Dense non-aqueous phase liquid (DNAPL) source zones can contribute to long-term groundwater contamination, thus remediation and management are of high importance. In spite of efforts towards understanding the fundamental processes affecting the fate of DNAPLs spilled or released in heterogeneous unconsolidated geologic materials, it is widely recognized that few, if any, sites contaminated by DNAPLs have been remediated with respect to either dissolved contaminants contained in the aqueous phase or removal of the DNAPL source. Further, there remains a paucity of knowledge on the behavior of DNAPLs spilled in fractured geologic media. The main objectives of this research were to: (1) develop computational tools for predicting aqueous-phase plume response to DNAPL source zone architecture and depletion for both porous and fractured geologic media; (2) conduct a suite of numerical experiments to investigate the relationship between DNAPL source-zone characteristics and dissolve-phase plume migration in porous and fractured media; (3) develop a stochastic information fusion (SIF) technology to define the DNAPL source and its characteristics by exploiting available hydraulic head and concentration data as well as signatures of stable isotope data of chlorinated solvents; (4) conduct laboratory experiments to validate the computational approaches; and (5) apply the technique at a well-characterized fractured rock site at Smithville, Ontario, Canada.

A data analysis environment has been developed through modification of an existing numerical model, CompFlow, to account for discrete fractures and stable isotope fractionation. Information that can be included in the data analysis environment include geologic information, well hydrographs, contaminant concentration data and isotopic signatures, and hydraulic property measurements.

The project has yielded robust, yet practical tools for predicting contaminant transport in porous and fractured geologic media. Key conclusions from the various project components are summarized below.

DNAPL-Involved Compound Specific Isotopic Analysis Modeling

A model that can simulate multi-phase and multi-component flow and transport with isotope fractionation was developed. The model is verified for DNAPL-aqueous phase equilibrium partitioning, aqueous phase multi-chain and multi-component reactive transport, and aqueous phase multi-component transport with isotope fractionation. Results from numerical simulations clearly indicate that the isotope signature can be significantly influenced by multiphase flow. They also illustrate that degradation and isotope enrichment compete with dissolution to determine the isotope signatures in the source zone: isotopic ratios remain the same as those of the source if dissolution dominates the reaction, while heavy isotopes are enriched in reactants along flow paths when degradation becomes dominant.

Modeling Flow in Fractured Media

The node bisection technique is effective in creating a mesh with fewer nodes than traditional discretization schemes.

Semi-Analytical Contaminant Transport Model Subject to Chain-Decay Reactions

A set of new, semi-analytical solutions to simulate three-dimensional contaminant transport subject to first-order chain-decay reactions and equilibrium sorption have been developed.

The analytical solutions can treat the transformation of contaminants into daughter products by first-order decay and the increasing concentrations of transformation species, leading to decay chains consisting of multiple contaminant species and various reaction pathways. The solutions in their current forms are capable of accounting for up to seven species and four decay levels and have been verified with a numerical model.

The ability of this model to consider decay chains consisting of multiple contaminant species, various reaction pathways, unique branching ratios, and retardation factors for different members makes it ideal for use in these screening studies.

Numerical Simulations of Source Mass Depletion in Fractured Porous Media

Trichloroethene (TCE) concentration and mass flux downstream and source depletion (dissolution) are strongly related. With increase in matrix permeability, NAPL TCE can migrate further vertically and aqueous TCE can transport further downstream.

It is concluded from the results that DNAPL source architecture and the partitioning of the source between the fracture and matrix domains are mainly functions of statistics of fracture network geometry and hydraulic characteristics of fracture/matrix; source depletion and the rate of DNAPL migration to downstream are closely related; and the downstream mass flux, however, is extremely difficult to lower under a certain level with a very small portion of remaining source in most fractured porous media.

Compound Specific Hydrogen Analysis

New method shows high accuracy and precision; quantification limit is as low as 400 ug/L.

Sorption Effects on Isotopic Fractionation

Results for adsorption and desorption experiments indicate higher fractionation of ³⁷Cl isotopes than ¹³C isotopes.

Biodegradation Effects on Isotopic Fractionation

cis-Dichloroethene (cis-DCE) production started to increase once TCE concentration decreased to about 120 mg/L. The type of bacteria responsible for degrading TCE to cis-DCE survived in the oxic environment, but the type of bacteria responsible for degrading cis-DCE to vinyl chloride (VC) and subsequently ethene was sensitive to oxygen and killed in the period when oxidation reduction potential (ORP) became positive. This is an ongoing project and the conclusions should be considered to be tentative.

Transient Hydraulic Tomography (THT) in Fractured Media

It is possible to delineate permeable fracture zones, their pattern and connectivity through the THT analysis of multiple pumping tests along with the inverse code successive linear estimator (SSLE). From the estimated hydraulic conductivity (K) and specific storage (SS) tomograms obtained from THT analysis of synthetic and laboratory data, it is evident that THT captured the fracture pattern quite well and they became more distinct with additional pumping tests. The results were validated using different methods. In particular, predicted drawdown from independent pumping tests captured observed behavior at later time while early time predicted drawdown deviated.

TCE Dissolution Modeling in Fractured Media

TCE field dissolution is expensive and time-consuming to model using the discrete fracture approach as it requires detailed deterministic and statistical information of the geometry of fractured zone and the spatial distribution of fracture apertures. This information is not typically available between boreholes.

On the other hand, the stochastic continuum approach could be comparatively less expensive and time consuming, as it does not require these detailed information about the spatial distribution of fractures.

TCE Attenuation Using Compound Specific Isotope Analysis

Along with redox and chemical data, the isotopic data from the Smithville site support the fact that biodegradation of TCE is occurring. In addition, numerical simulation studies suggest that the stability of the plume is due to first-order degradation. The dominant process is most likely reductive dechlorination of TCE. The further conversion of DCE to more degraded compounds is also supported by chemical and isotopic data.

DNAPL Simulations in Fractured Media

CompFlow simulations suggest that DNAPL penetration from the fracture into the matrix can take place in the carbonate units at the Smithville site. Imbibition is controlled by the capillary-saturation curves of the units. The penetration of DNAPL from the fracture into the matrix is different from the phenomenon of aqueous-phase contaminants diffusing from the DNAPL in the fracture into the matrix.

Substantial agreement with observed mass removal data and TCE plumes was achieved by modifying the composition of the DNAPL source and also by reducing the hydraulic conductivity in the source region of the Eramosa member.

Model results support earlier estimates that indicated that the pump-and-treat system has only recovered a small volume of TCE. It also suggests that the pump-and-treat system has been ineffective in controlling the plume and the stability of the plume is due to first-order degradation.

Application of multiphase compositional models (CompFlow) to realistic field-scale problems may be time-consuming and not currently feasible.

In summary, the CompFlow model has been modified to include the effects of isotope fractionation and discrete fractures. Numerical simulations provide important new insights on the utility of isotopes in revealing contaminant transport and reaction processes. Simulations have also revealed that DNAPL concentration and mass flux downstream and source depletion (dissolution) are strongly related. In addition, DNAPL source architecture and the partitioning of the source between the fracture and matrix domains are mainly functions of statistics of fracture network geometry and hydraulic characteristics of fracture/matrix. The downstream mass flux, however, is extremely difficult to lower under a certain level with a very small portion of remaining source in most fractured porous media. Laboratory adsorption and desorption experiments indicate higher fractionation of ³⁷Cl isotopes than ¹³C isotopes. Additional laboratory experiments with a rock block showed that Transient Hydraulic Tomography based on the SIF technology is a promising technology for mapping the spatial distributions of hydraulic conductivity, specific storage, and their uncertainty estimates. The field study at the Smithville site showed that isotopic and chemical data support the fact that biodegradation of TCE is occurring. In addition, numerical simulation studies of TCE plume transport suggest that the stability of the plume is due to first-order degradation. The dominant process is most likely reductive dechlorination of TCE. Model results support earlier estimates that indicated that the pump-and-treat system has only recovered a small volume of TCE. It also suggests that the pump-and-treat system has been ineffective in controlling the plume and the stability of the plume is due to first-order degradation.

The developed data analysis environment will improve our understanding of contaminant plume response to DNAPL source zone architecture and depletion in porous and fractured media. Thus, effective remediation strategies can be designed, which reduce the uncertainty and the cost of DNAPL source zone remediation.