Impacts of outliers and mis-specification of priors on Bayesian fisheries-stock assessment

Bayesian inference is increasingly used in estimating model parameters for fish-stock assessment, because of its ability to incorporate uncertainty an d prior knowledge and to provide information for risk analyses in evaluatin g alternative management strategies. Normal distributions are commonly used in formulating likelihood functions and informative prior distributions; t hese are sensitive to data outliers and mis-specification of prior distribu tions, both common problems in fisheries-stock assessment. Using a length-s tructured stock-assessment model for a New Zealand abalone fishery as an ex ample, we evaluate the robustness of three likelihood functions and two pri or-distribution functions, with respect to outliers and mis-specification o f priors, in 48 different combinations. The two robust likelihood estimator s performed slightly less well when no data outliers were present and much better when data outliers were present. Similarly, the Cauchy distribution was less sensitive to prior mis-specification than the normal distribution. We conclude that replacing the normal distribution with "fat-tailed" distr ibutions for likelihoods and priors can improve Bayesian assessments when t here are data outliers and mis-specification of priors, with relatively min or costs when there are none.