Evaluating surface soil moisture characteristics and the performance of remote sensing and analytical products in Central Asia

Reanalysis and remotely sensed soil moisture (SM) products are important for monitoring hydrology and the terrestrial carbon cycle, especially in water-limited arid regions. However, it is essential to assess the reliability of SM products before they are used. In this study, surface SM from the European Reanalysis-Interim v5 (ERA5), Essential Climate Variables (ECV), the Famine Early Warning Systems Network (FEWS) Land Data Assimilation System (FLDAS), and the Global Land Data Assimilation System (GLDAS) were assessed based on in-situ data and the Three-Cornered Hat (TCH) approach in a typical arid region, Central Asia (CA). The validation based on in-situ data indicated that all three products underestimated the observed surface SM, except for ERA5. The ERA5 SM was closest to the in-situ observations, averaging a higher correlation coefficient of 0.59, followed by the GLDAS, ECV, and FLDAS. The TCH method revealed low uncertainties (<0.05 m³/m³) for all four SM products, and high uncertainties were found at high altitudes. For ERA5 and ECV SM products, the uncertainty is high in the shrublands and forests, respectively. FLDAS and GLDAS SM products performed better in terms of uncertainty compared to ERA5 and ECV, with an average uncertainty below 0.03 m³/m³, except for forests. The above assessment results indicate that, compared with the other three SM products, ERA5 has a better performance and more potential for application in CA. There was an upward trend in surface SM from 1982 to 2018 in CA. The relationship between the four SM products and aridity index and drought events showed that GLDAS could capture the hydrothermal conditions and drought events, followed by FLDAS, ERA5, and ECV. Our research is intended to provide some reference for the application of gridded SM products to ecosystem and water resources monitoring in arid regions.