Click for next page ( 34


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



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

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

OCR for page 33
2 PROTECTING HUMAN HEALTH The primary objective of the proposed repository at Yucca Mountain is to dispose of high-level raciioactive defense waste and spent nuclear fuel in a safe manner. To determine whether the repository can be designed to protect the public health from the risks associates! with exposure to radiation from rarIionuclides that may be releasecI from the repository, it is necessary to establish stanciards against which to judge whether the design of the repository is acceptable. This target will be embocliec] in a racliation protection standard to be issued by EPA. In Section 801 of the Energy Policy Act of 1992, Congress directs that EPA set these standards by specifying the maximum annual effective close equivalent to indiviclual members of the public. In the same section, Congress also asks three questions, the first of which is: whether a health-based stanciard based on closes to individual members of the public from raclionuclide releases to the accessible environment . . . will provide a reasonable standard for the protection of the health and safety of the general public. This chapter addresses this question. As background, we first present a synopsis of the health effects of ionizing radiation and outline the development of radiation protection standards on a national and international basis. This discussion will illustrate the current status of scientific investigation ant] consensus of expert judgment on which most efforts to establish a stan~iard for high-level waste repositories are based. We then turn to the question of whether a standard for Yucca Mountain designed to protect individuals will, if met, also protect the general public. We conclucle that the answer to this question is "yes," given the particular characteristics of the site ant! assuming that policy makers ant! the public are prepared to accept that very low radiation doses pose a negligible risk. Because the current EPA standard for nuclear waste disposal in 40 CFR 191 takes an approach different from that required by Congress, 33

OCR for page 33
34 YUCCA MOUNTAIN STANDARDS :~ If::: however, addressing only the question posed in Section 801 is too narrow a response. Accordingly, we have expanded the discussion by recommending the use of a standard} designed to limit individual risk rather than individual dose and by ciescribing how a standard might be structured on this basis. We then address the specific question of protection of public health in the context of an individual-risk standard and compare this standard with the one currently user! by EPA for sites other than Yucca Mountain. Baser! on this analysis, we conclucle not only that an individual- risk standard would protect the health of the general public, but also that this form of standard is particularly appropriate for the Yucca Mountain site in light of the site's characteristics. Finally, standards are only useful if it is possible to make meaningful assessments of future repository performance with which the standarcis can be compared. In Chapter 3, we discuss our conclusion that it is feasible to conduct such compliance assessments against an inclividual- risk standard. Doing so, however, requires using the rulemaking process to arrive at a regulatory decision about certain assumptions as part of the standard, for example., about future human behavior. In the following discussion of the standard, we have indicated the assumptions for which this is required. THE HEALTH EFFECTS OF IONIZING RADIATION Cell and gene damage can be caused in humans exposed to ionizing radiation (NRC, 1990a), (also referred to as the BEIR V report). Extremely high closes of radiation can lead to quick death, as seen, for example, in Nagasaki, Hiroshima, and Chernobyl. However, even much lower levels of radiation can affect health. International scientific bodies currently accept what is called the linear, or no-threshold hypothesis for the tiose-response relationship. Most of what is known about effects of radiation on human health comes from studying people exposed to large doses of radiation. The empirical relationship between cancer induction and radiation dose appears linear at the high doses received by the atomic bomb survivors. The linear hypothesis postulates that this dose-response relationship continues when extrapolated to very low doses. The no- threshoic} hypothesis hoists that there is no dose, no matter how small, that does not have the potential for causing health effects. To explain this

OCR for page 33
PROTECTING HUMAN HEALTH 35 relationship of radiation to cancer, ant! other health effects, the following outlines the interaction between radiation and the human body. Radiation that is sufficiently energetic to clisIo~ige electrons from an atom is referred to as ionizing racliation. Impinging ionizing radiation, colliding with atoms en c! molecules in its path, gives rise to ions ant! free radicals that break chemical bonds and cause other molecular alterations in affected cells. Any molecule in the cell can be altered by racliation, but deoxyribonucleic acid (DNA), the clouble helix of base pairs that make up the genes to be passer! on to the next generation, is the most critical molecular target because of the uniquely important genetic information it contains. Damage to a single gene, which might consist of thousands of base pairs, can profouncily alter or kill the cell. Although millions of changes in DNA are proclucec! in the body of every person each year by exposure to natural background radiation ant] other influences, most of the changes are reparable. If unrepairec! or misrepaired, however, the damage might be expressed in the form of permanent genetic changes or mutations, the frequency of which approximates 10-5 to IO6 per gene per Sievert (Sv)~. Because the mutation rate tencIs to change in direct proportion to the dose, it is inferred that the interaction of the gene with a single ionizing particle might suffice in principle to mutate the gene. Damage to the genetic apparatus of a cell can also cause changes in the number or structure of its chromosomes, the thread-like structures on which the genes are arranged. Such changes increase in frequency in proportion to the (lose in the range below ~ Sv. Radiation ciamage to genes, chromosomes, or other vital organelles can be lethal to affected cells, especially dividing cells, which are highly radiosensitive as a class. The survival of dividing cells, measured in terms of their capacity to grow and divide, tencis to decrease exponentially with increasing close, 1-2 Sv generally sufficing to recluce the surviving cell population by about 50% (NRC, ~ 990a). The killing of cells, if sufficiently extensive, can impair the function of the affected organ or tissue. In general, however, too few cells are killed by a close below 0.5 Sv to cause clinically detectable impairment of function in most human organs other than those of the embryo. Because such effects on organ function are not produced unless the radiation dose exceeds an appreciable threshoici, they A unit of equivalent radiation dose, a Sievert is the product of the absorbed dose and the radiation weighting factor. 1 Sievert equals 100 rem.

OCR for page 33
36 YUCCA MOUNTAIN STANDARDS are commonly viewer! as nonstochastic (or deterministic) effects, in contradistinction to mutagenic effects of racliation, which are viewed as stochastic effects because they might have no thresholds (see Glossary). Carcinogenic effects of radiation, which can result from mutational changes in the affected cells, are likewise viewer] as stochastic effects, the frequency of which is assumed to increase as a linear, no-threshoIc] function of the dose, although the possible existence of a threshold for such effects cannot be excluded. Natural background! radiation is estimated by the National Council on Radiation Protection ant! Measurements (NCRP) to contribute 82% of the average annual radiation exposure to a Uniter] States citizen, and medical applications, an additional 15% (NCRP, 1987a). All other sources of radiation exposure together contribute approximately 3% (Table 2-1~. All sources combined give an average dose of 3.6 mSv/yr (360 mrem/yr). Background radiation levels are not uniform. For example, the average difference in background radiation between Denver, CO and Washington, DC, is 0.3 mSv/yr (30 mrem/yr). One cross-country plane ride contributes approximately 0.025 mSv (2.5 mrem) (NCRP, 1987a,b). At the low-dose rates characteristic of natural background radiation or occupational irradiation, the only health effects of radiation to be expected are stochastic effects; that is, mutagenic and carcinogenic effects. Although the risks of certain cancers have been significantly elevated in some cohorts of radiation workers, especially those employed in the era preceding modern safety standards, no definite or consistent evidence of carcinogenic effects has been observed in workers exposed within present maximum permissible dose limits or in populations residing in areas of high natural background radiation. Hence, assessment of any cancer risks attributable to irradiation in such populations must be based on extrapolation from observations of the effects of exposure at higher dose levels. Because a statistically significant increase in heritable abnormalities is yet to be demonstrated in human beings at any dose level, assessment of the risks of such effects must be based on extrapolation from observations on laboratory animals. Because of the assumptions inherent in the extrapolations that are involved, assessments of the carcinogenic and mutagenic effects of low-level irradiation are highly uncertain. The uncertainties notwithstanding, it has been possible to reach a reasonable consensus within the scientific community on the relationship between doses ant! health effects, that is generally considered to provide an

OCR for page 33
PROTECTING HUMAN HEALTH 37 acceptable basis for evaluating the risks attributable to a given dose or the degree of protection affordeci by a given limitation of exposure. Within recent years, the risks attributable to low-level irradiation have been assesses! in detail by the Uniter! Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR, 1988), the National Research Council Committee on the Biological Effects of Ionizing Radiation (NRC, 1 990a), ant] the International Commission on Radiological Protection (ICRP, 19911. The last of these assessments, which cirew on ant} extender} the previous two, arrived at risk assessments for carcinogenic effects and for heritable effects, which are shown in Table 2-2. Carcinogenic effects, which are expresser! only in exposed inclivi~iuals themselves, are estimated to account for the bulk (80%) of the overall risk of harm. The lifetime risk of developing a fatal cancer from irradiation is estimated to be 5 x 10~2/Sv for a member of the general population. Nonfatal cancers, although projected to be pro~iucect more frequently than fatal cancers, were judged to contribute less to the overall health impact of irradiation because of their lesser severity in affecter! individuals ant! were, therefore, weighted accordingly (Table 2-21. Of the total risk of heritable effects, about one-fourth is projected to be expressed in the first two generations alone, the remainder during subsequent scores of generations. This table indicates that if ~ 00 people were each to receive ~ Sv of radiation over their lifetimes, which is about 300 times greater than the overall average annual natural background level of radiation in the United States, five wouIct be expecter! to die from cancer induced by that radiation. Since it accounts for the great bulk of the potential harm that might be attributed to low-level radiation, the above risk estimate for fatal cancer is often used to calculate the expecter! number of fatalities attributable to low- dose irradiation in a population. For example, if one million persons were each exposed to a dose equivalent to that received from a transcontinental plane ride (0.025 mSv), the resulting collective dose (25 person-Sv) would be estimated to cause one extra fatal cancer in the population in addition to the 200,000 fatal cancers that would! be expected to occur in the same population from all other causes combined. Because the added risk, if any, is calculated to be such a small fraction of the total cancer risk, it is not surprising that epidemiological data have revealeci no significant differences in the rates of cancer or other diseases among populations exposed to far larger variations in natural background} radiation levels (NRC 1990a).

OCR for page 33
38 YUCCA MOUNTAIN STANDARDS Table 2-l Average Amounts of ionizing Radiation Received Yearly by a Member of the U.S. Populationa Source Doseb (mSv/yr) (/0) Natural Radons 2.0 55 Cosmic 0.27 Terrestrial 0.28 Internal 0.39 ~ ~ Total Natural 3.0 82 Anthropogenic Medical X-ray diagnosis 0.39 Nuclear medicine 0.14 Consumer products 0.10 Occupational ~ 0.01 Nuclear fuel cycle ~ 0.01 Nuclear fallout < 0.01 Miscellaneous ~ 0.01 Total anthropogenic Total Natural ant} Anthropogenic 11 <0.3 0.63 < 0.03 < 0.03 < 0.03 18 3.6 100 a From NRC (199Oa) and NCRP (1987a) b Average effective dose equivalent c Dose to bronchial epithelium alone d DOE facilities, smelters, transportation, etc.

OCR for page 33
PROTECTING HUMAN HEALTH 39 Table2-2. Estimated Frequencies of Radiation-Induced Fatal Cancers, Nonfatal Cancers, and Severe Hereditary Disorders, Weighted for the Severity of their Impacts on Affected Individualsa No. of cases per 100 per Svb Fatal cancers Nonfatal cancers Severe heredity disorders Total 5.0 1.0 1.3 7.3 a From ICRP (1 99 1 ~ b Numbers of cases, weighted for severity of their impacts on affected individuals over their lifetimes, attributable to low-level Radiation of a population of all ages. DEVELOPMENT OF RADIATION PROTECTION STANDARDS There is a worIc~wide interest in the development of radiation protection standards, including those for the disposal of high-level radioactive waste, and a consiclerable belly of analysis anti informer} judgment exists from which to draw in formulating a standard for the proposed Yucca Mountain repository. EPA's process for setting the Yucca Mountain standard is presumably not bound by this experience, but a sound technical approach should include a review of other relevant work to date. Accorciingly, we summarize below the status of relevant work on racliation protection standards both in the Uniter! States and abroad. General Consensus in Radiation Protection Principles and Standards A number of international ant! nongovernmental national bodies (such as the International Atomic Energy Agency (IAEA), ICRP and NCRP) have recommended radiation protection principles ant! standards.

OCR for page 33
40 YUCCA MOUNTAIN STANDERS These recommendations, in turn, usually are considered by the national agencies that set radiation protection standards, which then are codified into pertinent rules and regulations. Of the international bodies, the International Commission on Radiological Protection (ICRP) is perhaps the most influential. Its counterpart in the U.S. is the National Council on Radiation Protection and Measurements (NCRP). In the United States, several agencies establish radiation protection standards in their areas of responsibility. Among them are the following: the U.S. Environmental Protection Agency (EPA), the U.S. Nuclear Regulatory Commission (USNRC), ant! the U.S. Department of Energy (DOE). These three agencies play key roles in programs involving public health and safety, environmental protection, health ant! safety in the nuclear industry, and radioactive waste management and disposal. Recommendations for radiation standards to protect the public health and safety are frequently based on the analyses of radiation risks clevelopeci by the United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR) and the TCRP on the international level ant] by the Committees on Biological Effects of Ionizing Radiation (BEIR) in the United States. The most recent analyses are presented in the UNSCEAR (19X8) and NRC (1990a) reports, respectively. Concurrent with the development of radiation protection concepts internationally and in this country, a consensus has emerged among the organizations involves! in performing analyses and making recommendations (ICRP, NCRP, NRC's BEIR V, and UNSCEAR) anti those that promulgate regulations (EPA, USNRC, and DOE). This coalescence of views and resulting consensus can be seen in the general uniformity in the system of radiation close limitation, fundamental units and terminology, health effects factors, occupational and public dose limits, dose apportionment, anti use of the critical-group concept. The latter two concepts are defined and discussed later in this chapter. Consistent with the current understanding of the related consequences, ICRP, NCRP, IAEA, UNSCEAR, and others have recommended that radiation doses above background levels to members of the public not exceed 1 mSv/yr (100 mrem/yr) effective close for continuous or frequent exposure from radiation sources other than medical exposures. Countries that have considered national radiation protection standards in this area have endorsed the ICRP recommendation of 1 mSv per year radiation dose limit above natural background radiation for

OCR for page 33
PROTECTING HUMANHEALTH 41 members of the public. In the Uniter! States, DOE, in Order 5400.5, and USNRC, in 10 CFR 20, have set the close stanciarc! for public exposure to ionizing radiation at 1 mSv per year above natural background] level. EPA is in the process of cleveloping similar guidance for all U.S. federal agencies (EPA, 19931. This framework, with an effective dose limit of ~ mSv per year, is user] as a basis for protecting the public health from routine or expected anthropogenic sources of ionizing radiation (i.e., resulting from human activity) other than medical exposures. It inclucles any exposures to the public derives! from the management and storage of high-level radioactive defense waste ant! spent nuclear fuel. We note that guidance to date has been for expecter] exposures from actual routine practices. There is little guidance on potential exposures in the far distant future. {CRP (1985a) proposeci apportionment of the total allowable radiation close from all anthropogenic sources of racliation, excluding medical exposures. Thus, for radioactive waste management, including high-level radioactive defense waste and spent nuclear fuel, the national authorities coup! apportion, or allocate, a Faction of the 1 mSv per year to establish an exposure limit for high-level waste facilities. EPA in 40 CFR 191 notes} that its requirement for the WIPP transuranic waste facility, at a level of 0.15 mSv/yr (l 5 mrem/yr), is consistent with ICRP's concept of apportionment. Most other countries also have endorsed the principle of apportionment of the total allowed radiation close. Apportionment values that have been established by various countries for high-level radioactive waste range from 5% to 30%, corresponding to radiation doses ranging from 0.05 mSv (5 mrem) per year to 0.3 mSv (30 mrem) per year. . Table 2-3 presents the limits established by various countries on individual exposure from high-level waste disposal facilities. The information in this table suggests a general consensus among national authorities and agencies to accept and use the principle of radiation dose apportionment. THE FORM; OF THE STANDARD A standard is a societally acceptable limit on some aspect of repository performance that should not be exceeded if the repository is to

OCR for page 33
42 YUCCA MOUNTAIN STANDARDS be judges! safe. There is, however, a variety of ways in which this limit can be formulated. It can, for example, be imposer! at several points in the chain of events that might ultimately lead to adverse effects on public health. Thus, the limit could apply to the amount of radionucliries released from the repository, to the racliation doses to persons resulting from those releases, to the number of health effects associated with the doses, or to the level of risk. Risk, close, or health effect limits can be states! for inclivicluals or for populations. We recommend the use of a stanciard that sets a limit on the risk to inclividuals of adverse health effects from releases from the repository. In this context, risk is the probability of an individual receiving an adverse health effect. it is essential to define specifically how to calculate this risk, however, for otherwise it will not be clear what number to use to compare with the risk limit established in the standards. From the scientific perspective, the calculation of health risks should take into account all of the uncertainties involved in analyzing repository performance over very long time periods. Because many of the elements of the calculation are not well known, they must be dealt with by using distributions that represent the analysts's state-of-knowledge. The first step in calculating risk is therefore to clevelop a distribution of closes received by indivicluals, taking into account all of the events that go into determining whether a dose is receiveci.2 A probabilistic distribution of the health effects associated with these closes can then be cleveloped as the product of each value of close received and the health effects per unit dose. In this report, we choose to define risk as the expected value of the probabilistic distribution of health effects.3 2 This does not mean that every event needs to be treated probabilistically; some might be represented by a single bounding estimate, for example. The definition does require, however, that all of the parameters that determine the dose be considered in developing the probabilistic distribution of dose. 3 It is both easier and common practice to calculate doses received over an individual lifetime. One reason is that the effects of radiation might not appear until years after the dose is received. The lifetime calculation can be annualized by dividing by the duration of an average lifetime. Since this annualized risk is often more convenient for comparison to other risks, we recommend it be used.

OCR for page 33
PROTECTING HUAtINHEALTH Table 2-3 Quantitative High-Leve' Waste Disposal Objectives/Criteria at International Level and in OECD Countriesa Organization/Cou Main Other Main Comments ntry Objective/Crite Maturers) ria NEA (1984) Max. indiv. risk Individual No consensus on objective risk/dose= best ALARAJoptimiz 10-5/yr criterion to judge ation (all sources) long-term acceptability ICRP 1 mSv/yr Both prob. and ALARA useful, Publication 46 (normal doses should be notably to (1985) evolution taken into compare scenarios) account in alternatives, but 1 0~5/yr ALARA might not be the (probabilistic most important scenarios) for siting factor individuals (all sources) IAEA ICRP Also includes . Safety Series 99 Publication 46 qualitative . (1989) technical criteria . on disposal system features and role of safety analysis and quality assurance CANADA Max. indiv. risk Period of time Additional AECB regul. obj. 1 0~6/yr for qualitative, Document R. 104 demonstrating nonprescriptive (1987) 104yr requirement and No sudden and guidelines in dramatic regulatory increase for documents times > 1 04yr No explicit optimization required 43

OCR for page 33
56 YUCCA MOUNTAIN STANDARDS occur. We recognize that there are significant uncertainties in the supporting calculations and that the uncertainties increase as the time at which peak risk occurs increases. However, we see no technical basis for limiting the period of concern to a period that is short compared to the time of peak risk or the anticipated travel time. Nevertheless, we note that although the selection of a time perioc! of applicability has scientific elements, it also has policy aspects that we have not addresseci. For example, EPA might choose to establish consistent policies for managing risks from disposal of both long-lived hazardous nonradioactive materials and radioactive materials. Another time-relatec! regulatory concern can affect the formulation of the safety stanciarci. This is based on ethical principles, and is the issue of intergenerational equity (Berkovitz, 1992; Hol(lren, 1992; Okrent, 19941. Whether and how best to be fair to future generations is an important societal question. Although current generations are assumer} to have benefited from activities, such as electricity production or national defense programs that have causer] radioactive wastes to accumulate, far future generations will not benefit clirectly, but might be exposed to risks when any radioactive materials eventually escape the proposed repository. In crafting standards, EPA should as a matter of policy address whether future generations should have less, greater, or equivalent protection. The responsible institutions have considered the question of the protection to be afforded future generations. For example, in her presentation to us, Margaret Federline (USNRC, personal communication, May 27, 1993) spoke about a "societal pledge to future generations" that wouIc! "provide future societies with the same protection from radiation we wouic! expect for ourselves." The {AEA document, Safety Principles ant] Technical Criteria for HLW Disposal, Safety Series 99, has as one objective the "responsibility to future generations." Under this responsibility to future generations, IAEA recommends that "the degree of isolation of high-level radioactive waste shall be such so there are no predictable future risks to human health or effects on the environment that would not be acceptable today." In this IAEA establishes that "[tithe level of protection to be afforded to future individuals should not be less than that provided today." A health-based risk standard could be specified to apply uniformly over time and generations. Such an approach wouici be consistent with the principle of intergenerational equity that requires that the risks to future

OCR for page 33
PROTECTING HUMAN HEALTH 57 generations be no greater than the risks that wouIc! be accepted today. Whether to aclopt this or some other expression of the principle of intergenerational equity is a matter for social judgment. PROTECTING THE GENERAL PUBLIC Earlier in this chapter, we recommend the form for a Yucca Mountain standard! based on inclividual risk. Congress has asked whether standards intended to protect individuals wouIc! also protect the general public in the case of Yucca Mountain. We conclude that the form of the standards we have recommendecl would do so, provider] that policy makers and the public are prepared to accept that very low radiation doses pose a negligibly small risk. This latter requirement exists for all forms of the stan~iarcis, including that in 40 CFR 19 ~ . We recommenc! adciressing this problem by adopting the principle of negligible incremental risk to individuals. The question posed by Congress is important because limiting individual dose or risk lions not automatically guarantee that adequate protection is provider] to the general public for all possible repository sites or for the Yucca Mountain site in particular. As ciescribed in the previous section, the individual-risk stanciarc! should be constructed! explicitly to protect a critical group that is composed of a few persons most at risk from releases from the repository. The standards are then set to limit the risk to the average member of that group. Larger populations outside the critical group might also be exposed to a lower, but still significant, risk. It is possible that a higher level of protection for this population represented by a lower level of risk than the one established by the standards might be considered. For purposes of this discussion, the "general public" can be thought of as inciu~iing global (hemispheric or continental) populations that might receive very small risks from repository releases, as well as local populations that lie outside the critical group but that might still be expose(l to risks not much lower than those imposed on the critical group. The issues are different for these two types of populations, and we discuss them separately.

OCR for page 33
58 YUCCA MOUNTAII;STAND~=S PROTECTING THE GLOBAL POPULATION Radiation releases from a repository can in principle be distributed to a global, or other large and dispersed population, in several ways. For example, food contaminated by raclionuclicles could be shipped to regions far from the repository area, or contaminates! ground] water could enter a major river and the drinking water supplies that it serves. The global distribution of releases from a repository is assumed as the exposure scenario for the containment requirements in EPA's regulation 40 CFR 191. In the case of Yucca Mountain, there would be no releases to major rivers, ant! therefore the most likely pathways for global distribution are gaseous releases of carbon dioxide containing the radioactive isotope of carbon, i4C, that eventually will escape from the waste canisters, or by widespread distribution of foodstuffs grown with contaminated water. In general, the risks of radiation proclucec! by such wicle dispersion are likely to be several orders of magnitude below those to a local critical group. As noted earlier in this chapter, however, the "linear hypothesis" implies that even very small increments to background closes might cause effects from cancer induction in the same ratio (5xI0-2/Sv) as larger doses. Using the linear hypothesis to calculate the effects of very low doses on large populations requires multiplying this factor by the cumulative dose imposed on populations numbered] in the trillions over the life of the repository. There are, however, important cautions to be noted with this procedure. With respect to small increments to natural background] radiation levels, the BEIR V report (NRC 1990a) states that: Finally, it must be recognized that derivation of risk estimates for low doses and dose rates through the use of any type of risk mode] involves assumptions that remain to be validated. At low doses, a mode! dependent interpolation is involved between the spontaneous incipience and the incipience at the lowest doses for which data are available. Since the committee's preferrer! risk models are a linear function of dose, little uncertainty should be introducer! on this account, but departure from linearity cannot be excluder! at low closes below the range of observation. Such departures could be in the direction

OCR for page 33
PROTECTING HUMAN HEALTH of either an increased or decreased risk. Moreover, epidemiologic data cannot rigorously exclude the existence of a threshold in the millisievert dose range. Thus the possibility that there may be no risks from exposures comparable to external natural background radiation cannot be ruled out. At such low doses and dose rates, it must be acknowledged that the lower limit of the range of uncertainty in the risk estimates extends to zero.8 59 The doses to global populations involved in gaseous release from Yucca Mountain are likely to be well below the mSv range noted in BEIR V. For example, let us assume that the repository inventory of 91,000 Ci (3.37 x 10~5 Bq) (Wilson et al., 1994) of i4C is released into the air over 10,000 years. Using EPA's dose conversion factor 1.1 x 10-~ Sv/Bq (EPA, 1992), the population dose over 10,000 years would be 3.7 x 105 person-Sv, or an average of 37 person-Sv/year over the 10,000-year period (Nygaard et al., 19931. Assuming that the ~4C is well mixed with air over the globe, and for an average global population of 12 billion people during this period, the corresponding average individual dose rate is 3.1 x lO~9 Sv/yr (3.1 x 10-4 mrem/yr). For comparison, the dose set by EPA in 40 CFR 191 is I.5 x 10~4 Sv/yr (15 mrem/yr), and this is the limit to be applied for the persons likely to receive the highest doses from the repository. Therefore, there is great uncertainty about the number of health effects that would be imposed on the global population because of the difficulties in interpreting the risks associated with such small incremental risks from ~4C releases at Yucca Mountain. NEGLIGIBLE INCREMENTAL RISK To acidress scenarios of widespread but extremely low-level doses, the radiation protection community has introduced the concept of negligible individual dose. The negligible individual dose is defined as a level of effective dose that can, for radiation protection purposes, be dismissed from consideration. NCRP has recommended a value of 0.01 In this paragraph "low doses" applies to very small increments to the dose from the natural background.

OCR for page 33
60 YUCCA MOUNTAIN STANDARDS mSv/yr (I mrem/yr) per radiation source or practice (NCRP, 1993), which currently would corresponci to a projected risk of about 5 x 1 0~7/yr for fatal cancers, assuming the linear hypothesis. In its considerations, NCRP clecideri on this level of close or risk taking into account risk in relation to: l . Natural risk of the same health effects; 2. Risk to which people are accustomed; 3. Estimated risk for the mean and variance of natural background radiation exposure levels; 4. Perception of, and behavioral response to, risk levels; ant! 5. Difficulty in detection ant} measurement of dose anti health effects. Others over the years have advocated the use of a negligible otiose or risk level (Comer, 1979; Eisenbud, 1981; Schiager et al., 198619. The general consensus of these authorities was that a negligible value would be useful in many applications. Federal and state approaches for the regulation of chemical carcinogens are in keeping with this view, which generally take a 10-6 lifetime risk as an acceptable level (Travis et al., 1987; EPA, 1991), as are the exposure limits for radioactive waste adopted by most nations in the Organization for Economic Cooperation and Development (OECD) (Dejonghe, 1993~. The Federal German Radiation Protection Commission, for example, has recommender! ignoring individual closes of less than 0.003 mSv per year (Smith and Hocigkinson, 19881.~ We believe that the concept of a negligible incremental dose can be extended to risk anti can be applied to Yucca Mountain. Defining the level of incremental risk that is negligible is a policy judgment. We suggest the risk equivalent of the negligible incremental dose recommended by the NCRP as a reasonable starting point for cleveloping consensus in a rulemaking process. For example, the average dose to a member of the global population from exposure to i4C from the repository 9 Where authors use "negligible dose" or "negligible risk" the teens should be understood as increments to the unavoidable background radiation. In life, there is no zero dose and no zero risk. id Note that this is equivalent to an annual risk of fatal cancer of about 1 .SxlO~7/yr.

OCR for page 33
PROTECTING HUMAN HEALTH 61 is estimated to be about 3 x 10-9 Sv/yr, corresponding to a risk of fatal cancer of I.5 x 10~~/yr or about 1~8 per lifetime. As indicated earlier, NCRP has recommended a negligible incremental dose that corresponds to a risk of 5 x 10~7/yr (NCRP, 1993~. Therefore, if the NCRP recommendation were aclopted, the effects of gaseous ~4C releases on individuals in the global population would be considered negligible. PROTECTING LOCAL POPULATIONS Persons in some populations outside the critical group might be exposed to risk from repository releases in excess of the level of negligible incremental risk. As individuals, these persons would be (by definition and in practice) exposed to less risk than the risk limit establisher! by the stanciarc! for the critical group. If many persons were exposed to this incliviclual risk, however, the total number of health effects that couIci occur might be relatively large, particularly if integrated over a very long period of time. We know of no analysis that has addressed the spatial distribution of radiation closes and risks near Yucca Mountain at the distant future times when individual doses and risks would be at their maximum. It should be feasible to determine a spatial distribution of potential concentrations in ground water or air and a spatial distribution of individual doses ant! risks, employing the same types of exposure assumptions used for calculating doses ant} risks to members of a critical group (see Chapter 3~. However, the total number of fatal cancers cannot be known without knowledge of the number of future persons residing in the Yucca Mountain vicinity. This number is obviously unknowable. Even if EPA were to define it arbitrarily through a rulemaking process, comparing the total population risk against some cleaned figure-of-merit in order thereby to decide on whether to accept or reject a repository seems too arbitrary to be useful. Population-Risk Standard As an example of the difficulty of framing an absolute population- risk standard, we considered normalizing the population risks as a means to avoid the difficulty of not having a technical basis for knowing the total

OCR for page 33
62 YUCCA MOUNTAIN STANDARDS population at risk. Such a regulatory scheme might require that the integrated population risk over a given period (one generation, for example) be limited to some fractional risk in the affected population. A specific hypothetical example would be to require that the integrated population risk must produce fewer than x health effects per N people during a defined interval of time. Framed this way, however, the standard looks very much like an individual-protection standard: each person outside the critical group wouIci have an indiviclual lifetime risk limited to x/N. As a matter of policy, it is certainly legitimate to (lesire to protect a smaller group (the critical group) by limiting indiviclual risk to a certain level, ant! also to protect a larger group (the nearby population) with a different but still meaningful risk limit. However, this approach is not a collective-risk protection scheme - it is merely a two-tierec] inclividual-risk protection scheme. Spatial Gradient in Risk -An alternative approach that does have a technical basis is consideration of the spatial distribution of individual risks near the critical group, at the distant future time when the critical-group risk is highest. Such a spatial distribution has a technical significance because it depends on the characteristics not only of the Yucca Mountain physical site but also of the waste form and the engineered and geologic barriers of the repository (design. Furthermore, a risk distribution with a steep spatial gradient-that is, a distribution in which the indiviclual risks become smaller relatively quickly with increasing distance from the location of the highest individual risks seems obviously preferable to a distribution with a more gradual spatial gradient, all other things being equal. This is because a steeper spatial gradient implies smaller integrated population risks than does a more gradual gradient for the same spatial distribution of population. This observation cannot provide information for discriminating between an "acceptable" repository and an "unacceptable" one without an acceptable level of risk for comparison purposes. However, we have not been able to identify a technically based figure-of-merit that could be used to judge the compliance acceptability of a given spatial risk gradient. To

OCR for page 33
PROTECTING HUMAN HEALTH 63 use the gradient in an absolute sense, one is facet) with not only selecting a time interval of concern, which is arbitrary, but also defining the future nearby population. For the simpler task of adequately characterizing the exposure scenarios leacling to calculation of risks to a critical group, we have concluded that a feasible procedure can be developed using known distributions of physical and chemical parameters ant! defensible assumptions on lifestyles; in other worcis, there is a reasonable technical basis for a critical-group calculation. For identifying the size, the distribution and the varied lifestyles of a larger population, more assumptions of greater uncertainty would! be required. The resulting data for a risk assessment would become so arbitrary that no adequate decision basis would result. We therefore conclude that there is no technical basis for establishing a population-risk standard that would limit the risk to the nearby population for a Yucca Mountain repository. PREFERRED FORM OF THE STANDARD Although we have coucher! the discussion of the last two sections in terms of an individual-risk standard, we noted in an earlier section of this report that there are several possible forms of standard that could be used. We end this chapter by explaining why we conclude that the indiviclual-risk form has scientific advantages over the others. Release Limits. It is possible to state the standard in terms of a limitation on the amount of radionuclides crossing an imaginary boundary that encloses the repository. The limit generally would be placed on cumulative release over a specified time period. This is the approach used by EPA in 40 CFR 191, which relies primarily on a table of maximum allowable cumulative radioactive releases to the accessible environment for a period of 10,000 years. A release limit has the appearance of simplicity because it focuses on the amount of radionuclides released from the repository across some specified boundary. This form of standard does not provide any information about how these releases affect public health, however, ant! so is incomplete unless coupled with a calculation of inclividual (or population) risk (or dose or health effects). If one is interested! in this information on public health for a specific site, it is good scientific practice to incorporate specific data about the site into the calculation. If that is

OCR for page 33
64 YUCCA MOUNTAIN STANDARDS done, essentially all of the calculations clescribed in Chapter 3 are required. The advantage of our recommendation is that these calculations are to be clone using a methoclology approver! by a rulemaking, with all calculations explicit to the public. Hence, we conclude that a release limit for a site- specific standarci fines not reduce scientific complexity or uncertainty. Without calculations of dose or risk, a release standard appears arbitrary. Other than the appearance of simplicity, there seem to be no other advantages to a release-limit form of the standards. It floes not produce information that is easy to understand or to compare with other risks. Note that no other standard listen! in either Table 2-3 or 2-4 is expresser! as a release limit. A population standards, such as the one that appears to be the basis for the release limit in 40 CFR 191, establishes a total number of health effects permitted over some time period 1,000 in 10,000 years, in the case of 40 CFR 191. This form of standard does not provide a basis for assessing the risk to the individuals in the critical group, or for local populations nearby. Therefore, a population standard alone is insufficient to protect the population most at risk and, probably for this reason, 40 CFR 191 contains a parallel individual standard. Also, as discussed earlier in this chapter, assessing compliance with a standard designed to protect the global population involves highly uncertain calculations because of the extremely low incremental doses to which large numbers of persons may be exposed. We have recommended the use of the concept of negligible incremental risk to inclividuals as a preferable way of dealing with these uncertainties at the outset. An indivicI?val standarat is needed, however, and the issue is whether to state it in terms of dose, health effects, or risk. In Section 801, Congress directs EPA to use individual close. As mentioned above, we recommend using the risk form for the following reasons: A risk-based standard would not have to be revised In subsequent rulemaking if advances in scientific knowledge reveal that the dose-response relationship is different from that envisaged today. Such changes have occurred frequently in the past, ant! can be expected to occur in the future. For example, ongoing revisions in estimates of the Or, equivalently, a cumulative dose standard.

OCR for page 33
PROTECTING HUMAN HEALTH 65 radiation doses received by atomic bomb survivors of Hiroshima and Nagasaki may significantly modify the 2. apparent dose-response relationships for carcinogenic effects in this population, as have previous revisions in dosimetr (see Straume et al., 19921. Risks to human health from different sources, such as nuclear power plants, waste repositories, or toxic chemicals, can be compared in reasonably understandable terms. Doses or releases have to be states] in radiation units Sieverts or Becquerels that are not easily understood by the general public and that can only be compared conveniently with other sources of radiation or radioactivity. Although we recommend a risk-basec! standard rather than the dose-basec! standard in Section 801, they are closely related. We define risk as the expected value of the probabilistic distribution of health effects. The distribution of health effects is derived from a distribution of dose and the expected health effects per unit dose. Consequently, in answer to congressional question No. 1, we believe that a health-based individual standard will provide a reasonable standard! for protection of the general public. However, we recommenc! that this be a risk-based, rather than a dose-based standard.

OCR for page 33