Predictions and probabilistic forecasts from seasonal spatio-temporal models are valuable to support fisheries assessments and management

Explicitly accounting for seasonal variation in resource models allows for more insightful predictions to assist fisheries assessments and management. However, spatio-temporal models that simultaneously estimate seasonal, interannual, and spatial variation in fish data are often overlooked. Here, we encourage the wider use of these models through three illustrative applications. First, we employ fisheries catch rate data for west coast New Zealand snapper (Chrysophrys auratus), to show how a seasonally explicit spatio-temporal model can be utilized to assess stock structure hypotheses. Second, still using the snapper data, we develop a monthly explicit spatio-temporal model to help identify the best months for the West Coast North Island (WCNI) research survey. Finally, we employ Sub-Antarctic research trawl survey data for autumn (March–June) and summer (November–December) to design a seasonally explicit spatio-temporal model for hake (Merluccius australis) that includes a non-linear bottom temperature effect and generate probabilistic forecasts with this model. The hake seasonally explicit model delivered an index of relative biomass that was strongly correlated with the estimated spawning stock and total biomasses from the most recent Sub-Antarctic hake stock assessment. The spatio-temporal model exhibited reasonable predictive skill. We also produced probabilistic forecasts until 2050 with the model, which suggested that the index of relative biomass of Sub-Antarctic hake may increase in the future under the CMIP6 SSP3–7.0 shared socio-economic pathway scenario. Seasonal spatio-temporal models will hopefully be increasingly used to support fisheries assessments and management in the face of climate change, particularly to design optimal designs for research surveys.