Hightower Rockfish harvesting policies in Washington-Oregon^California trawl fisheries 



653 



[piS^ ,.er^ 



o 

 o 



O 1 



five~spp- 



Vorionce (total 'lotct-i! 



Figure 4 



Relationship between mean total rockfish catch and the variance of 

 total catch for the two- (bocaccio, chilipepper), three- (bocaccio, 

 chilipepper, widow rockfish), and five-species Osocaccio, chilipepper, 

 widow, splitnose, and shortbelly rockfish) models. Open symbols 

 represent Pareto Frontiers derived by simulating the fishery at 20, 

 40, 60. 80. and 100% of the constant fishing mortality rate (F) that 

 maximized harvest. Shaded and closed symbols represent the 

 catch:variance relationship for the variable F and multispecies 

 policies, respectively (equations 5-6). Simulation runs were made 

 assuming a 0. 25, or 50% coefficient of variation for annual estimates 

 of stock size. 



Mean yield was only slightly lower at higher biomass 

 CVs; the decrease was similar for constant F and 

 variable F policies (Tables 3-5). Thus, in this model at 

 least, there was no apparent disadvantage in using the 

 more complex policies, even when biomass was esti- 

 mated imprecisely. This was an interesting result 

 because, although all policies used biomass estimates 

 to calculate catch, the variable F policies also use the 

 estimates to calculate Fs. 



When biomass CVs were 50%, the variance for total 

 catch was in all cases lower for the single-species (5) 

 than for the multispecies policy (6). Apparently it was 

 important to readjust continuously for the changes in 

 stock size due to previous errors in estimating biomass. 

 Based on limited additional runs using the three-species 

 model, it appears that further gains in stability of total 



catch could l)e achieved by using a three-parameter 

 harvesting policy (Table 6): 



Fj[t] = b,,„ + b,, B,[t]/ 



Boo,, + b,, 1 B,[t] / 1 B„ 



(7) 



i^j 



i^'j 



The primary disadvantage of this policy is the computa- 

 tional expense required to obtain the policy parameter 

 estimates. 



Model limitations 



One of the primary assumptions underlying this model 

 is that trawlers can, on an annual basis, control the 

 relative proportions of the individual rockfish species 

 in the catch. That appears to be a reasonable assump- 

 tion for widow rockfish, which are caught primarily 

 in a directed midwater trawl fishery. Catches of other 

 rockfish species should be low in a midwater trawl 

 fishery for shortbelly rockfish (Lenarz 1980), and the 

 Pacific Fishery Management Council is unlikely to per- 

 mit a bottomtrawl fishery for that species. Peak abun- 

 dance of splitnose rockfish is in deeper water than for 

 the other species used in this model: however, some 



