GOSUD aims at assembling in-situ observations of the world ocean surface collected by a variety of ships and at distributing quality controlled datasets. At present time the variables considered by GOSUD are temperature and salinity. The GOSUD data are mostly collected using thermosalinographs (TSG) installed on research vessels, on commercial ships and in some cases on sailing exploration ships. GOSUD manages both near-real time (RT and NRT) data and delayed mode (DM-reprocessed) data. The GOSUD GDAC is hosted by the Coriolis data centre (France) and a back-up (permanent archived) is performed on a daily basis by NCEIS (NOAA's National Centers for Environmental Information).

The annually binned Sea Surface Salinity, Temperature and Density data set covers a large part of the North Atlantic subpolar gyre. It is based on T and S near-surface data collected since December 1895 to August 2018 from all data sources, including Voluntary Observing Ships subsequently validated. This annual product presents deviations from a climatological seasonal cycle and has furthermore been smoothed in time with a 1/4, 1/2, 1/4 running-mean filter. Density time series are estimated from those annual time series, considering winter (D-J-F) climatology. Uncertainty estimates due to sampling are also provided. Data for December 2017-August 2018 are also reported as 2018, but are temporary and should not be used without further check. Long gaps have been filled by linear interpolation, and are attributed largest error.

The gridded Sea Surface Salinity (SSS) data set covers the region between 95°W – 20°E and 50°N – 30°S in the Atlantic Ocean. It is based on available data collected from 1970 to 2016 mostly from Voluntary Observing Ships, PIRATA moorings and Argo profilers, and subsequently validated. This monthly SSS product is gridded using an objective mapping at the spatial resolution 1° x 1°. It is distributed with its associated error fields. It is an update of the SSS product presented in Reverdin et al (2007).

10 years of L-Band remote sensing Sea Surface Salinity (SSS) measurements have proven the capability of satellite SSS to resolve large scale to mesoscale SSS features in tropical to subtropical ocean. In mid to high latitude, L-Band measurements still suffer from large scale and time varying biases. Here, a simple method is proposed to mitigate the large scale and time varying biases. First, in order to estimate these biases, an Optimal Interpolation (OI) using a large correlation scale is used to map SMOS and SMAP L3 products and is compared to equivalent mapping of in situ observations. Then, a second mapping is performed on corrected SSS at scale of SMOS/SMAP resolution (~45 km). This procedure allows to correct and merge both products, and to increase signal to noise ratio of the absolute SSS estimates. Using thermodynamic equation of state (TEOS-10), the resulting L4 SSS product is combined with microwave satellite SST products to produce sea surface density and spiciness, useful to fully characterize the surface ocean water masses. The new L4 SSS products is validated against independent in situ measurements from low to high latitudes. The L4 products exhibits a significant improvement in mid-and high latitude in comparison to the existing SMOS and SMAP L3 products. However, in the Arctic Ocean, L-Band SSS retrieval issues such as sea ice contamination and low sensitivity in cold water are still challenging to improve L-Band SSS data.