Forecasting Halibut Biomass Using System Theoretic Time-Series Methods

A new procedure introduced by Masanao Aoki uses the ideas of linear systems theory to identify and estimate time-series models. A slightly modified version of this procedure is used to forecast halibut biomass in total and by regulatory area and subarea, and the out-of-sample forecasts are evaluated for eight years. All models produce highly accurate forecasts of biomass, with errors well within the bounds required for setting catch limits in the following year.