120 



Fishery Bulletin 94(1), 1996 



Q 



III 

 20 



M) 

 40 



su 

 Ml 

 -II 



(moo 0.050 oioo 



Fish/m 3 

 Figure 5 



Examples of the density and distribution offish schools recorded by 

 sonar at stations 1-3 during acoustic surveys conducted from Sep- 

 tember through October 1991, off Newport, Oregon. Bottom depth 

 is depicted by the solid line and fish densities are indicated by shaded 

 areas. 



species of rockfishes. Submersible observations con- 

 tributed greatly to the acoustic surveys because spe- 

 cies composition of large schools could be character- 

 ized once they had been located acoustically. 



Acoustic surveys enabled us to quantify the vertical 

 distribution of fishes and to determine the proportion 

 of fishes not detected by the submersible surveys. Al- 

 though several researchers have noticed an avoidance 

 of survey vessels by mid-water fishes (e.g. Olsen et al., 

 1983; Misund, 1990; Ona and God0, 1990), we observed 

 no difference between acoustic surveys conducted be- 

 fore and after submersible transects. The affinity that 

 most rockfishes have for bottom features may make 



them less likely to actively avoid a vessel pass- 

 ing overhead. At all stations, fishes recorded 

 by acoustic equipment were located primarily 

 in the lower third of the water column, often 

 above rock scarps or pinnacles. This pattern 

 occurred in all acoustic transects and is simi- 

 lar to observations of fish density increasing 

 with depth for both shallow and deep water 

 rockfishes (Hallacher and Roberts, 1985; 

 Leaman et al., 1990; Love et al., 1991; Richards 

 etal, 1991; Sullivan, 1991). 



The acoustic estimate of mean areal den- 

 sity in the "above sub" stratum was 2.3 times 

 greater than the mean submersible estimate 

 of fish density. Submersible surveys, there- 

 fore, accounted for almost one-third of the 

 fishes inhabiting Stonewall Bank during the 

 time of the study. However, the mean den- 

 sity estimates of both submersible and acous- 

 tic surveys were each greatly affected by one 

 transect. Submersible and "above sub" acous- 

 tic estimates offish density were more simi- 

 lar when two transects that skewed the mean 

 were removed from analysis (Table 3). On 

 dive 2604 at station 2, the submersible ob- 

 server recorded several large schools of ju- 

 venile rockfishes. Similarly, large, patchy 

 schools were detected on the accompanying 

 acoustic transect. We assumed that the high 

 fish densities associated with these transects 

 were due to the presence of schools of juve- 

 nile rockfishes. After removing these data 

 from the analysis, density estimates from the 

 submersible transects were 14% higher than 

 those generated acoustically for the "above 

 sub" stratum, suggesting that the relative 

 abundance of adult rockfishes was approxi- 

 mately equal above and below the level of the 

 submersible. 



The estimates of fish density generated 

 from submersible transects were more than 

 six times those from acoustic surveys for the 

 same depth stratum. This difference reflects the in- 

 herent limitations of acoustic sampling in rocky, high- 

 relief locations and demonstrates the influence of 

 sampling method on survey results. Pearcy et al. 

 ( 1989) and Stein et al. ( 1992) observed rockfish spe- 

 cies over a variety of benthic habitats ranging from 

 smooth mud to high-relief rock and also reported 

 schools of rockfishes in the water column. Given the 

 variety of habitats in which rockfishes are located, it 

 appears that several sampling tools are needed to 

 assess their relative abundance. 



Trawls have historically been used to assess fish 

 abundance over smooth bottoms. Weinberg ( 1994), 



