A Bayesian modelling framework for the estimation of catch-at-age of commercially harvested fish species

Abstract

A Bayesian hierarchical model was developed to estimate catch-at-age from commercial fishery data. Most common forms of data can be utilized: age and length, length-stratified ages, and length only. There is no need to construct an age–length key. Both landings and discards can be estimated, as can the effects of age reading errors. Estimates can be made for difficult to distinguish stocks, where stock identification is only possible in some fish, for example, by using otoliths and age determination. Uncertainty in stock identification can be included in this modelling approach which allows errors in the estimates to be fully captured in their posterior distributions. An important component of this model is the inclusion of random effects to account for positive correlation in both fish size and age within the sampling units.