Inversion of surface soil moisture under vegetated areas based on optical and microwave remote sensing data

Abstract

【Objective】 The effects of Vertical-Vertical (VV) polarization, Vertical-Horizontal (VH) polarization, and dual polarization on microwave remote sensing retrieval of surface soil moisture (0-10 cm) were investigated. And analyze different data sources (Landsat-8, L8; Sentinel-2, S2) Effect of the Normalized Difference Vegetation Index (NDVI) and the Normalized Difference Water Index (NDWI) on the accuracy of remote sensing estimation of surface soil moisture. 【Method】 Based on the single polarization of VV and VH and the dual polarization model, NDVI and NDWI calculated by S2 and L8 were used to estimate vegetation water content (VWC), eliminate the influence of vegetation on backscattering, and obtain the soil backscattering coefficient. Then the water cloud model was used to invert the surface soil moisture in Daxing District of Beijing. 【Result】 For VV polarization, the estimation accuracy of VV+S2NDWI was best (R2=0.763, RMSE=1.55%). For VH polarization, the estimation accuracy of VH+S2NDVI was best (R2=0.622, RMSE=1.66%); The dual polarization (DP) + S2NDVI had a highest accuracy in retrieving surface soil moisture (R2=0.895, RMSE=0.89%); Compared with the NDWI, the NDVI has a better effect on removing the influence of vegetation, and the NDVI can be used to characterize the VWC. 【Conclusion】 NDVI is more suitable for removing the influence of vegetation in water cloud model, and the dual polarization method has higher accuracy in retrieving surface soil moisture.

Keywords

• moisture
• sentinel-1
• dual polarization
• remote sensing
• ndvi
• ndwi