Fishery Bulletin 94(1), 1996 



gill nets but have detected large biases in species 

 and size composition of the catches (Matthews and 

 Richards, 1991). 



Acoustic surveys have proven useful in estimat- 

 ing distribution and abundance of many species of 

 fishes in mid-water (e.g. seeKarp, 1990), but the tech- 

 niques require some method of validation (Johan- 

 nesson and Mitson, 1983). Recently, there has been 

 an interest in developing acoustic methods to sur- 

 vey fishes in rocky areas (Leaman et al., 1990; 

 Richards et al., 1991; Kieser et al., 1993; Phillips, 

 1994). Assessing fish near the bottom, however, is 

 difficult with acoustic techniques (Mitson, 1983), 

 particularly over rocky habitats. Most echo integra- 

 tors have difficulty distinguishing between targets 

 near the bottom and bottom echoes in high-relief ter- 

 rain (Burczynski, 1979). The acoustic shadowing that 

 occurs in high-relief terrain also presents difficul- 

 ties for echo integrator signal processors. Fishes 

 in the acoustic lee of a rock are often not echo in- 

 tegrated. 



In the late 1980s, several investigators began 

 using submersibles to assess fishes inhabiting 

 rocky banks in the northeastern Pacific Ocean 

 (Richards, 1986; Pearcy et al., 1989; Stein et al., 

 1992; Krieger, 1993; O'Connell and Carlile, 1993; 

 Murie et al., 1994). With the exception of Krieger 

 ( 1993), these studies were designed to provide in- 

 formation only on fishes closely associated with 

 the bottom. These submersible surveys were suc- 

 cessful in assessing fishes associated with bottom 

 habitats but missed an unknown number of fishes 

 swimming above the bottom. For example, Pearcy 

 et al. ( 1989) noted large schools offish in the water 

 column that were near the bottom but not counted 

 by the submersible observers. 



In this study, we combined submersible and 

 hydroacoustic techniques to estimate more accu- 

 rately the distribution and relative abundance of 

 fishes on a rocky bank off Oregon. Our objectives 

 were to determine the proportion of fishes in the 

 water column that were not detected with 

 submersibles and to compare estimates of fish 

 density near the bottom from submersible surveys 

 with density estimates generated for the same 

 region from hydroacoustic surveys. 



Methods 



From 24 September to 3 October 1991, we sur- 

 veyed fish assemblages on Stonewall Bank, a large 

 (200 km 2 ), relatively flat, rocky bank located about 

 22 km southwest of Newport, Oregon (Fig. 1). 

 Stonewall Bank ranges in depth from 41 m to over 



75 m and comprises a gently sloping rock bottom with 

 dissected ridges of siltstone and mudstone. We used 

 the research submersible Delta to survey fixed sta- 

 tions on the top and side of Stonewall Bank. At three 

 stations, both submersible and acoustic surveys were 

 conducted (Fig. 1). 



Submersible surveys 



We completed three submersible transects at each 

 of three survey stations. In addition to the nine 

 transect dives, we completed three bounce dives to 

 evaluate large schools in the water column that were 

 detected acoustically. On the bounce dives, the sub- 

 mersible was launched and directed into schools in 

 the middle of the water column. Observers estimated 

 species and size composition of schooling fishes. 



Figure 1 



Location of Stonewall Bank off Newport, Oregon, where acous- 

 tic and submersible surveys were conducted from September 

 through October, 1991. Arrows indicate the length and direc- 

 tion of transects. 



