AUTHOR=Ma Teng , Han Ling , Liu Quanming 

TITLE=Retrieving the Soil Moisture in Bare Farmland Areas Using a Modified Dubois Model

JOURNAL=Frontiers in Earth Science

VOLUME=Volume 9 - 2021

YEAR=2021

URL=https://www.frontiersin.org/journals/earth-science/articles/10.3389/feart.2021.735958

DOI=10.3389/feart.2021.735958

ISSN=2296-6463

ABSTRACT=The accurate retrieval of soil moisture in farmland using Synthetic Aperture Radar (SAR) depends on the accurate description of ground surface roughness and other surface conditions. This study developed a modified Dubois model to retrieve the soil moisture. The model eliminates the influence of the incident angle, and takes a parameter established using Vertical transmit and Horizontal receive (VH) backscattering coefficient to describe surface roughness and systemic fluctuations. First, the Normalized Different Vegetation Index (NDVI) from 2017 to 2018 was calculated using Sentinel-2 data to obtain the 13 typical bare farmland plots with different soil textures in the Spanish Duero Basin, and the series SAR backscattering coefficient was obtained from Sentinel-1 data. The change trends of the VH polarization backscattering coefficients, Vertical transmit and Vertical receive (VV) polarization backscattering coefficients and the polarization ratio were analyzed under conditions of cultivated and uncultivated,  and rainfall were also taken into account. Then, the modified Dubois model was established. The model was verified using time series soil moisture observation data in situ. Finally, the retrieval accuracy of soil moisture under different conditions was discussed. The results show that the modified model can retrieve soil moisture with high precision, and the RMSE can reach 0.059 cm3cm-3. Under different soil texture conditions, clay has the lowest retrieval accuracy because its backscattering scattering coefficient has a low sensitivity to soil moisture, and sand has the highest retrieval accuracy.