Flux-based assessment at a manufacturing site contaminated with trichloroethylene

https://doi.org/10.1016/j.jconhyd.2006.02.011Get rights and content

Abstract

Groundwater and contaminant fluxes were measured, using the passive flux meter (PFM) technique, in wells along a longitudinal transect passing approximately through the centerline of a trichloroethylene (TCE) plume at a former manufacturing plant located in the Midwestern US. Two distinct zones of hydraulic conductivity were identified from the measured groundwater fluxes; a 6-m-thick upper zone (∼7 m to 13 m below the ground surface or bgs) with a geometric mean Darcy flux (q0) of 2 cm/day, and a lower zone (∼13 m to 16.5 m bgs) with a q0  15 cm/day; this important hydrogeologic feature significantly impacts any remediation technology used at the site. The flux-averaged TCE concentrations estimated from the PFM results compared well with existing groundwater monitoring data. It was estimated that at least 800 kg of TCE was present in the source zone. The TCE mass discharge across the source control plane (85 m × 38 m) was used to estimate the “source strength” (∼365 g/day), while mass discharges across multiple down-gradient control planes were used to estimate the plume-averaged, TCE degradation rate constant (0.52 year 1). This is close to the rate estimated using the conventional centerline approach (0.78 year 1). The mass discharge approach provides a more robust and representative estimate than the centerline approach since the latter uses only data from wells along the plume centerline while the former uses all wells in the plume.

Introduction

It is estimated that about 15,000–25,000 sites in the US are contaminated with dense non-aqueous phase liquids (DNAPLs) and that the life cycle costs for cleaning up these sites can be as high as $100 billion dollars (EPA, 2003). Cleanup of these sites has been driven by state and regulatory statutes that require treatment for maximum beneficial use, as in the achievement of maximum concentration levels (MCLs) for drinking water throughout the site. However, it has now been recognized (Einarson and Mackay, 2001, API, 2002, API, 2003, ITRC, 2003, EPA, 2003, NRC, 2004) that though MCL is indeed the ultimate goal, it is not a pragmatic measure of success, especially if failure to attain MCL serves as the primary reason for not undertaking site remediation other than risk reduction through source or plume containment (e.g., physical or hydraulic barriers).

There has been a growing consensus among technical groups and regulatory agencies (API, 2002, API, 2003, ITRC, 2003, EPA, 2003, NRC, 2004) that contaminant flux, J (M/L2T), and contaminant mass discharge, MD (M/T), should be used as alternate performance metrics for site assessment and remediation design at contaminated sites. Contaminant flux (Eq. (1)) is defined as the total mass of contaminants passing per unit time per unit area of a control plane that encompasses the plume and is orthogonal to the mean groundwater flow direction.J=q0C=KiCwhere q0 is the Darcy groundwater flux (L3/L2T), K is the saturated hydraulic conductivity (L/T), i is the hydraulic gradient (dimensionless), and C is the resident contaminant concentration (M/L3).

Contaminant mass discharge, MD (M/T), is the spatial integration of the contaminant flux over the control plane (Eq. (2)):MD=AJdAwhere A is the area of the control plane and J (M/L2T) is the spatially variable contaminant flux defined by Eq. (1). Note that for heterogeneous conductivity field or DNAPL distribution at a given control plane, q0 and J are spatially and temporally variable, while MD varies only over time.

The objective of flux-based design of a remediation system is to reduce contaminant discharge MD at some down-gradient compliance control plane (CCP) to achieve appropriate risk reduction goals. This is different from the conventional, and more stringent, designs of site remediation that strive to achieve MCLs throughout the site. Contaminant discharge, not the groundwater concentrations, define the human-health and ecological risks associated with a site (Einarson and Mackay, 2001, API, 2002, API, 2003, ITRC, 2003, EPA, 2003). Thus, it is more realistic to design remediation systems based on contaminant mass discharge.

A flux-based site assessment and management (FB-SAM) approach involves measurement of MD across multiple control planes down-gradient of the suspected source zone. This information can be used for the estimation of (1) source strength, (2) receptor loading, and (3) degradation rates in the plume. Contaminant discharge across a control plane just down-gradient of the suspected source zone can be defined as the “source strength”, while discharge across a compliance control plane is the best estimate of receptor loading. Degradation rates, averaged over the plume, can be estimated from MD at multiple control planes using the following equation (Bockelmann et al., 2003):MD,x=MD,Sexp(kmxv)where MD,S and MD,x are the mass discharge rates (MT 1) at the source control plane and at a distance x (L) down-gradient from the source respectively, v (= q0/n, n is the media porosity) is the groundwater velocity (L/T), and km (T 1) is the first-order TCE degradation rate constant estimated by the mass discharge technique (assuming that contaminant degradation can be represented as a first-order process).

Design of a site remediation strategy involves defining a target receptor loading based on the remedial objective. For example, if the water at the receptor is to be used for drinking, the target receptor loading MD,r would be (Einarson and Mackay, 2001):MD,r=(MCL)×q0Awwhere Aw is a representative area that is a function of the capture zone of the supply well. The target loading can be achieved by (1) decreasing source strength through source treatment, or (2) increasing plume degradation rates through plume treatment, or (3) a combination of source and plume treatment.

Source strength can be reduced by removing contaminant mass from the source. It is very difficult to characterize the magnitude and distribution of the source mass, but it is easier to characterize the source strength. It has been shown that source mass can be related to source flux or source strength (Rao et al., 2001, Rao and Jawitz, 2003, Parker and Park, 2004, Jawitz et al., 2005, Falta et al., 2005a, Falta et al., 2005b, Enfield et al., 2005, Wood et al., 2005) using a power function (Eq. (5)) of the formMDMD,0=(MM0)Γwhere, MD and MD,0 are the present and initial source strengths, M and M0 are the present and initial source masses, and Γ is an empirical constant that is a function of the heterogeneity of the flow field and the NAPL ‘architecture’ (Rao et al., 2001, Jawitz et al., 2005, Falta et al., 2005a, Falta et al., 2005b).

Source treatment can be designed to remove source mass such that source strength can be reduced to a target value, the target being a function of allowable receptor loading and natural attenuation capacity of the plume. Natural attenuation capacity of a plume encompasses all destructive (e.g., biotic/abiotic destruction) and non-destructive (e.g., sorption, dispersion, dilution from recharge, volatilization) processes that attenuate contaminant concentrations (Wiedemeier et al., 1999). Source treatment might be adequate by itself (1) if the plume is stable or shrinking, because the mass discharge rate from the source is equal to or less than the natural attenuation capacity of the plume, and (2) the advective transport velocity is fast enough that the reduced source flux front reaches the receptor location within the remediation time frame. However, in an advancing plume (i.e., source strength is larger than the natural attenuation rate) source treatment should be coupled with plume treatment at the leading edge of the plume (Falta et al., 2005a, Falta et al., 2005b).

Traditional methods for estimating contaminant fluxes and discharge involve individual measurement/estimation of each of the three terms on the right hand side of Eq. (1) (API, 2002, API, 2003). Multi-level monitoring well data are used to estimate contaminant concentrations while potentiometric maps are used to estimate hydraulic gradients. Aquifer pump tests and/or slug test data are used for estimating hydraulic conductivity. Tracer tests and groundwater models could be used to refine these estimates. However, under spatially and temporally varying hydrologic conditions in complex, heterogeneous aquifers, significant uncertainty is propagated from such indirect estimation of contaminant fluxes (Hatfield et al., 2004). Such variability is compounded by the spatial and temporal variability of the source mass, and patterns of its depletion either during active remediation or during dissolution with groundwater flow under natural gradient conditions and natural biogeochemical attenuation processes.

Two groups of innovative approaches for simultaneous measurement of groundwater and contaminant fluxes have been developed over the past 5 years. These are the integral pump tests (IPTs), developed in Germany at the University of Tuebingen (Schwarz et al., 1998, Teutsch et al., 2000, Bockelmann et al., 2001, Bockelmann et al., 2003) and the passive flux meter (PFM) technique developed at the University of Florida (Hatfield et al., 2001, Hatfield et al., 2002, Hatfield et al., 2004, Annable et al., 2005). IPT involves pumping groundwater from the contaminant plume using one or more wells (generally located along transects), and using numerical and analytical models for estimating contaminant discharge from the data collected during the transient phase of the pump test. This method may be expensive since disposal costs of the large volume of waste fluids generated have to be added to the pumping costs. PFM, on the other hand, relies on natural groundwater flow through a permeable sorbent deployed in a monitoring well. Provided it is applied properly, IPT potentially provides a measure of the integrated contaminant mass discharge (Bockelmann et al., 2001, Bockelmann et al., 2003), while multiple PFMs along a given transect can be used to estimate the flux distribution (Hatfield et al., 2004). At the site examined in the present study, the PFM technique was used since (a) it was important to characterize the flux distribution and (b) waste disposal would be minimal.

Development of the PFM and pertinent design criteria are documented in recent papers (Hatfield et al., 2002, Hatfield et al., 2004, Annable et al., 2005), and are briefly described below. The PFM technique involves the deployment of a permeable unit, which is a matrix of hydrophobic (or hydrophilic) permeable sorbents, into a well or boring such that it intercepts groundwater flow but does not retain it. The sorbent is selected such that it retains dissolved organic and/or inorganic contaminants present in the intercepted groundwater. The sorbent matrix is also impregnated with known amounts of a suite of water-soluble ‘resident tracers’ that are desorbed at rates proportional to the groundwater flux through the device. After a specified period of exposure to groundwater flow, the PFM is removed from the well and the sorbent extracted to quantify the mass of all contaminants intercepted and the mass of all resident tracers remaining. The residual resident tracer masses are used to calculate the cumulative, time-averaged groundwater flux (L3/L2T), while the contaminant masses are used to calculate the cumulative, time-averaged contaminant mass fluxes (M/L2T). Depth variations of both water and contaminant fluxes can be measured in an aquifer from a single PFM by vertically segmenting the exposed sorbent, and analyzing for resident tracers and contaminants. Note that multiple tracers with a range of partitioning coefficients are used to accommodate for the variability in groundwater flux with depth. In addition, at a given depth, multiple tracers can be used for the estimation of groundwater fluxes and correspondence between the results would prove the robustness of the flux estimates. For the PFM approach, it is assumed that vertical hydraulic gradients are insignificant. Previous site monitoring data confirmed this assumption to be true (personal communications, CRA Inc. 2004).

Laboratory (Hatfield et al., 2001, Hatfield et al., 2004) and field experiments (Annable et al., 2005) have been performed to validate the theory and design prototypes of PFM devices. Recent field-scale testing has been conducted by the University of Florida group at DNAPL sites at the Canadian Forces Base Borden, Ontario, Canada; Hill Air Force Base, Utah; Ft Lewis, Washington; Patrick Air Force Base and Cape Canaveral Air Station in Florida. Documentations of these field tests are in various stages of peer review.

These field tests have been primarily done at government sites with new monitoring wells installed specifically for flux measurements, as a part of the efforts to assess the performance of source remediation. The objective of the present study is to extend the technique to ‘real world’ sites that have an existing monitoring well network for site assessment as a component of site management and regulatory compliance. The challenges in PFM implementation at a site with an existing well network include: (1) choosing an optimal set of wells from the existing network for PFM deployment; and (2) using the added information gained from PFM deployment in conjunction with archived groundwater monitoring data for designing a remediation system that is more cost-effective than could be done by using traditional site characterization data alone. An additional challenge in deploying PFMs at sites with existing and possibly old monitoring wells is that PFM being a technique that relies on passive flow of groundwater through the monitoring well, results are potentially affected by clogging of the well screen. Clogging of well screen impacts the measurement of groundwater and contaminant flux due to flow divergence around the monitoring well. However, the flux-averaged concentration is unaffected by clogging provided there is no microbial growth on the screen that degrades the contaminant in situ. The reason behind this is discussed in greater details in later sections.

In this study, the utility of site assessment based on contaminant flux and mass discharge was evaluated at a TCE-contaminated site, a former electronic-parts manufacturing plant located in the Midwestern US. The specific objectives of the research reported here are: (1) Field demonstration of the PFM technique for direct, in situ, simultaneous measurement of groundwater and contaminant fluxes; (2) Comparison of PFM results with traditional flux estimation techniques; (3) Source and plume characterization using PFM and groundwater monitoring information; and (4) Evaluation of site remediation alternatives.

Section snippets

Site description

The PFM was field-tested at a former electronic-parts manufacturing facility and foundry in the Midwestern US. The following description is based on unpublished site characterization reports prepared by an environmental consulting agency (personal communications, CRA Inc., 2004) and submitted to the state regulatory agency. The unconfined sandy aquifer beneath the site is contaminated with trichloroethylene (TCE) and its degradation products namely cis-1,2-dichloroethylene (DCE) and vinyl

Contaminant source and plume characterization

The vertical and aerial extent of the groundwater contamination was investigated in March 2002 (personal communications, CRA Inc. 2004) through collection of 113 groundwater samples from 9 soil borings and 49 monitoring wells, out of which 35 belonged to a nested well series. Nested well series included two or three wells installed at the same location but screened at different depths (shallow, intermediate, and deep). These concentration data were interpolated to obtain three-dimensional maps

Summary and conclusions

The TCE site investigated has an unconfined shallow aquifer contaminated with TCE and its degradation products. Existing site data were re-interpreted to formulate hypotheses about hydrologic and biogeochemical processes occurring at the site. TCE depletion through dissolution into groundwater flowing through the source zone is estimated to result in a contaminant mass discharge of 950 mg/m2/day, but diminished rapidly within the plume to 1.2 mg/m2/day. Reductive dechlorination appears to be the

Acknowledgements

This research was funded by the Showalter Trust grant through Purdue University. The authors are grateful to Greg Carli, Phillip Lambert, and their colleagues at CRA Inc. for providing logistical support at the site, and for providing access to numerous technical reports documenting previous investigations at this site.

References (33)

  • API

    Estimating Mass Flux for Decision-Making: An Expert Workshop

    (2002)
  • API

    Groundwater Remediation Strategies Tool

    (2003)
  • J.F. Devlin et al.

    Hydrogeologic assessment of in situ natural attenuation in a controlled field experiment

    Water Resour. Res.

    (2002)
  • M.D. Einarson et al.

    Predicting impacts of groundwater contamination

    Environ. Sci. Technol.

    (2001)
  • EPA
  • EPA
  • Cited by (0)

    View full text