Paper presented at VALDOR 2003, June 9-13, Stockholm, Sweden Proceedings
Deconstructing Risk in the U.S. Nuclear Waste Repository Program
The Yucca Mountain, Nevada potential high-level nuclear waste repository site was designated, in July, 2002 by the U.S. President and Congress as the site for which the U.S. Department of Energy (DOE) is intended to apply for a repository license from the U.S. Nuclear Regulatory Commission (NRC). Investigation of the site leading to this decision spanned 24 years, with Yucca Mountain being the only site under consideration for the last 15 years of that period. In 1986, a suite of nine potential repository sites was screened by the Secretary of Energy and the President to three candidate sites for detailed site characterization. The screening was based on geologic criteria as well as environmental, socioeconomic, and transportation considerations . In 1987, without regard for the still effective screening criteria, Congress abandoned two of the candidate sites, along with the mandated screening program for a second repository location, and named Yucca Mountain as the only site to be considered for development of the nation's high-level nuclear waste repository. Political expedience, in an effort to quell strong opposition from all states hosting potential first and second repository sites, and rapidly escalating estimates of the cost to characterize candidate sites were the acknowledged incentives for the Congressional decision.
Thus, the first risk decision was made. Members of Congress from 49 states, none of whom wanted the repository in their state, ganged up on Nevada, with its low population in the state and in Congress, to eliminate the risk of being forced to accept the hazards of a nuclear waste repository in their state. Their decision belies the fact that transporting the used nuclear fuel from over 100 operating and shut down commercial power reactors to Yucca Mountain involves the risks of moving the waste through 43 states and more than 100 cities with populations greater than 100,000 residents. Nevadans have identified the full spectrum of risks being imposed on them and their state, which include risks to health and safety from expected and unexpected releases of radioactive material from the repository site; risks to the environment from the same causes; risks from releases due to transportation accidents; and risk to the state's economy which is heavily skewed toward the tourism industry .
What Nevadans did not realize with the 1987 decision was the extent to which they were at risk from the same institutions that earlier had falsely proclaimed that there was no danger to their health and well-being from the atmospheric nuclear weapons testing at the Nevada Test Site, adjacent to Yucca Mountain. The risk was that, with only one candidate site for the repository, institutional will and momentum would eclipse the government's promise public that "if it is not safe, we won't build it."
2. The Expectation of Safety
For a repository siting and development program to continue to progress it must have some level of acceptance, and in order for there to be acceptance it must meet the public's expectations about its safety. In the U.S. program, at the beginning the public was led to believe that deep geologic isolation meant that the radioactive wastes would be completely contained in the repository for the duration of their hazardous lifetime. The Nuclear Waste Policy Act of 1982 laid out a site screening process and the requirements for site selection criteria, safety standards, and license application review to be in place prior to screening of sites. The public's expectation was that the screening process and the siting criteria and standards would be applied consistently, rigorously, and fairly throughout the site selection process, in order to protect public safety. This included the understanding that is explicit in the Act that sites failing to meet the siting criteria would be disqualified from further consideration. The public was also led to believe that the government was looking for the best, i.e. the safest site for development of a nuclear waste repository. But these expectations were not to be fulfilled, despite the institutional knowledge of the public's deep concern about the risks associated with radioactive wastes.
2.1 The Meaning of Safety
The first blow to the public's expectation of safety, which relied on the complete geologic isolation of the waste, came with the understanding that the government's meaning of safety was that waste would not escape from the repository in amounts greater than regulatory limits, not that no waste would escape. The U.S Environmental Protection Agency (EPA), in 1985, established a safety standard for repositories that set maximum radionuclide release limits, at the accessible environment. The boundary of the accessible environment was to be established at a distance no greater than 5 km from the waste emplacement area. The release limit was not zero, but rather a calculated amount and rate that was expected to result in no more than 1,000 excess cancer fatalities over a 10,000 year period.
The public is not so naive as to believe zero risk is attainable in essentially any situation, but the basis of its expectation for risk from a repository is "as near zero as possible" (in keeping with the proclivity for things nuclear to be given inscrutable acronyms, we shall call this ANZAP). The demarcation between the public's zero risk goal and the government's quantitative determination of "acceptable risk" is clearly illustrated in this example.
The public's ANZAP expectation is grounded, at least in part, in the promotional rhetoric of government managers, nuclear industry leaders, and technologists whose general guarantees of repository safety were pervasive in their public presentations. But, while knowing of the disconnect, they never made clear what they meant by their use of the word "safety."
In its 1990 "Rethinking High-Level Radioactive Waste Disposal" Position Statement, the National Research Council Board on Radioactive Waste Management considered the question of the public's expectations about repository risk and the government's approach to assigning "acceptable" risk. The Board said:
"[S]afety is in part a social judgement, not just a technical one. How safe is safe enough? Is it safer to leave the waste where it is, mostly at reactor sites, or put it in an underground repository? In either case, safety cannot be 100 percent guaranteed [ital. added]. Technical analyses can provide background for answering such questions, but ultimately the answers depend on choices made by the citizens of a democratic society. The EPA has not based its standards (which must allow for these choices by the citizens) on social judgements derived from realistic considerations of these alternatives."
The Statement concludes that "The public has been told too often that absolute guarantees can be provided, but most citizens watching the human frailties of their governments and technologists know better."
One unfortunate implication of the Board's Statement is that the public's expectations of repository safety had been artificially heightened, and now must be realistically lowered.
2.2 The Meaning of Risk
Nuclear waste ranks very high on the public's list of perceived hazards or dangers to people and the environment. The reasons for this are not completely clear, but they involve people's inability to physically sense all but the most dangerous ionizing radiation, uncertainty and controversy about the dangers of low levels of radiation and how to respond to those dangers, and the association with consequences of atmospheric nuclear weapons use and tests.
The word risk is often substituted for hazard or danger in public conversation about nuclear waste, but this use does not coincide with the meaning of risk as it is used by those who assess risk and make regulatory decisions about activities involving risk. The U.S. Nuclear Regulatory Commission (NRC), in setting the basis for risk-informed and performance-based regulation  described risk as follows:
"The risk definition takes the view that when one asks, What is the risk? One is really asking three questions: What can go wrong? How likely is it? and What are the consequences? These three questions can be referred to as the risk triplet. The traditional definition of risk, that is, probability times consequence, is fully embraced by the triplet definition of risk."
The first question is usually answered in the form of a scenario of what can go wrong; the second is a quantification of the probability of the scenario being realized, and the uncertainty involved; and the consequence is the range of probable outcomes (eg. radiation dose to the public from repository releases) given the uncertainties.
When the public is told that the Yucca Mountain repository will be safe because the projected releases of radionuclides will not result in doses to an individual greater than the limit established by the EPA radiation protection standard for Yucca Mountain, they do not realize the conditions embedded in this seemingly straightforward statement. First, the dose referred to is actually a risk. The NRC requires that the dose, for regulatory decision purposes, be stated as the mean of a distribution of doses derived from the probabilistic range of scenarios evaluated. When the dose is probability weighted, it expresses a risk, and that risk can not be understood unless the range of possible doses is also expressed. Also, when the traditional risk definition, probability times consequence, is applied in extreme cases, such as the low probability disruption of the repository by volcanism , the risk is a meaningless, small number, while the actual calculated dose is potentially in the range of being lethal.
In its review of the models used to calculate dose, as part of DOE's Total System Performance Assessment for Yucca Mountain , a peer review team from the International Atomic Energy Agency (IAEA)  considered the way DOE calculated and presented the results (in doses to an individual) of the performance assessment. The International Review Team (IRT) made the following recommendation to help clarify the performance assessment results:
"The IRT suggests that even if probability-weighted dose is the main output required by the regulator, it would be desirable to present disaggregated information (doses and probabilities). This information would more clearly illustrate the nature of the potential impact, so as to better inform decision-makers and other interested audiences. For example, a useful complementary presentation would be the dose profiles conditional on an event occurrence at selected times."
Saying that the Yucca Mountain repository will be safe also includes the understanding, from the EPA Yucca Mountain radiation protection standards, that the dose (risk) from releases is calculated at a location 18 km (11 miles) from where the waste is emplaced beneath Yucca Mountain - allowing for dilution in a prescribed 3,000 acre feet (3.7 billion liters) of groundwater. The individual, who is said to be safe, is an age and gender composite who drinks two liters of groundwater from a well each day, and whose diet consists of some food grown in his irrigated garden.
In summary, when the public is told that the Yucca Mountain repository will be safe, the prevailing assumption is that the risk (meaning hazard or danger) is ANZAP. But the person who said it would be safe is basing the conclusion on the mean probability weighted dose to a composite individual, kilometers away from the waste source, after it is diluted in a fixed large amount of groundwater.
As described above, there is a startling dichotomy between the public's expectations of repository safety and the government's use of risk-informed, performance based regulation and the use of performance assessment as a measure of repository safety compliance.
The public wants, and deserves an understandable, credible, and transparent description of the three components of the "risk triplet" - What can go wrong? How likely is it? and What are the consequences? This is a difficult order for something as complex as a deep geologic repository intended for isolation of long-lived radioactive wastes. But, if people are to make any judgements about the safety of a repository beyond a simple inference of whether it is hazardous, dangerous, etc. or not, they must have information and answers for the three elements of the triplet. For the essential and primary barrier required for waste isolation, the geologic setting, there must be sufficient understanding to be able to enumerate beneficial and adverse characteristics and their consequences in terms of releases to the environment away from the waste area. The regulatory framework (unlike the current U.S. repository regulations) must include unacceptable characteristics, meaning that if the characteristic is known to be present at the site, the risk is too great for further consideration as a repository. There must also be answers to the questions of how likely is it that the understanding is, or can be good enough to make the necessary conclusions about consequences. To achieve "defense in depth," which is a safety net for "What if we are wrong about what could go wrong?" engineered barriers under consideration must undergo the same interrogation rigor as the geologic barrier. The ultimate risk decision on the part of the public is then not whether some obscurely derived "mean dose" is acceptable, but rather whether too much can go wrong, it's too likely that it will, or the consequences are too great - the disaggregation of the risk.
In contrast, the NRC's risk-informed, performance based Yucca Mountain repository regulation does not speak directly to the concern about risk, and understanding the necessary elements of a risk judgement. Instead, it is framed to fit an approach more aimed at engineered facilities designed (through experience) to have a defined operational lifetime, and (unlike a repository) able to have their functional operation terminated at any time. The NRC's description of risk-informed, performance-based regulation is as follows:
"Stated succinctly, a risk-informed, performance based regulation is an approach in which risk insights, engineering analysis and judgement including the principle of defense-in-depth and the incorporation of safety margins and performance history are used to (1) focus attention on the most important activities, (2) establish objective criteria for evaluating performance, (3) develop measurable or calculable parameters for monitoring system and licensee performance, (4) provide flexibility to determine how to meet the established performance criteria in a way that will encourage and reward improved outcomes, and (5) focus on the results as the primary basis for regulatory decision-making." 
Risk insights refer to the results and findings from risk assessments. For waste disposal facilities, these findings and results can be expressed as risk curves, which are estimates of the probability that a given consequence will be exceeded.
The U.S. high-level nuclear waste repository program has led the public to an unachievably high expectation of safety through its promotional rhetoric. Now, a candidate site, at Yucca Mountain, Nevada, has been designated for a repository license application and, if granted, repository development. The assurance of safety has moved from the realm of rhetoric into the hands of the regulatory system, where the public is expected to presume that if the repository is licensed, it must be safe.
But, the public is rightfully uneasy about this expectation. The National Research Council Board on Radioactive Waste Management summarized the logical basis for this unease when it stated, "[P]roof in the conventional sense cannot be available until we have experience with the behavior of an engineered repository system -- precisely what we are trying to predict." The risk-informed, performance-based regulatory approach does not help to mitigate this logical morass, because it does not explicitly and transparently answer the questions of the "risk triplet" - What can go wrong?; How likely is it to go wrong?; and What are the consequences? The public rightfully wants the risk, i.e. danger or hazard, of a repository to be ANZAP (as near zero as possible) so it can then, with the tools provided by answers to the risk triplet, make an informed social decision about whether the risk from this unproven activity is acceptable.
In an effort to overcome the unavailability of "proof" the U.S. repository program is now in the process of extracting and redefining elements of "adaptive staging" to attempt to create an impression that experience will be gained through flexibility in developing the repository in stages with an ongoing monitoring and performance confirmation program. There are even suggestions that the repository closure could be deferred for up to three hundred years to add confidence to a closure decision. But the experience sought is not relevant to "proof" of the long-term behavior of the Yucca Mountain repository system. Most importantly, it does not respond to the currently unfulfilled need for definitive information on what can go wrong with the engineered barrier system in the long-term closed repository chemical and physical environment. Waste isolation performance in a Yucca Mountain repository is heavily dependent on the metal waste containers remaining intact during the 10,000 year regulatory period. Also, contrary to "adaptive staging" the staged approach envisioned by the DOE is directed at development activities, not decision-making about safety of the repository based on ongoing investigations. Under U.S. nuclear waste repository law , the Secretary of Energy's recommendation of the site for repository development signals the completion of site characterization, which includes preliminary engineering specifications, and a description of the waste package (part of the engineered barrier) and its relationship to the chemical and physical environment of the site.
The U.S. repository program has never taken seriously the need to assure that the public has access to the necessary tools to understand potential repository risk in familiar terms and concepts. And further to its discredit, it has applied its own meanings to familiar terms and concepts in a self-serving effort to promote an illusion of assured safety.