Development of the ocean mid-term prediction system for the seas around Korea

Abstract

Due to a decline of total catch production and more frequent extreme phenomena in Korean waters during the recent decades, a mid-range (2~6 months) prediction of ocean conditions has emerged as an important tool for the management, recovery and disaster preparation of fishery resources. In this study, we report on a newly developed ocean mid-range prediction system for the Northwest Pacific focusing on the seas around Korea. Three-month reforecast experiments were conducted by using atmospheric forcing and open boundary condition from Climate Forecast System version 2 (CFSv2) operational forecast data. We analyzed the last month (March, June, September, and October 2016) of the simulation output and assessed the performance of the model’s prediction skill in terms of sea surface temperature. Although the reforecasts tend to overestimate SST especially in the warm current regions include the Kuroshio and the East Korea Warm Current. The spatial SST patterns are generally similar to observation. Based on the preliminary analysis for the three-month reforecast experiments in 2016, however, there is a need to improve the problem that CFSv2 does not cope well with Kuroshio and Oyashiocurrents. Therefore, by applying the delta method, we reduced the bias of CFS version 2 and improved the prediction performance of OMIDAS.