Premium
Optimum interpolation analysis of A quarius sea surface salinity
Author(s) -
Melnichenko Oleg,
Hacker Peter,
Maximenko Nikolai,
Lagerloef Gary,
Potemra James
Publication year - 2016
Publication title -
journal of geophysical research: oceans
Language(s) - English
Resource type - Journals
eISSN - 2169-9291
pISSN - 2169-9275
DOI - 10.1002/2015jc011343
Subject(s) - sss* , satellite , argo , buoy , interpolation (computer graphics) , environmental science , scale (ratio) , spurious relationship , grid , meteorology , geodesy , computer science , remote sensing , climatology , statistics , mathematics , geology , geography , physics , cartography , telecommunications , oceanography , astronomy , artificial intelligence , frame (networking)
A new high‐resolution sea surface salinity (SSS) analysis has been produced using Aquarius satellite observations from September 2011 to June 2015. The motivation for the new product is twofold: to produce Level‐4 SSS analysis that is consistent with existing in situ observations such as from Argo profile data, and to reduce the large‐scale satellite biases that have existed in all versions of the standard Level‐3 Aquarius products. The new product is a weekly SSS analysis on a nearly global 0.5° grid. The analysis method is optimum interpolation (OI) that takes into account analyzed errors of the observations, specific to the Aquarius instrument. The method also includes a large‐scale correction for satellite biases, filtering of along‐track SSS data prior to OI, and the use of realistic correlation scales of SSS anomalies. All these features of the analysis are shown to result in more accurate SSS maps. In particular, the method reduces the effects of relative biases between the Aquarius beams and eliminates most of the large‐scale, space‐varying, and time‐varying satellite biases relative to in situ data, including spurious annual signals. Statistical comparison between the weekly OI SSS maps and concurrent buoy data demonstrates that the global root‐mean‐square error of the analysis is smaller than 0.2 pss for nearly all weeks over the ∼4 year period of comparison. The utility of the OI SSS analysis is also exemplified by the derived patterns of regional SSS variability.