Alternatives for Managing the Nation's Complex Contaminated Groundwater Sites (2013)

National Academies Press: OpenBook

Chapter: 3 Remedial Objectives, Remedy Selection, and Site Closure

Visit NAP.edu/10766 to get more information about this book, to buy it in print, or to download it as a free PDF.

Suggested Citation:"3 Remedial Objectives, Remedy Selection, and Site Closure." National Research Council. 2013. Alternatives for Managing the Nation's Complex Contaminated Groundwater Sites. Washington, DC: The National Academies Press. doi: 10.17226/14668.

The issue of setting remedial objectives touches upon every aspect and phase of soil and groundwater cleanup, but none perhaps as important as defining the conditions for “site closure.” Whether a site can be “closed” depends largely on whether remediation has met its stated objectives, usually stated as “remedial action objectives.” Such determinations can be very difficult to make when objectives are stated in such ill-defined terms as removal of mass “to the maximum extent practicable.” More importantly, there are debates at hazardous waste sites across the country about whether or not to alter long-standing cleanup objectives when they are unobtainable in a reasonable time frame. For example, the state of California is closing a large number of petroleum underground storage tank sites that are deemed to present a low threat to the public, despite the affected groundwater not meeting cleanup objectives (California State Water Quality Control Board, 2010; Doyle et al., 2012). In other words, some residual contamination remains in the subsurface, but this residual contamination is deemed not to pose unacceptable future risks to human health and the environment. Other states have pursued similar pragmatic approaches to low-risk sites where the residual contaminants are known to biodegrade over time, as is the case for most petroleum-based chemicals of concern (e.g., benzene, naphthalene). Many of these efforts appear to be in response to the slow pace of cleanup of contaminated groundwater; the inability of many technologies to meet drinking water-based cleanup goals in a reasonable period of time, particularly at sites with dense nonaqueous phase liquids (DNAPLs) and complicated hydrogeology like fractured rock; and the limited resources available to fund site remediation.

This chapter focuses on the remedial objectives dictated by the common regulatory frameworks under which groundwater cleanup generally occurs. It first describes the phases of cleanup for the primary federal programs and their milestones, the gaining of which is often used as a metric of progress and ultimately success. The Comprehensive Environmental Response, Compensation, and Liability Act (CERCLA) and the Resource Conservation and Recovery Act (RCRA) guidance outline criteria for setting remedial objectives and points of compliance, and for selecting remedies to meet them. The chapter closes with a discussion of alternative strategies to address the current limitations on achieving groundwater restoration, such as CERCLA Technical Impracticability waivers for some portion of the site. This includes sustainability concepts that have become relevant to decision making regarding remedy selection and modification in the past few years.

The topic of setting cleanup objectives has a long history and was a significant component of the debates in the 1980s during the passage of the Superfund Amendments and Reauthorization Act (SARA) in 1986 and the establishment of the ARAR process in Section 121 of SARA. Several National Research Council (NRC) reports (1994, 2005) have provided insights and recommendations on improving the process of establishing objectives for groundwater cleanup. The DoD has also provided recommendations for setting objectives through reports published through the Environmental Security Technology Certification Program (e.g., Sale and Newell, 2011). Recently the Interstate Technology and Regulatory Council (ITRC) provided a comprehensive guidance document on setting objectives for remediation at DNAPL sites (ITRC, 2011). All these efforts have informed this overview of the objective setting process, which considers how that process might evolve in light of advances in our understanding of technical limitations to aquifer restoration.

The current regulatory framework for remediation of hazardous waste sites evolved from a complex collection of federal, state, tribal, and even local statutes, regulations, and policies. CERCLA and RCRA are the two federal programs that govern most subsurface cleanup efforts, and most state programs are similar to or even authorized under these federal models.

CERCLA provides federal authority for cleanup of sites with hazardous substances, usually excluding petroleum-only sites. At sites with no viable responsible party, EPA can fund remedial activities from the Superfund—a special account initially funded by a tax on petroleum and chemical compa-

nies, but presently derived from general tax revenues. However, at a majority of sites, the response is funded by private parties, either through a legally binding agreement to perform the remedy (e.g., an Administrative Order of Consent) or by reimbursing EPA for its remedial costs. At federal facilities cleanup is funded by the agency responsible for releasing contamination.

Initial Phases

A site regulated through CERCLA generally progresses through the Preliminary Assessment/Site Inspection, listing on the National Priorities List (NPL), site investigation (Remedial Investigation), remedial alternative assessment (Feasibility Study), remedy selection (Record of Decision), remediation implementation (remedial design followed by construction), and long-term monitoring and institutional controls until the site media concentrations are at or below unrestricted use levels (see Table 2-3). If there is an immediate threat to human health or the environment (“imminent and substantial endangerment”), the Preliminary Assessment/Site Inspection may trigger an interim emergency response.

The Remedial Investigation consists of detailed site characterization, while the Feasibility Study incorporates the evaluation of remedial alternatives that might meet remedial action objectives. The Remedial Investigation and Feasibility Study may be conducted concurrently, and, in any case, they influence each other. The Remedial Investigation generally includes a human health risk assessment and the determination of site-specific remedial action objectives. The Feasibility Study develops a series of remedial alternatives that describe the placement, timing, and remedial technology for cleanup activities, and it includes a detailed comparison of these alternatives with respect to criteria established under CERCLA regulations (see below).

Setting of Cleanup Goals and Selection of Remedies

CERCLA’s overarching groundwater remediation goal is to restore groundwater to its “beneficial use” “wherever practicable” (EPA, 2009a). A common beneficial use of groundwater, if conditions are appropriate, is that it be a source of drinking water. In addition, the groundwater plume “should not be allowed to migrate and further contaminate the aquifer or other media (e.g., vapor intrusion into buildings; sediment; surface water; or wetland)” (EPA, 2009a).

The alternative remedial strategies in the Feasability Study are evaluated based on a balancing of the nine criteria of the National Oil and Hazardous Substances Pollution Contingency Plan, usually called the National Contingency Plan (EPA, 1990):

1. Overall protection of human health and the environment (a threshold criterion that must be met by the chosen remedy)

2. Compliance with applicable or relevant and appropriate requirements (ARARs) (also a threshold criterion)

3. Long-term effectiveness and permanence (a balancing criterion)

4. Reduction of toxicity, mobility, or volume (a balancing criterion)

5. Short-term effectiveness (a balancing criterion)

6. Implementability (a balancing criterion)

7. Cost (a balancing criterion)

8. State acceptance (modifying criterion that is considered but not required to be met or balanced)

9. Community acceptance (modifying criterion)

Threshold Criteria. The first two criteria, called threshold criteria, must be met by the chosen remedy. The criterion “protective of human health” is sometimes embodied in a quantitative risk assessment and has been interpreted as having a calculated excess lifetime cancer risk between 10 –6 and 10 –4 or a hazard index < 1.0. 1 “Protective of the environment” is less clearly defined.

At most Superfund facilities with groundwater contamination, federal and state drinking water standards (such as maximum contaminant levels, MCLs, and non-zero maximum contaminant level goals) are established as ARARs and hence the groundwater cleanup goals. The designation of a drinking water standard as an ARAR is often independent of whether the particular groundwater is, in fact, currently used as a source of drinking water or is likely to be so used in the future, as long as it is capable of being used as a source of drinking water.

There is considerable variability in how EPA and the states consider groundwater as a potential source of drinking water. EPA has defined groundwater as not capable of being used as a source of drinking water if (1) the available quantity is too low (e.g., less than 150 gallons per day can be extracted), (2) the groundwater quality is unacceptable (e.g., greater than 10,000 ppm total dissolved solids, TDS), (3) background levels of metals or radioactivity are too high, or (4) the groundwater is already contaminated by manmade chemicals (EPA, 1986, cited in EPA, 2009a). California, on the other hand, establishes the TDS criteria at less than 3,000 ppm to define a “potential” source of drinking water. And in Florida, cleanup target levels

1 The hazard index (HI) is “the sum of more than one hazard quotient for multiple substances and/or multiple exposure pathways. The HI is calculated separately for chronic, subchronic, and shorter-duration exposures.” The hazard quotient is “the ratio of an exposure level to a substance to a toxicity value selected for the risk assessment for that substance (e.g., LOAEL or NOAEL)” http://www.epa.gov/oswer/riskassessment/glossary.htm.

for groundwater of low yield and/or poor quality can be ten times higher than the drinking water standard (see Florida Administrative Code Chapter 62-520 Ground Water Classes, Standards, and Exemptions). Some states designate all groundwater as a current or future source of drinking water (GAO, 2011). Although EPA generally defers to state or local groundwater classifications on these issues (EPA, 2009a), EPA policy recognizes that less stringent cleanup levels may be appropriate for groundwater that is not a current or reasonably expected future source of drinking water (GAO, 2011).

In addition to federal ARARs, states may propose requirements as state ARARs, subject to EPA acceptance. There is considerable variability between federal and some state ARARs, even for the same chemicals or situation, as described in Box 3-1. Table 3-1 demonstrates that the MCL for an individual compound can range over more than an order of magnitude, with some states being much more stringent than EPA. There are multiple reasons for these differences including differences in risk targets, different interpretations of technical feasibility, and different interpretations of toxicological findings.

Another example of variability among EPA and the states concerns the point of compliance. EPA has long directed that the point of compliance monitoring of the final cleanup levels for contaminated groundwater can apply “at and beyond the edge of the waste management area when waste is left in place” (EPA, 1988a, 1990, 1991a). (Note that the drinking water standard in this situation still defines whether the groundwater within the source area may be subject to unrestricted use.) At landfills the application of this policy is relatively straightforward, while at sites where DNAPL has migrated from the original area of release the application of this strategy may be more uncertain. 2 On the other hand, some states require that all points within a contaminated aquifer meet the state ARAR. All this variability can lead to different remedial objectives, different decisions about the chosen remedy, and different long-term outcomes.

Although the most commonly used ARAR, it is noteworthy that MCLs are not based on consideration of the vapor intrusion pathway, suggesting that there can be limitations to relying on ARARs based solely on drinking water ingestion in making decisions regarding remediation of groundwater contamination. Vapor intrusion is discussed further in Chapters 5 and 6.

Balancing Criteria. On a case-by-case basis, the remedy selection criteria (particularly the balancing criteria) are “balanced in a risk management

2 DNAPL may migrate within the area of waste management. At some CERCLA sites, the edge of the waste management area has been “flexibly applied,” while at others the edge of the waste management area has been “rigorously applied.”

BOX 3-1
State/Federal Differences in Goals for Groundwater Restoration

The differences between state and federal goals for groundwater restoration often hinge on the present and expected future use of the groundwater in question. However, even if the defined use of the groundwater is for drinking, there can still be differences in the actual numeric goals. This is because states have the option of developing their own, more restrictive MCLs that will replace the EPA’s MCL as the enforceable limit. Examples for different chemicals are given in Table 3-1, which provides a sense of the potential magnitude of state/federal differences but is not meant to be comprehensive.

In some cases, the difference between the federal MCL and the state MCL is more than an order of magnitude. For example, the federal drinking water limit for cis-1,2-dichloroethene (cis-1,2-DCE) is 70 ppb (1 ppb = 1 μg/L), whereas the California standard is 6 ppb. Both values are based on non-cancer liver toxicity in animals, with the differences mainly due to varying interpretations of toxicological findings. As another example, the federal MCL for carbon tetrachloride is 5 ppb, whereas the California standard is 0.5 ppb. Both carbon tetrachloride standards had similar conclusions regarding liver cancer in rodents as the critical endpoint. The differences for carbon tetrachloride are related to measurement feasibility and determination of the practical quantitation limit, rather than to differences in the underlying risk assessment (CalEPA, 2000).

In some cases, there are chemicals for which there are state standards but no federal standards. One example is perchlorate, where the Massachusetts standard is 2 ppb and the California standard is 6 ppb. Although both states chose the same toxicological study as the basis for establishing these limits, Massachusetts adopted a more conservative approach, both with respect to interpretation of the underlying human exposure study by Greer and coworkers (Zewdie et al., 2010), as well as with application of uncertainty factors to derive the non-cancer toxicity criterion (i.e., the reference dose or RfD). In addition, Massachusetts applied different assumptions regarding drinking water intake and other sources of perchlorate. Although the calculated health-based value for Massachusetts was 0.49 ppb, the state chose 2 ppb for risk management purposes to minimize compliance issues. In contrast, the California health-based value of 6 ppb is the same as the standard.

The reasons for differences in drinking water limits are varied and include the application of different toxicity studies to establish underlying health-based values, differences in application of uncertainty factors, variations in selection of exposure assumptions, and differences in risk management considerations. In some cases, the differences reflect the date when a standard was set, and does not always incorporate the new information that has become available for the more recent standard. 1 State/federal differences in drinking water limits may result in different levels of cost effectiveness and health protectiveness of remedial decisions across sites, as well as present risk communication challenges.

1 Due to lack of consideration of technical feasibility, advisory values can lower than mandated values, but they are not mandatory.

judgment as to which alternative provides the most appropriate solution for the site” (EPA, 1990). Under CERCLA, there is a preference for a permanent solution; indeed, EPA “expects to use treatment to address the principal threats 3 posed by a site, wherever practicable” (EPA, 1996a). However, there is “nothing in CERCLA §121 to suggest that selecting permanent remedies is more important than selecting cost-effective remedies” (Ohio v. EPA, 997 F.2d 1520, 1533, D.C. Cir. 1993). Rather, the emphasis on permanent solutions and treatment is balanced by the co-equal mandate that remedies be cost-effective through the addition of the wording to the maximum extent practicable (EPA, 1996a) (see Box 3-2). EPA believes that “certain source materials are generally addressed best through treatment because of technical uncertainties regarding the long-term reliability of containment of these materials, and/or the consequences of exposure should a release occur,” while other source materials generally can be reliably contained (EPA, 1996a).

An issue discussed in Chapter 7 but introduced here is that of the discount rate and its role in remedy selection in addressing one of the nine NPL criteria, namely cost effectiveness. During the feasibility study, cost estimates are developed for each remedial option to identify their relative cost effectiveness. Once costs are identified and quantified for each remedial option, they are discounted to a present value to adjust for differing annual costs across options. For example, some remedies may have large costs in the near future and other remedies may have large costs in the distant future; discounting is a mechanism to compare the costs of remedial options using a common dollar metric. The logic for discounting is that if firms were able to invest these funds they would earn a positive rate of return in the future, which means that expenditures in the present have a higher cost than expenditures in the future.

Currently, the annual cost of each option in EPA feasibility studies for private parties is discounted to present values using a presumptive value of 7 percent, which EPA argues reflects the long-term return to private capital in the United States (OMB, 2003; EPA and USACE, 2000; EPA, 2010a). Discount rates from Appendix C of OMB Circular A-94 (OMB, 2012), which currently are significantly lower than 7 percent, are generally used for all federal facilities.

Under the current approach to discounting, options with costs in the distant future will have lower present values than options with front-loaded costs. For example, with the discount rate of 7 percent, $1 next year is worth about 94¢ today and $1 in 50 years is worth about 3¢ when dis-

3 In addition to drum wastes and other similar source material, principal threats are where the toxicity and mobility of the source material combine to present an ingestion risk of greater than 10 –3 (EPA, 1991c).

TABLE 3-1 Examples of State versus Federal Maximum Contaminant Levels

Name	Tetrachloroethene (PCE)	Trichlorethene (TCE)	cis-1,2Dichloroethene (cis-1,2-DCE)	1,2,3-Trichloro-propane (1,2,3-TCP)
U.S. EPA	5 ppb	5 ppb	70 ppb	n/a
California	5 ppb	5 ppb	6 ppb (state MCL)	n/a
Florida	3 ppb (state MCL)	3 ppb (state MCL)	70 ppb	n/a
Massachusetts	5 ppb	5 ppb	70 ppb	n/a
New Jersey	1 ppb (state MCL)	1 ppb (state MCL)	70 ppb	n/a
New York	5 ppb	5 ppb	5 ppb (state MCL)	5 ppb (state MCL)

a EPA interim advisory level for perchlorate is 15 ppb.

b The Massachusetts MCL “is directed at the sensitive subgroups of pregnant women, infants, children up to the age of 12, and individuals with hypothyroidism. They should not consume drinking water containing concentrations of perchlorate exceeding 2 ppb. MassDEP [Massachusetts Department of Environmental Protection] recommends that no one consume

counted to the present. Thus, a cost-efficiency determination tends to favor selection of options that have larger costs in the future and lower near-term costs. Pump and treat, in particular, is an option that discounting favors because the remedy might operate for decades and the present-value calculation indicates the costs of this operation beyond 50 years is $0. A lower discount rate, such as the 3 percent social rate for public projects, would increase the present value of $1 in 50 years to 23¢ today, but it is still likely

Carbon Tetrachloride	Perchlorate	Source	Internet URL
5 ppb	n/a a	National Primary Drinking Water Regulations	http://www.epa.gov/safewater/contaminants/index.htm#listmcl
0.5 ppb (state MCL)	6 ppb (state MCL)	State Code of Regulations (Chapter 15, Title 22, Articles 4 and 5.5)	http://www.cdph.ca.gov/certlic/drinkingwater/Documents/Lawbook/dwregulations-01-01-2009.pdf
3 ppb (state MCL)	n/a	State Code of Regulations (Chapter 62-550)	http://www.dep.state.fl.us/legal/Rules/drinkingwater/62-550.pdf
5 ppb	2 ppb (state MCL) b	2008 Standards and Guidelines for Contaminant in Mass. Drinking Water	http://www.mass.gov/dep/water/drinking/standards/dwstand.htm
2 ppb (state MCL)	n/a	State Code of Regulations (N.J.A.C. 7:10)	http://www.state.nj.us/dep/watersupply/sdwarule.pdf
5 ppb	n/a	State Code of Regulations (Part 5, Subpart 5-1)	http://www.health.state.ny.us/environmental/water/drinking/part5/tables.htm

SOURCE: Modified, with permission, from Julie Blue, Cadmus Group, Inc. (2009).

that the alternative with higher future costs would be selected over options with high costs in the near future.

Most economists agree that discounting is necessary, because to not discount would overlook the differential time paths of costs across remedy options. There is a long-standing debate over what discount rate is appropriate for use in environmental cases where the costs may be intergenerational. While it is beyond the Committee’s charge to opine on the appropriate discount rate, discounting should be considered very carefully

BOX 3-2
Guidance on Definition and Application of “Maximum Extent Practicable”

The Committee was charged with answering the question: what should be the definition of “to the extent practicable” when discussing contaminant mass removal. Terms like “maximum extent practicable (MEP),” “to the extent practical,” “practicability,” etc., are routinely heard when discussing what can be achieved during groundwater remediation. For example, EPA groundwater remediation guidance, which applies to all EPA non-UST cleanup programs, repeatedly states that EPA’s goal is to attain drinking water standards “wherever practicable.” The UST regulations 40 CFR 280.64, which apply only to light nonaqueous phase liquid (LNAPL), requires removal of free product “to the maximum extent practicable” as determined by the implementing agency at sites where free product is present. These terms are not defined explicitly or quantitatively in the federal or state statutes, regulations, or settlements and administrative orders that dictate remediation requirements for soil and groundwater. That is, statements as explicit as “70% reduction in concentration” or “removal of mobile DNAPL” are not provided as definitions of “maximum extent practicable.”

The main statutory reference to the term “maximum extent practicable” is found in CERCLA in reference to practicability during remedy selection, where practicability reflects a balancing of the nine criteria specified in the NCP (EPA, 2009a, p. 4, footnote 9). EPA guidance states that CERCLA’s emphasis on permanent solutions and treatment should be balanced by “the co-equal mandate for remedies to be cost-effective” through the addition of the wording “to the maximum extent practicable” (EPA, 1996a). EPA considers cost to be relevant to technical impracticability because that term is “ultimately limited by cost,” although EPA policy is that cost should generally play a subordinate role in a technical impracticability determination unless compliance would be “inordinately costly” (EPA, 1996a).

For this limited use of the term “maximum extent practicable,” an explicit definition is already available. EPA has concluded that treatment is not practicable when

in the weighing of alternatives along with the other four National Contingency Plan (NCP) balancing criteria listed above. Specifically for projects whose duration exceeds 30 years, EPA and the Army Corps of Engineers (2000) recommend that the present value analysis include a “no discounting” scenario to demonstrate (for comparison purposes only) the impact of the discount rate on the total present value cost of the remedy and the relative amounts of future annual expenditures.

Modifying Criteria. Normally the lead agency evaluates a number of remedial alternatives against the first seven criteria and presents that evaluation, designating a preferred alternative to the public (i.e., community

(1) “treatment technologies are not technically feasible or are not available within a reasonable time frame;” (2) “the extraordinary volume of materials or complexity of the site may make implementation of the treatment technologies impracticable;” (3) “implementation of a treatment-based remedy would result in greater overall risk to human health and the environment due to risks posed to workers, the surrounding community, or impacted ecosystems during implementation (to the degree that these risks cannot be otherwise addressed through implementation measures);” or (4) “implementation of the treatment technology would have severe effects across environmental media” (EPA, 1997a). As an example of the second item above, the use of containment as a presumptive remedy for municipal landfills (EPA, 1997b) means that removal of waste from source areas in those situations can be interpreted as generally not practicable. This case-by-case application of the concept of practicability has been upheld in several court cases [State of Ohio v. U.S. Env’l’t Prot. Agency, 997 F.2d. at 1532 and U.S. v. Ottati & Goss, Inc., 900 F.2d 429 (1st Cir. 1990) (opinion by now Supreme Court Justice Breyer)]. Thus, as long as the remedy is chosen in accordance with the NCP and is performing in accordance with reasonable environmental engineering practices, that is the end of decision making with respect to what is practicable for remedy selection.

The term “maximum extent practicable” is often used informally as a measure of remediation progress even though it has no regulatory bearing in that context. In Chapter 7, the Committee suggests that remedies at complex sites be regularly assessed to determine whether they are being implemented in a manner consistent with good environmental engineering practice and their resulting performance. If a remedy reaches a point where continuing expenditures bring little or no reduction of risk prior to attaining drinking water standards, the Committee recommends that there should be a reevaluation of the future approach to cleaning up the site (called a Transition Assessment). When this point is reached, the chosen active remedy can be said, de facto, to have been operated to the “maximum extent practicable.”

stakeholders) in the form of a Proposed Plan. With regard to the two final, modifying criteria, neither the state nor the community have the legal authority to “veto” a remedy. The provision does mean that the lead agency must engage in a formal community involvement process and, at each NPL facility, provide a technical assistance grant to one eligible nongovernmental organization to hire an independent technical consultant to advise the community. EPA recognizes about 70 Community Advisory Groups at NPL facilities across the country. From 1988 to 2010, 323 technical advisory grants have been awarded (205 providing $50,000 or less and 15 providing a total of more than $250,000) (Catalogue of Federal Domestic Assistance, 2011). Following the public comment period, the lead agency selects a remedy and memorializes it in a Record of Decision.

After Remedy Selection

Following remedy selection decision, the remedy is designed, constructed, and operated. Once an active remedy is operating properly and successfully, it is considered to have met the EPA Construction Complete milestone. Operation and maintenance continue as long as an active remedy is needed to be protective. Optimization evaluations and five-year reviews are performed if chemical concentrations remain above unrestricted use levels in groundwater, soil, soil vapor, and other media (EPA, 2001a). As described in greater detail in Chapter 7, at these later stages monitoring data may be gathered, the remedy may be adjusted, and institutional controls (designed to minimize the potential for human exposure to residual contamination and/or protect the integrity of the remedy) are imposed. According to the NCP, institutional controls are supposed to supplement, not substitute for, active remediation “unless such active measures are determined not to be practicable, based on the balancing of trade-offs among alternatives that is conducted during remedy selection” [40 CFR § 300.430(a)(iii)(D)].

RCRA Corrective Action

Congress enacted RCRA in 1976 to regulate, by permit, the treatment, storage, and disposal of hazardous wastes. In 1984 it amended the law to regulate cleanup at facilities with RCRA permits (40 CFR section 264.101). Though RCRA is a federal law, most RCRA implementation is conducted by the states and territories. Today 43 states and territories have been delegated primacy over their RCRA Corrective Action programs. Therefore, there is more variation in RCRA oversight than under EPA’s CERCLA program.

The RCRA remedy selection process and criteria are generally similar to the CERCLA process (EPA, 1996b, 1997a, 2011a). Implementation of corrective action can vary from site to site (and state and state) but it invariably begins with an evaluation of site conditions through an RCRA facility assessment conducted by either EPA or the authorized state. Similar to the Preliminary Assessment/Site Inspection phase of CERCLA, this involves examination of the facility’s solid waste management units to determine if a release occurred or if the potential for a future release exists. Interim action to stop the spread of contamination or provide an alternate source of drinking water may be required during this stage. Additional information can be necessary to support interim actions and can be obtained by the site owner through an RCRA Facility Investigation. This investigation involves sampling and modeling to determine the nature and extent of contamination, the site hydrogeology, and the source zone architecture, similar

to the Remedial Investigation process under CERCLA. If it is determined that corrective action is required, the site owner will conduct a corrective measures study. Not unlike the feasibility study in CERCLA, a corrective measures study evaluates and selects the remedy and is conducted by the facility owner with oversight from the EPA or the state.

The RCRA program recommends that corrective action be based on risk (EPA, 1997c). EPA’s RCRA guidance specifies that cleanup levels be set at federal drinking water standards (where they exist) or be based on a residential drinking water exposure scenario where groundwater is currently used or may be reasonably expected to be used as a source of drinking water (EPA, 2004). RCRA regulations define the point of compliance as the “vertical surface located at the hydraulically down gradient limit of the waste management area that extends down to the uppermost aquifer underlying the regulated units” (EPA, 2004), which conceptually is the boundary of the waste disposal or other management area at the RCRA facility. The exact location is determined on a site-by-site basis.

The two primary RCRA milestones include the human exposures environmental indicator and the groundwater environmental indicator (see Chapter 2). The objectives that are frequently called for in site-specific agreements between owners and operators of treatment, storage, and disposal facilities and regulatory authorities are typically defined in terms of concentrations of particular contaminants as measured at the boundaries of given units of real property.

Public participation is a part of the corrective measures selection process, but community acceptance (the ninth NCP criterion) is not a statutory requirement for RCRA sites. While in many cases regulators may have established a robust community involvement process, in general this is less extensive than at sites regulated under CERCLA. For example, there are funding sources, such as Technical Assistance Grants, available for CERCLA public involvement that do not exist for RCRA, and regional Superfund programs have Community Involvement Coordinators.

While RCRA permits do not have a statutory requirement for five-year reviews, periodic reviews may be built into RCRA permits. EPA views RCRA permits as “living documents that can be modified to allow facilities to implement technological improvements, comply with new environmental standards, respond to changing waste streams, and generally improve waste management practices” (EPA, 2011b).

As part of RCRA, UST cleanup is also overseen by state and territories or their subjurisdictions. Of interest for this chapter is that the definition of UST “closure,” which is a major goal of UST programs, varies significantly from state to state. According to the ITRC 2009 report, historic cleanup goals for LNAPLs have been to remove them “to the maximum extent practicable (MEP),” although some states provide no interpretation

of MEP and others specify a maximum allowable amount of LNAPL in a monitoring well (e.g., no visible sheen or 1/8-inch thickness). Some state statutes include “LNAPL thickness-in-a-well requirements” and definitions for when LNAPL remediation efforts may be discontinued. Some states may be bound by statute to remove all LNAPL based on a law or policy stipulating nondegradation of waters.

Federal Facilities

Current and former federal facilities are subject to the same environmental cleanup laws as other properties (see Section 120 of CERCLA), but there are differences. For example, in 1986 Congress established the Defense Environmental Restoration Program (as part of the Superfund Amendments and Reauthorization Act, SARA, 1986), requiring the Defense Department to fund its own cleanups. Other federal agencies are similarly liable for the remediation of their properties.

In general, the Defense Department manages most of its facilities under CERCLA, whether or not they have been listed on the NPL. A major reason for this is that in 1987 President Reagan assigned lead agency status to federal responsible parties. At NPL sites, the lead agency is supposed to negotiate a Federal Facilities Agreement with EPA and its state counterparts. These agreements define the scope and timing of the cleanup, and they establish a dispute resolution mechanism whereby the EPA administrator is ultimately responsible for resolving differences between regulators and responsible parties. Federal responsible parties are responsible for conducting five-year reviews under CERCLA, but EPA must approve the finding of protectiveness.

The major federal responsible party agencies, the Departments of Defense and Energy, maintain robust community involvement programs, even at facilities that are not on the NPL. Currently the Defense Department sponsors 191 site-specific Restoration Advisory Boards covering 218 installations (DoD, 2010), and DOE hosts similar bodies at most of it major sites.

A fraction of contaminated federal facilities are regulated under RCRA. In 1992, Congress amended the law to make explicit that states have the legal authority to enforce RCRA cleanup requirements at federal facilities. In the Committee’s opinion, this may be a major reason that federal agencies prefer CERCLA, where they will maintain lead agency status, even though RCRA provides greater flexibility in establishing remedial objectives and points of compliance.

Federal facilities that are being transferred to non-federal ownership are subject to additional oversight under CERCLA Section 120(h). In most cases, remedies must be in place and operating properly and successfully

before a parcel can be transferred (EPA, 2010b), although groundwater concentrations need not meet drinking water standards prior to transfer. EPA and state regulators must issue a Finding of Suitability for Transfer, providing EPA with authority over federal cleanup at closing military bases and other properties, even if they are not on the NPL (DoD, 1994). There are also provisions for Leasing and Early Transfer, in which non-federal entities may use or take ownership of property before cleanup has been completed (EPA and DoD, 2005; DOE, 1998). In general, this means that regulators must approve of remedies if a transfer is to occur. However, properties that were transferred before the 1986 Superfund Amendments, such as the Defense Department’s Formerly Used Defense Sites and the former Atomic Energy Commission’s Formerly Used Site Remedial Action Program sites, are subject to CERCLA as managed by the Army Corps of Engineers. They are regulated only by the states and territories unless they are placed on the NPL, which gives EPA regulatory oversight as well.

Lessons Learned

The process outlined above for CERCLA and its counterparts occurs in a straightforward way at only relatively small or simple sites. In reality, the remedial action process is much more complex and nonlinear, particular for the type of sites that are the focus of this study. The process at a particular site can also be more flexible than implied in the description above. The Committee’s combined experience provides the following general observations about how cleanup can deviate from the idealized RCRA and CERCLA models. First, a significant amount of cleanup can be implemented through interim and emergency responses. Second, the study phase is often protracted, for several reasons. And third, at many complex sites attaining drinking water standards throughout the contaminated groundwater zone is difficult and unlikely for many decades, which can complicate the latter stages of remediation.

Interim and Emergency Responses

At most complex sites, actual cleanup activity begins long before the selection of a final remedy. First and foremost, easily accessible source materials can be and are quickly removed, such as piles of drums on the ground surface, leaking lagoons, and surface pits. Sites with surface contamination are typically fenced to prevent easy access. Second, measures are taken to interrupt exposure pathways. For example, in the San Gabriel Valley, California, wellhead treatment was provided to ensure that the public water supply, which derives from contaminated groundwater,

meets health standards. 4 At the Hopewell Precision Superfund facility in Hopewell Junction, New York, impacted homes were provided with water filtration and vapor mitigation systems. 5 People whose private wells were contaminated with perchlorate from the Olin plume in San Martin, California, were provided with bottled water (California Regional Water Quality Control Board, Central Coast Region, 2003). Containment remedies are often applied at the earliest stages of site response to prevent the spread of contamination. For example, at the MEW Superfund Study Area in Mountain View, California, responsible parties quickly installed slurry walls around the known source areas on their properties, and they found and plugged abandoned agricultural wells that served as vertical conduits for contamination to move between aquifers (EPA, 2009b).

Regulators and responsible parties often agree to conduct source removal or containment long before the full extent of contamination is even mapped. At the CTS Asheville site in North Carolina, EPA conducted soil vapor extraction as an emergency response prior to a remedial investigation (EPA, 2010c). At the MEW site responsible parties removed contaminated soil, conducted soil vapor extraction, and installed localized groundwater extraction and treatment systems long before the development of a regional remediation strategy (EPA, 2009b).

The lesson learned from existing case studies and the experience of the Committee is that in geographic locations where there are numerous separate sources affecting the same aquifer, a regional remediation strategy that addresses sources in a variety of federal and state programs (e.g., CERCLA, RCRA, and UST) early in the process and with the involvement of all stakeholders can allow for Interim and Emergency Responses to be implemented in a more effective manner. It is also consistent with EPA environmental justice program’s efforts to use “an integrated One EPA presence” to engage communities in the Agency’s work to protect human health and the environment (EPA, 2011c).

Protracted Study

Robust, reliable site characterization is essential to effective site cleanup. Without it, remedies may fail to address significant problems or may even spread contamination. For a number of reasons, investigation of a complex site is always protracted. First, it is inherently difficult to characterize groundwater contamination and develop an accurate conceptual model at complex sites. Once sampling schedules are established, it can be

difficult to change them. This is especially true at sites where assessment is an exercise in routine data gathering, rather than an attempt to improve the understanding of site conditions. In addition, at virtually all sites sampling results in new discoveries that may change the sampling strategy. Second, the nature of the study process is adversarial (i.e., where the work and funding come from the responsible party, but the final decision about moving forward rests, as it must, with the regulators). Third, the process of having a large number of government experts (both state and federal) review different portions of the responsible party’s submissions adds time. During the Remedial Investigation/Feasibility Study phase multiple documents are created, including specialized studies. Actual cleanup, of course, cannot proceed until regulators review responsible party documents, the responsible parties respond to regulator comments, and all outstanding issues are resolved. A lack of adequate staffing in state and federal agencies aggravates this situation (e.g., Sweeney, 2010). Finally, the interpretation of study data is nontrivial and often the subject of disputes between EPA and the potentially responsible parties.

In recent years, agencies have emphasized the establishment of data quality objectives to be certain that the quality of samples will be high enough to answer key questions about and to test hypotheses of the conceptual model. EPA’s Data Quality Objective Process (EPA, 2006) discusses how to “clarify study objectives, define the appropriate type of data [to collect], and specify tolerable levels of potential decision errors that will be used as the basis for establishing the quality and quantity of data needed to support decisions.” Nonetheless, in the Committee’s experience there is still a strong tendency to collect too much information for fear of missing a key data point, leading to protracted study at these complex sites.

There have been initiatives, such as the Air Combat Command’s “Streamlined Oversight” project (U.S. Air Force, 1995), piloted at Langley Air Force Base, Virginia, in which regulators and responsible parties have formed partnerships to jointly solve problems, eliminating much of the back-and-forth shuffle of documents, but those programs remain the exception rather than the rule.

The Limits of Aquifer Restoration

As shown in many previous reports (EPA, 2003; NRC, 1994, 1997, 2003, 2005), at complex groundwater contamination sites (particularly those with low solubility or strongly adsorbed contaminants), conventional and alternative remediation technologies have not been capable of reducing contaminant concentrations (particularly in the source area) to drinking water standards quickly. Because the history of groundwater cleanup is still relatively recent, in that few sites with remedies have been

operating for more than 25 years, the time to achieve restoration cannot be easily predicted based on empirical observations, but it likely extends for decades. As a practical matter, at both Superfund and RCRA sites a variety of strategies are being used, which recognize that drinking water standards are unlikely to be attained within source areas. These methods include the use of monitoring compliance points outside the source area, use of containment zones for petroleum and low-risk solvent sites (by the California Water Resources Board and the Regional Water Quality Control Board-San Francisco Bay Area, respectively), Texas’ Municipal Settings Designation, Florida’s Natural Attenuation Default Source Concentration, and EPA’s Technical Impracticability (TI) waivers, among others.

Because of the diversity of chemicals and conditions at sites, the limits of existing technologies, and the inevitable lack of agreement on the proper balance between the nine criteria of the NCP, there is no precise formula or clear trigger for determining when restoring an aquifer to drinking water standards is practicable or what is a reasonable remediation time frame in which to accomplish this, and these debates are likely to continue. Rather, general remedy selection principles (laid out in many EPA guidance documents and described in this chapter) should be applied to the specific conditions at a site to determine the remedy. The remedial alternatives should be reviewed to determine the timeframe, the cost, and the practicality of reducing the concentrations in groundwater to drinking water standards. This requires the transparent exchange of technical and cost information between regulators and responsible parties. Other implications of the limits of aquifer restoration are discussed more fully in Chapter 7.

The Committee assumes that drinking water standards will remain the long-term goal of groundwater remedies for the foreseeable future. Drinking water standards define unlimited groundwater use and unrestricted exposure, and until they are met, five-year reviews (at sites regulated under CERCLA) and institutional controls are needed. Despite these requirements, the Committee believes that new approaches to setting cleanup goals should be considered, to the extent permitted by law. These include giving more attention to site-specific risks, setting alternative concentration limits, seeking TI waivers, reclassifying groundwater, and considering sustainability, as discussed below.

A More Central Role for Risk Assessment

During the RI phase of CERCLA, information is collected that can be used to conduct human health and ecological risk assessments, usually

following EPA’s Risk Assessment Guidance for Superfund (RAGS) (EPA, 1989, 1991b). Site-specific risk assessment integrates information on the physical conditions at the site, the nature and extent of contamination, the toxicological and physicochemical characteristics of the contaminants, the current and future land use conditions, and the dose-response relationship between projected exposure levels and potential toxic effects. The end result is a numerical value of potential additional risk to the hypothetical receptor from the contaminant source under present conditions (i.e., the “no-action” scenario), along with a discussion of the attendant uncertainty. The calculated risk values are typically compared to the range of acceptable risk defined by EPA or by state regulations (often 10 –6 to 10 –4 for carcinogenic compounds). If the risk estimate is greater than the acceptable target risk level, target cleanup level objectives are identified for the site using the assumptions developed in the risk assessment related to potential levels of exposure.

In the Committee’s experience, EPA and state drinking water standards usually drive groundwater cleanup rather than the results of site-specific risk assessment. This can lead to responsible parties, regulators, and the public having an incomplete understanding of risk-related issues, including the plausibility of the scenarios that are driving decision making, the likely site risks at the present and in the future, and site risks reduced to date. Hamilton and Viscusi (1999) provided several examples, taken from Superfund, of the importance to risk estimates of assumptions regarding the selected scenarios (e.g., future on-site resident as compared to present-day off-site residents). These authors also provided estimates of both individual and population risks, demonstrating that at some sites population risks, reflected in the number of estimated cancer cases, can be small, often well below 1. Rarely is such information provided as part of the typical risk assessment process. Moreover, exposure pathways, such as inhalation of vapors from off-gassing during showering or inhalation of chemicals from vapor intrusion, are often not reflected in ARARs for groundwater, especially MCLs. Failing to consider these pathways could yield over- or underestimates of risks.

More Comprehensive Consideration of Time

The risk-based methods typically used at contaminated sites evaluate carcinogenic and non-cancer risk to a hypothetical individual over the course of the person’s lifetime. These methods do not factor in the changes in concentrations or exposure over the lifetime of a contaminant source. For example, the future potential risk is calculated over a 30-year timeframe 6

6 Thirty years is the typical exposure duration in the baseline risk assessment under RAGS.

based on the reasonable maximum exposure concentration determined in the remedy-selection risk assessment, even if the chemical concentration decreases going forward without any active remedy. In this example, the risk reduction predicted as a result of the selected remedy (and, therefore, the cost effectiveness and practicality of that remedy) would be overestimated since some of the risk reduction would have occurred even in the absence of active remediation. The only example known to the Committee of where time has been considered more explicitly is EPA’s guidance on remediation of polychlorinated biphenyl sites. This guidance recommends that the calculated risk consider concentration decreases over time from volatilization (e.g., a 72 percent reduction in concentration over 30 years) and biodegradation (e.g., a half-life of 50 years) (EPA, 1990).

Similarly, there is also no formal framework for considering the impact on risk of concentrations ceasing to decline once a remedy has been in place for an extended period (see Chapter 7 for more discussion). This is not an uncommon occurrence at complex sites, where contaminant concentrations may reach an asymptote beyond which there is very little, if any, further decline in concentration despite continued operation of an active remedy. In such situations, the reduction in potential risk has also plateaued (i.e., risk reduction ceases) and is not achieving its full extent as predicted in the Record of Decision.

The Committee believes that more formal consideration of the time element in risk assessment (i.e., by linking predicted changes in concentrations with and without a remedy to changes in risk, assuming that drinking water is a complete exposure pathway) can be important in understanding the cost effectiveness of a remedy. In addition, once a remedy is implemented, understanding the risk–concentration response function over time will provide risk managers and the community with a more complete understanding of the changing risk profile and, if restoration is not practical, facilitate decision making regarding the need for long-term management. These issues are discussed further in Chapter 7.

Population Impacts and Risks to Remediation Workers and Surrounding Residents

Population risk is commonly represented as the number of cases of disease or fatalities in a particular exposure setting. For cancer risks, this might be presented as the hypothetical estimated number of cancer cases associated with a particular exposure scenario. The value is calculated based on population size and average risk level. 7 Estimation of population

7 We note this is a simplification of the calculation and factors such as the age distribution of a population may also be considered.

risk over the lifetime of a hazardous waste site is rarely if ever conducted because there is no regulatory requirement to do so, nor is there a currently prescribed regulatory context for considering the results of such an evaluation. Because of the relatively small population size affected by a given Superfund facility, the total number of cancer cases associated with contaminant exposure is likely to be small (often less than 0.1 cancer cases per site according to Hamilton and Viscusi, 1999). In contrast, consideration of population risk is an important component of many federal rules in other settings, including other environmental exposures as well as occupational and pharmaceutical exposures. For example, in the setting of the arsenic MCL, EPA considered population size in reducing arsenic in drinking water from 50 μg/l to the options of 3, 5, or 10 μg/L. Benefits included the number of lung and bladder cancer cases averted in the potentially exposed population and the subsequent impact on reduction in costs of morbidity and mortality (EPA, 2001b).

Another risk component to remedial decision making related to population size that is infrequently quantified in any formal analytical way is that of short-term risks created as part of remedial activities. For every remedial alternative, there may be short-term risks to workers during implementation of the remedy (e.g., due to excavation of large volumes of waste and contaminated soil at landfills and/or treatment facilities), short-term risks of injury to local residents and populations along the transportation route due to traffic accidents during transportation of such wastes, and long-term risks to local residents who live near the redisposal site (e.g., Greenberg and Beck, 2011). Despite the existence of these risks, few remedy selection decisions consider them in a quantitative way (Leigh and Hoskin, 2000).

Population risks and risks from physical injuries to remediation workers (including on-site injuries) have been quantified in the site remediation context using the metric of years of potential life lost (YPLL) (Cohen et al., 1997). YPLL, which is used in public health decision making, considers the number of fatalities resulting from particular activities and the age of the individuals experiencing the fatalities. For some of the hypothetical case studies in the analysis by Cohen et al. (1997), the increase in YPLL to remedial workers was greater than the reduction in YPLL for the population at the site. Use of this concept in decision making, particularly when remediation extends over long periods, could be useful in selecting among remedial alternatives to find those with the largest overall benefit to public health (e.g., by reducing YPLL to the greatest extent practicable). An example of the practical application of the YPLL concept in environmental decision making (albeit not in the context of groundwater cleanup) is provided by Frost et al. (2002) who concluded that, in the analogous context of selecting among different strategies for reducing drinking water arsenic in Albuquerque, NM, the drinking water treatment approach of coagulation

and microfiltration yielded greater public health benefit (i.e., fewer YPLL) than other drinking water treatment approaches. Note that the Committee is not suggesting that population-size considerations should be used to select remedies that are less health protective—i.e., remedies should result in post-remedy risks within the target risk range for all populations.

Uncertainty and Variability in Risk Analyses

At complex sites, risk-based methods could be employed to more fully understand the nature of existing risk and expectations regarding future risks reduced under different scenarios through more explicit uncertainty analysis. The results of uncertainty analysis could help identify areas where additional data collection may be beneficial and provide risk managers and communities with a greater understanding of the implications of specific decisions.

In the context of risk assessment, variability refers to the natural variation that occurs across space and time in a population—including differences in exposure point concentrations, intake assumptions (e.g., breathing rates), and pharmacokinetic differences among individuals as a function of age, genetics, or other factors. For a given contaminated site, such differences yield a distribution of risks across the population of present and future residents living near a site. Use of probabilistic risk analysis methods, in which a distribution of risks is presented, provides a more complete understanding of variability.

Uncertainty refers more generally to a lack of knowledge about specific parameters. For example, the shape of the dose-response curve for a carcinogen at low doses (e.g., whether is it linear, nonlinear, sublinear) is often uncertain. Reasons for uncertainty may include a lack of experimental data in the dose range of interest or lack of understanding of the mode of action for carcinogenesis. In the case of groundwater contamination, an important source of uncertainty in risk assessment is often the choice of use scenarios for contaminated groundwater. If the water is presumed to be used for drinking water, then the relevant pathways would include ingestion, inhalation from off-gassing during showering and vapor intrusion, and dermal uptake. If the water is not used (and not reasonably expected to be used in the future) for drinking water, then only inhalation from vapor intrusion would constitute a complete exposure pathway.

Tools are available for formally characterizing uncertainty in risk assessment and could be applied more frequently at contaminated sites. For example, expert judgment could be used, based on assumptions regarding demographics and types of water usage in an area, to formally elicit determinations as to the likelihood of certain use scenarios. Sensitivity analyses could be conducted comparing different use scenarios, incorporating tem-

TABLE 3-2 Examples of Uncertainty and Variability in Risk Assessment for Contaminated Groundwater

Parameter	Source of Uncertainty	Source of Variability
Carcinogenicity of chlorinated solvent	Lack of toxicological understanding in low-dose region	Differences in metabolizing ability across individuals, resulting in differences in susceptibility to toxicity
Sampling for compliance with cleanup targets	Differences in detection limits as a function of changing technologies	Differences in concentrations across time and space
Groundwater use	Changes in water use patterns in the future can affect the plausibility of the use scenario	High variability in water ingestion rates
Exposed population	Changing demographics	Age distribution of population, which can affect water consumption patterns

porality into the analysis. More sophisticated modeling software tools are becoming available to conduct analyses of variability and uncertainty. For example, the Analytica program (Mansfield et al., 2009) provides general mathematical modeling language to develop uncertainty models, which can then be combined with probabilistic modeling. EPA has employed Analytica to combine information on variability in exposure along with uncertainty in dose-response function to evaluate the benefits and costs of air quality regulations. While such analyses can be complex and time consuming, they are likely to be worthwhile for certain situations at recalcitrant sites, and informative in the context of making decisions regarding the need for alternative endpoints. Table 3-2 presents some sources of uncertainty and variability in risk assessment for contaminated groundwater.

Additional Strategies for Goal Setting

Many strategies have been developed and accepted by regulators to acknowledge site complexity and inherent technical and cost barriers to achieving drinking water standards, yet provide a path forward that reduces risk and retains the ability to determine when unrestricted use is appropriate. Examples include applying for and being granted a TI waiver, groundwater reclassification, applying EPA’s flexible guidance on determining if a requirement is an ARAR (including applying the exceptions, exemptions, and variances associated with the federal or state requirement), use of

compliance monitoring points outside the presumed source area, and use of alternate concentration limits, among others (ESTCP, 2011; EPA, 1996c, 2009a).

There are multiple benefits of using these additional strategies, each of which is provided for in EPA guidance (EPA, 1996c, 2009e, 2011d). These strategies can meet most regulatory requirements, establish common expectations, protect public health through exposure control, provide a pathway toward meeting the DoD milestones of remedy in place and response complete, manage remedial project risks, and potentially use resources more efficiently. The challenges are also significant, and include regulatory reluctance to adopt such additional strategies because of (a) scientific disagreements on the fate of chemicals or technological performance; (b) disagreement about what is a reasonable time frame or what is cost effective; (c) community concerns; and (d) uncertainty about the ability to control long-term risks. Whether these alternative strategies can still be protective while leading to reductions in life-cycle costs is difficult to quantify, but intuitively cost savings seem likely.

ARAR Waivers

Under CERCLA, the selection of an ARAR requires a careful application of site-specific facts to the site of interest. A requirement under other environmental laws may be either applicable 8 (i.e., it would apply, but for this being a Superfund facility) or relevant and appropriate (i.e., it addresses problems or situations similar to the conditions at the site and is a requirement that is “well suited” to the site) [see 40 CFR §300.5 and 40 CFR §300.400(g)]. To determine if a requirement is well suited, one must assess the nature of the substances at the site; the physical, chemical, and microbial characteristics at the site; the circumstances surrounding the release; and the ability of the action to address the release. Thus, an ARAR must be determined on a case-by-case basis and the analysis may provide substantial flexibility.

ARARs, even if applicable, may also be waived, e.g., TI waivers. There is no rigid definition of what constitutes technical impracticability (EPA, 1993; AEC, 2004). Eighty-five TI waivers have been issued for ground-

8 “Applicable requirements” are those “cleanup standards, standards of control, and other substantive environmental requirements, criteria, or limitations promulgated under federal environmental or state environmental or facility sitting laws that specifically address a hazardous substance, pollutant, contaminant, remedial action, location, or other circumstance found at a CERCLA site.” 40 C.F.R. § 300.5; EPA, Draft ARARs Guidance at pp. 1-10. There should be a one-to-one correspondence between the requirement and the circumstances at the site. Id. “Applicability” implies that the remedial action or the circumstances at the site satisfies all of the jurisdictional prerequisites of a requirement. Id. at 1-10.

water through February 2012, based on a variety of site-specific factors. These factors, summarized in ESTCP (2011), include (1) complex geologic features, (2) confirmed presence of DNAPLs or other recalcitrant contamination, (3) a combination of the above, (4) excessive cost, (5) physical limitations due to surface structures, and (6) perceived technical limitations of remediation technologies.

Another less common ARAR waiver is the Greater Risk ARAR waiver, which applies if activities taken to meet an ARAR would cause greater harm (like remobilizing DNAPL or dewatering wetlands) than waiving the ARAR; an example is Onondaga Lake, which has elemental mercury as a DNAPL (ESTCP, 2011). The other strategies discussed below have been used much less frequently than waiving an ARAR due to technical impracticability.

Alternate Concentration Limits

Alternate concentration limits (ACLs), which apply at CERCLA and RCRA sites, allow the use of a remediation goal in groundwater that is protective of surface water into which contaminated groundwater discharges, rather than the drinking water standard. The basic concept is that when the groundwater plume enters surface waters, the remedial goal should be consistent with the permitted discharge program governing point source discharges into surface water, as regulated under the Clean Water Act. EPA (2005) clarified ACL policy at sites regulated under CERCLA by identifying a number of considerations. For example, one has to consider whether all plumes discharge to surface water (e.g., a deeper aquifer might not), whether there are potential degradation products between the source and the points of entry (e.g., trichloroethene degrades to vinyl chloride), and whether groundwater can be restored to beneficial use within a reasonable timeframe.

One example where an ACL was adopted is the Naval Surface Warfare Center Ammunition Burning Grounds in Crane, Indiana (ESTCP, 2011). Downgradient of the explosives-contaminated site are several springs that discharge into a nearby creek, which serves as a public water supply 11 miles downstream. Rather than setting the 3 μg/L drinking water standard for the chemical explosive RDX as the site remediation goal, an ACL for RDX of 140 parts per billion (ppb = μg/L) at the spring was set. It was based on ensuring that Indiana Water Quality Standard of 240 ppb would be achieved in the non-potable surface water and the 3 parts per billion RDX would be met at the public water supply. Note that because the ACL is less stringent than the contaminant level that would allow for unlimited use and unrestricted exposure, five-year reviews continue at this site.

Groundwater Management and Reclassification of Groundwater

Groundwater management or containment zones and reclassification of groundwater uses are similar to TI zones, in that they refer to a volume in the saturated zone that is allowed to exceed water quality standards, but the rationale may differ. Most regulators, as a matter of policy, have designated the goal for most groundwater as attaining the highest beneficial use (i.e., use as drinking water), even where the natural or background quality is relatively poor. At some sites, regulators have explicitly recognized that the groundwater in a particular area is unlikely to be used for drinking water now and in the future. A variety of terms have been developed for the affected area in such circumstances, such as plume management zone (Texas), groundwater management zone (Delaware, Illinois, New Hampshire), and containment zone (California Regional Water Quality Control Board – Region 2). An example is the Joliet Army Ammunition Plant, where explosives-contaminated groundwater has a cleanup timeframe of up to 340 years (ESTCP, 2011). There are three groundwater management zones at Joliet, for which the remedial objectives are higher concentrations than elsewhere. Contamination within the groundwater management zones is being addressed through a number of approaches including deed restrictions and continued groundwater and surface water monitoring.

Groundwater reclassification refers to changing the beneficial use of an aquifer such that it is no longer considered a potential source of potable water. At the Altus Air Force Base in Oklahoma, where DNAPL is likely present in fractured bedrock, the groundwater was reclassified to Class III, based primarily on the presence of elevated TDS in the aquifer. This classification will not allow the groundwater to be used for drinking water, although it does permit agricultural and industrial uses. The cleanup objective is to contain the plume, rather than restore it to maximum beneficial use, and the point of compliance is the base boundary (ESTCP, 2011). In New Jersey, there are groundwater classification exemption areas, which are used as an institutional control that provides for the protection of human health as long as the contaminant concentrations in the areas exceed the New Jersey groundwater quality standard.

The ESTCP (2011) report on alternative strategies for site management makes it clear that alternative remedial objectives are not used in many situations where they might apply, despite their attractiveness for dealing with complex sites. TI waivers have been approved by EPA, albeit at only a small percentage of NPL sites (3 percent) for which they could likely be used (ESTCP, 2011). (It should be emphasized that at Superfund facilities

groundwater restoration remains the goal outside of the agreed-upon TI zone.)

A similar initiative directed by the ITRC suggests that many state regulators wish to see a change in the process for overseeing groundwater cleanup activities. ITRC established a committee to develop an integrated strategy for cleanup of groundwater sites impacted by DNAPLs. That committee’s report (ITRC, 2011) includes recommendations on alternative procedures for setting objectives at these sites. For example, the report recommends that “setting of remedial objectives should be based on realistic assumptions and expectations”—a reference to the technical limitations on achieving MCLs. This is a clear indication of the desire to recognize technical limitations before a remedy is selected, equivalent to the use of a TI waiver prior to the completion of a Record of Decision. One of the most significant parts of the ITRC report is the emphasis on greater accountability in setting cleanup objectives. That is, it recommends that interim or functional objectives (see NRC, 2005, for an extensive discussion of “functional” versus “absolute” objectives) be established that can be observed within a 20-year timeframe in order to ensure that potentially responsible parties and engineering firms are held accountable, even where restoration remains the long-term goal. Timeframes beyond 20 years were felt to reduce the likelihood of holding parties accountable for remedial performance.

Despite these initiatives, there is still widespread reluctance by federal and state regulatory agencies to accept the concept of alternative remedial objectives. Reasons for this reluctance are not difficult to comprehend. Notwithstanding EPA’s written guidance, some regulators may seek more aggressive remedies. It is also likely that some regulators inherently make the most protective decision on cleanup objectives and are reluctant to accept the need to revise objectives. For example, it appears that one of the factors that may make issuance of TI waivers difficult is that when such a waiver is granted, an alternative groundwater remediation goal is set for the TI zone in lieu of the unrestricted use level. Thus, the perception is that in order to grant such a waiver, one must “abandon” achieving the unrestricted use of the groundwater in some portion of the aquifer. Some state regulatory bodies argue that as a matter of policy, state non-degradation policies should also be used to require cleanup to drinking water standards of groundwater already degraded and maintain that the goal of restoration is paramount regardless of the technical or economic constraints. Other causes of this reluctance to use alternative strategies on the part of regulatory agencies include, in the Committee’s experience, rotating project managers and lack of incentives to reach a compromise between the potentially responsible parties and the regulators. On the other hand, potentially responsible parties may be reluctant to accept an alternative remedial objective because

of the transaction costs associated with the process, or because of future litigation risks should residual contamination persist (see Chapter 5).

Sustainability as a Cleanup Objective

The historic definition of sustainable is “[d]evelopment that meets the needs of the present without compromising the ability of future generations to meet their own needs” (Brundtland, 1987). According to the Bruntdland report (1987), the most “sustainable” policies address environmental, economic, and social aspects of a problem (the so-called triple bottom-line approach)—a definition much broader than that encompassed by the federal and state hazardous waste laws. If sustainability is to be a remedial goal, this broad policy definition needs to be translated into concrete direction on how to clean up a site “sustainably.”

Incorporating sustainability concepts into remediation decision making is a developing, but still incomplete, practice at EPA and other agencies. EPA, DoD, the states, and others have “green” or sustainable remediation policies (DoD, 2009; Army Corps of Engineers, 2010; EPA, 2008; ITRC, 2011). All ten EPA Regions have adopted Clean and Green policies for contaminated sites, generally with green remediation goals including to minimize total energy use and to reduce, reuse, and recycle materials and wastes (EPA, 2011e). However, “green” remediation and even some of these agency guidance documents that use the word “sustainability” do not include all of the elements of sustainability found in the Brundtland report. For example, EPA’s definition of green remediation is the “practice of considering all environmental effects of remedy implementation and incorporating options to minimize the environmental footprints of cleanup actions” (EPA, 2011e). This is narrower than the concept of sustainable remediation as “balance[ing] outcomes in terms of the environmental, social, and economic elements of sustainable development” (see Table 3-3 below and Bardos et al., 2011; NRC, 2011). In fact, some argue that sustainable decisions should consider community improvements, jobs, and quality of life, and the benefits to the surrounding community (NRC, 2011). Several examples of sustainable remediation that illustrate the range of concepts that can be incorporated are given in Box 3-3.

Each of the Sustainable Remedy Selection environmental factors listed in Table 3-3 (i.e., column 1), and some of the social and economic factors (columns 2 and 3), fit into the standard EPA and state remedy selection criteria. For example, impacts on human health and safety (a social factor), impacts on various environmental media and natural resources, and community involvement can be assessed under existing remedy selection schemes. However, ethical and equity considerations, indirect economic costs and benefits, and employment and capital gain (among others) are