摘要：The Pacific Island countries have a substantial socio-economic dependency on fisheries. Skipjack tuna is one of the most important species in the Western Central Pacific (WCP) and its catches in this region exhibit a spatio-temporal variability influenced by ocean conditions, mainly the El Niño-Southern Oscillation (ENSO). This study investigates the relationship between skipjack tuna catch amounts and environmental variables in the equatorial Pacific during 1990–2014, and evaluates the potential predictability of the catches based on their statistical relationship. A series of regressed and reconstructed spatial patterns of upper-ocean temperature, salinity, currents and precipitation represent ENSO-like variability, and their principal component time series are used to estimate the predictability of skipjack tuna catches in the Federated States of Micronesia (FSM). ENSO-like variability depicted from 100 m temperature and 5 m salinity in the equatorial Pacific exhibit a significant predictability for the annual catch amount in the FSM for several years with a training period of > 20 years. This suggests that the subsurface temperature or near surface salinity can be a better predictor of ecosystem variability than widely used sea surface temperature. Applications of this result to other species could have broad implications for the fishery industry in the WCP.
