3
Performance of Conventional Pump-and-Treat Systems

Between 1982 and 1992, 73 percent of the cleanup agreements at Superfund sites where ground water is contaminated specified the use of pump-and-treat technology (Kelly, 1994). At most of these sites, the cleanup goal is to restore the aquifer so that the water extracted from it will be suitable for drinking without further treatment. Yet, within the past few years, studies of pump-and-treat systems have indicated that drinking water standards may be essentially impossible to achieve in a reasonable time frame at certain sites (Keely, 1989; Mackay and Cherry, 1989; EPA, 1989a; Mercer et al., 1990; Doty and Travis, 1991; Travis and Doty, 1990). This chapter presents the Committee on Ground Water Cleanup Alternatives' assessment of how well existing pump-and-treat systems have performed and whether it is reasonable to expect that they can achieve drinking water standards.

The analysis presented in this chapter is based on a review of 77 sites where pump-and-treat systems have been studied and the committee members' own extensive experience with ground water cleanup. Appendix A shows the sites the committee evaluated and summarizes the performance of pump-and-treat systems at each site. At 69 of the 77 sites, the pump-and-treat systems have not yet reached cleanup goals, as indicated in Appendix A. However, the committee also found eight sites where pump-and-treat systems have apparently achieved cleanup goals.

Throughout this chapter are brief case studies of sites where goals have been reached and those where they have not. Although the chapter highlights many success stories, the committee wishes to emphasize that



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 80
Alternatives for Ground Water Cleanup 3 Performance of Conventional Pump-and-Treat Systems Between 1982 and 1992, 73 percent of the cleanup agreements at Superfund sites where ground water is contaminated specified the use of pump-and-treat technology (Kelly, 1994). At most of these sites, the cleanup goal is to restore the aquifer so that the water extracted from it will be suitable for drinking without further treatment. Yet, within the past few years, studies of pump-and-treat systems have indicated that drinking water standards may be essentially impossible to achieve in a reasonable time frame at certain sites (Keely, 1989; Mackay and Cherry, 1989; EPA, 1989a; Mercer et al., 1990; Doty and Travis, 1991; Travis and Doty, 1990). This chapter presents the Committee on Ground Water Cleanup Alternatives' assessment of how well existing pump-and-treat systems have performed and whether it is reasonable to expect that they can achieve drinking water standards. The analysis presented in this chapter is based on a review of 77 sites where pump-and-treat systems have been studied and the committee members' own extensive experience with ground water cleanup. Appendix A shows the sites the committee evaluated and summarizes the performance of pump-and-treat systems at each site. At 69 of the 77 sites, the pump-and-treat systems have not yet reached cleanup goals, as indicated in Appendix A. However, the committee also found eight sites where pump-and-treat systems have apparently achieved cleanup goals. Throughout this chapter are brief case studies of sites where goals have been reached and those where they have not. Although the chapter highlights many success stories, the committee wishes to emphasize that

OCR for page 80
Alternatives for Ground Water Cleanup TABLE 3-1 Continuum of Possible Results for Pump-and-Treat Systems Result Explanation Unequivocal failure Fails to contain subsurface sources of contamination and to clean up the plume of dissolved contaminants emanating from source areas Prevention of additional exposure to contamination Contains subsurface sources of contamination and prevents the plume of dissolved contaminants from increasing in size Reduction of additional exposure and significant shrinkage of the area affected by the contamination Contains subsurface sources of contamination and possibly reduces the amount of contaminant mass in source areas; cleans up part or all of the plume of dissolved contaminants to healthbased standards Unequivocal success Fully removes sources of contamination and cleans up the plume of dissolved contaminants to health-based standards these successes are rare, as is evident in Appendix A. The committee also wishes to emphasize that whether a cleanup is labeled a success or a failure depends in part on the stringency of the cleanup goal. The success or failure of a cleanup should not be viewed as a simple ''yes'' or "no" but instead should be evaluated according to a continuum of possible results, from unequivocal failure to reduction in exposure to contaminants to unequivocal success, as shown in Table 3-1. HOW PUMP-AND-TREAT SYSTEMS WORK Conventional pump-and-treat systems are based on a theoretically very simple concept: contaminated ground water is extracted from the subsurface, and the extracted water is replaced with clean water. The clean water comes either from areas immediately adjacent to the contaminated zone or from water injected into the subsurface as part of the pump-and-treat process. (See Figure 1-1 in Chapter 1 for an example of a pump-and-treat system.) Occasionally, the extracted water is discharged directly into a surface water body, such as a stream. Direct discharge is acceptable where the surface water standards allow higher contaminant levels than do the ground water standards and where the contaminant concentration in the extracted ground water is low enough that surface water standards will not be exceeded. More often, however, the extracted water requires treat-

OCR for page 80
Alternatives for Ground Water Cleanup ment. The extracted water may be treated using any of a number of methods that have been well tested for removing contaminants from drinking water and industrial and municipal wastewater. For example, air strippers can remove volatile contaminants, granular activated carbon can remove dissolved organic contaminants, and biological systems can remove biodegradable contaminants. Once treated, the water may be discharged to a surface water body or reinjected underground. Reinjection can improve the system's efficiency and reduce cleanup time by speeding the flow of water and contaminants to the extraction wells. Pump-and-treat systems can be designed for two very different goals: containment, to prevent the contaminant from spreading, and restoration, to remove the contaminant. In pump-and-treat systems designed for containment, the extraction rate is generally established as the minimum rate sufficient to prevent enlargement of the contaminated zone. In pump-and-treat systems designed for restoration, the pumping rate is generally established to be much larger than that required for containment so that clean water will flush through the contaminated zone at an expedited rate. Because of their reduced pumping requirements, pump-and-treat systems designed for containment are much less costly to operate than pump-and-treat systems designed for restoration. In all other fundamental ways, the two types of systems are identical. However, pump-and-treat systems designed for restoration face a much greater technical challenge than those designed for containment. Even when these systems extract contaminated water and replace it with clean water, undissolved contaminants may remain underground. The remaining contaminants will dissolve slowly over time, making complete restoration of the ground water impossible until all of the contaminants can be removed PREVIOUS STUDIES OF PUMP-AND-TREAT SYSTEMS Before 1989, the limitations of pump-and-treat systems were not fully appreciated. No large-scale studies of the effectiveness of pump-and-treat systems were available because most of the systems were so new that their long-term performance could not be assessed. In 1989, however, the Environmental Protection Agency (EPA) released a study of pump-and-treat systems that caused concern in the regulatory community and among businesses paying for the cleanups. After a detailed review of 19 sites where pump-and-treat systems were operating, the EPA determined that at none of these sites had the aquifers been restored to drinking water standards (EPA, 1989a,b,c). In 1991, the EPA reassessed the data from these 19 sites and reviewed 5 additional sites. The agency found

OCR for page 80
Alternatives for Ground Water Cleanup Components of a pump-and-treat system at a former coal gas plant: pumped water is filtered through the activated carbon units shown here. Courtesy of the Johns Hopkins University, Department of Geography and Environmental Engineering. one site of the 24 where cleanup goals were apparently reached (EPA, 1992a,b). However, no follow-up monitoring was performed at the site to confirm the cleanup, and agency personnel have questioned the existing site data (Sutter and Glass, 1992). After the EPA studies, others conducted their own investigations. Researchers at the Oak Ridge National Laboratory reexamined data from 12 of the EPA sites and evaluated 4 additional sites (Doty and Travis, 1991). Like the EPA researchers, the Oak Ridge investigators concluded that pump-and-treat systems had not restored the aquifers to drinking water standards at any of the sites. More recently, the American Petroleum Institute (API) released a study of 13 sites not included in the EPA or Oak Ridge studies (API, 1993). The API's results were more promising: the study identified five sites, all gasoline stations, where pump-and-treat systems have reached cleanup goals. In a fourth study, researchers representing the California Regional Water Quality Board reviewed the records of 37 pump-and-treat systems at semiconductor manufacturing sites in California's Santa Clara Valley (Bartow and Davenport, 1992). Like the API study, this study yielded results somewhat more promising than previous studies: the researchers found two sites where pump-and-treat systems have reduced concentrations to below health-based standards for all of the contaminants; they identified an

OCR for page 80
Alternatives for Ground Water Cleanup additional eight sites where the pump-and-treat systems have reached health-based cleanup levels for some but not all contaminants. A common limitation of all the studies of pump-and-treat systems is that at most sites, systems have been operating for too short a time period to allow a final assessment of their effectiveness. The average starting year of all the systems in the EPA, Oak Ridge, API, and California Water Quality Control Board studies was 1985, which means these studies were based on only about five to seven years' worth of data, on average. A second problem with the studies is that for many of the early pump-and-treat systems, the designers did not fully appreciate the complexity of the subsurface and thus did not consider it in the system plans. For these systems, separating system success or failure from inadequate design is difficult. A third problem is that at many sites, surface sources of contamination such as heavily contaminated soils remain in place, raising questions about whether the inability to reach cleanup goals is due to continued leaching from these sources. As a consequence of these limitations, existing studies do not rule out the possibility that, given more time, optimal designs, and removal of surface sources of contamination, a larger number of pump-and-treat systems could reach cleanup goals. Amidst the uncertainty raised by the recent studies, some analysts have suggested that pumping and treating may be a wasted effort (Travis and Doty, 1990). Such critics question whether, given the poor record in meeting health-based cleanup goals, pump-and-treat systems are worth operating. These critics emphasize the enormous cost of pumping and treating large volumes of ground water over long time periods. On the other hand, others view the technology more favorably, contending that pump-and-treat systems can significantly reduce the risks of exposure to ground water contamination by removing contaminant mass and by containing the plume to keep it from points of water use, even if they cannot return all of the aquifer to near-pristine conditions. FEASIBILITY OF CLEANUP WITH PUMP-AND-TREAT SYSTEMS The effectiveness of pump-and-treat systems depends strongly on hydrogeologic and contaminant properties. As the complexity of the hydrogeologic conditions and the contaminants increases, the likelihood that the pump-and-treat system will meet stringent cleanup goals decreases. Table 3-2, developed by the committee, provides a framework for assessing the complexity of cleaning up contaminated ground water. In the table, the complexity of ground water cleanup increases with the complexity of contaminant chemistry, from left to right. The complexity

OCR for page 80
Alternatives for Ground Water Cleanup TABLE 3-2 Relative Ease of Cleaning Up of Contaminated Aquifers as a Function of Contaminant Chemistry and Hydrogeology   Contaminant Chemistry Hydrogeology Mobile, Dissolved (degrades/volatilizes) Mobile, Dissolved Strongly Sorbed, Dissolveda (degrades/volatilizes) Strongly Sorbed, Dissolveda Separate Phase LNAPL Separate Phase DNAPL Homogeneous, single layer 1b 1-2 2 2-3 2-3 3 Homogeneous, multiple layers 1 1-2 2 2-3 2-3 3 Heterogeneous, single layer 2 2 3 3 3 4 Heterogeneous, multiple layers 2 2 3 3 3 4 Fractured 3 3 3 3 4 4 a "Strongly sorbed" generally indicates contaminants for which the retardation coefficient is greater than 10. A retardation coefficient of 10 indicates that at any given time, 10 percent of the contaminant is dissolved in the water and 90 percent is sorbed to the aquifer solids. b Relative ease of cleanup, where 1 is easiest and 4 is most difficult.

OCR for page 80
Alternatives for Ground Water Cleanup of cleanup also increases with hydrogeologic complexity, from top to bottom. Conditions categorized as 1, shown in the upper left portion of the table, represent those that will be easiest to remediate. Conditions categorized as 4, shown in the lower right portion of the table, will pose the greatest technical challenge. Influence of Contaminant Chemistry and Site Geology As Table 3-2 shows, two types of contaminant characteristics can complicate ground water cleanup. The first characteristic is the tendency of the contaminant (organic or inorganic) to sorb to solid materials. As explained in detail in Chapter 2, chemical compounds dissolved in ground water interact with the solid media in the subsurface. As a result, at any given time, some of the chemical compound is dissolved in the ground water and some is attached to the solid media. Pump-and-treat systems can remove only dissolved contaminants. When a portion of the contaminant mass remains sorbed to solid media, it is possible that cleanup standards will not be met because the sorbed contaminants will desorb too slowly to be entirely removed but quickly enough to contaminate the clean ground water. The second contaminant characteristic complicating cleanup is the tendency for certain organic contaminants to remain undissolved as a nonaqueous phase. As explained in Chapter 2, these phases may be organic chemicals present as light nonaqueous-phase liquids (LNAPLs) that tend to float on the water table, such as gasoline, or as dense non-aqueous-phase liquids (DNAPLs) that tend to sink, such as chlorinated solvents. Contaminants dissolve slowly from these nonaqueous-phase liquids (NAPLs) into the passing ground water. As a result, it is likely that cleanup standards will not be met when NAPLs are present. As Table 3-2 shows, two types of contaminant characteristics can facilitate cleanup. As indicated in Chapter 2, certain chemicals degrade and/or volatilize. These processes may aid in cleanup when the processes occur naturally or when the remediation system takes advantage of them. For example, many LNAPLs are petroleum chemicals that degrade (when dissolved from the LNAPL) and/or volatilize. These processes, along with the fact that the LNAPLs usually rest above the water table, facilitate cleanup when a pump-and-treat system is combined with other technologies such as soil vapor extraction or bioventing (see Chapter 4). In addition to contaminant characteristics, Table 3-2 shows three types of geologic characteristics that can complicate cleanup: multiple layers, heterogeneities, and fractured rock. When such geologic features

OCR for page 80
Alternatives for Ground Water Cleanup are present, different regions of the contaminated zone will flush at different rates, with high-permeability zones cleaned up first. Attainment of cleanup standards will often be determined by how quickly the lower permeability zones flush. Furthermore, as explained in Chapter 2, heterogeneous regions with low permeability and regions with fractured rock can store significant quantities of contaminants that the bulk of the ground water cannot reach. Chemical transfer from these regions is slow and is controlled by diffusion. Regardless of the amount of contaminated water extracted, as long as significant diffusion occurs out of the low-permeability zones, it is possible that ground water cleanup standards will not be met. Geologic complexities and the presence of sorbed or nonaqueous-phase contaminants may affect the outcome of pumping and treating by causing progress toward cleanup to tail off above the cleanup goal. When the pump-and-treat system begins operation, contaminant concentrations may drop very rapidly, but after continued operation this rapid progress may cease at a level significantly above the cleanup goal. At such sites, there is no doubt that additional progress is still being made toward remediation, because the contaminant mass left in the aquifer is finite, and the pump-and-treat system continues to remove mass. In addition, the leveling effect often is observed only at some of the wells, near contaminant sources, while at other wells contaminant concentrations may continue decreasing. However, at wells where the concentration has leveled, continued progress toward reducing the concentration will be very slow, and the "final" stages of remediation may proceed for a very long time, as shown in Figure 3-1. Geologic and chemical complexities may also affect the result of pumping and treating by causing regrowth of the contaminant plume when pumps are turned off, even after the cleanup goal has been reached. Plume regrowth may occur when nonaqueous-phase contaminants that were not extracted with the pump-and-treat system dissolve in the clean water. It may also occur when contaminants in zones of low permeability that were not flushed with the pump-and-treat system diffuse into the clean water. Influence of the Quantity and Duration of Contamination In addition to the geologic and chemical characteristics depicted in Table 3-2, two other factors are very important in determining the difficulty of cleanup: (1) the mass of contaminant released and (2) the length of time the contaminant remained in the subsurface before cleanup. The easiest sites to remediate are those at which only a small mass of chemi-

OCR for page 80
Alternatives for Ground Water Cleanup FIGURE 3-1 The effect of tailing on cleanup time. The "theoretical removal" curve shows the number of aquifer volumes of ground water that must be pumped to remove the contamination, assuming all of it dissolves readily. The "removal with tailing" curve shows the number of aquifer volumes of ground water that must be pumped to remove the contamination when significant undissolved sources of contamination are present. Examples of such sources include contaminants present as pools of NAPLs and metals that have precipitated. SOURCE: Mercer et al., 1990. cal has been released in a small area and pumping and treating begin soon after the release. The length of time required for remediation generally increases with the amount of contaminant mass and the size of the source area. The size of the source influences cleanup time not only because larger quantities of contamination require more time to remove but also because the larger the source, the more difficult it is to identify and characterize the impact on ground water. The sites reviewed in Appendix A span a wide spectrum of source sizes and complexities. For example, at the Varian Associates site in Santa Clara, California, approximately 40 to 80 liters of 1,1,1-trichloroethane were lost in a one-time spill in 1984, whereas at the Aerojet site in Sacramento, California, potentially more than 4 million liters of chlorinated organic solvents were released at more than 100 source areas scattered about a 30-square-kilometer site starting in the early 1950s. For contaminants that resist degradation, the length of time for remediation increases with the length of time between the contaminant release and the start of remediation, because some processes that control ground water migration, such as diffusion, are time dependent. On the

OCR for page 80
Alternatives for Ground Water Cleanup Cleanup of crude oil from a burst pipeline near Bemidji, Minnesota. The equivalent of 8,000 barrels of oil was removed directly, but 2,500 barrels remained in the subsurface, forming a plume of contamination in the ground water. Courtesy of Hans-Olaf Pfannkuch, University of Minnesota. other hand, for contaminants that degrade to nontoxic products by chemical or biological processes, remediation may become easier as the time between the release and the start of remediation increases. Sites with ground water contamination range from coal gas generating facilities where releases occurred more than a century ago to service stations with ruptured underground storage tanks that are addressed within days of the release. Table 3-2 assumes that a "medium" amount of contaminant has resided in the subsurface for a "medium" length of time. CATEGORIZING SITES FOR CLEANUP Provided certain qualifications are kept in mind, the categories in Table 3-2 can provide a preliminary indication of the difficulty of clean-

OCR for page 80
Alternatives for Ground Water Cleanup ing up a particular site. The table does not encompass all of the factors that make ground water cleanup a complex task. As indicated above, it does not consider chemical mass released and duration of contamination. In addition, it does not consider cleanup goals, which influence whether the cleanup is perceived as a success or failure. Nevertheless, the categories in Table 3-2 are useful as subjective criteria for evaluating sites. Categorizing a site according to Table 3-2 requires information about the following site characteristics: site stratigraphy (i.e., a description of the geological layers), permeability of the layer(s), structural geology (especially information about fractures), types of chemicals in the subsurface, properties of chemicals in the subsurface, and estimates of the distribution of chemicals in the subsurface, including estimates of the potential or actual presence of LNAPLs or DNAPLs. Determining the appropriate row in Table 3-2 requires an assessment of whether the site hydrogeology is homogeneous, heterogeneous, or fractured, which can be determined by analyzing the first three types of information (stratigraphy, permeability, and structural geology). The ideal hydrogeologic environment for pump-and-treat systems is one that is a single layer with a hydraulic conductivity greater than about 10-5 cm/s (Mercer et al., 1990). Determining the appropriate column in Table 3-2 requires a judgment about the likely contaminant behavior in the subsurface, which requires the above types of information about the contaminants (types, properties, and distribution), as well as information about the composition of the solid media comprising the aquifer. For example, the retardation coefficient, which depends on properties of the contaminant and the solid media, indicates whether or not the contaminant will sorb strongly. In general, pump-and-treat systems are best suited to recover mobile chemicals that have retardation coefficients less than 10, which in general means that at any given time at least 10 percent of the contaminant in the plume is dissolved in the ground water (see Chapter 2). Relatively volatile organic compounds, indicated in the first and third columns of Table 3-2, have high vapor pressures and Henry's Law constants greater than 10-3 atm-m3/mole (EPA, 1990). Degradation, also shown in the first and third columns, is both chemical and site specific. As explained in Chapter 2, a wide variety of compounds—from gasoline and other fuels to chlorinated solvents—are potentially biodegradable, but whether they will degrade in the field depends on site conditions (especially on the presence of electron acceptors and other compounds necessary to support microbial activity).

OCR for page 80
Alternatives for Ground Water Cleanup TABLE 3-3 Pore Volumes Pumped per Year at 24 Sites     Plume Dimension Site Name Extraction Rate (liters/min) Area (ha) Thickness (meters) Pore Volumes (per year) Years to Pump 20 Pore Volumes Year Pumping Started Amphenol 980 3.6 30 1.55 13 1987 Black & Decker 38 4.5 12 0.12 164 1988 Des Moines 4,900 53 15 1.08 19 1987 DuPont Mobile 680 15 9 0.85 24 1985 Emerson Electric 110 1.2 15 1.08 19 1984 Fairchild San Jose 15,000 30 55 1.59 13 1982 General Mills 1,500 45 15 0.38 52 1985 GenRad 150 4.0 6 1.08 19 1987 Gilson Road 1,100 6.5 34 0.92 22 1981 Harris 1,200 24 27 0.31 65 1984 IBM Dayton 3,800 24 24 1.12 18 1978 IBM San Jose 23,000 310 76 0.17 118 1982 Lathrop 2,300 310 58 0.02 907 1982 Mid-South Wood Products 160 6.0 52 0.09 228 1985 Nichols Engineering 250 0.8 30 1.75 11 1988 Olin 24,000 97 24 1.74 12 1974 Ponders Corner 7,800 9.3 24 5.84 3 1984 Savannah River 2,100 420 46 0.02 1045 1985 Site A 140 0.3 6 14.59 1 1988 Tyson's Dump 450 26 110 0.03 735 1988 Utah Power and Light 760 3.6 53 0.68 29 1985 Verona Well Field 9,100 51 37 0.86 23 1984 Ville Mercier 2,800 3,100 24 0.01 3015 1983 Western Processing 830 5.7 20 1.30 15 1988

OCR for page 80
Alternatives for Ground Water Cleanup operation of pump-and-treat systems should be viewed as a long-term project, in which the system's design is modified in response to improved understanding of the site. In effect, operation of the pump-and-treat system should become part of a continued site characterization process. Continued monitoring of the system is essential to determine the performance of the original design and to predict and subsequently assess the improvement in performance that might result from modifications made during operation. What To Monitor The parameters to be monitored and the necessary frequency of monitoring vary from one situation to the next, but generally the minimum requirements for both are established by the lead regulatory agency. Typically, the following types of monitoring data are necessary to track performance of the cleanup system: water levels or piezometric heads at numerous sampling points throughout and around the contaminated zone to allow estimation of water flow directions and the portion of the aquifer that the extraction system is controlling; contaminant concentrations in ground water at numerous sampling points throughout and around the contaminated zone to allow estimation of the areal and vertical extent of contamination and the remaining dissolved contaminant mass; contaminant concentrations in the extracted ground water to assess progress toward the cleanup goal and to estimate the cumulative mass of contaminants removed from the aquifer; contaminant concentrations in the treatment system effluent to assess performance of the treatment system and compliance with discharge requirements; flow rates from the extraction wells and through the treatment system to confirm that the system is operating to specifications; and other operational parameters, such as line pressures, that indicate proper operation or incipient failure of pumps and filters or rising water levels in injection wells that may signal clogging. Given, as described throughout this report, that contaminated sites often have lingering subsurface sources of contamination, it would be advantageous to monitor the decrease or change in distribution of contaminant mass within source zones. Unfortunately, the tools currently available for source monitoring have not proven to meet the need or, in some cases, have been realized as potentially worsening the contamination problem. There is a great need for reliable, accurate techniques for

OCR for page 80
Alternatives for Ground Water Cleanup source mass monitoring, and several promising techniques are now under research and development. Modifying Existing Pump-and-Treat Systems Monitoring data may be used to assess performance of a pump-and-treat system by determining progress toward six general end points, which will vary with the site: elimination of contaminant migration beyond the extraction system, decrease in the size of the contaminated area, decrease in the contaminant concentration in the extracted ground water, decrease in the contaminant concentration remaining within the aquifer, increase in the cumulative mass of contaminants extracted from the subsurface, and minimization of the volume of water extracted for containment (and therefore the costs of treatment and disposal). If progress toward any of these objectives does not meet expectations, modification of the system's design or operation should be considered (or may be required). This iterative process will lead to a remedial program that changes over time, with the twin purposes of meeting remedial objectives and minimizing costs. For example, if the initial design does not meet the first goal, plume capture, the system may be modified by installing additional wells and/or by increasing the pumping rates in existing wells. The optimal modification should be predictable given the additional insight gained from head or water level data collected during initial system operation. Subsequent monitoring should show whether the modifications were adequate or whether the system will require further tuning. If progress toward goals 2 through 5, which indicate successful contaminant removal, is considerably slower than expected, then one or more assumptions used in the original system design are incorrect. Such disappointing progress may result because of the unanticipated discovery of the types of geologic and chemical complications discussed earlier in this chapter and shown in Table 3-2. Monitoring may provide insight about the most important of these complications. For example, researchers conducting field tests at the Rocky Mountain Arsenal cleanup site learned from multilevel monitoring that variations in hydraulic conductivity within the aquifer were an important cause of the unanticipated tailing of contaminant concentrations above cleanup goals after pumping and treating (Mackay and Thorbjarnarson, 1990). The consultants work-

OCR for page 80
Alternatives for Ground Water Cleanup Drilling rig used to install wells at a contaminated site. Courtesy of Rice University, Department of Environmental Science and Engineering. ing at this site had assumed, in essence, that the site belonged in category 2 according to Table 3-2 when in fact it was more appropriately categorized as 3 due to heterogeneities. Where monitoring reveals especially complex conditions (as in categories 3 and 4 in Table 3-2) and the tailing of contaminant concentrations at an asymptote, there may be a need to adjust the system design and to reevaluate the remedial objective and projected cleanup time. In such cases, as discussed earlier in this chapter, the most realistic remedial objective might be plume capture. Continued monitoring will be necessary

OCR for page 80
Alternatives for Ground Water Cleanup to confirm that the plume is completely contained and to determine whether changes in extraction rates and/or locations would conserve money and/or water. In the ideal case that progress in cleaning up the dissolved plume is initially relatively rapid or is enabled by source isolation, particularly for sites in category 1 or 2 in Table 3-2, subsequent monitoring is likely to reveal that portions of the aquifer have been cleaned adequately as remediation proceeds. In such cases, one or more of the extraction wells may no longer be removing significant amounts of contaminants. As monitoring confirms this situation, some of the wells may be shut down, or the extraction rate in some or all of the wells may be adjusted (some decreased and perhaps some increased, depending on the interactions between the wells). The goal of these adjustments is to minimize remaining costs while continuing the progress toward achieving cleanup goals. If the remediation has apparently reached the original cleanup goals or a relaxed cleanup criterion has been agreed upon, active remediation may cease. However, as illustrated graphically by the Dayton, New Jersey, case discussed in Box 3-3, monitoring should continue well beyond the time of active remediation. The goal is to ensure that residual contamination is not sufficient to cause the reestablishment of a contaminant plume. If monitoring reveals significant contamination, then new alternatives for remediation or isolation of the contamination will have to be evaluated. RESEARCH NEEDS FOR IMPROVING THE PERFORMANCE OF PUMP-AND TREAT SYSTEMS Much of the research needed to design better pump-and-treat systems exists (although this research has not all been transferred to practitioners in the field). Current research focuses on new technologies to couple with pump-and-treat systems and is discussed in Chapter 4. The key problems requiring further research to improve the performance of conventional pump-and-treat systems are all related to developing better methods for site characterization. Especially important is research to address the following questions: How can NAPLs, especially DNAPLs, be better characterized in the subsurface? How can partitioning of chemicals between the aqueous phase and NAPL and sorbed phases be more accurately quantified? How can this information be used to more accurately estimate cleanup times?

OCR for page 80
Alternatives for Ground Water Cleanup CONCLUSIONS Based on a review of the case studies listed in Appendix A and the experience of committee members, the committee reached the following conclusions about the performance of pump-and-treat systems: At some sites with simple geology and dissolved contaminants, pump-and-treat systems appear to be capable of cleaning up ground water to health-based standards in a relatively short time. Such simple sites are the exception rather than the rule. Nevertheless, the committee found examples of sites where pump-and-treat systems achieved health-based cleanup goals for one or more contaminants. At such sites, it is important to recognize that continued monitoring is necessary to verify the long-term effectiveness of the cleanup. Contaminants may remain attached to solid materials or stored in nonaqueous phases in the subsurface even when ground water from monitoring wells meets regulatory standards. At many of the sites where pump-and-treat systems have attained cleanup goals, the contaminants of concern are readily biodegradable . Pump-and-treat systems have achieved health-based goals at sites contaminated with gasoline, sites where contaminants are fully dissolved, and sites with both dissolved and LNAPL plumes (where the source has been removed) in shallow aquifers. The success of pump-and-treat systems at these sites may in part be due to biodegradation processes that convert contaminants to nontoxic products. The chemical nature of contaminants can prevent pump-and-treat systems from restoring aquifers to health-based standards in a relatively short time. Pump-and-treat systems cannot restore aquifers except over very long time periods (hundreds or thousands of years) where NAPL contaminants remain unless the NAPLs are contained or removed. For contaminants that strongly sorb to solid materials in the subsurface, cleanup times using pump-and-treat systems may also be very long. The geologic conditions of the site can prevent pump-and-treat systems from restoring aquifers to health-based standards in a relatively short time. Clay lenses and other heterogeneities, fractured bedrock, and zones of low hydraulic conductivity can trap contaminants and prevent the large-scale water circulation necessary for effective flushing of the subsurface. At sites where complete aquifer restoration to health-based standards is impossible or impractical due to the chemical nature of the contaminants or geologic complexity, pump-and-treat systems can prevent the contamination from spreading and can clean up or shrink the

OCR for page 80
Alternatives for Ground Water Cleanup dissolved portion of the contaminant plume. Pump-and-treat systems can prevent contaminant migration by establishing a hydraulic barrier around the site. They can shrink the contaminant plume by pumping out contaminated water, which is then replaced at the outer edges of the plume with clean water. These accomplishments reduce the risk posed by the contamination by minimizing the area affected by the contamination. Cleanup times for pump-and-treat systems vary widely depending on site conditions and pumping rates. For small sites with simple geology and dissolved contaminants, cleanup times may be relatively short, on the order of years. The presence of geologic heterogeneities, nonaqueous-phase contaminants, sorbed contaminants, and contaminant sources above the water table can extend cleanup times by anywhere from a few years to thousands of years and can make predicting the time highly uncertain. Because cleanup time also depends on the pumping rate (which system operators can control), evaluations of cleanup time should consider the number of pore volumes of ground water that must be extracted to achieve cleanup, in addition to the estimated cleanup time. The operation of pump-and-treat systems should be viewed as a long-term project, in which the system's design is modified in response to improved understanding of the site. Because of the complexity of the contaminated subsurface, the performance of a pump-and-treat system will always be uncertain until the system is tested by beginning the cleanup. Monitoring provides the information necessary to optimize the system's performance and ultimately determine whether it will be able to reach cleanup goals. NOTES 1.   See 51 Fed. Reg. 21,653, June 13, 1986; 51 Fed. Reg. 27,062, July 29, 1986; 51 Fed. Reg. 41,088, November 13, 1986. 2.   To prepare Figure 3-3, the committee assumed that the areal extent of the clay lenses is such that contaminant flow out of lenses is essentially one-dimensional and can be described by Fick's law: where .Jx is the contaminant flux from the clay lens, De is the porous media molecular diffusion coefficient, and C is the concentration of the contaminant in the clay lens. (The porous media diffusion coefficient is a function of the water molecular diffusion coefficient for the contaminant corrected for the porosity and the tortuosity of the day lens.) The committee further assumed that the relative concentration of the trichloroethene in the clay lens is one (unit concentration) as a result of the long period of contamination, while the trichloroethene concentration in the sands is zero because the pump-and-treat system has

OCR for page 80
Alternatives for Ground Water Cleanup     removed the contaminant from the water stored in the sand portion of the aquifer. Using these initial and boundary conditions, the committee followed a solution method for Fick's law similar to that described by Carslaw and Jaeger (1959) for heat flow from a solid bounded by parallel planes. The committee assumed that the porosity of the clay lens is 0.35, the retardation factor for trichloroethene in the clay lens is 2, and the water molecular diffusion coefficient for trichloroethene is 8.3 × 10-6 cm2/s, which results in a porous media molecular diffusion coefficient of 1 × 10-6 cm2/s. 3.   For this example, the committee based its computations on a 1 m3 volume of aquifer and used the following assumptions: (1) a porosity of 0.3, (2) a ground water flow rate of 0.03 meters per day (typical of conditions in fine-grained sands), (3) a dissolved trichloroethene concentration equal to 10 percent of the aqueous solubility of this compound (a value consistent with field observations but less than that indicated by theoretical calculations) (Hunt et al., 1988), and (4) a random distribution of the trichloroethene globules. The density of trichloroethene is 1.47 g/cm3, and its water solubility is about 1,100 mg/liter. With these assumptions and trichloroethene properties, the following calculations show the time required to dissolve the trichloroethene globules: Total contaminant mass = 30 liter/m3 × 1 m3 × 1.47 g/cm3 × (100 cm/m)3 × 10-3 m3/liter   = 44,100 g Concentration of dissolved trichloroetherne = 10% × 1,100 mg/liter   = 110 rag/liter Mass flux through 1-m2 area = 0.03 meter/day × 1 m2 × 110 mg/liter × 10-3 g/mg × 103 liter/m3 × 0.3   = 0.99 g/day Time required to dissolve residual trichloroethene = 44,100 g/(0.99 g/day)   = 122 years 4.   The rate at which the trichloroethene dissolves from the pool was estimated from the following equation, which is based on Hunt et al. (1988, 1989): where T is the thickness of DNAPL removed per unit time; Vx is the ground water velocity; Y is the aquifer thickness; Cs is the water solubility of the DNAPL; L is the length of the DNAPL

OCR for page 80
Alternatives for Ground Water Cleanup     pool in the direction of ground water flow; erfc is the complementary error function; ρ is the DNAPL density; n is porosity; αt is the transverse dispersivity (which describes dispersion normal to the ground water flow direction); and De is the porous media molecular diffusion coefficient. For a given DNAPL pool geometry, the rate of dissolution in most aquifers is a function primarily of the ground water velocity and the transverse dispersion length. For this example, the committee assumed that the aquifer has a porosity of 0.35, a thickness of 5 meters, and a bulk molecular diffusion coefficient of 1 × 10-6 cm2/s. REFERENCES API (American Petroleum Institute). 1993. Pump and Treat: The Petroleum Industry Perspective. Washington, D.C.: API. Bartow, G. W., and C. W. Davenport. 1992. A review of ground water remediation in Santa Clara Valley, California . Poster Presentation at National Ground Water Association Conference on Aquifer Restoration: Pump-and-Treat and the Alternatives, Las Vegas, September 30-October 2, 1992. Carslaw, H. S., and J. C. Jaeger. 1959. Conduction of Heat in Solids, 2nd ed. Oxford:Clarendon Press. Doty, C. B., and C. C. Travis. 1991. The Effectiveness of Groundwater Pumping as a Restoration Technology. Knoxville: University of Tennessee, Waste Management Research and Education Institute. EPA (Environmental Protection Agency). 1985. Modeling Remedial Actions at Uncontrolled Hazardous Waste Sites. EPA/540/2-85/001. Cincinnati, Ohio: EPA, Risk Reduction Engineering Laboratory. EPA. 1988. Guidance on Remedial Actions for Contaminated Ground Water at Superfund Sites. EPA/540/G-88/003. Directive 9283.1-2. Washington, D.C.: EPA, Office of Solid Waste and Emergency Response. EPA. 1989a. Evaluation of Ground-Water Extraction Remedies, Volume 1: Summary Report. EPA/540/2-89/054a. Washington, D.C.: EPA. EPA. 1989b. Evaluation of Ground-Water Extraction Remedies, Volume 2: Case Studies. EPA/540/2-89/054b. Washington, D.C.: EPA. EPA. 1989c. Evaluation of Ground-Water Extraction Remedies, Volume 3: General Site Data, Data Base Reports. EPA/540/2-89/054c. Washington, D.C.: EPA. EPA. 1990. Handbook on In Situ Treatment of Hazardous Waste-Contaminated Soils. EPA/540/2-90/002. Cincinnati, Ohio: EPA, Risk Reduction Engineering Laboratory. EPA. 1992a. Evaluation of Ground-Water Extraction Remedies: Phase II, Volume 1—Summary Report. Publication 9355.4-05. Washington, D.C.: EPA, Office of Emergency and Remedial Response. EPA. 1992b. Evaluation of Ground-Water Extraction Remedies: Phase II, Volume 2—Case Studies and Updates. Publication 9355.4-05A. Washington, D.C.: EPA, Office of Emergency and Remedial Response. EPA. 1993. Guidance for Evaluating the Technical Impracticability of Ground-Water Restoration. Directive 9234.2-25. Washington, D.C.: EPA, Office of Solid Waste and Emergency, Response. Harman, J., D. Mackay, and J. Cherry. 1993. Final Report to U.S. Air Force. Waterloo, Ontario, Canada: Waterloo Centre for Ground Water Research. Hunt, J. R., N. Sitar, and K. S. Udell. 1988. Nonaqueous phase liquid transport and cleanup 1: analysis of mechanisms. Water Resources Res. 24(8):1247-1258. Hunt, J. R., N. Sitar, and K. S. Udell. 1989. Correction to nonaqueous phase liquid transport and cleanup 1: analysis of mechanisms . Water Resource Res. 25(6):1450.

OCR for page 80
Alternatives for Ground Water Cleanup International Technology Corporation. 1989. Remedial Investigation: Former Firestone Facility. Salinas, Calif.: John Steinbeck Library. Jackson, R. E., S. Lesage, M. W. Priddle, A. S. Crowe, and S. Shikaze. 1991. Contaminant Hydrogeology of Toxic Organic Chemicals at a Disposal Site, Gloucester, Ontario. Scientific Series No. 181. Burlington, Ontario: Environment Canada. Keely, J. F. 1989. Performance Evaluations of Pump-and-Treat Remediations. EPA/540/4-89/005. Ada, Okla.: EPA, R. S. Kerr Environmental Research Laboratory. Kelly, M. M. 1994. Applying innovative technologies to site contamination: historical trends and future demand. Presented at HazMat South '94, Orlando, Fla., February 16, 1994. Mackay, D. M., and J. A. Cherry. 1989. Groundwater contamination: pump-and-treat remediation. Environ. Sci. Technol. 23(6):630-636. Mackay, D. M., and K. W. Thorbjarnarson. 1990. Flushing of Organic Contaminants from a Ground Water Plume at the Rocky Mountain Arsenal: Field and Laboratory Studies. Environmental Science and Engineering Program Technical Report No. 90-69, Vol. I (report) and Vol. II (appendices). Los Angeles: UCLA School of Public Health. Martel, R. 1988. Groundwater contamination by organic compounds in Ville Merrier: new developments. Presented at NATO Committee on Challenges of Modem Society Second International Conference: Demonstration of Remedial Action Technologies for Contaminated Land and Groundwater, Bilthoven, the Netherlands, November 7-11, 1988. Mercer, I. W., D.C. Skipp, and D. Giffin. 1990. Basics of Pump-and-Treat Ground-Water Remediation Technology. EPA/600/8-90/003. Ada, Okla.: EPA, R. S. Kerr Environmental Research Laboratory. Mercier Remediation Panel. 1993. Evaluation of long-term remedial measures for the subsurface contamination associated with the former Merrier Lagoons. Preliminary draft report submitted to Laidlaw Environmental Services (Mercier), Ltd. National Research Council. 1990. Ground Water Models: Scientific and Regulatory Applications. Washington, D.C.:National Academy Press. National Research Council. 1993. In Situ Bioremediation: When Does It Work? Washington, D.C.:National Academy Press. Pakdel, H., G. Couture, C. Roy, A. Masson, P. Gelinas, and S. Lesage. 1990. Method development for the analysis of toxic chemicals in soil and groundwater: the case of Ville Mercier, P.Q. In Ground Water Quality and Analysis at Hazardous Waste Sites, S. Lesage and R. E. Jackson, eds. New York: Marcel Dekker. Robertson, C. 1992. Performance of groundwater extraction systems at the IBM-Dayton site in Dayton, New Jersey. Presentation before the National Research Council's Committee on Ground Water Cleanup Alternatives, Washington, D.C., March 24, 1992. Smedes, H. W., N. Spycher, and R. L. Allen. 1993. Case history of one of the few successful Superfund remediation sites: a site at Salinas, California, USA. Engineering Geology 34:189-203. Stephanatos, B. N., K. Water, A. Funk, and A. MacGregory. 1991. Pitfalls associated with the assumption of a constant partition coefficient in modeling sorbing solute transport through the subsurface. Pp. 13-20 in Proceedings of the International Symposium on Ground Water, American Society of Civil Engineering (ASCE), Nashville, Tenn., July 29-August 2, 1991. New York: ASCE. Stipp, D. 1991. Throwing good money at bad water yields scant improvement. Wall Street Journal. May 15, 1991. A1. Sutter, J. L., and J. Glass. 1992. The EPA's 24-site study of pump-and-treat systems. Presentation to the Committee on Ground Water Cleanup Alternatives, National Research Council, Washington, D.C., March 24, 1992.

OCR for page 80
Alternatives for Ground Water Cleanup Travis, C. C., and C. B. Doty. 1990. Can contaminated aquifers at Superfund sites be remediated? Environ. Sci. Technol. 24:1464-1466. Zheng, C., G. D. Bennett, and C. B. Andrews. 1991. Analysis of ground-water remedial alternatives at a Superfund site. Ground Water 29(6):838-848. Zheng, C., G. D. Bennett, and C. B. Andrews. 1992. Reply to the preceding discussion by Robert D. McCaleb of ''Analysis of Ground-Water Remedial Alternatives at a Super-fund Site.'' Ground Water 30(3):440-442.