Inversion of Soil Moisture Profiles Under Tall Vegetation with Low-Frequency Radar: Theory and Experiment

Simultaneous estimation of deep and shallow soil moisture will provide a major breakthrough for estimation of key unobserved hydrologic fluxes, evapotranspiration and acquifer recharge, linking surface hydrologic processes with those in the subsurface. Given that more than 35 % of the global land surface is covered with substantial vegetation cover, exceeding 50 tons/ha, it is also imperative that the capability to estimate soil moisture under such amounts of vegetation is developed. We have previously presented the theoretical basis for estimating soil moisture under forest canopies using a combination of coherent and incoherent scattering models for radars in the UHF and VHF bands [1]. The present work focuses on sensitivity analyses and the application of the proposed method to experimental data. The results of this study directly impact retrieval of information from data obtained by low frequency radar systems on tower, airborne, and prospective spaceborne platforms. Representative of many natural subsurface structures, we assume the ground layers to be separated by rough interfaces of arbitrary depth and dielectric constant. Due to the large depth of penetration at VHF and to some extent at UHF, the waves can travel well into the ground surface and scatter from subsurface layers, even in the presence of dense forests. The wave scattering mechanisms, therefore, will include the interactions between the vegetation and subsurface layers in addition to the standard vegetation and top ground surface. Subsurface multiple scattering processes also need to be included. We have shown that the coherent scattering processes between the vegetation layers (crown and stems) and ground interfaces generally dominate the backscattere