Short-term Forecasting in Marine Fish Stocks

Abstract

The predictive power of stock production models and simple time series methods was considered for five marine fish stocks. The distinction between model fitting and forecasting future short-term catch is discussed, as is the difference between techniques to forecast short-term yield, and techniques to calculate long-term management practice. Fox's procedure for fitting Pella and Tomlinson's stock production model, Schnute's method for fitting Schaefer's model, and Gulland's method are all considered. We found that all methods except that of Gulland work well for some stocks, and the relative performance of the methods depends upon the exploitation history of the stock. In several instances one of the best forecasts of next year's catch per unit effort (CPUE) was the previous year's CPUE, emphasizing the fact that a good forecasting technique may have no utility in determining management policies.Key words: production models, catch and effort, fisheries, management, catch forecasting, time series