FIELD-SCALE SOIL MOISTURE ESTIMATION UNDER CORN AND SOYBEAN CROPS USING AIRBORNE SAR DATA

The capability of high-resolution L-band Synthetic Aperture Radar (SAR) to retrieve field-scale (< 30 m) soil moisture has been investigated over corn and soybean crop fields. The time-series retrieval algorithm inverts the look-up-table (LUT) representation of physics-based forward scattering model under the assumption of temporally invariant roughness condition. In order to improve the retrieval accuracy of both surface roughness and soil moisture, an enhancement in sensitivity of forward scattering model has been implemented through a linear scaling of LUT. The retrieval algorithm has been applied to the time-series UAVSAR data acquired during AMPM campaign over northern Arkansas in USA. The unbiased RMSE between the forward modeled and observed backscattering coefficients (sigma(0)) are 1.34 dB (HH), 1.83 dB (VV) for corn and 3.77 dB (HH), 3.53 dB (VV) for soybean respectively. The validation of retrieved soil moisture using multi-pol (HH & VV) inputs with in situ measurements shows an unbiased RMSE (correlation) of 0.061 m(3)/m(3) (0.71) and 0.081 m(3)/m(3) (0.47) for corn and soybean crop fields respectively.