A Monte Carlo study on impacts of the size of subsample catch on estimation of fish stock parameters

Abstract

Most models currently used in studying dynamics of fish stocks are age-structured stock assessment models. One of the essential input data requirements for this type of model is estimates of age composition by year. In practice, age composition is seldom estimated by aging all fish harvested; rather it is estimated from aging a small group of fishes (i.e. subsample catch) randomly subsampled from the total catch. Although many studies have noted errors in aging fish from the subsample catch with respect to the estimation of fish stock parameters, errors in estimating age composition resulting from this random subsampling has received little attention. Using a Monte Carlo simulation approach, this study evaluated the impacts of different sizes of subsample catch on estimating fish mortality rates, cohort sizes at recruitment, and age-specific selectivity coefficients with two cohort-based models. When the annual total catch and/or fishing effort were subject to small errors, the estimated stock parameters had much larger errors using a subsample catch with the size less than 1000 compared with those estimated without errors from randomly subsampling; and when the size of subsample was over 1000, the error differences in estimating stock parameters were insignificant. However, with increased errors in estimates of the annual total catch and fishing effort, impacts of the size of subsample on estimating stock parameters decreased. The impacts of the subsample size on estimating stock parameters varied among different parameters and among different age-structured models.