New insights into SMOS sea surface salinity retrievals in the Arctic Ocean

Type Article
Date 2020-11
Language English
Author(s) Supply Alexandre1, Boutin Jacqueline1, Vergely Jean-Luc2, Kolodziejczyk NicolasORCID5, Reverdin Gilles1, Reul NicolasORCID3, Tarasenko Anastasiia4, 5
Affiliation(s) 1 : Laboratoire d'océanographie et du climat : expérimentations et approches numériques - Institut Pierre Simon Laplace (LOCEAN-IPSL), Sorbonne Université-CNRS-IRD-MNHN, Paris, France
2 : ACRI-St, Guyancourt, France
3 : Laboratoire d'Océanographie Physique et Spatiale (LOPS), Univ. Brest, CNRS, Ifremer, IRD, Brest, France
4 : Arctic and Antarctic Research Institute, Saint-Petersburg, Russia
5 : Laboratoire d'Océanographie Physique et Spatiale (LOPS), Univ. Brest, CNRS, Ifremer, IRD, Brest, France
Source Remote Sensing Of Environment (0034-4257) (Elsevier BV), 2020-11 , Vol. 249 , P. 112027 (24p.)
DOI 10.1016/j.rse.2020.112027
WOS© Times Cited 13
Keyword(s) SMOS, Sea surface salinity, Arctic Ocean

Since 2010, the Soil Moisture and Ocean Salinity (SMOS) satellite mission monitors the earth emission at L-Band. It provides the longest time series of Sea Surface Salinity (SSS) from space over the global ocean. However, the SSS retrieval at high latitudes is a challenge because of the low sensitivity L-Band radiometric measurements to SSS in cold waters and to the contamination of SMOS measurements by the vicinity of continents, of sea ice and of Radio Frequency Interferences. In this paper, we assess the quality of weekly SSS fields derived from swath-ordered instantaneous SMOS SSS (so called Level 2) distributed by the European Space Agency. These products are filtered according to new criteria. We use the pseudo-dielectric constant retrieved from SMOS brightness temperatures to filter SSS pixels polluted by sea ice. We identify that the dielectric constant model and the sea surface temperature auxiliary parameter used as prior information in the SMOS SSS retrieval induce significant systematic errors at low temperatures. We propose a novel empirical correction to mitigate those sources of errors at high latitudes.

Comparisons with in-situ measurements ranging from 1 to 11 m depths spotlight huge vertical stratification in fresh regions. This emphasizes the need to consider in-situ salinity as close as possible to the sea surface when validating L-band radiometric SSS which are representative of the first top centimeter.

SSS Standard deviation of differences (STDD) between weekly SMOS SSS and in-situ near surface salinity significantly decrease after applying the SSS correction, from 1.46 pss to 1.28 pss. The correlation between new SMOS SSS and in-situ near surface salinity reaches 0.94. SMOS estimates better capture SSS variability in the Arctic Ocean in comparison to TOPAZ reanalysis (STDD between TOPAZ and in-situ SSS = 1.86 pss), particularly in river plumes with very large SSS spatial gradients.

Full Text
File Pages Size Access
24 10 MB Access on demand
Author's final draft 64 10 MB Open access
Top of the page