Harter et al : Assessment of fish populations and habitat on Oculina Bank 
197 
role of deep water corals. Several studies have concluded 
that deep water corals were no more important to fishes 
than other reef structures (Auster, 2005; Tissot et al., 
2006) suggesting an opportunistic fish association with 
deep corals. Ross and Quattrini (2007), however, found 
that deep reef habitats along the southeast United 
States slope contain a unique and possibly obligate 
assemblage of fish. No previous studies have examined 
whether O. varicosa supports a distinct assemblage of 
fish compared to other structure-forming, hardbottom 
habitats. 
In 2014, the South Atlantic Fishery Management 
Council will re-evaluate the effectiveness of the OECA. 
To aid the Council in making future management deci- 
sions, our goals for this project were to (1) compare fish 
assemblage composition, biodiversity, and grouper densi- 
ties among hardbottom reef habitat types to examine 
whether O. varicosa is an essential habitat structure 
compared to other structure-forming reef habitats; (2) 
compare fish assemblage composition, biodiversity, and 
grouper densities inside and outside managed areas to 
assess the effectiveness of the MPA; and (3) quantify 
the percent cover of all hardbottom habitat types. 
Materials and methods 
Sampling design 
In 2002 and 2005, multibeam maps (3-m resolution) were 
produced for a portion of the Oculina Bank. Coverage 
included 90% of O. varicosa bioherms thought to occur 
inside the OHAPC, and a portion of bioherms outside the 
OHAPC between the two satellite areas (Fig. 1). These 
maps were used to select remotely operated vehicle 
(ROV) transect stations (April-May 2003, October 2005) 
so that all habitat types and management areas were 
examined. Management areas sampled included open 
(any area outside the OHAPC open to fishing), OHAPC 
(areas where all bottom gear except hook and line are 
restricted, i.e., excluding the OECA), and OECA (inside 
the MPA where all bottom gear, including hook and line 
fishing, are restricted). 
Locations of ROV dive transects were non-random 
and were based on conducting an equal number of dives 
in each management area. Due to high current speeds, 
all dives were conducted in a northerly direction (drift- 
ing with prevailing Gulf Stream current with minimal 
east-west maneuvering). The starting points were cho- 
sen a priori in order to have each dive cover a range of 
the major substrate types (described below) as indicated 
from the multibeam maps. Dives ranged from 0.5 to 
3.5 hours. 
In addition to management area, fish assemblages 
were analyzed among five major hardbottom habitat 
types. Habitat types used were a subset of the Southeast 
Area Monitoring and Assessment Program (SEAMAP) 
habitat classification scheme and included pavement, 
rubble, rock outcrops, standing dead O. varicosa coral, 
and live O. varicosa coral. One difference between our 
habitat classification and that of SEAMAP is that we 
distinguished between live and dead coral. Pavement 
habitat was fairly flat rock pavement often with small 
cracks or crevices present. Rubble habitat consisted of 
small coral fragments exhibiting little to no relief. Rock 
outcrop habitat was small rock outcrops approximately 
0.3-0. 9 m relief, occasionally 1.2-1. 8 m relief. O. vari- 
cosa existed mostly as small individual heads (about 
0.3-0. 9 m relief), but occasionally as larger mounds 
and thickets. 
Collection methods 
The Phantom Spectrum II ROV (National Undersea 
Research Center, University of North Carolina at Wilm- 
ington) was used to conduct video and digital still tran- 
sects to estimate fish densities and characterize habitat. 
A downweight (~145 kg) was tethered to the umbilical 
cable of the ROV and the ROV was tethered to a 30-m 
leash, which allowed it to run just above the seafloor 
(<1 m) at a controlled over-the-ground speed of approxi- 
mately 0.39 m/s (range 0.26 to 0.77 m/s). The geographic 
position of the ROV was constantly recorded throughout 
each dive using a slant range positioning system linked 
to the ship’s Global Positioning System (GPS). The ROV 
was equipped with lights, lasers, forward-looking video 
camera, and down-looking still camera. Lasers projected 
parallel beams 10 cm apart for measuring fish and 
habitat features. The forward-looking color video camera 
provided continuous video while the down-looking high- 
resolution digital still camera captured images of fish 
and habitat. 
Fish population analyses 
Fishes were identified to the lowest discernable taxonomic 
level and counted and the habitat types were classified 
from video covering 50-m (±2.5 m) transects. Excluded 
from the analysis were sections of video recorded when 
the ROV was in non-hardbottom habitats, video clouded 
by stirred up sediment, video that zoomed in on a spe- 
cies of interest, or video recorded when the camera was 
elevated in the water column. 
Fish densities (numbers/hectare) were determined 
by estimating the area viewed during video transects 
from transect length (L) and width (W). Transect length 
was calculated from latitude and longitude recorded by 
the ROV tracking system. Width of each transect was 
calculated using the following equation: 
W = 2(tan(V2A))D, (1) 
where A - horizontal angle of view (a constant property 
of the video camera); and 
D = distance from the camera at which fishes 
could be identified with certainty. 
D was usually 5 m except for some dives in 2005 where 
visibility was reduced to 2-3 m. In 2003, a set of three 
lasers was mounted to the ROV. The lasers were set up 
