An Empirical Sea Ice Correction Algorithm for SMAP SSS Retrieval in the Arctic Ocean

Satellite observed sea surface salinity (SSS) reflects the spatial and temporal variability of surface freshwater, which is critical to monitoring the climate change in the Arctic Ocean. Our previous study found that SMAP SSS shows signatures consistent with the inter-annual anomalies of observed sea ice concentration and river discharge, but with large uncertainty (~1 psu) compared with limited in-situ data. One of error sources is the un-corrected sea ice contamination effect. The JPL SMAP algorithm retrieves SSS at each wind-salinity-cell if the matchup sea ice concentration (SIC) is less than 3% with no ice correction implemented yet. Since L-band brightness temperature (TB) of sea ice is much higher than that of seawater, SSS retrieved from TB from a field of view (FOV) mixed with water and ice will result in false fresh signature if the sea ice effect is not accurately accounted for. In this study, we develop an empirical observation-driven sea ice correction algorithm. We characterize the sea ice signature using SMAP TB and ancillary SIC data. The sea ice effect is corrected near ice edge where SIC is under a predetermined threshold for SSS retrieval. We consider the seasonal variation of TB over ice to be likely associated with seasonal change of physical temperature and summer melting pond. The sea ice fraction (ICEF) is calculated by integration of SIC over SMAP FOV weighted by antenna gain patterns. An empirical sea ice correction algorithm is proposed. The impact of sea ice correction is demonstrated by comparing TBs with or without sea ice correction.