Towards Estimation of Soil Moisture Using RF Polarimetric Responses with Topographical Data and Electromagnetic Scattering Models

High-resolution soil moisture measurements are of keen interest in a wide range of applications in atmospheric and hydrological sciences. This includes numerical weather modeling and seasonal climate predictions, flash flood forecasts, river flow forecasts and draught monitoring, agricultural productivity forecasts, water management, climate change research and fire potential prediction [1]. In particular, within the mesoscale research modeling communities, it is recognized that there is a lack of high resolution soil moisture estimates for data assimilation which poses substantial limitations to the reliability of operational model predictions [2]. This lack can be attributed to the scarcity of in-situ soil moisture measurements and to satellite-based observations at microwave frequencies with limited coverage and accuracy [3]. To address the need for field-scale soil moisture measurements, a new microwave remote sensing technology is being developed that can achieve field-scale sensing derived from microwave polarization mode dispersion responses, an approach that is known to exhibit sensitivities to soil moisture content [4]. However, the determination of soil moisture levels from these responses remains an area of active research. One approach is to employ site-specific calibration, an approach that has been proven in a laboratory setting, but that in practice is expected to involve measurements over a period of time to enable observations over a wide range of soil moisture and temperature conditions. In this work we consider a different approach based on modeling to estimate soil moisture levels that uses knowledge of the radio-frequency (RF) system topology, high-resolution terrain topography of the environment, and a characterization of the soil type. The approach involves comparing the measured polarization response with model estimates for the anticipated range of dielectric values to determine the most likely response. We employ 5m resolution topographic data and microwave measurements collected during the MATERHORN Fall 2012 campaign to investigate the approach applied to terrain near Sapphire and Granite mountains at Dugway Proving Ground in Utah. Although inaccuracies due to the coarse topographical data contribute to errors in the pursued approach, we demonstrate that trends in the approach suggest that this modeling methodology could provide a means of soil moisture estimation.