PREDICTION OF THE PATH TYPE AND OFFSHORE DISTANCE OF THE KUROSHIO CURRENT USING NEURAL-NETWORK

We built a forecasting system for the path types of the Kuroshio curre nt and the distance between the Kuroshio axis and Cape Iroh-zaki. A la yered type of the artificial neural network was used in the system. In put data sets included six months' precedent data of distances between the Kuroshio axis and major capes, occurrence rates of Kuroshio path types and deviations of sea surface temperature. The predicted values of the hydrographic conditions in the months of March and April were w ell matched to the observed values. The investigation of the interconn ection weight values suggests that the deviations of sea surface tempe rature and occurrence order of the Kuroshio path types significantly a ffected the predictions in the neural network. As the catch of sardine larvae in Sagami Bay during March and April is affected by the path t ypes of the Kuroshio current and the distance between the Kuroshio axi s and Cape Iroh-zaki, the predicted results will be used to forecast t he catch.