Large-scale three-dimensional hydraulic tomography analyses of long-term municipal wellfield operations

•Field application of hydraulic tomography using existing municipal well records.•Uncertainties regarding initial and boundary conditions are accounted in analyses.•Geostatistical model is advocated for revealing intralayer heterogeneity patterns. Hydraulic tomography (HT) has been shown to be an effective approach for subsurface heterogeneity characterization. However, only a few HT studies have been performed for large-scale field problems due to the difficulty in conducting dedicated HT surveys at large-scale sites, as well as the uncertainty regarding estimated initial and boundary conditions for inverse modeling. To overcome these issues, this study advocates the utilization of existing long-term municipal well records as alternative datasets for large-scale heterogeneity characterization, along with novel data processing and analyses strategies that are proposed to minimize the effect of uncertain initial and boundary conditions on inverse modeling. Specifically, geology-based zonation and geostatistical models are adopted for site heterogeneity characterization through HT analyses of municipal well records, and the estimated hydraulic parameters from both models are validated through the prediction of head response data that have not been used for calibration efforts. Our results reveal that existing field municipal well records could be utilized for large-scale subsurface heterogeneity characterization using the approach of HT when uncertainties regarding initial and boundary conditions are well addressed for inverse modeling. In comparison to geology-based zonation model, the geostatistical model reveals greater details of intralayer heterogeneity where head response data are available, yielding significantly improved calibration and validation results. Overall, this study provides a general framework of using existing hydrographs for large-scale heterogeneity characterization through HT, and advocates the utilization of geostatistical inverse modeling as the second step over traditional zonation modeling approach to reveal intralayer heterogeneity details of hydraulic parameters.