Comparison of Satellite-Derived Sea Surface Salinity Products from SMOS, Aquarius, and SMAP

Global sea surface salinity (SSS) has been obtained from space since 2009 by the Soil Moisture and Ocean Salinity (SMOS) mission and has been further enhanced by Aquarius in 2011 and Soil Moisture Active-Passive (SMAP) missions in 2015. Due to the differences between SMOS, Aquarius, and SMAP in the instruments used, retrieval algorithms, and error correction strategies, the quality of their gridded products are different. In this paper, we have assessed the accuracy of three satellite products using in situ gridded data and buoy data. Compared with gridded in situ salinity measurements, the monthly Aquarius data are of the best quality, reaching the mission target accuracy (0.2PSU) in the open ocean. SMOS and SMAP agree well with in situ data in the open ocean between 40 degrees S and 40 degrees N (root-mean-square deviation [RMSD]: SMOS 0.211PSU, SMAP 0.233PSU). The RMSD of SMAP is lower than that of SMOS at high latitudes, which may due to the fact that the roughness correction of SMAP is based on the Aquarius geophysical model function. Meanwhile, time series comparison of salinity measured at 1m by moored buoys indicates that satellite SSS captures variability of SSS at weekly time scales with reasonably good accuracy (RMSD: SMOS 0.25PSU, SMAP 0.26PSU), when excluding suspicious buoy data. Synergetic analysis of satellite SSS and Argo data indicates that satellite SSS can be applied as real-time quality control of buoy 1-m salinity data. Plain Language Summary It is important to analyze salinity variations and understand their relationship with the global hydrological cycle and climate changes. The performance of different sea surface salinity products from three salinity satellites is evaluated based on in situ salinity measurements in this study. The results show that the satellite salinity products agree well with in situ data in the open ocean and can be used as real-time quality control of buoy sea surface salinity data. However, large discrepancies can be found in high latitudes and in the regions along the coast.