Investigating the Potential of Downscaling Approaches for SMAP Radiometer Soil Moisture in Agroforestry Areas, China

Accurate high-spatial-resolution soil moisture content (SMC) datasets are crucial for applications, such as erosion modelling, flood forecasting, and agricultural production. Downscaling is an effective way to convert coarse satellite observations to a finer spatial resolution. Two downscaling approaches were proposed to improve the spatial resolution of the soil moisture active passive (SMAP) radiometer SMC. A downscaling method (method 1) based on the triangular feature space concept was developed to express SMAP L3 SMC as a polynomial function of the Global Land Surface Satellite leaf area index, preprocessed synthetic land surface temperature, and microwave polarization difference index. A second downscaling method (method 2) based on the simulated datasets was developed to express high-resolution SMC as a function of coarse-resolution SMC and sentinel-1 synthetic aperture radar observations. Downscaled SMC (1 km) was evaluated by the
in situ
measurements and compared by the SMAP L2 active and passive SMC product in agroforestry areas, China. The results showed that the two downscaling methods could effectively capture the spatial variability of soil moisture at 1-km spatial scales. The root-mean-square error (RMSE) of downscaled SMC for grass, shrub, and forestland is 0.052–0.055 cm
³
cm
⁻³
, 0.063–0.069 cm
³
cm
⁻³
, and 0.067–0.073 cm
³
cm
⁻³
, respectively. The accuracies of method 1, method 2, and SMAP L2 SMC in the grassland were higher than those in the shrubland and forestland. Overall, the
R
and RMSE between the downscaled soil moisture from method 1, method 2, and SMAP L2 SMC were 0.613, 0.626, and 0.619 and 0.051 cm
³
cm
⁻³
, 0.041 cm
³
cm
⁻³
, and 0.45 cm
³
cm
⁻³
, respectively. The active and SMAP passive microwave combination method has great potential for soil moisture downscaling in agroforestry areas in China.