Weinberg: Rockfish assemblages off Oregon and Washington 



631 



ized to facilitate comparisons among surveys. The 

 trawl employed a 89-mm mesh codend with a 32-mm 

 liner, capable of retaining juveniles and smaller-sized 

 species, such as rosethorn and greenstriped rockfish. 

 Smaller species may not be caught in similar pro- 

 portions in commercial catches. 



On the other hand, fishermen select, modify, and 

 operate their gear based on fishing strategy. These 

 strategies have changed over the years to adapt to 

 reduced allowable harvests and the imposed vessel 

 trip limits. To economize, fishermen target several 

 species over the course of a single trip, using more 

 than one type of gear (Tagart 5 ). For example, a ves- 

 sel may begin using midwater gear directed at catch- 

 ing a limit of widow rockfish, then switch to bottom 

 gear to target on yellowtail rockfish. Once the trip 

 limits for these two species are reached, the captain 

 may opt to fish flatfish or change to a more general- 

 ized rockfish strategy keeping larger specimens of 

 darkblotched, silvergray, redstripe, or sharpchin 

 rockfish. 



Because survey and fishery tactics differ greatly, I 

 assessed whether the assemblages found to persist 

 in summer surveys based on systematic sampling 

 also occurred in commercial collections characterized 

 by opportunistic sampling, market conditions, and 

 management-imposed restrictions. The literature 

 referred to previously describes various west coast 

 fish assemblages determined from different sampling 

 techniques, classification methods, and data types. 

 Although none of these studies used a time-series of 

 this duration, Rogers and Pikitch (1992) identified 

 two rockfish assemblages using year-round ( 1985- 

 87 ), prediscard, commercial trawl data from off the 

 coasts of Oregon and Washington. Briefly, they ex- 

 amined five fishing strategies, one of which targeted 

 demersal rockfishes using bottom trawl gear. Even 

 though the survey depth range was narrower than 

 that of the commercial operations, the results of 

 Rogers and Pikitch (1992) share much in common 

 with the present study. Both studies describe an as- 

 semblage of deepwater rockfishes with three species 

 in common: Pacific ocean perch, darkblotched rock- 

 fish, and splitnose rockfish. The present study also 

 includes shortspine thornyhead, whereas Roger and 

 Pikitch (1992) include sharpchin and yellowmouth 

 rockfish. 6 Their second rockfish assemblage combined 

 bocaccio, yellowtail, canary, yelloweye, and sharpchin 

 rockfish, species assigned to either one of two sepa- 

 rate assemblages or caught too inconsistently to be 

 considered part of any assemblage in the present study. 



5 Tagart. J. Washington Dept. Fish., Olympia, WA 98504-3144. 

 Personal commun., April 1991. 



6 For unknown reasons, fishery data had substantially greater 

 abundance of yellowmouth rockfish than did the survey data. 



Rogers and Pikitch ( 1992) included in their analy- 

 sis 11 of the 20 scorpaenids that I examined. These 

 were selected based on either an arbitrarily deter- 

 mined weight threshold (1% of the total commercial 

 catch from all five fishing strategies) or if fishermen 

 claimed to target a species (e.g. bocaccio and 

 yelloweye rockfish). Most of those selected repre- 

 sented larger proportions of the commercial catch; 

 however, other rockfishes were continuously affected 

 by trawling. When classifying assembages based on 

 biomass data, one takes the risk of failing to recog- 

 nize a less abundant member of the community. In 

 the study by Rogers and Pikitch ( 1992), several mi- 

 nor rockfish species were overshadowed by catches 

 of the area's dominant species. Nearly 60% of the 

 total catch sampled from all fishing strategies com- 

 bined was nonrockfish. Among the rockfishes, widow 

 rockfish accounted for the highest percentage ( 15%) 

 of the sample catch, about 75% of which was taken 

 by midwater trawl (Rogers 7 ). Assuming that the data 

 from their study is representative of commercial op- 

 erations, we can infer the impact that trawling has 

 on the overall rockfish community, including the 

 minor species, by examining prediscard catch data 

 from hauls where bottom rockfish were targeted (i.e. 

 bottom rockfish fishing strategy). In these data, 

 nonrockfish composed only 20% of the catch whereas 

 widow rockfish, taken by bottom trawl, still ac- 

 counted for 15% of the strategy's total. Of the spe- 

 cies assemblages identified in the present study, yel- 

 lowtail, canary, and greenstriped rockfish accounted 

 for 30% of the total; shortspine thornyhead. Pacific 

 ocean perch, darkblotched rockfish, and splitnose 

 rockfish accounted for 13%; and sharpchin, redstripe, 

 and rosethorn rockfish accounted for 10% of the to- 

 tal catch. Large marketable species typically seen in 

 the landings, like bocaccio, silvergray rockfish, and 

 yelloweye rockfish, composed only 4% of the pre- 

 discard catch. 



Conclusion 



The separation of deepwater species from shelf spe- 

 cies supports the division of the "Sebastes complex" 

 management category into smaller units of greater 

 ecological consequence, as also suggested by Rogers 

 and Pikitch ( 1992). There is enough variability, how- 

 ever, in the catches of shelf rockfish, as character- 

 ized by the relatively high levels of dissimilarity 

 among groupings (Fig. 5), to warrant the use of cau- 

 tion when designating assemblages. While this study 

 recognized two assemblages of shelf species vulner- 



7 Rogers, J. B. Natl. Mar. Fish. Serv., Southwest Fish. Sci. Cent., 

 Tiburon. CA 94920. Personal commun., Nov. 1993. 



