Evaluation of SMAP, SMOS, and AMSR2 Soil Moisture Products Based on Distributed Ground Observation Network in Cold and Arid Regions of China

Long-term surface soil moisture (SM) data are increasingly needed in water budget and energy balance analysis of watersheds. The performance nine remotely sensed SM products from AMSR2, SMOS SMAP missions, evaluated based on observations collected distributed observation networks the Heihe River Basin (HRB) China during 2013 to 2017. Results show that Level 3 Dual Channel Algorithm (DCA) retrievals reflect seasonal variations well with high temporal correlations ~0.7 accuracy within 0.04 m3/m3 terms unbiased Root Mean Squared Error (ubRMSE) over grassland HRB. level present increased underestimation ubRMSE ~0.10 as vegetation increases. newly published SMOS-IC product version 2 outperforms improved correlation coefficient above 0.4 lower ~0.05 m3/m3. AMSR2 Land Parameter Retrieval (LPRM) extremely large overestimation vegetated regions HRB, especially C-band products. Drastically biases observed Japan Aerospace Exploration Agency (JAXA) product. uncertainty analyses indicate different parameterization schemes optical depth (VOD) inputs could be one main reasons resulting systematic overestimation/underestimation AMSR2/SMOS/SMAP retrievals. findings aim provide insights into studies algorithms refinements fusions