Leveraging statistical models to improve preseason forecasting and in-season management of a recreational fishery

ABSTRACT Objective Effective management of recreational fisheries requires accurate forecasting of future harvests and real-time monitoring. Traditional methods that rely on historical catch data can be unreliable, particularly when regulatory changes alter angler behavior. This study aimed to develop and test statistical models to improve predictions of Gulf of Mexico Gag Mycteroperca microlepis harvests for both preseason planning and in-season monitoring. Methods We developed a statistical modeling framework to forecast Gag harvests and compared its performance to that of traditional historical average-based methods. Model accuracy was evaluated for both cumulative preseason projections and in-season monitoring. Additionally, we assessed the framework's ability to account for effort compression in shorter seasons. Results Our best-fitting model outperformed traditional methods in both preseason and in-season forecasts. The model also demonstrated higher accuracy in recent years, particularly for shorter seasons in which effort compression was apparent. Two key advantages of our framework are (1) the ability to account for changes in angler behavior as a result of effort compression and (2) the ability to quantify the probability of exceeding harvest quotas, providing managers with a tool to evaluate trade-offs between season duration and conservation objectives. Conclusions This study highlights the value of statistical modeling for improving fisheries management. By offering more accurate harvest predictions and explicit estimates of the probability of exceeding quotas, our approach provides a flexible tool to support decision making, particularly for vulnerable, highly targeted stocks.