Inverse estimation of empirical parameters used in a regional ocean circulation model

Significant impacts of the subgridscale parameterizations have been emphasized in modeling the ocean circulations, but various different parameter values are applied to similar numerical studies often without any justification. This study objectively estimates a set of empirical parameters along with their uncertainty for circulation modeling of the East Asian Marginal Seas. The solutions for 14 major parameters are obtained by using model Green's functions with constraints of climatological temperature and salinity data. The largest cost function reduction occurs in the eastern Japan Sea associated with the sharp gradient of the Polar Front. The calibrated parameters are also validated with realistic transport and path of the Kuroshio in the final experiment. The inverse estimation shows that freshwater discharges from small rivers can be attributed to the coastal precipitation over a strip of land 74-81 km wide. The thickness diffusion coefficient may be similar to the isopycnal and horizontal diffusion coefficients in their magnitude. The accelerated initial condition also contributes to the cost function reduction resulting in weaker trends of deep temperature. Most importantly, estimated scaling factors suggest a significant reduction of the reanalyzed wind stress data for more realistic ocean circulations. Possible reasons for the momentum missing are also discussed.