<p><strong>Fisheries management is very often supported by stock assessments. An adequate understanding of stock structure is key for accurately evaluating stock status with stock assessments. It is also important to understand the seasonal migration and connectivity patterns of substocks, as exposure and vulnerability to fishing pressure is likely to differ among substocks. The present research investigated seasonality and stock structure in snapper (Chrysophrys auratus) on the west coast of New Zealand using spatio-temporal models fitted to commercial fishery data implemented through the vector autoregressive spatio-temporal (VAST) modelling platform. We used Poisson-link delta models that combined models for fish numbers per unit of area and average weight per fish to predict snapper biomass-densities over time and space. Models were fitted to commercial bottom trawl biomass catch rate data separately for three seasons: spring (October–December), summer (January–April), and winter (May–September). The models included a fixed year effect, random effects for spatial and spatio-temporal variation, catchability covariates, and a vessel-year random effect. The models predicted snapper densities, which were used to produce summary metrics—including median log-density, interannual variability in log-density, and long-term abundance trend—as well as to conduct hierarchical clustering to group spatial areas into clusters. Results indicated clear seasonal differences in snapper dynamics between spring, summer, and winter. We also found noticeable differences between three spatial strata on the west coast of New Zealand, which is consistent with previous findings from stock assessments, genetic studies, and phenology studies.</strong></p>