| Sea surface salinity (SSS) is a key parameter for studying the effects of the ocean on global climate and ocean circulation, and satellite remote sensing detection functions as the most effective means to obtain SSS data. Currently, L-band SMOS and Aquarius / SAC-D satellites are being used to detect SSS based on observing data and the physical mechanism. However, in some near-shore areas, due to the inflow of freshwater and terrestrial radio frequency interference, the precision of salinity satellite products is relatively low. This paper uses the measured data from the "Dong Fang Hong 2" scientific expedition ship and SMOS data to predict SSS by the Bayesian network model for the first time in the South China Sea, and assesses the model with validation data sets (measured Argo salinity). Analysis results show that the model error and validation error is 0.47 psu and 0.45 psu, respectively, while the precision of SMOS Level 2 products is 1.90 psu and 1.82 psu, respectively. This model provides a new method to predict SSS. |
