Data Assimilation of Vertical Current Profiles Using 4D-Var

Abstract

Four dimensional variational (4D-Var) data assimilation is carried out using Regional Ocean Modeling System (ROMS) and the vertical current profiles which are measured by Osaka Bay Water Quality Observation System in this paper. The applicability of 4D-Var is tested through the sensitivity analyses in which parameters for 4D-Var and sets of control variables are changed. The RMSE of velocity at the observation stations decreases more with the number of inner-loop of 10, the radii of horizontal and vertical error covariance of 5 km and 3m respectively. The effect of the 4D-Var appears in the analyses of velocity near the surface at some stations where there is larger difference from the observations and those of surface salinity even though salinity is not assimilated directly.