Available sources of information on sablefish biology and the Gulf of Alaska sablefish fishery are reviewed and harvest recommendations are presented. A U.S. longline fishery for sablefish has existed in the Gulf of Alaska since the early 1900's. Gulf-wide catches escalated rapidly during the 1960's when the Japanese distant water fleet developed. Total Gulf of Alaska harvests peaked in 1972 at 36,199 metric tons and began to decline, with recent annual harvests at about 9,000 metric tons. The U.S. longline fishery traditionally fished only in the eastern Gulf of A1 aska, but is currently expanding westward. Foreign sabl efish harvests are currently restricted to the central and western Gul f of Alaska. Recent developments in sablefish aging techniques using "break and burn" methods describe very slow growth in adult sablefish. The slow growth rates imply that large sablefish are much older than previously calculated, with maximum ages frequently exceeding 40 years. A much lower natural mortality rate of 0.11 is estimated from the new growth information, substantially less than previous estimates of 0.2. Sablefish recruitment appears to be characterized by long periods of relatively low recruitment between very strong year classes. Because these periods of low recruitment may exceed 10 years, maintaining large number of older fish in the population may be critical to reproductive success. Electrophoretic studies of sablefish genetics indicate that there is some degree of spatial segregation of Gulf of Alaska sablefish stocks. However some tagging studies indicate substantial movements of Si sh, with small fish (< 66 cm) tending to move north and westward and larger fish moving south and eastward. Trends in sablefish abundance are derived from Japanese longline fishery statistics, a Gulf-wide U.S.-Japanese longline survey, pot indexing in Southeastern Alaska, and port sampling of the U.S. longline fleet in Southeastern Alaska. Sablefish stock biomass appears to be increasing in abundance throughout the Gulf, largely a result of the recruitment of the strong 1977 year class. Subsequent year classes have not been strong. A new stock production model which allows for increases in fishing power indicates that MSY is probably lower than calculated in previous models. Estimates of stock size are derived from stock production model parameters and from area- swept in a Bering Sea trawl survey, combined with comparative longline-trawl sampling in the Aleutian area. Yield-per-recruit analyses indicate that size limi ts wi 11 not increase the yield or landed value from the sablefish fishery. Population parameter estimates are used in an age and sex-structured simulation model to determine "equilibrium yields" (EY) which will maintain constant biomass from 1984-1991. However because of uncertainty in the parameter estimates, it is recommended that equilibrium yield (EY) calculated by other means be applied to the Gulf of Alaska. Optimum yield (OY) are recommended to remain at 75% of the EY, or 9,473 metric tons for the entire Gulf sf Alaska. A new method of spatial allocation of the Gulf-wide OY is presented, based on the spatial distribution of biomass estimated in the U. S.-Japanese longline survey.
KEY WORDS: sablefish, stock status, age, growth, migration, Gulf of Alaska, longline fishery, equilibrium yield, mortality, maximum sustained yield, yield-per-recruit.