Contaminant tailing in highly heterogeneous porous formations: Sensitivity on model selection and material properties

Summary Coupled impacts of slow advection, diffusion and sorption were investigated using two heterogeneity models that differ in structure and in the mathematical framework that was used to simulate flow and transport and to quantify contaminant tailing. Both models were built using data from a highly heterogeneous exposure of the Borden Aquifer at a site located 2 km north-west of the Stanford–Waterloo experimental site at Canadian Forces Base Borden, Ontario, Canada. The inclusions-based model used a simplified representation of the different materials found at the site, while the second model was based on transitional probability geostatistics of the formation. These two models were used to investigate sensitivity of contaminant tailing on model selection and on geometric and material properties. While simulations were based on data collected at Borden, models were exercised beyond the geometric and material properties that characterize the site. Various realizations have identified very low conductive silty clay, found at volume fraction of 23.4%, as the material with dominant influence on tailing, and vertical diffusion in and out of low conductive units, affected by sorption, as the dominant transport mechanism causing tailing. The two models yielded almost identical transport results when vertical correlation lengths of silty clay were matched. Several practical implications relevant for characterization of low conductive units were identified and briefly discussed.