Abstract: Soil surface water content, which exists at the land air interface, is an important variable to simulate the global hydrological, energy and carbon cycles and their interactions, and is crucial for climate and earth system research. In order to evaluate the accuracy and applicability of the soil moisture products of FY-3B, FY-3C and AMSR2 in Xilingol grassland and different land cover types, authors applied time series correlation (R), root mean square error (RMSE) , the average deviation (Bias) and unbiased root mean square error (ubRMSE) were to analyze the reliability on these three grid products. The results showed that: (1) on a daily scale, FY-3B and FY-3C soil moisture products overestimated the land surface water content, while AMSR2 underestimated the soil water content. Each evaluation index showed that the three soil water content products can capture the time change trend of surface soil water content in Xilingol grassland. (2) On a monthly scale, compared with AMSR2, the monthly distribution of FY-3B and FY-3C soil moisture products was more consistent with the measured soil moisture distribution. (3) In the growing season scale, the three satellite soil moisture products reflected the soil moisture status of each grassland subtype. FY-3B and FY-3C soil moisture products had good applicability in areas with high vegetation coverage, AMSR2 had good reliability in areas with low vegetation coverage, and the three products had complementary functions in monitoring soil moisture. FY-3B had the best consistency with the change trend of in-situ soil water content, FY-3B and FY-3C can better reflect the variation of soil moisture in the west and east of the study area, and AMSR2 can well capture the variation of soil moisture in the middle of the study area.

Key words: Microwave remote sensing, Soil moisture, Applicability, FY-3B, FY-3C, AMSR2