Presented at the 2014 Ocean Sciences MeetingWe validate Aquarius v2.0 level-2 sea surface salinities against Argo 5 m salinities for 27 Aug 2011 through 31 Jul 2013, using 16,625 collocated Argo/Aquarius data pairs. The glΔbal mean of the salinity difference is +0.003 psu but the standard deviation of ΔS ranges from 0.35 to 0.78 for increasing latitudes up to 60 degrees. Low temperatures, high winds, and rain account for 40% of the standard deviation. We discuss three potential collocation errors due to horizontal, temporal and vertical displacements of measurement. We show that none of these are significant for global validation of Aquarius. In particular, we use TRMM data to analyze the contribution of vertical salinity stratification to validation error. We show evidence of strong vertical stratifications >0.1 psu lasting ~3-6 h, but show that these events occur infrequently and contribute less than 0.03 psu bias in the tropics and 0.02 psu globally. We also argue that subsurface sampling provides a less noisy in-situ reference than surface sampling and should therefore be preferred for large-area validation of satellite-based salinity measurement.