Real-time multi-month forecasting of skipjack tuna (Katsuwonus pelamis) habitat in the western and central Pacific Ocean for improved fishing efficiency and fisheries management

Abstract

Abstract Marine fish experience shifts in their distribution due to changes in the physical and biological environments. These shifts pose challenges for fishery businesses and management in international fisheries organizations. In the western and central Pacific Ocean, changes in the spatial distribution of skipjack tuna (Katsuwonus pelamis) due to climate variability often influence fishing activities and economic benefits. This study provides an operational forecast to enhance the decision-making process for fishery managers and fishermen by informing them about fish distributions in the coming months. Monthly forecasts of skipjack tuna habitat are generated utilizing a species distribution model (boosted regression trees) in combination with real-time environmental forecasts. An alternative method (dynamic time warping) is proposed to improve monthly chlorophyll forecasts, which are crucial for enhancing skipjack tuna habitat forecasting. To assess the prediction skills of the model, a retrospective analysis was conducted, forecasting skipjack tuna distribution over 9-month periods and comparing the forecasts with observed occurrences. The results demonstrate that 67.9% of the observed skipjack tuna catches occurred within the forecasted habitats, indicating that the forecast model skillfully predicted skipjack tuna movements 9 months in advance. These monthly updated forecasts can serve as a potential tool for resource managers in developing effective management strategies and fishers in planning fishing operations sustainably and responsibly.