Impact of assimilating surface salinity from SMOS on ocean circulation estimates

Abstract

In a pilot attempt, the GECCO2 synthesis system is being used to investigate the impact of SMOS sea surface salinity (SSS) observations on estimates of SSS and freshwater fluxes. The paper focuses on the period 2010-2011, during which, in addition to traditional in situ and satellite observations, SMOS SSS is assimilated. A prior SMOS SSS error field is inferred through a comparison of the satellite data with in situ salinity data and reveals large biases (>1 g/kg) in the SMOS product near continents and in the Southern Ocean. Employing this error estimate in the assimilation procedure leads only to an insignificant impact of SMOS SSS on the estimated ocean state. However, when reducing the error artificially by a factor of 10, the SMOS data can be reproduced well in the interior ocean. In this case, the previously remaining positive model bias with respect to in situ salinity is changed to a negative bias while the misfit slightly increased. The clear freshening can be attributed to the SMOS bias with respect to in situ data. The associated increase in fresh-water input in the tropical oceans enhances slightly the correspondence of the estimated fluxes to the independent satellite-based estimate from HOAPS except for the South Pacific and South Atlantic. On short-timescales, changes in the estimated surface salinity result primarily from changes in surface freshwater fluxes, while over longer periods ocean dynamics become increasingly more important for changing the near-surface salinity.