Forecasting herring biomass using environmental and population parameters

Forecasting abundance and understanding year class strength is key to the ecosystem-based fisheries management of herring and other small pelagic fish. Using the San Francisco Bay herring population, we tested the hypothesis that abundance (spawning stock biomass; SSB) could be predicted prior to the commencement of annual fisheries using estimates of i) previous SSB (SSBlag1), ii) young-of-the-year production (YOY), and iii) environmental conditions. A time series model including SSBlag1, YOY lagged 3 years (YOYlag3), and environmental conditions in the season before spawning explained 67% of the variance in annual biomass, with better predictive error in comparison with simpler models. YOYlag3 was by far the strongest predictor. It was robust over the entire study period (1980-2017) and also for a more limited period (1991-2017) when observed variance in SSB increased. We attribute the predictive power of YOY productivity to age structure, as almost 70% of the population is comprised of young fish. We hypothesize that an age truncation effect, probably resulting from a combination of long-term environmental effects and fisheries impacts, supports this model, which effectively predicts year class strength. Assuming the population age structure remains the same in the future, our model provides management with an early warning indicator of upcoming SSB with a 3-year lead, which could be applied in harvest control rules.