Geophysical imaging of near-surface structure using electromagnetic and seismic waves

Date of Completion

January 2008






This thesis includes three different studies of geophysical imaging: (1) inference of plume moments from tomograms with cross-hole radar; (2) simulated annealing inversion for near-surface shear-wave velocity structure with microtremor measurements; and (3) time-lapse GPR imaging of water movement in the vadose zone. Although these studies involve different geophysical approaches, they are linked by a common theme—using geophysical imaging to understand hydrologic phenomena or subsurface structure. ^ My first study in this thesis is concerned with the identification of plume moments from geophysical tomograms. Previously geophysical imaging has been applied to characterize contaminant plume migration in groundwater, and to determine plume mass, extent, velocity, and shape. Although tomograms have been used for quantitative inference of plume moments, the reliability of these inferred moments is poorly understood. In general, tomograms represent blurry and blunted images of subsurface properties, as a consequence of limited data acquisition geometry, measurement error, and the effects of regularization. In this thesis, I investigated the effect of tomographic resolution on the inference of plume moments from tomograms. I presented a new approach to quantify the resolution of inferred moments, drawing on concepts from conventional geophysical image appraisal, and also image reconstruction from orthogonal moments. This new approach is demonstrated by synthetic examples in radar tomography. My results indicated that moments calculated from tomograms are subject to substantial error and bias. For example, for many practical survey geometries, crosshole radar tomography (1) is incapable of resolving the lateral center of mass, and (2) severely underpredicts total mass. The degree of bias and error varies spatially over the tomogram, in a complicated manner, as a result of spatially variable resolution. These findings have important implications for the quantitative use of tomograms to interpret plume morphology. ^ In my second study I developed a passive-seismic method to image shear-wave velocity, which is an important geotechnical property commonly correlated with soil type or lithology. I inverted shear-wave velocity profiles from the phase velocity dispersion of Rayleigh waves based on passive seismic observations (microtremors). I used several sets of microtremor data which were collected at different sites. I obtained the phase velocity dispersion curve by the Extended Spatial Autocorrelation (ESPAC) method. I used simulated annealing method is used to invert the subsurface shear-wave velocity profile from the fundamental phase velocity dispersion curve. The field-experimental and synthetic results indicated that the microtremor approach can provide valuable information for quantitative geotechnical and hydrologic characterization. ^ In my third study I developed a method to image vadose-zone dynamics using GPR. Flow in the unsaturated zone is important for predicting groundwater recharge, contaminant migration, and chemical/microbiological processes. However, it is difficult to characterize or monitor with conventional hydrologic measurements, which provide information at sparse locations. The purpose of this study was to image changes in moisture content, as well as aquifer structure based on the relation between dielectric constant and water content. The objective was to calibrate a flow model to field-experimental, time-lapse GPR data collected during an infiltration experiment. To this end, (1) I constructed a VS2DT model based on aquifer structure interpreted from static GPR reflection profiles; (2) I manually calibrated the model to reproduce observed changes in GPR data during infiltration; and (3) I used a time-domain electromagnetic finite-difference model to simulate experimental observations for comparison. The results of this work indicate that time-lapse GPR can monitor changes in water content on the order of a few percent. ^ Keywords. geophysical imaging, hydrogeologic characterization, geometric moments, orthogonal moments, central mass, regularization, measurement errors, unsaturated zone, moisture content, water infiltration, VS2DT, GPR, EM forward model, simulated-annealing, shear wave velocity, microtremor, Rayleigh wave, Extended Spatial Autocorrelation (ESPAC) method. ^