Comparative predictions of solute transport with hydraulic tomography in a laboratory sandbox aquifer: the importance of the injection data

Abstract

Over the past two decades, hydraulic tomography (HT) has been proven as a robust method for subsurface heterogeneity characterization with high resolution, which is critical for predicting solute transport. In this paper, HT was used to characterize the hydraulic conductivity (K) distribution within a laboratory sandbox aquifer, following established workflows from previous studies. The influence of estimated K fields on model calibration and validation performance was evaluated. Additionally, tracer injection data were considered as a key factor in the analytical framework. The simulated concentrations were compared to the observed data using inverse results derived from multiple modeling approaches, coupled with the classical advection–dispersion equation. The analysis yielded the following results: 1) geostatistical inversion provided better heterogeneity characterization compared to geology based zonation model, particularly in terms of hydraulic head data; 2) geostatistical inversion exhibited enhanced performance over the geology-based zonation model in predicting solute transport as evidenced by tracer concentration data, when injection data was incorporated into the inverse modeling framework; 3) breakthrough curve analysis revealed that solute transport predictions derived from HT still exhibited notable limitations, highlighting the need for further improvements; 4) overestimation issue identified in the HT results is linked to factors beyond observational error. Overall, this study highlights that the advantages of geostatistical inversion are obvious in heterogeneity characterization, and the involving of the tracer injection data is critical for improving solute transport prediction.