Elements of a Decision Protocol for Source Remediation
One perception of source remediation conveyed to the committee by the Army and other responsible parties is that investments in source depletion technologies have often failed to achieve the desired reductions in risk and/or site care requirements. There is no doubt that investing time and money into remedies without apparent progress creates a scenario that is sure to frustrate all parties. Although part of this challenge arises from the fact that historical releases of DNAPLs and explosives are technically difficult to clean up, the problem is also attributable to how source zones are managed.
The design and implementation of a successful source remediation project involves the iterative characterization of the source zone, development of remediation objectives, and evaluation of technologies, each of which is emphasized in previous chapters of this report. The resultant process is sufficiently complex to warrant a formal protocol to ensure that future projects do not skip essential elements—a need that was explicitly recognized by the Army in requesting this study. This chapter describes the elements of a protocol to assist project managers in designing, implementing, and assessing the effects of source remediation.
A protocol is defined as a strategy and methodology to be followed for accomplishing a stated purpose—in this case, the remediation (through removal, transformation, or isolation) of source material from the subsurface. The elements of a protocol presented in this chapter focus on decision making rather than on how to collect information, which is different from other commonly used protocols that provide extensive appendixes on field-sampling techniques, analytical methods, and data interpretation (e.g., the natural attenuation protocol of
Wiedemeier et al., 1996). Rather, this chapter is intended to help standardize the conceptual process for evaluating source remediation, including data gathering and analysis, setting objectives, and selecting remedial actions. The guidance is general, in that all possibilities are examined and no technology or endpoint is advocated over others.
Decision tools and protocols for general site cleanup date back to the early days of the Comprehensive Environmental Response, Compensation, and Liability Act (CERCLA or Superfund) and the Resource Conservation and Recovery Act (RCRA). For example, the nine criteria of CERCLA discussed in Chapter 4 are meant to enable the remedial project manager to select among various alternative remedies. Since the early 1980s, many detailed cleanup protocols have been created, most of which focus on specific types of contamination problems (for example, underground storage tanks, fuel hydrocarbon sites, LNAPLs) other than the recalcitrant chlorinated solvent sites that led the Army to request this study. There are also numerous protocols for using individual cleanup technologies, such as pump-and-treat, soil vapor extraction, or monitored natural attenuation, as well as for using engineering controls (such as containment) and institutional controls.
Protocols for remediation of DNAPL sources are still under development and are a focus of a number of ongoing studies. For example, the Texas Risk Reduction Program’s NAPL management decision process, on its eighth draft, will be useful for remedy selection. The U.S. Environmental Protection Agency (EPA) recently sponsored a white paper outlining key issues for DNAPL cleanup, authored by an independent panel of scientists (EPA, 2003). The Air Force Center for Environmental Excellence and the Strategic Environmental Research and Development Program (SERDP) of the Department of Defense (DoD) are both supporting research into decision tools for source remediation, notably the SERDP project on Decision Support Systems to Evaluate the Effectiveness and Cost of Source Zone Treatment (Newell, 2003). A review of these ongoing efforts reveals that flexibility to incorporate the diverse conditions that exist at individual sites (natural and anthropogenic) is a common theme in many of the protocols. Furthermore, rigorous predictions of the performance of source remediation technologies are not available, such that the predictive tools under development provide only a general sense of how things work and an ability to make relative comparisons between options (e.g., the LNAPL Dissolution and Transport Screening Tool—Huntley and Beckett, 2002). Finally, none of the current protocols under development outline what to expect in a given setting, from a given technology, and with a specific contaminant. These limitations, as well as the lengthy examination of the necessary elements of a natural attenuation protocol described in NRC (2000), were kept in mind as the committee developed the elements described below.
The decision protocol for source remediation takes the form of a six-step process (Figure 6-1) that includes activities (white boxes), data and information collection (gray boxes), and decision points (gray diamonds). The steps are pre-
sented as sequential; in actuality, however, there may be multiple iterations of each step until a decision can be made to proceed to the next step. As an example, the limits of what can be achieved with proven technologies may require a revisiting of the objectives. One of the distinguishing features of Figure 6-1 is its focus on identifying absolute and functional objectives that are clearly articulated and verifiable (Steps 2 and 3). The attention given to these steps is a reflection of the
observation that the absence of focused objectives can result in solutions that fall short of both expectations and needs.
A second theme identified in Figure 6-1 is managing data gaps and uncertainty (as exemplified by “collect data and refine site conceptual model” shown in gray boxes). As discussed in Chapters 2 and 3, geologic complexities, the subsurface behavior of contaminants, and historical changes in land use contribute to the considerable uncertainty that exists at almost every hazardous waste site with respect to the location and extent of contamination. Thus, the protocol focuses users on recognizing the limitations of their current understanding, on the importance of continually collecting the necessary information to effectively make decisions, and on managing plausible variation from perceived conditions. Although the iteration of data collection activities shown throughout Figure 6-1 may seem administratively burdensome, complex, and time consuming, the committee cannot envision source remediation occurring successfully in its absence. Indeed, the committee’s experience at many dozens of sites is that not making this iteration explicit in the remedy selection process merely adds to transaction costs, as remedies are found to be unsatisfactory to major stakeholders, Records of Decision (RODs) are reopened, new contractors are hired, new studies are conducted, and new remedies are chosen. Such a mode of operation must be more transactionally complex than relying on an iterative approach from the start. A primary goal of this report is to portray iterative site characterization as essential to the success of source remediation, and thus increase its acceptability to regulators and the regulated community.
Although one could potentially limit use of Figure 6-1 to only the most complex DNAPL sites, it was designed to encompass all sites with source zones, regardless of their complexity. The potential technologies considered in Step 4 can include all those discussed in Chapter 5, including presumptive remedies like excavation and physical or hydraulic containment. While the protocol was intended to be applicable to all DNAPL and explosives source sites, this does not mean that aggressive remediation has to occur at every site. Indeed, the degree of complication posed by the protocol is dependent on the desired objectives and the feasibility of finding technologies that can meet them. For example, at the thousands of contaminated dry cleaner sites across the United States, going through the elements of Figure 6-1 could be a one-day exercise, and it could be consistent with state programs that list presumptive remedies. Thus, although the questions in Figure 6-1 should be asked at every site, the amount of effort spent on each step will vary by orders of magnitude depending on the site conditions.
It is important to note that the suggested elements for a source protocol shown in Figure 6-1 are consistent with relevant CERCLA regulations and guidance documents. The suggestions presented herein reflect shifts in emphasis from and logical refinement of currently used approaches, as opposed to wholesale modifications of them. As a general matter, the National Contingency Plan (NCP) contemplates that the development of remedial objectives, the collection
and analysis of pertinent data, and the selection of an appropriate remedy will be accomplished in the context of a remedial investigation/feasibility study (RI/FS). Thus, most of the steps called for in Figure 6-1 should (and permissibly may) take place in that phase of the CERCLA process. At the same time, however, Step 6 in Figure 6-1 is likely to be carried out during the remedial design/remedial action (RD/RA) stage of a CERCLA-based cleanup. Under the NCP, all RD/RA activities are generally expected to conform with the remedy set forth in the Record of Decision (ROD). Nonetheless, the NCP very clearly allows for flexibility during the RD/RA, with respect to important details of remedial activities at sites. Thus, where a remedial action “differs significantly from the remedy selected in the ROD with respect to scope, performance or cost,” the lead agency must publish an explanation of all significant differences that do not “fundamentally alter” the remedy previously described in the ROD. In contrast, where differences in an RD/RA stage remedial action will fundamentally alter the basic features of an earlier-selected remedy, the lead agency is required to follow a prescribed procedure for amending the ROD. These NCP provisions clearly allow for the continuing analysis and supplementary data collection, referred to in Figure 6-1, that will take place during the RD/RA phase of source remediation. The iterative nature of source characterization shown in Figure 6-1 is consistent with current EPA guidance on site characterization (EPA, 2004; http://clu-in.com/download/char/dynwkpln.pdf) and with the Triad approach (EPA, 2001) to environmental data collection.
Not all elements of cleanup (particularly those not specific to source remediation) are discussed in this chapter. For example, how to conduct long-term monitoring or the need for contingency plans are not discussed. This should not be interpreted as suggesting that certain activities be omitted.
Although not explicitly shown in Figure 6-1, early and ongoing involvement of potentially affected parties is a necessary part of source remediation to gain consensus on appropriate actions. Without adequate participation, critical elements of solutions may be missed, a subset of the involved parties may feel that their needs have been ignored, and/or false expectations may develop as to what can be achieved. Furthermore, remedial actions may be viewed as being less aggressive and less costly alternatives that offer advantages to responsible parties without fully protecting human health and the environment. Another concern is that the presence of a source area may affect community property values. For these reasons, a comprehensive source remediation protocol needs to reflect ongoing input from the affected community at all stages. In some cases, no nearby residential or commercial community per se may exist, so flexibility on this issue is warranted.
The following sections describe each of the steps identified in Figure 6-1.
REVIEW EXISTING SITE DATA AND PRELIMINARY SITE CONCEPTUAL MODEL
Whether investigating a new suspected source or a known subsurface source, the first step is always to review existing site data. In the case of a potential new source, historical land use, waste management practices, regional geologic reports, and aerial photos all provide critical initial clues as to the nature and extent of sources. Chapter 3 discusses in greater depth the characteristics of subsurface sites that should be measured at this stage and the necessary tools. The most common scenario will be that the suspected source has been investigated previously and that release-specific documentation including a site conceptual model exists, which should be carefully reviewed.
During the review of existing site data, a key question to ask is: “Are there enough data to determine if a source exists?” If this question cannot be answered in the affirmative, additional data should be collected. If enough data do exist, then it is appropriate to ask whether there is a high probability that a source exists. The intent of this question is to ensure that source remediation does not proceed in the absence of a reasonable basis for expecting a source to be present (as has been observed at some sites).
Building on the above, guidances on data mining, on resolving the existence of a source, and on comprehensive source characterization are critical supplements to the source remediation protocol. This could include, for example, guidance on how much information is needed to understand the hydrogeology controlling contaminant transport and fate at the site, criteria for delineating the source zone, tabulations of common data sources, suggestions for the development of preliminary site conceptual models, lists of common exposure scenarios to consider, and descriptions of risk assessment methodologies.
IDENTIFY ABSOLUTE OBJECTIVES
Identifying absolute objectives is specifically denoted in Figure 6-1 because in many cases examined by the committee, it was not clear that this was done properly or that it was done at all. Although shown in Figure 6-1 as Step 2 because it is critical at that point in the framework, identifying objectives can begin as soon as a site is labeled as a potential concern.
The development of remedial objectives for a site is inherently a social valuation process, to which stakeholders will bring differing (and perhaps irreconcilable) points of view (Presidential/Congressional Commission on Risk Assessment and Management, 1997). A key requirement in this process is to differentiate between absolute objectives (which are not substitutable) and functional objectives (substitutable alternatives to meeting those absolute objectives). Making this distinction requires careful communication during stakeholder discussions. This is especially important where a particular objective, such as attaining
maximum contaminant levels (MCLs) at a particular point in time and space, may be a functional objective for one stakeholder (i.e., the stakeholder’s real objective can be obtained in another way) but an absolute objective for another stakeholder. During these discussions, it must be kept in mind that some stakeholders may have already begun to translate absolute objectives into particular functional objectives and may not even be aware of having done so. In addition, many stakeholder objectives may be independent of the conclusion that there is a source on the site, while others will correspond directly to the identification of a source contributing to ongoing contamination of the groundwater flowing under the site.
It is not the purpose of stakeholder discussions to necessarily achieve consensus on absolute objectives or even necessarily on a common list of absolute objectives for site cleanup. Rather, the process is intended to ensure that the objectives are explicit and to remove confusion regarding the status of objectives for different stakeholders. For example, it needs to be made clear how certain absolute objectives (e.g., protecting human health) will be defined operationally and how progress will be measured. Other sources of common confusion involve the relevant temporal and spatial scales for certain objectives (e.g., calls to protect the environment must be accompanied by some notion of how much to protect and for how long).
Another issue to be clarified as part of the process of setting absolute objectives is the meaning of partial success. For example, if an EPA objective is to get a ROD signed by the end of the fiscal year, and it turns out that this cannot be done, will it be considered a partial success if the ROD is signed in November or the following August? Does it matter if it is done in the next fiscal year, rather than taking five more years? As another example, is opening a river to catch-and-release fishing (but not to catch-and-eat) valued as a partial success, or does only total removal of restrictions on fishing represent meeting the goal? Finally, to the extent possible, trade-offs among objectives for each stakeholder need to be clarified to the extent possible. Typically, those responsible for cleanup costs will have very different trade-offs of performance vs. cost than will those without such responsibility.
In an attempt to illustrate the principles of developing absolute objectives (including the conflicting objectives that often are voiced at initial stakeholder meetings), a theoretical case study is presented in Box 6-1. The point of the case study is to show that the identification and selection of absolute objectives is a multistakeholder exercise that often reveals conflicting as well as similar desires.
IDENTIFY FUNCTIONAL OBJECTIVES AND PERFORMANCE METRICS
As discussed in Chapter 4, the relationship between absolute and functional objectives is not always simple. One functional objective may serve multiple absolute objectives. For example, achieving a particular contaminant concentration at
Fort Alpha, a hypothetical vehicle maintenance facility, began operation in 1943 and ceased active operations in 2001. The primary cause of concern is the engine maintenance center, which used large amounts of chlorinated solvents as degreasers. An estimated 100 55-gallon drums were disposed of in an open-pit during the 1960s. The contamination is known to be a DNAPL consisting of chlorinated solvents and waste oil (by weight, 70% TCE, 10% 1,1,1-TCA, 5% PCE, 15% waste oil and grease). The Army intends to turn over the site, which is unoccupied, to local authorities as a potential industrial park.
The site has two major hydrogeologic settings—6 m of fluvial alluvium overlying a fractured (weathered) shale. The alluvium (Type III) transitions from fine-grained sand (k=10–12 m2, K=10–5 m/s) at the water table to coarse channel sands at the base (k=10–10 m2, K=10–3 m/s) with sparse low permeability silt beds (k=10–15 m2, K=10–8 m/s). The shale (Type V) has horizontal and vertical fractures associated with release of overburden pressure in marine shale, with a bulk permeability of 10–14 m2 (K=10–7 m/s). The fracture porosity in the shale is 0.001 and the matrix porosity is 0.3. Vertical hydraulic gradients between the alluvium and shale are negligible. The depth to groundwater is 2 m below ground surface (bgs), and the average horizontal hydraulic gradient in the alluvium is 0.001. From the top to the bottom of the alluvium, the apparent range of groundwater flow velocities is 1–100 m/yr.
Limited information is available on the DNAPL source zone, which is estimated to be 0.40 hectare (1 acre). The vertical extent of DNAPL source zone is based on soil samples taken 2 m bgs (at the water table) and 8 m bgs (2 m into the shale). There is a plume of contaminated groundwater extending downgradient from the disposal area that contains not only parent chlorinated hydrocarbons, but also degradation products like DCE and vinyl chloride—all of which exceed standards by orders of magnitude 30 m downgradient of the site. Downgradient from the site is a residential neighborhood, Duke Estates, home primarily to persons formerly employed at Fort Alpha and their families. After passing beneath the residential neighborhood, the plume discharges to Halftrack Memorial State Wetlands Reserve 500 m from the source. At this location, TCE concentrations exceed drinking water standards by less than 1 order of magnitude. Absent the influence of the site, the aquifer would be Class I under state groundwater regulations, suitable as a source of drinking water. There is a well field 5 km upgradient of the site serving the regional water company, but no exceedances of standards for site-related contaminants have been observed in this location.
The various stakeholders at the site have come to the first meeting with vastly different priorities for site cleanup, as listed below. This list is not intended to be complete, and other stakeholders with additional interests could be identified as well.
Army Remedial Project Manager (RPM) – Prime responsibility is to terminate Army responsibility for the site. This includes providing for protection of public
health (i.e., no complete exposure pathways to the population at unacceptable concentrations) and protection of the environment (i.e., no demonstrable impact on endangered or threatened species). Cleanup must meet Army standards for transfer to civilian authority. Also responsible for overall costs of cleanup, within context of an annual budgeting process.
State Environmental Agency Site Manager – Prime responsibility is to ensure closure/transfer consistent with state law. Groundwater classification as Class I (suitable for drinking water) is key driver. Thus, per existing state regulations, there is no alternative other than to meet MCLs at all aquifer locations, including source areas. There is flexibility only in the time to achieve the objective.
U.S. EPA Site Manager – Responsible for consistency with the nine Superfund criteria and Applicable or Relevant and Appropriate Requirements (ARARs). For the protection of public health, there can be no complete exposure pathways to the population at unacceptable concentrations. For protection of the environment, there must be no demonstrable impact on endangered or threatened species; EPA also wants to satisfy natural resources trustee. With respect to ARARs, MCLs are not applicable because contaminant concentrations are not exceeded in drinking water supply. The Superfund Amendments and Reauthorization Act (SARA) dictates a preference for source removal, but this is not a decision driver.
State Fish and Wildlife District Officer – Natural Resource Trustee responsible for the wetlands. Protection of the environment is the primary concern, including the reversal of existing adverse effects on wetlands ecosystem and prevention of recurrence. Here, the agency’s criterion of protection embodies permanent removal of threat.
County Commissioners – Concerned about continuing tax base from active site use, avoidance of incurring liability by acquiring site, and employment opportunities. In particular, if transfer requires more than three years, the economic viability of site will be endangered. The commissioners defer to others on criteria for protection of health and environment, meaning that both the electorate in the county and relevant regulators must accept.
Jane Doe (Nearby Resident) – Concerned about protection of family health. She lacks confidence in active remedies; her concern remains as long as contamination in groundwater exceeds MCLs. Not convinced that use of public water supply removes risk. She is also concerned about property values and wants to be able to sell her house for same price as “comps” in nonaffected parts of community.
T. P. Jones (Nearby Farmer) – Concerned about use of aquifer water on farm in the future; potential impacts on certification of produce as “organic.” As long as contaminants are identified in water, his concern remains. He is unable to run the business he wants, or pass on the farm to his grandchildren as a viable enterprise (in his terms).
At the first stakeholders meeting following the identification of a source area at Fort Alpha, a long list of absolute objectives was put forward in discussions among the stakeholders, touching on each stakeholder’s concerns. These included cleanup to background, attainment of MCLs, and prevention of contaminant intrusion into the wetlands, as well as more globally expressed objectives such as protection of human health. Some of the confusion noted in Chapter 4 is experienced during the stakeholder meetings. For example, it is not immediately clear if the call for meeting MCLs is an absolute or functional objective or if it is viewed differently by the federal and state EPA representatives. It is not clear how to measure success of the objective of “giving our grandchildren a clean environment.” Nor are the time elements associated with the absolute objectives clarified.
Through further discussion, the following clarifications were made that allowed the various objectives to be redefined in the following way:
Meeting MCLs. This is an absolute goal only for the state site manager, as the other stakeholders have used it as a metric for public health protection. The EPA site manager is able to determine that it is not “appropriate” under SARA. The state site manager may have flexibility regarding when it is met in order to approve a remedial decision.
No contaminant intrusion into wetlands. This turns out to be both unnecessary in the short term because the Trustee can accept intrusion that does not exceed the “capacity” of a healthy ecosystem. Over the long term, the objective is inadequate because the Trustee cannot consider the wetland protected as long as there is a source. Accordingly, this is not included in the list of absolute objectives for the site.
Having the Army clean up the mess it made. This objective is determined to be equivalent to removal of the source by some stakeholders and to protecting human health and the environment by others. In either case, it is redundant and is not included in the final list of absolute objectives.
Buy-out of homes sitting over the middle of the plume. This is viewed by some stakeholders as one of many ways to remedy economic damage due to presence of groundwater contamination. Similarly, it is one of many ways to prevent adverse effects on health and is agreed to be deferred to the consideration of functional objectives.
Stakeholder discussions led to the following list of noneconomic absolute objectives, with the understanding that not every objective is absolute for each stakeholder, that there remains considerable overlap, and that not every stakeholder’s objectives are viewed as attainable by the others:
Functional objectives that address some these absolute objectives can be defined and are the subject of Box 6-2. Any given functional objective may address a larger or smaller set of these absolute objectives.
a point in time and space may serve to protect human and environmental receptors, to meet an absolute statutory requirement, and to meet the programmatic requirements of an agency. Similarly, a given absolute objective may be capable of translation into alternative functional objectives. For example, prevention of human exposure to site contaminants can be achieved by removing the contaminants or by removing the people from a given area.
The difficult part of setting functional objectives, and identifying alternative functional objectives, is generally the integration of multiple absolute objectives. When a given functional objective (e.g., preventing migration of contaminants off-site) meets multiple absolute objectives (e.g., protecting human health and preventing damage to identified environmental receptors, by interrupting the exposure pathway), alternative functional objectives may be fungible with regard to one absolute objective but not with regard to others. For example, a buy-out of neighbors could protect human health, but not environmental receptors.
In specifying functional objectives, it is important to continue elaborating until a metric can be identified by which it will be possible to measure progress toward achieving the functional objective and, hence, the absolute objective. For example, in the case of a property buy-out that was noted above as a way to interrupt an exposure pathway by moving receptors away from the contaminant, one might need to specify the secondary functional objectives of (1) moving current residents out and (2) preventing anyone else from moving in. The first of these could have as metrics such things as signed purchase contracts for each property and documented vacating of the property. The second might be more complex, involving factors such as destruction of residential structures, backfilling of wells, fencing the property, posting warning notices, establishing deed restrictions applicable to any future transfers of the property, and establishing a procedure for ensuring that these land-use controls remain in effect.
This last point illustrates an aspect of functional objectives that is often given insufficient attention—how well the functional objectives are matched to the temporal dimension of the absolute objectives they are intended to serve. If the characterization of a site suggests that conditions will persist for an extended period (predictions for some sites have extended to many centuries), then either the functional objectives must be stated so as to address that timeframe or it must be made clear to all stakeholders that the absolute objective is only being addressed within a limited period.
As shown in Figure 6-1, prior to determining functional objectives, it should be asked whether there are enough data to make such a determination. If not, then additional data collection and analysis are warranted.
The translation of absolute objectives into functional objectives is illustrated for the hypothetical Fort Alpha hazardous waste site in Box 6-2.
Functional Objectives to Achieve Interruption of the Human and Environmental Exposure Pathways
Most or all of the noneconomic absolute objectives for many of the stakeholders in Box 6-1 can be obtained by any functional objective that interrupts the pathway between the source and various receptors. For example, the absolute objective specified by some of the stakeholders was to protect the public from excessive concentrations of site contaminants by disrupting human exposure pathways (at present, this is only relevant to residents of Duke Estates). This can clearly be accomplished in several ways: moving the humans, decreasing the concentrations to which they might be exposed, interspersing barriers to contaminant transport, or preventing release of contaminants from a confined area.
Barriers to exposure could be implemented at any of several levels and locations, depending upon site conditions. At one extreme, residents could be supplied with personal protective equipment or (as has occurred elsewhere) have vapor strippers installed on their wells. However, there are no wells at present in Duke Estates, and public water supplies are in place.
For Fort Alpha, the barrier approach could address the limited human health risk, but it does not address other absolute objectives (like protecting the environment, specifically the wetlands). Alternative functional objectives, focused on reducing concentrations or transport of contaminants in groundwater, may be capable of addressing both absolute objectives. Any of a number of objectives can be specified for reducing concentrations in the exposure zone (or, alternatively, instituting a barrier to contaminant transport). These could include defining con-
centration limits in the exposure area, specifying a geographic boundary to the contaminant plume (e.g., the property fence line), or, in a more nuanced fashion, limiting the entrance of contaminants to the capacity of the ecosystem to absorb it. Each alternative definition will impose a different set of associated metrics and inference chains in order to establish that (1) the defined functional objective meets the absolute objective and (2) the measured quantity is in fact reflective of the functional objective.
Interestingly, previous cleanup efforts at Fort Alpha illustrate some of the potential problems in clearly specifying functional objectives and metrics. In an attempt to decrease contaminant concentrations under Duke Estates, a pump-and-treat system was installed, and it was carefully documented that several hundred kilograms of solvents were recovered. However, subsequent investigations showed no reductions in plume concentrations under the houses or in the wetlands, the ostensible goals of the technology. Thus, the chain of inference between metric and objective, however reasonable a priori, apparently involved some inappropriate assumptions for the site. The metric of mass removal was deemed to be inappropriate for documenting concentration reduction in this case.
Functional Objectives to Achieve Source Elimination
As can be seen from Box 6-1, only complete elimination of the source area will meet the absolute objectives of Mr. Jones, of the Natural Resources Trustee, and of the state site manager. There are likely significant differences in the expected time to elimination required to fulfill their absolute objectives (with Mr. Jones having the more immediate need). Clearly, functional objectives that address concentrations of contaminants in Duke Estates or the wetlands are not relevant to this absolute objective.
As in the case of interrupting exposure pathways, there are multiple alternatives for specifying functional objectives and metrics in support of the objective of eliminating site contamination. One alternative is to focus on mass removal from the source. A straightforward metric is mass of contaminant removed. It must be remembered, however, that it may be very difficult to relate this objective/metric to other objectives, such as the time during which source material will remain in the subsurface.
An alternative functional objective could be defined in terms of source strength—that is, the ability of the source to continue to contribute to groundwater contamination. This, in turn, could be further specified in terms of downgradient concentrations, some measure of mass flux, or as performance in a series of extraction tests. As in the case of interrupting exposure pathways, it is important that the connection between the stated absolute objective and the actual measurements is well established.
IDENTIFY POTENTIAL TECHNOLOGIES
Once functional objectives have been decided on, the next issue is to identify technologies that are viable means of achieving those objectives (Step 4 in Figure 6-1). As described in Chapter 5, the potential efficacy of a given technology is dependent on the contaminant type, the hydrogeologic setting at the site, and the chosen functional objectives. This suggests that a multidimensional screening approach can be used to identify promising technologies. This is represented by the cube shown in Figure 6-2, which has independent variables of hydrogeologic setting, functional objectives, and source remediation technologies. The details on these three factors are presented in Chapters 2, 4, and 5, respectively. Obviously, there must be sufficient information on these three factors to enable use of the cube. Otherwise, further data collection and analysis should be undertaken prior to using the cube, as shown in Figure 6-1. The dependent output of the cube is a “potential for success” rating of high, medium, low, or unknown. These
ratings could correspond to those given in Table 5-7, such that the entries to that table can be considered as elements of the cube, if the physical objectives used to create Table 5-7—mass removal, concentration reduction, mass flux reduction, reduction of source migration potential, and change in toxicity—are relevant. If not, then other functional objectives can be used. No matter what the chosen objectives on the cube axis are, the cube entries are not necessarily site-specific, but rather are gleaned from the literature, from other case studies, etc.
Use of the cube begins by considering the source contaminant or contaminants. As conceptualized in Figure 6-3 (and as noted in the second column of Table 5-7), contaminant type can constrain the list of applicable technologies. For example, soil vapor extraction is a technology best applied to volatile compounds and not to semivolatile or nonvolatile compounds.
Next, the physical setting is imposed. The efficacy of specific technologies is often highly dependent on the environmental setting. For example, many flushing technologies (e.g., surfactant-enhanced recovery) are not applicable to low-permeability media due to slow rates of fluid delivery and recovery. The output envisioned in Figure 6-4 describes the potential of those technologies appropriate for a given contaminant and in a given setting to achieve different functional objectives. The example in the figure indicates that for both technologies, objective 5 would be unlikely to be achieved, while objective 3 is far more likely to be achieved. What can be achieved provides a rational point for resolving what should be undertaken. In the event that none of the applicable technologies can
achieve a desired objective, then the user must cycle back to Step 2, as indicated in Figure 6-1.
It is important to understand the scope of this exercise. The cube is intended only to identify those technologies that could be used to achieve certain objectives in certain hydrogeologic settings, and thus it is a screening tool rather than a definitive solution. Furthermore, the limitations and caveats associated with the data in Table 5-7 must be kept in mind when using the cube. Not all of the table entries are based on published performance data. If a technology that is not well documented in the scientific and engineering literature is chosen, then the primary functional objective may need to shift to demonstrating an innovative technology as opposed to achieving specific objectives for the site itself. Ideally, only technologies that are rated “high” should be carried further through the evaluation process. As a fallback position, given an inability to achieve a desired outcome, technologies that are rated “medium” can be considered with the recognition that the likelihood of success is less certain.
Returning to our hypothetical case study, an example of using the cube is presented in Box 6-3.
Box 6-1 describes contamination at Fort Alpha as being a chlorinated solvent DNAPL and the hydrogeologic setting as a combination of Type III and Type V. Taking these factors into consideration, the following table was developed from Table 5-7 and screened for options that include a score of “low-high,” “medium-high,” or “high” for Type III and V settings. The table is equivalent to the plane of the cube shown in Figure 6-4 (although the exact table entries do not correspond with circles in Figure 6-4).
A critical feature of the table is that there appear to be no applicable technologies for achieving any of the five physical objectives for the fractured shale other than physical containment. Due to the adjacency of the shale to the alluvium, not treating the shale will diminish the likelihood of achieving the noted local aqueous concentration reduction and/or mass flux reduction in overlying adjacent alluvium. One notable exception to this is the soil mixing/chemical reduction technology. In this case, clay is usually used with the iron to aid in delivering the iron and reducing the horsepower required to mix the soil. Use of the soil mixing/chemical reduction
technology might cap the contaminated shale with low-permeability reactive media that could prevent back diffusion of contaminants from the shale into the alluvium.
Next, the functional objectives identified in Box 6-2 are compared to what can be achieved. None of the identified options will “eliminate” the source because at a minimum, some source material will be left in the fractured shale. This objective may need to be eliminated or modified to reflect mass removal to the extent practical using best available technology. The objective of interrupting human exposure pathways was discussed in Box 6-2 as being achievable in several ways: moving the humans, decreasing the concentrations to which they might be exposed, interspersing barriers to contaminant transport, or preventing release of contaminants from a confined area. While none of the objectives used in the cube and Table 5-7 precisely match the two absolute objectives at Fort Alpha, mass removal and mass flux reduction were interpreted to be closest to source elimination and elimination of the human exposure pathway, respectively. At least for the portion of the subsurface that is Type III, this narrows the list of technologies to excavation, surfactant flushing, conductive heating, and chemical reduction.
SELECT AMONG TECHNOLOGIES AND REFINE METRICS
Step 4 of the protocol may have revealed three or four different remedies or combinations of remedies that have the potential to succeed in meeting the functional objectives at a given site. For example, one option might be containment coupled with institutional controls, while another might be enhanced bioremediation followed by monitored natural attenuation. Given the suite of promising technologies revealed by the cube, the next step is to conduct site-specific evaluations of these options so that an informed decision on which technology to choose can be made. This is essentially a data collection, data analysis, and modeling exercise to better determine whether the various technologies will work at a given site. In addition to technical information, site-specific evaluations commonly involve estimates of cost, time to complete, endpoints, vendor performance, and side effects (e.g., noise, air emissions, or vehicle traffic). Depending on the scale of the exercise, this may be akin to CERCLA Feasibility Studies and RCRA Corrective Measure Studies.
A potential pitfall at this point is for the “best available technology” to be chosen without rigorously determining whether it will work at the site or questioning whether the absolute objectives will be addressed. Decisions of this nature typically reflect a desire to do the “best thing.” Unfortunately, actions that are not tied closely to functional objectives often fall short of expectation and/or needs. This can lead to the all-too-common scenario in which multiple technologies are sequentially deployed with no sense of meaningful progress toward closure. Thus, the site-specific data collection exercise should be considered an investment of time and resources that can prevent future problems associated with a faulty or inappropriate technology selection.
Once site-specific data on the technologies revealed by the cube have been collected and analyzed, a decision has to be made about which remedy to pursue (assuming that more than one possibility exists after Step 4). One approach to reaching this decision in a systematic way is by constructing a matrix of objectives vs. candidate technologies, as shown in Table 6-1. Technologies that were identified using the cube are listed in the vertical column, and functional objectives are listed across the top of the matrix. Each intersection represents the ability of a particular technology to meet a particular objective. The entries used in the matrix should be informed primarily by the site-specific evaluation exercise discussed above. All of the entries in a given row should be considered to give a total rating for each technology (although how this should be done—that is, qualitatively, via simple addition, or using weighted combination rules—is not specified here). The two or three highest-rated technologies are considered the leading candidates for the remedial action. Creation and use of this matrix should involve all stakeholders and it should be well documented.
The listed objectives in Table 6-1 should be those used in the cube as well as others that could not be meaningfully included in the cube but which are refined
TABLE 6-1 Source Remediation Alternatives Analysis Matrix
to the site-specific situation and/or are more practical. These might include minimizing life cycle costs, complying with regulations, maintaining positive relationships, eliminating liability, and ensuring worker safety. If a functional objective is found to have similar ratings for all remedial actions under consideration, then it is likely to have little impact on the decision to be made and might be removed from the matrix. Box 6-4 illustrates the use of such a site-specific matrix for the Fort Alpha site, using qualitative (low, medium, high), narrative ratings rather than numerical scores.
One of the concerns that could be taken under consideration when using Table 6-1 is how robust technologies are during implementation. Robustness refers to the ability of a technology to operate properly despite unfavorable or unexpected conditions. This requires understanding many complicating factors, such as the potential impacts of site heterogeneity, mixtures of contaminants, the potential formation of transformation byproducts, variability with climate and/or groundwater temperature, and the potential adverse effects of other remedial actions or land-use changes.
As for any complicated technical issue, making and implementing good decisions about source remediation requires a considerable amount of training, experience, and expertise. These are also essential to ensure that the protocol elements recommended in this report are implemented properly. To maximize success, it is important that the chosen vendors have documented experience using the specific technologies being considered, ideally in hydrogeologic settings similar to the site being studied. Qualifications for the major stakeholder groups are described in more detail in Box 6-5.
One possible outcome of Step 6 is that none of the specific technologies will remain viable, given the objectives of Table 6-1 and site-specific conditions. In this case, it is necessary to revisit the absolute and functional objectives (Steps 2 and 3) and subsequently develop new options. As a general matter, EPA regulations allow for changes in remedial objectives, although this requires that site managers engage the relevant stakeholders to ensure that the new objective is
Four technologies were brought forward for treating contamination in the overburden at Fort Alpha, while one (containment) was deemed to be potentially applicable for the fractured shale bedrock. In addition to the functional objectives of mass removal and mass flux reduction chosen for the site, the Army also desires to have a safe working environment, to be in compliance with all regulations, to have the community accept the remedial action, and to minimize the life cycle cost of the remedy. All of these objectives appear in the matrix below, which is filled in
with site-specific data and information. Note that the table entries are strictly hypothetical and were created using the information gleaned from Table 5-7 and the committee’s best professional judgment.
A qualitative reading of this matrix leads the Army and its advisors to believe that excavation and soil mixing/chemical reduction appear to be most likely to meet the Army’s list of objectives. Once all stakeholders agree to this evaluation, those two technologies should then be carried further through a highly detailed feasibility analysis. A laboratory evaluation and a field pilot of the soil mixing/ chemical reduction technology should seriously be considered if that technology is likely to be the Army’s final selection. Because excavation is a mature technology, there would be no need for a field pilot project if it were the final selection.
There is a substantial body of knowledge needed to conduct successful source remediation. The needed expertise differs depending on an individual’s role—be it regulator, responsible party, remediation consultant, or community-based organization.
Responsible parties need to be able to evaluate the work of remediation consultants and make remediation decisions based on the consultant’s work. These individuals should have sufficient technical background and experience to actively manage their consultants and to negotiate responsibly with environmental regulators. Responsible party representatives should ideally have a relevant technical degree and understand the fundamentals of risk evaluation and community involvement.
Contractors and Consultants
Remediation companies should have an expert in source remediation with a thorough understanding of the relevant science and engineering principles. Experience in related technologies and similar site conditions is invaluable.
State and Federal Regulators
State and federal regulators are charged with evaluating the merits of the various remediation proposals they receive and making judgments on whether source remediation proposals have sufficient technical justification. Because these regulators must be able to understand a wide variety of technical information, desirable skills include knowledge of environmental engineering, geology, chemistry, and biology principles, as well as how to use conceptual models, mathematical tools and models to estimate contaminant movement and degradation, and fundamentals of risk evaluation.
The members of communities affected by contaminated sites where source remediation is being considered should be included in the decision-making process as early as possible and should have the resources necessary to participate in this process. Potential training might include an introduction to contaminant behavior in the environment, basic risk assessment, a review of strengths and weaknesses of source remediation technologies, a review of how site conditions and contaminants affect source remediation, and a discussion of verifying the effectiveness of the various source remediation methods.
SOURCE: Adapted from NRC (2000)
acceptable and compatible with the expected future land use at the site. One option that is sometimes pursued at DNAPL sites in heterogeneous (Type III and V) media is a determination of Technologically Impracticability (TI). TI waivers, which are granted for groundwater contamination only, result in the selection of a less strict remedial objective (such as a higher allowable contaminant concentration). However, in order to be granted a TI waiver, the responsible party must provide their best estimation of a technically practicable alternative remedial strategy that will prevent migration of contamination beyond the source zone (see EPA, 1993, and NRC, 2003 for greater explanation of technical impracticability).
DESIGN AND IMPLEMENT CHOSEN TECHNOLOGY
In large part, the design and implementation of chosen remedies follow standard engineering practice, accompanied by essential administrative elements such as qualified staff, sufficient funds, and appropriate schedules. In both design and construction, it is advisable to be as adaptive and flexible as possible, which is the premise of the observational approach developed by Karl Terzaghi (e.g., Bjerrum et al., 1960) and Ralph Peck (e.g., Peck, 1962, 1969, 1980). This suggests that remediation systems should be designed and built per known and predicted future conditions, while at the same time anticipating plausible variations in site conditions and having contingency plans in place. This is similar to the adaptive site management concept espoused in NRC (2003), which suggests that higher risk, less certain remedies be explored in parallel with the chosen remedy in the event that the remedy fails to achieve cleanup objectives.
It is important to ask prior to designing a remedy whether there are enough data to design and implement the remedy. Further information is frequently needed to develop specifications, verify the extent of the source zone, and/or optimize treatment processes, similar to the conducting of a treatability study under CERCLA. Under select circumstances, this activity may include extensive field investigations and/or implementation of a pilot system or “first module.”
Once a remedy has been implemented, the monitoring data that accumulate will eventually be used to determine whether a remedy has performed as expected and whether objectives have been met. These monitoring data should provide the agreed upon scientific evidence needed to determine whether the remedy is meeting the objectives. Figure 6-1 highlights the question “is there sufficient information to resolve if the objectives have been achieved?” because this is a test of the adequacy of the monitoring system. If there is insufficient information after the remedy is allowed to operate for the prescribed period of time, then the monitoring program must be redesigned to allow the correct type of information to be collected. (This generally should not be a problem if practical performance metrics were identified in the previous step.) A second interpretation of the question involves how much time has passed since implementation of the remedy. Although it is highly specific to the chosen technology (and ranges from weeks to
decades), sufficient time must have passed before the remedy’s efficacy can be judged. For example, in the case of bioremediation, enough time must pass for the microorganisms to acclimate to changes in redox conditions and seasonal conditions (in temperature, for example). Aggressive source remediation strategies like in situ chemical oxidation produce changes within weeks. However, in most cases additional monitoring (on the order of months) is needed to determine the permanence of the result and the potential for rebound. Years might be required for hydraulic containment and air sparging.
If there is sufficient information, then the final step is to determine whether the absolute and functional objectives have been achieved. If the objectives have not been achieved, then it may be necessary to repeat the steps outlined in the above text and in Figure 6-1. The likelihood of this undesirable setback can be greatly reduced by employing a protocol that addresses the critical elements outlined herein.
* * *
As mentioned previously, the elements of a source remediation protocol described above should be applied to all current hazardous waste sites, regardless of their state of maturity. The amount of time spent at each step will be a reflection of the sufficiency of the information that has been acquired to date, such that those sites where source characterization has occurred commensurate with the chosen objectives and remedy can proceed rapidly through the protocol.
Documentation is not discussed in a separate section in this chapter. However, good documentation is an essential part of the decision-making and remedy selection process. The performance of a remedy can be properly evaluated only in relation to the objectives that were set before it was implemented. Likewise, the quality of decisions can be fairly evaluated in retrospect only if the context in which they were made is preserved. Therefore, each stage of discussions and each selection—whether of objectives or of technologies—should be documented along with the reasons that the selections were made. Establishing a central and complete record of the performance of the Army’s source zone remedies vs. their objectives would be very helpful for understanding how source zone technologies perform in a wide range of settings, and it would be a good reference for Army program managers faced with new source zone remediation decisions. When a remedial action is evaluated, it would be desirable for its degree of success to be evaluated by a party with no stake in the outcome.
CONCLUSIONS AND RECOMMENDATIONS
The six critical elements of a source remediation protocol are (1) review of existing site data, (2) identification of absolute objectives, (3) identification of functional objectives and metrics, (4) elucidation of potential technologies given
site hydrogeologic and contaminant characteristics, (5) selection of an appropriate technology, and (6) design and implementation of the chosen technology. Site-specific data collection informs each step of the process and is used to refine the site conceptual model. If all of these steps are not included, source remediation at an individual site will have a low probability of success.
The Army should develop and use a detailed protocol consistent with the elements prescribed in this chapter. A protocol specific to source zones is needed to aid stakeholders in optimizing the benefits derived from investments in remediating source zones. The key attributes that need to be addressed are pursuing actions that effect intended changes, understanding the extent to which objectives are attainable, and being able to measure progress toward desired objectives. The protocol will need to be integrated into the existing remedy selection frameworks used by the Army at individual sites, including Superfund, RCRA, relevant state laws, or the Base Realignment and Closure program.
Improved technology transfer and guidance on DNAPL source remediation technologies are desirable. Guidance is needed to help responsible parties determine whether source remediation is appropriate for the sites under consideration and to aid in the selection of technical approaches that are most appropriate for the site-specific conditions and remedial action objectives. Army personnel should be thoroughly trained in the use of this guidance. Particular attention should be paid to justifying and documenting remedial decisions as they are made, establishing success metrics at the time of remedy selection, and explicitly documenting the degree of success or failure of remedies by using those preestablished metrics.
Involvement of potentially affected parties is essential to the success of source remediation. Stakeholder participation is needed to better understand the range of absolute objectives at a given site, to develop functional objectives, and to gain consensus on appropriate actions. Without adequate public participation, critical elements of solutions may be missed, a subset of the involved parties may feel that their needs have been ignored, and/or false expectations may develop as to what can be achieved. As for all relevant stakeholders, knowledge acquisition by the public is essential to making decisions about source remediation.
One of the goals of this study was to be able to make definitive statements about the future use of source removal as a cleanup strategy. An important conclusion that can be made from reviewing source zone remediation attempts to date is that the data are inadequate to determine how effective most technologies will be in anything except the simpler hydrogeologic settings. (Indeed, many entries in Table 5-7 were based on the committee’s best professional judgment in lieu of having relevant data from field studies.) Furthermore, it is unlikely that
available source remediation technologies will work in the most hydrogeologically complex settings such as karst.
The committee believes that by following the elements of a source remediation protocol illustrated in Figure 6-1, project managers will be able to make critical decisions regarding whether and how to attempt source remediation and thereby accomplish a more beneficial distribution of resources. It is evident from the review of source zone remediation projects at Army facilities and elsewhere that the steps presented in Figure 6-1—determining whether a source exists; developing clear absolute and functional objectives and their metrics; selecting, designing, and implementing a technology; and collecting data to support all these decisions—have seldom been conducted in the manner described in this report. The efforts of potentially responsible parties to date suggest that in some cases, source remediation technologies are being prematurely scaled up at poorly characterized sites, at sites where there is known complex hydrogeology, and at sites where there is no clear reason for proceeding with the project.
Finally, Chapter 5 suggests that several technologies show enough promise, in terms of demonstrated mass removal and concentration reduction in monitoring wells, to warrant further investigation to determine their long-term effects on water quality—especially if objectives other than MCLs, such as mass flux reduction, become more prevalent. For almost all of the technologies discussed, their effectiveness is more uncertain in the more complex hydrogeologic settings. Thus, future work should attempt to determine the full range of conditions under which these technologies can be successfully applied, and to better understand how mass removal via these technologies affects water quality.
Bjerrum, L., A. Casagrande, R. B. Peck, and A. W. Skempton (eds.). 1960. From theory to practice in soil mechanics: selections from the writings of Karel Terzaghi. New York: Wiley.
Environmental Protection Agency (EPA). 1993. Guidance for evaluating the technical impracticability of ground-water restoration. OSWER Dir. No. 9234.2-25. Washington, DC: EPA.
EPA. 2001. Using the Triad approach to improve the cost-effectiveness of hazardous waste site cleanups. EPA-542-R-01-016. Washington, DC: EPA.
EPA. 2003. The DNAPL Remediation Challenge: Is there a Case for Source Depletion? EPA 600/R-03/143. Washington, DC: EPA Office of Research and Development.
EPA. 2004. Improving Sampling, Analysis, and Data Management for Site Investigation and Cleanup. EPA 542-F-04-001a. Washington, DC: EPA Office of Solid Waste and Emergency Response.
Huntley, D., and G. D. Beckett. 2002. Evaluation of Hydrocarbon Removal from Source Zones and its Effect on Dissolved Plume Longevity and Concentration. Publication 4715. Washington, DC: American Petroleum Institute.
NRC. 1994. Alternatives for Ground Water Cleanup. Washington, DC: National Academy Press.
NRC. 1997. Innovations in Ground Water and Soil Cleanup: From Concept to Commercialization. Washington, DC: National Academy Press.
NRC. 1999. Groundwater Soil Cleanup. Washington, DC: National Academy Press.
NRC. 2000. Natural Attenuation for Groundwater Remediation. Washington, DC: National Academy Press.
NRC. 2003. Environmental Cleanup at Navy Facilities: Adaptive Site Management. Washington, DC. National Academies Press.
Newell, C. 2003. Presentation to the NRC Committee on Source Removal of Contaminants in the Subsurface. April 14, 2003. Washington, DC.
Peck, R. B. 1962. Art and Science in Subsurface Engineering. Geotechnique 12(1):60–66.
Peck, R. B. 1969. Advantages and limitations of the observational method in applied soil mechanics, Ninth Rankin Lecture. Geotechnique 19(2):171–187.
Peck, R. B. 1980. Where has all the judgment gone? The Fifth Laurits Bjerrum Memorial Lecture, Norwegian Geotechnical, Oslo, Norway.
Presidential/Congressional Commission on Risk Assessment and Risk Management. 1997. Framework for environmental health risk management. Volumes 1, 2. Washington, DC: U.S. Government Printing Office.
Wiedemeier, T. H., M. A. Swanson, D. E. Moutoux, E. Kinzie Gordon, J. T. Wilson, B. H. Wilson, D. H. Kampbell, J. E. Hansen, P. Haas, and F. H. Chapelle. 1996. Technical protocol for evaluating natural attenuation of chlorinated solvents in groundwater. Report prepared for Air Force Center for Environmental Excellence.