Modeling dense nonaqueous phase liquid mass removal in nonuniform formations: Linking source-zone architecture and system response

Abstract

Dense nonaqueous phase liquid (DNAPL) source zones comprise persistent sources of groundwater contamination that are recalcitrant to complete remediation using conventional (e.g., pump and treat) or emerging (e.g., surfactant flushing) technologies. Increased attention to the assessment of the benefits of partial mass removal from such contaminant source zones has intensified efforts to model multiphase flow and transport behavior. This paper describes the simulated recovery of a tetrachloroethene (PCE) spill in a statistically homogeneous but nonuniform aquifer, incorporating nonuniformity in both nonaqueous phase liquid saturation and pore velocities. We developed a ganglia-to-pool metric to quantify DNAPL source-zone architecture, and explored the correlation of this metric with dissolved mass flux behavior in response to partial DNAPL mass removal. Dissolution of 20%‐70% of PCE mass from models exhibiting low ganglia-to-pool ratios resulted in a larger predicted reduction of dissolved contaminant mass flux than models with high ganglia-to-pool ratios. Results of this study suggest that DNAPL source-zone characterization at field sites with homogeneous, nonuniform aquifers would benefit from inclusion of an estimate of the overall ganglia-to-pool ratio. Simulations demonstrate that flux reduction behavior depends on the source-zone architecture, which is not readily predictable using a priori assumptions about the spatial correlation of physical aquifer parameters. Model results further suggest that stochastic investigations of DNAPL source remediation at field sites should avoid reliance upon Leverett scaling of capillary entry pressures to permeability fields, which can artificially narrow the range of simulated behaviors.