Editor's note: This is a lamentably long article, but possibly the most important so far published in Science for Democratic Action. So please read it, and send us your comments.

The most recent in a long list of studies on the setting of standards was prepared by an ad hoc committee of the National Research Council of the National Academy of Sciences (NAS), chaired by Robert Fri of Resources for the Future in Washington, D.C. The NAS committee's report was mandated by the 1992 Energy Policy Act, in which Congress directed the Environmental Protection Agency (EPA) to develop a set of standards for high-level waste disposal specific to Yucca Mountain in Nevada.(1)

In the NAS report, Technical Bases for Yucca Mountain Standards,(2) fourteen of the committee's fifteen members recommend a method for assessing risk to future populations that has never been used before in radiation protection. (3) One committee member, Professor Thomas H. Pigford, among the country's leading nuclear engineers and one of the founders of the nuclear engineering department of the Massachusetts Institute of Technology, disagreed with the recommendation and filed a vigorous dissent.

The recommendations of the committee majority and the dissent by Pigford are an important part of the current national debate over science, risk, and environmental policy. In this article, I will review both the majority view and the dissent as they concern some aspects of radiation protection standards and the technical assumptions underlying the standards. The centerfold in this issue contains additional information on the report and related matters.

Section II: Background on Setting Exposure Standards
The principal basis for radiation protection until now has been to set limits on the maximum allowable exposures to individuals from man-made sources. For example, the overall individual dose limit for the general population from all sources of radiation (other than medical) is 100 millirem per year. The limit for exposure due to emissions from specific facilities is generally in the range of 5 to 25 millirem per year.

Setting limits on general population exposure is much more difficult due to the logistics involved in measuring doses to all individuals. In addition, the large number of sources of radiation, both natural and artificial, make it very difficult to pin down exposure to any particular source, unless it were large relative to all the others. Therefore, radiation protection of the general population, while aimed at limiting doses, is, in practice, often based on two key concepts: limiting the total releases of individual radionuclides (or groups of radionuclides); and limiting the concentrations of radionuclides in air, water or soil. Both of these concepts are incorporated into current standards for high-level waste repositories, codified in federal regulations 40 CFR 191.

The first practice, limiting the total releases of individual radionuclides (or groups of radionuclides), limits the dose received by the total population. The second practice, limiting the concentrations of radionuclides in air, water or soil, limits the dose to the maximally exposed individual. If this individual is not exposed over a certain limit, then it follows that essentially all of the rest of the population would be exposed to health risks lower than those created by the upper limit of exposure. The "maximally exposed individual" is a hypothetical construct, corresponding to a set of "reasonable" assumptions about human needs and activities. People who may be unusually sensitive to radiation or who have unusual habits are not used for standard setting. For example, a British inquiry omitted people who subsisted mainly on clams from its definition of the affected population because this diet was considered unusual. (4)

When the main route of exposure over long time periods is expected to be via the use of water for drinking and subsistence farming, it is the general practice to use the "subsistence farmer scenario" for calculating exposure. This approach assumes that a person would unknowingly use contaminated water for drinking and farming and would grow all their own food.

For the purposes of calculating radiation dose, a small, homogeneous group of individuals is used to define a "critical group." The International Commission on Radiological Protection (ICRP) explicitly states that the critical group "represents an extreme" of radiation exposures within the entire population in the area in order "to ensure that no individual doses are unacceptably high."(5) (emphasis added) It recommends that critical groups be small so that they are homogenous, with the upper limit to size usually being a "up to a few tens of persons" and they could be as small as only one person. (6)

The device of a small critical group is used to represent the maximally exposed individual for regulatory purposes. Once the exposure scenario for the maximally exposed individual is selected, then it is possible to derive secondary standards for limiting concentrations of radionuclides in air, water, and soil. These secondary standards, if adhered to, would result in compliance with the primary dose standard.

Since it is difficult or impossible to measure radiation doses and risks to the general population from particular radiation facilities, secondary standards that limit concentrations of radionuclides are essential to ensuring compliance with dose or risk limits. Setting secondary standards to protect the general population from radioactive contaminants is used throughout the world in radiological protection, including in the United States.

Section III: Standards Suggested by the NAS Committee
The NAS committee majority proposes to set aside the concept of setting secondary measurable standards in favor of limiting the risk to a critical group as defined in a new way (see below for further details). The principal argument for such a standard is that it directly addresses the thing we most want to limit: risk of damage to health, including cancer risk. In fact, the NAS committee is explicit that it does not include the current goal of protecting ground water as a resource in its recommendations. The report states that the current EPA regulation for high-level waste disposal,

"40 CFR 191, includes a provision to protect ground water from contamination with radioactive materials that is separate from the 40 CFR 191 individual-dose limits. These provisions have been added to 40 CFR 191 to bring it into conformity with the Safe Drinking Water Act, and have the goal of protecting ground water as a resource. We make no such recommendation, and have based our recommendations on those requirements necessary to limit risks to individuals." (p. 121)

The possibility of very high radiation doses, far above allowable limits, from consumption and agricultural use of water contaminated by a high-level waste repository at Yucca Mountain is real. Since water is scarce in the area, there is only a relatively small volume available (compared to other repository locations) to dilute leaking radionuclides (pp. 27-28 of the NAS report). A 1983 National Academy of Sciences study on repository disposal of nuclear waste estimated that peak doses could range from a low on the order of one rem (perhaps less) to about 1,000 rem per year depending on assumptions about the behavior of the waste and water travel time. (7) More recent studies done by Sandia National Laboratory and INTERA, both Yucca Mountain Project contractors, also estimate that peak doses from using water contaminated by a Yucca Mountain repository could be high. (See centerfold.)

Section IV: What About the Subsistence Farmer?
Could the NAS committee's recommendation of limiting risk to individuals be compatible with allowing high doses of radiation to maximally exposed individuals, and in particular to subsistence farmers? And are the committee majority's recommendations in conformity with the recommendations of the ICRP? These questions are at the heart of the dispute between the committee majority and the lone dissenter, Professor Pigford. Appendix C of the report, where the majority specifies the method to be used for calculating exposures, also appears to make contradictory statements (see below).

Here is my understanding of the eight-step process of determining the exposure of the critical group as described in Appendix C by the NAS committee majority. My own comments are italicized in parentheses.

1. Identify the population which contains the people at risk of getting the highest doses. "For purposes of illustration, this example assumes a farming community in the Amargosa Valley." (The term "farming community" could include many occupations, not just subsistence farmers. It could be a large, inhomogeneous group, which would be incompatible with ICRP's recommendation for a critical group, or a small, homogenous group. For instance, it may consist of farmers, casino operators, and defense workers or it may have farmers only (see NAS report page 145. The farmers may or may not be subsistence farmers.)
2. Quantify the demographic and geographical characteristics of the population so as to determine what areas in the region "have the potential for farming and groundwater use." If possible, limit the area for exposure analysis by excluding some areas, such as those not likely to be farmed or where groundwater might be too deep. (On this basis, the area and groundwater in the immediate vicinity of the Yucca Mountain repository could be excluded from the calculations.)
3. Identify the intersections of those areas that might be farmed and those beneath which radioactively contaminated water would be present at some time.
4. Model the release of radionuclides from the repository and take into account that the plume of contamination passes through various areas at different times, limiting exposure in this way. Model various possible ways in which the contaminated plume of groundwater might travel (these are called "plume realizations"). People living in such areas before the plume is directly under them will be "at no risk" during these periods.
5. Calculate doses for a large variety of possible conditions and times, sampling from among the various plume realizations. (This step acknowledges, in contradiction to the one just above, that people "outside the area overlying the plume" could be exposed due to "local export of water or food.")
6. Calculate the times at which the groundwater under various exposed populations would be most contaminated.
7. Divide the results of each groundwater contamination ("plume realization") into geographical subareas in which doses are to be arithmetically averaged. The population of each subarea should be large enough "to allow computation of a meaningful average dose." Then define a "critical subgroup" consisting of all subareas with average risks within a factor of ten of the "maximum average" subarea risk. (The term "meaningful average" is not defined. This requirement could, in some cases, conflict with the ICRP recommendation that the critical group be small.)
8. Average the average doses for the critical subgroups in Step 7 for each plume realization. This final average of averages is defined by the committee majority to be the "technically appropriate representation for the critical-group risk."

"We recommend that the critical-group approach be used in the Yucca Mountain standards. (Text underlined in original.)

"The critical group has been defined by the International Commission on Radiological Protection (ICRP) as a relatively homogeneous group of people whose location and habits are such that they are representative of those individuals expected to receive the highest doses..."

"critical group for risk should be representative of those individuals in the population who, based on cautious, but reasonable, assumptions, have the highest risk resulting from repository releases. The group should be small enough to be relatively homogenous with respect to diet and other aspects of behavior that affect risks." (p. 53)

Section V: Professor Pigford's Dissent
The central points of Pigford's dissent (in Appendix E of the report) are as follows:

• The committee majority has abandoned the subsistence farmer scenario which is the surest, most conservative method for protecting all future populations. It is in conformity with the recommendations of the ICRP. It is also the radiation protection method of choice worldwide, including in the United States, Britain, Sweden, Switzerland, Finland, and Canada. He cites the British National Radiological Protection Board's advice, for instance, according to which the critical group would consist of people "at the place where the relevant environmental concentrations are highest, and [who] have habits such that their exposure is representative of the highest exposures that might reasonably be expected."
• "There is consensus that the subsistence-farmer approach is consistent with the critical group concept." Pigford cites several examples, including one in which the U.S. Nuclear Regulatory Commission used a subsistence farmer family of three as the critical group.
• The probabilistic critical group approach recommended by the majority "is demonstrably less stringent in protecting public health than the subsistence farmer approach." The example of the farming community in the Amargosa Valley used by the committee majority would contain part-time farmers, but "the full-time subsistence farmer will not be found on that distribution." The probabilistic critical group method is not in conformity with the recommendations of the ICRP.
• The method is subject to manipulation because it permits arbitrary choices of parameters such as population characteristics and sizes of subareas. Such choices could lower the calculated doses which would provide "an illusion of safety, but with a serious loss of credibility."
• The "[c]alculational techniques described in Appendix C are not mathematically valid."

Indeed, the calculational procedure set forth by the committee majority could allow for the exclusion of the subsistence farmer entirely (see below). In that case, the NAS committee would extend the definition of people with "unusual habits" from those whose diet consists almost entirely of clams to subsistence farmers, which is one of the most common occupations in the world today.

Section VI: Not Mathematically Valid?
In some ways, Pigford's charge that the method in Appendix C is not mathematically valid is a very surprising one to remain standing after the work of the committee was complete. Even more astonishing, Pigford has stated that none of the members of the committee or any of the reviewers even responded to his claim of the mathematical invalidity of the method during the course of the study. (9) The lack of a response to such a basic charge is most unusual and raises serious questions about the integrity of the scientific process by which the committee majority decided that its recommendation of a new, complex, and untried method in radiation protection was a workable one. Therefore, I explore the matter at some length here.

The essence of Pigford's claim of mathematical invalidity is that the method in Appendix C does not result in a critical group that corresponds to a critical group as defined by the ICRP, as the committee would claim. This is because Step 7 of the calculational process divides the "region into subareas, with no homogeneity requirement for the subarea." This means the doses to individuals within the subarea can be very different. A few individuals with high doses could be averaged in with a large number of individuals with low doses, resulting in a low average dose (or risk) for the area.

ICRP recommendations require that the individuals with the highest dose (or risk) be part of the critical group. But in the method of Appendix C, the averaging process over a subarea could result in the highest exposed individuals being in a subarea that has a low average dose. This could result in their exclusion from the critical group defined in Step 8 of Appendix C because there may be many subareas with a higher average dose (or risk), but which do not include the individuals with the highest dose (or risk).

The disagreement has not been resolved. My own preliminary conclusion is that Pigford's is right to conclude that Appendix C is not a mathematically valid approach to creating a critical group in conformity with ICRP's recommendations. While Pigford was a minority of one in the committee, the widespread use of the subsistence farmer approach and its clear conformity with the ICRP method shows that he is, in fact, in the majority in the scientific world at large on this issue.

While disagreements about policy abound, and there is wide room for legitimate debate about how future generations ought to be protected from what we do today, the lack of a clear resolution to the mathematical question is very troubling.

A first reading of the report led me to the conclusion that despite some differences of detail, the method in Appendix C had a close relationship to that described in a report by the Electric Power Research Institute (EPRI). The EPRI report, cited by Pigford in his dissent but not in Appendix C, shows calculations based on one variant of the probabilistic critical group method. (10) This variant is not in conformity with the ICRP critical group method and is not claimed to be.

I had assumed that the calculations by EPRI were an adequate description of possible results of actually trying out the suggested approach to risk evaluation. However, in a review of an early draft of this article, NAS committee chairman Fri explicitly denied the connection:

...it is simply untrue to suggest that the approach in Appendix C derives from an EPRI report. If it is the report I think you may have in mind, any careful comparison of the calculations involved would quickly show no relationship. (11)

Section VII: Implications Beyond High-Level Waste Disposal
The abandonment of explicit groundwater protection, if adopted by the EPA for its Yucca Mountain standard, would set a dangerous precedent. Industry will likely begin to clamor for its extension to all radioactive waste disposal and to cleanup standards for the nuclear weapons complex and other contaminated areas. That could mean the abandonment of clean water standards for vast sections of the country. In the current anti-regulatory climate, it is not at all out of the question that the approach may be extended to cover all toxic materials.

To treat groundwater, and by implication all other water, as if it is not a common resource for humanity is a sad abandonment of basic principles of ecology and of environmental protection. Any extension of such a philosophy would negate a central ecological presumption of radiation protection: that all other forms of life would be protected if human health is protected. While one may assume technology will save humanity if we all earn our living as casino operators and defense workers, other living beings do not have the same options.

It is relevant to note here that Yucca Mountain is claimed by the Western Shoshone people as their land. The NAS committee chose to entirely ignore not only their claims, but also their customs and their idea of what is to be protected. The recommendation of the NAS committee, which ignores groundwater protection as an explicit goal, along with the majority's probabilistic critical group method could provide the first step along a disastrous road to the abandonment of protection of other living beings.

Here is how Corbin Harney, an elder and spiritual leader of the Western Shoshone people, sees life and environmental protection on that same land:

"We've been taught this from the beginning of our lives: take care of this land and everything that's on it; take care of it well in order to bring good to all the plant life and all the things that are here. We have to take care of them all." (13)

Institute for Energy and Environmental Research

Comments to Outreach Coordinator, Pat Ortmeyer: ieer@ieer.org
Takoma Park, Maryland, USA

Revised March 21, 1996