It is accepted that the best way to monitor sea surfacesalinity (SSS) on a global basis is by means of L-band radiometry.However, the measured sea surface brightness temperature(TB) depends not only on the SSS but also on the sea surfacetemperature (SST) and, more importantly, on the sea state, whichis usually parameterized in terms of the 10-m-height wind speed(U10) or the significant wave height. It has been recently proposedthat the mean-square slope (mss) derived from global navigationsatellite system (GNSS) signals reflected by the sea surface couldbe a potentially appropriate sea-state descriptor and could be usedto make the necessary sea state TB corrections to improve theSSS estimates. This paper presents a preliminary error analysis ofthe use of reflected GNSS signals for the sea roughness correctionand was performed to support the European Space Agency’sSoil Moisture and Ocean Salinity (SMOS) mission; the orbit andparameters for the SMOS instrument were assumed. The accuracyrequirement for the retrieved SSS is 0.1 practical salinity unitsafter monthly averaging over 2◦ × 2◦ boxes. In this paper, potentialimprovements in salinity estimation are hampered mainlyby the coarse sampling and by the requirements of the retrievalalgorithm, particularly the need for a semiempirical model thatrelates TB and mss.