A three-dimensional variational data assimilation system for the South China Sea: preliminary results from observing system simulation experiments

Abstract

A three-dimensional variational data assimilation (3DVAR) system based on the Regional Ocean Modeling System (ROMS) is established for the South China Sea (SCS). A set of Observing System Simulation Experiments (OSSEs) are performed to evaluate the performance of this data assimilation system and investigate the impacts of different types of observations on representation of three-dimensional large-scale circulations and meso-scale eddies in the SCS. The pseudo-observations that are examined include sea surface temperatures (SSTs), sea surface heights (SSHs), sparse temperature/salinity (T/S) profiles, sea surface velocities (SSVs), and sea surface salinities (SSSs). The results show that SSHs can extend their impacts into the subsurface or even the deep ocean while other surface observations only have impacts within surface mixed layer. SSVs have similar impacts though confined to their spatial coverage, suggesting that SSVs could be a substitute of SSHs nearshore where SSHs are of poor quality. Despite their sparseness, the T/S profiles improve the representation of the temperature and salinity structures below the mixed layer, and a combination of T/S profiles with surface observations leads to a better representation of the meso-scale eddies. Based on the OSSE results, an affordable observing network for the SCS in the near future is proposed.