598 



Fishery Bulletin 97(3), 1999 



Males 



40 46 52 58 64 80 95 



40 46 52 58 64 80 95 



40 46 52 58 64 80 95 



Females 



0.0 i-^ 



40 46 52 58 64 70 85 



40 46 52 58 64 70 85 



Length class (cm) 



Figure 6 



Obsen-ed i ) and expected ( i length compositions. Data were collected 



during all annual surveys from 1979 to 1995. but only a subset is displayed for 

 brevity. See Table 1 for length-class definitions. 



because the estimates are linear functions of the sur- 

 vey abundance index. Exploitable biomass follows 

 every up and down in the abundance index. All mea- 

 surement error is reflected in the recruitment esti- 

 mate. The trend from the age-structured analysis is 

 a complicated function of the abundance index and 

 the age and length data and smooths annual changes 

 in the abundance index. 



The simulations in this paper imply that the age- 

 structured model provides reliable estimates of sable- 

 fish absolute biomass. The assumption that trawl 

 surveys provide absolute biomass estimates required 

 for the sablefish delay-difference analysis can be 

 dropped. Starting with the 1997 fishing year, the age- 



structured analysis has supplanted the delay-differ- 

 ence analysis in the stock assessment for estimating 

 biomass(Fujioka et al., 1996). 



Comparison to other 



estimates of sablefish abundance 



Sablefish absolute abundance was estimated in two 

 other studies, but the estimates were unreasonable. 

 Sablefish longline catchability was estimated by 

 depletion experiments (Clausen et al., 1997). The 

 computed longline catchabilities were applied to the 

 sablefish longline survey results for the Gulf of 

 Alaska; the resultant biomass estimates of 50,000- 



