Coupled Hydrogeophysical Inversion For Soil Hydraulic Property Estimation From Time Lapse Geophysical Data PDF Download

Vegetation and water on the landscape are directly linked, with leaf area controlling the partitioning of evapotranspiration, a process which creates a microclimate in the atmosphere above the plants. Therefore, widespread alterations to land use has potentially large implications for the water balance at regional and global scales. Unfortunately, many challenges persist that limit our ability to model with high confidence the biophysical constraints on evapotranspiration. One of the largest unknowns in this complex system is the root distribution, which is highly heterogeneous and dependent on both internal factors like the species and age of the plant, as well as external factors such as the climate and soil conditions. The physical limitations to studying properties below Earth's surface demands innovative approaches to improve our understanding of the interplay between roots, water, and the soil. Geophysical tools, such as electrical resistivity, have been employed for decades to study the properties of the Earth at scales from tens to hundreds of meters. Advancements in observing features such as oil and gas reserves, aquifer properties, and contaminant plumes has led the way to more recent work monitoring shallow soil properties such as water content and salinity. In this dissertation, I advance the field of hydrogeophysics with the development of a novel modeling approach that utilizes electrical resistivity data directly to parameterize root properties of site-specific hydrological models. Building on prior research using coupled hydrogeophysical inversion methods to estimate soil hydrological properties, the model presented and applied here addresses the pressing need for new tools to study root properties at the field scale. Chapter 1 provides a high-level introduction of the background and motivation for this area of research. Chapter 2 establishes the feasibility of the proposed modeling framework at a biofuel research site. Using site-specific soil and climate forcing data, I generated synthetic hydrological and electrical resistivity datasets using fixed soil and root parameters for a plot of maize. I then tested how well the model estimated those parameters under increasing levels of uncertainty. I found that even in the most data-poor scenario, the coupled hydrogeophysical inversion estimated the synthetic parameters with a high degree of accuracy. Chapter 3 proceeds to use the now-established model approach to estimate the root properties of two contrasting biofuel treatments, an annual rotation and a perennial grass. We again found the model returned reasonable estimates of the root distribution and evapotranspiration estimates for both crop types. In Chapter 4 I take advantage of the unique ability afforded by this modeling approach to test whether a theoretical coarse root fraction crossing the plane of the electrode array could produce an amplified resistivity measurement akin to what has been observed in the field. Given those estimates were within reason, subsequent estimates of coarse root fraction in a forested ecosystem were then validated against an index of above ground biomass. A statistically significant relationship was found, providing evidence in the absence of excavated root data that resistivity data can be used to non-invasively estimate the extent and relative quantity of coarse roots. Chapter 5 concludes this work by exploring the statistical relationship between above ground vegetation indices and the spatial and temporal heterogeneity in the observed resistivity data, providing the groundwork for future work modeling coarse root mass in a wide array of forest ecosystems.