Presented at the 2014 Ocean Salinity Science and Salinity Remote Sensing WorkshopSatellite imagery has revealed a series of ocean structures as meandering fronts, eddies and filaments. A high percentage of ocean energy has been found to accumulate at a range of scales (ranging from tens to a few hundred kilometers) known as the ocean mesoscale. Thanks to the SMOS and Aquarius missions, more than four years of satellite-derived sea surface salinity (SSS) data are available. For the first time, satellite data have been providing quasi global, synoptic information of the spatial variability of the ocean surface salinity. However, validation of the satellite data relies on sparse in-situ observations mostly provided by a single observing system, i.e. the Argo autonomous profilers. The aim of this work is to compare the point-to-point validation of satellite-derived SSS products by respect Argo data and the spectral validation through the help of the wavenumber power density spectra (PDS). Several remote sensing SSS products are analyzed: SMOS Level 3 (L3) binned, SMOS L3 optimally interpolated, SMOS L4 data fusion, and Aquarius L3. The study focuses in the North Atlantic Ocean, a challenging region because of the presence of radio frequency interferences (RFI). Analysis of the slope of the various PDS distributions indicates that, despite their different error characteristics, all products share similar spectral properties (k-3 slope) at large scales. Spectral differences arise at shorter scales. The comparison with Argo SSS data indicates that agreement with in situ point-wise data does not imply a realistic spectrum at short scales. Indeed, although Aquarius L3 and the climatology provide the closest values to the in-situ SSS, their slope display a flattening below 300 km scales and strong variance decay below 1500 km scales, respectively, far from the expected theoretical slopes. Theoretical arguments exist to expect similarities between the salinity and temperature PDS. The debate about if, at short scales, salinity fields should display a k-3 or a k-4 slope is highlighted by the results reported here. The SMOS L4 SSS PDS displays a k-3 slope at all scales. On the contrary, OSTIA SST, and the SMOS L3 OI display PDS slope switch from k-3 to k-4 at short scales.