Improvement of SMAP sea surface salinity in river-dominated oceans using machine learning approaches

TLDR

Sea salinity is a key indicator of the global water cycle that influences ocean circulation, yet passive microwave sensors suffer large errors from radio frequency interference and low sea surface temperature, especially in river‑dominated coastal seas. The study aimed to improve the Soil Moisture Active Passive (SMAP) sea surface salinity product over five river‑dominated oceans worldwide by applying three machine‑learning techniques—random forest, support vector regression, and artificial neural network. Four SMAP products and four ancillary data sets were used as input variables for the machine‑learning models. The models lowered SMAP SSS root‑mean‑square error by up to 28 %, with random forest achieving calibration/validation RMSEs of 0.203 and 0.556 psu versus 0.774 psu for the original product, and the corrected SSS accurately reproduced spatiotemporal patterns across all five regions, demonstrating the approach’s potential for operational global SMAP SSS improvement, including coastal and near‑polar areas.

Abstract

Sea salinity is one of the indicators of the global water cycle and affects the surface and deep circulation of the ocean. While passive microwave satellite sensors have been used to monitor sea surface salinity (SSS), the uncertainties from radio frequency interference (RFI) and low sea surface temperature often result in large errors, especially in river-dominated coastal seas. This study investigated the improvement of the Soil Moisture Active Passive (SMAP) SSS over five river-dominated oceans over the globe using three machine learning approaches (i.e., random forest (RF), support vector regression (SVR), and artificial neural network (ANN)). Four SMAP products and four ancillary data used in the SMAP SSS retrieval algorithm were used as input variables to the machine learning models. The results showed that all models improved the SMAP SSS product by up to 28% reduced in the root mean square error (RMSE) for validation, and RF yielded better performance than SVR and ANN. The calibration and validation RMSEs by RF were 0.203 and 0.556 practical salinity unit (psu), while those of SMAP SSS were 0.774 psu. The improved SSS well captured the spatiotemporal patterns of SSS for not only low but also high salinity water for all five regions. The proposed approach can be used to operationally improve the global SMAP SSS product including other coastal areas and the near Polar regions in the future.