Ranking Hazardous-Waste Sites for Remedial Action (1994)

The need for a unified national approach for setting remedial action priorities for sites contaminated by hazardous substances became evident during the committee's review of the different approaches currently used. Many steps in the present processes for setting priorities are not open to public scrutiny, and the ranking models used in those processes that were reviewed by the committee often lacked sufficient scientific rigor and validation. The situation calls for the development of a scientifically and environmentally sound, publicly acceptable, and consistent process that is commensurate with the enormous costs required for adequate site remediation (let alone complete restoration). This chapter discusses the advantages and disadvantages of a unified national process for setting priorities and proposes one such unified process.

During the past 15 years, a set of complex institutional arrangements have evolved in the United States for selecting and managing hazardous waste-site remediations. EPA, DOD, DOE, the states, and legally responsible parties all play major roles. Other federal agencies, local governments, environmental and legal consultants, the mass media, professional and industrial organizations, the judicial courts, and environmental advocacy groups are also important participants.

At the present time, there is no consistent relationship between the hazard present at a site and the process by which it is screened and evaluated for remediation. For example, EPA works closely with DOE and the states to develop plans to remediate DOE sites. Other sites are the responsibility of a single agency. Clearly, the processes involved in evaluating and remediating sites far outstrip the relatively simple model that the creators of Superfund had envisioned.

The remediation process is also much more expensive than expected. The number of high-priority sites is already three times more than the 400 required by the original Comprehensive Environmental Resource, Compensation, and Liability Act (CERCLA) of 1980, and current cost estimates in the hundreds of billions of dollars for remediating all active and inactive contaminated sites during the next 30 years dwarfs the original and amended CERCLA authorization of about $10 billion. The committee's analysis of the EPA, DOD, and DOE ranking models shows major differences in the history of site ranking model development, differences in the underlying logic and science supporting model development and testing, and differences in the use of science-based results to allocate scarce resources. In short, each agency has developed its own unique protocol. However, the use of inde-

pendent unique processes might not be in the best interests of the United States as a whole, especially during a period of pressure to accomplish more with fewer resources.

The committee recommends that the United States considers a common process of scientific analysis of sites to replace the existing multiplicity of approaches. Three alternative strategies for reducing inconsistencies have been identified, based respectively on: greater consultation, scientific consistency, and decision-making consistency. Greater consultation is the least intrusive to existing agency approaches. EPA, DOD, and DOE would form an interagency task force to review existing mathematical site-ranking methods and determine how the agencies can better share data, expertise, quality control, validation procedures, and other scientific and mathematical information. For example, data availability and data quality are major concerns. Any agreement about the kind of data that should be gathered and how its quality should be addressed can only be helpful to all agencies in the long run. The three agencies would also share their informal processes for including local social and economic concerns and their processes for communicating with tribal, state and local governments, interested parties, and the general public. Greater consultation would change existing site-ranking, decision-making, and budgetary processes only if and insofar as the agencies agreed to make changes and the federal administration agreed to the change.

The second alternative, a strategy of scientific consistency, requires that each site be subjected to the same scientific protocol for evaluating health and safety, environmental impact, and economic costs and benefits. For example, whatever scientific protocol is applied to a DOD site in Wyoming would also be applied to a DOE site in Maryland and to an EPA site in California. The committee believes that such use of a unified scientific approach would make the scientific input into the political process more explicit and more thorough. This uniform scientific process would be embedded as a common component in each agency's

larger process of deciding on resource allocations for site identification, ranking, and remediation. A unified national approach that standardizes the scientific elements of remediation decision-making would not replace or diminish the political parts of the process—that is, the parts that require bargaining among the major parties.

Decision-making consistency defines the third and most unifying alternative, incorporating scientific consistency, it goes further to add process and geographical consistency. All agencies would apply the same scientific protocol to each of its sites and would allocate remediation resources on the basis of the outcome of that protocol. In other words, resources would not be influenced by which agency was responsible for the site. For example, the remediation of a solvent spill on factory grounds in Illinois would be treated in the same way as solvent spills at a military base in Arizona or at a DOE facility in Ohio. Priority would be assigned by a central interagency group, not exclusively by the parties currently charged with remediating the site. The committee recognizes this alternative would involve major reorganization of responsibility, authority, and budgetary resources among the three major federal agencies charged with cleanup of hazardous waste sites.

Although all three approaches merit consideration, this chapter focuses primarily on scientific consistency because the committee's charge and expertise concentrate on the scientific factors that influence hazardous waste site assessment and management. First, the advantages and disadvantages of a unified national scientific and decision-making approach will be discussed from five policy perspectives, and then a particular proposal for a unified national priority-setting process will be presented.

The five policy perspectives addressed in the following discussion of advantages and disadvantages of a uniform scientific approach to aid in decision making for site remediation are: protection of health and the environment, investment of funds, organization acceptability, consistency, and adaptability.

The major advantage of requiring scientific consistency is that it would focus previously fragmented efforts, leading the best scientists and policy analysts to a concerted collaboration in developing the models necessary to give the most credible estimates of health, environmental, and welfare (e.g., land value) impacts and costs. This should, in turn, lead to the best scientific protocol now possible as a basis for decision-making, and to the best possible programs of data gathering and research for continued improvement of the protocol. Scientists from EPA, DOD, DOE, and the Agency for Toxic Substances and Disease Registry (ATSDR) would all be involved in selecting the best scientific approach for estimating risk to human health, human welfare, and the environment. For example, instead of each agency heading in its own direction to assess the impacts of contaminants on water quality and on the people who drink the contaminated water, the agencies could combine their efforts as well as get input from scientific experts outside the agencies and from public interest groups. They could start by scrutinizing their existing approaches in order to isolate those discrepancies in the assumptions, values, and scientific methods used in the models that result in different ratings and rankings for the same set of sites (see Chapter 8). They could establish a uniform definition of "site," join together to build a formal site discovery program, and formulate uniform criteria for an emergency cleanup. A better decision-making protocol would also help ensure that protection for human health and the environment is adequate.

A disadvantage is that the existing methods embody the investment of a great deal of time and money. Although these methods would undergo gradual refinement, the development of a national protocol might lead to a more rapid displacement for some of them. Resources would be needed to retrain technical personnel to implement methods of the protocol.

Only one of the methods considered by the committee, DOE's priority-setting approach, explicitly incorporates a consideration of costs and benefits. In light of the enormous costs of alternative site-remediation processes, a single consistent national process that formally includes representations of economic costs and benefits would be advantageous to decision-makers. In essence, such a process would recognize the reality that costs and benefits are always factored in some way into decisions. The committee believes that an explicit treatment of costs and benefits of remediation would increase the credibility of the process by providing estimates that could be compared with actual site costs and benefits—in a sense, a kind of cost and benefits accounting. A clearly documented costs and benefits protocol element should greatly reduce the possibilities for inadvertent or intended skewing of cost and benefit considerations for reasons that have nothing to do with hazard or remediation outcomes.

One disadvantage is the possibility of principled opposition by those who believe that economic costs and benefits should not be a consideration in protecting public health and environment, and would be distressed by such explicit consideration of costs and benefits. Also, it would take time to build an economic protocol that would gain broad acceptance to practitioners, interested parties, and theoreticians.

There are obvious organizational disadvantages for the major federal agencies, and perhaps their established consultants, all of whom would have some incentive to resist the imposition of a national scientific approach. This resistance could lead to a long

and drawn out effort to develop a unified system. The existing priority-setting processes have led to legal agreements that already promise remediation, so that signatories would feel threatened by the imposition of an approach that may be perceived as trying to set the clock back.

The previous point is countered by assessments showing that limitations on financial resources will probably make it infeasible to fulfill all existing legal agreements concerning site remediation. If and when this situation becomes recognized reality, a single consistent process will be invaluable in aiding the major federal agencies to bargain with each other, as well as with the states, local governments, private sectors, citizens groups, and other stake-holders because all will be negotiating from the same data base and on a more level playing field. Ideally, such a uniform approach would lead the federal agencies to develop a joint strategic plan for remediation under a variety of resource-constrained scenarios. In other words, it would lead to a hierarchical set of objectives to accommodate financial and technical limitations and advances.

There appears to be geographical inconsistency in the distribution of site-remediation resources. Some states have strong environmental protection programs and others do not. The result is a situation where not every American community and ecosystem is protected on a consistent basis by the current priority-setting or resource-allocation methods. A major advantage of the uniform scientific approach is that every analysis will treat every person the same and every forest the same, regardless of whether they are located in an urban area of New Jersey or Louisiana, or a rural area of Maine or Arizona. Going further, decision-making consistency

would allocate remediation resources on the basis of costs, benefits, and need for cleanup rather than on the basis of the ability of a responsible party, state, business, or federal agency to pay.

A disadvantage stems from the fact that many state governments under the policy of new federalism have been given back much of the authority to choose how they spend their limited resources. Many can be expected to oppose a national process that would ascribe high benefits compared with costs for remediation that is not a high priority for them. Likewise, under decision-making consistency, federal agencies can be expected to resist a decrease in their authority to decide priorities for their sites.

Some state opposition might be overcome by allowing states to use their own method of assessing cleanup priorities as long as they also take into account the results produced by the national method as a baseline. In other words, states would, at a minimum, use the national method. They could also use their own method and therefore would have two sets of information upon which to inform their discussion. In addition, states could be given the right to change the weights assigned to different environmental elements of models used in a national priority-setting approach (e.g., groundwater, forests, and housing). Each state could change the outcome for its sites by changing the weights of importance, but these changes would have to be explicit, documented as to justification and process of derivation, and open to public scrutiny. A similar accommodation might be made for federal agencies. The state and federal agencies would receive the same information and be able to use it to guide, but not dictate, their final decisions.

The committee believes that a national approach to setting pri-

orities would better accommodate changes in the scientific, technological, economic, and political processes in the United States and abroad than do the existing multi-organizational processes. However, some states are more innovative and adaptable than is the federal government. These few states would have the disadvantage of waiting for a federal process that would probably be slower. On the other hand, these states would doubtless take the lead and press the federal government and other states for change.

Much has been learned about evaluating risks from site contamination over the past 10 years, and new technologies have been developed that make site remediation more economical. In addition, the limits of the nation's technical and economic ability to remediate sites have become more widely recognized. Because of the enormous cost of remediation to attain unrestricted land use in many instances, and the technical near-impossibility of attaining this goal in others, there is growing opposition to the original goal of complete restoration of sites (see Chapter 1).

Furthermore, even though remediation is likely to go on for decades, many resources will be allocated throughout the 1990s as well as beyond. If the United States is ever to adopt a uniform national scientific and decision-making process, it makes sense to do it soon.

The three major federal agencies involved in site restoration use different approaches for evaluating site risks and setting site cleanup priorities. It is extremely difficult to compare the consistency in degree of cleanup and level of protection being provided by the different agencies. Nor have the risk-based procedures currently in use been adequately validated, partly because the effort, ex-

pense, and time to do so appears excessive. However, the committee considers such validation attempts essential for a program presenting such enormous costs, and thus demanding maximal assurance that funds are being spent wisely. With use of a unified scientific approach by all agencies, the cost and effort in development and validation can be shared, and a national consensus on the directions to be taken becomes more supportable and likely.

The committee proposes serious consideration of a three-tiered unified procedure for setting priorities for hazardous waste site remediation. The approach proposed draws heavily on procedures already being used either explicitly or implicitly by state and federal agencies, including EPA, DOD, and DOE, so that no radical change in thinking or development is required. A general outline of this three-tiered approach is presented in Figure 9-1.

The first tier of the unified approach embodies a procedure for screening candidate hazardous waste sites. Here, a site is evaluated simply to determine whether it (1) should be moved to the second tier for more detailed characterization, (2) should be eliminated from further consideration, or (3) should be held for further analysis or characterization so a decision can be made in a timely manner to move it to Tier Two or eliminate it from further consideration. The decisions in Tier One would necessarily be based on limited data regarding the degree of hazard to human health and the environment. The Hazard Ranking System of EPA and DOD's Defense Priority Model are examples of models that have been developed for the type of evaluations that would be performed in Tier One.

In Tier Two, health and environmental risks and remediation costs are estimated. A detailed site investigation would be conducted to determine the extent of contamination present at a site and the various environmental media and populations that might be affected by the contaminants. The data obtained here should be adequate to conduct a formalized assessment of the relative risk posed to human health and the environment. The objective here

is to obtain, through consensus, a single well-documented and validated model that not only provides a relative ranking of sites based upon risk to human health and the environment, but also denotes the risk reduction achieved by alternative levels of remediation. The model to be used for this purpose should be tested with validation studies at actual sites, and the uncertainty in the model output should be characterized so that the degree of confidence in the site rankings and the rankings of effectiveness for alternative levels of remediation can be estimated and considered. As an example, see the discussion of such analysis in Chapter 5. Such a model could also help to track progress during remediation.

The site data obtained under Tier Two should be adequate to perform a cost evaluation for site remediation. Here, present value costs would be estimated for each of three levels of remediation. The first level would involve sufficient remediation to contain the hazardous contaminants so that they would not present significant risk to human health and the environment. A no-action alternative might be equivalent to this level at some sites; at other sites, all contaminants would not necessarily be removed, and land use controls and restricted access might be required. The second level of remediation considered would restore the site to the point where no land use restrictions would be necessary. This level of cleanup would be equivalent to "permanence," in that no continuing costs for contaminant containment would be required. The third level of cleanup would be more extensive, comparable to returning the site to precontamination quality. Present value costs to achieve each of these three levels of control would be estimated. The three levels of control suggested here are similar to the three levels considered by Russell et al. (1991) in their evaluation of costs to the nation of hazardous waste remediation.

In Tier Three, a ranking resulting from the Tier Two assessment of risk to human health and the environment, together with the cost estimated for each of the three levels of control, would be

determined by an independent process. By giving a ranking to the site, a determination could be made regarding what sites to address first and what levels of control to instigate. The committee does not recommend a particular framework for doing this, but dearly one is needed. This process might involve some mathematical formalization, but that should be coupled with broader political, social, and economic considerations. Decisions at this level should be made in a well defined process. The ranking process could be centralized or decentralized, and individuals or groups who have an explicit objective in mind might contribute to the process. Today, such decisions on ranking and site cleanup level are generally made by an informal and ill-defined process that is peculiar to the agency involved. The process used needs to be more explicit than is the current practice, so that national resources can be expended in a more open and cost-effective manner.

There are several advantages to such a three-tiered unified national scientific approach. First, this process for priority setting proposed is similar overall to approaches currently being used among federal and state agencies. Thus, no radical change in thinking is required. It does not involve any greater degree of complexity than currently used procedures, and indeed would take advantage of knowledge gained from application of current processes. The main advantage is that the same procedure would be used in Tier One and Tier Two by all agencies. This would allow greater effort to be placed in evaluating and improving the scientific basis for the mathematical procedures being used and would be more cost efficient for determining how the models perform with respect to their intended purposes, for evaluating the validity of the approach used, and for determining the sensitivity of the model to data inputs. The overall cost for developing good model documentation and acquiring appropriate input coefficients would be reduced. The steps required for model development and validation can therefore be more readily implemented in a unified ap-

proach, so that the result can attain the credibility necessary for broad confidence in its use.

The three levels of cost for site remediation distinguished under Tier Two would allow better judgments to be made concerning the degree of cleanup that should be pursued at a given site. For example, suppose for a given site (Site One), the estimated cost for the first level of cleanup is $2 million and for the second level it is $3 million. For a second site (Site Two), the estimated cost for the first level of remediation is $2 million and for the second level it is $100 million. If the risks posed by the two sites are similar and funds available for remediation are limited, it would be understandable and reasonable to suggest that Site One be cleaned at least to the second level, while the first level of cleanup may be the best alternative for Site Two.

The benefits of cleanup would also be calculated. The derision for the two sites in the example above might change if the benefits of the second level remediation of Site Two were $400 million compared to $100 million in costs.

Assuming that all three levels of cleanup would be sufficiently protective of human health and the environment, the degree of permanence and the extent of land use restrictions might be quite different. As with any priority-setting approach, a continuing program would be needed for discovering new sites and for evaluating changes over time in the potential indicated hazards from discovered sites not undergoing remediation. Such a monitoring program would require efforts commensurate with the potential hazard of specific waste sites. For example, closer monitoring of "no-action" sites and "containment-only" sites would be needed compared with monitoring unlisted sites. Such a program would be greatly facilitated by a uniformly-applied and validated model.

Because the approach proposed by the committee is not dependent on a certain period, it could be used to incorporate considerations of long-term risk. For example, longer periods could be used in the analysis performed under Tier Two of the

recommended approach where risk is assessed for each site. Also, sites rejected from further consideration in Tier One could be fed back into the process periodically. However, how long a site should be monitored and assessed for risk and the timing of remedial actions are issues that are beyond the charge of this committee, but should be the subject of a separate study. OTA (1989) discussed some of the benefits and concerns regarding a priority-setting process that explicitly addresses future risk.

Under the proposed three-tiered system, hazards posed by various sites and the relative costs for different degrees of cleanup would be provided explicitly, and in a manner that is understandable by the public and decision-makers. This would make it easier for decision-makers and affected parties, both public and private, to arrive at rational decisions for setting priorities and levels for cleanup of contaminated sites. If used as a rational process for isolating and elucidating data limitations and scientific uncertainties, and for articulating and implementing societal values, it will enhance our ability to make prudent decisions.