94 



Fishery Bulletin 91(1). 1993 



merits in response to trawling have 

 been reported for other semi-demer- 

 sal fish species (Wardle 1983 and 

 1986, Ona & Godo 1990). The low 

 catch rates at the two marginally- 

 trawlable sites may reflect decreased 

 trawl efficiency in rugged habitat. 



Catch densities of small rockfish 

 were also highly correlated with ob- 

 served densities from the submers- 

 ible (r=0.89). Catch densities were 

 similar to observed densities at all 

 nine sites (Fig. 5); the ratio estimate 

 of catch-to-observed densities was 

 1.3:1 (SE=0.3). Two possible reasons 

 that ratio estimates were lower for 

 small rockfish than for large rock- 

 fish are that ( 1 ) small rockfish es- 

 cape through net meshes at a greater 



higher than observed densities at 

 the seven trawlable sites (Fig. 5) 

 and lower at the two marginally- 

 trawlable sites; the ratio estimate 

 of catch to observed densities was 

 2.2:1 (SE=0.4). This high ratio 

 estimate indicates that bottom 

 trawls are very efficient for cap- 

 turing large rockfish, resulting in 

 density estimates approximately 

 twice those observed from the 

 submersible. 



Herding of large rockfish to- 

 ward the net opening may ex- 

 plain the high catch rates of large 

 rockfish. Submersible counts in- 

 cluded rockfish to 7 m off bottom, 

 whereas the trawl sampled to 

 only 2 m off bottom. Rockfish ap- 

 parently moved downward in re- 

 sponse to the trawl gear. Down- 

 ward movements alone would 

 have resulted in a catch-to- 

 observed ratio of about 1:1. The 

 ratio of 2.2:1 indicates large rock- 

 fish also moved inward toward 

 the net opening to avoid the 

 bridles and otter doors. The 

 bridles and doors extend approxi- 

 mately 7 m on each side of the 

 14m horizontal net opening, in- 

 creasing by twofold the seafloor 

 area swept by the trawl gear. 

 Downward and inward move- 



THORNYHEAD 



13 



SAMPLEO 



7 10 11 12 13 14 17 



STTE 

  OBSERVED • SAMPLED 



FLATFISH 

 r. .79 



POLLOCK 

 r- .72 



SITE 



Figure 4 



Percent composition of fish groups observed from a submersible and sampled with bot- 

 tom trawls, and correlation coefficients (r) at nine sites in the eastern Gulf of Alaska. 



