[Federal Register Volume 70, Number 161 (Monday, August 22, 2005)]
[Proposed Rules]
[Pages 49014-49065]
From the Federal Register Online via the Government Publishing Office [www.gpo.gov]
[FR Doc No: 05-16193]



[[Page 49013]]

-----------------------------------------------------------------------

Part II





Environmental Protection Agency





-----------------------------------------------------------------------



40 CFR Part 197



Public Health and Environmental Radiation Protection Standards for 
Yucca Mountain, Nevada; Proposed Rule

  Federal Register / Vol. 70, No. 161 / Monday, August 22, 2005 / 
Proposed Rules  

[[Page 49014]]


-----------------------------------------------------------------------

ENVIRONMENTAL PROTECTION AGENCY

40 CFR Part 197

[OAR-2005-0083; FRL-7952-1]
RIN 2060-AN15


Public Health and Environmental Radiation Protection Standards 
for Yucca Mountain, NV

AGENCY: Environmental Protection Agency (EPA).

ACTION: Proposed rule.

-----------------------------------------------------------------------

SUMMARY: We, the Environmental Protection Agency (EPA), are proposing 
to revise certain of our public health and safety standards for 
radioactive material stored or disposed of in the potential repository 
at Yucca Mountain, Nevada. Section 801(a) of the Energy Policy Act of 
1992 (EnPA, Pub. L. 102-486) directed us to develop these standards. 
These standards (the 2001 standards) were originally promulgated on 
June 13, 2001 (66 FR 32074). Section 801 of the EnPA also required us 
to contract with the National Academy of Sciences (NAS) to conduct a 
study to provide findings and recommendations on reasonable standards 
for protection of the public health and safety. The health and safety 
standards promulgated by EPA are to be ``based upon and consistent 
with'' the findings and recommendations of NAS. On August 1, 1995, NAS 
released its report (the NAS Report), titled ``Technical Bases for 
Yucca Mountain Standards.'' In promulgating our standards, we 
considered the NAS Report as the EnPA directs.
    On July 9, 2004, in response to a legal challenge by the State of 
Nevada and the Natural Resources Defense Council, the U.S. Court of 
Appeals for the District of Columbia Circuit vacated portions of our 
standards that addressed the period of time for which compliance must 
be demonstrated. The Court ruled that the time frame for regulatory 
compliance was not ``based upon and consistent with'' the findings and 
recommendations of the NAS and remanded those portions of the standards 
to us for revision. These remanded provisions are the subject of 
today's action.
    Today's proposal incorporates multiple compliance criteria 
applicable at different times for protection of individuals and in 
circumstances involving human intrusion into the repository. Compliance 
will be judged against a standard of 150 microsievert per year (15 
millirem per year) committed effective dose equivalent at times up to 
10,000 years after disposal and against a standard of 3.5 millisievert 
per year (350 millirem per year) committed effective dose equivalent at 
times after 10,000 years and up to 1 million years after disposal. 
Today's proposal also includes several supporting provisions affecting 
DOE's performance projections. DOE will measure the median of the 
distribution of doses against the dose standard beyond 10,000 years, 
will calculate doses using updated scientific factors, and will 
incorporate specific direction on analyzing features, events, and 
processes that may affect performance.
    Section 801(b) of the EnPA requires the Nuclear Regulatory 
Commission (NRC) to modify its technical requirements for licensing of 
the Yucca Mountain repository to be consistent with the standards 
promulgated by EPA. NRC did incorporate EPA's Yucca Mountain standards 
into its licensing regulations and the regulatory time frame provision 
of these was similarly struck down by the Court of Appeals. Once 
revised regulatory time frame components of our standards have been 
promulgated, NRC must revise its licensing regulations to be consistent 
with our revised standards. The Department of Energy (DOE) plans to 
submit a license application providing a compliance demonstration. The 
NRC will determine whether DOE has demonstrated compliance with NRC's 
licensing regulations, which must be consistent with our standards, 
prior to granting or denying the necessary licenses to dispose of 
radioactive material in Yucca Mountain.

DATES: Comments must be received on or before October 21, 2005.

ADDRESSES: Submit your comments, identified by Docket ID No. OAR-2005-
0083, by one of the following methods:
    1. Electronically. If you submit an electronic comment as 
prescribed below, EPA recommends that you include your name, mailing 
address, and an e-mail address or other contact information in the body 
of your comment. Also include this contact information on the outside 
of any disk or CD-ROM you submit, and in any cover letter accompanying 
the disk or CD-ROM. This ensures that you can be identified as the 
submitter of the comment and allows EPA to contact you in case we 
cannot read your comment due to technical difficulties or we need 
further information on the substance of your comment. EPA's policy is 
that we will not edit your comment, and any identifying or contact 
information provided in the body of a comment will be included as part 
of the comment that is placed in the official public docket, and made 
available in EPA's electronic public docket. If EPA cannot read your 
comment due to technical difficulties and cannot contact you for 
clarification, we may not be able to consider your comment.
    i. Federal eRulemaking Portal: http://www.regulations.gov. Follow 
the on-line instructions for submitting comments.
    ii Agency Web site: EPA's preferred method for receiving comments 
is via its website, EDOCKET. EDOCKET is an ``anonymous access'' system, 
which means EPA will not know your identity, e-mail address, or other 
contact information unless you provide it in the body of your comment. 
Go directly to EDOCKET at http://www.epa.gov/edocket, or, from the EPA 
Internet Home Page (www.epa.gov), select ``Information Sources'' (in 
the left column), then ``Dockets,'' then ``EPA Dockets'' (in the first 
paragraph). For either route, then click on ``Quick Search'' (in the 
left column). In the search window, type in the docket identification 
number OAR-2005-0083. Please be patient, the search could take about 30 
seconds. This will bring you to the ``Docket Search Results'' page. At 
that point, click on OAR-2005-0083. From the resulting page, you may 
submit a comment by clicking on the balloon icon in the ``Submit 
Comment'' column and following the subsequent directions.
    iii. E-mail: Comments may be sent by electronic mail (e-mail) to [email protected], Attention Docket ID No. OAR-2005-0083. In 
contrast to EPA's electronic public docket, EPA's e-mail system is not 
an ``anonymous access'' system. If you send an e-mail comment directly 
to the Docket without going through EPA's electronic public docket, 
EPA's e-mail system automatically captures your e-mail address. E-mail 
addresses that are automatically captured by EPA's e-mail system are 
included as part of the comment that is placed in the official public 
docket, and made available in EPA's electronic public docket.
    2. Surface Mail. Send your comments to: EPA Docket Center (EPA/DC), 
Air and Radiation Docket, Environmental Protection Agency, EPA West, 
Mail Code 6102T, 1200 Pennsylvania Avenue, NW., Washington, DC 20460. 
Attention Docket ID No. OAR-2005-0083.
    3. Hand Delivery or Courier. Deliver your comments to: Air and 
Radiation Docket, EPA Docket Center, (EPA/DC) EPA West, Room B102, 1301 
Constitution Ave., NW., Washington, DC, Attention Docket ID No. OAR-
2005-0083. Such deliveries are only

[[Page 49015]]

accepted during the Docket Center's normal hours of operation.
    4. Facsimile. Fax your comments to: 202-566-1741, Attention Docket 
ID. No. OAR-2005-0083.
    Instructions for submitting information to EDOCKET: Direct your 
comments and information to Docket ID No. OAR-2005-0083. It is 
important to note that EPA's policy is that public comments, whether 
submitted electronically or in paper, will be made available for public 
viewing in EPA's electronic public docket as EPA receives them and 
without change, unless the comment contains copyrighted material, CBI, 
or other information whose disclosure is restricted by statute. When 
EPA identifies a comment containing copyrighted material, EPA will 
provide a reference to that material in the version of the comment that 
is placed in EPA's electronic public docket. The entire printed 
comment, including the copyrighted material, will be available in the 
public docket.
    Certain types of information will not be placed in EDOCKET. 
Information claimed as CBI and other information whose disclosure is 
restricted by statute, which is not included in the official public 
docket, will not be available for public viewing in EPA's electronic 
public docket. EPA's policy is that copyrighted material will not be 
placed in EPA's electronic public docket but will be available only in 
printed, paper form in the official public docket. To the extent 
feasible, publicly available docket materials will be made available in 
EPA's electronic public docket. When a document is selected from the 
index list in EPA Dockets, the system will identify whether the 
document is available for viewing in EPA's electronic public docket. 
Although not all docket materials may be available electronically, you 
may still access any of the publicly available docket materials through 
the docket facility. EPA intends to work towards providing electronic 
access to all of the publicly available docket materials through EPA's 
electronic public docket.
    The EPA EDOCKET and the federal regulations.gov websites are 
``anonymous access'' systems, which means EPA will not know your 
identity or contact information unless you provide it in the body of 
your comment. If you send an e-mail comment directly to EPA without 
going through EDOCKET or regulations.gov, your e-mail address will be 
automatically captured and included as part of the comment that is 
placed in the public docket and made available on the Internet. If you 
submit an electronic comment, EPA recommends that you include your name 
and other contact information in the body of your comment and with any 
disk or CD-ROM you submit. If EPA cannot read your comment due to 
technical difficulties and cannot contact you for clarification, EPA 
may not be able to consider your comment. Electronic files should avoid 
the use of special characters, any form of encryption, and be free of 
any defects or viruses.
    Public comments submitted on computer disks that are mailed or 
delivered to the docket will be transferred to EPA's electronic public 
docket. Public comments that are mailed or delivered to the docket will 
be scanned and placed in EPA's electronic public docket. Where 
practical, physical objects will be photographed, and the photograph 
will be placed in EPA's electronic public docket along with a brief 
description written by the docket staff.
    For additional information about EPA's electronic public docket 
visit EPA Dockets online or see 67 FR 38102, May 31, 2002.
    Docket: The official docket is the collection of materials that is 
available for public viewing at the Air and Radiation Docket in the EPA 
Docket Center (EPA/DC), EPA West, Room B102, 1301 Constitution Ave., 
NW., Washington, DC. The EPA Docket Center Public Reading Room is open 
from 8:30 a.m. to 4:30 p.m., Monday through Friday, excluding legal 
holidays. The telephone number for the Public Reading Room is 202-566-
1744. The telephone number for the Air and Radiation Docket is 202-566-
1742. As provided in EPA's regulations at 40 CFR part 2, and in 
accordance with normal EPA docket procedures, if copies of any docket 
materials are requested, a reasonable fee may be charged.
    All documents in the docket are listed in the EDOCKET index at 
http://www.epa.gov/edocket. Although listed in the index, some 
information is not publicly available since it will not be placed in 
EDOCKET. That is, although a part of the official docket, EDOCKET does 
not include Confidential Business Information (CBI) or other 
information whose disclosure is restricted by statute. Information 
claimed as CBI and other information whose disclosure is restricted by 
statute, which is not included in the official public docket, will not 
be available for public viewing in EPA's EDOCKET. In addition, EPA 
policy is that copyrighted material will not be placed in EPA's 
EDOCKET, but will be available only in printed, paper form in the 
official public docket. To the extent feasible, publicly available 
docket materials will be made available in EPA's EDOCKET. When a 
document is selected from the index list in EDOCKET, the system will 
identify whether the document is available for viewing. Although not 
all docket materials may be available electronically, you may still 
access any of the publicly available docket materials through the 
docket facility. EPA intends to work towards providing electronic 
access to all of the publicly available docket materials through EPA's 
electronic public docket.

FOR FURTHER INFORMATION CONTACT: Ray Clark, Office of Radiation and 
Indoor Air, Radiation Protection Division (6608J), U.S. Environmental 
Protection Agency, 1200 Pennsylvania Ave., NW., Washington, DC 20460-
0001; telephone number: 202-343-9601; fax number: 202-343-2305; e-mail 
address: [email protected].

SUPPLEMENTARY INFORMATION:

I. General Information

A. Does This Action Apply to Me?

    The DOE is the only entity regulated by these standards. Our 
standards affect NRC only because, under Section 801(b) of the EnPA, 42 
U.S.C. 10141 n., NRC must modify its licensing requirements, as 
necessary, to make them consistent with our final standards. Before it 
may accept waste at the Yucca Mountain site, DOE must obtain a license 
from NRC. DOE will be subject to NRC's modified regulations, which NRC 
will implement through its licensing proceedings.

B. What Should I Consider as I Prepare My Comments for EPA?

    1. Submitting CBI. If you submit CBI, clearly mark the part or all 
of the information that you claim to be CBI. For CBI information on a 
disk or CD-ROM that you mail to EPA, mark the outside of the disk or 
CD-ROM as CBI and then identify electronically within the disk or CD-
ROM the specific information that is claimed as CBI. In addition to one 
complete version of the comment that includes information claimed as 
CBI, a copy of the comment that does not contain the information 
claimed as CBI must be submitted for inclusion in the public docket. 
Information so marked will not be disclosed except in accordance with 
procedures set forth in 40 CFR part 2.
    2. Tips for Preparing Your Comments. You may find the following 
suggestions helpful for preparing your comments:
    1. Explain your views as clearly as possible.
    2. Describe any assumptions that you used.

[[Page 49016]]

    3. Provide any technical information and/or data you used that 
support your views.
    4. If you estimate potential burden or costs, explain how you 
arrived at your estimate.
    5. Provide specific examples to illustrate your concerns.
    6. Offer alternatives.
    7. Make sure to submit your comments by the comment period deadline 
identified.
    8. Respond to specific questions from the Agency.
    9. To ensure proper receipt by EPA, identify the appropriate docket 
identification number in the subject line on the first page of your 
response.

C. How Can I View Items in the Docket?

    1. Information Files. EPA is working with the Lied Library at the 
University of Nevada-Las Vegas (http://www.library.unlv.edu/about/hours.html#desks) and the Amargosa Valley, Nevada public library 
(http://www.amargosavalley.com/Library.html) to provide information 
files on this rulemaking. These files are not legal dockets, however 
every effort will be made to put the same material in them as in the 
official public docket in Washington, DC. The Lied Library information 
file is at the Research and Information Desk, Government Publications 
Section (702-895-2200). Hours vary based upon the academic calendar, so 
we suggest that you call ahead to be certain that the library will be 
open at the time you wish to visit (for a recorded message, call 702-
895-2255). The other information file is in the Public Library in 
Amargosa Valley, Nevada (phone 775-372-5340). As of the date of 
publication, the hours are Monday, Wednesday, and Friday (9 a.m.-5 
p.m.); Tuesday and Thursday (9 a.m.-7 p.m.); and Saturday (9 a.m.-1 
p.m.). The library is closed on Sunday. These hours can change, so we 
suggest that you call ahead to be certain when the library will be 
open.
    2. Electronic Access. An electronic version of the public docket is 
available through EPA's electronic public docket and comment system, 
EPA Dockets (EDOCKET). You may use EDOCKET to submit or view comments, 
access the index listing of the contents of the official public docket, 
and to access those documents in the public docket that are available 
electronically. To access the docket either go directly to http://www.epa.gov/edocket/ or, from the EPA Internet Home Page (www.epa.gov), 
select ``Information Sources'' (in the left column), then ``Dockets,'' 
then ``EPA Dockets'' (in the first paragraph). For either route, then 
click on ``Quick Search'' (in the left column). In the search window, 
type in the docket identification number OAR-2005-0083. Please be 
patient, the search could take about 30 seconds. This will bring you to 
the ``Docket Search Results'' page. At that point, click on OAR-2005-
0083. From the resulting page, you may access the docket contents 
(e.g., OAR-2005-0083-0002) by clicking on the icon in the ``Rendition'' 
column.

D. Can I Access Information by Telephone or Via the Internet?

    Yes. You may call our toll-free information line (800-331-9477) 24 
hours per day. By calling this number, you may listen to a brief update 
describing our rulemaking activities for Yucca Mountain, leave a 
message requesting that we add your name and address to the Yucca 
Mountain mailing list, or request that an EPA staff person return your 
call. In addition, we have established an electronic listserv through 
which you can receive electronic updates of activities related to this 
rulemaking. To subscribe to the listserv, go to https://lists.epa.gov/read/all_forums. In the alphabetical list, locate ``yucca-updates'' 
and select ``subscribe'' at the far right of the screen. You will be 
asked to provide your e-mail address and choose a password. You also 
can find information and documents relevant to this rulemaking on the 
World Wide Web at http://www.epa.gov/radiation/yucca. We also recommend 
that you examine the preamble and regulatory language for the earlier 
proposed and final rules, which appeared in the Federal Register on 
August 27, 1999 (64 FR 46976) and June 13, 2001 (66 FR 32074), 
respectively.

E. What Documents Are Referenced in Today's Proposal?

    We refer to a number of documents that provide supporting 
information for our Yucca Mountain standards. All documents relied upon 
by EPA in regulatory decisionmaking may be found in our docket (OAR-
2005-0083). Other documents, e.g., statutes, regulations, proposed 
rules, are readily available from public sources. The documents below 
are referenced most frequently in today's proposal.

Item No. (OAR-2005-0083-xxxx)
    0044 ``Safety Indicators in Different Time Frames for the Safety 
Assessment of Underground Radioactive Waste Repositories,'' 
International Atomic Energy Agency
TECDOC-767, 1994
    0045 ``Regulatory Decision Making in the Presence of Uncertainty in 
the Context of Disposal of Long Lived Radioactive Wastes,'' 
International Atomic Energy Agency
TECDOC-975, 1997
    0046 ``The Handling of Timescales in Assessing Post-Closure Safety: 
Lessons Learnt from the April 2002 Workshop in Paris, France,'' Nuclear 
Energy Agency (Organisation for Economic Co-operation and Development), 
2004
    0051 ``Geological Disposal of Radioactive Waste,'' International 
Atomic Energy Agency Draft Safety Requirements (DS154), April 2005
    0061 ``Principles and Standards for Disposal of Long-Lived 
Radioactive Wastes,'' Neil Chapman and Charles McCombie, Elsevier 
Press, 2003
    0062 ``An International Peer Review of the Yucca Mountain Project 
TSPA-SR,'' Joint Report by the OECD Nuclear Energy Agency and the 
International Atomic Energy Agency, OECD, 2002
    0076 Technical Bases for Yucca Mountain Standards (the NAS Report), 
National Research Council, National Academy Press, 1995
    0077 ``Assessment of Variations in Radiation Exposure in the United 
States,'' EPA Technical Support Document, July 2005
    0085 ``Assumptions, Conservatisms, and Uncertainties in Yucca 
Mountain Performance Assessments,'' EPA Technical Support Document, 
July 2005
    0086 DOE Final Environmental Impact Statement, DOE/EIS-0250, 
February 2002

Acronyms and Abbreviations

    We use many acronyms and abbreviations in this document. These 
include:

BID--background information document
CED--committed effective dose
CEDE--committed effective dose equivalent
DOE--U.S. Department of Energy
DOE/VA--DOE's Viability Assessment
EIS--Environmental Impact Statement
EnPA--Energy Policy Act of 1992
EPA--U.S. Environmental Protection Agency
FEIS--Final Environmental Impact Statement
FEPs--features, events, and processes
FR--Federal Register
GCD--greater confinement disposal
HLW--high-level radioactive waste
HSK--Swiss Federal Nuclear Safety Inspectorate
IAEA--International Atomic Energy Agency

[[Page 49017]]

ICRP--International Commission on Radiological Protection
KASAM--Swedish National Council for Nuclear Waste
LLW--low-level radioactive waste
MCL--maximum contaminant level
MTHM--metric tons of heavy metal
NAPA--National Academy of Public Administration
NAS--National Academy of Sciences
NEA--Nuclear Energy Agency
NEI--Nuclear Energy Institute
NRC--U.S. Nuclear Regulatory Commission
NRDC--Natural Resources Defense Council
NTS--Nevada Test Site
NTTAA--National Technology Transfer and Advancement Act
NWPA--Nuclear Waste Policy Act of 1982
NWPAA--Nuclear Waste Policy Amendments Act of 1987
OECD--Organization for Economic Cooperation and Development
OMB--Office of Management and Budget
RMEI--reasonably maximally exposed individual
SSI--Swedish Radiation Protection Authority
SNF--spent nuclear fuel
SR--Site recommendation
TRU--transuranic
TSPA--Total System Performance Assessment
UK--United Kingdom
UMRA--Unfunded Mandates Reform Act of 1995
U.S.C.--United States Code
WIPP LWA--Waste Isolation Pilot Plant Land Withdrawal Act of 1992

Outline of Today's Action

I. What is the History of Today's Action?
    A. Promulgation of 40 CFR part 197 in 2001
    1. What are the Elements of EPA's 2001 Standards?
    a. What is the Standard for Storage of the Waste? (Subpart A, 
Sec. Sec.  197.1 through 197.5)
    b. What Are the Standards for Disposal? (Subpart B, Sec. Sec.  
197.11 through 197.36)
    i. What is the Standard for Protection of Individuals? 
(Sec. Sec.  197.20 through 197.21)
    aa. Who Represents the Exposed Population?
    bb. How Far Into the Future Must Performance be Assessed?
    ii. What is the Standard for Human Intrusion? (Sec. Sec.  197.25 
through 197.26)
    iii. What are the Standards to Protect Ground Water? (Sec. Sec.  
197.30 through 197.31)
    c. What is ``Reasonable Expectation''? (Sec.  197.14)
    B. Legal Challenges to 40 CFR part 197
    1. Challenges by the State of Nevada and Natural Resources 
Defense Council
    2. Challenge by the Nuclear Energy Institute
    C. Ruling by the U.S. Court of Appeals for the District of 
Columbia Circuit
    1. What Did the Court of Appeals Rule on the Issue of Compliance 
Period?
    a. What Were NAS's Findings (``Conclusions'') and 
Recommendations on the Issue of Compliance Period?
    2. What Did the Court of Appeals Rule on Other Issues Related to 
EPA's Standards?
II. How Will EPA Address the Decision by the Court of Appeals?
    A. How Will Elements of the Disposal Standards be Affected?
    1. Individual-Protection Standard
    2. Human-Intrusion Standard
    3. Ground-Water Protection Standards
    4. Reasonable Expectation
    5. Effects of Uncertainty
    B. How Does the Application of ``Reasonable Expectation'' 
Influence Today's Proposal?
    C. How Is EPA Proposing to Revise the Individual-Protection 
Standard (Sec.  197.20) to Address Peak Dose?
    1. Multiple Dose Standards Applicable to Different Compliance 
Periods
    2. What Other Options Did EPA Consider?
    a. Maintain the 10,000-year Standard Alone Without Addressing 
Peak Dose
    b. Dose Standard To Apply at Peak Dose Alone
    c. Peak Dose Standard Varying Over Time
    d. Standard Expressed as a Dose Target, Rather Than Limit
    e. Standard Expressed as a Statistical Distribution
    3. What Dose Level is EPA Proposing for Peak Dose?
    4. What Other Peak Dose Levels Did EPA Consider?
    a. Maintain the 15 mrem/yr Standard at Peak Dose
    b. 100 mrem/yr Standard at Peak Dose
    c. Peak Dose Standard Based on Regional Background Radiation 
Levels
    5. How Will NRC Judge Compliance?
    6. How Will DOE Calculate the Dose?
    D. How Will Today's Proposal Affect the Way DOE Conducts 
Performance Assessments?
    1. Performance Assessments Up To 10,000 Years After Disposal
    2. Performance Assessments for Periods Longer Than 10,000 Years 
After Disposal
    a. Consideration of Likely, Unlikely, and Very Unlikely FEPs
    b. Consideration of Seismic FEPs
    c. Consideration of Igneous (Volcanic) FEPs
    d. Consideration of Climatological FEPs
    E. How Is EPA Proposing To Revise the Human-Intrusion Standard 
(Sec.  197.25) To Address Peak Dose?
    F. Summary of Today's Proposal by Section
III. Statutory and Executive Order Reviews
    A. Executive Order 12866: Regulatory Planning and Review
    B. Paperwork Reduction Act
    C. Regulatory Flexibility Act
    D. Unfunded Mandates Reform Act
    E. Executive Order 13132: Federalism
    F. Executive Order 13175: Consultation and Coordination with 
Indian Tribal Governments
    G. Executive Order 13045: Protection of Children from 
Environmental Health & Safety Risks
    H. Executive Order 13211: Actions that Significantly Affect 
Energy Supply, Distribution, or Use
    I. National Technology Transfer and Advancement Act

I. What Is the History of Today's Action?

    Radioactive wastes result from the use of nuclear fuel and other 
radioactive materials. Today, we are proposing to revise certain 
standards pertaining to spent nuclear fuel (SNF), high-level 
radioactive waste (HLW), and other radioactive waste (we refer to these 
items collectively as ``radioactive materials'' or ``waste'') that may 
be stored or disposed of in the Yucca Mountain repository. (When we 
discuss storage or disposal in this document in reference to Yucca 
Mountain, we note that no decision has been made regarding the 
acceptability of Yucca Mountain for storage or disposal as of the date 
of this publication. To save space and to avoid excessive repetition, 
we will not describe Yucca Mountain as a ``potential'' repository; 
however, we intend this meaning to apply.) Pursuant to Section 801(a) 
of the Energy Policy Act of 1992 (EnPA, Pub. L. 102-486), these 
standards apply only to facilities at Yucca Mountain.
    Once nuclear reactions have consumed a certain percentage of the 
uranium or other fissionable material in nuclear reactor fuel, the fuel 
no longer is useful for its intended purpose. It then is known as 
``spent'' nuclear fuel (SNF). It is possible to recover specific 
radionuclides from SNF through ``reprocessing,'' which is a process 
that dissolves the SNF, thus separating the radionuclides from one 
another. Radionuclides not recovered through reprocessing become part 
of the acidic liquid wastes that the Department of Energy (DOE) plans 
to convert into various types of solid materials. High-level waste 
(HLW) is the highly radioactive liquid or solid wastes that result from 
reprocessing SNF. The SNF that does not undergo reprocessing prior to 
disposal remains inside the fuel assembly and becomes the final waste 
form.
    In the U.S., SNF and HLW have been produced since the 1940s, mainly 
as a result of commercial power production and defense activities. 
Since the inception of the nuclear age, the proper disposal of these 
wastes has been the responsibility of the Federal government. The 
Nuclear Waste Policy Act of 1982 (NWPA, 42 U.S.C. Chapter 108) 
formalizes the current Federal

[[Page 49018]]

program for the disposal of SNF and HLW by:
    (1) Making DOE responsible for siting, building, and operating an 
underground geologic repository for the disposal of SNF and HLW;
    (2) Directing us to set generally applicable environmental 
radiation protection standards based on authority established under 
other laws \1\; and
---------------------------------------------------------------------------

    \1\ These laws include the Atomic Energy Act of 1954, as amended 
(42 U.S.C. 2011-2296) and Reorganization Plan No. 3 of 1970 (5 
U.S.C. Appendix 1).
---------------------------------------------------------------------------

    (3) Requiring the Nuclear Regulatory Commission (NRC) to implement 
our standards by revising its licensing requirements for SNF and HLW 
repositories to be consistent with our standards.
    This general division of responsibilities continues for the Yucca 
Mountain repository. Thus, today we are proposing to establish or 
revise specific aspects of our public health protection standards at 40 
CFR part 197 (which are, pursuant to EnPA Section 801(a), applicable 
only to Yucca Mountain, rather than generally applicable). The NRC will 
issue implementing regulations for these standards. The DOE plans to 
submit a license application to NRC. The NRC then will determine 
whether DOE has met NRC's regulations and whether to grant or deny a 
license for Yucca Mountain.
    In 1985, we established generic standards for the management, 
storage, and disposal of SNF, HLW, and transuranic (TRU) radioactive 
waste (see 40 CFR part 191, 50 FR 38066, September 19, 1985), which 
were intended to apply to any facilities utilized for the storage or 
disposal of these wastes, including Yucca Mountain. In 1987, the U.S. 
Court of Appeals for the First Circuit remanded the disposal standards 
in 40 CFR part 191 (NRDC v. EPA, 824 F.2d 1258 (1st Cir. 1987)). As 
discussed below, we later amended and reissued these standards to 
address issues that the court raised. Also in 1987, the Nuclear Waste 
Policy Amendments Act (NWPAA, Pub. L. 100-203) amended the NWPA by, 
among other actions, selecting Yucca Mountain, Nevada, as the only 
potential site that DOE should characterize for a long-term geologic 
repository. In October 1992, the Waste Isolation Pilot Plant Land 
Withdrawal Act (WIPP LWA, Pub. L. 102-579) and the EnPA became law. 
These statutes changed our obligations concerning radiation standards 
for the Yucca Mountain candidate repository. The WIPP LWA:
    (1) Reinstated the 40 CFR part 191 disposal standards, except those 
portions that were the specific subject of the remand by the First 
Circuit;
    (2) Required us to issue standards to replace the portion of the 
challenged standards remanded by the court; and
    (3) Exempted the Yucca Mountain site from the 40 CFR part 191 
disposal standards.

We issued the amended 40 CFR part 191 disposal standards, which 
addressed the judicial remand, on December 20, 1993 (58 FR 66398). The 
EnPA, enacted in 1992, set forth our responsibilities as they relate to 
Yucca Mountain. In the EnPA, Congress directed us to set public health 
and safety radiation standards for Yucca Mountain. Specifically, 
section 801(a)(1) of the EnPA directed us to ``promulgate, by rule, 
public health and safety standards for the protection of the public 
from releases from radioactive materials stored or disposed of in the 
repository at the Yucca Mountain site.'' Section 801(a)(2) directed us 
to contract with the National Academy of Sciences (NAS) to conduct a 
study to provide us with its findings and recommendations on reasonable 
standards for protection of public health and safety from releases from 
the Yucca Mountain disposal system. Moreover, it provided that our 
standards shall be the only such standards applicable to the Yucca 
Mountain site and are to be based upon and consistent with NAS's 
findings and recommendations. On August 1, 1995, NAS released its 
report, ``Technical Bases for Yucca Mountain Standards'' (the NAS 
Report) (Docket No. OAR-2005-0083-0076).

A. Promulgation of 40 CFR Part 197 in 2001

    Following the direction in the EnPA, we developed standards 
specifically applicable to releases from radioactive material stored or 
disposed of in the Yucca Mountain repository. In doing so, we gave 
special weight to both the NAS Report and our generic standards in 40 
CFR part 191, and also considered other relevant information, 
precedents, and analyses.
    We evaluated 40 CFR part 191 because those standards were developed 
to apply to any site selected for storage and disposal of SNF and HLW, 
and would have applied to Yucca Mountain had Congress not directed 
otherwise. Thus, we believed that 40 CFR part 191 already included the 
major components of standards needed for any specific site, such as 
Yucca Mountain. However, we recognized that all the components would 
not necessarily be directly transferable to the situation at Yucca 
Mountain, and that some modification might be necessary. We also 
considered that some components of the generic standards would not be 
carried into site-specific standards, simply because not all of the 
conditions found among all sites are present at each site. See 66 FR 
32076-32078, June 13, 2001 (Docket No. OAR-2005-0083-0042), for a more 
detailed discussion of the role of 40 CFR part 191 in developing 40 CFR 
part 197.
    We also considered the findings and recommendations of the NAS in 
developing standards for Yucca Mountain. In some cases, provisions of 
40 CFR part 191 were already consistent with NAS's analysis (e.g., 
level of protection for the individual). In other cases, we used the 
NAS Report to modify or draw out parts of 40 CFR part 191 to apply more 
directly to Yucca Mountain (e.g., the stylized drilling scenario for 
human intrusion). See the NAS Report for a complete description of 
findings and recommendations.
    Because our standards are intended to apply specifically to the 
Yucca Mountain disposal system, in a number of areas we tailored our 
approach to consider the characteristics of the site and the local 
populations. Yucca Mountain is in southwestern Nevada approximately 100 
miles northwest of Las Vegas. The eastern part of the site is on the 
Nevada Test Site (NTS). The northwestern part of the site is on the 
Nellis Air Force Range. The southwestern part of the site is on Bureau 
of Land Management land. The area has a desert climate with topography 
typical of the Basin and Range province. Yucca Mountain is made of 
layers of ashfalls from volcanic eruptions that happened more than 10 
million years ago. There are two major aquifers beneath Yucca Mountain. 
Regional ground water in the vicinity of Yucca Mountain is believed to 
flow generally in a south-southeasterly direction. The DOE plans to 
build the repository about 300 meters below the surface and about 300 
to 500 meters above the water table. For more detailed descriptions of 
Yucca Mountain's geologic and hydrologic characteristics, and the 
disposal system, please see chapter 7 of the 2001 BID (Docket No. OAR-
2005-0083-0050) and the preamble to the proposed rule (64 FR 46979-
46980, August 27, 1999, Docket No. OAR-2005-0083-0041).
    We proposed standards for Yucca Mountain on August 27, 1999 (64 FR 
46976). In response to our proposal, we received more than 800 public 
comments and conducted four public hearings. After evaluating public 
comments, we issued final standards (66 FR 32074, June 13, 2001). See 
the Response to Comments document from that rulemaking for more 
discussion of

[[Page 49019]]

comments (Docket No. OAR-2005-0083-0043).
1. What Are the Elements of EPA's 2001 Standards?
    We are issuing today's proposal to respond to a ruling by the U.S. 
Court of Appeals for the District of Columbia Circuit (``the Court'') 
that vacated portions of 40 CFR part 197. Sections I.B (``Legal 
Challenges to 40 CFR part 197'') and I.C (``Ruling by U.S. Court of 
Appeals for the District of Columbia Circuit'') discuss aspects of the 
legal challenges on which the Court ruled. This section summarizes some 
of the key provisions and concepts in 40 CFR part 197 to provide a 
context to better understand the basis for the legal actions and 
today's proposed action, which is described in Section II of this 
document (``How Will EPA Address the Decision by the Court of 
Appeals?'').
    The standards issued in 2001 as 40 CFR part 197 included the 
following:
     A standard to protect the public during storage operations 
at the Yucca Mountain site;
     An individual-protection standard to protect the public 
after disposal from releases from the undisturbed repository;
     A human-intrusion standard to protect the public after 
disposal from releases caused by a drilling penetration into the 
repository;
     A set of standards to protect ground water from 
radionuclide contamination caused by releases from the repository after 
disposal;
     The requirement that compliance with the disposal 
standards be shown for 10,000 years;
     The requirement that DOE continue its projections for the 
individual-protection and human-intrusion standards beyond 10,000 years 
to the time of peak (maximum) dose, and place those projections in the 
Environmental Impact Statement (EIS);
     The concept of the Reasonably Maximally Exposed Individual 
(RMEI), defined as a hypothetical person whose lifestyle is 
representative of the local population, as the individual against whom 
the disposal standards should be assessed; and
     The concept of a ``controlled area,'' defined as an area 
immediately surrounding the repository whose geology is considered part 
of the natural barrier component of the overall disposal system, and 
inside of which radioactive releases are not regulated.
    We emphasize that today's proposal is narrowly focused to respond 
to the Court ruling. Most sections of our 2001 rule are unaffected by 
the Court's ruling and are not implicated in today's proposal. We are 
requesting and will respond to comments only on those provisions we are 
proposing to change today.

a. What Is the Standard for Storage of the Waste? (Subpart A, 
Sec. Sec.  197.1 Through 197.5)

    Section 801(a)(1) of the EnPA calls for EPA's public health and 
safety standards to apply to radioactive materials ``stored or disposed 
of in the repository at the Yucca Mountain site.'' The repository is 
the excavated portion of the facility constructed underground within 
the Yucca Mountain site. The storage standard, therefore, applies to 
waste inside the repository, prior to disposal.
    The DOE also will handle, and might store, radioactive material 
outside the repository prior to subsurface emplacement. Therefore, our 
standards will provide public health and safety protection for surface 
management and storage activities on the surface of the Yucca Mountain 
site and in the Yucca Mountain repository. The combined doses incurred 
by any individual in the general environment from these activities must 
not exceed 150 [mu]Sv (15 mrem) committed effective dose equivalent per 
year (CEDE/yr).

b. What Are the Standards for Disposal? (Subpart B, Sec. Sec.  197.11 
Through 197.36)

    Subpart B of our 2001 rule consisted of three separate standards 
(or sets of standards) that apply after disposal, which are discussed 
in more detail in the appropriate sections of this document (e.g., 
Section II.A, ``How Will Elements of the Disposal Standards be 
Affected?''). For additional detail, see the preamble to the June 2001 
rulemaking (66 FR 32074, June 13, 2001). The disposal standards are:
     An individual-protection standard;
     A human-intrusion standard; and
     Ground-water protection standards.
i. What Is the Standard for Protection of Individuals? (Sec. Sec.  
197.20 Through 197.21)
    The first standard is an individual-protection standard. It 
specifies the maximum dose that a reasonably maximally exposed 
individual (RMEI) may receive from releases from the Yucca Mountain 
repository.
    Our individual-protection standard set a limit of 150 [mu]Sv (15 
mrem) CEDE/yr. This limit corresponds to an annual risk of fatal cancer 
within the range that NAS suggested as a ``reasonable starting point 
for EPA's rulemaking'' (NAS Report p. 5, Docket No. OAR-2005-0083-
0076). The NAS s suggested risk range corresponds to approximately 2 to 
20 mrem CEDE/yr.
    The standard described above applies for a period of 10,000 years 
after disposal, and is to be measured against exposures to the RMEI at 
a location outside the controlled area (in the ``accessible 
environment'').

aa. Who Represents the Exposed Population?

    To determine whether the Yucca Mountain disposal system complies 
with our standard, DOE must calculate the dose received by some 
individual or group of individuals exposed to releases from the 
repository and compare the calculated dose with the limit established 
in the standard. The standard specifies, therefore, the representative 
individual for whom DOE must make the dose calculation as the RMEI. It 
was left to NRC to define the details, beyond those which we specified, 
necessary for the dose calculation. NRC has further defined the RMEI as 
an adult (10 CFR 63.312(e)) and specified that the average 
concentration of radionuclides in well water ingested by the RMEI be 
based on a water demand of 3,000 acre-feet per year (10 CFR 63.312(c)).

The Reasonably Maximally Exposed Individual (RMEI)

    The approach we chose (the RMEI) embodies the intent of the 
internationally-accepted concept to protect those individuals most at 
risk from the proposed repository but specifies one or a few site-
specific parameters at their maximum values. The characteristics of the 
RMEI are defined from consideration of current population distribution 
and ground-water usage, and average food consumption patterns for the 
population downgradient from Yucca Mountain in Amargosa Valley, Nevada.
    Our RMEI is a theoretical individual representative of a future 
population group or community termed ``rural-residential'' (see Chapter 
8 of the 2001 BID for a description of this concept, Docket No. OAR-
2005-0083-0050). We assume that the rural-residential RMEI is exposed 
through the same general pathways as a subsistence farmer. However, 
this RMEI would not be a full-time farmer. Rather, the RMEI might do 
personal gardening and earn income from other sources of work in the 
area. Under our standard, the RMEI will have food and water intake 
rates, diet, and physiology similar to those of individuals living in 
Amargosa Valley, Nevada. We assume that all of the drinking water and 
some of the food (based upon surveys) consumed by the RMEI is from the 
local area. Similarly, we assume that local food production

[[Page 49020]]

will use water contaminated with radionuclides released from the 
disposal system. We believe this lifestyle is conservative but similar 
to that of most people living in Amargosa Valley today.
    Location of the RMEI. The location of the RMEI is a basic part of 
the exposure scenario. We require that the RMEI be located in the 
accessible environment (i.e., outside the controlled area) above the 
highest concentration of radionuclides in the plume of contamination. 
Based upon a review of available site-specific information (see Chapter 
8 of the 2001 BID, Docket No. OAR-2005-0083-0050), we identified the 
southern edge of the Nevada Test Site as the southernmost extent of the 
controlled area. The actual compliance point will be determined through 
the licensing process. (Even if the RMEI were to be located north of 
this line of latitude, the RMEI must still have the characteristics 
described in Sec.  197.21.) As discussed in Section I.B (``Legal 
Challenges to 40 CFR part 197'') and I.C (``Ruling by the U.S. Court of 
Appeals for the District of Columbia Circuit''), the location of the 
RMEI was a subject of the Court decision, was upheld, and is not a 
subject of today's proposal.

bb. How Far Into the Future Must Performance Be Assessed?

    In 2001, we established a compliance period of 10,000 years. Under 
the 2001 standards, the peak dose within 10,000 years after disposal 
would be required to comply with the individual-protection standard. In 
addition, we required calculation of the peak dose beyond 10,000 years, 
but within the period of geologic stability. We required DOE to include 
the results and bases of the additional analyses in the EIS for Yucca 
Mountain as an indicator of the future performance of the disposal 
system. The rule did not, however, require that DOE meet a specific 
dose limit after 10,000 years. The compliance period was a subject of 
the Court decision and is the primary subject of today's proposal.
ii. What Is the Standard for Human Intrusion? (Sec. Sec.  197.25 
Through 197.26)
    We adopted NAS's suggested starting point for a human-intrusion 
scenario. As NAS recommended, our standard required a single-borehole 
intrusion scenario based upon Yucca Mountain-specific conditions. The 
intended purpose of analyzing this scenario ``* * * is to examine the 
site- and design-related aspects of repository performance under an 
assumed intrusion scenario to inform a qualitative judgment'' (NAS 
Report p. 111). The assessment would result in a calculated RMEI dose 
arriving through the pathway created by the assumed borehole (with no 
other releases included). Consistent with the NAS Report, we also 
required ``that the conditional risk as a result of the assumed 
intrusion scenario should be no greater than the risk levels that would 
be acceptable for the undisturbed-repository case'' (NAS Report p. 
113). We interpreted NAS's term ``undisturbed'' to mean that the Yucca 
Mountain disposal system is not disturbed by human intrusion but that 
other processes or events that are likely to occur could disturb the 
system.
    The DOE is not required to use probabilistic performance assessment 
for the human-intrusion analysis, as it is for the individual-
protection standard. However, if it chooses to do so, we required that 
the human-intrusion analysis of disposal system performance use the 
same methods and RMEI characteristics for the performance assessment as 
those required for the individual-protection standard, with the 
exception that the human-intrusion analysis would exclude unlikely 
natural features, events, and processes (FEPs).
    The DOE must determine when the intrusion would occur based upon 
the earliest time that current technology and practices could lead to 
waste package penetration without the drillers noticing the canister 
penetration. In general, we believe that the time frame for the 
drilling intrusion should be within the period that a small percentage 
of the waste packages have failed but before significant migration of 
radionuclides from the engineered barrier system has occurred because, 
based upon our understanding of drilling practices, this period would 
be about the earliest time that a driller would not recognize an impact 
with a waste package.
    The compliance standard for human intrusion parallels that for the 
individual-protection scenario. If the intrusion were to occur at or 
earlier than 10,000 years after disposal, DOE must demonstrate a 
reasonable expectation that annual exposures incurred by the RMEI 
within 10,000 years as a result of the intrusion event would not exceed 
150 [mu]Sv (15 mrem) CEDE. However, if the intrusion occurred after 
10,000 years, or when earlier intrusions result in exposures projected 
to occur after 10,000 years, DOE would not have to compare its results 
against a numerical standard, but would have to include those results 
in its EIS.
iii. What Are the Standards To Protect Ground Water? (Sec. Sec.  197.30 
Through 197.31)
    We established separate ground-water standards as a means to 
protect the aquifer as both a resource for current users and a 
potential resource for larger numbers of future users either near the 
repository or farther away in communities comprised of a substantially 
larger number of people than presently exist in the vicinity of Yucca 
Mountain. The standards DOE must meet are equivalent to the 
radionuclide Maximum Contaminant Levels (MCLs) established for drinking 
water.
    To implement the ground-water protection standards in Sec.  197.30, 
we required that DOE use the concept of a ``representative volume'' of 
ground water (Sec.  197.31). Under this approach, DOE must project the 
concentration of radionuclides or the resultant doses within a 
``representative volume'' of ground water for comparison against the 
standards. We selected a value of 3,000 acre-ft/yr as a ``cautious, but 
reasonable'' figure for the representative volume. Section 197.31 also 
describes two methods by which DOE may calculate radionuclide 
concentrations in ground water. See the preamble to the 2001 rulemaking 
for more discussion of the representative volume and approaches for 
calculating radionuclide concentrations for compliance purposes.
    As with the individual-protection standard, compliance with the 
ground-water protection standards must be determined at the point of 
highest concentration in the plume of contamination in the accessible 
environment. The controlled area was defined in the same way as for the 
individual-protection standard. The ground-water protection standards 
were a subject of the Court decision, were upheld, and are not a 
subject of today's proposal.

c. What Is ``Reasonable Expectation''? (Sec.  197.14)

    An important provision of our standards is the establishment of the 
principle of ``reasonable expectation'' to guide implementation of our 
standards and provide context for evaluating projections against the 
numerical compliance standards discussed above. It is a critical 
element in implementing our standards, but its importance might easily 
be overlooked or misunderstood. We use the concept of ``reasonable 
expectation'' in these standards to reflect our intent regarding the 
level of ``proof'' necessary for NRC to determine whether the projected 
performance of

[[Page 49021]]

the Yucca Mountain disposal system complies with the standards (see 
Sec. Sec.  197.20, 197.25, and 197.30). In issuing our 2001 standards, 
we noted that this term is meant to convey our position that 
unequivocal numerical proof of compliance is neither necessary nor 
likely to be obtained for geologic disposal systems. We believe 
unequivocal proof is not possible because of the extremely long time 
periods involved and because disposal system performance assessments 
require extrapolations of conditions and the actions of processes that 
govern disposal system performance over those long time periods.
    The primary means for demonstrating compliance with the standards 
is the use of computer modeling to project the performance of the 
disposal system under the range of expected conditions. These modeling 
calculations involve the extrapolation of site conditions and the 
interactions of important processes over long time periods, 
extrapolations that involve inherent uncertainties in the necessarily 
limited amount of information that can be collected through field and 
laboratory studies and the unavoidable uncertainties involved in 
simulating the complex and time-variable processes and events involved 
in long-term disposal system performance. Overly conservative 
assumptions made in developing performance scenarios can bias the 
analyses in the direction of unrealistically extreme situations, which 
in reality may be highly improbable, and can deflect attention from 
questions critical to developing an adequate understanding of the 
expected features, events, and processes (``Assumptions, Conservatisms, 
and Uncertainties in Yucca Mountain Performance Assessments,'' Sections 
11 and 12, July 2005, Docket No. OAR-2005-0083-0085). The reasonable 
expectation approach focuses attention on understanding the 
uncertainties in projecting disposal system performance so that 
regulatory decision making will be done with a full understanding of 
the uncertainties involved. Thus, realistic analyses are preferred over 
conservative and bounding assumptions, to the extent practical.

B. Legal Challenges to 40 CFR Part 197

    Various aspects of our standards were challenged in lawsuits filed 
with the U.S. Court of Appeals for the District of Columbia Circuit in 
July 2001. Oral arguments were conducted on January 14, 2004. These 
challenges and the outcome are described in the following sections.
1. Challenges by the State of Nevada and Natural Resources Defense 
Council
    The State of Nevada, the Natural Resources Defense Council (NRDC), 
and several other environmental and public interest groups challenged 
several aspects of our final standards on the grounds that they were 
insufficiently protective and had not been adequately justified. 
Specifically, they claimed that:
     EPA's promulgation of standards that apply for 10,000 
years after disposal violates the EnPA because such standards are not 
``based upon and consistent with'' the findings and recommendations of 
the NAS. NAS recommended standards that would apply to the time of 
maximum risk and stated that there is ``no scientific basis for 
limiting the time period of the individual-risk standard to 10,000 
years or any other value.''
     The size of the controlled area defined by EPA, which 
represents the maximum extent of the disposal system and inside which 
DOE need not demonstrate compliance with the EPA standards, rests on 
inappropriate assumptions regarding the ability of people to live 
closer to the repository and violates the Safe Drinking Water Act 
provisions against endangering sources of drinking water.
     EPA's definition of ``disposal'' in 40 CFR 197.12 deviates 
from the definition in the NWPA by inserting the qualifying phrase 
``for as long as reasonably possible,'' suggesting that the Yucca 
Mountain disposal system would be held to a lesser standard of 
protection because it would not have to provide ``permanent 
isolation.''
2. Challenge by the Nuclear Energy Institute
    The Nuclear Energy Institute (NEI) is a trade organization 
representing nuclear power producers, who collect a surcharge from 
ratepayers for the Nuclear Waste Fund (established by the NWPA, see 42 
U.S.C. 10222). NEI challenged the ground-water protection provisions in 
40 CFR 197.30 on several grounds, including that:
     They conflict with the direction in the EnPA that EPA 
issue standards ``based upon and consistent with the findings and 
recommendations of'' NAS and that EPA's ``standards shall prescribe the 
maximum annual effective dose equivalent * * * from releases * * * from 
radioactive materials stored or disposed of in the repository.'' NEI 
argued that EPA's ground-water standards: (1) were in a form other than 
effective dose equivalent (EDE); (2) were not recommended by NAS, which 
stated that such standards were not ``necessary to limit risks to 
individuals'' (NAS Report p. 121); and (3) were not limited to releases 
from the repository because they require that DOE consider natural 
background when determining compliance.
     The science underlying the ground-water standards uses the 
outdated ``critical organ'' methodology, which results in inconsistent 
risk estimates and is inconsistent with other radiation-protection 
standards.
     EPA justified its ground-water standards on cost grounds 
without conducting a thorough cost-benefit analysis; NEI believes such 
an analysis would show that the ground-water standards provide no 
benefit to public health but will increase the cost and slow the 
construction of the repository.
     EPA is inappropriately applying drinking water standards, 
which were derived to apply to customers of public water supplies 
(i.e., ``at the tap'') to ground water.

C. Ruling by the U.S. Court of Appeals for the District of Columbia 
Circuit

    Oral arguments for the challenges described above were heard on 
January 14, 2004. The challenges to EPA's standards were consolidated 
with challenges to NRC's licensing requirements, DOE's siting 
guidelines, and the Presidential recommendation of the Yucca Mountain 
site and the subsequent Congressional resolution. The Court's ruling 
was handed down on July 9, 2004. The Court upheld EPA's Yucca Mountain 
rule in all respects, save for the regulatory compliance period.
1. What Did the Court of Appeals Rule on the Issue of Compliance 
Period?
    The Court upheld the challenge to EPA's 10,000-year compliance 
period, ruling that EPA's action was not ``based upon and consistent 
with'' the NAS Report, and that EPA had not sufficiently justified its 
decision to apply compliance standards only to the first 10,000 years 
after disposal on policy grounds. Nuclear Energy Institute v. 
Environmental Protection Agency, 373 F.3d 1 (D.C. Cir. 2004) (NEI) 
(Docket No. OAR-2005-0083-0080). On that point, the Court stated that:

    NAS's conclusion that EPA ``might choose to establish consistent 
policies'' is of little importance * * * And although our case law 
makes clear that a phrase like ``based upon and consistent with'' 
does not require EPA to hew rigidly to NAS's findings, EnPA Section 
801(a) cannot reasonably be read to allow a regulation wholly 
inconsistent with NAS recommendations. (NEI, 373 F.3d at 30.)

    Similarly, the Court rejected EPA's reasoning that the requirement 
of 40

[[Page 49022]]

CFR 197.35 that DOE project performance to the time of peak dose and 
place those projections in the Environmental Impact Statement (EIS) 
addressed the intent of the NAS recommendation by ensuring that 
assessments would not be arbitrarily cut off at some earlier time:

    Although EPA's addition of this provision might well represent a 
nod to NAS, it hardly makes the agency's regulation consistent with 
the Academy's findings. NAS recommended that the compliance period 
extend to the time of peak risk, yet EPA's rule requires only that 
DOE calculate peak doses and expressly provides that ``[n]o 
regulatory standard applies to the results of this analysis.'' (Id. 
at 31, emphasis in original)

    While the Court suggested that under different circumstances the 
Agency's standard might have been upheld, it nevertheless rejected the 
Agency's limitation of the compliance period to 10,000 years:

    In sum, because EPA's chosen compliance period sharply differs 
from NAS's findings and recommendations, it represents an 
unreasonable construction of section 801(a) of the Energy Policy 
Act. Although EnPA's ``based upon and consistent with'' mandate 
leaves EPA with some flexibility in crafting standards in light of 
NAS's findings, EPA may not stretch this flexibility to cover 
standards that are inconsistent with the NAS Report. Had EPA begun 
with the Academy's recommendation to base the compliance period on 
peak dosage and then made adjustments to accommodate policy 
considerations not considered by NAS, this might be a very different 
case. But as the foregoing discussion demonstrates, EPA wholly 
rejected the Academy's recommendations. We will thus vacate part 197 
to the extent that it requires DOE to show compliance for only 
10,000 years following disposal. (Id. at 31.)

    Finally, the Court concluded that ``we vacate 40 CFR part 197 to 
the extent that it incorporates a 10,000-year compliance period'' * * * 
(Id. at 100.) The Court did not address the protectiveness of the 150 
Sv/yr (15 mrem/yr) dose standard applied over the 10,000-year 
compliance period, nor was the protectiveness of the standard 
challenged. It ruled only that the compliance period could not be found 
consistent with or based upon the NAS findings and recommendations, and 
therefore was contrary to the plain language of the EnPA.

a. What Were NAS's Findings (``Conclusions'') and Recommendations on 
the Issue of Compliance Period?

    As the Court noted, NAS stated that it had found ``no scientific 
basis for limiting the time period of the individual-risk standard to 
10,000 years or any other value,'' and that ``compliance assessment is 
feasible * * * on the time scale of the long-term stability of the 
fundamental geologic regime--a time scale that is on the order of 10\6\ 
years at Yucca Mountain.'' As a result, and given that ``at least some 
potentially important exposures might not occur until after several 
hundred thousand years * * * we recommend that compliance assessment be 
conducted for the time when the greatest risk occurs'' (NAS Report pp. 
6-7).
    However, NAS also stated ``although the selection of a time period 
of applicability has scientific elements, it also has policy aspects 
that we have not addressed. For example, EPA might choose to establish 
consistent policies for managing risks from disposal of both long-lived 
hazardous nonradioactive materials and radioactive materials' (NAS 
Report p. 56).
2. What Did the Court of Appeals Rule on Other Issues Related to EPA's 
Standards?
    The Court did not sustain any of the other challenges lodged by 
Nevada, NRDC, or NEI. Instead, the Court found that:
     In defining the controlled area, EPA's conclusions 
regarding the likely extent of the future population and their 
exposures were reasonable. Further, the provisions of the Safe Drinking 
Water Act do not apply at Yucca Mountain (by virtue of the EnPA 
statement that EPA's standards ``shall be the only standards applicable 
to the Yucca Mountain site''). (NEI, 373 F. 3d at 32-38.)
     EPA is not bound to follow the NWPA definition of 
``disposal'' because the enabling authority for this action is the 
EnPA, which does not require that NWPA definitions be used and does not 
itself define ``disposal.'' Therefore, EPA acted reasonably ``in 
filling that statutory gap.'' (Id. at 38-39.)
     EPA's interpretation of the EnPA as permitting separate 
ground-water standards is reasonable because: (1) The EnPA does not 
restrict EPA to establish only EDE standards, but requires that EPA 
``establish a set of health and safety standards, at least one of which 
must include an EDE-based, individual-protection standard''; (2) NAS 
made no ``finding or recommendation'' either for or against a ground-
water standard, so consistency with NAS is not at issue; and (3) ``Part 
197 * * * does not regulate background radiation * * * the rule 
requires only that DOE take background levels into account when 
measuring permissible releases of radionuclides from the repository. 
Therefore, part 197 could not possibly run afoul of EnPA's focus on 
released radiation.'' (Id. at 43-48.)
     NEI's arbitrary and capricious arguments in NEI were the 
same as the arguments that NEI had raised in a challenge to EPA's 
radionuclide MCLs under the Safe Drinking Water Act, which the Court 
had rejected only one year previously in City of Waukesha v. EPA. (Id. 
at 48-49.)
     EPA ``unremarkably'' concluded that ground-water 
protection standards represent sound pollution prevention policy and 
will encourage a more robust repository design. This reasoning 
prevailed with the Court on both the cost-effectiveness and ``at the 
tap'' challenges. (Id. at 49-50.)

II. How Will EPA Address the Decision by the Court of Appeals?

    As promulgated, 40 CFR part 197 contained four sets of standards 
against which compliance would be assessed. The storage standard 
applies to exposures of the general public during the operational 
period, when waste is received at the site, handled in preparation for 
emplacement in the repository, emplaced in the repository, and stored 
in the repository until final closure. The three disposal standards 
apply to releases of radionuclides from the disposal system after final 
closure, and include an individual-protection standard, a human-
intrusion standard, and a set of ground-water protection standards.
    In today's action, we are not proposing to revise all of these 
standards, only those affected by the Court decision. Therefore, we are 
proposing to revise only the individual-protection and human-intrusion 
standards, along with certain supporting provisions related to the way 
DOE must consider features, events, and processes (FEPs) in its 
compliance analyses. In addition, we are proposing to adopt updated 
scientific factors for calculating doses to show compliance with the 
storage, individual-protection, and human-intrusion standards, as 
described in more detail in Section II.C.6. We are not proposing to 
change any aspect of the ground-water protection standards. We are 
providing notice and requesting public comment only on our proposed 
revisions to 40 CFR part 197. With the exception of the updated factors 
for calculating doses for the storage standard, we are not requesting 
and will not consider public comment on either the storage or ground-
water protection standards. Furthermore, we are not requesting, nor 
will we consider, comments on those aspects of the individual-
protection and

[[Page 49023]]

human-intrusion standards to which no changes are proposed.
    We are proposing to address the Court's decision by revising 
elements of our standards to incorporate the time of peak dose into the 
determination of compliance. We are also proposing to further delineate 
how DOE should incorporate features, events, and processes that may 
take place over very long times into its calculation of peak dose, 
consistent with our ``reasonable expectation'' standard.

A. How Will Elements of the Disposal Standards be Affected?

    The Court's ruling vacated only one aspect of 40 CFR part 197, the 
10,000-year compliance period. Thus, we considered the language and 
reasoning of the Court's decision to determine its applicability to 
each element of the disposal standards. The three main components of 
the standards are discussed in the following sections. We also 
considered the need to modify certain other aspects that would 
influence how DOE would conduct its performance assessments beyond 
10,000 years. These aspects are discussed in more detail in Section 
II.D (``How Will Today's Proposal Affect the Way DOE Conducts 
Performance Assessments?'').
1. Individual-Protection Standard
    The Court's decision clearly affects the compliance period for the 
individual-protection standard, which is the primary standard for 
public health and safety called for by the EnPA. The legal challenge 
and the Court's response left no doubt that the compliance period for 
the individual-protection standard was at issue and the decision 
centered on the NAS's recommendation regarding the compliance period 
for the individual-protection standard. Therefore, as described in 
Section II.C, we are proposing today to modify the individual-
protection standard to incorporate a compliance measure effective at 
the time of peak dose, in addition to the 15 mrem/yr standard 
applicable for the first 10,000 years after disposal, which we are 
retaining.
    Section I.A.1.b.i discusses other elements of the individual-
protection standard, specifically the definition of the controlled area 
and the use of the RMEI as the representative exposed person. We are 
not modifying the definition of the controlled area, which was upheld 
by the Court. We have described the maximum extent of the area, using 
current conditions and relatively near-term plans for development. The 
actual compliance point will be determined through the licensing 
process, and DOE will have to justify its reasons for selecting a 
particular location to NRC.
    Similarly, we are not proposing to alter the description of the 
RMEI as a person having a ``rural-residential'' lifestyle as reflected 
in today's population. We have described at length our reasons for 
using current characteristics as an appropriate means to avoid 
excessive speculation about which of the infinite number of possible 
future lifestyles would be most representative over very long periods 
(see 66 FR 32088-32094 (Docket No. OAR-2005-0083-0042) and Section 4 of 
the Response to Comments document for the 2001 rulemaking (Docket No. 
OAR-2005-0083-0050)). Some comments on our 1999 proposal disagreed with 
our reasoning and choice of RMEI. We recognize that interested parties 
may see an extension of the compliance period as justifying a different 
description for the RMEI, at least for time frames well beyond 10,000 
years. They may point to climate change scenarios as potentially making 
the ``rural-residential'' lifestyle as it is defined in our 2001 rule 
incompatible with climate change assumptions. It may be argued that 
climate change could significantly affect the types of locally grown 
food in the RMEI's diet, as well as the use of contaminated ground 
water for irrigation or watering livestock, which would ultimately 
influence exposures. NAS alluded to such a possibility, noting that one 
effect of climate change could be ``a shift in the distribution and 
activities of human populations'' (NAS Report p. 92). However, NAS also 
concluded that ``there is no simple relation between future climatic 
conditions and future population'' (NAS Report p. 92). We agree that it 
is difficult to predict exactly how climate change, or other 
evolutionary scenarios, would influence lifestyles, nor can we predict 
the viability or distribution of agricultural activities compared with 
those pursued today. In fact, we believe that the RMEI as a current 
``rural-residential'' individual may be among the more conservative 
possibilities. Given the importance of irrigation and other uses of 
ground water in the Amargosa Valley region, it is likely that potential 
exposures to contaminated ground water would be lower under many wetter 
climate change scenarios where greater precipitation could reduce the 
use of ground water for irrigation and other practices.
    Some commenters might question whether it is important to have 
internal consistency between climate/biosphere characteristics and RMEI 
lifestyle and characteristics. We believe that it would be highly 
speculative to select RMEI characteristics to correspond to some future 
climate state. We require that DOE consider climate change within 
10,000 years, and are proposing today also to require consideration of 
climate change for much longer times (see Section II.D.2.d, 
``Consideration of Climatological FEPs''). As noted above, we believe 
the present-day RMEI represents a conservative choice if, as seems 
likely, future climate in the Yucca Mountain region tends to be cooler 
and wetter. Under wetter conditions, agricultural activities around the 
site area would rely less on irrigation using well water. With less use 
of contaminated ground water for irrigation, the contribution to the 
RMEI dose from contaminated food would presumably be lowered or perhaps 
eliminated. In counterpoint, under wetter conditions, it is possible to 
speculate that individuals could live closer to the repository than is 
considered for present-day conditions and potentially tap contaminated 
ground waters closer to Yucca Mountain than at the RMEI location. We 
believe that the RMEI, as presently defined for present-day conditions, 
is a reasonably conservative approach for the dose assessments, and is 
appropriate for wetter climate conditions. Assumptions regarding the 
possible uses of ground water are quite speculative and have been 
avoided to the extent possible in the setting of the standards (66 FR 
32111). Therefore we are not redefining the RMEI characteristics in any 
attempt to correlate them with climatic variations, primarily due to 
speculation regarding the uses of ground water by man. As noted above, 
this approach is consistent with the NAS's conclusion that there is no 
exact correlation between potential climate changes and shifts in the 
distribution and activities of human populations. Comments on the 
definition of the controlled area and specification of the RMEI are 
outside the scope of today's proposal. We will not consider or respond 
to comments on these topics.
2. Human-Intrusion Standard
    While the Court did not specifically address the human-intrusion 
standard, we believe it is logical and defensible to modify it to 
parallel the individual-protection standard. Like the individual-
protection standard, our provisions for human intrusion envisioned some 
consideration of performance beyond 10,000 years. The 2001 standard 
required that DOE determine when an intrusion by drilling would be 
possible and assess the consequences. The resulting exposures

[[Page 49024]]

were then subject to the same compliance standard as the individual-
protection standard (15 mrem/yr at 10,000 years or earlier and dose 
projections beyond 10,000 years to be compiled in the EIS). In 
proposing revisions to the human-intrusion standard to conform to 
changes we are proposing to make to the individual-protection 
provisions, we are adhering to the NAS recommendation that ``EPA 
require that the estimated risk calculated from the assumed intrusion 
scenario be no greater than the risk limit adopted for the undisturbed-
repository case'' (NAS Report p. 12). In light of this recommendation, 
and the Court's interpretation of how closely we must align with the 
NAS recommendations to be deemed ``based upon and consistent,'' we 
believe it is both prudent and reasonable to propose to revise the 
human-intrusion standards to incorporate peak dose compliance measures 
that conform to the proposed revisions for individual protection.
    Aside from the application of dose standards at both 10,000 years 
and the time of peak dose, the foundation of the proposed revised 
human-intrusion standard is unchanged. DOE must determine the earliest 
time at which it would be possible to penetrate waste packages by 
drilling. The scenario described in Sec.  197.26 would still apply 
(i.e., penetration of a single package, direct pathway to ground water, 
etc.). The decision to apply a regulatory standard for the period of 
geologic stability does not in any way affect the reasoning underlying 
the selection of this scenario. It remains fully consistent with the 
NAS conclusion that at Yucca Mountain ``there is no scientific basis 
for estimating the probability of intrusion at far-future times'' (NAS 
Report p. 106). Instead, NAS recommended that ``the result of the 
analysis should not be integrated into an assessment of repository 
performance based on risk, but rather should be considered separately. 
The purpose of this consequence analysis is to evaluate the resilience 
of the repository to intrusion'' (NAS Report p. 109). NAS further 
suggested that EPA describe a ``stylized'' intrusion scenario based on 
current drilling technologies, an approach we adopted in Sec.  197.26 
and which will remain unchanged by today's proposal.
    The circumstances of the intrusion scenario in Sec.  197.26 are 
required to be developed based on present-day practices, in accordance 
with the NAS recommendation. This approach was fully justified for the 
reasons given by NAS and unchallenged for the 10,000-year time frame. 
We find that maintaining the approach beyond 10,000 years is also fully 
justified and consistent with the NAS for the same reasons. If 
anything, it would be even more speculative to attempt to project 
changes to the circumstances of the intrusion at time frames 
potentially out to 1 million years. Furthermore, in keeping with the 
purpose of the human-intrusion analysis as a test of repository 
resilience, it is appropriate to continue to exclude unlikely natural 
events and processes from the analysis.
    The intrusion scenario requires consideration of package 
degradation, premised on the assumption that drillers encountering an 
intact package would cease drilling and releases would be avoided. We 
believe that this assumption is equally valid both within and beyond a 
10,000-year time frame. In our 2001 rule, DOE would not have been 
required to demonstrate compliance with a dose limit if packages were 
determined not to degrade sufficiently within 10,000 years to permit 
intrusion (or, in any event, if the consequences of the intrusion were 
not calculated to occur within 10,000 years). We are proposing to 
modify our rule to require that DOE show compliance with a dose limit 
regardless of when the consequences of the intrusion occur. Consistent 
with the proposed revised individual-protection standard, DOE will have 
to show compliance with a peak dose standard beyond 10,000 years, in 
addition to a 150 [mu]Sv/yr (15 mrem/yr) standard applicable up to 
10,000 years. The dose standard that applies to exposures to the RMEI 
through the period of geologic stability will be the same as for the 
individual-protection standard (see Section II.C.3, ``What Dose Level 
is EPA Proposing for Peak Dose?''). Overall, this scenario continues to 
represent a reasonable test that ``can provide useful insight into the 
degree to which the ability of a repository to protect public health 
would be degraded by intrusion'' (NAS Report p. 108). We are not 
soliciting, and will not consider, comments on the overall intrusion 
scenario or other aspects of the human-intrusion standard that are not 
proposed to be changed.
3. Ground-Water Protection Standards
    The Court's decision does not affect the ground-water protection 
standards. The Court upheld our statutory reading of the EnPA as 
providing the authority to establish such standards as the Agency 
deemed necessary to supplement the individual-protection standard, as 
well as the scientific basis of those standards. (See NEI, 373 F.3d at 
43-48, Docket No. OAR-2005-0083-0080.) The Court further concluded that 
our reasoning for including such a standard as a means to protect the 
ground-water resource was sound and consistent with the Agency's 
overall pollution prevention policies. Regarding consistency with the 
NAS recommendations, the Court stated that:

    Although we concluded earlier in this opinion that EPA violated 
section 801's ``based upon and consistent with'' requirement by 
adopting a 10,000-year compliance period, we reach the opposite 
conclusion here because NAS treated the compliance-period and 
ground-water issues quite differently. Whereas NAS expressly 
rejected a 10,000-year compliance period, it said nothing at all 
about the need to add a separate ground-water standard * * * Put 
another way, NAS made no ``finding'' or ``recommendation'' that 
EPA's regulation could fail to be ``based upon and consistent 
with.''

NEI, 373 F.3d at 46-47.

    As a result, we do not believe the Court's ruling regarding the 
10,000-year compliance period applies to the ground-water protection 
standards, which have the same compliance period. Further, unlike the 
individual-protection and human-intrusion standards, we never 
envisioned that DOE would project its compliance with the ground-water 
protection standards beyond 10,000 years, even for inclusion in the 
EIS. The Court decision leaves EPA with discretion in formulating the 
provisions for ground-water standards. We believe (and the Court 
agreed) that the application over 10,000 years of limits equivalent to 
MCLs is a conservative but reasonable regulatory scheme that represents 
sound pollution prevention policy. Furthermore, protection of public 
health from releases to ground water over times beyond 10,000 years 
will be provided by extending the individual-protection standard to the 
time of peak dose, which accounts for transport and exposure through 
all pathways. For these reasons, we are not proposing to modify the 
ground-water protection standards, either by extending the period of 
compliance or in any other respect. We are not requesting, and will not 
consider, comments regarding any aspect of the ground-water protection 
standards.
4. Reasonable Expectation
    ``Reasonable expectation'' is the compliance concept underlying our 
disposal standards. That is, we require that DOE show a ``reasonable 
expectation'' that the standards will be met. As discussed extensively 
in our 2001 Yucca Mountain rulemaking, ``proof'' of disposal system 
performance

[[Page 49025]]

in the traditional sense of the word cannot be attained for periods 
extending into the thousands or hundreds of thousands of years (66 FR 
32101-32103, June 13, 2001, Docket No. OAR-2005-0083-0042). In such 
situations, it is a natural tendency to give greater emphasis to 
aspects that may not be the most likely to occur, but have the 
potential to significantly affect performance. This may be particularly 
true in areas where physical data are limited. However, assessments 
that are built around conservative assumptions at every decision point 
may in fact result in highly unrealistic performance projections. 
Simplifications and assumptions are involved out of necessity because 
of the complexity and time frames involved, and the choices made will 
determine the extent to which modeling simulations realistically 
simulate the disposal system's performance. If choices are made that 
make the simulations very unrealistic, the confidence that can be 
placed on modeling results is very limited. The uncertainties involved 
with these simplifications must be recognized. Overly conservative 
assumptions made in developing performance scenarios can bias the 
analyses in the direction of unrealistically extreme situations, which 
in reality may be highly improbable, and can deflect attention from 
questions critical to developing an adequate understanding of the 
expected features, events, and processes. ``Reasonable expectation'' 
encourages the use of ``cautious, but reasonable'' assumptions and 
discourages the reliance on highly conservative assumptions. It 
recognizes that projections of disposal system performance over very 
long times are best viewed as indicators of performance, rather than as 
firm predictions. It further requires the applicant and regulator to 
focus on the full range of outcomes and not to give greater weight to 
certain projections simply because they are more conservative.
    The concept of ``reasonable expectation'' was a guiding principle 
in the formulation of our 2001 standards. We believe the concept is 
equally applicable for periods well beyond 10,000 years, and is in fact 
more important for very long time periods. In our view, it is 
``reasonable'' to consider approaches for uncertainties in calculations 
at several hundred thousand years that may differ from the approach for 
uncertainties considered within 10,000 years after disposal. An 
approach applying standards ``acceptable today for the period of 
geologic stability would ignore this cumulative uncertainty and the 
extreme difficulty of using highly uncertain assessment results to 
determine compliance with that standard'' (66 FR 32098, June 13, 2001, 
Docket No. OAR-2005-0083-0042). We therefore emphasize the primacy of 
``reasonable expectation'' in compliance with 40 CFR part 197 and 
retain it without change. However, we have considered how DOE and NRC 
might need to approach the concept to account for the much greater 
overall uncertainty in projections over periods as long as 1 million 
years. Section II.B describes the overall concept of ``reasonable 
expectation'' and our thoughts for today's proposal in more detail.
5. Effects of Uncertainty
    We believe that the most problematic aspect of extending the 
compliance period to peak dose is the uncertainty involved in making 
projections over such long time frames, which we discussed in some 
detail in our proposed and final rulemakings in 1999 and 2001, 
respectively. This remains a critical factor in formulating today's 
proposal, which we feel must be emphasized and explored in detail. 
Although we refer generally to ``uncertainties'' throughout this 
document, it may not always be clear to readers exactly what we mean by 
this term, why their effects are difficult to manage, and why they 
should have an impact on the decision-making process. It may be useful 
to consider an analogous situation that will be readily familiar, such 
as the tracking of hurricanes.
    The strength and path of hurricanes are functions of factors such 
as temperature, humidity, barometric pressure, and wind speed. There is 
natural variation in these parameters, and their variation can make the 
difference between a Category 5 storm (the most severe) striking a 
populated coastal area and a tropical storm that remains out in the 
ocean. When one views the projected path of a storm, the surrounding 
envelope of possible paths expands as one looks into the future and may 
spread over several hundred miles. The critical task in tracking the 
storm is identifying which populated areas are in the path of the 
storm, and whether they must be evacuated.
    By this analogy, a 10,000-year dose projection might be comparable 
to selecting a single town to evacuate when the storm is still two 
hundred miles from landfall, while a peak dose projection might be more 
like pinpointing the correct location when a tropical depression first 
forms thousands of miles away, which may be weeks earlier. Regardless 
of the level of rigor that can be applied to the technical calculation, 
it is simply not possible to place the same level of confidence in the 
two selections. We see similar difficulties in ``predicting'' the 
``true'' behavior of the Yucca Mountain disposal system, or the 
multiple engineered and natural components of that system, for periods 
on the order of hundreds of thousands of years.
    We are aware that some stakeholders dispute our position that 
uncertainties increase significantly with time, and therefore believe 
that uncertainty offers little justification for placing less 
confidence in very long-term projections than can be placed in those 
that apply over the relatively near term. Some stakeholders, for 
example, suggest that uncertainty should have little impact on peak 
dose projections and that DOE should be required to identify where 
uncertainty, rather than reasonably expected performance, influences 
dose projections (Docket No. OAR-2005-0083-0029 and 0033). They have 
pointed to statements in the NAS Report to bolster this position, such 
as: ``analyses that are uncertain at one time might not be so uncertain 
at a later time; for example, the uncertainties about cumulative 
releases to the biosphere that depend on the rate of failure of the 
waste packages are large in the near term but are smaller later, when 
enough time has passed that all of the packages will have failed'' (NAS 
Report pp. 29-30); ``Because there is a continuing increase in 
uncertainty about most of the parameters describing the repository 
system farther in the distant future, it might be expected that 
compliance of the repository in the near term could be assessed with 
more confidence. This is not necessarily true'' (NAS Report p. 72); 
``Detailed estimates of time for canister failure are less important 
for much longer-term estimates of individual dose or risk'' (NAS Report 
p. 85).
    Although NAS pointed out that uncertainties associated with some 
disposal system components will decrease over time (e.g., at some time 
all waste packages will be degraded), our view, and the view of many 
others (including NAS, as should be clear from the above citation: 
``Because there is a continuing increase in uncertainty * * *''), is 
that uncertainties generally increase with time, at least to the time 
of peak dose. (See, for example, IAEA Draft Safety Requirements DS154, 
``Geological Disposal of Radioactive Waste,'' Section A.7, page 37, 
April 2005 (Docket No. OAR-2005-0083-

[[Page 49026]]

0051), which states, ``It is recognized that radiation doses to people 
in the future can only be estimated and the uncertainties associated 
with these estimates will increase farther into the future''; the 
Nuclear Energy Agency report on ``The Handling of Timescales in 
Assessing Post-Closure Safety,'' pp. 13-14 (Docket No. OAR-2005-0083-
0046), which states, ``These events and changes are subject to 
uncertainties, which generally increase with time and must be taken 
into account in safety assessments. Eventually, but at very different 
times for different parts of the system, uncertainties are so large 
that predictions regarding the evolution of the repository and its 
environment cannot meaningfully be made''; and the Swiss National 
Cooperative for the Disposal of Radioactive Waste (Nagra), which 
states, in Technical Report 02-05 (pp. 27-28) (Docket No. OAR-2005-
0083-0075), ``HSK-R-21 [Swiss disposal regulation] acknowledges that 
there is inevitable uncertainty in model calculations and the further 
into the future predictions are made, the greater the uncertainty. The 
implementer has to show what processes and events could affect the 
repository over the course of time and then to derive and evaluate 
potential evolution scenarios from these.'') For some aspects of the 
system, such uncertainties can increase dramatically (``Assumptions, 
Conservatisms, and Uncertainties in Yucca Mountain Performance 
Assessments,'' Section 12.3, July 2005, Docket No. OAR-2005-0083-0085). 
To repeat, we are in agreement with NAS that such projections can be 
performed and even ``bounded'' to some extent. However, the central 
question here is how the results of very long-term assessments can have 
sufficient meaning to provide an adequate basis for a licensing 
decision that the repository should or should not be approved.
    NAS demonstrated some concern with this issue by recognizing that 
the level of confidence that could be placed in projections was of key 
importance, and offered constructive guidance in limiting or 
considering the effects of uncertainties. Unfortunately, the NAS 
statements on decreasing uncertainty regarding some disposal system 
components do not draw a clear relationship to the time of peak dose at 
which it recommended compliance be measured. While we generally agree 
with these statements, we find that they are most relevant to times 
after peak dose and, therefore, after the time frame most important 
from a regulatory perspective. Returning to our hurricane analogy, it 
is true that uncertainties eventually decrease; one might be able to 
predict with equal confidence both the storm's location in two hours 
and that in two weeks it will have completely dissipated. In this 
sense, one can agree with the NAS's conclusion that ``it is not 
necessarily true'' that long-term projections are more uncertain than 
near-term projections. Nevertheless, relatively high confidence about 
the endpoint of the hurricane has little impact on the ability to 
predict where and when it might cause the greatest damage along its 
path. Similarly, for Yucca Mountain, increasing confidence in certain 
aspects of the system's components (e.g., the endpoint of the waste 
packages, much like the endpoint of the hurricane) does not necessarily 
inform estimates of peak dose.
    NAS notes that ``uncertainties about cumulative releases'' that 
``depend on the rate of failure of the waste packages'' will be 
lessened at far future times when ``all of the packages will have 
failed'' (NAS Report p. 28-29). The emphasis here on eventual failure 
cannot help us when the direction is to assess peak dose. It is self-
evident and non-controversial that the engineered barrier system cannot 
be expected to last forever. However, assumptions regarding ``the rate 
of failure of waste packages'' are exactly the critical element in 
estimating the timing and magnitude of the peak dose (``Assumptions, 
Conservatisms, and Uncertainties in Yucca Mountain Performance 
Assessments,'' Sections 12.3 and 12.4, July 2005, Docket No. OAR-2005-
0083-0085). Thus, identifying factors that would decrease overall 
system uncertainty at times approaching 1 million years does not 
adequately support a conclusion that uncertainties can be equally well 
managed at the time of peak dose, even if that time is much less than 1 
million years.
    In addressing this larger question of how to consider long-term 
projections in a regulatory process, we have considered guidance and 
precedents from international programs. NAS provided important 
scientific and technical reasoning for evaluating compliance at peak 
dose, which we augment with guidance from sources who approached the 
problem of uncertainty from the regulatory perspective. For regulatory 
compliance over 10,000 years, we were able to identify several (albeit 
limited) analogous regulatory programs in the U.S., including those for 
the WIPP and EPA's underground injection control program (see the 
preamble to the 2001 rulemaking, 66 FR 32098, Docket No. OAR-2005-0083-
0042). For time frames extending potentially to 1 million years, there 
are no precedents in U.S. regulation. In response to the Court 
decision, therefore, important sources for guidance and models for 
contemplating regulations at such long times were other international 
programs grappling with the same issues, namely disposal of highly 
radioactive and long-lived waste. Throughout this document, we quote 
extensively from a number of international sources, from both 
multinational organizations (such as IAEA) and individual countries 
(such as Sweden). We do this because we find ourselves in a situation 
that is, if not unique, shared by a rather small circle. We have found 
it useful to consult the ideas of those faced with a similar situation. 
In general, they reinforce two points we emphasize throughout this 
document. The first, which we have already discussed, is that 
uncertainties generally increase with time. The second point is that 
projections at those longer times cannot be viewed with the same level 
of confidence as shorter-term projections, and may in fact be viewed as 
more qualitative indicators of disposal system performance.
    For example, the IAEA has stated that, for periods lasting from 
about 10,000 to 1 million years, ``While it may be possible to make 
general predictions about geological conditions, the range of possible 
biospheric conditions and human behaviour is too wide to allow reliable 
modelling * * * Such calculations can therefore only be viewed as 
illustrative and the `doses' as indicative'' (IAEA-TECDOC-767, ``Safety 
Indicators in Different Time Frames for the Safety Assessment of 
Underground Radioactive Waste Repositories,'' p. 19, 1994, Docket No. 
OAR-2005-0083-0044). Also, ``[t]he utility of individual numerical 
indicators will vary greatly and, given the large uncertainties, 
considerable caution is needed to avoid any suggestion or expectation 
that any given indicator of disposal system performance can be an 
accurate estimate of future reality. Such an indicator typically 
provides only an estimate of what might happen under certain assumed 
conditions * * * The aim of the assessment is not to predict the actual 
performance of the disposal system * * * but rather to reach reasonable 
assurance that it will provide an adequate level of safety'' (IAEA-
TECDOC-975, ``Regulatory Decision Making in the Presence of Uncertainty 
in the Context of the Disposal of Long Lived Radioactive Wastes,'' pp. 
22, 24,

[[Page 49027]]

1997, Docket No. OAR-2005-0083-0045). Finally, ``[c]are has to be 
exercised in applying the criteria for periods beyond the time where 
the uncertainties become so large that the criteria may no longer serve 
as a reasonable basis for decision making'' (IAEA Draft Safety 
Requirements DS154, ``Geological Disposal of Radioactive Waste,'' 
Section A.7, p. 37, April 2005, Docket No. OAR-2005-0083-0051).
    The Nuclear Energy Agency (NEA) states that ``[t]here is an 
increasing consensus among both implementers and regulators that, in 
carrying out safety assessments, calculations of dose and risk should 
not be extended to times beyond those for which the assumptions 
underlying the models and data can be justified * * * Eventually, but 
at very different times for different parts of the system, 
uncertainties are so large that predictions regarding the evolution of 
the repository and its environment cannot meaningfully be made'' (``The 
Handling of Timescales in Assessing Post-Closure Safety,'' pp. 10, 13, 
2004, Docket No. OAR-2005-0083-0046). Similarly, the Swedish Radiation 
Protection Authority (SSI) has proposed draft guidance for the disposal 
of SNF, stating that ``[f]or very long periods * * * [t]he intention 
should be to shed light on the protective capability of the repository 
and to provide a qualitative picture of the risks'' (p. 7, Docket No. 
OAR-2005-0083-0048). This draft guidance is intended to supplement 
SSI's standards (SSI FS 1998:1, September 28, 1998, Docket No. OAR-
2005-0083-0047), which require that ``[f]or the first thousand years 
after disposal, the assessment of the repository's protective 
capability shall be based on quantitative analyses of the impact on 
human health and the environment'' (Sec.  11), but do not specify 
quantitative analyses as the basis for longer-term assessments (``shall 
be based on various possible sequences for the development of the 
repository's properties, its environment and the biosphere,'' Sec.  
12).
    We acknowledge that detailing the effects of uncertainty is itself 
uncertain. We recognize that knowledge is not absolute up to 10,000 
years, with uncertainties burgeoning shortly beyond that time. We also 
recognize that there can be considerable uncertainty in measurements of 
current conditions. Further, we concur with NAS that uncertainties can 
be qualitatively different for different aspects of the assessment. For 
example, NAS points out that human behavior can be projected for a few 
decades at most, while the geologic record can be studied for evidence 
of processes that have occurred over millions of years (and are still 
occurring today). However, the assessment of Yucca Mountain's 
performance depends not only on the ability to project large-scale 
geologic processes, such as seismicity and volcanism, but also the 
gradual evolution of complex saturated and unsaturated zone 
characteristics, such as the chemistry of infiltrating water or the 
direction and connectivity of a fracture-flow system.

B. How Does the Application of ``Reasonable Expectation'' Influence 
Today's Proposal?

    Under today's proposal, projecting disposal system performance 
involves the extrapolation of physical conditions and the interaction 
of natural processes with the wastes for unprecedented time frames in 
human experience, i.e., possibly hundreds of thousands of years. In 
this sense, the projections of the disposal system's long-term 
performance cannot be confirmed. Not only is the projected performance 
of the disposal system not subject to confirmation, the natural 
conditions in and around the repository site will vary over time and 
these changes are also not subject to confirmation, making their use in 
performance assessments equally problematic over the long-term. In 
light of these fundamental limitations on assessing the disposal 
system's long-term performance, we believe that the approach used to 
evaluate disposal system performance must take into account the 
fundamental limitations involved and not hold out the prospect of a 
greater degree of ``proof'' than in reality can be obtained.
    There are several fundamental components to be established in 
setting up and analyzing disposal system performance scenarios. A model 
must be created that translates the physical processes operating at the 
site into mathematical statements, such as ground-water flow equations, 
that can calculate the movement of radionuclides through the various 
components of the disposal system and into the accessible environment. 
A model may be very generic or highly sophisticated and tailored to 
capture distinct aspects of a particular site. Two additional steps are 
necessary in order to develop dose projections. First, the possible 
performance scenarios themselves and associated assumptions must be 
established, and second, the distribution of expected values for the 
parameters involved in the performance calculations must be determined. 
The scenarios are developed from an understanding of the natural 
processes, the engineered barrier design, and the interactions of the 
engineered barrier system with the repository environment. The range of 
expected parameter values for the analyses is based upon the results of 
site characterization studies, laboratory testing, and expert judgment. 
For both of these components, unrealistic and perhaps extreme choices 
can be made that would, in effect, give false expectations of disposal 
system performance, or hide important uncertainties that would, in 
reality, have important consequences on the performance projections 
(the model itself may also have conservatisms built into it, which may 
be even more difficult to identify). If extreme assumptions are made in 
defining the scenario, a de facto ``worst-case'' scenario is developed 
at the outset and analyses using the upper end of the range of 
parameter values result in performance projections that are in fact 
extreme cases, rather than representing the full range of expected 
performance. Effectively, such a restrictive approach results in 
emphasis on what would be the conservative extremes of the probability 
distributions for the performance assessments and analyses rather than 
if a realistic approach were taken. In such a case, the regulatory 
judgment would be focusing on extreme situations, rather than on 
evaluating safety under reasonably expected conditions. On the other 
hand, if the scenario were defined more realistically and the same 
distribution of parameter values used, the resultant distribution of 
doses would be closer to the actual expected performance and regulatory 
decisions could be made with confidence that the assessments represent 
a more realistic range of expected performance. Including multiple 
``worst-case'' assumptions in setting up the performance scenarios, 
combined with selecting conservative values for site-related parameter 
distributions, actually corresponds to assessing very low-probability/
high-consequence scenarios that can then easily be mistaken as 
expected-case analyses. Under the reasonable expectation approach, 
expected case as compared to conservative and worst-case assessments 
are more explicitly identified and the uncertainties presented more 
directly so that the reasoning behind regulatory decisions can be more 
easily understood and defended. We note that this approach was also 
recommended by a joint NEA-IAEA peer review of DOE's TSPA to support 
its site recommendation, which states in Section 4.1.3 (``Realism or 
conservatism''):


[[Page 49028]]


    At a fundamental level, it is useful to resort to a 
probabilistic analysis of a system evolution in time if a realistic 
model can be attempted but legitimate uncertainties persist. 
However, if the starting model is built a priori to be conservative, 
exercising it probabilistically has little or no added value, as one 
would still obtain conservative results. In the TSPA-SR a hybrid 
conservative/probabilistic methodology is used, which causes 
assumptions and reality to be mixed in a confusing way. In the 
future it may be appropriate to present: (i) A probabilistic 
analysis based on a realistic or credible representation; and (ii) a 
set of complementary analyses with different conservatisms, in order 
to place the best available knowledge in perspective. These 
ancillary analyses could be given a probabilistic weight as well. 
This should satisfy the regulatory requirements whilst providing a 
better basis for dialogue and decision-making.

    ``An International Peer Review of the Yucca Mountain Project TSPA-
SR,'' pp. 54-55, 2002, Docket No. OAR-2005-0083-0062, emphasis in 
original.
    In making its decisions, the primary task for NRC is to examine the 
projections put forward by DOE to determine ``how much is enough'' in 
terms of the information and analyses presented, i.e., how NRC 
determines when the analyses provide an acceptable level of confidence 
and the results can be interpreted in a way meaningful for regulatory 
compliance. In 40 CFR part 197 as originally promulgated, we did not 
have specific measures in our standards on how to make that judgment. 
NRC, as the implementing agency, must be satisfied with DOE's 
presentation; therefore, we concluded those specific measures of 
satisfaction were appropriate for NRC to determine. Neither did EPA 
specify: (1) Confidence measures for such judgments or numerical 
analyses; (2) analytical methods that must be used for performance 
assessments; (3) quality assurance measures that must be applied; (4) 
statistical measures that define the number or complexity of analyses 
that should be performed; or (5) any assurance measures in addition to 
the numerical limits in the standards. We specified only that the mean 
of the dose assessments must meet the exposure limit.
    We anticipate that if these very long-range performance projections 
(beyond 10,000 years) indicate that repository performance would 
degrade dramatically under a wide range of conditions at some point in 
time, that this would become a concern in the licensing decision. If 
such a dramatic deterioration were projected to occur close to the 
regulatory time period it would be a more pressing concern for 
licensing decisions than if it were to occur many hundreds of thousands 
of years into the future (remembering that the uncertainty in 
performance projections increases with time). With the initial issuance 
of 40 CFR part 197, EPA elected to leave the handling of the very long-
term projections of performance as an implementation decision for the 
regulatory authority, but to impose the requirement that such analyses 
be performed and reported in the EIS. The degree of ``weight'' that 
should be given to these very long-term assessments, we said, is an 
implementation decision that should be left to NRC to determine, by 
balancing the projected performance and the inherent uncertainties in 
these projections against the projected dose levels (2001 Response to 
Comments, p. 7-13, Docket No. OAR-2005-0083-0043).
    We propose to continue this general approach of not specifying the 
bases or mechanisms for a compliance decision, except that the post-
10,000-year analyses are now proposed to be part of the 40 CFR part 197 
standards with a quantitative limit imposed.
    As noted earlier, the conceptual framework of ``reasonable 
expectation'' as promulgated in our 2001 rulemaking is applicable even 
when extending the compliance period to peak dose. In fact, we believe 
it becomes even more important as the level of confidence that can be 
placed in numerical projections decreases over time. However, we are 
not proposing to expand or modify the definition in Sec.  197.14 to 
account for the greater uncertainty between 10,000 years and the time 
of peak dose (within 1 million years of disposal). The existing 
definition describes principles that are applicable for both shorter 
and very long time frames (although the implications of these 
principles may be different, depending on the time frame). To provide 
insight into our interpretation of reasonable expectation at very long 
times, we provide additional information in the remainder of this 
section and throughout our discussion of the proposed changes for NRC 
to consider as it implements our peak dose standard. We believe such 
guidance will be useful, particularly in the context of handling long-
term FEPs, as discussed in Section II.D of this document.
    We emphasize that parameters and scenarios should be included in 
the performance assessment even if they are not among the more highly 
conservative approaches. There is a tendency in long-term assessment to 
introduce conservatisms and to focus on the higher-end dose 
projections, while discounting lower dose projections that may actually 
be just as probable or perhaps represent higher-probability scenarios. 
We stress that DOE should work to ensure that the results express the 
full range of possible outcomes within the bounds of credible scenarios 
and parameter values. Less conservative scenarios (i.e., lower 
projected doses) should not be eliminated unless they are deemed to be 
highly improbable. Of course, the compliance measure will be expressed 
as a specific statistical measure of the results, not the entire range 
of results. The entire range of results is context to be used to assist 
the licensing authority in judging the likelihood of the facility to 
meet the standards. In that context, the results of the performance 
assessments are not to be biased by an overemphasis on low-probability 
scenarios at the expense of results for the entire spectrum of 
reasonably credible and supportable scenarios and parameter values. Our 
position is that the reasonable expectation approach accounts for the 
inherent uncertainties involved in projecting disposal system 
performance by taking into account a large spectrum of possible 
parameter values rather than making assumptions that reflect only 
conservative to very conservative values. We also emphasize that the 
uncertainties in site characteristics over long time frames, and how 
the long-term projections of expected performance of the disposal 
system were made, need to be well understood before regulatory 
decisions are made. We stress again the purpose of the assessments as 
expressed by IAEA: ``The aim of the assessment is not to predict the 
actual performance of the disposal system * * * but rather to reach 
reasonable assurance that it will provide an adequate level of safety'' 
(IAEA-TECDOC-975, p. 24, Docket No. OAR-2005-0083-0045). NAS agrees 
that ``[t]he results of compliance analysis should not, however, be 
interpreted as accurate predictions of the expected behavior of a 
geologic repository'' (NAS Report p. 71, Docket No. OAR-2005-0083-
0076).
    In Section II.D of this document (``How Will Today's Proposal 
Affect the Way DOE Conducts Performance Assessments?''), we propose to 
limit speculation over the long compliance period now being addressed 
by requiring compliance within a performance assessment that continues 
to emphasize the most significant features, events, and processes. The 
purpose is to provide a reasonable test of performance over a range of 
conditions. To do so, we propose to eliminate very unlikely features, 
events, and processes, and the scenarios

[[Page 49029]]

including them, from consideration and specify this in the standards. 
We believe this is consistent with a finding of the NAS: ``It is always 
possible to conceive of some circumstance that, however unlikely it may 
be, will result in someone at some time being exposed to an 
unacceptable radiation dose * * * The challenge is to define a standard 
that specifies a high level of protection but that does not rule out an 
adequately sited and well-designed repository because of highly 
improbable events'' (NAS Report pp. 27-28). We have chosen to do this 
by continuing to place reasonable constraints on the scenarios that 
need to be examined. We believe this is consistent with another finding 
of the NAS: ``We conclude that the probabilities and consequences of 
modifications generated by climate change, seismic activity, and 
volcanic eruptions at Yucca Mountain are sufficiently boundable so that 
these factors can be included in performance assessments that extend 
over periods on the order of about 10\6\ years'' (NAS Report p. 91). 
Typically, as we discuss elsewhere in this document, the term 
``boundable'' implies a ``worst case'' approach (i.e., a ``bounding 
analysis'') to assessing the limits of disposal system performance. We 
do not believe such an approach is appropriate and are not proposing to 
adopt it. Instead, in this context, we interpret ``boundable'' as 
referring to limits that may be placed on the scenarios so that they 
will represent a reasonable test of disposal system performance over 
the very long term, but not be driven by extreme assumptions or endless 
speculation. Thus, we view our treatment of these ``modifiers'' as 
comparable to our specification of a ``stylized'' scenario for human 
intrusion, and consistent with the NAS statement that ``[i]t is 
important that the `rules' for the compliance assessment be established 
in advance of the licensing process'' (NAS Report p. 73).
    In our 1999 preamble to proposed 40 CFR part 197, we said that if 
we were to regulate longer than 10,000 years, we would expect the 
licensing judgment to be less strict in relying on dose projections 
compared to 10,000 years (64 FR 46998, August 17, 1999, Docket No. OAR-
2005-0083-0041): ``We note that if the compliance period for the 
individual-protection standard extended to the time of peak dose within 
the period of geologic stability (which NAS estimated to be 1 million 
years for the Yucca Mountain site), this [reasonable expectation] test 
would allow for decreasing confidence in the numerical results of the 
performance assessments as the compliance period increases beyond 
10,000 years. For example, this means that the weight of evidence 
necessary, based upon reasonable expectation, for a compliance period 
of 10,000 years would be greater than that required for a compliance 
period of hundreds of thousands of years.'' Given the increased 
uncertainty that is unavoidable in the capabilities of science and 
technology to project and affect outcomes over the next 1 million 
years, the concept of reasonable expectation underlying our standards 
implies that a dose limit for that very long period that is higher than 
the 15 mrem/yr limit that applies in the relatively ``certain'' pre-
10,000-year compliance period could still provide a comparable judgment 
of overall safety. See Section II.C.3 (``What Dose Level is EPA 
Proposing for Peak Dose?'') for a specific discussion of the dose limit 
in today's proposal.
    In formulating an approach to compliance out to the time of peak 
dose, we have established 10,000 years as an indicator for times when 
uncertainties in projecting performance are more manageable and for 
which comparisons can be made with other regulated systems. We realize 
that uncertainties exist within the initial 10,000-year period and that 
10,000 years does not represent a strict dividing point between periods 
over which projections can be made with certainty or not. Clearly, we 
believe that calculations beyond 10,000 years have value, or we would 
not have previously required DOE to include them in its EIS. However, 
we also believe that over the very long periods leading up to the time 
of the peak dose, the uncertainties in projecting climatic and geologic 
conditions become extremely difficult to reliably predict and a 
technical consensus about their effects on projected performance in a 
licensing process would be very difficult, or perhaps impossible, to 
achieve. This is one of the major reasons that the 10,000-year time 
frame was originally selected in the generic standard for land disposal 
of the types of waste intended for the Yucca Mountain repository (40 
CFR part 191) (2001 Response to Comments, p. 7-17, Docket No. OAR-2005-
0083-0043). In such a situation, one might conclude that little or no 
weight should be given to highly uncertain projections as a basis for a 
licensing decision. Conversely, others might conclude that the 
inability to produce highly reliable performance estimates should 
preclude the possibility of licensing at all. Such a conclusion would 
be inconsistent with any concept of permanent disposal, which 
necessarily requires examination of time frames and events that cannot 
be predicted with certainty. We believe that the performance 
projections at Yucca Mountain, if constructed and interpreted 
consistent with the concept of ``reasonable expectation,'' can provide 
useful information on the facility's performance and can form a key 
part of the basis for a licensing decision. Clearly NAS agreed, since 
it recognized that significant uncertainties exist, yet nonetheless 
recommended that projections to peak dose form the basis for EPA's 
standards to be used in judging compliance for licensing the Yucca 
Mountain disposal system. NAS further recognized that an approach akin 
to reasonable expectation is warranted: ``No analysis of compliance 
will ever constitute an absolute proof; the objective instead is a 
reasonable level of confidence in analyses that indicates whether 
limits established by the standard will be exceeded'' (NAS Report p. 
71).

C. How Is EPA Proposing To Revise the Individual-Protection Standard 
(Sec.  197.20) To Address Peak Dose?

    In considering how to revise the individual-protection standard, we 
have sought an approach that would be:
     Responsive to the Court ruling;
     Protective of public health and safety;
     Reflective of the best science and cognizant of the limits 
of long-term projections;
     Implementable by NRC in its licensing process; and
     Limited in scope and focused on aspects critical to 
accomplishing the above goals.
    In balancing these goals, we have carefully examined the NAS 
recommendations and looked more broadly to international models and 
guidance on long-term radioactive waste disposal. We believe today's 
proposal satisfies these goals. We believe the first three are 
straightforward and our reasoning outlined in the next sections will 
clearly show how they influenced our proposal. The fourth point relates 
to an essential purpose of our action that can sometimes be 
overshadowed by emphasis on the NAS recommendations and the Court 
ruling. As NAS stated, ``standards are only useful if it is possible to 
make meaningful assessments of future repository performance with which 
the standards can be compared'' (NAS Report p. 34). Ultimately, NRC 
must be able to use our standards to judge whether DOE has provided 
sufficient evidence that the disposal system will be protective of 
public health and safety. While there are

[[Page 49030]]

significant scientific aspects to this decision, regulatory judgment 
must bridge the gap between what science can show and the unprecedented 
time frames involved. The licensing process must consider the 
confidence that can be placed in performance assessments used to 
represent disposal system evolution and the information necessary to 
make a decision. Our ``reasonable expectation'' standard is critical to 
making this judgment.
    The last point above refers to the legal status of our rule. 
Today's proposal is specifically targeted toward addressing the Court 
ruling regarding the compliance period. Many other aspects of our rule 
were either upheld by the Court or not challenged. As discussed in 
Section II.A, we are not revisiting those issues.
    In a similar vein, when considering potential approaches to address 
the Court's decision, we did not feel constrained by our actions in the 
2001 rulemaking. Nor do we believe that rejecting certain approaches in 
that rulemaking creates a legal barrier to incorporating them into 
today's proposal. Our preferred approach was rejected by the Court in 
favor of a compliance standard applicable at the time of peak dose, 
whenever it might occur within the period of geologic stability. In our 
2001 rulemaking, we considered, discussed, and accepted comment on the 
length of the compliance period, including consideration of the time of 
peak dose. We ultimately chose not to establish a compliance period 
applicable throughout the period of geologic stability. Thus, it is 
difficult to see how we could satisfy the Court's ruling if we were not 
permitted to reconsider or revise our previous conclusions.
1. Multiple Dose Standards Applicable to Different Compliance Periods
    In balancing the considerations described above, the central 
problem is to determine what is achievable in terms of the reliability 
of dose projections. Our task was clearly presented by the Court, and 
our starting position is to fulfill that task by proposing a compliance 
standard at the time of peak dose, whenever it might occur within the 
period of geologic stability. We have discussed at length our concerns 
regarding the quality of very long-term projections and their 
application in a licensing process; even in light of the Court 
decision, those concerns remain. However, we also believe it is clear 
that shorter-term projections do have sufficient reliability to serve 
as the basis for regulatory decision-making. On the one hand, we do not 
want to place more regulatory emphasis on peak dose projections than 
can be justified; on the other, a standard effective at relatively 
short times, where we believe such emphasis is warranted, is unlikely 
on its own to be responsive to the Court ruling. We have sought to 
reconcile these two extremes in order to satisfy all of the goals 
outlined earlier.
    In what we see as the best solution to this difficulty, today we 
are proposing that the individual-protection standard consist of two 
parts, which will apply over different time frames. One part of the 
standard, which will apply over the initial 10,000 years after 
disposal, consists of the 15 mrem/yr individual-protection standard 
promulgated in 2001 as 40 CFR 197.20. The other part other part of the 
standard, which is being proposed today, will apply beyond 10,000 years 
to the time of peak dose up to a limit of 1 million years. We believe 
this approach appropriately recognizes the relative manageability of 
uncertainties at such disparate times, and the resulting level of 
confidence that can be derived from performance projections.
    There is no disagreement internationally that quantitative 
projections are the most direct means of evaluating disposal system 
performance, or that comparison of such projections with an acceptable 
level of performance is a straightforward and transparent method of 
assessing disposal system safety. However, there is also a general 
consensus that reliance on quantitative projections to determine safety 
may be misleading and incomplete, becoming more so at times very far 
into the future. IAEA notes that ``[q]uantitative analysis is 
undertaken, at least over the time period for which regulatory 
compliance is required, but the results from detailed models of safety 
assessment are likely to be more uncertain for time periods in the far 
future'' (DS154, Section 3.48, p. 25, Docket No. OAR-2005-0083-0051). 
Also, ``an indication that calculated doses could exceed the dose 
constraint, in some unlikely circumstances, need not necessarily result 
in the rejection of a safety case * * * In general, when irreducible 
uncertainties make the results of calculations for the safety 
assessment less reliable, then comparisons with dose or risk 
constraints have to be treated with caution'' (DS154, Sections A.7, 
A.8, pp. 36-37, Docket No. OAR-2005-0083-0051). As suggested by the 
discussion of reasonable expectation in Section II.A.4, at longer time 
periods, the quantitative projections should be considered less for 
their strict numerical outcomes and more as one component in a 
qualitative evaluation of the overall safety case.
    In their book ``Principles and Standards for the Disposal of Long-
Lived Radioactive Wastes'' (2003, Docket No. OAR-2005-0083-0061), 
Chapman and McCombie state that ``[a]n approach commonly used is to 
calculate releases, doses or risks out to peak consequences--but to use 
different approaches to judging acceptability in different time frames. 
At far future times (>10 ka) [>10,000 years] * * * calculated doses may 
then be more appropriately compared with less stringent limits than the 
typical limits at shorter times'' (p. 79). They also present the 
concept of ``time-graded containment objectives'' in which the first 
1,000 years or so is characterized by ``total containment of all 
activity in the repository.'' For the ``next one (or a few) hundred 
thousand years * * * doses * * * are below the range of natural 
background radiation.'' Finally, ``after this time * * * there is no 
further containment objective: doses may be envisaged in the range of 
those from natural background radiation.'' (p. 114)
    Different countries have approached this situation in various ways, 
and many national regulations are still evolving. For example, as 
summarized by Chapman and McCombie in Table 5.1 (Docket No. OAR-2005-
0083-0061): Canada at one time limited quantitative compliance to 
10,000 years, to be followed by qualitative evaluation, with special 
attention to the rate of increase in projected risk; Germany takes a 
similar approach in official guidance, but does not specify a time 
frame in regulation; France requires quantitative compliance for 
100,000 years, with the situation becoming ``hypothetical'' afterward; 
Switzerland requires numerical compliance at all times. The Swedish 
draft guidance referred to in Section II.A.5 states that ``[f]or long 
periods of time, thousands of years and even longer, the risk analysis 
should be successively regarded as an illustration of the protective 
capability of the repository assuming certain conditions'' (p. 7, 
Docket No. OAR-2005-0083-0048). We believe the approach proposed today, 
outlined in the paragraphs below, is consistent with that trend.
    First, we are retaining the standard promulgated in 2001 as Sec.  
197.20, which requires that DOE demonstrate a reasonable expectation 
that the RMEI will not incur annual exposures greater than 150 [mu]Sv 
(15 mrem) (expressed as a committed effective dose equivalent) from 
releases of radionuclides from the Yucca Mountain disposal system for 
10,000 years after disposal. DOE will make this demonstration using the 
arithmetic mean of performance

[[Page 49031]]

assessment results (see Section II.C.5, ``How Will NRC Judge 
Compliance?'' for further discussion of the mean). We believe this is 
appropriate, protective, and will maintain consistency with our generic 
standards (now applied to the WIPP) and other precedents described 
earlier. Further, NAS stated that the ``range [of 10-\5\ to 
10-\6\ per year for risk] could therefore be used as a 
reasonable starting point for EPA's rulemaking'' (NAS Report p. 49, 
emphasis in original). By maintaining the 15 mrem/yr standard for 
10,000 years we clearly establish a ``starting point'' for assessing 
compliance that is consistent with both NAS and our overall risk 
management policies, and serves as a logical foundation for us to 
incorporate concerns regarding far future projections.
    Because of the emphasis on peak dose as the key benchmark of safety 
in both the NAS Report and the Court decision, some commenters may 
question not only the need for a standard at such relatively short 
times, but also whether it is legally permissible, given the Court's 
decision. We believe there is ample justification for a separate 
10,000-year standard on both counts. Taking the legal questions first, 
there was no legal challenge and the Court made no ruling on the 
protectiveness of our standard up to 10,000 years. Further, the Court 
ruled that we must address peak dose, but did not state, and we do not 
believe intended, that we could not have additional measures to bolster 
the overall protectiveness of the standard. As the Court noted, the 
EnPA requires that EPA ``establish a set of health and safety 
standards, at least one of which must include an EDE-based, individual 
protection standard'' (NEI, 373 F.3d at 45, Docket No. OAR-2005-0083-
0080), but does not restrict us from issuing additional standards. 
Thus, as long as we issue ``at least one'' standard addressing the NAS 
recommendation regarding peak dose, we are not precluded from issuing 
other, complementary, standards to apply for a different compliance 
period. The Court's concern was whether we had been inconsistent with 
the NAS recommendation by not extending the period of compliance to 
times longer than 10,000 years. NAS itself did not address the idea of 
having separate standards to apply over different time periods. We 
believe such a decision falls well within our policy discretion and in 
that context the 10,000-year standard is analogous to our ground-water 
protection standards.
    An important reason for retaining a standard applicable for the 
first 10,000 years is to address the possibility, however unlikely, 
that significant doses could occur within 10,000 years, even if the 
peak dose occurs significantly later, as DOE currently projects.
    Examination of DOE's Total System Performance Assessments (TSPA) 
for the site shows that the time of peak dose occurs in the hundreds of 
thousands of years (FEIS, DOE/EIS-0250, Appendix I, Section 5.3, 
February 2002, Docket No. OAR-2005-0083-0086). The waste packages 
assessed in the TSPA are heavily engineered to provide corrosion 
resistance under the conditions expected in the repository, and are 
projected to remain essentially unbreached for periods well beyond 
10,000 years. The scientific data that underlie these corrosion 
resistance projections are laboratory tests on the metals, under 
conditions intended to stress the metals and simulate their performance 
in the repository. These testing methods are typical ``state-of-the-
art'' techniques for corrosion testing. However, it must be recognized 
that the extrapolation of laboratory test results in a predictive sense 
involves significant uncertainties, and our experience in verifying 
such projections is only for time frames of decades in the case of 
industrial applications (``Assumptions, Conservatisms, and 
Uncertainties in Yucca Mountain Performance Assessments,'' Section 5, 
July 2005, Docket No. OAR-2005-0083-0085). While DOE projects, based 
upon the results of laboratory testing, that the waste containers will 
maintain their integrity for thousands to tens of thousands of years, 
it is not possible to claim unequivocally that no information will come 
to light that might cause a reassessment of the containers' behavior 
and its effect on disposal system performance. Although we believe that 
significant doses within 10,000 years are highly unlikely, we also 
believe it important to structure our regulations to preclude the 
chance that protection at Yucca Mountain would be less than that 
provided for WIPP or the Greater Confinement Disposal facility (GCD, 
which is a group of 120-feet deep boreholes, located within NTS, which 
contain disposed transuranic wastes). It would be inappropriate to 
apply a standard designed to accommodate the uncertainties in 
projections many tens to hundreds of thousands of years into the future 
to projections within 10,000 years, when uncertainties are much more 
manageable. The 15 mrem/yr dose limit is the measure against which 
compliance would be judged during the initial 10,000-year period.
    In today's action, we are proposing to add a standard of compliance 
that would apply at the time of peak dose, if DOE determines that the 
peak occurs at any time beyond 10,000 years but within 1 million years 
(as recommended by NAS). Specifically, in addition to retaining the 15 
mrem/yr standard applicable up to 10,000 years, we are proposing to 
establish a separate numerical compliance standard against which the 
median of peak dose projections would be compared (see Section II.C.3 
for a discussion of the proposed dose limit and Section II.C.5 for a 
discussion of the arithmetic mean and median). As discussed earlier, we 
recognize that there is strong consensus in the international 
radioactive waste community that dose projections extending for periods 
into the many tens to hundreds of thousands of years can best be viewed 
as qualitative indicators of disposal system performance, rather than 
as firm predictions that can be compared against strict numerical 
criteria. The primary concern, which we have also expressed, is 
managing the uncertainties that become more prominent at longer time 
frames.
    Nevertheless, we believe that the best way to address the Court 
decision is to establish a numerical compliance standard for the time 
of peak dose so that a clear test for compliance decision-making can be 
applied to the results of quantitative performance assessments. What we 
are proposing is unprecedented in our national regulatory schemes, and 
we remain greatly concerned about the ability of the implementing 
agencies to manage the uncertainties in very long-term projections in 
order to make comparisons with a numerical standard meaningful. We 
discuss elsewhere in this document (see Sections II.B and II.D.2, for 
example) ways in which NRC and DOE might temper the effects of 
uncertainty in dose projections, e.g., through the selection of 
parameter distributions or scenarios.
    Some readers may note that we rejected similar approaches offered 
in comments on our 1999 proposed rule. One commenter in particular 
suggested that the dose standard could be increased over time, i.e., 15 
mrem/yr up to 10,000 years, 150 mrem/yr from 10,000 to 100,000 years, 
and 1.5 rem/yr from 100,000 to 1 million years (Docket A-95-12, Item 
IV-D-35). As stated in our Response to Comments document published in 
conjunction with the 2001 final rulemaking (p. 3-8, Docket No. OAR-
2005-0083-0043), we considered that our approach accomplished the same 
goal as that offered by the commenter. While we did state that ``no 
regulatory body that we are aware of considers doses of 150 mrem to be 
acceptable,'' we also stated that ``the

[[Page 49032]]

uncertainties involved in very long-term assessments would make it more 
difficult to judge compliance with any numerical standard,'' which we 
still believe is true. It is clear that we struggled to reconcile the 
competing claims of confidence in projections and intergenerational 
equity. We sought an approach that would account for what we see as 
potentially unmanageable uncertainties, but did not depart from levels 
of risk that are considered protective today. Nevertheless, the Court's 
decision puts us in the position of establishing a quantitative 
standard at the time of peak dose. It is necessary for us to re-
evaluate potential approaches to doing so, including whether and under 
what conditions a higher dose standard can be justified. We discuss an 
approach similar to that offered by the commenter in Section II.C.2.c 
(``Peak Dose Standard Varying Over Time'').
    We are not requesting comment on the 15 mrem/yr standard or its 
applicability for the initial 10,000-year period. The public record 
reflects an exhaustive level of comment and consideration on these 
points (see our 1999 proposed and 2001 final rulemakings, as well as 
Sections 3 and 4 of the 2001 Response to Comments Document (Docket Nos. 
OAR-2005-0083-0041, 0042, 0043, respectively). The Court did not 
question the scientific basis of the 15 mrem/yr dose standard, the 
protective nature of that limit, or its well-established precedents in 
regulation for periods as long as 10,000 years (including its 
implementation at WIPP and GCD), nor indeed were any of these aspects 
of the rule challenged. Further, as noted above, the Court did not rule 
that the 10,000-year compliance period had no value, only that it was 
not by itself consistent with the NAS recommendation (``We will thus 
vacate part 197 to the extent that it requires DOE to show compliance 
for only 10,000 years following disposal,'' NEI, 373 F.3d at 31, Docket 
No. OAR-2005-0083-0080).
    We are requesting comment on the combination of the 15 mrem/yr 
standard with a separate standard applicable beyond 10,000 years 
through the period of geologic stability. We believe we have provided a 
rational basis for taking this approach and that it is consistent with 
the Court's position that we could have ``taken the Academy's 
recommendations into account and then tailored a standard that 
accommodated the agency's policy concerns.'' NEI, 373 F.3d at 26, 
Docket No. OAR-2005-0083-0080.
2. What Other Options Did EPA Consider?
    We considered a number of other approaches to respond to the 
Court's decision, each of which had attractive qualities, as well as 
disadvantages. These disadvantages generally relate to the difficulty 
of implementation given the increasing complexity and uncertainty of 
much longer-term projections.

a. Maintain the 10,000-Year Standard Alone Without Addressing Peak Dose

    The Court suggested that, ``[h]ad EPA begun with the NAS 
recommendation to base the compliance period on peak dosage and then 
made adjustments to accommodate policy considerations not considered by 
NAS,'' the 40 CFR part 197 standards issued in 2001 might have been 
accorded more deference. NEI, 373 F.3d at 31, Docket No. OAR-2005-0083-
0080. However, it is not clear how EPA's earlier explanation of its 
policy concerns might be reconciled with NAS's technical 
recommendation. In view of this, we believe that the most direct and 
responsive action to address the Court ruling is to revise our 
standards to include consideration of the time when peak dose occurs. 
Therefore, although we are retaining the previous 10,000-year 
provisions as one component of our revised standards, we are also 
proposing an additional measure to address the time of peak exposure 
within the period of geologic stability beyond 10,000 years. We believe 
that this approach, coupled with the selection of the dose standard to 
apply at the time of peak dose (see Section II.C.3) and specification 
of certain aspects of DOE's performance assessment (see Section II.D), 
will adequately address our policy concerns.

b. Dose Standard To Apply at Peak Dose Alone

    The second option we considered is simply to replace the 10,000-
year standard with one that applies at the time of peak dose, whenever 
it might occur. This approach is attractive primarily because it would 
be straightforward in responding to the Court decision. Although we 
believe that 10,000 years has value as a precedent for safety 
assessments, and are retaining that element of the standards, it is not 
intrinsically significant as a demarcation point for addressing a peak 
dose standard beyond 10,000 years. A peak dose standard alone (i.e., 
not in conjunction with the 10,000-year standard we are retaining) 
would remove confusion on that point, but introduces additional 
difficulties, as described in the following sections.
    As discussed in Section II.C.4.a, we do not believe it is 
reasonable or justifiable simply to extend the application of a 15 
mrem/year dose limit over the entire period up to the time of peak 
dose. Rather, at the time of peak dose, which could potentially occur 
hundreds of thousands of years into the future, we believe rising 
uncertainties justify adopting a different (higher) dose level. 
However, as discussed in Section II.C.3, this approach, while more 
cognizant of the effect of uncertainties and the dangers of relying on 
specific numerical indicators at very long times, departs from our 
previous standards of protectiveness in the event that peak doses occur 
within relatively short time periods. Specifically, if peak doses occur 
within 10,000 years, we would be in the position of measuring safety 
against a dose level that we have explicitly rejected as not 
sufficiently protective over that time frame, both in our generic 
standards and in our earlier Yucca Mountain rulemaking. Further, there 
would be a clear contrast between the level of protection offered to 
the population in the vicinity of the WIPP and that offered the 
population affected by Yucca Mountain. We recognize that our insistence 
on maintaining a 15 mrem/yr standard over the initial 10,000 years 
might appear inconsistent with our proposal, which could allow peak 
doses shortly after 10,000 years at levels well above 15 mrem. However, 
as discussed previously, we believe NRC has the authority, as part of 
its licensing process, to consider the timing and magnitude of peak 
dose in assessing the safety of Yucca Mountain. Furthermore, we do not 
believe it is prudent to disregard the usefulness of a stringent 
10,000-year measure simply because uncertainties at longer time frames 
make it infeasible to conduct a performance assessment with the same 
level of rigor. Our view on this point is discussed in Section II.A.1.

c. Peak Dose Standard Varying Over Time

    We also considered a variation on our proposed approach, in which 
the post-10,000-year dose level would rise incrementally as time and 
the effects of uncertainty increase. This approach would provide 
greater continuity with the 10,000-year standard and a gradual 
transition as the role of uncertainty increases. The difficulty in this 
approach is identifying criteria to define the timing and level of 
these transitions, which would have to incorporate some appraisal and 
comparison of the effects

[[Page 49033]]

of uncertainty at various times. Some of the advantages of this 
approach are also captured by the statistical approach discussed in 
Section II.C.2.e. We have not identified a defensible way to derive 
transition levels or the times at which these dose level changes could 
be made.

d. Standard Expressed as a Dose Target, Rather Than Limit

    Although we have chosen to add a standard extending the compliance 
period beyond 10,000 years, we believe that the most problematic aspect 
of doing so is the uncertainty involved in making projections over such 
long time frames, which we discussed in some detail in our proposed and 
final rulemakings for 40 CFR part 197 in 1999 and 2001, respectively 
(Docket Nos. OAR-2005-0083-0041 and 0042). To repeat, we are in 
agreement with NAS that such projections can be performed and even 
``bounded'' to some extent. However, we remain concerned about whether 
and under what conditions results of very long-term assessments can 
have sufficient meaning to provide the basis for a licensing decision 
that the repository should or should not be approved.
    One way to take these uncertainties into account is to establish a 
more flexible compliance benchmark for very long time periods, one that 
would represent a more qualitative ``target'' for dose assessments 
rather than a strict numerical limit. This approach would be generally 
consistent with several international programs. For example, the 
Swedish Radiation Protection Authority (SSI) has proposed draft 
guidance for the disposal of SNF, stating that ``[f]or very long 
periods * * * [t]he intention should be to shed light on the protective 
capability of the repository and to provide a qualitative picture of 
the risks'' (p. 7, Docket No. OAR-2005-0083-0048). The Swedish 
regulations themselves are not detailed regarding the way different 
time periods should be addressed, although it is clear that times 
beyond 1,000 years are seen differently than the period up to 1,000 
years. For the first thousand years after closure, ``the assessment of 
the repository's protective capability shall be based on quantitative 
analyses of the impact on human health and the environment,'' but for 
longer periods that assessment ``shall be based on various possible 
sequences for the development of the repository's properties, its 
environment and the biosphere'' (Sections 11 and 12, respectively, 
Docket No. OAR-2005-0083-0047).
    In some cases, this reasoning is also applied to near-surface 
disposal facilities involving much shorter time frames. For example, in 
the United Kingdom, ``[t]he Government therefore considers it 
inappropriate to rely on a specified risk limit or risk constraint as 
an acceptance criterion for a disposal facility after control is 
withdrawn. It is, however, considered appropriate to apply a risk 
target in the design process. However, if the estimated risk is above 
the target, the Agency will need to be satisfied not only that an 
appropriate level of safety is assured, but also that any further 
improvements in safety could be achieved only at disproportionate cost 
* * * In the very long term, irreducible uncertainties about the 
geological, climatic and resulting geomorphological changes that may 
occur at a site provide a natural limit to the timescale over which it 
is sensible to attempt to make detailed calculations of disposal system 
performance. Simpler scoping calculations and qualitative information 
may be required to indicate the continuing safety of the facility at 
longer times'' (UK Environment Agencies, ``Disposal Facilities on Land 
for Low and Intermediate Level Radioactive Waste: Guidance on 
Requirements for Authorisation,'' sections 6.14 and 8.23, Docket No. 
OAR-2005-0083-0063). Thus, in the UK approach, estimated risks may be 
allowed to exceed the numerical target if it is determined that further 
restrictions in risk are impossible or impractical.
    Our approach in the 2001 rulemaking, which required peak dose 
projections to be placed in the Environmental Impact Statement, was 
based on similar reasoning. It allowed NRC to evaluate those results 
qualitatively, but did not prescribe that they be compared against a 
dose limit. We also believe such an approach would be consistent with 
our ``reasonable expectation'' standard, which intends to avoid a 
narrow focus on numerical calculations and encourages consideration of 
the totality of the assessment in the context of the overall safety 
case (ICRP took the same view in its Publication 81, ``Radiation 
Protection Recommendations as Applied to the Disposal of Long-Lived 
Solid Radioactive Waste,'' stating that ``as the time frame increases, 
some allowance should be made for assessed dose or risk exceeding the 
dose or risk constraint. This must not be misinterpreted as a reduction 
in the protection of future generations and, hence, a contradiction 
with the principle of equity of protection, but rather as an adequate 
consideration of the uncertainties associated with the calculated 
results'' (Docket No. OAR-2005-0083-0087); similarly, IAEA states 
``that calculated doses are less than the dose constraint is not in 
itself sufficient for acceptance of a safety case * * * Conversely, an 
indication that calculated doses could exceed the dose constraint * * * 
need not necessarily result in the rejection of a safety case,'' DS154, 
Section A.7, pp. 36-37, Docket No. OAR-2005-0083-0051). In considering 
how to address peak dose in this standard, however, we believe it is 
more implementable and will be viewed as more rigorous to set a 
specific dose limit and provide direction concerning assumptions and 
methodologies for peak dose calculations, and leave it to NRC to 
consider the quantitative projections of peak dose as a particularly 
important part of the ``full record before it'' that it will consider 
in determining whether there is a reasonable expectation that the dose 
limit will be achieved.

e. Standard Expressed as a Statistical Distribution

    Finally, we considered a standard of compliance that would combine 
features of the qualitative and quantitative approaches described 
earlier. Rather than incorporating a specific numerical limit that must 
be met by a single compliance measure (such as the median or arithmetic 
mean of a distribution), this approach would be based upon the 
characteristics of the distribution itself. It would take into account 
the range of results for performance assessment by examining multiple 
representative dose estimates such as upper and lower percentile 
values. Under this formulation, DOE might have to show that some 
percentage of the peak dose projections would remain within a certain 
range of a reference dose level. For example, this standard might say 
that at least 10% of peak annual dose results must be 15 mrem or lower, 
and that no more than 10% of results can exceed some upper limit. Using 
these parameters and assuming that DOE ran 100 assessments of system 
performance using probabilistically-sampled input parameter values, 
each resulting in a separately calculated ``peak'' dose, at least ten 
of those results would have to be 15 mrem or lower and no more than ten 
could be above the ``upper limit''.
    This approach seems to address some of our concerns. First, it 
recognizes growing uncertainties but constrains how much is acceptable 
by specifying characteristics of the distribution that must apply at 
all times without being overly affected by ``outliers.'' In fact, the 
value of the projected peak dose is considered only in determining 
where it falls in relation to the designated upper

[[Page 49034]]

and lower percentile measures. In this example, no more than 10% of the 
results may exceed the ``upper limit'', but the amount by which they 
exceed that limit is not taken into account (and similarly for doses 
below 15 mrem/yr). Thus, projected doses of 1 rem/yr (1,000 mrem/yr) 
would carry the same significance as much lower projected doses, as 
long as both were higher than the ``upper limit''. As a result, this 
approach might provide additional flexibility in judging the level of 
conservatism appropriate to addressing uncertainties (and perhaps 
compensate for conservatism) across a range of scenarios because the 
results would not be disproportionately affected by low-probability 
scenarios resulting in very high doses, as the arithmetic mean would 
be. In addition, the lower dose threshold acts as a conservative 
performance requirement in that it requires that the disposal system 
provide a specified level of performance tied to the 15 mrem/yr dose 
standard applicable to performance up to 10,000 years.
    A firm base of assessments at lower levels (e.g., 15 mrem/yr) would 
tie DOE's results to, and provide continuity with, the 10,000-year 
projections. It could be reasonable to allow a small number of results 
to exceed the ``upper limit,'' so long as the ``expected'' performance 
remains within a given range (within about an order of magnitude of 15 
mrem, if we were to use as the ``upper limit'' the value of 350 mrem/yr 
we are proposing today). It should be kept in mind that even using the 
mean of the distribution as the compliance measure allows for a 
percentage of results to exceed the limit, depending to some extent on 
how the distribution is skewed; this statistical approach offered for 
discussion is simply more precise in specifying the percentage.
    Second, while accounting for uncertainties, it can be linked to the 
standards of safety established for geologic repositories at earlier 
time frames. Percentile curves could be compared against reference 
levels based upon well-established limits within the U.S. and 
internationally, such as 15 mrem/yr, 25 mrem/yr, 30 mrem/yr, or 100 
mrem/yr, or the 350 mrem/yr we are proposing today. This could provide 
continuity with our approach at 10,000 years. It is reasonable to 
assume that uncertainties will tend to become less manageable as time 
increases, but there is no clear and predictable demarcation for when 
uncertainties become ``unmanageable.''
    Third, this approach would be consistent with our ``reasonable 
expectation'' standard, which is intended to encourage DOE to focus on 
``cautious, but reasonable'' scenarios and examine the full range of 
results to obtain the best possible understanding of the long-term 
behavior of the disposal system. In applying a standard that must 
address times from 10,000 years up to 1 million years, it might be more 
representative of system behavior to consider the entire distribution 
of results that may occur over those times than to focus on a single 
number as indicative of acceptable performance. Using this approach, 
NRC would be assured that the bulk of the results will fall within 
reasonable limits, may be better able to understand why results fall at 
certain points along the continuum, and would have additional 
flexibility to determine compliance within those limits.
    We used a somewhat similar approach in developing the containment 
requirements in 40 CFR 191.13(a). In that section of our generic 
regulations, we required that calculations show that a disposal system 
have no more than one chance in ten of exceeding the release limits, 
and no more than one chance in 1,000 of exceeding ten times the release 
limits. In establishing those requirements, we explained that the 
release limits applied to ``those processes that are expected to occur 
as well as relatively likely disruptions.'' The release limits 
multiplied by ten applied to ``more likely natural disruptive events * 
* * [and the] range of probabilities was selected to include the 
anticipated uncertainties in predicting the likelihood of these natural 
phenomena. Greater releases are allowed for these circumstances because 
they are so unlikely to occur.'' In part 191, no release limits were 
applied to even lower-probability (i.e., ``very unlikely'') events, 
analogous to our approach of screening out very unlikely events at 
Yucca Mountain: ``the Agency believes there is no benefit to public 
health or the environment from trying to regulate the consequences of 
such very unlikely events' (50 FR 38071, September 19, 1985, Docket No. 
OAR-2005-0083-0064). We have successfully implemented this regulation 
at WIPP.
    While we see several potential positive aspects of this statistical 
approach, we also have concerns regarding both the overall approach and 
the ways in which it could give a misleading impression of disposal 
system performance in a compliance demonstration. First, there is a 
difficulty in defining exactly where percentile limits should be placed 
and how they should be justified. Second, while the criteria we have 
suggested would apply to the entire distribution of results, they would 
essentially give the ``tails'' of the distribution a strong role in 
determining whether the disposal system should be licensed. As we 
discuss later in Section II.C.5 (``How Will NRC Judge Compliance?''), 
we believe it is appropriate to consider an indicator of the ``central 
tendency'' of the results as demonstrative of performance.
    Our second concern relates to the idea that the calculated peak 
dose values themselves are not explicitly incorporated into the 
compliance determination through calculation of a separate statistical 
measure, such as the mean. While this offers an advantage insofar as 
the overall measure is not overly influenced by very high results, for 
any defined set of cut-offs there is always the possibility that the 
distribution will fall just outside the acceptable criteria. While 
strictly speaking only the number of doses above the higher cut-off 
level enters into the compliance demonstration, the magnitude of those 
doses would also be important in the regulator's confidence in the 
overall acceptability of the disposal system. Similarly, a distribution 
that falls just outside the cut-offs could be judged ``better'' than a 
distribution that meets the criteria, if a different measure such as 
the mean or median were used for comparison. In considering a series of 
100 realizations, for example, a distribution with 11 above, but only 
slightly above, the ``upper limit'' and only nine at 15 mrem/yr or 
lower (but with the next highest at only 16 mrem) would fail the test, 
even if the bulk of the results were relatively low (say, below 100 
mrem). However, a distribution with ten realizations significantly 
higher than the ``upper limit'' (e.g., 500 mrem/yr and higher), ten at 
15 mrem/yr, and most of the remaining doses well above 100 mrem/yr, 
would pass the test, even though it is likely that the arithmetic mean 
would be noticeably higher in the second case. Such a disparity might 
also indicate the presence of high-dose scenarios in one distribution 
that were not included in the other.
    Therefore, we have chosen not to propose this approach for Yucca 
Mountain. We are concerned that it will be less transparent to the 
public and not give a clear indication of the necessary level of 
performance. Further, upper and lower percentiles and dose limits must 
be selected, as in the example above; the selection of all these values 
would need to account for risk management and policy considerations. It 
is difficult to identify a specific set of criteria that would lead to 
the selection of one set of values over another.

[[Page 49035]]

3. What Dose Level is EPA Proposing for Peak Dose?
    Having determined that it would be appropriate to propose a 
numerical peak dose standard for the period of geologic stability 
beyond 10,000 years, we must then determine the appropriate level for 
that standard. We considered several factors in selecting the level 
proposed today. First, and most significant, is the issue of 
uncertainty in long-term projections. Uncertainties are problematic not 
only because they are challenging to quantify, but also because their 
impact will differ depending on initial assumptions and the time at 
which peak dose is projected to occur. Further, the natural tendency in 
modeling long-term processes is to introduce additional conservatisms 
to help ensure that actual performance will be no worse than projected 
performance. Thus, excessive conservatism in addressing uncertainty 
drives assessments away from ``cautious, but reasonable'' assumptions 
and may result in an unrealistic, overly pessimistic view of disposal 
system performance. As we stated in our earlier rulemaking, ``[s]etting 
a strict numerical standard at a level of risk acceptable today would 
ignore this cumulative uncertainty and the extreme difficulty of using 
highly uncertain assessment results to determine compliance with that 
standard'' (66 FR 32098, June 13, 2001, Docket No. OAR-2005-0083-0042).
    This raises a broader point regarding the significance of very-long 
term projections and how they should be considered in the context of 
repository safety. Leaving aside the uncertainties inherent in 
projecting geologic characteristics over such periods, even a well-
characterized site will display natural variability in the parameters 
that influence radionuclide transport. This natural variability exists 
at every possible site and can be reduced (or at least better 
estimated) by site characterization, but can never be eliminated, no 
matter how stable the site. As assessments extend to longer time 
periods, this natural variability will lead to an increasing spread of 
results even if conditions do not change significantly (it may be 
useful again for the reader to refer to the hurricane analogy discussed 
in Section II.A.5, where the range of possible storm paths increases as 
forecasts look farther ahead in time). Therefore, given the difference 
in the level of confidence regarding the ``real'' performance of the 
disposal system for projections at 250,000 years as at 10,000 years, we 
believe that emphasizing incremental dose increases when such increases 
are overwhelmed by fundamental uncertainties inappropriately takes 
attention away from an evaluation of the overall safety of the disposal 
system and its ability to contain and isolate wastes or respond to 
disturbances. On that point, we have argued against viewing projections 
as ``predictions'' of disposal system performance and have emphasized 
that assessments should aim to provide a ``reasonable expectation'' 
that performance will be within acceptable limits (on this point, see 
the NAS Report, for example p. 71: ``The results of compliance analysis 
should not, however, be interpreted as accurate predictions of the 
expected behavior of a geologic repository''). While there is a body of 
experience in applying the ``reasonable expectation'' concept for 
10,000 years, we are also considering its implications for time periods 
in the hundreds of thousands of years (see Section II.B, ``How Does the 
Application of ``Reasonable Expectation'' Influence Today's 
Proposal?'').
    We have also considered the potential impacts to future generations 
that would be represented by a dose standard applied to periods up to 1 
million years. Impacts on future generations could come in the form of 
economic cost, health impacts, or a reduction in the options available 
to make decisions to address the problems faced by those generations. A 
number of regulatory and scientific bodies suggest that it is 
appropriate to relate longer-term standards to background radiation 
levels. NEA, for example, suggests that consideration of future 
generations ``implies that the safety implications of a repository need 
to be assessed for as long as the waste presents a hazard'' but that 
such assessments need not focus on exposures: ``In view of the way in 
which uncertainties generally increase with time, or simply for 
practical reasons, some cut-off time is inevitably applied to 
calculations of dose or risk. There is, however, generally no cut-off 
time for the period to be addressed in some way in safety assessment, 
which is seen as a wider activity involving the development of a range 
of arguments for safety'' (``The Handling of Timescales in Assessing 
Post-Closure Safety,'' p. 39, 2004, Docket No. OAR-2005-0083-0046, 
emphasis in original). This reasoning supports the idea that dose 
projections should be given progressively less weight in the overall 
decision as time passes. We note that ICRP recently discussed a similar 
concept. Specifically, ICRP suggests that future projected doses can be 
weighted to take into account a variety of factors, and that 
``[w]eights can also be assigned according to the time at which the 
exposure will occur'' (``The Optimisation of Radiological Protection,'' 
draft for consultation, p. 29, April 2005, Docket No. OAR-2005-0083-
0052). Such an approach could involve giving doses in the far future 
less weight, either in a numeric sense or in the context of the overall 
safety case.
    The National Academy of Public Administration (NAPA), in its 1997 
report ``Deciding for the Future: Balancing Risks, Costs, and Benefits 
Fairly Across Generations'' (Docket No. OAR-2005-0083-0087), recognizes 
that each generation must consider not only how its actions will affect 
future generations, but also the extent to which inaction will 
compromise its own interests and negatively affect those same future 
generations.
    To inform decision-making, NAPA defined four basic principles:
     Trustee: Every generation has obligations as trustee to 
protect the interests of future generations;
     Sustainability: No generation should deprive future 
generations of the opportunity for a quality of life comparable to its 
own;
     Chain of Obligation: Each generation's primary obligation 
is to provide for the needs of the living and succeeding generations. 
Near-term concrete hazards have priority over long-term hypothetical 
hazards;
     Precautionary: Actions that pose a realistic threat of 
irreversible harm or catastrophic consequences should not be pursued 
unless there is some countervailing need to benefit either current or 
future generations.
    Under NAPA's approach, there is no absolute freedom of succeeding 
generations to escape the effect of the preceding generations' 
decisions. Rather, it is the responsibility of each generation to 
consider those decisions and their consequences in the light of new 
knowledge, technology, societal attitudes, and economic or other 
factors. NAPA terms this the ``rolling present.'' As it relates to the 
management of spent nuclear fuel, there is no question that the next 
several generations may incur societal as well as economic costs, 
whether it involves continued development of the Yucca Mountain 
repository, development of interim storage facilities or expanded 
storage at reactor sites, or decisions regarding the future use of 
nuclear power. Application of the NAPA principles would lead each 
generation to an approach that would best address the problem without 
unduly limiting the options available to succeeding generations to 
modify that approach or

[[Page 49036]]

to take other actions to address their needs.
    In general, while there is wide agreement that future generations 
should not be unduly compromised by the decisions of the current 
generation, there is no clear consensus regarding the extent of the 
claims held by future generations on the current generation (i.e., how 
many generations should be considered, how to compare their interests 
to those of the current generation, or what it means to ``compromise'' 
their ability to take action). The Swedish National Council for Nuclear 
Waste (KASAM) concludes that increasing uncertainties ``means that our 
capacity to assume responsibilities changes with time. In other words, 
our moral responsibility diminishes on a sliding scale over the course 
of time'' (Nuclear Waste State-of-the-Art Reports 1998, p. 27, Docket 
No. OAR-2005-0083-0056). KASAM suggests that for the next 5 or 6 
generations (roughly 150 years), we can apply a ``Strong Principle of 
Justice'' so that these generations can be expected to achieve a 
quality of life equivalent to that of the current generation. For a 
further 5 or 6 generations, we may only be able to apply a ``Weak 
Principle of Justice'' to ensure that these generations can at least 
satisfy their basic needs. Beyond that point, the best we can do is 
conduct ourselves today so as not to jeopardize future generations' 
possibilities for life (the ``Minimal Principle of Justice''). In the 
case of spent fuel disposal, these considerations lead to the idea that 
a repository must provide reasonable protection and security for the 
very far future, but this may not necessarily be at levels deemed 
protective (and controllable) for the current or succeeding 
generations.\2\
---------------------------------------------------------------------------

    \2\ This sentiment, however, is not universal. Chapman and 
McCombie point out that the Swiss radiation protection regulations 
make the argument ``that since the current generation is the 
beneficiary of nuclear power future doses should be less'' (p. 53). 
They then acknowledge, however, that such arguments are complex, 
noting that ``it has been pointed out that future generations do 
indeed benefit from nuclear technology through the technical 
advances made, the conservation of fossil reserves, the reduction in 
greenhouse gases, etc.'' Further, they go on to write:
    In addition, the inability to guarantee long-term or effectively 
permanent institutional control over long-lived uranium mining 
wastes disposed of at the earth's surface or over historical 
``legacy wastes'' in countries where defence programmes have 
resulted in large-scale contamination, means that we are implicitly 
accepting (for this type of waste, and some NORM wastes) that future 
generations may have lower levels of protection than today. This is 
causing re-examination of the appropriate balance of radiological 
protection standards for the future for these materials. The most 
commonly accepted principle today for disposal of nuclear fuel cycle 
wastes is that future generations must be protected for very long 
times (at least 10,000 years) to at least reach the level of 
protection expected by today's generations; for extremely long times 
the growing tendency is to then make comparisons with natural 
sources of radiation, such as ore bodies.
    ``Principles and Standards for the Disposal of Long-Lived 
Radioactive Wastes,'' pp. 53-54, 2003, Docket No. OAR-2005-0083-
0061.
---------------------------------------------------------------------------

    In any case, it is clear that quantitative regulatory limits cannot 
be applied indefinitely. There is general agreement that assessments 
(and corresponding regulatory safety limits or reference points) for 
periods longer than 1 million years are of limited value in any case 
(e.g., IAEA states that ``little credibility can be attached to 
assessments beyond 106 years. Even qualitative assessments 
will contribute little to the decision making process'' (``Safety 
Indicators in Different Time Frames for the Safety Assessment of 
Underground Radioactive Waste Repositories,'' IAEA-TECDOC-767, p. 19, 
1994, Docket No. OAR-2005-0083-0044), and Sweden's draft guidance 
states that ``[n]o account need be given for periods beyond a million 
years after closure, even if'' peak exposures would be expected after 
that time (p. 7, Docket OAR-2005-0083-0048).
    In addition to examining international guidance and precedents, we 
also reviewed the NAS's statements on the subject. As discussed in 
detail later in this section, NAS refrained from recommending any 
specific dose or risk limit for regulations, but instead suggested a 
range of risks as a ``starting point'' for EPA's consideration. 
Further, while NAS stated that a standard that ``could * * * apply 
uniformly over time and generations * * * would be consistent with the 
principle of intergenerational equity,'' it also recognized that other 
approaches are possible: ``Whether to adopt this or some other 
expression of the principle of intergenerational equity is a matter for 
social judgment'' (NAS Report pp. 56-57).
    In determining an appropriate level of protection for periods up to 
1 million years, we believe it is appropriate to consider potential 
exposures from the Yucca Mountain disposal system in the context of 
exposures incurred by residents of other areas of the United States 
from natural sources. Specifically, we believe it is reasonable to set 
a standard that would represent a level of incremental radiation 
exposure such that the total annual exposure of the RMEI could be 
comparable to the total natural radiation exposures incurred now by 
current residents of well-populated areas. Given the large 
uncertainties surrounding the outcomes at these unprecedented time 
frames, we believe such an action is justifiable and protective. Using 
this approach, we are proposing to establish a standard of 350 mrem 
(3.5 mSv) per year, which will limit total radiation exposures of the 
RMEI to levels comparable to those incurred today from natural sources 
by residents of a nearby western State.
    We believe this level of protection appropriately blends the 
concerns outlined above with current and historical thinking regarding 
the acceptability of risks associated with background radiation, while 
recognizing the conceptual difficulties inherent in regulating at times 
potentially hundreds of thousands of years into the future. NAS 
recognized that the level of protection was a matter best left to EPA 
to establish through rulemaking: ``We do not directly recommend a level 
of acceptable risk'' (NAS Report p. 49). Thus, the NAS Report does not 
bind us to apply any particular dose limit in our Yucca Mountain 
standards.
    We note that a number of international scientific and regulatory 
bodies and programs suggest natural sources of radioactivity serve as a 
point of comparison when uncertainties become significant. For example, 
the IAEA has stated that, for time frames extending from about 10,000 
to 1 million years, ``it may be appropriate to use quantitative and 
qualitative assessments based on comparisons with natural radioactivity 
and naturally occurring toxic substances'' (``Safety Indicators in 
Different Time Frames for the Safety Assessment of Underground 
Radioactive Waste Repositories,'' IAEA-TECDOC-767, p. 19, 1994, Docket 
No. OAR-2005-0083-0044). IAEA also suggests that ``[i]n very long time 
frames * * * uncertainties could become much larger and calculated 
doses may exceed the dose constraint. Comparison of the doses with 
doses from naturally occurring radionuclides may provide a useful 
indication of the significance of such cases'' (``Geological Disposal 
of Radioactive Waste,'' DS154, Section A.7, p. 37, April 2005, Docket 
No. OAR-2005-0083-0051). Similarly, in summarizing the results of a 
workshop to assess long-term assessments, the NEA suggests that at time 
frames when the ``system [is] responding to external change,'' a key 
performance indicator could be ``comparison with background radiation 
levels.'' At that workshop, the idea was presented that up to 100,000 
years, ``a dose constraint derived from natural background levels is 
prescribed'' and beyond that point ``the eventual redistribution of the 
residual activity by natural processes remains indistinguishable from 
natural regional variations in radiation levels'' (``The

[[Page 49037]]

Handling of Timescales in Assessing Post-Closure Safety: Lessons Learnt 
from the April 2002 Workshop in Paris, France,'' pp. 33, 35, 2004, 
Docket No. OAR-2005-0083-0046). Further, as regards low- and 
intermediate-level waste disposal, the UK Environment Agencies 
(consisting of the Environment Agency of England and Wales, the 
Scottish Environment Protection Agency, and the Department of the 
Environment for Northern Ireland) state that ``At times longer than 
those for which the conditions of the engineered and geological 
barriers can be modelled or reasonably assumed * * * Comparisons with 
the ambient levels of radioactivity in the environment may also be 
appropriate'' (``Disposal Facilities on Land for Low and Intermediate 
Level Radioactive Wastes: Guidance on Requirements for Authorisation,'' 
section 6.22, 1996, Docket No. OAR-2005-0083-0063).
    We therefore considered which natural sources of radioactivity in 
the United States might provide similar reference points for a dose 
standard beyond 10,000 years. Natural background radiation in the U.S. 
averages roughly 300 mrem/yr, but varies significantly across the 
country, from a low of about 100 mrem/yr in coastal areas to above 1 
rem/yr (1,000 mrem/yr) in certain localized regions. For purposes of 
this discussion, natural background radiation consists of external 
exposures from cosmic and terrestrial sources, and internal exposures 
from indoor exposures to naturally-occurring radon. Altitude and 
geology are two of the primary variables accounting for regional 
variations; however, there can be tremendous fluctuation even within a 
city or county, primarily due to variations in radon emissions. These 
fluctuations introduce some uncertainty in estimates of localized 
background radiation levels, which are also affected by factors such as 
the number and distribution of samples within a geographic area, 
whether the samples are short-term or averaged over a longer period, 
the structure of the building, the location of the sampling point(s) 
within a building, and assumptions in translating measured 
concentrations to estimated doses.
    In order to assess total exposures and derive a dose limit, it is 
necessary to establish levels of natural background radiation already 
experienced in the vicinity of Yucca Mountain. We selected Amargosa 
Valley as the point of comparison for this analysis. We believe this is 
an appropriate approach, as the RMEI is defined as having a lifestyle 
and diet representative of current residents of Amargosa Valley. It is 
reasonable to consider total exposures in light of exposures already 
incurred by people in the immediate vicinity of Yucca Mountain. 
However, there are varying estimates of exposures from natural 
background sources in that area. DOE estimates that the natural 
background in Amargosa Valley is equivalent to the average across the 
U.S., or 300 mrem/yr (FEIS, DOE/EIS-0250, Table 3-28, Docket No. OAR-
2005-0083-0086). However, that overall figure is highly dependent on 
the radon contribution, which DOE also assumes is equivalent to the 
average across the U.S., or 200 mrem/yr. Based on EPA radon studies, we 
believe it is reasonable and somewhat conservative to assume that radon 
exposures to residents of Amargosa Valley would be slightly higher (say 
25%) than the national average (and possibly as much as 100 mrem/yr 
higher than the statewide average), resulting in a radon contribution 
to those residents of about 250 mrem/yr. Thus, combined with the cosmic 
and terrestrial exposures estimated by DOE, we estimate total annual 
natural background radiation at Amargosa Valley to be approximately 350 
mrem/yr.\3\
---------------------------------------------------------------------------

    \3\ Data from EPA studies in 1993 indicate that the total 
average natural background exposure in the State of Nevada is 222 
mrem/yr (``Assessment of Variations in Radiation Exposure in the 
United States,'' 2005, Docket No. OAR-2005-0083-0077), which is 
roughly 75% of the national average. Because data were not available 
specifically for Amargosa Valley, we used the statewide average as a 
starting point to estimate background radiation at Amargosa Valley. 
The overall statewide average is significantly affected by estimated 
exposures in Clark County (where Las Vegas is located), and not 
necessarily representative of exposures closer to Yucca Mountain. 
Clark County accounts for roughly two-thirds of the state's 
population (Census Bureau, Nevada State Data Center, http://dmla.clan.lib.nv.us/docs/nsla/sdc/). As outlined above, data support 
the conclusion that average exposures in Clark County would be 
significantly lower than in the rest of the state, primarily because 
of indoor radon exposures. EPA's map of radon zones developed in the 
early 1990s found Clark County to be the only county in Nevada 
placed into the lowest emission category, in which average exposure 
potential is less than 200 mrem/yr (``EPA Map of Radon Zones,'' EPA-
402-R-93-071, Docket No. OAR-2005-0083-0065). Most of the other 
counties, including Nye County (where Yucca Mountain and Amargosa 
Valley are located), fell into the intermediate category, in which 
average exposure potential is estimated in the range between 200 and 
400 mrem/yr.
---------------------------------------------------------------------------

    To make the comparison with total exposures, it is also necessary 
to consider what total exposures provide a reasonable reference point 
for limiting releases from Yucca Mountain. As noted above, our goal is 
to ensure that releases from Yucca Mountain will not cause total 
exposures to the RMEI to exceed natural background levels with which 
other populations live routinely. We selected the State of Colorado as 
the reference point in meeting this goal. We considered several factors 
in this selection. First, we must recognize that some incremental 
exposure will be allowed; that is, it is a foregone conclusion that 
even the most protective standard cannot be expected to reduce natural 
background exposures, and clearly we cannot establish a negative 
standard. Thus, the reference point would have to have a higher level 
of background than does the area near Yucca Mountain. In addition, 
because of the aforementioned complications in estimating localized 
background radiation (due primarily to the radon component), we chose 
to examine statewide averages, which are less uncertain. Of the states 
with sufficient data, 32 have average background radiation levels 
higher than Nevada. In selecting among these, we considered 
characteristics such as geographic location and population. Our 
preference is to choose a state in the western part of the country that 
is fairly well-populated and might otherwise have characteristics 
considered reasonably comparable to Nevada (such as radon potential, 
surface water/coastal features, or size of major cities). We find that 
Colorado best fits those criteria. According to the population data 
(U.S. Census Bureau Statistical Abstract of the United States, July 1, 
2004, http://www.census.gov/statab/ranks/rank01.html), Colorado ranks 
22nd among all states in total population (Nevada is 35th). Colorado's 
average annual background radiation is estimated at 700 mrem/yr (see 
``Assessment of Variations in Radiation Exposure in the United 
States,'' 2005, Docket No. OAR-2005-0083-0077, for both background 
radiation and population information). Other states have comparable or 
higher radon potential and higher background levels with which people 
live routinely (background levels in North Dakota, South Dakota, and 
Iowa, for example, are 789 mrem/yr, 963 mrem/yr, and 784 mrem/yr, 
respectively), and might also be used for comparison. However, we 
believe Colorado is more representative of the characteristics 
exhibited by Nevada (and Amargosa Valley).
    In view of these factors, we selected Colorado as our point of 
reference. Thus, comparing Colorado's estimated average annual 
background radiation of 700 mrem/yr to our estimate for Amargosa 
Valley, we derive an incremental exposure level of 350 mrem/yr, which 
we are proposing to establish today as the dose limit to

[[Page 49038]]

apply to the time of peak dose beyond 10,000 years.
    The limit we are proposing today is somewhat higher than the 
average natural background level of 300 mrem/yr across the U.S., which 
places it above two other options we considered (see Sections II.C.4.b 
and II.C.4.c). One option is the limit of 100 mrem/yr based on 
international guidance for all sources of exposure except natural, 
accidental, and medical. The other is 200 mrem/yr, which we derived 
through a somewhat different way of looking at total background levels 
nationwide. In our view, the 350 mrem/yr level and these other values 
are within a range of values for which projections might well be 
indistinguishable after several hundred thousand years. That is, when 
taking increasing uncertainties into account in the very long term, the 
effects of factors that would distinguish projections of 100, 200, and 
350 mrem/yr within a 10,000-year time frame are more difficult to 
identify clearly at very long times, so that such projections may be 
qualitatively identical to each other and to the level of performance 
represented by projections of 15 mrem/yr at 10,000 years. That is, 
modest differences in basic modeling assumptions regarding such factors 
as temperature inside the repository over the first few hundred years 
after disposal can lead to differences in projected doses. Such 
differences reflect uncertainties and changes in models, and should not 
be interpreted as representing meaningful differences in the level of 
safety that can be expected to be achieved. Given the difficulty in 
estimating performance in the very far future, we would also view 350 
mrem/yr as representing a satisfactory level of performance should it 
be the ``true'' value at such long times.
    We recognize that a standard based on variations in natural 
background radiation would be higher than previous non-occupational 
standards in the U.S. In our 2001 rulemaking, we justified the dose 
limit of 15 mrem/yr and the 10,000-year compliance period in part 
because they were consistent with other EPA policies. In particular, a 
peak dose standard of 350 mrem/yr (and the time frame of up to 1 
million years over which that standard could apply) may appear to some 
to be a departure from the risk-management policies EPA has adopted and 
applied in a variety of Agency programs, most notably in the Superfund 
cleanup program. We believe the circumstances involved in today's 
proposal are significantly different from the situations addressed 
under Superfund or any other existing U.S. regulatory program, and that 
it should be clear that comparisons between the two are inappropriate.
    It should be clear that we are not arguing that most people take 
into account levels of background radiation when deciding where to live 
or work, or that it in any way plays a major role in their decision-
making. Rather, in establishing a standard to apply to the RMEI over 
unprecedented times, we believe it is reasonable to consider exposures 
incurred routinely today by people in other locations, which in our 
view do not ``pose a realistic threat of irreversible harm or 
catastrophic consequences'' to those people.
    In that context, we note that EPA does not consider the risks from 
such exposures to be excessive in the context of radon occurrence in 
residences. As described earlier, radon exposures can vary widely even 
in localized areas for a number of reasons. While average radon doses 
are estimated to be roughly 200 mrem/yr, measurements indicate that 
some exposures could be more than ten times that level in unique 
situations. The concentration at which EPA recommends action be taken 
to mitigate exposures is 4 pCi/l, which translates roughly to 800 mrem/
yr. The Agency further recommends that homeowners consider taking 
action only if the measured concentration is between 2 and 4 pCi/l 
(i.e., above 400 mrem/yr) (``A Citizen's Guide to Radon: The Guide to 
Protecting Yourself and Your Family from Radon,'' EPA 402-K-02-006, May 
2004, Docket No. OAR-2005-0083-0058). It should be understood that this 
recommendation is not based solely on risk, but considers factors such 
as the voluntary nature of the exposure, the application to private 
property, and the capabilities of mitigation technology. The dose limit 
proposed today is well below the ``action level'' recommended for 
radon.
    One way to provide context for comparisons with natural 
radioactivity is to evaluate the radiotoxicity of the waste itself. In 
particular, it has been suggested that assessment time frames could be 
tied to the time necessary for the waste to decay to levels roughly 
comparable to the uranium ore from which the fuel was derived, which is 
often on the order of several hundred thousand years. For example, IAEA 
states that ``[r]adiotoxicity indices are useful in putting the 
potential hazards of radioactive waste disposal into perspective * * * 
they are qualitative indicators of the time-scales of interest for 
safety analysis'' (``Safety Indicators in Different Time Frames for the 
Safety Assessment of Underground Radioactive Waste Repositories,'' 
TECDOC-767, p. 15, 2004, Docket No. OAR-2005-0083-0044). NEA takes a 
similar position: ``radiological toxicity and comparison with natural 
systems such as uranium ores offer a basis for a safety indicator that 
can usefully complement dose and risk'' (``The Handling of Timescales 
in Assessing Post-Closure Safety,'' p. 30, 2004, Docket No. OAR-2005-
0083-0046). Standards developed in Finland explicitly incorporate this 
comparison by defining the ``farthest future'' for assessments as the 
period when the activity in spent fuel becomes less than that in the 
natural uranium from which the fuel was fabricated (NEA, p. 34, Docket 
No. OAR-2005-0083-0046). Draft guidance for the Swedish program states 
that assessments ``need not be extended beyond the point in time when 
the initial content of the radioactive substances in the repository has 
decayed to a level at which the potential of causing harmful effects or 
other environmental consequences has decreased to insignificant 
levels'' (p. 7, Docket No. OAR-2005-0083-0048). One technical paper 
presented in the U.S. concludes that ``regardless of the assumptions 
used, the risk to public health from a HLW or spent fuel waste 
repository will always become less than that of the original uranium 
ore deposit'' and that ``[c]onsidering the nature of the many barriers 
to release that are included in the repository design, [it] should 
easily be the case'' that this ``crossover time'' (the time at which 
the radiotoxicity, or overall hazard, of the remaining waste will be 
equivalent to that of the original ore used to make the fuel) will be 
less than 10,000 years (``An Assessment of Issues Related to 
Determination of Time Periods Required for Isolation of High Level 
Waste,'' Proceedings of the Symposium on Waste Management at Tucson, 
Arizona, February 26-March 2, 1989, Docket No. OAR-2005-0083-0049).
    While it is clear that consideration of natural radioactivity is a 
widely accepted concept for supporting safety assessments over very 
long times, it should also be clear that we believe regulatory 
standards for the Yucca Mountain disposal system based on background 
exposures can be reconciled with considerations of impacts on future 
generations, as outlined earlier in this section. Some international 
statements regarding natural radioactivity reflect the lack of 
consensus on what constitutes an undue burden. For example, NEA notes 
that when ``the repository has become comparable to a natural system in 
certain important aspects, this does not necessarily indicate a return 
to unconditionally safe

[[Page 49039]]

conditions'' (NEA, p. 30, Docket No. OAR-2005-0083-0046).
    However, Chapman and McCombie directly address this question, 
stating that, at these very long times, ``There is no logical or 
ethical reason for trying to provide more protection than the 
population already has from Earth's natural radiation environment, in 
which it lives and evolves * * * it must be recognized that man cannot 
be expected over infinite times to do much better than nature. The 
potential exists for natural uranium ore deposits, or spent fuel or HLW 
repositories, to give rise locally to doses that are higher than the 
global average for natural radiation, particularly if they are 
eventually eroded in the near-surface environment. However people exist 
today in many locations where doses are tens, even up to a hundred 
times higher than the average. Thus, a repository is not providing, 
globally, a novel source of exposure and does not at these long times 
represent any unusual anomaly in the global environment'' (``Principles 
and Standards for Disposal of Long-Lived Radioactive Wastes,'' pp. 114-
115, 2003, Docket No. OAR-2005-0083-0061).
    We do not mean to suggest that uranium ore bodies are benign 
entities, and there is certainly a difference between exposures 
incurred by direct contact with the material and those incurred at a 
distance after environmental transport of material has provided some 
lowering of potential exposures by natural retardation processes. These 
comparisons are relevant in the sense that exposures from longer-term 
releases from the Yucca Mountain disposal system would not be expected 
to be worse than those from natural features that are fairly common in 
parts of the country. The exposures that might result from ore body 
releases are highly dependent on the characteristics of the ore body 
and surrounding environment, as well as the other assumptions applied 
(measurements of releases from unmined ore bodies are limited; however, 
some surficial radiation measurements from unmined ore bodies suggest 
that a person at the site could easily receive several hundred mrem/yr 
(``The Uranium District of the Texas Gulf Coastal Plain'', U.S. 
Department of Energy Symposium Proceedings, CONF-780422, Vol. 2, 1978, 
Docket No. OAR-2005-0083-0081). On this point, we stated in our 1985 
final rulemaking for 40 CFR part 191 that ``estimates of the risks from 
unmined ore bodies ranged from about 10 to more than 100,000 excess 
cancer deaths over 10,000 years. Thus, leaving the ore unmined appears 
to present a risk to future generations comparable to the risks from 
disposal of wastes covered by these standards'' (50 FR 38083, September 
19, 1985, Docket No. OAR-2005-0083-0064). In the terms of the 
Precautionary Principle as presented by NAPA, exposures of this 
magnitude that are projected to occur several hundred thousand years 
into the future should not be considered to ``pose a realistic threat 
of irreversible harm or catastrophic consequences'' (Docket No. OAR-
2005-0083-0087).
    We recognize that meaningful distinctions are made today between 
natural background radiation and additional incremental (and 
involuntary) exposures caused by human activity. However, at long time 
frames (potentially as long as 1 million years into the future), such 
distinctions are less meaningful, and natural radiation levels can 
serve as a reasonable and logical reference point for assessing 
radiological impacts. We agree with NEA that a reasonable overall aim 
``is to leave future generations an environment that is protected to a 
degree acceptable to our own generation * * * this level of protection 
will ensure that any radiological impacts due to disposal will not 
raise levels of radiation above the range that typically occurs 
naturally'' (NEA, p. 9, Docket No. OAR-2005-0083-0046). Our proposed 
approach limits doses from the Yucca Mountain disposal system in the 
far future to levels that represent variations in natural background 
and are far below doses that can be projected from uranium ore bodies 
or natural radiation in some locations in the U.S. and worldwide. Our 
proposed limit is somewhat higher than the annual average background 
radiation in the U.S. Using the reasoning described above, under this 
standard the additional radiation exposure at the time of peak dose to 
a resident of Amargosa Valley from the Yucca Mountain disposal system 
would be no greater than what would be incurred if that person moved 
today from the vicinity of Yucca Mountain to a nearby state. Using the 
NAS suggestions as a starting point, and considering international 
guidance and examples, we have derived the proposed dose limit to 
balance competing factors highlighted by NAS and acknowledged by us as 
important: the dual objectives to effectively address the effects of 
uncertainty on compliance assessment and to adhere as closely as 
possible to the relevant ethical principles, including a consideration 
of impacts on future generations. We believe that our selection of a 
350 mrem standard is reasonable and effectively addresses the factors 
it is necessary to consider when projecting exposures very far into the 
future. By applying over the entire period of geologic stability beyond 
10,000 years (up to 1 million years), it will capture the peak dose 
during that period. By doing so, our proposal is consistent with the 
NAS recommendation to have a standard with compliance measured ``at the 
time of peak risk, whenever it occurs within the limits imposed by the 
long-term stability of the geologic environment, which is on the order 
of one million years'' (NAS Report p. 2).
    In all of our discussion of potential dose standards, we have 
emphasized the importance of perspective in evaluating dose projections 
at very long times. It is important to distinguish between effects that 
are meaningful in assuring public health and safety and those that 
simply illustrate a modeling exercise. We are proposing an approach to 
setting a dose level derived from variations in current natural 
background radiation in the U.S. that would relate potential exposures 
to the RMEI to exposures incurred today by people in other locations 
from sources of natural background radiation. Given the long times 
involved in dose projections, and the significant uncertainties, we 
believe that comparisons with natural sources of radiation are 
appropriate.
    Finally, from a regulatory perspective, we have also considered 
that the peak dose limit would apply at any time after 10,000 years. 
The limit we select must be credible both at times close to 10,000 
years and times much further into the future. Readers may also question 
whether a 350 mrem/yr standard can be considered credible at times 
beyond but closer to 10,000 years. (We have acknowledged that 
uncertainties are not immediately overwhelming and unmanageable for a 
period up to 10,000 years.) We think it unlikely that the peak would 
occur at a relatively early time beyond 10,000 years. However, should 
that be the case, we believe that NRC has the authority to consider not 
only the magnitude of the peak, but also the timing and overall trends 
of dose projections as it evaluates the license application. NRC will 
examine the full record before it in determining whether there is a 
reasonable expectation that the standards will be met. As an 
alternative, we might identify a sliding scale of compliance limits 
applicable at different times, but, as discussed in Section II.C.2.c, 
we do not believe there is a clear basis for doing so.
    In addition to our proposed level of 350 mrem/yr, we took into 
account the

[[Page 49040]]

factors described above in considering various options for the peak 
dose limit, as discussed in the next section. Clearly, the competing 
considerations described above are not easily resolved. While the final 
standard may not be identical to any of these options, we believe that 
they encompass the range of values we might reasonably select. We 
request comment upon our proposed annual peak dose limit of 350 mrem 
applicable beyond 10,000 years through the period of geologic 
stability, the reasoning outlined above, and other ways in which we 
might reconcile the various influential factors at very long times.
4. What Other Peak Dose Levels Did EPA Consider?
    We considered several other dose options before selecting 350 mrem 
as the value to propose. We request comment on the dose levels 
discussed in the following sections.

a. Maintain the 15 Mrem/Yr Standard at Peak Dose

    One approach would be simply to apply the same level deemed 
protective at 10,000 years to peak exposures, whenever they might 
occur. This approach has been recommended by some stakeholders (Docket 
No. OAR-2005-0083-0022). Stakeholders have suggested defining the 
``compliance period'' as the time from disposal to peak dose, stating 
that ``[t]his new compliance period is absolutely required by [NAS], 
which rejects any 10,000-year limitation on the compliance period.'' 
However, for the reasons discussed earlier, while we are proposing to 
extend the compliance period throughout the period of geologic 
stability, we have concerns that an approach that applies the same dose 
level throughout that period would not adequately recognize the 
complexities inherent in projections that could extend for several 
hundred thousand years. As a result, we believe a 15 mrem/yr standard 
at very long times would not be a meaningful indicator of disposal 
system performance, and may in fact be misleading.
    We disagree with the view that the Court's decision requires us to 
amend our standards by extending both the compliance period and the 
dose limit applicable at 10,000 years to the time of peak dose up to 1 
million years, and forbids us to establish standards applicable at 
intermediate times. The Court's decision reflected its judgment that 
our 2001 standards were not consistent with the recommendations of NAS 
as they related to the compliance period. Our goal in today's proposal 
is to amend our standards so that they are clearly consistent with the 
NAS recommendations, but also address the policy and other concerns we 
raised in our 2001 rulemaking. Extending the compliance period directly 
addresses NAS's primary recommendation. Regarding the dose limit 
applicable at the time of peak dose, NAS stated ``we do not directly 
recommend a level of acceptable risk'' (NAS Report p. 49). Similarly, 
NAS offered no opinion on approaches involving multiple dose standards, 
stating only that it viewed a 10,000-year standard by itself as 
insufficient (NAS Report pp. 54-56). As the Court made clear in its 
consideration of the ground-water protection standards, where ``NAS 
made no `finding' or `recommendation' that EPA's regulation could fail 
to be `based upon and consistent with' * * * [the EnPA] left [EPA] 
free'' to promulgate standards to address its policy concerns. (NEI, 
373 F.3d at 47, Docket No. OAR-2005-0083-0080.) In our view, the 
standard applicable for the first 10,000 years and the derivation of a 
different dose limit applicable beyond 10,000 years are both 
permissible under our EnPA authority.
    From a regulatory perspective, a compliance standard on the order 
of 15 mrem/yr implies far more precision in projections for very long 
times than can be supported and, as such, is inconsistent with the 
``reasonable expectation'' approach. We have also discussed at length 
the general agreement among international bodies and programs that 
longer-term standards should recognize the uncertainties involved and 
projections must be considered in a more qualitative manner, as one 
element in the overall safety case. As such, we believe it is 
inappropriate to portray that projections of incremental doses at such 
low levels can be precisely controlled at far future times and to give 
them excessive influence when they are not critical to improvements in 
health and safety. Here again, we believe our statement from the 2001 
rulemaking bears repeating: ``[s]etting a strict numerical standard at 
a level of risk acceptable today would ignore this cumulative 
uncertainty and the extreme difficulty of using highly uncertain 
assessment results to determine compliance with that standard'' (66 FR 
32098). From that perspective, holding the Yucca Mountain disposal 
system to a 15 mrem/yr standard would not merely be an issue of 
``fairness'' to very far future generations. Instead, by not 
recognizing the factors that fundamentally affect dose projections at 
such times, it would place those generations' interests in a much 
higher regard, and by doing so would unreasonably constrain the current 
and succeeding generations' abilities to pursue achievable solutions 
they deem best suited to meet the interests of all generations. It 
would, in other words, undermine the Chain of Obligation Principle by 
giving ``long-term hypothetical hazards'' precedence over ``near-term 
concrete hazards'' (``Deciding for the Future: Balancing Risks, Costs, 
and Benefits Fairly Across Generations,'' 1997, Docket No. OAR-2005-
0083-0087). It is not simply a question of whether a 15 mrem/yr 
standard could be met in actuality; rather, the question is whether 
holding probabilistic assessments to such a level over hundreds of 
thousands of years, when rising uncertainties exist in performance 
projections and various high-consequence (but necessarily somewhat 
speculative) scenarios must be considered in the analyses, represents a 
reasonable test of the disposal system. We believe it does not.

b. 100 Mrem/Yr Standard at Peak Dose

    In evaluating dose limits that might be responsive to the concerns 
outlined above, we also considered 100 mrem/yr as a value that may be 
appropriate for peak dose calculations. The value of 100 mrem/yr has 
potential benefits in terms of precedent. The ICRP has since 1985 
(Publication 45, ``Quantitative Bases for Developing a Unified Index of 
Harm,'' Statement from the 1985 Paris Meeting of the ICRP, Docket No. 
OAR-2005-0083-0087) recommended 100 mrem/yr as the principal overall 
dose limit for public exposures from all sources excluding natural 
background, medical, occupational, and accidental (these three man-made 
sources can involve higher exposures, can involve greater understanding 
of the reasons for exposure, and may require informed consent from the 
exposed person). NRC requires that its licensees conduct operations so 
that individual members of the public are not exposed above this level 
(10 CFR 20.1301). We view this figure as representing a national and 
international precedent as a generally-accepted benchmark for annual 
public exposures from various sources.
    The use of 100 mrem/yr can also be interpreted as protective of 
future generations' interests, yet not so restrictive as to represent 
an unreasonable standard for the very far future. We recognize that in 
practice today, doses from any particular source of radiation are 
generally kept to a fraction of the 100 mrem overall limit, in 
recognition that a person may be exposed to more than one practice or 
source. Given our current responsibility

[[Page 49041]]

to propose a peak dose standard, however, we would argue that 
allocation to a single source at the time of peak dose could be 
reasonable, as other contributors currently in the Yucca Mountain area 
are negligible by comparison (FEIS, DOE/EIS-0250, Section 8.3.2, Docket 
No. OAR-2005-0083-0086). Moreover, to assume (or even to estimate the 
chance) whether, how, or where other radiation facilities could develop 
in the far future would require immense speculation about the long-term 
evolution of government programs, population demographics, and 
technology. Relying on current conditions rather than speculating on 
future sources of exposure to the local population, as recommended by 
NAS, would justify allocating the entire 100 mrem to Yucca Mountain.\4\
---------------------------------------------------------------------------

    \4\ This approach would also be consistent with the recent ICRP 
draft for consultation on optimization of radiological protection, 
which states ``the choice of the relevant dose constraint for 
protection against exposures from the licensed facility under 
consideration will depend largely on whether or not this facility is 
the dominant source to the exposed public under consideration. If 
the facility is the dominant source, a dose constraint of 1 mSv/a 
[100 mrem/yr] would be the appropriate starting point for 
optimisation of protection'' (``The Optimisation of Radiological 
Protection,'' p. 45, April 2005, Docket No. OAR-2005-0083-0052).
---------------------------------------------------------------------------

    Nevertheless, we have decided not to propose a peak dose standard 
of 100 mrem/yr because over the very long-term, we believe that natural 
background levels to which individuals are or could be currently 
exposed provides a more reasonable context in which to judge the 
performance of the Yucca Mountain disposal system, and because our 
proposed approach appropriately reflects international guidance and 
consensus on this issue. See Section II.C.3 (``What Dose Level Is EPA 
Proposing for Peak Dose?'').

c. Peak Dose Standard Based on Regional Background Radiation Levels

    We also considered an alternative approach also based on an 
examination of natural background radiation levels across the country. 
In this approach, rather than examining total background radiation at a 
specific location (or State), as we did to derive the 350 mrem/yr level 
we are proposing today, we have looked at average levels across many 
States (``Assessment of Variation in Radiation Exposure in the United 
States,'' 2005, Docket No. OAR-2005-0083-0077). One reason for taking 
this approach is that it compares statewide averages calculated on a 
common basis. Even so, the cautions we expressed in Section II.C.3 
regarding the uncertainties and variation in background radiation 
values remain relevant.
    Using this approach, we arrived at a dose limit of 200 mrem/yr. As 
with our proposed approach, our overall policy goal is to establish a 
standard that would keep total exposures to the RMEI within the range 
of exposures incurred by residents of other locations today from 
natural background sources alone. We would not view 200 mrem/yr as 
excessive in the context of exposures routinely encountered today, 
particularly when considering the uncertainties in projecting potential 
doses over the very long times involved (i.e., 10,000 to 1 million 
years).
    We started by considering States with higher average background 
levels than Nevada. As with our proposed approach, we believe this is 
reasonable because the limit we establish must represent some positive 
incremental exposure to the RMEI. The data compiled in ``Assessment of 
Variation in Radiation Exposure in the United States'' (Docket No. OAR-
2005-0083-0077) show that 32 States have higher average background 
levels than Nevada's 222 mrem/yr. Rather than identify any particular 
State as the reference point, as we did in the direct comparison with 
Amargosa Valley, we averaged the values for those 32 States and 
obtained an average background radiation level of 429 mrem/yr. We 
compared this value to the statewide average for Nevada as an indicator 
of more regional, rather than localized, differences. Thus, on average, 
residents of those 32 States today receive roughly 200 mrem/yr more 
from natural background radiation sources than a resident of Nevada. 
Considering all of the factors involved in very long-term projections, 
such a limit would represent a level of exposure consistent with that 
routinely and normally incurred in other locations. However, we have 
decided not to propose this approach today because our preference is to 
use Amargosa Valley (and the RMEI as the person our standards are 
designed to protect) as a point of reference, but we welcome comment on 
both the approach and the dose level of 200 mrem/yr derived from it.
5. How Will NRC Judge Compliance?
    We require that DOE use probabilistic performance assessment to 
demonstrate compliance with the individual-protection standard in Sec.  
197.20 (DOE may, but is not required to, use the same technique to show 
compliance with the human-intrusion and ground-water protection 
standards). With this method, DOE will obtain a distribution of 
calculated dose results. This distribution will be influenced by 
variations in parameter values as well as fundamental uncertainties and 
the assumptions underlying the conceptual model(s) of disposal system 
evolution. In making a compliance demonstration, DOE must satisfy NRC 
that a specified portion of that distribution satisfies the dose 
criterion. In our 2001 rulemaking, we specified in Sec.  197.13 that 
the mean of the distribution of results be used to demonstrate 
compliance with Sec.  197.20. In doing so, we intended that the 
arithmetic mean (commonly known as the average) of the distribution be 
used, but did not feel the need to be so specific. The arithmetic mean 
is a well-understood measure that is used in many compliance 
applications, including at WIPP. As discussed later, we intend to 
retain the arithmetic mean for the compliance measure for the first 
10,000 years after disposal.
    However, for the period beyond 10,000 years, for which we must 
consider assessing performance for as long as 1 million years (the 
NAS's estimated period of ``geologic stability''), we realize that some 
additional specification is necessary. Although we do not believe that 
the basic approach to performance assessment should be affected, we 
discuss in Section II.D certain aspects of that approach that may need 
to be modified or given special attention to effectively address these 
much longer times in a meaningful way. Similarly, we must consider 
whether the arithmetic mean used for compliance at 10,000 years remains 
the appropriate measure of compliance, particularly at very long times, 
or whether another measure is more appropriate.
    We believe that for these very long-term projections, a measure 
that represents a ``central tendency'' in the distribution of 
calculated doses is most appropriate to adequately consider the range 
of uncertainty in making dose projections over such very long time 
spans. Such a measure should not be strongly influenced by high or low-
end projections that represent low probability situations. Today we are 
proposing to specify that compliance with the standard that will apply 
beyond 10,000 years should be measured against the median of the 
distribution of projected doses. The remainder of this section 
discusses our rationale for this approach.
    In general, the compliance measure we select must be meaningful and 
representative of the entire distribution of calculated doses, but, as 
we have stated, not easily influenced by results either at the very 
high or very low end of the distribution. In conducting

[[Page 49042]]

performance assessments many assumptions and uncertainties must be 
incorporated into the complex calculations. In constructing scenarios 
for repository performance, there are uncertainties in describing how 
the disposal system will perform and evolve over time, under the 
influence of natural conditions and the effects of the repository 
itself on the surrounding host rock. There are significant 
uncertainties in predicting when discrete events, such as seismic 
activity, will occur at and around the immediate repository location 
and the possible effects of these events. Some scenarios incorporating 
these uncertainties would be of low probability, and the results could 
vary from relatively poor performance to exceptionally good performance 
of the disposal system. The results of such low-probability situations 
with dramatically different results than the majority of the 
projections would show up in the ``tails'' of the dose results 
distribution. While we believe such low-probability situations should 
not be ignored in compliance decisions, neither do we believe they 
should be given undue influence in judging compliance. Therefore, we 
believe that the appropriate compliance measure should represent a 
central measure for the dose projections, and should not be defined in 
a way that it can be strongly affected by extreme results 
(``outliers'') in the dose projections (``Assumptions, Conservatisms, 
and Uncertainties in Yucca Mountain Performance Assessments, Sections 
12.1 and 12.2, July 2005, Docket No. OAR-2005-0083-0085).
    Today we are retaining, and more clearly specifying, the arithmetic 
mean of the dose projections for compliance within the initial 10,000-
year period. We believe the arithmetic mean is a familiar and well-
understood statistical concept, and that its application in 
probabilistic risk assessment is sufficiently established to support 
our decision. In addition, while uncertainties are present in 
performance assessments during this time frame, we believe that the 
uncertainties increase in importance as the assessments stretch into 
the extremely long time frames beyond 10,000 years but within the 
period of geologic stability. In this sense, we believe that the 
arithmetic mean (average value) of the dose projections can still be a 
reasonably reliable measure of the total dose distribution during the 
10,000-year period. More importantly, however, we believe it is 
valuable to maintain consistency between the compliance measure used 
for the first 10,000 years of disposal system performance for the Yucca 
Mountain repository and the measure applied for any other geologic 
disposal application under the authority of our generic regulation for 
geologic disposal, 40 CFR part 191. We believe that the Yucca Mountain 
disposal system should be required to meet the same level of 
protection, and be evaluated under the same regulatory compliance 
framework, as any other geologic disposal application for the 10,000-
year period considered in part 191, which has been applied to the WIPP 
facility specifically and would apply to any other disposal system in 
the future. In the unlikely event that performance assessments show 
that the peak dose would occur within the 10,000-year period, we 
believe that the same compliance measure and evaluation should be 
applied for the Yucca Mountain disposal system as for any other 
geologic disposal system.
    However, we have noted repeatedly that extending the compliance 
period to time frames well in excess of 10,000 years introduces 
additional uncertainty in making disposal system performance 
projections, since the natural system will continue to change through 
time (see ``Assumptions, Conservatisms, and Uncertainties in Yucca 
Mountain Performance Assessments,'' Section 12.5, July 2005, Docket No. 
OAR-2005-0083-0085, and the 2001 BID, section 7.3.11, Docket No. OAR-
2005-0083-0050). We believe probabilistic calculations are the most 
appropriate approach to assess the range of potential doses over very 
long time frames, both for the period up to 10,000 years and after. The 
probabilistic approach examines a spectrum of possible site conditions, 
and allows the construction of conceptual models that address 
reasonable alternative models of performance within that range of 
possible physical and chemical conditions at the site. A distribution 
of projected peak doses will result from these analyses, each 
representing a possible ``future'' and a dose associated with the 
specific parameter values chosen for each calculation. Only by 
examining the relative effects of variations in the parameter values on 
the calculated dose can the important Adriver'' parameters be 
identified. Without these types of analyses, an understanding of the 
disposal system's behavior in the long term would not be possible, and 
a compliance case supporting a decision about the protectiveness of the 
disposal system might not be a reasonable representation of potential 
risks. We are proposing to require that DOE apply this general approach 
for assessments regardless of time frame, although, as we have 
discussed earlier, there is much agreement that the level of confidence 
or meaning that can be placed in such analyses decreases over very long 
periods. The challenge lies in defining a performance measure that 
balances the uncertainties inherent in very long term projections and 
provides a reasonable level of protectiveness.
    Similarly, some discussion is warranted on the role of conservatism 
in performance assessment. Excess conservatism in constructing 
scenarios, i.e., making assumptions to include or exclude specific FEPs 
and defining parameter value ranges, can easily lead to very high dose 
estimates due to a compounding effect of very conservative assumptions. 
Such excessive conservatism is misleading if incorporated in 
assessments described as the Anominal'' or Abase case'' performance 
projections. A simple arithmetic mean calculated for an excessively 
conservative analysis would suggest that the ``most likely'' dose is 
higher than if a more reasonable and realistic approach were taken to 
framing the assessments. Recognizing that conservatism in long-term 
performance projections may be unavoidable in practice, as discussed 
below, we believe that a regulatory performance measure should not give 
undue emphasis to high-end projections. It is always possible to 
propose scenarios where releases are high, even though the probability 
of these particular scenarios may be extremely small or very difficult 
to estimate. The same reasoning also applies to scenarios that result 
in very low releases in the very long term. The ``bounding'' approach 
to assessments plays an important role in the light of the increasing 
uncertainties. Once the time frame for performance projections is 
extended into the very long term, the confidence that can be placed on 
either the high- or low-end release scenarios becomes progressively 
more difficult to estimate even though a ``bounding'' approach may 
simplify calculations. Consequently, we believe that a performance 
measure for these very long term assessments should not over emphasize 
high-end release scenarios or ignore low-end release scenarios under 
the motivation for conservatism in the assessments.
    In addition, uncertainty and conservatism can influence one 
another. Characterization of the site today yields a range of values 
for important parameters that would have a dominant effect on 
projecting doses from contamination plumes eventually released from the 
repository, and these

[[Page 49043]]

data form the base of the parameter value distributions input to the 
dose calculations. Attempting to project the evolution of these 
parameter values over the 1 million year geologic stability period adds 
additional uncertainty in their variations. To address these 
uncertainties in parameter value estimation and scenario construction, 
analyses of disposal system performance may be done Aconservatively,'' 
i.e., by selecting parameter values, constructing scenarios, and making 
assumptions that deliberately overestimate projected doses. This 
approach provides some confidence that uncertainties and other 
potential negative influences have been adequately considered and that 
regulatory decisions will not be based on overly optimistic views of 
disposal system performance. However, the distribution of doses, as 
well as peak doses, from such an approach will be biased toward high-
end values. As a result of making conservative assumptions and 
parameter distributions, there is a very real possibility that high-end 
projections could represent highly improbable situations in a physical 
sense (``Assumptions, Conservatisms, and Uncertainties in Yucca 
Mountain Performance Assessments,'' Sections 1 through 12, July 2005, 
Docket No. OAR-2005-0083-0085). For such cases, arriving at a 
compliance decision becomes complex and speculative. Thus, we believe 
the appropriate measure of compliance for peak dose estimates is a 
``central tendency'' measure which is not strongly influenced by low-
probability realizations giving either very high-end or low-end dose 
estimates (``Assumptions, Conservatisms, and Uncertainties in Yucca 
Mountain Performance Assessments,'' Sections 12.1 and 12.2, July 2005, 
Docket No. OAR-2005-0083-0085).
    The NAS also found this approach to have merit. NAS recommended 
that regulatory decision making should consider the period when risks 
are at their highest, whenever that occurs, i.e., the time of peak dose 
(NAS Report pp. 2, 6). In defining ``risk,'' the NAS used the term 
Aexpected value'' in referring to a probabilistic distribution of 
projected doses (NAS Report p.65). NAS did not further define this term 
in a statistical context, but elsewhere provided qualitative language 
describing the overall goal: ``define the standard in such a way that 
it is a useful measure of the degree to which the public is to be 
protected from releases from a repository'' and ``does not rule out an 
adequately sited and well-designed repository because of highly 
improbable events'' (NAS Report pp. 27-28). NAS in its recommendations 
did not speak explicitly to any particular performance measure to be 
used in determining compliance with regulatory standards. This decision 
was to be left to EPA in the course of rulemaking.
    Disposal programs abroad also have to consider the role of 
uncertainty in developing performance assessments. The U.S. is ahead of 
most other geologic repository programs abroad in terms of having a 
specific site that has been extensively characterized and for which 
detailed performance assessments have been done. While other programs 
have not reached the stage of developing specific regulatory criteria 
for judging the acceptability of a particular repository site and 
design, there is a general consensus that multiple lines of evidence 
and analysis are desirable in establishing a safety case, and that 
judgment plays a critical role in assessments of disposal system 
performance as well as establishing and applying regulatory criteria 
(IAEA-TECDOC-975, Docket No. OAR-2005-0083-0045). The joint NEA-IAEA 
International Peer Review for DOE's TSPA-SR modeling highlighted the 
difficulty of specifying the statistical measure of compliance, noting 
that ``the TSPA nominal case is treated probabilistically yet it 
involves a mixture of embedded conservatism and statistical analyses to 
determine the mean, median and the various percentiles of the dose 
distribution. The reported ``mean'' is therefore not the true mean in a 
statistical sense. Moreover, that value is reported in the Executive 
Summary of the TSPA-SR and elsewhere as the expected value of effective 
dose, without any qualification. This stretches credibility especially 
as the discrete numerical values are given for times in the far future. 
The USDOE needs to indicate that, for compliance purposes, a 
performance indicator has been chosen that is meant to illustrate the 
safety of the system and argue the compliance with regulation.'' The 
Peer Review Team further recommended that ``when a best estimate/best 
knowledge probabilistic analysis is performed, the best estimate or the 
most probable range of the calculated `dose' should also be given.'' 
(pp. 54-55, Docket No. OAR-2005-0083-0062)
    In determining the ``expected value'' of performance, some 
international efforts have considered the possibility of viewing the 
performance assessment as separate representations of scenarios driven 
by their relative likelihood, and which might be compared to different 
regulatory standards. For example, regulatory agencies of France and 
Belgium have developed a joint document that suggests preparation of 
``reference evolution'' and ``altered evolution'' scenarios 
(``Geological Disposal of Radioactive Waste: Elements of a Safety 
Approach,'' p. 24, 2004, Docket No. OAR-2005-0083-0066). The reference 
evolution scenarios would consider ``the most likely effects of certain 
or very probable events or phenomena,'' while the altered evolution 
scenarios ``take into account the least likely effects of these events 
or phenomena'' as well as considering ``the consequences of events or 
phenomena that are not integrated into the reference scenario, as the 
likelihood of occurrence is lower.'' Under this approach, the reference 
evolution scenarios might be directly compared to the dose constraint, 
while the altered evolution scenarios ``must be appraised on a case by 
case basis depending on'' various factors, and may then be ``compared 
to different references * * * without this comparison constituting an 
absolute acceptance criterion.'' This approach appears to go further 
than that recommended by the TSPA-SR Peer Review Team (and discussed in 
relation to our reasonable expectation principle in Section II.B). DOE 
similarly identifies ``nominal'' and ``disruptive'' scenarios, but 
aggregates the results for comparison with the relevant criteria.
    As stated earlier, we required in our 2001 rulemaking that DOE use 
the arithmetic mean of the distribution of results to demonstrate 
compliance with the 10,000-year dose limit (and are today proposing to 
clarify the use of that measure). However, in considering the much 
longer times, we are concerned that the arithmetic mean is too easily 
influenced by extremes in the distribution, particularly very high dose 
projections resulting from scenarios that are unlikely to occur. A 
conservative approach to constructing and evaluating performance 
scenarios tends to generate high-end results and a simple averaging of 
these results would drive the arithmetic mean to higher values that 
would not be as representative overall of the actual distribution of 
projected doses. Therefore, we do not believe the arithmetic mean will 
satisfy the goals laid out earlier in this section for a performance 
measure for periods well in excess of 10,000 years.
    While typically the ``average'' of a series of values (i.e., a 
distribution) is thought of as near the midpoint between the highest 
and lowest values, if a somewhat conservative approach is taken or 
there are significant outliers, it

[[Page 49044]]

is not unusual for the arithmetic mean to approach significantly higher 
percentiles. In such cases, the regulatory compliance decision can be 
influenced by the high-end doses of an overall set of very conservative 
performance assessment results. In fact, for early occurrences of 
disruptive events (human intrusion or igneous intrusion), DOE 
assessments show that at some periods of time the arithmetic mean of 
the projected doses can exceed the 95th percentile of the distribution 
of TSPA results (FEIS, DOE/EIS-0250, Appendix I, Section 5.3, Docket 
No. OAR-2005-0083-0086). While conservatism in assumptions is not the 
only reason for the arithmetic mean to occur at a relatively high 
percentile, in general we do not believe this can be reasonably 
interpreted to be an adequate representation of central tendency for 
the purpose of judging the performance of the Yucca Mountain disposal 
system.
    Thus, we found it necessary to consider what other statistical 
measures might better represent the central tendency for performance 
assessments at very long time frames. The identification of appropriate 
statistical measures for central tendency of a dose distribution is 
influenced by the shape of the distribution, when these estimates are 
plotted for a particular point in time, or more specifically for the 
peak dose values from each computer modeling simulation in the disposal 
system performance assessments. We have examined three measures of 
central tendency: the arithmetic mean, the geometric mean, and the 
median. The degree to which they reliably represent the central 
tendency of a particular distribution, and more importantly how well 
they could serve as compliance measures, is discussed below. Like the 
arithmetic mean we have discussed above, each measure has advantages 
and disadvantages, and is dependent on the actual shape of the dose 
distribution as to how well it would represent the central tendency 
(``Assumptions, Conservatisms, and Uncertainties in Yucca Mountain 
Performance Assessment,'' Sections 12.1 and 12.2, July 2005, Docket No. 
OAR-2005-0083-0085).
    The most familiar shape for a distribution is the bell-shaped 
``normal'' distribution. In a normal distribution, the ``peak'' occurs 
in the center of the distribution and the remaining values lie evenly 
on both sides of the center value. A normal distribution is often seen 
when values are relatively close together (i.e., the range of values 
does not cover many orders of magnitude), and are produced from a 
continuous function composed of additive terms. Because the values of 
the distribution are evenly spread out around the central peak, the 
distribution can be seen to be symmetrical; that is, one side is the 
``mirror image'' of the other. The arithmetic mean can be easily 
determined from such a distribution because an equal number of values 
are found at the same distance from the peak (e.g., if the peak is at 
10, there will be equal occurrences at 9 and 11, at 8 and 12, and so 
on). Thus, the center line in a purely normal distribution represents 
the arithmetic mean of the distribution. From the discussion earlier in 
this section, it should be clear that the performance results do not 
represent a purely normal distribution. In a purely normal 
distribution, the arithmetic mean cannot be as high as the 60th 
percentile, much less the 70th, 80th, or 95th percentile. It must 
always be the 50th percentile. For this reason, we believe it likely 
that at long times the arithmetic mean will be more strongly influenced 
by higher-end estimates (estimates lower than zero are not possible) 
and less representative of the overall distribution.
    As an alternative, we considered whether the geometric mean of the 
distribution would be an appropriate statistical measure. Referring 
back to the shape of the distribution as an indicator of the measure of 
central tendency, we noted earlier that the bell-shaped curve is the 
most familiar shape. However, many measured quantities in nature show a 
distribution skewed toward higher-end values, i.e., there is no 
symmetrical distribution around a central value (``The Lognormal 
Distribution in Environmental Applications,'' EPA/600/S-97/006, 
December 1997, Docket No. OAR-2005-0083-0057). When data like these are 
transformed by taking their logarithms and plotted on a logarithmic 
scale, the data can appear ``normally'' distributed. Such distributions 
are referred to as log-normal. For such ``transformed'' data, the 
geometric mean is used as the measure of central tendency (that is, the 
geometric mean of a log-normal distribution has a comparable 
significance to the arithmetic mean of a normal distribution).\5\ The 
fact that the arithmetic mean has been significantly higher than the 
50th percentile in DOE's published performance assessment results 
suggests those distributions might be log-normal in nature, which would 
indicate the geometric mean as the appropriate statistical measure of 
central tendency. As a point of comparison, the geometric mean of a 
log-normal distribution is always lower than the arithmetic mean. This 
makes the geometric mean attractive if we are concerned about the undue 
influence of high-end estimates, as the geometric mean will always show 
less influence than the arithmetic mean.
---------------------------------------------------------------------------

    \5\ The formula for calculating the geometric mean (GM) for a 
series of values, x1, x2, x3 . . . 
. Xn, is GM = \n\ [radic] x1 * x2 * 
x3 . . . . Xn, while the formula for 
calculating the arithmetic mean (AM) is AM = (x1 + 
x2 + x3 . . . xn)/n. For the GM 
calculation no zeros are permissible, and the GM is always less than 
the AM. For parameter values in a skewed distribution (skewed to 
high-end values) that is transformed into a log-normal distribution, 
the formula for the GM is expressed as ln GM = (1/n)(1n 
x1 + 1n x2 + 1n x3 . . . . + 1n 
xn). It can be seen that the GM of the log-transformed 
values in a log-normal distribution is calculated in the same 
fashion as the AM for a normal distribution. Both the AM and the GM 
are measures of central tendency for their respective distributions 
and equivalent to the median of the distributions as long as the 
distributions are truly normal or log-normal.
---------------------------------------------------------------------------

    However, there are some difficulties in using the geometric mean 
that must be considered. One difficulty is related to the nature of the 
geometric mean itself. Because the calculation involves the taking of 
the logarithm, the distribution values are expressed in terms of their 
exponential values, which may then be ``averaged.'' For example, the 
logarithm of 100 is 2, because 100 = 10\2\ (or 10 to the 2nd power). 
Similarly, the logarithm of numbers less than 1 are expressed as 
negative numbers (e.g., the logarithm of 0.01 = -2, because 0.01 can 
also be written as 10-2). Thus, in the same way that the 
arithmetic mean might be affected by a few very large values in a 
distribution, the geometric mean can be affected by very small numbers 
whose logarithms are expressed as very large negative numbers.
    In practical applications, this means that a distribution that 
generally appears log-normal can contain some very small numbers 
(outliers) that affect the geometric mean as a measure of central 
tendency. Depending on how many and how small these outliers are, the 
calculated geometric mean can also be very different from the 50th 
percentile of the distribution. For Yucca Mountain, this situation 
could represent a case where the waste packages remain essentially 
unbreached through the geologic stability period, leading to very small 
doses (and correspondingly high negative logarithms of those dose 
values). This scenario might have a very low probability in reality, 
but could influence the geometric mean, possibly causing it to be lower 
than the 50th percentile of results calculated from all the performance 
scenarios taken in total (and possibly very much lower). Alternatively, 
extremely pessimistic scenarios for waste package releases could give 
high-end outliers, although

[[Page 49045]]

the high-end projections may not affect the geometric mean as much 
because the site's characteristics will not easily allow orders of 
magnitude increase in releases to reach the RMEI. In terms of the 
logarithmic values, the difference between 0.001 mrem and 0.1 mrem is 
exactly the same as the difference between 1 mrem and 100 mrem (two 
orders of magnitude), yet the difference in actual site performance is 
clearly more significant between 1 mrem and 100 mrem. Thus, while it is 
possible to have very low-dose estimates, micro-rem/yr and below, which 
have large negative logarithms, there will not be correspondingly high 
dose estimates in the tens to hundreds of thousands of rem/yr (with 
equally high positive logarithms) to counterbalance the very low 
numbers, and therefore these very low numbers could exert a stronger 
effect on the geometric mean as an indicator of central tendency. In 
such cases, the values of the geometric mean as a central tendency 
performance measure could be compromised.
    An additional complication exists for the regulator using the 
geometric mean to judge compliance. Because the logarithm of the value 
must be taken, dose projections of zero must be removed from 
consideration altogether (the logarithm cannot be calculated). However, 
extremely low (and highly influential) non-zero values may be retained 
in the analyses, simply because computers are able to track them. That 
is, projected doses that are in reality essentially indistinguishable 
from zero can be calculated and carried through the analysis. If care 
is not taken, projections could include doses such as 10-20 
mrem/yr, which is meaningless in actuality (and clearly the logarithmic 
value of -20 cannot be offset by any single high-end estimate). The 
regulatory analyst is then faced with the prospect of ignoring 
simulations that yield no dose, eliminating values below a certain 
level (for very low dose estimates), or assigning some arbitrary value 
to them simply to calculate a geometric mean. Eliminating small values 
from consideration would not be consistent with our cautions (see 
discussions on reasonable expectation) that low-end projections should 
not be discounted in favor of higher estimates.
    It is also not proven that the distribution of performance 
assessment results is truly log-normal. As noted above, DOE's 
previously published TSPA results indicate that the distribution of the 
peak dose values is skewed to one side, so that values are not evenly 
distributed around a central point (FEIS, DOE/EIS-0250, Appendix I, 
Section 5.3, Docket No. OAR-2005-0083-0086). We have mentioned the role 
of conservatism in framing dose assessments and biasing them to high-
end values, so this skewed distribution is not surprising. Such skewed 
distributions often appear to be log-normal, for which the geometric 
mean represents the central tendency. However, while we have some 
confidence that future DOE results will be skewed toward the high end, 
we cannot predict with certainty that the distributions are truly log-
normal, although we can say that they display two prominent 
characteristics of log-normal distributions. First, the results span 
several orders of magnitude, making the use of logarithmic conversions 
effective in putting the values on a convenient scale. Second, its 
derivation involves multiplicative functions which are imbedded in the 
performance simulations, while normal distributions arise from simpler 
functions that are additive in nature. Their actual shape will be 
affected by DOE's modifications to the TSPA as it works through the 
licensing process. The geometric mean may not actually represent the 
best measure of central tendency if the distribution is not log-normal.
    For these reasons, we are not proposing to use the geometric mean 
as the measure of compliance at the time of peak dose. This brings us 
to the third statistical measure we considered for these very long 
times, the median of the distribution, for which 50% of the values lie 
above and 50% lie below. The median is a simpler measure of central 
value for any distribution of dose estimates. It is independent of the 
shape of the distribution and therefore avoids concerns about how well 
the performance assessment results may or may not strictly conform to 
the normal or log-normal profiles, and attendant uncertainty about how 
close the respective ``means'' are to the center of the distribution. 
In this respect, the median is an attractive alternative to the 
geometric or arithmetic means as a measure of central tendency that 
will not be strongly influenced by high or low-end outliers in the 
calculated projections. There is no need to eliminate or truncate 
results at the low end, as there may be for the geometric mean. 
Further, if the distribution includes many very low estimates, the 
median could actually be higher than the geometric mean. As such, it is 
also consistent with our reasonable expectation principle.
    As an additional advantage, if the distribution ultimately falls 
close to either a normal or log-normal distribution, the median 
converges with the arithmetic or geometric mean, respectively. It can 
be clearly seen that the median and arithmetic mean are identical for a 
normal distribution, as the ``mirror image'' around the arithmetic mean 
also shows that exactly half of the results fall on either side. For a 
log-normal distribution, the same result can be seen when the initial 
values are transformed by taking their logarithms. Since by definition 
the transformed data takes on the shape of the normal distribution, the 
geometric mean assumes the role of the arithmetic mean for that 
transformed distribution and is coincident with the median. From the 
formulas in footnote 5, it is evident that the geometric mean for log-
transformed data (a log-normal distribution) is calculated in the same 
manner as the arithmetic mean for non-transformed data in a normal 
distribution. This means that, if the performance assessment results 
align closely with the defined normal or log-normal distributions, the 
median will converge with the other statistically defined measures of 
central tendency for those distributions. If the results are very 
highly skewed toward a true log-normal distribution, the geometric mean 
essentially equates to the median, but without the calculational issues 
described earlier. If less conservatism is incorporated into the 
analyses and the resulting distribution is less skewed so that it more 
closely resembles a normal distribution, the arithmetic mean 
essentially converges with the median and the performance measure 
approaches that used to show compliance within 10,000 years.
    These relationships between the arithmetic and geometric means and 
the median are strictly correct only as long as the distributions fit 
the profiles for either the normal or log-normal distributions. If the 
actual shapes of the distributions differ to some degree from the ideal 
defined shapes, the means, either arithmetic or geometric, will not be 
coincident with the median values for the distributions, the degree of 
departure being dependent on exactly how much the distributions depart 
from the ideal ``normal'' or log-normal'' shapes. Deviations from the 
ideal normal and log-normal distribution shapes and the effects on 
these measures as representative of the central tendency for the 
calculated dose projections, are of critical importance in selecting 
the compliance measure. The likelihood of deviations discourages our 
use of either the arithmetic or geometric mean at the time of peak 
dose, but has limited effect on the use of the median.
    Therefore, we propose to use the median of the dose distribution as 
the

[[Page 49046]]

performance measure for compliance in the post-10,000-year period and 
request comment on that decision. Readers may note that our 1999 
proposal, as well as 40 CFR part 191, specified that DOE use the 
(arithmetic) mean or median, whichever was higher. We determined that 
the arithmetic mean would always be higher for periods up to 10,000 
years. Thus, we specified the more conservative measure to apply up to 
10,000 years. However, as noted above, the arithmetic mean may be 
overly influenced by higher-end estimates. Therefore, we do not 
consider it the appropriate measure for times in excess of 10,000 
years.
    In summary, we propose to maintain and clarify the use of the 
arithmetic mean for compliance with the 10,000-year standard. We 
believe this is appropriate because the shorter-term projections are 
not as influenced by the uncertainties or variability in performance 
scenarios seen at much longer times. Fewer very high-end estimates are 
expected and, therefore, overall the distribution of doses would be 
less skewed and more representative of ``expected'' performance. 
Further, in the unlikely event that the peak dose is found to occur 
within the first 10,000 years, the arithmetic mean would be consistent 
with the statistical measure used in all other applications for 
geologic disposal, i.e., 40 CFR parts 191 and 194 for the 10,000-year 
time frame. We request comment on the clarification of the arithmetic 
mean as the 10,000-year compliance measure. For the period extending 
beyond 10,000 years, we propose to use the median of the distribution 
of doses calculated from the performance assessments as the compliance 
measure, and we request comment on this choice.
6. How Will DOE Calculate the Dose?
    Our 2001 standards required DOE to calculate doses as an annual 
committed effective dose equivalent (annual CEDE) to demonstrate 
compliance with the storage, individual-protection, and human-intrusion 
standards. Today we are proposing to modify that requirement in a way 
that would incorporate updated scientific factors necessary for the 
calculation, but would not change the underlying methodology. 
Specifically, we are proposing to require DOE to calculate the annual 
CEDE using the radiation- and organ-weighting factors in ICRP 
Publication 60 (``1990 Recommendations of the ICRP''), rather than 
those in ICRP Publication 26 (``1977 Recommendations of the ICRP''). 
This point may seem straightforward to many readers. We wish to 
incorporate the most recent science into the calculation of dose, so 
why should we not do so? The complication arises from the terminology 
employed in the EnPA and ICRP 60 (and the follow-on implementing 
Publication 72, ``Age-Dependent Doses to Members of the Public from 
Intake of Radionuclides: Part 5 Compilation of Ingestion and Inhalation 
Dose Coefficients,'' 1996, Docket No. OAR-2005-0083-0087). Section 
801(a)(1) of the EnPA explicitly requires our standards to ``prescribe 
the maximum annual effective dose equivalent to individual members of 
the general public.'' Thus, we are required by law to issue an 
individual-protection standard presented as an effective dose 
equivalent. The Court agreed with this reasoning when it stated that 
the EnPA ``require[s] EPA to establish a set of health and safety 
standards, at least one of which must include an EDE-based, individual 
protection standard.'' (NEI, 373 F.3d at 45, Docket No. OAR-2005-0083-
0080.)
    ICRP is an independent body formed to develop consensus 
recommendations on appropriate radiation protection measures. In doing 
so, ICRP considers the principles and scientific bases involved in 
practices that involve the generation of radiation and radioactive 
materials, as well as the use of those materials. Over the years, ICRP 
recommendations have been adopted by regulatory authorities and other 
scientific and advisory bodies, and have helped to provide a consistent 
basis for national and international regulatory standards.
    In 1977 and 1979, ICRP published Report Nos. 26 and 30 (``Limits 
for Intake of Radionuclides by Workers''), respectively (Docket Nos. 
OAR-2005-0083-0087). These two reports reflect advances in the state of 
knowledge of radionuclide dosimetry and biological transport of 
radionuclides in humans that occurred over the 20 years since ICRP's 
1957 dose methodology recommendation (ICRP 2). This methodology, known 
as the effective dose equivalent (EDE) methodology, is the basis for 
dose calculations performed to demonstrate compliance with 40 CFR part 
191 and envisioned to be applied (although not specified) in the 2001 
version of 40 CFR part 197. The EDE methodology was first incorporated 
into Federal Guidance in 1987, in ``Radiation Protection Guidance to 
Federal Agencies for Occupational Exposure'' (52 FR 2822, January 27, 
1987; Docket No. OAR-2005-0083-0078).
    The basis of the EDE methodology is that each organ in the body 
reacts to radiation differently, i.e., some are more likely than others 
to react by developing a cancer. This methodology takes these 
differences into account by assigning a ``weighting factor'' to each 
organ relative to the whole body. The weighting factor reflects the 
likelihood, that is, risk, of fatal cancer developing in that organ per 
unit of dose. When added together, the risk-weighted doses incurred by 
the individual organs of the body become the ``effective dose 
equivalent.'' In this manner, the risk of radiation exposure to various 
parts of the body can be regulated through use of a single numerical 
standard.
    ICRP 26/30 uses the term ``effective dose equivalent.'' ICRP 60/72, 
which offers some improvements to the biokinetic models used in ICRP 30 
and thereupon updates the organ-weighting factors based on more recent 
scientific studies, uses the term ``effective dose.'' It may appear 
from this difference in terminology that we cannot both fulfill our 
statutory mandate and specify the use of the ICRP 60/72 factors.
    However, we do not believe this is the case. First, ICRP made it 
clear in Publication 60 that it was adopting the shorter nomenclature 
for ease of use, but did not intend to change the underlying approach 
of ICRP 26/30: ``The weighted equivalent dose (a doubly weighted 
absorbed dose) has previously been called the effective dose equivalent 
but this name is unnecessarily cumbersome, especially in more complex 
combinations such as collective committed effective dose equivalent. 
The Commission has now decided to use the simpler name effective dose, 
E'' (ICRP Publication 60, p. 7, Docket No. OAR-2005-0083-0087).
    Second, we have used the different terms interchangeably in various 
applications over the years. Historically, this concept has been 
referred to as effective dose equivalent, effective dose, and total 
effective dose equivalent, depending on when the terms were used and 
the weighting factors applied. The concept of a ``committed'' dose is 
inherent in the methodology (and recognized by ICRP, as in the previous 
citation), but we have applied the term to more explicitly acknowledge 
the continuing dose contribution over a period of years from 
radionuclides taken into the body through ingestion, inhalation, or 
absorption.
    For example, our standards in 40 CFR part 191 are written in terms 
of committed effective dose (CED). These standards were finalized in 
1993, after the publication of ICRP 60 and passage of the EnPA. At that 
time, our most recent Federal Guidance Report No. 11, ``Limiting Values 
of Radionuclide Intake and Air Concentration and Dose Conversion 
Factors for Inhalation,

[[Page 49047]]

Submersion, and Ingestion'' (EPA-520/1-88-020, September 1988, Docket 
No. OAR-2005-0083-0071), which was issued to serve as the basis for 
regulations setting upper bounds on exposures in the workplace, 
specified the ICRP 26/30 method to calculate CEDE. Appendix B of 40 CFR 
part 191 also specified use of the ICRP 26/30 weighting factors, but to 
calculate CED. Thus, we used two different (albeit similar) terms to 
represent the use of an identical methodology and associated weighting 
factors. From this, it should be clear that we have historically 
considered CED and CEDE to represent essentially the same approach, 
regardless of the weighting factors used.
    In today's proposal, we are specifying in the definition of 
effective dose equivalent in Sec.  197.2 that DOE will calculate annual 
CEDE using the radiation- and organ-weighting factors in ICRP 60/72, 
which we are proposing to be incorporated into a new Appendix A. This 
represents the most recent science and dose calculation approaches in 
the area of radiation protection, which we previously endorsed in our 
Federal Guidance Report No. 13 (``Cancer Risk Coefficients for 
Environmental Exposure to Radionuclides,'' EPA 402-R-99-001, September 
1999, Docket No. OAR-2005-0083-0072). We believe this change is 
appropriate and reflective of the direction of the international 
radiation-protection community as well as EPA's own guidance. 
Furthermore, we believe this approach is consistent with the intent and 
direction of the EnPA. The EnPA directs us to prescribe our standard 
for protection of individuals in the form of a general class of 
standards known as effective dose equivalent standards. We have done 
that by using a standard in the form of committed effective dose 
equivalent, which is a member of the class of effective dose equivalent 
standards. We request comment on this proposed change.
    Regardless of the preferences of radiation experts, the public may 
be unfamiliar with the differences between the two methods and ask 
whether a given dose level (for example, 15 mrem/yr) is equally 
protective when expressed under each method. The calculation of dose 
from individual radionuclides may be affected in different ways, 
depending on which organs they tend to affect and the pathway through 
which they enter the body. For example, consider two radionuclides that 
occur in the expected inventory at Yucca Mountain, such as technecium-
99 and neptunium-237. For a given intake, the dose from technecium-99 
is higher using the ICRP 60/72 system than it is using the ICRP 26/30 
system. On the other hand, the dose from a given intake of neptunium-
237 is lower using the latter system compared to that calculated using 
the former. However, in the majority of cases, the effect of changing 
from one system to the other is small (``Dosimetric Significance of the 
ICRP's Updated Guidance and Models, 1989-2003, and Implications for 
Federal Guidance,'' ORNL/TM-2003/207, August 2003, Docket No. OAR-2005-
0083-0070). Further, the overall factors used to convert dose to risk 
remain unchanged by today's proposal. Therefore, the estimated risk 
from a given radiation dose remains the same. Therefore, the 15 mrem/yr 
standard incorporated into today's proposal represents the same level 
of protection as the originally promulgated standards.
    We have also considered whether to allow for the use of future 
updates to the organ weighting or other factors. We believe this may be 
appropriate because DOE will continue to perform projections for many 
years, and the final demonstration before repository closure and 
license termination may be decades or even more than one hundred years 
into the future. A provision allowing such updates ensures that the 
most current science can be applied at all times. Therefore, we are 
including a provision in our proposed Appendix A allowing DOE to use, 
with NRC approval, updated dose calculation factors. We have tried in 
today's proposal to make clear the basis for our acceptance of the ICRP 
60/72 factors as sufficiently validated to be incorporated into 
rulemaking. To ensure that such factors that might be considered in the 
future have been appropriately reviewed and accepted by the scientific 
community, we propose that NRC may only approve factors that have been 
issued by independent scientific bodies (such as ICRP and its successor 
bodies) and incorporated by EPA into Federal Guidance. To ensure 
compliance with the EnPA, we would also require that the new approach 
be compatible with the effective dose equivalent methodology 
incorporated into today's proposal. We request comment on this 
approach.
    Commenters may be aware that the NAS released in June 2005 the 
latest in a series of studies on the Biological Effects of Ionizing 
Radiation (BEIR VII, Docket No. OAR-2005-0083-0087). EPA is a major 
sponsor of these studies, which we consider in developing our 
regulations and Federal Guidance. The BEIR VII report reaffirmed that 
evidence exists that even the smallest radiation dose may convey some 
risk of incurring a cancer, and that risk increases proportionally to 
the dose (i.e., if the dose doubles, the risk also doubles). This 
approach, known as the ``linear non-threshold'' hypothesis, has served 
for many years as the basis for all radiation protection regulation and 
guidance, including those issued by EPA. Further, the linear non-
threshold approach is the source of the assumptions regarding the dose-
risk relationship underlying both our 2001 rulemaking and today's 
proposal. Thus, the primary conclusion of the BEIR VII study does not 
affect the revision of our Yucca Mountain standards.
    For a detailed discussion of potential health effects related to 
exposure to radiation, as well as further explanation of the underlying 
relationship between radiation dose and cancer risk, see the preamble 
to the 1999 proposed rule (64 FR 46978-46979, August 27, 1999, Docket 
No. OAR-2005-0083-0041) and Chapter 6 of the 2001 BID (Docket No. OAR-
2005-0083-0050).

D. How Will Today's Proposal Affect the Way DOE Conducts Performance 
Assessments?

    We find it important to emphasize certain key aspects of the 
performance assessment that will apply regardless of the time frame 
involved. First, the overall purpose of our standards is to provide a 
reasonable test of disposal system performance. The overall purpose of 
the performance assessment is to provide a reasonable test for 
compliance with those standards. A major part of providing that 
reasonable test is determining which features, events, and processes 
(FEPs) are to be included in the performance assessment performed by 
DOE. Regardless of time frame, we find it reasonable to limit the 
consideration of FEPs and scenarios (sequences or combinations of FEPs) 
to those reasonably likely to occur and to affect the disposal system 
during the compliance period. Finally, in addressing those scenarios, 
it is also reasonable to further prescribe certain aspects of the way 
they are considered (``stylizing''), particularly when their 
characteristics are difficult to establish with confidence. This 
section provides an overview of the performance assessment process and 
addresses in more detail the key aspects just mentioned.
    The long-term performance of the disposal system is assessed 
through complex probabilistic computer simulations aimed at quantifying 
the behavior of the disposal system over time. The change in the 
compliance period does not affect fundamentally how the disposal system 
performance

[[Page 49048]]

assessment simulations are constructed and executed. The performance 
assessment takes into consideration the physical and chemical 
characteristics of the disposal system, and imposes on that 
characterization the events and processes expected to occur during the 
compliance period. The DOE has already conducted and published many of 
its performance assessment results focusing on periods up to 10,000 
years to support its Viability Assessment, FEIS, and site 
recommendation. While many of those projections did cover times up to 1 
million years, DOE did not focus as much attention on the assumptions 
and characterization of those longer-term processes and events, or 
necessarily conduct those projections in a way suitable for 
demonstrating compliance with a regulatory standard because there was 
no quantitative standard past 10,000 years. Today we are proposing 
certain provisions that will affect DOE's treatment of longer-term 
projections for compliance purposes, but will not alter the 
requirements issued in 2001 for compliance within 10,000 years.
    The performance assessment is developed by first compiling listings 
of features (characteristics of the disposal system, such as the 
description of the disposal system geologic setting), events (discrete 
events that can occur at the site, such as seismic events, i.e., 
earthquakes), and processes anticipated to be active during the 
performance period of the disposal system (such as corrosion processes 
operating on the metallic waste package). These items are collectively 
referred to as ``FEPs'' (features, events and processes). These FEPs 
are then used in combination to construct scenarios, which are 
essentially potential ``futures'' for the disposal system. A scenario 
describes one possible path along which the disposal system could 
evolve from the time of closure through the time of peak dose. 
Individual FEPs are components of scenarios and can be combined in 
various ways; while some FEPs, such as infiltration of water through 
the repository, will be included in nearly all scenarios, low-
probability FEPs may appear in only a few. Thus, a scenario can be 
visualized as a time history for the disposal system, beginning, for 
example, with precipitation over Yucca Mountain and water infiltration 
into the subsurface above the repository, and ending with a dose 
assessment for the down-gradient RMEI making use of the ground water 
moving from beneath the site. Natural parameter variations (such as 
differing ground-water movement rates through the repository and in the 
aquifers below the repository) give rise to many potential ``futures'' 
for a particular scenario, depending on the exact parameter values 
chosen from the distribution of possible values, for each computer 
simulation of repository performance. For ease of calculations, 
scenarios with similar characteristics may be grouped into scenario 
classes. More extensive descriptions of the scenarios used to assess 
disposal system performance for Yucca Mountain are detailed in DOE 
documents supporting such analyses for various purposes (see the 
Viability Assessment, DOE/RW-0508/V.3, Vol. 3, Chapter 1.3, December 
1998, Docket No. OAR-2005-0083-0086, and the Science and Engineering 
Report, DOE/RW-0539, Chapters 4.3 and 4.4, May 2001, Docket Nos. OAR-
2005-0083-0069).
    Scenarios have differing probabilities, depending on the likelihood 
of particular FEPs included in them. The dose results calculated for 
individual scenarios are weighted as a function of their probability to 
develop an overall distribution of doses with time that is the final 
product of the analyses. From this distribution of doses, compliance 
with the regulatory standard is determined in the licensing process.
    In considering how to approach assessments potentially out to 1 
million years, we have considered international consensus on the 
qualitative nature of such calculations. Although also true at the 
10,000-year time frame, for peak dose it is even more evident that the 
performance assessment cannot be viewed as a predictor of future events 
and resultant releases. Instead the goal is to design an assessment 
that is a reasonable test of the disposal system under a range of 
conditions that represent the expected case, as well as relatively less 
likely (but not wholly speculative) scenarios with potentially 
significant consequences. The challenge is to define the parameters of 
the assessment so that they demonstrate whether or not the disposal 
system is resilient and safe in response to meaningful disruptions, 
while avoiding extremely speculative (and in some cases, fantastical) 
events. As NAS notes, ``The results of compliance analysis should not 
be interpreted as accurate predictions of the expected behavior of a 
geologic repository'' (NAS Report p. 71).
    We recognize that many uncertainties can be bounded, and methods 
exist to take these uncertainties into account in evaluating compliance 
of the disposal system. Examples include the use of cautious, but 
reasonable, parameter values and assumptions that ensure the models err 
on the side of conservatism, and the use of probabilistic models in 
order to explore the range of possibilities of total system evolution. 
We further recognize that it can be difficult to determine when 
conservatism is appropriate and when it is excessive. However, as 
discussed earlier in this preamble, we are concerned that systematic 
conservatism in the face of uncertainties is inconsistent with the 
concept of reasonable expectation embodied in our standards. This view 
is also shared at the international level. A joint report by the IAEA 
and the NEA concludes that ``[w]hen uncertainty exists there is a 
tendency to skew the model or values of parameters towards 
conservatism,'' which ``results in embedded conservatism'' (``An 
International Peer Review of the Yucca Mountain Project TSPA-SR,'' p. 
52, 2002, Docket No. OAR-2005-0083-0062). However, those aspects of the 
disposal system and waste behavior that depend upon physical and 
geological properties can be estimated within reasonable limits of 
uncertainty.
    Still, ``[e]ven in the initial phase of the repository lifetime, a 
compliance decision must be based on a reasonable level of confidence 
in the predicted behavior rather than any absolute proof'' (NAS Report 
p. 72). For performance projections made past 10,000 years, the 
confidence that can be placed in those projections decreases as time 
increases. While NAS indicated that analyses of the performance of the 
Yucca Mountain system dealing with the far future can be bounded, ``the 
uncertainties in some of the calculations that might be required could 
render further calculation scientifically meaningless'' (NAS Report p. 
29). What is more, a different panel convened by NAS has recently 
stated that uncertainties often become so large that the results of a 
risk assessment must be deemed indeterminate (``Risk and Decisions 
About Disposition of Transuranic and High-Level Radioactive Waste,'' 
NAS, p. 91, 2005, Docket No. OAR-2005-0083-0060). Regarding natural 
processes and/or events that can occur during a large period of time, 
it becomes necessary to restrict the scenarios available to include in 
a performance assessment by not including events or processes that have 
a nearly negligible probability of occurrence over the period of 
geologic stability, or that introduce additional uncertainty without 
providing significantly new or different information about the 
performance of the disposal system.
    It is neither useful nor necessary for EPA to require DOE to 
predict or model every conceivable scenario that could occur at Yucca 
Mountain. Rather, we

[[Page 49049]]

believe that it is our responsibility to design a reasonable test of 
the disposal system's performance over a very long time period. This 
implies that some possible performance scenarios should not be included 
in the performance assessment because their probability of occurrence 
is extremely low. As a means of restricting scenarios, in setting the 
standards in 40 CFR part 197, the Agency outlined how to identify FEPs. 
For purposes of the performance assessment, the value of considering a 
particular FEP (or series of FEPs) diminishes if either its likelihood 
of occurrence or its potential consequence is insignificant. Therefore, 
a time frame and probability cut-off measure are needed to limit the 
range of FEPs that could be included as candidates for the performance 
assessment. Without such measures, the list of FEPs would be limitless, 
bounded only by the imagination. The Agency determined that FEPs that 
could occur with a probability equal to or greater than 1 in 10,000 
over a period of 10,000 years would be sufficiently likely to occur, so 
that they should be included among the FEPs available for selection in 
any particular scenario. FEPs with lower probabilities could be 
excluded from the analyses. This probability limit represents an annual 
probability of occurrence of 10-\8\ (1 in 100 million). This 
means, for example, an event with this minimum probability has only a 
one-hundredth of one percent chance of happening in any given 10,000-
year period. This is an extremely conservative screening criterion. 
Extending the regulatory compliance period to as much as 1 million 
years and maintaining the annual probability cut-off of 
10-\8\ would still mean that FEPs with only a one percent 
chance of occurring over this time period must be considered. This 
probability cut-off for screening FEPs for inclusion in the disposal 
system performance assessment provides a robust test of compliance, in 
that even FEPs with very low probabilities are not a priori excluded 
from the assessments.
    Given the conservative nature of this low probability cut-off for 
initial FEPs screening efforts, the application of the screening 
criteria still produce a large number of scenarios that could be 
postulated, presenting perhaps an unmanageable task for the analyses 
and ultimately the regulatory compliance decision. In the generic rule 
for geologic disposal, 40 CFR part 191, and the 2001 rule for Yucca 
Mountain, we provided a means to manage the situation, by allowing 
individual FEPs or scenarios to be deleted from the licensing 
performance assessment if they contribute little to the dose received 
by the RMEI, i.e., their consequences are low--either due to the low 
probability of the FEPs or the low doses calculated for the scenario. 
In extending the regulatory performance period in the regulation to the 
time of peak dose, a similar provision aimed at managing the scope of 
the analyses is called for.
    The need to maintain the assessment within a reasonable scope as a 
way to manage uncertainties leads us to conclude that a strict 
extension of the approach for 10,000-year assessments would not 
accomplish this overall goal. If, for example, we required 
consideration of events with a probability of occurrence of 
10-\4\ over 1 million years `` an approach that has been 
suggested by some stakeholders `` it would equate to an annual 
probability of 10-\10\ (one in 10 billion), which 
encompasses events nearly as remote as the ``big bang'' that created 
our universe. No disposal system, and perhaps not even our planet 
itself, would be expected to survive the effects of such an event, and 
we therefore do not find it a useful indicator to distinguish between 
safe or unsafe performance of the disposal system. There are an 
unlimited number of possible futures, some of which would involve risks 
from a repository and others that would not. We must balance these 
factors to ``define a standard that specifies a high level of 
protection but that does not rule out an adequately sited and well-
designed repository because of highly improbable events'' (NAS Report 
p. 28).
    In addition, NAS recommended ``against an approach under which a 
large number of future scenarios are specified for compliance 
assessment, since such an approach could be seen as putting both the 
regulator and the applicant in the indefensible position of claiming to 
have considered a sufficient number of scenarios and that all 
reasonable future situations are represented in the analysis'' (NAS 
Report p. 98). NAS explicitly recognized that ``[i]t is important that 
the `rules' for the compliance assessment be established in advance of 
the licensing process; that is, that the scenarios that might be 
excluded from the integrated risk analysis be identified'' (NAS Report 
p. 73). We emphasize that the purpose of making exposure scenario 
assumptions is not to identify exhaustively every possible future, but 
to construct a reasonable (or, as NAS put, a ``fair'') test of disposal 
system performance for the protection of public health. This is the 
case regardless of the time frame involved, and from that perspective 
today's proposal will not alter the way in which DOE will approach its 
performance assessments.
    In addition to placing limits on the probability of FEPs that 
should be considered, an additional tool to construct the test (or set 
``the `rules' for the compliance assessment,'' as NAS stated) is to 
specify how certain scenarios should be assessed. This ``stylizing'' of 
scenarios is similar to the approach we took (and NAS recommended) to 
defining the human-intrusion scenario. In a more general sense, NAS 
acknowledged that establishing the ``rules'' ``requires using the 
rulemaking process to arrive at a regulatory decision about certain 
assumptions as part of the standard'' (NAS Report p. 34). The NEA has 
also recommended exploring the possibility of using a similar stylized 
approach to address uncertainties in the evolution of the surface 
environment and the nature of future human actions (``The Handling of 
Timescales in Assessing Post-Closure Safety,'' pp. 22-23, 2004, Docket 
No. OAR-2005-0083-0046). This approach would avoid speculation 
regarding the evolution of the geologic environment at times when the 
hazards associated with the waste are reduced compared to when the 
waste is emplaced.
    Stylized approaches can be utilized to address associated 
uncertainties in order to allow consideration of events that are deemed 
potentially important to performance but whose characteristics are 
difficult to establish with certainty. There is international consensus 
that this approach may be used to define assumptions that are too 
difficult to bound (NEA, p. 22, Docket No. OAR-2005-0083-0046). This 
approach could therefore be used for the determination of the evolution 
of the geological environment over long periods. As noted above, this 
approach is similar to that recommended by NAS, and utilized by EPA in 
examining human intrusion (NAS Report p. 108). The NAS determined that 
it was technically infeasible to assess the probability of human 
intrusion into a repository over the long term. It concluded that it 
was not scientifically justified to incorporate a myriad of alternative 
scenarios of human intrusion into a fully risk-based compliance 
assessment that requires knowledge of the character and frequency of 
various intrusion scenarios. Accordingly, NAS recommended that we 
specify in our standards a typical intrusion scenario to be analyzed 
for its consequences on the performance of the repository. The intent 
of this ``stylized scenario'' is to avoid non-productive speculation on 
the forms and frequencies of intrusion that can never be predicted, 
while

[[Page 49050]]

allowing the ``robustness'' of the containment properties of the 
repository to be evaluated by a scenario that is plausible, and 
potentially causes some levels of exposure. The same factors must be 
balanced in considering how to assess key geologic and other features 
over very long time frames when it is exceedingly difficult to 
establish exact parameters--or even distributions of parameter values--
with any certainty.
    The modifications proposed in Section II.C (``How is EPA Proposing 
to Revise the Individual-Protection Standard to Address Peak Dose?'') 
would require DOE to project exposures to the RMEI until the time of 
peak dose and subject them to a compliance determination. The key 
aspects emphasized at the beginning of this section guide our 
requirements for the scope of performance assessments both at 10,000 
years and over times extending through the entire period of geologic 
stability. However, their implementation carries different implications 
for those different time periods, given the nature of uncertainties and 
the types of events that can be envisioned to occur. To address these 
implications, we are proposing four provisions that will affect the 
judgment of compliance when that judgment is extended to periods up to 
1 million years. Specifically, we are proposing:
     A separate compliance standard for the peak dose beyond 
10,000 years;
     That compliance beyond 10,000 years be demonstrated using 
the median of the distribution of results;
     That FEPs and scenarios not included in the 10,000-year 
analysis because of their limited consequence during that period need 
not be considered in the peak dose calculations;
     That scenarios involving climate change, seismic activity, 
igneous activity, and general corrosion be explicitly considered in the 
peak dose calculations.
    We have already discussed the peak dose standard and the use of the 
median to demonstrate compliance (see Sections II.C.3 and II.C.5). The 
selection of FEPs (including general corrosion) is discussed in detail 
in Section II.D.2.a (``Consideration of Likely, Unlikely, and Very 
Unlikely FEPs''). Discussion of climate, seismic, and igneous scenarios 
is included in Sections II.D.2.b, c, and d, respectively.
1. Performance Assessments Up To 10,000 Years After Disposal
    Our 2001 rulemaking established a framework within which DOE would 
conduct its performance assessments to show compliance with the 10,000-
year standard. The previous section touched on various aspects of this 
framework. Essentially, the performance assessment involves three basic 
steps: (1) Identify the FEPs and scenarios that might affect the Yucca 
Mountain disposal system, along with their probabilities of occurrence; 
(2) examine the effects of those FEPs and scenarios on disposal system 
performance; and (3) estimate the dose consequences from those FEPs and 
scenarios, weighted by their probabilities of occurrence. Today's 
proposal will not affect this framework in any way.
    We supplemented this basic framework with two additional 
provisions. The first, the underlying principle of reasonable 
expectation, we have discussed in detail in Sections II.A.4 and II.B. 
The other important provision, touched on in the previous section, 
establishes the approach to identifying FEPs and scenarios and their 
probability of occurrence. We specified that FEPs or scenarios with a 
probability of occurrence lower than 1 in 10,000 over 10,000 years need 
not be considered in the performance assessment. FEPs or scenarios with 
a higher probability of occurrence also need not be considered if they 
would not significantly change the results of the performance 
assessment. We are not proposing to alter this provision as it applies 
to the 10,000-year standard. The standards in 40 CFR part 191 (the EPA 
regulation that addresses geologic disposal generically) also used this 
formulation as the means of determining FEPs for any prospective 
disposal system. In developing 40 CFR part 197 in 2001, the Agency 
determined that there was no reason, on a site-specific basis, to 
depart from this conservative screening criterion. We also note that 
NAS endorsed this same probability level in its specific discussion of 
volcanism, and suggested that such a level ``might be sufficiently low 
to constitute a negligible risk [of occurrence]'' (NAS Report p. 95). 
Probabilities below this level are associated with events such as the 
appearance of new volcanoes outside of known areas of volcanic activity 
or a cataclysmic meteor impact in the area of the repository. We 
believe there is little or no benefit to public health or the 
environment from trying to regulate the effects of such very unlikely 
events.
2. Performance Assessments for Periods Longer Than 10,000 Years After 
Disposal
    As discussed in the previous sections, we do not believe that DOE's 
performance assessments need be changed fundamentally to accommodate an 
extended compliance period. The general framework described in the 
previous section applies equally well to periods beyond 10,000 years, 
although we are proposing specific provisions to apply to this longer 
period. We recognize, however, that there may be some confusion 
regarding the conduct of assessments to demonstrate compliance at two 
different times. DOE will not necessarily conduct one set of 
assessments to show compliance with the 10,000-year standard, and a 
separate set of assessments to show compliance with the peak dose 
standard applicable at times beyond 10,000 years. Rather, DOE's overall 
approach could be to run its dose assessments from the time of facility 
closure to the end of the period of geologic stability (1 million years 
after closure). The FEPs and scenarios selected for each individual run 
would continue to operate, and the disposal system to evolve, over that 
entire time period. DOE would extract the results necessary for 
comparison with our regulatory standards.
    As it is with the 10,000-year standards, the main purpose of the 
post-10,000-year standards is to provide a reasonable test of the 
performance of the disposal system. The NAS stated it another way: 
``The challenge is to define a standard that specifies a high level of 
protection but that does not rule out an adequately sited and well-
designed repository because of highly improbable events'' (NAS Report 
p. 28).
    In formulating our approach to an extended compliance period, we 
began by reviewing the NAS report. NAS concluded that several gradual 
and episodic natural processes or events have the potential to modify 
the properties of the repository and the processes by which 
radionuclides are transported. NAS concluded that the probabilities and 
consequences of modifications generated by volcanic eruptions 
(volcanism), seismic activity, and climate change are sufficiently 
boundable so that these ``modifiers,'' as it termed them, can be 
included (along with an undisturbed scenario) in performance 
assessments that extend over the expected period of geologic stability 
(on the order of 1 million years) in the Yucca Mountain region (NAS 
Report p. 91). NAS considered the ``long-term stability of the geologic 
environment at Yucca Mountain'' to describe the situation where 
geologic processes such as earthquakes (and similar physical and 
geological processes that could affect the performance assessment at 
the Yucca Mountain site) are sufficiently quantifiable and the related

[[Page 49051]]

uncertainties boundable that the performance can be assessed (NAS 
Report p. 67). Furthermore, NAS acknowledged that, conceptually, there 
is a need for screening criteria to distinguish significant FEPs from 
those that can be considered to have negligible effects (NAS Report, 
for example, pp. 59, 61, 72, 95, 98). NAS suggested that certain levels 
(including a probability cut-off of 10-8 per year) might be 
appropriate, but made no recommendation on this issue.
    We believe the three categories of FEPs identified by NAS deserve 
special attention. We will require that DOE consider these FEPs in its 
long-term projections. However, we are proposing to apply the same 
overall probability threshold and handle such events in a stylized 
manner to address only their most significant effects. In essence, DOE 
must include such FEPs in the peak dose assessment, but need not assess 
in detail every conceivable variation on those events. Thus, our 
approach would require that DOE develop reasonable igneous, seismic and 
climate change scenarios and assess the most likely and significant 
impacts, with appropriate variability in its assumptions, on dose 
projections. The NAS did not identify any other ``modifiers'' that it 
expected could be addressed in a quantitative risk assessment covering 
the period of geologic stability. In addition, NAS specifically 
mentioned potential effects of these modifiers, but also noted that, 
while possible, many of these effects would be so unlikely or limited 
that they would not be expected to significantly affect disposal system 
performance (NAS Report pp. 91-95). These igneous, seismic, and 
climatological FEPs are discussed in more detail in the following 
sections. We propose to specify certain significant aspects or 
characteristics of the event or process to which DOE may limit its 
analyses, and DOE will assess reasonable variations within those 
bounds, considering such basic assumptions as severity and time of 
occurrence. DOE must then evaluate the consequences on the disposal 
system and resulting impacts to the RMEI. By varying the time of 
occurrence within the probability framework, DOE can also address the 
effects of these FEPs on the peak dose.
    Having identified particular natural FEPs that should be considered 
throughout the period of geologic stability, we then considered whether 
there are FEPs affecting the engineered barrier system that should also 
be identified. In reviewing DOE's published TSPAs and other relevant 
information, we conclude that general corrosion of the waste packages 
has been shown to be a potentially significant failure mechanism at 
times in the hundreds of thousands of years (Yucca Mountain Science and 
Engineering Report, DOE/RW-0539, Section 4.2.4, May 2001, Docket No. 
OAR-2005-0083-0069). Unlike certain other corrosion processes, as 
discussed in the next section, which may be more likely or faster-
acting at earlier times, general corrosion may not be a significant 
factor within 10,000 years and could potentially be removed from 
consideration at those times because of its limited consequence. Were 
we simply to state that FEPs not included in the 10,000-year analyses 
should not be included in the post-10,000-year analyses, there might be 
some question as to whether DOE would need to consider general 
corrosion at all. We believe it has been shown potentially to be of 
sufficient importance that it should be included in those projections. 
Therefore, we are proposing to remove any ambiguity by specifying that 
DOE must consider general corrosion in its projections throughout the 
period of geologic stability.
    In general, we continue to believe that it is reasonable to require 
DOE to exclude from performance assessments those FEPs whose likelihood 
of occurrence is so small that they are very unlikely, or whose 
consequence is minimal, as described above. We propose that this 
probability threshold as expressed in our 2001 rule for the 10,000-year 
compliance period be extended throughout the entire period to peak dose 
(i.e., FEPs included in the 10,000-year assessments are included in the 
assessments beyond 10,000 years), but with the inclusion of the long-
term impacts of seismicity, volcanism, and long-term climate change, as 
consistent with the probability screening criteria described herein 
(NAS Report p. 94). These are the natural events and processes that NAS 
determined were reasonably boundable when compliance time frames at 
Yucca Mountain are extended out to the period of geologic stability. We 
also propose that DOE must consider the long-term effects of general 
corrosion on the engineered barriers, particularly on waste package 
integrity. This is an extremely inclusive standard. It captures 
significant events in the life of the repository, and yet is not so 
restrictive that no repository could ever pass, given that there would 
be no limit to the speculation of scenarios that could occur during the 
period of geologic stability.
    As discussed further in the following sections, we have examined a 
variety of events and feel confident that the screening analysis for 
10,000 years--with the assurance that seismic, igneous, climate change, 
and general corrosion scenarios are included--includes the appropriate 
range and severity of FEPs to also serve as a reasonable test of 
disposal system performance throughout the period of geologic 
stability. We have not (and have not claimed to) conducted an 
exhaustive or detailed analysis of variations or permutations of 
scenarios out to the time of peak dose. In fact, this is precisely the 
sort of unrestrained and speculative exercise we wish to avoid. We 
recognize that some commenters may believe it is appropriate to 
consider whether further analysis or new data could reveal that an 
event excluded from the 10,000-year screening is important to 
performance of the disposal system over the geologic stability period. 
As discussed later, we do not believe such scenarios are either very 
likely or very important to performance. Nor do we believe that this 
approach inappropriately constrains NRC, as the licensing authority. 
Rather, we consider this approach to be consistent with the NAS 
position that conducting compliance assessments ``requires using the 
rulemaking process to arrive at a regulatory decision about certain 
assumptions as part of the standard'' (NAS Report p. 34).

a. Consideration of Likely, Unlikely, and Very Unlikely FEPs

    Our individual-protection standards (Sec.  197.20) as promulgated 
in 2001 required DOE to consider in the performance assessment FEPs 
with a one in 10,000 or greater chance of occurring during 10,000 
years. FEPs below this probability threshold are considered ``very 
unlikely'' and can be discounted based on probability alone. We also 
allowed NRC and DOE to remove from consideration FEPs with a higher 
probability if their effects on performance assessment results were 
determined to be insignificant. In addition, performance assessments 
conducted to show compliance with the human-intrusion and ground-water 
protection standards may exclude FEPs considered ``unlikely.'' We 
specified that NRC was to determine the probability below which FEPs 
would be considered unlikely. NRC set that figure at a probability of 
occurrence of 1 in 10 over 10,000 years (equivalent to an annual 
probability of 10-5) (67 FR 62634, October 8, 2002, Docket 
No. OAR-2005-0083-0059).
    In extending the period of compliance, we must consider whether our 
threshold for probability screening

[[Page 49052]]

of ``very unlikely'' FEPs remains appropriate. We believe it does, and 
are proposing to retain it for the extended compliance period. While we 
are retaining the compliance standard of 150 [mu]Sv/yr (15 mrem/yr) 
applicable to 10,000 years, we are also proposing to introduce a second 
compliance standard of 3.5 mSv/yr (350 mrem/yr) for the peak dose 
beyond 10,000 years, which could potentially apply up to 1 million 
years. This may lead some commenters to suggest that the formulation 
for FEPs screening should simply be extended by two orders of 
magnitude, i.e., that very unlikely FEPs would have less than a one in 
10,000 chance of occurring over 1 million years. This would recognize 
that very low-probability FEPs would become more likely to be seen 
simply with the passage of time (essentially by looking at many 10,000-
year periods, the cumulative probability, rather than annual 
probability, would be increased). However, in our view, such a 
formulation would be unjustified and unreasonable.
    It is important to consider the real meaning of these probability 
thresholds. A FEP screened in at the existing lower probability 
threshold would have only a 0.01% chance of occurring through 10,000 
years, yet still must be included in the FEPs considered for the 
performance assessment. We question, then, whether the effort involved 
in incorporating even less likely events into the ``FEP pool,'' with 
the level of speculation likely to be attached to them, would be 
rewarded with even minimal contribution to safety.
    The threshold for very unlikely events suggested by NAS was an 
annual probability of 10-8 (1 in 100 million per year), 
which NAS equated to 1 in 10,000 over 10,000 years, stating that this 
level ``might be sufficiently low to constitute a negligible risk'' 
(NAS Report p. 95). We consider these two expressions to be 
functionally equivalent (and have explicitly included both in our 
proposal today), but adopted the latter as more clearly tied to the 
10,000-year compliance period. Even though the NAS statement above was 
referring to volcanism, we believe that this probability threshold is a 
generic consideration that is applicable to any risk at Yucca Mountain, 
not just volcanism. If one extends the time period of the assessment to 
1 million years, a FEP at this level would still have only a 1 in 100, 
or 1%, chance of occurring within that time, but would still be 
considered in the performance assessment process. We believe this is a 
``cautious, but reasonable'' level, especially when considering the 
confounding effects of uncertainties at such long time periods. In 
fact, we are unaware of any international precedents for scrutinizing 
FEPs of this low probability. Thus, we are proposing to retain the 
10-8 annual probability threshold for very unlikely FEPs for 
both the 10,000-year and post-10,000-year assessments.
    Application of this screening criterion deserves some additional 
discussion. For FEPs involving the natural barrier, an annual 
probability of 10-8 theoretically indicates that to compile 
a definitive list of all FEPs involving the natural barrier, the 
geologic record at the site would have to be examined back to a time 
frame of 100 million years to identify FEPs that would be projected to 
occur at least once in that time period. For the Yucca Mountain site, 
the volcanic rocks containing the repository are only on the order of 
10 million years in age, indicating that essentially any FEP that could 
be identified in the geologic record during the 10 million year time 
frame would have an annual probability higher than 10-8, and 
would be included in the list of FEPs for scenario construction. We 
believe that the Quaternary period, extending back approximately 2 
million years, is a sufficiently long period of the geologic record to 
allow DOE to make reasonable estimates of natural FEPs (see 66 FR 
32100). Observed FEPs from that period, as well as other that can be 
inferred, would be included in a 10-8 cut-off.
    For FEPs involving the engineered barrier, a similar logic applies. 
However, the ``record'' to be examined to identify FEPs for the 
performance of man-made materials and systems is much shorter than the 
geologic record. Application of the 10-8 annual limit 
ensures all relevant FEPs are considered for inclusion. For example, 
corrosion processes for which there is accelerated testing and analog 
information at longer time frames, could still be included in scenario 
development. Even when such processes would have a low probability, the 
conservative probability cut-off threshold would still assure they are 
considered in scenario development. For such processes, however, when 
probabilities of occurrence over long times may be difficult to assign, 
the decision to consider them may be based solely on consequence.
    By contrast, were we to stretch the probability threshold by two 
orders of magnitude, to an annual probability of 10-10 (one 
in 10 billion per year), we would be introducing an unprecedented level 
of conservatism into the performance assessments. At such a level, the 
performance assessment would be required to consider geologic events 
likely to have never happened, since the age of the Earth itself is 
estimated at about 4.5 billion years (http://pubs.usgs.gov/gip/geotime/age.html). Further, an event of this annual probability will not reach 
even a 50% cumulative probability for another 500 million years (a 
total of 5 billion years), or 500 times the period of geologic 
stability at Yucca Mountain (defined by NAS as on the order of 1 
million years). A probability threshold at that level would sweep in 
cataclysmic volcanic and seismic events, as well as meteor impacts of 
the type that extinguished the dinosaurs 65 million years ago. We 
simply find it inconceivable that such events could be considered a 
reasonable test of the repository, or that requiring them to be 
analyzed would provide any benefit to public health and safety. To look 
at it another way, an event at our current probability threshold of one 
chance in 100 million per year would still be likely to occur only a 
few times over an incremental 500 million year period, and roughly 50 
times over the entire history of the earth, of which humans have been 
present only 0.0001% of the time. Examining the geologic record at the 
Yucca Mountain site for such a time period to identify FEPs would not 
be meaningful. Even looking at the geologic record with the 
10-8 probability is challenging. In fact, the volcanic rocks 
that contain the repository were formed by very extensive volcanism 
over an area of thousands of square kilometers. Using the annual 
probability figure alone, it can be argued that such extensive 
volcanism should be included in the list of FEPs for the performance 
assessment. We strongly disagree. As emphasized by NAS, we reasonably 
must confine ourselves to assessing performance of the existing 
geologic setting. To remove such extreme assumptions, we addressed this 
particular difficulty by recommending the geologic record through the 
Quaternary (a period of approximately 2 million years) as the basis for 
identifying FEPs for the performance assessment (66 FR 32100). Based on 
this period as compared to the probability threshold we have 
established, DOE must consider for its performance assessments events 
that can be shown or reasonably inferred to have occurred during the 
Quaternary, based on the physical conditions of the site and disposal 
system.
    If the same probability threshold applies at all times, as we are 
proposing, then the FEP screening performed by DOE for its 10,000-year 
projections would be expected to adequately represent those longer time 
periods. We

[[Page 49053]]

believe it will, and do not believe it should be necessary for DOE to 
re-examine its results to ``screen in'' FEPs it has previously analyzed 
and rejected, or FEPs that might be expected to be more probable at 
longer times, if such exist. Further, our view is that it would be an 
endless task for DOE to analyze every FEP postulated to occur several 
hundred thousand years into the future, simply because a scenario can 
be invented to support it. Even if DOE were to exhaustively pursue each 
nominated FEP, their effects are likely to be minimal at best, 
especially when considering what are likely to be the much larger 
effects of increasing uncertainties and large-scale scenarios such as 
climate change. It should be clear, however, that FEPs selected for the 
analysis will continue to unfold as the assessment continues, up to 1 
million years. That is, for all FEPs included in the 10,000-year 
analysis, DOE must project the effects of these FEPs continuing to 
evolve over the course of the period of geologic stability, and account 
for their contributions to the peak dose.
    If we are starting from the basic screening for 10,000 years, it is 
reasonable to examine the reasons why FEPs might have been excluded 
from that screening when considering whether any warrant further 
evaluation in the post-10,000-year performance analysis. We see three 
general categories of FEPs (as opposed to the more specific seismic, 
igneous, and climatic FEPs, which are addressed separately in the 
following sections of this document) that may have been eliminated from 
the full analysis:

FEPs Screened Out by Probability

    The first category consists of FEPs determined to be ``very 
unlikely'' to occur. As described above, these are FEPs that would have 
a chance of occurrence of less than one in 10,000 over 10,000 years, or 
an annual probability less than 1 in 100 million (10-8). We 
see no reason to re-consider FEPs removed from the assessment based on 
this criterion. Such a FEP would have to be more likely to occur at 
some time in the future than it is now. This does not simply mean that 
the cumulative expectation of an event or process having already 
occurred is higher as time extends from 10,000 to 1 million years, 
which would be the case for any low-probability FEP; rather, it means 
that the probability itself would have to be higher at some later time 
(for example, 10-9 annual probability until year 50,000, 
then a 10-8 probability thereafter). We have not identified 
natural FEPs that would be very unlikely for the first 10,000 years, 
but would rise above that threshold within the period of geologic 
stability (FEPs whose probability of occurrence is related to the 
condition of the engineered barrier system are discussed later in this 
section). It may be argued that a FEP may become more likely if certain 
other FEPs have altered the site's characteristics in a particular way. 
As a basis for requiring additional FEP screening, we would find such a 
claim to be unreasonable and highly speculative. FEP probabilities are 
derived in large part from examinations of the historic geologic and 
climatic record going back millions of years. We suggested that the 
Quaternary period might be an appropriate benchmark for such an 
examination (66 FR 32100). Probabilities derived from such evaluations 
are not amenable to that level of fine-tuning. Furthermore, DOE has 
currently included FEPs which are at the boundary of the 
10-8 threshold, such as volcanic events (estimated at 1.6 x 
10-8). We would not view such an exercise as useful or of 
value in the licensing process. We do not believe it is necessary or 
appropriate for NRC to re-consider the probability criterion.

FEPs Screened Out by Consequence Within 10,000 Years

    Our 2001 standards allow NRC to eliminate FEPs whose effects would 
not significantly change the performance assessment results within 
10,000 years. We are proposing today to take the same approach to the 
peak dose projections, giving special attention to changes to the 
magnitude of the peak dose. There is no reason for DOE to re-consider 
FEPs for their effects on the 10,000-year projections, and we are aware 
that some FEPs have been included whose effects are manifest at times 
slightly beyond 10,000 years to give perspective on the shorter-term 
evolution of the disposal system, such as slower-acting corrosion 
mechanisms.
    At issue, then, would be FEPs whose effects might not be evident or 
as prominent until several tens or hundreds of thousands of years have 
passed. Such FEPs might be considered to be either gradual, continuing 
processes or episodic, disruptive events and processes. In general, we 
believe that the 10,000-year assessments should adequately address the 
more gradual processes and that the more significant of those processes 
have been included in those assessments (for example, infiltration of 
water through the repository and the processes leading to early failure 
of waste packages heavily influence the 10,000-year assessments and 
would do the same for peak dose projections). By the time those more 
gradual processes would take effect, it is likely that the effects of 
other processes would already be felt at a much higher level. One 
fundamental purpose of probabilistic performance assessment is to give 
proportionate emphasis to highly improbable events and processes. With 
one exception (discussed below), we find it unlikely that any gradual, 
continuing processes not already included through the screening for the 
10,000-year assessments under our proposed rule could significantly 
affect the projections over such long time periods. It is more likely 
that their effects would be overwhelmed by other, higher-probability 
(or faster-acting) processes operating over the same period.
    The single such slow-acting process we have decided to include in 
today's proposal is general corrosion of the engineered barriers, 
particularly its effects on the waste packages. We recognize that DOE 
has included general corrosion in its previous analyses for both the 
10,000-year period and over the longer term. However, even though 
general corrosion is significant to performance at longer times, it 
might reasonably be considered insignificant within the first 10,000 
years and could, thus, be screened out of the analysis to demonstrate 
compliance with the 10,000-year standard. Under our overall approach, 
were DOE to exclude general corrosion on the basis of consequence 
within the first 10,000 years, longer-term projections could also 
exclude this factor. We think such an exclusion over the period of 
geologic stability would ignore a crucial factor in long-term 
performance at Yucca Mountain. As we have noted, DOE's own analyses 
point to general corrosion as the dominant waste package failure 
mechanism, either alone or in combination with disruptive events 
(igneous events are assumed to be less dependent on prior degradation 
of waste packages). Without general corrosion assumed to act, a large 
proportion of the waste packages could be assumed to remain intact even 
up to or beyond 1 million years. Other corrosion mechanisms, such as 
localized corrosion, are highly correlated with temperature and would 
be expected to operate early in the assessment period, when 
temperatures inside the repository are likely to be very much higher. 
Stress-corrosion cracking is another mechanism that is somewhat 
correlated with temperature, but is of more importance in situations 
involving mechanical failure, such as rockfall resulting from seismic 
events. Their longer-term impact is likely to be

[[Page 49054]]

greatly reduced after the repository cools. The same is not true for 
general corrosion. The rate of general corrosion is somewhat influenced 
by temperature, but this process is expected to continue even when the 
temperature is lower. Our proposed approach would eliminate any 
questions regarding whether general corrosion should be considered for 
the longer-term assessments.
    Although general corrosion was not called out by NAS, as were the 
three natural FEPs, we believe this approach to general corrosion is 
consistent with NAS's overall expectations for the evolution of the 
disposal system. We have already discussed in the context of 
uncertainty NAS's expectation that a significant proportion of the 
waste packages would fail over the period of geologic stability and 
that, while peak doses might occur much later, significant releases 
could be anticipated within the first 10,000 years (see Section II.A.5, 
``Effects of Uncertainty''). For example, NAS suggested that some 
uncertainties will be lower ``when enough time has passed that all of 
the packages will have failed'' (NAS Report p. 29-30); that 
``uncertainties in waste canister lifetimes might have a more 
significant effect on assessing performance in the initial 10,000 years 
than in performance in the range of 100,000 years'' (NAS Report p. 72); 
that ``[d]etailed estimates of time for canister failure are less 
important for much longer-term estimates of individual dose or risk'' 
(NAS Report p. 85); and that ``[i]nflow of air through failed canisters 
and oxidation of waste prior to infiltration of water * * * would 
probably affect estimates of 10,000-year cumulative releases more than 
estimates of longer-term doses and risks'' (NAS Report p. 86). Further, 
NAS clearly identified corrosion as the dominant process leading to 
waste package failure and recognized its importance in projecting peak 
dose: ``Radionuclide releases from an undisturbed repository * * * can 
occur through * * * degradation and failure of the waste canister 
through corrosion'' * * *'' (NAS Report p. 26--see also pp. 68, 82, 
85). We also believe our proposed approach to general corrosion is 
consistent with both NAS's advice to use ``cautious, but reasonable'' 
assumptions and our principle of reasonable expectation, as general 
corrosion represents a potentially significant failure mechanism 
leading to radionuclide releases.
    Regarding natural FEPs, we are proposing that DOE explicitly 
evaluate the effects of seismic, volcanic, and climatological FEPs in 
its assessments beyond 10,000 years, as discussed in the following 
sections. It should be understood, however, that these FEPs may also be 
considered within the 10,000-year period if warranted by probability or 
consequence. The probabilities of seismic and igneous events beyond 
10,000 years will be the same as those probabilities within 10,000 
years. Events that DOE judges fall below the 10-8 
probability threshold need not be included in either the 10,000-year or 
post-10,000-year assessments. Such events might include seismic 
episodes above a certain magnitude. There is more certainty that the 
climate will experience significant changes over the period of geologic 
stability, and therefore we require it to be considered at all times. 
The effects of climate change on Yucca Mountain's performance, however, 
are likely to be minimal within 10,000 years, and potentially more 
significant at longer times when most of the waste packages are 
breached.

FEPs Screened Out by Condition of the Engineered Barrier System Within 
10,000 Years

    We are aware that DOE has identified certain FEPs that were 
eliminated from consideration within 10,000 years because it was deemed 
impossible for them to occur while the engineered barrier system 
remains intact. We believe such FEPs should be considered as a special 
case, as they depend on the condition of the engineered barrier system 
rather than a strict probability of occurrence.
    The prime example of the FEPs in this category is in-package 
nuclear criticality. The possibility of this occurring at Yucca 
Mountain was discounted within 10,000 years on the basis that the waste 
packages would remain largely intact during that time (although a 
certain level of premature failures was assumed). DOE stated that ``One 
of the required conditions is the presence of a moderator, such as 
water, in the waste package. The waste packages will be designed to 
make the probability of a criticality occurring inside the waste 
package extremely small'' (FEIS, DOE/EIS-0250, section I.2.12, p. I-21, 
Docket No. OAR-2005-0083-0086). At some point beyond 10,000 years, 
however, packages are anticipated to degrade sufficiently to allow 
water inside, so the reason for screening out this FEP is no longer 
credible. We understand that NRC has evaluated this possibility and 
initial results suggest that the effects would not be significant 
(``System-Level Performance Assessment of the Proposed Repository at 
Yucca Mountain Using the TPA Version 4.1 Code,'' CNWRA 2002-005, 
September 2002, Revised March 2004, Appendix G, Docket No. OAR-2005-
0083-0067). More recently, NRC staff analyses regarding the potential 
effects of a criticality event within the waste package indicated that 
the effects would be more significant within the first 10,000 years 
after disposal than at longer times (``Estimating In-Package 
Criticality Impact on Yucca Mountain Repository Performance,'' 
International High Level Radioactive Waste Management Conference, Las 
Vegas, Nevada, March 30-April 2, 2003, Docket No. OAR-2005-0083-0082). 
Therefore, we do not require that DOE consider in-package criticality 
beyond 10,000 years if it has not been considered for the first 10,000 
years. To the extent DOE's waste package assumptions make such a 
scenario credible within the initial 10,000 years, however, it would be 
appropriate to include it in the post-10,000-year projections.
    There may be other FEPs that fall within this category. However, 
this illustrates the very possibility we wish to avoid. It is possible 
to generate complex and vaguely-defined circumstances and insist that 
DOE analyze them thoroughly. We see such an exercise as being of no 
value. Rather, we believe it would be detrimental to the licensing 
process, as well as contrary to our ``reasonable expectation'' concept 
and the idea that performance assessments should represent credible 
projections of disposal system safety.
    Having considered the various types of FEPs that may have been 
excluded from the 10,000-year analysis, our goal is to require an 
appropriate consideration of FEPs in the analyses beyond 10,000 years. 
We considered an approach that would provide NRC with broader 
flexibility to consider previously excluded FEPs that it believes 
should be included in the peak dose analyses, perhaps based on the 
effect of the FEP on the magnitude of the peak dose. However, we 
believe that any potential FEPs to be included are likely to be 
overwhelmed by increasing uncertainties or larger-scale FEPs such as 
climate change. For this reason, we do not believe the inclusion of 
such FEPs will add materially to the understanding of the disposal 
system's performance or will lead to a safer disposal system. 
Furthermore, as stated earlier, we are guided by our reasonable 
expectation principle in not requiring an exhaustive and completely 
accurate prediction of repository conditions over a million-year 
period. See Sections II.A, II.B, and II.C for discussions of the

[[Page 49055]]

relative confidence in calculations at very long times, and the need to 
view those calculations in a more qualitative way. We aim to construct 
a reasonable test of the disposal system that accounts for the possible 
occurrence of significant FEPs at Yucca Mountain, and the system's 
response to those stresses. We believe that proposing the continued 
exclusion from peak dose calculations of events that are 
inconsequential for 10,000 years, with the exception of general 
corrosion and those identified by NAS, is consistent with this 
approach.
    To summarize our proposal for Sec.  197.36, we propose that DOE 
continue to use the FEPs selected for compliance with the 10,000-year 
projections in its projections for peak dose. This does not require 
that DOE continue to define the characteristics of those FEPs in 
exactly the same way it has previously (for example, in the FEIS). 
Rather, DOE may continue to refine its representation of FEPs in the 
analyses as its understanding of the factors involved improves. The 
contribution to dose estimates of FEPs selected for the analyses must 
be assessed throughout the period of geologic stability. We do require 
that DOE explicitly consider the effects of seismic, igneous, and 
climate change scenarios, within the overall probability constraints, 
as described in more detail in the following sections. We also require 
that DOE consider the effects of general corrosion throughout the 
period of geologic stability. We have considered two approaches for 
doing so. Under the first approach, consistent with our approach to 
climate change outlined in Section II.D.2. DOE may apply a constant 
representative corrosion rate throughout the period of geologic 
stability. Under the second approach, consistent with our approach to 
seismic and igneous FEPs outlined in Sections II.D.2.b and c, DOE may 
apply corrosion rates as derived for the 10,000-year period, which may 
be dependent on other factors, such as temperature within the 
repository.
    We have stated our concerns that the screening process should not 
be used to put forward highly speculative and implausible situations 
for DOE to analyze. It is our belief that the relevant FEPs are already 
captured within the 10,000-year screening process, and that any others 
would be overshadowed by other aspects of the longer-term modeling. We 
believe our proposal to explicitly include certain FEPs important to 
the longer-term projections appropriately balances these 
considerations. We request comment on this approach.

b. Consideration of Seismic FEPs

    The NAS stated, and we agree, that the effects of seismicity in the 
area on (1) the repository and (2) the hydrologic regime are key 
aspects to be considered during the period of geologic stability (NAS 
Report p. 93). The effects of seismicity may result in (most 
significantly) early waste package failure, an increase or decrease in 
conductivity (movement of water) in the saturated or vadose zones, or a 
shift in direction of fluid movement in the area (NAS Report pp. 92-
93). In addition, we believe the potential effects of seismic activity 
on the structural stability of the repository itself (i.e., drift 
collapse) may be important in projecting the failure of waste packages.
    In order to reasonably assess the effects of seismicity at the 
site, and yet also address the increasing uncertainty associated with 
magnitudes of seismic events over the greatly increased time period, we 
expect that DOE will take the rate of occurrence of seismic events 
originally derived for the 10,000-year time period and extend the 
calculations throughout the period of geologic stability. We are 
proposing that DOE may limit its assessment of seismicity to the 
effects on the disposal system of drift collapse and waste package 
failure, i.e., effects on the engineered barriers that comprise an 
essential component of the disposal system. At times sufficiently far 
into the future, a wide range of possibilities could be proposed, and 
some (for example, an earthquake of such an extreme magnitude that it 
collapses all the drifts of the repository, allowing for complete 
destruction of the facility), no matter how remote the probability, 
could have far-reaching implications for the disposal system. By using 
this approach, we can adhere to the basic premise that the risk 
calculations reasonably predict the geologic environment at the 
repository out to peak dose. We can also capture the potential effects 
of seismicity and faulting at Yucca Mountain. By extending the 
performance period to 1 million years, it is expected that more events 
will occur, consistent with the established seismic hazard curve for 
the site. No new types or classes of seismic or fault displacement 
disruptive events can reasonably be anticipated. In the case of 
seismicity, earthquakes are most likely to occur on the existing 
network of active seismogenic fault sources under current tectonic 
conditions. In the case of the fault displacement hazard, it is more 
likely that fault slip will occur on existing faults that on newly 
created ones.
    DOE has developed a seismic hazard curve that describes the 
seismicity to be expected at the site (``Seismic Consequence 
Abstraction,'' MDL-WIS-PA-00003-Rev 00, 2003, Docket No. OAR-2005-0083-
0073). A seismic hazard curve determines what the probability is of any 
particular strength of ground shaking. The goal of probabilistic 
seismic hazard analysis is to quantify the rate (or probability) of 
exceeding various ground-motion levels at a site (or a map of sites), 
given all possible earthquakes. It is reasonable to assume that seismic 
events will continue with activity rates and magnitudes predicted by 
the seismic and fault displacement hazards for the site over the period 
of geologic stability because the geologic record indicates relative 
tectonic stability of the region over the past 10 million years. This 
implies that there is continuity in the behavior of major geologic 
events (such as earthquakes) over that entire time frame. Further, the 
geologic record extending back millions of years has been used to 
establish the hazard curves. There is not further data that 
appropriately can be incorporated into the analysis, or used to justify 
an adjustment of the estimates simply because they are to be projected 
further into the future. It is expected that more events, such as 
earthquakes and fault displacements, will occur with the extended 
performance period, but that these events are much more likely to occur 
on existing faults and seismic sources than on newly created ones. 
Therefore, the rates and magnitudes considered in the probabilistic 
calculations for 10,000 years can also be used to generate estimates of 
seismicity out to the period of peak dose. These events should be 
defined on an annual probability of occurrence. The magnitudes and 
frequencies of potential seismic events should remain the same as in 
the 10,000-year analysis; however, the analysis would be expected to 
show greater consequences as potentially more major seismic events are 
incorporated into the assessment as a result of extending the analysis 
throughout the period of geologic stability as events occur at times 
when packages are expected to be largely degraded and thus more easily 
damaged.
    The NAS stated that seismologic effects on the hydrology at Yucca 
Mountain can also be bounded over the period of stability due to the 
fact that the hydrology has been influenced by many similar seismic 
events in the past (NAS Report p. 93). Seismic activity can account for 
a number of changes in the

[[Page 49056]]

hydrology of the area, from the opening or closing of fractures and 
large-scale changes in water levels to a shift in the direction of 
ground-water flow in the region. It could also increase the potential 
for enhanced movement of the radionuclides in the waste, because the 
potential for increased rate of water movement could contribute to a 
greater velocity of the ground water in the aquifer, which could reduce 
the travel time of radionuclides out to the boundary of the controlled 
area. However, we are proposing today that DOE's analysis for seismic 
events may exclude the effects of seismicity on the hydrology of the 
Yucca Mountain disposal system. In making this decision, we considered 
the NAS's guidance as well as the relative effects of climate change on 
the hydrology of the disposal system.
    In its report, NAS observed that seismicity potentially can affect 
the hydrologic regime by causing displacements and increasing 
conductivity along existing fractures. NAS noted that such 
displacements are likely to occur along existing fractures (as opposed 
to creating new ones) and, further, that hydrology near Yucca Mountain 
``has been conditioned by many similar seismic events over geologic 
time'' (NAS Report p. 93). Since no major new flow paths would be 
created, these statements imply that the most likely hydrologic effects 
are changes in conductivity or a localized shift in the ground-water 
flow. Nevertheless, NAS concluded that ``such displacements have an 
equal probability of favorably changing the hydrologic regime'' (NAS 
Report p. 93). We agree, and also conclude that predicting the 
magnitude of changes in hydraulic conductivity--whether favorable or 
unfavorable--or the details of localized changes in the direction of 
ground-water flow is highly speculative, especially in view of the 
highly fractured nature of the geology at Yucca Mountain.
    However, we also agree with NAS that ``the effect of seismicity on 
the hydrologic regime could probably be bounded'' (NAS Report p. 93). 
The endpoint of most concern resulting from changes in flowpaths or 
hydraulic conductivity would be the potential for greater movement of 
water through the disposal system. As previously mentioned, this could 
enhance movement of radionuclides from the waste. Importantly, this is 
also the endpoint of concern for climate change scenarios. As discussed 
in more detail in Section II.D.2.d, we are proposing that DOE must 
consider climate change scenarios that result in an increased flow of 
water through the disposal system. Unlike seismic events, such climate 
change scenarios do not have the potential to favorably affect (i.e., 
reduce) the ground-water flow through the disposal system (at best, 
they would have a neutral effect on overall performance). In addition, 
the effects on water flow from climate change would be expected to 
exceed any such effects resulting from seismicity. Thus, we believe 
that our proposed requirements for DOE to consider climate change over 
the period of geologic stability effectively bound the potential 
hydrologic effects and no further analysis is required separately as 
part of the seismic scenarios.
    In contrast, the potential effects on waste package failure through 
physical impact with other elements of the engineered barrier system or 
drift collapse (rockfall) are not clearly captured in analyses of other 
scenarios. Waste package failure is generally of importance because it 
is the immediate step allowing water to contact the waste, leading to 
release of radionuclides. Waste packages may be more vulnerable to 
seismic effects if corrosion processes have weakened them. Seismic 
events may cause the failure of the structures supporting the waste 
packages, allowing them to be physically damaged through impacts with 
other objects (i.e., if waste packages are no longer held in place, 
they could collide with other packages or elements of the engineered 
barrier system). The collapse of the emplacement drift itself could 
also be significant at these longer times as pieces of rock fall onto 
the already-weakened waste packages. Regarding waste package failure 
caused by seismicity, NAS concluded that the rocks in the Yucca 
Mountain area are so extensively fractured that future seismic events 
are likely to occur along existing fractures rather than new ones (NAS 
Report p. 93). By knowing the location of major fractures, DOE may be 
able to minimize the adverse effects of seismicity. For example, DOE 
can place waste packages away from these areas (fault avoidance), 
thereby decreasing the risk of failure by seismic induced rock falls. 
As can be seen by examples at the Waste Isolation Pilot Plant (WIPP), 
engineering practices at repositories can be successful in reducing the 
probability of adverse effects on isolation capabilities and DOE has 
criteria for such practices at Yucca Mountain. Because faults are being 
avoided by design, we do not believe DOE must assume they are not. In 
the end, DOE might be able to show that seismic effects on waste 
package failure ``could be reduced sufficiently to result in boundable 
and probably very low risk,'' as postulated by NAS (NAS Report p. 93). 
Our proposal would require that DOE specifically address waste package 
failure resulting from seismic events causing damage to the engineered 
barrier system, either through physical impacts within the drifts 
through failure of the supporting structures or drift collapse so that 
the significant effects identified by NAS will be fully considered.
    There are other effects that can be envisioned from seismic events 
near Yucca Mountain. Beyond the key aspects of seismicity discussed 
above, however, we do not believe there are others that would be 
expected to significantly affect performance (for example, from events 
that are of low magnitude or sufficiently distant from the disposal 
system), and NAS similarly identified none. The consideration of such 
effects would unnecessarily complicate the development of the 
performance assessment and the licensing process without contributing 
information on the protective capabilities of the Yucca Mountain 
disposal system. We believe they can reasonably be excluded from 
analysis over the period of geologic stability.
    Therefore, in conclusion, we propose that DOE evaluate the effects 
of seismic activity throughout the period of geologic stability, but 
limit those effects to those resulting in damage to the engineered 
barrier system and ultimately the waste packages. The probability of 
seismic events of different magnitude and duration for the period of 
geologic stability will be the same as determined for the period within 
10,000 years. We request comment on this approach.

c. Consideration of Igneous (Volcanic) FEPs

    EPA recognizes that a volcanic intrusion into the repository, 
although an unlikely event, could release a portion of the radioactive 
inventory. We agree with the NAS that this possibility exists over the 
period of geologic stability (NAS Report p. 94). While acknowledging 
the complexity of the release of radionuclides from the repository, 
given the known effects of the various types of past volcanic events 
and the study of the cinder cones in the area, we believe it is 
possible to develop reasonable estimates of the probability of 
radionuclide release via volcanic episodes through the repository 
through the period of geologic stability.
    We agree with NAS that the probability of igneous events may be 
great enough, and the potential

[[Page 49057]]

consequences significant enough, that they must be considered over the 
period of geologic stability. An analysis of the probability is based 
on extrapolations into the future of volcanic activity from the 
geologic record, and on assumptions about the spatial distribution of 
future volcanic eruptions in the Yucca Mountain region. Volcanism by 
nature is an episodic event. In the Yucca Mountain region it has been 
characterized as involving intermittent concentrated activity followed 
by long periods of quiescence (NAS Report p. 94). For example, the 
repository block tuffs are in the age range of approximately 11-12 
million years old and were generated by large-scale volcanism involving 
a large area around the site (``Site Environmental Report for the Yucca 
Mountain Project Calendar Year 2003,'' PGM-MGR-EC-000005-REV 00, 
Section 1.1, October 2004, Docket No. OAR-2005-0083-0086). This 
material is made of layers of ashfalls from volcanic eruptions that 
consolidated into the rock (of a type known as ``tuff''). Tuff has 
varying degrees of compaction and fracturing depending on the degree of 
``welding'' caused by temperature and pressure when the ash was 
deposited. An event of this nature is not likely to be repeated during 
the geologic stability period. It has been suggested by NAS, and fits 
within our FEPs screening, that a probability of 10-8/yr, 
which is a 1 in 10,000 possibility of a disruption (affecting the 
repository, not simply a volcanic event in the region) over 10,000 
years ``might be sufficiently low to constitute a negligible risk'' 
(NAS Report p. 95). Based on available information generated by DOE in 
its TSPA (Yucca Mountain Science and Engineering Report, DOE/RW-0539, 
Section 4.4.3, May 2001, Docket No. OAR-2005-0083-0069), the mean 
annual probability of an igneous event within the Yucca Mountain 
repository footprint is estimated at 1.6 x 10-8 per year 
(which is slightly higher than a one in 10,000 possibility of a 
disruption over 10,000 years). This probability, though extremely low, 
is just within the regulatory threshold for inclusion of events with 
very low probability of occurrence, but it can be assumed that this 
probability will hold throughout the period of geologic stability (NAS 
Report p. 94). For this reason, we are proposing to require that DOE 
include consideration of igneous FEPs extending over the period of 
geologic stability.
    We also agree with NAS that reasonable estimates of the effects can 
be developed (NAS Report p. 95). As with the seismic FEPs, we believe 
this is best accomplished by limiting the analysis to those effects 
most significant for performance. As we stated in our 2001 rule, the 
geologic record is the best source of evidence for the frequency and 
magnitude of natural features, events, and processes that could affect 
repository performance, and the geologic record is best preserved in 
the relatively recent past (66 FR 32100). Studies of the volcanic 
history of the area in the recent past indicate a different type of 
volcanic activity other than the intermittent layering volcanic 
activity that produced Yucca Mountain has occurred (FEIS, DOE/EIS-0250, 
Appendix I, Section 2.10, Docket No. OAR-2005-0083-0086). Basalt 
volcanism, exemplified by the Lathrop Wells volcano, and other features 
near the repository, appears to be the type of igneous activity, though 
unlikely, that has some probability of occurring within the period of 
geologic stability. By narrowing the type of events most plausible 
during the period of stability, we can attempt to constrain the 
uncertainty involved in using probabilistic analyses. The NAS noted 
that the most significant effects are related to future events that 
could intersect the repository (NAS Report p. 94).
    Existing DOE calculations provide an example of analysis of such 
disruptive igneous events. DOE states that, if igneous activity 
occurred at Yucca Mountain, possible effects on the repository could be 
grouped into three areas (FEIS, DOE/EIS-0250, Appendix I, Section 2.10, 
Volcanism, Docket No. OAR-2005-0083-0086):
     Igneous activity that would not directly intersect the 
repository (can be shown to have no effect on dose from the 
repository);
     Volcanic eruptions in the repository that would result in 
waste material being entrained in the volcanic magma or pyroclastic 
material, bringing waste to the surface (resulting in atmospheric 
transport of volcanic ash contaminated with radionuclides and 
subsequent human exposure downwind); or
     An igneous intrusion intersecting the repository (no 
eruption but damage to waste packages from exposure to the igneous 
material that would enhance release to the ground water and, thus, 
enhance transport to the biosphere).
    Based on studies of past activity in the region, probabilities for 
different types of igneous activity have been estimated by DOE. Each 
type of event was described in detail based on observation of effects 
of past activities as embodied in the geologic record of the region. 
These descriptions include geometry of intrusions, geometry of 
eruptions, physical and chemical properties of volcanic materials, 
eruption properties (velocity, power, duration, volume, and particle 
characteristics). Most of the parameters describing the igneous 
activity were entered in the modeling as probability distributions 
(FEIS, DOE/EIS-0250, Appendix I, Section 2.10, Volcanism, Docket No. 
OAR-2005-0083-0086).
    DOE's current igneous activity scenario contains two separate 
possible events: a volcanic eruption that includes exposure as a result 
of atmospheric transport and deposition on the ground, and an igneous 
intrusion ground-water transport event. In the volcanic eruption event, 
a dike (or dikes) would intersect the repository and compromise all 
waste packages in the conduit. Then, an eruptive conduit of an 
associated volcano would intersect waste packages in its path. Waste 
packages in the path of the conduit would be sufficiently damaged that 
they provide no further protection, and the waste in the packages would 
be entrained in the eruption and subject to atmospheric transport. In 
the igneous intrusion ground-water transport event, the analysis 
calculated releases caused by a dike (or dikes) intersecting 
emplacement drifts, causing varying degrees of waste-package damage and 
making the contents of the containers available for transport to the 
RMEI through ground water. We believe these are the most significant 
consequences that would result from a volcanic event through the 
repository. Other results from igneous events--the occurrence of 
distant events, potential drift instability, or changes in rock 
fracturing--are secondary to the direct releases of radionuclides. In 
addition, the response of the disposal system to such effects would 
likely be captured by consideration of other FEPs (such as seismicity 
or climate change). Therefore, we are proposing that DOE's 
consideration of igneous events over the period of geologic stability 
may be limited to events that intersect the repository, damage the 
waste packages, and cause releases of radionuclides either directly to 
the atmosphere and biosphere (i.e., an extrusive event) or to the 
ground water. We expect that the same probability of occurrence for 
these events used in the 10,000-year analysis be applied over the 
period of geologic stability. Using this probability, it is very 
unlikely that more than one igneous event would be included in a single 
realization. However, the two types of events are very different in 
terms of their potential effects and when those effects would be 
greatest. We

[[Page 49058]]

believe this approach is appropriate, as described in the next 
paragraph.
    DOE's analysis of releases from waste packages entrained by magma 
erupted on the surface assume the waste containers are breached by the 
eruption itself and the wastes are available for dispersal by the 
eruption. In this scenario, the doses would be highest if the eruption 
happened early in the geologic stability period (before significant 
decay of short-lived radionuclides that provide a dose through 
inhalation as well as through deposition and uptake by plants), and are 
lower if the event occurs at later times. Assuming waste packages are 
breached during the event provides that the assessment is a ``worst 
case'' in terms of potential doses because it does not depend on 
assumptions regarding other waste package failure mechanisms, such as 
corrosion. However, other analyses and laboratory experiments have been 
presented suggesting that intact waste containers can withstand the 
temperatures of the molten magma without melting or otherwise 
sustaining significant damage (``Evaluation of the Igneous Extrusive 
Scenario,'' Presentation to the Nuclear Waste Technical Review Board, 
September 20, 2004, Docket No. OAR-2005-0083-0074). These analyses 
suggest that an early eruption might not produce the highest doses 
since the wastes could not be dispersed as easily. Under these 
assumptions, an eruption considerably later in the geologic stability 
period, when the waste containers have degraded considerably from 
corrosion processes, is more likely to result in widespread dispersal 
of the wastes. However, at the later times, the radionuclide inventory 
in the wastes would have decreased from decay, and projected releases 
would probably not exceed those estimated for the early eruption 
scenario DOE performed. The existing assessments of the eruptive event 
based on our previously issued regulations contain a number of 
assumptions, which we believe has led to conservative assessments. 
Under DOE's assumptions, the highest dose as a result of volcanic 
eruptions would occur within the first 10,000 years because that is 
when the radionuclide inventory is at its highest. We are not assuming 
this approach will be retained in all details, and have structured our 
proposed rule accordingly to ensure that igneous events are considered 
over the period of geologic stability. However, we acknowledge that the 
current approach, if retained, would meet our requirements and be 
conservative. We request comment on our proposal.

d. Consideration of Climatological FEPs

    The average of weather conditions over a long period of time is the 
climate (www.cogsci.princeton.edu/cgi-bin/webwn), and it has been well 
documented that climate can vary significantly over geologic time (NAS 
Report p. 91). Climate controls the range of precipitation and 
temperature conditions at Yucca Mountain. There are a number of 
impacts, particularly on the hydrologic regime, that must be taken into 
account. Run-on, run-off, and evapotranspiration of precipitation 
influence the rate of infiltration into the subsurface. The greater the 
amount of infiltration, or recharge, the greater the potential for an 
increase in ground water to infiltrate into the repository, allowing 
for an increase in the dissolution of the radionuclides. This could 
lead to higher release rates from the waste. Consequently, it is 
important to examine the effects of climate change throughout the 
period of geologic stability.
    At present the Earth is in an interglacial phase (NAS Report p. 
91). Climate change historically has been cyclical: ``Over a million-
year time scale, however, the global climate regime is virtually 
certain to pass through several glacial-interglacial cycles * * *'' 
(NAS Report p. 91). Similarly, the Yucca Mountain FEIS states: ``The 
record shows continual variation, often with very rapid jumps, between 
cold glacial climates (* * * pluvial periods) and warm interglacial 
climates similar to the present. Fluctuations average 100,000 years in 
length'' (FEIS, DOE/EIS-0250, p. 5-12, Docket No. OAR-2005-0083-0086). 
NAS stated the following with regard to climate change at Yucca 
Mountain:

    During the past 150,000 years, the climate has fluctuated 
between glacial and interglacial status. Although the range of 
climatic conditions has been wide, paleoclimatic research shows that 
the bounding conditions, the envelope encompassing the total 
climatic range have been fairly stable (Jannik et al., 1991; 
Winograd et al., 1992; Dansgaard et al., 1993). Recent research has 
indicated that the past 10,000 years are probably the only sustained 
period of stable climate in the past 80,000 years (Dansgaard et al., 
1993). Based on this record, it seems plausible that the climate 
will fluctuate between glacial and interglacial states during the 
period suggested for the performance assessment calculations. Thus, 
the specified upper boundary, or the physical top boundary of the 
modeled system, would be a conservative approach that captures the 
most severe, detrimental performance effects of these variations 
(especially in terms of ground-water recharge).

(NAS Report pp. 77-78.)

    We are concerned about the possibility of over-speculation of 
climatic change over such extremely long time periods, possibly out to 
the next 1 million years. The NAS recognized this fact in its report, 
stating ``Although the typical nature of past climate changes is well 
known, it is obviously impossible to predict in detail either the 
nature or the timing of future climate change. This fact adds to the 
uncertainty of the model predictions' (NAS Report p. 77).
    EPA agrees with the NAS statement and takes the position that it is 
not useful to have unconstrained speculation on future climate during 
the period of geologic stability, because it is possible to assume any 
number of scenarios of climate over this large amount of time, and 
there is very little evidence available to accept or refute most of 
them. Because it is not possible to predict every situation that could 
occur over such a long time, we feel that the best course, as outlined 
below, is to construct a climate scenario that assumes reasonable 
temperature and precipitation values, and allow this scenario to run 
throughout the period of geologic stability.
    Climate change differs from seismic and igneous events in that its 
effects would not occur instantaneously, and it can affect multiple 
portions of the disposal system with a very direct effect on 
performance since the movement of water through the site is the primary 
means for transporting radionuclides. These effects can persist for 
very long time periods, even longer than the period of geologic 
stability. Seismic events and volcanism, in contrast, are episodic 
events; though the events occur relatively quickly and deliver their 
consequences over the short term, the consequences themselves can be 
very long-lasting and fundamentally change the geologic setting.
    There are three major effects that climate change can impart on the 
disposal system (NAS Report p. 91). The first is that increases in 
erosion might significantly decrease the burial depth of the 
repository. NAS pointed out that site-specific studies performed by DOE 
indicate that an increase in erosion to the extent necessary to expose 
the repository within the period of geologic stability is extremely 
unlikely (NAS Report p. 91). Therefore, we do not believe it is 
important or necessary to require DOE to assess the potential for 
erosion from climate change.
    The second change might be a shift in the distribution and 
activities of human populations (NAS Report p. 92). A cooler, wetter 
climate may provide a more hospitable environment,

[[Page 49059]]

increasing the population, and (some have argued) possibly changing the 
parameters we have outlined for the RMEI. We are not proposing to 
change the definition or characteristics of the RMEI. We have discussed 
our reasoning for taking this approach in greater detail in Section 
II.A.1 of this document. We do not believe that fixing the climate to 
present-day characteristics is the appropriate way to circumvent the 
difficulties in defining a biosphere applicable for 1 million years. 
Our view is that evaluation of reasonable climate change is critical to 
the integrity and meaning of peak dose projections. Further, as NAS 
noted, ``there is no simple relation between future climatic conditions 
and future population'' (NAS Report p. 92).
    Finally, for extremely long time periods, major changes in the 
global climate, for example a transition to a glacial climate, could 
affect ground-water movement. NAS states ``Change to a cooler, wetter 
climate at Yucca Mountain would likely result in greater fluxes of 
water through the unsaturated zone'' (NAS Report pp. 91-92). NAS 
observed that a doubling of the effective wetness (the ratio of 
precipitation to effective evapotranspiration) could cause a 
significant increase in recharge (NAS Report p. 91). This could affect 
the rates of radionuclide release from the waste and transport to the 
water table, although the location of the repository in the subsurface 
would provide a time lag for climate change effects. NAS states, ``The 
time required for unsaturated zone flux changes to propagate down to 
the repository and then to the water table is probably in the range of 
hundreds to thousands of years. The time required for saturated flow-
system responses is probably even longer. For this reason, climate 
changes on the time scale of hundreds of years would probably have 
little if any effect on repository performance, and the effects of 
climate changes on the deep hydrogeology can be assessed over much 
longer time scales'' (NAS Report p. 92).
    In its current analysis of future climate states (``Future Climate 
Analysis,'' ANL-NBS-GS-000008-Rev 00, 2000, Section 6.2, Docket No. 
OAR-2005-0083-0068), DOE assumed that all future climates were similar 
to current conditions or wetter than current conditions. The climate 
model provides a forecast of future climates based on information about 
past patterns of climates. The model represents future climate shifts 
as a series of instant changes. During the first 10,000 years, there 
are three changes, in order of increasing wetness, from present-day to 
a monsoon and then to a glacial-transition climate. Between 10,000 
years and 1 million years there are 45 changes between six climate 
states incorporated in the TSPA model:
     Interglacial Climate (same as present day)
     Intermediate Climate (same as the Glacial-transition)
     Intermediate/Monsoon Climate
     Three stages of Glacial Climate of varying infiltration 
rates
    Precipitation that is not returned to the atmosphere by evaporation 
or transpiration enters the unsaturated zone flow system. Water 
infiltration is affected by a number of factors related to climate, 
such as an increase or decrease in vegetation on the ground surface, 
total precipitation, air temperature, and runoff. The infiltration 
model uses data collected from studies of surface infiltration in the 
Yucca Mountain region. It treats infiltration as variable in the 
region, with more occurring along the crest of Yucca Mountain than 
along its base. The results of the climate model affect assumed 
infiltration rates. For each climate, there is a set of three 
infiltration rates (high, medium, low) and associated probabilities. 
This forms a discrete distribution that is sampled in the probabilistic 
modeling. Whenever a particular climate state is in effect, the 
associated infiltration rate distribution is sampled for each 
realization of the simulation.
    One of the issues associated with DOE's existing modeling efforts 
on climate at very long times is that the analysis assumed 
instantaneous changes between climate states. In other words, the 
entire flow field was assumed to immediately switch from one climate 
state to another. This approach is unrealistic because, as noted above, 
it would likely take hundreds or thousands of years for increased 
infiltration from a wetter climate to reach the underlying aquifer and 
affect transport and flow patterns. DOE also assumed that the climate 
change occurred at the same time for all realizations, which magnified 
the effect of the instantaneous change of climate when looked at as a 
probabilistic analysis. The result is that the doses calculated were 
the product of the conservatism of the assumptions noted above (e.g., 
instantaneous climate shift, which was assumed to occur at the same 
time for all realizations). Such assumptions are unlikely to produce 
meaningful or realistic results.
    We believe that an approach should be developed to answer several 
basic questions about how climatological effects realistically will 
impact the proposed repository until the time to peak dose. The 
questions that concern us are:
    1. How much total water will infiltrate into the repository over 
this large amount of time?
    2. Will more water infiltrate the repository over time when modeled 
as a wave function (current DOE modeling) or as total average?
    The answers to these questions assist in identifying conservative, 
yet reasonable, conditions the repository may encounter over the period 
of geologic stability. The amount of net infiltration into Yucca 
Mountain has an effect on the disposal system performance because 
higher net infiltration leads to the possibility that a greater 
proportion of the repository will experience ground-water seepage. For 
solubility-limited radionuclides in the waste, an increase in net 
infiltration could lead to a higher release rate of radionuclides from 
the disposal system, thereby affecting the potential dose to the RMEI 
in the accessible environment. We do not believe that it is important 
to know or predict with certainty precisely when the climate states 
with peak precipitation occur during the modeling. There are too many 
uncertainties and permutations available in trying to project a future 
set of climate conditions, and it is difficult to place specific times 
on when discrete pulses of precipitation should be injected into the 
modeling (NAS Report p. 77). Instead, we believe that it is reasonable 
to assume an average increase in precipitation over the entire time 
from 10,000 years through the period of geologic stability, and to 
model those consequences. An increase in average precipitation 
throughout the period of geologic stability is a more reasonable 
approach because it assumes a constant source of precipitation, 
creating more downward flow that will eventually reach the repository. 
This scenario need not be dominated by highs or lows in precipitation 
over the time period and does not require speculation about the exact 
timing or transient effects of shifts in climate. Rather, setting a 
constant value somewhat higher than today's average annual rainfall and 
extending it out to the time of peak dose would account for the greater 
potential for available fluids at the time of the failure of the waste 
packages. We believe that this approach provides a reasonable test of 
the repository conditions out to the time of peak dose, and will give a 
more conservative idea of potential fluid flow, as well as potential 
for migration of radionuclides out of the repository.

[[Page 49060]]

    We are proposing today that DOE, based on past climate conditions 
in the Yucca Mountain area, should determine how the disposal system 
responds to the effects of increased water flow through the repository 
as a result of climate change. We believe that the nature and extent of 
climate change can be reasonably represented by constant conditions 
taking effect after 10,000 years out to the time of geologic stability. 
We are proposing to explicitly require that DOE assume water flow will 
increase as a result of climate change. We leave it to NRC as the 
licensing authority to specify the values to be used to represent 
climate change. However, we expect that a doubling of today's average 
annual precipitation beginning at 10,000 years and continuing through 
the period of geologic stability would provide a reasonable scenario, 
given NAS's statements regarding potential effects on recharge (NAS 
Report p. 92). NRC could also use the range of projected precipitation 
values for different climate states and specify a reasonable long-term 
average precipitation based on the duration of each climate state over 
the period of geologic stability. We believe that either approach will 
allow for a reasonable estimate of how water will impact the site 
without subjecting the assessments to speculative assumptions that may 
well be unresolvable, while providing a reasonable indicator of 
disposal system compliance. NRC might choose to express the ground-
water flow effects directly as infiltration rates or other 
representative parameters, avoiding the necessity of translating 
precipitation and other climate-related parameters (e.g., temperature 
or evapotranspiration rates) into infiltration.
    Finally, we note that there are other potential effects of climate 
change such as the formation of surficial ponds or changes in fauna and 
flora (which could affect infiltration through changes in 
evapotranspiration rates). NAS did not identify these as significant, 
and also reiterated that speculation on the evolution of the biosphere 
(aside from climate) is unwarranted and unproductive. We agree fully. 
Therefore, in summary, we are proposing that DOE must include 
consideration of climate change in its performance assessment for 
compliance with the dose standard for the period of geologic stability. 
The assessment may be limited to the effects of increased water flow 
through the repository as a result of climate change. Climate change 
may be represented by constant conditions, which NRC would specify in 
regulation. We request comment on this proposal.

E. How Is EPA Proposing To Revise the Human-Intrusion Standard (Sec.  
197.25) To Address Peak Dose?

    As discussed in Section II.A.2, we believe it is logical and 
defensible to modify the human-intrusion standard in Sec.  197.25 to 
parallel the revisions we are proposing for the individual-protection 
standard. We described in some detail in that section the reasons why 
we believe that course of action to be appropriate, and briefly 
summarize our proposal here. Like the individual-protection standard, 
our provisions for human intrusion in the 2001 rule envisioned some 
consideration of performance beyond 10,000 years. The exposures 
resulting from the event were subject to the same compliance standard 
as the individual-protection standard (15 mrem/yr at 10,000 years or 
earlier coupled with compilation in the EIS if doses were projected to 
occur after 10,000 years). In deciding to propose revisions to the 
human-intrusion standard to conform to changes we are proposing to make 
to the individual-protection provisions, we kept in mind the NAS 
recommendation that ``the figure-of-merit for [the human-intrusion] 
calculation should be the same as in the undisturbed case * * * EPA 
should require that the conditional risk as a result of the assumed 
intrusion scenario should be no greater than the risk levels that would 
be acceptable for the undisturbed-repository case'' (NAS Report pp. 
112-113).
    The 2001 standard required that DOE determine when an intrusion by 
drilling would be possible and assess the consequences. We believe it 
is still appropriate for DOE to determine the time at which the 
intrusion could occur. However, under our proposal today, consequences 
at any time within the period of geologic stability would be subject to 
a compliance demonstration. We are proposing to apply the same dose 
limits to the human-intrusion scenario as we are proposing for the 
individual-protection scenario. Thus, exposures incurred by the RMEI 
within 10,000 years after disposal as a result of the intrusion must 
comply with a standard of 150 [mu]Sv/yr (15 mrem/yr). Exposures after 
that time within the period of geologic stability must comply with a 
standard of 3.5 mSv/yr (350 mrem/yr). DOE must still use the same 
assumptions regarding the RMEI as it used for the individual-protection 
analysis.
    We are not proposing to modify in any way the circumstances of the 
intrusion described in Sec.  197.26. We believe those circumstances 
continue to reflect two key points emphasized by NAS. First, ``there is 
no scientific basis for estimating the probability of intrusion at far-
future times'' (NAS Report p. 106). Second, like future society, future 
exploration technology cannot be predicted (NAS Report p. 107). 
Therefore, there is no basis for assuming a different set of 
circumstances to apply to intrusions beyond 10,000 years.
    We request comment on our proposed changes to the human-intrusion 
standard. We are not soliciting, and will not consider, comments on the 
overall intrusion scenario or other aspects of the human-intrusion 
standard that are not proposed to be changed.

F. Summary of Today's Proposal by Section

    Today's proposal is limited in scope. We are proposing to amend 
provisions only as necessary to address the Court ruling. Because of 
the unique nature of the challenge facing us, in which we must craft a 
regulatory standard to apply to times up to 1 million years, we have 
chosen to discuss many aspects of our 2001 rule in this document. We 
have done so because we believe it important that the public clearly 
understand what actions we are proposing to take and why, as well as 
reasons for not amending other provisions. In the listing that follows, 
we identify only those provisions of the rule that we are proposing to 
change today. We request public comment only on these proposed 
amendments. We are not proposing to change any other provisions. 
Therefore, we are not requesting, and will not respond to, public 
comments related to those provisions, since they have been previously 
established in rulemaking and are outside the scope of today's 
proposal.

Subpart A--Public Health and Environmental Standards for Storage

    Sec.  197.2, What definitions apply in subpart A?--Amends the 
definition of Effective Dose Equivalent to specify that calculations be 
performed using organ weighting factors in Appendix A.

Subpart B--Public Health and Environmental Standards for Disposal

    Sec.  197.12, What definitions apply in subpart B?--Modifies the 
definition of Performance Assessment to remove reference to 10,000 
years. Modifies the definition of Period of Geologic Stability as 
ending 1 million years after disposal.
    Sec.  197.13, How is subpart B implemented?--Specifies that the 
arithmetic mean of the distribution of projected doses is used to 
determine

[[Page 49061]]

compliance within 10,000 years. Specifies that the median of the 
distribution of projected doses is used to determine compliance beyond 
10,000 years but within the period of geologic stability (for 
Sec. Sec.  197.20 and 197.25 only).
    Sec.  197.15, How must DOE take into account the changes that will 
occur during the next 10,000 years after disposal?--Replaces references 
to 10,000 years with ``period of geologic stability.''
    Sec.  197.20, What [individual-protection] standard must DOE 
meet?--Retains the standard of 15 mrem/yr to apply up to 10,000 years 
after disposal. Adds a standard of 350 mrem/yr to apply beyond 10,000 
years within the period of geologic stability.
    Sec.  197.25, What [human-intrusion] standard must DOE meet?--
Retains the standard of 15 mrem/yr to apply up to 10,000 years after 
disposal. Adds a standard of 350 mrem/yr to apply beyond 10,000 years 
within the period of geologic stability. Removes references to time of 
intrusion and to placement of results in EIS.
    Sec.  197.35, What other projections must DOE make?--Section to be 
deleted.
    Sec.  197.36, Are there limits on what DOE must consider in the 
performance assessments?--Addresses probability of features, events, 
and processes in assessments used to comply with proposed Sec.  
197.20(b). Adds provisions to address climate change, igneous, seismic, 
and general corrosion scenarios.
    Appendix A, Calculation of Committed Effective Dose Equivalent--
describes the method to calculate the dose for comparison with the 
appropriate standards.

III. Statutory and Executive Order Reviews

A. Executive Order 12866: Regulatory Planning and Review

    Under Executive Order 12866, [58 Federal Register 51735 (October 4, 
1993)] the Agency must determine whether the regulatory action is 
``significant'' and therefore subject to OMB review and the 
requirements of the Executive Order. The Order defines ``significant 
regulatory action'' as one that is likely to result in a rule that may:
    (1) Have an annual effect on the economy of $100 million or more or 
adversely affect in a material way the economy, a sector of the 
economy, productivity, competition, jobs, the environment, public 
health or safety, or State, local, or tribal governments or 
communities;
    (2) Create a serious inconsistency or otherwise interfere with an 
action taken or planned by another agency;
    (3) Materially alter the budgetary impact of entitlements, grants, 
user fees, or loan programs or the rights and obligations of recipients 
thereof; or
    (4) Raise novel legal or policy issues arising out of legal 
mandates, the President's priorities, or the principles set forth in 
the Executive Order.
    Pursuant to the terms of Executive Order 12866, it has been 
determined that this rule is a ``significant regulatory action'' 
because it raises novel legal or policy issues arising out of the 
specific legal mandate of Section 801 of the Energy Policy Act of 1992. 
As such, this action was submitted to OMB for review. Changes made in 
response to OMB suggestions or recommendations will be documented in 
the public record.

B. Paperwork Reduction Act

    This action does not impose an information collection burden under 
the provisions of the Paperwork Reduction Act, 44 U.S.C. 3501 et seq. 
We have determined that this rule contains no information collection 
requirements within the scope of the Paperwork Reduction Act.
    Burden means the total time, effort, or financial resources 
expended by persons to generate, maintain, retain, or disclose or 
provide information to or for a Federal agency. This includes the time 
needed to review instructions; develop, acquire, install, and utilize 
technology and systems for the purposes of collecting, validating, and 
verifying information, processing and maintaining information, and 
disclosing and providing information; adjust the existing ways to 
comply with any previously applicable instructions and requirements; 
train personnel to be able to respond to a collection of information; 
search data sources; complete and review the collection of information; 
and transmit or otherwise disclose the information.
    An agency may not conduct or sponsor, and a person is not required 
to respond to a collection of information unless it displays a 
currently valid OMB control number. The OMB control numbers for EPA's 
regulations in 40 CFR are listed in 40 CFR part 9.

C. Regulatory Flexibility Act

    The Regulatory Flexibility Act (RFA) generally requires an agency 
to prepare a regulatory flexibility analysis of any rule subject to 
notice and comment rulemaking requirements under the Administrative 
Procedure Act or any other statute unless the agency certifies that the 
rule will not have a significant economic impact on a substantial 
number of small entities. Small entities include small businesses, 
small organizations, and small governmental jurisdictions.
    For purposes of assessing the impacts of today's rule on small 
entities, small entity is defined as: (1) A small business as defined 
by the Small Business Administration's (SBA) regulations at 13 CFR 
121.201; (2) a small governmental jurisdiction that is a government of 
a city, county, town, school district or special district with a 
population of less than 50,000; and (3) a small organization that is 
any not-for-profit enterprise which is independently owned and operated 
and is not dominant in its field.
    However, the requirement to prepare a regulatory flexibility 
analysis does not apply if the Administrator certifies that the rule 
will not, if promulgated, have a significant economic impact upon a 
substantial number of small entities (5 U.S.C. 605(b)). The rule 
proposed today would establish requirements that apply only to DOE. 
Therefore, it does not apply to small entities. Accordingly, I hereby 
certify that the rule, when promulgated, will not have a significant 
economic impact upon a substantial number of small entities. We 
continue to be interested in the potential impacts of our proposed 
rules on small entities and welcome comments on issues related to such 
impacts.

D. Unfunded Mandates Reform Act

    Title II of the Unfunded Mandates Reform Act of 1995 (UMRA), Pub. 
L. 104-4, establishes requirements for Federal agencies to assess the 
effects of their regulatory actions on State, local, and tribal 
governments and the private sector. Under section 202 of the UMRA, EPA 
generally must prepare a written statement, including a cost-benefit 
analysis, for proposed and final rules with ``Federal mandates'' that 
may result in expenditures to State, local, and tribal governments, in 
the aggregate, or to the private sector, of $100 million or more in any 
one year. Before promulgating an EPA rule for which a written statement 
is needed, section 205 of the UMRA generally requires EPA to identify 
and consider a reasonable number of regulatory alternatives and adopt 
the least costly, most cost-effective or least burdensome alternative 
that achieves the objectives of the rule. The provisions of section 205 
do not apply when they are inconsistent with applicable law. Moreover, 
section 205 allows EPA to adopt an alternative other than the least 
costly, most cost-effective or least burdensome alternative if the 
Administrator publishes with the final rule an explanation why that 
alternative was not adopted. Before EPA establishes

[[Page 49062]]

any regulatory requirements that may significantly or uniquely affect 
small governments, including tribal governments, it must have developed 
under section 203 of the UMRA a small government agency plan. The plan 
must provide for notifying potentially affected small governments, 
enabling officials of affected small governments to have meaningful and 
timely input in the development of EPA regulatory proposals with 
significant Federal intergovernmental mandates, and informing, 
educating, and advising small governments on compliance with the 
regulatory requirements.
    Today's proposed rule contains no Federal mandates (under the 
regulatory provisions of Title II of UMRA) for State, local, or tribal 
governments or the private sector. The proposed rule implements 
requirements specifically set forth by the Congress in section 801 of 
the EnPA and proposes radiological protection standards applicable 
solely and exclusively to the Department of Energy's potential storage 
and disposal facility at Yucca Mountain. The rule imposes no 
enforceable duty on any State, local or tribal governments or the 
private sector. Thus, today's rule is not subject to the requirements 
of sections 202 and 205 of UMRA.

E. Executive Order 13132: Federalism

    Executive Order 13132, entitled ``Federalism'' (64 FR 43255, August 
10, 1999), requires EPA to develop an accountable process to ensure 
``meaningful and timely input by State and local officials in the 
development of regulatory policies that have federalism implications.'' 
``Policies that have federalism implications'' is defined in the 
Executive Order to include regulations that have ``substantial direct 
effects on the States, on the relationship between the national 
government and the States, or on the distribution of power and 
responsibilities among the various levels of government.''
    This proposed rule does not have federalism implications. It will 
not have substantial direct effects on the States, on the relationship 
between the national government and the States, or on the distribution 
of power and responsibilities among the various levels of government, 
as specified in Executive Order 13132. Thus, Executive Order 13132 does 
not apply to this rule. In the spirit of Executive Order 13132, and 
consistent with EPA policy to promote communications between EPA and 
State and local governments, EPA specifically solicits comment on this 
proposed rule from State and local officials.

F. Executive Order 13175: Consultation and Coordination With Indian 
Tribal Governments

    Executive Order 13175, entitled ``Consultation and Coordination 
with Indian Tribal Governments'' (65 FR 67249, November 9, 2000), 
requires EPA to develop an accountable process to ensure ``meaningful 
and timely input by tribal officials in the development of regulatory 
policies that have tribal implications.'' This proposed rule does not 
have tribal implications, as specified in Executive Order 13175. The 
rule proposed today would regulate only DOE on land owned by the 
Federal government. The rule proposed today does not have substantial 
direct effects on one or more Indian tribes, on the relationship 
between the Federal Government and Indian tribes, or on the 
distribution of power and responsibilities between the Federal 
Government and Indian tribes. Thus, Executive Order 13175 does not 
apply to this rule. EPA specifically solicits additional comment on 
this proposed rule from tribal officials.

G. Executive Order 13045: Protection of Children From Environmental 
Health & Safety Risks

    Executive Order 13045: ``Protection of Children from Environmental 
Health Risks and Safety Risks'' (62 FR 19885, April 23, 1997) applies 
to any rule that: (1) Is determined to be ``economically significant'' 
as defined under Executive Order 12866, and (2) concerns an 
environmental health or safety risk that EPA has reason to believe may 
have a disproportionate effect on children. If the regulatory action 
meets both criteria, the Agency must evaluate the environmental health 
or safety effects of the planned rule on children, and explain why the 
planned regulation is preferable to other potentially effective and 
reasonably feasible alternatives considered by the Agency.
    This proposed rule is not subject to Executive Order 13045 because 
it is not economically significant as defined in Executive Order 12866, 
and because the Agency does not have reason to believe the 
environmental health risks or safety risks addressed by this action 
present a disproportionate risk to children. The public is invited to 
submit or identify peer-reviewed studies and data, of which EPA may not 
be aware, that assessed results of early life exposure to radiation.

H. Executive Order 13211: Actions That Significantly Affect Energy 
Supply, Distribution, or Use

    This rule is not a ``significant energy action'' as defined in 
Executive Order 13211, ``Actions Concerning Regulations That 
Significantly Affect Energy Supply, Distribution, or Use'' (66 FR 28355 
(May 22, 2001)) because it is not likely to have a significant adverse 
effect on the supply, distribution, or use of energy. The rule proposed 
today would apply only to DOE. Construction, operation, and closure of 
the repository at Yucca Mountain would fulfill the Federal government's 
commitment to manage the final disposition of spent nuclear fuel from 
commercial power reactors. However, there is no direct link between 
operation of the repository and an increased use of nuclear power. 
Other economic, technical, and policy factors will influence the extent 
to which nuclear energy is utilized.

I. National Technology Transfer and Advancement Act

    Section 12(d) of the National Technology Transfer and Advancement 
Act of 1995 (``NTTAA''), Public Law 104-113, 12(d) (15 U.S.C. 272 note) 
directs EPA to use voluntary consensus standards in its regulatory 
activities unless to do so would be inconsistent with applicable law or 
otherwise impractical. Voluntary consensus standards are technical 
standards (e.g., materials specifications, test methods, sampling 
procedures, and business practices) that are developed or adopted by 
voluntary consensus standards bodies. The NTTAA directs EPA to provide 
Congress, through OMB, explanations when the Agency decides not to use 
available and applicable voluntary consensus standards.
    In our original proposal (64 FR 46976, August 27, 1999), we 
requested public comment on potentially applicable voluntary consensus 
standards that would be appropriate for inclusion in the Yucca Mountain 
rule. We received no comments on this aspect of the rule. The closest 
analogy to consensus standards for radioactive waste disposal 
facilities are our regulations at 40 CFR part 191. As discussed above 
in this preamble, Congress expressly prohibited the application of the 
40 CFR part 191 standards to the Yucca Mountain disposal facility, and, 
therefore, the standards promulgated in 2001 and today's proposed 
revisions are site-specific and developed solely for application to the 
Yucca Mountain disposal facility.

[[Page 49063]]

List of Subjects in 40 CFR Part 197

    Environmental protection, Nuclear energy, Radiation protection, 
Radionuclides, Uranium, Waste treatment and disposal, Spent nuclear 
fuel, High-level radioactive waste.

    Dated: August 9, 2005.
Stephen L. Johnson,
Administrator.

    The Environmental Protection Agency is hereby proposing to amend 
part 197 of title 40, Code of Federal Regulations, as follows:

PART 197--PUBLIC HEALTH AND ENVIRONMENTAL RADIATION PROTECTION 
STANDARDS FOR YUCCA MOUNTAIN, NEVADA

    1. The authority citation for part 197 continues to read as 
follows:

    Authority: Sec. 801, Pub. L. 102-486, 106 Stat. 2921, 42 U.S.C. 
10141n.

Subpart A--Public Health and Environmental Standards for Storage

    2. Section 197.2 is amended by revising the definition of 
``Effective dose equivalent'' to read as follows:


Sec.  197.2  What definitions apply in subpart A?

* * * * *
    Effective dose equivalent means the sum of the products of the dose 
equivalent received by specified tissues following an exposure of, or 
an intake of radionuclides into, specified tissues of the body, 
multiplied by appropriate weighting factors. Annual committed effective 
dose equivalents shall be calculated using weighting factors in 
accordance with appendix A of this part.
* * * * *

Subpart B--Public Health and Environmental Standards for Disposal

    3. Section 197.12 is amended by revising paragraph (1) of the 
definition of ``Performance assessment'' and the definition of ``Period 
of geologic stability'' to read as follows:


Sec.  197.12  What definitions apply in subpart B?

* * * * *
    Performance assessment means an analysis that:
    (1) Identifies the features, events, processes, (except human 
intrusion), and sequences of events and processes (except human 
intrusion) that might affect the Yucca Mountain disposal system and 
their probabilities of occurring;
* * * * *
    Period of geologic stability means the time during which the 
variability of geologic characteristics and their future behavior in 
and around the Yucca Mountain site can be bounded, that is, they can be 
projected within a reasonable range of possibilities. This period is 
defined to end at 1 million years after disposal.
* * * * *
    4. Section 197.13 is revised to read as follows:


Sec.  197.13  How is subpart B implemented?

    (a) The NRC will determine compliance based upon the arithmetic 
mean of the projected doses from DOE's performance assessments for the 
period within 10,000 years after disposal:
    (1) For Sec.  197.20 of this subpart; and
    (2) For Sec. Sec.  197.25 and 197.30 of this subpart, if 
performance assessment is used to demonstrate compliance with either or 
both of these sections.
    (b) NRC will determine compliance based upon the median of the 
projected doses from DOE's performance assessments for the period after 
10,000 years of disposal and through the period of geologic stability:
    (1) For Sec.  197.20 of this subpart; and
    (2) For Sec.  197.25, if a performance assessment is used to 
demonstrate compliance.
    5. Section 197.15 is revised to read as follows:


Sec.  197.15  How must DOE take into account the changes that will 
occur during the period of geologic stability?

    The DOE should not project changes in society, the biosphere (other 
than climate), human biology, or increases or decreases of human 
knowledge or technology. In all analyses done to demonstrate compliance 
with this part, DOE must assume that all of those factors remain 
constant as they are at the time of license application submission to 
NRC. However, DOE must vary factors related to the geology, hydrology, 
and climate based upon cautious, but reasonable assumptions of the 
changes in these factors that could affect the Yucca Mountain disposal 
system during the period of geologic stability, consistent with the 
requirements for performance assessments specified at Sec.  197.36.
    6. Section 197.20 is revised to read as follows:


Sec.  197.20  What standard must DOE meet?

    (a) The DOE must demonstrate, using performance assessment, that 
there is a reasonable expectation that the reasonably maximally exposed 
individual receives no more than the following annual committed 
effective dose equivalent from releases from the undisturbed Yucca 
Mountain disposal system:
    (1) 150 microsieverts (15 millirems) for 10,000 years following 
disposal; and
    (2) 3.5 millisieverts (350 millirems) after 10,000 years, but 
within the period of geologic stability.
    (b) The DOE's performance assessment must include all potential 
pathways of radionuclide transport and exposure.
    7. Section 197.25 is revised to read as follows:


Sec.  197.25  What standard must DOE meet?

    (a) The DOE must determine the earliest time after disposal that 
the waste package would degrade sufficiently that a human intrusion 
(see Sec.  197.26) could occur without recognition by the drillers.
    (b) The DOE must demonstrate that there is a reasonable expectation 
that the reasonably maximally exposed individual will receive an annual 
committed effective dose equivalent, as a result of the human 
intrusion, of no more than:
    (1) 150 microsieverts (15 millirems) for 10,000 years following 
disposal; and
    (2) 3.5 millisieverts (350 millirems) after 10,000 years, but 
within the period of geologic stability.
    (c) The analysis must include all potential environmental pathways 
of radionuclide transport and exposure.


Sec.  197.35  [Removed and Reserved]

    8. Section 197.35 is removed and reserved.
    9. Section 197.36 is revised to read as follows:


Sec.  197.36  Are there limits on what DOE must consider in the 
performance assessments?

    (a) Yes, there are limits on what DOE must consider in the 
performance assessments. The DOE's performance assessments conducted to 
show compliance with Sec. Sec.  197.20(a)(1), 197.25(b)(1), and 197.30 
shall not include consideration of very unlikely features, events, or 
processes, i.e., those that are estimated to have less than one chance 
in 10,000 of occurring within 10,000 years of disposal (less than one 
chance in 100,000,000 per year). In addition, unless otherwise 
specified in these standards or NRC regulations, DOE's performance 
assessments need not evaluate the impacts resulting from any features, 
events, and processes or sequences of events and processes with a 
higher chance of occurrence if the results of the performance 
assessments would not be changed significantly in the initial 10,000 
year period after disposal.

[[Page 49064]]

    (b) For performance assessments conducted to show compliance with 
Sec. Sec.  197.25(b) and 197.30, DOE's performance assessments shall 
exclude unlikely features, events, or processes, or sequences of events 
and processes. The DOE should use the specific probability of the 
unlikely features, events, and processes as specified by NRC.
    (c) For performance assessments conducted to show compliance with 
Sec. Sec.  197.20(a)(2) and 197.25(b)(2), DOE's performance assessments 
shall project the continued effects of the features, events, and 
processes included in paragraph (a) of this section beyond the 10,000-
year post-disposal period through the period of geologic stability. The 
DOE must evaluate all of the features, events, or processes included in 
paragraph (a) of this section, and also:
    (1) The DOE must assess the effects of seismic and igneous 
scenarios, subject to the probability limits in paragraph (a) of this 
section for very unlikely features, events, and processes. Performance 
assessments conducted to show compliance with Sec.  197.25(b)(2) are 
also subject to the probability limits for unlikely features, events, 
and processes as specified by NRC.
    (i) The seismic analysis may be limited to the effects caused by 
damage to the drifts in the repository and failure of the waste 
packages.
    (ii) The igneous analysis may be limited to the effects of a 
volcanic event directly intersecting the repository. The igneous event 
may be limited to that causing damage to the waste packages directly, 
causing releases of radionuclides to the biosphere, atmosphere, or 
ground water.
    (2) The DOE must assess the effects of climate change. The climate 
change analysis may be limited to the effects of increased water flow 
through the repository as a result of climate change, and the resulting 
transport and release of radionuclides to the accessible environment. 
The nature and degree of climate change may be represented by constant 
climate conditions. The analysis may commence at 10,000 years after 
disposal and shall extend to the period of geologic stability. The NRC 
shall specify in regulation the values to be used to represent climate 
change, such as temperature, precipitation, or infiltration rate of 
water.
    (3) The DOE must assess the effects of general corrosion on 
engineered barriers. The DOE may use a constant representative 
corrosion rate throughout the period of geologic stability or a 
distribution of corrosion rates correlated to other repository 
parameters.
    10. Appendix A to part 197 is added to read as follows:

Appendix A to Part 197--Calculation of Annual Committed Effective Dose 
Equivalent

    Unless otherwise directed by NRC, DOE shall use the radiation 
weighting factors and tissue weighting factors in this Appendix to 
calculate committed effective dose equivalent for compliance with 
sections 20 and 25 of this part. NRC may allow DOE to use updated 
factors issued after the effective date of this regulation. Any such 
factors shall have been issued by consensus scientific organizations 
and incorporated by EPA into Federal radiation guidance in order to 
be considered generally accepted and eligible for this use. Further, 
they must be compatible with the effective dose equivalent dose 
calculation methodology established in ICRP 26/30 and continued in 
ICRP 60/72, and incorporated in this Appendix.

I. Equivalent Dose

    The calculation of the committed effective dose equivalent 
(CEDE) begins with the determination of the equivalent dose, 
HT, to a tissue or organ, T, listed in Table A.2 below by 
using the equation:
[GRAPHIC] [TIFF OMITTED] TP22AU05.002

where DT,R is the absorbed dose in rads (one gray, an SI 
unit, equals 100 rads) averaged over the tissue or organ, T, due to 
radiation type, R, and wR is the radiation weighting 
factor which is given in Table A.1 below. The unit of equivalent 
dose is the rem (sievert, in SI units).

             Table A.1.--Radiation weighting factors, wR \1\
------------------------------------------------------------------------
            Radiation type and energy range \2\                wR value
------------------------------------------------------------------------
Photons, all energies......................................            1
Electrons and muons, all energies..........................            1
Neutrons, energy:
  < 10 keV.................................................            5
  10 keV to 100 keV........................................           10
  > 100 keV to 2 MeV.......................................           20
  > 2 MeV to 20 MeV........................................           10
  > 20 MeV.................................................            5
Protons, other than recoil protons, > 2 MeV................            5
Alpha particles, fission fragments, heavy nuclei...........          20
------------------------------------------------------------------------
\1\ All values relate to the radiation incident on the body or, for
  internal sources, emitted from the source.
\2\ See paragraph A14 in ICRP Publication 60 for the choice of values
  for other radiation types and energies not in the table.

II. Effective Dose Equivalent

    The next step is the calculation of the effective dose 
equivalent, E. The probability of occurrence of a stochastic effect 
in a tissue or organ is assumed to be proportional to the equivalent 
dose in the tissue or organ. The constant of proportionality differs 
for the various tissues of the body, but in assessing health 
detriment the total risk is required. This is taken into account 
using the tissue weighting factors, wT in Table A.2, 
which represent the proportion of the stochastic risk resulting from 
irradiation of the tissue or organ to the total risk when the whole 
body is irradiated uniformly and HT is the equivalent 
dose in the tissue or organ, T, in the equation:

E = [Sigma] wT [sdot] HT.

                Table A.2.--Tissue Weighting Factors, wT
------------------------------------------------------------------------
                      Tissue or organ                          wT value
------------------------------------------------------------------------
Gonads.....................................................         0.20
Bone marrow (red)..........................................         0.12
Colon......................................................         0.12
Lung.......................................................         0.12
Stomach....................................................         0.12
Bladder....................................................         0.05
Breast.....................................................         0.05
Liver......................................................         0.05
Esophagus..................................................         0.05
Thyroid....................................................         0.05
Skin.......................................................         0.01
Bone surface...............................................         0.01
Remainder..................................................    a,b 0.05
------------------------------------------------------------------------
a Remainder is composed of the following tissues: adrenals, brain,
  extrathoracic airways, small intestine, kidneys, muscle, pancreas,
  spleen, thymus, and uterus.
b The value 0.05 is applied to the mass-weighted average dose to the
  Remainder tissues group, except when the following ``splitting rule''
  applies: If a tissue of Remainder receives a dose in excess of that
  received by any of the 12 tissues for which weighting factors are
  specified, a weighting factor of 0.025 (half of Remainder) is applied
  to that tissue or organ and 0.025 to the mass-averaged committed
  equivalent dose equivalent in the rest of the Remainder tissues.

III. Annual Committed Tissue or Organ Equivalent Dose

    For internal irradiation from incorporated radionuclides, the 
total absorbed dose will be spread out in time, being gradually 
delivered as the radionuclide decays. The time distribution of the 
absorbed dose rate will vary with the radionuclide, its form, the 
mode of intake and the tissue within which it is incorporated. To 
take account of this distribution the quantity committed equivalent 
dose, HT([tau]) where [tau] is the integration time in 
years following an intake over any particular year, is used and is 
the integral over time of the equivalent dose rate in a particular 
tissue or organ that will be received by an individual following an 
intake of radioactive material into the body:
[GRAPHIC] [TIFF OMITTED] TP22AU05.000

for a single intake of activity at time t0 where 
HT(t) is the relevant equivalent-dose rate in a tissue or 
organ at time t. For the purposes of this rule, the previously 
mentioned single intake may be considered to be an annual intake.

[[Page 49065]]

IV. Annual Committed Effective Dose Equivalent

    If the committed equivalent doses to the individual tissues or 
organs resulting from an annual intake are multiplied by the 
appropriate weighting factors, wT, from table A.2, and 
then summed, the result will be the annual committed effective dose 
equivalent, E([tau]):
[GRAPHIC] [TIFF OMITTED] TP22AU05.001

[FR Doc. 05-16193 Filed 8-19-05; 8:45 am]
BILLING CODE 6560-50-P