Data Assimilation of Argos profiles in North-west Pacific Model

Abstract. Based on a novel specification of the background error covariance applying to Argos profiles assimilation, an oceanographic three-dimensional variational (3DVAR) data assimilation scheme is set up in the Regional Oceanic Model system (ROMs). Temperature and salinity data extracted from Argos profiles in 2006 have been assimilated into the North-West Pacific Model (NWPM). The quality control is done by comparing background estimation with observations in 2006.Firstly, the assimilated results are compared with merged in-situ data, Sea Surface Temperature (SST) derived from satellite data and reanalysis salinity data. It is found that assimilation of Argos profiles can improve the model results of SST and salinity. Secondly, the Root Mean Square (RMS) difference between model and Argos profiles is analyzed. For the tropic Pacific, the range of RMS temperature (salinity) error are less than 0.83 ℃ (0.11 PSU), decreasing ~ 23.2 % (~ 18.8 %) by comparing with the experiments without data assimilation. For the sub-tropic Pacific Ocean, the RMS of temperature (salinity) is less than 1.43 ℃ (0.135 PSU) and it also shows a decreasing trend after assimilation. It's indicated that the 3DVAR method works well in ROMs and can be used for the operational forecasting systems.