Introduction

The American Nuclear Society is a not-for-profit, international, scientific and technical organization made up of 11,000 members who are engineers, scientists, administrators, physicians and educators working in a wide variety of settings. The Society's main objective is to promote the advancement of engineering and science related to the atomic nucleus. ANS currently has 61 local sections (52 in the U.S. and 9 non-U.S.), 23 plant branches, 42 student sections, and approximately 100 organization members.

This is the report of a Task Group appointed by the President of the American Nuclear Society (ANS) in October 1999 to review and comment on the Environmental Radiation Protection Standards proposed by the Environmental Protection Agency (EPA) for Yucca Mountain, Nevada. Yucca Mountain has been proposed as a potential repository site for United States high level radioactive waste.

The Task Group reviewed the EPA Proposed Rule, 40 CFR Part 197, and Supplementary Information as published in the Federal Register, Vol. 64, No. 166, August 27, 1999. We had available the proposed implementing Rule published February 22, 1999, by the Nuclear Regulatory Commission, "Disposal of High-Level Radioactive Waste in a Proposed Geologic Repository at Yucca Mountain, Nevada." We also had available the report of the National Research Council for the National Academy of Sciences, "Technical Bases for Yucca Mountain Standards."

Our Task Group was composed of individual professionals working in nuclear science and engineering fields who are familiar with the various technical and regulatory aspects of radioactive waste disposal. This report is a consensus of the views of the task group members and does not necessarily represent the views of the organizations by which they are employed. On a few issues the conclusions of individual Task Group members varied, and this is indicated in the text.

The report focuses primarily on the technical aspects of the proposed EPA standard and on the rationale provided by EPA on various aspects of the standard. We also note several policy issues that derive from the legislative mandate given by Congress to EPA regarding the proposed standard.

Technical Aspects of the Proposed Standard

EPA Appropriately Bases the Proposed Standard on Individual Dose EPA Appropriately Bases the Proposed Standard on Individual Dose

The National Academy of Sciences (NAS) in its report, "The Technical Bases for Yucca Mountain Standards," recommended that the standard for individual protection be expressed as a risk, that is, the probability of adverse health effects, rather than as an individual dose limit. We agree with EPA and the Nuclear Regulatory Commission (NRC) that an individual dose limit is more easily understood by the public and easier to implement and is therefore appropriate in this case. At the present time, the risk per unit dose is highly uncertain. Future changes in our understanding of the relation between dose and risk would need to be codified in a rulemaking of some sort no matter how the rule is currently expressed. However, the individual dose limit in the standard must be set in a fully risk-informed manner. As discussed below, we conclude that EPA has not always properly reflected risks in the proposed individual dose limits.

The Individual Dose Limit that EPA is Recommending is Not Appropriate

EPA points out that the proposed dose limit of 15 mrem per year is far below the level of background radiation (about 300 mrem per year) and that any hypothesized effects of background radiation are not detectable against the rate of health effects in the general public. While this is certainly true, we believe that the Nuclear Regulatory Commission has a better basis in scientific logic than EPA. The individual dose limit that the NRC has proposed (25 mrem per year) is also lower than warranted. The NRC bases its radiation protection standards on the concept, recognized internationally, that 100 mrem per year is adequately protective. We agree.

We also conclude that releases of radioactivity from current sites in the area other than the Yucca Mountain repository should be considered as adding to individual dose when setting a standard. However, 100 mrem per year is already conservative in NRC's estimation. Subsequent proposed waste disposal sites should be examined to assure that individual exposures from postulated releases will not cause the combined exposure of an individual to approach 100 mrem per year, just as any existing effluent source is considered when siting a new, potentially polluting facility. In addition, for the Yucca Mountain site, the national need for a high level waste repository should give Yucca Mountain priority with respect to subsequent regulated disposal in the Yucca Mountain area of radioactive material. (We note that any postulated unregulated disposal in the area of radioactive material could far exceed any pathway doses calculated for the proposed Yucca Mountain repository and are irrelevant to a finding of site adeqacy.)

An individual dose limit for Yucca Mountain well above 25 mrem per year but below 100 mrem per year is therefore appropriate. Considering the above discussion and the unique national priority for use of the Yucca Mountain area for high level radioactive waste disposal, we conclude that a dose standard of 70 mrem for the repository alone is appropriate, conservative, and adequately protective. This will assure that other current sources, when added to the repository source, will not exceed 100 mrem per year. This individual Yucca Mountain dose is also as limiting as certain aspects of the groundwater standard proposed by EPA. Consumption of groundwater resulting in an individual health risk acceptable to EPA would result in the same net risk to the individual as our proposed 70 mrem/year dose standard. The 100 mrem all source dose is still approximately one-third of natural background.

When one considers the 10,000-year time frame for a demonstration of compliance, no scientific evidence of harmful effects of radiation at these small doses, the location of the Yucca Mountain next to the Nevada Nuclear weapons test site, and the national importance of a facility for the disposal of commercial, research and national defense nuclear wastes, we believe that the 70 mrem individual dose limit is safe and realistic.

Use of the Expected Annual Dose rather than the Mean (Arithmetic Average) Dose is Essential

Whether the EPA concept of a Reasonably Maximally Exposed Individual (RMEI) or the NRC and NAS concept of the Average Member of the Critical Group (CG) is used as the receptor for the calculated dose, it is very important that the value calculated not be skewed by high calculated values that have extremely low probabilities. While EPA has attempted to address this possibility in their human uptake guidelines, the greatest uncertainty in the calculated doses will lie in the pathways through geologic media (for example, fracturing after a geologic event). The NRC proposal of an Expected Annual Dose that is the sum of the performance assessment calculation spectrum, with each calculated dose weighted by its probability, will help avoid distortions in risk perception. The EPA proposal of using the higher of the median or the mean will almost certainly give unrealistic results, as an arithmetic mean will be dramatically skewed to high values by low probability events.

Individual Dose Limits for Intrusion are Inappropriate

In its report on the "Technical Bases for Yucca Mountain Standards," the National Academy of Sciences concluded that it was unrealistic to make scientifically supportable predictions of the probability of human intrusion over a period of 10,000 years. We concur. However, they recommended a stylized intrusion scenario for the purpose of evaluating the consequences of such an intrusion over the life of the repository. It is not clear how their stylized intrusion should be considered in setting a standard for the repository. It appears that EPA has interpreted the guidance as drilling into a failed canister sometime in the far future and still meeting the individual dose limit for an unpenetrated canister system. This then becomes a limiting design requirement for the repository as has been interpreted by EPA. We believe that it is instructive to understand the consequences of such a stylized scenario, no matter how unlikely, but using it as a design requirement does not make sense. Thus, setting a dose limit is unnecessary for such an intrusion.

In our opinion, if there were such an intrusion, a driller capable of drilling that deep would have some technical knowledge and awareness of what he was looking for which means he would have the capability to analyze what was coming out of the bore hole. It is thus reasonable to expect that the individual would also recognize the significance of drilling into a waste repository and what would be necessary to assure containment of the waste. This would include properly sealing the hole.

We conclude that consequence calculations for intrusion scenarios penetrating waste packages after degradation of engineered barriers should be included in the EIS only. It would also be useful to present cases in the EIS for near-repository drilling, or drilling that stops at engineered barriers, and for which proper closure is not achieved. The impacts of postulated early failure of a few waste packages after several thousand years should be examined to illustrate the robustness of the repository and site in the EIS.

We do not believe that an individual dose standard should be set for this scenario since it is at best speculative and realistic assumptions using existing technological capability would address human intrusion scenarios postulated.

Use of Random Sampling Techniques in Performance Assessment is Appropriate

We note that in the Supplementary Information for Section III.B.4, EPA argues against the Critical Group approach, in part, because selection of such a group "requires" use of a theoretical population distribution which is then sampled randomly (Monte Carlo sampling, the term used by EPA, is unconstrained random sampling). As EPA notes, indiscriminate use of Monte Carlo sampling or its more constrained forms like Latin Hypercube Sampling (LHS), and improper coupling or decoupling of input variables to the distribution, can lead to artificial and unrealistic results. However, random sampling properly used is a technique that EPA itself used as the technique by which uncertainty was incorporated in the Waste Isolation Pilot Plant (WIPP) performance assessment parameters. The technique has, since 1980, been the preferred method for analytically reflecting uncertainty in performance assessment parameters.

As discussed below, most of the ANS group preferred the RMEI approach rather than the CG approach for selecting the appropriate receptor population. However, we wish to make clear that our preference is not based at all on the need to use Monte Carlo sampling to identify a critical group or any other population. Unconstrained Monte Carlo and LHS (where appropriate) are the preferred methods for including uncertainty in repository performance assessment. We recommend that EPA and NRC recognize this methodological preference in any implementation guidance.

EPA Standard Should Not Specify Specific Receptor Location

Most of our Task Group favors the EPA formulation of a Reasonably Maximally Exposed Individual over the NRC and NAS choice of the Average Member of the Critical Group. Our reasoning is that the RMEI would be an easier receptor for the public to understand and would reduce by one the probabilistic overlays required to calculate the receptor dose. We recognize that the Average Member of the Critical Group is a scientifically sound concept.

While we can agree with the RMEI approach, we disagree with attempting to choose, in advance, the most limiting location of the receptor. Whatever group, groups, or individuals are to be analyzed for compliance against the standard, they must be representative of the existing population and exposure pathways at Yucca mountain in order to be credible. Because this is a unique standard, specific to Yucca Mountain, the unique features of the site geology and hydrology need to be considered in determining the exposure scenarios and locations of representative groups or individuals.

At Yucca Mountain, the primary exposure pathway is through ingestion of groundwater. There are several different groundwater pathways for consideration in the compliance (repository and site acceptability) determination. The direction of flow ranges from about 90 degrees (east) to 180 degrees (south). The flow paths are complex, chemically and hydrologically, and could require significantly different approaches to calculating radionuclide transport to a given point.

The flow path most used by DOE, including the total system performance assessment in the Viability Assessment, appears to be based on hydraulic head data only. Very different results in terms of concentration over time may be obtained, depending on the technical understanding of the flow field and on the hydraulic parameters assigned. Since the actual flow paths of the radionuclides are currently uncertain, the analysis of dose received should be calculated under the scenarios considered most likely and other credible flow paths. To date, DOE Performance Analysis evaluations have not apparently examined flow paths that would result in more time in fractured tuff rather than through valley fill sediments. Less time in the sediments and more time in fractured tuff could result in less sorption of radionuclides and higher transport velocities. We understand that DOE is currently revising its saturated zone flow models to better analyze these pathways.

No specific locations or receptors should therefore be specified in the EPA standard. DOE should calculate several different pathways and receptor locations (with receptor uptake characteristics appropriate to the receptor location). An NRC determination that each of the locations, with appropriate associated receptors, met the dose standard would be a demonstration of compliance more readily understood by the public. Meeting the standard at various credible locations under lifestyle assumptions and hydrologic conditions appropriate to the location would provide a more credible demonstration of site acceptability. Precisely how this is done must be left as an implementation matter to NRC.

A Groundwater Standard is Unnecessary

As EPA notes in the Supplementary Information of the Proposed Rule, past groundwater standards were set in an attempt to influence siting choices for disposal sites (away from sites where large dilution factors might be used to dispose of large quantities of material). Aside from whether this was an appropriate motivation, EPA correctly notes that this rationale is not applicable to Yucca Mountain.

The remaining EPA reasons for applying a groundwater standard appear to stem from a desire to influence the engineering design of the repository and to reduce collective dose to the general population, neither of which is appropriate. Both approaches are inconsistent with the National Academy of Sciences conclusion that an individual dose standard is adequately protective. The National Academy of Sciences had a well reasoned basis for considering the health and safety of the individual in an integrated assessment from all sources of exposure including groundwater.

The quotation EPA cites from the NCRP (National Council on Radiation Protection) report to support their position on a separate groundwater standard appears to be misapplied.

"The most reasonable risk assessment that can be made for such situations is to calculate potential individual doses for a range of scenarios in order to: (1) evaluate protective measures and (2) to try to place some boundaries on estimates of future individual risks. For the few very long-lived nuclides that are metabolically regulated in the body and more or less uniformly distributed within the biosphere (e.g., 14C and 129I), the future average individual doses may be estimated from total quantities in the environment..."

We note that the NCRP quotation used to support the EPA position relates to calculation of future individual doses from material accumulated in the environment, not health effects of collective dose. As shown by NAS in the case of carbon-14, these individual doses will be extraordinarily small. We therefore disagree with the use of collective dose estimates in the NEPA process or in any standard setting.

As noted previously, very small individual doses are not meaningful in assessing public health impacts (for example in comparison to background radiation). In addition, the Linear Non-Threshold theory of radiation health effects is being questioned with increasing intensity, and a body of scientific opinion exists today that holds it to be without scientific basis. Extrapolation of health effects at lower and lower exposures should be avoided.

The assumptions proposed by EPA for groundwater compliance are less stringent (undisturbed performance of the geology) than the individual standard (events having one chance in 10,000 over a 10,000 year period). Because of this factor alone, it is unclear which standard will be more limiting. Because of the wide isotopic variability of the EPA MCL (and the resultant implied individual dose variability), and because individual dose rather than groundwater concentration is the relevant measure, application of a groundwater standard is without scientific basis in this case. We conclude that the groundwater calculation specified by EPA lacks scientific meaning due to outdated maximum concentration limits of isotopes and is an unnecessary compliance complication.

Policy Aspects of the Proposed Standard

The EnPA Requires EPA to Set Only an Individual Dose Standard

The EPA legislative mandate in this matter, the EnPA (Energy Policy Act of 1992), specifically requires EPA to ". . . prescribe the maximum annual effective dose equivalent to individual members of the public." There is no legislative direction to set separate groundwater standards or attempt to influence the engineering design of the repository by the way that the EPA standards are framed. The separate groundwater standard is not necessary for the protection of human health since it is included in the integrated dose standard and performance assessment of the repository.

EPA May Specify General Receptor Characteristics (RMEI vs. CG) but not the Location of Receptors

The EPA role (under current legislation) is to set the individual health-based dose standard and perhaps specify the general characteristics of the receptors (for example whether a Reasonably Maximally Exposed Individual or Average Member of a Critical Group is to be used). However, the location of the receptors, the geologic and human uptake pathways, and the dose conversion factors are a matter for NRC implementation. (We commented earlier in this report that a specification of a single receptor location is premature.) With respect to dose conversion models, we note that the conversion factor (latent cancer fatalities per rem) cited in the proposed rule is different than that specified in ICRP 60 by about 15 percent. We recommend that whatever version of Part 20 of the NRC regulations is current at the time of licensing be used. This will assure that the agency with the most expertise in this area has adopted appropriate dose conversion factors and calculational methods through the most current public rulemaking process.

Reasonable Expectation should be Better Defined

The performance assessment specified to show compliance with the dose standard will quantify the judgments made in far more detail than in past NRC compliance proceedings. The use of "Reasonable Assurance" (the NRC preference) or "Reasonable Expectation" (the EPA preference) will therefore be primarily in unquantified or high uncertainty aspects of the evaluation. Neither term has been adequately expressed in terms of the degree of confidence desired. We conclude that the use of "Reasonable Expectation" is a better choice for the high uncertainty areas of geologic performance provided that the term is defined as the median value of the spectrum of probabilities estimated. In most cases we expect that this will involve areas where expert judgment must be applied. Another alternative would be to use neither term, and simply address use of median values in remaining unquantified areas in the calculation of the expected annual dose.

As a historic matter, the differences in the NRC and EPA approaches are rooted in their philosophical approach to setting limits. EPA has tended to set very aggressive goals (often based on best technology), but has been very forgiving when best efforts at compliance with the goals are made (thus: "Reasonable Expectation"). NRC, on the other hand, has set more achievable, science-based, standards and has been very strict in enforcing the standards once set (thus: "Reasonable Assurance"). Overall, we much prefer the NRC approach and make an exception for Yucca Mountain as discussed above.

A 10,000-Year Compliance Period is Appropriate

We agree with the EPA and NRC rationale for a 10,000-year time period over which compliance is to be judged. We recognize that uncertainties over this long period will be high, but it is hoped that the licensing process employing "reasonable expectations" will be implemented with that knowledge and intent. We also conclude that an assessment at longer time periods should be performed with median values to illustrate the robustness of the repository and site. The primary location for this assessment should be the Environmental Impact Statement.

The Probability of Events to be Considered Appears Unrealistic

If our interpretation is correct, events with a lower than one in 10,000 chance of occurring within 10,000 years means an annual occurrence rate of less than one in 100 million (10-8) per year. No rationale for this choice is presented. This is another implementation area that should be left to NRC.

Any Guidelines on the Use of Expert Opinion should be Developed in an Industry Standard

Use of expert elicitation by NRC or DOE will be subject to the public process used in the licensing proceeding. A separate EPA guideline would be inappropriate. NRC should have the lead in this implementation matter. The Industry Standards process would seem best suited to the development of a guideline on expert elicitation. We note also that the National Technology Transfer and Advancement Act and an OMB circular require federal agencies to cooperate with private industry and private national consensus organizations in developing standards in preference to developing agency guidance.

Responses to Specific EPA Questions

1.	Q.	The NAS recommended that we base the individual protection standard upon risk.  Consistent with this recommendation and the statutory language of the EnPA, we are proposing a standard in terms of annual CEDE incurred by individuals.  Is our rationale for this aspect of our proposal reasonable?
	A.	As discussed earlier in this report, our Task Group agrees with EPA and NRC that the standard should be based on individual dose rather than risk.  The terminology used in NRC regulations TEDE (Total Effective Dose Equivalent) should be used to avoid confusion.  The NRC approach of annual expected dose should be adopted.
2.	Q.	We are proposing an annual limit of 150 micro-Sv (15 mrem) CEDE to protect the RMEI and the general public from releases from waste disposed of in the Yucca Mountain disposal system.  Is our proposed standard reasonable to protect both individuals and the general public?
	A.	The value proposed by EPA, 15 mrem per year, is far too low.  A number well above 25 mrem, but below 100 mrem is appropriate.  As explained earlier in this report, a value of 70 mrem per year for the repository alone is appropriate, conservative and adequately protective.  This will assure that other current sources, when added to the repository source, will not exceed 100 mrem per year.
3.	Q.	To define who should be protected by the proposed individual protection standard, we are proposing to use an RMEI as the representative of the rural-residential CG.  Is our approach reasonable?  Would it be more useful to have DOE calculate the average dose occurring within the rural-residential CG rather than the RMEI dose?
	A.	Most of our Task Group favors a Reasonably Maximally Exposed Individual but we recognize the scientific validity of the Average Member of the Critical Group approach.
4.	Q.	Is it reasonable to use RMEI parameter values based upon characteristics of the population currently located in proximity to Yucca Mountain?  Should we promulgate specific parameter values in addition to specifying the exposure scenarios?
	A.	It is reasonable to use values based on the current population in the Yucca Mountain vicinity.  Neither specific parameter values nor exposure scenarios should be specified by EPA.  These are implementation matters for NRC.  The values accepted by NRC should reflect actual, rather than postulated, dietary habits.
5.	Q.	Is it reasonable to consider, select, and hold constant today's known and assumed attributes of the biosphere for use in projecting radiation-related effects upon the public of releases from the Yucca Mountain disposal system?
	A.	Yes, for the reasons expressed in the NAS report.
6.	Q.	In determining the location of the RMEI, we considered three geographic subareas and their associated characteristics.  Are there other reasonable methods or factors which we could use to change the conclusion we reached regarding the location of the RMEI?  For example, should we require an assumption that for thousands of years into the future people will live only in the same locations that people do today?  Please include your rationale for your suggestions.
	A.	Neither EPA nor NRC should specify a single location for the receptors in advance.  DOE should calculate results for several locations using relevant geologic and pathway parameters.  Our rationale for this conclusion is expressed earlier in this report.
7.	Q.	The NAS suggested using an NIR [Negligible Incremental Risk] level to dismiss from consideration extremely low, incremental levels of dose to individuals when considering protection of the general public.  For somewhat different reasons, we are proposing to rely upon the individual protection standard to address protection of the general population.  Is this approach reasonable in the case of Yucca Mountain?  If not, what is an alternative, implementable method to address collective dose and the protection of the general population?
	A.	The individual protection standard also assures protection of the general population.  Standards using collective doses for very small exposures spread over millions of people are meaningless.  Addressing collective dose is not appropriate as discussed earlier in this report.
8.	Q.	Is our rationale for the period of compliance reasonable in light of the NAS recommendations?
	A.	Yes.  Longer time periods than 10,000 years should be evaluated in the EIS to illustrate site resiliency.
9.	Q.	Does our requirement that DOE and NRC determine compliance with § 197.20 based upon the mean of the distribution of the highest doses resulting from the performance assessment adequately address uncertainties associated with performance assessments?
	A.	As discussed earlier in this report, we conclude that it is essential to adopt the NRC formulation of the expected annual dose rather than the EPA proposal.
10.	Q.	Is the single-borehole scenario a reasonable approach to judge the resilience of the Yucca Mountain disposal system following human intrusion?  Are there other reasonable scenarios which we should consider, for example, using the probability of drilling through a waste package based upon the area of the package versus the area of the repository footprint or drilling through an emplacement drift but not through a waste package?  Why would your suggested scenario(s) be a better measure of the resilience of the Yucca Mountain disposal system than the proposed scenario?
	A.	The single-borehole scenario is adequate for long time periods at a time when the canister and engineered barriers have deteriorated for purposes of demonstrating site resiliency in the EIS.  As discussed earlier in this report, it may also be useful, in the EIS, to evaluate the consequences of nearby boreholes which do not intersect or penetrate canisters to illustrate the effect of these intrusions on postulated early failures of repository packages.  We do not believe that setting a standard for dose using any stylized scenario is useful.  It would be useful to assess various scenarios for inclusion in the EIS to assess what possible consequences may result and if additional design features need to be added.  We would credit existing technological capabilities to deal with any inadvertent intrusion since it is so hard to predict its likelihood over the 10,000 year period.
11.	Q.	Is it reasonable to expect that the risks to future generations be no greater than the risks judged acceptable today?
	A.	Yes.  In addition, we expect research into the effects of low levels of radiation to demonstrate that the risks to future generations are far lower than projected by current methodologies.
12.	Q.	What approach is appropriate for modeling the groundwater flow system downgradient from Yucca Mountain at the scale (many kilometers to tens of kilometers) necessary for dose assessments given the inherent limitation of characterizing the area?  Is it reasonable to assume that there will be some degree of mixing with uncontaminated groundwater along the radionuclide travel paths from the repository?
	A.	This aspect should be left to DOE evaluation and NRC review and approval, as it is an implementation matter.
13.	Q.	Which approach for protecting groundwater in the vicinity of Yucca Mountain is the most reasonable?  Is there another approach which would be preferable and reasonably implementable?  If so, please explain the approach, why it is preferable, and how it could be implemented.
	A.	As noted earlier in this report, there is neither a technical nor a policy reason for establishing a separate groundwater standard.  In addition, the EnPA directs EPA to set only an individual protection standard.  The objective is to protect individuals over the course of the repository and this is accomplished by the individual dose standard which includes the groundwater pathway.
14.	Q.	Is the 10,000-year compliance period for protecting the RMEI and groundwater reasonable or should we extend the period to the time of peak dose?  If we extend it, how could NRC reasonably implement the standards while recognizing the nature of the uncertainties involved in projecting the performance of the disposal system over potentially extremely long periods?
	A.	The time period for compliance with the individual dose standard is reasonable.  The groundwater standard is inappropriate as discussed previously.
15.	Q.	As noted by NAS, some countries have individual protection limits higher than we have proposed.  In addition, other Federal authorities have suggested higher individual dose limits with no separate protection of groundwater.  Therefore, we request comment upon the use of an annual CEDE of 250 micro-Sv (25 mrem) with no separate groundwater protection, including the consistency of such a limit with our groundwater protection policy.
	A.	The NRC rationale for 25 mrem was set forth in the proposed 10 CFR Part 63 in February of 1999.  EPA has not addressed the substantive case set forth by NRC for 25 mrem as a single adequately protective standard.  To refer to the NRC case only obscurely in one paragraph of the Supplementary Information (Question 15 above) of the proposed EPA rule does not advance a science-based discussion of this issue.  Our views on the appropriate single, adequately protective, standard (70 mrem per year) and on the lack of any basis on which to promulgate a separate groundwater standard are given earlier in this report.
16.	Q.	We are proposing to require, in the individual protection standard, that DOE must project the disposal system's performance after 10,000 years.  Are the specified uses of the projections appropriate and adequate?
	A.	Although it is not clear that a specification to project doses at long time periods is needed in the EPA rule, we agree that projections for the purpose of illustrating the resiliency of the site are appropriate for inclusion in the EIS.

Contributors to this Report Contributors to this Report