Abstract.- Optimal harvesting 

 policies were developed for a model 

 of the Washington-Oregon-Califor- 

 nia trawl fisheries for bocaccio Sebas- 

 tes paiwispinus, chilipepper S. goodei, 

 widow rockfish S. entomelas, split- 

 nose rockfish S. diploproa, and short- 

 belly rockfish S. Jordan i. In the simu- 

 lated management system, the annual 

 quota for each species was based on 

 an intended fishing mortality rate (F) 

 and an estimate of stock biomass. 

 Traditional constant F policies were 

 compared with variable F policies in 

 which F was a linear function of 

 either stock biomass (single-species) 

 or the combined biomass of all other 

 species (multispecies). Variable F 

 policies were ineffective for increas- 

 ing total harvest of all species com- 

 bined, when compared with constant 

 F policies. However, both single- 

 species and multispecies variable F 

 policies were found to reduce the 

 variance for total harvest substan- 

 tially with no significant loss in yield. 

 The reduction in variance was rough- 

 ly equal to the reduction that would 

 be achieved by lowering the coeffi- 

 cient of variation (CV) for biomass 

 estimates from .50% to 25% or from 

 25% to 0%. The reduction in vari- 

 ance generally was greater for the 

 multispecies than for the single- 

 species variable F policy. Compared 

 with the case where CV = 0%, aver- 

 age total yield was reduced slightly 

 for both constant F and variable F 

 policies when CVs for biomass esti- 

 mates were greater than 0%. 



Multispecies Harvesting Policies for 

 Washington-Oregon-California 

 Rockfish Trawl Fisheries 



Joseph E. Hightower 



Tiburon Laboratory, Southwest Fisheries Science Center 

 National Marine Fisheries Service, NOAA 

 3150 Paradise Drive. Tiburon, Calfiornia 94920 



Manuscript accepted 4 .June i;i!K). 

 Fishery Bulletin, U.S. 88:64.5-6.56. 



In the Washington-Oregon-Califoi'iiia 

 (WOC) trawl fisheries for rockfish 

 {Sebastes spp.), the landings are com- 

 prised of many species, including 

 bocaccio S. paucispinus, chilipepper 

 S. goodei, Pacific ocean perch S. alu- 

 tus, canary rockfish S. pinninger, 

 widow rockfish S. entomelas, and yel- 

 lowtail rockfish S. flavidus (PFMC 

 1982). There is httle evidence of sub- 

 stantial biological interaction among 

 the adults of these species. Techno- 

 logical interactions (Pope 1979) are 

 significant in the shelf rockfish fish- 

 ery (Pikitch 1987), which accounts for 

 most of the landings of bocaccio, 

 chilipepper, and canary and yellow- 

 tail rockfish. Fishermen have some 

 control over the species composition 

 of the catch in this fishery by vary- 

 ing the areas or depths fished. The 

 other significant rockfish fishery is 

 the midwater trawl fishery for widow 

 rockfish. Catches of other rockfish 

 species are negligible in that fishery. 

 Most currently marketed Seh<tstes 

 species have similar or equal value to 

 fishermen; consequently, fishermen 

 are more concerned about the aggre- 

 gate yield than about the species 

 composition of the catch. Thus, it 

 seems reasonable to ask whether ef- 

 forts to vary the annual proportions 

 harvested for these species would 

 result in benefits to the resource or 

 to the fishing industry. For example, 

 reducing the variability of annual 

 quotas has been an important con- 

 cern for the west coast fishing in- 

 dustry. The objective of this study 

 was to evaluate multispecies harvest- 



ing policies for the WOC trawl fish- 

 eries for rockfish to determine whe- 

 ther total yield (all species combined) 

 can be increased or the variance for 

 total yield can be reduced. Specif- 

 ically, I compared the performance 

 of policies obtained using (1) risk- 

 neutral and risk-averse objective 

 functions, (2) constant and variable 

 fishing mortahty rates (F), and (3) 

 abundance estimates for a single 

 species or multiple species. 



Except for widow rockfish, rela- 

 tively little is known about the pop- 

 ulation dynamics of these stocks. For 

 that reason, this study should be 

 viewed as an exploratory analysis of 

 alternative harvesting policies, and 

 not as a basis for regulatory changes 

 by the Pacific Fishery Management 

 Council. It may be worthwhile to 

 develop similar harvesting policies 

 for Council review once we have ade- 

 quate life-history information about 

 the commercially important ground- 

 fish stocks. The harvesting policies 

 examined in this study should be 

 useful in other cases where fishermen 

 derive income from several species 

 and have some control over the spe- 

 cies composition of the catch. 



Methods 



This section provides an overview of 

 the model structure, harvesting 

 policies, and life-history data. A de- 

 tailed description of methods is pro- 

 vided elsewhere (Hightower 1990). 

 I developed optimal harvesting 

 policies for two-, three-, and five- 



645 



