Environmental Cleanup at Navy Facilities: Adaptive Site Management (2003)

The Navy’s request for guidance on its environmental cleanup program was prompted in part by the ineffectiveness of current remedies to meet cleanup goals at their major remaining hazardous waste sites, particularly those high-risk sites contaminated with recalcitrant chlorinated solvents and metals, often in complex hydrogeologic or sediment settings. Remediating and reaching closure for these types of sites has proved to be elusive in the context of current technologies. In addition to the ineffectiveness of many remedies, the Navy is also struggling with how to balance and meet different remediation goals, such as risk reduction, attainment of drinking water standards, and complete removal of the source of contamination. This chapter first explores these two basic problems (the multiobjective nature of cleanup and the ineffectiveness of current remedies to meet cleanup goals) and then introduces adaptive site management—an approach that can address these problems while encompassing all stages of cleanup.

Contaminated sites can pose multiple hazards to human and ecological health, natural resources, and the economic and social welfare of surrounding communities. In a similar vein, the objectives for site cleanup and restoration are multidimensional and often evolve over time. The eight key objectives are:

1. To protect the health and safety of those on the site and in surrounding communities,

2. To ensure the ecological viability and health of native plants and animals, and migratory species,

3. To protect and restore natural land and water resources,

4. To promote positive economic value and development in the area of the site,

5. To comply with all applicable laws and regulations governing the site and the cleanup process,

6. To promote positive participation and communication with the local community and other affected stakeholders,

7. To advance the understanding of site contamination and cleanup processes (technical, managerial and social), and

8. To accomplish each of these objectives in an affordable, cost-effective, and efficient manner.

These objectives are usually pursued with different emphasis and urgency in different phases of a site cleanup effort. Following site discovery, the first priority is to eliminate immediate threats to human health and safety. Open contamination is enclosed, leaks are plugged, and, if necessary, local residents are switched to alternative sources of drinking water and other protective measures are implemented. Similarly, acute risks to wildlife and aquatic species are controlled to eliminate fish kills, animal poisonings, and other effects that could threaten the viability of ecosystem populations on or near the site. Virtually all Department of Defense (DoD) and other federal sites in the United States have passed beyond this initial phase of site discovery and “emergency response.” Following the control of immediate site hazards, cleanup and management can emphasize different remediation goals and objectives. A broad range of operational objectives have evolved over the last 20 years, from complete soil, aquifer, or sediment restoration to use of a technology-based approach to goals based on minimizing long-term risk to humans and the environment (“risk-based” objectives).

The objectives listed above are closely related to the set of nine criteria established in the National Contingency Plan (NCP) for the evaluation of a proposed remediation plan.¹ Thus, for example, the first NCP criterion of overall protection of human health and the environment is

divided into two separate objectives—one for human health (objective 1 above) and one for ecosystem protection (objective 2 above)—because the steps needed to pursue these objectives are not always fully coincident.

The NCP criterion for compliance with the chemical-specific standards that are considered the statutorily required “applicable or relevant and appropriate requirements” (ARARs) is equivalent to this report’s fifth objective for regulatory compliance with Superfund and applicable state requirements. For contaminants in groundwater, a typical ARAR would be the maximum contaminant level (MCL) for that compound, if one exists. Some states may adopt a “complete restoration” and thus more stringent goal as the site-specific cleanup objective, with which the Navy would need to comply. Superfund and some state regulatory programs allow nonresidential land use assumptions to be considered in the selection of cleanup levels and remedies, so long as selected remedies are protective of human health and the environment.

The three NCP criteria of long-term effectiveness and permanence, short-term effectiveness, and implementability are not specifically noted in the list on page 2 because these features of a remediation plan are all essential to ensure that the other objectives are met. Similarly, reduction of toxicity, mobility, or volume through the use of treatment is the operational objective of a site cleanup needed to accomplish the broader objectives, and is addressed in a subsequent discussion.

The seventh NCP criterion regarding cost is equivalent to this report’s eighth and final objective, because it constrains the extent to which all other objectives can be met. Cost minimization is a key objective in any public or private endeavor, although the weight placed on cost depends on the relevant statute and site-specific factors (EPA, 1996a, 1997a). Given the long-term requirements of site cleanup and stewardship at many sites, estimating costs over the full life cycle of a project is difficult. Approaches that appear cost-effective because of lower capital and initial operating costs may in the long term be more costly, especially if unanticipated problems arise in remediation performance and/or site conditions. Better anticipation of such problems, both initially and through ongoing data collection and evaluation, and ensuring that flexibility is maintained for improving or changing remediation technologies when needed, are key elements of the adaptive site management approach proposed later in this chapter. As discussed in the recent Guidance for Optimizing Remedial Action Operation (RAO) report for the Navy (NAVFAC, 2001), careful assessment of operation and maintenance costs for site cleanup plans can reveal many opportunities for cost

reduction (see Box 2-1).

The eighth and ninth criteria of the NCP (state acceptance and community acceptance) are related to this report’s fifth objective for regulatory compliance and sixth objective for positive participation and communication with the local community and other affected stakeholders. Positive participation includes community involvement in the development of remediation proposals rather than implying (as “community acceptance” does) that the community be involved only after remedial plans are proposed. As recognized by EPA (1999a, 2000a), effective public participation and input into the planning of a site cleanup are both a means—to ensure that the remediation plan can be implemented without costly delays and conflict—and an end—because public participation is a core value of a democratic society. Promoting participation and communication with the local community and other affected stakeholders applies to all aspects of a military site’s operation, but it is especially important in dealing with health and safety risks to the public, where trust is easy to lose, but very difficult to regain (Slovic, 1993).

Our list of objectives includes three that are not explicitly mentioned in the NCP remedy criteria: to protect and restore natural land and water resources (objective 3 above), to promote positive economic value and development in the area of the site (objective 4 above), and to advance site-specific and more general scientific knowledge (objective 7 above). The inclusion of objectives 3 and 4 is based on the fundamental importance of these issues in environmental and economic policy and on the committee’s professional experience as to what is important to states and communities. These objectives are especially likely to arise as a key component of long-term site stewardship and management efforts, once the more immediate threats to health and safety are addressed. The restoration of land and water resources and the return of economic value may or may not be linked for a given site. For some uses, ensuring that there are no (or minimal) risks to health and safety may be sufficient to allow surrounding economic development (including use of lands for recreational or species preservation purposes, if this is the locally targeted objective) to proceed, even though some land and water (especially groundwater) contamination remains. In other locations, planned uses may dictate a more complete cleanup. When site cleanup is critical to an economic or community development plan for a region, strong community and political pressure will be brought to bear both to identify cleanup criteria that can be met in a timely manner and to proceed with the needed effort to reach this objective.

Our seventh objective of advancing knowledge during a site cleanup effort—both knowledge of the site itself and broader insights applicable to other sites—is usually secondary, and as a practical matter may be hard to justify to site managers and the public alike. However, because the science of cleaning up hazardous waste sites is often insufficient to attain even risk-based remediation goals, advancing scientific knowledge must be a component of site remediation. That is, such learning is essential if the other cleanup objectives are to be met in an effective manner. Although scientific study cannot be the principal driver for site cleanup (taking precedence over essential health, safety, and economic objectives), failure to take advantage of opportunities to use data and experiences acquired as part of a cleanup program to enlighten and guide subsequent efforts is in itself wasteful and dooms many of these later efforts to repeat mistakes that could otherwise have been avoided. Indeed, for responsible parties with large numbers of hazardous waste sites, the benefits that accrue from scientific study can be captured by using what is learned in one place at other sites and in future decisions. More focused and explicit building, cataloging, and transmission of knowledge

during remedial investigation, remedy implementation, and monitoring is a key feature of the proposed adaptive site management process.

Given the close correspondence and dependence of the objectives put forth here with those identified in the NCP for remedial selection, the Navy and other agencies can view this charge for ongoing management as fully consistent with the existing directives and goals.

The eight broad objectives discussed above can help guide the overall context and goals of a site remediation plan. However, it can be difficult to translate these into specific programs and activities for site cleanup (especially all at once). To assist in this effort, two more specific site cleanup metrics, contaminant mass removal and risk reduction, are often used to define the specific operational objectives of a remediation program. Like the eight broader objectives identified above, these two metrics are consistent with previous NCP guidelines (i.e., for the reduction of toxicity, mobility, or volume) and with well-established procedures already used by the military and other federal agencies to track and evaluate cleanup. These specific operational objectives can promote the broader goals of site cleanup to different degrees. For example, contaminant mass reduction may (in some cases) be especially important for achieving objectives 3 and 4 (natural resource protection and economic development), with risk reduction being central to the first two objectives (protecting human and ecological health and safety).

Evaluating the potential for risk reduction—central to a risk-based approach to site cleanup—has been used with increasing frequency in recent years. This approach defines the objectives for site cleanup as solely, or at least principally, to minimize human health and/or ecological risk (NRC, 1999a). Although socioeconomic impacts and risks to community welfare are sometimes considered in a broader framing of risk issues, they are rarely included in a formal risk assessment.² Depending on the current or potential hazard to human and/or ecological health and safety, a risk-based approach may lead to full-scale remedial activities (e.g., complete removal of the contaminant source); to more limited onsite engineering and control activities (e.g., containment measures); or

even to no onsite remediation (e.g., use restrictions and other institutional controls). Thus, risk-based approaches as defined do not place inherent value in soil and groundwater resources, unless human or ecological health is directly threatened by contamination of those resources. As a consequence, risk-based approaches are more likely than strategies aimed at natural resource restoration to result in remedies that leave contamination in place.

Box 2-2 describes risk-based approaches, which have gained growing acceptance as a basis for site management decisions despite controversy surrounding the risk assessment process. As noted in the box, a broader definition of risk, combined with effective community input and communication, can help to ensure that risk assessments are appropriately structured and implemented to include the key values and concerns of the affected parties at a site.

Although risk reduction at a site is always sought, source removal (also called mass removal) at the site can be an objective in and of itself. This is because a site that contains significant quantities of remnant contamination may require continued limitations on human use or ecological function, leading to a loss of natural resource value and economic benefit. Surrounding property values and local or regional development may be impaired by the presence and stigma of remaining contamination and perceived risks, even if the actual risks of exposure have been minimized (e.g., Edelstein, 1988; Zeiss and Atwater, 1991; Gregory et al., 1995). Furthermore, breaches of the containment or loss of institutional controls could lead to actual exposures and risks for future generations.

It is sometimes the case that technologies that achieve some (or even a high degree of) mass removal may have little effect on exposure concentrations. Aggressive mass removal can even lead to increased pollutant release and mobilization in the surrounding environment, at least for the short term. This concern is especially important when considering large-scale excavation of contaminated soils or active dredging of sediments.

Although risk reduction and mass removal are not the only targets for hazardous waste cleanup, they are common operational objectives; thus, their relationship is explored in greater detail. There are distinct tradeoffs between different treatment strategies in terms of meeting mass removal and risk reduction goals of cleanup over time. Figure 2-1 schematically shows both the contaminant mass removal (A) and the exposure or potential risk (B) using alternative cleanup and management methods over time. Two major types of remediation strategies are illustrated in Figure 2-1. The first type, designated as an “M” strategy, seeks

first to remove contaminant mass. M1 represents an in situ mass removal strategy, which is typified by an asymptotically diminishing capture, since the last portions of remaining mass are often difficult to access and remove and may thus remain in place (Figure 2-1A). As shown in Figure 2-1B, M1 also displays a limited reduction in exposure concentrations and potential risk until a significant portion of the onsite mass is removed. The M1 curve represents the known behavior of certain in situ technologies, such as chemical oxidation or active bioremediation. Vapor extraction and conventional pump-and-treat might also yield results of this type of curve in the case where much (even if not all) of the pollutant mass is found in, or is readily transferred to, the captured fluid. Strategy M2 represents a more aggressive mass removal strategy, such as soil excavation or sediment dredging, which achieves results over a shorter time period (Figure 2-1A) but which might lead to short-term increases in exposure and risk during the period of implementation (Figure 2-1B).

The second type of remediation strategy, indicated with an “E,” places first priority on reducing exposure to contamination. E1 represents an approach like plume containment, reactive barrier walls, or natural attenuation where the contaminant source zone is not targeted and the focus is on exposure reduction at some compliance point. E2 represents a pathway intervention strategy that would be implemented through institutional controls or onsite containment leaving the bulk of the contamination in place. The dotted upward arrow for E2 in Figure 2-1B signifies the possibility that the remaining onsite contamination could become exposed and impose a potential risk in the future, if the containment is breached or the institutional control is lost.

Figure 2-2 combines the progress in mass removal and risk reduction into a single, multiobjective graph. With this representation, the origin represents the starting point of site cleanup, when there is significant contaminant mass present at the site and affected populations are subject to significant exposure and potential risk. Progress in achieving cleanup and restoration is shown by moving along one of the paths over time toward the upper right-hand corner (i.e., total risk reduction and complete mass removal). Although the ultimate goal of cleanup and restoration is to move to this point in as rapid and efficient a manner as possible, this is not always feasible. Indeed, in many cases the costs are prohibitive, and the objectives of complete mass removal and/or exposure and risk elimination may simply be unachievable with current technology and policy options.

As discussed in Chapter 3, visualizations like Figures 2-1 and 2-2

FIGURE 2-2 A multiobjective representation of alternative remediation and site management strategies in terms of contaminant mass removal and exposure or risk reduction. This graph combines Figures 2-1 A and B. The question mark refers to the possibility that containment or an institutional control fails, leading to a sudden change in exposure and potential risk.

can be used as a means for assessing and tracking the effectiveness of facility management options as a program for site restoration and stewardship evolves over time. By collecting the data and information necessary to record progress to date and by predicting (using mathematical models) the possible future outcomes for the objectives displayed in these figures, a more coherent and responsive effort can be planned and executed for the adaptive site management program recommended in this study.

Adaptive site management is needed not only to handle multiple and sometimes conflicting objectives, but also to provide flexibility when

cleanup remedies are not entirely effective. Particularly at sites characterized by complex hydrogeology and contaminated with recalcitrant hydrophobic compounds and metals, the potential need for changing a remedy over time can be high. The following section describes conditions typically seen at complex hazardous waste sites in which the chosen remedy has been unsuccessful in meeting cleanup goals.

Many methods for remediation of contaminated soil, sediments, and groundwater are characterized by an initial phase of relatively high effectiveness, followed by a prolonged period of much lower effectiveness. An obvious reason why mass removal rates of a remedial system may decrease over time is that contaminant mass is depleted in the vicinity of the extraction or treatment points. However, observing this behavior does not mean that complete mass removal across the entire site has been achieved (unless the initial contaminant mass is known with a high degree of certainty, which is highly unlikely). The more likely causes of decreased mass removal rates and hence decreased remedy effectiveness over time have been well documented for pump-and-treat systems (Mackay and Cherry, 1989; NRC, 1994) and soil vapor extraction (Travis and Macinnis, 1992). These technical factors are summarized below and include geological heterogeneity, flow heterogeneity, slow desorption, slow dissolution from nonaqueous phase liquids (NAPLs), erosion–deposition processes in contaminated sediments, and thresholds for microbial degradation. Although some of the factors described below apply across the range of contaminated media (e.g., slow desorption), others (e.g., geological and flow heterogeneity) apply only to in situ remediation of soil and groundwater.

Geological heterogeneity. The subsurface environment is heterogeneous, and soil permeability can vary orders of magnitude over short spatial scales. Substantial quantities of contaminants can thus be trapped in lower-permeability strata that are bypassed during conventional pump-and-treat or soil vapor extraction. Transfer out of these strata into the faster-moving fluid is controlled by molecular diffusion, which is a very slow process that can take years. Heterogeneities can also exert a strong influence on NAPL migration following spills. For example, the controlled field study reported by Kueper et al. (1993) showed that downward perchloroethylene (PCE) migration was hindered by small, finer-

grain lenses on the centimeter scale. These lenses enhanced lateral spreading and caused a highly variable distribution of PCE pools and residuals that were trapped in relatively coarser-grained horizontal lenses. Both geological and flow heterogeneity (discussed below) can limit the effectiveness of many strategies for active remediation of soil and groundwater, such as pump-and-treat, soil vapor extraction, in situ chemical oxidation, and air sparging.

Flow heterogeneity. In addition to geological heterogeneity, there are several other important factors that can cause spatial variability of groundwater flow. Even in relatively homogeneous porous media, the spatial arrangement of extraction or injection wells can result in flow that bypasses certain regions of the aquifer (Javandel and Tsang, 1986; Christ et al., 1999). Water flow through material having high dense nonaqueous phase liquid (DNAPL) residual saturation will be hindered, and in fact the water flow patterns will change as the DNAPL saturation decreases due to dissolution (Nambi and Powers, 2000). Flow heterogeneity can have a significant impact upon the performance of recently developed technologies for aggressively treating DNAPL source zones. Some of these techniques require injection of fluid containing reacting chemicals (e.g., potassium permanganate, surfactants) that must mix with ambient groundwater or trapped DNAPL. Because of flow heterogeneity, this mixing may be incomplete, and in some instances the injected fluid can push contaminated groundwater outside of the treatment zone. For in situ air sparging, injection of air bubbles through a series of regularly spaced wells results in removal of volatile organic compounds that are dissolved in the groundwater. However, several studies have demonstrated that the bubbles may follow preferential airflow channels, thus bypassing a significant fraction of the contaminated zone (Brooks et al., 1999; Elder and Benson, 1999).

Slow desorption. Many hazardous compounds tend to sorb onto aquifer solids and sediments. The degree of sorption depends upon the properties of both the contaminant and the solid phase; however, certain classes of compounds like metals, polychlorinated biphenyls (PCBs), and polyaromatic hydrocarbons (PAHs) tend to sorb much more strongly than others (e.g., light petroleum compounds) (NRC, 2003). Although some of the sorbed mass is readily desorbed, a significant fraction of some organic compounds that have been in long-term contact with aquifer materials containing diagenetically aged carbon will undergo very slow desorption as a result of hindered molecular diffusion through microporous

solids (Weber et al., 1991; Pignatello and Xing, 1996; Luthy et al., 1997). This mechanism may be partially responsible for the long tailing observed in contaminant breakthrough curves during active remediation of groundwater and soil, and it may also affect the performance of ex situ separation operations like soil washing. Slow desorption may be a benefit for certain kinds of contaminated sediment management strategies (e.g., capping) that aim to isolate contaminants in situ. Many studies suggest that the slowly desorbing fraction of an organic contaminant pool has reduced bioavailability and thus may present less of a potential risk (NRC, 2003a).

Slow dissolution from nonaqueous phase liquids. Many organic contaminants were released into the subsurface as a separate organic liquid phase. Liquids such as light petroleum products are less dense than water and tend to form pools above the water table, whereas organic solvents and coal tar that are more dense than water can sink below the water table and be trapped as small ganglia and lenses. These pools, ganglia, and lenses serve as reservoirs of contaminants that dissolve very slowly into the flowing groundwater. It is well known that NAPL-contaminated sites are among the most difficult to remediate. NRC (1994), using simplified dissolution rate equations, computed that it could take more than 100 years for moderately sized NAPL spills to completely dissolve.

Erosion–deposition processes in contaminated sediments. Natural attenuation processes in sediments may lead to decreases in water column contaminant concentrations over time. These processes include deposition of clean sediment, which tends to stabilize and separate contaminated sediment from the overlying water, as well as contaminant degradation and transformation processes. However, other processes such as erosion or resuspension of the sediment may serve to reintroduce contamination to the water column from the sediment. Erosion of sediment during high-flow conditions may lead to replenishment of surficial sediment concentrations, such that significant water column concentrations are maintained over long periods.

Threshold for microbial degradation. At low concentrations, some contaminants may no longer be able to act as a carbon or energy source or as an electron acceptor to support microbial communities. This is generally because the enzymatic machinery responsible for contaminant transformation requires a certain contaminant concentration in order

to be activated. Thus, microbial degradation may slow or stop, leading to an asymptotic limit on the decrease in concentration. Sokol et al. (1998), for example, reported that below 35–45 ppm, microbial transformation of PCBs in Hudson River sediments effectively ceased. This value is generally above both screening levels and remedial goals frequently set for these compounds (EPA, 1997b; Buchman, 1999).

There are numerous case studies that document how the effectiveness of remediation decreases over time, one of which is highlighted in Box 2-3. Also, see Appendix B and over 270 additional case studies described at the Federal Remediation Technologies Roundtable (FRTR) website (http://www.frtr.gov) and in FRTR (1995, 1997, 1998a, 2000, 2001). The case presented here exhibits the so-called “asymptote effect” where the contaminant concentration or mass decreases over time and

levels off at a finite value. For a treatment technology, such an asymptote may be indicative of system ineffectiveness if the asymptotic value is greater than the cleanup goal. (The many technical reasons for this were described above.) However, for a containment technology, such an asymptote could be indicative of effectiveness. Some technologies (e.g., permeable reactive barriers or some forms of institutional controls) do not exhibit the asymptote effect, because contaminant concentration or exposure may decrease “instantaneously” to acceptably low values.

The behavior shown for the Livermore site in Box 2-3 is typical of that of many additional case studies (see the FRTR website noted above and Appendix B) and consistent with the experience gained in attempting to clean up sites using many different approaches. The general trend is shown conceptually in Figure 2-4, which demonstrates how the effect-

FIGURE 2-4 Schematic graphs showing typical changes in cumulative mass removed and the mass removal rate over time. In cases where costs increase linearly with time, the x-axes may also represent cost.

effectiveness of typical cleanup technologies in removing contaminant mass decreases over time. On the left is shown the cumulative mass removal over time, and on the right is shown the mass removal rate. Not all of the case studies reviewed by the committee reached an asymptotic value as shown in Figure 2-4. However, many did, and others are likely to do so in the future. Most of the pump-and-treat systems that did reach such a limit did so within five years of commencing operations (FRTR, 1998b,c).

Congress was not aware of the level of complexity involved in attaining health-based cleanup goals when the Comprehensive Environmental Response, Compensation, and Liability Act (CERLA) was enacted in 1980 and amended in 1986. Thus, there is limited guidance in the law regarding what to do when remedies are not able to meet cleanup goals. For Superfund sites, Congress requires that each remedy attain the requirements provided in environmental law, for example, health-based drinking water standards. (There are exceptions to this requirement as discussed in Chapter 6—e.g., if compliance can be shown to be technically impracticable or if the remedy will lead to greater risk.) In many cases, treatment technologies reach an asymptote prior to reaching this goal, although protectiveness can often be maintained by cutting off exposure pathways via containment or institutional controls. Except for guidance on technical impracticability, there is minimal if any guidance on how to deal with situations in which cleanup goals are not being met after prolonged operation of the remedy (which necessarily prevents site closeout).

Private sector cleanups may shed some light on how sites are dealt with when remedies are no longer effective. The U. S. Environmental Protection Agency (EPA) requires private sector potentially responsible parties (PRPs) to implement the remedial action “until the Performance Standards [remedial action goals] are achieved and for so long thereafter as is otherwise required.”³ EPA may modify the work when it “is necessary to achieve and maintain the Performance Standards or to carry out and maintain the effectiveness of the remedy set forth in the ROD [Record of Decision]”; a “modification may only be required…to the extent

that it is consistent with the scope of the remedy selected in the ROD.” Clearly, then, changes are allowed if a remedy is proving to be ineffective in meeting cleanup goals. However, the PRP is allowed to seek a remedy change only if EPA determines that the remedial action is not protective of human health and the environment. This creates a problem because many remedies that are ineffective in reaching cleanup goals might be very effective at protecting human health and the environment through plume containment. The situation is different for Resource Conservation and Recovery Act (RCRA) corrective action sites. Here, EPA policy is to consider implementing additional, more effective remedial technology if it becomes available (EPA, 1993). The committee, however, could find no examples of this policy having yet been implemented.

Base Realignment and Closure policy also limits the implementation of additional remedial action to situations where the selected remedy is no longer protective of human health and the environment because the remedy (including institutional controls) did not perform as expected, or because there has been a discovery of additional contamination attributable to the Department of Defense (DoD) (DoD, 1997). This same guidance does not explicitly state whether failure to attain a health-based groundwater or soil remedial action goal constitutes a failure to “perform as expected.”

Thus, except for technical impracticability, there is no widely accepted policy for addressing situations where cleanup goals are not being met after extended operation of the remedy. The lack of explicit policies for addressing the large number of hazardous waste sites reaching this “point of diminishing return” is most likely a reflection of the fact that such situations have not become common until recently.

***

The need to respond to a set of multiple, sometimes conflicting, objectives; the ineffectiveness of current technologies in reaching cleanup goals for contamination at complex, high-risk sites; and the limited guidance on what to do when an asymptote is reached prior to meeting cleanup goals as long as the remedy remains protective of human health and the environment are pressing problems at many federal facilities. The remainder of this chapter describes an approach to cleanup that can accommodate these issues, that provides guidance at key decision-making periods, and that deals with the uncertainty inherent in many remedial strategies—both engineered technologies and institutional controls.

The predominant paradigm for site restoration in the United States has until relatively recently involved a highly linear, unidirectional march from site investigation to remedial action and eventually to site closure, which reflects our natural and understandable desire to “deal with the contamination and put it behind us.” The paradigm is implicit in a recent DoD description of key milestones for site restoration programs, as shown in Figure 2-5. This figure depicts three targeted and sequential milestones of “remedy in place,” “response complete,” and “site closeout” that are consistent with this linear approach to site cleanup. Nowhere does the schematic allow for sites to cycle back through previous stages, although it indicates that some sites may need to be reevaluated prior to closeout. As sites have advanced through the restoration process, there has been a growing recognition that more iterative procedures are needed, with ongoing site stewardship and reevaluation of monitoring and remediation efforts at many sites. Because of the complexity of the subsurface environment, often incomplete identification of contaminant sources, and the long timeframes required for remediation, site cleanup must not be viewed as a one-time event or an action that ends once a remedy is implemented.

The need for iterative, adaptive approaches to site restoration and stewardship is supported by much of the recent literature on risk assessment and risk management. Most recently published frameworks and approaches to understanding and addressing risks to human health or the environment incorporate a high level of public participation and deliberation in which iterative steps are proposed for problem formulation, process design, option identification, information gathering, synthesis, decision, implementation, and evaluation (NRC, 1996). Such an iterative risk management framework was developed by the Presidential– Congressional Commission on Risk Assessment and Risk Management (1997a) (see Box 2-4). It is applicable to federal and other facilities and was recommended for use on PCB-contaminated sediments (NRC, 2001a). Hallmarks of such frameworks are that ongoing learning and feedback are used to address and incorporate scientific knowledge, new technological capabilities, and changing socioeconomic conditions into action plans over time, thereby informing new analysis, institutional learning, and public participation.

The overall environmental planning and management system for federal site restoration involves a hierarchy of decisions, many of which are stimulated by the changing conditions (economic, technological, pub-

lic needs) encountered over the years. Choices among management alternatives may be constrained by the limitations of technical and institutional capabilities; however, these constraints are not necessarily fixed. Rather, investments in research and experimentation may allow development or improvement of technology to overcome current limitations, and thus provide DoD with more alternatives for risk management than existed initially. A broader systems approach that promotes effective knowledge generation (monitoring and fundamental research) and use of that knowledge (adaptation) can provide a wider range of decision options and thereby improve site management over the long term.

The characteristics of a broader systems approach described above are embodied in the concept of adaptive management. Adaptive management is an innovative approach to resource management in which policies are implemented with the express recognition that the response of the system is uncertain, but with the intent that this response will be monitored, interpreted, and used to adjust programs in an iterative manner, leading to ongoing improvements in knowledge and performance (Holling, 1978; Walters, 1986, 1997; Walters and Holling, 1990; Lee, 1993).

As noted by Lee (1999), “Adaptive management is learning while doing; it does not postpone action until ‘enough’ is known but acknowledges that time and resources are too short to defer…action.” As such, adaptive management provides a structured approach for addressing uncertainty, making decisions in the face of it, and seeking to improve these decisions in an iterative manner by actively acquiring the knowledge necessary to reduce uncertainty. Part of this process can involve formal hypothesis testing, which is an important theme in the adaptive management literature. That is, hypotheses are formulated about the future events of interest, and then experiments or other activities (sometimes including statistical analysis) are conducted that will either confirm or reject the hypotheses. Adaptive management is also enhanced both by formal analysis/optimization methods (e.g., Williams, 2001) and by public participation (Shindler and Cheek, 1999).

The committee has coined the term “adaptive site management” (ASM) to refer to the application of the adaptive management concept to hazardous waste cleanup. Within the environmental arena, adaptive management concepts are relatively new but are particularly timely, especially given the observed limitations in remediation effectiveness and the increased use of remedies that will leave residual contamination in place for long periods. ASM is a flexible and iterative approach designed to allow decision makers to evaluate new information as it is re-

ceived and adjust cleanup procedures or other management options over time.

ASM formalizes questions and decisions that the remedial project manager (RPM) and remediation team should address to readily adapt to changes in technology, remedy effectiveness, and other external influences that impact the management of contaminated sites. It follows closely the approach outlined in the President’s Risk Commission Framework, and it builds on proposed Navy guidance for maximizing remedial effectiveness and cost efficiency (NAVFAC, 2001). It is also consistent with recent Department of Energy (DOE) interpretation of EPA guidance—Using Remedy Monitoring Plans to Ensure Remedy Effectiveness and Appropriate Modifications (DOE, 1998). At its heart is a formal decision process that stresses the collection and evaluation of information on remedial performance, provides options in the face of uncertainty, and embeds linked feedback loops so that action is taken quickly to change or optimize remedies that are unlikely to attain site-specific cleanup goals within a reasonable period of time. The main tenets of ASM are that it:

• is applicable at various stages of site restoration,

• is applicable to a wide variety of sites regardless of the contaminants being addressed or remedies envisioned,

• provides a mechanism for the optimization of existing remedies, changing ineffective remedies, and refining the site conceptual model,

• formalizes the routine examination of monitoring data and how to act upon the data,

• incorporates public participation,

• recognizes uncertainty and suggests approaches to dealing with it, especially when institutional controls are used,

• stimulates the search for new, innovative technologies to replace older or inefficient approaches,

• stresses the need for pilot programs to test both new technologies as well as modifications of existing technologies that might enhance their effectiveness, and

• recognizes the increasing role of long-term stewardship (which is synonymous with long-term management used in DoD terminology and in Figure 2-5).

Adaptive management provides a way of moving forward with site cleanup and stewardship programs in the face of uncertainty. As noted in

Chapter 3, uncertainty is inevitably present in the type, amount, and location of contaminants; in the response of contaminants to changes in physical, chemical, or biological conditions brought about by the remediation technology; in the response of on- and offsite species to these changes; and in the evolving socioeconomic conditions of the surrounding community that can affect the community’s preferences. Chapter 4 advocates that more certain steps be taken in the short term based on available data, while data collection and evaluation of more uncertain elements of the overall plan proceed in parallel, in order to make progress toward cleanup.

There is growing experience with adaptive management approaches in a number of other public and private domains. The principal use of adaptive management has been for applications to wildlife and ecosystem management (see, for example, provisions for the U.S. Forest Service’s Land and Resources Management Plans,⁴ recent decisions by the Alaska Department of Fish and Game,⁵ and the EPA/Environment Canada-sponsored Lake Superior Lakewide Management Plan.⁶ NRC (1999b) reviewed the growing use of adaptive management for ecosystem resources in the Grand Canyon. In particular, it noted the need for (1) a long-term monitoring program in the Grand Canyon and (2) a strategy for scientific evaluation of policy alternatives, both in terms of ecological outcomes and the values of stakeholder groups. The authors recognized that effective adaptive management in the Grand Canyon will require tradeoffs among objectives favored by different groups as well as mechanisms for equitable weighting of these objectives. Similarly, NRC (2002) supported adaptive management to enhance scientific inquiry and policy formulation about the Missouri River ecosystem. The key tenets of the approach were outlined as (1) programs to maintain and restore ecosystem resilience, (2) recognizing and adapting to uncertainty, (3) interdisciplinary collaboration, (4) models to support collaboration and decisions, (5) meaningful representation of a wide array of interest groups, and (6) ecosystem monitoring to evaluate the impacts of management actions.

Adaptive management is also an important element of water re-

sources planning (e.g., Mays and Tung, 1992) and strategies for dealing with global climate change (McCarthy et al., 2001). Thus, for example, NRC (2001b) recommends an adaptive implementation of strategies for determining and implementing Total Maximum Daily Loads under the Clean Water Act, including immediate actions, an array of possible long-term actions, success monitoring, and experimentation for model refinement.

Adaptive management has a similarly strong foundation in business strategies and planning approaches that utilize “feedback loops” to help respond to new information (Ayres, 1969). Iterative programs with ongoing performance evaluation and improvement are also a key component in the recent design of corporate environmental management systems (Crognale, 1999). These environmental management systems focus on planning for continuous improvement in environmental compliance and include the identification of performance metrics, comparison to goals, consideration of costs and benefits, and institutional and personnel steps needed to ensure that the process is self-sustaining. A review of a recent advance in methods used for adaptive management of private and public investment decisions is presented in Box 2-5.

Like the business management models discussed above, adaptive management requires explicit identification of future decision points and alternative actions that can be taken depending upon ongoing performance assessment and outcome evaluation, all of which are embodied in the ASM approach. Nonetheless, there are important issues related to applying adaptive management to hazardous waste cleanup that make it different from business applications. Site management is constrained by particular legal requirements, though consideration of changes in regulations that could better enable flexibility is also important. Furthermore, as noted before and expounded upon below, site management is a multiobjective problem, often involving incommensurate measures of health, safety, ecological quality, cost, natural resources value, and the social and economic well-being and satisfaction of the surrounding community. Finally, federal site management cannot afford obvious failures in remediation. Every site is unique and important, and its aggregate management is expected to be successful. However, not every individual project explored for the site need necessarily be successful. This is both expected and accepted when the full range of possible outcomes for each option is considered, and effective contingency plans and alternatives are available for program adjustment. Absent the valuable data collected during evaluation and experimentation outlined in the ASM approach, it is unlikely that new and innovative approaches to cleanup will be developed and implemented.

The concept of adaptation is not foreign to CERCLA and RCRA activities. There are certainly cases where project managers have modified remedial activities in response to poor system performance. Over the last decade, a number of formal approaches have been developed to introduce adaptation specifically into data collection and site characterization activities. Examples of these include Expedited Site Characterization (ASTM, 1999), Adaptive Sampling and Analysis Programs (DOE, 2001), the Observational Approach (Baecher and Ladd, 1997), and, most recently, EPA’s TRIAD program (Crumbling et al., 2001). Expedited Site Characterization has emphasized the application of appropriate onsite technical expertise, rapid data collection activities, and dynamic work plans to subsurface characterization problems. Adaptive Sampling and Analysis Program work has primarily focused on the application of in-field decision making and field analytics to contaminated soil issues. The Observational Approach has its roots in geotechnical engineering. As discussed in Box 2-6, it explicitly acknowledges that environmental decision making often involves irresolvable uncertainty. This uncertainty is addressed by contingency planning and flexible designs that can be

modified based on changing field conditions as they are encountered. Finally, the EPA’s TRIAD program mixes systematic planning with dynamic work plans and onsite analytical tools to streamline environmental data collection programs. Many of these sampling techniques are discussed in greater detail in Chapter 3. ASM builds on these experiences to provide a framework that weaves adaptability into the remedial design and implementation process as a whole.

By explicitly recognizing the role of ongoing research and information collection, remedial performance assessment, and evaluation of alternative remedies, ASM provides greater flexibility than the current, more linear approach to site management, with the expectation of an

overall improvement in the performance of the management system.

ASM involves two main steps. Step 1 focuses on understanding and conceptualizing a site problem and identifying risks to human health and the environment. Step 2 focuses on the selection and implementation of a remedy, monitoring performance of the remedy, adapting the remedy or management goals to accommodate changing conditions and improve cost-effectiveness, and, finally, completing the remedy and closing out the site. Although many Navy sites may have progressed past the point of selecting a remedy, the entire process (Steps 1 and 2) is shown because it is important that it be visualized by and articulated to all remediation team members so that key tasks are understood and not overlooked. This is absolutely essential at the onset of the Problem/Context phase and should continue regularly throughout implementation of the remedy. Note that the text gives greater weight to those steps within ASM that differentiate the approach from current practice. This is not meant to impart greater importance to any particular activity, but rather to provide detail where there is otherwise less information.

Recent Air Force (2001) and Navy (NAVFAC, 2000, 2001) guidance documents contain a number of the precepts imbedded in the proposed ASM approach. For example, NAVFAC (2001) requires evaluation of system performance (how well a remedy is meeting design criteria), system suitability (how likely it is the remedy will attain cleanup goals), and whether there are life-cycle limitations (i.e., whether the remedy will reach the point of diminishing returns). The guidance calls for an alternative strategy when a plot of cumulative mass removed versus time exhibits “an asymptotic condition” prior to attainment of the cleanup goal. The alternative strategy may include (1) modifying an existing system to improve cost-efficiency and cost-effectiveness, (2) implementing different remediation (including the sequencing of several remedial technologies to achieve cleanup goals) when the remedial action cannot be modified to achieve cleanup goals, or (3) changing the cleanup goals. Various places in the following discussion highlight where ASM builds upon or reinforces NAVFAC (2001) and where it differs.

There is also new and explicit guidance from the DoD Under Secretary of Defense that supports ASM. In particular, updated Defense Environmental Restoration Program guidance states that the evaluation does not end once a response action is implemented (DoD, 2001). Continued

activities should include optimizing the overall performance and effectiveness of the remedy, assessing if the ROD objectives have been achieved and if the treatment system is still needed, and determining if a different remediation goal is needed or if an alternative technology or approach is more appropriate. The guidance also suggests that technology development efforts should be supported, with a goal of increasing the overall effectiveness of response activities, including the validation and certification of emerging technologies.

It should be noted that “optimization” is used in the remediation literature in different contexts. In the broadest sense, optimization means implementation of any change to a remedial system to make it work more efficiently toward the cleanup goal, where operating costs are reduced as a result of making a change, or where the desired asymptotic cleanup condition is reached more quickly. The change could involve making a single technology more efficient, adding components to a treatment train (as discussed in Chapter 5), or switching cleanup remedies. Extensive literature, including the military documents cited above, is available to provide guidance and criteria for such “optimization.” Within this report and within ASM, however, “optimization” is used specifically to refer to making a single technology more efficient.

Step 1 of the ASM process (Figure 2-6) is specific to those tasks that occur before a remedy is selected. The importance of this pre-remedy selection step is also discussed in existing Navy guidance (e.g., NAVFAC, 2001), internal Navy memos (e.g., Pirie, 1999), and previous NRC reports (e.g., NRC, 1999a). In some cases, sites may have progressed to Step 2—remedy selection and implementation—without adequately completing Step 1, which presents another series of problems that ASM can help to address. The individual tasks of Step 1 include developing the problem/context and the site conceptual model and conducting risk assessments. It should be noted that there is nothing about the Step 1 process outlined in Figure 2-6 that is not already encompassed by the CERCLA cleanup paradigm.

During the problem/context formulation phase, decision makers,

FIGURE 2-6 Step 1 of adaptive site management: pre-remedy selection.

stakeholders, and other interested parties identify and define important issues to help formulate the problems at an individual site. Even for sites that have progressed beyond the problem/context phase, it may be necessary to return to this point if new public issues, risks, or problems arise. That is, if site conditions and other important influences on management actions should change, the remediation team will need to return to the problem/context phase. As shown in Figure 2-7, ASM allows users to revisit Step 1 as new information is obtained, as site conditions change, or where it is otherwise necessary to review previous tasks.

One of the first issues that should be defined during the problem/context phase is the amount of time available for undertaking the requisite studies and information gathering, as well as any resource con-

straints (personnel or financial) that might hinder this activity. It is also important to understand any regulatory requirements that may influence the kinds of data and information collected and data analysis needed. There may be more than one set of regulations operating at a given site, and some requirements may be mandatory while others may be left to the discretion of site managers.

Another important issue is to determine and clearly understand what the future land use will be for the site, or whether there might be multiple land uses envisioned. The latter may occur where large sections of a site may be relatively unimpacted by operations and require relatively little restoration before they can be utilized for commercial or recreational purposes. The Navy has issued memoranda (Navy, 2001) to ensure that future land use activities remain compatible with the land use restrictions imposed on the property during the remediation and restoration process. Future land use has a direct influence on the type of data needed from environmental or engineering studies, on the long-term stewardship undertaken, as well as on the level of cleanup that is required. In addition, there is a need to work with local stakeholders and the public, particularly since these groups will likely have a vested interest in the disposition of the land once the remedial action has been completed and/or the site is closed out.

Both active military bases and those that are closing are of substantial interest to the local public and, in some cases, to the larger public in a particular geographic region. Thus, it is imperative that site managers identify individuals or groups that should be consulted on future land use issues and on the remedies being considered to achieve cleanup consistent with the planned land use. Stakeholders will need to understand their role in the decision-making process and have a mechanism with which to articulate their expectations to the RPM.

The site conceptual model (SCM) is an “illustration” of the site that details the location, concentration, and pathways by which contaminants are thought to be moving through the environment and how/why humans and ecological receptors are being exposed to those contaminants (NRC, 1997). That is, the SCM provides an initial assessment of the hydrogeologic environment, contaminant sources and sinks, and key processes (such as environmental fate) that are potentially operating at the site, and illustrates how these could influence the need for or types of remediation.

With regard to contaminant sources, the SCM should identify both known sources such as landfills and underground storage tanks as well as impacted media such as groundwater, surface water, soils, sediments, and air. The SCM documents exposure pathways or routes through which humans, non-humans, and habitats come in contact with contaminant(s) and links those pathways with a list of contaminants of concern found at the site that might pose potential risks. Processes that may alter contaminant form such as hydrolysis, photolysis, and microbial degradation should be assessed (Peyton et al., 2000). Finally, the SCM should identify those receptors—both human and non-human—that are exposed to the contaminants, including organisms in affected aquatic and terrestrial habitats. Further guidance on developing a site conceptual model can be found in NRC (1994), USACE (1996), EPA (1997c), and NAVFAC (2001).

From a more formal perspective, the site conceptual model and its companion risk assessment can be thought of as a set of linked hypotheses, with the remediation process that follows constituting a test of the validity of these hypotheses. If the experience acquired during remediation brings some of these hypotheses into question, the site conceptual model may need to be revised. For example, the SCM will need to be changed to reflect a discovery of new contaminants during implementation of the remedy.

Although not a separate activity per se from problem/context formulation and site conceptual model development, data collection is nonetheless a key element of ASM. Quantitative data are needed to support remedy selection, to determine if the remedy is effective, and to reveal how the remedy should be implemented or modified to achieve optimal performance (NAVFAC, 2000, 2001). During this activity, the remediation team collects and analyzes data on contaminant fate and effects in media. This includes determining what contaminants are present (at what levels and in which media, considering air, surface water, soils, sediments, and groundwater) as well as which receptors (human, non-human, habitats) are likely to be exposed to and adversely impacted by these contaminants. More detailed information can be found in NRC (1999a) and EPA (1997c).

Risk assessment is an important step in the site cleanup processes proscribed under CERCLA and RCRA. Briefly, the process results in a mathematical estimate of the potential risk faced by humans or ecological receptors exposed to the contaminants of concern. A similar framework applies to both human health and ecological risk assessment. The risk assessment can be purely quantitative, purely qualitative, or some combination of the two. Risk assessments integrate information on the physical conditions at the site, the nature and extent of contamination, the toxicological and chemical/physical characteristics of the contaminants, the current and/or future land use conditions, and the dose– response relationship between projected exposure levels and potential toxic effects. The results of the risk assessment, coupled with the expected future land use, are key in determining what level of cleanup will be necessary for a particular site. Because risk assessment is not the focus of this report, the reader is referred to the seminal report describing this process (NRC, 1983) and to other reports (EPA, 1992; Calabrese and Baldwin, 1993; Maughan, 1993; Suter, 1993; EPA, 1997c; Committee on Environment and Natural Resources, 1999; NRC, 1999a).

As shown in Figure 2-7, the second part of ASM involves selecting, implementing, monitoring, adapting, and completing the remedy. Step 2 of ASM also links readily with the elements of Step 1 because it is expected that there will be situations where it will be necessary to return to Step 1 to refine the site conceptual model, collect additional information, refine the risk assessment, or change the remedial goal.

A key element of ASM is the formalization of “management decision periods” (MDP) at which decisions are made based on pilot-scale work, on changes in land use or stakeholder needs, and on monitoring data and other intelligence that may lead the RPM to refine and/or revise a management decision. Such management decisions provide an opportunity for periodic check-ups to determine if the remedial technology is meeting its objectives and, if not, whether adjustments are needed. In this respect, ASM differs from recent cleanup guidance for Navy facilities (NAVFAC, 2001), which does not formalize these decisions nor stress the need for stakeholder involvement. These management decision periods are similar to “scientific management decision points” detailed in

EPA guidance for Superfund (EPA, 1997c).

The management decision periods are designed to take advantage of the feedback loops embedded in ASM, such that uncertainties in site restoration can be addressed. They also represent formal opportunities for the RPM and other project mangers, regulators, and interested stakeholders to evaluate incoming and existing data and to reach agreement on what additional management steps, if any, need to be taken. Having formal management decision periods does not preclude routine discussion among groups or individuals involved with the remediation process; in fact, it should encourage greater dialogue among groups so that important issues are managed as they arise. For simplicity, the management decision periods are focused on the four main categories of actions associated with Step 2: implementation, monitoring, adaptation, and long-term stewardship.

The purpose of MDP1 is to ensure that the remedy selected is still practicable and implementable under site-specific conditions and that an appropriate, well-designed monitoring plan is developed. At the onset of Step 2, a remedial goal is chosen that takes into consideration the risk assessment results, expected or desired future land use, public and stakeholder concerns, and technological capabilities. The goal can be established to protect human health, ecological receptors, or both, and will be site-specific. In some cases there may be multiple goals set for sites where there will be multiple future land uses, as may be the case at larger federal facilities. ASM affords the flexibility to return to Step 1 in order to revise the remedial goal in cases where the goal has been determined to be unattainable (see discussion below).

Subsequently, an initial remedy is selected and either published as a ROD or finalized under another regulatory context (perhaps as a Consent Agreement, Consent Order, or federal facilities cleanup agreement). Details on specific remedies of importance at complex sites including the use of treatment trains can be found in Chapter 5, along with recommendations on when specific remedies are applicable and when they are not.

Designing the remedy requires data on contaminant concentrations and movement and other exposure-related parameters that are generally collected before remedy selection. However, because there can be a long lag time (years) between remedy selection and implementation, there

may be a need for additional information in these same areas or in areas specific to engineering or construction needs so that final implementation is based on valid assumptions. Thus, the remedy design should take into account new information on the types of contaminants detected in affected media and on specific engineering or construction requirements that was not available when the initial remedy was selected. For example, placing a sheet-pile wall in a shallow aquifer, coupled with a pumping regime to gain control over the contaminant plume, could be viewed as the most likely remedy for groundwater remediation. The ROD could be written to reflect this view and performance standards put in place when the ROD is issued. A finite amount of data would have been collected to support this remedy decision, but there may have been uncertainties with the data set. Later, as remedy design begins, it is discovered that the geological conditions or hydrologic scheme are such that the wall either will not work as envisioned or will require extensive modification for proper performance. That is, the hypothesis that the wall would be the appropriate remedy for the conditions present is disproved by subsequent data collection. The RPM and remediation team now have a specific engineering hurdle to overcome prior to implementing a remedy. The value of the MDP1 analysis prior to full-scale implementation is that it helps to ensure that the remedy will indeed meet the requirements set forth in the ROD.

Another early decision point not specifically noted in Figure 2-7 but available within the CERCLA process is called value engineering. Value engineering is not a regulatory process per se, but entails the application of proven engineering analysis to the implementation and functioning of a remedy. In this process the full details of what the remedy is expected to do, the conditions at the site, and all other relevant data are presented to a group of experts who may have had little or no prior contact with the site. These experts “peer review” the proposed remedy and provide an analysis of the potential for the remedy to work and, more importantly, they identify the situations they believe could cause the remedy to fail. The review can help to determine the level of uncertainty in the remedy’s performance and the steps that should be taken to reduce that uncertainty. For any independent review to be successful, site managers and regulators must agree beforehand to abide by the results of the expert panel. In this way both groups gain the benefit of additional advice absent the potential bias that may occur were the review to be conducted by experts from one group or the other. [It should be noted that EPA created a National Remedy Review Board to provide technical and policy review of remedial decisions (EPA, 1996c, 2001a), although the independence of

this review process is questionable because the team is composed of the EPA supervisory personnel involved in the Agency’s cleanup program.]

Another important task prior to implementation is to design an appropriate monitoring program (see Chapter 3 for details), especially in situations where the remedy is likely to require significant operation and maintenance costs over an extended period of time. Long-term monitoring, which is likely when pump-and-treat is the preferred remedy for groundwater, will be an important source of quantitative data highly relevant to determining whether the remedy is effective or not. The monitoring plan should be designed in light of the site conceptual model so that the data not only will illuminate the remedy’s performance, but will also be valuable in refining the site conceptual model at later stages of site restoration.

Developing a monitoring program will require that the site managers and regulators agree on the kinds of parameters to be monitored and how those parameters fit with the performance standards established in a ROD or other regulatory document. An even more important element of MDP1 is reaching agreement on how the monitoring data will be used in decision making, particularly because the potential future decisions involve modifying or changing the remedy based on the results of the monitoring program.

Subsequent to MDP1 and once the remedy is implemented, several actions can potentially occur as part of ASM. In addition to operation of the remedy, there are ongoing monitoring activities, as discussed below under MDP2. A third activity—evaluation and experimentation—is denoted in Figure 2-7 by a dashed line. This activity is unique to ASM and is one of the hallmarks of adaptive management in general. It refers to conducting experiments and other research activities in parallel with implementation of the chosen remedy. This activity may occur at the level of an individual site, in which portions of the site are devoted to experimentation while others are undergoing the chosen remedy, or it may refer to collecting information about experiments going on elsewhere, the results of which are relevant to specific sites. Evaluation and experimentation may consist of pilot-scale studies done at one of the national demonstration sites. The evaluation and experimentation track is an opportunity to test innovative, less certain, sometimes riskier remedies that were not well enough established to be chosen as the initial remedy in the ROD. The data and information gathered on this parallel track can then be used later to optimize or change the remedy at MDP3 if performance standards or the remedial goal are not being met. This explicit evaluation and experimentation track of ASM is discussed in greater detail below

under MDP3.

MDP2 consists of a series of key questions regarding the output of the monitoring program—quantitative information on the effectiveness of the remedy. Affirmative responses to these questions lead to “response complete” and eventually to MDP4, whereas negative responses lead to MDP3. MDP2 was developed to ensure that at regular intervals, the monitoring data are evaluated and judged against the operational parameters agreed to or imposed by a regulatory construct. Throughout the monitoring period, and specifically at these intervals, the RPM should ask three sequential questions: (1) is the remedy meeting the performance standards (as set forth in the ROD or other binding document), (2) are the operational expectations of the remedy being met (whether cost or some other parameter that the RPM and remediation team have set), and (3) is the remedial goal being met.

The first question listed above is perhaps the simplest—is the remedy meeting performance standards? The data to make this judgment can derive both from the monitoring program and from other relevant studies that might have been completed (e.g., geotechnical studies to further define site conditions). If the remedy is not meeting performance standards, the RPM should initiate a thorough review of the remedy, the site conceptual model, and other relevant data to ascertain why. An important consideration is to determine whether the remedy has been in place long enough to have had time to equilibrate and operate properly. If not, then the RPM may wish to continue with the remedy and collect performance data for an additional period to ensure the remedy is not changed prematurely. The decision to continue the remedy and monitor it should have a finite time limit so that an ineffective remedy is not operated indefinitely.

If the system is meeting the performance standards, the second question is whether it is meeting operational expectations. These expectations may take the form of cost, up time, or some other metric, depending on the remedy and the needs of the RPM with respect to long-term stewardship and fiscal responsibility. For this comparison, the latest scientific information on technology performance could be combined with updated site characterization data to determine operational expectations. Generally these will be metrics selected by the RPM and others to ensure that the system is functioning as they want it to, above and beyond what

might be required by the ROD or other binding regulatory provision. Indeed, a system may be meeting regulation-specific performance standards, yet may not be particularly cost-effective, may require excessive maintenance or down time, or may require constant reoptimization.

If the remedy is found to be meeting performance standards and is operating as efficiently as expected, then the third question is whether remedial goals are being met. For example, goals may be risk-based, such as reduction of contaminant concentration to a level below that which poses a risk to human health or the environment. Remedial goals may be based on containment, as in preventing a groundwater plume from moving offsite, or they may be based on mass removal, as for many groundwater extraction and sediment and soil excavation remedies. If the remedy is not meeting the remedial goal, then site managers are faced with two important considerations discussed in MDP3: how the remedy can be modified, optimized, or changed to meet the performance standards, and whether the remedial goal is inappropriate and in need of change. If sufficient data are available to indicate the remedy has achieved its stated remedial goal, then another decision period (MDP4) is reached.

It is important to allow sufficient time to pass before deciding whether a system is meeting remedial goals. The necessary amount of time will be highly dependent on the chosen remedy and even more so on the hydrogeological conditions at the site. For example, in the case of bioremediation, enough time must pass for the microorganisms to acclimate to seasonal changes (in temperature, for example). Some aggressive source removal strategies like dredging and in situ chemical oxidation produce changes within weeks. However, in most cases additional monitoring is needed to determine the permanence of the result and the potential for rebound. This can be problematic with in situ oxidation, which usually consists of only one injection and perhaps an additional injection if rebound occurs within months. If rebound occurs later or is not monitored for at all, false conclusions about the performance of the technology may arise. Months are needed to properly evaluate the performance of soil/vapor extraction, and years are required for pump and treat, air sparging, monitored natural attenuation, or permeable reactive barriers. Time frames for performance determinations may also be imposed by the regulatory authority.

Monitoring data used in this fashion allow initial hypotheses about remedy performance to be tested and either confirmed or rejected. As discussed previously, these data should also be used to evaluate whether the hypothesis of environmental conditions represented by the site con-

ceptual model remains valid. Updating the model will reduce uncertainty in the current remedy’s effectiveness, and it may lead to the development of new management options. As a matter of course, changes in the site conceptual model should stimulate site managers to review ASM Step 1 to ensure that other potential issues (e.g., changes in risk assumptions) are not overlooked.

MDP3 illustrates most clearly the adaptive nature of ASM. Prior to this point, the site managers are focused on obtaining monitoring data and asking specific questions about remedy performance. Now, in MDP3, they must turn their attention to analyzing those data and any other relevant information to determine what future management steps are appropriate.

Reviewing, Modifying or Changing a Remedial Goal. After determining that the remedy is not meeting remedial goals, the RPM and remediation team, in coordination with regulatory agencies and interested stakeholders, should address whether the remedial goal is still appropriate. If it is not, the site managers would return to Step 1 of ASM. There are a number of issues that must be taken into consideration to ensure that this decision is made deliberately. First, it is possible that the goal was inappropriate in the first place. For example, a remedial goal may be inappropriate when a health-based cleanup goal, such as an MCL for an organic contaminant, is applied to groundwater that is nonpotable because of naturally occurring aquifer constituents. Second, a remedial goal could become inappropriate over time because changes occurring subsequent to the signing of the ROD—for example, a land use change or the discovery of new contamination—have altered the site conceptual model. In the case of a land use change, a goal designed for commercial/industrial land use would be inappropriate if in fact the future land use is residential. A third issue is whether a sufficient amount of time has been allowed for the remedy to function before reaching a determination of “goal inappropriate.” As discussed above, the time period will vary depending on the site, regulatory constraints, stakeholder concerns, the chosen remedies, and other influences. Site managers should recognize that changing a remedial goal is not a simple task and will require that they engage the relevant stakeholders to ensure that the new remedial goal is acceptable and compatible with the expected future land use.

Assuming that site conditions have not changed and sufficient time has passed since remedy implementation, the inability to attain remedial goals is largely controlled by the complexity of the site and the types of technologies that can be applied. At sites with either highly heterogeneous stratigraphy or fractured media aquifers and highly refractory contaminants such as DNAPLs, cleanup to background or health-based standards may be technically infeasible. Thus, one option at MDP3 is to pursue a technical impracticability (TI) determination, which for the most part documents the inability to achieve a particular remedial goal regardless of the technology applied (see EPA, 1993, and Chapter 6 for details on technical impracticability waivers). TI waivers, which are granted for groundwater contamination only, result in the selection of a new least-cost remedial goal (such as a new feasible concentration level). The remedy is then modified to achieve that new goal (e.g., to achieve containment rather than removal). This option is not the equivalent of site closeout because activities to protect human and ecological receptors such as land use restrictions, groundwater monitoring, and five-year reviews must continue.

If it is determined that the remedial goal should be modified, prompting a return to Step 1, a number of additional changes will result that should be discussed openly so there are no misconceptions. Where no existing technology can meet the initial remedial goal, the result is likely to be a less stringent goal, which may result in greater contaminant concentrations being left on site. This situation may necessitate a change in the future land use, perhaps from residential use to commercial/industrial use, which may in turn require modifications to deeds or restrictions on current or future property access agreements. Because changing the remedial goal automatically leads site managers back to remedy selection (see Figure 2-7), the remedy and monitoring program will have to be revisited.

Optimizing, Changing, or Adding a Remedy. If it is determined that the remedial goal is appropriate, but the remedy is not achieving the goal after sufficient time has passed, then the remedy should be optimized, added to, or changed entirely. Details on optimizing, modifying, and adding remedies are provided in Chapter 5 and NAVFAC (2001). At a minimum, optimization techniques, particularly for groundwater systems, should be undertaken whenever a remedy is not performing appropriately. Optimization of an existing remedy leads the site manager back to the “design process: remedy and monitoring” box. For those remedies that do not perform appropriately even after optimization, the RPM

should consult with the remediation team, recognized experts, and others to help them ascertain if the remedy can be further modified to achieve proper performance. If after further evaluation and consultation it is determined that the remedy cannot be modified adequately, the remedy may be unsalvageable and require wholesale replacement. Such a determination will, as illustrated in Figure 2-7, require a return to the “select remedy” box in Step 2.

Although a wide array of tools can be used to evaluate whether an additional remedial action or change is warranted once the point of diminishing returns has been reached, a relatively simple, graphical test can be used. In the case of groundwater contamination, contaminant concentration within the source area can be plotted over time; the need for a change may be evident when the slope of the line tangent to the performance curve approaches zero (the so-called asymptote) but the concentration remains above the site-specific remedial action goal. Such plots can also make it clear when continued operation of the existing remedy may incur substantial per-unit costs with relatively little improvement in mass removal. In order for these plots to be useful, the remediation team and the regulatory agencies must agree on a unit cost for the continued operation of the remedial action at the site under investigation, above which the existing remedy is no longer considered a tenable option. Information on the types of data to plot and their analysis are found in Chapter 3 and NAVFAC (2001).

Evaluation and Experimentation. Site managers should strive to incorporate new information collected during the evaluation and experimentation track of ASM into decisions about optimizing, adding, or changing remedies. As discussed previously, laboratory studies, pilot-scale activities conducted on- or offsite, expert panel evaluations, literature reviews, or newly acquired experience from other federal or private-sector sites should be assessed on a regular basis to determine if a more effective remedy applicable to the site of concern exists. For example, a selected remedy of containment might be replaced with an innovative treatment technology that would allow unrestricted use⁷ of the site and the associated economic and health benefits if the incremental costs were reasonable. This parallel track in which site managers adapt remedies to

information from internal and external sources is critical to overcoming the stalemate encountered at many sites where cleanup goals cannot be achieved. However, in order for this to succeed, potentially responsible parties, the Navy in particular and the federal government more generally, would have to make evaluation and experimentation an integral part of their overall remedial program. This adaptive feature of ASM differentiates it from the recent Navy guidance (NAVFAC, 2001), which does not specify an explicit need for ongoing evaluation and experimentation. It is also an extension of the report’s seventh objective, which stresses the role of knowledge generation and transmittal.

As discussed in Chapter 4, there are numerous mechanisms for undertaking evaluation and experimentation at an individual site and for obtaining relevant information and data externally. Some could involve current DoD agreements with EPA laboratories or offices, extramural grants with academic institutions or other nongovernmental groups, or collaborative activities such as those conducted through the Remediation Technology Development Forum (RTDF), a joint effort between EPA and private industry. Adoption of ASM would encourage the Navy to build stronger networks to the scientific and engineering communities in order to stay abreast of new technological developments that might prove applicable to existing or future cleanup scenarios.

The committee recognizes that time will be required to test ideas and new technologies prior to a full-scale implementation. It is not the committee’s intent that ASM be used as an argument for delaying important decisions while extensive analysis takes place (so-called “paralysis by analysis”). In fact, a definable characteristic of adaptive management is that more certain and sometimes simple actions are taken immediately while information is being gathered about potentially more effective but less certain technologies. That information should then be used to periodically revise the original action. In order for the concept to succeed, both tracks must operate simultaneously. However, in recognition of the many existing sites for which remedies have been ongoing for some time and for which no evaluation and experimentation were done, Figure 2-7 shows an upward arrow from “optimize, change, and add to remedy” to “evaluation and experimentation.” This suggests that it can be useful to conduct the latter activities even after a chosen remedy has stalled in meeting cleanup goals. While evaluation and experimentation take place, the temporary inability to meet performance standards or other regulatory requirements should not be used as a basis for notices of deficiency or enforcement action. Ideally, ASM should foster frequent interactions between site managers, regulatory agencies, and other stake-

holders that will improve overall communications, build trust and credibility, improve flexibility, and ultimately lead to greater efficiency in the restoration of federal sites.

The purpose of MDP4 is to provide a clear road map and describe actions necessary to achieve a site closeout designation. Crucial elements of MDP4 include planning for long-term stewardship and monitoring (if required) and agreeing on time intervals at which the site status can be formally reviewed by the RPM, remediation team, regulators, and interested stakeholders. The CERCLA process provides an explicit mechanism for doing this through the five-year review process, and five years should be viewed as the maximum time period between reviews. For non-CERCLA sites, the RPM and remediation team may wish to establish their own timetable for formal reviews and ensure that the regulatory agencies and interested stakeholders are involved.

Once a remedy has been in place for sufficient time and monitoring data provide measurable evidence that the remedial goal has been met, a site is designated (under military terminology) as “response complete.” Depending on the nature of the site, a variety of actions may still be required. If the remedial goal was based solely on mass removal and the required mass was removed, then the RPM is no longer required to take additional action if long-term monitoring is not required and if residual contamination (if any) poses no risk to human health or the environment. That is, any remaining contamination must be present at levels below those that allow for unrestricted use. In many cases where contamination is left in place, additional action is needed to ensure protection of human health or the environment. Examples include where a series of remedies are in place and only one of several has reached completion, where the remedy has changed from an active one (pump-and-treat) to a more passive remedy (monitored natural attenuation), or where a passive remedy was selected initially as the remedy of choice and a long-term monitoring plan was put in place.

The monitoring and oversight actions necessary to ensure protectiveness at such sites are lumped under the umbrella activity of long-term stewardship. It is the primary required activity once it has been determined that residual contamination has been left in place at levels above those required for unrestricted use, or when the remedy is one that requires monitoring and maintenance (pump-and-treat/containment of

groundwater plume; institutional controls). MDP4 presents an opportunity to make forward progress with long-term stewardship and eventually reach site closure. For sites where there is still some ongoing remediation, such as passive technologies or monitored natural attenuation, contaminant concentrations may steadily decrease, albeit slowly. In these cases, the site manager should periodically ask whether there is still residual contamination present in amounts that preclude unrestricted use and thus pose a human or ecological health risk. If not, it may be possible to proceed to site closure, as described below.

MDP4 introduces an opportunity to modify actions at those sites where residual contamination persists above unrestricted use levels. Periodically during long-term stewardship, the site manager should ask whether there is a reason to revisit the chosen remedy. If the answer is yes, then ASM affords the flexibility to optimize, modify, or replace a remedy with something that may be more effective in removing the residual contamination. As shown in Figure 2-7, this may eventually lead (via remedy selection and implementation) to site closeout if the new treatment is successful. There are several reasons that site managers should reconsider remedies in place during long-term stewardship. State law may require complete restoration; therefore, attaining these goals may be mandatory. Also, considerable cost savings may be possible if a new technology can alleviate the need for continual monitoring and/or maintenance. In addition, there are substantial economic benefits to returning a site to unrestricted land uses. This path is most likely to succeed if site managers stay abreast of recent developments in new treatment technologies, as discussed previously under the evaluation and experimentation track. As discussed in Chapter 6, the five-year review process currently does not support reconsideration of remedies during long-term stewardship if the remedies are maintaining protectiveness of human health and the environment.

It is important to clarify the meaning of the term “site closeout” in Figure 2-7. The term can have many connotations within the hazardous waste cleanup world, and may imply sites that have been cleaned up to, for example, industrial land uses. Indeed, at some sites that are, from EPA’s standpoint, considered “closed,” “deleted from the National Priorities List (NPL)⁸,” and perhaps redeveloped, a variety of remedial actions

(some very expensive) may continue until or unless the site meets unrestricted use (EPA, 2000b). Such actions include continued operation and maintenance of the remedy (e.g., the cover inspected and repaired, pumps replaced) and monitoring. Maintenance of institutional controls requires inspections and verification that land use has not changed. However, throughout this report the term “site closeout” is used to mean that residual contamination has been removed to levels below that which allows for unrestricted use, which is consistent with the DoD usage of the term. More specifically, “site closeout” refers to the “point at which the DoD will no longer engage in active management or monitoring at an environmental restoration site, and no additional environmental restoration funds will be expended unless the need for additional remedial action is demonstrated.” According to interagency guidance, “for practical purposes site closeout occurs when cleanup goals have been achieved that allow unrestricted use of the property (i.e., no further long-term monitoring, including institutional controls, is required).”

Key activities during site closeout (that are sometimes overlooked) include the decommissioning of monitoring wells, treatment systems, and pipelines, and the termination of institutional use controls. Sufficient evidence should have been collected to ensure that conditions will not be reversed in the future. For example, quantitative evidence of the absence of a reversal in site conditions—such as a rebound in groundwater plume concentrations when the pump-and-treat system is shut down—should be gathered. If this requires additional monitoring, there should be agreement on the timing of this monitoring so it is well defined and finite. Only when the above conditions are addressed to the satisfaction of the RPM, remediation team, regulatory agencies, and interested stakeholders does the site move to the status of closed.

Adaptive management approaches are now being used by a number of public and private organizations to improve the quality of their operations and decisions. Adaptive management recognizes that uncertainty is inherently present when predicting the effects of new policies and programs, and that including directed testing, evaluation, and learning as part of these programs can build the knowledge base for ongoing improvements in decisions. Like the domains of natural resource and business management where the principles of adaptive management have been applied, site cleanup planning and stewardship involve significant

uncertainty in system response. Given this, and the strong support for adaptive approaches already present in recent federal guidance on monitoring and remediation, we propose adopting ASM for site cleanup and long-term stewardship.

ASM is an iterative, flexible approach to improving federal site cleanup. It builds on recently developed guidance for the Navy (NAVFAC, 2001), but provides a much broader and more well-defined series of tasks to ensure that remediation is cost-effective. Whereas the recent Navy guidance also recommends close scrutiny of existing remedies and monitoring data, ASM goes further to suggest how to interpret the monitoring data, when to consider using new technologies, and how to reach site closure for all types of sites. The differences between current cleanup practice and ASM with regard to monitoring and data analysis, evaluation and experimentation, and long-term stewardship are elaborated in Chapters 3, 4, and 6, respectively.

A critical aspect of ASM is its call for evaluation and experimentation, and the coupling of that information with the adaptation of remedial programs so that ineffective or inefficient remedies are replaced quickly. This approach presents a way to manage uncertainty while moving forward with the cleanup process because conventional remedies can be implemented first while additional information is being gained on innovative but more risky technologies.

ASM formalizes discrete management decision periods to provide an explicit mechanism for communication (between the RPM and remediation team, the regulatory agencies, and interested stakeholders), to allow for critical evaluation of information, and to guide the determination of new management actions. MDP1 ensures that a selected remedy is indeed the right one to implement, while MDP2 details how to assess remedy effectiveness based on monitoring data. MDP3 draws upon monitoring data as well as information from evaluation and experimentation and stakeholder input to optimize, modify, or replace remedies or revise remedial goals. MDP4 provides the road map for long-term stewardship and site closure for sites where residual contamination remains in place. In addition, MDP4 suggests how to make forward progress at sites where remedies, such as pump-and-treat systems and monitored natural attenuation, require substantial financial resources for monitoring, operation, and maintenance. Feedback loops are present throughout in order to revisit different points when new information warrants such an examination. A final important feature of ASM is its applicability to sites at any stage of cleanup.

There is little more than anecdotal evidence about the difficulty or

ease which with remedies can be changed, although it is likely that there are disincentives for RPMs to optimize or change in-place remedies in some cases. ASM institutionalizes the concept of being open-minded about chosen remedies, and its success will depend on the creation of incentives to promote this mindset. The recent trend within the Navy of optimizing existing remedies and changing ineffective remedies, but without the benefit of evaluation and experimentation (see example in Box 2-7), suggests that the Navy and other federal agencies would have a need for and an interest in ASM.

Given its many discrete decision points and the evaluation and experimentation track, it is possible that ASM will be time-consuming and (over the short-term) more expensive than the current practice. Thus, full-scale ASM that includes public participation during all decision periods should be targeted to the more complex (e.g., multiple contaminants and stressors, heterogeneous hydrogeology) and high-risk sites where projected large costs are at stake. An example would be where DNAPL contamination threatens a sole source aquifer. Indeed, these are the sites where cleanup goals are not being achieved and where innovative technologies are needed to provide new avenues for treatment. A substantial number of DoD sites, including Navy sites, fall into this high-risk/high-cost category. If targeted in this manner, the ASM approach is expected to lead to an optimum solution from both a cost and performance perspective, as well as to a solution acceptable to stakeholders. The benefits of ASM are expected to be less at smaller, low-risk, low-cost remediation sites (e.g., a BTEX spill at an UST site) where there is greater certainty about the ability of remedies to reach cleanup goals.

Because of the enormous variability in site conditions across Navy facilities, it is not appropriate for this report to suggest a distinct cost basis for assessing whether or not to use ASM—for example, transactional costs should only represent a certain percentage of the total costs— although this may be possible in the future following more in-depth analysis by the Navy. Nonetheless, it is anticipated that up-front cost increases associated with implementing ASM will be balanced by the benefits of evaluation and experimentation, which include optimization of remedies and more expeditious achievement of cleanup goals. In many cases, the costs associated with ASM may be exceeded by the long-term savings that result from switching to a more efficient and effective technology or by overall life-cycle savings. These issues, along with pertinent examples, are discussed in greater detail in Chapters 4 and 6, respectively.

Finally, current understanding within the military of what and for

how long information needs to be collected, catalogued, and maintained is too inconsistent at present to fully support the ASM concept (M. Ierardi, Air Force Base Conversion Agency, personal communication, 2002). Information is currently found in many different information systems for records management, financial management, contract management, real-estate management, progress reporting, and technical data systems, most of which are not integrated. Greater understanding of the value of the information associated with the cleanup program will be needed to support ASM. Specific recommendations regarding data collection on remedy performance at federal facilities are discussed in Chapters 3, 4, and 5.

The Navy and other federal agencies should adopt adaptive site management. To our knowledge, ASM has never been formally used for hazardous waste cleanup. ASM will enable site managers to use new data and innovative technologies when they become available, both during active implementation of remedies and during long-term stewardship. The Navy is currently drafting policy that will require periodic reviews of remedies, as prescribed by the recent NAVFAC (2001) guidance on optimization (R. Kratke, NFESC, personal communication, 2003). Because ASM is broader in scope than that guidance, it will be necessary for the federal agencies to develop guidance to further define the management decision periods that are inherent to ASM.

Full-scale ASM that includes public participation during each decision period should be targeted to the more complex and high-risk sites where projected large costs are at stake. ASM is particularly appropriate for sites with multiple or recalcitrant contaminants and multiple stressors and heterogeneous hydrogeology because progress at such sites is likely to have stalled prior to reaching cleanup goals. Prior to widespread adoption, the Navy should consider pilot testing ASM at a limited number of high-risk, complex sites to allow Navy managers to better understand any transactional costs and delays that may accompany ASM implementation.

Air Force. 2001. Final remedial process optimization handbook. Prepared for Air Force Center for Environmental Excellence, Technology Transfer Division, Brooks Air Force Base, San Antonio, Texas, and Defense Logistics

Agency, Environmental Safety Office, Fort Belvoir, VA. Available at http://www.afcee.brooks.af.mil/er/rpo/rpohandbook.pdf.

Andrews, R.1997. Risk-based decision making. Pp. 208–230 In: Environmental policy in the 1990s. N. Vig and M. Kraft (eds.). Washington, DC: Congressional Quarterly.

ASTM. 1999. Standards on design, planning and reporting of ground water and vadose zone investigations (2nd Edition). Philadelphia, PA: American Society for Testing and Materials.

Ayres, R. U. 1969. Technological forecasting and long-range planning. New York: McGraw Hill.

Baecher, G. B., and C. C. Ladd. 1997. Formal observational approach to staged loading. Transportation Research Record, Annual Issue.

Brooks, M. C., W. R. Wise, and M. D. Annable. 1999. Fundamental changes in in situ air sparging flow patterns. Ground Water Monitoring and Remediation, Spring:105–113.

Buchman, M. F. 1999. NOAA screening quick reference tables. NOAA HAZMAT Report 99-1. Seattle, WA: NOAA Coastal Protection and Restoration Division.

Calabrese, E. J., and L. A. Baldwin. 1993. Performing ecological risk assessments. Chelsea: Lewis Publishers.

Christ, J. A., M. N. Goltz, and J. Huang. 1999. Development and application of an analytical model to aid design and implementation of in situ remediation technologies. J. Contam. Hydrol. 37(3–4):295–317.

Collins, J., and J. Porras. 1997. Built to last. New York: Harper Business Press.

Committee on Environment and Natural Resources. 1999. Ecological risk assessment in the federal government. CENR/5-99/001. Washington, DC: Committee on Environment and Natural Resources of the National Science and Technology Council .

Crognale, G. 1999. Environmental management strategies: the 21st century perspective. Upper Saddle River, NJ: Prentice Hall.

Crumbling, D. M., C. Groenjes, B. Lesnik, K. Lynch, J. Shockley, J. Van Ee, R. Howe, L. Keith, and J. McKenna. 2001. Managing uncertainty in environmental decisions. Environ. Sci. Technol. 35:405–409.

deGeus, A. 1997. The living company. Boston: Harvard Business School Press.

Deming, W. E. 1993. The new economics for industry, government, and education. Cambridge, MA: MIT Press.

Department of Defense (DoD). 1997. Guidance responsibility for additional environmental cleanup after transfer of real property. Washington, DC: DoD Office of the Deputy Under Secretary of Defense (Acquisition and Technology).

DoD. 2001. Management guidance for the DERP. Washington, DC: Office of the Deputy Under Secretary of Defense (Installations and Environment).

Department of Energy (DOE). 1998. Using remedy monitoring plans to ensure

remedy effectiveness and appropriate modifications. RCRA/CERCLA Information Brief, DOE/EH-413 9809. Washington, DC: DOE Office of Environmental Policy and Assistance.

DOE. 2001. Adaptive sampling and analysis programs (ASAP). Innovative Technology Summary Report DOE/EM-0592. Washington, DC: U.S. Department of Energy.

Dixit, A., and R. Pindyck. 1994. Investment under uncertainty. Princeton, NJ: Princeton Univ. Press.

Edelstein, M. R. 1988. Contaminated communities: the social and psychological impacts of residential toxic exposure. Boulder, CO: Westview Press.

Elder, C. R., and C. H. Benson. 1999. Air channel formation, size, spacing, and tortuosity during air sparging. Ground Water Monitoring and Remediation, Summer:171–181.

Environmental Protection Agency (EPA). 1989. Methods for evaluating the attainment of cleanup standards. Volume 1: soils and solid media. EPA 230/02-89-042. Washington, DC: EPA.

EPA. 1990. National oil pollution and hazardous substances contingency plan. Federal Register 56:8666–8728.

EPA. 1992. Framework for ecological risk assessment. EPA/630/R-92/001. Washington, DC: EPA Risk Assessment Forum.

EPA. 1993. Guidance for evaluating the technical impracticability of groundwater restoration. OSWER Dir. No. 9234.2-25. Washington, DC: EPA Office of Solid Waste and Emergency Response.

EPA. 1996a. The role of cost in the Superfund remedy selection process. Washington, DC: EPA.

EPA. 1996b. Soil screening guidance: user’s guide (2^nd edition). Publication 9355.4-23. Washington, DC: EPA.

EPA. 1996c. Memorandum from Stephen D. Luftig, Director, Office of Emergency and Remedial Response, to Regional Directors, Re: National Review Board (September 26, 1996). Available at http://www.epa.gov/superfund/programs/nrrb/9-26-96.htm.

EPA. 1997a. Rules of thumb for Superfund remedy selection. Washington, DC: EPA.

EPA. 1997b. Incidence and severity of sediment contamination in surface waters of the United States. Volume 1, Appendix D. Washington, DC: EPA Office of Water OST.

EPA. 1997c. Ecological risk assessment guidance for Superfund: process for designing and conducting ecological risk assessments. EPA/540-R97-006. Washington, DC: EPA Office of Solid Waste and Emergency Response.

EPA. 1998. The incidence and severity of sediment contamination in surface waters of the United States. Volume 1: National sediment quality survey. EPA 823-R-97-006. Washington, DC: EPA Office of Water OST.

EPA 1999a. Risk assessment guidance for Superfund. Volume 1–Human health evaluation manual, supplement to part A: community involvement in Superfund risk assessments. EPA 540-R-98-042. Washington, DC: EPA

Office of Solid Waste and Emergency Response.

EPA. 1999b. The benefits and costs of the Clean Air Act, 1990 to 2010. EPA-410-R-99-001. Washington, DC: EPA Office of Air and Radiation.

EPA. 2000a. Draft public involvement policy. Federal Register 65:82,335.

EPA. 2000b. Close-out procedures for National Priority Sites. EPA 540-R-98-016. Washington, DC: EPA Office of Emergency and Remedial Response.

EPA. 2001a. EPA National Remedial Review Board. Available at: http://www.epa.gov/superfund/programs/nrrb/index.htm

EPA. 2001b. Comprehensive five-year review guidance. EPA 540-R-01-007. Washington, DC: EPA Office of Solid Waste and Emergency Response.

Federal Remediation Technologies Roundtable (FRTR). 1995, 1997, 1998a, 2000, 2001. Abstracts of Remediation Case Studies, Vol. 1, March, 1995; Vol. 2, July 1997; Vol. 3, Sept 1998; Vol. 4, June 2000; Vol. 5, May 2001. Washington, DC: FRTR.

FRTR. 1998b. Remediation case studies: groundwater pump-and-treat (chlorinated solvents), Volume 9. EPA-542-R-98-013. Washington, DC: FRTR.

FRTR. 1998c. Remediation case studies: groundwater pump-and-treat (nonchlorinated solvents), Volume 10. EPA-542-R-98-014. Washington, DC: FRTR.

General Accounting Office (GAO). 2000. Environmental Protection Agency: use of precautionary assumptions in health risk assessments and benefits estimates. GAO-01-55. Washington, DC: GAO.

Gregory, R., J. Flynn, and P. Slovic. 1995. Technology stigma. American Scientist 83(3):220–223.

Hax, A. C., and D. L. Wilde II. 1999. The Delta model: adaptive management for a changing world. MIT Sloan Management Review 40(2), Reprint 4021.

Holling, C. S. (ed.). 1978. Adaptive environmental assessment and management. New York: John Wiley & Sons.

Javandel, I., and C. F. Tsang. 1986. Capture-zone type curves: a tool for aquifer cleanup. Ground Water 24(5):616–625.

Kaslow, T. W., and R. S. Pindyck. 1994. Valuing flexibility in utility planning. The Electricity Journal 7(2):60–65.

Kasperson, R. E., D. Golding, and S. Tuler. 1992. Social distrust as a factor in siting hazardous facilities and communicating risks. Journal of Social Issues 48:161–187.

Kueper, B. H., J. D. Redman, R. C. Starr, S. Reitsma, and M. Mah. 1993. A field experiment to study the behaviour of tetrachloroethylene below the water table: spatial distribution of residual and pooled DNAPL. Journal of Ground Water 31(5):756–766.

Kulatilaka, N. 1993. The value of flexibility: the case of a dual-fuel industrial steam boiler. Financial Management 22(3):271–280.

Lee, K. 1993. Compass and gyroscope: integrating science and politics for the environment. Covelo, CA: Island Press.

Lee, K. N. 1999. Appraising adaptive management. Conservation Ecology

3(2):3.

Luthy R. G., G. R. Aiken, M. L. Brusseau, S. D. Cunningham, P. M. Gschwend, J. J. Pignatello, M. Reinhard, S. J. Traina, W. J. Weber, and J. C. Westall. 1997. Sequestration of hydrophobic organic contaminants by geosorbents. Environ. Sci. Technol. 31(12):3341–3347.

Mackay, D. M., and J. A. Cherry. 1989. Groundwater contamination: pump-and-treat remediation. Environ. Sci. Tech. 23(6):630–36.

Maughan, J. T. 1993. Ecological assessment of hazardous waste sites. New York: Van Nostrand Reinhold.

Mays, L. W., and Y-K. Tung. 1992. Hydrosystems engineering and management. New York: McGraw-Hill.

McCarthy, J. M., O. F. Canziani, F, Osvaldo, N. A. Leary, D. J. Dokken and K. S. White (eds.). 2001. Climate change 2001: impacts, adaptation, and vulnerability. Report of Working Group II to the Third Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge, England: Cambridge University Press.

Merton, R. 1973. The theory of rational option pricing. Bell Journal of Economics and Management Science 4:141–183.

Montgomery, D. C. 1996. Introduction to statistical quality control. New York: John Wiley & Sons.

Nambi, I., and S. E. Powers. 2000. NAPL dissolution in heterogeneous systems: an experimental investigation in a simple heterogeneous system. J. Contam. Hydrol. 44(2):161–184.

National Research Council (NRC). 1983. Risk assessment in the federal government: managing the process. Washington, DC: National Academy Press.

NRC. 1989. Improving risk communication. Washington, DC: National Academy Press.

NRC. 1994. Alternatives for ground water cleanup. Washington, DC: National Academy Press.

NRC. 1996. Understanding risk: informing decisions in a democratic society. Washington, DC: National Academy Press.

NRC. 1997. Innovations in ground water and soil cleanup: from concept to commercialization. Washington, DC: National Academy Press.

NRC. 1999a. Environmental cleanup at Navy facilities: risk-based methods. Washington, DC: National Academy Press.

NRC. 1999b. Downstream: adaptive management of Glen Canyon Dam and the Colorado River ecosystem. Washington, DC: National Academy Press.

NRC. 2001a. A risk management strategy for PCB-contaminated sediments. Washington, DC: National Academy Press.

NRC. 2001b. Assessing the TMDL approach to water quality management. Washington, DC: National Academy Press.

NRC. 2002. The Missouri River ecosystem: exploring the prospects for recovery. Washington, DC: National Academy Press.

NRC. 2003. Bioavailability of contaminants in soils and sediments: processes,

tools, and applications. Washington, DC: National Academy Press.

Navy. 2000. Naval Air Station Pensacola. Optimization of remedial action operation to treat chlorinated hydrocarbons in groundwater. Available at http://enviro.nfesc.navy.mil/erb/erb_a/support/wrk_grp/raoltm/case_studies/rao_pensacola.pdf

Navy. 2001. Memorandum from Goodman, S. W., Re: Policy on land use controls associated with environmental restoration activities. January 17^th.

NAVFAC. 2000. Guide to optimal groundwater monitoring. Prepared for the Naval Facilities Engineering Research Center by Radian International. White Rock, NM: Radian International.

NAVFAC. 2001. Guidance for optimizing remedial action operation (RAO). Special Report SR-2101-ENV. Prepared for the Naval Facilities Engineering Service Center. Research Triangle Park, NC: Radian International.

Nonaka, I., and H. Takeuchi. 1995. The knowledge creating company. New York: Oxford Univ. Press.

O’Brien, M. 2000. Making better environmental decisions: an alternative to risk assessment. Cambridge, MA: MIT Press.

Peyton B. M., T. P. Clement, and J. P. Connolly. 2000. Modeling of natural remediation: contaminant fate and transport. Pp. 79–120 In: Natural remediation of environmental contaminants: its role in ecological risk assessment and risk management. M. Swindoll, R. G. Stahl, and S. J. Ells (eds.). Pensacola, FL: SETAC Press.

Pierce, F. J., and W. E. Larson. 1993. Developing criteria to evaluate sustainable land management. Pp. 7–14 In: Proceedings of the eighth international soil management workshop: utilization of soil survey information for sustainable land use, May 3, 1993. J. M. Kimble (ed.). Lincoln, NE: USDA Soil Conservation Service National Soil Survey Center.

Pignatello, J. J., and B. S. Xing. 1996. Mechanisms of slow sorption of organic chemicals to natural particles. Environ. Sci. Technol. 30(1):1–11.

Pirie, R. B.1999. Memorandum 99-02—Land Use Controls. 25 May 1999.

Pittinger, C. A., R. Bachman, A. L. Barton, J. R. Clark, P. L. deFur, S. J. Ells, M. W. Slimak, R. G. Stahl, and R. S. Wentsel. 1998. A multi-stakeholder framework for ecological risk management: summary from a SETAC Technical Workshop. Environmental Toxicology and Chemistry Supplement 18(2):1–20.

Presidential/Congressional Commission on Risk Assessment and Risk Management. 1997a, b. Framework for environmental health risk management, volumes 1 and 2. Washington, DC: U.S. Government Printing Office.

Renn, O. 1999. A model for an analytic-deliberative process in risk management. Environ. Sci. Technol. 33(18):3049–3055.

Sexton, K. 1999. Setting environmental priorities: is comparative risk assessment the answer? Pp. 195–219 In: Better environmental decisions: strategies for governments, businesses, and communities. K. Sexton, A. A. Marcus, K. W. Easter and T. D. Burkhardt (eds.). Washington, DC: Island Press.

Shindler, B., and K. A. Cheek. 1999. Integrating citizens in adaptive management: a prepositional analysis. Conservation Ecology 3(1):9.

Simpson, R. D., and N. L. J. Christensen. 1997. Ecosystem function and human activities: reconciling economics and ecology . New York: Chapman and Hall.

Slovic, P.1993. Perceived risk, trust, and democracy. Risk Analysis 13(6):675–681.

Sokol, R. C., C. M. Bethoney, and G. Y. Rhee. 1998. Effects of Aroclor 1248 concentrations on the rate and extent of polychlorinated biphenyl dechlorination. Environ. Toxicol. Chem. 17(10):1922–1926.

Stahl, R. G., Jr., R. A. Bachman, A. L. Barton, J. R. Clark, P. L. deFur, S. J. Ells, C. A. Pittinger, M. W. Slimak, and R. S. Wentsel. 2001. Risk management: ecological risk-based decision-making. Pensacola, FL: SETAC Press.

Stahl, R. G., Jr., C. A. Pittinger, and G. R. Biddinger. 1999. Developing guidance on ecological risk management (guest perspectives). Risk Policy Report May 14:38–40.

Suter, G. W. 1993. Ecological risk assessment. Chelsea: Lewis Publishers.

Travis, C. C., and J. M. Macinnis. 1992. Vapor extraction of organics from subsurface soils: is it effective? Environ. Sci. Technol. 26(10):1885–87.

Trigeorgis, L. 1996. Real options: managerial flexibility and strategy in resource allocation. Cambridge, MA: MIT Press.

U.S. Army Corps of Engineers (USACE). 1996. Risk assessment handbook. Volume II: environmental evaluation. EM 200-1-4. Washington, DC: U.S. Army Corps of Engineers.

Walters, C. 1986. Adaptive management of renewable resources. New York: Macmillan.

Walters, C. 1997. Challenges in adaptive management of riparian and coastal ecosystems. Conservation Ecology 1(2):1.

Walters, C. J., and C. S. Holling. 1990. Large-scale management experiments and learning by doing. Ecology 71: 2060–2068.

Weber, W. J., P. M. McGinley, and L. E. Katz. 1991. Sorption phenomena in subsurface systems: concepts, models and effects on contaminant fate and transport. Water Research 25(5):499–528.

Williams, B. K. 2001. Uncertainty, learning, and the optimal management of wildlife. Environmental and Ecological Statistics 8:269–288.

Woodhouse, E. J. 1995. Can science be more useful in politics? The case of ecological risk assessment. Human and Ecological Risk Assessment 1(4):395–406.

Zeiss, C., and J. Atwater. 1991. Waste disposal facilities and community response: tracing pathways from facility impacts on community attitude. Can. J. Civ. Eng. 18(1):83–96.