Soil moisture estimation using full-wave inversion of near- and far-field ground-penetrating radar data: A comparative evaluation

A new near-field ground-penetrating antenna model was applied for characterization of soil moisture in an agricultural transect in central Belgium. The measurement system consists of a vector network analyzer connected with a horn antenna and a differential GPS mounted on a motorcycle for quick data acquisition. Numerical experiments show that near-field GPR data are more sensitive to the soil dielectric properties than far-field data due to nearer distance between the antenna and medium. For the field measurements, the modeled GPR data from far-field and near-field configurations agree very well with the measured ones. However, soil water content estimated from near-field data is higher and more in agreement with Theta Probe measurements than far-field owning to the deeper penetration depth, smaller foot print and larger sensitivity with soil permittivity of near-field data. The results also show that the spatial pattern of the soil moisture is mainly controlled by the topography, while temporal variability is influenced by the rainfall intensity and time lag between rainfall event and experiment. The proposed approach shows a promising potential for temporal and spatial characterization of soil moisture at the field scale.