This study aims to comprehensively evaluate the Coastal-KOOS Level 2 model’s ability to predict near-surface currents throughout the Northwest Pacific, using an extensive dataset of 153 surface drifters deployed since 2020. These drifters were released in diverse oceanographic settings, including the Yellow Sea, East China Sea, and East/Japan Sea, which feature strong tidal forcing, boundary currents (e.g., the Kuroshio).
We utilized drifter trajectories to derive Eulerian velocities, which were then compared with model outputs. To further validate the drifter-derived velocities, we contrasted them with in situ measurements from multiple fixed observation sites in offshore Korean waters. We also examined the influence of Stokes drift and direct wind forcing by attempting to remove these components from the drifter velocities; however, the corrections had negligible impact on most trajectories, likely because the majority of each drifter’s body remained submerged, thereby limiting its wind exposure.
Forecast accuracy was quantified using several statistical metrics, including root-mean-square error (RMSE), correlation coefficient, and complex correlation (separating magnitude and directional agreement). The results indicate that Coastal-KOOS exhibits robust performance in regions dominated by tidal currents, such as parts of the Yellow Sea, where the model’s operational focus aligns well with actual conditions. In contrast, performance degrades in areas strongly influenced by the Kuroshio and in the East/Japan Sea, likely due to strong eddies and persistent warm currents that are not fully resolved by the current model configuration. In the Korea Strait and along the Chinese coast, the model generally captures flow directions more reliably than current magnitudes, underscoring the importance of regional calibration and higher-resolution modeling.
Future improvements will focus on integrating additional observational data, including temperature and sea surface height, to better capture the complex dynamics of offshore regions. These findings underscore the importance of regional calibration and high-resolution modeling in refining operational ocean predictions. Ultimately, a strengthened Coastal-KOOS framework will enhance the accuracy of search-and-rescue operations, environmental monitoring, and disaster response in waters around the Korean peninsula.