Estimation of soil water content for engineering and agricultural applications using ground penetrating radar

Near-surface water content is important for a variety of applications in engineering, agriculture, ecology, and environmental monitoring and is an essential input parameter for hydrological and atmospheric models. Water content is both spatially and temporally variable and is difficult to characterize using conventional measurement techniques, which are invasive, time-consuming to collect, and provide only a limited number of point measurements. The purpose of this study is to investigate ground penetrating radar (GPR) techniques for improved estimation of water content. GPR techniques have potential for providing accurate, high-resolution estimates of water content quickly and non-invasively, but the efficacy of these techniques for field-scale applications has not been previously determined.; This study begins with a literature review of the application of GPR techniques for water content estimation, followed by a description of the principles employed in GPR surveying and the general methodology for converting electromagnetic GPR measurements to water content estimates. Next, a pilot experiment using GPR techniques for water content estimation is described; this experiment was performed under very controlled conditions and used common-offset GPR reflections to estimate the water content in sandy test pits. This experiment showed that GPR techniques can estimate water content very accurately (within 0.017 cm3/cm3 of the volumetric water content estimates obtained gravimetrically) and provided motivation for the second, less-controlled experiment. The second study used common-offset GPR reflections to estimate water content in a transportation engineering application, where the GPR data were used to monitor the water content in sub-asphalt aggregate layers and to estimate deformation under dynamic loading. This experiment showed that GPR data could be used to accurately monitor changes in the horizontal and vertical distributions of sub-asphalt water content with time, with an average error of 0.021 cm3/cm3 between the water content estimates derived from GPR and gravimetric measurements. The third experiment used multi-frequency GPR groundwave data to estimate the near-surface water content in the uncontrolled, heterogeneous environment of a California vineyard. Comparison of water content estimates from GPR data and from gravimetric measurements collected during this experiment showed that GPR data could accurately estimate water content in a heterogeneous environment (average error was less than 0.02 cm3/cm3) and could be used to guide precision vineyard management. Lastly, geostatistical analysis was performed on the high-resolution water content estimates collected in the vineyard, and the results of this analysis showed that the variability of near-surface water content is a function of measurement depth, season, and shallowly-rooted vegetation.