Development of a state space assessment model for Cod (Gadus morhua) in NAFO subdivision 3Ps

Abstract

A state-space assessment model (SAM) that incorporates commercial catches and survey indices is developed for the 3Ps cod stock. It allows for process and observations errors and uses random walks in fishing mortality rate to estimate time variant selectivity. The model provides estimates of stock size, fishing mortality and recruitment together with their uncertainties and would eventually have the ability to provide stochastic forecasts. The objective of the study is to achieve an accurate estimation of the stock dynamics and better account for uncertainty. Four runs of the model are presented here with different options on mortality and survey inputs. Runs produce comparable estimation but lead to different stock status in the recent period. They also perform differently in terms of retrospective analysis, residual patterns, and description of latent processes. Three runs out of four produced outputs with concerning issues, so they could not be accepted as basis for assessment advice. The last one (called sentinel run) was kept as a “control model,” which means it would be use for comparison purpose alongside with the accepted model (i.e., state-space model [HYBRID]).