Wang, Chenxi2025-11-112025-11-112025-11-112025-10-30https://hdl.handle.net/10012/22622Hydraulic Tomography (HT) has been demonstrated to be a robust approach to characterize the subsurface heterogeneity, which employs the inversion of head data collected from multiple pumping tests to estimate the spatial distribution of hydraulic properties (e.g., hydraulic conductivity (K) and specific storage (Ss)). However, the resolution of K and Ss tomograms is reduced when the number of pumping tests and observation density gradually decrease. Another issue is that HT’s ability to predict solute/heat tracer transport behavior has not been rigorously examined for complex aquifer systems. Previous studies showed that different types of data (e.g., geological, geophysical, and other hydraulic testing data) carrying non-redundant information can be integrated with HT analysis to improve the mapping of K heterogeneity. This thesis evaluates the effectiveness of integrating geophysical logging data with HT analysis for improved imaging of K distributions. Furthermore, a heat tracer test was conducted in a highly heterogeneous glaciofluvial deposit to investigate the feasibility of reproducing the spatial distribution of observed temperature responses based on a heat transport model with HT K estimates. Five sequential studies are documented in this thesis to explore the integration of borehole geophysical logging with HT analysis for improved mapping of K and porosity heterogeneity, which enables enhanced predictions of groundwater flow and solute/heat tracer transport: (1) Study I integrated two conventional geophysical logging surveys, including electrical conductivity (EC) and gamma ray (GR) logging, with HT analysis to yield 2D K fields in a numerical sandbox experiment. A new spatial conditioning term was proposed to better delineate the hydrostratigraphy from geophysical logging data, which was used to derive the initial guess of K fields for geostatistical inverse modelling of HT analysis. The HT K models with comparative initial guesses of K distributions were evaluated based on their predictive capabilities for groundwater flow and solute transport. After demonstrating the effectiveness of integrating geophysical logging with HT analysis for improved K estimation in a numerical sandbox study, the subsequent studies were conducted at the North Campus Research Site (NCRS) underlain by a highly heterogeneous glaciofluvial deposit. (2) Study II conducted nuclear magnetic resonance (NMR) logging at the NCRS. Compared to conventional geophysical logging surveys, NMR logging can directly provide K estimates, as well as total porosity and effective porosity measurements. The petrophysical relationship between NMR signals and K was site-specifically optimized, and the NMR-derived downhole K profiles were compared with a variety of hydraulic measurements to evaluate their accuracy and resolution along boreholes. (3) Study III constructed 3D K fields based on downhole NMR K profiles and evaluated the representativeness of these K models. Various spatial interpolation approaches were employed to generate spatial K patterns. A multi-level heterogeneity characterization approach was proposed to better represent the layered porous medium at the NCRS. The model performance to predict groundwater flow was examined through simulating the observed drawdown responses from multiple pumping tests. (4) Study IV integrated NMR logging with HT analysis for improved characterization of subsurface heterogeneity. To highlight the importance of incorporating high-resolution initial K distributions to reduce the smoothness of K tomograms, a limited HT calibration dataset with fewer pumping tests and decreased observation density was utilized for model calibration. The effectiveness of this integration was evaluated through a comparative case study using varying numbers of head data for calibration and different spatial interpolation techniques for constructing initial NMR K models. (5) Study V conducted a heat tracer test at the NCRS, in which a dense monitoring network was installed to record temperature responses. The ability of various characterization approaches (e.g., HT analysis) to accurately map K heterogeneity was investigated by reproducing the complex spatial distribution of the temperature breakthrough curves (BTCs). Additionally, NMR-derived effective porosity was used to map a heterogeneous porosity field. Lastly, the sensitivities of heat tracer plume migration to flow, transport, and thermal parameters were investigated. The main contributions of these studies are: (1) conventional geophysical logging survey can provide hydrostratigraphic information to improve the resolution and accuracy of HT estimates; (2) NMR logging yields reliable downhole K estimates for interbedded layers of gravel, sand, silt and clay; (3) after spatial interpolation, 3D K models can be constructed based on NMR logging, which offers reasonable drawdown predictions to pumping tests; (4) integrating NMR logging with HT analysis can provide more representative K estimates consistent with the depositional environment, and the integrated models can still yield reliable K estimates at high resolution when only a limited head data is available for calibration; and (5) the complex temperature response from a heat tracer test can be best reproduced using HT analysis and NMR logging to represent the heterogeneous K and effective porosity fields. Based on their robust performance in predicting groundwater flow and solute/heat transport at different scales, this work advocates the joint use of HT analysis and borehole geophysical logging to characterize subsurface heterogeneity.enhydraulic tomographysubsurface heterogeneity characterizationgroundwater flowgeostatistical inverse modelingtracer transporthydraulic conductivityborehole geophysical loggingnuclear magnetic resonanceDoctoral Thesis