A Novel Method for Making Stochastic Catch Projections in Association with a State-Space Bayesian Production Model: Application to the Gulf of Mexico Smoothhound Complex

A novel method for making stochastic catch projections based on MCMC was developed in association with a state-space Bayesian surplus production model (SSSPM) that was used to assess the Gulf of Mexico complex of smoothhound sharks (Mustelus spp.). The main advantage of this projection method, compared with other existing methods developed for other production model formulations, is that the assessment model fitting and projections are integrated into a single MCMC-based process, which makes assumptions and approximations for the posterior distribution and covariance of the estimated parameters unnecessary. Unknown parameters and unobservable states were estimated given the data and priors during model fitting and the estimated parameter values were subsequently used to make projections on stock status under alternative fixed catch scenarios, thus allowing for variability in estimated parameters to propagate into the future through each MCMC iteration. We also present a phase plot depicting the probability of the stock being overfished and of overfishing occurring that could be used as a convenient management tool to implement overfished/overfishing policies at a variety of risk tolerance levels. We will discuss challenges and propose directions to integrate environmental information into the SSSPM in the spirit of the Next Generation of Stock Assessments framework.