We propose a new approach to estimate the vertical variability of the salinity field. The method is based on combined vertical modes of temperature T and salinity S and reconstructs salinity profiles via a weighted least squares procedure. The major advantages of this new approach over using a climatological T-S relationship are that seasonal-to-interannual variability is better taken into account and that the method combines in a consistent way different sources of information such as T and S profiles and sea surface height. The present results estimate the salinity along 165 ¿E in the western Pacific Ocean for the 1993--1998 period and emphasize the importance of the salinity on sea level variability. The reverse problem of estimating the salinity variability along the water column from the satellite-derived sea level and sea surface fields is also investigated. Finally, comparison with in situ salinity observations demonstrates the possibility of extracting useful information about the salinity variability from the TOPEX/Poseidon altimeter data. ¿ 2000 American Geophysical Union