Estimation of quasi-full spatial coverage soil moisture with fine resolution in China from the combined use of ERA5-Land reanalysis and TRIMS land surface temperature product

Knowledge of surface soil moisture (SSM) at various scales is essential for understanding energy and water exchange between the atmosphere and land surface. In this study, we propose a practical approach based on the synergy of all-weather land surface temperature (LST) and reanalysis data to generate a daily/1-km SSM with quasi-full spatial coverage. In this approach, the day/night LST difference and SSM derived from the ERA5-Land with a spatial resolution of 0.1° were first used to develop linear relationships between them under various vegetation densities (sparse, medium, and dense) for four dominant land cover types (forest, cropland, grassland, and barren) in China. The relationships were subsequently used in an all-weather 1-km LST product to generate an initial daily/1-km SSM dataset in 2019. In addition, two blended microwave-based SSM products were applied to correct the initial dataset by assuming that the aggregated SSM values within the coarse scale were unbiased when compared with the original microwave-based SSM products. Furthermore, two blended microwave-based SSM products were used to correct the initial dataset with the assumption that aggregated SSM values within the coarse scale are unbiased when compared to the original microwave-based SSM products. Finally, the estimated SSM were preliminarily evaluated against the ground in-situ measurements at cropland under different aridity conditions. A satisfactory accuracy was obtained with an unbiased root mean square error (ubRMSE) of ∼0.05 m3/m3, which is not only close to the ubRMSE requirements of 0.04 m3/m3 for the SSM estimates in most practical applications, but is also characterized with quasi-full spatial coverage. •Reanalysis data and land surface temperature product were integrated to estimate soil moisture.•Soil moisture measured under different aridity conditions were used to assess estimated value.•The proposed method reveals satisfactory accuracy with ubRMSE of ∼0.05 m3/m3.