A NONPARAMETRIC FORECAST MODEL OF INSHORE ATLANTIC COD (GADUS-MORHUA)LANDINGS BASED ON BIOMASS, CUMULATIVE LANDINGS, AND WATER TEMPERATURE

Studies have suggested that cod biomass and water temperature might be major factors influencing inshore landings in the Newfoundland northe rn cod fishery. However, the relationship between inshore landings and these factors may be nonlinear. Modelling such a process using parame tric approaches may not be adequate. We used a nonparametric method to forecast inshore gill-net landings. Two probability density function models based on biomass and cumulative landings were tested with the 1 974-1998 data, as well as a third model that combines the first two mo dels. The residuals from the combined forecast were then modelled usin g a third probability density function to account for the effect of wa ter temperature. Model performances were evaluated by comparing the cr oss-validated prediction sums of squares. The model based on biomass, cumulative landings, and water temperature had the least error in fore casting gill-net landings, a feature useful for fisheries management a nd for projecting processing and marketing capacity.