A Method to Assess the Performance of SAR-Derived Surface Soil Moisture Products

• Main Authors: John Beale, Toby Waine, Jonathan Evans, Ronald Corstanje
• Format: Article
• Language: English
• Published: IEEE 2021-01-01
• Series: IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
• Subjects: Land surface; moisture measurement; remote sensing; soil moisture
• Online Access: https://ieeexplore.ieee.org/document/9397254/

Synthetic aperture radar (SAR) is a remote sensing technique for mapping of soil moisture with high spatial resolution. <italic>C</italic>-band SAR can resolve features at field scale, or better, but responds to moisture only within the top 1 to 2&#x00A0;cm of the soil. When validating SAR-derived soil moisture products against standard <italic>in situ</italic> measurements at 5 to 10&#x00A0;cm depth, the greater moisture variability at the soil surface may be inaccurately categorized as measurement error. An alternative method was developed where the <italic>C</italic>-band SAR product is validated against soil moisture simulated at 2&#x00A0;cm depth by the HYDRUS-1D model. This reproduces soil moisture depth profiles from daily meteorological observations, leaf area index, and soil hydraulic parameters. The model was fitted at 13 COSMOS-UK sites so that the model output at 10&#x00A0;cm depth closely reproduced the cosmic ray neutron sensor data. At ten of the sites studied, there was an improvement of up to 8% in root-mean-squared difference by validating the Copernicus surface soil moisture (SSM) product at 2&#x00A0;cm compared to 10&#x00A0;cm. This suggests that Copernicus SSM and other <italic>C</italic>-band SAR surface soil moisture algorithms may be more accurate than have hitherto been acknowledged.