1
HAZARDOUS-WASTESITEPROBLEMSINTHEUNITEDSTATES

INTRODUCTION AND CHARGE TO THE COMMITTEE

This report discusses the ranking methods and overall priority-setting approaches used by or developed for the U.S. Environmental Protection Agency (EPA), the U.S. Department of Defense (DOD), the U.S. Department of Energy (DOE), and some state governments to choose sites for remediation among the tens of thousands of hazardous-waste sites. As complex components of the priority-setting approaches, the ranking methods combine available information about waste sources; air, water, and soil pathways for contaminants; toxicants; and population and resources at risk to attempt to produce integrated numerical values. The resulting rankings can be used together with other social, economic and political factors to set priorities for cleanup. Reflecting the complexity of the



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



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

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

OCR for page 23
Ranking Hazardous-Waste Sites for Remedial Action 1 HAZARDOUS-WASTESITEPROBLEMSINTHEUNITEDSTATES INTRODUCTION AND CHARGE TO THE COMMITTEE This report discusses the ranking methods and overall priority-setting approaches used by or developed for the U.S. Environmental Protection Agency (EPA), the U.S. Department of Defense (DOD), the U.S. Department of Energy (DOE), and some state governments to choose sites for remediation among the tens of thousands of hazardous-waste sites. As complex components of the priority-setting approaches, the ranking methods combine available information about waste sources; air, water, and soil pathways for contaminants; toxicants; and population and resources at risk to attempt to produce integrated numerical values. The resulting rankings can be used together with other social, economic and political factors to set priorities for cleanup. Reflecting the complexity of the

OCR for page 23
Ranking Hazardous-Waste Sites for Remedial Action ranking procedures, much of the report addresses technical issues such as accuracy of toxicity data bases, appropriateness of mathematical operations performed and applicability to various kinds of waste sites. The National Research Council's Committee on Remedial Action Priorities for Hazardous-Waste Sites was formed to assess the principal methods that federal and state agencies are using or developing to rank sites for remedial priority. The committee was asked to consider, as part of its analysis, issues such as the following, and others that it considers relevant: the intended technical and policy purposes and actual use of the method in the ranking-decision process, and its effectiveness in achieving the intended purposes; the types and levels of uncertainty of input data resulting in the method's limitations; the method's assumptions (explicit and implicit) and the appropriateness of the assumptions for the method's intended purpose; source, magnitude, and treatment of significant uncertainties in each method; the sensitivity of the resulting score to the method's computation process; the method's flexibility for follow-up evaluation of site assessments (e.g., regarding changes in risk) or for comparative analyses of the costs and effectiveness of remediation techniques. The committee was asked to make recommendations regarding information and research needed to establish standards of performance and consistency for nationally applicable ranking methods for hazardous-waste sites and to provide a basis for refining existing methods to improve the decision process for hazardous-waste site management and remediation in the future. As background before addressing the scientific content of the

OCR for page 23
Ranking Hazardous-Waste Sites for Remedial Action methods, this chapter briefly reviews the recent history of the issue. Hazardous-waste site-remediation programs began with the goal of quickly cleaning up a limited number of highly contaminated and highly politicized sites attributed to American industry. As will be seen, however, these programs have evolved into endeavors whose full implementation in their present forms might not be possible because of technical limitations and costs. The chapter addresses the EPA Superfund program, the DOD and DOE remediation programs, and other federal and state programs. Their mandates, program sizes, and cleanup costs are addressed. Priority setting is then discussed. In short, this chapter tries to answer two questions: (1) Why is the hazardous-waste site-management problem so much more complex than it was perceived to be a decade ago? (2) What role might a science-based ranking system play in the overall priority-setting process, considering the increasing political, legal, and economic pressures being placed on decision makers? This chapter is not intended to present a comprehensive enumeration of all the legal mandates and political forces. (Some additional details are provided in Chapters 4-6 and in the citations therein.) Rather this chapter focuses on the programs that set the tone for hazardous-waste site remediation in the United States. The committee has tried to present an objective and balanced sample of the issues. U.S. ENVIRONMENTAL PROTECTION AGENCY SUPERFUND PROGRAM EPA is directly responsible for or indirectly involved in hazardous-waste cleanups of private and government sites. It is the agency of last resort if no party assumes responsibility for site cleanup. Some sites no longer operate; some are still operating;

OCR for page 23
Ranking Hazardous-Waste Sites for Remedial Action and some facilities have both operating and closed sites. This section focuses on the Superfund program because of its historical, legal, and symbolic importance. Other EPA responsibilities are briefly reviewed elsewhere in the chapter. Initial Mandate and Funding In 1980, Congress mandated EPA to cleanup abandoned hazhazardous-ardous waste sites. The Comprehensive Environmental Response, Compensation, and Liability Act (CERCLA) of 1980 (P.L. 96-510) established a $1.6 billion "Superfund" to identify sites contaminated by hazardous waste, to determine responsible parties, and to finance cleanups when responsible parties could not. EPA was required to develop a National Priority List (NPL) with a minimum of 400 sites for cleanup. At least one site from each state had to be on the NPL. Superfund was extended for 5 years in 1986 by the Superfund Amendments and Reauthorization Act (SARA)(P.L. 99-499), and another $8.5 billion was added to finish the job. In addition, EPA was given guidance about additional risks to consider, research needs, modifications to its hazard-ranking system, and relationships with states and with the U.S. Agency for Toxic Substances and Disease Registry (ATSDR). (Some of that guidance is reviewed in this chapter. Chapter 4 provides a detailed presentation of priority setting within EPA's Superfund program.) Original Hazard-Ranking System EPA defines a hazardous-waste site in terms of the risks presented to human health and the environment at a specific location. The extent of the site depends upon the extent of the con-

OCR for page 23
Ranking Hazardous-Waste Sites for Remedial Action tamination. As assessment and remediation move forward, a site overseen by EPA is likely to be split into "operable units" that each require a specific remediation. Later, once a Superfund Record of Decision is filed and remediation has occurred, any natural resource damage must be reevaluated. For that purpose, a variety of sites or operable units can be combined for consideration by the Natural Resources Trustees. CERCLA required the identification of 400 sites for an NPL, but provided very limited guidance for a hazardous-site ranking system. Section 105(8)A mentioned the following as relevant factors: population at risk, potential for drinking-water contamination, direct human contact, and destruction of sensitive environments. EPA retained the MITRE Corporation to develop a method for ranking and choosing 400 NPL sites. Because the Hazard Ranking System (HRS) that MITRE developed has had a major impact on priority-setting approaches used by other government agencies, a brief review of the original HRS (Chang et al., 1981) follows; a detailed presentation of the revised HRS is provided in Chapter 4. MITRE's scientists reviewed existing methods of rating sites' relative hazards. They concluded that the existing models focused on water pollution impact, had no air pollution, soil pollution, or direct human contact elements, and had no method of integrating the different impacts (Chang et al., 1981). The need to meet CERCLA's mandate clearly demanded a state-of-the-art advance in priority-setting models. The original HRS expanded consideration of hazardous-waste site impacts to five components: groundwater, surface water, air, fire and explosion, and direct contact. The model required data on waste characteristics, quantities, releases, and targets for each of the five components for each site; thus the original HRS provided a method of integrating site-related data into a single site score (Chang et al., 1981). In 1981, a decade before this National Research Council committee was convened to study priority-setting methods, the

OCR for page 23
Ranking Hazardous-Waste Sites for Remedial Action MITRE staff identified limitations of their work that remain as unresolved issues today (see Chapter 4 for greater detail). The model is not amenable to an economic analysis because the hazard-ranking scores are only relative on a scale from 1 to 100. A site score of 60 is not twice as bad as one of 30. Second, quality control of the data and sensitivity tests of the results (with respect to data and model uncertainties) were limited. Third, the model did not take into account socioeconomic factors; these were to be evaluated by EPA after the model had been used to rank sites. The required 400 sites were selected by using the original HRS. The cutoff score for the top 400 sites turned out to be 28.5. With only slight modifications described in Chapter 4, that cutoff score has continued to be the criterion used to determine if a site should be placed on the NPL The use of a scoring model such as the HRS to dichotomize sites has been criticized (OTA, 1989; Hird, 1990). The major concern is that adequate procedures be used to ensure that dichotomizing sites will not exclude some sites that should be on the NPL Program Size and Cost: EPA Superfund has become a massive program. One reason is that the number of sites that require cleanup has proved to be much larger than the initial estimates. In 1977, 1 year before Love Canal made national headlines and 3 years before the passage of the Superfund, EPA reported findings of hazardous conditions at 421 disposal sites (EPA, 1977). The principle problem was clearly chronic pollution of water, not imminence of catastrophic explosions and fires. Groundwater pollution was observed at 61% of the sites and surface water pollution at 40%. Fires and explosions were reported at 4% One year later, after the identification of contamination at Love

OCR for page 23
Ranking Hazardous-Waste Sites for Remedial Action Canal, Thomas Jorling, then EPA's assistant administrator for water and waste management, addressed the concern that government might be called upon to fund other Love Canal-type cleanups by requesting that the ten EPA regional administrators estimate the numbers of hazardous-waste sites and those that might be imminent hazards (Greenberg and Anderson, 1984). The response was 32,000 sites, of which 838 that might be imminent hazards. The estimating methods were, however, imprecise and nonuniform across the regions. In 1979, Fred C. Hart Associates, under contract to EPA, used a more systematic method to estimate the number of hazardous-waste sites at 50,664, of which 2,027 might pose a significant threat (Greenberg and Anderson, 1984). After CERCLA was passed, EPA began to keep records of the number of sites identified as hazardous or potentially hazardous. In January 1983, for example, 13,392 such sites had been reported (Greenberg and Anderson, 1984). By June 1986, the year Superfund was reauthorized, EPA had 24,269 sites in its inventory with 951 on the NPL (Conservation Foundation, 1987). That is, in 3 years the number of identified sites had increased over 80% and the number of priority sites already was more than twice the original 400 required by Superfund legislation. As of April 1994, EPA reported 37,987 sites in its CERCLIS (CERCLA Information System) inventory of potentially hazardous sites. The number of sites will almost surely continue to increase. The question is how many more will be added to the NPL. EPA expects to add about 100 new sites a year for the foreseeable future (S. Caldwell, EPA, pers. comm., January 1992). EPA (1990a) estimated that the number of NPL sites could reach 2,000. The Congressional Office of Technology Assessment (OTA, 1985, 1989) expects there will be 4,000 NPL sites by the year 2000 and that the NPL might eventually contain 10,000 sites (EPA's estimate of 2,000 sites; another 1,000 currently active hazardous-waste sites; 5,000 solid waste sites; and 2,000 sites to be found because of bet-

OCR for page 23
Ranking Hazardous-Waste Sites for Remedial Action ter site identification). Russell et al. (1991) predicted a range of 2,100 to 6,000 NPL sites. Initial underestimation of the average cost of cleaning up sites is another major reason why the Superfund program now appears so much more costly. Enacted on December 11, 1980, the $1.6 billion trust fund was to come from taxes imposed on oil (raw materials) and on 42 specific compounds. About 13% was to come from general revenues (Superfund Fact Sheet, 1981). This formula was a matter of concern for the chemical industry, but appeared to be a manageable way of collecting $1.6 billion. Speaking in early 1981, Senator Robert Stafford (1981) estimated the average cost of cleaning up a site at $3.6 million; which would make the Superfund's $1.6 billion enough to cleanup the stipulated 400. In 1990, EPA reported an average cost of $26 million to cleanup an NPL site, yielding a projection of $27 billion to cleanup existing NPL sites (EPA, 1990a,b). The estimate excludes costs for remediating sites not listed at the end of fiscal year 1990. Other estimates place the cost closer to $100 billion or more for the EPA program. OTA (OTA, 1989; Passell, 1991) estimated the cost of cleaning up 4,000 NPL sites to be $80 to 120 billion, and a potential cost of $500 billion spread out over 30 to 50 years to cleanup 10,000 sites without major changes in the program goals and technological innovations. The "best guess" estimate of Russell et al. (1991) is that $151 billion will be spent from 1990 to 2020 to remediate 3,000 non-federal NPL sites, at an average site cleanup cost of $5 to 15 million. The general picture emerging from these more recent projections is that Superfund cleanup will cost the U.S. 10 to 50 times the amount of money allocated to EPA under CERCLA and its reauthorization. The Congressional Budget Office (CBO) reported a base-case estimate of $75 billion to cleanup Superfund sites from fiscal year 1993 onward (CBO, 1994). A low-case estimate of $42 billion and a high-case estimate of $120 billion were also provided. CBO indicated that its estimates are lower than

OCR for page 23
Ranking Hazardous-Waste Sites for Remedial Action comparable EPA and Russell et al. (1991) estimates, primarily because of the assumptions about the future incidence of mega-sites and the costs saved in private-sector cleanups. The Setting of Cleanup Goals Another aspect of the Superfund program that bears upon priority setting is the slow pace of cleanups. This is related to a definition of a clean site, a matter on which the initial legislation, CERCLA, offered no guidance. Initially, EPA generally accepted containment of pollutants at sites or removal and transportation of wastes to permitted landfills. However, paving over wastes and building day barriers around them failed to conciliate residents who wanted the wastes removed, not contained. Removing and transporting wastes may satisfy residents near a site, but other interest groups often objected to transporting wastes from one town to another (OTA, 1985). Furthermore, some local and national interest groups were not satisfied with EPA's ten regional administrators, who typically decided the extent of cleanup required at a site. That permitted outcomes that seemed inequitable to the groups—sites with residual cancer risks of 1 in 10,000 in one region would not be declared high risk, while the same site might have been considered a major risk in another region (Harris and Wrenn, 1988). In 1986, SARA provided specific guidance that had not been in CERCLA. The original legislation contained only a few paragraphs on health impacts; SARA had 10 pages. The Agency for Toxic Substances and Disease Registry (ATSDR) was mandated to perform health assessments for each NPL site and for sites proposed for the NPL before the remedial investigation and feasibility study phase was completed by EPA. In other words, ATSDR was responsible for providing a second opinion based on health criteria. To assist in its efforts, ATSDR requested

OCR for page 23
Ranking Hazardous-Waste Sites for Remedial Action the NRC to review current knowledge of human-health effects by exposure to hazardous-waste sites (NRC, 1991). In addition, Congress relegated capping and removal and transfer methods to the status of last-resort remedies. It called for on-site use of permanent remedies that reduced the volume, toxicity, or mobility of hazardous substances. At a minimum, on-site treatment must achieve the groundwater quality goals of the Safe Drinking Water Act. These changes reportedly have tended to slow cleanup and to make it more costly (Harris and Wrenn, 1988). Gathering the data required for on-site cleanups resulted in increased expenditure of time and funds. Most important, requiring cleanup to conform to the federal drinking-water standards can be costly. This has become a highly controversial issue. The January 1991 CERCLA Superfund Inventory System reported the status of more than 33,000 sites. Preliminary assessments and site inspections indicated that 58% of the sites did not pose significant risk. Another 12% awaited site inspection; 18% had completed site inspection; 8% awaited final assessment; and 4% were on the NPL list. EPA reported that remedial actions had been initiated at over 2,000 sites, with cleanup in progress at 400 NPL sites. Yet, EPA has been severely criticized (e.g., Mazmanian and Morell, 1992) because the cleanup had proceeded too slowly. As of February 1994, EPA had deleted from the NPL 57 sites that had been cleaned up and completed remedy construction at another 167 sites, for a total of 224 sites (Federal Register, 1994). In response to the slow rate of cleanup, EPA has proposed a shift to standardized cleanup plans for different types of sites and to shift EPA personnel from nonsite to site cleanup activities. That step, which is likely to take several years to plan and implement, might relieve some of the pressure felt by EPA, but runs the risk that generic programmatic responses might be ill-suited to some sites and communities. The slow pace of remediation has at least provided the United States with an opportunity to assess the economic implications of

OCR for page 23
Ranking Hazardous-Waste Sites for Remedial Action different definitions of "clean." Had the pace been faster, more resources might have been committed without an opportunity for reconsideration. For example, Russell et al. (1991) discuss three cleanup scenarios. Each cleanup option assumes equal protection of public health. The goal of the "less stringent" case is to isolate the waste so that people will not be exposed and the uncontaminated environment cannot be contaminated. The "more stringent" case aims to destroy Wastes, unless destruction is infeasible or excessively costly. The "current policy" scenario falls between the less and more stringent cases. Russell et al. (1991) estimate a cost of $90 billion to dean up 3,000 NPL sites with the less-stringent criteria. The cost of the current-policy option for the same sites is estimated to be $151 billion, while the estimate for cleanup under the more-stringent criterion is $352 billion. Based upon these estimates, the goal of destroying wastes (more-stringent criterion) costs 3.9 times as does their isolation (less-stringent criterion) and 1.7 times as much as the current-policy option. Technical Limitations The Superfund program is unquestionably larger, more complicated, and much more expensive than was envisioned in 1980. In the context of 1980, hazardous-waste sites were repulsive and frightening eyesores of unchecked industrial proliferation that had to be addressed. In the context of the 1990s, the image is unchanged, and perhaps the fear of those sites has even been reinforced. However, the realities of extremely high costs and a program that might take decades to complete have begun to counter the desire to fully clean every contaminated site. When Superfund was first authorized, there was little experience with cleanup of contaminated soils and groundwater. It was often assumed that existing technologies would meet the need.

OCR for page 23
Ranking Hazardous-Waste Sites for Remedial Action (3) surface water pollution that threatens water used for drinking and recreation; (4) impact on wildlife; and (5) danger of fire or explosion. The state then combines the data without a great deal of quantification to set priorities for cleanups. Michigan developed its own site assessment model in 1983 (Environmental Response Division, 1990). A later revision scores sites according to six factors: (1) up to 20 points for environmental contamination; (2) up to 5 points for substance mobility; (3) up to 3 points for damage to sensitive environmental resources; (4) up to 4 points for population exposure; (5) 1 point for institutional population; and (6) up to 15 points for the toxicity of waste. Each site can score up to 48 points, but environmental contamination and substance toxicity obviously drive the system. New York State uses the HRS scores, but has developed separate health and "biothreat" models. The health model emphasizes human exposure and the biothreat model, natural resources damage. Chapter 7 will describe some of the state priority-setting approaches in greater detail. SETTING PRIORITIES Do DOD, DOE, EPA, and other agencies need scientifically based priority-setting systems? The purpose of such systems is to provide a consistent and scientifically based framework to catalog and compare potential risks to aid in resource allocation, to evaluate progress, and to serve as the basis for communications with affected parties. Hazardous-waste site-remediation efforts can be started and operated without a scientifically based system. One way this is done is for states to sign legal agreements with the federal agencies or the PRPs that would stipulate remediation based on whatever criteria the parties choose. A second alternative is to allocate clean-

OCR for page 23
Ranking Hazardous-Waste Sites for Remedial Action up funds directly in proportion to some indirect measure of potential risk. For example, some state-by-state ratio of the number of people dependent upon groundwater to the number of hazardous-waste sites in areas with groundwater might serve as a quick guide for apportioning remedial action funds among state and local areas. Third, the United States has a long history of allocating funds through political processes. Congressional leaders advised by lobbying groups are capable of writing legislation that will assign remediation funds to provide jobs as well as a cleaner environment in their districts. The disadvantage of these nonscientifically based approaches is that places with the largest political influence, because of economic powers or seniority of their elected officials on important political committees are likely to receive a share disproportionate by nonpolitical criteria. In short, without an objective and scientifically based approach for evaluating the relative risk of different sites to human health and the environment, there is no possibility to match risk and urgency with the allocation of remediation funds. One can then only resort to the political process. Another illustration of how priorities may be set without a scientifically based system is provided in the recommendations of the FFER (Federal Facilities Environmental Restoration) Dialogue Committee (1993). When funding shortfalls result from insufficient appropriations from Congress to meet existing cleanup obligations, the FFER Dialogue Committee recommends a flexible strategy for applying "fair share" principles to allocating funding shortfalls. That is, all federal waste sites subject to outside supervision should share equally in the total amount of the funding shortfall. When a funding shortfall is caused by unanticipated program growth (e.g., due to new circumstances or new data), the FFER Dialogue Committee recommends a greater emphasis on absorbing the shortfall at those sites where it arose and a more general set of principles of fair share relative to shortfalls caused by insufficient appropriations.

OCR for page 23
Ranking Hazardous-Waste Sites for Remedial Action EPA, DOD, and DOE have shown different needs for scientifically based priority-setting models. The differences, the committee believes, are manifested in the agencies mathematical models. Chapters 4-6 describe models in some detail. This section suggests how differences among the three agencies, in political pressures and in the numbers and types of sites, are reflected in their models. EPA funded some engineering studies and emergency remedial actions before Superfund was enacted (Greenberg and Anderson, 1984), but the majority of its remediation funds have been distributed in response to CERCLA and SARA. The idea that sites could be quickly identified and remediated is embedded in CERCLA. EPA required a formal approach that could quickly compare tens of thousands of large sites with small sites, sites in cities with sites in rural areas, and sites in states with strong public support for cleanups with sites in states with little interest in identifying and remediating contaminated sites. Under these circumstances, EPA had neither the time nor funds for 30,000 thorough site reviews and risk assessments. A quick sorting had to be done with a limited amount of data. Indeed, the ATSDR mandate under SARA to provide EPA with a second opinion about sites can be viewed as an effort to alleviate this shortcoming in the CERCLA mandate to EPA. Until lately, DOD and DOE have not been under as much external political pressure to dean up every site in a few years (FFER Dialogue Committee, 1993). Furthermore, by the time these departments were ready to use a formal mathematical model to set priorities for cleanup of sites, a good deal of their funds were already tied up by legal agreements (Whelan et al., 1987). DOD's stated goals for setting priorities are to remove imminent health threats, to address the worst sites first, to meet SARA requirements at NPL sites, and to use resources effectively and efficiently (DOD communication with committee, April 10, 1991). DOE's goals call for an allocation system that must be seen

OCR for page 23
Ranking Hazardous-Waste Sites for Remedial Action to be technically defensible, fair, and not subject to political manipulation (DOE communication with committee, April 10, 1991). In other words, the federal agencies want to remediate sites, use resources efficiently, and protect funds not already committed from being divided by political influence. DOD's and EPA's models have some obvious resemblances. DOD's Defense Priority Model (DPM) and EPA's HRS are similar in that they are risk-based, include both human health and ecological impacts (though with significantly greater weight on human health); can be used to rank sites; are not amenable to a cost-benefit interpretation; and deliberately omit social and economic impacts from their formal calculations. Like the HRS, the DPM assigns a relative numerical score based on combining and normalizing separate scores for groundwater, surface water, air, and soil pathways. Indeed, both models are also similar insofar as a single contaminant and environmental pathway combination can cause a site to be included on a priority list. The models tend to flag a site by choosing the most serious likely threat rather than precisely assessing risk among the sites. There are important differences, however. For example, DOD has been able to provide some funding to all sites. Many of these sites are in relatively remote areas, and national security considerations have discouraged any public outcry. DOD has experienced much less pressure than EPA to rapidly make yes or no decisions on every site. Consequently, DOD's model could assume the availability of time and resources to obtain more field data before using its model. The mathematical modeling approach developed for DOE is different from EPA's and DOD's. In January 1993, DOE estimated that 62 percent of its environmental management activities for fiscal year 1994 are legally driven; 24 percent are other environment, safety, and health activities (required by internal DOE order); and 14 percent are allocated for other desirable program activities (DOE, 1993). T. Cotton (J.K. Research Associates, Arling-

OCR for page 23
Ranking Hazardous-Waste Sites for Remedial Action ton, Va., pers. comm. January 1992) and Rezendes (1992) note that results of the DOE modeling approach to setting priorities would not supersede these prior agreements. Cotton also indicated that the numerical results can be used when no agreement is in place and that they can be used to provide a uniform baseline against which all installation requests can be measured. DOD also has agreements with EPA and the states, but the agreements stipulate that DOD can revert to its modeling results when funds are not adequate to cover all agreements (M. Read, DOD, pers. comm., January 1992). After the committee completed its analyses and was preparing this report, DOD and DOE decided not to use their modeling approach, referred to as the Defense Priority Model and the Environmental Restoration Priority System (ERPS), respectively. Despite those decisions, the committee chose to retain its discussion of these models, because such models serve to broaden the range of priority setting approaches under consideration. As Chapter 6 shows, the modeling approach DOE was considering (ERPS) included more factors, such as economic impacts and weightings based on public preference. In contrast to the EPA and DOD approaches, ERPS was an optimization-allocation model that attempted to economically optimize the use of resources subject to a series of constraints, most notably legal agreements. This former more complex approach to mathematical setting of priorities was consistent with a goal of maximizing a measure of total return from distributing scarce funds among a limited number of sites. ERPS was not a worst-case-first model—neither DOE's agreements with the EPA regions and the states nor the use of ERPS would necessarily lead to remediating those sites posing the greatest risk to public health and the environment.

OCR for page 23
Ranking Hazardous-Waste Sites for Remedial Action DISCUSSION Expenditures to cleanup all forms of pollution are estimated to be equivalent to about 40% of the U.S. defense budget and just over 2% of the gross domestic product (GDP) (Roberts, 1991). The proportion is expected to climb to 60% of the defense budget and 2.6% to 2.8% of the GDP by the year 2000. Paul Portney, an environmental economist with Resources for the Future, argues for the same public debate on environmental expenditures that occurs for defense spending (Roberts, 1991). There is little doubt that such a public debate has begun, and hazardous-waste remediation is likely to be central to it. Under a current-policy scenario, Russell et al. (1991) estimated plausible lower and upper bounds of $478 billion and $1.046 trillion, respectively, for all hazardous-waste site remediation in the country if it maintains the course on which it has embarked. Their best-guess estimate was $752 billion. Can and should the United States commit an estimated 400 billion to a trillion dollars over 30 years to remediate hazardous-waste sites? Policies to remediate hazardous-waste sites in the United States have evolved under extraordinary constraints of politics, time, and money. Scientific and engineering analyses have encountered technical problems that defy current consensus regarding applicability of methods of theory. Consequently, scientific and engineering research agendas and recommendations about hazardous-waste remediation often are negotiated and laden with political values. In addition, the growing awareness of the great limitations of available technology to meet the cleanup goals originally envisioned by Congress, and now anticipated by the public, raises serious questions about what actually would be achieved even if this enormous estimated pool of funds were spent for environmental restoration. The committee believes that the United States is not likely to

OCR for page 23
Ranking Hazardous-Waste Sites for Remedial Action fund the highest level of estimated costs of remediation discussed in this chapter. Congress and its state equivalents might set aside a relatively constant amount of money to be spent on remediation. In short, EPA, DOD, DOE, and the states are going to have to make difficult choices. One such choice could be that not all sites are going to be cleaned up to their precontamination levels. That is, agencies could decide to cleanup some sites to a level suitable for any type of public access, others could be remediated only for limited access, and still others could be declared simply as off limits. Second, the agencies might turn more emphatically to priority-setting approaches and formal models to help choose which sites deserve greater or lesser resources. As discussed in this report, developing a good scientifically-based ranking system is not easy, and much time is required for proper validation. Thus, there is some degree of urgency in starting its development if a sound ranking system is desired. During the last decade, in addition to high costs, five factors have become increasingly evident must be taken into account in prioritizing the cleanup of hazardous-waste sites: Many priorities are and will be set by legal action. The states as well as other interested parties, have been frustrated for years by the failure of the federal government to move more quickly to cleanup sites. DOE and DOD are now subject to many of the same laws as the private sector, and are thus subject to numerous enforcement orders. These orders, usually enforceable in court, not only set forth many procedural steps and regulatory limits, but also contain deadlines for each step and sometimes for ultimate compliance. Although such orders can be renegotiated if all parties agree, it is doubtful whether local citizens would regard a lower priority ranking from a model as sufficient reason for consent. Lack of availability of remediation alternatives that are technologically and economically feasible. Some projects ranking

OCR for page 23
Ranking Hazardous-Waste Sites for Remedial Action higher than others might not be good candidates for "ultimate" remediation or even for complete characterization because economically feasible technological solutions have not yet been developed. Logistics and coordination. Managers must use expensive cleanup crews wisely. On some sites, hundreds of people are deployed along with specialized and complex equipment to manage the sites. Effective planning will be needed to move from one site to another to maximize efficiency and ultimate success. Political intervention. It is inevitable, as many recognize, that scientifically selected priorities will not automatically be reflected in budgetary priorities. That is the nature of our system and always will be. Sometimes, this effect results in a type of fairness, but often it does not; a good priority-setting system should dampen enthusiasm for this type of administrative or congressional behavior. Lack of confidence in the existing system. As the body of this report details, there are real considerations that will make any practical system open to legitimate questions. For example, there are obvious reasons to favor a simple, clear scientifically based site ranking model that relies only on known facts, but such a model is found to be to some extent naive, failing to take into account many societal issues and unmeasured technical factors. A bias on the part of some toward cleaning up contaminated sites to a level of zero risk complicates the priority-setting process. Baron et al. (1993) found a preference in many of those surveyed toward cleaning up a waste site completely, even if the total number of lives saved as a result of this strategy would be lower than the number saved if the identical resources had been expended toward cleaning up two different waste sites partially. The purpose in discussing this list was certainly not to discourage the development and use of a ranking system based on risk information and modeling to aid in making priority-setting deci-

OCR for page 23
Ranking Hazardous-Waste Sites for Remedial Action sions. Rather, the purpose of this report is to show something of the context into which such mathematical ranking models must fit for best performance in the overall process of site prioritization and decision-making. Ranking models are not substitutes for political decisions on what to do first, but they can serve as valuable aids to better insure that decisions are not made on political considerations alone. Indeed, they can help decision-making by providing factual information of importance in an easy to understand way that will help them make the right choices for the public they serve. SCOPE OF THE REPORT The committee examined ranking and priority-setting models developed by EPA, DOD, DOE, and some state governments to help choose sites for remediation from among the tens of thousands of abandoned hazardous-waste sites. It also attempted to understand the larger processes by which these agencies choose sites to remediate and the level of cleanup at each site. Most of the discussions in this report focus on the site-ranking processes rather than on priority-setting processes, which consider ranked scores along with other factors to arrive at actual remediation priorities. Chapter 2 describes some of the desirable features of a priority-setting system and the analytic models used for site ranking. The chapter provides evaluation criteria applied to methods considered in later chapters. Chapter 3 describes five alternative approaches for evaluations performed as a part of priority setting: risk analysis, environmental impact analysis, structured value-scoring methods, cost-benefit and cost effectiveness approaches, and multiattribute decision making. Chapter 4 presents the background and history of EPA's Hazard Ranking System (HRS) in the context of priority setting and the legislative mandate and management pressures of the Super-

OCR for page 23
Ranking Hazardous-Waste Sites for Remedial Action fund program. The strengths and drawbacks of the HRS are discussed. Chapter 5 discusses DOD's development and use of its Defense Priority Model (DPM) to assist in priority setting. The chapter also evaluates the DPM's structure, scoring algorithms, validation, and sensitivity and uncertainty analyses. Chapter 6 provides an evaluation of DOE's Environmental Restoration Priority System, which can use the results of risk indicator models, such as DOE's Multimedia Environmental pollutant Assessment System (MEPAS). The discussion in that chapter is more descriptive and less analytical than discussions in chapters on the HRS and DPM as the ERPS was not as well documented as the other priority systems. Chapter 7 examines several of the state ranking models with respect to how they compare with each other and how well they achieve general objectives of a ranking system. The chapter does not describe how the various state ranking systems help to obtain a final priority for cleanup or the policy context of their application. In Chapter 8, the committee compares and contrasts the procedures used by EPA, DOD, and DOE to make decisions about remediation priorities. A scoring exercise using actual site data and the different agencies' models was performed to help the committee become familiar with the models' input data requirements, operating constraints, and characteristics that contribute to similarities and differences in model outputs. Chapter 9 discusses the advantages and disadvantages of a unified national process for setting priorities and proposes one such unified process. Chapter 10 presents the committee's general conclusions and recommendations for the overall priority-setting process and the mathematical models that are a part of the overall process. Conclusions and recommendations for the priority setting processes and ranking models of specific agencies are presented in Chapters 4-8.

OCR for page 23
Ranking Hazardous-Waste Sites for Remedial Action This page in the original is blank.