Study on collaborative inversion of surface soil moisture in farmland based on multisource remote sensing data

Soil moisture(SM) in farmland is vital for crop growth and yield. Accurately acquiring SM under crop cover is essential for agricultural production and drought monitoring. Vegetation cover can affect microwave data, reducing inversion accuracy. To address this, we improved the Water-Cloud Model (WCM) to better remove vegetation scattering signals from microwave backscattering coefficients. In this study, conducted in Dachang Hui Autonomous County, Langfang City, Hebei Province, we used Leaf Area Index (LAI) and Fractional Vegetation Cover (FVC) to enhance the WCM. The improved model was compared with the original one to extract the soil backscattering coefficient under crop cover. We then analyzed the relationship between soil backscattering coefficient, dielectric constant, and surface combined roughness using the soil dielectric model and Advanced Integral Equation Model (AIEM). Combining this with measured data from sampling points, we constructed and implemented an inversion model for surface SM under winter wheat cover. Results show that the vegetation water content inversion model built using several vegetation indices and least squares method has a determination coefficient of 0.755 with measured vegetation water content data. The WCM improvement based on vegetation water content empirical equations significantly enhanced the elimination of vegetation influence. The final SM inversion model achieved determination coefficients R² of 0.727, 0.784, and 0.751 with measured SM data, and root mean square errors of 0.02, 0.018, and 0.019 respectively, indicating its ability to reflect SM in the study area. This study combines semi-empirical and theoretical models to achieve quantitative inversion of SM under crop cover, offering a reference for microwave data application in agricultural remote sensing.