A hierarchical operational regional ocean forecasting system off the Korean peninsula

KHOA(Korea Hydrographic and Oceanographic Agency) has been operating KOOFS(Korea Ocean Observing and Forecasting System) since 2012. KOOFS runs 23 hierarchical prediction models not only for ocean forecasting but also for atmosphere and wave forecasting in various spatial scales(from bays to the North Pacific). Each prediction model provides 72 to 144 hours forecasts of temperature, salinity, sea-level heights and currents for public use. KOOFS uses WRF model for the atmospheric forecasting, ROMS and MOHID for the oceanic forecasting, and WW3 for the wave forecasting. To improve the accuracy of forecast products, we applied two data assimilation methods of 4DVAR and EnOI. Overall, both assimilation schemes show reasonable enhancement of accuracy on sea surface temperature prediction. 4DVAR produces generally better analyses than the EnOI, whereas EnOI shows advantages on computational time. In this presentation, we introduce the results from various scale forecasting models and several assimilation experiments.