Validation of alternative stock assessment hypotheses: north Atlantic shortfin mako shark

A multi-model approach for evaluating uncertainty is presented for North Atlantic Shortfin Mako. Several stock assessment methods were used to explore assumptions and uncertainty about biological parameters, reported catch, effort and length data. Methods include trend analysis, length-based indicators, and catch-only, Bayesian state space biomass dynamic, and integrated statistical age-based assessment models. A variety of diagnostics are available to examine goodness of fit, however, it is difficult to compare models with different data sets or structures. Particularly, as residual patterns can be removed by adding more parameters than justified by the data, and retrospective patterns removed by ignoring the data. Therefore, hindcasting was used to estimate prediction skill, a measure of an estimate’s accuracy compared to its observed value unknown by the model. Consideration of prediction skill allows data conflicts and model misspecification to be explored. The next steps are to use the methods developed in this study as an objective and transparent way to evaluate and weight stock assessment scenarios and to evaluate the impacts of uncertainty, and the benefits of reducing risk by improving data and knowledge.