Intercomparison of Global Sea Surface Salinity from Remote Sensing, Reanalysis and In-situ Products

Abstract

<p><em>Sea surface salinity (SSS) is an important indicator of hydrological cycle, oceanic processes and climate variability, and has been obtained from various methods including remote sensing, in-situ observations and numerical modelings. Due to the differences of instruments used, error correction algorithm and gridding strategy, each dataset has unique strengths and weaknesses. In this study, we conducted a multi-scale comparison of SSS among eight datasets, including satellite-based, in-situ-based and ocean reanalysis products from 2012 to 2020. Compared with WOA18 climatology, all products show good consistency in describing the dominant mode of global SSS distribution. Among eight datasets, the ISAS20 product is of the best quality, and observation-based products are generally more accurate than reanalysis products. Analysis on zonal average shows that positive bias appears in subtropic regions while negative bias distributes in subpolar areas. It was found that reanalysis products have significantly large negative biases at the polar region compared with satellite products and in-situ observations. On both the seasonal and interannual scales, high correlation coefficients (0.65-0.95) are found in the global mean SSSs between individual satellite products, in-situ analysis and ocean reanalysis products, with the differences relatively smaller among the same types of datasets. This analysis provides information on the consistency and discrepancy of different SSS products to guide future use, such as improvements to ocean data assimilation and the quality of satellite-based data.</em></p>