Fundamental Limits on Technical and Institutional Capabilities
“ . . . Policies do not implement themselves” (Weimer and Vining 1999, p. 401). Sites with significant levels of residual contamination will require long-term institutional management, and planning for such management will require realistic thinking. In particular, a realistic understanding is needed of technical and institutional capabilities and limitations and the way those capabilities and limitations may affect the institutional management of U.S. Department of Energy (DOE) waste sites over time.
It is also important to recognize that institutional management decisions and actions take place within a broader setting. While decisions and actions are the most visible component of residually contaminated site management, they must be supported with effective organizational, financial, and legal structures. These structures, in turn, are shaped by contextual factors such as legal and budgetary realities and political and economic pressures, as well as by societal and technological changes that can promote or inhibit the long-term success of institutional management.
Long-term institutional management of DOE's residually contaminated sites can be conceptualized as a system within which planning, decision making, and implementation of all segments of the system must work well for management to operate as anticipated. In Chapter 3, Chapter 4, and Chapter 5, the contaminant reduction, contaminant isolation, and stewardship measures (the “legs” of the “stool”) central to site management decisions and actions were described. In Chapter 6, various contextual factors (the “rungs” of the “stool”), also affecting the disposition of waste sites, were addressed. In the discussion that follows, we delve more deeply into technical and institutional limitations and the societal foundation upon which the entire management system rests.
TECHNICAL CAPABILITIES AND LIMITATIONS
Our collective capability to understand and manage the technical aspects of contaminant reduction and isolation has improved enormously over the past few decades, and there is every reason to think that improvements will continue. Nevertheless, those responsible for managing the investigation and remediation of a contaminated site often must make a remediation decision (or, more realistically, ongoing remediation decisions) in the absence of sufficient scientific and technological knowledge and experience. In addition to the broader challenges of institutional management that will be discussed later in this chapter, this fundamental dilemma is manifest in at least three broad areas:
inadequate site characterization;
inadequate understanding and monitoring of the behavior of chemicals and radionuclides in complex environments; and
performance uncertainties of the candidate technologies.
Adequate site characterization would result in sufficient knowledge of the site, its contaminants, and the surrounding environment to make an informed site disposition decision. Each aspect of site characterization, however, may be hampered by scientific and technological limitations.
Quantitative information concerning a residually contaminated site may be sparse or only partially available, in part because of the absence of records of what, where, and how much of hazardous and radioactive materials were disposed of into the environment, particularly during the early operations of the site. The zones of contaminated soil and groundwater and their heterogeneity and extent may be unknown, or known only in a general sense. If data are lacking and conceptual understanding of site dynamics as mediated by physical, chemical, and biological processes is poor, contaminant behavior will similarly be poorly understood. At the Hanford Site, the Idaho National Engineering and Environmental Laboratory (INEEL), and the Nevada Test Site, all sites once thought to possess relatively simple hydrologic and geologic characteristics, contaminant migration has recently been found to be much different from what had been expected (National Research Council, 2000b). Repeated discoveries of such “surprises” regarding the nature of contaminant transport have clear implications for the implementation of long-term institutional management plans at DOE sites.
In situ waste characterization remains difficult at DOE sites. As one example, consider the Subsurface Disposal Area at INEEL, where DOE is pursuing a pilot characterization approach for buried transuranic (TRU) waste (waste that is contaminated with alpha-emitting transuranium radionuclides with half-lives greater than 20 years and concentrations greater than 100 nCi/g at the time of assay). The focus is on a subsection of Burial Pit 9 of the Subsurface Disposal Area prior to recovery of the contaminants and remediation of the surrounding environment. The characterization approach consists of downhole logging of TRU radionuclide levels, together with sample core collection and analysis. Wastes will then be retrieved to verify the ability of the characterization approach to locate TRU. Plans call for the excavation of these wastes and the processing into acceptable waste forms for disposal in the Waste Isolation Pilot Plant (WIPP) in New Mexico. These efforts have brought out the need for technology development (e.g., remote sensing approaches) to characterize buried wastes.
Still another example is provided by differences in our understanding of acidic versus basic wastes. Most DOE wastes were produced in the operation of acidic separation systems such as PUREX (plutonium and uranium extraction, a solvent extraction process used at the Hanford and the Savannah River Sites); an exception was the metallurgical process used at Rocky Flats, Colorado. Current knowledge of the behavior of plutonium in such acidic wastes is derived from extensive investigations of acidic systems used for a long time in plutonium separations and processing. However, most of the DOE acidic waste in the United States was made highly basic for storage in mild steel tanks to avoid tank corrosion. Unfortunately, the chemical and reduction/oxidation (redox) behavior of plutonium in such basic media is very complicated and much less understood. Even though significant progress has been made in the last decade, more and better data on behavior in neutral and basic solutions is needed before reliable modeling of chemical separations and remediation of soil contaminated via tank leaks and overflows of these wastes can be achieved.
The Site Environment
In many cases, the inhomogeneity of the hydrology and geology of the site and its proximate environment are not sufficiently known to reliably forecast and model potential pathways for contaminant recovery and long-term migration. Pathways may include, for example, continuous and discontinuous fractures in the soils and rocks, permeable sand wedges that cut through strata, and folded and faulted rock strata. Scientific and technological limitations contributing to this characterization problem are discussed immediately below.
Behavior of Chemicals and Radionuclides in Complex Environments
In making decisions about contaminant reduction and isolation technologies, it is essential to understand the behavior of chemicals and radionuclides in complex environments. It is notable that most major DOE contaminated sites are complex due to the geology and hydrology, the waste composition and form, or both; of particular complexity is the unsaturated, or vadose, zone. To achieve the necessary understanding there are essentially two methods: empirical (gathering data and learning from experience) or analytical (use of descriptive or predictive models). The two methods are interrelated; models need data and are based in part on observation, and observation and data are often made intelligible by theory. Below, the capabilities and current limitations of the two methods and their interconnections are briefly discussed.
Learning From Experience
Since the late 1970s our understanding of the behavior of chemicals in complex environments has increased significantly. For example, waste site investigations through the 1970s and into the 1980s failed to recognize and take into account the degree to which the release of liquids with limited water solubility, such as chlorinated solvents (dense non-aqueous phase liquids, or DNAPLS), would complicate and exacerbate both subsurface site characterization and, ultimately, the effectiveness of an aquifer restoration strategy. Consequently, the pump-and-treat approach often was selected in the hope that this technology would restore contaminated groundwater to the desired quality in “reasonable” periods of time (a few years). In actuality, the mass transport limitations posed by DNAPLS, and found to be present to some degree in any heterogeneous subsurface environment, prevent pump-and-treat technologies from removing sources of contamination from many, if not most, contaminated aquifers in time periods less than tens to hundreds of years (National Research Council, 1994c).
This example illustrates on the one hand the limitations of science and technology, and on the other the capability to improve by learning from experience. In some instances, however, learning from experience is simply not possible or the risks are too great. Then especially, there is a temptation to rely on models (which are conceptual or mathematical expressions, simplified to some extent, of how one perceives a system) to complement what can be learned by experience about the behavior of chemicals and radionuclides in complex environments.
As discussed in Chapter 6, risk assessments typically are used to evaluate the hazards posed by sites where contamination will remain, while the term “performance assessment” usually refers to an evaluation of the extent to which an engineered system satisfies its predetermined design or performance criteria. Most performance or risk assessments use mathematical models, usually implemented on computers, to describe and predict the fundamental transport and fate processes of both the engineered system and its environment (see Sidebar 7-1). Because of the importance of modeling to much decision making concerning institutional management, the committee has included a discussion of mathematical models in Appendix G.
To be useful, mathematical models rely on the best information about the site, including its physical, chemical, geological, and hydrological properties, and the routes and timing of contaminant exposure to human and environmental receptors. This information determines what parts of the models are deemed relevant in a particular situation and what parameter values and forcing terms (source terms, initial and boundary conditions) are entered.
ROLE OF MODELS, SITE DATA, AND SCIENCE AND TECHNOLOGY IN RISK ASSESSMENT AND MANAGEMENT
(by Shlomo P. Neuman and Benjamin Ross; from Appendix G)
Nevertheless, it is now widely recognized that the subsurface is a complex, multi-scale, spatially variable natural environment that cannot be fully characterized. Hence, the results of even the most thorough site characterization and monitoring efforts are often ambiguous and uncertain. It is important that models reflect these ambiguities and uncertainties explicitly and, whenever possible, quantitatively.
DOE and other organizations often rely heavily on models for decisions about site remediation and waste disposal. Models have been used to “demonstrate” that a potential waste disposal site or remedial option complies with regulations and is therefore safe, an often fallacious inference (see Sidebar 7-2). Often, models have been used without a serious attempt to validate them against site data. This is especially true of one-dimensional “multimedia” or “multiple-pathway” dose and risk assessment models (such as RESRAD, MMSOILS, MEPAS, and DandD). These models, which are based on a limited menu of highly simplified conceptual frameworks, are used for screening as well as more advanced investigative purposes. They are often used with generic parameters and inputs rather than with site-specific data and are often insufficiently calibrated against actual site conditions. This is also true, albeit to a lesser extent, of more complex two-and three-dimensional subsurface flow and contaminant transport models that incorporate various details of site geology. The tendency to rely on models without detailed site investigations, site monitoring, and field experimentation is sometimes used to justify decisions that additional site or experimental data would be of little value for a project. The reasons for this practice are sometimes identified as regulatory and budgetary pressures.
EVALUATION OF NEVADA TEST SITE GROUNDWATER MODELING
(by Shlomo P. Neuman, member of the DOE/NTS External Expert Peer Review Panel)
In Sidebar 7-1 of this report, models are described as appropriate, often essential tools for risk assessment. The following describes an evaluation of major modeling work conducted at the Nevada Test Site (NTS), work that does not appear to meet many of the conditions listed in Sidebar 7-1. This sidebar was prepared by Shlomo P. Neuman, member of the U.S. Department of Energy (DOE) NTS External Expert Peer Review Panel, from material approved for release by DOE.
Over the past 40 years close to 900 nuclear devices were detonated underground at the NTS as part of the U.S. program of nuclear weapons testing. Many of these devices were detonated at depths near or below the water table so that there is a significant potential for groundwater contamination by radionuclides generated during underground explosions. The DOE Nevada Operations Office (DOE/NV) initiated the Underground Test Area (UGTA) Project to evaluate the effects of underground nuclear weapons tests on groundwater. The Nevada State Division of Environmental Protection regulates the corrective action activities of the UGTA Project through a Federal Facilities Agreement and Consent Order. The individual nuclear test sites have been grouped geographically into six different Corrective Action Units (CAU).
Phase I of the UGTA Project is a Data Analysis Task whose goals include the development of groundwater flow and tritium transport models for the NTS and assessment of risks to human health and the environment, at the regional level. An External Expert Peer Review Panel of scientists was appointed by DOE/NV to examine these modeling and risk assessment efforts. The panel concluded that three-dimensional groundwater flow and contaminant transport modeling of the kind developed under the UGTA Project are appropriate for use in evaluating risks at the regional level within a complex geological setting. The program has gone to considerable lengths to establish a geologic and hydrogeologic model that is reasonably true to data and observations. However, the model does not adequately address the large uncertainties associated with tritium source inventory, the geologic model, controlling flow and transport parameters, and associated risk factors. The panel therefore concluded that the summary statement in the DOE report concerning human health, according to which “… risk to members of the public from subsurface migration of tritium in groundwater is not expected to result in an unacceptable risk as long as human activities involving groundwater remain greater than 10 km from the detonation point during the next 30 years,” is not supported by the underlying information presented in the report.
The review panel noted that risk assessment was done only for tritium, while the risk associated with other radionuclides remains unknown. This is so despite the fact that the majority of these radionuclides are more “toxic” than tritium, will persist in the environment for thousands of years due to their long half-lives, and the potential exists for a few of them to migrate almost as rapidly as tritium in groundwater at the NTS. A risk management framework that incorporates spatial dimensions and levels of risk is needed but has not been developed for the NTS.
In the opinion of the review panel, greater emphasis should be placed on modeling uncertainty as a means for determining critical monitoring locations and additional field experiments that are needed to develop reliable observations and predictions at the scale of the CAUs. Wherever possible, model results and predictions should be evaluated against available monitoring data to provide overall weight of evidence for the assessment of risk. There is no indication that such comparisons were attempted. All in all, the panel believes that the project could benefit from a better balance between modeling and data collections efforts, with data collection supporting the modeling and modeling serving to identify where data would be most useful.
The panel looked specifically at the UGTA Project underground nuclear weapons tests in the Frenchman Flat basin, the southernmost CAU at the NTS. It found that, because of data limitations and ineffective modeling strategies, the very limited extent of contaminant migration (a few hundreds of meters) that was
predicted to occur in the alluvial aquifer, though possible, has not been established with the degree of confidence that would normally be expected at such contaminated sites. The panel concluded that uncertainty in model predictions was underestimated primarily because alternative geologic and hydrologic conceptual models were not adequately considered in the uncertainty analysis. The models were replete with assumptions that have not been adequately verified by field and laboratory measurements. There were also concerns that the existing data are not adequate to predict the rate of release of radionuclides from test sites or radionuclide reactions with the surrounding rocks. The exclusion from the study of classified radionuclides further increases the uncertainty in model predictions of future radionuclide doses in groundwater.
In the panel's opinion, the current level of problem identification in the Frenchman Flat CAU is not acceptable. Additional field data are needed simply to see whether problems exist or not. Given the current level of information, it is not possible to unequivocally determine the direction of groundwater flow, let alone whether any contaminant plumes have developed in the flow systems at the site. Current model predictions suggest that no such problems exist, but there is almost no field evidence to back up these claims. The panel knows of no precedent where a no-further-action recommendation has been reached at a potentially contaminated site without a much better understanding of the hydrogeological environment and some field confirmation of the model-generated predictions of contaminant distribution.
It may be tempting to use a model to support a decision that a given waste disposal or remedial option is safe, or that additional site data would be of little value, by basing the model on assumptions, parameters, and inputs that favor a predetermined outcome. An example is the assignment of lower permeability in a groundwater flow model than is warranted by available data. Similarly, it may be tempting to “cast the model in a good light” by basing it on a unique system conceptualization and by subjecting it to sensitivity and uncertainty analyses in which parameters and input variables are constrained to vary within narrower ranges than are warranted by the available information. Such practices ultimately detract from the credibility of those who employ them. The use of models is essential, but they need to be untainted by predetermined outcomes. Moreover, they need to be specially designed to the site at hand and supported with adequate, unbiased data. To the extent that they cannot be, their limitations need to be recognized when making remediation and waste disposal decisions.
Technology Performance Uncertainties
There are significant uncertainties regarding the performance of many remediation technologies. These uncertainties can result in selection of a technology that has no clear demonstration of its efficacy in a given environment. The choice may be between two or more currently available technologies, of which one technology does a more complete job but poses a higher risk of failure. Or the choice may be between technologies available today and the prospect of technological improvements in the future. A currently available technology may preclude using a more effective technology later, but waiting for further technology development may defer remedial actions that are needed now.
Use of Current and Future Technologies
In Chapter 4 the example was given of using multiple grout and cement barriers to fix waste remaining in a high-level waste tank after most of the contaminants have been removed. This example illustrates the difficulty of
using a technology that stabilizes residual wastes, but may effectively preclude using improved, more complete contaminant removal techniques later on. Similarly, there is serious concern about using more aggressive contaminant tank waste removal technologies currently available, such as oxalic acid solutions, because the aggressive technique may attack the tank shell and induce its partial failure. Yet the grout and cement approach to tank waste stabilization also raises concerns about isolation performance and about waste classification. Thus, the limitations of current technologies may force tradeoffs between, on the one hand, preserving access to the residual contaminants while developing more effective techniques, and on the other, “closing” by stabilizing residual waste that cannot now be safely extracted.
Weighing Uncertainties of Technology Performance
Performance assessments for waste isolation technologies (including both barrier and stabilization technologies) often depend on predicting waste transport. As noted above, however, understanding of the waste site, the contaminants, and the surrounding environment is often too rudimentary in comparison with the modeling accuracy needed to demonstrate compliance with current regulations. Consequently, decisions about waste isolation technologies often must be made under conditions of considerable uncertainty. In the face of these uncertainties, decisions can benefit from an estimate of the health consequences if the technology fails completely or to some degree.
Most regulatory criteria are set at exposure levels that are acceptable for licensed or approved activities. Nevertheless, future exposure to chemical and radioactive contaminants may exceed acceptable levels if there was insufficient allowance for uncertainty in the performance assessment. If so, it is important to distinguish whether the exceedances are likely to result directly in grave health consequences for the exposed persons, or whether the exceedances are more likely to go beyond acceptable levels into tolerable levels, that is, levels not likely to result in serious adverse effects. When considering alternate courses for remediation, some defense in depth can be provided if one knows whether failure of a barrier or of an institutional control can lead to radioactive and hazardous chemical exposures that may result in significant health and environmental risk. Sensitivity analyses that explore data and model uncertainties should be used for this purpose. There is a need to study systematically the scientific and technical aspects of contaminant reduction and isolation to reveal the capabilities and limitations with the accuracy and detail necessary to provide for and maintain a focused and relevant program of research and development.
INSTITUTIONAL CAPABILITIES AND LIMITATIONS
At most DOE contaminated sites there is a need to understand not simply how institutional management policies should be formally enunciated, but how they are likely to be implemented over time, and in particular how various factors may cause people to behave or not behave in accordance with official policies. This is sometimes called “forward and backward mapping.” The discussion that follows emphasizes that expectations for the fulfillment of institutional management policies should not be unduly optimistic. Instead, it should be recognized that institutional management policies are undergirded by broader organizational, financial, and legal structures that are not static; they can change. Although institutional capabilities and limitations have not received much systematic attention, there is a body of existing social sciences literature on issues of institutional capacity.
Much has been published in the academic literature in recent years of the notion of “government failure” (that, as for the case with markets, imperfections that are “built-in” frequently prevent government from realizing hoped-for aspirations and efficiencies). The studies, taken in aggregate, suggest that government is inherently better at some tasks than others. One line of particularly relevant interpretation, for example, is that government may work best when serving as a referee between parties of equal power (government's adjudicatory function), but less well
when it takes on the role of partisan “player.” Where feasible, such findings should be applied to designing management systems, thereby playing to inherent strengths of the governance system and avoiding its inherent weaknesses. The objectives should be to assign long-lasting problems to long-lasting organizations, and select organizational designs that maximize the chances for effectiveness over time.
Organizations, like people, can differ greatly in their personalities, competencies, and sources of motivation. For example, some organizations have operated nuclear power plants efficiently and safely; others have been less successful. Some organizations make a serious commitment to environmental protection; others simply go through the motions. Each organization often displays a consistent tendency, or organizational culture. An organizational culture transcends the characteristics of individual workers; moreover, it typically is resistant to change (see Short and Clarke, 1992; Lawless, 1991). Because of its importance, this point is amplified below.
Organizations tend to develop distinctive ways of viewing the world. In the words of Morgan (1986), “Organization rests in shared systems of meaning.” These shared systems can be helpful; they can simplify communication and improve cohesion and task coordination. They have the potential, however, of becoming deeply ingrained. This trait is troublesome when the organizational belief system includes what Clarke (1993) has termed the disqualification heuristic—the belief that “it couldn't happen here.”
With respect to the DOE defense complex, numerous studies have concluded the organization's culture and belief system contributed to the many health and environmental protection problems that arose as a result of site operations. The primary mission of the complex, nuclear weapons production, does appear to have been executed with great competence, and even DOE's severest critics often emphasize that some shortcomings in protecting human health and the environment may have been due to war-time urgency and subsequent cold-war concerns.1
It should be emphasized that DOE is by no means the only organization in which failures of institutions have led to increased risks. For example, the President's Commission on the Accident at Three Mile Island (1979) began its investigation looking for hardware problems that precipitated the 1979 incident at this nuclear power facility, but wound up concluding that the overall problem was one of humans, a problem of what the Commission called a pervasive mind-set, both at the Three Mile Island facility and in the nuclear power industry more broadly, that contributed substantially to the likelihood of accidents. Similarly, the 1986 explosion of the space shuttle Challenger has been attributed in large part to the “push” at NASA to get shuttle missions launched on a regular schedule (see, e.g., Vaughan, 1997), and the Exxon Valdez oil spill was described by the Wall Street Journal as reflecting a pervasive lack of concern by both Exxon and Alyeska with their own risk management plans (McCoy, 1989; for a more detailed assessment, see Clarke, 1993).
In addition, certain predictable tendencies appear to influence many large organizations: in particular, the bureaucratic attenuation of information flows and the diffusion of responsibility. The bureaucratic attenuation of information flows is a phenomenon wherein concerns expressed by on-the-scene workers are not heard by persons at the top. Among organizational analysts (see especially Vaughan, 1997), such a phenomenon is not necessarily seen as entailing a conscious cover-up. Instead, communication is always an imperfect process, and the more “links” in a communication chain, the more imperfect it is likely to be. This phenomenon sometimes is exacerbated, not counterbalanced, by the aforementioned tendency to develop “shared systems of meaning.” Not all kinds of information are equally likely to get through an organizational chain of communication, and bad news is particularly unwelcome.
The lack of a “culture of stewardship” has been noted by numerous analysts of past experience at DOE facilities: for example, at Fernald (Sheak and Cianciolo, 1993; Hardert, 1993), Hanford Site (Gerber, 1992; Jones, 1998; U.S. General Accounting Office, 1993, 1996), Pantex (Gusterson, 1992; Mojtabai, 1986), Rocky Flats (Lodwick, 1993), and Savannah River Site (Peach, 1988; Shrader-Frechette, 1993; U.S. General Accounting Office, 1989), as well as across the weapons complex as a whole (Dunlap, Kraft, and Rosa, 1993; Herzik and Mushkatel, 1993; Hooks, 1991; Jacob, 1990; Lawless, 1991; Morone and Woodhouse, 1989; Shrader-Frechette, 1993; Slovic, 1993; U.S. Congress Office of Technology Assessment, 1991; National Research Council, 1995b, 1996e, 1999b; The Washington Advisory Group, 1999).
The diffusion of responsibility phenomenon creates additional problems (Freudenburg, 1992). While the division of labor can enhance efficiency, it can also increase the likelihood that no one will take responsibility for broader or commonly shared problems. Both phenomena can have especially severe consequences for organizations that have been developed to manage advanced and potentially risky technologies, be they power plants, oil tankers, space shuttles, or nuclear weapons complexes (see Perrow, 1984; Sagan, 1994; LaPorte, 1996; Rochlin, 1996). Precisely because the technologies are complex, the organizations that manage them must be large and, consequently, prone to bureaucratic problems.
Even if an organization begins with a strong commitment to safety, a number of factors can cause this commitment to decline over time. One potentially important factor is mission change; another is the atrophy of vigilance. Over time, most organizations undergo subtle or dramatic mission changes. These changes may occur formally, through official pronouncements and commitments (e.g., the shift from weapons production to cleanup at sites such as Hanford and Rocky Flats), or they may occur informally when workers put energy into priorities that are rewarded while ignoring or giving little attention to other responsibilities that are seemingly less pressing. Since institutional management plans will generally be both repetitive in nature and of substantial duration, they can fall prey to formal and, especially, informal mission change.
The “atrophy of vigilance” (see, e.g., Freudenburg, 1992; Clarke, 1999) is a more subtle but important long-term tendency. To understand this tendency, two of its components—growing complacency and predictable cost control concerns—need to be understood. Growing complacency can be illustrated by the Exxon Valdez accident. Although ships coming in and out of the Alyeska pipeline terminal in Valdez had not been immune to problems, over 8,000 tankers had gone in and out of the port over more than a decade without a single catastrophe, that is until 11:59 p.m. on March 23, 1989. It may have been the very success of earlier trips in and out of Prince William Sound that helped create a situation in which a tanker was under the control of a third mate and the Coast Guard personnel on duty were not bothering to monitor even the lower-power radar screens that remained at their disposal after cost-cutting efforts a few years earlier.
This example also illustrates a second component of the atrophy of vigilance: predictable cost control concerns. Not just DOE today, but virtually all institutions, public or private, are likely to face periodic pressures to control costs. The sources of pressure may include responses to cost overruns, calls to “cut down on waste and inefficiency,” private-sector competition, or simply a desire to do more with less. Whatever the original pressure source and the nature of the organization, at least one response is likely to be consistent: organizations will seek to protect what they regard as their core functions and will cut back on those they regard as peripheral.
Unfortunately, safety measures such as long-term monitoring may be regarded as peripheral or “non-productive,” especially if there has been no demonstrated need for them. For example, planned installation of a larger permanent cap over the currently buried highly radioactive residues stored at the former DOE Niagara Falls Storage Site in Lewiston, NY, would have resulted in the loss of 13 inner perimeter sampling locations. The consequence of such an action would be to increase migration distance (and thus, time) before contaminants leaking in the groundwater from the containment structure would be detected, providing an increased risk to nearby residents and public facilities (National Research Council, 1995a).
A key feature of the legal structure governing the cleanup of DOE 's waste sites concerns the U.S. Constitution; there may be no such thing as a “binding” Congressional or federal commitment. Just as the U.S. Congress makes laws, it can unmake them. While its powers are not unchecked (by, e.g., the threat of political opposition, public outrage, or a presidential veto), the ability of Congress to reverse itself always remains. This power gives flexibility to undo laws that, in retrospect, were ill-advised or are no longer appropriate; however, it also means that, as noted in Chapter 6, even legally mandated arrangements are subject to change. Thus, if Congress passed
a law requiring systematic stewardship for DOE's waste sites, for example, it subsequently could amend or repeal the law or simply fail to appropriate funding to carry it out.
Proposed changes to an existing law are sometimes highly visible and well publicized, but they sometimes occur with relatively little congressional debate, simply by inserting an amendment into an apparently unrelated measure or into a final conference committee action on an appropriation. For example, within less than 10 years of the passage of the much debated and carefully crafted compromises that became the 1982 Nuclear Waste Policy Act, there were at least two major congressional amendments. One amendment, added to the Senate version of the Omnibus Budget Reconciliation Act of 1987, abandoned the specified site selection process for a second high-level and commercial spent nuclear fuel waste repository and established an official preference that the site for the first repository be located at Yucca Mountain, Nevada (Appendix E). The other amendment terminated the guarantee that Nevada could receive independent federal funding for its own scientific research on the suitability of the Yucca Mountain site and on the prospective impacts of the repository siting (Wald, 1992). Both of these amendments were considered quite important to citizens of the affected state, but each was inserted into the final legislation at the “last minute ” and passed with relatively little congressional debate (Easterling and Kunreuther, 1995). Such examples illustrate the need for the forthright recognition of the fragility of federal assurances, not just over the course of several centuries or decades, but also over a few years.
Congressional actions can have unanticipated consequences for agencies as well. For more than 20 years, DOE had responsibility for cleaning up sites that were within the Formerly Utilized Sites Remedial Action Program (FUSRAP). Under Public Law 105-62, however, Congress transferred the FUSRAP sites from DOE to the U.S. Army Corps of Engineers, beginning in FY 1998. The subsequent agreement between DOE and the Corps to effectuate the law put DOE in the position of reinheriting the FUSRAP sites two years after the Corps completed remediation. DOE now has only limited input into site remediation decisions but remains responsible for providing long-term institutional management of FUSRAP sites that remain residually contaminated following their remediation. This obligation, already difficult, becomes especially challenging when institutional controls must be specified in detail in decision documents for remedies that include institutional controls (see Chapter 5, Periodic Reevaluation of the Site Protective System).
Even if policies aren't changed dramatically, carrying out a long-term governmental commitment requires predictable funding. Nevertheless, unless funds are provided in advance, continued funding depends on continued congressional actions to authorize, appropriate, and otherwise see to the actual spending of the “promised” funds. In general, the traditional response to the threat of congressional reversal on funding is to rely on political pressure. Such techniques have been fairly successful to date in obtaining reasonably steady funding for site remediation, especially in politically influential states, but these techniques cannot be counted on over the longer term (and sometimes over the short term), and they thus do not provide a good basis for planning for ongoing institutional management. No matter how genuine a given agency's or official's intentions may be, governmental assurances of future funding are justifiably met with skepticism.
BROAD SOCIETAL FACTORS
Beyond site management activities themselves and their underlying organizational, legal, and financial structures (including reduced spending to lower taxes), broad societal factors also can exert important influences. Three categories of factors are particularly worthy of attention: (a) “beneficial reuse” pressures, (b) societal/technological changes, and (c) trust and credibility challenges.
“Beneficial Reuse” Pressures
Some DOE sites (land and/or facilities) are or will be attractive for economic, residential, or recreational purposes. Recent examples include the transfer of the Pinellas Site, Florida, to the Pinellas County Industry
Council, the sale of portions of the Mound Site, Ohio, to the Miamisburg Mound Community Improvement Corporation (see Sidebar 7-3), and the reindustrialization effort at the former K-25 site (now the East Tennessee Technology Park) at Oak Ridge Reservation, Tennessee. In addition, the potential for urban development pressures exists at sites such as Rocky Flats, Colorado, and the Nevada Test Site, both of which were once considered remote but are now experiencing rapid urban growth nearby.
The goal of beneficial reuse of otherwise under-used federal sites is laudable in many ways, but it can pose risks. The Hall Amendment (Section 31544 of the National Defense Authorization Act of 1994 amending Section 646 of the U.S. Department of Energy Organization Act), aimed at promoting that goal, allows DOE to lease its temporarily unneeded or excess acquired real property at closing or reconfigured weapons production facilities. Leases are for periods of up to 10 years, but they can be renewed for more than 10 years if the Secretary of Energy determines that renewal promotes national security or is in the public interest. Before leasing, the DOE is to consult with the EPA for sites listed on the National Priorities List (NPL), or the appropriate state official for sites
not on the NPL, to “determine whether the environmental conditions of the property are such that leasing the property, and the terms and conditions of the lease agreement, are consistent with safety and the protection of public health and the environment” (Department of Energy Organization Act, section 646, 42 U.S.C 7256).
Many of the proposed uses entail potential human exposure to contaminants that either remain on site or are migrating from nearby, still-contaminated areas. As discussed in Chapter 4 and Chapter 5, the ability to restrict uses under proprietary and governmental controls is questionable, especially over the long term, as is the long-term ability to maintain contaminant isolation barriers and monitoring systems. Moreover, as a day-to-day DOE presence is replaced by contractors, subcontractors, lessees, sublessees, etc., the careful supervision of activities that could result in exposure to residual contaminants is likely to diminish.
Nevertheless, state and local governments often face intense pressures to maximize jobs and development. Authors such as Krannich and Luloff (1991) have noted that in rural areas, leaders often desperately encourage development at virtually any cost (see also Freudenburg, 1991); the same could be said of depressed urban areas. Meanwhile, at the outer fringes of urban areas the pressures to minimize obstacles to economic development are so well known that urban scholars often refer to cities as “growth machines” (see Molotch, 1976). While local residents may have a range of views, often the most powerful and influential local advocates tend to be strongly in favor of the intensification of land use and seek to attract new economic activities, rather than seeing it go elsewhere (see Block, 1987; Edelman, 1964; Logan and Molotch, 1987; Stone, 1989). These advocates are often capable of exerting quiet, behind-the-scenes development pressure long after most members of the public have lost interest and long after records of residual contamination have become lost or forgotten. These pressures can make it difficult for governments to restrain development and to conduct vigorous oversight of residually contaminated sites. While development pressures are not necessarily suspect, they do need to be anticipated as having the potential to undercut present-day intentions.
One of the few general predictions about the future that can be made with confidence is that society and the available technology are likely to undergo changes that are difficult, if not impossible, to envision in advance. At present, for example, development pressures in Henderson, Nevada, a Las Vegas suburb that is now the state's second largest city, are requiring urgent remediation measures for mines that were abandoned mere decades ago. Nobody anticipated the recent population growth for the Las Vegas region and that humanity would move toward old mining areas (Craig Daily Press, December 25, 1999, p. 2). Another example of this urban sprawl is the encroachment of suburban Denver, Colorado, where population has increased five-fold since 1930, toward the contaminated DOE Rocky Flats Site (U.S. Department of Energy, 1999).
Given the rapid increase in suburban sprawl in the latter half of the twentieth century, residential development has begun to press in on many areas that were once considered remote. Indeed, the rates of sprawl have increased significantly in the past two decades, suggesting that conflicts with what are now considered to be remote locations may become problematic sooner than expected. Yet, as has been noted by Erikson, Colglazier, and White (1994), our ability to “see” far into the future may be similar to trying to understand a vast cavern from what can be seen through a small peephole. Realistically, perhaps all that can be done is to plan for the kinds of societal and technological futures that we are able to anticipate. But, when doing so, we need to be realistic about the limitations of our foresight. If the changes of the next century are as great as those of the century just ended, the implications could be quite dramatic. Just 100 years ago, for example, uranium was defined as a mineral with few uses (porcelain glaze, for one); as recently as 150 years ago no one had thought of drilling for oil or, for that matter, drilling more than a few dozens of feet in search of any mineral resource.
Trust and Credibility Challenges
As noted by a wide range of analysts (see, e.g., Dunlap, Kraft, and Rosa, 1993; Jacob, 1990; LaPorte and Metlay, 1996; Slovic, 1991; Secretary of Energy Advisory Board Task Force on Radioactive Waste Management, 1993), one of the central challenges facing the management of DOE waste sites is the legacy of distrust. It is now
well known that governmental organizations in general have suffered a decline of deference by the public in the past three decades. Trust and credibility are subject to “the asymmetry principle:” they are hard to gain but easy to lose (on this point, see especially Slovic, 1991, 1993.) At many DOE contaminated sites, past mistakes have led to a severe erosion of trust and credibility. As Rosa and Clark (1999, p. 22) have noted, such nuclear enterprises provide “a paradigmatic example, capturing the essential features of technological gridlock . . . producing a polarization between citizens, on the one hand, and policymakers, experts and managers, on the other hand, with the net result being impasse over technological choices.”
Trust and credibility issues have particularly important implications for the long-term management of DOE contaminated sites. Some of these sites are large and remote enough to be good locations for undertaking dangerous experiments. For example, approaches could be tested for cleaning up intractable non-nuclear contamination such as DNAPLs, but for such arrangements to be put into place, DOE or other responsible parties would need to reach understandings with regulators and other interested and affected parties (National Research Council, 1999b). While such understandings are not out of the question, they would be far easier to reach within a context of trust and credibility.
STRENGTHENING LINKS BETWEEN TECHNICAL AND INSTITUTIONAL CAPABILITIES
As is evident from the foregoing discussion, disposition of DOE's waste sites is hampered by difficult technical and institutional limitations. Nevertheless, corrections can be made if these limitations are acknowledged and if links between technical and institutional capabilities are strengthened. Two areas where technical and institutional capability can be mutually reinforcing include (1) periodic reevaluations of site disposition decisions, and (2) the development of new science and technologies.
As noted in Chapter 5, a comprehensive approach to long-term institutional management of residually contaminated sites includes periodic reevaluations. To date, there is insufficient evidence to predict whether these evaluations will be meaningful. Based on the EPA record of completing five-year reviews under the Comprehensive Environmental Response, Compensation, and Liability Act of 1980, as amended (CERCLA), there is reason for skepticism about whether periodic reevaluations will actually be conducted when needed, and in a thorough and effective manner. Nevertheless, periodic reevaluations offer an opportunity to reassess how well the total site disposition system, including its technical as well as its stewardship components, are working together to ensure an acceptable level of risk. Through periodic reevaluations, some of the negative impacts of the technical and institutional limitations discussed in this chapter can be reduced, even if they cannot be eliminated.
This chapter has provided examples of cases where vigilance or constancy have been degraded over periods as short as a decade or less, raising serious concerns about the ability of measures like diligent periodic reevaluations to persist for periods of 100 years or more. In part, however, the failures that have occurred have come about because initial policy planning failed to recognize or take into consideration some of the predictable ways the proposed implementation could be disrupted by parties having the incentive and ability to prevent full implementation. In addition, there have been at least some approaches to institutional management that have proven more successful that others, particularly over the short-to-medium term of between 1 and 100 years. Successful approaches appear to have been characterized by incentive structures that appear to be well suited for the types of performance needed. Such incentives may include adequate, stable resources for monitoring and maintenance of contamination, provisions for broad, effective oversight by the public, and establishment of appropriate public use for the area (e.g., nature reserve or park). For the future, accordingly, while there would clearly appear to be value in minimizing the need to rely on fallible human institutions where possible, there may prove to be considerable value in examining more systematically the types of institutional management measures that have proved to be somewhat more successful over the short to medium term.
The relatively high likelihood that institutional management measures will fail at some point underscores the need to assure that decisions made in the near term are based on the best available science. Where deficiencies in scientific understanding that inhibit present-day planning are recognized, incorporating strategies for improving the scientific and technical basis for future decisions increases the chances that those decisions will be soundly based. At the same time, the deficiencies in institutional performance pointed to in this chapter can work against the long-term interests of research and development as well, and have done so in the past (National Research Council, 1996c). At some level, the same fundamental limits affect both the scientific-technical and the institutional and organizational systems. For this reason, attention to the needs of both is necessary in the design and implementation of institutional management systems, a point pursued in the next chapter.
BASIC RESEARCH NEEDS IN SUBSURFACE SCIENCE
A recent report issued by the National Research Council (2000) concluded that basic research is needed in four areas of subsurface science: location and characterization of subsurface contaminants and characterization of the subsurface, conceptual modeling, containment and stabilization, and monitoring and validation. These recommendations are germane to the issues of long-term institutional management of contaminated sites.
Location and Characterization of Subsurface Contaminants and Characterization of the Subsurface
The research needs outlined above call for more hypothesis-driven experimental approaches that address how to integrate the understanding of system behavior.
Containment and Stabilization
New Science and Technology Development
The prospect of advances in contaminant reduction and isolation technologies gives reason for optimism, albeit not for complacency. There is a good possibility that future scientific and technological developments will
create new capabilities that can, in fact, improve waste and environmental characterization and monitoring. They can also resolve uncertainty and spur contamination reduction and isolation and stewardship that achieve the goal of protecting the public and the environment in the most efficient manner. Scientific and technological developments that are actively sought, however, have greater likelihood of being realized than are those that are merely hoped for. As noted in Chapter 5, science and technology developments can be sought either by monitoring the emergence of new remedial technologies in non-DOE settings (e.g., in the private sector or in other nations) or by direct sponsorship of new science and technology research and development (R&D), through peer-reviewed processes and rational, needs-based selection of R&D projects (National Research Council, 1996e; 1998c; 1999b,c). New science and technology developments entail an institutional commitment that is complementary to periodic reevaluations. Sidebar 7-4 gives some of the recommendations for basic research needs in the subsurface sciences from a report by the National Research Council (2000b). A reevaluation may suggest the need for further contaminant reduction or isolation; new sciences and technology development may suggest a way this can be accomplished. Moreover, both periodic reevaluations and new science and technology developments take advantage of present institutional capabilities to “buy time” to find longer-lasting solutions.