest backscatter (Fig. 1.5B, blue).
 

Data Analysis

Survey Period and Reef Tract Comparisons

ROV dives were divided into day and night periods, and data from crepuscular periods were excluded from this analysis.  A two-way chi-square goodness of fit test was used to compare abundance between survey period and reef tract.  A sequential Bonferroni procedure was used to adjust significance values for multiple tests (Sokal and Rohlf, 1995). At five reefs, Alabama Alps, Ludwig & Walton Pinnacles, Scamp, Yellowtail, and Roughtongue Reefs, both day and night data were obtained by ROV.  Although sample sizes were small, a paired t-test was used to test day-night differences in total fish abundance, number of taxa, and top six taxa abundance (Sokal and Rohlf, 1995).  All data were standardized by effort measured as video survey time in minutes.

Cluster Analysis

We used cluster analysis to help explore and define the structure of fish assemblages among reefs and biotopes.  Data were summarized by reef and survey period. Data from Ludwick and Walton Pinnacles 1, 2, A, B and C were combined into a single reef category.  Raw counts were standardized by effort expressed as number of fish per minute of video survey time for comparisons among reefs. Survey time was not recorded by biotope so data were double standardized for analysis. Taxa were restricted to those that occurred at two or more sites to reduce the potential effect of bias from rare species (Wolda, 1981). Two measures of similiarity, percent similarity and Morisita, were used for comparison as they emphasize different attributes (Bloom, 1981; Wolda, 1981). An agglomerative hierarchical clustering strategy was used with an unweighted pair-group mean average sorting algorithm to obtain grouping of reefs and taxa (Boesch, 1977). The COMPAH computer program was used to perform the analysis (Gallagher, 1998). Clusters were subjectively divided into groups to facilitate comparisons.  We compared cluters from the two similarity measures. If a reef or taxa was found in different groups between clusters it was considered poorly resolved and a weak member of the group; thus not indicative of the group structure.
 

Both sites (reefs) and taxa were clustered and nodal analysis was performed to enhance the interpretation of the data (Boesch, 1977). The relationship between reefs and taxa groups was measured as constancy, the faithfulness of a taxa group to a particular reef group and fidelity, the degree to which taxa are limited to reef groups.  Constancy and fidelity were arbitrarily categorized from high to low to better elucidate relationships between reefs and taxa.
 

Next Page