Comparison of multiple salinity datasets: upper ocean salinity and stratification in the tropical Pacific during the Argo period

Ocean salinity is an important variable that affects the ocean stratification. We compared the salinity and ocean stratification in the tropical Pacific derived from the Argo (Array for Real-time Geostrophic Oceanography data), EN4 (Ensemble 4 analysis), SODA (the Simple Ocean Data Assimilation reanalysis), IAP (Institute of Atmospheric Physics data), and ORAS4 (Ocean Reanalysis System 4) over 2005–2017. Results show that the spatial distribution of climatological mean of sea surface salinity (SSS) in all the products is consistent, and the low salinity region showed large deviation and strong dispersion. The Argo has the smallest RMSE and the highest correlation with the ensemble mean, while the IAP shows a high-salinity deviations relative to other datasets. All the products show high positive correlations between the sea surface density (SSD) and SSS with respect to the deviations of climatological mean from ensemble mean, suggesting that the SSD deviation may be mainly influenced by the SSS deviation. In the aspect of the ocean stratification, the mixed layer depth (MLD) climatological mean in the Argo shows the highest correlation with the ensemble mean, followed by EN4, IAP, ORAS4, and SODA. The Argo and EN4 show thicker barrier layer (BL) relative to the ensemble mean while the SODA displays the largest negative deviation in the tropical western Pacific. Furthermore, the EN4, ORAS4, and IAP underestimate the stability in the upper ocean at the depths of 20–140 m, while Argo overestimates ocean stability. The salinity fronts in the western-central equatorial Pacific from Argo, EN4, and ORAS4 are consistent, while those from SODA and IAP show large deviations with a westward position in amplitude of 0°–6° and 0°–10°, respectively. The SSS trend patterns from all the products are consistent in having ensemble mean with high spatial correlations of 0.95–0.97.