Implementation of a High-Resolution Regional Ocean Analysis System for Northwest Pacific Using an Ensemble Data Assimilation Method

Reliable monitoring of coastal ocean states is critical for understanding regional climate variability and managing marine resources. We developed a high-resolution regional ocean analysis system for the seas around Korea. The system is based on the Regional Ocean Modeling System (ROMS) with a horizontal resolution of approximately 5 km and 30 vertical layers. To resolve complex coastal physical processes, we incorporated tidal forcing from the TPXO9 model and atmospheric forcing from ECMWF ERA5, while boundary conditions were supplied by the global GLORYS-NRT product.

To minimize model errors and incorporate the observation data, we applied the Ensemble Optimal Interpolation (EnOI) method for data assimilation. The background error covariance was estimated from a long-term simulation (1980–2022) comprising 44 ensemble members. We implemented a localization radius of 50 km horizontally and 100 m vertically to eliminate spurious correlations. The system assimilates a wide range of observations, including Sea Surface Temperature (OSTIA), surface geostrophic currents from satellite altimetry, and in-situ vertical profiles of temperature and salinity CTD and Argo floats.

Comparison with independent observation data and the global ocean analysis (GLORYS-NRT) demonstrated the system's reliable performance. The analysis field showed a high correlation (0.99) for sea surface temperature and reduced RMSE compared to the global model. Notably, our system accurately reproduced the vertical structure of the Yellow Sea Bottom Cold Water (YSBCW) and tidal fronts in the Yellow Sea and the meandering path of the East Korea Warm Current and Kuroshio. Furthermore, validation of volume transport through the Korea and Jeju Straits confirmed that our system better captures seasonal variability compared to the global product, which tended to underestimate transport in the Korea Strait.

The regional ocean analysis system successfully tracked significant climate anomalies in 2025. The region experienced distinct warming, with surface temperatures 0.5–2.0°C higher than the climatological mean (1991–2020), a warming trend extending to 150 m depth. Additionally, surface freshening (0.1–0.3 psu decrease) was observed in the Yellow Sea. These results underscore the necessity of including tidal processes and assimilating high-resolution local observations for effective monitoring of ocean climate change in the coastal seas.