NAPL (Non Aqueous Phase Liquid) source characterization based on dissolved contaminant concentrations

Abstract

Subsurface contamination frequently originates from spatially distributed sources of Non-Aqueous Phase Liquids (NAPLs). Such chemicals are typically persistent sources of groundwater contamination that are difficult to characterize. This work addresses the feasibility of using dissolved concentration measurements to estimate the spatial distribution of an immobile (or residual) NAPL mixture in a saturated field-scale system. We first developed a state estimation algorithm for a one-dimensional point source problem. (Bennett 1992, McLaughlin 1995). The spatially and temporally varying states (i.e., the dissolved contaminant concentrations and the NAPL saturations) were estimated by taking advantage of the time-dependence of solute concentrations originating from competitive dissolution of a NAPL mixture. We applied the estimator to a point source problem using sixteen measurement strategies with dissolved concentration measurements placed at different points in space and time. The estimation results assessed the uncertainty reduction associated with each strategy, highlighting the importance of capturing a characteristic scale of the problem with the sampling design. The characteristic scale for a given problem is either a temporal or spatial spread of measurements large enough to capture the temporal change in the NAPL source. At least for the sample problem, sampling strategies that utilized multiple measurements at the same location were as effective as those that sampled over a wider spatial domain. In the second part of this research we developed a two-dimensional coupled flow/dissolution/transport model for a spatially distributed residual NAPL source. We modeled the distributed NAPL as a first order gradient driven source term of the dissolved phase transport equation. Since this model was specifically designed to describe a residual NAPL source, it is a useful tool for analyzing the sensitivity of dissolved measurements under various source regimes. Based on the results of the one-dimensional estimator as well as the analyses we performed with the two-dimensional model, we concluded that dissolved concentration measurements are sufficiently sensitive to NAPL source composition and location to make source estimation feasible. However, this analysis assumed that both the prior statistics of the NAPL saturation and the field scale mass transfer coefficient for NAPL dissolution were known, an unlikely assumption in practical/applied estimation.