The Assimilation Results of Ocean Temperature and Salinity Data from GTS in BCC_GODAS 2.0

Ocean temperature and salinity observations data from GTS (Global Telecommunication System) are assimilated in second generation global ocean data assimilation system of Beijing Climate Center (BCC_GODAS 2.0) and the results are analyzed. First, the comparison with SODA (Simple Ocean Data Assimilation) reveals the vertical distribution features of root mean square error (RMSE) of global temperature and salinity in model and assimilation system. The analysis shows that, for the RMSE of temperature with assimilation, compared with the results without assimilation, it has a slight decline with a range of 0—0.3 ℃ in the sea surface layer, and an obvious descent with a range of 0.1—0.7 ℃ in depth from about 100 m to deep layer, but has no obvious variation in depth from the middle and lower mixed layer to about 100 m. For the RMSE of salinity after assimilation, it has a descent with a range of 0—0.2 psu in depth from ocean surface to deep layer. Second, further comparison is made for some vertical cross sections, including the zonal cross section along equator, the meridional cross section along 165°E in Pacific Ocean, the meridional cross section along 30°W in Atlantic Ocean, the longitudinal cross section along 90°E in Indian Ocean. The results show that, generally speaking, the GTS data assimilation improves the temperature and salinity simulation in many aspects including the extension and central intensity of warm sector in mixed layer, the depth of temperature ridge and trough in thermocline, the temperature gradient near thermocline, the extension and central intensity of high and low salinity sector, and so on. Moreover, the further comparison with some ARGO (Array for Real time Geostrophic Oceanography) observation indicates that, in most cases, the RMSE of temperature and salinity profiles has an obvious descent after assimilation, leading to more accurate vertical distribution features of temperature and salinity simulations. For selected single point profiles in different ocean areas in January, after assimilation, the RMSE of temperature and salinity decrease by 0.49 ℃ and 0.19 psu, respectively; for selected profiles in July, the descent of RMSE of temperature and salinity is 0.87 ℃ and 0.18 psu, respectively. The comparison with TAO (Tropical Atmosphere Ocean) data also shows that the assimilation can improve the temperature and salinity features to a certain extent. The BCC_GODAS 2.0 has superiority in some degree in ocean data assimilation, however, there is still some deficiency, and the assimilation effect is not very good in some areas or periods, which may be induced by lack of observations, uncertainties of the data, the imperfection of assimilation system, as well as the simulation capability of model, and so on.