Press Release

Table of Contents

Acknowledgements

Main findings

Recommendations

I. Introduction

II. National Primary Drinking Water Regulations – Radionuclides
A. Bone dose estimation in ICRP 2
B. Bone dose estimation, present-day dose conversion factors
1. Bone doses according to FGR 11
2. Bone doses according to FGR 13

III. Conclusions

IV. Costs

V. Estimating the impact of residual radioactivity

VI. Other risks and radionuclides

References

II. National Primary Drinking Water Regulations – Radionuclides

In 1959, the National Bureau of Standards published its Handbook 69 (NBS 69), which established the maximum permissible average concentrations of radionuclides in air and water calculated on the basis of a 5 rem dose to the whole body, and a 15 rem dose to the most exposed organ, also called critical organ, for each pathway and solubility class.⁶ As discussed below, a somewhat different method was used for bone-seeking radionuclides like radium-226 and plutonium-239. All these limits were established for radiation workers.⁷

ICRP 2 and NBS 69 also set forth the scientific approach for calculating these maximum permissible concentrations, with ICRP 2 providing significantly greater detail. A table adding data and correcting some errors in the 1959 version of NBS 69 was published in 1963, along with the original 1959 NBS 69 publication. In the text that follows, the term NBS 69 refers to this 1963 publication, since the EPA based its drinking water standards on it.

In March 1975, the EPA proposed, for the first time, National Primary Drinking Water Regulations for public water systems.⁸ The proposed rules for radionuclides were published in August of that year.⁹ The regulations for contaminants other than radionuclides were promulgated in December 1975;¹⁰ the rules for radionuclides were promulgated in July 1976.¹¹ The MCLs and dose limits were originally codified in 40 CFR 141.15 and 40 CFR 141.16, both of which have since been renumbered and consolidated, without change, into 40 CFR 141.66.¹²

In the final rule of July 1976, the EPA promulgated Maximum Contaminant Levels (MCLs) for radionuclides in public water systems either by directly specifying the MCL values (in picocuries per liter) or by specifying dose limits, which implied MCLs for drinking water, based on an adult water intake of two liters per day. The science underlying the standards was published in NBS 69. The drinking water limit for alpha-emitting radionuclides excluding uranium and radon, but including radium-226, was set at 15 picocuries per liter. There was a separate sub-limit for radium-226 and radium-228 of 5 picocuries per liter. For beta and photon-emitters the dose limit was 4 millirem per year (mrem/year) to the most exposed organ. (For radionuclides that are approximately uniformly distributed in the body, such as cesium-137 and tritium, the most exposed organ is considered to be the whole body.) The MCLs for beta- and photon-emitters were set according to the 4 mrem/year criterion, with a slight variation from this being adopted for tritium and for strontium-90. The limits for these categories have remained the same since that time.¹³ Detection limits and analytical methods for radionuclides were set forth in 40 CFR 141.25.

The rule as originally promulgated discusses natural and man-made radionuclides separately. However, it does not explicitly discuss the alpha-emitting transuranic radionuclides that are the subject of this report, but specifies only a gross alpha MCL. The gross alpha limit excludes only uranium and radon and it automatically includes the alpha-emitting, long-lived transuranic radionuclides of concern here, as these radionuclides are explicitly listed in the tables in NBS 69.

The following statement indicates the intent of the regulation that first established maximum contaminant limits for man-made radionuclides in drinking water:

Man-made radioactivity may enter the public water systems from a variety of sources. Such contamination is usually confined to systems utilizing surface waters. Past deposition of fallout materials from nuclear weapons tests, particularly strontium-90 and tritium, is probably the most important source of contamination. The dose equivalent to individual users of public water systems in some areas of the United States from this pathway is in the range of 1 to 2 millirem (mrem) per year. At present, the dose equivalent from public water systems contaminated by effluents produced in the nuclear fuel cycle is probably only a fraction of that due to fallout materials, though perhaps ranging up to 0.5 mrem per year. The dose equivalent from effluents released by medical, scientific, and industrial users of radioactive materials that enter the public water systems has not been fully quantified. Taken as a whole these users handle much smaller amounts of radioactivity than nuclear power facilities but (with the exception of tritium) their liquid releases and the resultant doses to man may be somewhat comparable.

EPA recognizes that the national use of radionuclides in medicine and industry and the utilization of nuclear power to supply energy needs will unavoidably lead to some radioactivity entering the aquatic environment so that the quality of some surface waters is likely to decrease slightly in the future. Even though the increase of radioactivity in drinking-water will normally be small, the Agency believes that the risk of future contamination warrants vigilance. It is the intent of the proposed monitoring and compliance requirements to provide a mechanism whereby the supplier of water can be cognizant of changes in the level of radioactivity in its water sources, so that the appropriate remedial measures may be taken.¹⁴

While this passage does not explicitly mention nuclear-weapons-related activities and facilities, their inclusion is clearly indicated, notably from the fact that fallout from nuclear weapons testing is discussed as the most important source of surface water contamination. It is also clear from the discussion of fallout that the intent was to consider the most important sources of contamination. The mention of industrial users also does not exclude weapons facilities (which handle radioactivity in considerably smaller amounts when compared to reactor core and spent fuel inventories in the commercial nuclear power sector). It is implicit, therefore, that there was no intent to exclude alpha-emitting man-made radionuclides from the vigilance and concern of the regulations.

The level of doses at which concern and vigilance were warranted in regard to man-made radionuclides was a few millirem per year. The maximum contaminant level for photon- and beta-emitters was set to 4 millirem per year because they were considered to be the most important sources of man-made radioactivity:

Considering the sum of the deposited fallout radioactivity and additional amounts due to effluents from other sources currently in existence, the total dose equivalent from made-made radioactivity is not likely to result in a total body or organ dose to any individual that exceeds 4 millirem per year…¹⁵

This quote shows that the sum of the doses from military and civilian activities was considered in evaluating the limit of 4 millirem per year that was set for beta- and photon-emitters in 1976. In fact, fallout was the single most important component of the dose from man-made radionuclides evaluated by the EPA.

The cancer fatalities from whole body exposure to 4 millirem per year from man-made beta and photon sources of radioactivity were estimated at between 0.4 and 2.0 deaths per year per million people exposed. This was comparable to the exposure to natural radium-226 and radium-228 estimated at 0.7 to 3 fatal cancers per year per million persons at the level of 5 picocuries per liter selected as the maximum contaminant level. The slightly higher fatality rate for radium (a factor of 1.2 to 1.8) at the allowable limit of 5 picocuries per liter must be seen in the context that it is a ubiquitous, naturally occurring radionuclide, with considerable variation in drinking water concentrations (which the EPA estimated at the time to be between 0.1 and 60 picocuries per liter).¹⁶ The EPA imposed considerable costs on public water systems by requiring remediation of those systems that had levels of radium greater than 5 picocuries per liter in order to bring them to the regulatory level. Further, the EPA mandated testing of water supplies and established detection limits (at the 95 percent confidence limit) that were considerably below the MCLs set forth in the regulation.¹⁷ The detection limits were set in order to ensure that the mandated MCLs would not be exceeded. In considering the mandated MCLs and detection limits, the EPA took technical, health, and economic considerations into account.

In looking to the future, the EPA did not anticipate that man-made radionuclides would result in a dose of more than 4 millirem per year from drinking water, because it believed that fallout would remain the main source and that this source would decrease with time due to the ban on atmospheric tests:¹⁸

The 4 millirem per year standard for man-made radioactivity was chosen on the basis of avoiding undesirable future contamination of public water supplies as a result of controllable human activities. Given current levels of fallout radioactivity in public water supply systems and their expected future decline, and the degree of control on effluents from the nuclear industry that will be exercised by regulatory authorities, it is not anticipated that the maximum contaminant levels for man-made radioactivity will be exceeded except in extraordinary circumstances.¹⁹

There is no explicit exclusion of alpha-emitting transuranic radionuclides from this statement. Also, the National Primary Drinking Water regulations explicitly mention strontium-90 in fallout. Hence, the regulations explicitly took into account a man-made radionuclide from a military activity — nuclear weapons testing — in protecting public water supplies from radioactive contaminants. Further, the critical organ listed in NBS 69 for strontium-90 and for the transuranic radionuclides that are the subject of this report was the same — the bone.

The language of the regulation indicates that the MCL in the range of 10 to 15 picocuries per liter for the alpha-emitting, long-lived transuranic radionuclides set at the time would have corresponded approximately to a bone dose of a few millirem per year according to then-prevailing estimation methods. We show in the next section, A. Bone dose estimation in ICRP 2, that was indeed the case. However, present-day methods result in far higher dose estimates, as discussed below in the section after next, B. Bone dose estimation, present-day dose conversion factors.

A. Bone dose estimation in ICRP 2

Bone dose was estimated in ICRP 2 (and NBS 69) as dose to the skeletal bone without the marrow. The reference bone-seeking radionuclide used by ICRP 2/NBS 69 was radium-226 and the reference amount was 0.1 microcurie of radium-226 in the skeletal bone. The amount of energy deposited in the bone each year corresponded to an absorbed radiation dose rate of about 3 rad per year, not accounting for relative biological effectiveness (RBE) of alpha particles. ICRP 2 used an RBE = 10, thus yielding an annual dose for a 0.1 microcurie body burden of radium-226 of 30 rem per year, according to the then-prevailing method of estimation.²⁰ Doses were calculated by estimating a whole-body or organ burden of the radionuclide assuming lifetime ingestion or inhalation at the MCL, for which values were given either in the workplace (40-hour workweek) or continuously (168 hours per week).

Some radionuclides, such the beta-particle-emitting strontium isotopes, were recognized even then to behave somewhat differently than radium-226 in the body in that they tended to concentrate in certain parts of the bone, while radium-226 is distributed less unevenly. Research since that time has validated that observation. For instance, the alpha-emitting, long-lived transuranic radionuclides tend to concentrate adjacent to the endosteal cells on the bone surface. Hence, these radionuclides deliver a considerably higher dose to the endosteal cells than would be indicated by an assumption of uniform distribution over a marrow-free skeleton.

In order to account for non-uniform distribution of several bone-seeking radionuclides, ICRP 2 suggested (and used) a factor of safety of 5 for such radionuclides when estimating maximum permissible levels of radionuclides in air and water for workers.²¹ The effect of this safety factor was to reduce the maximum allowable dose for workers from alpha-emitting, long-lived transuranic radionuclides to 6 rem per year, compared to 30 rem per year for radium-226. Correspondingly, the maximum permissible concentrations were also reduced by a factor of five.

This intent to reduce the maximum permissible dose to the bone by a factor of about 5 can be confirmed by estimating the dose corresponding to the maximum permissible burden of plutonium-239 in the bone of 0.04 microcuries specified in NBS 69. Using a value of 5.15 MeV per alpha particle and an RBE = 10, the annual dose corresponding to a bone burden of 0.04 microcuries of plutonium-239 is about 5.5 rem per year. Since the whole body and organ burdens in NBS 69 are rounded, this is in close agreement with the figure of 6 rem inferred by applying the safety factor of 5 to the radium-226 dose of 30 rem.

The MCL for soluble plutonium-239 set in NBS 69 corresponding to the 6 rem per year bone dose would be 5 x 10^-5 µCi/cc, or 5 x 10^-2 µCi/liter, or 50,000 pCi/liter. The current drinking water limit of 15 picocuries per liter in the absence of radium-226 corresponds to a bone dose of about 1.8 millirem per year (or 1.2 millirem per year corresponding to 10 picocuries per liter, which is the MCL for plutonium-239 in the presence of radium-226 at its MCL of 5 picocuries per liter).²²

The bone doses corresponding to 15 picocuries per liter for various alpha-emitting, long-lived transuranic radionuclides are shown in Table 1, estimated according to the method in NBS 69 which was the prevailing scientific understanding in 1976, when the EPA first promulgated the MCLs for radionuclides. All of these calculations follow NBS 69 in assuming soluble radionuclides when estimating doses to the bone from drinking water. An assumption of soluble forms of the radionuclides is reasonable (and in keeping with the regulation as originally promulgated) since it is likely that the radionuclides will be in that form if they are present in drinking water. The presence of insoluble colloidal forms is not excluded, but the likely presence of soluble forms makes it necessary to use the uptake coefficient for that form, which has been done throughout this report.

The NBS 69 (ICRP 2) calculations for bone dose are not directly comparable to present-day methods of dose estimation. NBS 69 specifies annual doses to the "bone," defined as the marrow-free skeleton. But Federal Guidance Report 11, which lays out methods of dose estimation that are the basis of EPA regulations at the present time, defines committed doses to two different parts of the bone — the "red marrow" and the "bone surface."²³ The latter is defined as the most exposed organ in Federal Guidance Report 11 for alpha-emitting, long-lived transuranic radionuclides because they concentrate adjacent to the endosteal cells, which are located on the bone surface. In other words, the understanding of what is the most exposed organ for alpha-emitting, long-lived transuranic radionuclides has evolved along with the methods of dose estimation since the MCLs were promulgated in 1976.

As shown in Table 1, the range of doses to the bone using a limit of 15 picocuries per liter for alpha-emitting, long-lived transuranic radionuclides estimated according to NBS 69 is approximately from 1.8 to 3 millirem per year. This is about the same as the doses estimated from man-made radionuclides, notably in fallout, in the safe drinking water regulation as promulgated in 1976. Hence we can infer that the intent of the rule was to limit the dose from drinking water to the maximum exposed organ, defined then as the bone, to approximately 2 millirem per year.

While the bone surface was not specified as a target organ for dose calculations in 1976, when the safe drinking water regulations were promulgated, it is possible to estimate the dose to the endosteal cells at a level of drinking water contamination of 15 picocuries per liter based on the NBS 69 dose conversion factors. For plutonium-239, the annual dose to the endosteal cells would be about 26 millirem per year.²⁴ The bone surface dose for the other radionuclides shown in Table 1 are about the same, except for Np-237, for which the figure is about 44 mrem per year. These estimated doses, which take into account the evolution of scientific understanding in the years after 1976, are far higher than what the safe drinking water regulations allow. The implied dose to the endosteal cells is about a factor of 14.6 higher for plutonium-239. All of these calculations were done within the framework of NBS 69, which was (and continues to be) the scientific guidance for the safe drinking water regulation.

B. Bone dose estimation, present-day dose conversion factors

Scientific understanding of radiation doses and harm from intake of radionuclides has advanced considerably over the years. Regulations have also evolved to some extent, though at a slower pace. Specifically, in the 1970s, the International Commission on Radiological Protection (ICRP) published ICRP 26 and ICRP 30 followed by ICRP 48 in 1986. The scientific work in these publications was incorporated by the EPA into Federal Guidance Report 11 in 1988. The doses from alpha-emitting, long-lived transuranic radionuclides in the new guidance issued by the EPA are much higher than those estimated by NBS 69 methods. Federal Guidance Report 11 is the report that is the basis of current EPA regulatory dose estimation methods. We will estimate bone doses according to Federal Guidance Report 11 (FGR 11) in this section. Then we discuss the same problem using Federal Guidance Report 13 (FGR 13), which is the most recent EPA Guidance, but not yet in force for regulatory calculations for doses from air and water.

1. Bone doses according to FGR 11

As touched upon above, several major changes have transpired from NBS 69 to FGR 11 so far as this analysis is concerned:

• The quality factor, or RBE, was increased from 10 to 20.

• The bone was divided into two different target organs, the "bone marrow" and the "bone surface," as compared to a single organ, the marrow-free skeleton, in NBS 69.

• The division of the bone into two organs in FGR 11 allowed the omission of the safety factor of 5 that was used in NBS 69 to account for selective, non-uniform deposition in the bone of certain radionuclides.

• NBS 69 used annual doses, while FGR 11 provides the conversion factors for committed doses.²⁵

While these technical changes are complex, it is possible to estimate the effect of the changes from NBS 69 to Federal Guidance Report 11 on doses in several different ways, each of which raises some technical issues. The approaches and issues are set forth in Table 2 using plutonium-239 as the reference alpha-TRU radionuclide.

Of these approaches, the first one is the least persuasive scientifically because it compares cumulative annual doses to cumulative committed doses. Since plutonium is eliminated from the bone very slowly (with a biological lifetime of several decades), most of the dose from intakes in the last years of a 70-year reference lifetime would be delivered after the full lifetime of even a long-lived person (even if one considers a ~100 year life, for instance). Hence, only the latter two approaches are scientifically reasonable. Both yield values for MCLs for alpha-emitting, long-lived transuranic radionuclides that are far below 15 picocuries per liter. However, they yield values also an order of magnitude different from each other — 0.08 picocuries per liter and 1.2 picocuries per liter. The approach shown in the second row is the most close to the intent of the drinking water regulation because it compares cumulative dose over a lifetime to the most exposed organ as defined in 1976 (marrow-free skeleton) and the most exposed organ as currently defined (bone surface). The last approach compares dose to the same organ (bone surface), which has scientific merit. However, it is not in accord with the intent of the regulation to limit dose in that the prevailing views of the most exposed organ (marrow-free skeleton in 1976 and bone surface in 1988) are no longer being compared. Hence, the most appropriate value to use for a new standard based on Federal Guidance Report 11 would be 0.08 picocuries per liter. However, since this is no longer the most recent scientific guidance published by the EPA, this factor would also need to be considered in the review of MCLs for alpha-emitting, long-lived transuranic radionuclides when they are reviewed in 2006.

2. Bone doses according to FGR 13

The most recent regulatory guidance for estimating doses is based on dose conversion factors published in ICRP 72. These have been incorporated into Federal Guidance Report 13, including the compact disk supplement, which has dose conversion factors for various ages published in a database.²⁶ The dose conversion factors are age-dependent and can be used to estimate committed doses for the remainder of life from the age of intake to age 70 years. This allows the estimation of total dose over a lifetime corresponding to a water contamination at 15 picocuries per liter.

The dose conversion factors in Federal Guidance Report 13 are generally somewhat lower than those in Federal Guidance Report 11. Therefore the total dose to the bone surface using the newer dose conversion factors in Federal Guidance Report 13 is roughly a factor of two lower than that estimated using FGR 11. In addition to the change in the dose conversion factors, water intake variation with age also needs to be considered. The current drinking water MCLs are based on an adult intake of 2 liters of water per day, excluding the water content of food. However, the water intake of children is smaller and there is also some gender variation. Further, children have a greater intake of fluids, notably in the form of milk. Therefore, we have done the Federal Guidance Report 13-based dose calculation using two sets of intake rates for various ages that are published in the literature. The first set corresponds to fluid intakes, including milk. The second set includes only water intake. These assumptions about intake rates are show in Table 3 below:

When total fluid intake is considered (i.e., Case 1 above) the cumulative lifetime dose to the bone surface from plutonium-239 over a 70-year period is about 15,500 mrem. For Case 2, water intake only, the lifetime bone surface dose is about 12,000 mrem. The corresponding dose to the maximally exposed organ under NBS 69 (the marrow-free skeleton) is 126 mrem. These doses are calculated by applying dose conversion factors specified in the relevant publications to the intake of plutonium in drinking water over a 70-year period. This last figure of 126 mrem can be viewed as the intent of the original regulation in terms of the dose to the maximally exposed organ from drinking water contaminated with plutonium to the maximum allowable limit of 15 picocuries per liter. If we compare the value of 126 mrem to the dose to the maximally exposed organ as estimated by the methods specified in Federal Guidance Report 13, we find that for drinking water intakes corresponding to Case 1, the MCL of 15 picocuries per liter is about 123 times too high and for Case 2, it is about 95 times too high. Therefore the most recent science would indicate a tightening of the current MCL for plutonium-239 (15 pCi/L) by about 123 times to about 0.122 picocuries per liter in the case of fluid intake case (Case 1) and by over 95 times to about 0.157 picocuries per liter for water intake only (Case 2). The results for the other alpha-emitting, long-lived transuranic radionuclides are similar, since the dose conversion factors are quite close to those of plutonium-239, with the exception of neptunium-237, for which the dose conversion factors are about a factor of two lower.

Footnotes

⁷ Until 1958 there were no separate radiation exposure limits for the public. They were the same as for workers. In 1958, the dose limits for the public were set at one-tenth the maximum allowable doses for workers (NBS 59 Addendum, page 5).

⁸ Fed. Reg. 1975/03/14.

⁹ Fed. Reg. 1975/08/14.

¹⁰ Fed. Reg. 1975/12/24.

¹¹ Fed. Reg. 1976.

¹² The changed numbering can be found in the 2004 edition of 40 CFR 141.

¹³ The limits were first specified in 40 CFR 141.15 and 40 CFR 141.16. An MCL for uranium of 30 micrograms per liter was established on December 7, 2000, in 40 CFR 141.66 (e), based mainly on the heavy metal toxicity of uranium to the kidney. The revision to 40 CFR 141 was announced in Fed. Reg. 2000.

¹⁴ Fed. Reg. 1975/08/14, page 34324, emphasis added.

¹⁵ Fed. Reg. 1975/08/14, page 34325, emphasis added.

¹⁶ Fed. Reg. 1975/08/14, page 34325.

¹⁷ Fed. Reg. 1976, page 28404.

¹⁸ Of the nuclear weapons states, only China was testing in the atmosphere at the time. China conducted its last atmospheric nuclear test in 1980.

¹⁹ Fed. Reg. 1975/08/14, pages 34325-34326, emphasis added.

²⁰ ICRP-2, 1959, page 13 and FGR 11, 1988, page 18. The current value of the RBE, often called the quality factor in the regulatory context, for alpha particles is 20.

²¹ FGR 11, 1988, pages 16-19.

²² This assumes that no Ra-228 is present.

²³ There is more recent federal guidance on the subject in Cancer Risk Coefficients for Environmental Exposure to Radionuclides, Federal Guidance Report No. 13. Washington, D.C., Environmental Protection Agency, 1999 (hereafter cited as FGR 13). This report also uses the same two parts of the bone as the target organs for which doses are calculated.

²⁴ This estimate is derived by using a mass of 120 grams for the endosteal cells corresponding to an overall skeletal mass of 7,000 grams. Further, it is assumed that one-fourth of the energy is deposited in the 120-gram mass of the endosteal cells, with the rest being deposited in other parts of the bone. This mass of the endosteal cells is specified in Federal Guidance Report 11. This gives a ratio of dose to endosteal cells of (7000/120)*0.25 = 14.6. All calculations assume that the dose to the bone permitted under NBS 69 at the specified MCL was 6 rem per year. There is some imprecision associated with the fact that the MCLs were rounded to one significant figure in NBS 69, but this is not significant in the present context.

²⁵ "Annual dose" corresponds to the amount of energy from ionizing radiation deposited in the target organ per unit mass of the organ in a single year. The dose in rem is then calculated by applying the RBE to the deposited energy. "Annual committed dose" corresponds to the amount of energy that would be deposited in the organ over the entire time that the radionuclide is present in the organ due to the intake of the radionuclide in a single year. If a radionuclide is eliminated rapidly from the body (say in a few days or weeks), as for instance is the case with tritium, then annual dose and committed dose are usually the same. But if the radionuclide is slowly eliminated from the target organ, over years or even decades (the latter is the case for alpha-emitting, long-lived transuranic radionuclides, their target organ being the bone), the dose to the bone from an intake in any given year is delivered over a period of decades after that. With the annual committed dose, the intake is over a year but the dose is delivered over a different period of time — and, in the case of alpha-emitting, long-lived transuranic radionuclides to the bone, a much longer period of time. Hence, the actual dose delivered to the person in the case of an intake of an alpha-emitting, long-lived transuranic radionuclide late in life (say a few years before death) is less than the full committed dose and less than the dose that would be delivered from the same intake early in life.

²⁶ FGR 13, 1999 and 2002 (the latter for the CD supplement, rev.1).