Responding to climate-driven changes in growth in the modern stock assessment models
Abstract
Temporal variability in somatic growth is one of the main drivers of biomass fluctuations in fish stocks but is infrequently modeled explicitly in stock assessments. In recent years, state-space assessment models have been developed to estimate stochastic deviations in several biological parameters (e.g., recruitment, natural mortality). The Woods Hole Assessment Model (WHAM) is an age-structured state-space assessment model implemented in Template Model Builder (TMB) and used for several stocks on the U.S. east coast. WHAM also permits the incorporation of environmental covariates, treating the true, unobserved values as random effects. However, WHAM does not explicitly model somatic growth, which limits its use in other regions and stocks. This study expanded the base WHAM features to incorporate length information (i.e., marginal length compositions) and model somatic growth. We introduce these novel features and apply them to walleye pollock in the Gulf of Alaska, USA, a species that has shown substantial annual variability in mean length-at-age. We explored two approaches to account for variability in somatic growth 1) using a parametric growth equation and estimating random effects on parameters, and 2) estimating random effects on mean length-at-age. Moreover, we explored the incorporation of sea surface temperature (SST) as an environmental covariate linked to growth parameters. This study presents a new tool that can be applied to fish stocks worldwide, especially when length data and growth estimation are influential, or when linking climate variables to growth in hindcasts or forecasts.
