Presented at the 2014 Ocean Salinity Science and Salinity Remote Sensing WorkshopImproving the SSS (Sea Surface Salinity) constrain at various scales is an important issue for ocean forecasting. It concerns the short term meso-scale and the seasonal anomalies. Both strongly depend on the surface freshwater budget (evaporation, precipitation and runoff). Presently, it is not yet possible to fully remove SSS biases with the poorly sampled Argo network data near the surface (depth <5m). It encourages us to find the best way to deal with SSS data observed from space (SMOS and Aquarius). Taking into account different errors associated with different scales could fill the gap. In this context, Mercator Ocean and CLS give an overview of dealing with the SSS issues within their operational and reanalysis frameworks. Comparisons with SMOS/Aquarius and systematic biases found in reanalysis and operational results are first presented with the global 1/4Â° ocean forecasting system. They give us indications on how to take into account errors and/or to sort out corrections. We also show results of first SSS data assimilation experiments performed with this system. Finally, in the context of the operational INDESO project, comparisons of monthly SSS fields with SMOS and Aquarius in the Indonesian seas seem to corroborate a freshwater forcing bias in the representation of South China Sea upper waters.