SPATIAL-DISTRIBUTION OF PACIFIC HALIBUT (HIPPOGLOSSUS-STENOLEPIS) - AN APPLICATION OF GEOSTATISTICS TO LONGLINE SURVEY DATA

The spatial distribution of Pacific halibut (Hippoglossus stenolepis) in the Gulf of Alaska was analyzed using longline catch per unit of effort (CPUE) data collected during three grid surveys in 1984, 1985, and 1986. Geostatistical techniques were used: (i) a variographic analysis to model and estimate the spatial structure of halibut abundance and (ii) ordinary kriging to predict local abundance. Available small-scale information made it possible to model satisfactorily the spatial structure. Results show (i) a persistent large-scale east-west difference in average CPUE and (ii) spatially correlated CPUE data with an average covariance decreasing as the distance between observations increased, over a range of 0-20 nautical miles (nmi) in 1984 and 1985, and 50 nmi in 1986. The survey design had limitations in that it was too unbalanced, with stations very close together along north-south transects, and transects too far apart from each other. Consequently, prediction error was small close to the transects and large in between in a clear banded pattern. To achieve a more regular coverage of the same area, a new survey design was developed: the global variances obtained with this new design using the variogram parameters for 1985 and 1986 were 20% lower than those based on the old design.