Radiation Standards: Scientific Basis Inconclusive, and EPA and NRC
Disagreement Continues (Testimony, 07/18/2000, GAO/T-RCED-00-252).

Pursuant to a congressional request, GAO discussed the regulatory
standards used to protect the public from the risks of low-level nuclear
radiation, focusing on: (1) whether current radiation standards have a
well-verified scientific basis; (2) whether the Environmental Protection
Agency (EPA) and the Nuclear Regulatory Commission (NRC) have come
closer to agreeing on exposure limits (how much radiation people can be
safely exposed to) in the safety standards; and (3) how implementing
these standards and limits may affect the costs of nuclear waste cleanup
and disposal activities.

GAO noted that: (1) U.S. radiation standards for public protection lack
a conclusively verified scientific basis, according to a consensus of
recognized scientists; (2) below certain radiation exposure levels, the
effects of radiation are unproven, despite many years of research
efforts; (3) evidence of these effects is especially lacking at
regulated public exposure levels--levels of 100 millirem a year and
below from human-generated sources; (4) at these levels, scientists and
regulators assume radiation effects according to what is commonly known
as the "linear no threshold hypothesis," or model; (5) according to this
model, even the smallest radiation exposure carries a cancer risk, and
risks double as the exposure doubles; (6) research into low-level
radiation effects continues, including studies attempting to
statistically correlate natural background radiation levels in the
United States and around the world with local cancer rates; (7) lacking
conclusive evidence of low-level radiation effects, U.S. regulators have
in recent years set sometimes differing exposure limits; (8) in
particular, EPA and NRC appear no closer to agreeing on exposure limits
today than in 1994; (9) the two agencies continue to favor different
policies and regulatory approaches for various nuclear cleanup and waste
disposal applications, especially those relating to groundwater
protection; (10) the disagreement involves EPA- and NRC-preferred
protection levels that are both well below the range where radiation
effects have been conclusively verified; (11) in this regard, the
disagreement essentially involves policy judgments and has complicated
efforts to clean up facilities, as well as planning for the prospective
Yucca Mountain, Nevada, high-level waste repository; (12) costs of
implementing radiation protection standards at nuclear cleanup and waste
disposal facilities vary from site to site; (13) long-term overall costs
could be immense, although these costs have not been comprehensively
estimated; (14) an indication of the potential costs is that agencies,
especially the Department of Energy, expect to fund hundreds of billions
of dollars in nuclear cleanup and waste disposal projects over many
years in the future; (15) differences in the costs of the EPA and NRC
regulatory approaches to radiation protection have not been
comprehensively estimated; (16) however, agency analyses indicate that
more restrictive radiation standards cost more to implement, as might be
expected; and (17) these analyses also generally show accelerating costs
to achieve the most restrictive protection levels.

--------------------------- Indexing Terms -----------------------------

 REPORTNUM:  T-RCED-00-252
     TITLE:  Radiation Standards: Scientific Basis Inconclusive, and
	     EPA and NRC Disagreement Continues
      DATE:  07/18/2000
   SUBJECT:  Radiation safety
	     Nuclear waste disposal
	     Safety standards
	     Environmental research
	     Cost analysis
	     Standards evaluation
	     Radiation exposure hazards
	     Environmental monitoring
	     Interagency relations
	     Hazardous substances
IDENTIFIER:  Yucca Mountain (NV)
	     Superfund Program

******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO Testimony.                                               **
**                                                              **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced.  Tables are included, but    **
** may not resemble those in the printed version.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
******************************************************************

GAO/T-RCED-00-252

RADIATION STANDARDS

Scientific Basis Inconclusive, and EPA and NRC Disagreement Continues
Statement of Ms. Gary L. Jones, Associate Director, Energy, Resources, and
Science Issues, Resources, Community, and Economic Development Division

United States General Accounting Office

GAO Testimony Before the Subcommittee on Energy and Environment,

Committee on Science, U. S. House of Representatives

For Release on Delivery Expected at 10 a. m. EDT Tuesday, July 18, 2000

GAO/ T- RCED- 00- 252

1

Mr. Chairman and Members of the Subcommittee: We are pleased to be here
today to discuss the regulatory standards used to protect the public from
the risks of low- level nuclear radiation. The scientific basis for these
standards has been in question, as well as what level of protection is
appropriate and adequate for the public. As you know, historically federal
agencies, including especially the Environmental Protection Agency (EPA) and
the Nuclear Regulatory Commission (NRC), have sometimes disagreed over how
restrictive U. S. radiation standards should be. They have set differing
standards, which include varying limits on radiation exposure to the public.
The standards cover regulatory applications such as cleaning up major
weapons production sites, decommissioning commercial nuclear power plants,
and potentially constructing an underground repository for the disposal of
highly radioactive waste at Yucca Mountain, Nevada.

On the basis of our June 30, 2000, report to Senator Pete Domenici, 1 our
statement today addresses three issues: (1) whether current radiation
standards have a well- verified scientific basis, (2) whether federal
agencies, particularly EPA and NRC, have come closer to agreeing on exposure
limits (how much radiation people can be safely exposed to) in the standards
since we reported on this issue in 1994, 2 and (3) how implementing these
standards and limits may affect the costs of nuclear waste cleanup and
disposal activities. In regard to the scientific basis of radiation
standards, we examined many scientific studies and interviewed recognized
scientists in the fields of radiation protection and radiation research. In
addition, we employed an expert consultant to help review scientific
research correlating natural occurring (background) radiation levels around
the world with local cancer rates.

In summary:

ï¿½ U. S. radiation standards for public protection lack a conclusively
verified scientific basis, according to a consensus of recognized
scientists. Below certain radiation exposure levels, the effects of
radiation are unproven, despite many years of research efforts. Evidence of

1 See Radiation Standards: Scientific Basis Inconclusive, and EPA and NRC
Disagreement Continues (GAO/ RCED- 00- 152, June 30, 2000).

2

these effects is especially lacking at regulated public exposure levels-
levels of 100 millirem a year and below from human- generated sources. 3 At
these levels, scientists and regulators assume radiation effects according
to what is commonly known as the “linear no threshold
hypothesis,” or model. According to this model, even the smallest
radiation exposure carries a cancer risk, and risks double as the exposure
doubles. The model is useful and relatively simple, but controversial. Some
scientists argue that the model overestimates radiation risks. Others take
the position that the model underestimates these risks. Research into low-
level radiation effects continues, including studies attempting to
statistically correlate natural background radiation levels in the United
States and around the world with local cancer rates. A promising long- term
research area is focusing on low- level radiation effects within human
cells, including a 10- year Department of Energy (DOE) program begun in
1999. Also, a major National Academy of Sciences reassessment of the status
of research into lowlevel radiation risks, called BEIR VII, is under way,
for which U. S. regulators have set high expectations and which is due to
conclude in 2001.

ï¿½ Lacking conclusive evidence of low- level radiation effects, U. S.
regulators have in recent years set sometimes differing exposure limits. In
particular, EPA and NRC appear no closer to agreeing on exposure limits
today than in 1994. The two agencies continue to favor different policies
and regulatory approaches for various nuclear cleanup and waste disposal
applications, especially those relating to groundwater protection. At
nuclear sites, EPA favors applying restrictive standards--- originally
applicable to community drinking water-- to limiting groundwater
contamination. The drinking water standards include contamination limits for
a long list of radioactive substances, equivalent in some cases to fractions
of a millirem a year. On the other hand, NRC favors less restrictive
standards that treat groundwater as one of various potential exposure means,
or “pathways,” within an allpathway exposure limit of 25
millirem a year. 4 The disagreement involves EPA- and NRCpreferred
protection levels that are both well below the range where radiation effects
have been conclusively verified. In this regard, the disagreement
essentially involves policy judgments-- not strictly scientific judgments.
The disagreement has complicated efforts to

2 See Nuclear Health and Safety: Consensus on Acceptable Radiation Risk to
the Public Is Lacking (GAO/ RCED- 94- 190, Sept. 19, 1994). 3 A millirem is
a commonly used unit of measurement of the biological effect of radiation.
The radiation

from a routine chest X- ray is equivalent to about 6 millirem. 4 All-
pathway exposure refers to exposure through soil, water, and air.

3

clean up nuclear facilities, as well as planning for the prospective Yucca
Mountain, Nevada, high- level waste repository. It does not appear that EPA
and NRC will readily agree on appropriate groundwater protection approaches
for Yucca Mountain. Also, while the two agencies are working on a memorandum
of understanding to clarify their regulatory roles related to nuclear
facility decommissioning, they have made little progress on this matter
since 1994 and before. Our June 2000 report to Senator Domenici concludes
that intervention by the committees of jurisdiction may be needed to resolve
the policy differences and clarify the regulatory responsibilities between
the two agencies.

ï¿½ Costs of implementing radiation protection standards at nuclear cleanup
and waste disposal facilities vary from site to site. For all sites
nationwide, long- term overall costs could be immense, although these costs
have not been comprehensively estimated. An indication of the potential
costs is that agencies, especially DOE, expect to fund hundreds of billions
of dollars in nuclear cleanup and waste disposal projects over many years in
the future. Differences in the costs of the EPA and NRC regulatory
approaches to radiation protection have not been comprehensively estimated.
However, agency analyses indicate that more restrictive radiation standards
cost more to implement, as might be expected. These analyses also generally
show accelerating costs to achieve the most restrictive protection levels.

Background

U. S. radiation standards protect the public from very low radiation levels.
Specifically, the standards regulate human- generated exposures to the
public in the range of 100 millirem a year and below. This regulatory range
is in the lowest portion of the low- level radiation exposure range- which
extends up to about 10,000 total millirem. The low- level range includes
natural background radiation, which varies locally in the United States, but
averages about 300 millirem a year. At exposure levels equivalent to or
below background radiation levels, radiation is commonly considered to be a
relatively weak source of cancer risk, although there is limited
understanding of such causation. 5 Above about 10,000 total millirem, the
high- level radiation

5 Low- level radiation, along with many other environmental and biological
events, may mutate cell structure. To counter these mutations, which occur
by the thousands daily in each human cell, the human body has active cell-
repair processes, though such processes are not entirely error free.

4

exposure range begins. In this range, extending without limit into the
hundreds of thousands of millirem or even more, the cancer risks of
radiation are better understood, and other, immediate health effects become
apparent. Above about 30,000 total millirem, radiation exposure is a
wellknown cause of cancer. At about 200,000 millirem of instantaneous or
short- duration radiation exposure, there can be blood cell changes,
infections, and temporary sterility. Above about 400,000 millirem, short-
duration exposure can cause death within days or a few weeks.

EPA and NRC administer the majority of federal radiation standards. EPA
issues environmental radiation protection standards as mandated under
Presidential Reorganization Plan No. 3 of 1970. NRC issues standards as part
of its mandate to regulate civilian sources of nuclear radiation, under the
Atomic Energy Act. Under the same act, DOE has issued public and worker
protection standards applicable on- site at the department's nuclear
installations. Both EPA and NRC have major regulatory roles related to
nuclear site cleanup and decommissioning and nuclear waste disposal. In
regard to nuclear cleanup, EPA administers Superfund, the legislation that
governs cleanups of federal and nonfederal facilities, and NRC regulates the
decommissioning of commercial nuclear power plants, as well as other
commercial nuclear facilities, under the Atomic Energy Act. DOE is involved
in nuclear site cleanup as the manager of over a dozen major nuclear weapons
production sites. In regard to Yucca Mountain, EPA has the role of issuing
standards to protect the public from releases of radioactive materials from
the facility. NRC has the role of issuing technical requirements and
criteria and licensing the facility. DOE is involved as the developer and
potential operator of the facility.

EPA and NRC have historically implemented different regulatory approaches in
their radiation standards, as we reported in 1994. EPA has implemented a
risk- based radiation protection approach, setting a range of acceptable
risk- between 1 chance in 10,000 and one chance in a million of an
individual getting cancer. In association with this approach, the agency
addresses individual environmental contamination sources, co- regulates
chemicals and radioactive substances, and seeks to protect both human health
and environmental resources. EPA's approach has been described as
“bottom up,” setting a relatively restrictive risk goal to be
pursued through the best available technology- but allowing less restrictive
acceptable risks in site- specific situations. In contrast, NRC favors a
dose- based, radiation- specific protection approach. The commission's
regulations focus on human health protection and “all pathways”
of exposure in the environment. NRC's approach has been described as
“top down,” setting a

5

relatively less restrictive dose limit but reducing doses well below the
limit in site- specific situations where the reductions are
“reasonably achievable.” (In implementing its standards, DOE has
historically implemented the same “top down” protection
approach.)

U. S. Radiation Standards Lack a Conclusive Scientific Basis

U. S. radiation standards for public protection lack a conclusively verified
scientific basis, according to a consensus of recognized scientists. Below
certain exposure levels, the effects of radiation are unproven. At these
levels, scientists and regulators assume radiation effects according to the
“linear no threshold hypothesis,” or model, under which even the
smallest radiation exposure carries a cancer risk. However, the model is
controversial among scientists, and decades of research into radiation
effects have not conclusively verified or disproved the model, including
studies attempting to statistically correlate natural background radiation
levels in the United States and around the world with local cancer rates.
Research is continuing, including a promising 10- year DOE program begun in
1999, addressing the effects of low- level radiation within human cells.
Also, the National Academy of Sciences is conducting a major reassessment of
the status of research into low- level radiation risks, called BEIR VII, for
which the regulators requesting the work have set high expectations.

According to a consensus of recognized scientists, below about 5,000 to
10,000 total millirem of exposure, the effects of radiation are unproven.
Evidence of these effects is especially lacking at regulated public exposure
levels- levels of 100 millirem a year and below from humangenerated sources.
The consensus view that we encountered among scientists and in the
scientific literature is that the research data on low- level radiation
effects are inadequate to either establish a safety threshold or to exclude
the possibility of no effects. Individual viewpoints differed. Some
scientists and studies held that the data support the existence of a safety
threshold- an exposure level below which there are no risks from radiation.
Other scientists and studies held that there is no such threshold and there
can be risks at even the lowest exposure levels. In addition, other
scientists and studies noted that risks from low- level radiation are
complicated and variable, depending on factors such as the type and amount
of radiation involved, body organs exposed, sex of the person, and/ or age
at exposure. For example, some researchers hold that children and fetuses
may be more at risk from low- level

6

radiation than adults. Some scientists and studies held that there are
considerable data to support the view that low levels of radiation can
actually be beneficial to health- the highly controversial theory of
hormesis. Proponents of hormesis argue that research indicating beneficial
effects has not been adequately considered in the “consensus”
scientific community.

Although conclusive evidence of low- level radiation effects is lacking,
regulators still have the task of developing radiation standards to protect
the public. In doing so, regulators routinely assume that low- level
radiation effects exist, according to the “linear no threshold
hypothesis” or model. According to this model, even the smallest
radiation exposure carries a cancer risk, and risks double as the exposure
doubles. This model is endorsed by national and international radiation
protection organizations and is used as a preferred model by EPA, NRC, and
DOE. It is thought by many to be a conservative “fit” to the
data, unlikely to underestimate the risks of radiation. However, the model
is controversial. Some scientists argue that use of the model to assess
risks from radiation may result in either over- or underestimating radiation
risks.

Decades of radiation effects research have neither verified nor disproved
the linear model. The research data on low- level radiation effects
generally include two different types of studies. One type follows the long-
term health of a studied population, seeking statistically significant
cancer effects, and is called epidemiology. Another type subjects animals or
tissue or cell cultures to radiation, seeking biological evidence of
radiation effects, and is called radiobiology. Epidemiology has been a key
basis for the linear model, including research evidence accumulated on over
85,000 Japanese survivors of the Hiroshima and Nagasaki atomic bomb blasts.
The Japanese data have well established high- level radiation effects, and
scientists have extrapolated this relationship to the low- level radiation
range as well- with considerable inherent uncertainty. Extrapolating from
high- level exposures, delivered instantaneously or for a short duration, to
low- level exposures delivered over years, may be subject to question. Also,
the estimated doses received by the Japanese survivors are still subject to
re- evaluation, even after many years of effort devoted to determining these
doses.

In addition, epidemiological studies have attempted to statistically
correlate natural background radiation levels in the United States and
around the world with local cancer rates, with inconclusive results. A
premise relating to such studies is that if the linear model of low- level
radiation effects holds, places with significantly higher background
radiation levels should have elevated cancer rates. With the help of an
expert consultant, we examined 82 such studies, done

7

in the United States, Europe, Asia, and South America. In the United States,
areas of high natural background radiation include the Rocky Mountains,
where levels are over three times higher than along the Gulf Coast. Also, in
some areas of the world, mean annual doses can be more than double the
average U. S. levels. Such studies are subject to methodological
difficulties, including the small size of the studied population and the
pursuit of small radiationcaused cancer effects that are difficult to detect
among all cancers in the population. The results of the studies were
inconclusive overall. Studies differed, though they generally found little
or no evidence of elevated cancer risks from high natural background
radiation levels. One prominent U. S. statistical study, by Bernard Cohen,
University of Pittsburgh, in 1995, found a strong tendency for lung cancer
rates to decrease with increasing radon exposures in 1,601 counties
nationwide. However, the study has been criticized by epidemiologists as
methodologically flawed because it is a compilation of average statistics in
these counties-- not data on individuals.

Conclusive evidence of radiation effects may not soon be obtained, but
radiobiological studies may hold more future promise than epidemiological
studies, according to researchers and regulators. Recently, there has been
interest in research into the cellular processes through which radiation
causes cancer, and since fiscal year 1999, the Congress has funded a DOE
research program targeting the biological effects of low- level radiation at
the cellular level. Many scientists and regulators we interviewed said this
type of research could eventually help to determine more conclusively the
effects of low- level radiation and their potential link to causing cancer.
The DOE program projects total funding of almost $220 million over 10 years.
The program is considered unique in that it is designed specifically to
better validate the effects of very low radiation levels, in areas such as
cells' response to radiation damage, thresholds for low- dose radiation
effects, and features distinguishing radiation- caused cell damage from
damage from other, intra- cellular causes.

U. S. regulators have concluded that a major reassessment of the status of
epidemiological and radiobiological research into low- level radiation
effects is warranted. At their request, a committee of the National Academy
of Sciences is conducting such a reassessment, called BEIR VII. The last
such Academy study was done in 1990. The 1990 analysis, called BEIR V,
established risk estimates that have been influential for U. S. regulators
in setting radiation

8

standards. 6 It also gave the linear model a qualified endorsement, stating
that the model was not inconsistent with the available research data, but
that at low radiation exposures, risks either less or greater than expressed
in the linear model could not be excluded. U. S. regulators have set high
expectations for the BEIR VII effort, which to a degree may shed light on
the controversy concerning the linear model. The effort is due to conclude
in 2001. In requesting BEIR VII, the regulators set expectations that the
committee would focus on areas not necessarily emphasized in the 1990 study,
including (1) any clear indications of the weight of evidence for radiation
risks at low doses and dose rates, (2) epidemiological studies on nuclear
workers, (3) evidence of radiation effects specifically at the very low
levels where regulators set radiation standards, and (4) evidence of
hormesis. However, according to scientists and agency officials, and on the
basis of the research evidence to date, it may be too much to expect BEIR
VII to fully resolve the current controversy by either validating or
disproving the linear model.

EPA and NRC Continue to Disagree on Radiation Standards

In 1994, we reported that EPA and NRC disagreed on radiation standards.
Today, they appear no closer to agreement. In the absence of conclusive
scientific evidence of low- level radiation effects, the two agencies
continue to have a policy disagreement concerning how much radiation risk is
acceptable to protect the public. The disagreement essentially involves
groundwater protection, at dose levels well below the range where radiation
effects have been verified. EPA prefers a more restrictive approach than
does NRC. Essentially, EPA favors specially protecting groundwater at
nuclear sites, regulating groundwater to drinking water standards.
Conversely, NRC favors including groundwater and other exposure means, or
pathways, under an allpathway exposure limit. The all- pathway exposure
limit is less restrictive than the drinking water standards. The
disagreement is affecting the implementation of nuclear site cleanup
regulations and the development of regulations for the disposal of highly
radioactive waste at Yucca Mountain.

The EPA- NRC disagreement involves policy judgments, not strictly scientific
or technical differences. For both the cleanup and decommissioning of
nuclear facilities and the disposal of nuclear waste, EPA's standards
reflect the agency's “bottom up” protective approach, setting a

6 BEIR VI was a 1999 Academy assessment of risks from radon.

9

relatively restrictive risk goal to be pursued through the best available
technology. Also, EPA is attempting to implement a consistent regulatory
policy- for both chemical and radioactive pollutants-- of protecting
groundwater as a national resource. In this respect, at nuclear sites EPA
(1) sets a risk- based limit of 15 millirem a year for exposure from all
pathways and (2) favors additional, more restrictive groundwater protection,
to the same standards the agency applies to drinking water in community
water supplies. In relation to various radioactive substances, these
standards set exposure limits that are equivalent in some cases to fractions
of a millirem a year. Conversely, NRC's approach does not include special
groundwater protection. The commission's approach reflects its “top
down” strategy-- setting a relatively less restrictive dose limit but
pursuing lower doses where reasonably achievable. NRC prefers to set
radiationspecific, dose- based standards, and the Commission includes
groundwater and other exposure means (or “pathways”) under an
all- pathway exposure limit of 25 millirem a year. (DOE also prefers an all-
pathway approach.)

In specific cleanup and waste disposal applications, the differing EPA and
NRC approaches have been implemented as follows. In 1995, EPA drafted
cleanup standards reflecting 15- millirem- ayear all- pathway protection,
plus separate groundwater protection. The agency withdrew the standards
unfinalized in 1996, after other agencies objected to them, and implemented
the same approach in 1997 in the form of nonbinding Superfund guidance. Also
in 1997, NRC finalized its own cleanup standards, in the form of
decommissioning standards reflecting all- pathway 25- millirem- a- year
protection. Both EPA and NRC issued proposed standards for the Yucca
Mountain high- level waste repository in 1999. EPA's draft standards reflect
15- millirem- a- year all- pathway protection, plus extra groundwater
protection, while NRC's draft standards reflect 25- millirem- a- year all-
pathway protection. (Under the Energy Policy Act of 1992, NRC's final
standards are to be consistent with EPA's final standards.)

The differing EPA and NRC approaches have contributed to various regulatory
complications. For example, in the 1990s, perceived dual regulation by EPA
and NRC has complicated the cleanup and decommissioning process at some
sites where both agencies' standards may apply. Such situations can lead to
duplication of effort, regulatory delays, and added compliance costs. Also,
such situations can raise public questions about what cleanup levels are
appropriate and safe. For example, in 1999, in individual situations at NRC-
licensed sites, EPA has indicated that it might not view cleanups performed
to NRC's standards as adequately protective under its

10

Superfund guidance. EPA considers such situations to be the exception, not
the rule. However, licensees, including two New England power plants in
1999, have construed EPA involvement in such a situation as a warning that
EPA could reevaluate the adequacy of a cleanup that has met NRC's
requirements. Also, as we have reported, EPA and DOE have had historical
differences concerning standards and acceptable risks for cleanups at DOE
sites. These differences have contributed to regulatory delays and higher
regulatory and cleanup costs while raising public questions about what
cleanup levels are appropriate. 7

Further, the EPA- NRC disagreement over standards for Yucca Mountain is
complicating planning for the repository. How the disagreement is resolved
could affect the technical credibility and acceptability of the final
standards that are to be issued, prospectively in the summer of 2000. In
large part, the disagreement has centered on the technical basis for EPA's
extra groundwater protection approach for the repository. In particular, the
National Academy of Sciences, mandated to recommend standards for the
repository, has commented that EPA has not provided a technical rationale
for its groundwater approach and that the agency is proposing to apply
outdated drinking water concentration limits to groundwater at the
repository. The limits are based on 1970s- era dose estimation methods. NRC,
DOE, and other commenters have raised similar criticisms. NRC and DOE have
also pointed out that EPA has not done a comprehensive analysis of the
health benefits and costs of its groundwater approach for Yucca Mountain.

EPA says that its proposed groundwater protection approach for the
repository is justified on policy grounds and is technically justifiable as
well. The agency seeks to protect groundwater at the site as an
environmental resource in a region where the population has been growing
quickly. In addition, EPA says that the standards for Yucca Mountain should
be in accord with agency policy of coregulating chemicals and radionuclides
according to similar regulatory requirements. EPA believes that its
regulatory approach has fully addressed the pertinent overall technical
issues related to setting radiation standards for the site. EPA officials
recognize that the drinking water concentration limits to be applied to
groundwater at the repository are outdated, but they said the agency is in
the process of updating the limits by the

7 See for example Nuclear Cleanup: Completion of Standards and Effectiveness
of Land- Use Planning Are Uncertain( GAO/ RCED- 94- 144, Aug. 26, 1994); and
DOE: Accelerated Cleanup of Rocky Flats- Status and Obstacles( GAO/ RCED-
99- 100, Apr. 30, 1999).

11

fall of 2000. Further, EPA officials agreed that the agency has not done a
comprehensive analysis of the health benefits and costs of the agency's
groundwater approach for Yucca Mountain. However, they are developing a
regulatory impact analysis to accompany their final standards. While
according to EPA this analysis will not constitute a specific technical
rationale for its groundwater approach, and will not be a comprehensive
cost- benefit analysis, the analysis will address in detail various
technical and cost issues related to the standards' implementation.

It does not appear that EPA and NRC will readily agree on appropriate
groundwater protection approaches for Yucca Mountain. Also, while the two
agencies are working on a memorandum of understanding to clarify their
regulatory roles related to nuclear facility decommissioning, they have made
little progress on this matter since 1994 and before. Our June 2000 report
to Senator Domenici concludes that intervention by the committees of
jurisdiction may be needed to resolve the policy differences and clarify the
regulatory responsibilities between the two agencies.

Costs To Implement Radiation Standards Vary But Could Be Immense In the Long
Term

The costs of implementing radiation protection standards at nuclear cleanup
and waste disposal facilities vary from site to site. For all sites
nationwide, the long- term overall costs could be immense, likely in the
hundreds of billions of dollars, although these costs have not been
comprehensively estimated. Also, EPA, DOE, and NRC analyses indicate that
(1) more restrictive radiation standards are more costly to implement than
less restrictive standards and (2) costs accelerate to achieve the most
restrictive protection levels.

The costs of nuclear cleanup and waste disposal are largely radiation
standards driven. Over the long term, DOE, as well as regulated activities,
may spend hundreds of billions of dollars in nuclear cleanup and waste
disposal projects, in large part to help protect the public from radiation
exposure. For example, DOE has projected funding for environmental cleanup
at its nuclear sites from fiscal year 2000 through fiscal year 2070 to be
anywhere from $151 billion to $195 billion (in 1999 dollars). And this
estimate could go higher. In addition, the Nuclear Energy Institute has
estimated over $38 billion in costs to NRC licensees to decommission their
nuclear

12

facilities, including nuclear power plants, in coming decades. Further, DOE
has estimated longterm funding of over $43 billion, and potentially over $55
billion according to the latest projections, for the Yucca Mountain
repository system, in large part to help ensure that the public is protected
from the high- level waste stored there. This estimate could also go higher,
considering that since 1993 there have been repository- performance- related
cost increases of over $10 billion to achieve added confidence that
performance requirements and radiation protection requirements can be met
over thousands of years. (Furthermore, in planning for commercial low- level
nuclear waste disposal, state compacts and unaffiliated states have incurred
almost $600 million in costs, although no disposal sites have yet been
built.)

Differences in the costs of the EPA and NRC regulatory approaches to
radiation protection have not been comprehensively estimated. However,
agencies' cost analyses indicate that more restrictive radiation standards
cost more to implement, as might be expected. Agencies routinely do such
cost analyses to support their nuclear regulatory efforts. Many such
analyses estimate both the costs and health benefits from meeting radiation
standards, relying on the linear model. Such analyses sometimes determine
hypothetical cancer deaths averted from meeting different protection levels,
as well as dollars expended per hypothetical cancer death averted. 8 We
examined numerous DOE, NRC, and EPA cost analyses, which the agencies
provided to us as best available data. Most of the analyses were site-
specific, but EPA attempted a nationwide analysis in 1996. The analysis, to
support a prospective EPA cleanup standard, addressed potential nuclear
cleanup costs for 16 generic types of facilities around the country, based
on actual DOE, NRC- licensed, and Department of Defense sites. The analysis
did not address overall soil cleanup costs for these sites. Instead, it
estimated incremental costs to clean up soil at these sites below a 100
millirem a year baseline, as shown in table 1.

8 Such analyses can be potentially controversial, relying on the linear
model to estimate population risks- i. e., projecting low radiation
exposures across large populations to enumerate hypothetical cancer deaths.

13

Table 1: Potential Incremental Costs to Achieve Different Soil Cleanup
Levels- EPA Analysis, 1996

Cleanup level achieved, millirem a year Incremental cost, in billions of
dollars 25 1 15 1.5 3 3.2

As shown in table 1, EPA estimated significant nationwide cost differences
to achieve different cleanup levels, including differences between the 15
millirem a year and 25 millirem a year levels, favored by EPA and NRC,
respectively, as all- pathway protection levels. In addition, some DOE and
NRC analyses of individual soil cleanup sites- either actual or generic
sites- showed cost differences in the multiple millions of dollars per site
between the 25- millirem- ayear and 15- millirem- a- year levels. (According
to DOE and NRC officials, soil cleanup analyses do not represent overall
site cleanup costs, which may include additional expenditures, such as for
the decontamination and removal of structures, as well as liquid waste
treatment.)

The EPA, DOE, and NRC analyses also generally showed accelerating costs to
achieve the most restrictive protection levels, below 10 millirem a year.
For example, a 1995 DOE analysis of the plutonium- contaminated Nevada Test
Site and test ranges estimated, from a 100- millirem- a- year baseline, cost
increases of over three times to achieve 25 millirem a year, and over six
times to achieve 15 millirem a year, but over 28 times achieve the 5-
millirem- a- year level. These accelerating costs can be shown graphically
in the form of a cost curve, as depicted in figure 1:

14

Figure 1: Cleanup Costs As A Function of Cleanup Levels- Nevada Test Site
Analysis, 1995 Agencies generally did not have analyses showing the cost
differences between EPA's groundwater protection approach and NRC's all-
pathway approach. However, two DOE analyses showed potential multi- million
dollar added costs to meet EPA drinking water standards in onsite aquifers
through long- term “pump and treat” techniques, involving
pumping the water out of the ground, treating it, and discharging it back
into the ground. Less aggressive approaches, such as allowing natural
attenuation or dilution of the contamination, were less expensive.

Mr. Chairman, the cost data and analyses we examined indicate that
protecting the U. S. public from the risks of low- level radiation is a
costly undertaking. This is especially the case at the currently regulated
public exposure levels of 100 millirem a year and below from humangenerated
sources- in some cases, levels of fractions of a millirem a year. Protecting
at such levels, well below the levels where radiation effects have been
verified, is essentially a policy judgment by regulators. Such an approach
may be arguably prudent, using the linear model as its fundamental
scientific basis. To the extent that the linear model is in question, new
and better research evidence relating to the validity of this model could
alter regulatory policies. In

15

this regard, the National Academy of Sciences BEIR VII effort is important
and bears watching. However, according to scientists and regulators,
conclusive evidence either validating or disproving the linear model may not
be forthcoming for years, despite the promise of ongoing radiobiological
research.

--- Mr. Chairman, this concludes my prepared statement. I will be pleased to
respond to any questions that you or Members of the Subcommittee may have.

Contact and Acknowledgement

For future contacts regarding this testimony, please contact me at 202- 512-
3464. Individuals contributing to the testimony included Duane G. Fitzgerald
and Dave Brack.

16

Orders by Internet

For information on how to access GAO reports on the Internet, send an e-
mail message with “info” in the body to

Info@ www. gao. gov Or visit GAO's World Wide Web home page at http:// www.
gao. gov

To Report Fraud, Waste, and Abuse in Federal Programs

Web site: http:// www. gao. gov/ fraudnet. html E- mail: fraudnet@ gao. gov
Automated answering system: 1- 800- 424- 5454
*** End of document. ***