A high-resolution soil moisture content analysis product in Jing-Jin-Ji using the successive correction method

Abstract

Two soil moisture content analysis products, each with a spatial resolution of 0.0625° and a temporal resolution of 1 h from 1 October 2015 to 1 October 2021 in the Jing-Jin-Ji area, have been developed and evaluated. Using a quality control (QC) scheme based on the biweight method and a successive correction method (SCM) with different weighting functions introduced from the Cressman and/or Barnes methods, the soil moisture collected from China Land Data Assimilation System version 2.0 was taken as the background data. Two soil analysis products with induced hourly station soil observations were constructed and further compared with in situ observations and satellite products. Results show that with the biweight method, the outliers of soil moisture could be identified, and that the number of these was less than 10% of the actual reported samples. The analysis product is more accurate with the Cressman weighting function than with the Barnes as the root mean square error was reduced by 30% to 0.06, while the correlation coefficient increased by 73% to 0.59, compared to the control. These results indicate that the Cressman product is a useful and promising soil moisture data set. The proposed Cressman method is computationally simple and efficient, so the product can be easily updated to keep pace with significant increases in the volume of soil moisture data. The methods used here for QC, SCM, and comparison could help evaluate and improve land surface models and their soil moisture products.