266 



Fishery Bulletin 97(2), 1999 



long, dives to 365 m, and travels 2-6 km/h for 2-4 h. 

 It is equipped with halogen lights, internal and ex- 

 ternal video cameras, magnetic compass, directional 

 gyro compass, underwater telephone, and transpon- 

 der that allow tracking of the submersible from a sur- 

 face vessel. On each dive, the submersible descended 

 to 265-365 m and then usually traveled parallel to the 

 shelf break, followed by up-slope travel to less than 

 300 m before it ascended. The surface vessel recorded 

 LORAN fixes at the beginning and end of a dive, and 

 every 1-5 min during a dive. The submersible traveled 

 1.0-2.5 km/h, depending on the slope and ruggedness 

 of the terrain and the magnitude and direction of the 

 current. The degree of slope determined how observa- 

 tions were made from the submersible. When the slope 

 was less than 60°, the submersible remained in con- 

 tact with the seafloor while the scientist viewed the 

 water column parallel to the seafioor through a star- 

 board porthole 0.5 m above the base of the submers- 

 ible. When the slope was greater than 60°, the sub- 

 mersible traveled 2-3 m away from the seafloor while 

 the scientist viewed the water column almost per- 

 pendicular to the seafloor through portholes on the 

 starboard side and bow. The pilot sat above the ob- 

 server in a tower with a panoramic view and assisted 

 in locating fish, especially above the submersible. The 

 submersible lights provided constant illumination. 



Data analysis 



Observations of rockfish and their habitat were audio- 

 tape- and videotape-recorded for subsequent analysis 

 and verification. The senior author reviewed all video 

 tapes by 1-min segments ( 16-42 m travel distances ), 

 and four habitat parameters were estimated: 1 ) main 

 substrate, 2) secondary substrate, 3) slope, and 4) 

 boulder abundance. The main substrate made up 50- 

 100% of the substrate, whereas the secondary sub- 

 strate made up 10-50'^ of the substrate. Substrates 

 consisted of mud, sand, pebble, cobble, and bedrock. 

 Granular size used to separate pebble and cobble was 

 2.5 inches (64 mm), and to separate cobble and boul- 

 ders was 10 inches (256 mm). Mud would stay sus- 

 pended when disturbed by the submersible, whereas 

 sand would not. Size references for classifying pebble, 

 cobble, and boulders included the known length and 

 width of a video frame as well as the known, uni- 

 form size of invertebrates such as sea stars and 

 shrimp. The slope was classified into four categories: 

 1 = 0-5°, 2 = 6-20 \ 3 = 21-45°, and 4 = 46-90 . Slope 

 classification was based on estimates by the pilot and 

 on the view from a downward-aimed, mounted video 

 camera. Boulder abundance was classified into five 

 categories: - absent, 1 = scarce, 2 - scattered 

 patches, 3 = common (usually in view), and 4 = abun- 



dant (always in view). For each site, we calculated 

 the average slope, average boulder abundance, and 

 the densities of rockfish associated with each sub- 

 strate. For all sites combined, we calculated rock- 

 fish densities associated with each substrate, and the 

 percentage of rockfish associated with each slope 

 category and each boulder category. 



Rockfish observations included number, size, 

 grouping behavior, above-bottom distribution, and 

 movements. Three sizes of rockfish were estimated 

 visually: small (<30 cm), medium (30-60 cm), and 

 large (>60 cm). The observers had used laser beams 

 to measure rockfish lengths from submersibles and 

 were, therefore, experienced in sizing them. Rock- 

 fish were considered grouped if they were within 5 

 m of each other. Densities were estimated from counts 

 of rockfish and the total seafloor area surveyed. The 

 surveyed area was the distance the submersible trav- 

 eled (0.3-1.8 kin/dive) multiplied by the estimated 

 viewing distance from the submersible (4-10 m, de- 

 pending on water clarity). These estimates were cali- 

 brated against sonar readouts when the seafloor be- 

 came visible during descents and when the seafloor 

 disappeared from view during ascents. Estimated 

 distances were within 1 m of true distances accord- 

 ing to sonar readouts and distance calibrations in 

 previous studies (Krieger, 1993). Rockfish movement 

 rates were based on the estimated distance moved 

 during a specific time period. 



Results 



Submersible dives 



Fifteen submersible dives were completed and 

 104.900 m- of seafloor was surveyed at 262-365 m 

 depths (Table 1). Rockfish densities ranged from 1.2 

 to 14.8 rockfisli/1000 m" (mean, 5.8/1000 m-) at the 14 

 sites where they were observed. Of the 646 rockfish 

 observed, 188 were small, 289 medium, and 169 large. 



Above-bottom and on-the-bottom behavior 



We observed 115 rockfish 1-10 m above bottom and 

 531 on the bottom (Table 2). Sites 5 (80 above-bot- 

 tom rockfish) and 13 (19 above-bottom rockfish) ac- 

 counted for 86% of the above-bottom rockfish. Above- 

 bottom rockfish were medium-size (108 rockfish) or 

 large (7 rockfish), and they were observed descend- 

 ing at less than 10 m/min without detectable move- 

 ments of fins or body. They would contact the seafloor 

 without disturbing sediments and would orient broad- 

 side to the current, which would tilt them 10-45° 

 (Fig. 2). Currents were less than 1.0 km/h at all sites. 



