Assimilation of the surface geostrophic currents estimated from merged satellite altimeter data in the East Sea

Abstract

Data assimilation (DA) is the process to determine the best possible ocean state using the model and observation. Usually, ocean DA used only active tracer dataset as temperature and salinity. In this study, the surface geostorphic currents calculated from satellite altimeter data were assimilated using the Ensemble Kalman Filter (EnKF) method.To match frequency between the observed geostrophic current and model velocity, high frequency variability of the velocity from the model was filtered using the moving average. Optimal time window was 31 days, which was calculated using the observed current from ocean buoys and the surface geostrophic current. To examine the effects of data assimilation using geostrophic current, eight numerical experiments were designed: Free run (EXP 1), DA using only active tracer dataset (EXP 2), DA using active tracer and surface geostrophic current (EXP 4), shorter and longer time window (EXP 3, 5), lower and higher observation error (EXP 6, 7) than that in EXP 5, use of univariate covariance matrix in the EnKF (EXP 8). The improvement were evaluated using correlation and normalized root mean square error between the simulated variables from each experiment and observation data. Performances in hydrographic variables (temperature and salinity) of EXP 2, 8 and 7 were higher 17.2%, 15.3% and 11.2% than that from the free run, respectively. Performances in surface current velocity of EXP 6, 8 and 3 were higher 85.4%, 72.1% and 70.4% than the free run, respectively. Total performances of EXP 8, 2 and 7 improved 17.1%, 16.8% and 13.0% than the free run during the 4 summer months, respectively. Data assimilation of the surface geostrophic current improved the surface velocity. However, Performances in hydrographic variables were slightly reduced in the East Sea. The use of univariate background error covariance matrix in DA can limit the degradation in the hydrographic performances.