Non-Contact Prediction of Soil Moisture Profiles using Radio Wave Reflection

This work investigated the potential for the non-contact measurement of volumetric soil moisture profiles by detecting reflected VHF and UHF radio waves. The investigation included the development of an instrumentation system to measure radio wave reflection coefficients, a calibration routine for such a system, a model for simulating electromagnetic reflections from layered dielectric media, and an algorithm devised to resolve layers of moisture from radio wave reflections of multiple frequencies. Field trials of the system were conducted in which measured reflection coefficients were compared with simulated results, and predicted moisture gradients were compared with actual moisture profiles. The model that simulated reflection coefficients in the frequency range of 80 MHz to 1 GHz was tested using hypothetical and existent moisture profiles. Results of simulated profiles indicated that reflection coefficients could be used to distinguish between volumetric surface moisture and could detect subsurface moisture to a depth of 45.7 cm. Reflection measurements made in the field trials indicated that linear correlation could be made with volumetric moisture in the top 15.2 cm. The profile restoration algorithm closely predicted simulated near-surface moisture but had a high failure rate predicting deep subsurface moisture. Results of the study indicated that reflection coefficients could be used to detect soil moisture at depth, but the restoration algorithm did not effectively resolve moisture layers.