[1] Partitioning tracer techniques, initially developed in the petroleum industry, are promoted as a way to determine entrapped dense nonaqueous phase liquid (DNAPL) mass in source zones. A suite of chemical tracers, some of which partition into the DNAPL phase, is injected into the source zone to estimate the entrapped saturation of contaminant. Current application of the technique can determine DNAPL at residual saturation but does not provide reliable mass estimates of DNAPL in high-saturation areas (e.g., pools) that may occur in some source zones. To evaluate this technique's limitations in characterizing source zones with complex entrapment architectures that may contain pools, we investigated estimation errors associated with well-characterized single-isolated DNAPL pools. X-ray attenuation was used to determine the spatial distribution of saturation within the pool under different total DNAPL mass entrapment while the DNAPL was subjected to tracer tests. Model analysis used showed that the partitioning tracer breakthrough simulations based on equilibrium partition coefficient (Kp, from equilibrium batch tests) fit the experimental data poorly, suggesting nonequilibrium partitioning that is controlled by the water flow velocity through the NAPL entrapment zone. In order to relax the requirement of partitioning to be at equilibrium, we investigated the use of ''effective partition coefficients'' (Kp e ). A comparison of the saturation estimates using method of moments and inverse modeling with both equilibrium and effective partition coefficients was conducted. Results demonstrated that the incorporation of nonequilibrium behavior in the analysis can improve DNAPL mass estimation by tracers in high-saturation zones.Citation: Moreno-Barbero, E., and T. H. Illangasekare (2006), Influence of dense nonaqueous phase liquid pool morphology on the performance of partitioning tracer tests: Evaluation of the equilibrium assumption, Water Resour. Res., 42, W04408,

The primary goals for most ground water capture systems (i.e., pump-and-treat systems) are that (1) all contaminants within zones of interest will eventually be captured and (2) the extraction and reinjection wells are best located and operated at optimal flow rates, creating hydraulically efficient flow systems. A new tool, MODular ALLocation (MODALL), is presented to aid in the design and assessment of capture systems. MODALL uses the MODFLOW-calculated cell-by-cell flow terms to evaluate internodal flow balances to determine the percentage of flow in each cell which has either originated from a given source(s) or flows to a specified sink(s). Output from MODALL can be easily displayed in isopleths of "capture fraction" (CF) to indicate the certainty or strength of capture in various areas. MODALL results are compared to the results from an analytical solution, a pathline analysis using MODPATH, and solute transport simulation with MT3DMS. A brief case study is also presented where MODALL is used to optimize an existing pump-and-treat system to more effectively and more efficiently contain a 5000-m long plume.

Characterization of dense non-aqueous phase liquid (DNAPL) in heterogeneous media represents a major challenge in the remediation process due to the complexity of DNAPL distribution in the subsurface. The partitioning interwell tracer test (PITT), which evaluates the relative transport of DNAPL-phase partitioning and conservative (i.e., non-partitioning) tracers, has recently been promoted as a way to quantify entrapped DNAPL mass in source zones. In some cases, the technique has been successfully applied to sites where DNAPL is present primarily at residual saturations. However, the effects of the geologic heterogeneity and DNAPL architecture on the performance and reliability of this technique have not yet been thoroughly examined. A systematic two-dimensional simulation study was conducted to evaluate the influence of DNAPL vertical distribution, well location, and geologic heterogeneity on the performance of the PITT for a total of 60 stochastic aquifer realizations. The aquifer hydraulic conductivity (K) distribution was defined using three different geostatistical parameter values describing the variance of the log K (r 2 ln k ¼ [1.22, 0.75, 0.25]). A multiphase flow simulator was used to generate the source zone, allowing the DNAPL to distribute naturally. The tracer test simulations conducted using these various synthetic aquifers were specifically designed to quantify PITT performance for different degrees of heterogeneity. For cases where the aquifer is relatively heterogeneous and the DNAPL architecture is complex, PITT estimation efficiency averaged 48 percent (decreasing to 20 percent in certain scenarios) as a result of bypass flow and the limited tracer accessibility to lower hydraulic conductivity regions. While other studies have suggested the PITT technique may underestimate DNAPL saturation for these conditions, potential errors of this magnitude have not previously been reported. The results of this work underscore the importance of geological heterogeneity on PITT performance.

Partitioning interwell tracer tests (PITTs) have been proposed as a tool to characterize source zones of waste sites with chemicals entrapped in the form of nonaqueous‐phase liquids (NAPLs). This technique has the potential to be applied to NAPL source characterization of both the vadose and saturated zones, but the recent applications have been primarily focused on dense NAPLs (DNAPLs) entrapped below the water table. This study presents an intermediate‐scale experiment to investigate the applicability of the technique to heterogeneous sites with complex NAPL entrapment architecture. Tracer experiments were conducted in an intermediate‐scale test tank (4.87 by 1.21 by 0.05 m) with heterogeneity defined using the geostatistical parameters of a spatially correlated random field. Tetrachloroethene (PCE), a DNAPL, was spilled into the heterogeneous packing and was allowed to distribute naturally, creating zones of both residual and pooled entrapment. The saturation distribution of the DNAPL was determined in situ using a gamma attenuation system. Once the saturation distribution was characterized, a set of tracer tests was conducted. Results of this study demonstrate how the inversion of the tracer breakthrough concentrations provided valuable information to establish the validity of the use of equilibrium partition coefficients in situations where tracer partitioning might be controlled by nonequilibrium behavior. While our previous work quantified the estimation error introduced by the assumption of equilibrium partitioning in tracer data analysis, this study demonstrated that more accurate NAPL estimates can be obtained using effective partition coefficients in combination with inverse modeling methods upscaled for more realistic heterogeneous settings.