Development of a medium range ocean prediction model for the seas around Korea

Abstract

Thus we have developed the Ocean MIDium range forecAst System (OMIDAS) for the Northwest Pacific including the seas around Korea. To evaluate the forecast performance of the OMIDAS model, we conducted a 3-month re-forecast experiment with the initial conditions of January 1, April 1, July 1, and October 1 in 2016. For the initial conditions of this re-forecast experiments, we used the central ocean and atmosphere variables of CFSv2 including temperature, salinity, current, and sea surface height. We analyzed the last month (March, June, September and December of 2016) of the simulation output and verified that the model successfully re-forecasted the temperature anomaly distribution of Sea Surface Temperature. As observed, the model showed a positive anomaly compared to the normal year averaged for 1982-2010, especially in the regions with warm currents such as Kuroshio and Tsushima, East Korea Warm Current, and Yellow Sea. Overall, the OMIDAS model shows a detailed anomaly structure compared with CFSv2 forecast, with slightly intensified anomalies with respect to the OISST reanalysis. Our results suggest that the developed medium range ocean prediction model can be used to predict the marine environment at least three month in advance. For a future study, we will integrate a biological component with the OMIDAS for the medium range forecast in the context of fishery resources and marine ecosystem.he initial conditions of January 1, April 1, July 1, and October 1 in 2016. For the initial conditions of this re-forecast experiments, we used the central ocean and atmosphere variables of CFSv2 including temperature, salinity, current, and sea surface height. We analyzed the last month (March, June, September and December of 2016) of the simulation output and verified that the model successfully re-forecasted the temperature anomaly distribution of Sea Surface Temperature. As observed, the model showed a positive anomaly compared to the normal year averaged for 1982-2010, especially in the regions with warm currents such as Kuroshio and Tsushima, East Korea Warm Current, and Yellow Sea. Overall, the OMIDAS model shows a detailed anomaly structure compared with CFSv2 forecast, with slightly intensified anomalies with respect to the OISST reanalysis. Our results suggest that the developed medium range ocean prediction model can be used to predict the marine environment at least three month in advance. For a future study, we will integrate a biological component with the OMIDAS for the medium range forecast in the context of fishery resources and marine ecosystem.