Age-structured stock assessment models are key tools for managing Antarctic toothfish (Dissostichus mawsoni) fisheries within CCAMLR, yet the assumptions underlying future recruitment projections remain a source of uncertainty in yield calculations. This paper reviews current approaches for forecasting recruitment in fisheries stock assessment models, focusing on medium- to long-lived species such as toothfish. We examine methods ranging from simple historical averages to sophisticated environmental covariate models, evaluating their applicability across different projection timeframes.
Based on the review, for short-term projections (1-5 years) we recommend using average recent recruitment. For long-term projections (30+ years), incorporating environmental covariates into stock-recruitment models and employing ensemble modelling approaches may better capture potential trends and variability under changing climate conditions. Case studies from Southern Bluefin Tuna, North Atlantic Albacore, and Pacific Halibut demonstrate diverse current practices and highlight the importance of considering both biological mechanisms and management timeframes.
Key challenges include the breakdown of historical climate-recruitment relationships under climate change, the assumption of stationarity in traditional approaches, and the need to balance biological realism with practical management requirements. We recommend a multi-faceted approach that explicitly acknowledges uncertainty through stochastic simulations, regularly updates models as new data becomes available, and employs Management Strategy Evaluation frameworks to test robustness under various recruitment scenarios.
As climate change increasingly affects marine ecosystems, moving beyond simple historical averages toward more adaptive, scenario-based approaches will be essential for sustainable fisheries management, particularly for longer-lived species where recruitment assumptions significantly influence long-term stock projections and reference point calculations.