Direct velocity observations remain sparse in the western Pacific, limiting regional ocean analyses in a region shaped by energetic currents, eddies, and sharp thermohaline fronts. Here we test whether satellite-derived surface currents can improve a 4-km ROMS 4D-Var ocean analysis during 2020. Surface currents are inferred from 3-hourly Himawari-8 sea surface temperature gradient sequences using the GOFLOW method and assimilated together with conventional sea surface temperature, sea surface height, Argo, and cruise observations.

Adding the current observations introduces a dense but cloud-limited velocity constraint. The current-assimilating experiment slightly worsens the analysis fit to sea surface temperature and sea surface height, suggesting that the fixed inner-loop configuration does not fully reconcile the expanded observing system. However, independent validation shows clear benefits. Near-surface drifter velocity and temperature errors decrease, and independent hydrographic sections show seasonally dependent improvements in subsurface temperature and salinity. The dynamical adjustment is also coherent. Upper-ocean kinetic energy is reduced across most resolved wavelengths in summer and mainly at wavelengths longer than about 100 km in winter, with changes extending below the surface.

These results suggest that high-frequency geostationary satellite currents can provide useful dynamical information for regional ocean state estimation, even when observations are spatially incomplete.